Patellar tendon versus hamstring tendon autograft for anterior cruciate ligament rupture in adults

Nicholas GH Mohtadi; Denise S Chan; Katie N Dainty; Daniel B Whelan

doi:10.1002/14651858.CD005960.pub2

. 2011 Sep 7;2011(9):CD005960. doi: 10.1002/14651858.CD005960.pub2

Patellar tendon versus hamstring tendon autograft for anterior cruciate ligament rupture in adults

Nicholas GH Mohtadi ¹, Denise S Chan ^2,^✉, Katie N Dainty ³, Daniel B Whelan ⁴

PMCID: PMC6465162 PMID: 21901700

Abstract

Background

Reconstruction of the anterior cruciate ligament (ACL) commonly involves patellar tendon (PT) or hamstring tendon(s) (HT) autografts. There is no consensus with respect to the choice between these two grafts in ACL surgery.

Objectives

This review compared the outcomes of ACL reconstruction using PT versus HT autografts in ACL deficient patients.

Search methods

We searched the Cochrane Bone, Joint and Muscle Trauma Group Specialised Register (April 2008), the Cochrane Central Register of Controlled Trials (2008, Issue 2), MEDLINE (1966 to April 10 2008), EMBASE (1980 to April 10 2008), conference proceedings and reference lists. No language restrictions were applied.

Selection criteria

Randomized and quasi‐randomized controlled trials comparing outcomes (minimum two year follow‐up) following ACL reconstruction using either PT or HT autografts in skeletally mature adults, irrespective of the number of bundles, fixation method or incision technique.

Data collection and analysis

After independent study selection, the four authors independently assessed trial quality and risk of bias, and extracted data using pre‐developed forms. Trial authors were contacted for additional data and information. Risk ratios with 95% confidence intervals were calculated for dichotomous outcomes, and mean differences and 95% confidence intervals for continuous outcomes.

Main results

Nineteen trials providing outcome data for 1597 young to middle‐aged adults were included. Many trials were at high risk of bias reflecting inadequate methods of randomization, lack of blinding and incomplete assessment of outcome.

Pooled data for primary outcomes, reported in a minority of trials, showed no statistically significant differences between the two graft choices for functional assessment (single leg hop test), return to activity, Tegner and Lysholm scores, and subjective measures of outcome. There were also no differences found between the two interventions for re‐rupture or International Knee Documentation Committee scores. There were inadequate long‐term results, such as to assess the development of osteoarthritis.

All tests (instrumental, Lachman, pivot shift) for static stability consistently showed that PT reconstruction resulted in a more statically stable knee compared with HT reconstruction. Conversely, patients experienced more anterior knee problems, especially with kneeling, after PT reconstruction. PT reconstructions resulted in a statistically significant loss of extension range of motion and a trend towards loss of knee extension strength. HT reconstructions demonstrated a trend towards loss of flexion range of motion and a statistically significant loss of knee flexion strength. The clinical importance of the above range of motion losses is unclear.

Authors' conclusions

There is insufficient evidence to draw conclusions on differences between the two grafts for long‐term functional outcome. While PT reconstructions are more likely to result in statically stable knees, they are also associated with more anterior knee problems.

Keywords: Adult; Humans; Middle Aged; Young Adult; Anterior Cruciate Ligament Injuries; Anterior Cruciate Ligament; Anterior Cruciate Ligament/surgery; Patellar Ligament; Patellar Ligament/transplantation; Recovery of Function; Rupture; Rupture/surgery; Tendons; Tendons/transplantation; Thigh; Transplantation, Autologous; Treatment Outcome

Patellar or hamstring tendon grafts for ACL reconstruction in adults

The anterior cruciate ligament (ACL) is important for maintaining the stability in the knee, particularly in activities involving cutting, pivoting or kicking. People with ruptured ACLs have unstable knees that generally become more damaged over time. Reconstruction of ruptured ACLs commonly involves using autografts (grafts taken from the person undergoing surgery), obtained by removing part of the patellar tendon or the hamstring tendon. This review aimed to find out if one graft was better than the other.

This review included 19 studies reporting the outcomes of ACL reconstruction with patellar tendon versus hamstring tendon grafts in a total of 1597 young to middle‐aged adults. Many trials used flawed methods that might have affected their results.

The limited data available for functional outcomes including patient‐rated assessment did not show whether one graft was better than the other. Similarly, there were no differences found between the two types of graft for re‐rupture or in the results of an internationally used knee score. All tests for knee stability favoured patellar tendon grafts. Conversely, people had more anterior knee pain and discomfort with kneeling after patellar tendon reconstruction. After patellar tendon reconstruction, more people had some loss in their ability to straighten out their leg at the knee. In contrast, more people had some loss in their ability to bend their leg at the knee after hamstring tendon reconstruction. It is not clear how important these losses in range of motion of the knee were to the patients themselves.

The review concluded that the current evidence was insufficient to recommend which of the two types of graft was better for ACL reconstruction.

Background

The anterior cruciate ligament (ACL) of the knee acts to maintain joint stability by restraining anterior translation of the tibia relative to the femur (Seitz 1996). It also prevents abnormal rotational motion and varus/valgus angulation at full knee extension. Its role is particularly important in athletes, such as footballers, when performing activities involving cutting (i.e. sudden change in direction), pivoting, and kicking.

Description of the condition

ACL injury is a common orthopaedic problem with an annual incidence of approximately 200,000 cases per year in the United States (Miyasaka 1991). The classic mechanism of ACL injury involves non‐contact deceleration, such as sudden stopping or changes in direction. ACL tears are commonly associated with meniscal and articular cartilage injury (Shelbourne 1991; Smith 2001).

An ACL deficient knee is defined by the absence or loss of function of the ligament, resulting in biomechanical loss of stability. Though the natural history of ACL deficiency is poorly defined, studies have reported that a person with an ACL deficient knee presents with pain, recurrent symptoms of instability (Noyes 1983; Noyes 1989) and the discontinuation or limitation of pre‐injury sporting activities (Barrack 1990). ACL injury predisposes the knee to chronic instability, further meniscal and chondral damage and an impaired quality of life (Jackson 1993; Mohtadi 1998). It may also predispose to osteoarthritis (Daniel 1994; Sherman 1988).

Description of the intervention

Where available and clinically‐indicated, the most commonly recommended treatment for ACL deficiency is a surgical procedure called ACL reconstruction. This involves the use of a tendon graft to reconstruct the torn or deficient ligament. The primary goal of surgery is to achieve a functionally stable knee while minimizing morbidity and complications associated with the procedure. More than 70,000 ACL reconstructions are performed annually in the United States (Lyman 2009). While ACL reconstruction is a clinically accepted intervention, non‐operative management is often indicated for people who are less active, have minimal instability symptoms, and who are unable or unwilling to follow the demanding post‐surgical rehabilitation protocols. It is noteworthy that a Cochrane review comparing surgical versus non‐surgical intervention for ACL injuries found no evidence from randomized controlled trials to inform current practice (Linko 2005). Nevertheless, surgical ACL reconstruction remains the current standard of care.

An ACL reconstruction can use several options for the tendon graft. One option is to use allograft tissue from a cadaver donor (Frank 1997) or, less commonly, an artificial graft (Grontvedt 1996; Mirza 2000). More commonly, surgeons use the patient's own tendon tissue (autograft). The two most common autografts are the patellar tendon (PT) and hamstring tendon(s) (HT) (Johnson 1992). A PT graft involves surgically harvesting the central one third of the tendon with the attached bone from the patella and tibia. The HT graft involves harvesting the tendonous portion of the patient's semitendinosus and/or gracilis muscles.

Possible benefits and risks of patellar tendon and hamstring grafts

Proponents of the PT autograft cite superior graft strength, secure fixation and ease of harvest as advantages over the HT autograft, which is associated with increased graft incorporation time, possible hamstring weakness and inferior fixation. Those favouring the HT autograft cite smaller incisions, decreased donor site morbidity, multi‐bundled structure and larger surface area for incorporation. These proponents often state concern over patellar tendonitis and tendon rupture, patellar fracture and anterior knee pain with use of the PT graft. In the absence of specific contraindications to the use of one graft type over the other, the autograft offering the greatest likelihood of superior outcome and minimal morbidity and complication remains an unresolved issue (Biau 2006; Biau 2007; Grant 2003; Spindler 2004).

Why it is important to do this review

Thirteen recently published reviews (Biau 2006; Biau 2007; Dauty 2005; Forster 2005; Freedman 2003; Goldblatt 2005; Grant 2003; Herrington 2005; Prodromos 2005; Schultz 2002; Spindler 2004; Thompson 2005; Yunes 2001) compare PT and HT autografts in ACL reconstruction. These reviews vary in their methodology, resulting in potentially biased conclusions. For example, three reviews (Goldblatt 2005; Freedman 2003; Yunes 2001) used search strategies limited to MEDLINE and the English language, which may be evidence of a publication bias. In addition, these authors did not restrict their reviews to randomized clinical trials. The most recent publication addressing this topic analyzes the previously published systematic reviews (Poolman 2007). The purpose of this "review of systematic reviews" was to address the discrepancies and contradictory recommendations. Amongst their conclusions were that the existing reviews were of variable quality, sensitivity analyses were inconsistently applied, and that only two reviews (Biau 2006; Dauty 2005) were found to be methodologically sound. A key reason for performing this Cochrane review was to include more recent trials that utilize modern surgical techniques. In addition, the Cochrane Library facilitates a dynamic process for inclusion of future trials.

Our review compares the effectiveness of PT and HT autografts for the reconstruction of the ACL deficient knee using a comprehensive literature search to identify the evidence from recent and internationally published and unpublished randomized controlled trials. Given the controversy and uncertainty over which graft type should be used, our goal was to provide a definitive, unbiased and reproducible systematic review.

Objectives

The aim of this review was to compare the outcomes of patellar tendon versus hamstring tendon autografts for ACL reconstruction in people with ACL deficiency.

Methods

Criteria for considering studies for this review

Types of studies

We included randomized controlled trials (RCTs) and quasi‐randomized controlled trials (for example, allocation by hospital record number or date of birth) that compared PT and HT grafts as stated in the 'Objectives'.

Types of participants

The population of interest was skeletally mature patients with documented ACL deficiency of the knee, requiring ACL reconstruction.

Types of interventions

Arthroscopically assisted ACL reconstruction using either a patellar tendon (PT) or a hamstring tendon (HT) autograft. HT autografts could be double or multiple stranded grafts. Single or double incision techniques were included as well as any method of fixation of the graft. Excluded interventions included allografts, synthetic materials and revision ACL reconstructions.

Types of outcome measures

Specific outcomes were not used to include or exclude trials. The outcomes were reported as follows:

Primary outcomes

1. Functional assessments 2. Return to activity / level of sport participation (including Tegner (Tegner 1985) and Lysholm (Lysholm 1982) scores) 3. Subjective knee scores (patient satisfaction, Cincinnati score (Noyes 1983; Noyes 1989), Anterior Cruciate Ligament Quality of Life (Mohtadi 1998))

Secondary outcomes

1. Complications (i.e. infection, arthrofibrosis), adverse outcomes, recurrent injury with and without reoperation 2. Static stability measures (KT arthrometer or other stability assessment devices) 3. Clinical composite scores; i.e. IKDC (International Knee Documentation Committee) (Irrgang 2001) 4. Range of motion 5. Strength testing (Cybex muscle testing or equivalent) 6. Pain / anterior knee symptomatology

Timing of outcome assessment

Outcome assessment was analyzed based on long term follow‐up, which was defined as greater than two years following ACL reconstruction. Subsequent to the protocol, we stipulated that a minimum two year follow‐up was necessary for trial inclusion. This was to ensure that there was consistency across the trials in the report of functional outcomes.

Search methods for identification of studies

Electronic searches

We searched the Cochrane Bone, Joint and Muscle Trauma Group Specialized Register (April 2008), the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library, 2008 Issue 2), MEDLINE (1966 to April 10 2008), EMBASE (1980 to April 10 2008), and reference lists of articles. There were no constraints based on language or publication status.

In MEDLINE (OVID ONLINE), the first three levels of the optimal trial search strategy (Higgins 2005) were combined with the subject specific search. The complete search strategy is shown in Appendix 1. The search strategies used in the EMBASE (OVID ONLINE) database and The Cochrane Library (Wiley InterScience) are also shown in Appendix 1.

Searching other resources

The bibliographies of all papers identified by the search strategy were handsearched.

Specific proceedings of knee surgery, arthroscopic surgery and sport medicine meetings and conferences (1997 to 2006) were searched from the following organizations: European Society of Sports Traumatology Knee Surgery and Arthroscopy (ESSKA), American Orthopaedic Society for Sports Medicine (AOSSM), International Society of Arthroscopy, Knee Surgery and Orthopaedic Sports Medicine (ISAKOS), American Academy of Orthopaedic Surgeons (AAOS), World Congress on Orthopaedic Sports Trauma, and Arthroscopy Association of North America (AANA). To avoid publication bias, the investigators of the trials identified from the proceedings were contacted to obtain results and data of any unpublished studies.

Data collection and analysis

Selection of studies

All randomized or quasi‐randomized controlled trials directly comparing outcomes following ACL reconstruction using either PT or HT (gracilis and semitendinosus) autografts in adults were considered. Four review authors from two centres (NM and DC ‐ Calgary, DW and KD ‐ Toronto) independently applied the inclusion criteria to select citations in MEDLINE, other databases and reference lists for retrieval of full articles. Where there was disagreement or doubt, the full article was retrieved. Each centre independently assessed each full study report to see if it met the review inclusion criteria.

Data extraction and management

Four review authors (NM, DC, DW and KD) used pre‐developed data‐extraction forms to independently extract the data. The review authors compared the data extracted for each study to achieve consensus between the two centres. Where required, corresponding authors of individual trials were contacted for additional data or clarification of methodology.

Assessment of risk of bias in included studies

Risk of bias tables for the included studies were completed for the six domains of The Cochrane Collaboration's 'Risk of bias' tool: sequence generation, allocation concealment, blinding, completeness of outcome data, selective reporting and other bias. Many of the judgements drew on our ratings for methodological quality, which was assessed previously using the Cochrane Bone, Joint and Muscle Trauma Group's former quality assessment tool (Madhok 2006) (seeAppendix 2), and the Detsky scale (Detsky 1992) (seeAppendix 3), which has been used previously to grade orthopaedic RCTs (Bhandari 2002).

Four review authors (NM, DC, DW and KD) independently scored the methodological quality of included studies. The next level of review involved discussion within each centre and the third level of review was between centres. Consensus agreement was achieved between centres.

Measures of treatment effect

Risk ratios with accompanying 95% confidence intervals were calculated for dichotomous outcomes, and mean differences and 95% confidence intervals for continuous outcomes. In general, unfavourable outcome data are presented for dichotomous outcomes. However, for continuous outcomes such as strength and range of motion where higher values represent a better outcome, the descriptors 'Favours PT' or 'Favours HT' in the analyses were switched to reflect the correct direction of effect.

Dealing with missing data

Trial authors were contacted for missing data. Where standard deviations were not reported for a trial that otherwise could be included in a meta‐analysis, the mean standard deviations for the treatment groups from the other trials were used in an exploratory analysis.

Assessment of heterogeneity

Heterogeneity was evaluated by visual inspection and tested using a chi‐squared test (significance P < 0.10) and the I‐squared statistic (> 50%) (Higgins 2003).

Data synthesis

The results of comparable groups of trials were pooled using a fixed‐effect model. In the presence of heterogeneity, a random‐effects model was used.

Subgroup analysis and investigation of heterogeneity

Subgroup analyses were performed comparing number of strands for HT grafts (< 4 strands versus 4 strands), and the method of hamstring tendon femoral fixation (endobutton versus screws).

In the published protocol of this review, we specified the following subgroups:

1. Duration of ACL deficiency (acute: less than three months; or chronic: greater than three months) 2. Method used in graft fixation (e.g. screw, button, staple) 3. Type of hamstring graft preparation (double or quadruple stranded) 4. Surgical approach used in ACL reconstruction (single or two incisions)

Sensitivity analysis

Sensitivity analyses are a method of investigating the importance of some of the assumptions and decisions made during a systematic review. For this review, we planned sensitivity analyses to explore the effect of methodological quality on the aggregate estimate of treatment effect, assuming that studies of lower quality would produce an exaggerated estimate of treatment effect. We performed two types of sensitivity analyses, one comparing randomized to quasi‐randomized controlled trials and the other examining the effects of using imputed standard deviations.

Results

Description of studies

Results of the search

The electronic search of the databases resulted in a total of 326 references. An additional 39 references were identified by handsearching conference proceedings from the European Society of Sports Traumatology Knee Surgery and Arthroscopy (ESSKA), American Orthopaedic Society for Sports Medicine (AOSSM), International Society of Arthroscopy, Knee Surgery and Orthopaedic Sports Medicine (ISAKOS), American Academy of Orthopaedic Surgeons (AAOS), World Congress on Orthopaedic Sports Trauma, and Arthroscopy Association of North America (AANA). No new references were identified by handsearching the bibliographies of identified studies.

Seventy‐nine of the 365 identified references discussed patellar tendon (PT) and hamstring tendon (HT) autografts for ACL reconstruction in the form of reviews, meta‐analyses, prospective clinical trials, case series, or randomized and quasi‐randomized clinical trials. One Spanish study and three German studies were identified and required translation. After removing duplicate references and selecting only randomized and quasi‐randomized clinical trials, 26 studies were identified. Only 19 of these studies, which were reported in 22 publications, were included in the final analysis. Seven studies, four of which were insufficiently reported in conference abstracts, were excluded for reasons listed in the Characteristics of excluded studies. Of the three trials reported in full, Beard 2001 alone was excluded because of insufficient follow‐up, whereas Carter 1999 and Sato 2005 reported only on intermediate outcomes as well. One ongoing study (Taylor 2006) was not included in the review because minimum two year follow‐up data were not yet available at the cut‐off time for study selection.

Included studies

Nineteen studies were included. All 19 studies reported long‐term (24 months or more) results with an average follow‐up ranging from a minimum of 24 months to 102 months. Additionally, short‐term (< 6 months) follow‐up results were reported in three studies (Aglietti 2004; Feller 2003; Maletis 2007) and intermediate‐term (6 months to < 24 months) results in five studies (Aglietti 2004; Aune 2001; Beynnon 2002; Feller 2003; Maletis 2007). For three studies, the findings were reported in two publications. Aglietti (Aglietti 1994) reported two and five year follow‐up information on the same group of randomized patients (Aglietti 1994; Aglietti 1997). Jansson and Harilainen (Jansson 2003) reported results on the same population of patients at a minimum of 21 months and minimum of three years (median of five years) respectively (Jansson 2003; Harilainen 2006). O'Neill (O'Neill 1996) initially reported results on 125 patients at a mean follow‐up of 42 months in 1996, then on an additional 101 patients (one died) for a total of 225 patients (with minimum six year follow‐up) in 2001 (O'Neill 1996; O'Neill 2001). However, O'Neill's 2001 publication did not report on all of the original outcomes analyzed in the 1996 publication (O'Neill 1996; O'Neill 2001).

One relevant study was translated from the German language to provide information for use in the review (Ropke 2001). One trial, which was identified by handsearching conference proceedings, represented an unpublished manuscript received after contacting the author. This study was subsequently published in 2007 with the information being essentially identical (Maletis 2007). The remaining studies were published between 1994 and 2007.

Design

All included studies used some form of randomization. Eight studies used computer‐generated randomization, random numbers tables, or sealed envelopes for the randomization process (Anderson 2001; Aune 2001; Beynnon 2002; Ejerhed 2003; Eriksson 2001; Feller 2003; Laxdal 2005; Maletis 2007). The remaining 11 trials used a quasi‐randomization method for treatment allocation, including alternating sequences (Aglietti 1994; Aglietti 2004; Zaffagnini 2006), birth date (Ibrahim 2005; Jansson 2003; Marder 1991; Matsumoto 2006; O'Neill 1996; Shaieb 2002), surgery date or surgical register sequence (Ropke 2001; Sajovic 2006).

The majority of the studies compared two treatment groups. However, four studies compared three treatment groups with different variations on the reconstructive technique (Anderson 2001; Laxdal 2005; O'Neill 1996; Zaffagnini 2006). Anderson 2001 included a third comparison group that combined an intra‐articular double‐stranded semitendinosus/gracilis hamstring with an extra‐articular reconstruction using the ilio‐tibial band. This group of patients was not included in this review. Laxdal 2005 included two hamstring groups (3‐strand semitendinosus graft and a quadruple semitendinosus/gracilis graft), both of which were included in the review. O'Neill 1996 included two patellar tendon reconstruction groups, which differed only in the incision technique (single or two‐incision); both patellar tendon groups were included in the review. Zaffagnini 2006 also included a second hamstring treatment group with an extra‐articular component to the reconstructive technique; this group was not included in the review.

Sample sizes

The sample sizes for the 19 trials ranged from 40 participants (Ropke 2001) to 229 participants (O'Neill 1996; O'Neill 2001). Sample size was determined a priori in only six trials (Aglietti 2004; Anderson 2001; Aune 2001; Beynnon 2002; Jansson 2003; Maletis 2007). Five of these trials based sample size calculations on KT arthrometer measurements; however, there were significant discrepancies between the definitions for clinically important differences. Aglietti 2004 based their sample size calculations on the KT arthrometer measurement (134 N), with a side to side difference of 1.0 mm and a standard deviation of 1.5 mm. Their calculated sample size was 48 patients per group. They subsequently entered 60 patients per group to account for a 20% loss to follow‐up rate. Based on additional information provided by Anderson 2001, a calculated "sample size of 35 [patients per group] was necessary to determine a difference of 0.8 standard deviations for KT‐1000 values. For nominal data, such as the IKDC score, a power of 0.8 detected a 36% difference in the groups." Therefore, Anderson 2001 enrolled 105 patients into three separate groups; 102 were seen at an average of 35 months follow‐up. As described by Aune 2001, the calculated sample size of 22 patients per group was based on a 3 mm difference in joint laxity on the KT arthrometer (manual maximum) and standard deviation of 3 mm (alpha 0.05, beta 0.01). (However, the beta value claimed by Aune 2001 is unusually small and inconsistent with the small sample size.) Seventy‐two patients were enrolled in Aune 2001 to account for patients lost due to no shows or injuries to the uninvolved knee, with 61 patients analyzed at a minimum of 24 months follow‐up. The sample size estimate reported for Beynnon 2002 was based on the primary outcome of translation, as determined by KT arthrometer measurement at 133 N, with a side to side difference of 2.5 mm and a variation of 4.0 mm. This resulted in a calculated sample size of 24 patients per group. Beynnon 2002 entered 56 patients in total, but only evaluated 22 in each group at final follow‐up. Maletis 2007 also used the KT arthrometer measurement (manual maximum) as their primary outcome measure, but used a clinically important side to side difference of 1.5 mm with a standard deviation of 2.0 mm. They calculated a sample size of 39 patients per group, subsequently entered a total of 99 patients and evaluated 96 patients at a mean follow‐up of 26 months. Jansson 2003 used the Lysholm score as, "one of our main outcome measurements" to calculate sample size and defined a 10‐point difference as clinically significant. They did not report a standard deviation, but stated that a total sample size of 54 for both treatment groups was sufficient. They randomized 99 patients, but evaluated 89 patients at a minimum of 21 months follow‐up. We were unable to replicate all sample size calculations based on the available information. There was no consistency with respect to what was defined as the minimal clinically important difference, standard deviation of the KT arthrometer measurement, and the necessary power to show a difference between groups.

Setting

The included studies were truly international in scope. The 10 countries represented were Australia, Finland, Germany, Italy, Japan, Kuwait, Norway, Slovenia, Sweden and USA. All settings represented referral based practices.

Participants

A total of 1748 patients with a confirmed diagnosis of ACL deficiency were randomized between 1989 and 2003, where reported. Twenty‐five patients from one trial (Ibrahim 2005) were not accounted for and two patients died in O'Neill's trial (O'Neill 1996), but no group designation was provided for these. Data from these patients and a further 124 patients (lost to follow‐up = 89; re‐rupture = 24; contralateral rupture = 9; withdrawal = 1; and previous reconstruction = 1) were unavailable for analysis. This left a total of 1597 participants with data for analysis (seeTable 2). A comparison of the number of analyzed patients in each treatment group showed little difference in the proportions available in each group (seeAnalysis 1.25: risk ratio (RR) 0.99; 95% CI 0.96 to 1.03). Based on the definition of chronic ACL deficiency being greater than three months from time of injury, seven trials (Aglietti 1994; Beynnon 2002; Ibrahim 2005; Jansson 2003; Marder 1991; Matsumoto 2006; Ropke 2001) only included patients with chronic ACL deficiency. Seven studies provided baseline data for numbers of patients with acute ACL deficiency (seeAnalysis 1.26) but did not provide separate follow‐up data for these patients. Therefore, it was not possible to separate out acute reconstructions in the analysis.

Table 1.

Randomized versus analyzed

Study	Randomized	Lost to follow‐up	Re‐ruptures	Contralateral ruptures	Withdrawal	Previous ACL Recons	Analyzed
Aglietti 1994 / Aglietti 1997	63 Total	3 Total	NR	NR	NR	NR	30 PT 30 HT
Aglietti 2004	60 PT 60 HT	0	NR	NR	NR	NR	60 PT 60 HT
Anderson 2001	35 PT 35 HT	0 PT 2 HT	NR	NR	NR	NR	35 PT 33 HT
Aune 2001	35 PT 37 HT	8 Total	1 PT* 2 HT*	3 Total	NR	NR	29 PT 32 HT
Beynnon 2002	28 PT 28 HT	6 PT 6 HT	NR	NR	NR	NR	22 PT 22 HT
Ejerhed 2003	34 PT 37 HT	1 PT 1 HT	1 PT 2 HT	NR	NR	NR	32 PT 34 HT
Eriksson 2001	84 PT 80 HT	2 PT 2 HT	2 PT 3 HT + 1 extra surgery	NR	NR	NR	80 PT 74 HT
Feller 2003	31 PT 34 HT	4 PT 2 HT	1 PT 0 HT	NR	0 PT 1 HT	NR	26 PT 31 HT
Ibrahim 2005	110 total	25	NR	NR	NR	NR	40 PT 45 HT
Jansson 2003 / Harilainen 2006	51 PT 48 HT	13 PT 3 HT	0 PT 4 HT	1 PT 4 HT	NR	NR	37 PT 37 HT
Laxdal 2005	134 Total	9 Total	1 PT 2 HT + 2 early septic failure	1 PT 0 HT	NR	0 PT 1 HT	40 PT 78 HT
Maletis 2007	46 PT 53 HT	0 PT 2 HT	0 PT 1 HT	2 PT* 1 HT*	NR	NR	46 PT 50 HT
Marder 1991	40 PT 40 HT	3 PT 5 HT	1 PT* 1 HT*	NR	NR	2 PT* 0 HT*	37 PT 35 HT
Matsumoto 2006	40 PT 40 HT	3 PT 5 HT	NR	NR	NR	NR	37 PT 35 HT
O'Neill 1996 / O'Neill 2001	229 Total	2 PT 0 HT 2 Deaths	2 PT* 2 HT*	NR	NR	NR	150 PT 75 HT
Ropke 2001	20 PT 20 HT	NR	NR	NR	NR	NR	20 PT 20 HT
Sajovic 2006	32 PT 32 HT	2 PT 1HT	1 PT 1HT	3 PT 2 HT	NR	NR	26 PT 28 HT
Shaieb 2002	82 Total	12 Total	2 PT 2 HT	NR	NR	NR	31 PT 35 HT
Zaffagnini 2006	25 PT 25 HT	0	0	NR	NR	NR	25 PT 25 HT
TOTAL	1748	0* 89 + 2 deaths (excluded) 25 unaccounted (excluded)	9* 24 (excluded)	3* 9 (excluded)	0* 1 (excluded)	2* 1 (excluded)	1597
NR = None reported * = included in analysis; all other patients for lost to follow‐up, re‐ruptures, contralateral ruptures and withdrawals were excluded from the analysis

	Items (defined in Table 2)
Study ID	A	B	C	D	E	F	G	H	I	J	K	L
Aglietti 1994	0	0	0	2	0	0	2	2	2	2	2	2
Aglietti 2004	0	0	2	2	0	0	2	2	2	2	2	2
Anderson 2001	1	1	0	2	0	0	2	2	2	2	2	2
Aune 2001	2	1	0	2	0	0	2	1	2	2	2	2
Beynnon 2002	1	0	0	1	0	0	2	2	2	2	2	2
Ejerhed 2003	2	1	0	2	0	0	2	0	2	2	2	1
Eriksson 2001	1	1	0	1	0	0	2	2	2	2	1	1
Feller 2003	1	1	0	2	0	0	2	1	2	2	2	2
Ibrahim 2005	0	0	0	1	0	0	2	0	2	1	1	1
Jansson 2003	0	1	0	1	0	0	2	1	2	2	1	1
Laxdal 2005	2	1	0	2	0	0	2	0	2	2	2	2
Maletis 2007	2	2	2	2	0	0	2	2	2	2	1	2
Marder 1991	0	0	0	1	0	0	2	0	2	2	1	1
Matsumoto 2006	0	0	0	1	0	0	2	1	2	2	2	2
O'Neill 1996	0	1	0	2	0	0	2	1	2	2	1	2
Ropke 2001	0	0	0	1	0	0	2	1	2	2	2	1
Sajovic 2006	0	1	0	2	0	0	2	2	2	2	2	2
Shaieb 2002	0	1	0	1	0	0	2	1	2	2	1	2
Zaffagnini 2006	0	1	0	2	0	0	2	1	2	2	2	1

Study ID	Item
	1			2			3		4		5
	a)	b)	c)	a)	b)	c)	a)	b)	a)	b)	a)	b)	c)	d)
Aglietti 1994	1	1	0	1	1	0	1	1	2	2	1	2	0	0
Aglietti 2004	1	1	0	1	1	1	2	0	2	2	1	2	0	1
Anderson 2001	1	2	0	1	1	0	2	1	2	2	1	2	0	0
Aune 2001	1	1	0	1	1	0	1	2	2	2	1	1	0	0
Beynnon 2002	1	1	0	1	1	0	2	0	2	2	1	2	0	1
Ejerhed 2003	1	1	1	1	1	0	1	0	2	2	1	2	0	0
Eriksson 2001	1	0	0	1	1	0	2	1	2	2	1	2	0	0
Feller 2003	1	2	0	1	1	0	1	1	2	2	1	2	0	0
Ibrahim 2005	1	0	0	1	1	0	0	0	2	2	0	0	0	0
Jansson 2003	1	1	0	1	1	0	1	1	2	2	1	2	0	1
Laxdal 2005	1	1	1	1	1	0	0	1	2	2	1	2	0	0
Maletis 2007	1	2	1	1	1	1	2	2	2	2	1	1	0	1
Marder 1991	1	1	0	1	1	0	0	0	2	2	0	1	0	0
Matsumoto 2006	1	1	0	1	1	0	1	1	2	2	1	2	0	0
O'Neill 1996	1	1	0	1	1	0	1	2	2	2	1	1	0	0
Ropke 2001	1	1	0	1	1	0	1	0	2	2	1	2	0	0
Sajovic 2006	1	1	0	1	1	0	1	2	2	2	1	2	0	0
Shaieb 2002	1	1	0	1	1	0	1	1	2	2	1	1	0	0
Zaffagnini 2006	1	1	0	1	1	0	1	0	2	2	1	1	1	0

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Functional assessment ‐ single hop test: participants with < 90% of opposite side	5	534	Risk Ratio (M‐H, Fixed, 95% CI)	1.17 [0.84, 1.63]
2 Return to activity: participants returning to light or sedentary activity only	4	304	Risk Ratio (M‐H, Fixed, 95% CI)	1.23 [0.81, 1.85]
3 Tegner activity level (0: to 10: top activity)	7		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
3.1 Studies with full data	4	323	Mean Difference (IV, Fixed, 95% CI)	0.23 [‐0.12, 0.59]
3.2 Exploratory analysis: including studies with imputed SDs	7	537	Mean Difference (IV, Fixed, 95% CI)	0.10 [‐0.18, 0.38]
3.3 Randomized studies	3	273	Mean Difference (IV, Fixed, 95% CI)	0.13 [‐0.27, 0.52]
4 Lysholm Score: mean score	7		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
4.1 Studies with full data	5	482	Mean Difference (IV, Fixed, 95% CI)	0.00 [‐1.72, 1.72]
4.2 Exploratory analysis: including studies with imputed SDs	7	637	Mean Difference (IV, Fixed, 95% CI)	‐0.20 [‐1.76, 1.36]
4.3 Randomized studies	4	428	Mean Difference (IV, Fixed, 95% CI)	0.00 [‐1.87, 1.87]
5 Cincinnati Score: mean score	2	118	Mean Difference (IV, Fixed, 95% CI)	‐1.30 [‐4.95, 2.35]
6 Re‐rupture rate	13		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
6.1 All studies	13	1156	Risk Ratio (M‐H, Fixed, 95% CI)	0.78 [0.41, 1.50]
6.2 Quasi‐randomized studies	7	560	Risk Ratio (M‐H, Fixed, 95% CI)	0.83 [0.33, 2.06]
6.3 Randomized studies	6	596	Risk Ratio (M‐H, Fixed, 95% CI)	0.79 [0.31, 2.01]
6.4 HT Femoral fixation with endobutton	4	351	Risk Ratio (M‐H, Fixed, 95% CI)	0.83 [0.24, 2.80]
6.5 HT Femoral fixation with screw	6	509	Risk Ratio (M‐H, Fixed, 95% CI)	0.82 [0.31, 2.19]
6.6 HT Graft with < 4 strands	2	208	Risk Ratio (M‐H, Fixed, 95% CI)	2.30 [0.34, 15.58]
6.7 HT Graft with 4 strands	11	919	Risk Ratio (M‐H, Fixed, 95% CI)	0.67 [0.32, 1.38]
7 KT arthrometer: patients with > 5 mm side to side difference at 134 Newtons/30 lbs	4	311	Risk Ratio (M‐H, Random, 95% CI)	0.40 [0.07, 2.33]
8 KT arthrometer: patients with > 5 mm side to side difference at Maximum Manual Force	6		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
8.1 All studies	6	524	Risk Ratio (M‐H, Fixed, 95% CI)	0.53 [0.25, 1.09]
8.2 Quasi‐randomized studies	5	431	Risk Ratio (M‐H, Fixed, 95% CI)	0.49 [0.22, 1.09]
8.3 Randomized studies	1	93	Risk Ratio (M‐H, Fixed, 95% CI)	0.74 [0.13, 4.24]
8.4 HT Femoral fixation with endobutton	2	135	Risk Ratio (M‐H, Fixed, 95% CI)	0.36 [0.10, 1.25]
8.5 HT Femoral fixation with screw	2	139	Risk Ratio (M‐H, Fixed, 95% CI)	0.83 [0.27, 2.59]
8.6 HT Graft with < 4 strands	1	130	Risk Ratio (M‐H, Fixed, 95% CI)	0.44 [0.09, 2.11]
8.7 HT Graft with 4 strands	5	394	Risk Ratio (M‐H, Fixed, 95% CI)	0.55 [0.24, 1.25]
9 KT arthrometer: Mean side to side difference (mm) at 134 Newtons/30 lbs	7		Mean Difference (IV, Random, 95% CI)	Subtotals only
9.1 Studies with full data	3	195	Mean Difference (IV, Random, 95% CI)	‐0.65 [‐1.35, 0.04]
9.2 All studies: including those with imputed SDs	7	465	Mean Difference (IV, Random, 95% CI)	‐0.94 [‐1.66, ‐0.22]
9.3 Randomized studies	3	185	Mean Difference (IV, Random, 95% CI)	‐1.68 [‐3.03, ‐0.32]
9.4 HT Femoral fixation with endobutton	1	48	Mean Difference (IV, Random, 95% CI)	‐1.03 [‐1.84, ‐0.22]
9.5 HT Femoral fixation with screw	3	193	Mean Difference (IV, Random, 95% CI)	‐0.59 [‐1.32, 0.13]
9.6 HT Graft with < 4 strands	1	44	Mean Difference (IV, Random, 95% CI)	‐3.30 [‐4.43, ‐2.17]
9.7 HT Graft with 4 strands	6	421	Mean Difference (IV, Random, 95% CI)	‐0.60 [‐0.96, ‐0.24]
10 KT arthrometer: Mean side to side difference (mm) at Maximum Manual Force	7		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
10.1 Studies with full data	4	294	Mean Difference (IV, Fixed, 95% CI)	‐0.51 [‐0.95, ‐0.07]
10.2 All studies (including those with imputed SDs)	7	440	Mean Difference (IV, Fixed, 95% CI)	‐0.63 [‐0.99, ‐0.27]
10.3 Randomized studies	3	222	Mean Difference (IV, Fixed, 95% CI)	‐0.52 [‐1.04, 0.00]
10.4 HT Femoral fixation with endobutton	1	61	Mean Difference (IV, Fixed, 95% CI)	0.0 [‐1.08, 1.08]
10.5 HT Femoral fixation with screw	4	251	Mean Difference (IV, Fixed, 95% CI)	‐0.67 [‐1.13, ‐0.22]
10.6 HT Graft with < 4 strands	2	108	Mean Difference (IV, Fixed, 95% CI)	‐1.04 [‐1.82, ‐0.27]
10.7 HT Graft with 4 strands	5	332	Mean Difference (IV, Fixed, 95% CI)	‐0.52 [‐0.92, ‐0.11]
11 Lachman Test: patients with > 2 mm or positive test	10		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
11.1 All studies	10	856	Risk Ratio (M‐H, Fixed, 95% CI)	0.83 [0.71, 0.99]
11.2 Quasi‐randomized studies	4	261	Risk Ratio (M‐H, Fixed, 95% CI)	0.63 [0.40, 0.98]
11.3 Randomized studies	5	470	Risk Ratio (M‐H, Fixed, 95% CI)	0.89 [0.73, 1.07]
11.4 HT Femoral fixation with endobutton	3	289	Risk Ratio (M‐H, Fixed, 95% CI)	0.83 [0.62, 1.11]
11.5 HT Femoral fixation with screw	4	326	Risk Ratio (M‐H, Fixed, 95% CI)	0.93 [0.70, 1.22]
11.6 HT Graft with < 4 strands	3	246	Risk Ratio (M‐H, Fixed, 95% CI)	0.87 [0.65, 1.15]
11.7 HT Graft with 4 strands	7	584	Risk Ratio (M‐H, Fixed, 95% CI)	0.85 [0.69, 1.05]
12 Pivot Shift: patients with a positive test	14		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
12.1 All studies	14	1118	Risk Ratio (M‐H, Fixed, 95% CI)	0.70 [0.54, 0.89]
12.2 Quasi‐randomized studies	8	583	Risk Ratio (M‐H, Fixed, 95% CI)	0.78 [0.55, 1.10]
12.3 Randomized studies	6	535	Risk Ratio (M‐H, Fixed, 95% CI)	0.62 [0.44, 0.88]
12.4 HT Femoral fixation with endobutton	4	346	Risk Ratio (M‐H, Fixed, 95% CI)	0.69 [0.46, 1.02]
12.5 HT Femoral fixation with screw	4	319	Risk Ratio (M‐H, Fixed, 95% CI)	0.59 [0.32, 1.09]
12.6 HT Graft with < 4 strands	3	189	Risk Ratio (M‐H, Fixed, 95% CI)	0.53 [0.29, 0.98]
12.7 HT Graft with 4 strands	12	968	Risk Ratio (M‐H, Fixed, 95% CI)	0.71 [0.55, 0.93]
13 IKDC: patients with Normal and Nearly Normal scores	15		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
13.1 All studies (2000 version)	2	192	Risk Ratio (M‐H, Fixed, 95% CI)	0.94 [0.84, 1.06]
13.2 All studies (1995 version)	13	1150	Risk Ratio (M‐H, Fixed, 95% CI)	1.01 [0.95, 1.07]
13.3 Quasi‐randomized studies (1995 version)	6	542	Risk Ratio (M‐H, Fixed, 95% CI)	1.00 [0.93, 1.07]
13.4 Randomized studies (1995 version)	6	548	Risk Ratio (M‐H, Fixed, 95% CI)	1.00 [0.89, 1.12]
13.5 HT Femoral fixation with endobutton (1995 version)	5	381	Risk Ratio (M‐H, Fixed, 95% CI)	1.01 [0.88, 1.15]
13.6 HT Femoral fixation with screw (1995 version)	4	328	Risk Ratio (M‐H, Fixed, 95% CI)	0.92 [0.81, 1.04]
13.7 HT Graft with < 4 strands (1995 version)	4	407	Risk Ratio (M‐H, Fixed, 95% CI)	1.03 [0.94, 1.12]
13.8 HT Graft with 4 strands (1995 version)	8	655	Risk Ratio (M‐H, Fixed, 95% CI)	0.98 [0.90, 1.06]
14 IKDC Subjective score (1995 version): patients with Normal and Nearly Normal scores	6		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
14.1 All studies	6	427	Risk Ratio (M‐H, Fixed, 95% CI)	0.99 [0.94, 1.06]
14.2 Quasi‐randomized studies	2	94	Risk Ratio (M‐H, Fixed, 95% CI)	0.95 [0.83, 1.09]
14.3 Randomized studies	4	333	Risk Ratio (M‐H, Fixed, 95% CI)	1.01 [0.94, 1.07]
14.4 HT Femoral fixation with endobutton	2	97	Risk Ratio (M‐H, Fixed, 95% CI)	0.93 [0.80, 1.07]
14.5 HT Femoral fixation with screw	3	262	Risk Ratio (M‐H, Fixed, 95% CI)	0.99 [0.92, 1.07]
14.6 HT Graft with < 4 strands	3	182	Risk Ratio (M‐H, Fixed, 95% CI)	1.05 [0.93, 1.18]
14.7 HT Graft with 4 strands	4	283	Risk Ratio (M‐H, Fixed, 95% CI)	0.97 [0.91, 1.03]
15 IKDC Subjective score (2000 version): Mean score	2	192	Mean Difference (IV, Fixed, 95% CI)	‐2.78 [‐5.74, 0.18]
16 Range of motion: Heel height difference (mm)	3		Mean Difference (IV, Random, 95% CI)	Subtotals only
16.1 All studies	3	215	Mean Difference (IV, Random, 95% CI)	3.72 [‐1.38, 8.82]
17 Range of motion: Extension deficit > 3 degrees	14		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
17.1 All studies	14	1139	Risk Ratio (M‐H, Fixed, 95% CI)	1.71 [1.25, 2.33]
17.2 Quasi‐randomized studies	9	691	Risk Ratio (M‐H, Fixed, 95% CI)	1.49 [0.76, 2.91]
17.3 Randomized studies	5	448	Risk Ratio (M‐H, Fixed, 95% CI)	1.79 [1.26, 2.54]
17.4 HT Femoral fixation with endobutton	4	337	Risk Ratio (M‐H, Fixed, 95% CI)	1.58 [1.00, 2.49]
17.5 HT Femoral fixation with screw	5	357	Risk Ratio (M‐H, Fixed, 95% CI)	1.64 [1.04, 2.58]
17.6 HT Graft with < 4 strands	3	267	Risk Ratio (M‐H, Fixed, 95% CI)	1.67 [0.83, 3.35]
17.7 HT Graft with 4 strands	9	742	Risk Ratio (M‐H, Fixed, 95% CI)	1.78 [1.22, 2.59]
18 Range of motion: Flexion deficit > 5 degrees	12		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
18.1 All studies	12	1001	Risk Ratio (M‐H, Fixed, 95% CI)	0.88 [0.70, 1.10]
18.2 Quasi‐randomized studies	7	553	Risk Ratio (M‐H, Fixed, 95% CI)	2.08 [0.89, 4.87]
18.3 Randomized studies	5	448	Risk Ratio (M‐H, Fixed, 95% CI)	0.77 [0.61, 0.97]
18.4 HT Femoral fixation with endobutton	3	252	Risk Ratio (M‐H, Fixed, 95% CI)	0.68 [0.40, 1.16]
18.5 HT Femoral fixation with screw	4	304	Risk Ratio (M‐H, Fixed, 95% CI)	0.94 [0.74, 1.19]
18.6 HT Graft with < 4 strands	3	267	Risk Ratio (M‐H, Fixed, 95% CI)	1.08 [0.74, 1.55]
18.7 HT Graft with 4 strands	9	706	Risk Ratio (M‐H, Fixed, 95% CI)	0.88 [0.66, 1.19]
19 Strength testing: Mean flexion at 60 deg/s (as % torque of opposite knee)	5		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
19.1 Studies with full data	2	144	Mean Difference (IV, Fixed, 95% CI)	11.62 [6.44, 16.80]
19.2 All studies (including those with imputed SDs)	5	317	Mean Difference (IV, Fixed, 95% CI)	6.63 [3.12, 10.13]
19.3 Randomized studies (all with imputed SDs)	3	173	Mean Difference (IV, Fixed, 95% CI)	2.39 [‐2.38, 7.16]
19.4 HT Femoral fixation with endobutton	1	61	Mean Difference (IV, Fixed, 95% CI)	4.0 [‐4.07, 12.07]
19.5 HT Femoral fixation with screw	1	72	Mean Difference (IV, Fixed, 95% CI)	14.40 [7.51, 21.29]
19.6 HT Graft with < 4 strands	2	112	Mean Difference (IV, Fixed, 95% CI)	1.53 [‐4.38, 7.44]
19.7 HT Graft with 4 strands	3	205	Mean Difference (IV, Fixed, 95% CI)	9.39 [5.04, 13.75]
20 Strength testing: Mean flexion at 180 deg/s (as % torque of opposite knee)	3		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
20.1 All studies (including those with imputed SDs)	3	184	Mean Difference (IV, Fixed, 95% CI)	5.58 [1.52, 9.65]
21 Strength testing: Mean extension at 60 deg/s (as % torque of opposite knee)	5		Mean Difference (IV, Random, 95% CI)	Subtotals only
21.1 Studies with full data	2	144	Mean Difference (IV, Random, 95% CI)	‐4.60 [‐10.19, 0.98]
21.2 All studies (including those with imputed SDs)	5	317	Mean Difference (IV, Random, 95% CI)	‐2.97 [‐8.10, 2.16]
21.3 Randomized studies (all with imputed SDs)	3	173	Mean Difference (IV, Random, 95% CI)	‐1.49 [‐10.91, 7.93]
21.4 HT Femoral fixation with endobutton	1	61	Mean Difference (IV, Random, 95% CI)	0.0 [‐8.61, 8.61]
21.5 HT Femoral fixation with screw	1	72	Mean Difference (IV, Random, 95% CI)	‐5.90 [‐13.42, 1.62]
21.6 HT Graft with < 4 strands	2	112	Mean Difference (IV, Random, 95% CI)	‐1.98 [‐18.24, 14.27]
21.7 HT Graft with 4 strands	3	205	Mean Difference (IV, Random, 95% CI)	‐3.24 [‐7.92, 1.45]
22 Strength testing: Mean extension at 180 deg/s (as % torque of opposite knee)	3		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
22.1 All studies (including those with imputed SDs)	3	184	Mean Difference (IV, Fixed, 95% CI)	‐1.48 [‐6.45, 3.48]
23 Anterior knee symptomatology: Incidence (general)	8		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
23.1 All studies	8	459	Risk Ratio (M‐H, Fixed, 95% CI)	1.45 [1.05, 2.01]
23.2 Quasi‐randomized studies	4	232	Risk Ratio (M‐H, Fixed, 95% CI)	1.73 [1.08, 2.79]
23.3 Randomized studies	4	227	Risk Ratio (M‐H, Fixed, 95% CI)	1.22 [0.78, 1.92]
23.4 HT Femoral fixation with endobutton	3	158	Risk Ratio (M‐H, Fixed, 95% CI)	1.58 [0.93, 2.69]
23.5 HT Femoral fixation with screw	3	185	Risk Ratio (M‐H, Fixed, 95% CI)	1.26 [0.73, 2.15]
23.6 HT Graft with < 4 strands	2	84	Risk Ratio (M‐H, Fixed, 95% CI)	1.88 [0.89, 3.96]
23.7 HT Graft with 4 strands	4	256	Risk Ratio (M‐H, Fixed, 95% CI)	1.62 [1.07, 2.47]
24 Anterior knee symptomatology: Kneeling discomfort	4		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
24.1 All studies	4	355	Risk Ratio (M‐H, Fixed, 95% CI)	4.46 [2.97, 6.69]
25 Numbers of participants in analyses / Numbers entered in each trial	17	1456	Risk Ratio (M‐H, Fixed, 95% CI)	0.99 [0.96, 1.03]
26 Numbers with acute reconstructions at baseline	7	610	Risk Ratio (M‐H, Fixed, 95% CI)	1.07 [0.87, 1.33]
27 Numbers of males at baseline	17	1423	Risk Ratio (M‐H, Fixed, 95% CI)	0.98 [0.91, 1.05]

Methods	Treatment allocation by alternating Non blinded assessment Mean follow‐up: 28 months (1994 publication); 5.7 years (1997 publication) Loss to follow‐up: 3 (excluded from analysis) Graft re‐ruptures: not reported
Participants	Italy 63 participants with chronic (> 6 months), isolated unilateral ACL tears Group allocation data for 60 participants. PT: n = 30, mean 21.5 years (14 to 33 years), 16 males / 14 females HT: n = 30, mean 23.7 years (14 to 36 years), 20 males / 10 females
Interventions	Arthroscopic ACL reconstruction with: 1. Patellar tendon (number of incisions not specified; 20 mm bone blocks, proximal fixation ‐ sutures, cortical bone screw, washer; tibial fixation ‐ interference screw, sutures and cortical bone screw) versus 2. Hamstring tendon: doubled semitendinosus/gracilis tendon (4 strands; number of incisions not specified; proximal fixation ‐ cortical bone screw, washer; tibial fixation ‐ staple, sutures and cortical bone screw (in 24 cases), or sutures, screw and washer (in 6 cases))
Outcomes	Mean 28 months follow‐up for: Subjective evaluation, symptoms, range of motion, joint stability; extension loss; flexion loss; static stability with KT‐2000 (15 lb, 20 lb, 30 lb); muscle strength with Cybex II (60, 120, 180 deg/sec); level of activity (based on function, intensity, and return to sport); objective stability; patellofemoral symptoms; pivot shift (absent/glide/jerk/subluxation)
Notes	No additional methodological information or individual patient data were received from the authors.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Alternating patients.
Allocation concealment (selection bias)	High risk
Blinding (performance bias and detection bias) All outcomes	High risk
Incomplete outcome data (attrition bias) All outcomes	High risk	Three patients lost to follow‐up not accounted for in analysis.
Selective reporting (reporting bias)	High risk	Did not report re‐ruptures, contralateral ruptures or withdrawals.
Other bias	Unclear risk	Unclear

Methods	Treatment allocation by alternating Independent & blinded assessment Minimum follow‐up: 24 months Loss to follow‐up: None Graft re‐ruptures: not reported
Participants	Italy 120 patients with chronic (> 30 days), isolated unilateral ACL tears PT: n = 60, mean 25 years (16 to 39 years), 46 males / 14 females HT: n = 60, mean 25 years (15 to 39 yrs), 46 males / 14 females
Interventions	Arthroscopically assisted ACL reconstruction with: 1. Patellar tendon (single incision; proximal fixation ‐ Tunneloc screw; tibial fixation ‐ Soft threaded interference screw) versus 2. Hamstring tendon: Semitendinosus/gracilis tendons (4 strands; single incision; proximal fixation ‐ Bone Mulch screw; tibial fixation ‐ WasherLoc device)
Outcomes	Minimum 4, 12 and 24 months follow‐up for: Static stability with KT‐1000 (134N & MM); Lachman (defined by the IKDC 2000); pivot shift (defined by IKDC 2000); IKDC (2000); return to sports/activity; strength with Cybex (60, 120, 180 deg/sec); KOOS; range of motion; anterior knee pain
Notes	No additional methodological information or individual patient data were received from the authors.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Alternating.
Allocation concealment (selection bias)	High risk
Blinding (performance bias and detection bias) All outcomes	Low risk	Independent blinded assessor for follow‐up outcome assessment. Patients were not blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk
Selective reporting (reporting bias)	High risk	Did not report re‐ruptures, contralateral ruptures or withdrawals.
Other bias	High risk	It is difficult to understand how 60 patients randomly assigned to each group resulted in identical numbers of left and right sided injured knees, the same sex distribution and mean age. There were no patients lost to follow‐up despite an a priori sample size estimate, which allowed for a 20% lost to follow‐up. In their previous published trial, only 3 of 60 patients were lost to follow‐up, which suggests that accounting for a 20% lost to follow‐up rate in the sample size estimate was excessive.

Methods	Computer‐generated random assignment Non blinded assessment Mean follow‐up: 35.4 months Loss to follow‐up: 2 HT (these were excluded from the analysis) Graft re‐ruptures: not reported
Participants	USA Study recruitment period: 1991 to 1993 70 participants with acute (< 12 weeks) or chronic (> 12 weeks) ACL deficiency PT: n = 35, mean 23.6 years (14 to 44 years), 23 males / 12 females HT: n = 35, mean 20.1 years (14 to 38 years), 23 males / 12 females
Interventions	Arthroscopically‐assisted ACL reconstruction with: 1. Patellar tendon (single incision; proximal fixation ‐ interference screw (7 x 25 mm); tibial fixation ‐ barbed staples) versus 2. Hamstring tendon: semitendinosus/gracilis tendons (2 strands; single incision; proximal fixation ‐ barbed staples; tibial fixation ‐ non‐absorbable sutures)
Outcomes	Mean 35 months follow‐up for: Static stability with KT‐1000; range of motion; pivot shift; IKDC; return to sport/activity; strength with Cybex II (60 & 180 deg/s); patellofemoral crepitus
Notes	A third intervention group involving HT with extra‐articular procedure (semitendinosus/gracilis tendons with a Losee extra‐articular iliotibial band tenodesis) in 35 patients is not included in this review. Additional methodological information received from Anderson comprised a description of the sample size calculation, patient recruitment / inclusion / exclusion numbers, and the number or re‐ruptures and revisions. Individual patient data were unavailable.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer generated.
Allocation concealment (selection bias)	Unclear risk	No information
Blinding (performance bias and detection bias) All outcomes	High risk
Incomplete outcome data (attrition bias) All outcomes	Low risk
Selective reporting (reporting bias)	High risk	Did not report re‐ruptures, contralateral ruptures or withdrawals.
Other bias	Low risk

Methods	Randomization method: sealed envelopes Independent, non blinded assessment Minimum follow‐up: 24 months Loss to follow‐up: 8 patients in total (excluded from the analysis) Exclusions: 3 patients in total with ACL injury of the uninvolved knee (excluded from the analysis) Graft re‐ruptures: 1 PT and 2 HT had traumatic failures requiring revision sx (included in analysis)
Participants	Norway 72 participants with subacute (as soon as inflammation was over and full range of motion was achieved) & chronic (failed non‐operative treatment) ACL deficiency PT: n = 35, mean 25 years, 19 males / 16 females HT: n = 37, mean 27 years, 21 males / 16 females
Interventions	ACL reconstruction with: 1. Patellar tendon (single incision; proximal fixation ‐ Titanium interference screw (7 x 25 mm); tibial fixation ‐ interference screw (7 x 25 mm)) versus 2. Hamstring tendon: semitendinosus/gracilis tendons (4 strands; single incision; proximal fixation ‐ EndoButton and interference screw; tibial fixation ‐ barbed staples)
Outcomes	Minimum 24 months (also 6 and 12 months) follow‐up for: Cincinnati score; static stability with KT‐1000 (mm); stairs hopple test; single leg hop; patient satisfaction; kneeling test; strength with Cybex 6000 (60 & 240 degrees/sec)
Notes	Additional methodological information received from Aune comprised a description of allocation concealment and the sample size calculation. Individual patient data were unavailable.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details
Allocation concealment (selection bias)	Unclear risk	Sealed envelopes used for randomization, but allocation process to preserve concealment was not described.
Blinding (performance bias and detection bias) All outcomes	High risk	Independent assessor.
Incomplete outcome data (attrition bias) All outcomes	High risk	Eight patients were lost to follow‐up and three patients had contra‐lateral ruptures. The authors failed to report to which groups these patients belonged at baseline.
Selective reporting (reporting bias)	Low risk
Other bias	Low risk

Methods	Treatment allocation using a random numbers table Non blinded assessment Mean follow‐up: 39 months Loss to follow‐up: 6 PT and 6 HT (excluded from analyses) Graft re‐ruptures: not reported
Participants	USA 56 participants with chronic (> 6 months), isolated unilateral ACL tears PT: n = 28, mean 28.5 years (18 to 46 years), 18 males / 10 females HT: n = 28, mean 29.9 years (19 to 42 yrs), 13 males / 15 females
Interventions	Arthroscopic ACL reconstruction with: 1. Patellar tendon (2 incisions; proximal and tibial fixation ‐ interference screw) versus 2. Hamstring tendon: semitendinosus/gracilis graft (2 strands; 2 incisions; proximal fixation ‐ Belt‐buckle type double‐staple technique; tibial fixation ‐ staple)
Outcomes	Mean 12 and 39 months follow‐up for: Static stability with KT‐1000 (90 N & 133 N); Lachman (0/1+2+3+); pivot shift; knee function (single‐leg hop, weight bearing, squats, stairs, run in place, duck walk); anterior knee pain; Tegner; return to sport/activity; strength with Cybex (60, 180, 240 deg/s); stiffness
Notes	No additional methodological information or individual patient data were received from the authors.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Random number table.
Allocation concealment (selection bias)	High risk
Blinding (performance bias and detection bias) All outcomes	High risk
Incomplete outcome data (attrition bias) All outcomes	Low risk	Both groups had 6 patients lost to follow‐up.
Selective reporting (reporting bias)	High risk	Did not report re‐ruptures, contralateral ruptures or withdrawals.
Other bias	Low risk

Methods	Sealed envelopes, but randomization method not described Independent, but non blinded assessment Minimum follow‐up: 24 months Loss to follow‐up: 1 PT and 1 HT (excluded from analysis) Graft re‐ruptures: 1 PT and 2 HT (excluded from analysis)
Participants	Sweden 71 participants with chronic (> 2 months), unilateral ACL tears PT: n = 34, mean 26 years (14 to 49 years), 21 males / 11 females HT: n = 37, mean 29 years (15 to 59 years), 25 males / 9 females
Interventions	Arthroscopic ACL reconstruction with: 1. Patellar tendon (2 incisions; proximal fixation ‐ "Silk" interference screw (7 mm); tibial fixation ‐ "Silk" interference screw (9 mm)) versus 2. Hamstring tendon: semitendinosus tendon (3 or 4 strands; single incision; proximal and tibial fixation ‐ soft‐threaded RCI interference screw (7 mm))
Outcomes	Minimum 24 months follow‐up for: Lachman (0/1+2+3+); Lysholm; Tegner; range of motion; static stability with KT‐1000 (89N); single leg hop test; IKDC; anterior knee pain; knee walking test; strength with Cybex (60 deg/s)
Notes	Additional information received from Ejerhed comprised individual pre‐ and post‐operative patient data for the following outcomes: anterior knee pain, knee walking, kneeling, Lysholm, Cybex extension and flexion of injured and non‐injured sides, KT‐1000, range of motion (extension & flexion), IKDC, Tegner activity level.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details
Allocation concealment (selection bias)	Unclear risk	Sealed envelopes but allocation process to preserve concealment was not described.
Blinding (performance bias and detection bias) All outcomes	High risk	Independent assessor.
Incomplete outcome data (attrition bias) All outcomes	Low risk
Selective reporting (reporting bias)	High risk	Excluded re‐ruptures in final analysis. Did not report contra‐lateral ruptures or withdrawals.
Other bias	High risk	Lack of explicit inclusion and exclusion criteria, patient sampling strategy and 14 patient overlap with Laxdal's trial.

Methods	Closed envelopes, but randomization method not described Independent, but non blinded assessment Mean follow‐up: 33 months (median 31 months) Loss to follow‐up: 4 patients in total (excluded from analysis) Graft re‐ruptures (all traumatic): 2 PT (no revision surgery), 3 HT (no revision surgery), 1 HT requiring revision surgery (total of 10 patients excluded from the analysis)
Participants	Sweden 164 participants with chronic (> 2 months), isolated unilateral ACL tears and an overall mean age of 25.7 years (+/‐6.9 years), PT: n = 84, mean 25 years (15 to 40 years), 41 males / 43 females HT: n = 80, mean 27 years (16 to 45 years), 55 males / 25 females
Interventions	Arthroscopic ACL reconstruction with: 1. Patellar tendon (single incision; proximal fixation ‐ non‐absorbable interference screw (7 x 20 mm); tibial fixation ‐ non‐absorbable interference screw (9 x 20 mm)) versus 2. Hamstring tendon: semitendinosus tendon (4 strands; single incision; proximal fixation ‐ Endobutton; tibial fixation ‐ screws)
Outcomes	Mean 33 months follow‐up for: Stryker laxity test (9.08 kg, 18.16 kg); range of motion; pivot shift; Lachman (as defined by the IKDC 1995); single‐leg hop; Tegner activity level; Lysholm; patellofemoral pain; IKDC; patient satisfaction; knee function (VAS)
Notes	Additional information received from Eriksson comprised a description of the allocation concealment methodology, lack of a sample size calculation, and the mean age for each group. Eriksson also provided individual patient data for the following outcomes: Tegner activity level, Lysholm, single leg hop, IKDC, ROM of index and contralateral knees, flexion and extension deficits, Stryker strength (20lbs) for index and contralateral knees; Lachman; anterior and posterior drawer tests; pivot shift test; MCL and LCL stability; patellofemoral, medial and lateral crepitus; harvest site pathology; thigh circumference of the index and contralateral legs; pain.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details
Allocation concealment (selection bias)	Low risk	Closed envelopes opened within one hour of surgery.
Blinding (performance bias and detection bias) All outcomes	High risk	Independent assessor.
Incomplete outcome data (attrition bias) All outcomes	High risk	The authors excluded patients who were lost to follow‐up and those who suffered a re‐rupture, of the ACL graft.
Selective reporting (reporting bias)	Low risk
Other bias	High risk	Gender distribution was different between the two groups at baseline.

Methods	Computer‐generated random assignment Independent, but unblinded assessment Mean follow‐up: 36 months Loss to follow‐up: 4 PT and 2 HT (excluded from analysis) Withdrawals: 1 HT (excluded from analysis) Graft re‐ruptures (traumatic): 1 PT (excluded from analysis)
Participants	Australia 65 participants with acute (3 weeks to 12 months), isolated unilateral ACL tears PT: n = 31, mean 25.8 years (SD = 6 years), 23 males / 8 females HT: n=34, mean 26.3 years (SD = 6 years), 24 males / 10 females
Interventions	Arthroscopically‐assisted ACL reconstruction with: 1. Patellar tendon (single incision; proximal fixation ‐ EndoButton; tibial fixation ‐ Cannulated metallic interference silk screw) versus 2. Hamstring tendon: semitendinosus/gracilis tendons (4 strands; single incision; proximal fixation ‐ EndoButton; tibial fixation ‐ whip stitched suture and fixation post)
Outcomes	Minimum 36 months (also of 4, 8, 12 and 24 months) follow‐up for: Static stability with KT‐1000 (67N & 134N); range of motion; pivot shift; IKDC, Cincinnati; anterior knee pain (VAS); strength with Cybex II (60 & 240 deg/s); return to sport/activity; kneeling; radiographic assessment
Notes	Additional information received from Feller comprised individual pre‐ and post‐operative patient data for the following outcomes: IKDC, pain, effusion, pain with kneeling, flexion and extension deficits, heel height difference, KT‐1000 (67N, 134N, max manual), Cincinnati
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated random assignment.
Allocation concealment (selection bias)	Unclear risk	No details
Blinding (performance bias and detection bias) All outcomes	High risk
Incomplete outcome data (attrition bias) All outcomes	High risk	Patients who were lost to follow‐up, withdrew or suffered a re‐rupture were excluded from the final analysis.
Selective reporting (reporting bias)	Low risk
Other bias	Low risk

Methods	Treatment allocation by birth date Non blinded assessment Mean follow‐up: 81 months (60 to 96 months) Loss to follow‐up: 25 patients in total (information on group allocation unavailable) Graft re‐ruptures: not reported
Participants	Kuwait 110 participants Group allocation data for 85 male participants (mean age 22.3 years (17 to 34 years)) with chronic ACL deficiency (mean time from injury to surgery was 9.7 months (4 to 13 months)) PT: n = 40 (mean age unavailable) HT: n = 45 (mean age unavailable)
Interventions	Arthroscopically‐assisted ACL reconstruction with: 1. Patellar tendon (single incision; proximal fixation ‐ EndoButton; tibial fixation ‐ interference screw) versus 2. Hamstring tendon: semitendinosus/gracilis tendons (4 strands; single incision; proximal fixation ‐ EndoButton; tibial fixation ‐ screw and spiked washer or small plate 2 screws and staple)
Outcomes	Mean 81 months follow‐up for: Subjective questionnaire (return to work, satisfaction, giving way episodes, anterior knee pain, return to activity); loss of extension and flexion, Lachman (reported as the proportion of patients with a Grade 1 or negative test); pivot shift (reported as the proportion of patients a pivot glide or negative shift); anterior drawer sign (proportion of patients with grade one or negative test) ; KT‐1000 arthrometer (maximum manual test reported as a proportion of patients); Lysholm (proportion of patients with excellent, good, fair; and as a mean score); Tegner (reported as an average activity level); IKDC (1995); radiographic examination
Notes	No information on loss to follow‐up. Unable to get additional information or data from authors.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Birthdate.
Allocation concealment (selection bias)	High risk
Blinding (performance bias and detection bias) All outcomes	High risk
Incomplete outcome data (attrition bias) All outcomes	High risk	Did not account for lost to follow‐up patients or describe the group allocation for these lost to follow‐up patients.
Selective reporting (reporting bias)	High risk	Did not report standard deviations for mean values (e.g. Lysholm and Tegner scores). No detail with respect to how radiographic assessment was determined.
Other bias	High risk	No information presented to determine any selection bias since inclusion and exclusion criteria were not described. All patients with reported data and included in follow‐up were males. The description of the baseline characteristics (e.g. no average age for each group) between the two groups was insufficient to determine if the groups were comparable.

Methods	Treatment allocation by birth year Non blinded assessment Minimum follow‐up: 21 months (21 to 38 months, 2003 publication); 47 months (median 5 years, 2006 publication) Loss to follow‐up (excluded from analysis): 2003 ‐ 8 PT; 2006 ‐ 13 PT and 3 HT Graft re‐ruptures (high energy trauma): 2 HT (excluded from analysis) Failures: 2 HT (excluded from analysis)
Participants	Finland 99 participants with acute or chronic (1 week to 20 years), isolated unilateral ACL tears PT: n = 51 (no age or gender information) HT: n = 48 (no age or gender information)
Interventions	Arthroscopically‐assisted ACL reconstruction with: 1. Patellar tendon (2 incisions; proximal and tibial fixation ‐ interference metal screws (8 or 9 mm)) versus 2. Hamstring tendon: semitendinosus/gracilis tendons (one incision; 4 strands; proximal fixation ‐ titanium plate attached with a Dacron loop; tibial fixation ‐ cortical screw (4.5 mm) and spiked washer)
Outcomes	Minimum 21 months and median 60 months follow‐up for: Lachman (positive or negative test); pivot shift (positive or negative); stability with CA‐4000 arthrometer; strength with Lido Multijoint II dynamometer (60 & 180 degrees/second); IKDC; Lysholm; Tegner; Kujala patellofemoral score form; range of motion (mean extension deficit); radiographic assessment
Notes	No additional methodological information or individual patient data were received from the authors.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Treatment allocation by birth year
Allocation concealment (selection bias)	High risk
Blinding (performance bias and detection bias) All outcomes	High risk
Incomplete outcome data (attrition bias) All outcomes	High risk	Patients who were lost to follow‐up and who suffered graft re‐ruptures and failures were excluded.
Selective reporting (reporting bias)	High risk	For the Tegner Activity outcome score, data from Jansson 2003 were taken from the original two‐year publication and follow‐up period because this data for the later follow‐up period was not reported (Harilainen 2006).
Other bias	High risk	No baseline information on age and gender distribution was provided

Methods	Closed envelopes, computer‐generated randomization at time of surgery. Trained and experienced blinded assessment Mean follow‐up: 26.2 months Loss to follow‐up: 2 HT (no data for 24‐month analysis only) Exclusions: 2 PT + 1 HT (3 patients in total ruptured their contralateral ACL; therefore were excluded from the KT‐1000 and 1‐leg hop analyses only) Graft re‐ruptures: 1 HT with no follow‐up (excluded from analyses)
Participants	USA 99 patients with acute or chronic (0 to 3 months, 3.1 to 6 months, > 6 months), isolated unilateral ACL tears PT: n = 46, mean 27.2 years (15 to 42 years), 31 males / 15 females HT: n = 53, mean 27.7 years (14 to 48 years), 45 males / 8 females
Interventions	Arthroscopically‐assisted ACL reconstruction with: 1. Patellar tendon (single incision, proximal fixation ‐ bioabsorbable interference screw (7 x 23 mm); tibial fixation ‐ bioabsorbable interference screw (9 x 23 mm)) versus 2. Hamstring tendon: semitendinosus/gracilis graft (4 strands; single incision, proximal fixation ‐ 1 screw; tibial fixation ‐ 2 bioabsorbable screws)
Outcomes	Minimum 24 months (also 12 months) follow‐up for: Tegner; IKDC; Lysholm: ROM; single leg hop; KT‐1000 (manual maximum) (3 and 6 months reported; SF‐36; patient knee rating; pivot shift (0/1/2/3); kneeling; return to activity; Biodex strength testing (60, 180, 300, degrees/second) (6 months reported)
Notes	Additional information received from Maletis comprised individual pre‐ and post‐operative patient data for the following outcomes: IKDC, Lysholm, Tegner activity level, range of motion (flexion and extension deficits), heel height difference, single‐leg hop, KT‐1000 (maximum manual force and 134N), SF‐36 score, pivot shift test, Lachman; patellofemoral crepitus, strength testing (60 and 180 degrees/second).
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated randomization.
Allocation concealment (selection bias)	Low risk	Closed envelopes opened at time of surgery.
Blinding (performance bias and detection bias) All outcomes	Low risk	Independent and blinded assessor.
Incomplete outcome data (attrition bias) All outcomes	Low risk
Selective reporting (reporting bias)	Low risk
Other bias	Low risk

Methods	Treatment allocation by one‐to‐one alternating sequence Unblinded assessment Mean follow‐up: 29 months (24 to 42 months) Loss to follow‐up: 3 PT and 5 HT (excluded from analysis) Graft re‐ruptures: 1 PT and 1 HT (included in analysis)
Participants	USA 80 participants with chronic ACL deficiency, including 2 patients with previous ACL repair in the PT group PT: n = 37, mean 21.6 years (16 to 35 years), 24 males / 13 females HT: n = 35, mean 23.8 years (17 to 41 years), 26 males / 9 females
Interventions	Arthrocopic ACL reconstruction with: 1. Patellar tendon (2 incisions, proximal and tibial fixation ‐ posts and washers) versus 2. Hamstring tendon: semitendinosus/gracilis tendons (4 strands; 2 incisions; proximal and tibial fixation ‐ posts and washers)
Outcomes	Mean 29 months (24 to 42 months) follow‐up for: Zarins‐Rowe subjective rating score, average out of 50 points; proportion of patients with giving way or apprehension and knee pain; proportion of patients who returned to their pre‐injury activity level, had 2+ or greater patellofemoral crepitus or presence of an effusion; range of motion reported as a proportion of patients with a loss of extension or flexion; Lachman test (reported as a proportion of patients by grade (0/1+/2+/3+/4+) and as a mean value; pivot shift test (reported as a proportion of patients by grade (0/1/2/3/3+) and as a mean value; static stability with KT‐1000 at 90 N reported as a proportion of patients by grade and as a mean value; isokinetic testing using Cybex dynamometer at 60 degrees/sec.
Notes	No additional methodological information or individual patient data were received from the authors.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	One‐to‐one alternating sequence.
Allocation concealment (selection bias)	High risk
Blinding (performance bias and detection bias) All outcomes	High risk	No attempt to blind assessors or patients.
Incomplete outcome data (attrition bias) All outcomes	High risk	No information on lost to follow‐up patients.
Selective reporting (reporting bias)	High risk	Static stability testing using KT arthrometer only at 90 Newtons. Ordinal data were analysed as continuous and parametric statistics utilized (e.g. Lachman, pivot shift and KT arthrometer testing). No baseline information on patients lost to follow‐up.
Other bias	High risk	No inclusion or exclusion criteria were specified. Two patients in the patellar tendon group had a previous ACL repair prior to the reconstruction.

Methods	Treatment allocation by birth month Independent, non blinded assessment Follow‐up: mean 42 months (24 to 60 months) [1996 publication]; mean 102 months (72 to 121 months) [2001 publication] Loss to follow‐up: 1 PT (2 incision), 1 HT (2 patients in total; excluded from analysis) Graft re‐ruptures (traumatic): 2 PT (1 incision), 2 HT (4 patients in total; included in analysis)
Participants	USA 127 participants (1996 publication) / 229 participants (2001 publication) with acute (< 3 weeks) or chronic (> 3 weeks), isolated unilateral ACL tears Group allocation data for 125 participants (1996 publication) / 225 participants (2001 publication) PT (1 incision): n = 45, mean 28 years (14 to 56 years), 28 males / 17 females (1996 publication); n = 75 (2001 publication) PT (2 incision): n = 40, mean 26 years (15 to 49 years), 26 males / 14 females (1996 publication); n = 75 (2001 publication) HT: n = 40, mean 27 years (14 to 56 years), 27 males / 13 females (1996 publication); n = 75 (2001 publication) Loss to follow‐up: 1 PT (1 incision) and 1 PT (2 incision) and 2 deaths (group assignment not reported). Failures (traumatic, included in the analysis) : 2 PT (2 incision) and 2 HT (1996 publication) / 5 PT (1 incision), 4 PT (2 incision) and 1 HT(2001 publication).
Interventions	Arthroscopically‐assisted ACL reconstruction with: 1. Patellar tendon: 1 or 2 incisions, (proximal fixation ‐ interference screw (9 x 25 mm); tibial fixation ‐ interference screw, or barbed staples for long grafts (17 cases)) versus 2. Hamstring tendon: semitendinosus/gracilis tendons (2 strands; 2 incisions; proximal and tibial fixation ‐ barbed staples)
Outcomes	Mean 42 months follow‐up for: Lysholm, KT‐2000 (Maximum Manual), range of motion, Lachman (0/1/2/3); single‐leg hop, return to activity; failures; strength with Biodex, patellofemoral crepitus; IKDC; radiographs. Mean 102 months follow‐up for: KT‐2000 (Maximum Manual), return to activity; failures; strength with Biodex, patellofemoral crepitus; IKDC; radiographs.
Notes	No additional methodological information or individual patient data were received from the authors.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Allocation by birth month.
Allocation concealment (selection bias)	High risk
Blinding (performance bias and detection bias) All outcomes	High risk
Incomplete outcome data (attrition bias) All outcomes	High risk	The two patients lost to follow‐up and two deaths were excluded from the analysis.
Selective reporting (reporting bias)	High risk	The authors failed to report the following outcomes in their 2001 publication on the larger groups of patients: Return to activity; Lachman; range of motion; patellofemoral crepitus; Lysholm score and single leg hop.
Other bias	High risk	Recruitment in this trial continued beyond the point where the results of the original publication (1996) were likely analysed. This may have resulted in a selection bias in the 2001 publication.

Methods	Treatment allocation by surgery date Independent, non blinded assessment Minimum follow‐up: 2 years Loss to follow‐up: not reported Graft re‐ruptures: not reported
Participants	Germany 40 participants (total 32 males / 8 females) with chronic (> 3 months), isolated unilateral ACL tears PT: n = 20, mean 27.4 years (17 to 37 years) HT: n = 20, mean 27.7 years (15 to 43 years)
Interventions	Arthroscopically‐assisted ACL reconstruction with: 1. Patellar tendon (single incision; proximal and tibial fixation ‐ titanium interference screws (7 x 20 mm)) versus 2. Hamstring tendon: semitendinosus tendon (2 strands; single incision; proximal fixation ‐ Endobutton and MerseleneBand; tibial fixation ‐ ligament staples (8 mm))
Outcomes	Minimum 24 months follow‐up for: IKDC (subjective evaluation, symptoms, range of motion, joint stability); extension loss; flexion loss; static stability with KT‐2000 (15 lb, 20 lb, 30 lb); muscle strength with Cybex II (60, 120, 180 degrees/second); return to sports/activity; objective stability; patellofemoral symptoms; pivot shift
Notes	No additional methodological information or individual patient data were received from the authors.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Treatment allocation by surgery date
Allocation concealment (selection bias)	High risk
Blinding (performance bias and detection bias) All outcomes	High risk	Independent assessor.
Incomplete outcome data (attrition bias) All outcomes	Low risk
Selective reporting (reporting bias)	High risk	The authors did not provide any information on graft re‐ruptures, contra‐lateral ruptures and patients lost to follow‐up.
Other bias	Unclear risk	There was no information to determine gender distribution for each group.

Methods	Treatment allocation by operative registration position Non blinded, not independent assessment Minimum follow‐up: 60 months Loss to follow‐up: 2 PT and 1 HT (excluded from analysis) Exclusions from analysis (5 patients in total): 3 PT and 2 HT (due to contralateral ruptures) Graft re‐ruptures: 1 PT and 1 HT (revision surgery)
Participants	Slovenia 64 participants with acute or chronic isolated unilateral ACL tears PT: n = 32 randomized / n = 26 analyzed (4 acute, 22 chronic), mean 27 years (16 to 46 years), 14 males / 12 females HT: n = 32 randomized / n = 28 analyzed (4 acute, 24 chronic), mean 24 years (14 to 42 years), 13 males / 15 females
Interventions	Arthroscopically‐assisted ACL reconstruction with: 1. Patellar tendon (single incision; proximal fixation ‐ metal round‐head cannulated interference (RCI) screw; tibial fixation ‐ bioabsorbable interference screws) versus 2. Hamstring tendon: semitendinosus/gracilis tendons (4 strands; single incision; proximal fixation ‐ custon left‐threaded or right‐thread round‐head cannulated interference (RCI) screw; tibial fixation ‐ bioabsorbable interference screw)
Outcomes	Minimum 60 months follow‐up for: Lysholm, KT‐2000 (89 N and 134 N), range of motion, Lachman (A/B/C); pivot shift; single‐leg hop, anterior knee pain, IKDC, return to activity, radiographs
Notes	Additional information received from Sajovic comprised a description of the lack of a sample size calculation, and individual pre‐ and post‐operative patient data for the following outcomes: IKDC, Lysholm, anterior knee pain, patellofemoral crepitation, flexion and extension deficits, Lachman, pivot shift, KT‐2000, donor side morbidity, single‐leg hop, x‐ray arthrosis.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Treatment allocation by operative registration position.
Allocation concealment (selection bias)	High risk
Blinding (performance bias and detection bias) All outcomes	High risk
Incomplete outcome data (attrition bias) All outcomes	High risk	Patients were excluded due to graft re‐ruptures, contra‐lateral ruptures and lost to follow‐up.
Selective reporting (reporting bias)	High risk	Baseline information was not provided for the patients excluded from the final analysis.
Other bias	Unclear risk	Insufficient information was available to understand the eligible population and how patients were included and excluded from the trial.

Methods	Treatment allocation by birth date Independent, non blinded assessment Mean follow‐up: 33 months Loss to follow‐up: 12 patients in total (excluded from analysis) Graft re‐ruptures (4 patients in total, excluded from analysis): 1 PT (traumatic), 1 PT (atraumatic) and 2 HT (traumatic)
Participants	USA 82 participants with acute or chronic isolated unilateral ACL tears Group allocation data for 66 participants (after excluding lost to follow‐ups and graft re‐ruptures) PT: n = 31 (17 acute, 14 chronic), mean 32 years (14 to 48 years), 26 males / 7 females HT: n = 35 (14 acute, 21 chronic), mean 30 years (14 to 53 years), 21 males / 16 females
Interventions	Arthroscopically‐assisted ACL reconstruction with: 1. Patellar tendon (single incision, proximal fixation ‐ round‐head cannulated, noncutting, metal interference screw; tibial fixation ‐ interference screw) versus 2. Hamstring tendon: semitendinosus/gracilis tendons (4 strands; single incision, proximal fixation ‐ round‐head cannulated, noncutting, metal interference screw; tibial fixation ‐ interference screw)
Outcomes	Mean 33 months follow‐up for: Lysholm, KT‐1000 (89N,134N, manual maximum), range of motion, pivot shift; return to activity, patient satisfaction, Cincinnati, Lachman (mean side to side difference), return to sport, thigh circumference, patellofemoral pain
Notes	No additional methodological information or individual patient data were received from the authors.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Treatment allocation by birth date.
Allocation concealment (selection bias)	High risk
Blinding (performance bias and detection bias) All outcomes	High risk
Incomplete outcome data (attrition bias) All outcomes	High risk	Complete clinical data was not available for 13 of the patients available for follow‐up. These patients completed questionnaires by mail only.
Selective reporting (reporting bias)	High risk
Other bias	High risk	Baseline characteristics for all patients entered into the trial were not reported.

Methods	Treatment allocation by alternate systematic sampling Not independent, non blinded assessment Minimum follow‐up: 60 months Loss to follow‐up: none Exclusions from analysis: none Graft re‐ruptures: none
Participants	Italy 50 participants with acute or chronic (1 to 13 months) isolated unilateral ACL tears PT: n = 25, mean 30.5 years (22 to 47 years), 16 males / 9 females HT: n = 25, mean 31.3 years (26 to 49 years), 15 males / 10 females HT and extra‐articular plasty: n = 25, mean 26.7 years (15 to 44 years), 18 males / 7 females
Interventions	Arthroscopically‐assisted ACL reconstruction with: 1. Patellar tendon (single incision, proximal and tibial fixation ‐ interference screws) versus 2. Hamstring tendon: semitendinosus/gracilis tendons (4 strands; single incision; proximal fixation ‐ EndoButton; tibial fixation ‐ interference screw)
Outcomes	Minimum 60 months follow‐up for: Tegner, KT‐2000 (maximum manual force and 134 N); range of motion, Lachman (0/1+/2+/3+); pivot shift; single‐leg hop; time required to return to sport and activity, complications, anterior knee pain, IKDC, radiographic, thigh circumference
Notes	A third intervention group involving HT with extra‐articular plasty procedure in 25 patients is not included in this systematic review and meta‐analysis. Additional methodological information received comprised a confirmation of the number of re‐ruptures, graft failures and contralateral ruptures, and individual patient data for the following outcomes: Lachman, pivot shift test, re‐rupture/revisions.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Treatment allocation by alternate systematic sampling. The authors did not define what is meant by "alternate systematic sampling."
Allocation concealment (selection bias)	High risk
Blinding (performance bias and detection bias) All outcomes	High risk
Incomplete outcome data (attrition bias) All outcomes	Low risk
Selective reporting (reporting bias)	High risk	Did not report contralateral ruptures or withdrawals.

PERMALINK

Patellar tendon versus hamstring tendon autograft for anterior cruciate ligament rupture in adults

Nicholas GH Mohtadi

Denise S Chan

Katie N Dainty

Daniel B Whelan

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Patellar or hamstring tendon grafts for ACL reconstruction in adults

Background

Description of the condition

Description of the intervention

Possible benefits and risks of patellar tendon and hamstring grafts

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Primary outcomes

Secondary outcomes

Timing of outcome assessment

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Measures of treatment effect

Dealing with missing data

Assessment of heterogeneity

Data synthesis

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

Results

Description of studies

Results of the search

Included studies

Design

Sample sizes

Setting

Participants

Table 1.

Analysis 1.25.

Analysis 1.26.

Analysis 1.27.

Interventions

Outcome measures

Primary outcomes

Functional assessments

Return to activity/level of sport participation

Subjective knee scores

Patient satisfaction

Cincinnati and Anterior Cruciate Ligament Quality of Life (ACL‐QOL)

Secondary outcomes

Complications /adverse outcomes/recurrent injury with and without reoperation

Static stability measures

Clinical composite scores (IKDC ‐ International Knee Documentation Committee)

Range of motion

Strength testing

Pain /Anterior knee symptomatology

Risk of bias in included studies

Figure 1.

Allocation

Blinding

Incomplete outcome data

Selective reporting

Other potential sources of bias

Effects of interventions

Primary outcomes

Functional assessments

Study	Reason for exclusion
Bach 2000	Abstract only that reported interim results: two year follow‐up data for only 41% of the trial participants. No response from trial author.
Beard 2001	Twelve month follow‐up only.
Callaway 1994	Abstract only with insufficient data for analysis. No response from trial author.
Carter 1999	Only six month follow‐up and only reported on isokinetic testing.
Hantes 2004	Twelve month follow‐up only. Insufficient information available in abstract.
Mologne 1996	Abstract only reporting preliminary results at six months. No response from trial author.
Sato 2005	Less than two year follow‐up; only measured anterior tibial translation during isokinetic concentric contraction exercises.

Trial name or title	A comparison of patellar tendon and hamstring tendons for ACL reconstruction using similar femoral and tibial fixation methods: A randomized clinical trial
Methods	Randomized controlled trial
Participants	64 Patients: 32 PT 32 HT Military patients 11 females, 53 males
Interventions	PT 10 mm endobutton and bioscrew femoral fixation; Tibial fixation bioscrew and sutures over screw and washer HT quadruple ST and Grac with same fixation as PT
Outcomes	Lysholm; SANE; IKDC subjective and objective; KOOS; recurrence of instability; re‐injury; return to activity; heel height diff; ROM; thigh girth; single leg hop distance and time; Biodex at 60 and 300 deg/sec; KT 2000; X‐rays at 2 years;
Starting date	August 2000 to May 2003
Contact information	dean.taylor@duke.edu; Phone (919) 684‐6603
Notes	Still getting follow‐up information; not yet published (in 2008).