Osteoarthritis (OA) of the knee is a common chronic

Transcription

1 ORIGINAL ARTICLE A Systematic Review and Meta-Analysis of Medical Leech Therapy for Osteoarthritis of the Knee Romy Lauche, PhD, Holger Cramer, PhD, Jost Langhorst, MD, and Gustav Dobos, MD Objectives: Osteoarthritis of the knee is a common chronic disease among older adults. Therapeutic approaches mainly consist of physiotherapy or pharmacological therapy, but these approaches are limited over time by their cost and/or side effects. This paper presents a systematic review and meta-analysis of the effectiveness of medical leech therapy for osteoarthritis of the knee. Methods: The PubMed/MEDLINE, Cochrane Library, EMBASE, Scopus, and CAMBASE databases were screened in August 2012 to identify randomized (RCTs) and nonrandomized controlled clinical trials (CCTs) comparing leech therapy to control conditions. Main outcome measures were pain, functional impairment, and joint stiffness. For each outcome, standardized mean differences (SMD) and 95% confidence intervals were calculated. Results: Three RCTs and 1 CCT were found, in which a total of 237 patients with osteoarthritis were included. Three trials had a low risk of bias. There was strong overall evidence for immediate (SMD = 1.05; P < 0.01) and short-term pain reduction (SMD = 1.00; P < 0.01), immediate improvement in patients physical function (SMD = 0.72; P < 0.01), and both immediate (SMD = 0.88; P = 0.04) and long-term improvement in their joint stiffness (SMD = 0.62; P < 0.01). Moderate evidence was found for leech therapy s short-term effects on physical function (SMD = 0.46; P < 0.01) and long-term effects on pain (SMD = 0.45; P < 0.01). Leech therapy was not associated with any serious adverse events. Discussion: This systematic review found moderate to strong evidence for the reduction of pain, functional impairment, and joint stiffness after medical leech therapy in patients with osteoarthritis of the knee. Given the low number of reported adverse events, leech therapy may be a useful approach in treating this condition. Further high-quality RCTs are required for the conclusive judgment of its effectiveness and safety. Received for publication August 1, 2012; revised November 22, 2012; accepted December 20, From the Department of Internal and Integrative Medicine, Kliniken Essen-Mitte, Faculty of Medicine, University of Duisburg-Essen, Essen, Germany. They were not, however, involved in the extraction or analysis of the reviewed data or the risk of bias assessment. R.L. was responsible for conceiving and designing the review, performing the literature search, extracting and analyzing the data, conducting the risk of bias assessment, and drafting the manuscript. H.C. was responsible for conceiving and designing the review, performing the literature search, extracting and analyzing the data, conducting the risk of bias assessment, and critically revising the manuscript. J.L. and G.D. participated in conceiving and designing the study, reviewing trials, as well as critically revising the manuscript; it is important to note that 2 authors (J.L. and G.D.) were involved in the reviewed trials. All authors approved the final version of the manuscript. The authors declare no conflict of interest. Supported by a grant from the Rut- and Klaus-Bahlsen-Foundation, Hannover, Germany. This funding source had no influence on the review s planning, organization, management, or publication. Reprints: Romy Lauche, PhD, Kliniken Essen-Mitte, Klinik fu r Naturheilkunde und Integrative Medizin, Am Deimelsberg 34a, Essen, Germany ( r.lauche@kliniken-essen-mitte.de). Copyright r 2013 by Lippincott Williams & Wilkins Key Words: osteoarthritis of the knee, leech therapy, complementary therapies, meta-analysis, systematic review (Clin J Pain 2014;30:63 72) Osteoarthritis (OA) of the knee is a common chronic disease among older adults. In people aged over 55 years, about a quarter report a significant episode of knee pain in the last year and half experience mild to moderate disability. 1 OA results from the destruction of articular cartilage, with underlying bone changes at the joint margins. 2 The resulting pain and stiffness impair function in patients basic activities of daily living, in turn, severely affecting their quality of life. 3 Therapeutic approaches to treating OA, besides joint replacement, rely mainly on physiotherapy and/or pharmacological therapy 4 ; however, cost and the side effects of antiphlogistics over the long term makes their use limited. In light of the above, interest has revived in the use of a traditional complementary treatment, namely leech therapy. Leeches (Hirudo medicinalis) have been used therapeutically for many years, 5 but their use has declined as surgery and pharmacotherapy have advanced. In developed countries, leech therapy has mainly been used for treating phlebitis and thrombosis in recent years, as the hirudin in leech saliva has anticoagulant properties. 6 Leech saliva also contains anti-inflammatory and anesthetic substances, 7 although the latter are yet to be identified. In the last decade, clinical trials have evaluated the value of leech therapy in such conditions as OA of the knee 8 12 and lateral epicondylitis. 13 If proven effective, leech therapy might provide a useful alternative approach for patients with OA whose pain excludes them from physical exercise, as well as those who decline to undergo joint replacement. Leech therapy is less invasive than surgery and has the additional advantage of being low cost. A thorough literature search found no meta-analyses of leech therapy for OA of the knee. The current review sought to remedy this omission, assessing the literature systematically to produce such a meta-analysis. Randomized controlled trials (RCTs) and controlled clinical trials (CCTs), which compared leech therapy to other control conditions were included. MATERIALS AND METHODS Protocol and Registration This review was planned and conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines 14 and the recommendations of the Cochrane Back Review Group. 15,16 The protocol is not registered on any database. Clin J Pain Volume 30, Number 1, January

2 Lauche et al Clin J Pain Volume 30, Number 1, January 2014 Eligibility Criteria To be eligible for review, studies were required to meet the following conditions: (1) Types of study designs: RCTs and nonrandomized CCTs were eligible. (2) Types of participants: Studies of adults (18 y and older) with OA of the knee were eligible, regardless of the condition s duration or intensity, as long as patients diagnoses were radiologically confirmed. No further restrictions were applied. (3) Types of interventions: Studies that compared leech therapy with no treatment, usual care, placebo, or any active treatment were eligible. No restrictions were applied about the details of the leech therapy. Although studies with cointerventions were included, these studies were excluded in the subsequent sensitivity analyses. (4) Types of outcome measures: Studies were eligible for inclusion if they assessed at least 1 important patientcentred outcome, namely pain, functional impairment, or joint stiffness. If available, studies data on quality of life and safety issues served as secondary outcome measures. (5) Length of follow-up: Studies were eligible if they assessed study outcomes at least once, either within 7 days (immediate effects), within 1 month (short-term effects), or at 6 months (long-term effects) posttreatment. (6) Accessibility of data: Studies were eligible only if they were published as full papers. No restrictions on language or publication date were applied. Literature Search The following electronic databases were searched from their inception through to July 31, 2012: PubMed/MED- LINE, EMBASE, Scopus, the Cochrane Library, and CAMBASE. The literature search, which was constructed around search terms for leech and osteoarthritis, was adapted for each database as necessary. For example, the following search strategy was used on the MEDLINE database: (leeches[mesh Terms] OR leeching[mesh Terms] OR leech therapy[mesh Terms] OR hirudo medicinalis[mesh Terms] OR leech[title/abstract] OR hirudo medicinalis[title/abstract] OR leech therapy[title/ Abstract]) AND (osteoarthritis[mesh Terms] OR arthritis[mesh Terms] OR arthralgia[mesh Terms] OR joint disease[mesh Terms] OR osteoarthritis[title/abstract] OR osteoarthrosis[title/abstract] OR gonarthrosis[title/ Abstract] OR degenerative arthritis[title/abstract] OR arthritis[title/abstract]). The reference lists of identified original articles or reviews were also searched manually for relevant articles. Study Selection All duplicate papers were first removed. Two reviewers then screened the abstracts of remaining papers independently. They went on to obtain the full papers for all potentially eligible studies. These papers were then checked for eligibility, with eligible papers being included in the systematic review. Papers that provided data on relevant clinical outcomes were also included in the meta-analysis. Data Collection Two reviewers independently extracted data on studies characteristics (participants, interventions, control conditions, cointerventions, outcome measures, and results). Disagreements were submitted to a third reviewer and resolved by discussion. Where data could not be extracted from the original published papers, their authors were contacted. This step was undertaken as data on leech therapy are sparse and thus valuable in judging the treatment s effectiveness. Additional caution was thought advisable as leech therapy is seen by some as controversial. Outcomes Eligible treatment outcomes were defined as follows: (1) Pain intensity measured on a visual analogue scale or on the Western Ontario and McMaster Universities Arthritis Index (WOMAC) pain scale. (2) Functional disability was included where this was measured on a visual analogue scale, on the WOMAC physical function scale, or on another validated physical function index suitable for patients with OA. (3) Joint stiffness was included where this was measured on a visual analogue scale or on the WOMAC stiffness scale. (4) Quality of life was included where it was assessed using a validated generic measure, for example, the SF-36, or a validated specific measure, for example, the AIMS-2. (5) Patient safety outcomes were defined as any adverse event occurring during a study. Risk of Bias in Individual Studies The risk of bias was assessed, at study level, by 2 independent reviewers using the 2009 Updated Method Guidelines for Systematic Reviews in the Cochrane Back Review Group. 16 These guidelines recommend the imposition of 12 quality criteria, each of which is rated as YES, NO, or UNCLEAR. These criteria relate to the following risk of bias domains: selection bias, performance bias, attrition bias, reporting bias, and detection bias. If study data were considered inconclusive, then the original study authors were contacted for further details. Studies that scored positively on at least 6 of the 12 criteria, and had no serious flaws, were rated as having low risk of bias. Studies that met fewer than 6 criteria and/or showed a serious flaw were rated as having a high risk of bias. 16 This classification, as a low-risk or high-risk study, served as the basis for the sensitivity analyses conducted. Data Analysis Studies were analyzed separately for their immediate, short-term, and long-term effects. Immediate outcomes were defined as those from measures applied up to 7 days posttreatment, short-term outcomes came from measures applied closest to 28 days posttreatment, and long-term outcomes from measures applied closest to 6 months postrandomization. 16 Assessment of Overall Effect Size If at least 2 studies presented data on a specific outcome, then meta-analyses were undertaken using Review Manager 5 software (Version 5.1, The Nordic Cochrane Centre, Copenhagen). Where studies had >1 control group, the control groups for the assessment of the study s overall effect were selected in the following order of preference: no treatment, placebo, and active comparator. One study had 2 active treatment groups. 11 Although the randomization procedure used to divide the study population into treatment and control was adequate in this study, the treatment groups were further split into 2 according to participants symptom severity. As this rendered the treatment and control groups incomparable, data 64 r 2013 Lippincott Williams & Wilkins

3 Clin J Pain Volume 30, Number 1, January 2014 Leech Therapy for Knee Osteoarthritis from the 2 treatment groups were condensed. In order to merge the data, the reviewers contacted the trial authors for access to the original data. Standardized mean differences (SMD) with 95% confidence intervals (CI) were calculated as the mean group difference divided by the pooled SD. 17 Where no SDs were available, SEs, CIs, or t values were used to calculate them. A random effect model was applied. The magnitude of the overall effect size was judged according to Cohen categories: small effect size: SMD = 0.2 to 0.5; moderate effect size: SMD = 0.5 to 0.8; and large effect size: SMD > A negative SMD was defined to indicate the beneficial effects of leech therapy, as compared with the control interventions, for pain and functional impairment. For quality of life, a positive SMD corresponded to enhanced well-being. If necessary, patients scores were inverted and the mean score was subtracted from the instruments maximum score. 15,19 Levels of evidence were judged according to the recommendations of the Cochrane Back Review Group as follows: (1) strong evidence: consistent findings among multiple RCTs with a low risk of bias; (2) moderate evidence: consistent findings among multiple high-risk RCTs or CCTs and/or 1 low-risk RCT; (3) limited evidence: 1 RCT with a high risk of bias; (4) conflicting evidence: inconsistent findings among multiple RCTs and CCTs; (5) no evidence: no RCTs. 20,21 Assessment of Heterogeneity Statistical heterogeneity between the reviewed studies was quantified by determination of I 2. I 2 > 30%, I 2 > 50%, and I 2 > 75% were defined to indicate moderate, substantial, and considerable heterogeneity, respectively. 15 A P value r0.10 from the w 2 test was taken to indicate significant heterogeneity. 15 Subgroup and Sensitivity Analyses Where at least 2 studies existed in each subgroup, subgroup analyses were conducted for the type of control treatment delivered (ie, no treatment, usual care, placebo, or any active treatment). Sensitivity analyses to test the robustness of any significant results were conducted by removing studies based on the following methodological quality criteria: absence or inadequate randomization and/ or quality score <6. 17 In this meta-analysis, the following studies were removed in the sensitivity analysis: 1 CCT. 9 If statistical heterogeneity was present in the respective metaanalysis, the subgroup and sensitivity analyses were also used to explore the nature of this heterogeneity. Risk of Bias Across Studies If at least 10 studies were included in a meta-analysis, the risk of publication bias was assessed by visual analysis of funnel plots generated by Review Manager 5.1 software. Roughly symmetrical funnel plots indicate a low risk of publication bias, whereas asymmetrical funnel plots indicate a high risk of such bias. 22 RESULTS Study Selection The literature search retrieved 121 records, of which 55 were duplicates (Fig. 1). Of the remainder, 62 records either referred to studies with no controls (N = 60) or did not investigate OA of the knee (N = 2). The remaining 4 fulltext articles were assessed for eligibility 9 12 (Fig. 1). One FIGURE 1. Flowchart for the results of the literature search. r 2013 Lippincott Williams & Wilkins 65

4 Lauche et al Clin J Pain Volume 30, Number 1, January 2014 TABLE 1. Characteristics of the Included Studies References Type of Study Sample Size, No. Group Mean age (± SD) Inclusion Criteria Michalsen et al 9 CCT 16, ± 9.0 (treatment); 68.0 ± 8.0 (control) Primary osteoarthritis of the knee >6 mo No secondary osteoarthritis due to rheumatic disease No injection in the prior 3 mo Michalsen RCT 51, ± 10.2 (treatment); 65.5 ± 6.7 (control) Osteoarthritis of the knee as defined by the American College of Rheumatology Pain >40 mmvas No secondary osteoarthritis due to rheumatic disease No arthroscopy or surgery No injection in the prior 3 mo Andereya RCT 118, ± 10.0 Osteoarthritis of the knee at least 6 mo Kjellgran/ Lawrence stage III-IV No rheumatic diseases No surgery No injection in the prior 3 mo Treatment Group: Intervention Leeching, single application of 4 leeches (maximally painful periarticular sites) +14 to 16d inpatient integrative treatment program (health education, exercise, physiotherapy, behavioral, and relaxation techniques) Leeching, single application of 4-6 leeches (maximally painful points) Group 1: Leeching, single application of 4 leeches (proximal, medial, and lateral to the patella) Group 2: Leeching, 2 applications of 4 leeches (proximal, medial, and lateral to the patella) at days 0 and 28 Control group: Intervention Program Length, Duration, Frequency 14 to 16 d inpatient integrative treatment program (health education, exercise, physiotherapy, behavioral, and relaxation techniques) Topical diclofenac application twice daily for 28 d Sham leeching, single application of 1 sham leech Outcome Assessment a) Immediate Follow-up b) Short-term Follow-up c) Long-Term Follow-up a) 7 d b) 28 d a) 7 d b) 28 d c) 91 d a) 7 d b) 28 d c) 6 mo Outcome Measures 1) Pain 2) Function 3) Stiffness 4) Quality of life 5) Safety 1) VAS 2) NA 5) Safety 1) WOMAC pain score 2) WOMAC function score 3) WOMAC stiffness score 4) SF-36 5) Safety 1) WOMAC pain score; VAS; KOOS pain 2) WOMAC function scale; KOOS ADL 3) WOMAC stiffness scale 5) Safety Results* a) Immediate follow-up b) Short-term follow-up c) Long-term follow-up 1) Pain 2) Function 3) Stiffness 4) Quality of life 5) Safety a) 1) Sign. inner-group change 2) NA b) 1) Sign. inner-group change 2) NA 5) No serious adverse events a) 1) Treatment > control 2) Treatment > control 3) Treatment > control b) 1) NS 2) Treatment > control 3) Treatment > control 4) Treatment > control 5) NA c) 1) NS 2) Treatment > control 3) Treatment > control 4) NS 5) No serious adverse events a) 1) WOMAC pain: Group 1 > control; group ; VAS: Group 1 vs. control NS; group ; KOOS pain: Group 1 > control; group 2) WOMAC function: Group 1 vs. control NS; group ; KOOS ADL: Group 1 vs. control NS; group 3) Group 1 vs. control NS; group 66 r 2013 Lippincott Williams & Wilkins

5 Clin J Pain Volume 30, Number 1, January 2014 Leech Therapy for Knee Osteoarthritis Stange et al 12 RCT (crossover design) 52, ± 10.2 Osteoarthritis of 1 or both knees with stadium II according to Jaeger and Wirth No inflammatory diseases Leeching, single application of 8 leeches (above, below, medially, and laterally the patella) Transcutaneous electrical nerve stimulation, single application of 2 electrodes (above and below the patella) a) 7 d b) 21 d c) 42 d 1) VAS 2) Lequesne s index 5) Safety b) 1) WOMAC pain: Group 1 > control; group ; VAS: Group 1 vs. control NS; group ; KOOS pain: Group 1 > control; group 2) WOMAC function: Group 1 vs. control NS; group ; KOOS ADL: Group 1 vs. control NS; group 3) Group 1 vs. control NS; group c) 1) WOMAC pain: Group 1 > control; group ; VAS: Group 1 vs. control NS; group ; KOOS pain: Group 1 > control; group 2) WOMAC function: Group 1 vs. control NS; group ; KOOS ADL: Group 1 vs. control NS; group 3) Group 1 vs. control NS; group 5) No serious adverse events a) 1) NR 2) NR b) 1) Sign. inner-group change 2) Sign. inner-group change c) 1) Sign. inner-group change 2) Sign. inner-group change 5) Acute worsening (n = 1) KOOS indicates knee injury and osteoarthritis outcome score; LI, Lequesne-index of severity for osteoarthritis; NA, not assessed; NR, not reported; NS, not significant; SF-36, Short Form Health Survey; VAS, visual analogue scale; WOMAC, Western Ontario and McMaster Universities Osteoarthritis Index. * > indicates significantly better than ; sign, significant. r 2013 Lippincott Williams & Wilkins 67

6 Lauche et al Clin J Pain Volume 30, Number 1, January 2014 TABLE 2. Risk of Bias Assessment of the Included Studies Using the Cochrane Risk of Bias Tool Bias Selection Bias Performance Bias References Adequate Random Sequence Generation Adequate Allocation Concealment Similar Baseline Characteristics Adequate Participant blinding Adequate Provider Blinding Similar or no cointerventions Michalsen et al 9 No No Yes No No Yes* Michalsen Yes Yes Yes No No Yes Andereya Yes Yes Yes No No Yes Stange et al 12 Yes Yes Yes No No Yes* *Additional details provided upon request. was a nonrandomized clinical trial 9 and the other 3 were RCTs Of the latter, 1 study used a cross-over design. For this trial, only the first phase of this study was included. 12 In another study, the intervention group was divided into 2, with a single or double treatment application. In this study, both groups were condensed into 1 treatment group (see the Methods section). 11 One CCT and 3 RCTs, with a total of 237 patients, were included in the qualitative and quantitative analyses Study Characteristics The characteristics of the study samples, interventions, outcome measures, and results are shown in Table 1. Setting and Participant Characteristics All trials originated from Germany Patients were either inpatients in an integrative clinic, 10 outpatients, 12 or community-dwelling adults recruited through newspaper advertisements. 9,11 Two studies included people with mild to moderate OA 10,12 and 1 study patients with moderate to severe OA. 11 One study included patients who were primarily treated for back pain, but had concomitant OA. 9 Two of the studies explicitly excluded patients who previously had surgery for their OA 10,11 and all did so for patients with rheumatic diseases On average, patients were in their mid 60s and most were female (56%; range, 28% to 82%). No data on ethnicity were available. Intervention Characteristics Leech therapy was administered once in 3 studies 9,10,12 and once or twice in 1 study. 11 The data from the latter 2 intervention groups were condensed, as noted previously. In the single study that used a cross-over design, 12 only the data gathered before the therapy was changed over were entered into the meta-analysis. The studies control interventions differed widely. In 1 study, control patients received a 14-day course of integrative inpatient treatment, as did the intervention group. 9 In another study, leech therapy was compared with 28 days application of diclofenac gel. 10 A sham procedure was used as a control by Andereya and transcutaneous electrical nerve stimulation by Stange et al. 12 Patients received cointerventions in all 4 studies 9 12 ; however, according to the original papers authors, these interventions did not differ between the study groups. Outcome Measures Pain was assessed as an outcome measure in all 4 studies, with 3 measuring pain intensity on a visual analogue scale 9,11,12 and 1 using the WOMAC pain scale. 10 Three studies also measured functional impairment, with 2 using the WOMAC function scale 10,11 and the other the Lequesne Index. 12 Joint stiffness was measured in 2 studies using the WOMAC stiffness scale 10,11 and quality of life in 1 study. 10 Data on patient safety were reported in all 4 studies Immediate (7 d) and short-term (28 d) effects were evaluated in all studies, but long-term effects (6 mo) were only reported in 2 studies. 10,11 Risk of Bias in Individual Studies Three studies had a low risk of bias One study, although reaching a quality score of 6 points, was seriously flawed by its lack of randomization, and was thus considered at high risk 9 (Table 2). Studies risk of selection bias was mainly low, with only 1 having a high risk 9 and 3 a low risk The risk of performance bias was generally high, mostly due to a lack of blinding for the patients and care providers participating. One study 11 tried to blind patients by using a placebo leech, but this process proved unsuccessful. The studies risk of attrition bias was mixed, with only 1 study including an intention-to-treat analysis. 10 The risk of reporting bias was low, but the risk of detection bias was high, as no study reported adequate blinding of the outcome assessors. Analyses of Overall Effects Pain: Meta-analyses revealed strong evidence for leech therapy s immediate (SMD = 1.05; 95% CI, 1.76 to 0.34; P < 0.01; heterogeneity: I 2 = 80%; w 2 = 14.87; P < 0.01) and short-term effects (SMD = 1.00; 95%, CI 1.72 to 0.28; P < 0.01; heterogeneity: I 2 = 80%; w 2 = 14.88; P < 0.01) on patients pain, and moderate evidence for its long-term effects (SMD = 0.45; 95% CI, 0.78 to 0.13; P < 0.01; heterogeneity: I 2 = 0%; w 2 = 0.51; P = 0.48), compared with controls (Table 3). Functional impairment: Meta-analyses revealed strong evidence for the immediate effect of leech therapy on functional impairment (SMD = 0.72; 95% CI, 1.25 to 0.19; P < 0.01; heterogeneity: I 2 = 69%; w 2 = 6.4; P = 0.04) and moderate evidence for its short-term effects (SMD = 0.46; 95% CI, 0.73 to 0.18; P < 0.01; heterogeneity: I 2 = 0%; w 2 = 1.93; P = 0.38). No evidence was found for its long-term effects (SMD = 0.39; 95% CI, 68 r 2013 Lippincott Williams & Wilkins

7 Clin J Pain Volume 30, Number 1, January 2014 Leech Therapy for Knee Osteoarthritis Performance Bias Attrition Bias Reporting Bias Detection Bias Acceptable Compliance Acceptable and Described Drop-out Rate Inclusion of an Intention-to-Treat Analysis No Selective Outcome Reporting Adequate Outcome Assessor Blinding Similar Timing of Total Risk: Outcome Assessment (Max. 12) Yes* Yes* No Yes No Yes 6 Yes Yes Yes Yes No Yes 9 Yes Yes No Yes No Yes 8 Yes Yes No* Yes No Yes to 0.03; P = 0.07; heterogeneity: I 2 = 37%; w 2 = 1.58; P = 0.21) compared with controls (Table 3). Joint stiffness: There was strong evidence for the immediate (SMD = 0.88; 95% CI, to 0.04; P = 0.04; heterogeneity: I 2 = 81%; w 2 = 5.31; P = 0.02) and long-term effects (SMD = 0.62; 95% CI, 0.95 to 0.29; P < 0.01; heterogeneity: I 2 = 0%; w 2 = 0.38; P = 0.54) of leech therapy on joint stiffness compared with controls (Table 3). No evidence was found for its shortterm effects (SMD = 0.78; 95% CI, 1.72 to 0.17; P = 0.11; heterogeneity: I 2 = 85%; w 2 = 6.89; P < 0.01). Based on Cohen categories, leech therapy s immediate and short-term effects on patients pain were considered large, but the long-term effects were small. The therapy s immediate effects on functional impairment were moderate, but its short-term and long-term effects were small. Its immediate effects on joint stiffness were large, but its longterm effects were moderate. All 4 studies reported the existence of adverse events These were, however, characterized as mild to moderate, relating to local skin irritation, itching, burning, or oozing. No serious adverse events were observed or reported. Quality of life: Quality of life was cited in only 1 study, 10 hence no meta-analysis was undertaken. Subgroup Analyses No subgroup analyses were conducted due to insufficient trials in each subgroup. Sensitivity Analyses After the exclusion of the high-risk CCT, 9 leech therapy s immediate and short-term effects on patients pain remained (immediate: SMD = 0.74; 95% CI, 1.26 to 0.21; P < 0.01; heterogeneity: I 2 = 68%; w 2 = 6.34; P < 0.01; short-term: SMD = 0.64; 95% CI, 0.92 to 0.36; P < 0.01; heterogeneity: I 2 = 92%; w 2 = 1.64; P = 0.04). These effect sizes were moderate. No further sensitivity analyses were undertaken as the single CCT 9 reviewed did not include any other outcome measure. Risk of Bias Across Studies As <10 studies were included in each meta-analysis, funnel plots were not analyzed. DISCUSSION Summary of Main Results This meta-analysis found strong evidence that leech therapy reduced patients OA knee pain both immediately (within a week) and in the short-term (within a month), to reduce their functional impairment immediately and to reduce their joint stiffness, both immediately and in the long-term (after 6 mo). Moderate evidence was also found for the long-term reduction of pain and the short-term reduction of functional impairment. The available safety data did not link leech therapy to any serious adverse events. Patients withdrawal rates were low in the reviewed studies, suggesting that leech therapy was accepted well. Applicability of Evidence The reviewed trials were all conducted in primary, secondary, or tertiary care settings in Germany. Most participants (56%; range, 28% to 82%) were female adults in their mid 60s. The studies included patients at different stages of OA, from mild to severe. Although 2 studies included people who previously had surgery for their OA, all excluded people with rheumatic disease. Given the breadth of these inclusion criteria, this review s results potentially apply to many people with OA, at least in Germany. Leech therapy has a history of international use for the treatment of venous congestion in plastic and reconstructive surgery. 23 In Asia, it is also used for pain management. 24 This review suggests that leech therapy has much to offer in relieving OA pain. It is hoped that this meta-analysis will foster a broader recognition, in developed countries, of this link. Quality of Evidence Although the methodological quality of the 4 studies reviewed differed somewhat, leech therapy s effects on pain and functional impairment were robust against methodological bias. After excluding the single CCT found, these effects remained. Agreements and Disagreements With Other Systematic Reviews A thorough literature search found no other systematic review of leech therapy for OA of the knee. By comparison, a meta-analysis of the effectiveness of nonsteroidal anti-inflammatory drugs found an effect size of d = 0.4 (Cohen), a moderate effect. In addition, 2 weeks of treatment with nonsteroidal anti-inflammatory drugs proved no better than treatment with a placebo. 25 The current metaanalysis showed that leech therapy has moderate to large effect sizes, which persist at long-term (6 mo) follow-up, making it a potentially valuable treatment alternative. r 2013 Lippincott Williams & Wilkins 69

8 Lauche et al Clin J Pain Volume 30, Number 1, January 2014 TABLE 3. Effect Sizes of Leech Therapy on Selected Study Outcomes Outcome Timeframe References Leech Mean Leech SD N Control Mean Control SD N Weight SMD [95% Confidence Interval] % 0.52 [ 0.90, 0.13] % 0.61 [ 1.01, 0.22] % 0.37 [ 0.77, 0.02] Pain Immediate Andereya Michalsen % 3.07 [ 4.67, 1.48] et al. 9 Michalsen % 1.37 [ 2.00, 0.75] Stange et al % 0.42 [ 0.97, 0.13] Total % 1.05 [ 1.76, 0.34] Overall effect Test for overall effect: Z = 2.89 (P = 0.004) Heterogeneity Heterogeneity: w 2 = 14.87, df =3 (P = 0.002); I 2 = 80% Short-term Andereya et al. 11 Michalsen % 4.50 [ 6.56, 2.44] et al. 9 Michalsen % 0.94 [ 1.53, 0.35] Stange et al % 0.42 [-0.97, 0.13] Total % 1.00 [ 1.72, 0.28] Overall effect Test for overall effect: Z = 2.73 (P = 0.006) Heterogeneity Heterogeneity: w 2 = 14.88, df =3 (P = 0.002); I 2 = 80% Long-term Andereya Michalsen % 0.62 [ 1.19, 0.06] Total % 0.45 [ 0.78, 0.13] Overall effect Test for overall effect: Z = 2.74 (P = 0.006) Heterogeneity Heterogeneity: w 2 = 0.51, df =1 (P = 0.48); I 2 =0% Physical Function Immediate Short-term Long-term 51.8% 0.29 [ 0.68, 0.09] 60.9% 0.22 [ 0.62, 0.18] Andereya % 0.43 [ 0.82, 0.04] Michalsen % 1.34 [ 1.96, 0.73] Stange et al % 0.49 [ 1.04, 0.06] Total % 0.72 [ 1.25, 0.19] Overall effect Test for overall effect: Z = 2.65 (P = 0.008) Heterogeneity Heterogeneity: w 2 = 6.40, df =2 (P = 0.04); I 2 = 69% Andereya Michalsen % 0.78 [ 1.36, 0.21] Stange et al % 0.49 [ 1.04, 0.06] Total % 0.46 [ 0.73, 0.18] Overall effect Test for overall effect: Z = 3.23 (P = 0.001) Heterogeneity Heterogeneity: w 2 = 1.93, df =2 (P = 0.38); I 2 =0% Andereya Michalsen % 0.67 [ 1.24, 0.10] Total % 0.39 [ 0.82, 0.03] Overall effect Test for overall effect: Z = 1.81 (P = 0.07) Heterogeneity Heterogeneity: w 2 = 1.58, df =1 (P = 0.21); I 2 = 37% % 0.48 [ 0.87, 0.10] % 0.32 [ 0.70, 0.06] Stiffness Immediate Andereya Michalsen % 1.34 [ 1.96, 0.72] Total % 0.88 [ 1.72, 0.04] Overall effect Test for overall effect: Z = 2.05 (P = 0.04) Heterogeneity Heterogeneity: w 2 = 5.31, df =1 (P = 0.02); I 2 = 81% Short-term Andereya Michalsen % 1.29 [ 1.91, 0.68] Total % 0.78 [ 1.72, 0.17] Overall effect Test for overall effect: Z = 1.60 (P = 0.11) Heterogeneity Heterogeneity: w 2 = 6.89, df =1 (P = 0.009); I 2 = 85% (Continued) 70 r 2013 Lippincott Williams & Wilkins

9 Clin J Pain Volume 30, Number 1, January 2014 Leech Therapy for Knee Osteoarthritis TABLE 3. (continued) Outcome Timeframe References Long-term Leech Mean Leech SD N Control Mean Control SD N Weight SMD [95% Confidence Interval] Andereya % 0.55 [ 0.95, 0.15] Michalsen % 0.77 [ 1.35, 0.20] Total % 0.62 [ 0.95, 0.29] Overall effect Test for overall effect: Z = 3.71 (P = ) Heterogeneity Heterogeneity: w 2 = 0.38, df =1 (P = 0.54); I 2 =0% Strengths and Weaknesses This is the first review and meta-analysis of leech therapy for OA of the knee. Its results are strengthened by its use of patient-centred outcomes, as recommended by the Osteoarthritis Research Society task force. 26 The review s primary limitation is the paucity of eligible trials identified, which rendered subgroup analyses impossible. More studies comparing leech therapy to other guideline-endorsed therapies and/or placebo are urgently needed. The current lack of trials and patients may lead to a misperception of leech therapy s side effects. Serious side effects, such as sustained hemorrhage, bacterial infection, or allergic reaction might be rare, but might severely impact patients health. Three of the included studies had a low risk of bias, leading to little risk that the meta-analyses, themselves, were biased. Although statistical heterogeneity was mostly seen, the paucity of studies reviewed meant that sensitivity analyses could not account for this. Heterogeneity therefore remains a limiting factor, at least for all immediate and short-term future comparisons. Another limiting factor is the short follow-up timeframe. Although 2 studies followed patients for 6 months, only 1 did so beyond this. The Osteoarthritis Research Society s guidance, about the need for longer follow-up timeframes, 26,27 should be taken in future studies. CONCLUSIONS This systematic review found moderate to strong evidence for the reduction of pain, functional impairment, and joint stiffness after medical leech therapy in patients with OA of the knee. Given the low number of reported adverse events, leech therapy may be a useful approach in treating this condition. Further high-quality RCTs are required for the conclusive judgment of its effectiveness and safety. ACKNOWLEDGMENT The authors thank Dr Sue Rugg, PhD, Universities of Exeter and Plymouth, Peninsula College of Medicine and Denistry, Plymouth, UK; for language editing. REFERENCES 1. Peat G, McCarney R, Croft R. Knee pain and osteoarthritis in older adults: a review of community burden and current use of primary health care. Ann Rheum Dis. 2001;60: Altman R, Asch E, Bloch D, et al. Development of criteria for the classification and reporting of osteoarthritis. Classification of osteoarthritis of the knee. Diagnostic and Therapeutic Criteria Committee of the American Rheumatism Association. Arth Rheum. 1986;29: Reginster JY. The prevalence and burden of arthritis. Rheumatology. 2002;41(suppl.1): Pendleton A, Arden N, Dougados M, et al. EULAR recommendations for the management of knee osteoarthritis: report of a task force of the standing committee for international clinical studies including therapeutic trials (ESCISIT). Ann Rheum Dis. 2000;59: Fields WS. The history of leeching and hirudin. Haemostasis. 1991;21(suppl 1): Markwardt F. Hirudin as alternative anticoagulant a historical review. Sem Thromb Haemost. 2002;28: Michalsen A, Roth M, Dobos G. Medicinal Leech Therapy. Stuttgart: Thieme; Michalsen A, Deuse U, Esch T, et al. Effect of leeches therapy (Hirudo medicinalis) in painful osteoarthritis of the knee: a pilot study. Ann Rheum Dis. 2001;60: Michalsen A, Moebus S, Spahn G, et al. Leech therapy for symptomatic treatment of knee osteoarthritis: results and implications of a pilot study. Altern Ther Health Med. 2002; 8: Michalsen A, Klotz S, Lu dtke R, et al. Effectiveness of leech therapy in osteoarthritis of the knee: a randomized, controlled trial. Ann Intern Med. 2003;139: Andereya S, Stanzel S, Maus U, et al. Assessment of leech therapy for knee osteoarthritis: a randomized study. Acta Orthop. 2008;79: Stange R, Moser C, Hopfenmueller W, et al. Randomised controlled trial with medical leeches for osteoarthritis of the knee. Compl Ther Med. 2012;20: Ba cker M, Lu dtke R, Afra D, et al. Effectiveness of leech therapy in chronic lateral epicondylitis: a randomized controlled trial. Clin J Pain. 2011;27: Moher D, Liberati A, Teztlaff J, et al. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Ann Intern Med. 2009;51: Higgins JPT, Green S. Cochrane handbook for systematic reviews of interventions. Version The Cochrane Collaboration, Available at: org. Assessed February 09, Furlan AD, Pennick V, Bombardier C, et al updated method guidelines for systematic reviews in the Cochrane Back Review Group. Spine. 2009;34: Langhorst J, Klose P, Musial F, et al. Efficacy of acupuncture in fibromyalgia syndrome a systematic review with a meta-analysis of controlled clinical trials. Rheumatology. 2010;49: Cohen J. Statistical Power Analysis for the Behavioral Sciences. Hillsdale, NJ: Lawrence Erlbaum Associates; Cramer H, Lauche R, Paul A, et al. Mindfulness-based stress reduction for breast cancer a systematic review and metaanalysis. Curr Oncol. 2012;19:e343 e van Tulder M, Furlan A, Bombardier C, et al. Updated method guidelines for systematic reviews in the Cochrane Collaboration Back Review Group. Spine. 2003;28: r 2013 Lippincott Williams & Wilkins 71

10 Lauche et al Clin J Pain Volume 30, Number 1, January Langhorst J, Musial F, Klose P, et al. Efficacy of hydrotherapy in fibromyalgia syndrome a meta-analysis of randomized controlled clinical trials. Rheumatology. 2009;48: Egger M, Davey Smith G, Schneider M, et al. Bias in metaanalysis detected by a simple, graphical test. BMJ. 1997; 315: Whitaker IS, Oboumarzouk O, Rozen WM, et al. The efficacy of medicinal leeches in plastic and reconstructive surgery: a systematic review of 277 reported clinical cases. Microsurgery. 2012;32: Kalender ME, Comez G, Sevinc A, et al. Leech therapy for symptomatic relief of cancer pain. Pain Med. 2010;11: Lin J, Zhang W, Jones A, et al. Efficacy of topical nonsteroidal anti-inflammatory drugs in the treatment of osteoarthritis: meta-analysis of randomised controlled trials. BMJ. 2004;329: Altman R, Brandt K, Hochberg M, et al. Design and conduct of clinical trials in patients with osteoarthritis: recommendations from a task force of the Osteoarthritis Research Society. Results from a workshop. Osteoarthritis cartilage. 1996;4: Hochberg MC, Altman RD, Brandt KD, et al. Design and conduct of clinical trials in osteoarthritis: preliminary recommendations from a task force of the Osteoarthritis Research Society. J Rheumatol. 1997;24: r 2013 Lippincott Williams & Wilkins