To the Editor—Machado et al.  recently reported the results of a double-blind, placebo-controlled, randomized clinical trial that compared the effectiveness of pentoxifylline and antimony dual therapy with that of antimony monotherapy in patients with mucosal leishmaniasis. The authors concluded that dual therapy reduced healing time and was superior to antimony monotherapy for leishmaniasis eradication. However, the results should be interpreted cautiously. Critical methodologic limitations warrant attention to foster appropriate clinical interpretation.
The study by Machado et al.  illustrates the inability of small-sample randomization to allocate groups with comparable baseline risks for the outcome. Participation was limited to 23 participants overall—11 participants in the intervention group and 12 participants in the control group. Randomization of small study populations may introduce accidental bias from failure to maintain equal distribution of confounders between comparison groups, regardless of statistical significance . Machado et al.  reported a lack of statistically significant differences between the 2 groups. However, the study was insufficiently powered to observe statistically significant differences, and the observed clinically relevant differences between groups suggest the introduction of bias. Previous studies have reported that older age, male sex, longer duration of symptoms, and previous cutaneous leishmaniasis may be related to risk of mucosal leishmaniasis, disease severity, and time to cure [3, 4]. The antimony plus placebo group was older and had a greater proportion of male participants, longer duration of symptoms, and greater frequency of cutaneous leishmaniasis, compared with the antimony plus pentoxifylline group; these findings indicate that the pentoxifylline group had a greater probability for eradication at baseline than did the placebo group. Consequently, a greater probability of cure in the group who received dual therapy, compared with the group who received monotherapy, may have created a positive bias away from the null (i.e., overestimated the effectiveness of dual therapy). Accidental bias offers a plausible alternative explanation for the authors' findings, emphasizes the inadequacy of hypothesis testing to evaluate the comparability of small groups [5, 6], and underscores the necessity for a larger randomized trial.
It is unclear why Machado et al.  did not provide an effect estimate, such as risk difference or hazard ratio, using robust statistical techniques for small trials . The failure to report an effect estimate undermines comparison. Furthermore, effect estimates facilitate cost-effectiveness evaluations, which are particularly important for resource-limited regions where leishmaniasis commonly occurs. Effect estimates with corresponding confidence intervals for the outcome would have more appropriately elucidated the potential clinical use of dual therapy.
The small sample size used by Machado et al.  greatly limited potential inferences because of the failure to produce comparable groups after randomization and compromised statistical testing. Therefore, the study may not constitute decisive evidence for the superiority of pentoxifylline and antimony dual therapy for mucosal leishmaniasis. Future studies should recognize the limitations of hypothesis testing and P values to evaluate whether small groups are comparable after randomization [5, 6]. Ultimately, small sample sizes may lead to inconclusive inferences and warrant conservative interpretations from a clinical perspective until further evidence is available.
Potential conflicts of interest. All authors: no conflicts.