Australian Physiotherapy Journal

Appraisal Critically Appraised Papers Additional physiotherapy during acute care reduces falls in the first 12 months after hip fracture Synopsis min per day of physiotherapy and 800 IU per day vitamin D3 therapy. In addition, the additional physiotherapy groups Summary of: Bischoff-Ferrari HA, Dawson-Hughes B, Platz received an extra 30 minutes of home program instruction A, Orav EJ, Stahelin HB, Willett WC, et al (2010) Effect each day during acute care and an instructional leaflet at of high-dosage cholecalciferol and extended physiotherapy discharge. The high dose Vitamin D therapy groups also on complications after hip facture. Arch Intern Med 170: received an additional 1200 IU per day vitamin D3 therapy. 813–820. Prepared by Nora Shields, CAP Editor. Outcome measures: The primary outcomes were rate of falls and the rate of hospital readmission at 12 months, Question: Do additional physiotherapy and high dose vitamin assessed by monthly telephone calls and a patient diary. All D3 therapy reduce the rate of falls and hospital admissions in analyses were based on intention to treat and included 173 patients with hip fracture? Design: Randomised, controlled patients. Results: 128 participants completed the study. At trial with blinded outcome assessment. Setting: One large 12 months, the falls rate in the patients who had received hospital centre in Switzerland. Participants: 173 patients additional physiotherapy was 25% less (95% CI –44% to with acute hip fracture. All participants had to have a mini- –1%). High dose vitamin D3 therapy did not reduce the rate mental examination score of at least 15, have had no prior of falls. At 12 months, the rate of hospital readmission was hip fracture at the newly fractured hip, have undergone 39% less in patients who received the high dose vitamin D3 surgical repair, have creatinine clearance of more than 15 therapy (95% CI –62% to –1%). Additional physiotherapy mL/min and to have been able to walk 3 m before their hip did not reduce the rate of hospital admission. Conclusion: fracture. Key exclusion criteria included metastatic cancer Additional physiotherapy and supplementation with high or chemotherapy, kidney stones, hypercalcaemia, primary dose vitamin D3 therapy of 2000 IU per day had different parathyroidism, sarcoidosis, or severe vision or hearing benefits after hip fracture. Additional physiotherapy reduced impairment. Randomisation of 173 participants allocated 42 the rate of falls and supplementation with high dose vitamin to standard physiotherapy and high dose vitamin D3 therapy, D3 therapy reduced the rate of hospital readmission. These 44 to additional physiotherapy and high dose vitamin D3 two interventions may be useful together as they address therapy, 44 to standard physiotherapy and standard vitamin two distinct but important complications after hip facture. D3 therapy, and 43 to additional physiotherapy and standard vitamin D3 therapy. Interventions: Both groups received 30 (Sherrington et al 2008). In the brief description of the exercise program in this paper, there appears to be limited Commentary focus on balance ( standing on both legs then standing on one leg while holding a handrail’). Other successful falls Hip fractures are predicted to increase in incidence by 36% prevention exercise programs such as the Otago program by 2051 in Australia (Sanders et al 1999). Studies aiming to (Robertson et al 2002) have incorporated a stronger focus improve outcomes in this group with effective and relatively on specific balance activities. Given that falls in most low cost interventions have potentially substantial impact cases caused the hip fracture in these patients, and balance for the individual, their family, and costs to the health impairment is strongly implicated in falls, it will be worth system. This study is a valuable addition to the limited investigating if stronger focus on balance performance can evidence regarding effective interventions in reducing falls achieve even better outcomes. or improving associated outcomes in this high risk group. Keith Hill Importantly, this study adds to the substantial evidence La Trobe University and Northern Health, Australia available that exercise programs can reduce falls in at-risk older people, although few of these studies have investigated References high risk clinical groups such as patients with hip fracture or stroke. The 25% reduction in falls, and a non-significant Hill K, Williams S (2009) Aust J Physiother 55: 224. although substantial reduction in hospitalisations, and hip fracture-related hospitalisations are impressive outcomes. Robertson MC et al (2002) J Am Geriatr Soc 50: 905–911. One critical element for physiotherapists is the content of Sanders K et al (1999) Med J Aust 170: 467–470. the exercise program (Hill and Williams 2009), particularly given the findings of a recent meta-analysis that a critical Sherrington C et al (2008) J Am Geriatr Soc 56: 2234–2243. element of successful fall prevention exercise programs is that they incorporate challenges to the balance system Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010 201

Jones et al: Slow breathing training to reduce hypertension An inspiratory load enhances the antihypertensive effects of home-based training with slow deep breathing: a randomised trial Chulee U Jones1, Benjarat Sangthong1,2 and Orathai Pachirat1 1Khon Kaen University, 2Rangsit University Thailand Question: Can adding an inspiratory load enhance the antihypertensive effects of slow breathing training performed at home? Design: Randomised trial with concealed allocation. Participants: Thirty patients with essential hypertension Stage I or II. Intervention: Experimental groups performed slow deep breathing at home, either unloaded or breathing against a load of 20 cmH2O using a threshold-loaded breathing device. Participants trained for 30 min, twice daily for 8 weeks. A control group continued with normal activities. Outcome measures: Resting blood pressure and heart rate were measured at home and in the laboratory before and after the training period. Results: Compared to the control group, systolic and diastolic blood pressure decreased significantly with unloaded breathing by means of 7.0 mmHg (95% CI 5.5 to 8.5) and 13.5 mmHg (95% CI 11.3 to 15.7), respectively (laboratory measures). With loaded breathing, the reductions were greater at 18.8 mmHg (95% CI 16.1 to 21.5) and 8.6 mmHg (95% CI 6.8 to 10.4), respectively. The improvement in systolic blood pressure was 5.3 mmHg /+9?'$&je/$,]h[Wj[hj^Wdm_j^kdbeWZ[ZXh[Wj^_d]$>[WhjhWj[Z[Yb_d[ZXo.X[Wji%c_d/+9?,$+je'&$)m_j^kdbeWZ[Z Xh[Wj^_d]\"WdZ/X[Wji%c_d/+9?+$,je'($(m_j^beWZ[ZXh[Wj^_d]$L[hoi_c_bWhc[Wikh[ie\\XbeeZfh[iikh[WdZ^[WhjhWj[ were obtained by the patients at home. Conclusion: Home-based training with a simple device is well tolerated by patients and produces clinically valuable reductions in blood pressure. Adding an inspiratory load of 20 cmH2O enhanced the decrease in systolic blood pressure. Trial registration: NCT007919689. <+POFT
$6
4BOHUIPOH
#
1BDIJSBU
0

\"O
JOTQJSBUPSZ
MPBE
FOIBODFT
UIF
BOUJIZQFSUFOTJWF
FGGFDUT
PG
IPNFCBTFE
USBJOJOH
XJUI
TMPX
EFFQ
CSFBUIJOH
B
SBOEPNJTFE
USJBM
Journal of Physiotherapy

o> ,FZ
XPSET: Loaded breathing, Slow deep breathing exercise, Blood pressure, Heart rate, Therapeutic, Hypertension, Physiotherapy Introduction a recent study with a device that uses a tone to control the rate of breathing (Altena et al 2009). Hypertension is an important and common co-morbidity associated with stroke, diabetes mellitus, cardiac and We have recently developed a simple device to train the renal disease. In developing countries the problem of inspiratory muscles (Jones et al 2004) which was designed hypertension is increasing as it is elsewhere in the world. to be affordable and acceptable to a wide range of patients. Given the potential number of patients affected there is The device may be used to regulate the pattern and depth a pressing need for effective, accessible, and affordable of breathing but can also provide a load for the respiratory treatments. Whole body exercise is generally recommended muscles to work against. Evidence is accumulating that as a key component in the management of hypertension. resistance training, at least with moderate loads, has no While cycling, jogging, aerobic exercise, and dance may be adverse effects and may well result in modest reductions acceptable to younger urban patients, these may not be so in blood pressure for moderately hypertensive individuals suitable for older, poorer, and rural patients for a variety (Kelley and Kelley 2000, Cornelissen and Fagard 2005). of practical and cultural reasons. There are, however, It is possible, therefore, that a combination of deep, slow some other promising non-pharmacological possibilities, breathing and an inspiratory load may be more effective in including breathing training. Improvements in blood reducing blood pressure than just regulating the pattern of pressure have been seen with yoga training that emphasises breathing. slow and regular breathing (Patel and North 1975) and several studies have shown that patients who train with slow Therefore the specific research questions for this study and regular breathing over a period of about eight weeks were: benefit from a reduction of blood pressure (Schein et al 2001, Grossman et al 2001, Rosenthal et al 2001, Elliot et al 2002, 1. Does unloaded deep and slow breathing training Viskoper et al 2003, Meles et al 2004). In these studies the reduce both systolic and diastolic blood pressure for pattern of breathing was guided by music, a metronome, or people with mild to moderate essential hypertension? similar feedback devices, some of which are now available commercially. There is, however, some controversy in this 2. Does combining this with an inspiratory load using a area, since no improvements in blood pressure were seen in water pressure threshold loading device increase the effect? Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010 179

Research the patients began training and in the week following the last training session. Statistical analysis was carried out by Methods an investigator blinded to the identity of the intervention groups. Design Participants The study was a randomised trial with concealed allocation and partial blinding. Patients with essential hypertension Patients were recruited from those routinely attending Stage I or II were recruited from the Outpatients the hypertension clinic of Srinagarind Hospital and came Department, Srinagarind Hospital, Khon Kaen, Thailand. from mixed urban and rural areas around Khon Kaen in Following an initial assessment the patients were assigned the north east of Thailand. Inclusion criteria were: essential to one of three intervention groups by block randomised, hypertension Stage I or II (systolic blood pressure 140– concealed allocation: a control group, those training 179, diastolic blood pressure 90–109 mmHg) based on with unloaded breathing, and those training with loaded recommendations of JNC-VII (Chobanian et al 2003); age breathing (see Figure 1). One week before the study, 35–65 years; good understanding and communication; participants visited the Cardiovascular Research Room at independent ambulation. Exclusion criteria were: secondary Khon Kaen University to familiarise themselves with the hypertension; respiratory disease; diabetes mellitus; general procedures and to learn how to measure and record cardiac, renal or cerebrovascular disease; dyslipidemia; blood pressure and heart rate at home. The unloaded and pregnancy within the last 6 months. Medication was loaded breathing groups also learnt how to use the water continued unchanged for the duration of the study (10 pressure threshold loading device and practised their weeks). Recruitment was by medical staff and nurses of the allocated deep breathing technique (ie, unloaded or loaded). Hypertension Unit of Srinagarind Hospital. Measurements of resting heart rate and blood pressure were made both by the patients themselves in their home setting and by the investigators in the laboratory in the week before Screened for eligibility (n = 70) Excluded (n = 30) š Diabetes (n = 10) š Secondary hypertension (n = 8) š Coronary artery disease (n = 4) š Respiratory disease (n = 2) š Training ‘too onerous’ (n = 6) (n = 13) Randomised (n = 40) (n = 13) Practice and familiarisation with training and testing procedures (n = 14) Week 0 Measured resting blood pressure and heart rate daily 7 days a week at home and one occasion in the laboratory (n = 13) (n = 14) (n = 13) Loaded Lost to Unloaded Lost to Control Lost to breathing group breathing group follow-up follow-up šde follow-up š)&c_dibem šYedijhW_dji group šYedijhW_dji šZ _\\ÅYkbjje š)&c_dibem breathing deep breathing on time for on time for training attend (n = 3) deep with 20 cmH2O training (n = 2) breathing training (n = 2) inspiratory load šZ _\\ÅYkbjje š(%ZWon. šZ _\\ÅYkbjje š(%ZWon. attend (n = 1) weeks attend (n = 2) weeks Measured resting blood pressure and heart rate at home and in the laboratory: data analysis Week 9 (n = 10) (n = 10) (n = 10) 'JHVSF
 Design and flow of participants through the trial. 180 Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010

Jones et al: Slow breathing training to reduce hypertension AB arm blood pressure monitoring devicea. Two measurements were made in the morning between 7.00 and 9.00 am, after 'JHVSF
 The Water Pressure Threshold Bottle used at least 5 minutes rest while sitting in a comfortable chair. for loaded breathing training. A. Using the bottle for Subjects were asked to refrain from physical activity or loaded breathing. B. Diagram showing the flow of air and caffeine for at least 30 minutes before the measurement. how the depth of the right hand tube sets the threshold Resting heart rate was measured by the same device whilst inspiratory pressure. the blood pressure was being measured. Data were recorded daily in the week before training and likewise in the week Intervention after the training program had ended. Two measurements were made on each day and the values averaged to give For training, the patients used a new simple loaded single values for that day. The measurements made on the breathing device, the Water Pressure Threshold Bottle, seven days during each of these weeks were averaged to developed in our laboratory (Figure 2). The device consists give single values pre- and post-training for each patient. of a plastic bottle with two tubes passing through the lid. Patients were contacted once a week during the training to One tube provides an outlet through the top of the bottle monitor their well-being and compliance. and is connected with corrugated tube to a mouthpiece, while the other is a longer adjustable inlet tube passing Laboratory-based measurements: Laboratory-based blood into the water. The subjects breathed in through the pressure measurements were made on one occasion in the mouthpiece and out through their nose. Thus, inspiratory week before training and within 3 days of the end of the resistance was determined by the column of water that was training. Blood pressure was measured between 9.00 and displaced, set by the length of the inlet tube below the water 12.00 am with an automatic digital bedside monitorb after in the cylinder. The device is simple and easy to use and at least 15 minutes rest while sitting. Subjects were asked adjust. It has the added advantage that the inspired air is not to smoke or consume caffeine for 30 minutes before humidified and the bubbling sound acts as feedback helping the measurements. The electrocardiogram was recorded to establish a steady breathing pattern. A preliminary with bipolar limb leads and resting heart rate calculated study with healthy elderly subjects found no evidence of from averaged three consecutive R-R intervals. Two hypocapnia, no changes in blood pressure, and only a small measurements were made on each occasion and the values rise in heart rate while using the device (Jones et al 2004). were averaged to give single values pre- and post-training for Participants were trained by physiotherapists from Khon each patient. Participants were trained by physiotherapists Kaen University. from Khon Kaen University. Training protocols: Patients in the unloaded breathing Data analysis group inhaled deeply through the device with the inlet tube set just above the level of the fluid so the inspired air We sought to detect a difference of 10 mmHg in blood was humidified but there was no added resistance. For the pressure between groups. Assuming a standard deviation loaded breathing group, the water level was set to provide an of 7.5 mmHg, 10 participants per group would provide 80% inspiratory load of 20 cmH20. The patients were instructed power to detect as significant, at the two-sided 5% level, to adopt a breathing pattern with a controlled flow rate of a 10-mmHg difference in blood pressure between groups. about 200 ml/sec, an inspiratory time of 4 seconds, and a To allow for loss to follow-up, the total sample size was total respiratory time of 10 seconds. The paced breathing increased to 40 participants. was first practised using a metronome in the laboratory until it could be reliably performed without the metronome. Pulse pressure was taken as the difference between systolic Patients rested for 5 seconds after every 6 deep breaths. and diastolic pressures and mean arterial pressure was Training was performed at home for 30 minutes, twice a calculated as diastolic blood pressure plus one-third of day, every day for 8 weeks. Patients in the control group pulse pressure. A two-way AVOVA with post hoc analysis were asked to continue with their normal daily life. 5BCMF
. Baseline characteristics of participants. Outcome measures Characteristic Control Unloaded Loaded Home-based measurements: Subjects were taught to (n = 10) (n = 10) (n = 10) measure their blood pressure at home with a digital upper- Gender (n male) Age (yr) mean (SD) 3 4 4 Weight (kg) mean 50 (5) 53 (4) 51 (5) (SD) 63 (7) 68 (12) 76 (11) Height (cm) mean (SD) 156(5) 158 (10) 163 (5) BMI (kg/m2) mean (SD) 27 (2) 27 (5) 29 (5) Duration of hypertension (yr) 5.6 (2.5) 5.8 (3.4) 5.7 (2.3) mean (SD) Duration of 4.7 (1.6) 5.1 (2.8) 4.9 (1.7) medication (yr) mean (SD) Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010 181

Research into the trial, 10 dropped out very early in the training for a variety of reasons, mainly because of difficulty (Tukey’s test) was used to compare the mean values before attending the laboratory or finding the time to train. and after training within groups and differences in mean Details of the participants completing the study are given changes between groups. Data are presented as means and in Table 1. All participants in all groups were taking one standard deviations or 95% CIs. Statistical significance was or two of the following medications: enalapril, atenolol, or assumed at p ≤ 0.05. Statistical analysis was carried out by hydrochorothiazide. an investigator blinded to the identity of the intervention groups. $PNQMJBODF
XJUI
UIF
USJBM
NFUIPE Results No participants withdrew, or were withdrawn, for medical reasons or difficulty with the training. The 30 patients who 'MPX
PG
QBSUJDJQBOUT
UISPVHI
UIF
USJBM completed the full 10 weeks of the study showed excellent The flow of participants is presented in Figure 1. Of the 70 compliance ( 95%) with the training and data recording. patients who volunteered, 40 were included in the trial after The participants commented that the training, especially the the initial screening. Of the 40 patients initially accepted Change in SBP (mm Hg) 6A Unloaded Loaded Change in SBP (mm Hg) 6B Unloaded Loaded 3 * ** 3 * ** 0 0 -3 Control -3 Control -6 -6 -9 -9 -12 -12 -15 -15 -18 -18 -21 -21 Change in DBP (mm Hg) 6C Unloaded Loaded Change in DBP (mm Hg) 6D Unloaded Loaded 3 * * 3 * * 0 0 -3 Control -3 Control -6 -6 -9 -9 -12 -12 -15 -15 -18 -18 -21 -21 6E 6F Change in PP (mm Hg) 3 Unloaded Loaded Change in PP (mm Hg) 3 Unloaded Loaded 0 0 -3 Control -3 Control -6 * -6 -9 -9 -12 * * -12 * -15 -15 -18 -18 -21 -21 * 'JHVSF
. Mean and SD for changes in systolic (SBP), diastolic (DBP) and pulse pressure (PP) with training. A and B systolic blood pressure changes; C and D diastolic blood pressure changes; E and F pulse pressure changes in Control participants or following eight weeks of training with either Unloaded or Loaded breathing. A , C and E, measurements made by the patients at home; B, D and F, measurements made in the laboratory. * = Values that differ significantly from the control group. ** = Values that differ from the control group and also from the unloaded training group. 182 Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010

JournalofPhysiotherapy2010 Vol.56 – ©AustralianPhysiotherapyAssociation2010 5BCMF

Mean (SD) for each group at Week 0 and 8, mean (SD) difference within groups, and laborator y. Groups D SBP(mmHg) LB Week 0 Con LB Week 8 Con W Home n = 10 n = 10 n = 10 n = 10 ULB ULB L Laborator y 142 n = 10 142 130 n = 10 144 (8.9) (9.6) (8.0) (8.8) –1 DBP (mmHg) 137 141 131 122 134 135 (2. Home (12.7) (5.9) (9.1) (12.8) (5.9) (10.9) –1 136 126 (3. Laborator y 87 (12.6) 87 80 (13.7) 88 (5.2) (5.3) (5.8) (4.8) – PP (mmHg) 85 79 (2 Home 81 (4.4) 78 75 (3.9) 80 (8.8) (6.5) (7.6) (7.3) – Laborator y 80 75 (2. 55 (5.5) 56 50 (6.7) 57 MAP (mmHg) (9.4) (9.8) (8.1) (9.2) – Home 57 55 (3. 59 (4.9) 53 47 (5.4) 55 –1 Laborator y (14.2) (9.4) (14.9) (10.6) (9. 56 51 Heart rate (beats/min) 105 (13.7) 105 97 (12.4) 106 – Home (4.9) (5.3) (5.5) (5.0) (2. 101 104 98 Laborator y (7.9) (4.4) 95 90 (4.0) 98 –1 (6.0) (7.5) (7.1) (6. 75 99 92 (5.8) (5.5) 73 69 (7.5) 76 – (7.3) (5.9) (8.1) (3. 76 74 69 (4.2) (5.6) 75 68 (6.2) 76 – (9.6) (6.2) (7.6) (4. 73 65 (7.3) (7.1) LB = loaded breathing, ULB = unloaded breathing, Con = control groups, SBP = systolic blood pressure, D 183

mean (95%CI) differences between groups for all outcomes measured at home and in the Difference within groups Difference between groups Week 8 minus Week 0 Week 8 minus Week 0 LB ULB Con LB minus ULB LB minus Con ULB minus Con 12 –7 2 – 4.7 –14.0 – 9.3 .9) (2.2) (2.6) (–2.4 to –7.0) (–11.6 to –16.4) (–7.5 to –11.1) 15 –10 4 – 5.3 –18.8 –13.5 .3) (6.1) (2.8) (–1.0 to –9.6) (–16.1 to –21.5) (–11.3 to –15.7) –7 –5 1 –1.6 –7.5 – 5.9 2.1) (2.2) (3.2) (0.3 to –3.5) (–5.1 to –9.9) (–3.8 to –8.0) –6 –5 2 –1.6 – 8.6 –7.0 .2) (2.8) (1.9) (0.6 to –3.8) (– 6.8 to –10.4) (–5.5 to –8.5) –5 –2 1 – 3.1 – 5.7 –2.6 .2) (2.7) (1.6) (– 0.5 to –5.7) (–3.5 to –7.9) (–1.3 to –3.9) 12 –5 1 – 6.4 –12.9 – 6.5 .5) (5.9) (1.9) (0.5 to –13.3) (– 6.9 to –18.9) (– 4.7 to –8.3) –9 –6 1 –2.7 – 9.5 – 6.8 Jones et al: Slow breathing training to reduce hypertension .2) (1.9) (3.1) (– 0.9 to – 4.5) (–7.6 to –11.4) (– 4.8 to –8.8) 11 –6 3 – 4.8 –13.9 – 9.1 .6) (3.1) (2.2) (– 4.7 to –8.3) (–9.6 to –18.2) (–7.5 to –10.7) –7 –5 3 –1 –9 –8 .5) (1.6) (2.8) (1.4 to –3.4) (– 6.4 to –12.0) (– 6.3 to –10.1) –8 –8 1 –1 –9 –8 .5) (1.1) (2.8) (2.4 to –3.4) (–5.6 to –12.2) (– 6.5 to –10.3) DBP = diastolic blood pressure, PP = pulse pressure, MAP = mean arterial pressure

Research fell significantly in both the unloaded and loaded training groups of patients (Table 2, Figure 4). loaded breathing, was hard work but perfectly acceptable. Blood pressure and heart rate measures were made both Discussion by the participants themselves whilst at home and by the investigators when participants visited the laboratory. Controlled slow breathing training using a relatively There was good agreement between these two sets of simple threshold loading device resulted in significant and measurements, with similar changes evident in the two data clinically valuable reductions in systolic blood pressure, sets (Table 2). diastolic blood pressure, pulse pressure, and heart rate. Adding a resistive load to the inspiratory muscles generally Effect of intervention enhanced the benefits, significantly so, for systolic blood pressure. These findings were replicated in home-based Data for the cardiovascular parameters before and after blood pressure monitoring by the patients and in laboratory the 8-week training period are given in Table 2, together assessments, indicating that they were not affected by the with differences within and between groups. Participants white coat’ effect that can complicate the measurement of in the control group showed minimal change in any of the blood pressure and confuse the interpretation of trial results. measured parameters. Both the training groups showed significant reductions in systolic and diastolic blood The baseline characteristics of the participants, including pressures of 5 to 15 mmHg (Table 2, Figure 3) with very their medication use, were very similar between the similar changes seen in the measurements made at home by groups, with only slightly greater height and weight in the the patients and in the laboratory. The reductions in blood loaded breathing group. The pre-training cardiovascular pressure were somewhat greater for the loaded breathing parameters were very similar in the three groups. group, with the difference between the two groups reaching statistical significance for systolic blood pressure, measured The threshold loading device is very suitable for home use either at home or in the laboratory (Table 2, Figure 3A and and has the advantage that the air is humidified – avoiding B). The changes in systolic blood pressure were greater the unpleasant dry mouth and throat normally associated than those in diastolic blood pressure with the consequence with breathing through a mouthpiece. Such a relatively that pulse pressure was also reduced significantly when simple and inexpensive device could therefore be a valuable measured both at home and laboratory (Table 2, Figure 3E and F). Mean arterial pressure and resting heart rate also Change in MAP (mm Hg) 6A Unloaded Loaded Change in MAP (mm Hg) 6B Unloaded Loaded 3 * * 3 * * 0 -3 Control 0 Unloaded Loaded -6 Control -9 * * -12 -3 -15 -18 -6 Change in HR (bpm) 6C Change in HR (bpm) -9 4 -12 2 Unloaded Loaded 0 -15 -2 Control -18 -4 6D -6 4 -8 * 2 0 -10 -2 -4 Control -12 * -6 -8 -14 -10 -12 -14 Figure 4. Mean and SD for changes in mean arterial pressure (MAP) and heart rate (HR). A and B mean arterial pressure changes; C and D resting heart rate changes in Control participants or following eight weeks of training with either Unloaded or Loaded breathing. A and C, measurements made by the patients at home; B and D, measurements made in the laboratory. * = Values that differ significantly from the control group. 184 Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010

adjunct to conventional approaches for treatment of Jones et al: Slow breathing training to reduce hypertension hypertension in all communities. the 20-cmH2O threshold the minimal airflow maintained Although participants and assessors were not blinded, for the 4-s inspiratory time ensured a relatively large chest participants were not informed that there were loaded and expansion. This lung inflation and the negative intrathoracic unloaded breathing groups, so this may have reduced some pressures generated may have activated pulmonary stretch sources of bias due to lack of blinding on this comparison. receptors and the Hering-Breuer inflation reflex, which The potential problems of an unblinded study were further would reduce heart rate and systemic vascular resistance. minimised by the nature of the measurements since blood pressure and heart rate were recorded automatically and The mechanisms by which breathing training results required no particular skill or judgments to be made in reductions of blood pressure are not clear. It has been either by the participants at home or the experimenters in suggested that in essential hypertension there is enhanced the laboratory (Wood et al 2008). Furthermore, the post- sympathetic activity (Guzzetti et al 1988, Goldstein, training measurements were all made without either the 1993) pressor hyper-responsiveness (Goldstein 1993), participants or the experimenters having access to the pre- and reduced vagal activity at rest (Guzzetti et al 1988). training data measured some eight weeks earlier. Since the breathing training reduced resting systolic and diastolic blood pressure together with heart rate, one The consequences of unloaded breathing training for mechanism of its action may be that the training increased systolic and diastolic blood pressure were very similar cardiac vagal tone and reduced sympathetic activity to the to previous reports where breathing has been regulated cardiac and peripheral arterioles. It is known that resistive in various ways (Schein et al 2001, Grossman et al 2001, slow deep breathing at elevated tidal volumes – as in this Rosenthal et al 2001, Elliot et al 2002, Viskoper et al 2003), study – leads to decreased sympathetic excitation (Seals with the mean changes being 6 to 10 mmHg for systolic et al 1993). Hyperventilation and low end-tidal carbon and diastolic blood pressure for all the trials, including the dioxide pressures at rest have been described in essential present one. The reductions in blood pressure achieved in hypertension (Joseph et al 2005), which could enhance this way are clinically valuable and appreciably greater than peripheral chemoreflex sensitivity (Trzebski et al 1982) and those reported for aerobic physical training reductions of sympathetic activity. Slow breathing training may reduce 3.8 and 2.6 mmHg for systolic and diastolic blood pressure hyperventilation at rest, as seen in yoga practice, thereby (Whelton et al 2002) which is generally recommended as an altering the chemoreflex sensitivity (Spicuzza et al 2000). A adjunct to treatment for hypertension. change in baroreflex sensitivity is another possibility as the baroreflex-cardiac sensitivity is shown to be decreased in It is of particular interest that both training modes reduced hypertension (Goldstein 1993), and the effects of slow deep systolic blood pressure and pulse pressure. Systolic blood breathing reducing blood pressure have been suggested pressure is considered a better predictor of cardiovascular to be mediated via an increase in baroreflex sensitivity complications than diastolic blood pressure (Lewington et (Joseph et al 2005). Loaded breathing has also been shown al 2002). It has recently been suggested that systolic blood to increase the aortic baroreflex as a consequence of pressure should be the target of treatment in people aged augmented negative intrathoracic pressures (Angell-James over 50 years with hypertension (Williams et al 2008) but 1971). The greater response of systolic blood pressure found controlling systolic blood pressure with pharmacological with loaded slow deep breathing may be a consequence of measures is more difficult than controlling diastolic blood the load amplifying some of the mechanisms discussed pressure (Waeber and Mourad 2006). Pulse pressure above. has been reported to be a good predictor of myocardial infarction and cardiovascular risk in hypertensive subjects The results presented here suggest that the key factor (Benetos 1999) and was the only measure of blood pressure in reducing blood pressure is deep inspiration and lung found to be significantly and independently related to the inflation. However, one of the most common commercially incidence of myocardial infarction in treated and untreated available devices, RESPeRate, emphasises the control of hypertension (Fang et al 1995). The greater reduction in expiration. It may be the case that any form of controlled systolic blood pressure using loaded breathing training slow breathing rate is sufficient to reduce diastolic blood in the present study indicates that this method could be a pressure. Alternatively, although RESPeRate aims to valuable adjunct treatment for older hypertensive people control expiration, in order to be able to breathe out slowly and in cases of isolated systolic hypertension. subjects need to take a deep breath in, thus providing a degree of lung inflation. In either case it seems important Our findings differ from previous work involving breathing to have a high level of lung inflation in order to obtain the training in that there was a consistent reduction of 5 to 8 decreases in systolic pressure that we have observed. beats/min in resting heart rate as a result of both loaded and unloaded breathing whereas previous studies of breathing We conclude that controlled breathing using this novel training report no change in heart rate (Schein et al 2001, and simple device for 8 weeks is well tolerated by patients Grossman et al 2001, Rosenthal et al 2001, Viskoper et al for home-based training and provides clinically valuable 2003). These previous studies used devices which guided reductions in blood pressure. Adding an inspiratory load of the breathing rate but did not necessarily control the depth of 20 cmH2O enhanced the decrease in systolic blood pressure, inspiration, as is evident from the high variation in the ratio an important target for the reduction of cardiovascular risk of inspiratory to expiratory times during breathing training in people with hypertension. For such training to be widely with RESPeRate (Schein et al 2007). With the pressure used, however, further studies will be required to determine threshold device we have used, it is necessary to maintain the minimum duration and intensity of training needed to a certain inspiratory pressure to obtain any air flow. With produce useful changes and how long the effects last after the end of training so that the frequency with which patients need to train can be determined. Q Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010 185

Research Human Hypertension 15: 263–269. Ethics: The trial was approved by the Ethical Committee Jones C, Pongsuwan P, Khrisanapant W, Bunsawat W (2004) for Human Research of Khon Kaen University. Participants Cardiopulmonary responses in the elderly during inspiratory received full information about the nature of the study muscle loading. European Respiratory Journal 24: 243s. before providing written consent. Joseph CN, Porta C, Casucci G, Casiraghi N, Maffeis M, Rossi Support: This study was supported by grants from Thai M, et al (2005) Slow breathing improves arterial baroreflex Health Promotion Foundation, Ministry of Public Health, sensitivity and decreases blood pressure in essential Graduate School and Faculty of Associated Medical hypertension. Hypertension 46: 714–718. Sciences, Khon Kaen University, Thailand. Kelley GA, Kelley KS (2000) Progressive resistance exercise Acknowledgements: The authors are grateful to the and resting blood pressure: a meta-analysis of randomized patients, nurses and officers of the Hypertension Clinic controlled trials. Hypertension 35: 838–843. of Srinagarind Hospital for their assistance in the conduct of the present study. We thank Professor David Jones for Lewington S, Clarke R, Qizilbash N, Peto R, Collins R (2002) useful discussions and help with preparing the manuscript. Age-specific relevance of usual blood pressure to vascular mortality: a meta-analysis of individual data for one million Competing interests: None declared. adults in 61 prospective studies. Lancet 360: 1903–1913. Footnotes: a Ri-champion , Rudolf Riester GMBH Co, Meles E, Giannattasio C, Failla M, Gentile G, Capra A, Mancia Germany. b Nihon Kohden-life scope , Nihon Kohden G (2004) Non-pharmacologic treatment of hypertension by Corporation, Tokyo, Japan. respiratory exercise in the home setting. American Journal of Hypertension 17: 370–374. Correspondence: Dr Chulee U Jones, Assistant Professor, Department of Physical Therapy, Khon Kaen University, Patel C, North W (1975) Randomized controlled trial of yoga Thailand. Email: chulee kku.ac.th and biofeedback in management of hypertension. Lancet 19: 93–95. References Rosenthal T, Alter, Peleg E, Gavish B (2001) Device-guided Altena MR, Kleefstra N, Logtenberg SJ, Groenier KH, breathing exercises reduce blood pressure: ambulatory and Houweling ST, Bilo HJ (2009) Effect of device–guided home measurements. American Journal of Hypertension 14: breathing exercises on blood pressure in patients with 74–76. hypertension: a randomized controlled trial. Blood Pressure 18: 273–279. Schein MH, Gavish B, Herz M, Rosner-Kahana D, Naveh P, Knishkowy B (2001) Treating hypertension with a device Angell-James JE (1971). The effects of changes in extramural, that slows and regularises breathing: a randomised, double- ‘intrathoracic’ pressure on aortic arch baroreceptors. Journal blind controlled study. Journal of Human Hypertension 15: of Physiology (London) 214: 89–103 271–278. Benetos A (1999) Pulse pressure and cardiovascular risk. Schein MH, Alter A, Levine S, Baeysky T, Nessing A, Gavish Journal of Hypertension 17: s21–s24. B (2007) High blood pressure reduction in diabetics with interactive device-guided paced breathing: final results of Chobanian AV, Bakris GL, Black HR, Cushman WC, Green a randomized controlled study. Journal of Hypertension 25: LA, Izza JL, et al (2003) The seventh report of the joint S192. national committee on prevention, detection, evaluation, and treatment of high blood pressure. Journal of American Spicuzza L, Gabutti A, Porta C, Montano N, Bernardi L (2000) Medical Association 289: 2560–2572. Yoga and chemoreflex response to hypoxia and hypercapnia. Lancet 356: 1459–1496. Cornelissen VA, Fagard RH (2005) Effect of resistance training on resting blood pressure: a meta analysis of randomized Trzebski A, Tafil M, Zoltowski M, Przybylski J (1982) Increased controlled trials. Journal of Hypertension 23: 251–259. sensitivity of the arterial chemoreceptor drive in young men with mild hypertension. Cardiovascular Research 16: 163– Elliott W, Izzo JL, Rosing D, Synder C, White WB, Alter A 172. (2002) Hypertension reduction by device–guided breathing shows a dose–response relationship. American Journal of Viskoper R, Shapira I, Priluck R, Mindlin R, Chornia L, Hypertension 22: s116. Laszt A (2003) Nonpharmacologic treatment of resistant hypertensive by device-guided slow breathing exercises. Fang J, Madhaven S, Cohen H, Alderman MH (1995) Measures American Journal of Hypertension 16: 484–487. of blood pressure and myocardial infarction in treated hypertensive patients. Journal of Hypertension 13: 413–419. Waeber B, Mourad JJ (2006) Targeting systolic blood fh[iikh[0 j^[ a[o je Yedjhebb_d] YecX_d[Z ioijeb_Y%Z_Wijeb_Y Giannattasio C, Failla M, Meles E, Gentile G, Grappiolo A, hypertension. American Journal of Hypertension 19: 985– Mancia G (2002) Efficacy of self treatment of hypertension 986. at home with device-guided breathing. American Journal of Hypertension 15: 186A. Whelton SP, Chin A, Xin X, He J (2002) Effect of aerobic exercise on blood pressure: a meta-analysis of randomized, Goldstein DS (1983) Arterial baroreflex sensitivity, plasma controlled trials. Annals of Internal Medicine 136: 493–503. catercholamines and pressor responsiveness in essential hypertension. Circulation 68: 234–240. Williams B, Lindholm LH, Sever P (2008) Systolic pressure is all that matters. Lancet 371: 2219–2221. Grossman E, Grossman A, Schein MH, Zimlichman R, Gavish B (2001) Breathing-control lowers blood pressure. Journal of Wood L, Egger M, Gluud LL, Schulz KF, Jüni P, Altman DG, et al (2008). Empirical evidence of bias in treatment effect estimates in controlled trials with different interventions and outcomes: meta-epidemiological study. British Medical Journal 336: 601–605. 186 Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010

Appraisal Critically Appraised Papers Ankle exercises in combination with intermittent ice and compression following an ankle sprain improves function in the short term Synopsis following four weeks a standardised treatment consisting of ankle rehabilitation exercises was provided to both groups. Summary of: Bleakley CM, O’Connor SR, Tully MA, Outcome measures: The primary outcome was subjective Rocke LG, MacAuley D, Bradbury I, et al (2010) Effect of ankle function assessed by the lower extremity functional accelerated rehabilitation on function after ankle sprain: scale (0–80) at weeks 1 to 4. Secondary outcomes assessed randomised controlled trial. BMJ 340: c1964 doi:10.1136/ were: pain at rest and pain with activity with 10-cm visual bmj.c1964 Prepared by Margreth Grotle and K re Birger analogue scales, swelling by a modified version of the Hagen, CAP Editors. figure of eight method, and physical activity by a physical activity logger. Ankle function by the Karlsson score and Question: What is the effect of an accelerated intervention rate of reinjury were also assessed at 16 week follow-up. incorporating early therapeutic exercise as compared to a Results: 15 of the 101 patients dropped out during the trial, standard intervention of protection, rest, ice, compression, 11 in the exercise group and 4 in the standard group. An and elevation after acute ankle sprain? Design: Randomised, effect was found in favour of the exercise group with the controlled trial with blinded outcome assessment and lower extremity functional scale (0–80) at week 1 (MD 5.3, intention-to-treat analysis. Setting: An emergency 98.75% CI 0.3 to 10.3) and week 2 (MD 4.9, 95% CI 0.3 to department and sports injury clinic in Northern Ireland. 9.6). In addition, the exercise group was more active in the Participants: Men and women 16–65 years, with acute ( 7 first week as measured by time spent walking (0.4 hours per days) grade 1 or 2 ankle sprain. Key exclusion criteria were day, 95% CI 0.2 to 0.6). No between-group differences were complete (grade 3) rupture, bony ankle injury, and multiple observed for pain at rest, pain with activity, or swelling. injuries. Randomisation allocated 101 participants to an At 16 weeks there were no significant differences between accelerated intervention incorporating early therapeutic the groups in the Karlsson score or reinjury rate (2 in each exercises (exercise group) or a standard protection, rest, group). Conclusion: An accelerated exercise protocol ice, compression, and elevation intervention (standard during the first week after ankle sprain improved ankle group). Interventions: During the first week after baseline function and early return to weight bearing activity. both groups received written advice on using ice and compression. The exercise group also undertook 20 minutes Between-group difference in time spent walking per day of exercises three times a day focused on increasing ankle calculated by CAP editors range of movement, activation and strengthening of ankle musculature, and restoring sensorimotor control. In the the first week ( 20 min, 3 times a day’). However, based on the study protocol previously published (Bleakley et al Commentary 2007) we assume that the exercises were prescribed daily during the first week. For general practitioners, as well as This study is the first to describe the effect of early sports physicians and physiotherapists, seeing patients with mobilisation in combination with the standard PRICE acute ankle sprains in the clinic, these findings emphasise (Protection, Rest, Ice, Compression, Elevation) treatment the importance of prescribing exercises in combination after an acute ankle sprain using a randomised controlled with the PRICE protocol in the first week after injury to trial where, instead of rest, the intervention group performed optimise rehabilitation. However, the optimal dosage of therapeutic exercises aimed at increasing ankle movement, as treatment, including PRICE, choice of exercises, intensity well as static strengthening and stretching exercises (Knight and frequency of the exercise protocol, requires further 1995). The main finding was a significant improvement in investigation. short-term ankle function for those completing the exercise protocol during the first week following an ankle sprain. Kathrin Steffen, Agnethe Nilstad It is worth noting that the size of the effect (expressed Norwegian School of Sport Sciences, Oslo, Norway as change in the lower extremity functional score from baseline to week 1) was smaller than the change of 9 points References nominated as the clinically important change. However, as the adherence rate to the prescribed treatment was higher Bleakley CM (2007) BMC Musculoskelet Disord 8: 125. in the standard group compared to the exercise group (77.9% versus 67.8%), the effect for those adhering to the Knight KL (1995) Champaign IL, Human Kinetics: 220–229. exercise protocol might have been higher than confirmed by the published results. The authors did not describe in detail how often the exercises should be performed during 202 Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010

Shields and Taylor: Resistance exercise in Down syndrome A student-led progressive resistance training program increases lower limb muscle strength in adolescents with Down syndrome: a randomised controlled trial Nora Shields and Nicholas F Taylor La Trobe University, Australia Question: Does progressive resistance training improve muscle strength and physical function in adolescents with Down syndrome? Design: Randomised controlled trial with concealed allocation, assessor blinding, and intention-to-treat analysis. Participants: 23 adolescents with Down syndrome (17 boys, 6 girls; mean age 15.6 ± 1.6 years) were randomly assigned to either an experimental group (n = 11) or a control group (n = 12). Intervention: The intervention was a student-led progressive resistance training program, comprising 6 exercises using weight machines performed twice a week for 10 weeks. Participants completed 3 sets of 12 repetitions of each exercise or until they reached fatigue. The intervention took place in a community gymnasium. The control group continued with their usual activities. Outcome measures: The outcomes measured at baseline and immediately after the intervention phase were muscle strength (1 repetition maximum), a timed stairs test, and the grocery shelving task. Results: The experimental group attended 90% of their scheduled sessions. They demonstrated improvement in lower limb muscle strength compared to the control group (MD 36 kg, 95% CI 15 to 58). There were no significant differences between the groups for upper limb muscle strength or physical function measures. No major adverse events were recorded. Conclusion: Progressive resistance training is a feasible and safe exercise option that can improve lower limb muscle strength in adolescents with Down syndrome. Trial registration: ACTRN12608000261314. <4IJFMET
/
5BZMPS
/'

\"
TUVEFOUMFE
QSPHSFTTJWF
SFTJTUBODF
USBJOJOH
QSPHSBN
JODSFBTFT
MPXFS
MJNC
NVTDMF
TUSFOHUI
JO
BEPMFTDFOUT
XJUI
%PXO
TZOESPNF
B
SBOEPNJTFE
DPOUSPMMFE
USJBM
Journal of Physiotherapy

o> ,FZ
XPSET: Exercise, Randomized controlled trial, Resistance training, Down syndrome Introduction Shields et al 2008, Weber and French 1988). These studies found improved upper (Davis and Sinning 1987, Rimmer Good muscle strength is particularly important for young et al 2004, Weber and French 1988) and lower limb muscle people with Down syndrome because their workplace strength with training (Rimmer et al 2004, Weber and activities typically emphasise physical rather than cognitive French 1988). Only one of these studies investigated the skills (Shields et al 2008). The physical component of work effect of progressive resistance training in adolescents with tasks can be a problem because of muscle weakness. Muscle Down syndrome (Weber and French 1988), but it did not strength in the upper (Pitetti et al 1992) and lower limbs include a control group in its design, the assessors were not (Croce et al 1996) is up to 50% less in people with Down blind to group allocation, and it did not report the effects of syndrome compared to their peers with typical development the training on functional activities. Therefore, because of and also compared to their peers with an intellectual potential biases in research design, it is not known to what disability but without Down syndrome. Muscle weakness extent the reported effects are due to the intervention, or if can also impact their ability to perform everyday activities, any improvements in muscle strength carried over into an including walking (Carmeli et al 2002). Improvement improved ability to complete functional tasks. in strength has been associated with positive changes in functional activities in adults with Down syndrome Adolescence is a strategic time to implement an exercise (Carmeli et al 2002) and in work-related skills in people program as establishing good exercise habits early in life with intellectual disability ( etts et al 1995). is an important predictor of continued healthy activity patterns in adulthood (Telama et al 2005). Children with People with intellectual disability have the capacity to Down syndrome become less active during adolescence improve their muscle strength with progressive resistance (Shields et al 2009). It is especially important for young training (Shields and Dodd 2004). In progressive resistance people with Down syndrome to exercise because they training, high loads are lifted for a low number of repetitions have lower cardiovascular fitness than their peers without before muscular fatigue, and the load is progressed as the disability (Baynard et al 2008). The causes of their lower person gets stronger (American College of Sports Medicine fitness are unclear but are due in part to their low peak heart 2009). Only four trials have investigated the effects of rate (approximately 30% below expected) and may be due to progressive resistance training in people with Down their reduced physical activity levels, ventilatory difficulties, syndrome (Davis and Sinning 1987, Rimmer et al 2004, and reduced muscle strength (Khalili and Elkins 2009; Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010 187

Research Baynard et al 2008). People with Down syndrome are also and were fit and well enough to participate in the training predisposed to a higher incidence of cardiovascular disease program. The last inclusion criterion was ascertained by (Hill et al 2003), diabetes (Hermon et al 2001), osteoporosis asking parents to complete the 7-item Physical Activity and obesity, and so are more susceptible to a premature and Readiness questionnaire on behalf of their child. The level significant decline in function as they age (Rimmer et al of intellectual disability of each participant (described as 2004). It is also a pertinent time because future employment mild, moderate, or severe as perceived by their parent) may be dependent on their physical ability. was documented. Parent perceptions were used to give a general indication of the level of disability of their child Adolescents with Down syndrome should be encouraged to and because of concerns about formal intelligence testing engage in exercise as they transition to adulthood. However, in this population (American Association on Intellectual they face significant barriers to participation in exercise and Developmental Disabilities 2010). Participants were including a need for someone to exercise with (Heller et excluded if they had participated in a progressive resistance al 2002) and a need for suitable programs (Menear 2007). training program in the 6 months prior to the trial. Facilitators of exercise for this group include: the need for close supervision, motivational support, and the need to We adopted a 40% increase in 1RM leg press as the minimum ensure they exercise at the correct intensity (Shields et al clinically important difference based on a previous trial 2008), and to provide for social interaction (Menear 2007). by Rimmer et al (2004). The standard deviation in 1RM Exercising at a gym is a socially acceptable activity for leg press in a similar population was 41.5 kg (Rimmer et typically developing adolescents, and might be a reasonable al 2004). From this, we calculated that to maintain power recreation option for adolescents with Down syndrome. of 80% with a significance level of 0.05, we required 11 The aim of this trial therefore, was to determine the effects participants per group to complete the study. of a student-led community-based progressive resistance training program for adolescents with Down syndrome. A Intervention student-led program provides the supervision and social interaction adolescents with Down syndrome need to The experimental group completed progressive resistance exercise. training twice a week for 10 weeks at a community gymnasium located close to where each adolescent with The research questions were: Down syndrome lived. A 10-week program was selected as 1. Does a progressive resistance training program lead it fits in with the typical school term and therefore could be to increased muscle strength in adolescents with timetabled around the weekly schedule of the families of the Down syndrome? adolescents. The training program (including the duration 2. Does it lead to improved physical function in these and frequency of the program) was designed according to adolescents? the recommendations of the American College of Sports Medicine (American College of Sports Medicine 2009). Method The participants performed six exercises using weight machines; three for the upper limbs (lat pull-down, seated Design chest press, seated row) and three for the lower limbs (seated leg press, knee extension, calf raise). These exercises were We conducted a randomised controlled trial. Adolescents chosen because they would strengthen the major multi-joint with Down syndrome were recruited for the trial through a muscles of the upper and lower limbs. The exercises were community support group for people with Down syndrome conducted on pin-loaded weight machines as they were and their families. A flyer promoting the trial was mailed to considered safer for novice participants than free weights as members as part of the support group’s usual mail out and there was less chance of a weight being dropped on a body families were asked to contact the researchers if interested. part and causing injury. These exercises could be modified to suit the needs of the individual, or the availability of Participants were randomly allocated to the experimental training equipment at a particular gymnasium. All but or control group using a concealed method. Participants very minor modifications were completed by the student were randomised in blocks of four, generated from mentors in conjunction with the researchers. For example, a random numbers table with assignments sealed in if a participant found it difficult to do the standing calf sequentially numbered, opaque envelopes. Assignment was raise exercise, the exercise could be modified to a seated made after the recruiter had determined eligibility for the calf raise exercise. Participants performed up to 3 sets of 12 study and their parents had consented to the adolescent’s repetitions of each exercise, or until fatigue. A 2-minute rest participation. Group allocation was prepared and performed was taken between each set to allow for recovery, and the by a researcher not involved in recruitment or assessment resistance was increased when 3 sets of 12 repetitions of an by opening the next envelope in the sequence. exercise could be completed (American College of Sports Medicine 2009). The experimental group received 10 weeks of progressive resistance training and the control group continued The progressive resistance training program was led by with their usual activities. All participants completed student mentors recruited from the physiotherapy student assessments of muscle strength and upper and lower limb body at the university. Provision was made for the students physical function at baseline (week 0) and immediately to include the training experience as part of their clinical after the intervention phase of the study (week 11). The experience portfolio. To ensure consistency, the student assessments were completed by an assessor who was blind mentors received training on the program content, the to group allocation and who was not involved in any other exercise equipment, program progression, and motivational aspect of the trial. strategies. Each student mentor was contacted by a researcher every three weeks during training to monitor Participants progress and help solve any problems. The adolescents with Down syndrome were matched with a student mentor Participants were included if they were aged 13–18 years, were able to follow simple verbal instructions in English, 188 Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010

Shields and Taylor: Resistance exercise in Down syndrome based on the metropolitan suburb where they lived and, 5BCMF
. Baseline characteristics of participants. in some cases where parents requested this, based on gender. The student mentors also completed a progressive Characteristic Randomised resistance training program, completing their exercise set (n = 23) while the adolescent with Down syndrome was taking their 2-minute rest between sets. The mentors were responsible Exp Con for completing a log book for the adolescent with Down (n = 11) (n = 12) syndrome detailing each exercise performed, the weight lifted, the number of repetitions, and number of sets. Age (yr), mean (SD) 15.9 (1.5) 15.3 (1.7) Gender, n males (%) 8 (73) 9 (75) The control group participants continued with their usual Height (cm), mean (SD) 159 (11) 156 (7) activities, which may have included leisure and sporting Weight (kg), mean (SD) 63 (6) 58 (7) activities but did not include a progressive resistance training BMI program. After the trial was completed, these participants Level of perceived ID, n (%) 25.5 (4.4) 24.0 (3.2) were invited to complete the same program with a student mentor, but no further assessments were conducted. Mild 1 (9) 5 (42) Moderate 10 (91) 5 (42) Outcome measures Severe 0 (0) 2 (16) School or program, n (%) Primary outcome: Muscle strength was assessed using 1 Mainstream 0 (0) 3 (25) repetition maximum (1RM) force generation tests. These Specialist 11 (100) 9 (75) tests established the amount of weight each participant could lift in a single seated chest press and seated leg press Exp = experimental group, Con = control group, ID = intellectual respectively. Single 1RM chest press and leg press tests have disability, BMI = body mass index high levels of retest reliability (r ® 0.89) and demonstrated no systematic change when measured over 3 weeks in adults log book. At the beginning and end of each session the with neurologic impairment (Taylor et al 2004). Single 1RM student mentor asked the participant if they had experienced chest press and leg press tests were used as representative any injuries or other problems. measures of upper and lower limb strength, respectively, as they involve the major muscle groups exercising over Data analysis multiple joints. Intention to treat analysis was performed and outcomes were Secondary outcome: Lower-limb physical function was analysed using ANCOVA with the baseline measure of each measured using the Timed Up and Down Stairs test ( aino variable used as the covariate (Vickers 2005). Where data et al 2004). This test was chosen because it is a challenging were missing, the carry-forward technique was used, which test of mobility that would be expected to be related to assumes that missing data remained constant (Hollis and an improved ability to generate muscle force. It has also Campbell 1999). The mean difference within each group been implemented previously as an outcome measure in a and between the groups and their 95% CI were calculated. population of people with Down syndrome (Shields et al Standardised mean differences (SMD) (otherwise known as 2008). Participants were asked to ascend, turn, and descend effect sizes) were also calculated. SMDs were calculated a flight of stairs as quickly as possible. They could choose by subtracting the mean of the control group from the any method of traversing the stairs including alternating mean of the experimental group and dividing by the pooled steps, running up the stairs, or using handrails for support. standard deviation. The SMDs were interpreted as follows: The time taken to complete the task was recorded in seconds less than 0.2 was considered small, between 0.2 and 0.5 was using a stopwatch. The test was repeated twice and the considered moderate, and greater than 0.8 was considered fastest time was used in the analysis. Secondary analysis of large (Cohen 1977). data from our laboratory has demonstrated moderate retest reliability of the Timed Up and Down Stairs test in adults Results with Down syndrome (ICC3,1 = 0.74). 'MPX
PG
UIF
QBSUJDJQBOUT
UISPVHI
UIF
USJBM Upper-limb physical function was measured using the Grocery Shelving Task (Hill et al 2004). Participants started Twenty-three adolescents (17 boys, 6 girls) with Down from a seated position 2m from a bench. They were asked syndrome participated in the trial (Table 1). The participants to stand up and carry 2 grocery bags, each containing 10 had a mean age of 15.6 years (SD 1.6) and a mean body mass items weighing 410 g (total weight of each bag was 4.1 kg), index of 24.7 kg/m2 (SD 3.8, range 19.8 to 35.0). Eleven to the bench. The participants then took the items out of the participants were randomly allocated to the experimental bag and stacked them onto a shelf at shoulder height. The group and 12 participants to the control group. There were participants completed the task as fast as possible and the no apparent differences at baseline between the groups time taken was recorded. Participants were given a practice for most of the demographic factors or outcome measures trial before they completed two timed tests, the average (Tables 1 and 2). However, the proportion of adolescents of which was used in the analysis. Secondary analysis of with moderate/severe intellectual disability appeared to data from our laboratory has demonstrated moderate retest be greater in the experimental group compared with the reliability of the Grocery Shelving Task in adults with control group. Down syndrome (ICC3,1 = 0.76). $PNQMJBODF
XJUI
UIF
USJBM
NFUIPE Any adverse events that occurred during training (including minor events such as delayed onset muscle soreness) were Participants attended 90% (198/220) of the scheduled recorded by the student mentor in the participant’s exercise training sessions. No serious adverse events were recorded. Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010 189

Research Patients assessed for eligibility (n = 29) Excluded (n = 6) š location (n = 6) Measured limb strength, Timed Up and Down Stairs test, and Grocery Shelving Task Week 0 Randomised (n = 23) (n = 11) (n = 12) Lost to follow-up Experimental group Control group Lost to follow-up (n = 0) š usual daily activity (n = 1) š usual daily activity š progressive resistance jhW_d_d]\"(%ma\\eh'&ma Week 10 Measured limb strength, Timed Up and Down Stairs test, and Grocery Shelving Task (n = 11) (n = 11) Analysed (n = 11) (n = 12)* * The 12th participant in the control group was included in the analysis by the carry-forward approach 'JHVSF
. Design and flow of participants through the trial. Missed sessions were due to illness or vacation time. None timed stairs test was 0.5, and for the Grocery Shelving Task of the sessions was missed due to soreness, injury, or was 0.3, which represented moderate effects. illness as a result of the training program. Four participants complained of mild muscle soreness during training, No major adverse events were reported. Although five mostly during the early weeks of the program and all participants complained of muscle soreness during the recovered spontaneously. Three participants complained of initial weeks of training, this did not preclude them from sore hands as a result of using the weight equipment; one training. The reported symptoms were mild and were to be participant resolved this by wearing gloves during training. expected in a group of novice trainees completing moderate Over the course of the training program, the experimental to high intensity training. group progressed the amount of resistance lifted for each of the prescribed exercises by at least 95% of the initial Discussion training resistance. One participant in the control group was unavailable for reassessment but this participant was Several of the study’s findings indicate that progressive included in the intention to treat analysis via the carry- resistance training was feasible and safe for adolescents forward approach (Fig. 1). with Down syndrome when facilitated by a student mentor. Adherence to the program was excellent, adverse events Effect of the intervention were minimal, the reasons for missed sessions were unrelated to the intervention, and the only participant lost The average baseline 1RM for leg press was 88 kg, to follow-up was allocated to the control group. These data approximately 15% less than values for adolescents with suggest progressive resistance training was an acceptable typical development (Christou et al 2006). The experimental form of exercise to the participants, a finding consistent group increased lower limb muscle strength compared to the with previous literature concluding that this type of training control group (MD 36 kg, 95% CI 15 to 58, SMD 0.7). The is safe for people with a range of health conditions and lower limb strength increase represented a 42% increase disabilities (Taylor et al 2005). This is an important finding, in baseline strength in the experimental group compared as some people with intellectual disability and their carers to the control group. There were no significant differences are apprehensive about taking part in exercise and believe between the groups for upper limb muscle strength or upper they should not engage in exercise (Heller et al 2004). Our and lower limb physical function. Group data are presented results and future studies should alleviate this concern and in Table 2 and individual data in Table 3 (see eAddenda for may encourage people with Down syndrome to become Table 3). The SMD for the 1RM chest press was 0.6, for the more active. Given that people with Down syndrome are 190 Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010

Shields and Taylor: Resistance exercise in Down syndrome 5BCMF
. Mean (SD) score, mean (SD) difference within groups, and mean (95%CI) difference between groups for all outcomes for the experimental group and the control group. Outcome Score Difference within groups Difference between groups Chest press 1RM (kg) Week 10 minus Week 0 Baseline Post-intervention Week 10 minus Week 0* (week 0) (week 10) Exp Con (95 % CI) 11 5 Exp Con Exp Con (15) (8) Exp-Con 8 44 40 55 44 (18) (9) (24) (12) (–3 to 17) Leg press 1RM 87 89 132 97 45 8 36 (kg) (46) (44) (50) (43) (31) (17) (15 to 58) Timed Up and Down 18 20 15 22 –3 3 –6 (4) (17) Stairs test (sec) (8) (9) (6) (18) (–17 to 5) –6 –1 Grocery Shelving 80 88 74 87 (17) (25) –6 Task (sec) (38) (38) (31) (48) (–25 to 13) * = derived from ANCOVA with dependent variable at baseline as covariate. Con = control group, Exp = experimental group at risk of the health consequences of inactivity (Hill et An innovative aspect of this trial was that the progressive al 2003), it is necessary that we identify feasible exercise resistance training intervention was led by physiotherapy options for this group. These results suggest that progressive student-mentors. This feature provided the supervision and resistance training can be a safe, socially desirable, and the social interaction needed to encourage the adolescents feasible exercise and recreation option for adolescents with to exercise. Choosing physiotherapy students to act as Down syndrome. mentors was advantageous as they had an understanding of the principles of exercise training, and were also close Our data show that progressive resistance training was in age to the adolescents so that the social interaction effective in improving the strength of the major antigravity between the pair was meaningful. An additional benefit muscles of the lower limb (quadriceps and hip extensors) in was that the physiotherapy students had the opportunity to adolescents with Down syndrome. The average percentage gain a unique experience of disability, something that they increase in muscle strength was 42%, which was clinically may not necessarily have gained from their professional worthwhile and was similar to increases of 27–46% reported training due to a lack of appropriate clinical placements. in other populations (O’Shea et al 2007, Dodd et al 2004). Progressive resistance training is a program typical of Although it cannot be concluded with 95% confidence that those that members of the community might undertake there was a change in upper limb strength, the SMD was if they attended a community gym. The model developed similar in magnitude to what was observed for changes and implemented in this study has the potential to become in lower limb muscle strength. These findings are notable part of the on-going clinical experience of physiotherapy considering the relatively short duration of the program and students and therefore could be an avenue for the long term the fact that the majority of the participants had moderate sustainability of this type of community-based exercise to severe intellectual disability. An increased ability program. It could also provide on-going opportunities to generate force in the major muscles of the lower limb for people with Down syndrome and those with other may be important for adolescents with Down syndrome, disabilities who require a high level of support to exercise. whose vocational roles may be influenced by their physical It is anticipated that, like with all novices, after a period of capacity. supervised exercise it may be possible for adolescents with Down syndrome to continue with the program with a lesser Although no corresponding changes in physical function degree of supervision such as with a family member. were found, the observed SMDs for these variables (0.3 for the Grocery Shelving task and 0.5 for the timed stairs test) The main strength of the trial is that it was the first indicated a moderate observed effect size. Effect sizes of randomised controlled trial that assessed the effects of a this magnitude are encouraging and are similar to those progressive resistance training program among adolescents reported among adults with Down syndrome (Shields et with Down syndrome. Of the previous four studies al 2008). If these SMD results were confirmed on a larger published, three included adults with Down syndrome sample, then it is possible progressive resistance training (Davis and Sinning 1987, Rimmer et al 2004, Shields et might have clinically significant effects on the physical al 2008), and the other was a non-controlled trial of 14 functioning of adolescents with Down syndrome. The adolescents with Down syndrome (Weber and French 1988). SMDs for the physical functional measures were smaller An important aspect of the program was that it took place than for the muscle strength measures. This is expected as in an inclusive setting (a community gymnasium). This is muscle strength is only one component required for these noteworthy as adolescents with Down syndrome often have functional tasks; that is, there was less specificity of training restricted opportunities to participate in exercise programs for these functional tasks. Consistent with this, there are taking place in an integrated community setting (Menear some data in people with Down syndrome to suggest that 2007). muscle strength is an important but not the only variable important in completing functional tasks (Cowley et al While the trial was powered to detect changes in lower limb 2010). muscle strength, a limitation was the relatively small sample Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010 191

Research size, which required the effects of the intervention to be a future randomised trial to ascertain whether progressive large in order to detect any changes in task-related activities. resistance training carries over into an improved ability for However, the 95% CIs around the estimates of the effects on adolescents with Down syndrome to complete daily tasks task-related outcomes include clinically worthwhile effects. and physical activities. Q Therefore, the trial provides important pilot data for the conduct of a randomised trial to define more precisely the eAddenda: Table 3 available at www.jop.physiotherapy.asn. effect of the training on task-related outcomes. au Other factors in the design of the intervention that could Ethics: The trial received ethics approval from the La Trobe be considered are the duration and frequency of the University Human Ethics Committee (08-024). Written program. Given its relatively short duration, it is possible informed consent to the research was obtained from the that a larger effect might be obtained from continuing the parents of all participants. program for longer. A study on people with intellectual disability reported greater gains in muscle strength from Support: Windermere Foundation. programs of longer duration and frequency (Suomi 1998). However, the 10-week program, had the advantage of Acknowledgements: The authors acknowledge the fitting in with the typical school term and therefore could contributions of all the participants and their families. be timetabled around the weekly schedule of the families of the adolescents. Increasing the program frequency from Competing interests: None declared. twice to three times a week might change the outcome, as a previous study including adults with Down syndrome Correspondence: Dr Nora Shields, School of Physiotherapy, completed training three times per week and reported larger La Trobe University, Australia. Email: N.Shields latrobe. positive effects (Davis and Sinning 1987). However, it is edu.au not known what effect this change would have on program adherence in adolescents with Down syndrome. References There appeared to be a greater number of participants with American Association on Intellectual and Developmental moderate intellectual disability in the experimental group. Disabilities (2010) Intellectual disability: definition, It is possible that adolescents with moderate intellectual classification, and systems of supports (11th edn). disability might find it more difficult to follow instructions Washington, DC: American Association on Intellectual and and learn the exercises than adolescents with a mild Developmental Disabilities. intellectual disability, which could limit the benefit they obtain from the program. However, there was a very high American College of Sports Medicine (2009) Progression adherence rate in participation in the intervention program models in resistance training for healthy adults. Medicine & by participants with moderate intellectual disability Science in Sports & Exercise 41: 687–708. suggesting the intervention was well accepted and feasible. Baynard T, Pitetti K, Guerra M, Unnithan VB, Fernhall B (2008) A limitation of the study is that there was no follow-up as Age-related changes in aerobic capacity in individuals with to whether the effects of the intervention were maintained mental retardation: a 20–year review. Medicine & Science in and whether there were any longer term outcomes from Sports & Exercise 40: 1984-1989. engaging in regular progressive resistance training. Further studies are also necessary to help determine the long term Carmeli E, Kessel S, Coleman R, Ayalon M (2002) Effects sustainability of the program given the level of support of a treadmill walking program on muscle strength and that adolescents with Down syndrome need to begin to balance in elderly people with Down syndrome. Journals exercise. An additional outstanding issue that should also of Gerontology–Series A Biological Sciences and Medical be addressed in future studies is whether progressive Sciences 57: M106–110. resistance training alone can change physical activity levels. Christou M, Smilios I, Sotiropoulos K, Volak K, Piliandis T, Progressive resistance training is one possible exercise and Tokmakidis SP (2006) Effects of resistance training on the recreation option for adolescents with Down syndrome. physical capacities of adolescent soccer players. Journal of Previous studies have investigated the effectiveness of other Strength and Conditioning Research 20: 783–791. exercise options in this population such as aerobic training and circuit training (Khalili and Elkins 2009, Millar et al Cohen J (1977) Statistical Power Analysis for Behavioural 1993, Weber and French 1988). The predominance of males Sciences. New York: Academic Press. who volunteered to participate in the current study might suggest that it is more socially desirable for males to take Cowley P, Ploutz-Snyder L, Baynard T, Heffernan K, Young part in progressive resistance training. The prevalence of Jae S, Hsu S, et al (2010) Physical fitness predicts functional Down syndrome is approximately 10% higher among males tasks in individuals with Down syndrome. Medicine & than females (Shin et al 2009), so more males self-selected Science in Sports & Exercise 42: 388–393. into this study than would be expected on the basis of population distribution alone. Croce RV, Pitetti KH, Horvat M, Miller J (1996) Peak torque, Wl[hW][ fem[h\" WdZ ^Wcijh_d]i%gkWZh_Y[fi hWj_ei _d In conclusion, progressive resistance training led by nondisabled adults and adults with mental retardation. physiotherapy student mentors and performed in a Archives of Physical Medicine & Rehabilitation 77: 369–372. community gymnasium is a feasible, socially desirable, and safe exercise option for adolescents with Down syndrome Davis WE, Sinning WE (1987) Muscle stiffness in Down that can lead to improvements in lower-limb muscle syndrome and other mentally handicapped subjects: a performance. This trial provides important data that justify research note. Journal of Motor Behavior 19: 130–144. Heller T, Hsieh K, Rimmer J (2002) Barriers and supports for exercise participation among adults with Down syndrome. Journal of Gerontological Social Work 38: 161–178. Heller T, Hsieh K, Rimmer JH (2004) Attitudinal and psychosocial outcomes of a fitness and health education program on adults with Down syndrome. American Journal of Mental Retardation 109: 175–185. Hermon C, Alberman E, Beral V, Swerdlow A (2001) Mortality 192 Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010

and cancer incidence in persons with Down’s syndrome, Shields and Taylor: Resistance exercise in Down syndrome their parents and siblings. Annals of Human Genetics 65: 167–176. Shields N, Taylor NF, Dodd KJ (2008) Effects of a community- based progressive resistance training program on muscle Hill C, Denehy L, McDonald C (2004) Reproducibility, validity performance and physical function in adults with Down and responsiveness of a grocery shelving task: a measure syndrome: a randomized controlled trial. Archives of Physical of upper limb function for patients with chronic obstructive Medicine & Rehabilitation 89: 1215–1220. pulmonary disease. Respirology 9 (Supplement): A45. Shin M, Besser LM, Kucik JE, Lu C, Siffel C, Correa A, Hill D, Gridley G, Cnattingius S, Mellemkjaer L, Linet M, and the Congenital Anomaly Multistate Prevalence and Adami H, et al (2003) Mortality and cancer incidence Survival (CAMPS) Collaborative (2009) Prevalence of Down among individuals with Down Syndrome. Archives of Internal syndrome among children and adolescents in 10 regions of Medicine 163: 705–711. the United States. Pediatrics 124: 1565–1571. Hollis S, Campbell F (1999) What is meant by intention to treat Suomi R (1998) Self-directed strength training: its effect on leg analysis? Surveys of published randomized controlled trials. strength in men with mental retardation. Archives of Physical BMJ 319: 670–674. Medicine & Rehabilitation 79: 323–328. Khalili MA, Elkins MR (2009) Aerobic exercise improves lung Taylor N, Dodd K, Damiano D (2005) Progressive resistance function in children with intellectual disability: a randomised exercise in physical therapy: a summary of systematic trial. Australian Journal of Physiotherapy 55: 171–175 reviews. Physical Therapy 85: 1208–1223. Menear KS (2007) Parents’ perceptions of health and physical Taylor N, Dodd K, Larkin H (2004) Adults with cerebral palsy activity needs of children with Down syndrome. Down benefit from participating in a strength training programme Syndrome: Research & Practice 12: 60–68. at a community gymnasium. Disability & Rehabilitation 26: 1128–1134. Millar AL, Fernhall B, Burkett LN (1993) Effects of aerobic training in adolescents with Down syndrome. Medicine & Telama R, Yang. X, Viikari J, Välimäki I, Wanne O, Raitakari O Science in Sports & Exercise 25: 270–274. (2005) Physical activity from childhood to adulthood: a 21- year tracking study. American Journal of Preventive Medicine Pitetti KH, Climstein M, Mays MJ, Barrett PJ (1992) Isokinetic 28: 267–273. arm and leg strength of adults with Down syndrome: a comparative study. Archives of Physical Medicine & Vickers A (2005) Parametric versus non-parametric statistics in Rehabilitation 73: 847–850. the analysis of randomized controlled trials with non-normally distributed data. BMC Medical Research Technology 5: 35. Rimmer JH, Heller T, Wang E, Valerio I (2004) Improvements in physical fitness in adults with Down syndrome. American Weber R, French R (1988) Down’s syndrome adolescents and Journal of Mental Retardation 109: 165–174. strength training. Clinical Kinesiology 42: 13–21. Shields N, Dodd K (2004) A systematic review on the effects Zaino CA, Marchese VG, Westcott SL (2004) Timed up and of exercise programmes designed to improve strength for down stairs test: preliminary reliability and validity of a new people with Down syndrome. Physical Therapy Reviews 9: measure of functional mobility. Pediatric Physical Therapy 109–115. 16: 90–98. Shields N, Dodd K, Abblitt C (2009) Children with Down Zetts R, Horvat M, Langone J (1995) Effects of a community- syndrome do not perform sufficient physical activity to based progressive resistance training program on the maintain good health or optimize cardiovascular fitness. work productivity of adolescents with moderate to severe Adapted Physical Activity Quarterly 26: 307–320. intellectual disabilities. Education & Training in Mental Retardation 30: 166–178. Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010 193

Conclusion Editorial Not surprisingly, health literacy is starting to be addressed Chew LD, et al (2004) Am J Surg 188: 250–253. at both health policy and program levels in Australia. Both DeWalt DA (2007) N C Med J 68: 327–330. the Health and Hospitals Reform Commission Report DeWalt DA, et al (2004) J Gen Intern Med 19: 1228–1239. and the National Primary Health Care Strategy outline Ferreira M, et al (2005) J Clin Oncol 23: 1548–1554. key initiatives relating to health literacy. These include Jordan JE, et al (2010a) Patient Educ Couns 79: 36–42. health professionals supporting patients to improve their Jordan JE, et al (2010b) J Clin Epi (in press). health literacy skills to navigate the health system, engage Lerner EB, et al (2000) Am J Emerg Med 18: 764–766. in preventive activities, enhance self-management, and Liddle SD, et al (2009) Man Ther 14: 189–196. change risky lifestyle behaviours. Similar policy and Linton SJ, et al (2002) J Occup Rehab 12: 223–232. program initiatives are also in development by state Paasche-Orlow MK, et al (2005) J Gen Intern Med 20: 175–184. governments. For physiotherapists, who are recognised Schillinger D, et al (2002) JAMA 288: 475–482. primary care clinicians and spend considerable time with Schillinger D, et al (2003) Arch Intern Med 163: 83–90. patients delivering health information, particularly in the Scott TL, et al (2002) Med Care 40: 395–404. context of chronic condition management and post-surgical USA Department of Health and Human Services (2000) rehabilitation, these policy initiatives imply that clinicians will increasingly need to be familiar with health literacy Healthy People 2010: Understanding and improving health. concepts, measurement, and interventions to assist patients US Government Printing Office. Washington, DC. in seeking, understanding and utilising health information. USA National Academy on an Aging Society (1999) Low health literacy skills increase annual health care expenditures References by $73 billion. The Center for Health Care Strategies and National Academy on an Aging Society. Washington, DC. Baker DW, et al (1998) J Gen Intern Med 13: 791–798. Weinman J, et al (2009) BMC Family Practice 10: 43. Williams CM, et al (2010) Arch Intern Med 170: 271–277. Bishop A, et al (2008) Pain 135: 187–195. Briggs AM, et al (2010) Pain 150: 275–283. Erratum In Vol 55 No 3 there was an error in the results reported in predicted by revision hip arthroplasty. The regression the paper by Stevens et al (2009). The error occurred in the coefficient for being in the revision group was –1153.7 (95% final page make up. The last two paragraphs of Column 1 CI –2241.1 to –66.3). The regression coefficient for being in p. 188 should be corrected as follows (corrected text in bold the revision group of –912.8 (95% CI –1989.1 to 163.6) was type): no longer significant when age, gender, and Charnley group were added to the prediction equation, suggesting that these Linear regression analysis was also performed to determine additional predictors did confound the relation between whether total amount of physical activity was predicted by group and total intensity of physical activity (Box 3). revision hip arthroplasty. The regression coefficient for Revision group, age, gender, and Charnley group accounted being in the revision group was –394.3 (95% CI –701.1 to for 9% of the variance in total intensity of physical activity.’ –87.5). The regression coefficient for being in the revision group of –121.2 (95% CI –408.0 to –165.7) was no longer AJP apologises to the authors and to our readers. significant when age, gender, and Charnley group were added to the prediction equation, suggesting that these Reference additional predictors did confound the relation between group and total amount of physical activity (Box 2). Stevens M, Hoekstra T, Wagenmakers R, Bulstra SK, van Revision group, age, gender, and Charnley group accounted den Akker-Scheek I (2009) People who undergo revision for 18% of the variance in total amount of physical activity. arthroplasty report more limitations but no decrease in physical activity compared with primary total hip arthroplasty: Finally, linear regression analysis was performed to an observational study. Australian Journal of Physiotherapy determine whether total intensity of physical activity was 55: 185–189. Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010 151

Appraisal Clinimetrics Impact of Event Scale-Revised Description special training required to administer the questionnaire. The client is asked to report the degree of distress The Impact of Event Scale-Revised (IES-R) is a self-report experienced for each item in the past 7 days. The 5 points measure of current subjective distress in response to a on the scale are: 0 (not at all), 1 (a little bit), 2 (moderately), specific traumatic event (Weiss and Marmar 1997). The 22- 3 (quite a bit), 4 = (extremely). The sum of the means of item scale is comprised of 3 subscales representative of the each subscale instead of raw sums is recommended (Weiss major symptom clusters of post-traumatic stress: intrusion, and Marmar 1997). Thus, the scores for each subscale avoidance, and hyperarousal (American Psychiatric range from 0 to 4 and the maximum overall score possible Association 1994). The intrusion subscale includes 8 items is 12. There are no specific cut-off scores for the IES-R related to intrusive thoughts, nightmares, intrusive feelings, although higher scores are representative of greater distress. and imagery associated with the traumatic event. The Increased overall scores on all subscales may indicate the avoidance subscale includes 8 items related to avoidance of need for further evaluation. feelings, situations, and ideas. The hyperarousal subscale includes 6 items related to difficulty concentrating, anger Reliability, validity and sensitivity to change: Test-retest and irritability, psychophysiological arousal upon exposure reliability (r = –0.89 to 0.94) and internal consistency to reminders and hypervigilance. (Chronbach’s α) for each subscale (intrusion = 0.87 to 0.94, avoidance = 0.84 to 0.97, hyperarousal = 0.79 to 0.91) are The IES-R is a revised version of the Impact of Event Scale acceptable (Creamer et al 2003). IES-R scale scores have (Horowitz 1979) and was developed as the original version also been found to have moderate to strong correlations with did not include a hyperarousal subscale. IES-R responses one another (r = 0.52 to 0.87) (Beck et al 2008). Correlations were also modified so the client was requested to report have been found to be high between those of the IES-R and on the degree of distress rather than the frequency of the the original IES for the intrusion (r = 0.86) and avoidance (r symptoms. = 0.66) subscales which supports the concurrent validity of both measures (Beck et al 2008). Instructions to the client and scoring: The IES-R takes approximately 10 minutes to complete and score with no the IES (scores of 26 or more out of a possible 75). There has been several cut-off values suggested for a probable Commentary diagnosis of PTSD ranging from 22 to 24 in individuals with substance use disorders (Rash et al 2008) to 33 from The indications for using the IES-R remain largely similar a possible 88 in Vietnam veterans (Creamer et al 2003). to those of the original IES. The IES has been recommended However, these cut-off values have been based on specific for use as a measure of subjective distress in clinical population groups and also relate to the raw sum of scores. guidelines such as the NSW Government Guidelines for As both measures were intended to provide an indication the Management of Acute Whiplash). Similar to the IES, of a general level of distress related to an event and not to the IES-R is a valid measure of post-traumatic stress diagnose PTSD, cut-off points seem inappropriate. It would symptoms and is useful to monitor symptoms as well as seem unlikely the decision to provide psychological referral to track progress with interventions. When compared to would be based on the IES-R or IES alone and rather the the original version, the key strength of the IES-R is that IES-R is a tool which may aid the clinical reasoning process. it correlates better with DSM-IV criteria for PTSD through the inclusion of the hyperarousal subscale (American Helena Motlagh Psychiatric Association 1994). The University of Queensland, Australia Physiotherapists are commonly involved in the care of References individuals following a traumatic event such as a motor vehicle accident. In this area, it has been recommended that American Psychiatric Association (1994) Diagnostic and all three symptom clusters be considered (Buitenhuis et al statistical manual of mental disorders, 4th ed. Washington. 2006). Further, there is evidence suggesting a relationship between increased hyperarousal symptoms with persistent Beck et al (2008) J Anxiety Disorders 22: 187–198. pain and disability in chronic whiplash (Sterling et al 2003). Buitenhuis et al (2006) J Psychosom Res 61: 681–689. There has been some evidence to suggest the IES-R can discriminate between individuals with and without post- Creamer et al (2003) Behav Res Ther 41: 1489–1496. traumatic stress disorder (PTSD) (Beck et al 2008). However, there is insufficient evidence to support the Horowitz et al (1979) Psychosom Med 41: 209-218. IES-R as a diagnostic tool as well as conflicting evidence regarding its use as a screening tool for PTSD (Creamer Rash et al (2008) Addict Behav 33: 1039–1047. et al 2003, Beck et al 2008). As with the original IES, a diagnosis of PTSD cannot be made on the IES-R alone and Weiss D, Marmar C (1997) In: Wilson P, Keane T (eds) alternative measures should be considered if this condition Assessing psychological trauma and post traumatic stress is suspected (Weiss and Marmar 1997, Beck et al 2008). disorder: a handbook for practitioners. New York, Guilford Press. pp 399–411. Unfortunately, the IES-R does not have established cut-off points to suggest grounds for psychological referral as does Website ^jjf0%%mmm$cWW$dim$]el$Wk Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010 203

Cader et al: Inspiratory muscle training and weaning time Inspiratory muscle training improves maximal inspiratory pressure and may assist weaning in older intubated patients: a randomised trial Samária Ali Cader1,4, Rodrigo Gomes de Souza Vale1,4, Juracy Correa Castro1,2, Silvia Corrêa Bacelar3, Cintia Biehl4, Maria Celeste Vega Gomes4, Walter Eduardo Cabrera4 and Estélio Henrique Martin Dantas1,4 1University Federal do Estada do Rio de Janiero – LABIMH-UNIRIO, Brazil, 2Universidade de Trás-os-Montes e Alto Douro, Portugal, 3Grande Rio University – UNIGRANRIO, Brazil, 4Catholic University of Nuestra Senora de Asunción, Paraguay Questions: Does inspiratory muscle training improve maximal inspiratory pressure in intubated older people? Does it improve breathing pattern and time to wean from mechanical ventilation? %FTJHO Randomised trial with concealed allocation and intention-to-treat analysis. Participants: 41 elderly, intubated adults who had been mechanically ventilated for at least 48 hr in an intensive care unit. Intervention: The experimental group received usual care plus inspiratory muscle training using a threshold device, with an initial load of 30% of their maximal inspiratory pressure, increased by 10% (absolute) daily. Training was administered for 5 min, twice a day, 7 days a week from the commencement of weaning until extubation. The control group received usual care only. Outcome measures: The primary outcome was the change in maximal inspiratory pressure during the weaning period. Secondary outcomes were the weaning time (ie, from commencement of pressure support ventilation to successful extubation), and the index of Tobin (ie, respiratory rate divided by tidal volume during a 1-min spontaneous breathing trial). Results: Maximal inspiratory pressure increased significantly more in the experimental group than in the control group (MD 7.6 cmH20, 95% CI 5.8 to 9.4). The index of Tobin decreased significantly more in the experimental group than in the Yedjheb]hekfC:.$)Xh%c_d%B\"/+9?($/je')$-$?dj^ei[m^eZ_ZdejZ_[ehh[Y[_l[WjhWY^[eijeco\"j_c[jem[Wd_d]mWi significantly shorter in the experimental group than in the control group (MD 1.7 days, 95% CI 0.4 to 3.0). Conclusions: In intubated older people, inspiratory muscle training improves maximal inspiratory pressure and the index of Tobin, with a reduced weaning time in some patients. Trial registration: NCT00922493. [Cader SA, Vale RGS, Castro JC, Bacelar SC, Biehl C, (PNFT
.$7
$BCSFSB
8&
%BOUBT
&).

*OTQJSBUPSZ
NVTDMF
USBJOJOH
JNQSPWFT
NBYJNBM
JOTQJSBUPSZ
QSFTTVSF
BOE
NBZ
BTTJTU
XFBOJOH
JO
PMEFS
JOUVCBUFE
QBUJFOUT
B
SBOEPNJTFE
USJBM
Journal of Physiotherapy 
o> ,FZ
XPSET: Respiration, artificial; Weaning; Aged; Intensive care; Inspiratory muscle training; Physiotherapy Introduction Data from several other studies support the hypothesis that inspiratory muscle training improves inspiratory muscle The primary reason for admission to an intensive care strength in patients who are weaning from ventilatory unit is the need for mechanical ventilation (Tobin 2001). support and therefore results in an improvement in the Weaning from mechanical ventilation often accounts for likelihood of success of extubation (Chang et al 2005a, a large proportion of the total time spent on the ventilator Epstein et al 2002). However, most of the clinical studies that (Esteban et al 1994) and respiratory muscle weakness is a have examined the efficacy of inspiratory muscle training major determinant of failure to wean (Ambrosino 2005). in the intensive care setting have been performed with Failure to wean increases the risk of ventilator-associated tracheostomised participants (Aldrich et al 1989, Chang et al pneumonia and further respiratory muscle deconditioning 2005b, Martin et al 2002, Sprague and Hopkins 2003). One (Epstein 2006). With ageing, lung elastic recoil, chest wall study with intubated patients (Caruso et al 2005) delivered compliance, and respiratory muscle strength all decrease, the inspiratory muscle training intervention primarily while with resultant changes in static lung volumes and regional patients were still receiving controlled ventilation. The ventilation (Kim and Sapienza 2005, Krieg et al 2007). controlled ventilation was continued until approximately Therefore interventions to improve the success of weaning, one day before extubation. In our experience, however, a especially those targeting respiratory muscle strength, may longer weaning period’ (ie, spontaneously initiated breaths be particularly important in the older population. with pressure support only) is used before extubation. We are unaware of any clinical studies of inspiratory muscle Inspiratory muscle strength and the index of Tobin are training in critically ill, intubated patients during the recognised as predictors of the success of weaning patients weaning period. Therefore, the research questions were: from mechanical ventilation (Meade et al 2001). Maximal inspiratory pressure is used widely as a test of inspiratory 1. Does inspiratory muscle training during the weaning muscle strength (Green et al 2002). The index of Tobin is period improve maximal inspiratory pressure in the ratio of respiratory frequency to tidal volume ( ang and intubated older patients? Tobin 1991); it therefore quantifies the degree to which the breathing pattern is fast and shallow. 2. Does it improve the index of Tobin and the time to wean from mechanical ventilation? Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010 171

Research Screened for eligibility (n = 198) Excluded (n = 131) š Not meeting inclusion criteria (n = 128) š Refused to participate (n = 3) Eligible, awaiting start of weaning (n = 67) Excluded (n = 26) š tracheostomised (n = 20) š died (n = 5) š transferred out of the hospital (n = 1) Start of Measured maximal inspiratory pressure and index of Tobin weaning Randomised (n = 41) Lost to follow-up (n = 21) (n = 20) š died (n = 4) š tracheostomised Experimental group Control group Lost to follow-up š usual care (n = 3) š inspiratory muscle š died (n = 4) training 5 min twice daily š tracheostomised š usual care (n = 2) Measured maximal inspiratory pressure and index of Tobin Extubation (n = 14) (n = 14) 'JHVSF
 Design and flow of participants through the trial. Method Participants Design Patients were included in the study if they were aged at least 70 years, had undergone mechanical ventilation for at least A randomised trial was conducted between December 48 hours in a controlled mode (Chang et al 2005a), had been 2007 and November 2008. Participants were recruited intubated because of acute hypoxaemic (Type I) respiratory from the intensive care unit of one hospital in Brazil. failure, and were unable to generate greater inspiratory After undergoing confirmation of eligibility and baseline pressure than 20 cmH2O ( ang and Tobin 1991). Patients measurements, the participants were randomly allocated were excluded if they had a condition that could compromise into either an experimental group or a control group. The weaning, eg, cardiac arrhythmia, congestive heart failure or enrolling investigator contacted another investigator to unstable ischaemic cardiac disease, or that could prevent request an allocation for the participant from the concealed adequate performance of inspiratory muscle training, eg, list of random allocations that had been generated by neuropathy or myopathy. Patients were also excluded if drawing numbers from a bag. This investigator was not they had been tracheostomised before the commencement otherwise involved in the study. The experimental group of weaning, had a major neurological co-morbidity, were received usual care and also underwent inspiratory muscle morbidly obese, or were taking medication that could cause training twice daily throughout the weaning period. The a disorder of attention. control group received usual care only. The weaning period was defined as from the end of controlled ventilation (ie, the Intervention commencement of pressure support ventilation only) until extubation. Maximal inspiratory pressure and the index In the experimental group, inspiratory muscle training of Tobin were measured immediately before participants was commenced when the participant was changed from commenced pressure support ventilation, daily during the controlled to spontaneous (ie, pressure support) ventilation. weaning period, and immediately before extubation (Figure A threshold device was used because it provides resistance 1). to inspiration through the use of a flow-independent one-way 172 Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010

Cader et al: Inspiratory muscle training and weaning time valve, generating a linear pressure load. During expiration of 6 ml/kg was maintained, followed by use of a T-tube there is no resistance because the unidirectional valve for 30 minutes (Boles et al 2007). The extubation was opens, while during inspiration the valve closes, providing considered a failure if the patient returned to mechanical resistance to inspiration. The amount of resistance can ventilation within 48 h (Sprague and Hopkins 2003) or be adjusted by increasing the compression on a spring required a tracheostomy. mechanism in the device (Sprague and Hopkins 2003, Johnson et al 1996). Outcome measures At each training session, participants were positioned supine The primary outcome was maximal inspiratory pressure, with the backrest raised to 45 deg (Sprague and Hopkins measured using a vacuum manometer according to the 2003). The target regimen was to commence with a load method of Marini and colleagues (1986), which needs of 30% of the participant’s maximal inspiratory pressure little contribution from the patient. The manometer is (Chang et al 2005b), increasing daily by 10% (absolute), attached to the endotracheal tube via a connector with with training for five minutes (Cahalin et al 1997), twice an expiratory unidirectional valve, permitting expiration a day, seven days a week (Liaw et al 2000) throughout while inspiration is blocked. This causes the participant to the weaning period. Supplemental oxygen was provided make successive respiratory efforts as their lung volume as needed (Martin et al 2002). The training session was progressively approaches residual volume. Measurement of interrupted when the treating therapist observed any of the inspiratory pressures is maintained with the valve in situ following: respiratory rate greater than 35 breaths/min or for 25 seconds to obtain the best result (Caruso et al 1999). 50% higher than at the start of the session; oxyhaemoglobin Testing was performed once daily in both groups before saturation less than 90%; systolic pressure greater than 180 any inspiratory muscle training or other physiotherapy, with mmHg or less than 80 mmHg; heart rate more than 140 participants positioned supine with the backrest raised to 45 beats/min or 20% higher than at the start of the session; deg (Sprague and Hopkins 2003). paradoxical breathing; agitation; depression; haemoptysis; arrhythmia or sweating (Caruso et al 2005, Conti et al Secondary outcomes were the index of Tobin and 2004). When any of these signs occurred during a training weaning time. For the index of Tobin, the participant session, the load was maintained (ie, not increased by 10%) was disconnected from the ventilator and a ventilometer at the next session. measured the participant’s spontaneous ventilation for one minute ( ang and Tobin 1991). The index is calculated The control group did not undergo any training of the as the number of breaths per minute divided by the tidal respiratory muscles during the weaning period. Both volume in litres. Testing was performed once daily in both groups continued to receive all other usual care. This groups before any inspiratory muscle training or other included changes in ventilatory support settings (such as physiotherapy, with participants positioned supine with the positive end-expiratory pressure and supplemental oxygen) backrest raised to 45 deg (Sprague and Hopkins 2003). as needed by the patient, in accordance with arterial blood gas reports. Usual care also included regular physiotherapy Outcomes were measured or recorded by physiotherapists intervention including passive to active-assisted in the intensive care unit. Compliance with the training mobilisation of the limbs, chest compression with quick regimen was also noted daily. release at end-expiration, aspiration of the endotracheal tube, and positioning, with manual hyperinflation and Data analysis saline instillation where indicated (Blattner et al 2008, Lemes et al 2009). In the absence of an established minimum clinically important difference in maximal inspiratory pressure Decision to extubate: The decision to extubate was in this population, we nominated 10 cmH20. The best based on the presence of: improvement in the aetiology estimate of the standard deviation of maximal inspiratory that resulted in respiratory insufficiency; a cough reflex; pressure in a population of intubated elderly patients is 4.5 haemodynamic stability; normal body temperature; no cmH20 ( ang et al 1993). A total of 10 participants would vasoactive drugs (with the exception of Dopamine 5 mg/kg/ provide an 80% probability of detecting a difference of min); stable electrolytes (mainly calcium, magnesium, and 10 cmH20 in maximal inspiratory pressure at a two-sided phosphate); and normal radiological evaluation (without 5% significance level. We anticipated that a substantial pneumothorax, congestion, or major pleural effusion or proportion of these critically ill participants would die or atelectasis). In addition, the pH was required to be between receive a tracheostomy. We therefore increased the recruited 7.30 and 7.60; the partial pressure of carbon dioxide to be sample to 20 participants per group to allow for this. less than 60 mmHg; the fraction of inspired oxygen to be less than 40%; and the ratio of partial pressure of oxygen All participants with follow-up data were analysed to fraction of inspired oxygen to be at least 200. Also, the according to their group allocation, ie, using the intention- participant was required not to have paradoxical breathing, to-treat principle. Statistical significance was considered use of accessory musculature, a respiratory rate over 35 as p 0.05, therefore mean between-group differences br/min (or an increase of 50% compared with before the and 95% confidence intervals are presented for maximal training) and sweating (Martinez et al 2003). The decision inspiratory pressure, the index of Tobin, and weaning time. to extubate was also delayed until the patient could The Kappa test was used to evaluate the agreement between demonstrate maximal expiratory pressure of at least 20 the evaluators of maximal inspiratory pressure. Total cmH2O (Afessa et al 1999). The cut-off point for the index intubation time was analysed using a Kaplan-Meier curve. of Tobin to consider extubation was 100 br/min/L (Epstein In the event of death, tracheostomy, or self-extubation, and Ciubotaru 1996). participants were excluded from the independent t-tests of between-group differences and were treated as censored cases in the survival analysis. The protocol for extubation was to reduce the pressure support to 8 cmH2O ensuring that a minimum tidal volume Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010 173

Research 5BCMF
. Baseline characteristics of participants. Characteristic Randomised Lost to follow up Exp Con Exp Con (n = 21) (n = 20) (n = 7) (n = 6) Age (yr), mean (SD) 83 (3) 82 (7) 82 (6) 82 (5) Gender, n male (%) 9 (43) 10 (50) 3 (50) 3 (50) Weight (kg), mean (SD) 66 (5) 65 (6) 67 (3) 66 (2) Size OTT, n (%) 7.0 2 (10) 2 (10) 1 (14) 1 (17) 7.5 9 (43) 9 (45) 3 (43) 2 (33) 8.0 10 (48) 9 (45) 3 (43) 3 (50) APACHE II score, mean (SD) 20 (6) 20 (7) 19 (4) 20 (1) Causes of ARF, n (%) Postoperative 3 (14) 2 (10) 1 (14) 1 (17) Pneumonia 11 (52) 10 (50) 3 (43) 3 (50) Aspiration 5 (24) 5 (25) 2 (29) 1 (17) Trauma 1 (5) 2 (10) 1 (14) 1 (17) Sepsis 1 (5) 1 (5) 0 (0) 0 (0) Controlled ventilation period (d), 7 (2) 6 (2) 8 (1) 8 (1) mean (SD) Exp = experimental group, Con = control group, OTT = oro-tracheal tube, APACHE = Acute Physiology and Chronic Health Evaluation, ARF = acute respiratory failure Results 1.0 'MPX
PG
QBSUJDJQBOUT
BOE
UIFSBQJTUT
UISPVHI
UIF
Cumulative probability of still being intubated 0.8 trial 0.6 During the recruitment period, 198 patients were screened, of whom 67 were eligible and monitored daily 0.4 to assess readiness to start weaning. Of the 67, 20 were tracheostomised, 5 died, and 1 was transferred to another 0.2 centre before the start of weaning. The remaining 41 were randomised: 21 to the experimental group and 20 to the 0.0 5 10 15 control group. The baseline characteristics, ie, on the day weaning started, of the two groups are presented in Table 0 Days 1 and in the first two columns of Table 2. Four participants in each group died before extubation. Three participants in 'JHVSF
 Kaplan-Meier estimates of time to extubation. the experimental group and two in the control group were Blue line = experimental group; Black line = control group. tracheostomised before extubation. Effect of intervention The intensive care unit had a total of 24 beds, with 8 of these dedicated to postoperative patients. The physiotherapy Group data for all outcomes at the start of weaning and team comprised 11 physiotherapists working in three shifts, at extubation for the experimental and control groups are all with expertise in intensive care, of which two have presented in Table 2 while individual data are presented in doctoral and six have masters qualifications. Consistency Table 3 (see eAddenda for Table 3). Maximal inspiratory between the physiotherapists for the assessment of maximal pressure increased significantly more in the treatment group inspiratory pressure was good, with a Kappa value of 0.68. than the control group (MD 7.6 cmH20, 95% CI 5.8 to 9.4). The index of Tobin increased (ie, worsened) in both groups $PNQMJBODF
XJUI
USJBM
NFUIPE over the weaning period, but the increase was attenuated Participants in the experimental group underwent training on all days during their weaning period. The average training load of the participants in the experimental group increased from 3 cmH2O initially to 20 cmH2O at the end of the weaning period. 174 Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010

Cader et al: Inspiratory muscle training and weaning time 5BCMF
. Mean (SD) of outcomes for each group, mean (SD) difference within groups and mean difference (95% CI) between groups. Outcome Groups Difference within groups Difference between groups Pre-test Post-test Post-test minus pre-test Post-test minus pre-test MIP Exp Con Exp Con Exp Con Exp minus con (cmH2O) (n = 14) (n = 14) (n = 14) (n = 14) IT 9.9 2.3 7.6 (br/min/L) 15.1 15.3 25.0 17.6 (2.5) (2.1) (5.8 to 9.4) (2.6) (2.2) (3.9) (1.9) 6.1 14.4 (3.6) (9.1) -8.3 73.6 81.5 79.7 95.9 (–13.7 to –2.9) (8.8) (6.9) (11.2) (12.4) C?F3cWn_cWb_dif_hWjehofh[iikh[Yc>(E\"?J3_dZ[ne\\JeX_dXh%c_d%B\";nf3[nf[h_c[djWb]hekf\"9ed3Yedjheb]hekf\"Fh[#j[ij3ijWhj of weaning, Post-test = extubation, shaded row = primary outcome Table 4. Mean (SD) duration of the weaning period among uncensored participants. Outcome Groups Difference between groups Weaning period (days) Exp Con Exp minus Con Exp = experimental group, Con = control group (n = 14) (n = 14) 1.7 (0.4 to 3.0) 3.6 (1.5) 5.3 (1.9) significantly by the inspiratory muscle training (MD 8.3 br/ However, because informed consent was provided by min/L, 95% CI 2.9 to 13.7). the relatives of these critically ill patients, the potential for placebo and Hawthorne effects to operate within the Among uncensored participants (ie, those who did not patients was reduced. die or receive a tracheostomy), the weaning period was significantly shorter in the experimental group than in the Previous research suggests that imbalance between control group. The effect of inspiratory muscle training was the ventilatory load and the strength and endurance of to reduce the weaning period by 1.7 days (95% CI 0.4 to the respiratory muscles is an important determinant of 3.0), as presented in Table 4, with individual data in Table 5 dependence on mechanical ventilation. For example, (see eAddenda for Table 5). patients who have success in weaning have a significantly higher maximal inspiratory pressure than those who do not Prior to the weaning period, the controlled ventilation wean successfully (Epstein et al 2002). This relationship period (see Table 1) accounted for approximately half of is also reflected in our data, with the experimental group the total ventilation period. A Kaplan-Meier analysis of the showing both a significant increase in maximal inspiratory total intubation time (ie, the controlled ventilation period pressure and a reduction in the weaning period when plus the weaning period) did not identify a significant compared to the control group. difference between the experimental and control groups (p = 0.72, see Figure 2.) Our findings that inspiratory muscle training improved both inspiratory muscle strength and the weaning process are also Discussion similar to the findings of several other case series. Martin and colleagues (2002), Sprague and Hopkins (2003), and Although we screened 198 patients in the intensive care Chang and colleagues (2005b) delivered inspiratory muscle unit, a large proportion of these critically ill patients died training to tracheostomised patients with a long-standing or were tracheostomised either before or after commencing dependence on mechanical ventilation. All of these patients weaning. This is typical of research in inspiratory muscle showed improved inspiratory muscle strength and almost training in the intensive care setting (Caruso et al 2005, all weaned successfully within several weeks of starting Chang et al 2005a, How et al 2007, Sprague and Hopkins the training. However, the participants in the current study 2003). This loss to follow-up was one limitation of the differed by being intubated and not tracheostomised, and study. It was compounded by the wide variability in the the conclusions are more robust due to the inclusion of a condition of these patients, including modifications to their randomised control group. medication regimen, psychological state, haemodynamic stability, and degree of sepsis. Nevertheless, the sample size Our findings differ, however, from those of one randomised remained sufficient for statistically significant between- trial (Caruso et al 2005). In this trial, inspiratory muscle group differences to be identified on several outcomes. training was achieved by increasing the pressure required Another limitation of the study was the lack of blinding. to trigger pressure support, and the outcomes were the Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010 175

Research duration of the weaning period and the rate of re-intubation Ambrosino N (2005) Weaning and respiratory muscle in critically ill patients. The experimental and control dysfunction: The egg-chicken dilemma. Chest 128: 481–483. groups did not differ significantly in terms of the weaning period (p = 0.24) and the maximum inspiratory pressure WMA–World Medical Association (2008) Declaration of final value (p = 0.34). One possible explanation for the Helsinki. Ethical Principles for Medical Research Involving discrepancy between the studies is that inspiratory muscle Human Subjects. 59th WMA General Assembly, Seoul. training via reduction of sensitivity of the pressure support trigger only offers an initial resistance to the opening of the Barreto AC LYG, Ribeiro LG (2004) Determinação do tamanho valve of the system, while inspiratory muscle training with amostral. Fitness e Performance Journal 3: 124. a threshold device maintains resistance to the respiratory system for the period of the inspiration. Other studies Blattner C, Guaragna JC, Saadi E (2008) Oxygenation and have also reported differences in the clinical efficacy of static compliance is improved immediately after early manual inspiratory muscle training when delivered by a threshold hyperinflation following myocardial revascularisation: device versus another method (Johnson et al 1996). a randomised controlled trial. Australian Journal of Physiotherapy 54: 173–178. The beneficial effect of inspiratory muscle training on the index of Tobin in this study indicates a more relaxed Boles JM, Bion J, Connors A, Herridge M, Marsh B, Melot C breathing pattern. This is consistent with a study of et al (2007) Weaning from mechanical ventilation. European inspiratory muscle training in 23 healthy adults (Huang et Respiratory Journal 29: 1033–1056. al 2003). After training, a significant increase in maximum inspiratory pressure was observed, which had a significant Cahalin LP, Semigran MJ, Dee GW (1997) Inspiratory muscle negative correlation with the significant reduction in training in patients with chronic heart failure awaiting cardiac respiratory stimulation P0.1. These data suggest that a transplantation: results of a pilot clinical trial. Physical reduced time of P0.1 results in a reduction in the occurrence Therapy 77: 830–838. of dyspnoea. Caruso P, Denari SD, Ruiz SA, Bernal KG, Manfrin GM, Inspiratory muscle training in the experimental group was Friedrich C et al (2005) Inspiratory muscle training is found to contribute to a significant increase in maximum ineffective in mechanically ventilated critically ill patients. inspiratory pressure and to a reduction in the index of Tobin. Clinics (Sao Paulo) 60: 479–484. These are considered to be good predictors of weaning, which is consistent with our finding that inspiratory muscle Caruso P, Friedrich C, Denari SD, Ruiz SA, Deheinzelin training significantly reduces the weaning period in patients D (1999) The unidirectional valve is the best method to who did not die or receive a tracheostomy. determine maximal inspiratory pressure during weaning. Chest 115: 1096–1101. We conclude that inspiratory muscle training improves inspiratory muscle strength in older intubated patients. In Chang AT, Boots RJ, Brown MG, Paratz J, Hodges PW (2005a) patients who do not die or receive a tracheostomy, it may Reduced inspiratory muscle endurance following successful also reduce weaning time. Q weaning from prolonged mechanical ventilation. Chest 128: 553–559. eAddenda: Tables 3 and 5 available at www.jop. physiotherapy.asn.au Chang AT, Boots RJ, Henderson R, Paratz JD, Hodges PW (2005b) Case report: inspiratory muscle training in chronic Ethics: Committee of Ethics in Research Involving Human critically ill patients––a report of two cases. Physiotherapy Beings of the Euro-American Network of Human Kinetics Research International 10: 222–226. – REMH (protocol number: 005/2007). Informed consent was obtained from each participant’s relatives with no Conti G, Montini L, Pennisi MA, Cavaliere F, Arcangeli A, Bocci refusals, and the experimental procedures were executed MG et al (2004) A prospective, blinded evaluation of indexes in accordance with the Declaration of Helsinki from 1975. proposed to predict weaning from mechanical ventilation. Intensive Care Medicine 30: 830–836. Acknowledgements: We are grateful to the physiotherapists in the Center of Intensive Therapy for their help with Epstein CD, El-Mokadem N, Peerless JR (2002) Weaning older measurement. patients from long-term mechanical ventilation: a pilot study. American Journal of Critical Care 11: 369–377. Competing interests: None declared. Epstein SK, Ciubotaru RL (1996) Influence of gender and Correspondence: Professor Sam ria Ali Cader, LABIMH, endotracheal tube size on preextubation breathing pattern. Castelo Branco University, Brazil. Email: samariacader American Journal of Respiratory Critical Care Medicine 154(6 gmail.com Pt 1): 1647–1652. References Epstein S (2006) Complications in ventilator supported patients. In: Tobin M (ed) Principles and Practice of Mechanical Afessa B, Hogans L, Murphy R (1999) Predicting 3-day and Ventilation 2nd ed. New York: McGraw-Hill, pp 877–902. 7-day outcomes of weaning from mechanical ventilation. Chest 116: 456–461. Esteban A, Alia I, Ibanez J, Benito S, Tobin MJ, the Spanish Lung Failure Collaborative Group (1994) Modes of Aldrich TK, Karpel JP, Uhrlass RM, Sparapani MA, Eramo mechanical ventilation and weaning: a national survey of D, Ferranti R (1989) Weaning from mechanical ventilation: Spanish hospitals. Chest 106: 1188–1193. adjunctive use of inspiratory muscle resistive training. Critical Care Medicine 17: 143–147. Green M, Road J, Sieck GC, Similowski T (2002) Tests of respiratory muscle strength. American Journal of Respiratory and Critical Care Medicine 166: 528–547. How SC, McConnell AK, Taylor BJ, Romer LM (2007) Acute and chronic responses of the upper airway to inspiratory loading in healthy awake humans: an MRI study. Respiratory Physiology and Neurobiology 157: 270–280. Huang CH, Martin AD, Davenport PW (2003) Effect of inspiratory muscle strength training on inspiratory motor drive and RREP early peak components. Journal of Applied Physiology 94: 462–468. Johnson PH, Cowley AJ, Kinnear WJ (1996) Evaluation of the THRESHOLD trainer for inspiratory muscle endurance training: comparison with the weighted plunger method. European Respiratory Journal 9: 2681–2684. Kim J, Sapienza CM (2005) Implications of expiratory muscle 176 Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010

strength training for rehabilitation of the elderly: Tutorial. Cader et al: Inspiratory muscle training and weaning time Journal of Rehabilitation Research and Development 42: 211–224. Use of inspiratory muscle strength training to facilitate ventilator weaning: a series of 10 consecutive patients. Krieg S, Alison JA, McCarren B, Cowell S (2007) Position Chest 122: 192–196. affects distribution of ventilation in the lungs of older people: an experimental study. Australian Journal of Physiotherapy Martinez A, Seymour C, Nam M (2003) Minute ventilation 53: 179–184. recovery time: a predictor of extubation outcome. Chest 123: 1214–1221. Lemes DA, Zin WA, Guimarães FS (2009) Hyperinflation using pressure support ventilation improves secretion clearance Meade M, Guyatt G, Cook D, Griffith L, Sinuff T, Kergl C et and respiratory mechanics in ventilated patients with al (2001) Predicting success in weaning from mechanical pulmonary infection: a randomised crossover trial. Australian ventilation. Chest 120(6 Suppl): 400S–424S. Journal of Physiotherapy 55: 249–254. Sprague SS, Hopkins PD (2003) Use of inspiratory strength Liaw MY, Lin MC, Cheng PT, Wong MK, Tang FT (2000) training to wean six patients who were ventilator-dependent. Resistive inspiratory muscle training: its effectiveness in Physical Therapy 83: 171–181. patients with acute complete cervical cord injury. Archives of Physical Medicine and Rehabilitation 81: 752–756. Tobin MJ (2001) Advances in mechanical ventilation. New England Journal of Medicine 344: 1986–1996. Marini JJ, Smith TC, Lamb V (1986) Estimation of inspiratory muscle strength in mechanically ventilated patients: The Yang KL, Tobin MJ (1991) A prospective study of indexes measurement of maximal inspiratory pressure. Journal of predicting the outcome of trials of weaning from mechanical Critical Care 1: 32–38. ventilation. New England Journal of Medicine 324: 1445– 1450. Martin AD, Davenport PD, Franceschi AC, Harman E (2002) OWd] AB '//) ?dif_hWjeho fh[iikh[%cWn_cWb _dif_hWjeho pressure ratio: a predictive index of weaning outcome. Intensive Care Medicine 19: 204–208. Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010 177

Appraisal Clinical Practice Guidelines Juvenile idiopathic arthritis Clinical guideline for the diagnosis and management of juvenile idiopathic arthritis Latest update: August 2009. Date of next update: 2014. Description: This is a 43-page document that presents Patient group: Patients aged under 16 years presenting recommendations to assist with the early diagnosis and with arthritic symptoms and those diagnosed with multidisciplinary management of JIA in the primary care Juvenile idiopathic arthritis (JIA). Intended audience: setting. The guideline focuses on evidence underpinning Health professionals (general practitioners and allied four main areas: the diagnosis of JIA, treatment and health including physiotherapy) in the primary health care management of JIA in the early stage, during acute episodes, setting. Additional versions: Nil. Expert working group: and the long term management of JIA. It covers issues such Two working groups were involved: the Royal Australian as early and accurate diagnosis, care and referral pathways, College of General Practitioners (RACGP) Juvenile use of medications, non-pharmacological management Idiopathic Arthritis Working Group consisted of 8 health including evidence for land and water exercise, patient care professionals (representing medicine, nursing, public self-management education, and psychosocial support health, and physiotherapy) and a consumer representative. requirements. Two detailed algorithms are presented on The Australian Paediatric Rheumatology Working Group pages 8 and 9, covering the diagnosis and early management consisted of 7 medical fellows. Funded by: RACGP of JIA, and the management of JIA. A summary of the and the Australian Department of Health and Ageing. 21 recommendations is presented on pages 10–11, with Consultation with: Draft versions of the guidelines were more detailed explanation of the recommendation level available on the RACGP website for public consultation, and specific evidence contained in pages 12–24. Three and over 200 stakeholder groups were targeted specifically. pages of resources are provided on pages 35–37 including Approved by: National Health and Medical Research publications, electronic sources (websites), and a history and Council of Australia, RACGP. Location: http://www.racgp. clinical examination checklist to assist with examination org.au/guidelines/juvenileidiopathicarthritis and differential diagnosis. Sandra Brauer The University of Queensland The COPD-X plan \"VTUSBMJBO
BOE
/FX
;FBMBOE
HVJEFMJOFT
GPS
UIF
NBOBHFNFOU
PG
DISPOJD
PCTUSVDUJWF
pulmonary disease Latest update: May 2010. Date of next update: 2014. Description: The .pdf version is a 71-page document that Patient group: Individuals with chronic obstructive presents recommendations and the underlying evidence pulmonary disease (COPD). Intended audience: Health to assist with the diagnosis and management of patients professionals who manage patients with COPD. Additional with COPD. The key recommendations are summarised versions: This is the first update to the guidelines. The on page 10 in the COPD- plan: Confirm diagnosis, original guidelines were published in the Medical Journal Optimise function, Prevent deterioration, Develop a self- of Australia in 2003. (http://www.mja.com.au/public/ management plan, and manage eXacerbations. Information issues/178 06 170303/tho10508 all.html). Expert working is presented on the aetiology and natural history of COPD, group: The guidelines were developed by the Australian the role of history, physical examination, and spirometry Lung Foundation and the Thoracic Society of Australia in diagnosis, methods to assess severity, and indicators and New ealand. The guidelines evaluation committee for referral to specialist respiratory care. The evidence consisted of 8 Australian health professionals representing for the efficacy of medication and non-pharmacological medicine, public health, and physiotherapy. A larger group approaches to optimise function is discussed, including of 27 experts from Australia and New ealand including exercise, education and self management, pulmonary physiotherapists also contributed. Funded by: Australian rehabilitation, chest physiotherapy, psychosocial support, Lung Foundation. Consultation with: Draft versions of and nutrition. Likely co-morbidities and their management the guidelines were available on the RACGP website for are presented, and surgical options and palliative care public consultation and over 200 stakeholder groups were are discussed. Evidence and approaches for the reduction specifically targeted. Approved by: The Royal Australian of risk factors such as smoking cessation, medication, College of Physicians, The Royal College of Nursing vaccination, and oxygen therapy are presented. The section Australia, the Australian Physiotherapy Association, on self management promotes a multidisciplinary team Australian Asthma and Respiratory Educators Association, approach. Evidence underpinning the management of and the Asthma Foundation. Location: The website (http:// acute exacerbations is presented. This includes guidelines www.copdx.org.au/home) contains the guidelines spread to confirm the exacerbation and categorise its severity, over pages on the site, as well as a .pdf version. pharmacological and non-pharmacological interventions, indicators for hospitalisation or ventilation, and discharge planning. Appendices provide information on inhaler devices, and long-term oxygen therapy. Sandra Brauer The University of Queensland Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010 205

Ada et al: Mechanically assisted walking in stroke patients Mechanically assisted walking with body weight support results in more independent walking than assisted overground walking in non-ambulatory patients early after stroke: a systematic review Louise Ada, Catherine M Dean, Janine Vargas and Samantha Ennis The University of Sydney, Australia Question: Does mechanically assisted walking with body weight support result in more independent walking and is it detrimental to walking speed or capacity in non-ambulatory patients early after stroke? Design: Systematic review with meta-analysis of randomised trials. Participants: Non-ambulatory adult patients undergoing inpatient rehabilitation up to 3 months after stroke. Intervention: Mechanically assisted walking (eg, treadmill, electromechanical gait trainer, robotic device, servo-motor) with body weight support (eg, harness with or without handrail, but not handrail alone) versus assisted overground walking of longer than 15 min duration. Outcome measures: The primary outcome was the proportion of participants achieving _dZ[f[dZ[dj mWba_d]$ I[YedZWho ekjYec[i m[h[ mWba_d] if[[Z c[Wikh[Z Wi c%i Zkh_d] j^[ '&#c MWba J[ij WdZ mWba_d] capacity measured as distance in m during the 6-min Walk Test. Results: Six studies comprising 549 participants were identified and included in meta-analyses. Mechanically assisted walking with body weight support resulted in more people walking independently at 4 weeks (RD 0.23, 95% CI 0.15 to 0.30) and at 6 months (RD 0.23, 95% CI 0.07 to 0.39), faster walking Wj,cedj^iC:&$'(c%i\"/+9?&$&(je&$('\"WdZ\\khj^[hmWba_d]Wj,cedj^iC:++c\"/+9?'+je/,j^WdWii_ij[Z overground walking. Conclusion: Mechanically assisted walking with body weight support is more effective than overground walking at increasing independent walking in non-ambulatory patients early after stroke. Furthermore, it is not detrimental to walking speed or capacity and clinicians should therefore be confident about implementing this intervention. [Ada L, Dean $.
7BSHBT
+
&OOJT
4

.FDIBOJDBMMZ
BTTJTUFE
XBMLJOH
XJUI
CPEZ
XFJHIU
TVQQPSU
SFTVMUT
JO
NPSF
JOEFQFOEFOU
XBMLJOH
UIBO
BTTJTUFE
PWFSHSPVOE
XBMLJOH
JO
OPOBNCVMBUPSZ
QBUJFOUT
FBSMZ
BGUFS
TUSPLF
B
TZTUFNBUJD
SFWJFX
Journal of Physiotherapy

o> ,FZ
XPSET: Stroke, Treadmill, Walking, Systematic review, Meta-analysis, Randomised controlled trials Introduction treadmill walking with other mechanised walking (such as an electromechanical gait trainer) which may be expected After stroke, many individuals have residual walking to result in even more practice than treadmill walking. disability. Despite recent advances in medical and A systematic review examining electromechanical gait rehabilitation sciences, only half of those who cannot trainers only (Mehrholz et al 2010) found an increase in the walk on entering rehabilitation after stroke regain the likelihood of walking. We therefore planned a systematic ability to walk (Dean and Mackey 1992). Being able to review focusing broadly on any mechanically assisted walk independently is a major determinant of whether an walking, and comparing it with overground walking so that individual returns home following a stroke and has long therapists and health administrators would have evidence lasting implications for the person’s quality of life and to help guide decision making in terms of investing in ability to participate in activities of daily living. mechanical walking equipment. In particular, we were interested in whether any benefits of mechanically assisted For non-ambulatory stroke patients, mechanically assisted walking were still apparent in the long term or whether the walking with body weight support has been suggested as effect was short lived. a strategy to facilitate walking (Hesse 1998, Richards et al 1993) because it provides the opportunity to complete Clinicians still seem reluctant to implement treadmill more practice of the whole task than would be possible training for stroke patients due to a fear that an abnormal by assisting overground walking. A Cochrane Review walking pattern will be practised (Hesse 2008) resulting (Moseley et al 2005) found no statistically significant in abnormal overground walking (Davies 1999). We were difference between treadmill walking with body weight therefore interested in examining any aspects of walking support when compared with any other walking intervention commonly measured, such as speed and capacity, which in terms of amount of independent walking, walking speed, would shed some light on whether this fear is reasonable. or walking capacity. The main conclusion from this review The specific research questions for this review were: was an urgent need for well-designed large-scale studies to evaluate the effects of treadmill walking and body weight 1. In subacute, non-ambulatory patients after stroke, support on walking after stroke. Since then, more large scale does mechanically assisted walking with body weight trials have been completed. The inconclusive result of the support result in more independent walking than Cochrane review could be partially the result of comparing overground walking in the short term? 2. Is it detrimental in terms of walking speed or capacity? Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010 153

Research 3. Are any benefits maintained in the long term? scale rates the methodological quality of randomised trials between 1 and 10. The score is determined by In order to make recommendations based on the highest two independent raters, with a third rater resolving any level of evidence, this review included only randomised disagreements. Where a study was not included on the or quasi-randomised trials in which patients undergoing database, the PEDro scale was scored by two reviewers inpatient stroke rehabilitation to enable them to walk were independently with disagreements resolved by a third randomised to receive either mechanically assisted walking reviewer. with body weight support or assisted overground walking. Participants: Studies involving subacute, non-ambulatory, Method adult stroke survivors were included. Subacute was defined as within the first three months following stroke. Non- Identification and selection of studies ambulatory was defined as Functional Ambulatory Category Searches were conducted of the following databases: 3 (Holden et al 1984), Functional Independence Measure MEDLINE (1966 to August Week 4 2009), CINAHL (1982 (Keith et al 1987) walking subscale score 5, Item 5 Motor to August Week 4 2009), EMBASE (1980 to August Week Assessment Scale score 2, or equivalent. Even so, in many 4 2009) and PEDro (to August Week 4 2009), without trials, the ambulation status of the participants at baseline language restrictions for relevant articles. Search terms was not clear. Therefore, the measurement of independent included words relating to stroke, exercise therapy, and walking as an outcome was used as an inclusion criterion in locomotion (see Appendix 1 on the eAddenda for the full order to confirm that the trial investigated participants who search strategy). In addition, we contacted authors about were non-ambulatory at baseline. trials that we knew were in progress from trial registration. Title and abstracts were displayed and screened by one Intervention: The experimental intervention was any reviewer to identify relevant studies. Full paper copies of type of mechanically assisted walking (such as treadmill, relevant studies were retrieved and their reference lists were electromechanical gait trainer, robotic device or servo- screened. The methods of retrieved papers were extracted motor) with body weight support (provided by a harness so that reviewers were blinded to authors, journals and system, with or without handrail, but not handrail alone) outcomes and examined against predetermined inclusion regardless of the amount of therapist assistance. The control criteria (Box 1) by two independent reviewers. Conflict of intervention was overground walking and could include any opinion was resolved by consensus after discussion with a type of assistance from therapists or aids (such as orthoses third reviewer. or sticks). Training was required to be of a duration that could be expected to improve walking, ie, ®15 minutes per #PY
. Inclusion criteria. session. Design Outcome measures: The amount of independent walking š RCT or Q-RCT was the primary outcome measure. Independent walking š PEDro Scale score > 4 was defined as being able to walk without aids or physical Participants assistance (ie, Functional Ambulatory Category ≥ 3 or š Adults (> 18 year) equivalent). š Stroke patients (> 24 hr) š Subacute (≤ 3 months after stroke) Secondary outcomes were walking speed and walking š Non-ambulatory (FAC < 3 [0-5], FIM < 5 [1–7]) or capacity. Walking speed was measured in m/s during any short distance test (such as the 10-m Walk Test, Wade et al where mixed participants, data for non-ambulatory 1987). Walking capacity was measured as distance walked patients reported separately in m during a longer timed test (such as the 2-, 5-, 6- or 12- min Walk Test) and converted to the equivalent of a 6-min Intervention Walk Test (Guyatt et al 1984). For both secondary outcomes, š Experimental group received mechanically assisted only data from participants who could walk independently were used. walking with body weight support > 15 min š Control group received assisted overground walking Data analysis > 15 min Data were extracted from the included studies by one reviewer and cross-checked by a second reviewer. Outcomes measured Information about the method (ie, design, participants, š Independent walking (measured as proportion or intervention, measures) and outcome data (ie, number of participants who could walk independently, mean (SD) scale from which proportion could be determined) walking speed, and walking capacity) were extracted. Authors were contacted where there was difficulty Comparisons extracting and interpreting data from the paper. š Mechanically assisted walking with body weight The post-intervention scores were used to obtain the pooled support vs overground walking estimate of the effect of intervention at 4 weeks (short term) š Mechanically assisted walking with body weight and 6 months (long term). A fixed-effects model was used. In the case of significant statistical heterogeneity (I2 ®25%), support + assisted overground walking vs a random-effects model was applied to check the robustness overground walking of the results. The analyses were performed using the MI a program (Bax et al 2006, Bax et al 2008). Dichotomous RCT = randomised clinical trial, Q-RCT = quasi-randomised outcomes (ie, amount of independent walking) were clinical trial, FAC = Functional Ambulation Category, FIM = reported as risk difference (95% CI) whereas continuous Functional Independence Measure Assessment of characteristics of studies Quality: The quality of included studies was determined by obtaining PEDro scale scores from the Physiotherapy Evidence Database (www.pedro.org.au). The PEDro 154 Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010

Ada et al: Mechanically assisted walking in stroke patients Titles and abstracts screened Du et al 2006, Schwartz et al 2009, Tong et al 2006) were (n = 2425) included in the review. One trial was reported across two publications (Ada et al 2010, Dean et al 2010), so the seven Papers excluded after screening included papers provided data on six studies. See Figure j_jb[i%WXijhWYjid3().* 1 for flow of studies through the review. See Table 1 for a summary of the excluded papers (see eAddenda for Table Potentially relevant papers retrieved 1). for evaluation of full text (n = 43) š from electronic search (n = 41) Description of studies š from registry search (n = 2) Six randomised trials investigated the effect of mechanically assisted walking on independent walking. Five trials investigated the effect on walking speed. Two trials investigated the effect on walking capacity. The quality of the included studies is outlined in Table 2 and a summary of the studies is presented in Table 3. Papers excluded after evaluation Quality: The mean PEDro score of the included studies was of full text (n = 36)* 6.7. Randomisation was carried out in 100% of the studies, š research design not RCT or concealed allocation in 33%, assessor blinding in 66%, and intention-to-treat analysis in 83%. Only one trial reported Q-RCT (n = 17) a loss to follow up greater than 15% – and that was only š PEDro Scale score ≤ 4 (n = 3) 16%. No study blinded participants or therapists, due to the š not subacute (n = 3) inherent difficulties associated with these interventions. š ambulatory participants Participants: The mean age of participants across studies included (n = 7) ranged from 57 to 73 and they were on average within the š experimental intervention not first month after their stroke. Non-ambulatory was defined as Functional Ambulatory Category 3 (five studies) and mechanically assisted walking Motor Assessment Scale Item 5 score 2 (one study). (n = 3) š control did not include Intervention: Mechanically assisted walking included overground walking (n = 6) treadmill with harness (two studies), treadmill with š independent walking not an robotic device and harness (Lokomat) (one study) and outcome measure (n = 8) electromechanical gait trainer with harness (three studies). š same data reported in other The experimental group received some assisted overground study (n = 1) walking as well as mechanically assisted walking in the š not enough information majority of studies (four studies). The control group in all (n = 2) studies received overground walking assisted by therapists. š independent walking not Participants trained from 20 to 80 min/day, from 3 to 5 an outcome (n = 8) days/wk for 4 to 6 wk or until discharge from inpatient rehabilitation. The experimental group received the same Papers included in amount of walking training as the control group in all systematic review (n = 7) studies. Studies included in systematic review (n = 6) Outcome measures: Independent walking was identified as the ability to walk 15 m continuously with no aids and in 'JHVSF
 Flow of studies through the review. * = papers bare feet (one study), a Functional Ambulatory Scale score may have been excluded for failing to meet more than one ≥ 3 (two studies) or ®3 (three studies). Independent walking inclusion criterion, RCT = randomised clinical trial, Q-RCT data were available for six studies at 4 weeks and three = quasi-randomised clinical trial. studies at 6 months. Walking speed was measured during the 10-m Walk Test (three studies) and the 5-m Walk Test outcomes (ie, walking speed and capacity) were reported as (two studies) and all results were converted to m/s. Walking the weighted mean difference (95% CI). speed data were available for five studies at 4 weeks and three studies at 6 months. Walking capacity was measured Results using the 6-min Walk Test (two studies) and the 2-min Walk Test (one study) and these results were multiplied to equate 'MPX
PG
TUVEJFT
UISPVHI
UIF
SFWJFX to 6 min. Walking capacity data were available for two The search returned 2425 papers. After screening the titles studies at 4 weeks and at 6 months. and abstracts, 41 papers were retrieved for evaluation of full text. Another two papers were retrieved as a result of Effect of intervention searching trial registries. Thirty-six papers failed to meet the inclusion criteria and therefore seven papers (Ada et Independent walking: The short-term effect of al 2010, Dean et al 2010, Ng et al 2008, Pohl et al 2007, mechanically assisted walking on independent walking was examined by pooling data at 4 weeks from six studies (Ada et al 2010, Du et al 2006, Ng et al 2008, Pohl et al 2007, Schwartz et al 2009, Tong et al 2006) involving 539 participants. Mechanically assisted walking increased independent walking compared with overground walking Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010 155

Research Total YYY8 NYY 5 YNYNNNYYYY 6 Pohl et al (2007) Y Y Y N N Y Y Y Y Y 8 Schwartz et al (2009) Y N Y N N Y N Y Y Y 6 a) Ada/Dean 156 (0 to 10) Tong et al (2006) Y N Y N N Y Y Y Y Y 7 Du Point Ng estimate and Pohl Schwartz variability Tong reported Between- –1 –0.5 0 0.5 1 group b) Ada/Dean difference reported Ng Pohl Intention- to-treat analysis –1 –0.5 0 0.5 1 < 15% 'JHVSF

Risk difference (95% CI) of effect of dropouts mechanically assisted walking for independent walking by pooling data from a) six studies at 4 weeks (n = 539) Y and b) three studies at 6 months (n = 312). Y Assessor (RD = 0.23; 95% CI 0.15 to 0.30) (Figure 2a, see also blinding Figure 3a on eAddenda for detailed forest plot), with 55% of participants in the experimental group being able to walk Y against 32% of participants in the control group. N The long-term effect of mechanically assisted walking on Therapist N YNYNN independent walking was examined by pooling data at 6 blinding months from three studies (Ada et al 2010, Ng et al 2008, Pohl et al 2007), involving 312 participants. Mechanically Participant N assisted walking increased independent walking compared blinding with overground walking (RD = 0.24, 95% CI 0.13 to 0.34), with 70% of participants in the experimental group being Groups Y able to walk against 46% of participants in the control similar at group. There was, however, between-study heterogeneity baseline for this outcome at 6 months (I2 = 51%), indicating that the variation between the results of the studies is above that 5BCMF
. PEDro scores for included studies (n = 6). Random Concealed Y expected by chance. When a random-effects model was allocation allocation applied the results were similar (RD = 0.23, 95% CI 0.07 Y to 0.39) (Figure 2b, see also Figure 3b on eAddenda for detailed forest plot). Study Ada et al (2010) Du et al (2006) Ng et al (2008) Dean et al (2010) Walking speed: The short-term effect of mechanically assisted walking on walking speed was examined by pooling data from five studies (Dean et al 2010, Ng et al 2008, Pohl et al 2007, Schwartz et al 2009, Tong et al 2006), involving the 142 participants who could walk independently at 4 weeks. Mechanically assisted walking increased walking speed by 0.09 m/s (95% CI 0.01 to 0.17) more than overground walking. There was, however, between-study heterogeneity for this outcome at 6 months (I2 = 41%), indicating that the variation between the results of the studies is above that expected by chance. When a random-effects model was applied the results were similar (MD = 0.10 m/s, 95% CI 0.00 to 0.21) (Figure 4a, see also Figure 5a on eAddenda for detailed forest plot). The long-term effect of mechanically assisted walking on walking speed was examined by pooling data from Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010

JournalofPhysiotherapy2010 Vol.56 – ©AustralianPhysiotherapyAssociation2010 5BCMF
 Summary of included studies (n = 7). Study Design Par ticipants Walking intervention Ada et al RCT n = 126 Exp = treadmill with harness + (2010) RCT Age (yr) = 73 (SD 12) assisted overground walking Dean et al Time since stroke (wk) = 2.4 <h[gk[dYo3)&c_dn+%makd (2010) Non-ambulatory = MAS < 2 Con = assisted overground wa Du et al n = 128 <h[gk[dYo3)&c_dn+%makd (2006) Age (yr) = 57 (SD 6) Exp = treadmill with harness + Time since stroke (wk) = ≤ 12 assisted overground walking Ng et al RCT Non-ambulatory = FAC < 3 <h[gk[dYo3*&c_dn(%ZWon (2008) RCT n = 38 Con = assisted overground wa Pohl et al Age (yr) = 70 (SD 11) <h[gk[dYo3*&c_dn(%ZWon (2007) Time since stroke (wk) = 2.6 Exp = electromechanical gait t Non-ambulatory = FAC < 3 <h[gk[dYo3(&c_dn+%man* Schwar t z RCT n = 155 et al Age (yr) = 63 (SD 12) Con = assisted overground wa (2009) Time since stroke (wk) = 4.4 <h[gk[dYo3(&c_dn+%man* Non-ambulatory = FAC < 3 Exp = electromechanical gait t Tong et al RCT assisted overground walking (2006) n = 67 <h[gk[dYo3]W_jjhW_d[h(&c_ Age (yr) = 64 (SD 8) Wii_ij[Zel[h]hekdZmWba_d]( Time since stroke (wk) = 3.2 Non-ambulatory = FAC < 3 Con = assisted overground wa <h[gk[dYo3*+c_dn+%man* n = 35 Exp = treadmill with robotic de Age (yr) = 69 (SD 12) assisted overground walking Time since stroke (wk) = 2.7 <h[gk[dYo3jh[WZc_bb)&c_dn Non-ambulatory = FAC < 3 Wii_ij[Zel[h]hekdZmWba_d]) Con = assisted overground wa <h[gk[dYo3,&c_dn)%ma!) Exp = electromechanical gait t <h[gk[dYo3]W_jjhW_d[h(&c_ Con = assisted overground wa <h[gk[dYo3(&c_dn+%man* Exp = experimental group, Con = control group, BWS = Body Weight Support, FAC = Functional Ambulati 157

Walking outcome measures + Ability to walk independently = 15 m, no aid Ada et al: Mechanically assisted walking in stroke patients Speed = 10-m Walk Test (comfortable, no aids) dj_bZ_iY^Wh][ Capacity = 6-min Walk Test alking <ebbem#kf3'%makdj_bZ_iY^Wh][\"(,ma dj_bZ_iY^Wh][ + Ability to walk independently = FAC ≥ 3 Speed = not measured *ma Capacity = not measured alking Follow-up = 0, 4 wk *ma trainer with harness Ability to walk independently = FAC > 3 *ma Speed = 5-m Walk Test (fast, aids) alking Capacity = not measured *ma Follow-up = 0, 4, 26 wk trainer with harness + Ability to walk independently = FAC > 3 Speed = 10-m Walk Test (fast) _dn+%man*ma! Capacity = 6-min Walk Test (+c_dn+%man*ma Follow-up = 0, 4, 24 wk alking *ma Ability to walk independently = FAC ≥ 3 evice and harness (Lokomat) + Speed = 10-m Walk Test (fast, aids) Capacity = 2-min Walk Test n)%man,ma! Follow-up = 0, 6 wk )&c_dn+%man,ma alking Ability to walk independently = FAC > 3 )&c_dn(%man,ma Speed = 5-m Walk Test (fast, aids) Capacity = not measured trainer with harness Follow-up = 0, 4 wk _dn+%man*ma alking *ma ion Category, MAS = Motor Assessment Scale

Research a) Ada/Dean independently at 4 weeks. Mechanically assisted walking increased walking capacity by 35 m (95% CI –13 to 84) Ng more than overground walking (Figure 6a, see also Figure Pohl 7a on eAddenda for detailed forest plot). Schwartz Tong The long-term effect of mechanically assisted walking on walking capacity was examined by pooling data from two –0.6 –0.4 –0.2 0 0.2 0.4 0.6 studies (Ada et al 2010, Pohl et al 2007), involving the 152 participants who could walk independently at 6 months. m/s Mechanically assisted walking increased walking capacity by 55 m (95% CI 15 to 96) more than overground walking (Figure 6b, see also Figure 7b on eAddenda for detailed forest plot). b) Ada/Dean Discussion Ng The strength of this systematic review is that it has Pohl pooled data from randomised trials of mechanically assisted walking (and included both treadmill and electro- –0.6 –0.4 –0.2 0 0.2 0.4 0.6 mechanical gait trainers) with body weight support compared with the usual practice of overground walking m/s in non-ambulatory people during the subacute phase of stroke. It includes six studies of reasonable size that have 'JHVSF

Mean difference (95% CI) for effect of investigated the effect of mechanically assisted walking mechanically assisted walking on walking speed by with body weight support on independence, speed and pooling data from a) five studies at 4 weeks (n = 142) and capacity of walking. The review provides evidence that b) three studies at 6 months (n = 172). mechanically assisted walking with body weight support increases the amount of independent walking without being a) Schwartz detrimental to walking speed or capacity after 4 weeks of intervention. Furthermore, the benefits appear to be Pohl maintained at 6 months with walking speed and capacity being superior in patients who received mechanically –200 –200 0 100 200 assisted walking during inpatient rehabilitation. m The six studies included in this review were of moderate to b) Ada/Dean good methodological quality. Given that 8 was the likely maximum PEDro score achievable (because it is not usually Pohl possible to blind the therapist or the participants), the mean score of 6.7 suggest that the findings are credible. There were sufficient data for a meta-analysis to be performed on each outcome measure. The number of participants included in the meta-analyses of independent walking was 539 at 4 weeks and 312 at 6 months. Even the meta-analyses of walking speed and capacity, which were carried out only on those who could walk, included numbers ranging from 88 to 172. –200 –200 0 100 200 Meta-analysis indicated that, on average, 23% more patients (ie, 55% of participants in the experimental m group compared with 32% of participants in the control group) could walk after 4 weeks of mechanically assisted 'JHVSF

Mean difference (95% CI) of effect of walking with body weight support than could walk after mechanically assisted walking for walking capacity by assisted overground walking, ie, it decreased dependence pooling data from a) two studies at 4 weeks (n = 88) and b) for those patients who were non-ambulatory a few weeks two studies at 6 months (n = 152). after stroke. In addition, there were sufficient data from two trials to examine whether this benefit was maintained. three studies (Ada et al 2010, Ng et al 2008, Pohl et al At 6 months, there were still 24% more people (ie, 70% of 2007), involving the 172 participants who could walk participants in an experimental group compared with 46% independently at 6 months. Mechanically assisted walking of participants in a control group) walking having received increased walking speed by 0.12 m/s (95% CI 0.02 to 0.21) mechanically assisted walking as an inpatient compared more than overground walking (Figure 4b, see also Figure with those having received overground walking. Even 5b on eAddenda for detailed forest plot). though there was statistical heterogeneity between these studies suggesting caution, it is encouraging that the mean Walking capacity: The short-term effect of mechanically benefit was almost the same when a random effects model assisted walking on walking capacity was examined by was applied (23% more patients walking) and was also the pooling data from two studies (Schwartz et al 2009, Pohl same as it had been at 4 weeks when 539 participants were et al 2007), involving the 88 participants who could walk pooled over six studies. 158 Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010

Ada et al: Mechanically assisted walking in stroke patients One hypothesis for the increase in independent walking together, these reviews suggest that it is worthwhile to use with mechanically assisted walking is that this intervention some form of mechanical assistance to improve walking provides the opportunity to complete more whole task after stroke. walking practice than would be possible with overground walking alone. The allowable amount of time spent This review has some potential limitations. First, as is on walking was the same for the control group as the usual with studies of complex interventions, the outcome experimental group in all the studies. However, three measures were not the same, although they were similar. studies report more distance covered or steps taken by the Second, only half the studies measured the outcomes in the group receiving mechanically assisted walking than the long term. Finally, most systematic reviews are susceptible group receiving assisted overground walking. Ada et al to publication bias and we attempted to pre-empt this by (2010) report that in Week 1 the average distance walked including studies published in languages other than English. per session by the control group was only 20% of the experimental group and in the last week the distance was In conclusion, this systematic review provides evidence that still less than 50%. Similarly, Pohl et al (2007) report that mechanically assisted walking results in more independent the average steps taken per session by the control group was walking after 4 weeks of intervention in patients who cannot less than 20% of the experimental group, and Tong et al walk within the first month after stroke. Importantly, this (2006) report that the steps taken per session by the control increase is without detriment to walking speed or capacity. group were 10% of the experimental group. Therefore, for a Further, benefits appear to be maintained at 6 months, with similar therapy time, more walking was carried out. Given walking capacity and speed being superior in those who the evidence from a systematic review of randomised trials received mechanically assisted walking during inpatient that outcome after stroke is associated with the amount rehabilitation. Q of practice undertaken (Kwakkel et al 2004), the extra walking carried out during the same therapy time probably Footnote: aMix for Meta-Analysis Made Easy Version 1.7. explains why more patients receiving mechanically assisted www.mix-for-meta-analysis.info/ walking walked independently than those receiving assisted overground walking. eAddenda: Appendix 1 (search strategy), Figures 3, 5, and 7 available at www.jop.physiotherapy.asn.au. Meta-analysis revealed that mechanically assisted walking resulted in more walking without compromising the walking Acknowledgements: We are grateful to Jan Mehrholz and itself. For those participants who could walk independently Raymond Tong for providing information and/or data. at 4 weeks, mechanically assisted walking with body weight support tended to produce 0.09 m/s higher walking Correspondence: Assoc Prof Louise Ada, Discipline of speeds. The upper limit of the 95% CI only just spans a Physiotherapy, Faculty of Health Sciences, The University worthwhile effect which has been suggested as 0.16 m/s by of Sydney, Australia. Email: louise.ada sydney.edu.au Tilson et al (2010). However, it does strongly suggest that mechanically assisted walking is not detrimental to walking References speed. Furthermore, at 6 months, there was a statistically significant improvement in walking speed of 0.12 m/s for Ada L, Dean CM, Morris ME, Simpson JM, Katrak P (2010) participants who gained the ability to walk independently as Randomised trial of treadmill walking with body weight a result of mechanically assisted walking and body weight support to establish walking in subacute stroke: the support compared with overground walking. Furthermore, MOBILISE trial. Stroke 41: 1237–1242. the upper limit of the 95% CI spans a worthwhile effect. Baker B, Breen JC, Snyder D, Kelley T (2006) Body weight For those participants who could walk independently at 4 support treadmill training in community rehabilitation weeks, mechanically assisted walking with body weight program improves walking in severely disabled stroke. support tended to produce 35 m further walking distance, Journal of Neurologic Physical Therapy 30: 210–210. with the average capacity achieved by participants in the experimental group being 144 m compared with 110 m Bax L, Yu LM, Ikeda N, Tsuruta H, Moons KG (2006) achieved by participants in a control group. This strongly Development and validation of MIX: comprehensive free suggests that mechanically assisted walking is not software for meta–analysis of causal research data. BMC detrimental to walking capacity. Furthermore, at 6 months, Medical Reserach Methodology 13: 50. there was a statistically significant improvement in walking distance of 55 m for participants who gained the ability to Bax L, Ikeda N, Fukui N, Yaju Y, Tsuruta H, Moons SGM (2008) walk independently as a result of mechanised walking and More than numbers: the power of graphs in meta-analysis. body weight support compared with overground walking. American Journal of Epidemiology 169: 249–255. In the two studies that included a 6 month follow-up, the average distance walked in 6 min for the experimental Bogey R, Hornby GT (2007) Gait training strategies utilized in group was 203 m compared with 148 m in the control group. poststroke rehabilitation: are we really making a difference? Topics in Stroke Rehabilitation 14: 1–8. Our review reports similar findings to that of a recent Cochrane systematic review investigating the use of Chen G, Patten C (2006) Treadmill training with harness electromechanical gait trainers to improve walking after support: selection of parameters for individuals with post stroke. Mehrholz et al (2010) found that electromechanical stroke hemiparesis. Journal of Rehabilitation Research & gait training increased the odds of becoming independent in Development 43: 485–497. walking (OR 2.21, 95% CI 1.52 to 3.22) without detriment to walking speed (MD 0.04 m/s, 95% CI –0.05 to 0.14) da Cunha Jr IT, Lim PA, Qureshy H, Henson H, Monga T, Protas or walking capacity (MD 7 m, 95% CI –32 to 46). Taken EJ (2002) Gait outcomes after acute stroke rehabilitation with supported treadmill ambulation training: a randomized controlled pilot study. Archives of Physical Medicine and Rehabilitation 83: 1258–1265. Davies PM (1999) Weight-supported treadmill training. Neurorehabilitation and Neural Repair 13: 167–169. Dean CM, Ada L, Bampton J, Morris ME, Katrak P, Potts S (2010) Treadmill walking with body weight support in Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010 159

Research subacute non-ambulatory stroke improves walking capacity Kwakkel G, van Peppen R, Wagenaar R, Dauphinee SW, more than overground walking: a randomised trial. Journal of Richards C, Ashburn A, et al (2004) Effects of augmented Physiotherapy 56: 97–103. exercise therapy time after stroke: a meta-analysis. Stroke 35: 2529–2536. Dean CM, Mackey FH (1992) Motor assessment scale scores as a measure of rehabilitation outcome following stroke. Langhammer B, Stanghelle JK, Lindmark B (2009) An Australian Journal of Physiotherapy 38: 31–35. evaluation of two different exercise regimes during the first year following stroke: a randomised controlled trial. Du JB, Song WQ, Wang MB (2006) The application of Physiotherapy Theory and Practice 25: 55–68. partial body weight support treadmill training in hemiplegia rehabilitation after stroke. China Journal of Cerebrovascular Langhorne P, Coupar F, Pollock A (2009) Motor recovery after Diseases 3: 361–364. stroke: a systematic review. The Lancet Neurology 8: 741– 754. Eng JJ, Tang PF (2007) Gait training strategies to optimize walking ability in people with stroke: A synthesis of the Liang SQ, Peng XS, Yang JH (2005) Two-stage rehabilitative evidence. Expert Review of Neurotherapeutics 7: 1417–1436. training in the recovery of walking ability in stroke patients with hemiplegia. Chinese Journal of Clinical Rehabilitation Guyatt GH, Pugsley SO, Sullivan MJ, Thompson PJ, Berman 9: 8–9. L, Jones NL (1984) Effect of encouragement on walking test performance. Thorax 39: 818–822. Lord S, McPherson KM, McNaughton HK, Rochester L, Weatherall M (2008) How feasible is the attainment of Hakgüder A (2007) Bodyweight supported treadmill training in community ambulation after stroke? A pilot randomized stroke rehabilitation. Turkiye Fiziksel Tip ve Rehabilitasyon controlled trial to evaluate community-based physiotherapy Dergisi 53: 41–44. in subacute stroke. Clinical Rehabilitation 22: 215–225. Hesse S (1998) Treadmill training with partial body weight Marincek C, Bajd T (2007) Surface FES-assisted walking. support for the restoration of gait of hemiparetic subjects. Critical Reviews in Physical Rehabilitation Medicine 19: 275– Neurologie und Rehabilitation 4: 113–118. 294. Hesse S, Eich HJ, Mach H, Parchmann H, Werner C (2005) Mayr A, Kofler M, Quirbach E, Matzak H, Fröhlich K, Saltuari L Aerobic treadmill training plus physiotherapy improves (2007) Prospective, blinded, randomized crossover study of walking speed and capacity in subacute moderately affected gait rehabilitation in stroke patients using the Lokomat gait patients after stroke. Neurologie und Rehabilitation 11: 7–12. orthosis. Neurorehabilitation and Neural Repair 21: 307–314. Hesse S, Schmidt H, Werner C (2006) Machines to support McCain KJ, Pollo FE, Baum BS, Coleman SC, Baker S, motor rehabilitation after stroke: 10 years of experience in Smith PS (2008) Locomotor treadmill training with partial Berlin. Journal of Rehabilitation Research & Development body-weight support before overground gait in adults with 43: 671–678. acute stroke: a pilot study. Archives of Physical Medicine Rehabilitation 89: 684–691. Hesse S (2006b) Gait training after stroke: a critical reprisal. Annales de Réadaptation et de Médecine Physique 49: 621– Mehrholz J, Werner C, Kugler J, Pohl M (2010) 624. Electromechanical-assisted training for walking after stroke Cochrane Database of Systematic Reviews 4. Hesse S, Werner C, Schmidt H, Bardeleben A (2006c) Automated motor rehabilitation after stroke. Nervenheilkunde Moseley AM, Stark A, Cameron ID, Pollock A (2005) Treadmill 25: 137–141. training and body weight support for walking after stroke. Cochrane Database of Systematic Reviews 4. Hesse S (2008) Treadmill training with partial body weight support after stroke: A review. NeuroRehabilitation 23: 55– Ng MFW, Tong RKY, Li LSW (2008) A pilot study of randomized 65. clinical controlled trial of gait training in subacute stroke patients with partial body-weight support electromechanical Hidler J, Nichols D, Pelliccio M, Brady K, Campbell DD, gait trainer and functional electrical stimulation: Six-month Kahn JH, et al (2009) Multicenter randomized clinical trial follow-up. Stroke 39: 154–160. evaluating the effectiveness of the Lokomat in subacute stroke. Neurorehabilitation and Neural Repair 23: 5–13. Nilsson L, Carlsson J, Danielsson A, Fugl-Meyer A, Hellström K, Kristensen L, et al (2001) Walking training of patients with Holden MK, Gill KM, Magliozzi MR, Nathan J, Peihl–Baker hemiparesis at an early stage after stroke: a comparison of L (1984) Clinical gait assessment in the neurologically walking training on a treadmill with body weight support and impaired: reliability and meaningfulness. Physical Therapy walking training on the ground. Clinical Rehabilitation 15: 6: 35–40. 515–527. Hornby TG, Campbell DD, Kahn JH, Demott T, Moore JL, Peurala SH, Tarkka IM, Pitkänen K, Sivenius J (2005) The Roth HR (2008) Enhanced gait-related improvements after effectiveness of body weight–supported gait training and therapist- versus robotic-assisted locomotor training in floor walking in patients with chronic stroke. Archives of subjects with chronic stroke: a randomized controlled study. Physical Medicine and Rehabilitation 86: 1557–1564. Stroke 39: 1786–1792. Peurala SH, Airaksinen O, Jäkälä P, Tarkka IM, Sivenius J Husemann B, Muller F, Krewer C, Heller S, Koenig E (2007) (2007) Effects of intensive gait–oriented physiotherapy Effects of locomotion training with assistance of a robot- during early acute phase of stroke. Journal of Rehabilitation driven gait orthosis in hemiparetic patients after stroke: a Research and Development 44: 637–648. randomized controlled pilot study. Stroke 38: 349–354. Peurala SH, Airaksinen O, Huuskonen P, Jäkälä P, Juhakoski Ivey FM, Hafer-Macko CE, Macko RF (2008) Task-oriented M, Sandell K, et al (2009) Effects of intensive therapy using treadmill exercise training in chronic hemiparetic stroke. gait trainer or floor walking exercises early after stroke. Journal of Rehabilitation Research and Development 45: Journal of Rehabilitation Medicine 41: 166–173. 249–259. Pohl M, Werner C, Holzgraefe M, Kroczek G, Wingerdorf Keith RA, Granger CV, Hamilton BB, Sherwin FF (1987) I, Hoölig G, et al (2007) Repetitive locomotor training and The functional independence measure: a new tool for physiotherapy improve walking and basic activities of daily rehabilitation. In: Eisenberg MG, Grzesiak RC (Eds) living after stroke: a single-blind, randomised multicentre Advances in Clinical Rehabilitation. New York: Springer trial (DEutsche GAngtrainerStudie, DEGAS). Clinical Verlag, pp. 6–18. Rehabilitation 21: 17–27. Kosak MC, Reding MJ (2000) Comparison of partial body Richards CL, Malouin F, Wood-Dauphinee S, Williams JI, weight-supported treadmill gait training versus aggressive Bouchard JP, Brunet D (1993) Task-specific physical therapy bracing assisted walking post stroke. Neurorehabilitation and for optimization of gait recovery in acute stroke patients. Neural Repair 14: 13–19. 160 Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010

Archives of Physical Medicine and Rehabilitation 74: 612– Ada et al: Mechanically assisted walking in stroke patients 620. randomized controlled trial. Archives of Physical Medicine Scheidtmann K, Brunner H, Muller F, Weinandy-Trapp M, Wulf and Rehabilitation 87: 1298–1304. D, Koenig E (1999) Treadmill training in early stroke patients – do timing and walking ability matter? Neurorehabilitation Tong RKY, Ng MFW, Li LSW, So EFM (2006b) Gait training of and Neural Repair 5:198–202. patients after stroke using an electromechanical gait trainer combined with simultaneous functional electrical stimulation. Schmidt H, Werner C, Bernhardt R, Hesse S, Krüger J (2007) Physical Therapy 86: 1282–1294. Gait rehabilitation machines based on programmable footplates. Journal of NeuroEngineering and Rehabilitation van de Port IGL, Wood-Dauphinee S, Lindeman E, Kwakkel 4: 2. G (2007) Effects of exercise training programs on walking competency after stroke: a systematic review. American Schwartz I, Sajin A, Fisher I, Neeb M, Shochina M, Katz-Leurer Journal of Physical Medicine and Rehabilitation 86: 935–951. M, et al (2009) The effectiveness of locomotor therapy using robotic-assisted gait training in subacute stroke patients: Wade DT, Wood VA, Heller A, Maggs J, Langton Hewer R a randomized controlled trial. Physical Medicine and (1987) Walking after stroke. Measurement and recovery over Rehabilitation 1: 516–523. the first 3 months. Scandinavian Journal of Rehabilitation Medicine 19: 25–30. Tilson JK, Sullivan KJ, Cen SY, Rose DK, Koridia CH, Azen SP, et al (2010) Meaningful gait speed improvement during first Werner C, Pohl M, Holzgraefe M, Kroczek G, Mehrholz J, 60 days poststroke: minimally clinically important difference. Wingendorf I, Hölig G, et al (2006) Locomotor training in Physical Therapy 90: 196–208. subacute stroke patients: results of a multicenter study (DEGAS). Neurologie & Rehabilitation 12: 262–269. Tong RK, Ng MF, Li LS (2006) Effectiveness of gait training using an electromechanical gait trainer, with and without Wilson DJ (2007) Principles of gait rehabilitation and the functional electric stimulation, in subacute stroke: a efficacy of partial body-weight-supported training. Critical Reviews in Physical Rehabilitation Medicine 19: 169–194. Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010 161

Ng et al: Exercise training and Type 2 diabetes Minimal difference between aerobic and progressive resistance exercise on metabolic profile and fitness in older adults with diabetes mellitus: a randomised trial Cindy LW Ng1, Su-Yen Goh1, Rahul Malhotra2, Truls Østbye2 and E Shyong Tai1,2 1Singapore General Hospital, 2Duke-NUS Graduate Medical School Singapore Question: Is progressive resistance training as effective as aerobic training of similar duration in sedentary older adults with diabetes mellitus? Design: A randomised trial with concealed allocation, assessor blinding and intention-to-treat analysis. 1BSUJDJQBOUT Sixty people with Type 2 diabetes mellitus with glycosylated haemoglobin (HbA1c) between 8% and 10% in the past month. Intervention: One group undertook progressive resistance exercise and the other group undertook aerobic exercise. Both groups completed 18 sessions over 8 weeks. In each session, the progressive resistance exercise group did nine resistive exercises while the aerobic exercise group did 50 minutes of aerobic exercise. Outcome measures: HbA1c, blood glucose, lipid profile (total, high- and low-density cholesterol and triglycerides), weight, body mass index, body fat, waist circumference, waist:hip ratio, blood pressure, and peak oxygen consumption. Results: Forty-nine (82%) participants completed the intervention. HbA1c reduced by a similar amount in both groups (MD 0.1%, 95% CI –0.3 to 0.5). However, significant between-group differences occurred in change in waist circumference in favour of progressive resistance exercise C:¸'$.Yc\"/+9?¸&$+je¸)$'\"WdZ_dY^Wd][_df[Waeno][dYedikcfj_ed_d\\Wlekhe\\W[heX_Y[n[hY_i[C:+$(cb%a]\" 95% CI 0.0 to 10.4). Conclusions: Progressive resistance exercise has similar effects to aerobic exercise and therefore offers a useful alternative for patients unable to participate in aerobic exercise. Trial registration: NCT01000519. </H CLW, Goh 4:
.BMIPUSB
3
TUCZF
5
5BJ
&4

.JOJNBM
EJGGFSFODF
CFUXFFO
BFSPCJD
BOE
QSPHSFTTJWF
SFTJTUBODF
FYFSDJTF
PO
NFUBCPMJD
QSPmMF
BOE
mUOFTT
JO
PMEFS
BEVMUT
XJUI
EJBCFUFT
NFMMJUVT
B
SBOEPNJTFE
USJBM
Journal of Physiotherapy

o> ,FZ
XPSET: Diabetes mellitus, Type 2; Exercise; Hemoglobin A, glycosylated Introduction et al 2004, Sigal et al 2007, Snowling and Hopkins 2006, Tresierras and Balady 2009). More than 100 million people in Asia were living with diabetes mellitus in 2007 (Chan et al 2009). In Singapore, In spite of the reported beneficial effects of aerobic exercise the ageing of the population together with the rise in rates on cardiovascular and metabolic parameters, adoption of of obesity and sedentary lifestyle parallelled the rise of aerobic activities may be difficult for some patients with Type 2 diabetes mellitus. The prevalence of Type 2 diabetes Type 2 diabetes mellitus, especially those who are older and mellitus in 2004 was 8.2% in adults aged 18 to 69 years obese (Willey and Singh 2003). In the last decade, there has (Lim et al 2004). Diabetes doubles the risk of cardiovascular been increasing interest in the role of resistance exercise disease (Wang et al 2005) and, in Singapore, one-third of in the management of diabetes as it appears to improve patients developing cardiovascular disease were reported to insulin sensitivity (Tresierras and Balady 2009). While have underlying Type 2 diabetes mellitus (Lee et al 2001). the American College of Sports Medicine recommended Singaporeans have a higher percentage of body fat for the resistance exercise at least twice a week (Albright et al same body mass index as Caucasians (Deurenberg- ap 2000), the American Diabetes Association recommended et al 2003), and those with Type 2 diabetes mellitus have it three times per week. These recommendations were significantly higher body mass index and waist:hip ratio based primarily on findings from two trials comparing compared with healthy adults (Lim et al 2004). aerobic and resistance exercise (Cauza et al 2005, Dunstan et al 2002). However, neither study attempted to make the Aerobic exercise is known to reduce weight and maintain modes of exercise comparable in intensity or duration. good glycaemic control, and thus reduce the risk of Furthermore, some studies have included both modes in cardiovascular disease among diabetic patients (Lee the same intervention arm (Cuff et al 2003, Maiorana et al et al 2001). Studies involving exercise as a therapeutic 2000), thus limiting our ability to compare the two. Other intervention in patients with Type 2 diabetes mellitus data suggest that progressive resistance exercise has benefits have focused primarily on aerobic training (Boule et in the treatment of Type 2 diabetes (Neil and Ronald 2006, al 2003, Snowling and Hopkins 2006). The beneficial Irvine and Taylor 2009). However, two studies showed that effects of aerobic training on the metabolic profile include a home-based resistance exercise program had no effect on reduced HbA1c, lowered blood pressure and resting heart glycaemic control (Dunstan et al 2005, Cheung et al 2009). rate, improved cardiac output and oxygen extraction, Therefore, there is a need to further study the relative favorable lipid profile, and reduction of weight and waist benefits of aerobic exercise and progressive resistance circumference (Albright et al 2000, Boule et al 2001, Lim exercise in patients with Type 2 diabetes mellitus. Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010 163

Research The research question for this study was: assessing the number of repetitions that the participant was Is progressive resistance training as effective as able to complete at submaximal loads. The progressive aerobic training of similar intensity and duration in resistance exercise intervention is presented in Table 1. terms of glycaemic, metabolic, anthropometric, and cardiovascular variables in sedentary older adults with #PY
. Resistance exercises. Type 2 diabetes mellitus? Muscle group Description Method Quadriceps Seated leg press: Seated upright with Design feet onto a plate, the participant pushed against the load extending and flexing A randomised trial was conducted with participants the knee. recruited from the Diabetes Centre of Singapore General Hospital. After baseline measurements of glycaemic, Straight leg raise: Lying on the back metabolic, anthropometric, and cardiovascular profile with one leg bent and one leg straight were taken, participants were randomised to either an with the pelvis posteriorly tilted, the experimental (progressive resistance exercise) or a control participant lifted the straightened leg up (aerobic exercise) group, based on a computer-generated to approximately 45 degrees and slowly assignment schedule that was kept by a physician not lowered it back to the plinth. involved in the selection of the participants. Allocation was concealed by investigators making telephone contact with Hamstrings Hamstrings curl machine: Lying prone the physician who was the only person with access to the with hips flush against the bench, the calf assigned schedule. All outcome measures were taken at the was placed under the roller and the leg end of the 8-week intervention period by an independent curled the weight up to 90 degrees from assessor who was blinded to group allocation. Outcomes the machine and was then lowered down were measured between 36 and 48 hours after the last slowly. exercise session. All participants were specifically told not to discuss any aspect of their training with the assessor. The Biceps Biceps curls: The participant held the templates developed by the Research on Research group dumb-bells with palms faced out, elbows were used to facilitate communication with the statistician next to the body and curled the weights regarding data analysis and in the writing of the manuscript towards the shoulders and then lowered (Pietrobon et al 2004, Shah et al 2009). them slowly. Participants Triceps Triceps curls: Arms were raised straight overhead while keeping them close to Patients were included if they were aged 50 years or above, the ears and elbows bent, lowering the had glycosylated haemoglobin (HbA1c) levels between dumb-bells behind the participant’s head. 8% and 10% in the past month, and were able to walk The elbows were straightened to raise the continuously for at least 20 min and climb one flight of weights and bent to lower them again. stairs unaided without stopping. They were also required to be sedentary, defined as reporting never having participated Deltoids Lateral raises (middle deltoids): The in a structured exercise program or recreational physical dumb-bells were held in front of the hips activity or sport. Subjects were excluded if they had: with palms facing each other and elbows uncontrolled diabetes mellitus with HbA1c more than slightly bent. The weights were then 10% or if escalation of treatment of glycaemic control or raised out to the sides and upwards in a dyslipidaemia was likely to be necessary over the 8-week semi-circular manner to just above the trial period; congestive cardiac failure, unstable angina, or shoulder level and then lowered slowly. acute myocardial infarction within the last year; proliferative diabetic retinopathy; uncontrolled hypertension; advanced Front raises (anterior deltoids): The arthritis likely to limit mobility or participation in dumb-bells were held in front on the body prescribed exercises; respiratory co-morbidities; significant with palms facing each other and elbows proteinuria or chronic renal insufficiency; been prescribed slightly bent. The weights were then a very low caloric diet (less than 1000 kcal/day) or drugs raised out to the front and upwards in a for the treatment of obesity; renal disease; or inability semi-circular manner to just above the to monitor glucose level or to comply with the exercise shoulder level and then lowered slowly. program. Recruited participants gave informed consent and their physicians were informed about the research protocol Gluteus Hip abduction: The outside of the thigh and that there should be no escalation of medical treatment was placed against the roller pad and during the 8-week intervention period. raised against the roller pad to the side and returned to initial position while body Intervention weight was on the other leg. The experimental group (progressive resistance exercise) Hip extension: The back of the thigh undertook nine resistive exercises using a combination was placed against the roller pad and of machines and free weights (Box 1) at 65% of their raised against the roller pad to the back assessed one repetition maximum (1RM) as recommended by extending hip and straightening leg by American College of Sports Medicine (Ratamess et al and returned to initial position while body 2009). The 1RM for each muscle group was determined weight was on the other leg. using a prediction formula (Brown and Weir 2001) by The control group (aerobic exercise) underwent 50 minutes of aerobic training involving treadmill, elliptical cycle, and stationary bicycle exercise, with a target heart rate of 65% of their age-predicted maximum heart rate, after being assessed by a submaximal treadmill test according to American College of Sports Medicine guidelines. Both groups were progressed after 4 weeks of training to 70% of their predicted 1RM or age-predicted heart rate depending on grouping. The metabolic equivalents (METs) for both the aerobic exercise and progressive resistance 164 Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010

Ng et al: Exercise training and Type 2 diabetes 5BCMF
. Details of the aerobic exercise and progressive resistance exercise interventions. Progressive resistance exercise Aerobic exercise Intensity 65–70% of 1RM 65–70% maximum heart rate as determined by the Duration modified Bruce protocol test Jof[%CeZ[ One set of 10 repetitions for each of the 9 50 minutes (10 minutes on upright or recumbent bicycle resistive exercises completed in a circuit and 20 minutes each on the treadmill and elliptical Before exercise cycle) Warm up 3 rounds of the circuit were completed in Program a maximum of 50 minutes Frequency Quadriceps (seated leg press machine and Treadmill, stationary upright bicycle, stationary straight leg raises) recumbent bicycle, cross trainer (elliptical cycle) Hamstrings (hamstring curls machine) Biceps, triceps, anterior and middle deltoids (using free weights) Hip abductors and extensors (gluteal machine) Both interventions Heart rate, blood pressure and glucose level (pin prick) 3 min of stretches of quadriceps, hamstrings, calf, biceps, triceps and back, with each muscle group stretched twice, holding each stretch for 15 s 10 min of unloaded cycling 8-week group exercise program at Singapore General Hospital Physiotherapy Gym 2 to 3 times a week, Mondays, Wednesdays and Fridays exercise training were estimated to be approximately 3.5 for cholesterol is 1.9% at 3.22 mmol/L and 1.3% at 7.72 in accordance with the compendium of METs provided by mmol/L. The CV for triglyceride is 1.8% at 1.02 mmol/L the American College of Sports Medicine (ACSM 2000), a and 1.4% at 2.27 mmol/L. HDL-C was measured using value defined as moderate intensity (Pate et al 1995). The homogenous enzymatic colorimetric assay with a CV of 4.8 aerobic exercise intervention is presented in Table 1. % at 0.93 mmol/L and 3.7 % at 2.06 mmol/L. LDL-C was calculated using the Friedewald formula. All participants wore a heart rate monitor during the warm- up and exercise program and were supervised in their Anthropometric measurements included weight, body mass exercises in a group. Each participant was scheduled to index, body fat measured by skin fold and by bioimpedance, complete 18 exercise sessions over 8 weeks at a frequency waist circumference and waist:hip ratio. Body mass index of 2 to 3 times a week. was calculated as weight in kg divided by the square of height in m. Skin-fold thickness was measured at four sites: Outcomes measures biceps, triceps, sub-scapular, and suprailiac, on the right side of the body (Heyward 2002), and percentage body fat The primary outcome measure was HbA1c. Secondary was estimated using a formula applicable to Singaporeans outcomes included blood glucose, lipid profile, and (Deurenberg- ap et al 2003). Percentage body fat was also anthropometric and cardiovascular measures. Adverse measured using two-point bioimpedance analysisa and events were also recorded. All outcome assessors were a regression equation based on measured resistance and blinded to group allocation. reactance. Waist circumference was measured with a tape placed horizontally at the mid-point between the iliac crest HbA1c was measured using 10 ml of blood drawn from and the lower aspect of the floating ribs in the mid axillary participants who fasted at least 10 h from the night before and line, at the end of normal expiration. Hip circumference analysed at the Biochemistry Laboratory of the Pathology was measured at the mid point of the gluteal region. Department in Singapore General Hospital by laboratory assistants who were also blinded to the project. HbA1c was Cardiovascular measures included peak oxygen measured using high performance liquid chromatography consumption and resting blood pressure. Peak oxygen with a coefficient of variation (CV) of 2.4% at 5.1% (HbA1c) consumption was measured during a submaximal exercise and a CV of 1.9% at 9.6% (HbA1c). test using a Modified Bruce protocol (ACSM 2000) with 12-lead electrocardiogram and with monitoring of blood Glucose was measured using the glucose oxidase method pressure. The treadmill test was terminated if the participant with a CV of 1.6% at 3.3 mmol/L and a CV of 1.1% at 18.8 (i) reached his or her peak oxygen consumption or predicted mmol/L. maximum heart rate, (ii) indicated that he or she could not continue the testing, (iii) had systolic blood pressure above The lipid profile comprised total cholesterol, triglycerides, 220 mmHg or diastolic blood pressure above 100 mmHg, high-density lipoprotein cholesterol (HDL-C), and low- or (iv) developed abnormal electrocardiographic changes. density lipoprotein cholesterol (LDL-C). Total cholesterol and triglycerides were measured using enzymatic Data analysis colorimetric methods with cholesterol oxidase-peroxidase amino phenazone phenol and glycerol-3-phosphate For sample size calculation, we adopted a 1% difference oxidase-peroxidase amino phenazone phenol. The CV in HbA1c as clinically worthwhile because an increase of Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010 165

Research 5BCMF
 Baseline characteristics of participants, therapists and centres. Characteristic Participants Randomised Lost to follow-up (n = 60) (n = 11) Exp Con Exp Con (n = 30) (n = 30) (n = 5) (n = 6) Participants 57 (7) 59 (7) 56 (7) 56 (7) Age (yr), mean (SD) 11 (37) 8 (27) 3 (60) 2 (33) Gender, n male (%) 22 (73) 17 (57) 1 (20) 1 (17) Chinese ethnicity, n (%) Highest education level, n (%) 8 (27) 9 (30) 1 (20) 1 (17) Primary 15 (50) 16 (53) 4 (80) 4 (67) Secondary 7 (23) 5 (17) 0 (0) 1 (17) Tertiary 11 (9) 12 (9) 7 (9) 9 (4) Duration of diabetes (yr), mean (SD) Medication*, n (%) 15 (50) 18 (60) 0 (0) 0 (0) Sulphonylurea 24 (80) 25 (83) 3 (60) 6 (100) Metformin 10 (33) 8 (27) 0 (0) Glitazone 5 (17) 7 (23) 0 (0) 0 (0) Alpha glucosidase 8 (27) 8 (27) 2 (40) 0 (0) Combination of oral hypoglycaemic 0 (0) agent and insulin 15 (50) 15 (50) 2 (40) 15 (50) 15 (50) 3 (60) 3 (50) Therapists 3 (50) Participants treated, n (%) Therapist 1 Therapist 2 Exp = experimental group, Con = control group, * = not mutually exclusive 1% is associated with an 18% increase in the relative risk through the trial and reasons for exclusion are presented in of cardiovascular disease in patients with Type 2 diabetes Figure 1. The baseline characteristics of the participants mellitus (Selvin et al 2004). Most studies in the systematic who completed the study and those lost to follow-up are review by Irvine and Taylor (2009) reported a standard presented in Table 2. Both groups were comparable and the deviation of HbA1c between 1.0% and 1.7%. Therefore, participants lost to follow-up were comparable to those who we anticipated a standard deviation of 1.35%. A total of completed the study. 30 patients per group would provide an 80% probability of detecting a difference of 1% in HbA1c at a two-sided 5% Two physiotherapists with 3 years experience supervised significance level, assuming a standard deviation of 1.35%. the exercise sessions at the Physiotherapy Outpatient Therefore we sought to recruit 60 participants. Department in Singapore General Hospital. All participants with follow-up data were analysed $PNQMJBODF
XJUI
USJBM
NFUIPE according to their group allocation, ie, using an intention- to-treat analysis. Baseline values of the various outcome Forty-nine (82%) participants completed all 18 sessions parameters were carried forward for the 11 participants within 8 weeks. Eleven participants (5 in the progressive who dropped out during the intervention. The difference resistance exercise group and 6 in the aerobic exercise in change from baseline to post-intervention between the group) failed to attend for the full exercise program and aerobic exercise and progressive resistance exercise groups declined to attend for further measurement. No changes for each outcome was assessed using an independent t-test. in medication were prescribed for the study participants Statistical significance was set at p 0.05, so results are during the intervention period. presented as a mean difference (95% CI). Effect of intervention Results Group data for all outcomes are presented in Table 3. 'MPX
PG
QBSUJDJQBOUT
BOE
UIFSBQJTUT
UISPVHI
UIF
Individual data are presented in Table 4 (see eAddenda for trial Table 4). The change in HbA1c was similar in both groups. It reduced by 0.4% (SD 0.6) in the progressive resistance Five hundred and thirty patients diagnosed with Type 2 exercise group and by 0.3 % (SD 0.9) in the aerobic exercise diabetes mellitus attending the Diabetes Centre at Singapore group, which was not a statistically significant difference General Hospital were screened for eligibility between (MD –0.1%, 95% CI –0.5 to 0.3). October 2003 and October 2004. Sixty-eight patients met the eligibility criteria, of whom 60 patients gave informed Three of the secondary outcomes had significant consent to participate in the study and were randomised, with between-group differences: waist circumference, peak 30 being allocated to each group. The flow of participants oxygen consumption, and resting systolic blood pressure. The between-group difference in the change in waist 166 Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010

Ng et al: Exercise training and Type 2 diabetes Patients with Type 2 diabetes screened (n = 530) Excluded (n = 462) š HbA1c < 8% or > 10% (n = 320) š inability to walk independently (n = 142) Met inclusion criteria (n = 68) Excluded (n = 8) š refused to participate (n = 6) š other personal reasons (n = 2) Measured HbA1c, blood glucose, lipid profile, anthropometry and cardiovascular fitness Week 0 Randomised (n = 60) (n = 30) (n = 30) Lost to follow-up (n = 5) Experimental group Control group Lost to follow-up (n = 6) š declined to attend š progressive š aerobic š declined to attend testing session resistance exercise testing session (n = 6) (n = 5) exercise š 8-wk program š 8-wk program š 18 x 50-min š 18 x 50-min sessions sessions Week 8 Measured HbA1c, blood glucose, lipid profile, anthropometry and cardiovascular fitness (n = 25) (n = 24) 'JHVSF
 Design and flow of participants through the trial. circumference favoured the progressive resistance group study had sufficient power to exclude clinically worthwhile (MD –1.8 cm, 95% CI –0.5 to –3.1). The between-group differences between the therapies on the primary outcome. difference in the change in peak oxygen consumption Because very few significant between-group differences favoured the aerobic group, improving by a mean of 5.2 were identified and the confidence intervals around ml/kg (95% CI 0.0 to 10.4) more than in the progressive the between-group differences were generally narrow, resistance exercise group. The reduction in resting systolic progressive resistance exercise is likely to be a similarly blood pressure was significantly greater in the aerobic effective alternative to aerobic exercise. exercise group than in the progressive resistance exercise group (MD 9 mmHg, 95% CI 2 to 16). Two previous randomised trials comparing progressive resistance exercise and aerobic exercise reported better Discussion improvement in HbA1c with resistance exercise (Arora et al 2009, Cauza et al 2005). However, one trial did not describe Comparison of the two modes of exercise was the primary the training programs in terms of intensity or volume (Cauza aim of the study, so the exercise regimens were matched as et al 2005), so it is difficult to determine the source of the closely as possible for frequency, intensity, duration, and rate between-group differences. The other trial had a small of progression. Because all participants in both groups who sample size (n = 10) in each arm and a wide (5% to 10%) attended the exercise sessions were able to cope with the baseline HbA1c (Arora et al 2009), so the current trial may prescribed regimen, this strengthens the interpretation that provide more robust data. A recent meta-analysis showed between-group differences did reflect the relative effects that the patients randomised to resistance exercise tend of the two exercise modes. Furthermore, although there to spend more time on training than the aerobic exercise were some dropouts, the resulting reduction in statistical group (Gordon et al 2009). A systematic review showed power was offset by the smaller than anticipated standard that resistance exercise alone reduced HbA1c by 0.3% but deviation in HbA1c in our cohort, at 1.21%. Therefore the was not significantly different when compared to aerobic Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010 167

Research 5BCMF
. Mean (SD) of groups, mean (SD) difference within groups, and mean (95% CI) difference between groups. Outcome Groups Difference within groups Difference between groups Week 0 Week 8 Week 8 minus Week 0 Week 8 minus Week 0 Exp Con Exp Con Exp Con Exp minus Con (n = 30) (n = 30) (n = 30) (n = 30) HbA1c (%) 8.9 8.5 8.4 8.1 –0.4 –0.3 –0.1 (1.5) (0.9) (1.2) (1.1) (0.6) (0.9) (–0.5 to 0.3) Blood glucose 10.4 9.5 10.1 9.3 –0.3 –0.2 (mmol/L) (3.1) (2.5) (3.6) (2.2) (2.8) (1.7) –0.1 (–1.3 to 1.1) Lipid profile 5.4 5.1 5.5 5.1 0.0 0.0 (1.6) (0.9) (1.6) (1.1) (0.8) (0.6) 0.0 Total cholesterol 3.1 1.8 3.5 1.8 0.4 0.0 (–0.3 to 0.4) (mmol/L) (3.6) (0.8) (6.5) (0.8) (3.2) (0.4) 1.3 1.4 1.4 1.4 0.1 0.0 0.4 Triglycerides (0.4) (0.4) (0.4) (0.4) (0.2) (0.1) (–0.8 to 1.6) (mmol/L) 3.1 2.8 3.0 2.8 –0.1 –0.0 0.1 High density (1.5) (0.9) (1.4) (1.0) (0.5) (0.6) (0.0 to 0.2) lipoprotein (mmol/L) 69.5 70.3 69.7 70.2 0.2 0.0 0.0 (14.2) (13.8) (14.4) (13.6) (1.1) (1.4) (–0.3 to 0.2) Low density 27.4 27.8 27.5 27.8 0.1 0.0 lipoprotein (4.7) (5.2) (4.7) (5.2) (0.4) (0.5) 0.2 (mmol/L) 33.9 35.3 32.7 34.3 –1.3 –1.1 (–0.4 to 0.9) (7.8) (6.3) (7.4) (5.9) (2.2) (2.2) Anthropometric 33.1 33.9 31.6 32.8 –1.4 –1.1 0.1 (6.2) (5.2) (6.1) (5.3) (2.4) (2.2) (–0.2 to 0.3) Weight (kg) 90.8 91.9 89.2 92.1 –1.6 0.2 –0.2 (11.2) (11.6) (11.7) (11.0) (2.6) (2.4) (–1.3 to 0.9) BMI (kg/m2) –0.3 (–1.5 to 0.9) Percentage body fat – skinfold (%) –1.8 (–0.5 to–3.1) Percentage body fat – bioimpedance (%) Waist circumference (cm) Waist:hip ratio 0.92 0.91 0.90 0.91 –0.02 0.00 –0.02 (0.08) (0.06) (0.07) (0.05) (0.05) (0.03) (–0.04 to 0.00) Cardiovascular Peak volume of 32.8 32.3 37.3 42.0 4.5 9.8 –5.2 oxygen consumed (17.8) (15.5) (17.5) (12.8) (9.4) (10.7) (–10.4 to 0.0) (ml/kg) Resting systolic 123 133 123 123 –1 –9 9 blood pressure (12) (14) (11) (13) (12) (14) (2 to 16) (mmHg) Resting diastolic 76 78 76 74 0 –4 4 blood pressure (9) (11) (10) (11) (9) (12) (–1 to 10) (mmHg) Exp = experimental group (progressive resistance exercise), Con = control group (aerobic exercise), HbA1c = glycosylated haemoglobin, BMI = body mass index exercise (Irvine and Taylor 2009). Our study showed that, mechanisms of action of progressive resistance exercise and controlling for exercise volume, duration, and intensity, aerobic exercise may have also played a role. Progressive aerobic exercise and progressive resistance exercise had resistance exercise increases muscle strength or fat free similar improvement. The degree of change in HbA1c seen mass and mobilises visceral adipose tissue, thus enhancing in both groups in our study was similar to that seen with insulin sensitivity (Tresierras and Balady 2009). oral medications and diet (Irvine and Taylor 2009). Unfortunately, the greater reduction in waist circumference Despite similar effects on body fat percentage, progressive was not also associated with any additional benefit in terms resistance exercise resulted in a greater reduction in waist of blood pressure or lipid profile, all of which are closely circumference than aerobic exercise – a finding in line related parameters. A study on obese Japanese men with with a previous study showing that progressive resistance metabolic syndrome, which can be considered closest exercise reduced visceral and subcutaneous abdominal fat to our population, suggested that a reduction of at least (Ibanez et al 2005). The different exercise physiology and 3cm in waist circumference was required for any change 168 Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010

Ng et al: Exercise training and Type 2 diabetes in metabolic profile (Miyatake et al 2008). The average Correspondence: Ms Cindy Li Whye Ng, Department of reduction observed for the progressive resistance exercise Physiotherapy, Singapore General Hospital, Block 1 Level group in the present study was only about half of that, at 1.6 1, Outram Road, S169608, Singapore. Email: cindy.ng.l.w cm (SD 2.6). sgh.com.sg The effect of aerobic exercise on peak oxygen consumption References was significantly greater than that of progressive resistance exercise. Previous studies showed that resistance Albright A, Franz M, Hornsby G, Kriska A, Marrero D, Ullrich exercise can elicit modest improvement in peak oxygen I, Verity LS (2000) American College of Sports Medicine consumption, by approximately 6% (ACSM 1998). The Position Stand: Exercise and type 2 diabetes. Medicine and progressive resistance exercise group in our study improved Science in Sports and Exercise 32: 1345–1360. their peak oxygen consumption by approximately 14%, comparable to that observed in a previous 6-month study on American College of Sports Medicine (2000) Guidelines for progressive resistance exercise on cardiorespiratory fitness exercise testing and prescription (6th edn). USA: Lippincott in elderly men and women (Vincent et al 2003). This can M_bb_WciM_ba_di$ be attributed to increased lower limb strength (Vincent et al 2003). These improvements may be clinically important American Diabetes Association (1995–2010). ‘Living with as physical activity in patients with chronic conditions can Z_WX[j[i0 7'9$½ H[jh_[l[Z -j^ 7fh_b\" (&'&\" \\hec ^jjf0%%mmm$ reduce mortality (Martinson et al 2001, Sigal et al 2006). Z_WX[j[i$eh]%b_l_d]Um_j^UZ_WX[j[i%jh[Wjc[djUWdZUYWh[% XbeeZU]bkYei[UYedjheb%W'Y%$ The training duration of 8 weeks was brief compared to the 12-week regimens examined in earlier studies. The 8-week Arora E, Shenoy S, Sandhu JS (2009) Effects of resistance duration was chosen to minimise or avoid the influence of training on metabolic profile of adults with type 2 diabetes. any medication change during the course of the trial. HbA1c Indian Journal of Medical Research 129: 515–519. levels reflect glycaemic control over the previous 2 to 3 month period (American Diabetes Association 1995–2010), Boule NG, Kenny GP, Haddad E, Wells GA, Sigal RJ (2003) thus the observed change in HbA1c may not adequately Meta-analysis of the effect of structured exercise training reflect the effect of the interventions on glycaemic control. on cardio respiratory fitness in type 2 diabetes mellitus. The recommended frequency of 2 to 3 sessions per week Diabetologia 46: 1071–1081. was not adhered to for some participants for reasons such as public holidays, caring for family members, and feeling Boule NG, Haddad E, Kenny GP, Wells GA, Sigal RJ (2001) unwell. Nevertheless, meaningful differences in some Effects of exercise on glycemic control and body mass in parameters were demonstrated between the groups, as well type 2 diabetes mellitus. A meta-analysis of controlled as within each group, similar to those observed in other clinical trials. JAMA 286: 2941–2941. studies of longer duration. These included improvements in waist circumference and peak oxygen consumption Brown LE, Weir JP (2001) American Sport Education (Vincent et al 2003) and reduction in HbA1c (Boule et al Program (ASEP) Procedures Recommendation I: Accurate 2003, Boule et al 2001). assessment of muscular strength and power. Journal of Exercise Physiology 4: 1–21. As our inclusion criteria included a baseline HbA1c of 8% to 10%, the absence of exercise training would have Cauza E, Hanusch-Enserer U, Strasser B, Ludvik B, Metz- required an escalation of medical management. Thus, a Schimmert S, Pacini G, et al (2005) The relative benefits of non-intervention control group was excluded. Though this endurance and strength training on the metabolic factors limits our ability to assess the true benefits of exercise, it and muscle function of people with type 2 diabetes mellitus. was not the aim of the study since the benefits of exercise Archives of Physical Medicine and Rehabilitation 86: 1527– for Type 2 diabetes mellitus are well established. Q 1533. Footnotes: aBiodynamics Model 450, Biodynamic Chan JCN, Malik V, Jia W, Kadowaki T, Yajnik-Chittaranjan S, International, USA. Yoon K-H, Hu FB (2009) Diabetes in Asia. Epidemiology, risk factors, and pathophysiology. JAMA 301: 2129–2140. eAddenda: Table 4 available at www.jop.physiotherapy.asn. au Cheung N, Cinnadaio N, Russo M, Marek S (2009) A pilot randomised controlled trial of resistance exercise bands in Ethics: The study was approved by Singapore General the management of sedentary subjects with type 2 diabetes. Hospital (SGH) Institutional Review Board (IRB 253/2002). Diabetes Research and Clinical Practice 83: e68–e71. All participants provided informed consent before data collection began. Cuff DJ, Meneilly GS, Martin A, Ignaszewski A, Tildesley HD, Frohlich JJ (2003) Effective exercise modality to reduce Competing interests: Nil insulin resistance in women with type 2 diabetes. Diabetes Care 26: 2977–2982. Support: National Medical Research Council of Singapore (www.nmrc.gov.sg NMRC/0728/2003). Deurenberg-Yap M, Ng SA, Foo LL, Deurenberg P (2003) Development and validation of a prediction equation for Acknowledgements: Abbott Laboratories (Singapore) Pte. percentage body fat based on skin folds for Singaporeans Ltd. for supplying the OptiumTMglucose meter, lancets, and adults and adolescents. International Journal of Body glucose strips for daily monitoring of participants blood Composition Research 1: 103–109. glucose level. Dunstan DW, Daly RM, Owen N, Jolley D, De Courten M, Shaw J, et al (2002) High-intensity resistance training improves glycemic control in older patients with type 2 diabetes. Diabetes Care 25: 1729–1736. Dunstan DW, Daly RM, Owen N, Jolley D, Vulikh E, Shaw J, et al (2005) Home-based resistance training is not sufficient to maintain improved glycemic control following supervised training in older individuals with type 2 diabetes. Diabetes Care 28: 3. Gordon BA, Benson AC, Bird SR, Fraser SF (2009) Resistance training improves metabolic health in type 2 diabetes: A systematic review. Diabetes Research and Clinical Practice 83: 157–175. Heyward V (2002) Advanced fitness assessment and exercise prescription. USA: Human Kinetics. Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010 169

Research American College of Sports Medicine (1998) ACSM Position Stand: The recommended quantity and quality of exercise Ibanez J, Izquierdo M, Arquelles I, Forga L, Larrion JL, Garcia- for developing and maintaining cardiorespiratory and Unciti M, et al (2005) Twice-weekly progressive resistance muscular fitness, and flexibility in healthy adults. Medicine training decreases abdominal fat and improves insulin and Science in Sports and Exercise 30: 975–991. sensitivity in older men with type 2 diabetes. Diabetes Care 28: 662–667. American College of Sports Medicine (2009) ACSM Position Stand Progression models in resistance training for healthy Irvine C, Taylor N (2009) Progressive resistance exercise adults. Medicine and Science in Sports and Exercise 41: improves glycaemic control in people with type 2 diabetes 687–708. mellitus: a systematic review. Australian Journal of Physiotherapy 55: 237–246. Selvin E, Marinopoulos S, Berkenblit G, Rami T, Brancati FL, Powe NR, et al (2004) Meta-analysis: Glycosylated Lee J, Heng D, Chia KS, Chew SK, Tan BY, Hughes K (2001) hemoglobin and cardiovascular disease in diabetes mellitus. Risk factors and incident coronary heart disease in Chinese, Annals of Internal Medicine 141: 421–431. Malay and Asian Indian males: The Singapore Cardiovascular Cohort Study. International Journal of Epidemiology 30: Shah J, Shah A, Pietrobon R (2009) Scientific writing of novice 983–988. researchers: what difficulties and encouragements do they encounter? Academic Medicine 84: 511–516. Lim JG, Kang HJ, Stewart KJ (2004) Type 2 diabetes in Singapore: The role of exercise training for its prevention and Sigal RJ, Kenny GP, Wasserman DH, Castaneda-Sceppa management. Singapore Medical Journal 45: 62–68. 9\" M^_j[ H: (&&, F^oi_YWb WYj_l_jo%[n[hY_i[ WdZ jof[ 2 diabetes: A consensus statement from the American Maiorana A, O’Driscoll G, Cheetham C, Collis J, Goodman Diabetes Association. Diabetes Care 29: 1433–1438. C, Rankin S, et al (2000) Combined aerobic and resistance exercise improves glycemic control and fitness in type 2 Sigal RJ, Kenny GP, Boule NG, Wells GA, Prud’homme D, diabetes. Diabetes Research and Clinical Practice 56: 115– Fortier M, et al (2007) Effects of aerobic training, resistance 123. training, or both on glycemic control in type 2 diabetes. Annals of Internal Medicine 147: 357–369. Martinson BC, O’Conner PJ, Pronk NP (2001) Physical inactivity and short-term all-cause mortality in adults with chronic Snowling NJ, Hopkins WG (2006) Effects of different modes disease. Archives of Internal Medicine 161: 1173–1180. of exercise training on glucose control and risk factors for complications in type 2 diabetes patients: A meta-analysis. Miyatake N, Matsumoto S, Fujii M, Numata T (2008) Reducing Diabetes Care 29: 2518–2527. waist circumference by at least 3 cm is recommended for improving metabolic syndrome in obese Japanese men. Tresierras MA, Balady GJ (2009) Resistance training in Diabetes Research and Clinical Practice 79: 191–195. the treatment of diabetes amd obesity–Mechanisms and outcomes. Journal of Cardiovascular Rehabilitation and Neil DE, Ronald CP (2006) Resistance training and Type Prevention 29: 67–75. 2 diabetes: Considerations for implementation at the population level. Diabetes Care 29: 1933. Vincent KR, Vincent HK, Braith RW, Bhatnagar V, Lowenthal DT (2003) Strength training and hemodynamic response to Pate RR, Pratt M, Blair SN, Haskell WL, Macera CA, Bouchard exercise. American Journal of Geriatric Cardiology 12: 97– C, et al (1995) Physical activity and public health. A 106. recommendation from the centers for Disease Control and Prevention and the American College of Sports Medicine. Wang J, Armour T, Geiss LS, Engelgau MM (2005) Obesity JAMA 273: 402–407. and diabetes: dual epidemics on the rise. Current Opinion of Endocrinology and Diabetes 12: 174–180. Pietrobon R, Guller U, Martins H, Menezes AP, Jacobs DO, Higgins LD (2004) A suite of web applications to streamline Willey KA, Singh MAF (2003) Battling insulin resistance in the interdisciplinary collaboration in secondary data elderly obese people with type 2 diabetes. Diabetes Care analyses. BMC Medical Research Methodology 4: 29. 26: 1580–1588. 170 Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010

Blennerhassett et al: Risk factors for shoulder pain after stroke Reduced active control and passive range at the shoulder increase risk of shoulder pain during inpatient rehabilitation post-stroke: an observational study Jannette M Blennerhassett, Karen Gyngell and Rachael Crean Austin Health: Royal Talbot Rehabilitation Centre, Australia Question: What factors at admission are associated with shoulder pain during stroke rehabilitation? Design: Retrospective audit of medical histories and logistic regression. Participants: 94 people with primary diagnosis of stroke attending inpatient rehabilitation. Outcome measures: Predictors were a battery of impairments, stroke-related factors, and patient characteristics. The outcome of interest was shoulder pain. Results: Shoulder pain was present in 23% of patients at admission, and in a total of 35% of patients during inpatient stay. Patients with pain differed significantly (p ≤ 0.04) from those without pain for several factors including age, longer time until rehabilitation admission, impaired movement of the arm (Motor Assessment Scale items), reduced passive range of movement, subluxation, and altered tone and sensation. No differences were found for many factors including neglect, cognitive impairment, side of stroke, and body weight. Logistic regression exploring the association between four predictors (shoulder range, Motor Assessment Scale items, subluxation, and altered sensation) and shoulder pain (outcome of interest) found that shoulder pain was reliably associated with two factors: reduced passive shoulder range (OR 14%, 95% CI 3 to 64), and Motor Assessment Scale Upper Arm item score (OR 64%, 95% CI 43 to 96). The model accurately classified 85% of patients. Conclusion: Shoulder pain is common and occurs early after stroke. Reduced active control and passive range at the shoulder appear to be risk factors for shoulder pain during inpatient rehabilitation post-stroke. <#MFOOFSIBTTFUU
+.
(ZOHFMM
,
$SFBO
3

3FEVDFE
BDUJWF
DPOUSPM
BOE
QBTTJWF
SBOHF
BU
UIF
TIPVMEFS
JODSFBTF
SJTL
PG
TIPVMEFS
QBJO
EVSJOH
JOQBUJFOU
SFIBCJMJUBUJPO
QPTUTUSPLF
BO
PCTFSWBUJPOBM
TUVEZ
Journal of Physiotherapy 
o> ,FZ
XPSET: Shoulder pain, Stroke, Risk assessment, Medical audit Introduction body weight, and communication impairment, which may also contribute to risk and influence clinical management Post-stroke shoulder pain is a frequent and disabling (Ratnasabapathy et al 2003). The accuracy of this tool to condition that has been reported in up to 85% of people predict people with stroke who develop shoulder pain has who attend rehabilitation (Bender and McKenna 2001, not yet been investigated. It is also likely that relationships Turner-Stokes and Jackson 2002), and in one-third of stroke exist between proposed risk factors. Models used to assess survivors in general (Lingdgren et al 2007, Ratnasabapathy risk may therefore contain redundant factors and be overly et al 2003). Moderate to severe levels of pain are often complicated. However, knowledge is limited regarding the reported (Lingdgren et al 2007), which can restrict multivariate relationships for predictors of shoulder pain to participation in daily activities and rehabilitation, and guide the development of risk assessment tools. degrade quality of life (Bender and McKenna 2001, Chae et al 2007). Many factors are proposed to contribute to post- Given that existing knowledge about post-stroke shoulder stroke shoulder pain, but these are not well understood. This pain has generally been derived from low quality studies limits effective management of this disabling condition (Snels et al 2002) in small biased samples (Ratnasabapathy (Bender and McKenna 2001, Turner-Stokes and Jackson et al 2003, Turner-Stokes and Jackson 2002), more 2002). investigation is needed to identify predictors for this complex, multifactorial problem. Therefore the research Clinicians need a thorough understanding of the factors questions for this study were: that increase the risk of post-stroke shoulder pain in order to identify patients at risk and implement strategies 1. What is the incidence of post-stroke shoulder pain to prevent and manage this disabling condition (Nicks et during inpatient rehabilitation? al 2007, Turner-Stokes and Jackson 2002). Recently, the Management Tool for Acute Hemiplegic Shoulder was 2. What factors at admission to rehabilitation predict post- developed to assess risk during acute hospital settings stroke shoulder pain during inpatient rehabilitation? (Nicks et al 2007). This tool expanded on a best practice model implemented in a rehabilitation setting (Bernhardt Method and Griffin 2002) and was based on current evidence. The tool focuses on risk factors such as passive range of motion, Design subluxation, pain, limited shoulder function, and altered muscle tone. While these risk factors are consistent with A retrospective audit of medical histories was undertaken many outlined in the literature (Bender and McKenna 2001, to collate the presence of shoulder pain and potential Lingdgren et al 2007), the Management Tool for Acute predictors. Information about predictors was obtained Hemiplegic Shoulder omits several factors, such as age, from the initial physiotherapy and occupational therapy inco-ordination, altered sensation, dyspraxia, side of stroke, assessments, which were standardised and involved a comprehensive overview of impairments and activity limitations. Ninety-four histories were randomly selected from a possible 150 histories of all patients with a primary Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010 195

Research diagnosis of stroke discharged from Austin Health Royal Pain was reported at various frequencies for the 33 Talbot Rehabilitation Centre between July 2005 and June participants with pain (ie, median 33%, range 4% to 100%, 2008. Histories were excluded if the length of stay was 6 of entries per participant). For the 11 participants not days or less. admitted with shoulder pain, the first report of pain was at a median of 4 (range 1 to 14) weeks after admission. Several Participants events were noted that might have contributed to the onset of pain in these 11 participants. These included events or The 94 histories audited represented 63% of stroke patients poor postures that may have traumatised the shoulder (eg, admitted for inpatient rehabilitation over a 3-year period. whilst having investigations such as radiology), altered use The sample was intended to represent a broad cross-section of arm supports, change in pattern of motor recruitment for of people with and without shoulder pain, and included the arm, and a fall. people with cognitive and linguistic impairment who are often not represented in the literature due to inability to As no apparent differences were observed between the provide informed consent (Macrae and Douglas 2008). The 22 participants admitted with pain and the 11 who later sample audited (Table 1) was similar to those not audited developed pain for any of the variables collated, the two for age (mean 59 yr, range 17–80 versus 56 yr, range 18–81) subgroups were pooled for further analyses. This permitted and gender (61% males versus 60%) but had a somewhat a comparison between two groups: participants with (n = longer inpatient stay (mean 48 d, range 7–153 versus 27 d, 33) or without shoulder pain (n = 61). Several factors were 1–190). observed to differ between those with or without pain (Table 1). Those with pain tended to be younger, took longer Outcome measures to be admitted to rehabilitation after their stroke, and had lower Motor Assessment Scale (Carr et al 1985) scores for The variables collated included a range of impairments, the arm. They also tended to have limited passive range of patient characteristics, and stroke-related factors (Table shoulder motion, shoulder subluxation, impaired sensation, 1), the selection of which was guided by current literature and altered muscle tone. For this study, altered muscle (Lingdgren et al 2007, Ratnasabapathy et al 2003, Turner- tone included both hypotonia and hypertonia (Carr and Stokes and Jackson 2002). Passive range of shoulder Shepherd 1998). In contrast, no differences were observed movement was measured using either a goniometer or for several variables including the presence of inattention, visual observation. Sensation was measured using a range of communication impairment, or area and side of stroke clinical assessments including light touch, proprioception, (Table 1). two-point and temperature discrimination. Subluxation was measured by palpation or calipers when the arm was Prediction of shoulder pain unsupported in sitting. Shoulder pain was deemed present if documented in the weekly therapy reports, ward round, The four predictors selected for inclusion in logistic or case conference notes (eg, shoulder pain interfered with regression were Motor Assessment Scale Upper Arm item, dressing or sleeping, therapeutic exercises, or task-related passive range of shoulder flexion, subluxation, and altered practice, or required analgesia). When possible, information sensation. These were selected from the 10 variables that about events (eg, a fall, change in mobility, or use of arm differentiated between people with and without pain (Table supports) preceding the onset of shoulder pain was collated. 1) for several reasons. The predictors focused on primary and secondary impairments following the stroke rather Data analysis than those relating to hospital processes (eg, days between onset and admission to rehabilitation). When two similar Data were summarised for the sample, and subsamples with variables were moderately related, only one variable was and without pain. Data were then analysed using Mann- selected. For instance, the Motor Assessment Scale Upper Whitney (ordinal and interval data that was not normally Arm item was selected over the Hand item as it was distributed) and Chi-Square (categoric data) tests to considered more relevant to the shoulder. Passive range of determine how people with pain differed from those without shoulder flexion was chosen over external rotation as it was pain. To assist in interpreting the observed differences, odds considered easier to measure clinically given the reliance ratios and mean group differences (with 95% CIs) for all upon retrospective data. Although Nicks and colleagues variables were also calculated. Factors that differentiated (2007) suggested that less than 160 degrees shoulder the group with pain from those without pain were then flexion was a predictor for post-stroke shoulder pain, we explored in order to select predictors, and to reduce the used ≤ 150 degrees as a predictor due to the distribution of likelihood of muticollinearity and overfitting within the shoulder ranges observed. Altered tone was not selected as multivariate model (Tabachnick and Fiddell 2001). Given a predictor as it related to several variables including Motor the sample size, the multivariate analysis was restricted Assessment Scale scores, subluxation and shoulder range to a maximum of five predictors. Logistic regression was of motion. then conducted to explore factors associated with shoulder pain. The fit of the model was further explored by entering Logistic regression using the four predictors identified various combinations of predictors into the model. Level of shoulder pain as reliably associated with two predictors: statistical significance was 0.05 for all analyses. Motor Assessment Scale Upper Arm item and passive range of shoulder flexion (Box 1). These findings indicate Results that the odds of experiencing shoulder pain are, on average, 14% greater for people with 150 degrees passive shoulder Participants flexion relative to those with > 150 degrees. The average odds of shoulder pain increase by 64% for each unit lower The participants’ characteristics are summarised in Table on the Motor Assessment Scale Upper Arm item (ie, a 1. Of the 94 participants, 22 (23%) had shoulder pain score of 5 has a 64% greater chance of shoulder pain than when admitted to rehabilitation. A further 11 participants a score of 6). Based on the prediction equation, the mean developed pain during rehabilitation, leading to a total of 33 (35%) who experienced shoulder pain whilst hospitalised. 196 Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010

Blennerhassett et al: Risk factors for shoulder pain after stroke 5BCMF
. Comparison of characteristics of all participants and of the subgroups with and without shoulder pain. Characteristic All Groups Difference between groups participants Pain No pain Mean difference or odds ratio (n = 94) (n = 33) (n = 61) (95% CI) Age (yr), median (range) 59 54 63 MD 4.7 (17 to 80) (23 to 79) (17 to 80) (–1.8 to 11.2) Gender, n male (%) 61 (65) 22 (66) 39 (64) OR 0.89 Onset to admission (days), (0.36 to 2.16) median (range) 12 19 9 Length of inpatient stay (days), (3 to 181) (4 to 181) (3 to 51) MD 12.6 median (range) (3.6 to 21.8) MAS Upper Arm item (0 to 6), 48 70 37 median (range) (7 to 153) (17 to 153) (7 to 148) MD 26.8 MAS Hand item (0 to 6), (13.6 to 39.9) median (range) 5 1 5 PROM shoulder flexion (deg), (0 to 6) (0 to 5) (0 to 6) MD 3.1 median (range) (2.3 to 3.8) PROM shoulder ER (deg), 5 0 6 median (range) (0 to 6) (0 to 6) (0 to 6) MD 3.2 Altered Tone, n (%) (2.3 to 4.1) 180 130 180 Subluxation, n (%) (20 to 180) (20 to 180) (60 to 180) MD 43.3 (23.4 to 63.3) Sensory deficit, n (%) 35 30 40 (–90 to 40) (–90 to 40) (10 to 40) MD 14.8 ?dWjj[dj_ed%d[]b[Yj\"d (4.4 to 25.1) 36 (38) 26 (79) 10 (16) Cognitive impairment, n (%) OR 18.94 24 (25) 19 (58) 5 (7) (6.46 to 55.51) Impaired communication, n (%) 40 (43) 19 (58) 21 (34) OR 19.34 Side of stroke, n (%) (5.57 to 65.94) 24 (26) 10 (30) 14 (23) OR 2.59 48 (51) 17 (52) 31 (51) (1.08 to 6.17) 37 (39) 15 (46) 22 (36) OR 1.53 (0.59 to 3.97) 34 (36) R 15 (46) R 19 (31) R 51 (54) L 16 (48) L 35 (57) LV OR 1.03 9 (10) B 2 (6) B (0.44 to 2.40) 7 (12) B OR 1.48 (0.62 to 3.50) — Type of stroke, n (%) 74 (79) Infarct 27 (82) Infarct 47 (77) Infarct OR 0.76 Area of stroke, n 6 (18) Haem 14 (23) Haem (0.26 to 2.17) 20 (21) Haem Hand dominance, n right (%) 17 MCA 36 MCA — 53 MCA 3 BG 3 BG 6 BG OR 0.24 6 Brainstem 12 Brainstem (0.03 to 2.05) 18 Brainstem 2 PCA 1 PCA 3 PCA 3 ACA 6 ACA OR 2.55 9 ACA (0.63 to 10.22) 2 Multiple 3 Multiple 5 Multiple MD 3.7 32 (97) 54 (89) (–9.3 to 5.5) 86 (92) Previous shoulder problems, n (%) 9 (10) 5 (15) 4 (7) Weight (kg), median (range) 77 (47 to 116) 76 (48 to 116) 77 (47 to 108) Differences between those with and without pain were examined using the Mann-Whitney test for ordinal and continuous data and the Chi-Square test for dichotomous variables. MAS = Motor Assessment Scale, PROM = passive range of motion, ER = external rotation, R = right, L = left, B – bilateral, Haem = haemorrhagic, MCA= middle cerebral artery, BG = basal ganglia, PCA = posterior cerebral artery, ACA = anterior cerebral artery odds and probabilities for experiencing shoulder pain are gender, and altered tone). Similarly, different scoring estimated for the range of people with stroke admitted to methods were used for the passive range of shoulder rehabilitation (Table 2). flexion variable entered (ie, entering scores in degrees, a continuous variable, or a binary variable, ≤ 150 degrees or Goodness of fit of the model was confirmed statistically, not). After all of these variations, the overall interpretation and then further examined by varying the combination of of the model created remained unchanged, and indicated risk factors entered directly into regression. For example, that Motor Assessment Scale Upper Arm item and passive the logistic regression was repeated with an additional range of shoulder flexion were associated with post-stroke 5th variable (eg, days between onset and admission, age, shoulder pain. Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010 197

Research #PY
 Regression coefficients, mean (95% CI) odds ratio that shoulder pain is a multifactorial problem (Price 2002, of predictors, clinical prediction rule and accuracy of model Ratnasabapathy et al 2003). to predict post-stroke shoulder pain. People who experienced shoulder pain also had longer Regression coefficients of predictors periods of hospitalisation, in both the acute and rehabilitation Constant = 3.73 settings. These findings are likely to reflect the severity PROM shoulder flexion = –1.95 of stroke and associated complications. Nevertheless, the MAS Upper Arm item = –0.45 observations that risk of pain increases with the degree of motor impairment at the shoulder and anecdotal events .FBO

$*
PEET
SBUJPT
GPS
QSFEJDUPST
of trauma that preceded shoulder pain reaffirm that PROM shoulder flexion = 0.14 (0.03 to 0.64) the shoulder is highly vulnerable and requires careful MAS Upper Arm item = 0.64 (0.43 to 0.96) management. Given that one-quarter of patients were admitted to rehabilitation with shoulder pain, strategies to Clinical prediction rule identify risk and prevent shoulder pain should occur early Odds (shoulder pain) = e 3.73-1.95 (PROM shoulder flexion) –0.45 and within the acute hospital setting, as recommended by Nicks and colleagues. (MAS upper arm) Shoulder pain after stroke is a complex multifactorial Probability of pain = Odds (shoulder pain) phenomenon (Bender and McKenna 2001, Price 2002). Odds (shoulder pain) + 1 We used multivariate analyses to mathematically simplify a set of 10 factors to two predictors of shoulder pain. The Accuracy of prediction multivariate model had a good level of accuracy, and Nagelkerke R2 = 0.63 explained 63% of the variance in the dataset. Additional Overall accuracy in classifying cases = 85% factors, such as age and altered tone, did not enhance the Sensitivity = 73% model, which suggests that the fit of the model was good. Specificity = 92% Nevertheless, given that any model is highly dependent upon its derived dataset (Tabachnick and Fiddell 2001), the PROM shoulder flexion = Passive range of motion shoulder findings should be replicated in other samples before being flexion (0 = range is ≤ 150 degrees, 1 = range is > 150 degrees), recommended for wider use. MAS = Motor Assessment Scale Our findings support that shoulder pain post-stroke is 5BCMF
. Estimated odds and probability of experiencing heterogeneous in nature (Price 2002). Level of risk and shoulder pain for scores of Motor Assessment Scale underlying mechanisms are likely to vary according to the Upper Arm item, and the presence of restricted passive type and severity of impairments, and personal (eg, age and shoulder flexion. premorbid shoulder problems) and environmental factors (eg, trauma) (Ratnasabapathy et al 2003). It therefore seems MAS Odds Probability important to develop clearer diagnostic classifications in Upper order to direct clinical management. Our findings indicate Arm Shoulder flexion Shoulder flexion that the Motor Assessment Scale Upper Arm item score score > 150 deg ≤ 150 deg > 150 deg ≤ 150 deg may be helpful for this issue. For instance, a score of 0 5.9 41.7 0.86 0.98 4 indicates a high risk of developing shoulder pain, as 3.8 26.6 0.79 0.96 proposed in the Management Tool for Acute Hemiplegic 1 2.4 16.9 0.71 0.94 Shoulder (Nicks et al 2007). For this group of patients, who 1.5 10.8 0.61 0.91 are also more likely to have shoulder subluxation, clinical 2 1.0 6.9 0.50 0.87 management including use of arm support, electrical 0.6 4.4 0.38 0.81 stimulation, education, and active motor training to promote 3 0.4 2.8 0.28 0.74 shoulder girdle control, as outlined by Nicks and colleagues, 4 seems highly appropriate. However, despite the lower odds, patients admitted with a score of 4 or 5 in our study also had 5 shoulder pain. Physiotherapists would need to employ other 6 approaches to manage these people as different mechanisms for pain, such as shoulder impingement, are likely (Bender MAS = Motor Assessment Scale and McKenna 2001, Blennerhassett et al 2009). Discussion Despite the observed association with pain, reduced passive range and motor control at the shoulder cannot The findings from this study support that shoulder pain be considered the cause of post-stroke shoulder pain. is a common problem (Lingdgren et al 2007) that can Nevertheless, the findings suggest that clinical attention occur early after stroke (Dromerick et al 2008). Shoulder could be directed to improving pain free shoulder joint pain was noted in one in four participants at admission to range, or promoting active shoulder girdle control to align rehabilitation and one in three participants during inpatient the glenohumeral joint and enable arm elevation. Training rehabilitation. The incidence observed is consistent with should be carefully structured and monitored, given the other reports during stroke rehabilitation (Dromerick et al importance of highly co-ordinated muscular control within 2008) and the general population with stroke (Lingdgren the shoulder girdle (Dontalelli 2004), and the potential for et al 2007, Ratnasabapathy et al 2003). Several factors, impingement, wear and tear, inflammation, and subsequent including weakness, altered motor control, joint stiffness, pain at the shoulder – particularly when the muscles are and subluxation, differentiated people who developed pain weak or fatigued, or while performing overhead activities from those who did not. These factors have often been (Ludewig and Reynolds 2009). Education and training found to be associated with shoulder pain (Chae et al 2007, Turner-Stokes and Jackson 2002), supporting the notion 198 Journal of Physiotherapy 2010 Vol. 56 – © Australian Physiotherapy Association 2010