Single versus double bundle in posterior cruciate ligament (PCL) reconstruction: a meta-analysis

Posterior cruciate ligament (PCL) reconstruction can be performed using single bundle (SB) and double bundle (DB) techniques. The present study investigated whether DB PCL reconstruction is superior to SB reconstruction in terms of patient reported outcome measures (PROMs) and joint stability. In December 2021 Embase, Google Scholar, Pubmed, Scopus databases were accessed. All clinical trials comparing SB versus DB reconstruction to address PCL insufficiency in skeletally mature patients were considered. Data from 483 procedures were retrieved. The mean follow-up was 31.0 (28.0 to 107.6) months, and the mean timespan between injury and surgery was 11.3 (6 to 37) months. The mean age of the patients was 29.3 ± 3.8 years. 85 of 483 patients (18%) were women. At a mean of 31.0 months post reconstruction, ROM (P = 0.03) was slightly greater in the SB group, while the Tegner score (P = 0.03) and the Telos stress (P = 0.04) were more favorable in the DB cohort. Similarity was found in instrumental laxity (P = 0.4) and Lysholm score (P = 0.3). The current evidence does not support the use of DB techniques for PCL reconstruction. Both methods could restore knee stability and motion with satisfactory short term patient reported outcome measures. Further high quality clinical trials are required to validate these results on a larger scale.


Material and methods
Search strategy. This meta-analysis was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses: the PRISMA guidelines 50 . The PICO algorithm was followed: • P ( Methodology quality assessment. The methodological quality assessment was performed by a single author (A.P.) using the Coleman Methodology Score (CMS) The CMS is a reliable and validated tool to evaluate the methodological quality of articles included in systematic reviews and meta-analyses 52 , evaluating the population size, length of follow-up, surgical approach used, study design, description of diagnosis, surgical technique, and rehabilitation. Additionally, outcome criteria assessment and the subject selection process were also evaluated. The quality of the studies is scored between 0 (poor) and 100 (excellent), with values > 60 considered satisfactory.
Statistical analysis. The statistical analyses were performed by the main author (F.M.). To assess baseline comparability, the unpaired t-test was performed using the IBM SPSS version 25. Values of P > 0.05 indicated similarity between the two groups. The meta-analyses were performed using the Editorial Manager Software version 5.3 (The Nordic Cochrane Collaboration, Copenhagen). Continuous data were analyzed using the inverse variance method, with mean difference (MD) effect measure. Dichotomic data were analyzed using the Mantel-Haenszel method and odd ratio (OR) effect measure. The confidence interval was set at 95% in all the comparison. A fixed model effect was set as default. If moderate or high heterogeneity was detected, a random model effect was adopted. Heterogeneity was evaluated through Higgins-I 2 and χ 2 tests. Values of Higgins-I 2 were interpreted as low (< 30%), moderate (30% to 60%), high (> 60%). Forest and funnel plot were performed. Values of P > 0.05 were considered statistically significant.
Ethical approval. This study complies with ethical standards. Methodological quality assessment. According to the CMS, the study size and length of follow-up were adequate. Surgical approach, diagnosis, and rehabilitation were well described in most articles. Outcome measures and timing of assessment were frequently defined, providing moderate reliability. The procedures for assessing outcomes, along with subject selection were often biased and poorly described. Concluding, the CMS Risk of publication bias. The funnel plot of the most commonly reported outcome (Lysholm score) was used to investigate the risk of publication bias. The plot evidenced a very good symmetrical disposition of the referral points. No study was located outside the shapes, increasing the reliability of the plot. Concluding, the risk of publication bias was low (Fig. 2).

Results
Patient demographics. Data from 483 procedures were retrieved. The mean follow-up was 31.0 ± 49.4 months, and the mean timespan between injury and surgery was 11.3 ± 39.1 months. The mean age of the patients was 29.3 ± 3.8 years. 85 of 483 patients (18%) were women. Good comparability was found at baseline between the two groups in terms of length of follow-up (P = 0.9), timespan form injury to surgery (P = 0.9), mean age (P = 0.4), women (P = 0.08), Lysholm scale (P = 0.7), ROM (P = 0.6), Tegner (P = 0.9), Telos stress radiography (P = 0.7), arthrometer laxity (P = 0.9). Generalities and baseline characteristics of the included studies are shown in Table 2.

Discussion
This meta-analysis confirmed our hypothesis of similarity between PCL reconstruction using a SB or a DB graft. DB evidenced minimally greater Telos stress and Tegner score, along with a slightly lower range of motion than the SB. No difference was found in the instrumental laxity and Lysholm score. The Tegner activity scale is a validated PROM to evaluate the level of activity of the patients [53][54][55] . Stress radiographs with the Telos stress device are widely employed to evaluate laxity of cruciate ligaments [56][57][58] . Our results indicated that the Tegner score and the results of the Telos stress were favorable in to the DB cohort; however, the clinical impact of these differences was minimal. Indeed, the MD between the two groups did not overcome the minimum clinically important difference of the Tegner scale, which was estimated between 0.5 and 1 point [59][60][61] . The instrumental laxity using the AK-1000/2000 and the functional assessment using Lysholm score were also similar, suggesting comparability between the two techniques. Several biomechanical studies stated that DB PCL reconstruction better restore antero-posterior stability than SB techniques 9,12,24,25,[62][63][64][65] . A recent biomechanical study demonstrated that a DB PCL reconstruction could better restore knee stability across the full ROM, while SB leads to high graft tension during extension and laxity during flexion 66   www.nature.com/scientificreports/ closely the physiological knee biomechanics. These results explain partially the findings of the present study. DB reconstruction achieves greater stability according to the Telos stress test, allowing patients to increase their activity level or to quickly return to preinjury level of activity 67 , but also reducing the range of motion. However, the clinical relevance of these findings is questionable, especially in light of the similarity evidenced by the Lysholm score and instrumental laxity. Concluding, even though biomechanical results are encouraging, the clinical outcomes are similar for SB and DB PCL reconstruction. To establish the optimal number of bundles which should be reconstructed, the rate of complications should be investigated. Given the lack of quantitative data concerning the rate of complications experienced by patients after SB and DB PCL reconstruction, no further analyses can be inferred. Only two studies reported data concerning complications after PCL reconstruction surgery 48,49 . Jain et al. 49 reported four patients with residual laxity and persistent sensation of instability in the SB group; however, only the 5% (1 of 22 patients) underwent revision surgery. Yoon et al. 48 reported that four patients underwent additional surgeries: one in the SB, and three in DB group. In clinical practice, DB PCL reconstruction present some disadvantages which are worthy of discussion. DB techniques theoretically expose the patients to higher risk of complications, as four drill holes and four fixation devices are required, and pitfalls are possible. Moreover, surgeons must be aware that revision surgery after DB failure may be challenging. Indeed, removal of DB grafts results in bigger bone defects than in SB reconstruction: they can necessitate larger implants for revision, two stage surgeries, higher costs and patient morbidity. Thus, given the similar outcomes, a SB PCL reconstruction may be encouraged as primary choice. It is unclear whether DB should be reserved for revision surgeries. The retrospective nature of most of the included studies is an important limitation of this eneavour. Unfortunately, only three studies were randomized clinical trials [68][69][70] , which represents an important source of selection bias. Eligibility criteria and allocation concealment between SB and DB were not clearly stated, and often biased by the studies. The analyses were conducted irrespective of the type of graft used for reconstruction and the tensioning protocol associated with the procedure, representing other important limitations. Instrumental laxity was evaluated regardless to the type of arthrometer (KT-1000 and/or 2000); however, both the instruments provide a static force to the translational displacement of 134 N. The difference between the two instruments is the duration of the test (KT-1000: 2 min and 15 s versus KT-2000: 2 min and 3 s), and the methodology of saving the resulting data (KT-1000: manual versus KT-2000: digital). Postoperative rehabilitation pattern may also change the biomechanical results, especially at last follow up 71 . However, the rehabilitation process was often biased, and only minimal between-group differences were detectable. Given these limitations, the results from the present study must be interpreted with caution. Finally, further high-quality clinical trials providing longterm follow-up are strongly recommended to establish whether this minimal greater stability affects chondral degeneration, secondary meniscus lesions, onset of osteoarthritis, and to establish the rate of complications of failure of the two bundles.

Conclusion
Current evidence does not support the use of DB techniques for PCL reconstruction. Both methods could restore knee stability and motion with satisfactory short term patient reported outcome measures. Further high quality clinical trials are required to validate these results on a larger scale.

Data availability
The datasets generated during and/or analysed during the current study are available throughout the manuscript.