Association between passing return-to-sport testing and re-injury risk in patients after anterior cruciate ligament reconstruction surgery: a systematic review and meta-analysis

Data sources, search strategy, and selection criteria

This review was conducted and reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis Statement (Page et al., 2021). Studies that examined the impact of the RTS test on the risk of subsequent re-injury for athletes after ACLR were eligible for inclusion in our analysis, with no restrictions on the publication language. We searched electronic databases (MedLine, EmBase, and Cochrane Library) for articles published from the inception of the databases to September 2023. The following core search terms were used: “anterior cruciate ligament reconstruction” OR “ACL reconstruction” AND “return to sport” OR “return to sport criteria” OR “return to play” OR “return to play criteria” OR “functional testing” OR “return to athletic*”. The details of the search strategy are presented in Supplemental Material 1. Additionally, we manually searched the reference lists of relevant articles to identify additional eligible studies. The study selection was conducted based on the medical subject heading, methods, patient population, design, exposure, and outcome variables.

The literature search and study selection process were independently conducted by two authors (Wenqi Zhou and Qiaomei Hong) using a standardized approach. Disagreements were resolved by the first author until a consensus was reached. The inclusion criteria were based on the following predefined eligibility criteria: (1) Participants: post-ACLR athletes; (2) Exposure: passing RTS testing; (3) Control: failure in passing RTS testing; (4) Outcomes: knee injury (all knee injuries and ACL injury), secondary ACL injury (defined as contralateral ACL injury and graft rupture), contralateral ACL injury, or graft rupture; and (5) Study design: no restrictions were placed on study design, and prospective and retrospective design were eligible. Detailed methodology can be referred to previous words (Chen et al., 2023; Li et al., 2021; Zhang et al., 2021a; Zhang et al., 2021b). Animal experiments, reviews, and case reports were excluded as they could not provide sufficient data.

Data collection and quality assessment

The data collected included the first author’s name, publication year, study design, country, sample size, male percentage (%), mean age (years), physical condition of the participants, RTS test battery, RTS test time and pass rate, and reported outcomes. The methodological quality of each individual study in the meta-analysis was assessed using the Newcastle-Ottawa Scale (NOS), which consists of selection (four items), comparability (one item), and outcome (three items) criteria (Wells et al., 2009). A “star system” ranging from 0 to 9 was used to evaluate each study. These assessments were independently performed by two authors (Wenqi Zhou and Qiaomei Hong), and any disagreements were resolved through a review of the full-text of the original articles by an additional author (Xiaobing Luo), followed by a discussion (Dutaillis et al., 2003; Li et al., 2022; Zhang et al., 2021a; Zhang et al., 2021b).

Statistical analysis

We examined the association of RTS testing with the risk of knee injury, secondary ACL injury, contralateral ACL injury, and graft rupture considering the incidence rate and the sample size in each individual study. Then, the pooled odds ratios (ORs) and 95% confidence intervals (CIs) were calculated using the random-effects model (Ades, Lu & Higgins, 2005; DerSimonian & Laird, 1986). The heterogeneity across the studies was analyzed using the I² and Q statistic; the significance of heterogeneity was determined using I² >50% or P <0.10 (Higgins et al., 2003; Deeks, Higgins & Altman, 2008). The robustness of the pooled conclusions was assessed through a sensitivity analysis (Tobias, 1999). Subgroup analysis was also conducted based on study design, country, percentage male, mean age, and study quality. The differences between subgroups were compared using the interaction P test. The publication bias for each investigated outcome was assessed through visual inspections of funnel plots, and quantitatively assessed using the Egger’s and Begg’s tests (Begg & Mazumdar, 1994; Egger et al., 1997). All reported P- values are two-sided; P <0.05 was considered to indicate statistical significance. Statistical analyses were performed using STATA software (version 12.0; Stata Corporation, College Station, TX, USA).

Literature search

A total number of 1,591 articles were identified in our initial electronic search, 1,527 of which were excluded due to duplication and not fitting the inclusion criteria. The remaining 64 studies were retrieved for further full-text evaluations, of which 55 were excluded for the following reasons: inappropriate control (n = 26), insufficient data (n = 21), and reviews (n = 8). Further, nine studies were selected for the final meta-analysis (Graziano et al., 2017; Grindem et al., 2016; Kyritsis et al., 2016; Nawasreh et al., 2017; Paterno et al., 2022; Raoul et al., 2019; Sousa et al., 2017; Webster & Feller, 2019; Wellsandt, Failla & Snyder-Mackler, 2017). The manual search for reference lists yielded eight potentially relevant articles, but no eligible study was detected after detailed evaluations (Fig. 1). The baseline characteristics of the included studies and individuals are presented in Table 1.

Study characteristics

Of the nine included studies (a total number of 1,410 individuals), seven were prospective cohort studies, and the remaining two were retrospective studies. The mean age range of the athletes was 12.0–28.4 years, and 42–329 athletes were included in each individual study. One study included only male athletes, whereas the remaining eight studies included both male and female athletes. Five studies were conducted in USA, whereas the remaining four studies were performed in Europe or Australia. Study quality was assessed by using the NOS scale. Three of the studies obtained eight stars, three studies seven stars, and the remaining three studies received six stars.

Knee injury

A total number of six studies reported the association between passing RTS test and the risk of knee injury. RTS test passing by athletes was not associated with the risk of knee injury as compared with individuals failing the RTS test (OR: 0.95; 95% CI [0.28–3.21]; P = 0.929; Fig. 2). Significant heterogeneity was detected (I² = 82.6%; P <0.001). Our sensitivity analysis found the pooled conclusion was robustness and not altered by sequential excluding individual study (Supplemental Material 2). For pooled studies conducted in Europe, the subgroup analysis found that passing the RTS test was associated with a lower risk of knee injury. However, it was associated with an increased risk of knee injury in pooled studies performed in the USA or in studies with a lower quality (Table 2). The interaction test results showed that the passing of the RTS test with the risk of knee injury could affect by study design, country, percentage male, and study quality. There was no significant publication bias for knee injury (P- value for the Egger’s test: 0.558; P- value for the Begg’s test: 0.452; Supplemental Material 3).

Secondary ACL injury

After pooling all included studies, we found that passing a RTS test by athletes did not affect their risk of secondary ACL injury (OR: 0.98; 95% CI [0.55–1.75]; P = 0.945; Fig. 3). A significant heterogeneity among the included studies was detected (I² = 53.1%; P = 0.030). Thus, sensitivity analysis was conducted for secondary ACL injury. Our findings were not affected by the exclusion of any specific study from the pooled analysis (Supplemental Material 2). The results of subgroup analyses of pooled studies conducted in Europe or Australia showed that passing an RTS test was associated with a reduced risk of secondary ACL injury, whereas an increased risk was detected in the pooled studies performed in the USA. Additionally, study design, country, percentage male, and study quality could affect the role of passing RTS with the risk of secondary ACL injury (Table 2). No significant publication bias for secondary ACL injury was detected (P- value for Egger’s test: 0.772; P- value for Begg’s test: 0.917; Supplemental Material 3).

Contralateral ACL injury

A total number of eight studies reported the association between passing RTS test with the risk of contralateral ACL injury. Athletes passing RTS test was not associated with the risk of contralateral ACL injury (OR: 1.53; 95% CI [0.63–3.71]; P = 0.347; Fig. 4). Significant heterogeneity was observed across the included studies (I² = 46.0%; P = 0.073). Sensitivity analysis found that passing an RTS test might be associated with an elevated risk of contralateral ACL injury (Supplemental Material 2). Moreover, subgroup analysis results suggested that passing an RTS test was associated with an increased risk of contralateral ACL injury c (Table 2). There was no significant publication bias for contralateral ACL injury (P- value for Egger’s test: 0.550; P- value for Begg’s test: 0.711; Supplemental Material 3).

Graft rupture

A total number of eight studies established the existence of an association between passing an RTS test and an increase in the risk of graft rupture. Passing an RTS test was associated with a reduced risk of graft rupture in athletes (OR: 0.49; 95% CI [0.33–0.75]; P = 0.001; Fig. 5). We found no evidence of heterogeneity across the included studies (I² = 0.0%; P = 0.430). The pooled conclusion was validated by the sequential exclusion of individual studies (Supplemental Material 2). The subgroup analysis results of pooled studies with a prospective design, studies conducted in Europe or Australia, those with a male proportion ≥ 60%, with a mean age <25.0 years, and pooled studies with high quality revealed that passing an RTS test was associated with a reduced risk of graft rupture in athletes (Table 2). No significant publication bias for graft rupture was detected (P- value for Egger’s test: 0.944; P- value for Begg’s test: 0.711; Supplemental Material 3).