A Commentary on “Effectiveness of local anesthetic application methods in postoperative pain control in laparoscopic cholecystectomies; a randomized controlled trial” (Int J Surg 2021; 95: 106134)

Dear Editor, In a randomized double-blind controlled trial including 160 patients who underwent laparoscopic cholecystectomy, Ergin et al. studied the effectiveness of local anesthetic injection (LAI), transversus abdominis plane block (TAPB), intraperitoneal local anesthetic injection (IPLA) and control with no local anesthetic on postoperative pain [1]. Although the article clearly stated that TAPB resulted in better pain management than LAI and IPLA, attention should be paid to the methodological and clinical concerns of this article. First, a substantial deficiency in the design of this study is the lack of clear identification of primary and secondary outcomes. If all outcomes were treated as co-primary outcomes in this study, the threshold for type I error rate should first be adjusted for all the co-primary outcomes according to the Bonferroni-Holm correction i.e. if the number of co-primary outcomes is n, for a 2-tailed statistical test, the type I error rate should be adjusted to 0.05/n to maintain an overall familywise error rate of less than 0.05 [2,3]. Besides, even if the Visual Analog Scale (VAS) pain scores was considered as the only primary outcome, the threshold for type I error rate should also be adjusted due to repeated measurements and multiple comparisons according to the Bonferroni-Holm adjustment (i.e., 99% CI, P < 0.01), rather than using a P-value< 0.05 as the threshed of statistical significance for the whole study [3]. Second, there are major concerns on the unjustified and mis-specified use of statistical methods in this study. Although the authors stated that the distribution normality of the parameters was assessed using the Shapiro–Wilk’s test, this description is far from adequate. As statistical description should always be the first step in data analysis, the authors should provide details on the types of data description. For example, continuous data should be presented as mean ± standard deviation (SD) for normally distributed variables, and median with inter-quartile-range (IQR) for non-normally distributed data. Categorical variables should be summarized as numbers (proportions). Furthermore, a generalized estimating equation (GEE) model with robust standard error estimates should be more appropriate to use for repeated measurement of VAS data [4]. Also, whether all patients reported pain scores should be stated clearly by the authors; otherwise, sensitivity analyses should be performed to evaluate the statistical impact on missing data (VAS = 0 was considered as missing data). Analyzing data with defective statistical methods can yield biased estimates of treatment effects. Third, failure of the authors to interpret results based on minimal clinically important difference (MCID) is another concern as MCID is critical to clinical interpretation. To our knowledge, a clinical difference between treatments of 1–1.3 points is the established MCID for postoperative pain [3,5]. Given all the above concerns, re-quantification of the primary outcome (VAS pain scores) and re-evaluation of the conclusions drawn by the authors of this study were carried out by us based on MCID and statistical threshold corrections. The results showed that the different local anesthetic methods used in laparoscopic cholecystectomy produced significant clinical benefits in preventing postoperative pain when compared to the control group. Such findings are consistent with the results obtained by the authors (Table 1). However, the efficacy of pain relief at all time-points were no worse in the LAI than the TAPB group. Specific, integrative analyses showed that TAPB did not significantly reduced post-surgery pain scores at 1 hour (WMD, −0.22 cm; 99% CI, −1.72 to 1.28, p = 0.71), 2 hours (WMD, 0.17 cm; 99% CI, −0.89 to 1.33, p = 0.68), 4 hours (WMD, 0.13 cm; 99% CI, −1.04 to 1.30, p = 0.78), 6 hours (WMD, 0.00 cm; 99% CI, −1.26 to 1.26, p = 1.00), 12 hours (WMD, 1.02 cm; 99% CI, −0.03 to 2.07, p = 0.01) and 24 hours (WMD, 0.62 cm; 99% CI, −0.27 to 1.51, p = 0.07) when compared to the LIA group (see Table 1). Such findings are not in keeping with the findings of the authors. More importantly, these authors also reported that there was no significant difference between the LAI and TAPB groups in the need for additional analgesic (p > 0.05). Furthermore, when compared with IPLA, our analysis suggested that there was clinical superiority in using LAI and TAPB in prevention of postoperative pain (Table 1). Overall, comparison of the methods of local anesthetic administration showed different results from the findings by the authors that TAPB facilitated better pain management than the other two methods and the findings by the authors of this study has produced incorrect and misleading results.Table 1: Comparison of VAS among the study groups.We applaud the authors’ hard work. However, the review by Ergin et al. drew unreliable findings which can lead to exposure of patients to unreasonable treatments. Thus, correction of the stated faults is helpful for better interpretation of the reported results. Ethical approval Not Applicable. Sources of funding None. Author contribution Jing Dong: Conceptualization, Formal Analysis, Writing-Original Draft, Validation. Yi-Feng Ren: Conceptualization, Formal analysis, Writing-Original Draft, Writing-Review and Editing, Supervision, Software, Validation. Wei Shi: Conceptualization, Writing-Review and Editing, Supervision, Software, Validation. Research registration unique identifying number (UIN) Not Applicable. Guarantor Yi-Feng Ren and Wei Shi. Provenance and peer review Commentary, internally reviewed. Declaration of competing interest None.