This review found that the short-term clinical outcomes of anterior cruciate ligament reconstruction with allograft versus autograft did not significantly differ. The poor quality of the evidence from few patients implies that the reliability of the authors’ conclusions is unclear.

To compare the short-term clinical outcomes of autograft versus allograft anterior cruciate ligament reconstruction.

MEDLINE (from 1950), and EMBASE (from 1966), were searched to March 2009 for publications in any language; search terms were reported. Unpublished studies were not considered. The bibliography of each retrieved article was handsearched.

Studies with either a prospective or retrospective design were eligible for inclusion if they had a level of evidence of at least I, II, or III according to criteria described by Wright, et al. (2003, see Other Publications of Related Interest) and they compared any clinically relevant short-term clinical outcome of autograft versus allograft unilateral anterior cruciate ligament reconstruction. Eligible studies had to use the same anatomical graft, have at least 15 patients in each arm, and follow-up all patients for at least two years.

The included studies were set in the USA or Europe. Where reported, the mean patient age ranged from 22 to 46 years and the percentage of male patients ranged from 42 to 80. Most of the included studies compared autograft and allograft bone-patellar tendon-bone reconstruction; one study compared autograft and allograft quadruple-stranded hamstring reconstruction. In most of the studies allografts were not irradiated; in one study they were irradiated, in second study some were irradiated and some were not, in a third study they were treated with acetone and then irradiated, and in a fourth study irradiation status was not reported. Autografts were fresh-frozen in most of the studies. All studies involved one or two surgeons and the mean follow-up ranged from 24 to 71 months. Outcomes of interest included patient-orientated measures, results of physical examination tests, instrumental laxity measurements, complications, and failures.

The authors did not state how many reviewers performed the selection.

Methodological quality was assessed by two reviewers independently using criteria for study design, randomisation or patient choice, loss to follow-up, similarity of surgical technique, utilisation of independent examiners, and blinding. The levels of evidence criteria described by Wright, et al. were also applied; level I was the highest and included only randomised controlled trials (RCTs), level II included prospective cohort studies and poor-quality RCTs, and level III included retrospective and case-control studies.

The number of events for each outcome was extracted and percentages calculated. Two reviewers independently extracted the data, using worksheets, and any discrepancies were resolved by discussion and consensus.

The results were pooled using a random-effects model (Mantel-Haenszel) to produce odds ratios and 95% confidence intervals. Heterogeneity was determined using the χ² test or by comparing the frequency distribution by group, where heterogeneity was found if p was less than 0.1. One study was excluded from the meta-analysis (and synthesis) as it failed a qualitative assessment and a statistical test of homogeneity. In this study the allografts were treated with acetone and irradiated and the failure rate was 45%. Publication bias was assessed using funnel plots. Sensitivity analyses were performed to allow for factors, such as graft-type and secondary sterilisation technique.

Nine relevant studies were identified (n=746 patients); five were prospective comparative studies (level II; n=464, range 60 to 186) and four were retrospective comparative studies (level III; n=282, range 50 to 90). The overall level of evidence was considered to be level III. Five studies had more than 80% follow-up and two studies had less than 60% follow-up. For most studies, the selection of the treatment type was based on patient choice and/or allograft availability; in one study, randomisation occurred in 75% of patients. Four studies did not involve independent examiners. In five studies, the treatment groups differed at baseline.

There were no significant differences in patient-orientated outcome and composite scales between autograft and allograft (eight studies) and for a pooled analysis of Lysholm knee scores (six studies). There were no significant differences for: Lachman testing (four studies); pivot-shift testing (five studies); flexion deficit (seven studies); one-leg hop test (three studies); thigh circumference (five studies); and instrumental laxity of more than 0.5mm (seven studies). Two of six studies showed a significant benefit for autograft versus allograft for extension deficit.

No significant differences were found in the complications of: anterior knee pain (three studies); patellofemoral pain (one study); retropatellar pain (one study); deep infection rate (four studies); arthrofibrosis (four studies); and reoperation rates (three studies). One study found fewer incision site complaints in the allograft group (significance not reported) and the study, in which allografts were treated with acetone, found a lower rate of kneeling pain and paraesthesia in the allograft group. No significant differences were found in failures (seven studies).

There was no evidence of heterogeneity in the meta-analyses. Sensitivity analyses on the graft type, instrumental laxity cut-off value, secondary sterilisation technique, duration of follow-up, mean patient age, and study methodology did not change the overall results.

The short-term clinical outcomes of anterior cruciate reconstruction using autograft were not significantly different to those using allograft.

This review addressed a well-defined question in terms of participants, interventions, and study design, but included a wide variety of outcomes. Relevant databases were searched for articles in any language, but unpublished studies were not considered and some relevant studies might have been missed. There was little evidence of publication bias. Study quality was assessed using suitable criteria. Efforts were made to reduce error and bias in the data extraction and quality assessment, but a similar process was not reported for study selection. No relevant randomised controlled trials were identified and only a limited number of prospective and retrospective comparative studies, with a relatively low overall number of patients, were included. The relevant study details were reported. Meta-analysis was only possible for a small number of outcomes. The sensitivity analyses did not change the conclusions.

The review had no major flaws, but the poor quality of the evidence from a small number of patients implies that the reliability of the conclusions is unclear.

Practice: The authors suggested that surgeons could incorporate the limited results of this review into the informed-consent and shared-decision-making process to provide the best individual patient care. The method of sterilisation of allografts using acetone drying in addition to irradiation was not recommended.

Research: The authors identified a need for randomised controlled trials, but noted the ethical and practical concerns of using tissue from cadavers and recommended high-quality prospective comparative studies, as the next best option, with the use of multivariable modelling to adjust for possible confounding variables.

National Institutes of Health, and National Institute of Arthritis and Musculoskeletal and Skin Diseases, grant numbers R01 AR053684-01A1 and 5 K23 AR052392-04.

Wright JG, Swiontkowski MF, Heckman JD. Introducing levels of evidence to The Journal. Journal of Bone and Joint Surgery. American volume 2003; 85(1): 1-3.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.