BackgroundDupuytren’s disease is a slowly progressive condition of the hand, characterised by the formation of nodules in the palm that gradually develop into fibrotic cords. Contracture of the cords produces deformities of the fingers. Surgery is recommended for moderate and severe contractures, but complications and/or recurrences are frequent. Collagenase clostridium histolyticum (CCH) has been developed as a minimally invasive alternative to surgery for some patients.ObjectivesTo assess the clinical effectiveness and cost-effectiveness of collagenase as an alternative to surgery for adults with Dupuytren’s contracture with a palpable cord.Data sourcesWe searched all major electronic databases from 1990 to February 2014.Review methodsRandomised controlled trials (RCTs), non-randomised comparative studies and observational studies involving collagenase and/or surgical interventions were considered. Two reviewers independently extracted data and assessed risk of bias of included studies. A de novo Markov model was developed to assess cost-effectiveness of collagenase, percutaneous needle fasciotomy (PNF) and limited fasciectomy (LF). Results were reported as incremental cost per quality-adjusted life-year (QALY) gained. Deterministic and probabilistic sensitivity analyses were undertaken to investigate model and parameter uncertainty.ResultsFive RCTs comparing collagenase with placebo (493 participants), three RCTs comparing surgical techniques (334 participants), two non-randomised studies comparing collagenase and surgery (105 participants), five non-randomised comparative studies assessing various surgical procedures (3571 participants) and 15 collagenase case series (3154 participants) were included. Meta-analyses of RCTs assessing CCH versus placebo were performed. Joints randomised to collagenase were more likely to achieve clinical success. Collagenase-treated participants experienced significant reduction in contracture and an increased range of motion compared with placebo-treated participants. Participants treated with collagenase also experienced significantly more adverse events, most of which were mild or moderate. Four serious adverse events were observed in the collagenase group: two tendon ruptures, one pulley rupture and one complex regional pain syndrome. Two tendon ruptures were also reported in two collagenase case series. Non-randomised studies comparing collagenase with surgery produced variable results and were at high risk of bias. Serious adverse events across surgery studies were low. Recurrence rates ranged from 0% (at 90 days) to 100% (at 8 years) for collagenase and from 0% (at 2.7 years for fasciectomy) to 85% (at 5 years for PNF) for surgery. The results of the de novo economic analysis show that PNF was the cheapest treatment option, whereas LF generated the greatest QALY gains. Collagenase was more costly and generated fewer QALYs compared with LF. LF was £1199 more costly and generated an additional 0.11 QALYs in comparison with PNF. The incremental cost-effectiveness ratio was £10,871 per QALY gained. Two subgroup analyses were conducted for a population of patients with moderate and severe disease and up to two joints affected. In both subgroup analyses, collagenase remained dominated.LimitationsThe main limitation of the review was the lack of head-to-head RCTs comparing collagenase with surgery and the limited evidence base for estimating the effects of specific surgical procedures (fasciectomy and PNF). Substantial differences across studies further limited the comparability of available evidence. The economic model was derived from a naive indirect comparison and was hindered by a lack of suitable data. In addition, there was considerable uncertainty about the appropriateness of many assumptions and parameters used in the model.ConclusionsCollagenase was significantly better than placebo. There was no evidence that collagenase was clinically better or worse than surgical treatments. LF was the most cost-effective choice to treat moderate to severe contractures, whereas collagenase was not. However, the results of the cost–utility analysis are based on a naive indirect comparison of clinical effectiveness, and a RCT is required to confirm or refute these findings.Study registrationThis study is registered as PROSPERO CRD42013006248.FundingThe National Institute for Health Research Health Technology Assessment programme.