Journal of Physiotherapy 69 (2023) 93–99 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Research Self-administered stretching exercises are as effective as motor control exercises for people with chronic non-specific low back pain: a randomised trial Aline Mendonça Turci a, Camila Gorla Nogueira a, Helen Cristina Nogueira Carrer b, Thais Cristina Chaves b a Health Sciences Department, Ribeirão Preto Medical School, University of São Paulo, Ribeirão Preto, Brazil; b Department of Physical Therapy, Federal University of São Carlos, São Carlos, Brazil KEY WORDS ABSTRACT Randomised clinical trial Question: In people with chronic non-specific low back pain, what is the effect of self-administered Non-specific chronic low back pain stretching exercises relative to motor control exercises on pain intensity, disability, fear avoidance, global Self-stretching exercises perceived effect and flexibility? Design: Randomised trial with concealed allocation, intention-to-treat Motor control exercises analysis and blinding of assessors. Participants: One hundred people with chronic non-specific low back Physical therapy pain. Interventions: The self-stretching exercise group performed 6 stretches in 40-minute sessions. The motor control exercise group performed trunk stabilising exercises in 40-minute sessions. Both groups performed weekly supervised sessions for 8 weeks with one or more home sessions/week. Outcome measures: The primary outcomes were pain intensity (0 to 10 scale) and disability (Oswestry Disability Index). The secondary outcomes were the Fear Avoidance Beliefs Questionnaire, global perceived effect, and the fingertip-to-floor test. Measures were taken at baseline and at 8, 13 and 26 weeks. Results: On the 0 to 10 scale, the between-group difference in pain intensity was negligible, with a mean difference of roughly 0 (95% CI 21 to 1) at each time point. Similarly, the between-group difference on the 100-point disability scale was negligible: MD –1 (95% CI –3 to 1) at week 8, MD 1 (95% CI –1 to 3) at week 13 and MD 0 (95% CI –1 to 2) at week 26. The two interventions also had similar effects on the secondary outcomes. Conclusion: In people with chronic non-specific low back pain, self-stretching exercises had very similar effects to motor control exercises on pain intensity, disability, fear avoidance, global perceived effect and flexibility up to 18 weeks beyond the end of an 8-week program. Given the established effectiveness of motor control exercises, either intervention could be recommended to people with chronic low back pain. The choice of intervention might be directed by patient preference. Registration: NCT03128801. [Turci AM, Nogueira CG, Nogueira Carrer HC, Chaves TC (2023) Self-administered stretching exercises are as effective as motor control exercises for people with chronic non-specific low back pain: a randomised trial. Journal of Physiotherapy 69:93–99] © 2023 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Introduction to continue with everyday activities. In addition, exercises are Low back pain is common in all age groups, and it is frequently another non-invasive treatment recommended by the NICE guide- observed in high, middle and low-income countries.1 It is the leading line7 and were also recommended in a summary of 11 standard/ cause of disability and lost productivity, with a prevalence over a lifetime of up to 84% for adults.2 Low back pain can be defined as pain consistent recommendations obtained across high-quality clinical or discomfort located between the costal margin and the gluteal practice guidelines for musculoskeletal pain conditions.8 In an over- folds, with or without referred pain in the legs, and it is the most common type of chronic pain.3,4 It is estimated that 85% of low back view of Cochrane reviews about physical activity and exercise for pain does not present an identifiable pathoanatomical cause, and no reliable evidence of a cause can be observed through the available chronic pain, the results were inconsistent for pain intensity across imaging tests.5,6 When such low back pain lasts for . 3 months, it is described as chronic non-specific low back pain (CNSLBP). studies. However, the review reported small-to-moderate effect sizes for physical function.9 The National Institute for Health and Care Excellence (NICE)7 guideline recommends that self-management strategies should be Motor control exercises (MCE) are very commonly used to treat provided for patients with CNSLBP. Such strategies include informa- tion on the nature of low back pain and sciatica and encouragement CNSLBP. They are focused on activating the deep trunk muscles and targeting the restoration of control and coordination during exercises, progressing to more complex and functional tasks integrating the activation of deep and global trunk muscles.10 Previous systematic reviews with meta-analysis have demonstrated that MCE is more effective than a minimal intervention to reduce pain intensity and disability.10,11 Previous systematic reviews with meta-analysis have https://doi.org/10.1016/j.jphys.2023.02.016 1836-9553/© 2023 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).
94 Turci et al: Self-stretching and motor control exercises for back pain found no clinically important difference between MCE and other generated the allocation sequence and assigned participants to the types of exercise.10,12 Conversely, a network meta-analysis11 showed intervention groups was not involved in their recruitment, assess- that core strengthening exercises (ie, MCE) showed better effects for ment, intervention or any other task of the study. After the group pain intensity and disability when compared with other exercise assignment, a second blinded researcher administered baseline types, but these effects did not meet the minimum clinically impor- questionnaires (ie, appropriately validated questionnaires in Brazilian tant difference.13 Portuguese) and the fingertip-to-floor test. Afterwards, the partici- pants received the SSE or MCE interventions as one-to-one sessions Exercises involving therapeutic whole-body stretching, also called with the same physiotherapist, who was not involved in any other global postural re-education, can treat painful musculoskeletal con- task in the study. The treatment period took place over 8 weeks, with ditions such as chronic back pain. These exercises aim to promote the a weekly session23 of 40 minutes. In the first session, participants stretching of muscle chains and improve the contraction of the from each group took home handouts with clear written instructions antagonist muscles,14–16 that is, the isometric contraction of antago- and illustrations on how to perform the exercises. Participants were nist muscles in association with the active stretching of the target invited to perform at least one home session per week. Immediately muscles (agonists).17,18 after the last intervention, the participants were reassessed for the primary and secondary outcomes. Finally, the participants returned 4 In a systematic review,19 global postural re-education had similar and 12 weeks after the end of the treatment for follow-up assess- effects on pain intensity and disability as other treatments for pa- ment. Those who were unable to return face-to-face were contacted tients with a range of musculoskeletal conditions. A meta-analysis20 by telephone for the reassessments. This study is reported according on the effectiveness of global postural re-education for treating spi- to the recommendations of the CONSORT Statement.24 nal disorders concluded that this might be an effective method for treating spinal disorders, decreasing pain and improving function. Participants, therapists and centres Previous studies have shown that self-stretching exercises for The inclusion criteria were: age 18 to 60 years; diagnosis of CNSLBP showed better effects on pain intensity,16,21 function and CNSLBP in the last 3 months and/or pain located between T12 and quality of life16 compared with a conventional physiotherapy pro- gluteal folds on at least half of the days in the last 6 months;25 pain gram of exercises or a waiting list group. In one previous study,22 intensity 3 on a 0 to 10 numerical pain rating scale; mechanical which was not randomised, global stretching exercises were pain behaviour caused by postures, activities and movements such as compared with MCE in CNSLBP and better effects were found from flexion and functional movement tasks; and scores . 14% on the stretching exercises on pain intensity, function and global mobility. Oswestry Disability Index.26 The exclusion criteria were: red flags Additionally, a previous network meta-analysis11 reported that (neoplastic diseases or spinal tumours, inflammatory diseases, in- stretching showed greater effects for pain intensity and disability fections and fractures); presenting severe neurological (central or than minimal intervention in chronic low back pain. peripheral) symptoms; psychiatric, rheumatological and/or cardiac diseases; signs of radiculopathy, lumbar stenosis or spondylolisthesis; Global stretching exercises is a therapist-dependent approach, history of spinal surgery; pregnancy; and receipt of physiotherapy which is a disadvantage of the method. On the other hand, self- treatment in the 6 months before the evaluation period. stretching exercises (SSE), also called Global Active Stretching®, is an alternative that minimises dependence on the therapist, as the One physiotherapist who was not involved in the assessments patients are invited to adopt self-stretching postures, with a contin- treated the participants. Due to the nature of the interventions, the uous increase in autonomy throughout the treatment. Patients sus- therapist could not be blinded. The physiotherapist who delivered the tain the stretching postures through low-level sustained contraction interventions was an experienced clinician who was certified in the of the antagonist muscles to those being stretched. SSE approach with 10 years of clinical experience using the SSE approach regularly. The study was conducted in a primary care However, no previous study has investigated the effect of SSE setting from the Centro Saúde Escola Cuiabá - Ribeirão Preto Medical when compared with MCE for patients with CNSLBP. Therefore, the School, University of São Paulo, Brazil. primary aim of this study was to compare the effect of SSE with the effect of MCE on pain intensity and low back pain-related disability in Interventions patients with CNSLBP immediately and 8 weeks after the treatment programs. Considering that self-stretching posture exercises combine Self-stretching exercise stretching and contraction in the same approach, we hypothesised The SSE protocol was based on the study by Lawand et al,16 who greater benefits from an SSE program on pain intensity and disability compared with an MCE program. used a series of stretching postures from the Global Active Stretch- ing® program described by Souchard27 (Appendix 1 on the eAd- Therefore, the research question for this randomised trial was: denda). A certified therapist with experience in the Global Active Stretching® approach administered the sessions in a standardised In people with chronic non-specific low back pain, what is the way (Appendix 1), and each exercise posture was sustained for 10 to effect of self-administered stretching exercises relative to motor 20 minutes. The therapist verbally led and guided the treatment control exercises on pain intensity, disability, fear avoidance, postures. Furthermore, because it was an active technique, the par- global perceived effect and flexibility? ticipants were advised to progress each posture as they felt able during the sustained stretching period. The aim was to achieve Method maximal leg extension and maximal ankle dorsiflexion for each posture. Maximal shoulder abduction was the target for some pos- Design tures (Appendix 1, picture B weeks 3 and 4) and maximal shoulder adduction for others (Appendix 1, picture B weeks 1 and 2). In This study was a prospectively registered, two-arm, randomised addition, the participants were advised to decrease low back lordosis trial with concealed allocation, blinded assessment of some outcomes (trying to keep their lower back in contact with the surface). Use of a and intention-to-treat analysis. Participants were recruited from specific breathing pattern is a fundamental part of the approach and among people with CNSLBP who had been referred to a physio- was instituted during all the postures. During inspiration, the par- therapy clinic. An initial evaluation was conducted for eligibility ticipants were encouraged to inhale air through the nose, expanding assessment, in which participants answered questions about de- the region of the lower ribs. During expiration, the top of the chest mographic characteristics, clinical status (including medication use), was lowered and the abdomen was allowed to protrude, as shown in and red flags. A research assistant then conducted the group alloca- Appendix 1.27 tion procedure. Participants were randomly assigned to two treat- ment groups following simple computerised randomisation procedures using opaque sealed envelopes. The researcher who
Research 95 Motor control exercise Data analysis The MCE protocol that was used was described by Hicks et al,28 as The two primary outcome variables (pain intensity and disability shown in Appendix 2 on the eAddenda. It was directed by a single related to back pain) were each considered in the sample size therapist (the same therapist as directed the SSE group) in individual calculation, as described in the registered protocol. For the pain in- sessions. Progression of the exercises was based on specific criteria tensity outcome, the study was powered to detect a between-group (being able to maintain muscle contraction for 8 seconds for 30 difference of at least 2 units30 (SD 4.5) on the 0 to 10 numerical repetitions in bilateral exercises or 20 repetitions for each limb in pain rating scale. Statistical power of 90% and a = 0.05 were used in unilateral exercises). One of the first steps to start the MCE program the calculation, which resulted in 40 participants per group. For the was to teach the participants to maintain contraction of the trunk’s low back pain-related disability, the sample size calculation showed stabilising muscles while dissociating that from movements of the that 43 participants per group were necessary to detect a difference extremities.29,30 Movements were started in a single plane, pro- of 10 units30 with an anticipated SD of 16. The statistical analyses gressing to multidimensional planes.30 were conducted using commercial statistical softwarea. To allow for some loss to follow-up, the calculated sample size of 86 was Outcome measures increased to 100. The primary outcomes were pain intensity and disability related All statistical procedures were performed according to the to low back pain. The secondary outcomes were fear avoidance, intention-to-treat principle. First, descriptive statistics and histogram global perceived effect and the fingertip-to-floor test. All measures inspections were used to determine whether the data were normally were recorded at baseline (week 0), at the end of treatment (week 8) distributed. To assess between-group differences in response to and at two follow-up points (weeks 13 and 26), except global treatment at each post-baseline time point, the mean between-group perceived effect, which was only recorded at weeks 8, 13 and 26. difference and its associated 95% confidence interval (CI) were calculated. The statistical analysis was conducted by a researcher who Pain intensity was not involved in any of the phases of data collection and received The Numerical Pain Rating Scale (0 = no pain to 10 = worst data in coded form. The statistical analyses were conducted using commercial statistical softwareb. imaginable pain) evaluated the mean pain intensity in the last 7 days.31 The Numerical Pain Rating Scale is responsive to changes and Results the minimum clinically important difference has been nominated as 2.0 points among patients with chronic low back pain.13 Flow of participants through the study Disability Between February 2017 and April 2019, we screened 158 potential The Oswestry Disability Index32 comprises ten items – each one participants. Of these, 36 did not meet the inclusion criteria, 19 declined to participate, and three had acute health problems that has six response options. The first response option receives 0 points prevented them from participating in the study (one of them suffered and describes the absence or a small amount of low back pain and an acute myocardial infarction, and two others were affected by functional disability. In contrast, the sixth response option receives 5 dengue fever). One hundred participants with CNSLBP were rando- points and describes extreme pain or functional disability. The sum of mised into two groups: SSE (n = 50) and MCE (n = 50) (Table 1). The the points calculates the total score, the highest possible sum being flow of participants through the trial is shown in Figure 1. All par- 50 points. The result is transformed to a scale from 0 to 100 by ticipants were measured at the final time point so there was no multiplying the final score by two. A higher score on the question- permanent loss to follow-up in the trial, as shown in Figure 1. naire means more significant disability related to low back pain. However, one participant declined to be evaluated at week 8 and six Previous research has described 10 points as the minimum clinically participants declined at week 13 (Tables 2 and 3). important difference effect of the Oswestry Disability Index.13 Compliance with the study protocol Fear avoidance beliefs The Fear Avoidance Beliefs Questionnaire (FABQ),33 consisting of All registered outcome measures are reported in this manuscript. Two participants did not return to receive further intervention after 16 self-response items evaluated on a 7-point Likert scale, is scored on two subscales: work (FABQ-Work) and physical activity (FABQ- Table 1 Phys). A previous study34 reported the minimum clinically important Baseline characteristics of the participants. difference effect of 7 points for FABQ-Work and 4 points for FABQ- Phys. Characteristics SSE (n = 50) MCE (n = 50) Global perceived effect Age (yr), mean (SD) 37 (13) 37 (12) This measure is a single-item scale for participants to rate their Gender, n (%) 37 (74) 31 (62) perception of clinical change on an 11-point scale ranging from female 13 (26) 19 (38) extremely worse (–5) to completely recovered (15) with the male 76 (17) 80 (23) midpoint as no modification (0).31 The minimum clinically important Weight (kg), mean (SD) 167 (9) 168 (12) difference effect has been nominated as 2 points for low back pain35 Height (cm), mean (SD) 27.5 (5.8) 28.0 (5.8) and 1.7 points for chronic low back pain.36 Body mass index (kg/m2), mean (SD) 10 (10) Pain duration (y), mean (SD) 9 (8) Flexibility Pain frequency (episodes/wk), mean (SD) 5 (2) 5 (2) The fingertip-to-floor test is a global stretching assessment that Education, n (%) completed first degree 1 (2) 1 (2) was described and validated by Perret et al.37 The subject should incomplete high school 3 (6) 1 (2) stand barefoot on a 20-cm high platform with the feet together and completed high school 8 (16) 16 (32) bend the torso forward as far as possible, keeping the knees, arms and technician 1 (2) 0 (0) fingers extended for the test. The vertical distance between the tip of incomplete higher degree 7 (14) 8 (26) the middle finger and the platform should be measured with a flex- completed higher degree 25 (50) 22 (44) ible measuring tape, and the measurement expressed in centimetres. completed postgraduate studies 4 (8) 2 (4) The test is considered positive when the tip of the middle finger does Occupational activity, n (%) not reach the platform and negative when it goes past the plat- No manual handling at work 37 (74) 38 (76) form.22,38 It is a test that has good psychometric properties.37,38 The Loading weight at work 13 (26) 12 (24) minimum clinically important difference has not been established but the minimum detectable change for this test was 4.5 cm.38 MCE = motor control exercise group, SSE = self-stretching exercise group.
96 Turci et al: Self-stretching and motor control exercises for back pain Figure 1. Flow of participants through the trial. the third day of intervention and five participants did not return to Disability receive further intervention after the fourth day of intervention Disability improved in both groups during the intervention (overall abandonment rate 7%). Of these seven participants, five were in the SSE group (abandonment rate 10%) and two were in the MCE period; however, there was negligible difference in the effect of the group (abandonment rate 4%). The assessor remained blinded to all two interventions at the end of the intervention period: MD –1 (95% participants’ group allocation. All analyses followed the intention-to- CI –3 to 1). The gains achieved during the intervention period were treat principle. sustained thereafter. However, there was negligible difference in the effect of the two interventions at the 13-week and 26-week follow-up Characteristics of the participants periods: MD 1 (95% CI –1 to 3) and MD 0 (95% CI –1 to 2), respectively (Table 2). The individual participant data are presented in Table 4 on The randomly allocated groups were well matched on all de- the eAddenda. mographic characteristics measured in the study (Table 1). They were also well matched on all the baseline scores of the outcome measures Fear avoidance (Table 2). Scores on the FABQ-Work and FABQ-Phys domains improved in Effect of the intervention both groups during the intervention period.; however, there was negligible difference in the effect of the two interventions at the end Pain intensity of the intervention period: MD 1 (95% CI –3 to 5) and MD –1 (95% CI Although pain severity improved markedly in both groups during –3 to 2), respectively. At the week 13 follow-up assessment, the between-group differences for the FABQ-Work and FABQ-Phys do- the intervention period, there was negligible difference in the effect mains were both negligible. These estimates had confidence intervals of the two interventions at the end of the intervention period: MD where the upper limit equated to the published minimum clinically –0.2 (95% CI –1.1 to 0.7). The gains achieved during the intervention important difference. That is, the estimate for FABQ-Work was MD 2 period were largely sustained thereafter. However, there was still (95% CI –2 to 7); the upper limit of 7 shows that there is a small negligible difference in the effect of the two interventions at the 13- possibility that the true average difference in effect is worthwhile in week and 26-week follow-up periods: MD 0.0 (95% CI –1.0 to 1.1) and favour of MCE, but it is more likely that the true difference is negli- MD 0.2 (95% CI –0.9 to 1.2), respectively (Table 2). The individual gible. Similarly, the upper limit of 4 in the result for FABQ-Phys shows participant data are presented in Table 4 on the eAddenda. that there is a small possibility that the true average difference in effect is worthwhile in favour of MCE, but it is more likely that the true difference is negligible. At the 26-week follow-up, the
Research 97 Week 13 minus Week 26 minus Week 0 SSE minus MCE differences were again clearly negligible (Table 2). The individual 0.2 (–0.9 to 1.2) participant data are presented in Table 4 on the eAddenda. Between-group differences 0 (–1 to 2) Global perceived effect 2 (–2 to 6) The between-group difference in the rating of global perceived 1 (–2 to 3) 0 (–1 to 1) effect was negligible at the end of the intervention period: MD 0.1 (95% CI –0.5 to 0.7). The difference remained negligible at the two Week 0 SSE minus MCE follow-up assessment points (Table 3). The individual participant data 0.0 (–1.0 to 1.1) are presented in Table 4 on the eAddenda. 1 (–1 to 3) Fingertip-to-floor test 2 (–2 to 7) The between-group difference in the fingertip-to-floor test was 1 (–2 to 4) 0 (–1 to 1) the same at all time points: MD 0 cm (95% CI –1 to 1), which was clearly negligible (Table 2). The individual participant data are pre- Week 8 minus SSE minus MCE sented in Table 4 on the eAddenda. Week 0 –0.2 (–1.1 to 0.7) –1 (–3 to 1) 1 (–3 to 5) –1 (–3 to 2) 0 (–1 to 1) Week 26 minus MCE Discussion Week 0 –5.0 (2.8) This study compared the effects of SSE with the effects of MCE on –6 (4) pain intensity, disability, fear avoidance, global perceived effect and –8 (9) flexibility in people with CNSLBP. The effects of these two in- –6 (8) terventions were similar enough that we can interpret that they are –3 (4) equally effective for clinical purposes. This can be concluded with great confidence for two reasons. First, the methods of the rando- SSE Shaded cells = primary outcomes. Small anomalies in subtraction are due to the effects of rounding. Higher scores reflect a worse outcome for all outcome measures in this table. mised trial were very robust. The random allocation process was –4.8 (2.7) MCE = motor control exercise group, Phys = physical, SSE = self-stretching exercise group. protected by concealed allocation and produced very comparable groups at baseline. Only 7% of participants received less than the full –6 (6) a 1 missing intervention allocated to them. The outcome assessor was blinded –6 (12) b 2 missing and all analyses followed the intention-to-treat principle. All partic- –5 (7) c 3 missing ipants were followed up at the final assessment timepoint, with d 4 missing minimal missing data at interim assessment timepoints. Second, the 0 (5) e 5 missing trial generated very precise estimates (as indicated by the very nar- row CIs), which were able to exclude the possibility of clinically Within-group differences MCE important differences between the two interventions. That is, the Week 13 minus –5.0 (2.4) limits of the CIs were smaller in magnitude than the nominated Week 0 minimum clinically important difference. –7 (4) –8 (9) One exception was the FABQ at the week 13 follow-up assessment. –7 (8) The between-group differences for FABQ domains were both negli- –3 (5) gible but the upper limit of their CIs equalled the minimum clinically important difference effect. That is, the estimate for FABQ-Phys was SSE MD 1 (95% CI –2 to 4); the upper limit of 4 shows that there is a small –5.0 (2.6) possibility that the true average difference in effect is worthwhile in favour of MCE, but it is more likely that the true difference is negli- –6 (5) gible. Similarly, the upper limit of 7 in the result for FABQ-Work –6 (12) shows that there is a small possibility that the true average differ- –5 (7) ence in effect is worthwhile in favour of MCE, but it is more likely that –1 (5) the true difference is negligible. In either case, the effect was tran- sient; it was not evident at the next assessment at week 26. It is Week 8 minus MCE difficult to hypothesise why a treatment benefit would not be present Week 0 –5.3 (2.6) at the end of the treatment program, but then appear 1 month later, and then be absent 2 months after that. It seems more reasonable to –5 (3) interpret that the two treatments are equally effective overall. –6 (9) –5 (6) Given the above, what factors might be used to help physiotherapists –3 (5) decide which of these two treatments to recommend and to help people with CNSLBP determine which of these two treatments they prefer? The SSE choice of treatment should be based on the patient’s preference, as rec- –5.5 (2.0) ommended by the NICE Guideline.7 Patients may have a preference regarding the types of exercise involved; some may prefer learning the –6 (5) skills of MCE or dislike the sensation or monotony of prolonged –5 (11) stretching, whereas others may prefer the sustained stretching or dislike –5 (7) the concentration required to perform MCE. Patients may find one –1 (6) intervention easier to learn and therefore may be able to move to an independent program with fewer one-to-one sessions from the physio- Week 26 MCE therapist. In shared decision-making, the therapist should also consider (n = 50) their own skills at delivering each intervention. 1.3 (2.3) Because the trial results indicate comparable effects of the two 4a (4) interventions, we cannot confirm our a-priori hypothesis that the SSE 7 (9) program would show greater benefits in pain intensity and disability 7 (7) than a rehabilitation program focused on MCE in patients with 9d (7) CNSLBP. It is worthwhile to consider why that hypothesis was not Table 2 SSE Mean (SD) of groups, mean (SD) within-group differences and mean (95% CI) between-group differences. (n = 50) 1.5 (2.3) 5b (5) 11 (10) 8 (7) 10e (9) Groups MCE Week 13 (n = 47) 1.2 (2.0) 3 (3) 6 (8) 6 (6) 9a (7) SSE (n = 48) 1.3a (2.3) 5b (5) 11 (10) 8a (7) 10a (9) Week 8 MCE (n = 49) 0.9 (1.9) 5a (4) 8 (8) 8 (6) 9c (7) SSE (n = 50) 0.8 (1.7) 5a (4) 12 (10) 8 (7) 10b (8) Week 0 MCE (n = 50) 6.2 (1.8) 10 (4) 14 (10) 13 (7) 12 (8) SSE (n = 50) 6.3 (1.3) 11 (4) 17 (11) 13 (7) 10 (9) Outcomes Pain intensity (0 to 10) Disability (0 to 100) Fear avoidance - Work (0 to 24) Fear avoidance - Phys (0 to 42) Fingertip-to-floor test (cm)
98 Turci et al: Self-stretching and motor control exercises for back pain Table 3 Mean (SD) of groups and mean (95% CI) between-group differences. Outcome Groups Between-group differences Week 13 Week 8 Week 26 Week 8 Week 13 Week 26 SSE (n = 50) MCE (n = 49) SSE (n = 47) MCE (n = 47) SSE (n = 50) MCE (n = 49) SSE minus MCE SSE minus MCE SSE minus MCE Global perceived 4.2 (1.4) 4.1 (1.6) 3.8 (1.7) 4.3 (1.2) 3.7 (1.8) 4.2 (1.5) 0.1 (–0.5 to 0.7) –0.5 (–1.1 to 0.1) –0.5 (–1.2 to 0.2) effect (–5 to 5) Higher scores indicate a better outcome. MCE = motor control exercise group, SSE = self-stretching exercise group. confirmed. First, because SSE involved both active stretching and Footnotes: a GPower 3.0.10, University of Kiel, Kiel, Germany. active contractions and MCE involved active contractions only, we b SPSS 22.0 software, SPSS Inc, Chicago, USA. anticipated a greater beneficial effect from SSE. Conversely, the re- eAddenda: Table 4 and Appendices 1 and 2 can be found online at sults showed exercises based on stretching and active contraction https://doi.org/10.1016/j.jphys.2023.02.016 showed the same efficacy as exercises based in contractions only, and Ethics approval: The Ethics Committee Board from the School so no clear between-group differences were evident. Perhaps the Health Centre of the Ribeirão Preto Medical School – University of São nature of exercise is not important other than it means the patient is Paulo (FMRP/USP) approved this study (CAAE number: not receiving passive treatments or restricting their movement; in 55268116.9.0000.5414). All participants were informed about the this case, equivalent doses of SSE or MCE would be expected to give procedures of this study, agreed to participate, and signed the con- equivalent effects. In accordance with these findings, Hayden et al11 sent form. showed no differences that meet the minimum clinically important Competing interests: Nil. difference for pain intensity (2 out of 10)13 and functional limitation Source(s) of support: Nil. (10 out of 100)13 between core strengthening exercises (ie, MCE) Acknowledgements: Nil. when compared with stretching exercises in chronic low back pain. A Provenance: Not invited. Peer reviewed. final possibility is that, as the participants themselves controlled the Correspondence: Thais Cristina Chaves, Department of Physical progression of the stretching, this may have influenced the amount of Therapy, Federal University of São Carlos, São Carlos, Brazil. Email: stretching obtained during the treatment. [email protected] Both exercise programs brought immediate and long-term bene- References fits for pain and disability in chronic low back pain. Given the well-established efficacy of MCE,10–12 either exercise program can be 1. Ketenci A, Zure M. Pharmacological and non-pharmacological treatment ap- recommended for the treatment of CNSLBP. proaches to chronic lumbar back pain. Turk J Phys Med Rehabil. 2021;67:1–10. https://doi.org/10.5606/tftrd.2021.8216 Regarding treatment adherence, all rates were considered accept- able (, 20% abandonment).39 The abandonment rate was 7% of the total 2. Popescu A, Lee H. Neck pain and lower back pain. Med Clin North Am. sample, consisting of five participants (10%) in the SSE group sample 2020;104:279–292. https://doi.org/10.1016/j.mcna.2019.11.003 and two participants (4%) in the MCE group. However, 99% of the par- ticipants agreed to be re-evaluated at the end of the 8-week inter- 3. Suzuki H, Aono S, Inoue S, Imajo Y, Nishida N, Funaba M, et al. Clinically significant vention period, including those who left the treatment. Although 6% of changes in pain along the Pain Intensity Numerical Rating Scale in patients with the participants (n = 3 in each group) did not attend assessment at week chronic low back pain. PLoS One. 2020;15:e0229228. https://doi.org/10.1371/jour- 13, all participants attended the final assessment at week 26. nal.pone.0229228 A limitation of the study was that, due to the nature of the in- 4. Airaksinen O, Brox JI, Cedraschi C, Hildebrandt J, Klaber-Moffett J, Kovacs F, et al. terventions administered, it was impossible to blind the participants European guidelines for the management of chronic nonspecific low back pain. Eur and the therapist. Although not a limitation of the current trial per se, Spine J. 2006;15(Suppl 2):S192–S300. https://doi.org/10.1007/s00586-006-1072-1 the lack of studies on SSE as a therapeutic modality made it difficult to compare our results to comparable studies. Thus, future studies are 5. Goertz M, Thorson D, Bonsell J, Bonte B, Campbell R, Haake B, et al. Institute for encourage to further explore this modality of exercises. Clinical Systems Improvement. Adult acute and subacute low back pain. Updated November 2012. The results did not confirm the a priori hypothesis that the SSE program would have greater benefits on pain intensity and disability 6. Sheeran L, Sparkes V, Caterson B, Busse-Morris M, van Deursen R. Spinal position than MCE in patients with CNSLBP. As the effectiveness of MCE is well sense and trunk muscle activity during sitting and standing in nonspecific chronic established10–12 and the effect observed in the SSE and MCE groups low back pain: classification analysis. Spine. 2012;37:E486–E495. https://doi.org/ were so similar, we recommend that either intervention could be used 10.1097/BRS.0b013e31823b00ce for the treatment of CNSLBP and that patient (and therapist) prefer- ences regarding the different exercise modalities should be considered. 7. National Institute for Health and Care Excellence. Low back pain and sciatica in over 16s: assessment and management (NICE Guideline NG59); 2016. www.nice. What was already known on this topic: Motor control ex- org.uk/guidance/ng59. Accessed December 14, 2012. ercises have well-established moderate benefits for people with chronic non-specific low back pain. Although trials have shown 8. Lin I, Wiles L, Waller R, Goucke R, Nagree Y, Gibberd M, et al. What does best better effects on pain and disability from self-stretching exer- practice care for musculoskeletal pain look like? Eleven consistent recommenda- cises than from exercises based on stabilising the spine and tions from high-quality clinical practice guidelines: systematic review. Br J Sports correcting postural alignment, the methodological quality of Med. 2020;54:79–86. https://doi.org/10.1136/bjsports-2018-099878 those trials was not robust. What this study adds: In people with chronic non-specific low 9. Geneen LJ, Moore RA, Clarke C, Martin D, Colvin LA, Smith BH. Physical activity and back pain, self-stretching exercises had very similar effects to exercise for chronic pain in adults: an overview of Cochrane Reviews. Cochrane motor control exercises on pain intensity, disability, fear avoid- Database Syst Rev. 2017;4:CD011279. https://doi.org/10.1002/14651858.CD011279. ance, global perceived effect and flexibility up to 18 weeks pub3 beyond the end of an 8-week program. Given the established effectiveness of motor control exercises, either intervention 10. Saragiotto BT, Maher CG, Yamato TP, Costa LO, Costa LC, Ostelo RW, et al. Motor could be recommended to people with chronic low back pain. control exercise for nonspecific low back pain: a Cochrane review. Spine. 2016;41:1284–1295. https://doi.org/10.1097/BRS.0000000000001645 11. Hayden JA, Ellis J, Ogilvie R, Stewart SA, Bagg MK, Stanojevic S, et al. Some types of exercise are more effective than others in people with chronic low back pain: a network meta-analysis. J Physiother. 2021;67:252–262. 12. Smith BE, Littlewood C, May S. An update of stabilisation exercises for low back pain: a systematic review with meta-analysis. BMC Musculoskelet Disord. 2014;15:416. https://doi.org/10.1186/1471-2474-15-416 13. Ostelo RW, Deyo RA, Stratford P, Waddell G, Croft P, Von Korff M, et al. Interpreting change scores for pain and functional status in low back pain: towards interna- tional consensus regarding minimal important change. Spine. 2008;33:90–94. https://doi.org/10.1097/BRS.0b013e31815e3a10 14. Oliveri M, Caltagirone C, Loriga R, Pompa MN, Versace V, Souchard P. Fast increase of motor cortical inhibition following postural changes in healthy subjects. Neu- rosci Lett. 2012;530:7–11. https://doi.org/10.1016/j.neulet.2012.09.031 15. Barroqueiro C, Morais NV. The effects of a global postural reeducation program on an adolescent handball player with isthmic spondylolisthesis. J Bodyw Mov Ther. 2014;18:244–258. https://doi.org/10.1016/j.jbmt.2013.10.002
Research 99 16. Lawand P, Lombardi Júnior I, Jones A, Sardim C, Ribeiro LH, Natour J. Effect of a 28. Hicks GE, Fritz JM, Delitto A, McGill SM. Preliminary development of a clinical muscle stretching program using the global postural reeducation method for pa- prediction rule for determining which patients with low back pain will respond to tients with chronic low back pain: A randomized controlled trial. Joint Bone Spine. a stabilization exercise program. Arch Phys Med Rehabil. 2005;86:1753–1762. 2015;82:272–277. https://doi.org/10.1016/j.jbspin.2015.01.015 https://doi.org/10.1016/j.apmr.2005.03.033 17. Teodori RM, Negri JR, Cruz MC, Marques AP. Global Postural Re-education: a 29. Akuthota V, Nadler SF. Core strengthening. Arch Phys Med Rehabil. 2004;85(3 Suppl literature review. Rev Bras Fisioter. 2011;15:185–189. 1):S86–S92. https://doi.org/10.1053/j.apmr.2003.12.005 18. Di Ciaccio E, Polastri M, Bianchini E, Gasbarrini A. Herniated lumbar disc treated 30. Willson JD, Dougherty CP, Ireland ML, Davis IM. Core stability and its relationship with Global Postural Reeducation. A middle-term evaluation. Eur Rev Med Phar- to lower extremity function and injury. J Am Acad Orthop Surg. 2005;13:316–325. macol Sci. 2012;16:1072–1077. https://doi.org/10.5435/00124635-200509000-00005 19. Ferreira GE, Barreto RG, Robinson CC, Plentz RD, Silva MF. Global Postural Reedu- 31. Costa LO, Maher CG, Latimer J, Ferreira PH, Ferreira ML, Pozzi GC, et al. Clinimetric cation for patients with musculoskeletal conditions: a systematic review of ran- testing of three self-report outcome measures for low back pain patients in Brazil: domized controlled trials. Braz J Phys Ther. 2016;20:194–205. https://doi.org/10. which one is the best? Spine. 2008;33:2459–2463. https://doi.org/10.1097/BRS. 1590/bjpt-rbf.2014.0153 0b013e3181849dbe 20. Lomas-Vega R, Garrido-Jaut MV, Rus A, Del-Pino-Casado R. Effectiveness of Global 32. Vigatto R, Alexandre NM, Correa Filho HR. Development of a Brazilian Portuguese Postural Re-education for Treatment of Spinal Disorders: A Meta-analysis. Am J version of the Oswestry Disability Index: cross-cultural adaptation, reliability, and Phys Med Rehabil. 2017;96:124–130. https://doi.org/10.1097/PHM.00000000000 validity. Spine. 2007;32:481–486. https://doi.org/10.1097/01.brs.0000255075. 00575 11496.47 21. Castagnoli C, Cecchi F, Del Canto A, Paperini A, Boni R, Pasquini G, et al. Effects in 33. Abreu AM, Faria CD, Cardoso SM, Teixeira-Salmela LF. Versão brasileira do Fear Short and Long Term of Global Postural Reeducation (GPR) on Chronic Low Back Avoidance Beliefs Questionnaire. Cad Saude Publica. 2008;24:615–623. https://doi. Pain: A Controlled Study with One-Year Follow-Up. Sci World J. 2015;2015:271436. org/10.1590/s0102-311x2008000300015 https://doi.org/10.1155/2015/271436 34. Monticone M, Frigau L, Vernon H, Rocca B, Giordano A, Vullo SS, et al. Reliability, 22. Bonetti F, Curti S, Mattioli S, Mugnai R, Vanti C, Violante FS, et al. Effectiveness of a responsiveness and minimal clinically important difference of the two Fear ‘Global Postural Reeducation’ program for persistent low back pain: a non- Avoidance and Beliefs Questionnaire scales in Italian subjects with chronic low randomized controlled trial. BMC Musculoskelet Disord. 2010;11:285. https://doi. back pain undergoing multidisciplinary rehabilitation. Eur J Phys Rehabil Med. org/10.1186/1471-2474-11-285 2020;56:600–606. https://doi.org/10.23736/S1973-9087.20.06158-4 23. Costa LO, Maher CG, Latimer J, Hodges PW, Herbert RD, Refshauge KM, et al. Motor 35. Kamper SJ, Maher CG, Mackay G. Global rating of change scales: a review of control exercise for chronic low back pain: a randomized placebo-controlled trial. strengths and weaknesses and considerations for design. J Man Manip Ther. Phys Ther. 2009;89:1275–1286. https://doi.org/10.2522/ptj.20090218 2009;17:163–170. https://doi.org/10.1179/jmt.2009.17.3.163 24. Moher D, Hopewell S, Schulz KF, Montori V, Gøtzsche PC, Devereaux PJ, et al. 36. Ferreira ML, Ferreira PH, Latimer J, Herbert RD, Hodges PW, Jennings MD, et al. CONSORT 2010 Explanation and Elaboration: Updated guidelines for reporting Comparison of general exercise, motor control exercise and spinal manipulative parallel group randomised trials. J Clin Epidemiol. 2010;63:e1–e37. https://doi.org/ therapy for chronic low back pain: A randomized trial. Pain. 2007;131(1–2):31–37. 10.1016/j.jclinepi.2010.03.004 https://doi.org/10.1016/j.pain.2006.12.008 25. Deyo RA, Dworkin SF, Amtmann D, Andersson G, Borenstein D, Carragee E, et al. 37. Perret C, Poiraudeau S, Fermanian J, Colau MM, Benhamou MA, Revel M. Validity, Report of the NIH Task Force on research standards for chronic low back pain. reliability, and responsiveness of the fingertip-to-floor test. Arch Phys Med Rehabil. J Pain. 2014;15:569–585. https://doi.org/10.1016/j.jpain.2014.03.005 2001;82:1566–1570. https://doi.org/10.1053/apmr.2001.26064 26. Vibe Fersum K, O’Sullivan P, Skouen JS, Smith A, Kvåle A. Efficacy of classification- 38. Ekedahl H, Jönsson B, Frobell RB. Fingertip-to-floor test and straight leg raising based cognitive functional therapy in patients with non-specific chronic low back test: validity, responsiveness, and predictive value in patients with acute/subacute pain: a randomized controlled trial. Eur J Pain. 2013;17:916–928. https://doi.org/10. low back pain. Arch Phys Med Rehabil. 2012;93:2210–2215. https://doi.org/10.1016/ 1002/j.1532-2149.2012.00252.x j.apmr.2012.04.020 27. Souchard PE. Fundamentos do SGA: RPG a serviço do esporte. 2nd ed. Rio de Janeiro: 39. Amico KR. Percent total attrition: a poor metric for study rigor in hosted inter- É Realizações; 2004. vention designs. Am J Public Health. 2009;99:1567–1575.
Journal of Physiotherapy 69 (2023) 79–92 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Research Some magnetic resonance imaging findings may predict future low back pain and disability: a systematic review Christopher S Han a, Christopher G Maher a, Daniel Steffens b,c, Ashish Diwan d, John Magnussen e, Emma C Hancock f, Mark J Hancock g a Institute for Musculoskeletal Health, The University of Sydney and Sydney Local Health District, Sydney, Australia; b Surgical Outcomes Research Centre (SOuRCe), Royal Prince Alfred Hospital, Sydney, Australia; c Central Clinical School, Faculty of Medicine and Health, The University of Sydney, Sydney, Australia; d Spine Labs, St. George and Sutherland Clinical School, University of New South Wales, Sydney, Australia; e Department of Clinical Medicine, Faculty of Medicine, Health and Human Sciences, Macquarie University, Sydney, Australia; f University of New South Wales, Sydney, Australia; g Faculty of Medicine, Health and Human Sciences, Macquarie University, Sydney, Australia KEY WORDS ABSTRACT Magnetic resonance imaging Questions: Do magnetic resonance imaging (MRI) findings predict future low back pain (LBP), associated Predict disability and global recovery in people with current LBP? Do MRI findings predict these outcomes in people Diagnosis with no current LBP? Do MRI findings predict these outcomes in a mixed sample of people with and without Low back pain current LBP? Design: This review is an update of a previous systematic review investigating the relationship Musculoskeletal between lumbar spine MRI findings and future LBP. Participants: People with or without LBP with lumbar MRI scans. Outcome measures: MRI findings, pain and disability. Results: Of the included studies, 28 re- ported on participants with current LBP, eight reported on participants with no LBP and four reported on a mixed sample. Most results were based on single studies and did not demonstrate clear relationships be- tween MRI findings and future LBP. In populations with current LBP, pooling demonstrated that the presence of Modic type 1 changes alone or Modic type 1 and 2 changes were associated with slightly worse pain or disability outcomes in the short term, and the presence of disc degeneration was associated with worse pain and disability outcomes in the long term. In populations with current LBP, pooling demonstrated no evidence of an association between the presence of nerve root compression and disability outcomes in the short term, and no evidence of an association between the presence of disc height reduction, disc herniation, spinal stenosis, high-intensity zone and clinical outcomes in the long term. In populations with no LBP, pooling demonstrated that the presence of disc degeneration may increase the likelihood of experiencing pain in the long term. In mixed populations, no pooling was possible; however, single studies demonstrated that Modic type 1, 2 or 3 changes and disc herniation were each associated with worse pain in the long term. Conclusions: The results suggest that some MRI findings may have weak associations with future LBP; however, larger high-quality studies are needed to resolve uncertainty. Systematic review registration: PROSPERO CRD42021252919. [Han CS, Maher CG, Steffens D, Diwan A, Magnussen J, Hancock EC, Hancock MJ (2023) Some magnetic resonance imaging findings may predict future low back pain and disability: a systematic review. Journal of Physiotherapy 69:79–92] © 2023 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Introduction imaging (MRI) findings and their relevance to LBP could potentially be used to define subgroups that follow a different natural course and Low back pain (LBP) continues to be the leading cause of years to moderate the effects of specific treatments. lived with disability, according to The Global Burden of Disease Study.1 After excluding specific causes of LBP (eg, radiculopathy) and The importance of various MRI findings and their link to the serious pathologies (eg, lumbar spine fracture), the majority (90 to aetiology of LBP has been challenged, as structural changes on im- 95%) of people experiencing LBP are classified as having non-specific aging can be present in both symptomatic and asymptomatic people.3 LBP.2 This label acknowledges the inability to identify the specific cause of most LBP.2 One view is that if the nociceptive source of pain For example, one review (14 cross-sectional studies, n = 3,097) found could be identified, more specific and logical treatments could be developed and provided. A potential group of features that could be that disc degeneration on imaging was present in 57% of symptomatic used with this approach are structural changes observed on lumbar people, but also present in 34% of asymptomatic people.4 However, it spine imaging. A better understanding of magnetic resonance is important to note that the prevalence of disc degeneration was higher in people with a history of LBP (compared with those with no history of LBP), as was the prevalence of disc bulge, disc extrusion, disc protrusion, Modic type 1 changes and spondylolysis.4 Further https://doi.org/10.1016/j.jphys.2023.02.007 1836-9553/© 2023 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).
80 Han et al: MRI prediction of low back pain complicating interpretation of MRI findings is that the prevalence of Given that the review by Steffens et al6 used a follow-up time most imaging findings increases substantially with age.4 point of 3 months, all studies excluded at the full-text stage were searched to include any studies previously excluded for having a A limitation of most previous research studies investigating the follow-up time point , 3 months. It is unlikely that the review by relationship between structural changes on MRI and LBP is that Steffens et al6 would have excluded studies during the title and ab- they use cross-sectional5,6 designs, which only provide weak evi- stract screening stage when assessing study eligibility based on dence of a causal relationship.6 Most imaging studies have also minimum follow-up time points. been limited by small samples, and inconsistent measures of MRI findings and LBP status. Acknowledging the limitations of cross- When studies reported multiple measures from the domain of sectional designs, other research studies have used longitudinal pain outcomes, the Numeric Pain Rating Scale or Visual Analogue designs to investigate the relationship between MRI findings and Scale was preferentially used.7 For disability, any version of the future LBP. Oswestry Disability Index was preferentially used if available. If the Oswestry Disability Index was not reported, data were extracted Our previous systematic review investigating the relationship from the Roland Morris Disability Questionnaire.7 If neither were between MRI findings and future LBP found that definitive conclu- used, whichever disability outcome measure was reported was sions were not possible due to limitations such as a paucity of high- extracted. For global measure of recovery status, whichever outcome quality studies and significant heterogeneity between studies.6 measure was reported was extracted. Outcome data were re-scaled However, single studies in our previous review did report signifi- to a common 0-to-100 scale to facilitate between-study cant associations between Modic type 1 changes and future pain and comparisons. between disc degeneration and future disability, in samples with current LBP.6 Very few studies in our previous review investigated Association reported: The association between MRI findings and whether MRI findings predicted LBP in populations that were pain LBP outcomes must have been reported. If associations were not re- free at baseline.6 Given that our review is now 8 years old, an update ported, sufficient raw data to calculate a measure of association must of the review with recently published longitudinal studies was have been reported. Reporting just a p-value was insufficient to meet performed. this criterion. Therefore, the specific research questions for this systematic re- Exclusion criteria view were: Studies including participants with serious or specific pathol- 1. Do MRI findings predict future LBP, associated disability and global ogies such as tumours, fractures, inflammatory arthritis and cauda recovery in people with current LBP? equina were excluded. One reviewer (CSH) screened all titles to exclude irrelevant studies. Two reviewers (CSH and ECH) inde- 2. Do MRI findings predict these outcomes in people with no current pendently screened abstracts to exclude irrelevant studies. Two LBP? reviewers (CSH and ECH) then independently reviewed full texts for eligibility based on the inclusion criteria. Any discrepancies 3. Do MRI findings predict these outcomes in a mixed sample of were resolved through discussion and a third reviewer (CGM or people with and without current LBP? MJH). Relevant authors were contacted via email if any additional information was required. Methods Data extraction A review protocol was specified in advance and registered on PROSPERO (CRD42021252919). All relevant data were extracted independently by two reviewers. The data were categorised into two groups based upon whether study Search strategy participants had pain at baseline. In the mixed sample, if a clear majority ( 95%) of participants had either pain or no pain at base- MEDLINE, CINAHL and EMBASE were searched for the period line, the data were categorised based on the pain status of the ma- January 2012 to 24 November 2021 with the search strategy used in jority. Data extracted from the included studies were number of our earlier review.6 The complete search strategies from all databases participants, participant age, presence of pain at baseline, follow-up are presented in Appendix 1 on the eAddenda. The reference lists of duration, MRI findings, outcome measures, and strength of associa- identified papers were examined and forward citation searching was tions between MRI findings and outcome measures. When sufficient also performed. raw data were available, odds ratios and 95% confidence intervals were calculated. Inclusion criteria Methodological quality appraisal Studies were required to meet all the following criteria to be included: We originally intended to assess the methodological quality of each study using the QUality In Prognostic Studies tool; however, Longitudinal cohort study: This included secondary analyses of the Altman8 checklist was used as it was simpler to use and also randomised controlled trials. Trials where the intervention provided used in our previous review. The methodological quality of each was designed to alter lumbar spine MRI findings (eg, surgery or study was assessed using the standardised checklist of pre-defined antibiotic treatment for Modic changes) were excluded, unless the criteria9,10 used in the review by Steffens et al.6 This checklist is a data from the other treatment groups could be extracted separately. modified version based on theoretical considerations and method- We reviewed all studies excluded by the previous study authors at ological aspects described by Altman et al.8 These criteria are pre- the full-text stage from the review by Steffens et al6 to include any sented in Box 1. studies previously excluded for being retrospective studies. Two reviewers independently assessed each included study. Each MRI undertaken: Participants must have undergone a baseline MRI criterion was scored positive ‘yes’ or negative ‘no’. A positive score and had at least one lumbar spine finding reported; for example, but indicated sufficient information and a positive assessment. Any dis- not limited to, disc degeneration, disc herniation, disc protrusion, agreements in scores were resolved through discussion or a third facet joint arthropathy, Modic changes or high-intensity zone. reviewer (CGM or MJH). No overall methodological quality score was used as an inclusion criterion. Outcome measure relevant to LBP reported: Studies must have reported a clinically relevant outcome measure at follow-up, such as pain, disability and a global measure of recovery status. A follow-up time point of 4 weeks was accepted for both symptomatic and asymptomatic population groups.
Research 81 Box 1. Criteria used for appraisal of methodological quality. definition of study sample (description of participant source and inclusion and exclusion criteria) representative sample of the target sample (participants selected by random selection or as consecutive cases) follow-up rate 80% (outcome data were available for 80% of participants at 3-month follow-up or later) adequate follow-up time ( one prognostic outcome was followed up at 3 months or later) interpretable prognostic outcomes available (raw data, percentages, survival rates or continuous outcome reported) blinding (assessor unaware of one prognostic factor used to predict prognostic outcome, at the time prognostic outcome was measured) Data analysis Figure 1. Flow of studies through the review. LBP = low back pain, MRI = magnetic resonance imaging. Following data extraction, study level associations were collated. Data were pooled using meta-analysis with a random effects model primary reasons for exclusion are presented in Appendix 2 on the when possible. Commercial softwarea was used for meta-analysis. eAddenda. Due to the heterogeneity between studies (ie, MRI findings, clinical outcome measures, statistical measure used to assess association of MRI finding), some results were also presented descriptively. Study results were reported and analysed separately based on the presence or absence of pain at baseline. All positive scores for continuous measures (eg, mean difference) indicated that the exposure group had a worse outcome at follow-up compared with the comparison group. Odds ratios or hazard ratios . 1 indicated greater likelihood of poorer outcome in the exposure group compared with the comparison group. Short-term outcomes were defined as a follow-up time point , 12 months, and the closest time point to 3 months was taken. Long-term outcomes were defined as a follow-up time point 12 months, and the closest time point to 12 months was taken. Associations were defined as: ‘no evidence of an association’ if the effect sizes were small (MD , 10 points on a 100-point scale, OR/HR , 1.5) and confidence intervals spanned no association (or where confidence intervals were not reported, the result was not statistically significant). If the effect sizes were small but the con- fidence intervals did not span no association (or the result was statistically significant), this was defined as a ‘weak association’ and the term ‘slightly’ was used to imply a small effect. If the effect size was moderate to large (MD .10 points on a 100-point scale, OR/HR . 1.5) and confidence intervals did not span no association (or the result was statistically significant), this was defined as having an ‘association’. If the effect size was moderate to large but confidence intervals spanned no association (or the result was not statistically significant), the term ‘may’ was used to imply a potential association. Results Management of missing information Selection of studies Five authors of conference abstracts51–55 or protocols56 were contacted for full-text articles, and two full texts were received.52,57 After removing duplicates (n = 976) the search identified 9,949 Six authors14,17,21,26,47,58 were contacted regarding missing data and/ citations. Of these, 6,919 were deemed clearly irrelevant based on or discrepancies; three14,17,21 were able to provide additional data. screening by title and a further 2,882 studies were excluded based on title and abstract. Two studies11,12 originally excluded in the Methodological quality Steffens et al review6 were included for full-text screening as they potentially met the criteria. Following full-text review by two in- The results of the methodological quality assessment for all 39 dependent authors, 27 new studies13–39 and the 12 studies from the studies are presented in Table 1. Thirty-six studies provided an previous review5,40–50 were included for analysis (Figure 1). One adequate description of the study sample. Nineteen studies reported study presented data separately for populations with current LBP selecting participants by random selection or as consecutive cases. and populations with no LBP;25 therefore, this study was separated Twenty-six studies reported . 80% follow-up. All studies reported into two strata for analysis: Kasch 2022a and Kasch 2022b.25 A total adequate follow-up time. All studies reported a prognostic outcome of 39 studies were included for analysis. Excluded studies and the at 3 months. Twenty-nine studies reported assessor blinding.
82 Han et al: MRI prediction of low back pain Table 1 Methodological quality assessment of included studies. Author, year Definition of Representative Follow-up Adequate Interpretable prognostic Blindingf study samplea sampleb rate 80%c follow-up timed outcomes availablee Yes Boos, 200040 Yes No Yes Yes Yes Yes Borenstein, 20015 Yes No No Yes Yes No Elfering, 200243 No No Yes Yes Yes Yes Carragee, 200542 Yes Yes Yes Yes Yes No Jarvik, 200545 Yes Yes Yes Yes Yes Yes Modic, 200550 Yes No Yes Yes Yes Yes Carragee, 200641 Yes Yes Yes Yes Yes Yes Kleinstuck, 200648 Yes No Yes Yes Yes Yes McNee, 201149 Yes Yes No Yes Yes Yes Hellum, 201244 Yes No Yes Yes Yes Yes Jensen, 201246 Yes No Yes Yes Yes Yes Keller, 201247 Yes No No Yes Yes No Wilkens, 201339 Yes Yes Yes Yes Yes No el Barzouhi, 201418 Yes No Yes Yes Yes Yes Jensen, 201423 Yes No Yes Yes Yes Yes Peterson, 201432 Yes Yes No Yes Yes Yes Schistad, 201435 Yes No Yes Yes Yes Yes Hancock, 201520 Yes Yes Yes Yes Yes Yes Jarvinen, 201522 Yes Yes Yes Yes Yes Yes Jensen, 201524 Yes Yes Yes Yes Yes No Kerr, 201526 Yes No No Yes Yes Yes Määttä, 201531 Yes No No Yes Yes Yes Annen, 201614 Yes Yes Yes Yes Yes Yes de Schepper, 201617 Yes Yes Yes Yes Yes No Suri, 201636 Yes Yes Yes Yes Yes Yes Hancock, 201719 Yes Yes Yes Yes Yes Yes Koivisto, 201728 Yes No Yes Yes Yes Yes Romero-Munoz, 201733 No Yes Yes Yes Yes Yes Tonosu, 201737 Yes No No Yes Yes Yes Cai, 201815 Yes No Yes Yes Yes Yes Annen, 201813 Yes Yes No Yes Yes No Konstantinou, 201829 Yes Yes No Yes Yes No Chen, 201916 Yes No Yes Yes Yes Yes Kesikburun, 201927 Yes Yes No Yes Yes No Harrisson, 201954 Yes No No Yes Yes No Kuligowski, 202030 Yes No Yes Yes Yes Yes Sääksjärvi, 202034 Yes No No Yes Yes Yes Kasch, 202225 No Yes No Yes Yes Yes Udby, 202138 Yes Yes Yes Yes Yes a Description of participant source and inclusion and exclusion criteria. b Participants selected by random selection or as consecutive cases. c Outcome data were available for 80% of participants at 3-month follow-up or later. d At least one prognostic outcome was followed up at 3 months or later. e Raw data, percentages, survival rates or continuous outcome reported. f Assessor of MRI unaware of one prognostic factor, used to predict prognostic outcome, at the time prognostic outcome was measured. Study characteristics to be pooled for 19 different comparisons (which included 55 asso- A full description of included studies is presented in Table 2. ciations from individual studies) in total (Table 5). Only results of Twenty-eight studies13–18,21–30,32,33,35,36,38,39,44,46–50 included partici- pants with current LBP, eight studies5,19,20,25,37,40,43,45 included par- associations investigating different types of Modic changes compared ticipants with no LBP and four studies31,34,41,42 enrolled a mixed participant sample. The mean participant age ranged from 29 to 62 with no Modic changes for pain or disability outcomes are presented years. The follow-up period ranged from 4 weeks to 30 years. below. Results of all pooled associations are presented in Table 5. Association of MRI findings and clinical outcomes in people with current LBP Any Modic change: Pooling demonstrated those with Modic Twenty-eight studies13–18,21–30,32,33,35,36,38,39,44,46–50 reported a changes (type 1 and 2, type 1, and type 2) compared with no Modic total of 377 associations between MRI findings and clinical outcomes change had slightly worse pain (MD 6.6, 95% CI 4.3 to 8.9)13–16,32,35 in participants with current LBP (Appendix 3 on the eAddenda). The and disability outcomes (MD 4.6, 95% CI 0.3 to 8.9)14,16 in the short MRI findings presented in three or more studies or where associa- tions could be pooled across at least two studies are presented below. term. In the long term, pooling demonstrated no evidence of an as- Results of associations between MRI findings and pain or disability outcomes are presented in Tables 3 and 4, respectively. All MRI sociation between Modic changes (Modic type 1 and Modic type 2) findings and outcomes are presented in Appendix 3. compared with no Modic changes for pain (MD 0.0, 95% CI 26.5 to Association between Modic changes and pain and disability 6.4)14,23–25,35,47 and disability outcomes (MD 20.3, 95% CI 23.7 to Nineteen studies13–16,18,22–25,28,32,33,35,38,39,44,46–48 assessed the 3.2)14,23,39,47 (Table 5). associations between Modic changes (type 1, type 2, type 3, or Modic type 1: Pooling demonstrated that those with Modic type 1 type 1 and/or 2) and pain13–16,18,22–25,28,32,33,35,46–48 or disability14,16,22,23,28,33,38,39,44,47,48 (Tables 3 and 4). Results were able changes compared with no Modic changes had slightly worse pain outcomes (MD 4.4, 95% CI 1.1 to 7.6);13–16,35 however, there was no evidence of an association for disability outcomes (MD 3.1, 95% CI 24.3 to 10.5)14,16 in the short term. In the long term, pooling demonstrated no evidence of an association between Modic type 1 changes compared with no Modic changes and pain (MD 1.2, 95% CI 29.1 to 11.4)14,23–25,35,47 and disability outcomes (MD 22.4, 95% CI 28.8 to 4.0).14,23,39,47 Pooling also demonstrated no evidence of an association between Modic type 1 changes compared with no Modic changes and the presence of pain (OR 1.1, 95% CI 0.5 to 2.4)18,46 in the long term.
Research 83 Table 2 Individual study characteristics. Study Participants Follow-up duration Clinical outcome MRI finding N (% female) Age (y), mean (SD)a Canal stenosis, nerve root Modic, 200550 246 (58%) 43 (NR) Populations with current LBP compression, disc herniation Kleinstuck, 200648 53 (49%) 44 (not reported) 2 y Disability (RMDQ) High-intensity zone, disc bulge, disc degeneration, Modic type 1, Modic 1 y Pain last 2 wk (VAS) type 2 Disability (RMDQ) Number of MRI abnormalities, disc degeneration McNee, 201149 240 (54%) NR (range 20 to 60) 22 mo (mean) Pain . 14 d in past 4 wk Modic type 1, Modic type 2, disc Hellum, 201244 66 (58%) 42 (NR) 1y Disability past 4 wk (RMDQ) degeneration (height reduction), disc degeneration (signal intensity), facet Disability (ODI) joint arthropathy, high-intensity zone Jensen, 201246 96 (69%) 46 (range 21 to 60) 14 mo Pain (NPRS) Modic type 1, Modic type 1 (change Keller, 201247 269 (50%) 50 (range 20 to 60) in size) Wilkens, 201339 250 (48%) 1y Patient global impression of Modic type 1, Modic type 2 49 (11) improvement (PGI-I) el Barzouhi, 201418 1y High-intensity zone, Modic type 1, 95 (29%) 44 (10) Disability (RMDQ) Modic type 2, and Modic type 1, 2 or 4 wk 3 Jensen, 201423 141 (53%) 42 (11) 3 mo Disabling LBP 40/100 mm Vertebral endplate signal changes 6 mo Pain (VAS) type 1, vertebral endplate signal 1y changes type 2, vertebral endplate 1y Pain (LBP rating scale) signal changes type 1 and 2 Disc nucleus signal, lumbar Peterson, 201432 345 (50%) 60 (NR) 1 mo Patient impression of change osteophytes, disc height reduction, Schistad, 201435 235 (47%) 41 (11) (PGIC) high-intensity zone, disc herniation, 6 wk nerve root compression, spina Jarvinen, 201522 64 (87%) 44 (10) 6 mo Pain (VAS) stenosis, Modic type 1, Modic type 2, Jensen, 201524 96 (69%) 46 (range 21 to 60) 1y Pain experience (McGill sensory and Modic type 1 and 2 2y Modic type 1 and 2 pain score) 14 mo Modic type 1, Modic type 2 and 3 Pain (NRS) (Disability) ODI Change in Modic type 1 extent, change in Modic type 2 extent Change in pain (NRS) Modic type 1, large Modic change ( 25% vertebral height), large Modic Kerr, 201526 392 (41%) 44 (12) 8y Disability (ODI) type 1, severe disc degeneration ( 3 Annen, 201614 72 (23%) 42 (11) Pfirrmann), large disc herniation 1 mo Improvement (PGIC) Posterolateral herniation, other de Schepper, 201617 683 (47%) 50 (13) 3 mo Back pain (NRS) herniation location, extrusion, Suri, 201636 341 (62%) 44 (12) 6 mo Leg pain (NRS) protrusion, sequestration 1y Disability (ODI) Modic type 1, Modic type 2, Modic Koivisto, 201728 20 (45%) 52 (NR) 1y type 1 and 2 Romero-Munoz, 201733 70 (66%) 57 (range 43 to 67) Recovery (strongly improved/ 4y completely improved on GPE) Disc bulging, disc herniation, nerve Annen, 201813 112 (42%) 45 (16) root compression, spinal stenosis, 51 wk Back pain recurrence (LBP spondylolisthesis Cai, 201815 37 (39%) 62 (11) 10 y bothersome) Extrusion, sequestration, posterolateral disc herniation 1 mo Leg pain recurrence (Sciatic both- 3 mo ersome index) Change in Modic type 1 extent, 6 mo change in Modic type 2 extent Change in LBP intensity (VAS) Modic type 1, Modic type 2, Modic Disability (ODI) type 3 Back pain (VAS) Modic type 1, Modic type 2, and Leg pain (VAS) Modic type 1 and 2 Disability (ODI) Mild disc degeneration, severe disc Pain (NRS) degeneration, Modic type 1 Bournemouth Questionnaire Improvement (PGIC) Pain (VAS)
84 Han et al: MRI prediction of low back pain Table 2 (Continued) Study Participants Follow-up duration Clinical outcome MRI finding N (% female) Age (y), mean (SD)a Konstantinou, 201829 609 (63%) 50 (14) 4 mo Disability (RMDQ) Nerve root compression 12 mo Chen, 201916 129 (57%) 42 (range 21 to 67) 3 mo Pain (VAS) Modic type 1, Modic type 2 6 mo Disability (ODI) Kesikburun, 201927 40 (45%) 50 (14) 17 mo (SD 7) Pain (NRS) Disc herniation Disability (ODI) Harrisson, 201954 402 (63%) 50 (14) 4 mo Mean pain intensity (0 to 10) Nerve root compression Kuligowski, 202030 30 (60%) 29 (3) 1y Protrusion, extrusion 3y Pain (NRS) Disability (ODI) 4 wk Kasch, 2022a 1,997 (54%) 53 (14) 6 y Pain (NRS) Disc degeneration, disc height loss, disc herniation, high-intensity zone, Udby, 202138 170 (54%) 53 (range 31 to 70) 13 y Disability (RMDQ) spondylolisthesis, Modic type 1, Modic type 2 or 3, hypertrophy ligamentus flava, Schmorl lesion, spinal canal stenosis, 1 MRI finding, 2 MRI findings, 3 MRI findings, 4 MRI findings, 5 MRI findings Modic change (1, 2, 3), disc degeneration, facet joint degeneration Populations with no LBP Boos, 200040 46 (NR) NR (range 20 to 50) 62 mo (mean) Pain (NQ) Disc herniation, nerve root contact, disc degeneration Borenstein, 20015 67 (NR) 42 (NR) 7 y Pain (ordinal scale) MRI abnormalities (change) Elfering, 200243 41 (27%) NR (range 20 to 50) 62 mo (mean) Pain (NQ) Disc degeneration (change) Jarvik, 200545 148 (11%) 54 (range 36 to 71) 3y Pain (PFI) Disc herniation, nerve root contact, canal stenosis Hancock, 201520 76 (39%) 46 (13) 1y LBP recurrence (return of LBP Disc degeneration 3, high- Hancock, 201719 76 (39%) 46 (13) 52 wk lasting at least 24 h, pain intensity intensity zone, Modic change, disc 3/10 NRS) herniation, facet joint arthrosis, Tonosu, 201737 49 (51%) 45 (9) 10 y spondylolisthesis Recurrence of activity limited LBP Days of LBP recurrence (adapta- Pfirrmann DD score 3, disc height loss, disc annular tear, high-intensity tion of item 8 of SF-36) zone, disc herniation, spondylolisthesis, facet joint Back pain in the past 1 mo osteoarthritis, Modic change, and 1, 2 or 3, aggregate MRI findings Disc degeneration, disc bulge, high- intensity zone Kasch, 2022b 1,366 (46%) 53 (14) 6 y Pain (NRS) Disc degeneration, disc height loss, disc herniation, high-intensity zone, spondylolisthesis, Modic type 1, Modic type 2 or 3, hypertrophy ligamentus flava, Schmorl lesion, spinal canal stenosis, 1 MRI finding, 2 MRI findings, 3 MRI findings, 4 MRI findings, 5 MRI findings Populations with a mixed sample Carragee, 200542 100 (37%) 42 (NR) 63 mo (mean) Pain (remission) Disc herniation Carragee, 200641 200 (40%) 39 (NR) 5 y Pain (NRPS) Disc degeneration, endplate changes, Disability (ODI) canal stenosis Määttä, 201531 429 (98%) 54 (NR) 10 y Disabling low back pain lasting DD summary score (0 to 60), Schmorl Sääksjärvi 2020,34 35 (0%) 51 (1) 30 y . 1 mo node, Modic change Pain (VAS) Single intensity decrease . 60%, Disability (ODI) Modic change, disc protrusion or bulge, severe body endplate lesion, disc height reduction, annular fissure, high-intensity zone GPE = global perceived effect, LBP = low back pain, MRI = magnetic resonance imaging, NPRS = numeric pain rating scale, NQ = Nordic questionnaire, NRS = numeric rating scale, ODI = Oswestry Disability Index, PFI = Pain Frequency Index, RMDQ = Roland Morris Disability Questionnaire, SF-36 = 36 Item Short Form Survey, VAS = visual analogue score. a Unless otherwise stated. Fourteen studies14,18,22–25,28,33,35,38,39,44,47,48 investigated the in outcome measures or measures of association (Tables 3 and 4). remaining 46 associations between Modic change (any type) Point estimates in 40 of the 46 associations demonstrated no evi- compared with no Modic changes for pain or disability outcomes in dence of an association between Modic changes (type 1 and 2, type 1 populations with LBP. Pooling was not possible due to heterogeneity alone, type 2 alone, or type 3 alone) compared with no Modic
Research 85 Table 3 Difference in means (95% CI)b Association between MRI findings and clinical outcomes in populations with current low back pain in the short term. 0.1 (p = 0.6)d Exposure Comparison Clinical outcome (follow up)a 0.2 (p = 0.1)d 0.1 (p = 0.6)d Disc bulge48 No disc bulge Back pain (3 mo)c –8.5 (–23.4 to 6.4)e Disability (6 wk)c 0.1 (p = 0.7)d Disc bulge48 No disc bulge Back pain (3 mo)c 2.7 (1.4 to 5.1) (OR) Disc degeneration48 No disc degeneration Back pain (6 mo) 30.0 (NR)f Disc degeneration (severe)15 Disc degeneration (mild) Disability (3 mo)c 18.0 (NR)f Disc degeneration48 No disc degeneration Disability (6 wk) –0.1 (p = 0.4)d Disc herniation50 No disc herniation –0.2 (p = 0.1)d Back pain (1 mo) Disc protrusion30 Disc extrusion Disability (1 mo) –4.0 (–22.6 to 14.6) Disc protrusion30 Disc extrusion Back pain (3 mo)c 5.0 (1.4 to 8.6) Disability (3 mo)c High intensity zone48 No high intensity zone –1.6 (–16.2 to 13.0) High intensity zone48 No high intensity zone Back pain (3 mo) 7.0 (–5.3 to 19.3) Back pain (3 mo) Modic type 114 No Modic Back pain (3 mo) 0.3 (–15.7 to 16.3)e Modic type 116 No Modic Back pain (6 mo) –8.0 (–27.9 to 11.9) Modic type 113 No Modic Back pain (6 mo) 9.3 (–2.2 to 20.9) Modic type 135 No Modic Leg pain (3 mo) Modic type 115 No Modic Disability (3 mo) 0.8 (–1.2 to 2.8) Modic type 114 No Modic Disability (3 mo) Modic type 114 No Modic –7.7 (–26.9 to 11.5) Modic type 116 No Modic Back pain (3 mo) –5.0 (–9.8 to –0.2) Back pain (3 mo) –5.7 (–22.9 to 11.5) Modic type 114 Modic type 2 Back pain (3 mo) –18.4 (–43.9 to 7.1) Modic type 116 Modic type 2 Leg pain (3 mo) –1.3 (–16.5 to 13.8) Modic type 113 Modic type 2 Disability (3 mo) –5.0 (–7.5 to –2.5) Modic type 114 Modic type 2 Disability (3 mo) 8.0 (–1.0 to 17.0) Modic type 114 Modic type 2 Back pain (6 mo) Modic type 116 Modic type 2 7.3 (1.1 to 13.5) Modic type 135 Modic type 2 and 3 Back pain (1 mo) 0.7 (–13.2 to 14.6) Back pain (3 mo) Modic type 1 and 232 No Modic Back pain (3 mo) 8.0 (4.5 to 11.5) Modic type 1 and 214 No Modic Back pain (3 mo) 1.3 (–12.2 to 14.8) Modic type 1 and 216 No Modic Leg pain (3 mo) 2.7 (–12.9 to 18.3) Modic type 1 and 213 No Modic Disability (3 mo) 10.1 (0.3 to 20.0) Modic type 1 and 214 No Modic Disability (3 mo) Modic type 1 and 214 No Modic Back pain (3 mo)c 3.8 (2.0 to 5.7) Modic type 1 and 216 No Modic Disability (3 mo)c –0.3 (p = 0.2)d Modic type 1 and/or 248 No Modic 0.1 (p = 0.5)d Modic type 1 and/or 248 No Modic Back pain (3 mo) Back pain (3 mo) 3.7 (–11.7 to 19.1) Modic type 214 No Modic Back pain (3 mo) 10.0 (6.0 to 14.0) Modic type 216 No Modic Leg pain (3 mo) 4.1 (–12.8 to 21.0) Modic type 213 No Modic Disability (3 mo) 10.4 (–9.4 to 30.2) Modic type 214 No Modic Disability (3 mo) 10.7 (–0.4 to 21.7) Modic type 214 No Modic Back pain (6 mo) Modic type 216 No Modic 5.8 (3.8 to 7.8) Modic type 2 and 335 No Modic Back pain (4 mo) –1.0 (–8.1 to 6.1) Disability (4 mo) NRC21 No NRC Disability (6 wk) –3.0 (–13.0 to 7.0) NRC29 No NRC (all patients) 0.9 (0.4 to 2.0) (OR) NRC50 No NRC Disability (6 wk) 1.3 (0.7 to 2.5) (OR) Spinal stenosis (severe)50 No spinal stenosis 0.4 (0.2 to 1.2) (OR) Positive outcome for difference in means indicates exposure group had a worse outcome at follow-up compared with comparison group. OR . 1 indicates greater odds of poor outcome in exposure group compared with comparison group. Shaded rows indicate change in exposure or change in comparison MRI finding. NR = not reported, NRC = nerve root compression. a Short-term follow-up; outcome , 12 months (closest time point to 3 months) taken. b Unless otherwise indicated. c Average pain in the past 2 weeks. d Beta coefficient. e Change in pain scores. f Z score. changes and pain or disability outcomes at any follow-up time point. nerve root compression compared with the absence of nerve root Two studies22,28 found a relationship between greater increases in compression and pain outcomes in the short and long term21 and the volume of Modic type 1 changes between baseline and follow-up, disability outcomes in the long term29 (Tables 3 and 4). and worse pain and disability outcomes over the same period (Table 4). Point estimates of the remaining four associations from two Association between disc degeneration and pain and disability studies24,44 suggested that there may be an association between Eight studies15,23–25,38,44,48,49 investigated 12 associations be- those with Modic type 1 changes alone24 or Modic type 1 and 244 changes and worse pain24 or disability44 outcomes, respectively, in tween disc degeneration and pain15,23–25,48 or disability23,38,44,48,49 the long term (Table 4). outcomes in populations with LBP. Studies were only able to be pooled for two comparisons (which included four Association between nerve root compression and pain and disability associations) 23,25,44,49 due to different measures of association re- Four studies21,23,29,50 investigated the association between nerve ported. Pooling demonstrated that those with disc degeneration compared with no disc degeneration had slightly worse pain out- root compression and pain21,23 or disability23,29,50 outcomes in pop- comes (MD 5.3, 95% CI 1.3 to 9.3)23,25 and worse disability outcomes ulations with LBP. Pooling demonstrated no evidence of an association (OR 2.1, 95% CI 1.3 to 3.6)44,49 in the long term. Point estimates of the between the presence of nerve root compression compared with no remaining eight associations did not consistently suggest a relation- compression and pain (MD 21.7, 95% CI 211.0 to 7.5)21,23 or disability ship between those with disc degeneration compared with no disc outcomes (OR 1.1, 95% CI 0.7 to 1.8)29,50 in the long and short term, degeneration and clinical outcomes; all of these associations had respectively. Results from associations based on two single studies21,29 confidence intervals that included no association (or where they were demonstrated no evidence of an association between the presence of not reported, were not statistically significant).
86 Han et al: MRI prediction of low back pain Table 4 Association between MRI findings and clinical outcomes in populations with current low back pain in the long term. Exposure Comparison Clinical outcome (follow up)a Difference in means (95% CI)b Disc bulge48 No disc bulge Back pain (1 y)c 0.19 (p = 0.15)d 0.22 (p = 0.11)d Disc bulge48 No disc bulge Disability (1 y) –0.18 (p = 0.25)d Disc degeneration48 No disc degeneration Back pain (1 y)c 0.0 (p = 1.0)d Disc degeneration (severe)24 No disc degeneration (severe) Back pain (14 mo) Disc degeneration23 No disc degeneration Back pain (1 y) 8.1 (–1.2 to 18.1) Disc degeneration25 No disc degeneration Back pain (6 y) 4.7 (0.3 to 9.1) Disc degeneration23 No disc degeneration Disability (1 y) 10.4 (–9.6 to 29.7) Disc degeneration48 No disc degeneration Disability (1 y)c –0.24 (p = 0.15)d Disc degeneration49 No disc degeneration Disability (18.5 mo)e 2.3 (1.3 to 4.0) (OR) Disc degeneration38 No disc degeneration Disability (13 y) 0.06 (p = 0.41)d Disc degenerationf,44 No disc degeneration Disability (2 y) 1.3 (0.3 to 5.3) (OR) Disc extrusion36 Disc protrusion Back pain (4 y)g 1.0 (0.6 to 1.6) (HR) Disc extrusion26 Disc protrusion Disability (8 y) –2.4 (–5.2 to 0.4)h Disc extrusion36 Disc protrusion Leg pain (4 y)f 0.7 (0.5 to 1.2) (HR) Disc extrusion26 Disc sequestration Disability (8 y) 1.2 (–3.6 to 6.0)h Disc extrusion (non-regression)27 Disc extrusion (partial regression) Back pain (17 mo) 11.0 (–12.1 to 34.1) Disc extrusion (non-regression)27 Disc extrusion (partial regression) Disability (17 mo) 9.8 (–13.4 to 33.3) Disc extrusion (non-regression)27 Disc extrusion (complete regression) Back pain (17 mo) 18.0 (–2.1 to 38.1) Disc extrusion (non-regression)27 Disc extrusion (complete regression) Disability (17 mo) 14.2 (–4.9 to 33.3) Disc extrusion (partial regression)27 Disc extrusion (complete regression) Back pain (17 mo) 7.0 (–10.8 to 24.8) Disc extrusion (partial regression)27 Disc extrusion (complete regression) Disability (17 mo) 4.4 (–12.1 to 20.9) Disc height reduction23 No disc height reduction Back pain (1 y) 4.3 (–4.7 to 13.4) Disc height reduction25 No disc height reduction Back pain (6 y) –4.8 (–10.2 to 0.5) Disc height reduction23 No disc height reduction Disability (1 y) 5.3 (–11.6 to 22.2) Disc height reductionf,44 No disc height reduction Disability (2 y) 1.4 (0.5 to 4.2) (OR) Disc herniation23 No disc herniation Back pain (1 y) 11.2 (–0.7 to 23) Disc herniation25 No disc herniation Back pain (6 y) –7.0 (–16.3 to 2.4) Disc herniation (large)24 No disc herniation (large) Back pain (14 mo) Disc herniation (posterolateral)36 Other disc herniation Back pain (4 y)g 0.7 (p = 0.3)d Disc herniation (posterolateral)36 Other disc herniation Leg pain (4 y)g 0.7 (0.4 to 1.1) (HR) Disc herniation (posterolateral)26 Disc herniation (other herniation location) Disability (8 y) 0.6 (0.4 to 0.9) (HR) Disc herniation23 No disc herniation Disability (1 y) –3.7 (–6.4 to –1.0)h 22.0 (–0.2 to 44.2) Disc sequestration36 Disc protrusion Back pain (4 y)g Disc sequestration26 Disc protrusion Disability (8 y) 0.6 (0.2 to 1.4) (HR) Disc sequestration36 Disc protrusion Leg pain (4 y)g –3.6 (–8.4 to 1.2)h 0.9 (0.4 to 1.9) (HR) Facet joint degeneration38 No facet joint degeneration Disability (13 y) –0.02 (p = 0.80)d Facet joint arthropathy grade 2/3f,44 No facet joint arthropathy grade 2/3 Disability (2 y) 1.3 (0.3 to 5.3) (OR) Lumbar spine osteoarthritis23 No lumbar spine osteoarthritis Back pain (1 y) Lumbar spine osteoarthritis23 No lumbar spine osteoarthritis Disability (1 y) –1.8 (–9.5 to 5.9) –0.5 (–14.8 to 13.8) High-intensity zone48 No high-intensity zone Back pain (1 y)c High-intensity zone23 No high-intensity zone Back pain (1 y) –0.38 (p = 0.01)d High-intensity zone25 No high-intensity zone Back pain (6 y) 4.2 (–4.2 to 12.6) High-intensity zone39 No high-intensity zone Disability (1 y) –0.3 (–4.3 to 3.8) High-intensity zone23 No high-intensity zone Disability (1 y) High-intensity zone48 No high-intensity zone Disability (1 y)c –1.1 (?) High-intensity zone (posterior)f,44 No posterior high-intensity zone Disability (2 y) 5.0 (–9.5 to 19.4) –0.08 (p = 0.59)d Hypertrophy ligamentous flava25 No hypertrophy ligamentous flava Back pain (6 y) 0.6 (0.2 to 1.8) (OR) Modic type 114 No Modic Back pain (1 y) 8.4 (–3.7 to 20.4) Modic type 147 No Modic Back pain (1 y)i Modic type 135 No Modic Back pain (1 y) –4.5 (–25.9 to 16.9) Modic type 123 No Modic Back pain (1 y) 2.0 (–6.2 to 10.3) Modic type 147 No Modic Back pain (1 y)i 1.0 (–12.0 to 14.0) Modic type 146 No Modic Back pain (14 mo) Modic type 118 No Modic Back pain (1 y) –14.5 (–27.2 to –1.8) Modic type 1k,24 No Modic Back pain (14 mo) 1.0 (0.8 to 1.2) (HR) Modic type 1 (large)24 No Modic Back pain (14 mo) 1.0 (0.4 to 2.4) (OR) Modic type 1 (large)k,24 No Modic Back pain (14 mo) 1.5 (0.1 to 15.9) (OR) Modic type 1 (large)j,24 No Modic Back pain (14 mo) Modic type 1 (large)m,24 No Modic Back pain (14 mo) 17.8 (2.6 to 33.0)l Modic type 124 No Modic Back pain (14 mo) 0.1 (p = 0.9)h Modic type 1j,24 No Modic Back pain (14 mo) Modic type 1m,24 No Modic Back pain (14 mo) –0.7 (–14.6 to 13.2)l Modic type 125 No Modic Back pain (6 y) 7.0 (–2.7 to 16.7)l Modic type 133 No Modic Back pain (10 y) 14.3 (0.8 to 27.9)l Modic type 114 No Modic Leg pain (1 y) –1.8 (p = 0.03)d Modic type 133 No Modic Leg pain (10 y) 22.8 (12.2 to 33.4)l Modic type 114 No Modic Disability (1 y) 27.8 (12.8 to 42.8)l Modic type 147 No Modic Disability (1 y) –3.0 (–13.4 to 7.4) Modic type 139 No Modic Disability (1 y) –5.0 (–23.6 to 13.6) Modic type 123 No Modic Disability (1 y) –4.8 (–24.8 to 15.2) Modic type 133 No Modic Disability (10 y) 11.0 (–9.7 to 31.7) Modic type 147 No Modic Disability (1 y) 2.9 (–10.6 to 16.5) Modic type 1i,44 No Modic Disability (2 y) 0.60 (–5.9 to 7.1) –5.2 (–13.8 to 3.4) Modic type 114 Modic type 2 Back pain (1 y) –22.5 (–47.1 to 2.1) Modic type 118 Modic type 2 Back pain (1 y) –8.6 (–28.0 to 10.8) Modic type 133 Modic type 2 Back pain (10 y) 1.00 (0.98 to 1.04) (HR) 0.7 (0.2 to 1.9) (OR) –1.3 (–28.4 to 25.8) 1.0 (0.1 to 10.1) (OR) –8.0 (–26.6 to 10.6)
Table 4 (Continued) Research Clinical outcome (follow up)a 87 Exposure Comparison Leg pain (1 y) Difference in means (95% CI)b Leg pain (10 y) Modic type 114 Modic type 2 Disability (1 y) –18.9 (–43.5 to 5.7) Modic type 133 Modic type 2 Disability (10 y) 14.0 (–8.6 to 36.6) Modic type 114 Modic type 2 –6.9 (–28.6 to 14.9) Modic type 133 Modic type 2 Back pain (1 y) 4.0 (–20.2 to 28.2) Modic type 135 Modic type 2 and 3 Back pain (10 y) 0.0 (–10.5 to 10.5) Leg pain (10 y) Modic type 133 Modic type 3 Disability (10 y) 18.0 (–19.1 to 55.1) Modic type 133 Modic type 3 40.0 (–2.9 to 82.9) Modic type 133 Modic type 3 Back pain (1 y) 38.0 (–6.6 to 82.6) Back pain (2 y) Modic type 1 change28 Not applicable as change measure Disability (1 y) 0.3 (NR)o Modic type 1 change22 Not applicable as change measure Disability (2 y) 0.26 (p = 0.04)d Modic type 1 change28 Not applicable as change measure Modic type 1 change22 Not applicable as change measure Back pain (1 y) 0.3 (NR)d Back pain (1 y) 0.30 (p = 0.02)a Modic type 1 and 214 No Modic Back pain (1 y) Modic type 1 and 223 No Modic Leg pain (1 y) –3.7 (–19.5 to 12.1) Modic type 1 and 218 No Modic Disability (1 y) –18.0 (–29.2 to –6.8) Modic type 1 and 214 No Modic Disability (1 y) 1.4 (0.3 to 6.3) (OR) Modic type 1 and 214 No Modic Disability (2 y) Modic type 1 and 223 No Modic 7.1 (–8.0 to 22.2) Modic type 1 and 2f,44 No Modic type 1 and 2 Disability (2 y) 7.3 (–3.5 to 18.0) Back pain (1 y)c –45.4 (–66.7 to 24.1) Modic type 1 or 2f,44 No Modic type 1 or 2 Disability (1 y)c 1.0 (0.3 to 3.5) (OR) Modic type 1 and/or 248 No Modic Back pain (14 mo) 3.4 (0.6 to 17.7) (OR) Modic type 1 and/or 248 No Modic Back pain (14 mo) 0.17 (p = 0.26)d Back pain (14 mo) 0.20 (p = 0.23)d Modic type 1, 2, 3 (large)24 Modic type 1, 2, 3 (small) Back pain (14 mo) Modic type 1, 2, 3 (large)k,24 Modic type 1, 2, 3 (small) Back pain (10 y) 1.2 (p = 0.13)h Modic type 1, 2, 3 (large)m,24 Modic type 1, 2, 3 (small) Leg pain (10 y) –12.3 (–27.4 to 2.8)l Modic type 1, 2, 3 (large)n,24 Modic type 1, 2, 3 (small) Disability (10 y) 2.5 (–12.2 to 17.2)l Modic type 1, 2, 333 No Modic Disability (13 y) –4.7 (–15.2 to 5.8)l Modic type 1, 2, 333 No Modic Disability (13 y) –1.0 (–14.3 to 12.3) Modic type 1, 2, 333 No Modic Disability (1 y) 0.0 (–15.4 to 15.4) Modic type 1, 2, 338 No Modic –13.2 (–28.3 to 1.9) Modic type 1, 2, 338 No Modic Back pain (1 y) Modic type 1, 2 or 339 No Modic Back pain (1 y)i 0.3 (NR) (R2) Back pain (1 y) 0.2 (p = 0.03)d Modic type 214 No Modic Back pain (1 y) 3.8 (–3.0 to 10.7) Modic type 247 No Modic Back pain (1 y)i Modic type 223 No Modic Back pain (10 y) –3.2 (–20.0 to 13.6) Modic type 218 No Modic Leg pain (1 y) 1.0 (–5.1 to 7.6) Modic type 247 No Modic Leg pain (10 y) Modic type 233 No Modic Disability (1 y) –6.1 (–13.7 to 1.4) Modic type 214 No Modic Disability (1 y) 1.5 (0.3 to 7.1) (OR) Modic type 233 No Modic Disability (1 y) 0.9 (0.8 to 1.1) (HR) Modic type 214 No Modic Disability (1 y) 3.0 (–12.2 to 18.2) Modic type 247 No Modic Disability (10 y) 14.1 (–5.01 to 33.2) Modic type 223 No Modic Disability (2 y) –3.0 (–20.5 to 14.5) Modic type 239 No Modic 9.8 (–0.5 to 20.0) Modic type 233 No Modic Back pain (1 y) Modic type 2f,44 No Modic Back pain (6 y) 0.5 (–6.0 to 7.0) –6.3 (–19.5 to 6.9) Modic type 2 and 335 No Modic Back pain (10 y) Modic type 2 or 325 No Modic Leg pain (10 y) 0.2 (–5.1 to 5.5) Disability (10 y) –12.6 (–29.7 to 4.5) Modic type 233 Modic type 3 2.1 (0.7 to 5.9) (OR) Modic type 233 Modic type 3 Back pain (1 y) Modic type 233 Modic type 3 Back pain (2 y) 1.0 (–7.1 to 9.1) Disability (1 y) 1.2 (–4.1 to 6.4) Modic type 2 change28 Not applicable as change measure Disability (2 y) Modic type 2 change22 Not applicable as change measure 26.0 (–11.4 to 63.4) Modic type 2 change28 Not applicable as change measure Back pain (10 y) 26.0 (–20.6 to 72.6) Modic type 2 change22 Not applicable as change measure Leg pain (10 y) 34.0 (–16.4 to 84.4) Disability (10 y) Modic type 333 No Modic –0.2 (NR)o Modic type 333 No Modic Back pain (1 y) –0.24 (p = 0.05)d Modic type 333 No Modic Back pain (1 y) Disability (1 y) –0.3 (NR)o NRC21 No NRC Disability (1 y) –0.13 (p = 0.31)d NRC23 No NRC NRC23 No NRC Back pain (6 y) –23.0 (–63.9 to 17.9) NRC29 No NRC (all patients) –29.0 (–74.1 to 16.1) Back pain (1 y) –46.6 (–87.5 to –5.7) Schmorl lesion25 No Schmorl lesion Back pain (6 y) Disability (1 y) –7.0 (–17.0 to 3.0) Spinal stenosis23 No spinal stenosis Back pain (6 y) 2.5 (–5.5 to 10.4) Spinal stenosis25 No spinal stenosis 2.9 (–11.1 to 16.8) Spinal stenosis23 No spinal stenosis Back pain (6 y) 0.9 (0.5 to 2.0) (OR) Spondylolisthesis25 No spondylolisthesis Back pain (18.5 mo)e Back pain (18.5 mo)e –3.6 (–7.6 to 0.4) 1 MRI abnormality25 No MRI abnormality Back pain (6 y) 1 MRI abnormality49 No MRI abnormality Back pain (18.5 mo)e 2.7 (–6.9 to 12.2) 1 MRI abnormality49 No MRI abnormality Back pain (18.5 mo)e –2.4 (–7.7 to 2.9) 2 MRI abnormalities25 No MRI abnormality Back pain (6 y) –6.0 (–23.5 to 11.5) 2 MRI abnormalities49 No MRI abnormality Back pain (18.5 mo)e –16.6 (–42.4 to 9.1) 2 MRI abnormalities49 No MRI abnormality Back pain (18.5 mo)e 3 MRI abnormalities25 No MRI abnormality Back pain (6 y) –0.6 (–5.4 to 4.2) 3 MRI abnormalities49 No MRI abnormality 1.0 (0.4 to 3.1) (OR) 3 MRI abnormalities49 No MRI abnormality 1.3 (0.5 to 3.0) (OR) 4 MRI abnormalities25 No MRI abnormality –1.9 (–7.8 to 4.0) 1.0 (0.4 to 2.7) (OR) 1.3 (0.6 to 3.1) (OR) 0.4 (–5.5 to 6.3) 1.0 (0.4 to 2.7) (OR) 1.5 (0.6 to 3.8) (OR) 2.6 (–6.7 to 11.8)
88 Han et al: MRI prediction of low back pain Table 4 (Continued) Comparison Clinical outcome (follow up)a Difference in means (95% CI)b Exposure 4 MRI abnormalities49 No MRI abnormality Back pain (18.5 mo)e 2.6 (1.0 to 7.1) (OR) 5 or more MRI abnormalities25 No MRI abnormality Back pain (6 y) –3.0 (–9.9 to 3.8) Positive outcome for difference in means indicates exposure group had a worse outcome at follow-up compared with comparison group. OR . 1 indicates greater odds of poor outcome in exposure group compared with comparison group. HR . 1 indicates greater incidence of poor outcome in exposure group compared with comparison group. Shaded rows indicate change in exposure or change in comparison MRI finding. LBP = low back pain, MRI = magnetic resonance imaging, NRC = nerve root compression, RMDQ = Roland Morris Disability Questionnaire, R2 = coefficient of determination, ? = the confidence interval reported in this study was implausible. a Long-term follow-up; outcome 12 months (closest time point to 12 months) taken. b Unless otherwise indicated. c Average pain in the past 2 weeks. d Beta coefficient. e Disabling LBP in past 4 weeks (RMDQ . 11). f Rehab group. g Recurrence of pain. h Change in disability score. i Pain at rest during previous week. j Rest 1 exercise group. k Exercise group. l Change in pain scores. m Rest group. n Rest 1 exercise group. o Pearson’s r. Association between all other MRI findings and pain and disability arthropathy,44,38 high-intensity zone,23,25,39,44,48 hypertrophy liga- Thirteen studies23–27,30,36,38,39,44,48–50 investigated 68 associations mentous flava,25 lumbar osteoarthritis,23 Schmorl lesion,25 spinal stenosis,23,25,50 spondylolisthesis,25 one or multiple MRI abnormal- between the remaining 11 MRI findings (disc herniation,23–27,30,36,50 ities25,49) and clinical outcomes in populations with LBP. We were disc bulge,48 disc height reduction,44,23,25 facet joint Table 5 Pooled estimates for MRI findings and clinical outcomes for populations with current LBP and no LBP. Exposure Comparison Clinical outcome No. of Difference in (follow-up) participants means (95% CI)a Pooled estimates for Modic changes (regardless of type) compared with no Modic in populations with LBP 6.6 (4.3 to 8.9) 4.6 (0.3 to 8.9) Modic (type 113,14,15,16,35, type 213,14,16, and type 1 and 232) No Modic Back pain (ST) 852 0.0 (–6.5 to 6.4) Modic (type 114,16, type 214,16) No Modic Disability (ST) 273 –0.3 (–3.7 to 3.2) Modic (type 114,23,24,25,47,35, type 214,23,47) No Modic Back pain (LT) 2,440 Modic (type 114,23,47,39, type 214,23,47,39) No Modic Disability (LT) 784 4.4 (1.1 to 7.6) 3.1 (–4.3 to 10.5) Pooled estimates for MRI findings and clinical outcomes in populations with LBP 1.5 (0.6 to 3.7) (OR) 1.2 (–9.1 to 11.4) Modic type 113,14,15,16,35 No Modic Back pain (ST) 308 –2.4 (–8.8 to 4.0) Modic type 114,16 No Modic Disability (ST) 129 1.1 (0.5 to 2.4) (OR) Modic type 113,14 No Modic Improvement (ST) 95 7.2 (4.3 to 10.1) Modic type 114,23,24,25,47,35 No Modic Back pain (LT) 2,171 5.1 (0.2 to 10.0) Modic type 114,23,47,39 No Modic Disability (LT) 334 –12.0 (–25.8 to 1.8) Modic type 118,46 No Modic Back pain (LT) 140 –18.3 (–69.9 to 33.3) 1.5 (1.0 to 2.2) (OR) Modic type 1 and 213,14,16,32 No Modic Back pain (ST) 624 –5.2 (–9.7 to –0.6) Modic type 1 and 214,16 No Modic Disability (ST) 201 –4.9 (–7.4 to –2.5) Modic type 1 and 213,23 No Modic Back pain (LT) 142 1.0 (0.4 to 2.6) (OR) Modic type 1 and 213,23 No Modic Disability (LT) 142 9.4 (5.6 to 13.3) Modic type 1 and 213,14,32 No Modic Improvement (ST) 495 5.9 (3.9 to 7.9) Modic type 113,14,16 Modic type 2 Back pain (ST) 165 1.5 (0.6 to 3.6) (OR) Modic type 114,16 Modic type 2 Disability (ST) 119 Modic type 113,14 Modic type 2 Improvement (ST) 86 –2.1 (–6.8 to 2.7) 0.9 (–3.4 to 5.2) Modic type 213,14,16 No Modic Back pain (ST) 199 –1.7 (–11.0 to 7.5) Modic type 214,16 No Modic Disability (ST) 144 1.1 (0.7 to 1.8) (OR) Modic type 213,14 No Modic Improvement (ST) 101 5.3 (1.3 to 9.3) Modic type 214,23,47 No Modic Back pain (LT) 269 2.1 (1.3 to 3.6) (OR) Modic type 214,23,47,39 No Modic Disability (LT) 450 –1.0 (–9.8 to 7.9) 6.0 (–9.3 to 21.3) NRC21,23 No NRC Pain (LT) 437 –0.6 (–3.1 to 4.2) NRC29,50 No NRC Disability (ST) 598 1.1 (–7.6 to 9.8)b Disc degeneration23,25 No disc degeneration Pain (LT) 2,138 –1.2 (–5.8 to 3.4) Disc degeneration44,49 No disc degeneration Disability (LT) 305 1.9 (0.8 to 4.6) (OR) Disc height reduction23,25 No disc height reduction Pain (LT) 2,138 Disc herniation23,25,27 No disc herniation Pain (LT) 2,172 High-intensity zone23,25 No high-intensity zone Pain (LT) 2,247 High-intensity zone23,39 No high-intensity zone Disability (LT) 391 Spinal canal stenosis23,25 No spinal canal stenosis Pain (LT) 2,138 Pooled estimates for MRI findings and clinical outcomes in populations with no LBP Disc degeneration43,37 No disc degeneration Pain (LT) 90 Positive outcome for difference in means indicates exposure group had a worse outcome at follow-up compared with comparison group. OR . 1 indicates greater odds of poor outcome in exposure group compared with comparison group. Shaded rows indicate change in exposure or change in comparison MRI finding. LT = long-term follow-up (outcome 12 months, closest time point to 12 months taken, NRC = nerve root compression, ST = short-term follow-up (outcome , 12 months, closest time point to 3 months taken). a Unless otherwise indicated. b Confidence intervals for Wilkens39 for high-intensity zone were implausible therefore the standard deviation (SD) for Jensen23 was used for the purpose of meta-analysis.
Research 89 Table 6 Association between MRI findings and clinical outcomes in populations with no LBP. Exposure Comparison Clinical outcome (follow-up duration) HR (95% CI)a Disc annular tear20 No disc annular tear Back pain (1 y)b 1.0 (0.5 to 2.0) No disc annular tear Back pain (1 y)c Disc annular tear19 1.2 (0.5 to 2.7) No disc degeneration Back pain (6 ys) 0.3 (–0.1 to 0.7)d Disc degeneration25 Disc degeneration (no change) Back pain (5 y)e 2.9 (0.8 to 10.3) (OR) Disc degeneration (worsening)43 Disc degeneration Pfirrmann , 3 Back pain (1 y)b 2.6 (0.6 to 10.6) Disc degeneration Pfirrmann 320 Disc degeneration Pfirrmann , 3 Back pain (1 y)c 3.6 (0.5 to 26.2) Disc degeneration Pfirrmann 319 No disc degeneration Back pain (5 y)e 1.4 (0.4 to 4.6) (OR) Disc degeneration40 No disc degeneration Back pain (10 y) 1.3 (0.4 to 4.7) (OR) Disc degeneration37 No disc bulge Back pain (10 y) 0.6 (0.2 to 2.4) (OR) Disc bulge37 0.2 (–0.4 to 0.7)d No disc height loss Back pain (6 y) 1.3 (0.9 to 1.8) Disc height loss25 No disc height loss Back pain (1 y)b 3.3 (0.8 to 13.9) Disc height loss20 No disc height loss Back pain (1 y)c 0.4 (–0.5 to 1.2)d Disc height loss19 No disc herniation Back pain (6 y) 1.1 (0.3 to 4.2) (OR) Disc herniation25 No disc herniation Back pain (5 y)e 0.9 (0.4 to 2.3) Disc herniation40 No disc herniation Back pain (1 y)b 1.8 (0.4 to 7.5) Disc herniation20 No disc herniation Back pain (1 y)c 0.5 (0.3 to 0.9) Disc herniation19 No disc protrusion Back and leg pain (3 y)d Disc protrusion45 1.2 (0.4 to 3.4) No disc extrusion Back and leg pain (3 y)d 0.3 (0.0 to 0.7)d Disc extrusion45 1.9 (1.0 to 3.5) No disc high-intensity zone Back pain (6 y) 2.5 (1.2 to 5.1) Disc high-intensity zone25 No disc high-intensity zone Back pain (1 y)b Disc high-intensity zone20 No disc high-intensity zone Back pain (1 y)c 1.0 (0.5 to 1.9) Disc high-intensity zone19 No facet joint arthrosis Back pain (1 y)b 0.7 (0.3 to 1.5) Facet joint arthrosis20 –0.4 (–1.7 to 0.9)d No facet joint osteoarthritis Back pain (1 y)c Facet joint osteoarthritis19 1.0 (0.5 to 2.3) No hypertrophy ligamentous flava Back pain (6 y) 1.4 (0.6 to 3.4) Hypertrophy ligamentous flava25 0.1 (–0.8 to 1.1)d No Modic Back pain (1 y)b 0.8 (–0.2 to 1.8)d Modic type 1, 2, 320 No Modic Back pain (1 y)c Modic type 1, 2, 319 1.9 (0.6 to 5.8) No Modic Back pain (6 y) 2.3 (0.6 to 8.6) (OR) Modic type 125 No Modic Back pain (6 y) Modic type 2 or 325 0.7 (0.0 to 1.4)d No nerve root contact Back and leg pain (3 y)d 0.7 (0.1 to 5.4) Nerve root contact45 No neural compromise Back pain (5 y)e 1.9 (0.8 to 4.8) Neural compromise40 1.1 (–0.1 to 2.3)d No spinal canal stenosis Back pain (6 y) 1.3 (0.6 to 2.8) Spinal canal stenosis25 No spinal canal stenosis Back pain (1 y)b 1.6 (0.7 to 3.6) Spinal canal stenosis20 No spinal canal stenosis Back and leg pain (3 y)d –0.5 (–0.9 to –0.2)d Spinal central stenosis45 No spondylolisthesis Back pain (6 y) 3.5 (NR) (RR) Spondylolisthesis25 No spondylolisthesis Back pain (1 y)b 0.3 (–0.1 to 0.6)d Spondylolisthesis20 No spondylolisthesis Back pain (1 y)c 2.8 (0.4 to 22.0) Spondylolisthesis19 0.2 (–0.2 to 0.6)d No Schmorl lesion Back pain (6 y) 5.9 (0.8 to 44.8) Schmorl lesion25 12.2 (1.3 to 118.2) No worsening abnormalities on MRI Back pain (7 y) 0.5 (–0.2 to 1.2)d Worsening abnormalities on MRI5 0.8 (0.0 to 1.5)d 0 findings Back pain (6 y) 1.2 (0.0 to 2.4)d 1 MRI finding25 0 findings Back pain (1 y)c 1 finding present19 0 findings Back pain (6 y) 2 MRI findings25 0 findings Back pain (1 y)c 2 findings present19 0 findings Back pain (1 y)c 3 findings present19 0 findings Back pain (6 y) 3 MRI findings25 0 findings Back pain (6 y) 4 MRI findings25 0 findings Back pain (6 y) 5 MRI findings25 Positive outcome for difference in means indicates exposure group had a worse outcome at follow-up compared with comparison group. HR . 1 indicates greater incidence of poor outcome in exposure group compared with comparison group. OR . 1 indicates greater odds of poor outcome in exposure group compared with comparison group. RR . 1 indicates that the poor outcome is more likely to develop in the exposure group. Shaded rows indicate change in exposure or change in comparison MRI finding. LBP = low back pain, MRI = magnetic resonance imaging, NR = not reported. a Unless otherwise indicated. b LBP (yes/no). c Days to recurrence of activity limiting LBP. d LBP recurrence. e Beta coefficient. only able to pool results for five different comparisons (which may have worse clinical outcomes at any follow-up time point but included 11 associations from individual studies) in total (Table 5). had confidence intervals that included no association (or where confidence intervals were not reported, were not statistically signif- Pooling demonstrated no evidence of an association between disc icant) (Tables 3 and 4). The remaining three associations from single height reduction (MD 21.0, 95% CI 29.8 to 7.9),23,25 disc herniation studies demonstrated statistically significant results. The presence of (MD 6.0, 95% CI 29.3 to 21.3),23,25,27 and spinal stenosis (MD 21.2, a high-intensity zone compared with no high-intensity zone was 95% CI 25.8 to 3.4)23,25 compared with the absence of the respective associated with slightly better pain outcomes in the short term MRI finding and pain outcomes in the long term. Pooling demon- strated no evidence of an association between the presence of a high- (b 20.38, p = 0.01).48 The presence of disc herniation compared with intensity zone compared with the absence of a high-intensity zone no disc herniation was associated with worse disability outcomes (OR and pain (MD 0.6, 95% CI 23.1 to 4.2)23,25 and disability outcomes 2.7, 95% CI 1.4 to 5.1)50 in the short term but slightly better disability (MD 20.4, 95% CI 25.3 to 4.6)23,39 in the long term (Table 5). outcomes (MD 23.7, 95% CI 26.4 to 21.0)26 in the long term. We were unable to pool data for the remaining 57 associations. Association of MRI findings and clinical outcomes in people with no LBP Point estimates in 44 of the 57 associations demonstrated no evi- Eight studies5,19,20,25,37,40,43,45 investigated 97 associations be- dence of an association between those with an MRI finding at base- tween MRI findings and pain outcomes in participants with no line and clinical outcomes at any follow-up time point, and 10 of the 57 associations suggested that those with an MRI finding at baseline
90 Han et al: MRI prediction of low back pain Table 7 Association between MRI findings and clinical outcome in mixed low back pain population samples. Exposure Comparison Clinical outcome (follow-up duration) OR (95% CI)a Disc herniation42 No disc herniation Back pain (5 y)b 4.81 (p = 0.01) Disc degeneration (MRI group)34 No disc degeneration (MRI group) Back pain (30 y) –0.3 (–2.3 to 1.7) (MD) Disc degeneration41 No disc degeneration Back pain (5 y)c 4.40 (p = 0.08) Disc degeneration31 No disc degeneration Back pain (10 y)d 1.1 (1.1 to 1.1) Disc degeneration (MRI group)34 No disc degeneration (MRI group) Disability (30 y) Disc degeneration41 No disc degeneration Disability (5 y)c –2.9 (–14.0 to 8.2) (MD) 1.1 (0.5 to 2.3) Canal stenosis (moderate/severe)41 No canal stenosis Back pain (5 y)c Canal stenosis (moderate/severe)41 No canal stenosis Disability (5 y)c 2.90 (p = 0.09) 2.1 (0.9 to 5.1) Schmorl Node31 No Schmorl node Back pain (10 y)d 1.1 (0.7 to 1.8) Modic (moderate/severe)41 No Modic Back pain (5 y)c Modic (type 1, 2 or 3)31 No Modic Back pain (10 y)d 2.5 (p = 0.1) 2.2 (1.4 to 3.3) Positive outcome for difference in means indicates exposure group had a worse outcome at follow-up compared with comparison group. OR . 1 indicates greater odds of poor outcome in exposure group compared with comparison group. Shaded rows indicate change in exposure or change in comparison MRI finding. LBP = low back pain, MRI = magnetic resonance imaging, NPRS = numerical pain rating scale. a Unless otherwise indicated. b Remissions of LBP for . 6 months. c Serious LBP (NPRS 6 for at least 1 week). d Disabling low back pain lasting . 1 month. current LBP (Appendix 3). No studies reported on disability outcomes aggregate MRI findings19 compared with the absence of the respected in populations with no current LBP. The MRI findings presented in MRI finding and worse pain outcomes at 1 year (OR 2.5, 95% CI 1.2 to three or more studies or where associations could be pooled across at 5.1 and OR 12.2, 95% CI 1.3 to 118.2, respectively)19 (Table 6). However, least two studies are presented below. Results are presented in results for three or more aggregate MRI findings19 should be inter- Table 6 and Appendix 3. Only one comparison (which included two preted with caution due to the imprecise confidence intervals. associations) was able to be pooled. The remaining studies were unable to be pooled due to different outcome measures or measures Association of MRI findings and clinical outcomes in mixed LBP of association reported across individual studies, or data from indi- population samples vidual studies were based on the same study populations.19,20,40,43 Four studies31,34,41,42 investigated 14 associations between MRI Association between disc degeneration and pain findings and clinical outcomes populations with mixed samples Six studies19,20,25,37,40,43 reported on the association between disc (Appendix 3); the results are presented in Table 7 and Appendix 3. No associations were able to be pooled due to different outcome mea- degeneration and pain outcomes in populations with no LBP. Pooling sures or measures of association reported. demonstrated that the presence of disc degeneration compared with the absence of disc degeneration may increase the likelihood of expe- One study31 demonstrated that the presence of Modic type 1, 2 or riencing pain in the long term (OR 1.9, 95% CI 0.8 to 4.6).37,43 The 3 changes compared with no Modic changes was associated with remaining five studies19,20,25,40,43 were unable to be pooled due to worse pain outcomes (OR 2.2, 95% CI 1.4 to 3.3) in the long term (10 different outcome measures and measures of association having been years). One study42 demonstrated that the presence of disc herniation used. Their associations suggested a consistent relationship of worse compared with no disc herniation was associated with worse pain pain outcomes in those with disc degeneration compared with no or outcomes (OR 4.8, p = 0.01) in the long term (5 years) (Table 7 and milder disc degeneration in the long term; only one study reported a Appendix 3). The remaining associations were not statistically sig- confidence interval that excluded no association and even in this study nificant. Point estimates in five of the 12 remaining associations the results were imprecise (OR 7.3, 95% CI 1.2 to 45.3)43 (Appendix 3). suggested that those with an MRI finding at baseline may have worse clinical outcomes at any follow-up time point, and seven of the 12 Association between disc herniation (all types) and pain remaining associations suggested no evidence of an association be- Five studies19,20,25,40,45 reported on the association between the tween those with an MRI finding at baseline and clinical outcomes at any follow-up time point. presence of disc herniation (including disc protrusion or extrusion) and pain outcomes in populations with no LBP. No studies were able to be Discussion pooled due to different outcome measures and measures of association used. Results of comparisons based on the five studies suggested no The available evidence is insufficient to judge whether most of the consistent relationship between the presence of disc herniation MRI findings that were tested were or were not associated with compared with no disc herniation and pain outcomes in the long term. future pain and disability. The exceptions were Modic changes and No studies reported statistically significant associations (Table 6). disc degeneration, which were associated with slightly worse pain and disability outcomes. In populations with current LBP, the pres- Association between all other MRI findings and pain ence of Modic type 1 and 2 changes was associated with slightly Five studies5,19,20,25,37,40,45 investigated 36 associations between worse pain and disability in the short term, the presence of Modic type 1 changes alone was associated with slightly worse pain in the the remaining 13 MRI findings (disc annular tear,19,20 disc bulge,25,37 short term, and the presence of disc degeneration was associated disc height loss,19,20,25 high-intensity zone,19,20,25 hypertrophy liga- with slightly worse pain and worse disability outcomes in the long mentous flava,25 facet joint osteoarthritis,19,20 Modic type 1, 2 and/or term. In populations with no LBP, pooled estimates demonstrated that 3 changes,19,20,25 nerve root contact,40,45 Schmorl lesion,25 spinal ca- the presence of disc degeneration may increase the likelihood of nal stenosis,20,25,45 spondylolisthesis,19,20,25 worsening abnormalities experiencing pain in the long term (Table 5). on MRI,5 one or more MRI finding19,25) and clinical outcomes in populations with no pain (Table 6). Point estimates from 26 of the 36 This updated systematic review included 27 new studies and the associations demonstrated no evidence of an association between 12 previous studies,6 resulting in a total of 39 studies that investi- those with an MRI finding at baseline and clinical outcomes, and gated MRI findings and their relationship to future LBP. This review eight of the 36 associations suggested that those with an MRI finding included a substantial number of studies (n = 39) in comparison with at baseline may have worse clinical outcomes at any follow-up time our previous review6 (n = 12). Despite the increase in studies, pooling point (1 to 10 years). Only two associations from one study19 reported associations between the presence of a high-intensity zone19 or three
Research 91 was limited due to the heterogeneity of the included studies. Studies use standardised MRI protocols and sequences; MRI protocols should varied in sample sizes (participants ranging from 20 to 1,997, me- also include standardised descriptions and thresholds for positive dian = 106) and used different outcome measures, different measures MRI findings. of associations, and different MRI measures. The results should be interpreted with caution, as they were based mostly on single studies, Our results suggest that some MRI findings may have weak as- typically with small sample sizes and most associations did not sociations with future LBP; however, for most of the MRI findings we demonstrate a statistically significant relationship between MRI evaluated it was unclear whether there was an association. Larger findings and future LBP. high-quality studies are still needed to resolve uncertainty for most MRI findings. This review used the same sensitive search strategy that was previously used by Steffens et al6 and followed a pre-specified pro- What was already known on this topic: The link between tocol with clear inclusion criteria. A limitation of this review was the various findings on magnetic resonance imaging and low back heterogeneity of the studies included, particularly the different pain is unclear because: structural changes on imaging can be outcome measures, MRI measures investigated and measures of as- present in both symptomatic and asymptomatic people; imaging sociations reported. This limited the number of comparisons where a findings become more prevalent with age; and much past meta-analysis could be conducted, and the number of studies in the research has used cross-sectional designs. possible meta-analyses. However, the previous review was unable to What this study adds: Based on longitudinal studies, some perform any meta-analyses. Some studies included participants from magnetic resonance imaging findings may have weak associa- RCTs13216,24,26,28,30,32,38,44,46,48,50 and it was not possible to determine tions with future low back pain. The imaging findings that are the influence of the intervention on the relationships investigated; predictive differ depending on whether pain is present at base- however, studies were excluded where the intervention was designed line. Although clinicians can incorporate this information into to change lumbar spine MRI findings (eg, surgery or antibiotic estimation of a patient’s prognosis when imaging findings are treatment for Modic changes). available, it is not recommended to obtain imaging purely to gain this prognostic information. A previous systematic review investigated the relationship be- tween 11 different MRI findings and LBP in cross-sectional studies.4 Footnotes: a Comprehensive Meta-Analysis software Version 3, Many MRI findings (disc herniation, disc degeneration, disc bulge, Biostat, Englewood, NJ, USA. Modic type 1 changes and spondylolysis) were more prevalent in those with LBP (ORs ranging from 2.2, 95% CI 1.2 to 4.2 to 7.5, 95% CI 1.3 to eAddenda: Appendices 1 to 3 can be found online at https://doi. 44.6).4 Another review investigated cross-sectional associations be- org/10.1016/j.jphys.2023.02.007 tween Modic changes and LBP.59 A statistically significant association was found between Modic change (any type) and the presence of LBP Ethics approval: Not applicable. in 15 of the 30 studies (ORs ranging from 1.5, 95% CI 1.0 to 2.3 to 83.1, Competing interests: The authors have no competing interests to 95% CI 4.9 to 1,424.1).59 The findings from these two reviews cannot be declare. directly compared with the current results as they investigated cross- Source(s) of support: Nil. sectional associations; however, the associations between Modic Acknowledgements: Nil. changes and pain in the cross-sectional studies4,59 are consistent with Data sharing: The authors will make all relevant data available the relationship to future LBP that the current review identified. upon reasonable request. Provenance: Not invited. Peer reviewed. For those imaging findings (eg, Modic changes, disc degeneration) Correspondence: Christopher S Han, Institute for Musculoskeletal where some evidence of an association with future LBP was found, it Health, The University of Sydney and Sydney Local Health District, is recommended that clinicians incorporate this information into Australia. Email: [email protected] their overall impression of a patient’s likely prognosis when imaging findings are available, but imaging is not currently recommended References purely to gain this prognostic information. For most MRI findings, it remains uncertain whether there is any relationship with future LBP. 1. Wu A, March L, Zheng X, Huang J, Wang X, Zhao J, et al. Global low back pain This review also suggests that the relationship between some imag- prevalence and years lived with disability from 1990 to 2017: estimates from the ing findings and future LBP may be different in people with and Global Burden of Disease Study 2017. Ann Transl Med. 2020;8:299. without LBP. For example, the presence of a high-intensity zone was associated with better outcomes in those with LBP but worse out- 2. van Tulder M, Becker A, Bekkering T, Breen A, del Real MT, Hutchinson A, et al. comes in those without LBP. This is important, as studies combining European guidelines for the management of acute nonspecific low back pain in participants with and without LBP may miss a relevant association in primary care. Eur Spine J. 2006;15(Suppl 2):S169–S191. one population. Our results should be interpreted with caution, as pooling mostly involved a small number of studies with varied 3. Brinjikji W, Luetmer PH, Comstock B, Bresnahan BW, Chen LE, Deyo RA, et al. sample sizes. There is still a paucity of large high-quality studies in Systematic literature review of imaging features of spinal degeneration in this field of research. asymptomatic populations. Am J Neuroradiol. 2015;36:811–816. Future research should focus on conducting high-quality studies 4. Brinjikji W, Diehn FE, Jarvik JG, Carr CM, Kallmes DF, Murad MH, et al. MRI findings to determine the clinical relevance of MRI findings and LBP. Analyses of disc degeneration are more prevalent in adults with low back pain than in should investigate the impact of age and psychosocial factors on as- asymptomatic controls: A systematic review and meta-analysis. Am J Neuroradiol. sociations between MRI findings and clinical outcomes. Future 2015;36:2394–2399. studies should aim to standardise the assessment and reporting of clinical outcomes based on the recommended core outcome mea- 5. Borenstein DG, O’Mara JW, Boden SD, Lauerman WC, Jacobson A, Platenberg C, sures for trials in patients with non-specific LBP (eg, using the et al. The value of magnetic resonance imaging of the lumbar spine to predict low- Oswestry Disability Index or Roland Morris Disability Questionnaire back pain in asymptomatic subjects. J Bone Joint Surg Am. 2001;83:1306–1311. to measure physical functioning).7 Standardising clinical outcome measures allows for better comparisons between study results and 6. Steffens D, Hancock MJ, Maher CG, Williams C, Jensen TS, Latimer J. Does magnetic improves the interpretability of results in this area. Studies should resonance imaging predict future low back pain? A systematic review. Eur J Pain. also investigate the impact of the way clinical outcomes (eg, pain) and 2013;18:755–765. MRI findings are defined or categorised, as many studies continue to dichotomise clinical outcomes or MRI findings (present or absent) 7. Chiarotto A, Boers M, Deyo RA, Boers M, Buchbinder R, Corbin TP, et al. Core and do not account for different degrees of severity. This may result in outcome measurement instruments for clinical trials in nonspecific low back pain. missing important data or relationships. Future studies should also Pain. 2018;159:481–495. 8. Altman DG. Systematic reviews of evaluations of prognostic variables. BMJ. 2001;323(7306):224–228. 9. da C Menezes Costa L, Maher CG, Hancock MJ, McAuley JH, Herbert RD, Costa LOP. The prognosis of acute and persistent low-back pain: a meta-analysis. CMAJ. 2012;184:E613–E624. 10. Pengel LHM, Herbert RD, Maher CG, Refshauge KM. Acute low back pain: sys- tematic review of its prognosis. BMJ. 2003;327(7410):323. 11. Buttermann GR, Mullin WJ. Pain and disability correlated with disc degeneration via magnetic resonance imaging in scoliosis patients. Eur Spine J. 2008;17:240– 249. 12. Mitra D, Cassar-Pullicino VN, McCall IW. Longitudinal study of vertebral type-1 end-plate changes on MR of the lumbar spine. Eur J Radiol. 2004;14:1574–1581.
92 Han et al: MRI prediction of low back pain 13. Annen M, Peterson C, Humphreys BK. Comparison of treatment outcomes in analysis of data from the spine patient outcomes research trial. Phys Med Rehabil. nonspecific low-back pain patients with and without modic changes who receive 2016;8:405–414. chiropractic treatment. J Manipulative Physiol Ther. 2018;41:561–570. 37. Tonosu J, Oka H, Higashikawa A, Okazaki H, Tanaka S, Matsudaira K. The associa- tions between magnetic resonance imaging findings and low back pain: A 10-year 14. Annen M, Peterson C, Leemann S, Schmid C, Anklin B, Humphreys BK. Comparison longitudinal analysis. PLoS One. 2017;12:e0188057. of outcomes in MRI confirmed lumbar disc herniation patients with and without 38. Udby PM, Ohrt-Nissen S, Bendix T, Brorson S, Carreon LY, Andersen MO. The as- modic changes treated with high velocity, low amplitude spinal manipulation. sociation of MRI findings and long-term disability in patients with chronic low J Manipulative Physiol Ther. 2016;39:200–209. back pain. Global Spine J. 2021;11:633–639. 39. Wilkens P, Scheel IB, Grundnes O, Hellum C, Storheim K. Prognostic factors of 15. Cai G, Laslett LL, Aitken D, et al. Effect of zoledronic acid and denosumab in patients prolonged disability in patients with chronic low back pain and lumbar degener- with low back pain and modic change: A proof-of-principle trial. J Bone Miner Res. ation in primary care: a cohort study. Spine. 2013;38:65–74. 2018;33:773–782. 40. Boos N, Semmer N, Elfering A, Schade V, Gal I, Zanetti M, et al. Natural history of individuals with asymptomatic disc abnormalities in magnetic resonance imaging. 16. Chen Y, Yang H, Zhang L, Wang Y, Zou J. Analyzing the influence of modic changes Spine. 2000;25:1484–1492. on patients with lower back pain undergoing conservative treatment. Pain Res 41. Carragee E, Alamin T, Cheng I, Franklin T, Hurwitz E. Does minor trauma cause Manag. 2019;2019:8185316. serious low back illness? Spine. 2006;31:2942–2949. 42. Carragee EJ, Alamin TF, Miller JL, Carragee JM. Discographic, MRI and psychosocial 17. de Schepper EIT, Koes BW, Oei EHG, Bierma-Zeinstra SMA, Luijsterburg PAJ. The determinants of low back pain disability and remission: a prospective study in added prognostic value of MRI findings for recovery in patients with low back pain subjects with benign persistent back pain. Spine. 2005;5:24–35. in primary care: a 1-year follow-up cohort study. Eur Spine J. 2016;25:1234–1241. 43. Elfering A, Semmer N, Birkhofer D, Zanetti M, Hodler J, Boos N. Risk factors for lumbar disc degeneration: A 5-year prospective MRI study in asymptomatic in- 18. el Barzouhi A, Vleggeert-Lankamp CLAM, van der Kallen BF, à Nijeholt GJ, van den dividuals. Spine. 2002;27:125–134. Hout WB, Koes BW, et al. Back pain’s association with vertebral end-plate signal 44. Hellum C, Johnsen LG, Gjertsen Ø, Berg L, Neckelmann G, Grundnes O, et al. Pre- changes in sciatica. Spine. 2014;14:225–233. dictors of outcome after surgery with disc prosthesis and rehabilitation in patients with chronic low back pain and degenerative disc: 2-year follow-up. Eur Spine J. 19. Hancock MJ, Kjaer P, Kent P, Jensen RK, Jensen TS. Is the number of different MRI 2012;21:681–690. findings more strongly associated with low back pain than single MRI findings? 45. Jarvik JG, Hollingworth W, Heagerty PJ, Haynor DR, Boyko EJ, Deyo RA. Three-year Spine. 2017;42:1283–1288. incidence of low back pain in an initially asymptomatic cohort: Clinical and im- aging risk factors. Spine. 2005;30:1541–1548. 20. Hancock MJ, Maher CM, Petocz P, Lin CW, Steffens D, Luque-Suarez A, et al. Risk 46. Jensen RK, Leboeuf-Yde C, Wedderkopp N, Sorensen JS, Jensen TS, Manniche C. Is factors for a recurrence of low back pain. Spine. 2015;15:2360–2368. the development of Modic changes associated with clinical symptoms? A 14- month cohort study with MRI. Eur Spine J. 2012;21:2271–2279. 21. Harrisson SA, Ogollah R, Dunn KM, Foster NE, Konstantinou K. Prevalence, char- 47. Keller A, Boyle E, Skog TA, Cassidy JD, Bautz-Holter E. Are Modic changes prog- acteristics, and clinical course of neuropathic pain in primary care patients nostic for recovery in a cohort of patients with non-specific low back pain? Eur consulting with low back-related leg pain. Clin J Pain. 2020;36:813–824. Spine J. 2012;21:418–424. 48. Kleinstück F, Dvorak J, Mannion AF. Are “structural abnormalities” on magnetic 22. Järvinen J, Karppinen J, Niinimäki J, Haapea M, Grönblad M, Luoma K, et al. As- resonance imaging a contraindication to the successful conservative treatment of sociation between changes in lumbar modic changes and low back symptoms over chronic nonspecific low back pain? Spine. 2006;31:2250–2257. a two-year period. BMC Musculoskelet Disord. 2015;16:1–8. 49. McNee P, Shambrook J, Harris EC, Kim M, Sampson M, Palmer KT, et al. Predictors of long-term pain and disability in patients with low back pain investigated by 23. Jensen OK, Nielsen CV, Sorensen JS, Stengaard-Pedersen K. Type 1 Modic changes magnetic resonance imaging: a longitudinal study. BMC Musculoskelet Disord. was a significant risk factor for 1-year outcome in sick-listed low back pain pa- 2011;12:234. tients: a nested cohort study using magnetic resonance imaging of the lumbar 50. Modic MT, Obuchowski NA, Ross JS, Brant-Zawadzki MN, Grooff PN, Mazanec DJ, spine. Spine J. 2014;14:2568–2581. et al. Acute low back pain and radiculopathy: MR imaging findings and their prognostic role and effect on outcome. Radiology. 2005;237:597–604. 24. Jensen RK, Kent P, Hancock M. Do MRI findings identify patients with chronic low 51. Beall DP, Wilson GL, Bishop R, Tally W. VAST Clinical Trial: Safely supplementing back pain and Modic changes who respond best to rest or exercise: a subgroup tissue lost to degenerative disc disease. Int J Spine Surg. 2020;14:239–253. analysis of a randomised controlled trial. Chiropr Man Therap. 2015;23:1–9. 52. Berry DB, Rodriguez-Soto AE, Englund EK, Shahidi B, Parra C, Frank LR, et al. Multiparametric MRI characterization of level dependent differences in lumbar 25. Kasch R, Truthmann J, Hancock MJ, Maher CG, Otto M, Nell C, et al. Association of muscle size, quality, and microstructure. Spine. 2020;3:e1079. lumbar MRI findings with current and future back pain in a population-based 53. Gornet MF, Bailey AS, Eltahawy HA, Sasso RC, Cheng WK, Palmer S, et al. 250. cohort study. Spine. 2022;47:201–211. Interspinous process distraction compared to nonoperative care for moderate lumbar degenerative disc disease: results of the FDA IDE clinical trial. Spine. 26. Kerr D, Zhao W, Lurie JD. What are long-term predictors of outcomes for lumbar 2019;19:S122–S123. disc herniation? A randomized and observational study. Clin Orthop Relat Res. 54. Harrisson SA, Ogollah R, Dunn KM, Foster NE, Konstantinou K. Prognosis of low 2015;473:1920–1930. back-related leg pain patients with neuropathic pain: clinical course and prog- nostic factors. Physiotherapy. 2019;105:e10. 27. Kesikburun B, Eksioglu E, Turan A, Adiguzel E, Kesikburun S, Cakci A. Spontaneous 55. Mushlin AI. In patients with sciatica, MRI at 1 year did not differentiate between a regression of extruded lumbar disc herniation: Correlation with clinical outcome. favorable or unfavorable outcome. Ann Transl Med. 2013;159:JC12. Pak J Med Sci. 2019;35:974–980. 56. Brøgger HA, Maribo T, Christensen R, Schiøttz-Christensen B. Comparative effec- tiveness and prognostic factors for outcome of surgical and non-surgical man- 28. Koivisto K, Järvinen J, Karppinen J, Haapea M, Paananen M, Kyllönen E, et al. The agement of lumbar spinal stenosis in an elderly population: protocol for an effect of zoledronic acid on type and volume of modic changes among patients observational study. BMJ open. 2018;8:e024949–e. with low back pain. BMC Musculoskelet Disord. 2017;18:274. 57. Beall DP, Wilson GL, Gershon S, Davis TT, Bishop R, Tally WC, et al. 58. Viable disc tissue allograft supplementation in the treatment of degenerated intervertebral 29. Konstantinou K, Dunn KM, Ogollah R, Lewis M, van der Windt D, Hay EM, et al. discs the one-year results of a randomized control trial. Spine. 2021;21:S29. Prognosis of sciatica and back-related leg pain in primary care: the ATLAS cohort. 58. el Barzouhi A, Vleggeert-Lankamp CL, Lycklama à, Nijeholt GJ, Van der Kallen BF, Spine. 2018;18:1030–1040. van den Hout WB, Koes BW, et al. Influence of low back pain and prognostic value of MRI in sciatica patients in relation to back pain. PloS One. 2014;9:e90800. 30. Kuligowski T, De¸ biec-Ba̧k A, Skrzek A. Mobilisation efficacy in young patients with 59. Herlin C, Kjaer P, Espeland A, Skouen JS, Leboeuf-Yde C, Karppinen J, et al. Modic different stages of degenerative disc disease. J Back Musculoskelet Rehabil. changes—their associations with low back pain and activity limitation: A sys- 2020;33:913–918. tematic literature review and meta-analysis. PloS One. 2018;13:e0200677. 31. Määttä JH, Wadge S, MacGregor A, Karppinen J, Williams FMK. Vertebral endplate (modic) change is an independent risk factor for episodes of severe and disabling low back pain. Spine. 2015;40:1187–1193. 32. Peterson CK, Pfirrmann CWA, Hodler J. Are modic changes related to outcomes in lumbar disc herniation patients treated with imaging-guided lumbar nerve root blocks? Eur J Radiol. 2014;83:1786–1792. 33. Romero-Muñoz LM, Barriga-Martín A, Segura-Fragoso A, Martín-González C. Are Modic changes in patients with chronic low back pain indicative of a worse clinical course? 10 years of follow-up. Rev Esp Cir Ortop Traumatol. 2018;62:274–281. 34. Sääksjärvi S, Kerttula L, Luoma K, Paajanen H, Waris E. Disc degeneration of young low back pain patients: A prospective 30-year follow-up MRI study. Spine. 2020;45:1341–1347. 35. Schistad EI, Espeland A, Rygh LJ, Roe C, Gjerstad J. The association between Modic changes and pain during 1-year follow-up in patients with lumbar radicular pain. Skeletal Radiol. 2014;43:1271–1279. 36. Suri P, Pearson AM, Scherer EA, Zhao W, Lurie JD, Morgan TS, et al. Recurrence of pain after usual nonoperative care for symptomatic lumbar disk herniation:
Journal of Physiotherapy 69 (2023) 68–69 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Editorial Women’s health Cinara Sacomori a, Mark R Elkins b,c a Facultad de Medicina, Clínica Alemana, Universidad del Desarrollo, Santiago, Chile; b Editor, Journal of Physiotherapy; c Faculty of Medicine and Health, University of Sydney, Sydney, Australia This Editorial introduces another of Journal of Physiotherapy’s assessment). Meanwhile, the PFM forces (dynamometry) during rest article collections.1–4 The articles in this collection relate to the and during a maximal voluntary contraction were higher in standing physiotherapy management of conditions that are common among compared with supine. women, such as incontinence and breast cancer. The featured articles have been curated to: facilitate access to recent important findings in Facilitating access to individual PFMT is still a challenge world- the field; highlight trends in the study designs, methodology, pop- wide. Delivering group-based approaches could be more efficient. ulations and interventions addressed by the research; and provide a The first paper in this collection qualitatively examining group-based scoping overview of avenues for further research. PFMT was by Fernandes et al.12 The analysis of the interviews revealed satisfaction with the content and usefulness of the PFMT Research into physiotherapy interventions in women’s health has group education, which was reinforced by the fact that many par- a strong history. Clinical trials in the women’s health subdiscipline are ticipants shared their new knowledge with others. In addition, their amongst the higher-quality trials on the Physiotherapy Evidence perception of the benefits persisted over time.12 Database (PEDro).5 Trials of interventions for incontinence and cancer provide more complete descriptions of the experimental and control The randomised trial by Cacciari et al13 indicated that group-based interventions than trials in any other subdiscipline of physiotherapy.6 PFMT was less expensive and as clinically effective when compared The research summarised in this article collection is no exception, as with standard individual PFMT for older women with stress or mixed shown by the mean PEDro score of 8 among the included trials. urinary incontinence. The authors used a widely accepted economic analysis – the incremental cost-effectiveness analysis – for each of Women might benefit from physiotherapy throughout their life- the effectiveness outcomes and considered two perspectives span. Some critical moments are during pregnancy, postpartum, (participant and provider). menopause and older age.6 This article collection discusses some of the health conditions for which physiotherapy is of great value: pelvic Brennen et al14,15 conducted a systematic review of cost- floor dysfunctions, breast cancer, gestational diabetes, and psycho- effectiveness analyses of various ways of providing PFMT to prevent social issues related to sexual assault. or treat postpartum incontinence. Their review showed that during pregnancy, group-based PFMT is more efficient than individually Pelvic floor dysfunction supervised PFMT to prevent urinary incontinence. Furthermore, it has the added benefit of preventing and treating faecal incontinence. Urinary incontinence is a prevalent and bothersome pelvic floor dysfunction that affects women of all ages.7,8 The invited topical re- Despite the favourable effects of individual and group PFMT view by Professor Kari Bø9 deftly summarises the available evidence described in the above studies, the effect of perioperative PFMT was about the burden associated with the condition, its treatment and less beneficial. This was demonstrated in the assessor-blinded, prevention (with a particular focus on physiotherapy interventions), randomised controlled trial of Duarte et al16 that included 96 and future directions for research and clinical practice. It also contains women undergoing surgery for pelvic organ prolapse. Perioperative a large section on group pelvic floor muscle training (PFMT), which is PFMT was estimated to be trivially helpful or harmful on manometric also the focus of three other papers in this collection. testing of PFM strength and relevant quality of life questionnaires. Participants who received the additional PFMT perceived their global However, some women are unable to voluntarily contract their improvement as marginally higher, although the estimates of this pelvic floor muscles (PFM) at all. Antonio et al10 enrolled a group of effect showed that the extra improvement may be trivially small. such women in a study of intravaginal electrical stimulation. This assessor-blinded, randomised controlled trial showed that those who Breast cancer received eight sessions of intravaginal electrical stimulation with instructions to contract their PFM achieved greater ability to contract Cancer is the leading cause of death in many countries and breast their PFM voluntarily and diminished urinary incontinence symp- cancer is the most common type of cancer among women.17 Lym- toms compared with those that did not receive any treatment.10 phoedema is a common sequela of breast cancer. Fortunately for readers of Journal of Physiotherapy, Professor Elizabeth Dylke’s invited The PFM are commonly assessed in supine, but assessment in the topical review expertly summarises the available evidence about standing position potentially allows a more functional evaluation physiotherapy measurement of breast cancer-related lymphoedema, because urine loss usually occurs in the upright position. To address including circumference measurements, whole limb volume mea- this gap, Gimenez et al11 used an innovative protocol to assess PFM surement, bioimpedance spectroscopy and tissue dielectric constant in standing. They found that during a maximal voluntary PFM assessment.18 This is the first of a series of three invited topical re- contraction, women had lower scores in the standing position views on breast cancer; the others will cover physiotherapy treat- compared with supine for most of the functional PFM assessments ment of breast cancer-related lymphoedema and physiotherapy (vaginal pressures, electromyographic activity and digital management of other sequelae of breast cancer. https://doi.org/10.1016/j.jphys.2023.02.013 1836-9553/© 2023 Published by Elsevier B.V. on behalf of Australian Physiotherapy Association. This is an open access article under the CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/).
Editorial 69 The International Society of Lymphology recommends that lym- by considering patients’ needs, being trustworthy, offering options, phoedema be treated with decongestive lymphatic therapy, which collaborating and empowering them. consists of skin care, manual lymphatic drainage, multi-layer bandaging and exercise therapy in phase one; and skin care, educa- All the articles included in this collection suggest avenues for tion regarding self-management, a compression sleeve, exercises and further research. Other potential areas for further development are manual lymphatic drainage in phase two.19 Although it has been the role that physiotherapy has in improving functioning and quality applied in many countries for decades, a Cochrane review with six of life for women with other health conditions, such as occupational randomised trials could not demonstrate any added value of manual diseases, cancer, chronic pain and cardiovascular diseases. Trans- lymphatic drainage.20 The rigorous multicentre randomised trial of gender health is another important area for research. In addition, we De Vrieze et al21 confirmed this lack of benefit. It showed that manual must deal with bigger challenges, such as how to make evidence- lymphatic drainage with or without fluoroscopy guidance did not based women’s health physiotherapy more accessible and cost- improve the effect of decongestive lymphatic therapy in people with effective to implement worldwide and how to enhance adherence chronic breast cancer-related lymphoedema (stage I to IIb). Manual to physical activity/exercise. lymphatic drainage did not provide any clinically important benefits on any outcomes in the study. In conclusion, this collection shows that physiotherapy in- terventions can be effective in many aspects of women’s health. Gestational diabetes Physiotherapists seeking to offer such services might consider joining the International Organization of Physical Therapists in Pelvic and Gestational diabetes mellitus is a frequent complication of preg- Women’s Health (IOPTPWH) – an initiative trying to integrate efforts nancy. The systematic review and meta-analysis of Harrison et al22 to improve health and wellbeing in pelvic and women’s health. demonstrated that adding exercise to usual care reduces post- prandial blood glucose, fasting blood glucose and glycated haemo- Competing interests: Nil. globin. Aerobic and resistance exercise programs with at least Source(s) of support: Nil. moderate intensity, for 20 to 30 minutes, three to four times per Acknowledgements: Nil. week, proved to be safe and effective.22 Provenance: Invited. Peer reviewed. Correspondence: Mark R Elkins, Centre for Education & Workforce In the face of this evidence, the challenge was how to promote Development, Sydney Local Health District, Sydney, Australia. Email: adherence to physical exercise/activity recommendations among [email protected] these women. Harrison et al23 provided new insights signalling that women with gestational diabetes want clear, simple, physical activity References messages directly related to pregnancy outcomes from a credible source, but adaptable to their daily life routine. Aiming to fulfil this 1. Blanquero J, et al. J Physiother. 2021;67:3–4. requirement, the same authors24 proposed a consumer co-created 2. Dale MT, et al. J Physiother. 2021;67:84–86. infographic about physical activity, which was tested in a rando- 3. Campos TF, et al. J Physiother. 2021;67:158–159. mised trial. The infographic generated clinically important short-term 4. Goff AJ, et al. J Physiother. 2021;67:240–241. improvements in women’s knowledge and self-efficacy to engage in 5. Moseley AM, et al. Physiother Can. 2014;66:36–43. physical activity compared with standard education.24 6. Yamato TP, et al. Physiotherapy. 2016;102:121–126. 7. Zaid F, et al. J Pak Med Assoc. 2022;724:603–604. Psychosocial aspects of violence and sexual assault 8. Milsom I, et al. Epidemiology of urinary incontinence and other lower urinary tract Women are at increased risk of all forms of sexual assault symptoms, pelvic organ prolapse and anal incontinence. Abrams P, Cardozo L, compared with men and are sometimes also exposed to distinct Wagg A, Wein A (Eds.). Incontinence, Vol. 1, 6th International Consultation on forms of domestic violence.25 In the time of the Me Too Movement, we Incontinence, Tokyo. 2017:1–141. should reflect on the role that physiotherapists may play, either 9. Bø K, et al. J Physiother. 2020;66:147–154. explicitly or implicitly, in the management of people who have un- 10. Antonio FI, et al. J Physiother. 2022;68:37–42. dergone sexual assault. Physiotherapists who work in the women’s, 11. Gimenez MM, et al. J Physiother. 2022;68:51–60. men’s and pelvic health subdiscipline of the profession have a role to 12. Fernandes ACNL, et al. J Physiother. 2021;67:210–216. play with women who are survivors of sexual assault because they 13. Cacciari L, et al. J Physiother. 2022;68:191–196. are more likely to have multiple pelvic floor complaints compared 14. Brennen R, et al. J Physiother. 2021;67:105–114. with those without a history of sexual assault.26 15. Brennen R, et al. J Physiother. 2021;67:161. 16. Duarte TB, et al. J Physiother. 2020;66:27–32. The thoughtful editorial by Stirling et al27 recommends that 17. Sung H, et al. CA Cancer J Clin. 2021;71:209–249. women’s health physiotherapists: screen for sexual trauma to sup- 18. Dylke E, et al. J Physiother. 2022;68:238–243. port assessment, treatment and referral; be aware that some exer- 19. International Society of Lymphology. Lymphology. 2020;53:3–19. cises or manual therapy may trigger uncomfortable somatic 20. Ezzo J, et al. Cochrane Database Syst Rev. 2015;5:CD003475. experiences related to the trauma; and adopt the trauma-informed 21. De Vrieze T, et al. J Physiother. 2022;68:110–122. approach in cases where a patient reports suffering from a sexual 22. Harrison AL, et al. J Physiother. 2016;62:188–196. assault. The latter includes providing emotional and physical safety 23. Harrison AL, et al. J Physiother. 2019;65:37–42. 24. Harrison AL, et al. J Physiother. 2020;66:243–248. 25. Acierno R, et al. Behav Med. 1997;23:53–64. 26. Beck JJ, et al. J Sexual Med. 2009;6:193–198. 27. Stirling J, et al. J Physiother. 2021;67:1–2. Websites PEDro www.pedro.org.au International Organization of Physical Therapists in Pelvic and Women’s Health www.wcpt.org/ioptwh
Search
