In Reference to <i>NAD+ Enhanced on Hearing Recovery in Sudden Sensorineural Hearing Loss: Randomized Controlled Trial</i>

We read with interest the article by Gao et al. [1] who explored nicotinamide adenine dinucleotide (NAD+) as adjunctive therapy for idiopathic sudden sensorineural hearing loss (SSNHL). The study addresses an important time-sensitive otologic emergency, where earlier treatment is crucial. According to AAO-HNSF guidelines, corticosteroids should begin within days of onset, as recovery declines markedly after 7–14 days [2-5]. Meta-analyses show maximal benefit when therapy starts within 3 days (odds ratio 1.5–2.5 vs. ≥ 7 days), underscoring the need for strict temporal control in randomized trials. While Gao et al. [1] advance an interesting therapeutic concept, several methodological refinements could strengthen future studies. First, reasons for loss to follow-up were not reported. Because attrition in SSNHL is rarely random, outcome validity may be compromised even if accounted for in sample size. The control group could appear worse if early recoveries were disproportionately lost. An explicit intent-to-treat (ITT) analysis—recommended by CONSORT—would reduce bias from incomplete data [6]. We also noted heteroscedasticity in “onset-to-treatment time” (F = 4.01; p ≈ 0.006), indicating unequal variance of this prognostic factor. Second, clarification on blinding integrity would strengthen internal validity. It remains unclear whether control participants received an identical intramuscular placebo, potentially affecting outcomes through differential placebo effects. Third, hypothesis testing of baseline variables in Table 1 of the original article contradicts reporting standards [6, 7]. Absence of statistical difference does not imply equivalence; detecting a 2-day onset difference would require ~186 subjects (93 per group). For the primary PTA outcome, ~60 participants would be needed for similar power. Ordinal variables such as Siegel's recovery categories [8] are better analyzed using proportional-odds logistic models, avoiding penalization of poorer baselines. This NAD+ RCT used fixed Siegel criteria (I–IV) for outcomes. Given the small sample (n = 38) and potential baseline imbalance—milder cases in the NAD+ arm—this fixed dichotomy may have inflated Type I error. Generally, the absence of a sliding dichotomy increases Type II error, but here the imbalance likely inflated both. Simulation showed α rising from 5% to ~16%, while a Siegel-based sliding dichotomy (adjusted by baseline PTA and WRS) restored α ≈ 5% [9]. In conclusion, Gao et al. [1] contribute meaningfully to SSNHL research. Future RCTs should adopt ITT analysis, temporal adjustment, sliding dichotomy evaluation, and ordinal outcome modeling to improve validity and error control. The authors have nothing to report. The authors declare no conflicts of interest. The data that support the findings of this study are available from the corresponding author upon reasonable request.