Response by Mossavar-Rahmani and Wassertheil-Smoller to Letters Regarding Article, “Artificially Sweetened Beverages and Stroke, Coronary Heart Disease, and All-Cause Mortality in the Women’s Health Initiative”

We acknowledge that we cannot establish causality in this prospective study and that there may be residual confounding as raised by Pyrogianni et al. We clearly state these points in our paper and we have controlled for confounders such as age, race, education, health risk variables (diabetes, cardiovascular disease, high cholesterol requiring pills, hypertension, body mass index) and behaviors (smoking, alcohol, physical activity and diet) in our analyses.

Pyrogianni et al. incorrectly state that there was no trend with dose in our model 4 analyses in Table 2. In fact there was a statistically significant rising trend (P<0.011) with dose for coronary heart disease and (P<0.007) with dose for all-cause mortality.¹

We were surprised that Drs. Khan, Malik and Sievenpiper suggest our finding may be spurious, since Dr. Malik was first author on the paper published just a month after ours that supported our findings for all-cause mortality in the analyses related to the Nurses’ Health Study (NHS).² That study also found a significant trend (p=0.01). The HR, 95% CI for ≥2 artificially sweetened beverage (ASB)/day reported by Dr. Malik and colleagues for the NHS cohort was 1.10 (1.04–1.16) [multivariate model 2 in Table 4] compared with our data: 1.19 (1.07–1.32) [Table 3 with exclusions for diabetes & heart disease]. These new findings add to the evidence of an adverse effect of high ASB consumption on health outcomes.