Abstract
We read with great interest the article titled “Comparable Instrumented Knee Joint Laxity and Patient-Reported Outcomes After ACL Repair With Dynamic Intraligamentary Stabilization or ACL Reconstruction: 5-Year Results of a Randomized Controlled Trial,” by Glasbrenner et al. 2 We congratulate the authors on their work and the effort put into this prospective randomized controlled trial comparing dynamic intraligamentary stabilization (DIS), a novel technique for anterior cruciate ligament (ACL) repair, with traditional ACL reconstruction (ACLR).
From an initial cohort of 85 patients (DIS, n = 43; ACLR, n = 42), instrumented knee joint laxity measurements and patient-reported outcomes (PROs) were obtained 5 years postoperatively. A total of 64 patients (75%) completed final follow-up; 35% of the patients in the DIS group and 20% in the ACLR group had recurrent instability. Increased failure rates were observed in young and active patients. The overall rate of revision surgery was 38% (29% ACL revision) in the DIS group and 27% (17% ACL revision) in the ACLR group. After exclusion of all patients with recurrent instability, a statistically significant difference between groups was found only with the International Knee Documentation Committee subjective knee form, albeit within the minimal clinically important difference. The authors concluded that 5 years after ACL repair with DIS, instrumented knee laxity and PROs were comparable with those of ACLR, but recurrent instability and revision surgery rates were high. The authors further concluded that ACL repair with DIS is a “feasible option to treat acute ACL tears in patients aged >25 years with low to moderate activity levels (Tegner score <7).”
After reading the article, we have several concerns about the analysis and interpretation of the data, as well as the conclusions drawn.
First, it is important to note that based on the 2009 Oxford level of evidence guidelines, a therapeutic randomized controlled trial can be considered level 1 evidence only if the follow-up rate is >80%. 4 In evidence-based medicine, a correct statement of the level of evidence is of utmost importance, for future studies citing or analyzing the work and to appropriately direct therapeutic decisions. Based on the aforementioned guidelines, the present study corresponds to level 2b evidence.
Furthermore, young age and high activity level were mentioned as risk factors for recurrent instability in the DIS and ACLR groups, but no statistical analysis was provided to support this. Particularly given the small sample sizes in this study, it is important to demonstrate statistical significance rather than make assumptions based on numbers so that surgeons are appropriately guided in their decision-making processes.
Further concerns arise with the overall 25% loss to follow-up. Of the remaining 64 patients, 18 were excluded because of recurrent instability or reoperation. This left only 46 patients (54.1%) for final analysis and represents a failure to follow the concept of intention to treat. While stated in the limitations, these numbers were not highlighted in the Results section, figures, or tables and could therefore distort interpretation of the results. An additional analysis of the patients lost to follow-up should have been performed to provide a more complete picture of the entire patient cohort and to contextualize the results.
Finally, in the abstract, the authors hypothesized that “ACL repair would lead to comparable knee joint stability and PROs at 5 years postoperatively in comparison with ACL reconstruction,” and they concluded that “instrumented knee joint laxity after ACL repair with DIS was not inferior to that after ACL reconstruction, with ΔATT [anterior tibial translation] <3 mm in both groups at 5 years postoperatively.” Yet, the authors performed a sample size analysis based on a superiority hypothesis and expected a mean side-to-side difference in anterior tibial translation of 3 mm with an SD of 2 mm and loss to follow-up of 15%. The effect size for their sample size analysis was 1.5, which is unreasonably large, resulting in a sample size with insufficient power to detect clinically meaningful differences in anterior tibial translation. We performed a sample size analysis for noninferiority, aligning with the authors’ hypothesis, and found that the sample size was insufficient to demonstrate noninferiority of ACL repair versus ACLR. Specifically, for a target power of 80%, a noninferiority margin of 1 mm (half the original assumed standard deviation), and the assumed SD of 2 mm, the total number of patients required to demonstrate noninferiority would be 128 at 5-year follow-up (1-sided test; α = .025). Based on an assumed dropout rate of 20%, the number of patients required for initial enrollment would need to have been 160, nearly twice as many as were randomized in this trial.1,3
Given the data derived from their initial sample size analysis, the authors concluded that instrumented knee laxity and PROs are comparable between patients undergoing DIS and ACLR. They further concluded that young age and high activity level are risk factors for recurrent instability, without providing a supporting statistical analysis. We find it troubling that an underpowered study could draw such strong conclusions and, moreover, place these conclusions in the title and abstract. It is important for readers to understand that the failure to find significant differences with an underpowered study does not imply that the results of ACL repair utilizing DIS are comparable with standard ACLR. At best, all that the authors can conclude is that with the available sample size, they were unable to detect significant differences between the approaches, which does not imply that the 2 interventions were comparable. In fact, we would argue that the observed absolute difference of 15% for recurrent instability (35% vs 20%) and 11% for revision surgery (38% vs 27%) in favor of ACLR over DIS is clinically meaningful even though it was not statistically significant.
From a scientific standpoint, we would have liked to see a clear statement of the total number of patients being analyzed for each outcome variable, an accurate representation of the study's level of evidence and power, and a more neutral title to appropriately convey the study findings to the reader. This is particularly important given the clinical implications of the study's conclusions.
Once again, we congratulate the authors on their valuable work on an important topic. We encourage the authors to update their article and look forward to reading their future work to advance the field of ACL treatment and ACL repair.
Footnotes
The authors declared that they have no conflicts of interest in the authorship and publication of this contribution. AOSSM checks author disclosures against the Open Payments Database (OPD). AOSSM has not conducted an independent investigation on the OPD and disclaims any liability or responsibility relating thereto.