Study selection
Primary search identified 513 papers from PubMed (n = 125), Embase (n = 130), Web of Science (n = 132), Cochrane Library (n = 126). After removing 259 duplicated studies, 254 studies were screened by titles, abstracts and full texts.
During the screening phase, two authors removed 219 articles based on title and abstract. The reasons for excluding these studies were as follows: not in English, follow-up less than 12 months, in vivo experiment, not level I RCTs, not anatomical DB versus SB trials, missing full texts, replicated reports, ongoing studies.
Full texts were evaluated to find eligible data in remaining 35 studies, 12 studies were excluded because techniques were not anatomical ACL reconstruction [5, 54, 77, 85, 1, 2, 29, 44, 55, 63, 65, 87], 1 article used allograft [66], 1 article did not report full outcomes [11] and 1 article had none surgical technique description [10]. Finally, 20 RCTs [9, 51, 49, 3, 81, 36, 60, 43, 39, 12, 79, 6, 67, 32, 68, 8, 61, 33, 38, 37] were included in this study. The inclusion strategy of eligible studies is presented in Fig. 1. The publication year of the included articles was from 2007 to 2022 with a minimum follow-up of 12 months.
Characteristics of included studies
Patients were randomly assigned to primary ACL reconstruction with anatomical DB or SB in each of the included studies, and they were followed for at least 12 months. The anatomical DB reconstruction consisted of two distinct bundles to replicate the AMB and PLB of native ACL, and both methods used autografts. The included studies' demographic details and surgical-related data are presented in Tables 1 and Table 2, respectively.
Risk of bias
Information describing the risk of bias for each study is presented in Fig. 2 and Fig. 3, 3 studies had unclear risk of bias [81] [60] [61]. Other 17 studies had low risk of bias [9, 51, 49, 3] [36] [43, 39, 12, 79, 6, 67, 32, 68, 8] [33, 38, 37]. Lack of information about random sequence generation, allocation concealment, blinding of participants and personnel resulted in the unclear risk of bias. Random sequence generation was adequate in 20 studies. Allocation concealment was carried out adequately in 18 studies. Blinding of participants and personnel was adequate in 17 studies.
Results of included studies
The overall outcomes reported in the included studies are represented in the Table 3.
Subjective outcomes
No statistically significant difference was found in IKDC subjective score according to the pooled results of nine studies (P = 0.37; SMD 0.17, 95% CI -0.38 to 0.72), while the I2 test of heterogeneity was high as 65% and needed to be make sensitivity analysis (Fig. 4).
The outcome of Lysholm score showed no statistically significant difference between two surgical techniques (P = 0.65; SMD 0.03, 95% CI -0.10 to 0.16). The I2 = 0% of heterogeneity analysis among twelve studies showed no heterogeneity (Fig. 5).
The difference in the outcome of Tegner activity score was statistically significant (P = 0.02; SMD − 0.20, 95% CI -0.37 to -0.04). The I2 of heterogeneity test was 24%, which showed little heterogeneity exists between seven studies (Fig. 6).
Objective outcomes
Seven studies were pooled for outcomes of IKDC objective grade, there was no statistically significant difference between the anatomical DB and the SB operation (P = 0.43; RR 1.14, 95% CI 0.82–1.60), without significant heterogeneity for the I2 was 21% (Fig. 7).
When comes to the Pivot-shift test, the analysis of thirteen trials showed that no statistical significance of the difference exists between the anatomical DB and the SB technique (P = 0.90; RR 0.98, 95% CI 0.69–1.39), and the I2 of 0% showed no heterogeneity among the included studies (Fig. 8).
A statistically significant difference was observed between two surgical techniques in Lachman test (P = 0.04; RR 0.80, 95% CI 0.65–0.99) within six trials, I2 = 41% of heterogeneity analysis showed middle heterogeneity among these enrolling studies (Fig. 8).
To assess Side-to-side differences, data from 12 studies (1062 participants) were combined. The total side-to-side difference was not statistically significant with an SMD of 0.05 (95% CI -0.07-0.17). I2 = 66% indicates that there was significant heterogeneity between the trials (Fig. 10).
The outcomes of OA change at the final follow-up examination were reported in five trials. There was no discernible difference between the anatomical DB group and the SB group, according to the results of the pooled analysis (P = 1.00; RR 1.00; 95% CI 0.74–1.35). Between these trials, there was only a little amount of heterogeneity (I2 = 11%) (Fig. 11).
Sensitivity analysis
In outcomes of IKDC subjective score, Tegner activity score and Side-to-side difference, sensitivity analyses were conducted to identify the source of heterogeneity.
After removing the study Sergi Sastre 2010 [61] in the sensitivity analysis of IKDC subjective score, the I2 decreased significantly to 2%, which also suggested that no significant difference exists between two techniques (P = 0.89; SMD= -0.01; 95% CI -0.14-0.12) (Fig. 12).
When removing the study Mattias Ahlden 2013 [6] in the sensitivity analysis of Tegner activity score, the I2 changed to 0%, and the outcome showed no significant difference between anatomical DB and SB techniques (P = 0.21; SMD= -0.12; 95% CI -0.30-0.07) (Fig. 13).
We removed the study S. A. R. Ibrahim 2009 [33] in the sensitivity analysis of Lachman test, the I2 moved to 0%, and no significant difference was found between anatomical DB and SB techniques (P = 0.38; SMD = 0.91; 95% CI 0.74–1.12) (Fig. 14).
The sensitivity analysis of the Side-to-side difference removed the study Yan Xu 2014 [79], I2 changed from 66–45%, which was middle heterogeneity, and the result suggested no significant difference between the anatomical DB and SB technique (P = 0.09; SMD = 0.11; 95% CI -0.02-0.23) (Fig. 15).