Clarifying power‐mediated aerosol effects and enhancing statistical rigor in <i>in utero</i> E‐cigarette exposure research

We thank Wang et al. (2025) for their interest in our work and suggestions to further contextualize the evidence we reported in Mills et al. (2025) with a primary focus to clarify the dose-dependent relationship between electronic cigarette (E-cig) device power and aerosol mass concentration (mg/m3). Herein, as suggested, we provide an analysis of covariance (ANCOVA) using the maximum acetylcholine (ACh)-induced vasodilatory response as the dependent variable, power (5 vs. 30 W) and nicotine (0 vs. 50 mg/ml) as independent variables, and in-chamber aerosol concentration as a covariate. We also include partial η2 for power and nicotine for each age strata as well as overall results with all ages combined (Table 1). In general, we observed the main effect for E-cig device power was lost when including aerosol concentration as a covariate, suggesting that the power setting has a significant influence on exposure burden resulting in reduced middle cerebral artery (MCA) reactivity. Based on this result it may be tempting to assume that E-cig device power has a linear dose-dependent effect. However, in a previous report (Burrage et al., 2021), we used an ∼17 W exposure paradigm with maternal exposure during pregnancy, and when comparing the deficits in adult offspring MCA reactivity we find that 17 W exposure equates more closely 30 W exposure (∼50% vs. ∼55% deficit, respectively) compared to 5 W exposures (∼30% deficit). As the literature is replete in studies showing the non-linear correlations for thermal breakdown compounds and chemicals found in E-cig aerosols at varying wattage/heating temperatures, it remains difficult to conclude that our effects are solely an aerosol dose-related mechanism, but rather that there is probably a threshold effect – where below the threshold aerosol concentration (i.e. dose) may have a more dominant influence, and above the threshold the increasing aerosol concentration has no further effect (i.e. perhaps because the maximum vascular insult might have been achieved). Unfortunately, since the methodology in characterizing maternal aerosol exposure was different between Burrage et al. (2021) using 17 W and Mills et al. (2025) using 5 and 30 W, we are not able to provide a robust statistical characterization using a common aerosol concentration outcome. Comparison between these studies is further complicated in the different e-liquid constituents (flavour vs. no flavour additive) and the differing length of maternal exposure between the studies. Wang et al. (2025) also suggested that conducting a two one-sided test (TOST) for equivalence might provide a more robust interpretation with respect to the non-significant nicotine effects we report using the more conventional null hypothesis testing. In Table 2, we provide the results of TOST analysis and Hedge's g (effect size) across each age strata from our data (Caldwell, 2022). We used ±1 SD as the equivalence bounds. In most cases (i.e. four of the seven conditions), we observed the TOST analysis (comparing no nicotine vs. nicotine-exposed offspring) showed results at the 0.10 level or higher, suggesting equivalence, and therefore agreeing with Welch's two-sample t test (of no significant difference). The three instances where results showed differing interpretation for P-values were not consistent with age strata or E-cig power condition (5 vs. 30 W), and therefore we believe this additional analysis supports our interpretation that nicotine does not appear to have a significant influence on the impaired MCA reactivity we observed. In addition, tests across all age groups yielded significant TOST results (P < 0.01 for both 5 and 30 W), indicating equivalence. The average absolute value of Hedge's g effect size across age strata was 0.395 for 5 W offspring and 0.618 for 30 W offspring, suggesting small to medium effects (assuming >0.2 for small and >0.5 for medium effects). Lastly, it was suggested we normalize the max ACh response to the passive maximum diameter obtained for myogenic response curves to strengthen the conclusion of functional endothelial impairment over morphological differences. While this is an intriguing suggestion, this approach is complicated by the very different bath conditions (with and without Ca2+) the vessels are subject to between these tests. We suggest, in addition to the active and passive myogenic data we showed (indicating no difference between exposure groups), that a more straightforward and well-established approach would be to calculate the ‘elastic modulus’ from the circumferential wall stress and strain that can be obtained from myogenic response data. Here we find in 12-month-old offspring that the elastic modulus for the MCA was not significantly different between the Air and Ecig50 groups (i.e. Air = 4.28 ± 2.26; Ecig5W = 3.76 ± 0.51, Ecig30W = 6.26 ± 4.62, mean ± SD, one way ANOVA main effect, P = 0.400), indicating that the inherent vessel stiffness or resistance to elastic deformation is not different between the exposure conditions. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article. The authors declare no conflicts of interest. Writing – original draft: I.M.O., P.D.C. Data organization and statistical analysis: I.M.O., A.M., C.P.G. Writing – review and editing: A.M., C.P.G., P.D.C. NIH NIEHS: I. Mark Olfert, Paul Chantler, 1R21 ES033026-01; NIH NIEHS: Amber V Mils, F31 ES034646-01; American Heart Association (AHA): I. Mark Olfert, Paul Chantler, CSA20 35320107; HHS | NIH | National Institute of General Medical Sciences (NIGMS): I. Mark Olfert, u54 GM104942-05S1