Physiotherapy Rehabilitation Post First Lumbar

SPINE Volume 36, Number 14, pp E961–E972 ©2011, Lippincott Williams & Wilkins

LITERATURE REVIEW

Physiotherapy Rehabilitation Post First Lumbar Discectomy A Systematic Review and Meta-Analysis of Randomized Controlled Trials Alison Rushton, EdD,* Chris Wright, BSc,* Peter Goodwin, PhD,† Melanie Calvert, PhD,* and Nick Freemantle, PhD,‡

Study Design. Systematic review and meta-analysis. Objective. To evaluate effectiveness of physiotherapy intervention in patients post first lumbar discectomy on clinically relevant outcomes short (3 months) and longer term (12 months). Summary of Background Data. Physiotherapy intervention is recommended post discectomy, although the most beneficial intervention and the effectiveness of physiotherapy management is unclear. Methods. Randomized Controlled Trials (RCTs) published in English before December 31, 2009 investigating physiotherapy outpatient management of patients (>16 years), post first single level lumbar discectomy were included. Measurements reported on ≥1 outcome of disability, function, and health were included. Two reviewers independently searched information sources, assessed studies for inclusion, and evaluated risk of bias. Quantitative synthesis was conducted on comparable outcomes across studies with similar interventions and no clearly identified overall risk of bias. Results. Sixteen RCTs (1336 participants) from 11 countries were included. Interventions were categorized as intervention versus control/sham, and less versus more intensive comparisons. Eight of 16 trials were evaluated as high risk of bias, 7 as unclear and 1 as low. Six hundred and thirty-five participants were incorporated in the meta-analysis on eight trials. Although evidence from two trials suggested that intervention might reduce disability short-term, From the *School of Health and Population Sciences, College of Medical and Dental Sciences, University of Birmingham, Edgbaston, Birmingham B15 2TT, UK; †Department of Professional Registration, Faculty of Health, Psychology & Social Care, Manchester Metropolitan University, Elizabeth Gaskell Campus, Manchester M13 0JA, UK; and ‡Department of Primary Care and Population Health Upper Third Floor UCL Medical School (Royal Free Campus) Rowland Hill Street London NW3 2PF. Acknowledgment date: May 13, 2010. First revision date: June 22, 2010. Acceptance date: July 8, 2010. The manuscript submitted does not contain information about medical device(s)/drug(s). No funds were received in support of this work. No benefits in any form have been or will be received from a commercial party related directly or indirectly to the subject of this manuscript. Address correspondence and reprint requests to Alison Rushton, School of Health and Population Sciences, College of Medical and Dental Sciences, 52 Pritchatts Road, University of Birmingham, Edgbaston, Birmingham B15 2TT, UK; E-mail: [email protected] DOI: 10.1097/BRS.0b013e3181f0e8f8

and more intensive intervention may be more beneficial than less intensive, the pooled effects (−0.89, 95% CI −1.84 to 0.06 for intervention vs. control/sham; −0.27, 95% CI −0.80 to 0.25 for more vs. less intensive) did not show statistically significant effects. There was no evidence that intervention changes range of movement flexion (ROM) or overall impairment short term, or disability or back pain longer term. There was no evidence that intensity of intervention affects back pain short or longer term, ROM short term, or patients’ satisfaction with outcome longer term. Substantial heterogeneity was evident. Conclusion. Inconclusive evidence exists for the effectiveness of outpatient physiotherapy post first lumbar discectomy. Best practice remains unclear. Key words: lumbar discectomy, spinal surgery, physiotherapy, rehabilitation, systematic review, meta-analysis. Spine 2011;36: E961–E972

RATIONALE Surgery is the largest single component of UK National Health Service expenditure in managing back problems, with 9677 lumbar discectomies performed in 2008/2009 for a primary indication of leg pain.1 This contributes to an international problem estimating 10,000 to 11,000 operations in the Netherlands annually,2 and surgery rates in the United States twice that of other developed countries.3 There is wide variation (60%–90%) in the reported success for lumbar discectomy,4–7 leaving up to 40% patients experiencing residual problems, and 3% to 12% seeking further surgery.8 Differences in surgical management could account for some variation, but with estimates of only 70% patients fit to return to work 12 months post surgery,9 and 30% to 70% continuing to experience pain,10 postoperative rehabilitation is a key issue. A UK audit of surgeons and a survey of current physiotherapy practice at spinal surgery centers identified significant variability in routine treatment and advice in outpatient care. McGregor et al11 found advice on functional restrictions after discharge varied considerably within and between surgeons, for example, 31% (n = 16) advised “no sitting” ranging between 2 and 42 days. Fifty-five percent (n = 28) of surgeons referred patients routinely for physiotherapy after discharge. www.spinejournal.com E961

Spine

Copyright © 2011 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

BRS204057.indd E961

26/05/11 12:48 AM

LITERATURE REVIEW

Physiotherapy Rehabilitation Post First Lumbar Discectomy • Rushton et al

Forty-four percent (n = 33) of centers had access to outpatient physiotherapy routinely12 and 46% (n = 35) “if required,” with a range of 1 to 20 treatment sessions. The content of physiotherapy management was variable and the most common interventions included: mobility exercises (53% centers), core stability (47%), education/advice (38%), general rehabilitation (37%), and treatment according to assessment findings (32%). Increasing evidence to support physical management of low back and leg pain13 is emerging. A recent Cochrane systematic review,14 evaluated the effects of active rehabilitation after first lumbar discectomy, including functional rehabilitation exercise programs, but excluded interventions aimed at improving pain relief, strength, and mobility. Ostelo et al14 found low-quality evidence (three trials, n = 156, WMD −11.13, 95% CI −18.44 to −3.82, 0–100 Visual Analogue Scale) supporting exercise as more effective than no treatment for pain, and moderate evidence (two trials, n = 136, WND −6.50, 95%CI −9.26 to −3.74, 0–50 modified Oswestry index) supporting exercise for improved functional status, both in the short term. They found low-quality evidence (two trials, n = 103, WMD 10.67, 95%CI −17.04 to −4.30) supporting high-intensity exercise being slightly more effective than low intensity for pain, and moderate evidence (two trials, n = 103, SMD −0.77, 95% CI −1.17 to −0.36) supporting high-intensity exercises for improved functional status, both in the short term. The findings supported previous conclusions encouraging patients to return to activity and work as soon as they felt able, with no increase in complication rates.15,16 The type of exercise and effectiveness of other physiotherapy interventions most beneficial to this population remains unclear. In particular, the effectiveness of individualized physiotherapy reflective of current practice in countries including the UK12 is unclear owing to Ostelo et al’s inclusion of inpatient trials and non-physiotherapist interventions, and their exclusion of physical interventions aimed at improving pain relief, strength, and mobility. It is therefore important to review the more recent and broader evidence for physiotherapy intervention in this population.

(±laminectomy, no fusion). Studies not written in English were excluded rather than restricting the inclusion of studies, thereby providing information of potential bias.21 No restrictions were placed on publication date.

Objectives

Search

To investigate the short- and longer-term effectiveness of physiotherapy outpatient management in terms of disability, function, and health17 in patients aged more than 16 years.

Predefined terms were used. Table 1 details an example of searches used: the Medline OvidSP search.

MATERIALS AND METHODS

Information sources were searched independently by two subject experts (AR/PG), who independently assessed identified studies for inclusion, facilitated by grading each criterion (Table 2) as eligible/not eligible/might be eligible.18 A study was considered potentially relevant and its full text reviewed when, after discussion between the two independent reviewers, it could not be unequivocally excluded on the basis of its Title and Abstract.21 For abstracts with insufficient information or in a situation of disagreement, full text was obtained. A study was included when both reviewers independently assessed it as satisfying the prespecified criteria from the full text. A third

A systematic review was conducted according to a protocol following method guidelines by the Back Review Group of the Cochrane Collaboration,18 and Cochrane handbook.19 It is reported in line with the PRISMA statement.20

Eligibility Criteria Studies RCTs evaluating the effectiveness of physiotherapy outpatient management of patients after first time lumbar discectomy

Participants Patients who had undergone first time lumbar discectomy, with no complications or additional surgery, aged more than 16 years. Interventions Any physiotherapy outpatient management intervention. Outcome measures Measurements were reported on ≥1 outcome of disability, function, and health,17 short term (approximately 3 months post surgery/intervention) and/or longer term (12 months).

Information Sources Sensitive topic based search strategies designed for each database were employed (to end of December 2009): • The Cochrane Library: Controlled Trials Register, Health Technology Assessment Database, NHS Economic Evaluation Database • CINAHL, EMBASE, MEDLINE, PEDro, ZETOC databases • Selected Internet sites and Indexes: Turning Research into Practice, Health Services/Technology Assessment, PUBMED • National Research Register, Current Controlled Trials website (York) • Cochrane Back Review Group • Hand searches key journals • Science Citation Index and Social Science Citation Index • Unpublished research21: British National Bibliography for Report Literature, Dissertation Abstracts, Index to Scientific and Technical Proceedings, National Technical Information Service, System for Information on Gray Literature.

Study Selection

E962 www.spinejournal.com

June 2011


BRS204057.indd E962

26/05/11 12:48 AM

LITERATURE REVIEW TABLE 1. Example of Medline OvidSP Search

Strategy

Searches 1

TABLE 2. Criteria for Inclusion and Exclusion of

Studies in the Review

Results microdiscectomy.mp. (mp = title, original title, abstract, name of substance word, subject heading word, unique identifier)

2


physiotherapy.mp. (mp = title, original title, abstract, name of substance word, subject heading word, unique identifier)

3

physical therapy.mp. (mp = title, original title, abstract, name of substance word, subject heading word, unique identifier)

4

1 and 3

5

lumbar discectomy.mp. (mp = title, original title, abstract, name of substance word, subject heading word, unique identifier)

Criteria Inclusion criteria Study design

RCT

Population Age

16 yr or older

Subjects

Human; male or female; outpatients

Condition

Post first time single level lumbar discectomy

Intervention

Conservative physiotherapy outpatient management

Comparison group(s)

At least one comparison group, either placebo/other intervention/no intervention

Outcome

Measurement on at least one of the following outcomes: disability; functional status; physical impairment; impact on social and occupational levels of fitness; pain; quality of life

6

Exercise.mp. (mp = title, original title, abstract, name of substance word, subject heading word, unique identifier)

7

post lumbar surgery.mp. (mp = title, original title, abstract, name of substance word, subject heading word, unique identifier)

8

randomized controlled trial.mp. (mp = title, original title, abstract, name of substance word, subject heading word, unique identifier)

9

1 and 2

10

5 and 8

11

1 and 6 (physical approach or therapy or intervention).af.

Participant characteristics

• Multiple pathology

12 13

7 and 12

Intervention

None

14

5 and 12

Outcome

None

15

home exercises.mp. (mp = title, original title, abstract, name of substance word, subject heading word, unique identifier)

Language

Full article/report not written in English

16

5 and 15

17

7 and 15

18

1 and 15

19

diskectomy.mp. (mp = title, original title, abstract, name of substance word, subject heading word, unique identifier)

20

15 and 19

reviewer (CW, methodologic expert) mediated in the event of disagreement following discussion.18 The same reviewers independently assessed the risk of bias for each included study. A study was categorized as potentially relevant for inclusion in the meta-analysis when there was agreement that a low or unclear risk of bias existed, or no high risk of bias in findings relating to the key outcomes of interest. In line with Cochrane,19 homogeneity of participants, interventions, and outcomes was a further consideration, to

Measurement of short term outcome (approx 3 mo post surgery) and/or longer term outcomes (≥1 yr post surgery) Exclusion criteria Study design

Initial search: • Studies stated as RCTs but do not have a comparison group or random allocation to groups

ensure meaningfulness of findings from a clinical perspective. The third reviewer again mediated.18 Agreement between reviewers was evaluated using Cohen κ.22 The process and tools were piloted.

Data Collection Process Using a standardized form, two reviewers (CW/AR) independently extracted the data.21,23 A third reviewer (XX) checked for consistency and clarity.

Data Items Design, participants and indication, interventions and study setting, outcome measures, and timing of assessments were extracted for each included study. Key outcome measures were predefined as valid tools to measure disability, pain, function, physical impairment, social impact, and patient satisfaction, reflecting domains from the International Classification of Functioning, Disability and Health.17 www.spinejournal.com E963

Spine


BRS204057.indd E963

26/05/11 12:48 AM

LITERATURE REVIEW


Risk of Bias in Individual Studies

Additional Analyses

The internal validity of each included study was appraised using the Cochrane ‘Risk of bias’ assessment tool24 that is informed by empirical research,24 in contrast to the majority of quality scales used in health research.25–27 Each component was reported independently and considered with regard to each key outcome measure.24,28 The component of ‘blinding’ the treating therapist has been acknowledged as generally impossible24 and this formed part of the appraisal by the reviewers as the Cochrane tool permits evaluation of the likely influence of any lack of blinding.

A series of post hoc supportive analyses were conducted. These comprised syntheses for intensive versus less intensive rehabilitation, short and longer term, on disability and pain scores from nonspecific locations, and short-term scores on overall physical impairment. Post hoc analyses were also conducted on pain scores from nonspecific locations for intensive versus control rehabilitation.

Summary Measures

Study Selection

Quantitative synthesis was conducted according to the protocol on comparable key outcomes across studies that had similar interventions (nature and intensity of intervention, and timing of assessments at approximately 3 months and/or 12 months post surgery or intervention) and no clearly identified overall risk of bias. Comparable outcomes were defined as tools developed to measure the same underlying domain. One subject expert and two methodologic experts identified the combinations of studies and outcomes on which to conduct meta-analyses. Meta-analyses compared standardized differences in means using DerSimonian–Laird random effects29 for the principal analyses to allow for systematic differences in effects estimated across the included trials,21,29 and Hedges–Olkin fixed effects30 as the supportive analyses; using StatsDirect software.31 A comparison of odds ratios was performed for outcomes reported on a dichotomized scale. The 95% confidence intervals were reported for summary statistics. Standardized mean differences were selected to make comparisons across studies that used different tools to measure the same outcome,21 or reported a mixture of final value scores and change from baseline scores.

In total, 20 articles and 16 trials34–49 from 11 countries were included. For two trials, other articles retrieved presented further data to the original trial and were considered as part of the main trial. This included one further article for Hakkinen et al50; and three further articles for Ostelo et al.51–53 Most retrieved trials were published in English with only two in another language. No relevant unpublished studies were found. Figure 1 presents the numbers of studies at each stage of selection. Substantial (>0.61) inter-reviewer agreement was achieved on study inclusion (Cohen κ 0.64, P = 0.001).22 Less than perfect agreement arose from differences in interpretation of wording on key aspects in the reports of the studies; for example, clarity of random allocation.

RESULTS

Study Characteristics Descriptive data for the 16 included trials are summarized in Table 3 (see Supplemental Digital Content 1, http://links.lww.com/BRS/A516).

Planned Methods of Analysis Analyses were conducted on final summary statistics when reported or the raw data where supplied. When necessary, standard deviations were estimated from reported confidence intervals or percentiles.32 In-line with the use of random effects as primary analyses,29 change scores were used for studies when no other data were forthcoming. Ordinal patient satisfaction scales were dichotomized. Heterogeneity in treatment effects was considered by comparing findings from random and fixed effects models, and computation of I2 or Cochrane Q as appropriate. Analysis of the quality of the interventions was central to interpretation of heterogeneity.21,33

Risk of Bias Across Studies Potential publication bias could not be assessed visually using Funnel plots21 because of the retrieval of too few studies reporting comparable outcome measures. A summary assessment for risk of bias was tabulated across studies, and consensus agreed concerning the overall potential risk of bias.

Figure 1. Study selection flow diagram (from Moher et al.20).

E964 www.spinejournal.com

June 2011


BRS204057.indd E964

26/05/11 12:48 AM

LITERATURE REVIEW


TABLE 4. Summary Assessment of the Overall Risk of Bias for Each Trial Components of Risk of Bias Trial (Authors, Year, Country)

Summary

Decision to Include or Exclude Trial from Meta-analysis

1

2

3

4

5a

5b

6

Alaranta et al (1986) (Finland)

U

U

U

U

N/A

U

U

Unclear (6)

Excluded: No high-risk components. All components unclear. Primary outcome not a key outcome of interest.

Choi et al (2005) (Korea)

U

U

U

H

U

U

H

High (2) Unclear (5)

Excluded: Two high-risk components. Some subjects excluded owing to increased pain.

Danielson et al (2000) (Norway)

L

L

L

L

U

U

H

High (1) Unclear (2) Low (4)

Excluded: One high-risk component. Problematic randomization, group allocation n = 39 and 24 and significant age difference across groups at baseline.

Dolan et al (2000) (UK)

U

U

U

U

U

U

H

High (1) Unclear (6)

Excluded: One high-risk component. Preliminary findings on first 20 subjects reported, hence power limited. Gender differences between groups.

Donaldson et al (2006) (New Zealand)

L

L

L

U

U

U

U

Unclear (4) Low (3)

Included: No high-risk components. Authors provided requested raw data.

Erdogmus et al (2007) (Austria)

L

L

L

L

U

U

U

Unclear (3) Low (4)

Included: No high-risk component. Authors provided requested raw data.

Filiz et al (2005) (Turkey)

L

L

L

L

U

U

L

Unclear (2) Low (5)

Included: No high-risk components. No response from authors re request for data.

Hakkinen et al (2003, 2005) (Finland)

U

U

L

H

U

U

H


Excluded: Two high-risk components with reasons potentially linked to outcomes: large drop out and differences between groups at baseline (duration LBP and leg pain).

Johannsen et al (1994) (Denmark)

L

U

U

H

U

U

L


Excluded: One high-risk component: large drop out with reasons possibly linked to outcome.

Johansson et al (2009) (Sweden)

L

L

L

L

U

U

L

Unclear (2) Low (5)

Included: No high-risk component. Authors provided summary statistics.

Kjellby-Wendt et al (1998) (Sweden)

L

U

L

H

U

U

L


Included: One high-risk component are drop out. However, rate