Australian Journal of Physiotherapy

["228 Malliaras: Physiotherapy management of Achilles tendinopathy Table 2a (Continued) Categories of Speci\ufb01c Short term (\u0014 8 weeks) Medium term (. 8 to 16 weeks) Longer term (. 16 weeks) interventions interventions compared compared Exercise Eccentric versus Favours acupuncture Favours acupuncture Favours acupuncture acupuncture Pain\/disability Pain\/disability Pain\/disability versus adjunctse Eccentric versus Favours eccentric Favours eccentric traditional caref Pain\/disability Pain\/disability Eccentric versus Favours heel wedge Favours heel wedge heel wedge Pain\/disability Pain\/disability Favours heel wedge Favours heel wedge Pain Pain Neutral Favours heel wedge Global Global Eccentric versus Neutral Neutral Neutral night splint Disability, pain Disability, pain Disability, pain Eccentric versus Neutral ESWT Pain\/disability, pain Neutral Global Eccentric versus Neutral Neutral AirHeel Pain Pain Eccentric versus US Neutral Neutral Pain Pain Eccentric versus Neutral Neutral Neutral pressure massage Pain\/disability Pain\/disability Pain\/disability Adjunct Radial ESWT versus Favours ESWT versus wait-and-see Pain\/disability, pain control Focused ESWT Neutral versus placebo Global change Neutralb Pain LLLT versus Neutral Neutral placebo Pain\/disability, pain Pain\/disability, pain Acupuncture Neutral Neutral versus placebo Pain\/disability, pain Pain\/disability, pain Evidence certainty is based on GRADE. Clinically meaningful bene\ufb01ts are bolded. ESWT = extracorporeal shock wave therapy, LLLT = low-level laser therapy, US = ultrasound. a Stasinopoulos 201396 removed in sensitivity analysis. b Mixed cohort of insertional and midportion Achilles tendinopathy. c Unspeci\ufb01ed pain from Knobloch 2008114 (see Appendices 3 and 4). d Foot and Ankle Outcome Score (FAOS) pain from Knobloch 2008114 (see Appendices 3 and 4). e Some exercise arms contained the adjuncts tested in the comparator. However, the fundamental question was still the same: How do adjuncts alone compare with exercise (6 adjuncts)? f Deep friction massage and therapeutic US was \u2018traditional\u2019 in both studies and Herrington et al 2007116 also included these adjuncts in the exercise arm. Moderate certainty evidence of a between-group difference. Low certainty evidence of a between-group difference. Very low certainty evidence of a between-group difference. Moderate certainty evidence favouring neither intervention. Low certainty evidence favouring neither intervention. Very low certainty evidence favouring neither intervention. management period may be reached. Otherwise, secondary man- data and GRADE assessment (Appendices 3 and 4). An outline of the agement can be initiated (Figure 4); this involves medical referral for search, trial selection, quality appraisal and data pooling methods is consideration of imaging and specialist care (usually a sports or provided in Appendix 4. Forty-four trials addressing the comparisons musculoskeletal physician), with or without a repeat of the primary of interest were identi\ufb01ed and the PEDro scores ranged from 3 to 10. treatments. Although uncommon, people who suffer continued im- Tables 2a and 2b provide an evidence summary about education, pairments may consider surgery. exercise and adjunct interventions for Achilles tendinopathy. It is beyond the scope of this review to evaluate the evidence Education and advice about all management options for Achilles tendinopathy. The following sections discuss the evidence for primary physiotherapy Developing an individual\u2019s understanding of their health condi- interventions versus important comparators, including: education or tion facilitates shared decision-making and successful self- advice versus control or other education or advice; exercise versus management. Education (imparting knowledge) and advice control or other exercise; exercise plus an adjunct versus exercise or (counsel regarding action the patient can take in management) for exercise versus an adjunct; and adjunct versus control. See the eAd- Achilles tendinopathy is guided by themes arising from qualitative denda for between-group effects and Grading of Recommendations studies26-29 and clinical practice guidelines53 in this population. Assessment, Development and Evaluation (GRADE) assessment People affected by this condition express a desire to understand (Appendix 1), PEDro scores (Appendix 2) and forest plots of pooled","Research 229 Table 2b Longer term Evidence summary for insertional Achilles tendinopathy: exercise and adjunct interventions. (. 16 weeks) Favours instrumented Categories of interventions compared Speci\ufb01c interventions compared Short term Medium term Pain\/disability (\u0014 8 weeks) (. 8 to 16 weeks) Exercise 1 adjunct versus Adding instrumented manual therapy to exercise Neutral exercise Neutral Favours instrumented Pain Pain\/disability Pain\/disability Adding ESWT or placebo to exercise Neutral Neutral Neutral Pain\/disability Pain Pain Neutral Neutral Neutral Pain Pain\/disability Global Neutral Neutral Pain\/disability Pain Netural Exercise versus adjunct Eccentric versus ESWT Pain Eccentric exercise versus multimodala Favours ESWT Pain\/disability, pain, global Neutral Pain Evidence certainty is based on GRADE. Clinically meaningful bene\ufb01ts are bolded. ESWT = extracorporeal shock wave therapy. a Stretch, massage, bilateral heel lifts, night splint included as \u2018multimodal\u2019 in both groups. Low certainty evidence of a between-group difference. Very low certainty evidence of a between-group difference. High certainty evidence favouring neither intervention. Moderate certainty evidence favouring neither intervention. Low certainty evidence favouring neither intervention. Very low certainty evidence favouring neither intervention. causes, what the pathology is, expected outcomes and what is recommendations should be tailored to individual needs identi\ufb01ed needed to overcome the condition. They also express: fears about during assessment. making their condition worse or a tendon rupture; maladaptive beliefs (eg, rest is needed to overcome their pain); and a desire for Two trials9,85 have speci\ufb01cally investigated education or advice individualised care and support. Suggested education and advice versus other education or advice for Achilles tendinopathy (Appendix topics and strategies to improve engagement are shown in Figure 5. 1). Chimenti et al85 compared pain science education (focusing on \u2018Education beyond words\u201929 is important and can be achieved with neurophysiology of pain and addressing fear of movement) with the use of engaging and patient-friendly demonstration of data that control education (focusing on pathophysiology and pathoanatomy). challenge beliefs, and multimedia resources (websites, videos) (ex- There was low certainty of evidence that these two interventions did amples provided in Figure 5). Education and advice not substantially differ in their effects on pain (short term), composite pain and disability, or global change outcomes at short and medium Patient-centred approach What is Explain individual risk factors \u2022 consider individual wrong with Explain tendon pathology (important to patients (Mallows et al, 202127)) impact me? \u2022 Share data of asymptomatic pathology is common (Docking et al, 202145) \u2022 empathy and Explain local nociceptive pain but influence of biopsychosocial factors What can \u2022 Use multimedia resources (eg, www.tamethebeast.org, Why Things Hurt video) reassurance we do for it? \u2022 Use pain models (eg, Mature Organism Model (Gifford, 2013177)) \u2022 shared decisions Explore and address belief that pain may signal tendon damage What can I considering expect? \u2022 Share data indicating minimal risk of rupture (Yasui et al, 20175) preferences and \u2022 Use individual pain\/apprehension with loading to reinforce biopsychosocial pain model lifestyle \u2022 active learning Explain primary and secondary treatment rationale, benefits and harms approaches Provide specific strategies for activity modification \u2022 understand individual patient goals and \u2022 Use step count, volume (eg, time) or intensity (speed) data from smart devices motivations Facilitate understanding of acceptable pain and adaptive pain behaviours \u2022 Share data showing pain during exercise is safe and acceptable (Smith et al, 2017133) \u2022 Discuss consequences of \u2018endurance\u2019 or \u2018fear-avoidance\u2019 strategies Provide specific strategies to progress and regress activities and exercise \u2022 Develop agreed criteria that are perceived as realistic Explain that \u2018recovery\u2019 means return to activities but possibly with mild pain \u2022 Use data to show meaningful improvement (on average) with primary treatment at 12 weeks (Murphy et al, 201833) \u2022 Use data to show that 60% may still have some symptoms at 5 years (Van der Plas et al, 201238) Discuss that it will take 6 to 12 months, challenging and hard to stay motivated \u2022 Set meaningful agreed goals \u2022 Provide feedback and monitoring based on individual preferences Explain that flare-ups are part of the process \u2022 Develop an agreed plan for managing flare-ups Figure 5. Education strategies for Achilles tendinopathy.","230 Malliaras: Physiotherapy management of Achilles tendinopathy term (Table 2a). Both groups experienced improved self-ef\ufb01cacy and Behaviour change strategies should be considered to optimise reduced fear of movement, which may be explained by education implementation of self-management exercise and activity modi\ufb01ca- about acceptable pain with exercise and activity delivered to both tion strategies, including: addressing beliefs and outcome expecta- groups, which is likely to be very important to allay movement- tions; setting shared goals that are meaningful to the patient (which related fears.86 Silbernagel et al investigated exercise with the addi- requires understanding their motivations to achieve these goals); tion of either continued sport activity (if pain during activity was action intention strategies, feedback and monitoring dependent on acceptable) or forced cessation of sport activity for 6 weeks individual needs and preferences; and practical strategies to improve (regardless of pain). There was low certainty of evidence for no exercise self-ef\ufb01cacy (eg, providing videos of exercises).87 between-group differences for composite pain and disability or pain outcomes at all timepoints (Table 2a) but evidence of worse physical Exercise function at 12 months (relative harm) in the group that was forced to stop sport. Clinical practice guidelines, therefore, recommend Thirty-nine trials (40 reports)32,88-125 were identi\ufb01ed that inves- continued activity within acceptable pain limits rather than forced tigated exercise (versus control, other exercise or adjuncts) for rest from activity.53 Achilles tendinopathy. However, most trials were underpowered and included diverse interventions and outcomes, which limits data In practice, activity modi\ufb01cation is the most effective strategy to pooling (see Appendix 4); these issues limit the certainty of conclu- manage pain and therefore critical to the success of physiotherapy sions. The most common exercise programs were the Afredson management. Removing provocative (ascertained from the history) eccentric program (or modi\ufb01ed versions), the Silbernagel program, and higher-intensity stretch-shorten cycle activities (faster running or the Stanish and Curwin program and the heavy slow resistance walking or hills) and reducing volume by \u0015 50% is often necessary for program, as described in Table 3. a period of 2 to 6 weeks when load tolerance is severely compro- mised (Figure 2). It is important to appreciate that although un- Is exercise superior to wait-and-see or placebo? common, people with very high pain levels may bene\ufb01t from ceasing provocative activity (100% reduction) in the short term (eg, stopping No studies compared exercise to placebo, and only two trials32,90 sport or wearing a boot if walking is the provocative activity). compared the eccentric program with a wait-and-see approach Explaining the rationale for activity modi\ufb01cation, especially for peo- (Table 2a, Appendix 1). There was low to very low certainty of evi- ple who adopt an \u2018activity endurance strategy\u2019 despite pain,69 will dence for bene\ufb01t favouring exercise at the medium term (clinically facilitate setting shared and effective activity goals. People who adopt important for pain\/disability). A recent network meta-analysis found the opposing coping strategy (ie, activity avoidance,69 often driven by that most of the treatments being assessed (ie, exercise alone or fear that activity will make their condition worse28,29) will bene\ufb01t combined with extracorporeal shockwave therapy [ESWT] or from gradual activity exposure coupled with acceptable pain educa- tion and strategies to monitor pain and activity (Figure 6). Graded exposure = progression based on pain Graded walking\/running and perceived threat or apprehension Stretch-shorten cycle criteria Graded increase in volume\/ Graded increase in range and \u2022 Acceptable pain during the activities intensity at comfortable range volume\/intensity \u2022 Adequate strength (see text) Isometric criterion Isotonic criterion Details \u2022 Start here if isotonic load is \u2022 Commence as soon as pain during the \u2022 Develop volume based on individual isotonic exercise is acceptable unacceptably painful demands then intensity (eg, speed, Details hop or jump intensity) Details \u2022 Progress over a step to achieve peak \u2022 Position in range does not matter Achilles force in ankle dorsiflexion unless the goal is range-specific \u2022 Mix of straight and bent knee is optimal strength gains \u2022 Select 1-3 exercises based on patient \u2022 Select 1-3 exercise based on patient preference preference Plyometric training Sports specific jumping, high-speed running and change of direction Strength and power training principles Develop range-specific strength Address capacity = progression based on capacity Figure 6. Graded exposure and addressing capacity exercise approaches in Achilles tendinopathy.","Table 3 Characteristics of common exercise interventions for Achilles tendinopathy. Characteristic Exercise intervention Modi\ufb01ed Alfredson32,97,98,103,119,124 Stanish and Curwin174,175 Alfredson176 1 to 2\/d Frequency 1\/d 1\/d Not speci\ufb01ed Intensity Not speci\ufb01ed Not speci\ufb01ed Sets 5 3 3 Reps 10 to 20 15 15 Duration (wk) 12 12 Tempo 12 Not speci\ufb01ed Not speci\ufb01ed Rest Not speci\ufb01ed Not speci\ufb01ed Fast eccentric phase Exercise and Isolated eccentric Isolated eccentric, standing contraction Not speci\ufb01ed standing calf calf raise single leg, knee type bent and knee straight 1. Warm-up raise, single leg, knee Progression 2. Static stretch bent and knee straight Load 3. Concentric-eccentric single Load Increasing sets or reps as leg, standing calf raise exercise tolerated97,103,125 or based 4. Static stretching on time32,119,124 or 1 x\/d98 5. Ice application Enough load to achieve Speed then load up to moderate pain Home-based Progression Add speed then load to Add load to achieve rules and achieve pain target pain target Allowed if can be performed tailoring with mild symptoms Pain Enough load to be Enough load to achieve painful in the third set up to moderate pain Setting Home-based Activity advice Home-based Not speci\ufb01ed Allowed if can be performed with mild symptoms HSR = heavy slow resistance, Reps = repetitions, RPE = Rating of perceived exertion.","4,125 Silbernagel9 Heavy slow resistance (HSR)91 Comment 1\/d 3\/wk \u000f 2\/d to 3\/wk Research 231 Goal of 15 reps but intensity target 15RM wk 1 \u000f One (20%) program sets intensity (eg, fatigue, RPE, reps in 12RM wk 2 to 3 10RM wk 4 to 5 targets reserve) not clearly stated \u000f HSR is the most intense program 8RM wk 6 to 8 3 6RM wk 9 to 12 \u000f Reps\/wk = 107 (average for HSR) to \u0014 15 1,260 (Alfredson) 12 to 26 3 to 4 Faster calf raise is a progression 15 to 6 \u000f One (20%) program is . 12 wks Not speci\ufb01ed \u000f Two (40%) programs add speed 12 \u000f One (20%) program describes rest 1. Concentric-eccentric 3 s\/phase (6 s total) double leg standing calf raise 2 to 3 min between sets, time 5 min between exercises \u000f Most (60%) use concentric-eccentric 2. Single leg Concentric eccentric, double 3. Isolated eccentric leg gym exercises (seated combined 4. Faster calf raises calf raise, standing calf raise, \u000f Silbernagel is the most varied, 5. Plyometric training calf raise in a leg press machine) 1. Increasing range of motion although \u2018plyometric\u2019 not described 2. Over a step Load 3. Increasing repetitions \u000f Silbernagel is the most progressive 4. Adding load Time-based intensity progression \u000f Silbernagel and modi\ufb01ed Alfredson Intensity and number of are the most \ufb02exible reps based on patient status Pain is not a goal but acceptable \u000f 60% of programs try to achieve some Pain is not a goal but acceptable Gym-based pain Home-based No sport for \ufb01rst 3 wks then \u000f 80% of programs are home-based Allowed if can be performed allowed if can be performed \u000f Most programs (60%) allow sport if with mild symptoms with mild symptoms symptoms are mild","232 Malliaras: Physiotherapy management of Achilles tendinopathy supplements, ESWT, acupuncture) were superior to wait-and-see at Phase one involves isometric exercise, which may be a necessary 12 weeks.126 starting point for patients with unacceptable symptoms during isotonic exercise. This is because a point in range can usually be found Is there an exercise approach that can be recommended based on where symptoms are tolerable. Although early evidence suggested the evidence? that isometric may also achieve superior immediate pain reduction compared with isotonic loading, these \ufb01ndings have been chal- Nine trials91-99 compared different exercise programs among lenged.135 Phase two, involving isotonic (in terms of external load not people with Achilles tendinopathy (Table 2a and Appendix 1). A muscle-tendon unit load) exercise should be the starting point if version of the Alfredson program was used in each of these trials. tolerable; otherwise, the goal should be to progress to this phase as There was evidence (moderate certainty at longer term and very low soon as possible. Exercise in this phase is generally performed slowly certainty at medium term) for global change bene\ufb01t of Silbernagel or (using a metronome can help to ensure it is slow with good heavy slow resistance programs over Alfredson. There was also con- technique - set metronome at 60 bpm and aim for 2 to 3 seconds for \ufb02icting bene\ufb01t of other outcomes (only at medium term for pain\/ each of the concentric and eccentric phases) through full range of disability and short term for pain) favouring alternative programs motion to ensure that tolerance and con\ufb01dence is achieved through (heterogeneous programs but all involved lower volume than range. Progressing as tolerated to full ankle dorsi\ufb02exion range of Alfredson, which may be more convenient) over the Alfredson pro- motion is important, as this is where greatest Achilles tendon loads gram Murphy et al58 reported a similar outcome in a recent meta- occur136 (see caveat below for insertional Achilles tendinopathy). analysis. There were very low certainty bene\ufb01ts for eccentric over Programs such as the modi\ufb01ed Alfredson, Silbernagel and heavy slow concentric exercise at short (pain) and medium term (global change, resistance (which involve exercise between 3x\/week and daily) are which was also clinically important). However, the trial94 reporting recommended (based on pooled data above) (Table 3). Phase three bene\ufb01t at medium term investigated a very low volume and intensity involves graded introduction of stretch-shorten cycles such as concentric exercise program compared with a very high volume and walking and running activities based on agreed goals (generally two intensity eccentric exercise program, so it is unclear which variable(s) to three times per week with rest days in between). This phase may explain the bene\ufb01t. Clinicians should consider the modi\ufb01ed involves higher tendon loads and risk of pain \ufb02are than prior phases, Alfredson, Silbernagel and heavy slow resistance programs as alter- so careful monitoring of load tolerance is needed (Figure 2). To ensure natives to the Alfredson program. Given that they are heterogeneous it is gradual, introduction of running or walking loads should be (Table 3), clinicians should adopt an individualised and patient- guided by objective data such as running time and intensity (eg, pace) centred approach to exercise prescription, with consideration of pa- and step count data, respectively. tient preferences (eg, for exercise frequency or support) and moti- vation, as well as resources (eg, time and equipment). How long an individual takes to progress through each phase is highly variable and depends on their load tolerance and progression Implications from the exercise literature regarding exercise (eg, if load tolerant, some patients may progress to restoration of mechanisms stretch-shorten cycle loads even in the \ufb01rst session). The isotonic and stretch-shorten cycle phases are generally performed two to three The biomedical view is that exercise for Achilles tendinopathy times per week each. Ideally, stretch-shorten cycle exercise is per- (and tendinopathy more generally) improves outcomes (such as pain, formed prior to (eg, in the morning of) isotonic exercise to enable disability and function) via local (muscle-tendon) mechanical mech- complete rest days (only possible with isotonic programs that are anisms.127 Potential mechanisms include disrupting pain-producing performed 3 times per week), or an alternative is performing stretch- neurovascular ingrowth into the tendon128 or stimulating tendon shorten cycle exercise on the days off from isotonic loads. healing via collagen synthesis.129 Development of exercise strength (eg, neuromuscular activation or hypertrophy) may improve Addressing speci\ufb01c capacity disability and function, although pathways linking strength with Despite improved pain following rehabilitation, some patients improved pain are less clear.130 Local mechanical mechanisms need to be questioned because exercise does not appear to consistently in- with Achilles tendinopathy have persisting calf strength and power \ufb02uence markers of neurovascular ingrowth,131 tendon healing capacity impairments.137 This is important because calf strength observed with imaging132 or muscular strength,61 despite improve- impairment is a risk factor for onset of Achilles tendinopathy138 and ments in self-reported pain, disability and other outcomes. Compa- may be related to recurrence. The calf needs to contract strongly to rable bene\ufb01ts from vastly different exercise approaches observed in enable the tendon to store and release energy during stretch-shorten this review and others suggest that other mechanisms such as cycle activities.13 Tendon energy storage and release in turn reduces improved self-ef\ufb01cacy or reduced movement-related fear may energy cost of locomotion and enables greater power production.13 mediate exercise outcomes. In a recent meta-analysis, Smith et al133 Therefore, speci\ufb01c considerations for recovering calf strength and found low certainty evidence that pain outcomes may be improved power may assist some patients to regain optimal function. in the short and medium term with painful rather than pain-free exercises. Advising painful exercise normalises pain that patients Suggestions for how the graded exposure phases can be adapted with Achilles tendinopathy typically experience with exercise and to consider capacity are provided in Figure 6. In the isotonic phase, may positively impact on movement-related fear.86 progression to higher intensity exercise (as in the heavy slow resis- tance program), including lower repetition ranges, can target Suggested exercise approaches based on the existing literature maximal strength, and adding speed (or intent to move quickly) can address rate of force development.139 Eccentric overload training Graded exposure in Achilles tendinopathy rehabilitation (performed with a load that cannot be shifted concentrically) can be Given this proposed view of exercise mechanisms, exercise for considered to maximise mechanical stimulus, tissue adaptation po- tential, and neural adaptations from resistance training.140,141 Achilles tendinopathy can be re-conceptualised as \u2018graded exposure\u2019. Per-session training intensity can be monitored and adjusted using The goal is to expose the person to activities that are \u2018graded\u2019 based repetitions in reserve or exercise intensity scales (eg, OMNI scale).142 on perceived threat and apprehension (useful for those with kinesi- It is important to train both sides individually (good principle for all ophobia and fear-avoidance) but also based on patient-reported level isometric and isotonic loading unless the individual can only manage of pain (Figure 6).134 Progression generally involves increasing dif\ufb01- bilateral exercise). Often the load lifted can be higher on the unaf- cult of exercise (eg, adding external load; introducing low and then fected side to achieve the same intensity target. Inclusion of straight high intensity stretch-shorten cycle), which provides an opportunity and bent-knee options will ensure adequate focus on soleus to reconceptualise pain and threat value.130 (Figure 6). Load can be added using a backpack containing weights in standing or with 20 kg sandbags (accessible at hardware stores) in seated (although the practical limit seems to be 60 kg [3 3 20 kg sandbags] with this option, so it is only useful for patients with","Research 233 poorer strength). Utilising gym machines (based on patient prefer- Evidence for adjuncts as an alternative to primary treatment ence and access) such as the Smith, standing or seated calf raise Thirteen trials32,88,105,116-125 investigated exercise versus adjuncts machines can enable higher load intensity progression. and a further four investigated an adjunct versus control32,147-149 A consideration related to higher intensity exercise is tendon (Table 2a, Appendix 1). Although limited, there is some evidence adaptation. Healthy tendon adapts (becomes more load resistant, eg, for standalone alternatives to exercise care. There was low certainty increased stiffness) with high-intensity exercise (. 70% maximal evidence that acupuncture may be superior to exercise for pain\/ voluntary contraction (MVC), equating to 4.5 to 6.5% strain) regard- disability across all timepoints (clinically meaningful in the short and less of contraction type (isometric or isotonic).127 It is unknown how medium term). The only other positive evidence was for heel wedges, this translates to pathological tendons. Regardless, ensuring that at for which there was very low certainty evidence for bene\ufb01t over least one exercise achieves a high intensity may increase the chance exercise for pain\/disability, pain and global change outcomes in the of tendon adaptation, assuming that this is possible for pathological medium term, yet con\ufb02icting bene\ufb01t in the short term. Very low tendon.127 Issues with this approach are that there are no reliable certainty evidence suggested that traditional physiotherapy including means of assessing MVC in a clinical setting and tendon strain for this frictions and ultrasound (pain\/disability at short and medium term) MVC level varies between individuals.127 A pragmatic clinical may be less bene\ufb01cial than exercise, so is an unsuitable alternative approach is to utilise an OMNI scale rating of 8 to 9 out of 10,143 based treatment. There was low to very low certainty evidence for no dif- on individual comfort. ference (for a majority of comparisons) between other adjuncts (ESWT, night splint, AirHeel brace, pressure massage, ultrasound) and In the stretch-shorten cycle phase, reactive strength capabilities exercise for various timepoints and outcomes (Table 2a). Although can be developed via progressive plyometric interventions (landing, within-group bene\ufb01ts over time for these adjuncts are generally jumping and hopping; see here for an example for recreational run- clinically important, the estimates of the between-group difference ners144). Depending on individual goals, development of sports- often have a lot of inherent uncertainty so the interventions should speci\ufb01c high-speed running and change of direction competencies not be considered equivalent without careful consideration of the can also be considered. Given that adequate strength is required for con\ufb01dence interval around the between-group difference. Given that power, muscle strength targets should be considered prior to power exercise is a more established treatment (with bene\ufb01ts over wait- progression. For example, achieving (or approaching) 1.5 times body and-see), the best alternative options are those with some evidence weight 6RM in seated calf raise (or 0.5 times body weight in standing) of superiority over exercise. is desirable for recreational runners prior to running and submaximal hopping.74 Adequate maximal strength should be the goal prior to Comparisons between adjuncts and placebo showed that LLLT and commencing maximal plyometric or high-speed running acupuncture may not show bene\ufb01t (very low certainty evidence for progressions. pain\/disability in the short and medium term) (Table 2a). The acupuncture \ufb01nding cautions against overenthusiasm in relation to A common capacity impairment is inability to perform a single leg the acupuncture bene\ufb01t over exercise derived from a single study calf raise (at all or with adequate height), which indicates that (described above). There was low certainty evidence for bene\ufb01t of standing calf raise 1RM is lower than the individual\u2019s body weight. radial ESWT compared with wait-and-see (for two out of three out- This is a challenging impairment that requires a speci\ufb01c exercise comes, Table 2) but no bene\ufb01t of focused ESWT versus placebo. Given strategy. One option is to start with double leg concentric-eccentric the differences in the type of ESWT and comparator it is dif\ufb01cult to loading. Utilising bathroom scales under each leg can ensure that explain these \ufb01ndings. the unaffected leg is not being favoured and the goal is to eventually move towards 60 to 100% body weight on the affected side (ensuring Insertional Achilles tendinopathy adequate planter\ufb02exor height prior to progression). Another option There is much less evidence speci\ufb01cally related to insertional that enables more speci\ufb01c and measurable load application is per- forming a seated single-leg calf raise with a gradually increasing Achilles tendinopathy, including no trials comparing exercise with proportion of body weight based on individual capacity. These op- control or different exercise programs. Three trials103,110,111 investi- tions can be supplemented by performing isometrics near full plan- gated the addition of adjuncts to exercise, and two117,118 compared tar\ufb02exion to achieve strength gains speci\ufb01c to this part of the adjuncts with exercise (Table 2b, Appendix 1). There was low cer- range145 or performing eccentric-only exercise, which may be tainty evidence for bene\ufb01t when adding instrumented manual ther- possible (when concentric-eccentric is not) given it utilises fewer apy to exercise for pain\/disability at medium and longer term active motor units than concentric contraction when external load (clinically important for medium term), but not for pain\/disability at and speed are constant.146 Progression to concentric is required, as short term, global change at medium term or pain at any timepoint. strength adaptation may be speci\ufb01c to contraction type.146 There was high certainty of evidence (pain) and moderate certainty evidence (pain\/disability) of no worthwhile bene\ufb01t from adding Adjuncts ESWT to exercise, but low certainty evidence that ESWT alone was more effective than exercise (range of outcomes at medium term, Evidence for adjuncts added to primary treatment Table 2b). Interestingly, these data contrast with midportion Achilles Eighteen trials (19 reports)98,100-115,119,124 investigated whether tendinopathy, where the addition of ESWT to exercise may be war- ranted but not as a standalone treatment. A possible explanation is various adjuncts improved outcomes when added to exercise poorer exercise outcomes, which have been reported for insertional (Table 2a, Appendix 1). There was low certainty evidence that, when Achilles tendinopathy.150 There was very low certainty evidence for added to exercise, radial ESWT may improve pain\/disability and pain no difference between stretch, massage, bilateral heel lifts, night outcomes at all time points and global change at medium term splint with or without Alfredson eccentric exercise over a step (pain (Table 2a). Importantly, the bene\ufb01ts were still present even when a at medium term). This may be because Alfredson eccentrics into full placebo comparator was used (used in one112 of the two studies108). range of dorsi\ufb02exion, stretching and night splints may increase In contrast, there was very low certainty evidence for con\ufb02icting or no insertional tendon compression and provoke symptoms. bene\ufb01t when adding focused or combined radial\/focused ESWT compared with a placebo (pain and global change across various Despite the lack of speci\ufb01c evidence, the management principles timepoints, Table 2a). Current evidence favours the addition of radial outlined in this review apply to insertional Achilles tendinopathy. over other forms of ESWT but, given the low certainty of evidence, Based on case series evidence of poor outcomes with a traditional large trials comparing this modality to placebo are needed. There was Alfredson eccentric training (exercise over a step into ankle dorsi- low to very low certainty evidence that the addition of other adjuncts \ufb02exion),150 there are suggestions to limit ankle dorsi\ufb02exion (and to exercise (including low-level laser therapy [LLLT], AirHeel brace, thereby compression) in insertional Achilles tendinopathy.151 Care is night splint, needling, pressure massage and custom orthotics) did needed, as complete avoidance of ankle dorsi\ufb02exion may have a not add bene\ufb01t (Table 2a). nocebo effect (prolonging recovery)152 and may reduce outcomes,","234 Malliaras: Physiotherapy management of Achilles tendinopathy Figure 7. Progression of ankle dorsi\ufb02exion during calf raise exercise for insertional Achilles tendinopathy. given that tendon force is greatest in ankle dorsi\ufb02exion.136 Advice in Acknowledgements: Sincere thanks to Sanam Tavakkoli Oskouei the isotonic phase should be to limit ankle dorsi\ufb02exion if it is painful who assisted with data quality appraisal, risk of bias assessment, data but to aim for graded introduction of dorsi\ufb02exion as symptoms allow. extraction and pooling and data checking. Custom-made small steps of increasing size (10 mm, 20 mm, 30 mm, Figure 7) can be used to progress dorsi\ufb02exion range based on toler- Provenance: Invited. Peer reviewed. ance (avoiding extreme end range, so not everyone will progress to Correspondence: Peter Malliaras, Department of 20 mm or 30 mm). Physiotherapy, Monash University, Melbourne, Australia. Email: [email protected] Current research trends and future directions References More Achilles tendinopathy research is being published than ever 1. Scott A, Squier K, Alfredson H, Bahr R, Cook JL, Coombes B, et al. Icon 2019: in- before (28 of the 44 trials identi\ufb01ed in this review were published ternational scienti\ufb01c tendinopathy symposium consensus: clinical terminology. within the last decade); however, these efforts are not translating to Brit J Sports Med. 2020;54:260\u2013262. improved knowledge and patient bene\ufb01t. Most studies were under- powered and focused on the use of adjunct interventions (32 of 41, 2. Lopes AD, Hespanhol LC, Yeung SS, Costa LOP. What are the main running-related 78%) rather than different exercise types or education and advice musculoskeletal injuries? Sports Med. 2012;42:891\u2013905. interventions. They investigated heterogeneous outcomes, which hampered pooling in meta-analyses. Large, adequately powered trials 3. Kujala UM, Sarna S, Kaprio J. Cumulative incidence of achilles tendon rupture and are needed to improve knowledge of primary physiotherapy in- tendinopathy in male former elite athletes. Clin J Sport Med. 2005;15:133\u2013135. terventions (education, advice and exercise) for insertional particu- larly but also midportion Achilles tendinopathy. This should include 4. de Jonge S, Van den Berg C, de Vos R-J, Van Der Heide HJ, Weir A, Verhaar JA, et al. investigation of the bene\ufb01t of education as a standalone treatment Incidence of midportion Achilles tendinopathy in the general population. Brit J and new exercise approaches (eg, stretch-shorten cycle loading for Sports Med. 2011;45:1026\u20131028. certain athletic groups and use of technology to monitor and improve exercise adherence and \ufb01delity). International consensus efforts are 5. Yasui Y, Tonogai I, Rosenbaum AJ, Shimozono Y, Kawano H, Kennedy JG. The risk of currently underway to raise standards of Achilles tendinopathy Achilles tendon rupture in the patients with Achilles tendinopathy: healthcare research (eg, developing a core outcome set153 and consensus on database analysis in the United States. BioMed Res Int. 2017;2017. diagnosis) as well as addressing unmet needs (eg, developing stake- holder informed education for this condition). Despite the recogni- 6. Chimenti RL, Cychosz CC, Hall MM, Phisitkul P. Current concepts review update: tion of biopsychosocial aspects of Achilles tendinopathy, basic insertional Achilles tendinopathy. Foot Ank Int. 2017;38:1160\u20131169. biological knowledge is still lacking and an exciting new area of enquiry is focusing on subtendon (from the muscle components of 7. Kvist M. Achilles tendon injuries in athletes. 1991:188\u2013201. the calf) function and implications for pathology and treatment (eg, 8. Fahlstr\u00f6m M, Lorentzon R, Alfredson H. Painful conditions in the Achilles tendon the study by Yin et al154). Understanding how biopsychosocial factors in\ufb02uence risk, impairments and treatment outcomes will be critical region in elite badminton players. Am J Sports Med. 2002;30:51\u201354. to drive development of new patient-centred and more effective 9. Silbernagel KG, Thome\u00e9 R, Eriksson BI, Karlsson J. Continued sports activity, using a treatments. pain-monitoring model, during rehabilitation in patients with Achilles tendinop- eAddenda: Appendices 1, 2, 3 and 4 can be found online at https:\/\/ athy: a randomized controlled study. Am J Sports Med. 2007;35:897\u2013906. doi.org\/10.1016\/j.jphys.2022.09.010 10. Gr\u00e4vare Silbernagel K, Crossley KM. A proposed return-to-sport program for pa- tients with midportion Achilles tendinopathy: rationale and implementation. Ethics approval: Not applicable. J Orthop Sports Phys Ther. 2015;45:876\u2013886. Competing interests: Nil. 11. Lichtwark GA, Wilson A. Interactions between the human gastrocnemius muscle Source(s) of support: Nil. and the Achilles tendon during incline, level and decline locomotion. J Exp Biol. 2006;209:4379\u20134388. 12. Komi PV. Stretch-shortening cycle: a powerful model to study normal and fatigued muscle. J Biomech. 2000;33:1197\u20131206. 13. Roberts TJ, Konow N. How tendons buffer energy dissipation by muscle. Ex Sport Sci Rev. 2013;41:4. 14. Willy RW, Halsey L, Hayek A, Johnson H, Willson JD. Patellofemoral joint and Achilles tendon loads during overground and treadmill running. J Orthop Sports Phys Ther. 2016;46:664\u2013672. 15. Almonroeder T, Willson JD, Kernozek TW. The effect of foot strike pattern on Achilles tendon load during running. Ann Biomech Eng. 2013;41:1758\u20131766. 16. Clement D, Taunton J, Smart G. Achilles tendinitis and peritendinitis: etiology and treatment. Am J Sports Med. 1984;12:179\u2013184. 17. J\u00e4rvinen TA, Kannus P, Maffulli N, Khan KM. Achilles tendon disorders: etiology and epidemiology. Foot Ank Clin. 2005;10:255\u2013266.","Research 235 18. Di Caprio F, Buda R, Mosca M, Giannini S. Foot and lower limb diseases in runners: 48. Reiman M, Burgi C, Strube E, Prue K, Ray K, Elliott A, et al. The utility of clinical assessment of risk factors. J Sports Sci Med. 2010;9:587. measures for the diagnosis of Achilles tendon injuries: a systematic review with meta-analysis. J Athl Train. 2014;49:820\u2013829. 19. Mersmann F, Bohm S, Schroll A, Boeth H, Duda G, Arampatzis A. Muscle and tendon adaptation in adolescent athletes: a longitudinal study. Scand J Med Sci Sports. 49. Vallance P, Hasani F, Crowley L, Malliaras P. Self-reported pain with single leg heel 2017;27:75\u201382. raise or single leg hop offer distinct information as measures of severity in men with midportion and insertional Achilles tendinopathy: an observational cross- 20. Kj\u00e6r M, Magnusson P, Krogsgaard M, M\u00f8ller JB, Olesen J, Heinemeier K, et al. sectional study. Phys Ther Sport. 2021;47:23\u201331. Extracellular matrix adaptation of tendon and skeletal muscle to exercise. J Anat. 2006;208:445\u2013450. 50. Cook JL, Khan K, Kiss ZS, Purdam CR, Grif\ufb01ths L. Reproducibility and clinical utility of tendon palpation to detect patellar tendinopathy in young basketball players. 21. Heinemeier K, Olesen J, Haddad F, Langberg H, Kjaer M, Baldwin KM, et al. Brit J Sports Med. 2001;35:65\u201369. Expression of collagen and related growth factors in rat tendon and skeletal muscle in response to speci\ufb01c contraction types. J Physiol. 2007;582:1303\u20131316. 51. Docking SI, Ooi CC, Connell D. Tendinopathy: is imaging telling us the entire story? J Orthop Sports Phys Ther. 2015;45:842\u2013852. 22. Cook JL, Purdam C. Is compressive load a factor in the development of tendinop- athy? Brit J Sports Med. 2012;46:163\u2013168. 52. Lewis JS, Cook CE, Hoffmann TC, O\u2019Sullivan P. The elephant in the room: too much medicine in musculoskeletal practice. J Orthop Sports Phys Ther. 2020;50:1\u20134. 23. Malliaras P, O\u2019Neill S. Potential risk factors leading to tendinopathy. Apunts Medicina de l\u2019Esport. 2017;52:71\u201377. 53. Martin RL, Chimenti R, Cuddeford T, Houck J, Matheson JW, McDonough CM, et al. Achilles pain, stiffness, and muscle power de\ufb01cits: Midportion Achilles tendinop- 24. Van Der Vlist AC, Breda SJ, Oei EH, Verhaar JA, de Vos R-J. Clinical risk factors for athy revision 2018: Clinical practice guidelines linked to the International Classi- Achilles tendinopathy: a systematic review. Brit J Sports Med. 2019;53:1352\u2013 \ufb01cation of Functioning, Disability and Health from the Orthopaedic Section of the 1361. American Physical Therapy Association. J Orthop Sports Phys Ther. 2018;48:A1\u2013A38. 25. McAuliffe S, McCreesh K, Culloty F, Purtill H, O\u2019Sullivan K. Can ultrasound imaging 54. Jozsa L, Kannus P. Histopathological \ufb01ndings in spontaneous tendon ruptures. predict the development of Achilles and patellar tendinopathy? A systematic re- Scand J Med Sci Sports. 1997;7:113\u2013118. view and meta-analysis. Brit J Sports Med. 2016;50:516\u20131523. 55. Malliaras P, Cook J, Purdam C, Rio E. Patellar tendinopathy: clinical diagnosis, load 26. Ceravolo ML, Gaida JE, Keegan RJ. Quality-of-life in Achilles tendinopathy: an management, and advice for challenging case presentations. J Orthop Sports Phys exploratory study. Clin J Sports Med. 2020;30:495\u2013502. Ther. 2015;45:887\u2013898. 27. Mallows A, Head J, Goom T, Malliaras P, O\u2019Neill S, Smith B. Patient perspectives on 56. Thome\u00e9 R. A comprehensive treatment approach for patellofemoral pain syn- participation in exercise-based rehabilitation for Achilles tendinopathy: A quali- drome in young women. Phys Ther. 1997;77:1690\u20131703. tative study. Musculoskelet Sci Pract. 2021;56:102450. 57. Von Korff M, DeBar LL, Krebs EE, Kerns RD, Deyo RA, Keefe FJ. Graded chronic 28. Mc Auliffe S, Synott A, Casey H, Mc Creesh K, Purtill H, O\u2019Sullivan K. Beyond the pain scale revised: mild, bothersome, and high impact chronic pain. Pain. tendon: experiences and perceptions of people with persistent Achilles tendin- 2020;161:651. opathy. Musculoskelet Sci Pract. 2017;29:108\u2013114. 58. Murphy MC, Travers MJ, Chivers P, Debenham JR, Docking SI, Rio EK, et al. Ef\ufb01cacy 29. Turner J, Malliaras P, Goulis J, Mc Auliffe S. \u201cIt\u2019s disappointing and it\u2019s pretty of heavy eccentric calf training for treating mid-portion Achilles tendinopathy: a frustrating, because it feels like it\u2019s something that will never go away.\u201d A quali- systematic review and meta-analysis. Brit J Sports Med. 2019;53:1070\u20131077. tative study exploring individuals\u2019 beliefs and experiences of Achilles tendinop- athy. PloS one. 2020;15:e0233459. 59. Oskouei ST, Malliaras P, Hill KD, Clark R, Perraton L. Evaluating daily physical ac- tivity and biomechanical measures using wearable technology in people with 30. Visser TSS, Van Der Vlist AC, Van Oosterom RF, Van Veldhoven P, Verhaar JA, de Achilles tendinopathy: A descriptive exploratory study. Musculoskeletal Sci Prac. Vos R-J. Impact of chronic Achilles tendinopathy on health-related quality of life, 2022;58:102534. work performance, healthcare utilisation and costs. BMJ Open Sport Exerc Med. 2021;7:e001023. 60. Turk DC. The biopsychosocial approach to the assessment and intervention for people with musculoskeletal disorders. Handbook of Musculoskeletal Pain and Disability 31. Williams A, Kamper SJ, Wiggers JH, O\u2019Brien KM, Lee H, Wolfenden L, et al. Disorders in the Workplace. Springer; 2014:341\u2013363. Musculoskeletal conditions may increase the risk of chronic disease: a systematic review and meta-analysis of cohort studies. BMC Med. 2018;16:1\u20139. 61. Hasani F, Vallance P, Haines T, Munteanu SE, Malliaras P. Are Plantar\ufb02exor Muscle Impairments Present Among Individuals with Achilles Tendinopathy and Do They 32. Rompe J, Nafe B, Furia J, Maffulli N. Eccentric loading, shock-wave treatment, or a Change with Exercise? A Systematic Review with Meta-analysis. Sports Med Open. wait-and-see policy for tendinopathy of the main body of tendo Achillis: a ran- 2021;7:1\u201318. domized controlled trial. Am J Sports Med. 2007;35:374\u2013383. 62. McAuliffe S, Tabuena A, McCreesh K, O\u2019Keeffe M, Hurley J, Comyns T, et al. Altered 33. Murphy M, Travers M, Gibson W, Chivers P, Debenham J, Docking S, et al. Rate of strength pro\ufb01le in Achilles tendinopathy: a systematic review and meta-analysis. improvement of pain and function in mid-portion Achilles tendinopathy with J Athl Train. 2019;54:889\u2013900. loading protocols: a systematic review and longitudinal meta-analysis. Sports Med. 2018;48:1875\u20131891. 63. Ogbonmwan I, Kumar BD, Paton B. New lower-limb gait biomechanical charac- teristics in individuals with Achilles tendinopathy: a systematic review update. 34. Sayana MK, Maffulli N. Eccentric calf muscle training in non-athletic patients with Gait Posture. 2018;62:146\u2013156. Achilles tendinopathy. J Sci Med Sport. 2007;10:52\u201358. 64. Sancho I, Malliaras P, Barton C, Willy RW, Morrissey D. Biomechanical alterations in 35. Ram R, Meeuwisse W, Patel C, Wiseman DA, Wiley JP. The limited effectiveness of a individuals with Achilles tendinopathy during running and hopping: A systematic home-based eccentric training for treatment of Achilles tendinopathy. Clin Invest review with meta-analysis. Gait Posture. 2019;73:189\u2013201. Med. 2013:E197\u2013E206. 65. Hanlon SL, Pohlig RT, Silbernagel KG. Beyond the diagnosis: using patient char- 36. Lagas IF, Fokkema T, Bierma-Zeinstra SM, Verhaar JA, van Middelkoop M, de Vos RJ. acteristics and domains of tendon health to identify latent subgroups of Achilles How many runners with new-onset Achilles tendinopathy develop persisting tendinopathy. J Orthop Sports Phys Ther. 2021;51:440\u2013448. symptoms? A large prospective cohort study. Scand J Med Sci Sports. 2020;30:1939\u20131948. 66. Mallows A, Debenham J, Walker T, Littlewood C. Association of psychological variables and outcome in tendinopathy: a systematic review. Brit J Sports Med. 37. Silbernagel KG, Brorsson A, Lundberg M. The majority of patients with Achilles 2017;51:743\u2013748. tendinopathy recover fully when treated with exercise alone: a 5-year follow-up. Am J Sports Med. 2011;39:607\u2013613. 67. Stubbs C, Mc Auliffe S, Mallows A, O\u2019sullivan K, Haines T, Malliaras P. The strength of association between psychological factors and clinical outcome in tendinopathy: 38. Van der Plas A, de Jonge S, de Vos R-J, Van Der Heide HJ, Verhaar JA, Weir A, et al. A systematic review. PloS one. 2020;15:e0242568. A 5-year follow-up study of Alfredson\u2019s heel-drop exercise programme in chronic midportion Achilles tendinopathy. Brit J Sports Med. 2012;46:214\u2013218. 68. Robinson J, Cook JL, Purdam C, Visentini PJ, Ross J, Maffulli N, et al. The VISA-A questionnaire: a valid and reliable index of the clinical severity of Achilles ten- 39. Lagas I, Tol J, Weir A. One in four patients with midportion achilles tendinopathy dinopathy. Brit J Sports Med. 2001;35:335\u2013341. has persisting symptoms after 10 years: a prospective cohort study. Am J Sports Med. 2019. 69. Hasenbring MI, Verbunt JA. Fear-avoidance and endurance-related responses to pain: new models of behavior and their consequences for clinical practice. Clin J 40. Gajhede-Knudsen M, Ekstrand J, Magnusson H, Maffulli N. Recurrence of Achilles Pain. 2010;26:747\u2013753. tendon injuries in elite male football players is more common after early return to play: an 11-year follow-up of the UEFA Champions League injury study. Brit J Sports 70. H\u00e9bert-Losier K, Wessman C, Alricsson M, Svantesson U. Updated reliability and Med. 2013;47:763\u2013768. normative values for the standing heel-rise test in healthy adults. Physiotherapy. 2017;103:446\u2013452. 41. Docking SI, Rio E, Cook J, Orchard JW, Fortington LV. The prevalence of Achilles and patellar tendon injuries in Australian football players beyond a time-loss de\ufb01ni- 71. Albracht K, Arampatzis A, Baltzopoulos V. Assessment of muscle volume and tion. Scand J Med Sci Sports. 2018;28:2016\u20132022. physiological cross-sectional area of the human triceps surae muscle in vivo. J Biomech. 2008;41:2211\u20132218. 42. de Vos R-J, Van Der Vlist AC, Zwerver J, Meuffels DE, Smithuis F, Van Ingen R, et al. Dutch multidisciplinary guideline on Achilles tendinopathy. Brit J Sports Med. 72. Reid D, McNair PJ, Johnson S, Potts G, Witvrouw E, Mahieu N. Electromyographic 2021;55:1125\u20131134. analysis of an eccentric calf muscle exercise in persons with and without Achilles tendinopathy. Phys Ther Sport. 2012;13:150\u2013155. 43. Matthews W, Ellis R, Furness J, Hing WA. The clinical diagnosis of Achilles ten- dinopathy: a scoping review. PeerJ. 2021;9:e12166. 73. Habets B, Smits H, Backx F, Van Cingel R, Huisstede B. Hip muscle strength is decreased in middle-aged recreational male athletes with midportion Achilles 44. Rio EK, Mc Auliffe S, Kuipers I, Girdwood M, Alfredson H, Bahr R, et al. ICON PART-T tendinopathy: a cross-sectional study. Phys Ther Sport. 2017;25:55\u201361. 2019\u2013International Scienti\ufb01c Tendinopathy Symposium Consensus: recommended standards for reporting participant characteristics in tendinopathy research (PART- 74. Sancho I, Morrissey D, Willy RW, Tayfur A, Lascurain-Aguirrebe\u00f1a I, Barton C, et al. T). Brit J Sports Med. 2020;54:627\u2013630. Recreational runners with Achilles tendinopathy have clinically detectable im- pairments: A case-control study. Phys Ther Sport. 2022;55:241\u2013247. 45. Docking SI, Hart HF, Rio E, Hannington MC, Cook JL, Culvenor AG. Explaining Variability in the Prevalence of Achilles Tendon Abnormalities: A Systematic Re- 75. O\u2019Neill S, Watson PJ, Barry S. Why are eccentric exercises effective for achilles view With Meta-analysis of Imaging Studies in Asymptomatic Individuals. J Orthop tendinopathy? Int J Sports Phys Ther. 2015;10:552. Sports Phys Ther. 2021;51:232\u2013252. 76. Lyght M, Nockerts M, Kernozek TW, Ragan R. Effects of foot strike and step fre- 46. Hutchison A-M, Evans R, Bodger O, Pallister I, Topliss C, Williams P, et al. What quency on Achilles tendon stress during running. J Appl Biomech. 2016;32:365\u2013372. is the best clinical test for Achilles tendinopathy? Foot Ank Surg. 2013;19: 112\u2013117. 77. Plinsinga ML, Van Wilgen CP, Brink MS, Vuvan V, Stephenson A, Heales LJ, et al. Patellar and Achilles tendinopathies are predominantly peripheral pain states: a 47. Maffulli N, Kenward MG, Testa V, Capasso G, Regine R, King JB. Clinical diagnosis of blinded case control study of somatosensory and psychological pro\ufb01les. Brit J Achilles tendinopathy with tendinosis. Clin J Sport Med. 2003;13:11\u201315. Sports Med. 2018;52:284\u2013291.","236 Malliaras: Physiotherapy management of Achilles tendinopathy 78. Mc Auliffe S, Whiteley R, Malliaras P, O\u2019Sullivan K. Central sensitisation in different 105. Petersen W, Welp R, Rosenbaum D. Chronic Achilles tendinopathy: a prospective tendinopathies: are we comparing apples and oranges? Brit J Sports Med. randomized study comparing the therapeutic effect of eccentric training, 2019;53:142\u2013143. the AirHeel brace, and a combination of both. Am J Sports Med. 2007;35:1659\u2013 1667. 79. Vallance P, Crowley L, Vicenzino B, Malliaras P. Contralateral mechanical hyper- algesia and altered pain modulation in men who have unilateral insertional 106. Tumilty S, McDonough S, Hurley DA, Baxter GD. Clinical effectiveness of low-level Achilles tendinopathy: a cross-sectional study. Musculoskeletal Sci Prac. 2021;52: laser therapy as an adjunct to eccentric exercise for the treatment of Achilles\u2019 102353. tendinopathy: a randomized controlled trial. Arch Phys Med Rehabil. 2012;93:733\u2013 739. 80. Uddin Z, MacDermid JC. Quantitative sensory testing in chronic musculoskeletal pain. Pain Med. 2016;17:1694\u20131703. 107. Tumilty S, Munn J, Abbott JH, McDonough S, Hurley DA, Baxter GD. Laser therapy in the treatment of Achilles tendinopathy: a pilot study. Photomed Laser 81. Vicenzino B, de Vos R-J, Alfredson H, Bahr R, Cook JL, Coombes BK, et al. ICON Surg. 2008;26:25\u201330. 2019\u2014International Scienti\ufb01c Tendinopathy Symposium Consensus: There are nine core health-related domains for tendinopathy (CORE DOMAINS): Delphi study 108. Abdelkader NA, Helmy MNK, Fayaz NA, Saweeres ES. Short-and intermediate- of healthcare professionals and patients. Brit J Sports Med. 2020;54:444\u2013451. term results of extracorporeal shockwave therapy for noninsertional Achilles tendinopathy. Foot Ank Int. 2021;42:788\u2013797. 82. Mallows AJ, Debenham JR, Malliaras P, Stace R, Littlewood C. Cognitive and contextual factors to optimise clinical outcomes in tendinopathy. Brit J Sports Med. 109. Gatz M, Schweda S, Betsch M, Dirrichs T, de la Fuente M, Reinhardt N, et al. Line- 2018;52:822\u2013823. and point-focused extracorporeal shock wave therapy for Achilles tendinopathy: a placebo-controlled RCT study. Sports Health. 2021;13:511\u2013518. 83. F\u00e4rnqvist K, Morrissey D, Malliaras P. Factors associated with outcome following exercise interventions for Achilles tendinopathy: a systematic review. Physiother 110. Mansur NSB, Matsunaga FT, Carrazzone OL, Dos Santos BS, Nunes CG, Res Int. 2021;26:e1889. Aoyama BT, et al. Shockwave therapy plus eccentric exercises versus isolated eccentric exercises for Achilles insertional tendinopathy: a double-blinded ran 84. Mallows A, Jackson J, Littlewood C, Debenham J. The association of working alli- domized clinical trial. JBJS. 2021;103:1295\u20131302. ance, outcome expectation, adherence and self-ef\ufb01cacy with clinical outcomes for Achilles tendinopathy: A feasibility cohort study (the MAP study). Musculoskel 111. Pinitkwamdee S, Laohajaroensombat S, Orapin J, Woratanarat P. Effectiveness of Care. 2020;18:169\u2013176. extracorporeal shockwave therapy in the treatment of chronic insertional Achilles tendinopathy. Foot Ank Int. 2020;41:403\u2013410. 85. Chimenti RL, Post AA, Rio EK, Moseley GL, Dao M, Mosby H, et al. The effects of pain science education plus exercise on pain and function in chronic Achilles tendin- 112. Rompe JD, Furia J, Maffulli N. Eccentric loading versus eccentric loading plus opathy: A blinded, placebo-controlled, explanatory randomized trial. Pain. 2022: shock-wave treatment for midportion achilles tendinopathy: a randomized 10.1097. controlled trial. Am J Sports Med. 2009;37:463\u2013470. 86. Littlewood C, Malliaras P, Bateman M, Stace R, May S, Walters S. The central ner- 113. Vahdatpour B, Forouzan H, Momeni F, Ahmadi M, Taheri P. Effectiveness of vous system\u2013an additional consideration in \u2018rotator cuff tendinopathy\u2019and a po- extracorporeal shockwave therapy for chronic Achilles tendinopathy: a tential basis for understanding response to loaded therapeutic exercise. Man Ther. randomized clinical trial. J Res Med Sci. 2018;23. 2013;18:468\u2013472. 114. Knobloch K, Schreibmueller L, Longo UG, Vogt PM. Eccentric exercises for the 87. Bassett SF. Bridging the intention-behaviour gap with behaviour change stra- management of tendinopathy of the main body of the Achilles tendon with or tegies for physiotherapy rehabilitation non-adherence. N Z J Physiother. without the AirHeelTM Brace. A randomized controlled trial. A: effects on pain and 2015;43:105\u2013111. microcirculation. Disabil Rehabil. 2008;30:1685\u20131691. 88. Rabusin CL, Menz HB, McClelland JA, Evans AM, Malliaras P, Docking SI, et al. 115. Koszalinski A, Flynn T, Hellman M, Cleland J. Trigger point dry needling, manual Ef\ufb01cacy of heel lifts versus calf muscle eccentric exercise for mid-portion therapy and exercise versus manual therapy and exercise for the management of Achilles tendinopathy (HEALTHY): a randomised trial. Brit J Sports Med. Achilles tendinopathy: a feasibility study. J Man Manip Ther. 2020;28:212\u2013221. 2021;55:486\u2013492. 116. Herrington L, McCulloch R. The role of eccentric training in the management of 89. Stergioulas A, Stergioula M, Aarskog R, Lopes-Martins RA, Bjordal JM. Effects of Achilles tendinopathy: a pilot study. Phys Ther Sport. 2007;8:191\u2013196. low-level laser therapy and eccentric exercises in the treatment of recreational athletes with chronic achilles tendinopathy. Am J Sports Med. 2008;36: 117. Kedia M, Williams M, Jain L, Barron M, Bird N, Blackwell B, et al. The effects of 881\u2013887. conventional physical therapy and eccentric strengthening for insertional achilles tendinopathy. Int J Sports Phys Ther. 2014;9:488. 90. Horstmann T, Jud HM, Fr\u00f6hlich V, M\u00fcndermann A, Grau S. Whole-body vibration versus eccentric training or a wait-and-see approach for chronic Achilles ten- 118. Rompe JD, Furia J, Maffulli N. Eccentric loading compared with shock wave dinopathy: a randomized clinical trial. J Orthop Sports Phys Ther. 2013;43:794\u2013 treatment for chronic insertional achilles tendinopathy: a randomized, controlled 803. trial. JBJS. 2008;90:52\u201361. 91. Beyer R, Kongsgaard M, Hougs Kj\u00e6r B, \u00d8hlenschl\u00e6ger T, Kj\u00e6r M, Magnusson SP. 119. Roos EM, Engstr\u00f6m M, Lagerquist A, S\u00f6derberg B. Clinical improvement after 6 Heavy slow resistance versus eccentric training as treatment for Achilles tendin- weeks of eccentric exercise in patients with mid-portion Achilles tendinopathy\u2013a opathy: a randomized controlled trial. Am J Sports Med. 2015;43:1704\u20131711. randomized trial with 1-year follow-up. Scand J Med Sci Sports. 2004;14:286\u2013295. 92. Gatz M, Betsch M, Dirrichs T, Schrading S, Tingart M, Michalik R, et al. Eccentric and 120. Wiedmann M, Mauch F, Huth J, Burkhardt P, Drews BH. Die Behandlung der isometric exercises in Achilles tendinopathy evaluated by the VISA-A score and Midportion-Achillessehnentendinopathie mit exzentrischem Krafttraining und shear wave elastography. Sports Health. 2020;12:373\u2013381. dessen Auswirkung auf die Neovaskularisation. Sports Orthop Traumatol. 2017;33:278\u2013285. 93. Habets B, van Cingel RE, Backx FJ, van Elten HJ, Zuithoff P, Huisstede BM. No dif- ference in clinical effects when comparing alfredson eccentric and silbernagel 121. Benli MD, Tatari H, Balc\u0131 A, Peker A, S\u00b8 ims\u00b8 ek K, Y\u00fcksel O, et al. A comparison combined concentric-eccentric loading in achilles tendinopathy: a randomized between the ef\ufb01cacy of eccentric exercise and extracorporeal shock wave therapy controlled trial. Orthop J Sports Med. 2021;9:23259671211031254. on tendon thickness, vascularity, and elasticity in Achilles tendinopathy: A randomized controlled trial. Turk J Phys Med Rehabil. 2022;68:372\u2013380. 94. Ma\ufb01 N, Lorentzon R, Alfredson H. Superior short-term results with eccentric calf muscle training compared to concentric training in a randomized prospective 122. Chester R, Costa ML, Shepstone L, Cooper A, Donell ST. Eccentric calf muscle multicenter study on patients with chronic Achilles tendinosis. Knee Surg Sports training compared with therapeutic ultrasound for chronic Achilles tendon pain\u2014 Traumatol Arthrosc. 2001;9:42\u201347. a pilot study. Man Ther. 2008;13:484\u2013491. 95. N\u00f8rregaard J, Larsen C, Bieler T, Langberg H. Eccentric exercise in treatment of 123. Solomons L, Lee JJ, Bruce M, White LD, Scott A. Intramuscular stimulation vs sham Achilles tendinopathy. Scand J Med Sci Sports. 2007;17:133\u2013138. needling for the treatment of chronic midportion Achilles tendinopathy: a ran- domized controlled clinical trial. PloS one. 2020;15:e0238579. 96. Stasinopoulos D, Manias P. Comparing two eccentric exercise programmes for the management of Achilles tendinopathy. A pilot trial. J Bodyw Mov Ther. 124. Stefansson SH, Brandsson S, Langberg H, Arnason A. Using pressure massage for 2013;17:309\u2013315. Achilles tendinopathy: a single-blind, randomized controlled trial comparing a novel treatment versus an eccentric exercise protocol. Orthop J Sports Med. 2019;7: 97. Stevens M, Tan C-W. Effectiveness of the Alfredson protocol compared with a lower 2325967119834284. repetition-volume protocol for midportion Achilles tendinopathy: a randomized controlled trial. J Orthop Sports Phys Ther. 2014;44:59\u201367. 125. Zhang B-m, Zhong L-w, Xu S-w, Jiang H-r, Shen J. Acupuncture for chronic Achilles tendnopathy: a randomized controlled study. Chin J Integr Med. 98. Tumilty S, Mani R, Baxter GD. Photobiomodulation and eccentric exercise for 2013;19:900\u2013904. Achilles tendinopathy: a randomized controlled trial. Lasers Med Sci. 2016;31:127\u2013 135. 126. Van Der Vlist AC, Winters M, Weir A, Weir A, Ardern CL, Welton NJ, et al. Which treatment is most effective for patients with Achilles tendinopathy? A living 99. Yu J, Park D, Lee G. Effect of eccentric strengthening on pain, muscle strength, systematic review with network meta-analysis of 29 randomised controlled trials. endurance, and functional \ufb01tness factors in male patients with achilles tendin- Brit J Sports Med. 2021;55:249\u2013256. opathy. Am J Phys Med Rehabil. 2013;92:68\u201376. 127. Merry K, Napier C, Waugh CM, Scott A. Foundational Principles and Adaptation of 100. de Jonge S, de Vos R-J, Van Schie HT, Verhaar JA, Weir A, Tol JL. One-year follow- the Healthy and Pathological Achilles Tendon in Response to Resistance Exercise: up of a randomised controlled trial on added splinting to eccentric exercises A Narrative Review and Clinical Implications. J Clin Med. 2022;11:4722. in chronic midportion Achilles tendinopathy. Brit J Sports Med. 2010;44:673\u2013 677. 128. \u00d6hberg L, Alfredson H. Effects on neovascularisation behind the good results with eccentric training in chronic mid-portion Achilles tendinosis? Knee Surg Sports 101. De Vos R, Weir A, Visser R, de Winter T, Tol J. The additional value of a night splint Traumatol Arthrosc. 2004;12:465\u2013470. to eccentric exercises in chronic midportion Achilles tendinopathy: a randomised controlled trial. Brit J Sports Med. 2007;41:e5-e5. 129. Langberg H, Ellingsgaard H, Madsen T, Jansson J, Magnusson SP, Aagaard P, et al. Eccentric rehabilitation exercise increases peritendinous type I collagen 102. McAleenan M, McVeigh J, Cullen M, Sayers F, McCrea K, Baxter D. The effective- synthesis in humans with Achilles tendinosis. Scand J Med Sci Sports. ness of night splints in achilles tendinopathy: a pilot study. Physiother Prac Res. 2007;17:61\u201366. 2010;31:28\u201333. 130. Smith BE, Hendrick P, Bateman M, Holden S, Littlewood C, Smith TO, et al. 103. McCormack JR, Underwood FB, Slaven EJ, Cappaert TA. Eccentric exercise versus Musculoskeletal pain and exercise\u2014challenging existing paradigms and intro- eccentric exercise and soft tissue treatment (Astym) in the management of ducing new. Brit J Sports Med. 2019;53:907\u2013912. insertional Achilles tendinopathy: a randomized controlled trial. Sports Health. 2016;8:230\u2013237. 131. Tol JL, Spiezia F, Maffulli N. Neovascularization in Achilles tendinopathy: have we been chasing a red herring? Arthroscopy. 2012;10:1891\u20131894. 104. Munteanu SE, Scott LA, Bonanno DR, Landorf KB, Pizzari T, Cook JL, et al. Effec- tiveness of customised foot orthoses for Achilles tendinopathy: a randomised controlled trial. Brit J Sports Med. 2015;49:989\u2013994.","Research 237 132. Drew BT, Smith TO, Littlewood C, Sturrock B. Do structural changes (eg, collagen\/ 156. Lagas IF, van der Vlist AC, van Oosterom RF, van Veldhoven PL, Reijman M, matrix) explain the response to therapeutic exercises in tendinopathy: a sys- Verhaar JA, et al. Victorian Institute of Sport Assessment-Achilles (VISA-A) tematic review. Brit J Sports Med. 2014;48:966\u2013972. Questionnaire\u2014Minimal Clinically Important Difference for Active People With Midportion Achilles Tendinopathy: A Prospective Cohort Study. J Orthop Sports 133. Smith BE, Hendrick P, Smith TO, Bateman M, Moffatt F, Rathleff MS, et al. Should Phys Ther. 2021;51:510\u2013516. exercises be painful in the management of chronic musculoskeletal pain? A sys- tematic review and meta-analysis. Brit J Sports Med. 2017;51:1679\u20131687. 157. Mallows A, Littlewood C, Malliaras P. Measuring patient-reported outcomes (PROs\/PROMs) in people with Achilles tendinopathy: how useful is the VISA-A? 134. Coronado RA, Brintz CE, McKernan LC, Master H, Motzny N, Silva FM, et al. Psy- Brit J Sports Med. 2018;52:1221\u20131221. chologically informed physical therapy for musculoskeletal pain: current ap- proaches, implications, and future directions from recent randomized trials. Pain 158. Comins J, Siersma V, Couppe C, Svensson RB, Johansen F, Malmgaard-Clausen NM, Rep. 2020;5. et al. Assessment of content validity and psychometric properties of VISA-A for Achilles tendinopathy. PloS one. 2021;16:e0247152. 135. Clifford C, Challoumas D, Paul L, Syme G, Millar NL. Effectiveness of isometric exercise in the management of tendinopathy: a systematic review and meta- 159. Abate M, Salini V. Mid-portion Achilles tendinopathy in runners with metabolic analysis of randomised trials. BMJ Open Sport Ex Med. 2020;6:e000760. disorders. Eur J Orthop Surg Traumatol. 2019;29:697\u2013703. 136. Yeh CH, Calder JD, Ant\ufb02ick J, Bull AM, Kedgley AE. Maximum dorsi\ufb02exion in- 160. Ahn HS, Kim HJ, Kang TU, Kazmi SZ, Suh JS, Young Choi J. Dyslipidemia Is asso- creases Achilles tendon force during exercise for midportion Achilles tendinop- ciated with increased risk of Achilles tendon disorders in underweight individuals athy. Scand J Med Sci Sports. 2021;31:1674\u20131682. to a greater extent than obese individuals: A nationwide, population-based, longitudinal cohort study. Orthop J Sports Mede. 2021;9:23259671211042599. 137. Silbernagel KG, Thome\u00e9 R, Eriksson BI, Karlsson J. Full symptomatic recovery does not ensure full recovery of muscle-tendon function in patients with Achilles 161. Gaida J, Alfredson L, Kiss Z, Wilson A, Alfredson H, Cook J. Dyslipidemia in Achilles tendinopathy. Brit J Sports Med. 2007;41:276\u2013280. tendinopathy is characteristic of insulin resistance. Med Sci Sports Ex. 2009; 41:1194. 138. Mahieu NN, Witvrouw E, Stevens V, Van Tiggelen D, Roget P. Intrinsic risk factors for the development of achilles tendon overuse injury: a prospective study. Am J 162. Holmes GB, Lin J. Etiologic factors associated with symptomatic achilles tendin- Sports Med. 2006;34:226\u2013235. opathy. Foot Ank Int. 2006;27:952\u2013959. 139. Bird SP, Tarpenning KM, Marino FE. Designing resistance training programmes to 163. Watt FE. Musculoskeletal pain and menopause. Post Reproduct Health. enhance muscular \ufb01tness. Sports Med. 2005;35:841\u2013851. 2018;24:34\u201343. 140. Enoka RM. Eccentric contractions require unique activation strategies by the 164. Oliva F, Piccirilli E, Berardi AC, Frizziero A, Tarantino U, Maffulli N. Hormones and nervous system. J Appl Physiol. 1996;81:2339\u20132346. tendinopathies: the current evidence. Brit Med Bull. 2016;117:39\u201358. 141. Agyi TH, Devita P, Money J, Barrier J. Effects of standard and eccentric overload 165. Owens BD, Wolf JM, Seelig AD, Jacobson IG, Boyko EJ, Smith B, et al. Risk factors strength training in young women. Med Sci Sports Exerc. 2001;33:1206\u20131212. for lower extremity tendinopathies in military personnel. Orthop J Sports Med. 2013;1:2325967113492707. 142. Naclerio F, Rodr\u00edguez-Romo G, Barriopedro-Moro MI, Jim\u00e9nez A, Alvar BA, Triplett NT. Control of resistance training intensity by the OMNI perceived exer- 166. Friedman DJ, Tulloh L, Khan KM. Peeling off musculoskeletal labels: sticks and tion scale. J Strength Condit Res. 2011;25:1879\u20131888. stones may break my bones, but diagnostic labels can hamstring me forever. Brit J Sports Med. 2021;55:1184\u20131185. 143. Morishita S, Tsubaki A, Takabayashi T, Fu JB. Relationship between the rating of perceived exertion scale and the load intensity of resistance training. Strength 167. Kraemer R, Wuerfel W, Lorenzen J, Busche M, Vogt PM, Knobloch K. Analysis of Condit J. 2018;40:94. hereditary and medical risk factors in Achilles tendinopathy and Achilles tendon ruptures: a matched pair analysis. Arch Orthop Trauma Surg. 2012; 144. Sancho I, Morrissey D, Willy RW, Barton C, Malliaras P. Education and exercise 132:847\u2013853. supplemented by a pain-guided hopping intervention for male recreational runners with midportion Achilles tendinopathy: a single cohort feasibility study. 168. Wheeler PC. Nearly half of patients with chronic tendinopathy may have a Phys Ther Sport. 2019;40:107\u2013116. neuropathic pain component, with signi\ufb01cant differences seen between different tendon sites: a prospective cohort of more than 300 patients. BMJ Open Sport Ex 145. Oranchuk DJ, Storey AG, Nelson AR, Cronin JB. Isometric training and long-term Med. 2022;8:e001297. adaptations: Effects of muscle length, intensity, and intent: A systematic review. Scand J Med Sci Sports. 2019;29:484\u2013503. 169. Wang HK, Lin KH, Su SC, Shih TF, Huang YC. Effects of tendon viscoelasticity in A chilles tendinosis on explosive performance and clinical severity in athletes. Scand 146. Lieber RL. Skeletal muscle structure, function, and plasticity. Lippincott Williams & J Med Sci Sports. 2012;22:e147\u2013e155. Wilkins; 2002. 170. Firth BL, Dingley P, Davies ER, Lewis JS, Alexander CM. The effect of kinesiotape on 147. Costa M, Shepstone L, Donell S, Thomas T. Shock wave therapy for chronic Achilles function, pain, and motoneuronal excitability in healthy people and people with tendon pain: a randomized placebo-controlled trial. Clin Orthop Relat Achilles tendinopathy. Clin J Sport Med. 2010;20:416\u2013421. Res. 2005;440:199\u2013204. 171. Mc Auliffe S, Bisset L, Chester R, Coombes BK, Fearon A, Kirwan P, et al. ICON 148. Hutchison A, Pallister I, Evans R, Bodger O, Topliss CJ, Williams P, et al. Intense 2020\u2014International Scienti\ufb01c Tendinopathy Symposium Consensus: A Scoping pulsed light treatment of chronic mid-body achilles tendinopathy: a double blind Review of Psychological and Psychosocial Constructs and Outcome Measures randomised placebo-controlled trial. Bone Jy J. 2013;95:504\u2013509. Reported in Tendinopathy Clinical Trials. J Orthop Sports Phys Ther. 2022;52: 375\u2013388. 149. Kishmishian B, Richards J, Selfe J. A randomised feasibility study using an acupuncture protocol to the Achilles tendon in Achilles tendinopathy. Physiother 172. Alghamdi NH, Pohlig RT, Lundberg M, Silbernagel KG. The impact of the degree of Prac Res. 2019;40:59\u201367. kinesiophobia on recovery in patients with achilles tendinopathy. Phys Ther. 2021;101:178. 150. Fahlstr\u00f6m M, Jonsson P, Lorentzon R, Alfredson H. Chronic Achilles tendon pain treated with eccentric calf-muscle training. Knee Surg Sports Traumatol Arthrosc. 173. Chester R, Jerosch-Herold C, Lewis J, Shepstone L. Psychological factors are asso- 2003;11:327\u2013333. ciated with the outcome of physiotherapy for people with shoulder pain: a multicentre longitudinal cohort study. Brit J Sports Med. 2018;52:269\u2013275. 151. Jonsson P, Alfredson H, Sunding K, Fahlstr\u00f6m M, Cook J. New regimen for eccentric calf-muscle training in patients with chronic insertional Achilles tendinopathy: 174. Niesen-Vertommen S. The effect of an eccentric-type exercise versus a results of a pilot study. Brit J Sports Med. 2008;42:746\u2013749. concentric-type exercise in the management of chronic Achilles tendonitis. University of British Columbia; 1989. 152. Testa M, Rossettini G. Enhance placebo, avoid nocebo: How contextual factors affect physiotherapy outcomes. Man Ther. 2016;24:65\u201374. 175. Stanish WD, Rubinovich RM, Curwin S. Eccentric exercise in chronic tendinitis. Clin Orthop Relat Res. 1986;208:65\u201368. 153. Gr\u00e4vare Silbernagel K, Malliaras P, de Vos R-J, Hanlon S, Molenaar M, Alfredson H, et al. ICON 2020\u2014International Scienti\ufb01c Tendinopathy Symposium Consensus: a 176. Alfredson H, Pietil\u00e4 T, Jonsson P, Lorentzon R. Heavy-load eccentric calf muscle systematic review of outcome measures reported in clinical trials of Achilles training for the treatment of chronic Achilles tendinosis. Am J Sports Med. 1998; tendinopathy. Sports Med. 2021:1\u201329. 26:360\u2013366. 154. Yin N-H, Fromme P, McCarthy I, Birch HL. Individual variation in Achilles tendon 177. Gifford L. The mature organism model. Whiplash\u2014Science and Management: Fear, morphology and geometry changes susceptibility to injury. Elife. 2021;10: Avoidance Beliefs and Behaviour Topical Issues in Pain. 2013;1:45\u201356. e63204. Websites 155. Chimenti RL, Post AA, Silbernagel KG, Hadlandsmyth K, Sluka KA, Moseley GL, et al. Kinesiophobia Severity Categories and Clinically Meaningful Symptom Tame the Beast www.tamethebeast.org Change in Persons With Achilles Tendinopathy in a Cross-Sectional Study: Why Things Hurt www.youtube.com\/watch?v=gwd-wLdIHs Implications for Assessment and Willingness to Exercise. Front Pain Res. 2021:57.","Journal of Physiotherapy 68 (2022) 215\u2013217 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 Reasons to become a volunteer rater for the Physiotherapy Evidence Database (PEDro) Junior Vitorino Fandim a, Laura Crowe-Owen b, Mykola Romanyshyn c, Stephen WW Chan d aMasters and Doctoral Programs in Physical Therapy, Universidade Cidade de S\u00e3o Paulo (UNICID), Sao Paulo, Brazil; bTherapy for Life, Mount Gambier, Australia; cDepartment of Rehabilitation of Patients with Consequences of Diseases and Injuries of the Nervous System, Communal non-commercial enterprise \\\"Kyiv Regional Clinical Hospital\\\", Kyiv, Ukraine; dAllied Health Department, Hong Kong Children\u2019s Hospital, Hong Kong, China Evidence-based physiotherapy involves integrating the clinician\u2019s PEDro raters are primarily physiotherapy clinicians, educators and practice knowledge and the patient\u2019s preferences with high-quality students who have undertaken training and demonstrated pro\ufb01- clinical research about the effects of clinical interventions.1 High- ciency in using the PEDro scale. A key task performed by these vol- quality clinical research minimises biases by using rigorous study de- unteers is to help rate the indexed trials, to assist users to quickly signs such as randomised trials. Although most physiotherapists appraise each trial\u2019s validity and the completeness of reporting of its recognise that evidence-informed clinical decision-making is essential, results. Over the past decade, the number of active volunteer raters there are many barriers to accessing high-quality clinical research.2\u20134 has increased four-fold. PEDro had 25 volunteer raters in 2013 and The Physiotherapy Evidence Database (PEDro) is a database that can this increased to 99 in 2022 (Figure 2). help physiotherapists access the least biased evidence to inform their clinical decision-making.3,5 This editorial describes PEDro, the role of What are the bene\ufb01ts of being a rater for PEDro? PEDro raters and the process of becoming a volunteer PEDro rater. Numerous reasons can inspire someone to be a PEDro volunteer Physiotherapy Evidence Database rater. Being a PEDro rater allows the individual to contribute to the worldwide research community and a multinational educational Launched in 1999, PEDro is now an eminent global resource, initiative from any location and at any time. There are some advan- providing rapid access to high-quality clinical research across all tages of being a PEDro rater that apply to all volunteers and other subdisciplines of physiotherapy. PEDro is considered the most advantages that are applicable to students, bilingual volunteers, and comprehensive resource to search for research into the effects of those in research and clinical roles (Box 1). physiotherapy interventions.5\u20138 It indexes randomised trials of physiotherapy interventions, systematic reviews of such trials, and For students, being a trained rater can enhance academic reading evidence-based clinical practice guidelines. These types of research skills and help develop excellent critical thinking. This process may are identi\ufb01ed using sensitive search strategies without restriction by speed up the student\u2019s development as a healthcare professional or the language or journal of publication (https:\/\/pedro.org.au\/english\/ future researcher. It can also supply the student with more research learn\/faq\/ question 1). If they meet the criteria for inclusion, they ideas and equip them with appraisal skills. are then indexed in PEDro (https:\/\/pedro.org.au\/english\/learn\/ indexing-criteria-and-codes\/). A unique feature of PEDro is that all For clinicians, being a rater can provide a critical lens for under- indexed trials are rated for methodological quality by at least two standing the complexity and generalisability of the research and how PEDro-trained raters. Raters use the PEDro scale to assess the quality it applies to a patient\u2019s health. Regular rating may transform a barrier of the article out of a score of 10, so that PEDro users can quickly into a facilitator by increasing familiarity with research jargon and identify the trials that are likely to be valid and provide suf\ufb01cient data improving reading skills, which may be particularly important for to guide clinical decision-making. PEDro currently provides links to volunteers from non-English-speaking countries. Regular rating may free full-text access for about 60% of the research it indexes.3 also help clinicians to solidify their own clinical questions and update their knowledge with current evidence. It may also improve their The PEDro database has shown exponential growth in the number skills in critical appraisal, knowledge of research methodologies, and of research articles indexed over time (Figure 1). Thousands of articles ability to bring the best evidence to clinical practice. As such this are indexed for each subdiscipline of physiotherapy, with the highest increases the empowerment and autonomy of the clinician. proportions in the musculoskeletal, cardiothoracic and neurological subdisciplines (https:\/\/pedro.org.au\/english\/learn\/pedro-statistics\/). For researchers, regular use of the PEDro scale can develop In August 2022, PEDro had 55,812 indexed articles, comprising 42,673 enhanced appraisal skills and improve knowledge of research randomised trials, 12,418 systematic reviews and 721 clinical practice methods, hopefully leading them to avoid common mistakes. For some guidelines.9 researchers, it may help to avoid common errors in assessing the methodological quality of trials included in systematic reviews. PEDro What do PEDro raters do? rater training can help researchers to optimise the design and conduct of their research and to communicate their research transparently and During the past two decades, over 41,000 randomised trials completely. By keeping abreast of recently published evidence, re- indexed in PEDro have been quality rated.9 This has been achieved by searchers can better focus on generating evidence that is centred on a collaboration of volunteer PEDro raters from around the world. patient needs and priorities, thereby reducing research waste. In some countries, physiotherapists\u2019 access to high-quality research in physiotherapy may be limited by the fact that most https:\/\/doi.org\/10.1016\/j.jphys.2022.09.006 1836-9553\/\u00a9 2022 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\/).","216 Editorial Figure 1. The cumulative number of trials, reviews and guidelines indexed on the Box 1. Benefits obtained from being a volunteer PEDro rater. Physiotherapy Evidence Database (PEDro). Benefits relevant to all raters 125 \u000f Keep up to date with new clinically relevant research in Number of PEDro raters 100 your area(s) of interest 75 \u000f Enhance your ability to quickly appraise research into physiotherapy interventions 50 \u000f Receive training in how to apply the PEDro scale 25 \u000f Receive confirmation of your competence with the PEDro 0 2014 2016 2018 2020 2022 scale 2012 Year \u000f Develop a greater understanding of research methods \u000f Reinforce that physiotherapy is an evidence-based Figure 2. The number of active Physiotherapy Evidence Database (PEDro) volunteer raters each year. profession \u000f Guide physiotherapists, other healthcare professionals and research is published in English. This may also mean that they must use evidence that has been generated in other countries, which may other users towards better evidence not always be appropriate to their local culture or healthcare system. \u000f Contribute to a non-profit initiative that supports better Bilingual volunteer PEDro raters have the chance to assess the methodological quality of non-English-language trials and thereby patient management globally facilitate decision-making in countries with languages other than Additional benefits for clinicians English. Additionally, as the profession tackles extensive research \u000f Learn about new research to improve your own waste, predatory journals and massive pre-print research, it is vitally important that we appraise all published evidence and not merely management of patients that published in English. \u000f Maintain a database of evidence to help inform your clinical Together to make a difference practice \u000f Prove your time spent on professional development or One of the great motivations to be a volunteer PEDro rater is to contribute to the strengthening of evidence-based physiotherapy continuing education around the world, which is a necessary goal for the profession. For \u000f Increase your autonomy to identify effective interventions example, only 42% of physiotherapists reported having a clear under- \u000f Demonstrate the efficacy of your clinical interventions to standing of the use of research \ufb01ndings in clinical practice.10 Also, the evidence-based practice training provided at the undergraduate level funders may be insuf\ufb01cient.10 Some healthcare systems and healthcare organi- Additional benefits for researchers sation websites may make recommendations that are not aligned with \u000f Keep abreast of new research relevant to your own the available evidence.11 Some healthcare professionals are unaware of the bene\ufb01ts of evidence-based practice and have a misconception about research how evidence-based practice works. One way to overcome this problem \u000f Help maintain a comprehensive database of trials for is to support freely accessible educational initiatives dedicated to strengthening evidence-based practice, such as PEDro. systematic reviews \u000f Identify when new trials make a topic suitable for a How to become a PEDro volunteer rater systematic review Online rating training program \u000f Ensure you can use the PEDro scale correctly in your own To become a PEDro rater, you must demonstrate pro\ufb01ciency with systematic reviews the PEDro scale. PEDro now offers an online training program for \u000f Learn how to achieve each item on the PEDro scale with your own trials \u000f Learn how to report your trials\u2019 methodology unambiguously Additional benefits for students \u000f Increase your knowledge of new research in the profession \u000f Enhance your critical thinking skills and your understanding of research jargon \u000f Generate your own ideas and independence to respond to clinical questions \u000f Practise some of the steps of evidence-based physio- therapy for use on clinical placements \u000f Know which interventions are not supported by evidence \u000f Gain a comprehensive view of the breadth of clinical physiotherapy practice Additional benefits for bilingual raters \u000f Promote awareness of physiotherapy research published in languages other than English \u000f Ensure PEDro recognises the breadth of international clinical physiotherapy practice \u000f Support a non-profit organisation to strengthen evidence used around the world physiotherapists, other healthcare professionals and students who want to improve their appraisal skills. The online PEDro scale training program may be completed at your own pace no matter where you live. The PEDro scale summarises the methodological quality and completeness of statistical reporting in published clinical trials.12 The PEDro scale comprises 11 items, 10 of which contribute to the total PEDro score (range 0 to 10 points), which is used to rank search re- sults when users search the database.13 The online PEDro scale training program is available in English and Portuguese. Each sub- scription is valid for 3 months for a single user. The program provides step-by-step training with instructional videos, real-world examples, appraisal practice and feedback. Subscribers who pass the \ufb01nal","Editorial 217 accuracy exam receive a certi\ufb01cate. More details can be accessed at Ethics approval: Not applicable. the PEDro training website (https:\/\/training.pedro.org.au\/). Competing interests: Nil. Source(s) of support: Nil. Volunteering as a PEDro rater Acknowledgements: Nil. Provenance: Invited. Peer reviewed. You will be invited to be a PEDro rater after demonstrating pro- Correspondence: Junior Vitorino Fandim, Masters and Doctoral \ufb01ciency in using the PEDro scale by passing the PEDro training pro- Programs in Physical Therapy, Universidade Cidade de S\u00e3o Paulo gram accuracy test. Volunteer PEDro raters can select their preferred (UNICID), Brazil. Email: [email protected] subdiscipline or area of physiotherapy practice, rate trials in other languages in which they may have pro\ufb01ciency and nominate how References many trials to rate each month. All rating is performed using a password-protected website and can be done from anywhere in the 1. Herbert R, et al. Practical evidence-based physiotherapy. Edinburgh: Churchill Liv- world. The assessment of one trial using the PEDro scale usually takes ingstone; 2011. 5 to 20 minutes, depending on the complexity of the article and the rater\u2019s level of experience. 2. da Silva TM, et al. Manual Therapy. 2015;20:388\u2013401. 3. Moseley AM, et al. Braz J Phys Ther. 2022;26:100392. In conclusion, PEDro helps physiotherapists and other users 4. Paci M, et al. Int J Qual Health Care. 2021;33:mzab093. to ef\ufb01ciently identify high-quality clinical research about the 5. Michaleff ZA, et al. Phys Ther. 2011;91:190\u2013197. effects of physiotherapy interventions. This research is accumu- 6. Aromataris E, Munn Z. JBI Manual for Evidence Synthesis. Joanna Briggs Institute; lating rapidly so volunteer raters are needed if PEDro is to be maintained. PEDro provides an online training program for inter- 2020. Available from https:\/\/jbi.global\/ebp#jbi-manuals. ested volunteers and others to learn to use the rating scale. Each 7. Furlan AD, et al. Spine. 2015;40(21):1660\u20131673. PEDro volunteer rater has individual characteristics and expertise, 8. Moseley AM, et al. Braz J Phys Ther. 2020;24:384\u2013391. but all volunteers are proud to support PEDro to keep strength- 9. PEDro Partnership. Physiotherapy Evidence Database (PEDro) 2022. https:\/\/pedro. ening evidence-based physiotherapy practice around the world. Rating is made as \ufb02exible as possible for volunteers. Working org.au\/english\/pedro-update-1-august-2022\/. Accessed August 29, 2022. together, we all make a difference by sustaining this valuable 10. Silva TM, et al. Braz J Phys Ther. 2015;19:294\u2013303. resource. 11. Santos RP, et al. Braz J Phys Ther. 2022;26:100389. 12. Yamato TP, et al. J Clin Epidemiol. 2017;86:176\u2013181. 13. PEDro Partnership. Physiotherapy Evidence Database (PEDro) 2022. https:\/\/pedro. org.au\/english\/resources\/pedro-scale. Accessed August 24, 2022. Websites PEDro website https:\/\/pedro.org.au PEDro scale training program https:\/\/training.pedro.org.au","Journal of Physiotherapy 68 (2022) 283\u2013287 Appraisal 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 Note: Treatment effect moderators Randomised controlled trials (RCTs) provide evidence on the respond best to manipulation),6 but many are not (eg, Ottawa ankle average effect (difference in outcomes) of being allocated to the rule).7 Equally, a treatment effect moderator does not need to be a intervention group, compared with control group, across the range of clinical prediction rule, as it can also be a single patient characteristic participants included in the study. This information is critical to (eg, age). While a single characteristic like age or biological sex can be evidence-based medicine and informing clinical practice guidelines. an effect moderator, some have argued that effect moderators based However, the validity of applying the results from an RCT to an in- on a number of patient factors in combination are more likely to be dividual patient in clinical practice is dependent on how closely the clinically useful.8 participants in the trial match the patient in front of a clinician, and therefore whether the potential effects in the patient are likely to be Treatment effects happen to individuals, but they cannot be similar to the average effect for participants in the trial.1 It is also observed in individuals dependent on an assumption that the effect of the intervention in the trial is relatively consistent across the participants in a trial. This may Critical to understanding treatment effect moderators is to \ufb01rst often be a reasonable assumption, but in other cases there may be a understand treatment effects. Treatment effects are the difference in strong theoretical rationale to believe that the effect of an interven- outcomes between receiving an intervention and not receiving an tion substantially varies across participants in a trial, sometimes intervention. However, it is not possible to measure treatment effects referred to as heterogeneity of treatment effects.2 For example, it in individuals,9 for example: an individual cannot receive manipula- could be that an intervention that has a moderate size average effect tion for their spinal pain and at the same time not receive this across all participants may in fact have large effects in some partici- treatment. Therefore, to truly know the effect of an intervention for pants (eg, in those with a more severe condition) and may have small an individual is not possible. Once an individual has received an effects in some other participants (eg, those with a less severe con- intervention, we don\u2019t know what their outcome would have been if dition). In a more extreme example, a certain intervention could have they did not receive the intervention. RCTs attempt to overcome this no effect on average but could be moderately effective for some problem by using other people as the control (equivalent to the in- participants and moderately harmful for other participants. If this dividual not getting the intervention). Therefore, it is critical that was the case, it would clearly be important to understand the char- people not receiving the intervention are as similar as possible to acteristics of the individuals who either bene\ufb01tted or were harmed those who did receive the intervention. Randomisation is the best and this information would inform treatment decision-making. method to balance both known and unknown patient factors that might impact outcomes, and we become more con\ufb01dent that the Characteristics that identify patients who respond differently to participants in the intervention group in large RCTs are on average certain interventions are called treatment effect moderators. This very similar to participants in the control group. Therefore, the effect Research Note discusses principles of well-designed studies to found in a large well-conducted trial validly represents the average investigate treatment effect moderators and how to interpret the effect of the intervention across the participants in the trial. However, results of those studies. It is beyond the scope of this paper to cover standard analysis of an RCT does not provide any evidence for or the progress or lack of progress that has been made in identifying against the possibility that the effect of the intervention compared treatment effect modi\ufb01ers for conditions commonly seen by physio- with control is similar or different across the participants in the study. therapists; however, progress has generally been limited, as described Figure 1 provides a visual explanation of this challenge and demon- in previous papers.3,4 strates that variation in outcomes does not necessarily imply varia- tion in treatment effects. The language use varies and can be confusing Several different terms are used in this \ufb01eld of research, some- Study designs and analyses required to investigate treatment times correctly and sometimes incorrectly, which can contribute to effect moderators misunderstanding or confusion. Some of these terms include mod- erators\/modi\ufb01ers, treatment effect moderators\/modi\ufb01ers, subgroups, Study design treatment subgroups, interactions and clinical prediction rules. Importantly, the terms subgroups and clinical prediction rules are not The investigation of treatment effect moderators always requires a limited to identifying the characteristics of patients who best respond control group comparison, as occurs in an RCT. Without this it is to a speci\ufb01c intervention. These terms can equally be used when impossible to be con\ufb01dent if a patient characteristic being investi- identifying patients with characteristics that are associated with gated is an effect modi\ufb01er or simply associated with outcomes prognosis or diagnosis.5 When the intent is to investigate character- regardless of treatment.10 A simple example of how incorrect con- istics of patients who best respond to a speci\ufb01c intervention, the term clusions can be draw when not using a control group comparison is treatment effect moderator accurately describes the intent, and we provided in Figure 2. In this example the potential effect modi\ufb01er therefore recommend its use. A clinical prediction rule may be (acute pain duration) is clearly associated with different (better) developed to identify a treatment effect moderator (eg, patients who outcomes (panel A); however, when we compare with a control h t t p s : \/ \/ d o i . o r g \/ 10.1016 \/ j . j p hys .2 0 2 2 . 0 8 . 0 0 6 1836-9553\/\u00a9 2022 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\/).","284 Appraisal Research Note Figure 1. In a randomised controlled trial the same outcomes and average effect (red line) could result from homogenous treatment effects or heterogenous treatment effects. group we can see that this could represent a situation where: acute 2. An appropriately conducted test of interaction is performed. pain duration is not an effect modi\ufb01er (panel B); acute pain duration is an effect modi\ufb01er and identi\ufb01es patients who get most bene\ufb01t from When investigating whether a patient characteristic (eg, biological treatment (panel C); or acute pain duration is an effect modi\ufb01er and sex) is an effect modi\ufb01er, it is critical to demonstrate that the effect of identi\ufb01es patients who get less bene\ufb01t from treatment (panel D). the intervention (compared with control) is greater in those with a Clearly, clinical decisions based on which patients most bene\ufb01t from certain characteristic (eg, female, higher baseline pain intensity) than interventions cannot be made from studies without a control group. in those without this characteristic (eg, male, lower baseline pain Despite this, many publications continue to make these inappropriate intensity). It is inadequate to show that the intervention has a sta- conclusions.11,12 The best studies of effect modi\ufb01ers only use data tistically signi\ufb01cant effect in just one subgroup (eg, females).14,16 from RCTs to test moderation, but there is a limited role for the use of There is strong evidence from simulation studies that purely by other research designs with control groups, such as propensity-score chance a statistically signi\ufb01cant effect will commonly occur in one matched cohort data, as a means to generate hypotheses about subgroup but not in the other group, even when no true treatment moderation that can subsequently be tested in RCTs. Propensity moderation exists.14 Therefore, it is strongly recommended not to scores match participants on measured patient factors but, unlike investigate the effect of the intervention in just one subgroup.13,14,16,17 RCTs, cannot balance participants on unmeasured factors (there is a The simple and recommended solution to this problem is to use an greater risk of residual confounding). interaction test.13,16,17 The interaction test evaluates whether the ef- fect of the intervention is greater in those with a certain character- Analysis\/credibility criteria istic\/s than in those without this characteristic\/s. The size of the interaction describes how much bigger the effect of the intervention Even when a well-designed RCT is used to investigate treatment is in one group compared with the other (or how much bigger the effect moderators, there is a high risk of spurious \ufb01ndings.13\u201315 effect is for each unit of the effect moderator if the moderator of Several aspects of the study design and analysis can reduce this risk interest is continuous). So for example, if the effect of intervention and increase the credibility of the \ufb01ndings. A detailed description is compared with control is 3 points in women and 1 point in men, the provided by Sun et al15 but some key aspects are brie\ufb02y covered here. interaction will be 2 points. Figure 3 provides a visual description of what an interaction term is using group level data. For the interested 1. Limit the number of moderator variables investigated and state reader, we have previously demonstrated that the interaction effects hypothesis and direction of effect a priori and their con\ufb01dence intervals calculated using group level data are the same as those calculated using interaction terms in regression Investigating a large number of potential moderators, especially models.17 When analysing individual-level data (eg, using regression when no strong biological rationale exists, raises the chance that one models), interactions directly estimate the size of effect moderation or more variables will appear to be an effect moderator simply by with a con\ufb01dence interval. chance. This risk can be minimised by a priori de\ufb01ning a small number of effect moderators that will be investigated and providing a How to interpret a well-conducted treatment effect moderator hypothesis and direction of effect for each effect moderator. This in- study formation would ideally be recorded in the clinical trial registration or a published trial protocol. If this is not done and the authors intend When interpreting results from a paper investigating treatment to conduct secondary analyses of an RCT to explore effect modi\ufb01ers, effect moderation, a reader needs to determine: whether there is evi- this information should be published in a publicly available location dence suggesting that the investigated characteristic is a treatment such as Open Science Framework prior to initiating analyses. We effect moderator, and whether the treatment moderation effect is large recommend a priori reporting the moderator\/s, measurement tool enough to be clinically important (likely to change decisions about who used to evaluate the moderator, threshold for dichotomising moder- would receive a speci\ufb01c intervention).17 This is analogous to assessing ator (if appropriate), direction of hypothesised moderation effect, and whether a trial suggests that an intervention is more effective than the rationale for hypothesis. control, and then deciding whether the size of the effect is large enough to make the intervention clinically worthwhile. When \ufb01rst assessing","Appraisal Research Note 285 Figure 2. Comparing prognosis with effect modi\ufb01cation, demonstrating that without a control group it is impossible to determine whether a prognostic factor is also a treatment effect modi\ufb01er. Dashed lines represent the group receiving the intervention.","286 Appraisal Research Note Figure 3. Interaction term demonstrating that the interaction term is the difference in the effect of the intervention (compared with control) in people positive on the moderator compared with people negative on the moderator. whether there is evidence that the variable of interest is a treatment treatment effects in the two subgroups and therefore the clinical effect moderator, the focus needs to be on the interaction term and implication varies greatly. con\ufb01dence intervals, as described above. This is relatively straightfor- ward, except that most studies of treatment effect moderators are To help a reader interpret the clinical importance of a moderator, underpowered, as they typically require large sample sizes, so the manuscripts should report the effects within each subgroup, if there chance of Type 2 error is elevated. Readers should carefully assess the is evidence (based on the interaction CIs) suggesting that the inves- CIs for the interaction when deciding on the likelihood of treatment tigated characteristic is a treatment effect moderator. As mentioned effect moderation, and think of this on a continuum rather than as an above, a statistically signi\ufb01cant effect can be found in one subgroup absolute yes or no. This is consistent with recent broad recommenda- (eg, females) but not in the other subgroup (eg, males), even when no tions to limit the reliance on p-values.18,19 In addition to the con\ufb01dence true treatment moderation exists,14 so the effects within each sub- interval, other characteristics such as the presence of a strong biological group should only be investigated after assessing the interaction and rationale, limited number of moderators, and direction of effect CIs, and then determining that treatment effect moderation is likely consistent with hypothesis can all contribute to increased con\ufb01dence to be occurring. The effects within each subgroup are probably more that true effect moderation is present.15 useful than the interaction itself when considering if the moderator effect is clinically important and should inform who does or does not The clinical importance of the moderator should be assessed only receive the intervention (Figure 4).17 if the interaction and CIs suggest that treatment moderation is likely to be occurring. The size of the interaction term alone is not enough to Concluding remarks make an informed judgement on the clinical importance of a treatment effect moderator.17 As an example, an interaction term of 3 points Studies of treatment effect moderation have the potential to could indicate that the intervention produces a 3 point effect in those improve patient outcomes by identifying patients who get the most with the baseline characteristic and no effect in the control group, but bene\ufb01t from speci\ufb01c interventions. However, these studies are also at it could also mean that the intervention is effective in both groups but high risk of spurious \ufb01ndings and inappropriate interpretation. This more effective in the subgroup (eg, 5 points versus 2 points) or that Research Note has summarised key elements of study design, analysis the intervention has no effect in the subgroup but is harmful in those and interpretation to assist researchers and clinicians. A number of not in the subgroup (0 versus \u20133 points). Figure 4 demonstrates how important issues related to treatment effect moderators are beyond the same size for the interaction term can result from very different the scope of the manuscript but should be considered when","Appraisal Research Note 287 Figure 4. The same interaction effect (2 points) can result from very different effects (intervention compared with control) in patients who are positive or negative for a dichotomous moderator. Main effect is the average effect across everyone in the trial. interpreting this literature. First, effect moderators are not necessarily References causally related to the difference in effects they help to identify.20 Despite this, they can still be used to predict or identify patients 1. Lee H, et al. J Physiother. 2021;67:315\u2013318. who are likely to respond better or worse to a speci\ufb01c intervention. 2. Varadhan R, et al. J Clin Epidemiol. 2013;66:818\u2013825. Second, interaction effects can be reported on either an additive or 3. Saragiotto BT, et al. J Orthop Sport Phys Ther. 2017;47:44\u201348. multiplicative scale, and this can alter the results and interpreta- 4. Stanton TR, et al. Phys Ther. 2010;90:843\u2013854. tion.21 Third, no single study is likely to provide strong evidence of 5. Moons KG, et al. Ann Int Med. 2015;162:W1\u2013W73. moderation. It will usually be necessary to replicate the observation 6. Childs JD, et al. Ann Int Med. 2004;141:920\u2013928. of moderating effects in other well-conducted studies to provide 7. Bachmann LM, et al. BMJ. 2003;326:417. strong evidence of effect moderation. 8. Kent DM, et al. Ann Int Med. 2020;172:W1. 9. Herbert RD. J Physiother. 2020;66:273\u2013277. Provenance: Invited. Peer reviewed. 10. Hancock M, et al. Phys Ther. 2009;89:698\u2013704. 11. Hides JA, et al. Eur Spine J. 2019;28:2462\u20132469. Mark J Hancocka and Peter Kentb 12. Iverson CA, et al. J Orthop Sport Phys Ther. 2008;38:297\u2013309. aFaculty of Medicine, Health and Human Sciences, Macquarie University, 13. Lagakos SW. N Eng J Med. 2006;354:1667. 14. Brookes ST, et al. Health Technol Assess. 2001;5:1\u201356. Australia 15. Sun X, et al. BMJ. 2010;340:850\u2013854. bCurtin School of Allied Health, Curtin University, Australia 16. Klebanoff MA. Am J Obstet Gynecol. 2007;197:119\u2013122. 17. Hancock MJ, et al. Phys Ther. 2013;93:852\u2013859. 18. Herbert R. J Physiother. 2019;65:178\u2013181. 19. Wasserstein RL, et al. Am Stat. 2016;70:129\u2013133. 20. VanderWeele T. Explanation in causal inference: methods for mediation and inter- action. Oxford University Press; 2015. 21. VanderWeele TJ, et al. Epidemiol Methods. 2014;3:33\u201372.","Journal of Physiotherapy 68 (2022) 244\u2013254 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 conservative interventions are more effective than others for people with chronic non-speci\ufb01c neck pain: a systematic review and network meta-analysis Greta Castellini a,1, Paolo Pillastrini b,1, Carla Vanti b, Silvia Bargeri a, Silvia Giagio b, Elena Bordignon c, Francesco Fasciani b, Francesco Marzioni b, Tiziano Innocenti d,e, Alessandro Chiarotto d,f, Silvia Gianola a,2, Lucia Bertozzi g,h,2 aIRCCS Istituto Ortopedico Galeazzi, Unit of Clinical Epidemiology, Milan, Italy; bDepartment of Biomedical and Neuromotor Sciences (DIBINEM) \u2013 University of Bologna, Italy; cS.S.M.F.R. Ospedale San Bassiano, Bassano del Grappa, Italy; dDepartment of Health Science \u2013 Faculty of Science \u2013 Vrije Universiteit Amsterdam, Amsterdam Movement Sciences, The Netherlands; eGIMBE Foundation, Bologna, Italy; fDepartment of General Practice \u2013 Erasmus MC - University Medical Center Rotterdam, The Netherlands; gDivision of Occupational Medicine, IRCCS Azienda Ospedaliero-Universitaria di Bologna, Bologna, Italy; hDepartment of Biomedical and Neuromotor Sciences (DIBINEM), Alma Mater Studiorum University of Bologna, Bologna, Italy KEYWORDS ABSTRACT Chronic neck pain Question: Which is the most effective conservative intervention for patients with non-speci\ufb01c chronic neck Conservative interventions pain (CNSNP)? Design: A systematic review and network meta-analysis of randomised clinical trials. Pharmacological interventions Participants: Adults with CNSNP of at least 3 months duration. Interventions: All available pharmacolog- Pain ical and non-pharmacological interventions. Outcome measures: The primary outcomes were pain intensity Disability and disability. The secondary outcome was adverse events. Results: Overall, 119 RCTs (12,496 patients; 32 interventions) were included. Risk of bias was low in 50.4% of trials, unclear in 22.7% and high in 26.9%. Compared with inert treatment, a combination of active and\/or passive multimodal non-pharmacological inventions (eg, exercise and manual therapy) were effective for pain on a 0-to-10 scale at 1 month (MD range 0.84 to 3.74) and at 3 to 6 months (MD range 1.06 to 1.49), and effective on disability on a 0-to-100 scale at 1 month (MD range 10.26 to 14.09) and 3 to 6 months (MD range 5.60 to 16.46). These effects ranged from possible to de\ufb01nite clinical relevance. Compared with inert treatment, anti-in\ufb02ammatory drugs alone or in combination with another non-pharmacological treatment did not reduce pain at 1 month or 3 to 6 months. At 12 months, no superiority was found over inert treatment on both outcomes. Most mild adverse events were experienced following acupuncture\/dry needling intervention. On average, the evidence varied from low to very low certainty. Conclusions: While multimodal non-pharmacological interventions may reduce pain and disability for up to 3 to 6 months of follow-up when compared with inert treatment, the evidence was very uncertain about their effects. Better quality and larger trials are needed to improve the certainty of evidence. Registration: PROSPERO CRD42019124501 [Castellini G, Pillastrini P, Vanti C, Bargeri S, Giagio S, Bordignon E, Fasciani F, Marzioni F, Innocenti T, Chiarotto A, Gianola S, Bertozzi L (2022) Some conservative interventions are more effective than others for people with chronic non- speci\ufb01c neck pain: a systematic review and network meta-analysis. Journal of Physiotherapy 68:244\u2013 254] \u00a9 2022 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 Neck pain has a multifactorial aetiology; it might be related and Neck pain is a prevalent cause of pain and disability worldwide.1 It modulated to ergonomic or individual factors such as age, behavioural is a serious public health issue2 causing a heavy burden, with a attitude or psychosocial distress such as anxiety or job satisfaction.4 prevalence of 223 million people (95% uncertainty interval 179 to 281) and 22 million (95% uncertainty interval 15 to 32) years lived Since most episodes of neck pain are of unknown origin, this is usually with disability globally.3 In the latest Global Burden of Disease Study, labelled as non-speci\ufb01c neck pain.5,6 In the United States, together with neck pain ranked 19th as measured by disability-adjusted life years low back pain, it is the leading cause of healthcare spending.7 The burden (DALYs) for ages 25 to 49 years.2 and costs of neck pain are arguably driven by patients with chronic non- 1 Co-\ufb01rst authors. 2 Co-last authors. speci\ufb01c neck pain (CNSNP). According to symptom duration, CNSNP is usually classi\ufb01ed as pain lasting or recurring for . 3 months.8 Quality of life, mood, ability to cope, social participation, employment rates and job income are reduced and in\ufb02uenced by CNSNP, both for those who are affected by it and their spouses.9 Nevertheless, neck pain has received https:\/\/doi.org\/10.1016\/j.jphys.2022.09.007 1836-9553\/\u00a9 2022 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\/).","Research 245 very little attention in terms of research efforts compared to burden: 0.12 Studies that compared the same class of treatment (eg, low- trials per million DALYs.10 Few large and powered randomised clinical intensity versus high-intensity exercise) are less informative for this trials have focused only on neck pain11,12 and recommendations for its question and were excluded. Studies involving comparisons with management are sometimes extrapolated from evidence on general multiple arms (eg, three-arm trials) with two arms investigating the musculoskeletal pain.13 same intervention at different dosage\/intensity compared with a control group were included following Cochrane guidance for ana- Several conservative interventions for patients with CNSNP are lyses, splitting the sample size of the shared group into two or more commonly used in clinical practice: education, manual therapy, groups with equal sample size.28 therapeutic exercise, electrotherapy, acupuncture, medication such as non-steroidal anti-in\ufb02ammatory drugs (NSAIDs), and a combination Outcomes and time points of these.14\u201320 However, heterogeneity among current guideline rec- The primary outcomes were pain intensity (eg, measured with a ommendations21,22 does not facilitate the clinician\u2019s decision-making numeric rating scale or a visual analogue scale) and disability (eg, because it leaves uncertainty about which treatment options are measured with the Neck Disability Index or the Neck Pain and likely to be the most effective. Disability Scale). The secondary outcome was any adverse event re- ported. According to the Methodological Expectations of Cochrane Furthermore, published systematic reviews have focused only on Intervention Reviews, information was gathered at the following time the pairwise comparison of different treatments,14\u201320 and it is points: short-term (closest to 1 month assessment), intermediate- believed that none have investigated pain, disability and adverse term (closest to 3 to 6 months), and long-term follow-up (closest to events involving all comprehensive evidence on conservative in- 12 months).29 Comparisons with immediate-term follow-ups (eg, 60 terventions in CNSNP. The goal of this study was to compare the seconds after intervention) were excluded. available choices for patients with CNSNP in terms of bene\ufb01ts and harms, via a systematic review with network meta-analysis (NMA). Study selection Therefore, the research question for this systematic review and Two independent authors screened the title and abstract list network meta-analysis of randomised clinical trials was: obtained with the search strategy and assessed full-text copies of potential papers for eligibility. Disagreements were resolved through Which is the most effective conservative intervention for patients discussion or consulting a third author. We used Endnotea for with chronic non-speci\ufb01c neck pain? removing duplicates and Rayyan QCRI to manage the screening phase.30,31 Further details about selection criteria are provided in the Methods published protocol.23 Protocol and registration Data extraction and risk of bias assessment The systematic review protocol was registered in the PROSPERO A pre-de\ufb01ned standardised data extraction in an Excel spread- database and published.23 The Preferred Reporting Items for Systematic sheet was used to collect the data from the included studies. Two Reviews and Meta-Analyses (PRISMA) extension for NMA was followed authors independently extracted data about general characteristics for reporting.24 Additional sections speci\ufb01c to NMA have been consid- and outcomes of interest from the included studies and assessed the ered according to Chaimani et al25 (see Appendix 1 on the eAddenda). risk of bias using the Cochrane Collaboration\u2019s Risk of Bias tool (version modi\ufb01ed by the Cochrane Back and Neck Group).32 For the Data sources and searches clinical context, since blinding is implausible for most of the interventions that were considered, we judged the importance of The following electronic databases were searched from their potential bias using a proxy of the overall assessment for \u2018low\u2019, \u2018high\u2019 inception up to 8 February 2019 and updated on 3 May 2021: or \u2018unclear\u2019 risk of bias with the following domains: \u2018selection bias\u2019, PubMed, Cochrane Controlled Trials Register (CENTRAL), CINAHL, \u2018detection bias\u2019 and \u2018outcome reporting bias\u2019. Disagreements were Scopus, ISI Web of Science and PEDro, using the appropriate resolved via discussion with another member of the reviewing team. thesaurus and free-text terms (Appendix 2 on the eAddenda). No The following characteristics were extracted from each study: name restriction regarding year of publication and language was applied. of \ufb01rst author, country, year of publication, setting, number of centres, References lists of all eligible studies and any systematic reviews population characteristics (eg, age, sex, pain duration), number of retrieved with the search strategy were also checked for eligibility. In participants, percentage of dropouts at each follow-up, type of case of non-English studies for which a translation could not be ob- experimental\/control interventions with details (eg, length of treat- tained, studies were classi\ufb01ed awaiting assessment as potentially ment, frequency), and primary and secondary outcomes. The eligible but not considered in the full analyses. Their contribution was outcome measures of interest were collected at post-treatment as- considered irrelevant when , 5% of the whole included sample; sessments. Corresponding authors were contacted in cases of missing otherwise their \ufb01nding was discussed. data. In cases of no response, missing standard deviations were imputed using their baseline value for a small proportion of studies.33 Eligibility criteria Data from intention-to-treat analysis were used. When outcome data were available only in graphs, they were extracted as numerical data Participants using WebPlotDigitiser.34 Further details about outcome and missing This review included only randomised controlled trials assessing data are provided in Appendix 2 on the eAddenda. adults (. 18 years old) with CNSNP with no known cause,26 de\ufb01ned Data synthesis and analysis as pain lasting for \u0015 3 months at the time of intervention.27 Strategies to ensure inclusion of CNSNP patients are reported in Appendix 2 on Systematic review expressed as qualitative synthesis the eAddenda. We excluded studies involving patients with a speci\ufb01c Overall, the main study and patients\u2019 characteristics were sum- diagnosis (eg, radicular pain, fracture, tumour, in\ufb02ammatory disease), whiplash-associated disorders or \ufb01bromyalgia, and studies involving marised using descriptive synthesis and risk of bias assessment. mixed populations where data for patients with CNSNP were not Studies with unavailable outcome data to allow the NMA (eg, data presented separately. expressed in median due to skewed distribution, p-value, time points not reported) were narratively summarised. Interventions and comparisons We included trials that allocated participants to any of the conservative therapies (non-pharmacological and pharmacological) listed in Box 1, irrespective of modality, frequency, intensity, and length of treatment. Surgical treatments were excluded.","246 Castellini et al: Network meta-analysis for chronic neck pain Box 1. Eligibility criteria. Population \u000f Adults \u000f Chronic non-specific neck pain (. 3 months) Interventions \u000f Any conservative therapies (non-pharmacological and pharmacological) irrespective of modality, frequency or intensity, and length of treatment, alone or combined (maximum two treatments): \u000f Acupuncture\/dry needling: acupuncture; electroacupuncture; needling on trigger points without medication \u000f Cognitive: cognitive behavioural therapy \u000f Education: advice; neck pain educational program; pain education; self-care group \u000f Exercise: active therapeutic exercise (eg, deep cranio-cervical muscle training; exercise with balance devices; free exercise; motor control; postural control; proprioceptive training; supervised exercise); muscle stretching; strength and endurance training \u000f Mind-body practices: Pilates; Qi Gong; Tai-Chi; Yoga; other types of common gymnastics, Alexander technique \u000f Manual therapy: high-velocity low-amplitude\/thrust manipulation (applied to any cervical, thoracic or spinal level); mobilisation (eg, passive mobilisation, mobilisation-with-movement); soft tissue techniques (massage; myofascial techniques; trigger point manual treatment; tuina); manual traction \u000f Usual care: any type of common treatment used in primary care by general practitioner as minimal intervention (advice to stay active and\/ or to take drugs as needed) \u000f Non-steroidal anti-inflammatory drugs (NSAIDs): any kind of NSAIDs drug, including COX-2 inhibitors (eg, ibuprofen, naproxen, sulindac, ketoprofen, tolmetin, etodolac, fenoprofen, diclofenac, flurbiprofen, piroxicam, ketorolac, indomethacin, meloxicam, nabumetone, oxap- rozin, mefenamic acid, diflunisal) \u000f Paracetamol: any dose of paracetamol \u000f Physical agents: cryotherapy; electrotherapy; electromagnetic therapy; electro neuro adaptive regulator therapy device (ENAR); heat therapy; infrared radiation; laser therapy; phonophoresis; transcutaneous electrical nerve stimulation (TENS); ultrasound \u000f Relaxation: guided imagery; relaxation training; stress management \u000f Taping: any type of taping, both elastic (kinesio taping) and non-elastic \u000f Traction: any type of neck or spinal traction (eg, gravitational; mechanical; underwater traction) performed alone without any other manual intervention. Comparators \u000f Inert treatment: any type of intervention described as no intervention; placebo; sham; waiting list. \u000f All against all Outcomes \u000f Pain intensity (eg, Numeric Rating Scale or a Visual Analogue Scale) \u000f Disability (eg, Neck Disability Index or the Neck Pain and Disability Scale) \u000f Adverse events Time points \u000f Short term (closest to 1 month) \u000f Intermediate term (closest to 3 to 6 months) \u000f Long term (closest to 12 months) Study design \u000f Randomised controlled trials Pairwise meta-analysis (direct evidence) (s2) estimated by using network meta-analysis models. Heterogeneity We performed conventional pairwise meta-analysis for each was calculated across all treatment comparisons, accounting for correlations induced by multi-arm studies.44-46 outcome using a random effects model for each treatment compari- son with at least two studies.35 We assessed statistical heterogeneity The design-by-treatment interaction model was used in the whole using the I2 statistic. When the I2 value was . 90% (ie, high degree of network (global c2 test) and the node splitting function for each pairwise heterogeneity), we did not perform meta-analysis.36,37 comparison was used to evaluate the global and local consistency (ie, statistical manifestation of transitivity), respectively.39-41 If the data were Summary of the network consistent with the possibility that both global and local inconsistency According to the PRISMA-NMA guideline,24 the process leading to parameters were equal to zero, we \ufb01t a consistency model. When sub- stantial global inconsistency was found,39,40 multiple strategies were node grouping38 and rationale for node adopted are displayed in explored, such as checking data, splitting nodes to possibly remove Appendix 3 on the eAddenda. sources of the problem, inspecting the in\ufb02uence of effect modi\ufb01ers using the meta-regression random effects within a frequentist framework Assumption of transitivity and geometry of the network (metareg command in Stata).47,48 If these strategies did not resolve the inconsistency, we presented only pairwise comparisons.40 We \ufb01rst evaluated the transitivity assumption,39-41 de\ufb01ned as the balance between the distribution of potential effect modi\ufb01ers (age, Network meta-analysis (mixed and indirect evidence) sex, mean pain duration, presence of widespread pain, presence of Pain and disability estimates were calculated using the stand- somatisation, length of treatment, number of randomised partici- pants, and baseline pain intensity)42 across pairwise comparisons. ardised mean differences (SMDs) alongside 95% con\ufb01dence intervals Transitivity was judged exploring these variables by trials, in- (CIs). The SMDs were back-translated to a typical scale (ie, 0 to 10 for terventions, pairwise comparisons and single networks (eg, outcome pain and 0 to 100 for disability) by multiplying the SMD by the and follow-up). Then, the connection of treatments was evaluated average standard deviation of the sample,49,50 as reported in graphically by a network plot for each primary outcome at follow- Appendices 4 and 5 on the eAddenda. ups: 1 month, 3 to 6 months, and 12 months.43 The outcome data were \ufb01rst carefully checked to detect unusually Assumption of inconsistency (heterogeneity and coherence) large effect estimates. We de\ufb01ned a \u2018large effect size\u2019, visually inspecting pairwise meta-analysis, when SMDs were . 1.5.51 Further The assessment of statistical heterogeneity in the entire network details are provided in Appendix 6 on the eAddenda. was based on the magnitude of the heterogeneity variance parameter After checking the shared nodes in the compared interventions and covariates for any effect modi\ufb01ers, it was assumed that people","Research 247 with CNSNP meeting the inclusion criteria were, in principle, equally effect) to high (ie, a lot of con\ufb01dence that the true effect is similar to likely to be randomised to any of the eligible interventions. the estimated effect).61,62,63 Study limitations, reporting bias, indi- rectness, inconsistency (ie, heterogeneity and incoherence), impre- We assessed direct (ie, pairwise comparisons) and indirect evi- cision and publication bias domains were evaluated using CINeMA dence by network forest plots (mixed evidence) for each primary (Con\ufb01dence in Network Meta-Analysis) to interpret all the \ufb01ndings.64 outcome using a random-effects model within a frequentist frame- Imprecision, heterogeneity and incoherence of network were work.35,52,53 We presented effect sizes of all interventions against a assessed in relation to the clinical relevance.65 The frameworks for reference standard (ie, inert treatment, including sham therapy, pla- pain and disability are reported in Appendices 4 and 5 on the eAd- cebo drugs, waiting list control, no intervention) in each outcome denda, respectively. network at all follow-ups according to the certainty of evidence (ie, from high to low) and the clinical relevance in ad hoc tables. The Results clinical relevance for both pain and disability was achieved consid- ering 25% relative improvement based on baseline values of the Deviations from the study protocol dataset, which meant 1.3 points for pain (0-to-10 scale) and 7.4 for disability (0-to-100 scale) (Appendices 4 and 5 on the eAddenda).54- Deviations from the protocol are reported in Appendix 8 on the 56 Then, clinical relevance was interpreted considering the categories eAddenda. proposed by Man-Son-Hing et al57 (ie, de\ufb01nite, probable, possible, de\ufb01nitely not). All details for interpretation are shown in Appendix 7 Study selection on the eAddenda. Then, we reported the effect size of all available interventions against each other intervention for each outcome at After removing duplicates, the whole search strategy retrieved each follow-up in a league table. In order to identify the superiority of 6,468 records. Screening of titles and abstracts led to 6,170 irrelevant each intervention we estimated: the relative ranking probabilities of hits. The remaining 337 full-text articles were assessed, of which 119 being the best, the mean rank and the surface under cumulative studies met the inclusion criteria (Appendix 9 on the eAddenda). Of ranking (SUCRAs), which expresses the percentage of effectiveness of these, 18 were not included in the quantitative synthesis (16 had an intervention ranked \ufb01rst without uncertainty.58 Stata 16 softwareb outcome reporting bias\/unusable data and two had interventions was used in the analyses.48,59,60 with unclear eligibility). One study provided unusable data for the pain outcome but was included in quantitative analysis for disability. Sensitivity analyses For a more detailed description of the screening process, see the PRISMA \ufb02ow diagram in Figure 1 and full reasons for exclusion at OSF The sensitivity analysis was conducted excluding studies with https:\/\/osf.io\/ac653\/. high risk of bias and missing information (eg, imputed SD). The robustness of evidence (ie, same direction of results and global General characteristics inconsistency) was evaluated to assess any change from primary analyses. A total of 12,496 participants were included in 119 trials (277 arms, 32 different interventions) published between 1986 and 2021. Certainty of evidence Most of studies were conducted in Europe (42.0%) and Asia (37.8%). The sample size per arm ranged from 5 to 1,886 participants, with a For all consistent networks, using the Grading of Recommenda- median of 24 participants. Overall, 108 (90.8%) trials were mono- tion, Assessment, Development and Evaluation (GRADE) approach, centric. The median of the arms\u2019 mean ages was 44.3 years (IQR we evaluated four levels of certainty of evidence from very low (ie, 38.6 to 48.1), whereas the median percentage of males was 26.5 (IQR the true effect is probably markedly different from the estimated Figure 1. Flow of studies through the review. a One study was assessed in the qualitative synthesis only for pain outcome due to not usable data, whereas disability was assessed in the quantitative synthesis.","248 Castellini et al: Network meta-analysis for chronic neck pain 14.4 to 37.0). The median pain duration was 73.7 months (IQR 40.8 to comparison are summarised in Appendix 12 on the eAddenda. No 95.8) and the median length of treatments was 4 weeks (IQR 3 to 8). important concerns were raised regarding the violation of the tran- Baseline mean pain was 53.8 on a 0-to-100 scale; baseline mean sitivity assumption when the potential effect modi\ufb01ers were evalu- disability was 29.6 on the 0-to-100 NDI scale. The most studied in- ated. The transitivity assumption was guaranteed in terms of clinical terventions were exercise (n = 60 arms) and manual therapy (n = 38 and methodological features, except for eight trials that were dis- arms) (Table 1). Characteristics of all included studies are reported in similar in terms of length of treatment (ie, intervention performed Appendix 10 on the eAddenda. within 1 week); thus, these trials were excluded from the quantita- tive analyses at the short term (Appendix 4 on the eAddenda). The Risk of bias assessment inconsistency assessment is reported globally and locally in Appendix 13 on the eAddenda. The risk of bias assessments are summarised in Appendix 11 on the eAddenda. Regarding the overall risk of bias across studies Outcome: Pain intensity (n = 119), 50.4% trials were at low risk of bias (n = 60), 22.7% trials at Pain intensity was investigated by 73 studies at 1 month, 43 unclear risk of bias (n = 27) and 26.9% at high risk of bias (n = 32). studies at 3 to 6 months, and seven studies at 12 months. Table 2 Quantitative synthesis reports the overall summary of the estimates back-translated to a typical scale (ie, 0-to-10 scale) along with the clinical relevance Transitivity assumption interpretation. Studies and participants\u2019 characteristics strati\ufb01ed by network (eg, Short term: After checking data and connection of the network, the outcome and follow-up), intervention nodes, trial and pairwise network meta-analysis of pain at 1 month (55 studies; 4,206 partic- ipants) (Figure 2 and Appendix 14 on the eAddenda) provided data on Table 1 N (%) 40 direct comparisons and 236 indirect comparisons between 24 General characteristics of the included studies. different treatment nodes. No local inconsistency was found. Under 2 (1.68) global consistency, the network meta-analysis showed that manual Characteristic 45 (37.82) therapy was effective compared with inert treatment, with low cer- 50 (42.02) tainty of evidence (SMD \u20130.42, 95% CI \u20130.82 to \u20130.01), followed by six Countrya 15 (12.60) other conservative treatments (SMD range 0.77 to 1.87) with very low Africa certainty of evidence (Appendix 15 on the eAddenda). The forest plot Asia 7 (5.88) of the network meta-analysis is presented in Appendix 16 on the Europe eAddenda. The ranking of treatments based on cumulative proba- America 2 (1.68) bility plots and SUCRAs is presented in Appendix 17 on the eAddenda. Oceania 4 (3.36) The most effective treatment to reduce pain at 1 month was exercise 23 (19.33) with kinesio taping (93.2%), followed by acupuncture\/dry needling Year of publicationa 72 (60.50) with manual therapy (92.6%). All NMA estimates of all interventions 1980 to 1989 18 (15.13) compared with each other intervention are shown in Appendix 18 on 1990 to 1999 the eAddenda. 2000 to 2009 107 (90.76) 2010 to 2019 10 (8.40) Intermediate term: After checking data and connection of network, 2020 to 2021 1 (0.84) the NMA on pain at 3 to 6 months (38 studies, 3,782 participants) provided data on 28 direct and 125 indirect comparisons between 18 Study settinga 23 (8.30) different treatment nodes (Appendix 14 on the eAddenda). No local monocentre 4 (1.44) inconsistency was found but global inconsistency was found; thus, multicentre 1 (0.36) different strategies were followed to explore it (Appendix 13 on the not stated 1 (0.36) eAddenda). Pairwise meta-analyses and NMA are presented in 1 (0.36) Appendix 6 on the eAddenda. The forest plot of the pairwise meta- Interventionsb 1 (0.36) analysis showed that three conservative treatments (education with acupuncture\/dry needling 2 (0.72) exercise, mind body practices and physical agents) are effective acupuncture\/dry needling 1 exercise 3 (1.08) compared with inert treatment (SMD range 0.53 to 0.75). acupuncture\/dry needling 1 manual therapy 2 (0.72) acupuncture\/dry needling 1 NSAIDs 8 (2.89) Long term: After checking data and connection of network, the acupuncture\/dry needling 1 physical therapy 9 (3.25) network meta-analysis on pain at 12 months (7 studies, 1,417 par- acupuncture\/dry needling 1 usual care 2 (0.72) ticipants) provided data on 11 direct and 25 indirect comparisons cognitive 60 (21.66) between nine different treatment nodes (Appendix 14 on the eAd- cognitive 1 exercise 2 (0.72) denda). No local inconsistency was found. Under consistency, the cognitive 1 manual therapy 19 (6.86) network meta-analysis showed that no intervention was substantially education 11 (3.97) better than inert treatment (Appendix 15 on the eAddenda). The education 1 exercise 1 (0.36) certainty of the evidence ranged from very low to low. The forest plot education 1 manual therapy 44 (15.88) of network meta-analysis is presented in Appendix 16 on the eAd- exercise 3 (1.08) denda. The ranking of treatments based on cumulative probabilities exercise 1 kinesio taping 1 (0.36) plots and SUCRAs is presented in Appendix 17 on the eAddenda. In exercise 1 manual therapy 38 (13.72) terms of ef\ufb01cacy, the most effective treatments to reduce pain at 12 exercise 1 physical agents 3 (1.08) months were acupuncture\/dry needling (79.7%) and exercise (73.1%). exercise 1 relaxation 8 (2.89) All network meta-analysis estimates of all interventions against each inert treatment 1 (0.36) other intervention are shown in Appendix 18 on the eAddenda. kinesio taping 1 (0.36) kinesio taping 1 manual therapy 3 (1.08) Outcome: Disability manual therapy 1 (0.36) Disability was investigated by 61 studies at 1-month follow-up, 33 manual therapy 1 physical agents 14 (5.05) mind-body practices 4 (1.44) studies at 3 to 6 months, and eight studies at 12 months. Table 2 mind-body practices 1 paracetamol 1 (0.36) reports the overall summary of the estimates back-translated to a mind-body practices 1 usual care 2 (0.72) typical scale (ie, 0-to-100 scale) along with the clinical relevance NSAIDs 3 (1.08) interpretation. paracetamol physical agents relaxation traction usual care unclear NSAIDs = non-steroidal anti-in\ufb02ammatory drugs. a Out of 119 studies b Out of 277 arms","Research 249 Table 2 Summary of all estimates (back-translated). Treatments vs inert Pain 1 month Pain 3 to 6 monthsa Pain 12 months Disability 1 montha Disability 3 to 6 months Disability 12 months treatment MD (95% CI) MD (95% CI) MD (95% CI) MD (95% CI) MD (95% CI) MD (95% CI) Acu\/dn 0.60 (\u20130.02 to 1.22) 0.63 (\u20130.22 to 1.47) 0.26 (\u20130.42 to 0.94) 0.77 (\u20131.44 to 2.99) 6.94 (2.02 to 11.87) 2.58 (\u20131.23 to 6.38) Acu\/dn 1 exercise 1.24 (\u20130.36 to 2.84) 16.46 (4.93 to 28.00) Acu\/dn 1 manual therapy 3.24 (1.12 to 5.36) 1.31 (0.13 to 2.48) 0.04 (\u20130.86 to 0.94) 0.29 (\u20134.85 to 5.42) \u20136.50 (\u201312.66 to \u20130.34) Acu\/dn 1 NSAIDs 0.60 (\u20131.76 to 2.96) 10.26 (5.40 to 15.12) 11.42 (\u20130.45 to 23.30) 1.46 (\u20133.70 to 6.50) Acu\/dn 1 usual care 1.06 (0.22 to 1.90) \u20130.32 (\u20131.04 to 0.42) 14.09 (7.80 to 20.38) 13.44 (\u20132.13 to 28.90) Cognitive 1.62 (\u20130.52 to 3.74) \u20131.34 (\u20132.78 to 0.10) 0.08 (\u20130.40 to 0.56) 12.54 (\u20131.68 to 26.88) \u20133.25 (\u20137.28 to 0.90) Cognitive 1 exercise 1.46 (\u20130.76 to 3.68) 1.49 (0.66 to 2.33) \u20130.10 (\u20130.70 to 0.52) \u20131.57 (\u20134.26 to 1.12) Cognitive 1 manual 2.26 (\u20130.16 to 4.68) \u20130.07 (\u20130.56 to 0.42) 0.11 (\u20137.17 to 7.39) \u20132.02 (\u20135.38 to 1.46) therapy \u20130.54 (\u20131.20 to 0.14) 3.36 (\u20131.12 to 7.84) Education \u20130.52 (\u20131.98 to 0.92) 8.18 (2.35 to 14.11) \u20135.60 (\u20139.41 to \u20131.79) Education 1 exercise 0.40 (\u20130.64 to 1.44) \u20130.08 (\u20130.56 to 0.40) \u20136.72 (\u201312.99 to \u20130.56) Exercise 0.80 (\u20130.04 to 1.64) \u20130.68 (\u20131.68 to 0.34) 9.07 (\u20132.13 to 20.27) Exercise 1 kinesio taping 3.74 (1.36 to 6.12) 4.70 (\u20131.79 to 11.20) \u20131.57 (\u20134.26 to 1.23) Exercise 1 manual therapy 1.62 (0.60 to 2.62) 5.60 (0.22 to 10.98) \u20139.74 (\u201315.46 to \u20134.03) Exercise 1 physical agents 1.60 (0.48 to 2.72) Exercise 1 relaxation 1.06 (\u20131.12 to 3.22) 7.28 (0.34 to 14.22) Kinesio taping 1.28 (\u20130.22 to 2.80) 1.34 (\u20137.62 to 10.19) Manual therapy 1 physical 1.72 (0.18 to 3.26) 5.82 (\u20136.94 to 18.70) agents Manual therapy 0.84 (0.02 to 1.64) Mind-body 1.16 (\u20130.16 to 2.46) Mind-body 1 usual care NSAIDs 0.10 (\u20132.24 to 2.46) Physical agents 1.54 (0.64 to 2.44) Relaxation 0.98 (\u20130.28 to 2.24) Traction 1.44 (\u20130.72 to 3.60) Usual care 0.28 (\u20131.92 to 2.46) De\ufb01nite Probable Possible Not probable Not possible De\ufb01nitely not Abbreviations: Acu\/dn = acupuncture or dry needling, Mind-body = mind body practices, NSAIDs = non-steroidal anti-in\ufb02ammatory drugs. SMDs were back-translated to a typical scale (ie, 0 to 10 for pain and 0 to 100 for disability) by multiplying the SMD by the average standard deviation of the sample as reported in Appendices 17 and 18. The clinical relevance for both pain and disability was achieved considering 25% relative improvement based on baseline values of the dataset, which means 1.3 points for pain (0-to-10 scale) and 7.4 for disability (0-to-100 scale). Positive MDs favour the row treatments. Interpretation of clinical relevance was graded in the categories reported in the coloured bar (Appendix 7). Darker green indicates that the intervention listed in the row is more likely to be clinically relevant, whereas darker orange indicates that the intervention is not likely to be clinically relevant. Estimates with con\ufb01dence intervals that do not include zero are bolded. a Pairwise meta-analysis (network meta-analysis inconsistent). Short term: After checking data and connection of network, the interventions against each other intervention are shown in Appendix network meta-analysis on disability at 1 month (54 studies, 3,979 18 on the eAddenda. participants) (Appendix 14 on the eAddenda) provided data on 40 direct comparisons between 24 different treatment nodes. No local Long term: After checking data and connection of network, the inconsistency was identi\ufb01ed but global inconsistency was found; thus, network meta-analysis on disability at 12 months (eight studies, 1,934 different strategies were followed to explore inconsistency (Appendix participants) provided data on 14 direct comparisons and 41 indirect 13 on the eAddenda). Pairwise meta-analyses are presented in comparisons between 11 intervention nodes (Appendix 14 on the Appendix 6 on the eAddenda. The forest plot of the pairwise meta- eAddenda). No local inconsistency was found. Under consistency, the analysis showed that two conservative treatments (physical agents, NMA showed that inert treatment was statistically signi\ufb01cant relaxation) are effective compared with inert treatment (SMD range compared with four conservative treatments (Appendix 15 on the 0.92 to 1.26). eAddenda) with low certainty of evidence. The forest plot of network meta-analysis is presented in Appendix 16 on the eAddenda. The Intermediate term: After checking data and connection of network, ranking of treatments based on cumulative probabilities plots and the network meta-analysis on disability at 3 to 6 months (27 studies, SUCRAs is presented in Appendix 17 on the eAddenda. In terms of 6,636 participants) provided data on 23 direct comparisons and 82 ef\ufb01cacy, the most effective treatments to reduce disability at 12 months indirect comparisons between 15 different treatment nodes were acupuncture\/dry needling (94.1%) and cognitive behavioural (Appendix 14 on the eAddenda). No local inconsistency was found. treatment (86.6%). All NMA estimates of all interventions against each Under consistency, the network meta-analysis showed that \ufb01ve other intervention are shown in Appendix 18 on the eAddenda. conservative treatments were statistically signi\ufb01cant compared with inert treatment, with certainty of the evidence ranging from very low Outcome: Adverse events to low (SMD range 0.50 to 1.47) (Appendix 15 on the eAddenda). The Overall, 54 of 119 studies collected adverse events. Since reporting forest plot of network meta-analysis is presented in Appendix 16 on the eAddenda. The ranking of treatments based on cumulative of adverse events was heterogeneous for number of people and probabilities plots and SUCRAs is presented in Appendix 17 on the number of events, these data could not be quantitated. The majority eAddenda. In terms of ef\ufb01cacy, the most effective treatment to reduce of mild events that were reported were experienced by the disability at 3 to 6 months was acupuncture\/dry needling with ex- acupuncture\/dry needling arm (1,288 events in 10 studies on 2,353 ercise (88.2%) followed by cognitive behavioural treatment with ex- randomised participants). No serious event was reported. Further ercise (77.3%). All network meta-analysis estimates of all details about adverse events are provided in Table 3 and Appendix 10 on the eAddenda.","250 Castellini et al: Network meta-analysis for chronic neck pain Figure 2. Network graph for pain at 1-month follow-up. Acu-dn = acupuncture\/dry needling, NSAIDs = non-steroidal anti-in\ufb02ammatory drugs, Mind-body = mind-body practices The size of the nodes is proportional to the number of studies evaluating each intervention, and the thickness of the edges is proportional to the precision (the inverse of the variance) of each direct comparison. Sensitivity analysis therapy, manual therapy with acupuncture\/dry needling, manual therapy with physical agents, and physical agents may reduce pain Sensitivity analysis was conducted on all outcome networks at 1 (mixed and indirect evidence), whereas physical agents and relaxa- month and 3 to 6 months follow-ups (Appendix 19 on the eAddenda). tion may reduce disability (direct evidence). At 3 to 6 months, edu- Overall, exclusion of missing information (eg, imputed SD) and high- cation with exercise, physical agents and mind-body practices may risk of bias studies did not affect the robustness of evidence. reduce pain (direct evidence), whereas acupuncture\/dry needling, acupuncture\/dry needling with exercise, exercise with manual ther- Certainty of evidence apy, physical agents and mind-body practices may reduce disability (mixed and indirect evidence). At 12 months, no superiority was The GRADE judgments are reported in Appendix 4 for the pain found between interventions and inert treatment on pain (mixed and network meta-analysis and Appendix 5 for the disability network indirect evidence), whereas a few conservative interventions may meta-analysis on the eAddenda. In the analysis of pain at 1 month have a less helpful effect on disability than inert treatment (indirect (n = 276 comparisons), the evidence was mainly downgraded with evidence). major concerns for incoherence (due to disagreement about the range of clinical relevance) (n = 250) and imprecision (due to wide con\ufb01- These interventions can generally be safely provided by clinicians; dence intervals) (n = 122), and with some concerns for within-study however, acupuncture\/dry needling may elicit some mild adverse bias (due to selection, detection and outcome reporting bias) events. (n = 185) and indirectness (due to the majority of evidence coming from indirect comparisons) (n = 218). In the analysis of pain at 12 Generally, network meta-analysis \ufb01ndings (mixed and indirect months (n = 36 comparisons), the evidence was mainly with some evidence) were also con\ufb01rmed in the probability of being the best concerns for indirectness (n = 25). In the analysis of disability at 3 to 6 treatment. However, readers and stakeholders developing or updat- months (n = 105), there were major concerns for imprecision (n = 72) ing guidelines should take caution in the interpretation of the results, and some concerns for within-study bias (n = 74). In the analysis of given the low and very low certainty evidence, mainly due to mostly disability at 12 months (n = 55), there were some concerns for within- indirect evidence, imprecision in the estimates (ie, small nodes and study bias (n = 23) and indirectness (n = 41). No publication bias was large con\ufb01dence intervals), outcome non-reporting bias and unclear evident assessing small-study effects in both pain and disability reporting of effect modi\ufb01ers. Further research may have a consider- outcome at each follow-up (Appendix 20 on the eAddenda). Addi- able impact on the results. tional references are presented in Appendix 21 on the eAddenda. Short-term follow-up can be considered the most reliable follow- Discussion up as a proxy of the treatment effects because it is the closest mea- surement to the length of the majority of treatments (median 4 It is believed that this is the largest systematic review with weeks). Consequently, intermediate term is the \ufb01rst follow-up that network meta-analysis regarding comparative effectiveness of 32 evaluates persistence of effects, whereas the long-term follow-up is different conservative interventions in 12,496 patients with CNSNP. the latest. This could be the reason why treatment effects at the long- Overall, it was found that multimodal non-pharmacological in- term follow-up seem to be equal; previous Cochrane systematic re- terventions were more ef\ufb01cacious than inert treatment for reducing views displayed similar results for long-term outcomes.16,18 both pain and disability outcomes, with possible to de\ufb01nite clinical importance. At 1 month of follow-up, exercise with kinesio taping, Clinical implications exercise with manual therapy, exercise with physical agents, manual The main \ufb01ndings support the current recommended in- terventions reported in the latest published guidelines, in which multimodal interventions (ie, rehabilitative programs including two interventions) are recommended for patients with CNSNP.21,66","Research 251 Table 3 Eventsa Participants Studies Description of events Adverse events. (n) (n) (n) 10 Muscle soreness; pain at acupoint, dizziness, local bleeding, numbness; Node 1,288 2,353 haematoma; fainting; bruise at the site; chest discomfort; neck pain, 2 headache; swelling of the hand, pain and ulcer of the ear; vegetative Acu\/dn 0 28 1 symptoms; euphoria 0 47 1 \u2013 Acu\/dn 1 exercise 1 15 1 \u2013 Acu\/dn 1 manual therapy 10 173 Flushing, skin rash Acu\/dn 1 NSAIDs 2 Bruising, swelling, or numbness; muscle spasms; pain; and respiratory Acu\/dn 1 usual care 1 130 1 problems 0 16 1 Increased pain Cognitive 1 exercise 0 48 4 \u2013 Cognitive 1 manual therapy 30 193 \u2013 Education 1 Muscle soreness, upper extremity symptoms, headache, back pain, jaw Education 1 exercise 9 32 22 pain, nausea, and dizziness 127 758 Discomfort or pain, soreness, nausea Education 1 manual therapy Muscle soreness; muscle tension, aching muscles; transient limb pain, Exercise worsening of neck pain; increase radicular pain; migraine; vertigo; nausea and vomiting; dizziness, fainting; headache, back pain, jaw pain; knee pain and myogelosis, worsening of tinnitus Exercise 1 kinesio taping 0\u02c7 48 2\u2013 Exercise 1 manual therapy 99 405 10 Dizziness, fainting, nausea and vomiting; muscle soreness, increases in Exercise 1 physical agents 0 78 neck or headache pain; upper extremity symptoms, back pain, jaw pain. Exercise 1 relaxation 0 35 3\u2013 Inert treatment 57 688 1\u2013 16 Increased pain; headache; nausea; tingling; \u2018spaced-out\u2019 feeling; Kinesio taping 2 76 Manual therapy 49 453 sleepiness; tiredness; skin sensitivity; jaw pain; stiffness; depression; numbness; aching; fainting; swelling of the hand; muscle soreness; Mind-body 49 205 myogelosis; vertigo; other pain; thirst; engorged hands; twinge in the neck; urinary urgency; bursitis; cephalea; euphoria, dizziness, itching Mind-body 1 paracetamol 0 32 palm, warm feeling Mind-body 1 usual care 3 172 2 Cutaneous irritations NSAIDs 3 21 13 Muscle fatigue; headache, local soreness; increases in neck pain, thoracic Physical agents 21.9 b 107 pain; aching muscles; tensions, dizziness; sleepiness; mood swings; painful point; nausea; \u2018head not movable\u2019 Relaxation 3 70 6 Aching muscles, muscle tension; worsening of neck pain; muscle soreness Usual care 2 172 after practice; transient limb pain, migraine, nausea; vertigo; other Total 1,754.9 musculoskeletal, pain; Achilles tendon pain; meniscal tear; low back pain; \u2013 myogelosis; headache; thirst; engorged hands; twinge in the neck, urinary urgency; bursitis. 1\u2013 1 Pain and incapacity, knee injury, and muscle spasms 1 Gastric symptoms 4 Increased pain; headache; nausea; light-headed\/dizzy; tingling in extremity \u2018spaced-out\u2019 feeling; sleepiness; tiredness; skin sensitivity; stiffness depression; worsening symptoms 2 Dizziness; headache; tinnitus 1 Pain and incapacity \u2013\u2013 Acu\/dn = acupuncture or dry needling, Mind-body = mind body practices, NSAIDs = non-steroidal anti-in\ufb02ammatory drugs, a Number of adverse events or number of patients experiencing adverse events. More than one event may occur in the same patient. b For one study (Chow 2006) adverse events were collected calculating the mean of overall events reported. Nevertheless, this review provides some guidance on which a com- Another network meta-analysis for CNSNP was recently published, bination of interventions may be the most effective. While guidelines but this review displayed a different goal because the authors focus more on active interventions (eg, education and exercise), this investigated only the effectiveness of exercise interventions,68 review shows that a combination of active and passive treatment (eg, con\ufb01rming the current results and thus the effectiveness of some exercise and manual therapy), or a combination of two passive mo- types of exercise (eg, motor control, strengthening exercises) dalities (eg, acupuncture and manual therapy) may also be among the compared with no treatment, even if no type of exercise was most effective intervention options. Neck pain guidelines also superior to others. recommend, with weak evidence, the use of some painkillers.66 Our network meta-analysis highlights negligible differences between Strength and limitations NSAIDs with or without another conservative treatment and inert treatment for pain at 1 month. However, only a minority of phar- Some strengths of this review include a comprehensive systematic macological interventions were included in this systematic review review methodology that followed the Cochrane Handbook and (six arms) and we cannot exclude the presence of co-interventions PRISMA-NMA reporting guidance.24,33 It transparently reported the (eg, drug as needed). differences between protocol and review, as well as the assessment of transitivity in all networks. Additionally, discrepancies with recommended interventions are present because this review was based on the most recent evidence There were several limitations to this study. First, disability at 1 that has emerged in the last few years, whereas review teams of month and pain at 3 to 6 months showed inconsistency, which pre- clinical guidelines did not take this evidence into account; as an vented network meta-analysis from being performed. In fact, a high example, 18 trials (19.4%) published from 2019 to 2021 were percentage of sparse nodes (32% at 1 month for disability and 55% at 3 included in the quantitative synthesis in this review. Additional to 6 months for pain) was found (treatment investigated by one evidence can improve the precision of the estimated effect size.67","252 Castellini et al: Network meta-analysis for chronic neck pain study), leaving analyses poorly informative.40 To be more conserva- data are challenging and more prone to missing outcome data,72,78 tive, pairwise meta-analysis was reported. trials should report all outcome measurements (eg, mean and SD at each follow-up) avoiding missing evidence for quantitative anal- Second, some studies (21.9% of the whole sample) did not ysis70,79 (eg, graphs, p-values) following the CONSORT statement to contribute to the network meta-analysis because of non-transitive ensure transparency of results.80 Lastly, future systematic reviews follow-up (within 1 week), missing outcome data, or unusable should specify the duration of treatment in the study question measurements (eg, skewed distribution, p-value, time points not because this element can in\ufb02uence the timepoint for the post- reported).69 The last two points re\ufb02ect a potential for selective non- treatment therapy effects and persistence of effects. Furthermore, reporting bias that can distort quantitative analyses because the power of meta-analyses may bene\ufb01t from the existence of a core available results differ systematically from missing results.70 A recent outcome set for CNSNP, like those existing for low back pain tri- meta-research study found that at least one important outcome was als.81,82 All these considerations may help to improve the certainty of missing for 63% of randomised trials, but this waste was avoidable.71 the evidence and, consequently, future network meta-analyses in This percentage increased when also considering the missing SD and people with CNSNP. imputations for some outcomes in the included trials. However, the current review did not exclude these studies from the primary ana- Conclusion lyses because, under some circumstances, it can reduce precision in the estimated treatment effects and produce biased results; we This network meta-analysis showed that multimodal non- transparently presented the sensitivity analyses instead (excluding pharmacological interventions may be safe and effective for pain imputation of missing data).72 and disability at short-term and intermediate-term follow-up when compared with inert treatment in patients with CNSNP. Overall, given Third, we did not separate no treatment\/waiting list from sham\/ the generally unclear certainty of evidence, all results should be placebo interventions because too few studies reported a no- interpreted with caution. intervention or waiting-list control to allow the analyses. By not separating them, we also avoided a sparse network.73 A recent meta- Author contributions: Conceptualisation: PP, SG, AC, CV, LB; data analysis showed that placebo interventions are more effective than curation: LB, FF, FM, CV, CG, SG, SB; formal analysis: GC, SG, SB; no intervention on chronic low back pain in the short term.74 To investigation: LB, FF, FM, CV, SG, TI; methodology: GC, SG, AC, LB, SB; acknowledge this limitation of our approach, comparisons with a resources: CG, SG, SB; software: CG, SG, SB; supervision: LB, PP; combined inert treatment node (no treatment\/waiting list and sham\/ writing \u2013 original draft: GC, SG, CV, LB, SB; writing \u2013 review and placebo interventions) were downgraded in the certainty of the editing: GC, PP, CV, SB, SG, EB, FF, FM, TI, AC, SG, LB. evidence. Availability of data and materials: The full dataset is freely Fourth, the evaluation of the transitivity assumption was chal- available online in OSF (https:\/\/osf.io\/ac653\/), a secure online re- lenging because the majority of the included trials did not report the pository for research data. presence of widespread pain, somatisation and pain catastrophising, even though these may be treatment effect modi\ufb01ers.75 Most of the What was known on this topic: Chronic non-specific neck included interventions were combined interventions (eg, exercise pain is common and can affect quality of life, mood, social with manual therapy) and the usual care node was broadly de\ufb01ned by participation and employment. Clinical practice guidelines focus study authors, increasing the chance of incorporating more hetero- on active interventions (eg, education and exercise) and recom- geneity within the node and widening the con\ufb01dence intervals across mend with weak evidence the use of analgesic medication. the whole analysis. However, we carefully appraised the usual care What this study adds: A combination of active and\/or passive intervention to select only those that met our a priori de\ufb01nition. multimodal non-pharmacological treatments (eg, exercise and Further dif\ufb01culties in the categorisation of treatments were due to manual therapy) may be the most effective options compared poor reporting in the description of interventions. In addition, we did with inert treatment whereas anti-inflammatory drugs with or not consider the head-to-head study design con\ufb01guration comparing without another non-pharmacological treatment are not superior different characteristics of delivery (eg, low-intensity versus high- to inert treatment to reduce pain in the short term. Mild adverse intensity exercise, different drugs dosages) because they were less events were mainly experienced during acupuncture\/dry needling informative for the review question. treatment. Clinicians should opt for the best intervention balancing benefits and harms according to patients\u2019 preferences. Fifth, very few studies assessed long-term follow-up analysing persistence of effects. We did not select studies based on Footnotes: aEndNote, Version EndNote X9, Clarivate, Philadelphia, treatment effect duration. Currently, only one systematic review USA. based on individual patient data has evaluated the persistence of effects.76 bStata Statistical Software, v 15. StataCorp LLC, College Station, USA. Research implications eAddenda: Appendices 1 to 21 can be found online at https:\/\/doi. This is the \ufb01rst attempt to make quantitative comparison of in- org\/10.1016\/j.jphys.2022.09.007 terventions in the absence of pairwise comparison trials and to pro- vide ranking of each treatment being the best. Overall, while network Ethics approval: Nil. meta-analysis is an attractive statistical tool, this review highlights Competing interests: Nil. that the limitations of the evidence base (eg, trials with small sam- Source(s) of support: The work was supported by the Italian ples, high risk of bias, poor reporting, risk of bias due to missing Ministry of Health \u201cEf\ufb01cacia degli interventi riabilitativi nel back evidence) may hamper its applicability. To facilitate future network pain: una network meta-analisi\u201d (L3026). The funding sources had no meta-analyses in the \ufb01eld of CNSNP, higher quality and well-planned controlling role in the study design, data collection, analysis, inter- larger trials are necessary. For instance, one-third of the sample was pretation, or report writing. at high risk of bias and most studies were monocentric (90.8%) with a Acknowledgements: We wish to thank Dr Angeliki Veroniki for median of 24 participants. Randomised trials should be adequately her precious statistical contribution, Dr Pierluigi Pelliccioni, Dr Riikka designed using sample size calculation, well-conducted, appropri- Sailola, and Dr Eliisa L\u00f6yttyniemi for their help in Finnish translation ately reported using checklists like the TIDieR on the reporting of and Mrs Simindokht Yazdani for the Persian translation. interventions77 and well-described in all effect modi\ufb01ers. This could Provenance: Silvia Bargeri, IRCCS Istituto Ortopedico Galeazzi, help generate a better con\ufb01guration of the network nodes, assuring Unit of Clinical Epidemiology, Milan, Italy. Email: silvia.bargeri@ similarity of interventions and also allowing the inspection of grupposandonato.it different modalities of the same interventions (eg, inclusion of low- Correspondence: Not invited. Peer reviewed. intensity versus high-intensity exercise). Since continuous outcome","Research 253 References 28. Higgins JPT TJ, Chandler J, Cumpston M, Li T, Page MJ, Welch VA (editors). Cochrane Handbook for Systematic Reviews of Interventions version 6.3 (updated February 1. Sa\ufb01ri S, Kolahi AA, Hoy D, Hill C, Bettampadi D, Mansournia MA, et al. Global, 2022). Available from https:\/\/handbook-5-1.cochrane.org\/chapter_16\/16_5_4_ regional, and national burden of neck pain in the general population, 1990\u20132017: how_to_include_multiple_groups_from_one_study.htm. systematic analysis of the Global Burden of Disease Study 2017. BMJ. 2020;368:m791. 29. McKenzie JE, Brennan SE, Ryan RE, Thomson HJ, Johnston RV. Chapter 9: Sum- marizing study characteristics and preparing for synthesis. In: Higgins JPT, Thomas 2. Global burden of 369 diseases and injuries in 204 countries and territories, J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA (editors). Cochrane Handbook 1990\u20132019: a systematic analysis for the Global Burden of Disease Study 2019. for Systematic Reviews of Interventions version 6.2 (updated February 2021). Lancet. 2020;396(10258):1204\u20131222. Cochrane, 2021. Available from www.training.cochrane.org\/handbook [accessed 12 December 2020]. 3. Cieza A, Causey K, Kamenov K, Hanson SW, Chatterji S, Vos T. Global estimates of the need for rehabilitation based on the Global Burden of Disease study 2019: a 30. EndNote [computer program]. Version EndNote X9. Philadelphia, PA: Clarivate; systematic analysis for the Global Burden of Disease Study 2019. Lancet. 2013. 2020;396(10267):2006\u20132017. 31. Ouzzani M, Hammady H, Fedorowicz Z, Elmagarmid A. Rayyan-a web and mobile 4. Malchaire JB, Roquelaure Y, Cock N, Piette A, Vergracht S, Chiron H. Musculoskel- app for systematic reviews. Syst Rev. 2016;5:210. etal complaints, functional capacity, personality and psychosocial factors. Int Arch Occup Environ Health. 2001;74:549\u2013557. 32. Furlan AD, Malmivaara A, Chou R, Maher CG, Deyo RA, Schoene M, et al. 2015 Updated Method Guideline for Systematic Reviews in the Cochrane Back and Neck 5. Hogg-Johnson S, van der Velde G, Carroll LJ, Haldeman S, Holm LW, Carragee EJ, Group. Spine. 2015;40:1660\u20131673. et al. The burden and determinants of neck pain in the general population: results of the Bone and Joint Decade 2000\u20132010 Task Force on Neck Pain and Its Associ- 33. Higgins J, Deeks J, Altman D. Chapter 16: Special topics in statistics. In: Higgins JPT, ated Disorders. Spine (Phila Pa 1976). 2008;33(4 Suppl):S39\u2013S51. Green S (editors). Cochrane Handbook for Systematic Reviews of Interventions. Version 5.1.0 (updated March 2011). Cochrane Collaboration. Available from www. 6. Hush JM, Lin CC, Michaleff ZA, Verhagen A, Refshauge KM. Prognosis of acute cochrane-handbook.org. 2011. idiopathic neck pain is poor: a systematic review and meta-analysis. Arch Phys Med Rehabil. 2011;92:824\u2013829. 34. Li T HJ, Deeks JJ (editors). Chapter 5: Collecting data. In: Higgins JPT, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA (editors). Cochrane Handbook for 7. Dieleman JL, Cao J, Chapin A, Chen C, Li Z, Liu A, et al. US Health Care Spending by Systematic Reviews of Interventions version 6.2 (updated February 2021). Payer and Health Condition, 1996\u20132016. JAMA. 2020;323:863\u2013884. Cochrane, 2021. Available from www.training.cochrane.org\/handbook. 8. Treede RD, Rief W, Barke A, Aziz Q, Bennett MI, Benoliel R, et al. Chronic pain as a 35. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. symptom or a disease: the IASP Classi\ufb01cation of Chronic Pain for the International 1986;7:177\u2013188. Classi\ufb01cation of Diseases (ICD-11). Pain. 2019;160:19\u201327. 36. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta- 9. Cohen SP, Hooten WM. Advances in the diagnosis and management of neck pain. analyses. BMJ. 2003;327:557\u2013560. BMJ. 2017;358:j3221. 37. Deeks JJ HJ, Altman DG (editors). Chapter 10: Analysing data and undertaking 10. Emdin CA, Odutayo A, Hsiao AJ, Shakir M, Hopewell S, Rahimi K, et al. Association meta-analyses. In: Higgins JPT, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, between randomised trial evidence and global burden of disease: cross sectional Welch VA (editors). Cochrane Handbook for Systematic Reviews of Interventions study (Epidemiological Study of Randomized Trials\u2013ESORT). BMJ. 2015;350: version 6.3 (updated February 2022). Cochrane, 2022. h117. 38. James A, Yavchitz A, Ravaud P, Boutron I. Node-making process in network meta- 11. Kjellman GV, Skargren EI, Oberg BE. A critical analysis of randomised clinical trials analysis of nonpharmacological treatment are poorly reported. J Clin Epidemiol. on neck pain and treatment ef\ufb01cacy. A review of the literature. Scand J Rehabil Med. 2018;97:95\u2013102. 1999;31:139\u2013152. 39. White I, Barrett J, Jackson D, Higgins J. Consistency and inconsistency in network 12. Rubinstein SM, van Eekelen R, Oosterhuis T, de Boer MR, Ostelo RW, van meta-analysis: model estimation using multivariate meta-regression. Res Synth Tulder MW. The risk of bias and sample size of trials of spinal manipulative therapy Methods. 2012;3:111\u2013125. for low back and neck pain: analysis and recommendations. J Manip Physiol Ther. 2014;37:523\u2013541. 40. Veroniki AA, Vasiliadis HS, Higgins JP, Salanti G. Evaluation of inconsistency in networks of interventions. Int J Epidemiol. 2013;42:332\u2013345. 13. Lin I, Wiles L, Waller R, Goucke R, Nagree Y, Gibberd M, et al. What does best practice care for musculoskeletal pain look like? Eleven consistent recommenda- 41. Jansen JP, Naci H. Is network meta-analysis as valid as standard pairwise meta- tions from high-quality clinical practice guidelines: systematic review. Br J Sports analysis? It all depends on the distribution of effect modi\ufb01ers. BMC Med. Med. 2020;54:79\u201386. 2013;11:159. 14. Driessen MT, Lin CW, van Tulder MW. Cost-effectiveness of conservative treat- 42. Artus M, Campbell P, Mallen CD, Dunn KM, van der Windt DA. Generic prognostic ments for neck pain: a systematic review on economic evaluations. Eur Spine J. factors for musculoskeletal pain in primary care: a systematic review. BMJ Open. 2012;21:1441\u20131450. 2017;7:e012901. 15. Miller J, Gross A, D\u2019Sylva J, Burnie SJ, Goldsmith CH, Graham N, et al. Manual 43. Chaimani A, Salanti G. Visualizing assumptions and results in network meta- therapy and exercise for neck pain: a systematic review. Man Ther. 2010;15:334\u2013 analysis: The network graphs package. Stata J. 2015;15:905\u2013950. 354. 44. Salanti G. Indirect and mixed-treatment comparison, network, or multiple- 16. Gross A, Miller J, D\u2019Sylva J, Burnie SJ, Goldsmith CH, Graham N, et al. Manipulation treatments meta-analysis: many names, many bene\ufb01ts, many concerns for the or mobilisation for neck pain: a Cochrane Review. Man Ther. 2010;15:315\u2013333. next generation evidence synthesis tool. Res Synth Methods. 2012;3:80\u201397. 17. Gross A, Forget M, St George K, Fraser MM, Graham N, Perry L, et al. Patient ed- 45. Miladinovic B, Hozo I, Chaimani A, Djulbegovic B. Indirect treatment comparison. ucation for neck pain. Cochrane Database Syst Rev. 2012:CD005106. Stata J. 2014;14:76\u201386. 18. Monticone M, Cedraschi C, Ambrosini E, Rocca B, Fiorentini R, Restelli M, et al. 46. Jackson D, Barrett JK, Rice S, White IR, Higgins JP. A design-by-treatment interac- Cognitive-behavioural treatment for subacute and chronic neck pain. Cochrane tion model for network meta-analysis with random inconsistency effects. Stat Med. Database Syst Rev; 2015:CD010664. 2014;33:3639\u20133654. 19. Tough EA, White AR, Cummings TM, Richards SH, Campbell JL. Acupuncture 47. Sharp S. Meta-analysis regression. Stata Technical Bulletin 42:16\u201322. Reprinted in and dry needling in the management of myofascial trigger point pain: a Stata Tech Bull Reprints. College Station, TX: Stata Press;1998. p. 148\u2013155. systematic review and meta-analysis of randomised controlled trials. Eur J Pain. 2009;13:3\u201310. 48. White I. Multivariate random-effects meta-regression: updates to mvmeta. Stata J. 2011;11:255\u2013270. 20. Machado GC, Maher CG, Ferreira PH, Day RO, Pinheiro MB, Ferreira ML. Non-ste- roidal anti-in\ufb02ammatory drugs for spinal pain: a systematic review and meta- 49. Sch\u00fcnemann HJ OA, Vist GE, Higgins JPT, Deeks JJ, Glasziou P, et al (editors). analysis. Ann Rheum Dis. 2017;76:1269\u20131278. Chapter 12: Interpreting results and drawing conclusions. In: Higgins JPT, Churchill R, Chandler J, Cumpston MS (editors). Cochrane Handbook for Systematic Reviews 21. Parikh P, Santaguida P, Macdermid J, Gross A, Eshtiaghi A. Comparison of CPG\u2019s for of Interventions version 5.2.0 (updated June 2017). Cochrane, 2017. Available from the diagnosis, prognosis and management of non-speci\ufb01c neck pain: a systematic training.cochrane.org\/handbook\/archive\/v5.2. review. BMC Musculoskelet Disord. 2019;20:81. 50. Murad MH, Wang Z, Chu H, Lin L. When continuous outcomes are measured using 22. Corp N, Mansell G, Stynes S, Wynne-Jones G, Mors\u00f8 L, Hill JC, et al. Evidence-based different scales: guide for meta-analysis and interpretation. BMJ. 2019;364:k4817. treatment recommendations for neck and low back pain across Europe: A sys- tematic review of guidelines. Eur J Pain. 2020;25:275\u2013295. 51. Faraone SV. Interpreting estimates of treatment effects: implications for managed care. P T. 2008;33:700\u2013711. 23. Pillastrini P, Castellini G, Chiarotto A, Fasciani F, Marzioni F, Vanti C, et al. Comparative effectiveness of conservative and pharmacological interventions for 52. Higgins JP, Jackson D, Barrett JK, Lu G, Ades AE, White IR. Consistency and incon- chronic non-speci\ufb01c neck pain: Protocol of a systematic review and network meta- sistency in network meta-analysis: concepts and models for multi-arm studies. Res analysis. Medicine. 2019;98:e16762. Synth Methods. 2012;3:98\u2013110. 24. Hutton B, Salanti G, Caldwell DM, Chaimani A, Schmid CH, Cameron C, et al. The 53. White IR, Barrett JK, Jackson D, Higgins JP. Consistency and inconsistency in PRISMA extension statement for reporting of systematic reviews incorporating network meta-analysis: model estimation using multivariate meta-regression. Res network meta-analyses of health care interventions: checklist and explanations. Synth Methods. 2012;3:111\u2013125. Ann Intern Med. 2015;162:777\u2013784. 54. Cleland JA, Childs JD, Whitman JM. Psychometric properties of the Neck Disability 25. Chaimani A, Caldwell DM, Li T, Higgins JP, Salanti G. Additional considerations are required when preparing a protocol for a systematic review with multiple in- Index and Numeric Pain Rating Scale in patients with mechanical neck pain. Arch terventions. J Clin Epidemiol. 2017;83:65\u201374. Phys Med Rehabil. 2008;89:69\u201374. 55. Young BA, Walker MJ, Strunce JB, Boyles RE, Whitman JM, Childs JD. Responsive- 26. InformedHealth.org [Internet]. Cologne, Germany: Institute for Quality and Ef\ufb01- ciency in Health Care (IQWiG); 2006-. What can you do about non-speci\ufb01c neck ness of the Neck Disability Index in patients with mechanical neck disorders. Spine pain? 2015 Dec 2 [Updated 2019 Feb 14]. Available from https:\/\/www.ncbi.nlm.nih. gov\/books\/NBK338118\/ J. 2009;9:802\u2013808. 56. Young IA, Cleland JA, Michener LA, Brown C. Reliability, construct validity, and 27. van Tulder M, Becker A, Bekkering T, Breen A, del Real MT, Hutchinson A, et al. Chapter 3. European guidelines for the management of acute nonspeci\ufb01c low back responsiveness of the neck disability index, patient-speci\ufb01c functional scale, and pain in primary care. Eur Spine J. 2006;15(Suppl 2):S169\u2013S191. numeric pain rating scale in patients with cervical radiculopathy. Am J Phys Med Rehabil. 2010;89:831\u2013839. 57. Man-Son-Hing M, Laupacis A, O\u2019Rourke K, Molnar FJ, Mahon J, Chan KBY, et al. Determination of the clinical importance of study results. J Gen Intern Med. 2002;17:469\u2013476.","254 Castellini et al: Network meta-analysis for chronic neck pain 58. Salanti G, Ades AE, Ioannidis JP. Graphical methods and numerical summaries for 70. Page MJ, Bero L, Kroeger CM, Dai Z, McDonald S, Forbes A, et al. Investigation of Risk presenting results from multiple-treatment meta-analysis: an overview and Of Bias due to Unreported and SelecTively included results in meta-analyses of tutorial. J Clin Epidemiol. 2011;64:163\u2013171. nutrition research: the ROBUST study protocol. F1000 Res. 2019;8:1760. 59. Chaimani A, Higgins JP, Mavridis D, Spyridonos P, Salanti G. Graphical tools for 71. Yordanov Y, Dechartres A, Atal I, Tran VT, Boutron I, Crequit P, et al. Avoidable waste network meta-analysis in STATA. PLoS One. 2013;8:e76654. of research related to outcome planning and reporting in clinical trials. BMC Med. 2018;16:87. 60. White IR. Network meta-analysis. Stata J. 2015;15:951\u2013985. 61. Puhan MA, Schunemann HJ, Murad MH, Brignardello-Petersen R, Singh JA, 72. Mavridis D, White IR, Higgins JP, Cipriani A, Salanti G. Allowing for uncertainty due to missing continuous outcome data in pairwise and network meta-analysis. Stat Kessels AG, et al. A GRADE Working Group approach for rating the quality Med. 2015;34:721\u2013741. of treatment effect estimates from network meta-analysis. BMJ. 2014;349: g5630. 73. Bagg MK, Salanti G, McAuley JH. Comparing interventions with network meta- 62. Yepes-Nunez JJ, Li SA, Guyatt G, Jack SM, Brozek JL, Beyene J, et al. Development of analysis. J Physiother. 2018;64:128\u2013132. the Grading of Recommendations Assessment, Development and Evaluation (GRADE) Summary of Findings (SoF) Table for Network Meta-analysis. J Clin Epi- 74. Wilhelmus Strijkers RH, Schreijenberg M, Gerger H, Koes BW, Chiarotto A. Effec- demiol. 2019. tiveness of placebo interventions for patients with non-speci\ufb01c low back pain: a 63. Sch\u00fcnemann H, Broz_ ek J, Guyatt G, Oxman A, editors. GRADE handbook for grading systematic review. Pain. 2021;162:2792\u20132804. quality of evidence and strength of recommendations. Updated October 2013. The GRADE Working Group, 2013. Available from guidelinedevelopment.org\/ 75. May S, Runge N, Aina A. Centralization and directional preference: An updated handbook. systematic review with synthesis of previous evidence. Musculoskelet Sci Pract. 64. Salanti G, Del Giovane C, Chaimani A, Caldwell DM, Higgins JP. Evaluating the 2018;38:53\u201362. quality of evidence from a network meta-analysis. PLoS One. 2014;9:e99682. 65. Nikolakopoulou A, Higgins JPT, Papakonstantinou T, Chaimani A, Del Giovane C, 76. MacPherson H, Vertosick EA, Foster NE, Lewith G, Linde K, Sherman KJ, et al. The Egger M, et al. CINeMA: An approach for assessing con\ufb01dence in the results of a persistence of the effects of acupuncture after a course of treatment: a meta- network meta-analysis. PLoS Med. 2020;17:e1003082. analysis of patients with chronic pain. Pain. 2017;158:784\u2013793. 66. Corp N, Mansell G, Stynes S, Wynne-Jones G, Mors\u00f8 L, Hill JC, et al. Evidence-based treatment recommendations for neck and low back pain across Europe: A sys- 77. Hoffmann TC, Glasziou PP, Boutron I, Milne R, Perera R, Moher D, et al. Better tematic review of guidelines. Eur J Pain. 2021;25:275\u2013295. reporting of interventions: template for intervention description and replication 67. Kanters S, Ford N, Druyts E, Thorlund K, Mills EJ, Bansback N. Use of network meta- (TIDieR) checklist and guide. BMJ. 2014;348:g1687. analysis in clinical guidelines. Bull World Health Organ. 2016;94:782\u2013784. 68. de Zoete RM, McAuley JH, Arm\ufb01eld NR, Sterling M. The comparative effec- 78. Spineli LM, Kalyvas C, Papadimitropoulou K. Continuous(ly) missing outcome data tiveness of physical exercise interventions in individuals with chronic non- in network meta-analysis: A one-stage pattern-mixture model approach. Stat speci\ufb01c neck pain: protocol for a network meta-analysis. BMJ Open. Methods Med Res. 2021;30:958\u2013975. 2020;10:e034846. 69. Deeks JJ, Higgins JPT, Altman DG (editors). Chapter 9: Analysing data and under- 79. Chiocchia V, Nikolakopoulou A, Higgins JPT, Page MJ, Papakonstantinou T, taking meta-analyses. In: Higgins JPT, Green S (editors). Cochrane Handbook for Cipriani A, et al. ROB-MEN: a tool to assess risk of bias due to missing evidence in Systematic Reviews of Interventions Version 5.1.0 (updated March 2011). The network meta-analysis. BMC Med. 2021;19:304. Cochrane Collaboration, 2011. Available from www.handbook.cochrane.org. Accessed December 12, 2020. 80. Schulz KF, Altman DG, Moher D, Group C. CONSORT 2010 statement: updated guidelines for reporting parallel group randomised trials. BMJ. 2010;340:c332. 81. Chiarotto A, Boers M, Deyo RA, Buchbinder R, Corbin TP, Costa LO, et al. Core outcome measurement instruments for clinical trials in nonspeci\ufb01c low back pain. Pain. 2018;159:481\u2013495. 82. Chiarotto A, Deyo RA, Terwee CB, Boers M, Buchbinder R, Corbin TP, et al. Core outcome domains for clinical trials in non-speci\ufb01c low back pain. Eur Spine J. 2015;24:1127\u20131142."]