I read the study by Bolia et al.
1
with great interest. As the annual number of hip arthroscopy procedures performed is increasing, there is a clear need for studies investigating outcomes of this procedure. Because the technique itself is still evolving, it is important to investigate surgery-related factors to optimize the outcome of this procedure. One controversial issue is the management of capsulotomy performed during the surgery.2
,3
Although the hip joint is an inherently stable joint, there is emerging concern about the effect of capsulotomy on postoperative stability of hip joint, especially if the capsulotomy is left unrepaired.4
Capsulotomy resulting from a violation of ligamentous structures is suggested to result in long-term deterioration in hip function ultimately requiring a conversion to a total hip arthroplasty (THA).5
Current evidence for this is conflicting and the exact pathomechanism of this process is poorly established but is under extensive investigation.Bolia et al. sought to examine the conversion rate to THA in patients having undergone a repair of capsulotomy compared with those without a repair. Their study was a case-control study in which 42 patients without a repair were matched to 84 patients with a repair. The authors reported that 6 of 42 patients (14%) without a repair and 3 of 84 (4%) patients with repair were converted to a THA. This result was accompanied by a P value of .01. This result was further communicated as “Patients in the nonrepair group were 6.8 times (95% confidence interval, 1.2-52) more likely to require THA than patients in the repair group.”
I have 2 concerns regarding their main outcome result. First, expression of “times more likely” is a probabilistic statement that is inadequate in case-control studies. Second, and most important, the authors do not report any test to handle this binary data in the Methods section.
Using Fisher’s 2-sided exact test, the associated P value is .059. Chi-square test with and without Yates correction results in P values of .067 and .028, respectively. Using univariate logistic regression, patients without a repair has a crude (unadjusted) odds ratio of 4.5 (95% confidence interval, 1.1-22.3, P = .041) for conversion to THA.
As can be seen, statistical significance of the main result is located just around the borderline value of .05 depending on the test used. Interpretation of such a result is full of caveats, such as labeling the finding as “significant” or “not significant.”
6
Small sample size and low event rate indicate high uncertainty around the point estimate for THA conversion, which should be appreciated to avoid P value fallacy. Methods to handle binary data, however, are not reported by the authors, which is very concerning. Thus, the robustness of the main outcome result about THA conversion cannot be evaluated.Supplementary Data
- ICMJE author disclosure forms
References
- Midterm outcomes following repair of capsulotomy versus nonrepair in patients undergoing hip arthroscopy for femoroacetabular impingement with labral repair.Arthroscopy. 2019; 35: 1828-1834
- Should the capsule be repaired or plicated after hip arthroscopy for labral tears associated with femoroacetabular impingement or instability? A systematic review.Arthroscopy. 2018; 34: 303-318
- Hip arthroscopic capsulotomy techniques and capsular management strategies: a systematic review.Knee Surg Sports Traumatol Arthrosc. 2017; 25: 9-23
- Arthroscopic capsulotomy, capsular repair, and capsular plication of the hip: Relation to atraumatic instability.Arthroscopy. 2013; 29: 162-173
- Patient-reported outcomes of capsular repair versus capsulotomy in patients undergoing hip arthroscopy: Minimum 5-year follow-up-a matched comparison study.Arthroscopy. 2018; 34: 853-863
- Toward evidence-based medical statistics. 1: The P value fallacy.Ann Intern Med. 1999; 130: 995-1004
Article info
Footnotes
Note: The author reports that he has no conflicts of interest in the authorship and publication of this article. Full ICMJE author disclosure forms are available for this article online, as supplementary material.
Identification
Copyright
© 2019 by the Arthroscopy Association of North America
ScienceDirect
Access this article on ScienceDirectLinked Article
- Midterm Outcomes Following Repair of Capsulotomy Versus Nonrepair in Patients Undergoing Hip Arthroscopy for Femoroacetabular Impingement With Labral RepairArthroscopyVol. 35Issue 6
- Author Reply to “Regarding ‘Midterm Outcomes Following Repair of Capsulotomy Versus Nonrepair in Patients Undergoing Hip Arthroscopy for Femoroacetabular Impingement With Labral Repair’”ArthroscopyVol. 35Issue 11
- PreviewWe appreciate Dr. Reito’s comments on our recent publication. We apologize for the incorrect numbers in the manuscript. These numbers were for 2 failures in the repair group; however, as the final paper stated, we had 3 failures. We used Pearson’s chi square for our P value, which was .028, using SPSS, version 25. We updated the data and changed the odds ratio. The odds ratio should have been 4.5 (95% confidence interval, 1.1-19). We will submit a correction to Arthroscopy immediately. In addition, we agree that the statement “times more likely to” is a suboptimal statement; however, it is commonly used in the literature.
- Full-Text
- Preview
