Alcohol Consumption in Adolescence on Early-Adulthood Hypertension or Prehypertension

The authors were pleased to read the recent study on the association between alcohol consumption in adolescence and early-adulthood hypertension or prehypertension in 1556 adolescent participants in China [1]. Building on the findings from our previous work involving 5114 adolescents in the United States [2], this new study reinforces the impact of heavy alcohol consumption on hypertension or prehypertension. Specifically, the study identified associations of hypertension or prehypertension with alcohol consumption: (1) equal to or over two times per week and (2) more than 96 mL per week. Additionally, the study found an association between hypertension or prehypertension and the intake of two or more types of alcohol (beer, wine, and liquor) [1].

Methodologically, the study adopted a novel approach by applying the restricted cubic spline model. Particularly when alcohol intake exceeded 25 mL per week, the model demonstrated a dose-response relationship, showing an increasing risk of hypertension or prehypertension as alcohol consumption increased [1]. For males, the effect size—hazard ratio (HR)—remained above 1, suggesting a higher risk of health concerns regardless of the amount of alcohol consumed. Unfortunately, we could not find the effect size (HR) and the 95% confidence interval for the dose-response relationship derived from the restricted cubic spline model.

Statistically, we are curious as to why the Cox regression model was chosen to analyze the associations between alcohol consumption and health concerns rather than logistic regression, which is more commonly used when dealing with a binary outcome such as hypertension or prehypertension. Of note, the Cox regression model is typically applied to time-to-event scenarios; however, an important bias that needs to be addressed in such analyses is lead time bias. In the study, lead time bias could suggest that adolescent participants who consumed alcohol may have had a period of time that was not accounted for when analyzing the time to the development of hypertension or prehypertension in the Cox regression model. We point this out because, given the baseline comparison of age, the drinking group was significantly older than the non-drinking group (16 vs. 14 years old; p < 0.005) [1]. To address this bias, researchers might need to examine and compare the follow-up periods between the groups. Even though nonsignificant findings may be found for the follow-up periods, the nondrinking group may not have reached the comparative age to develop hypertension/prehypertension as they were younger. This issue could also arise when using logistic regression. In addition to adjusting the statistical model for age, we wonder if propensity score matching could improve the validity of the analysis by matching baseline characteristics, including age, between the comparison groups.

This study makes important strides in establishing a link between adolescent alcohol consumption and the risk of hypertension in adulthood. Given the presented study design, few research questions remain unanswered among the Chinese population, particularly regarding the impact of alcohol consumption behavior changes between survey periods on health concerns. In our previous work, we analyzed the risk of hypertension given the drinking behaviors between measurement periods and found a strong association between binge drinking both in adolescence and early adulthood [2]. The absence of such an analysis is a missed opportunity to capture the potential long-term cumulative impact of sustained alcohol consumption across different life stages. Considering the role of evolving drinking patterns, future studies should incorporate longitudinal tracking of alcohol consumption behaviors into the assessment of their cumulative impact on evolving risk patterns for hypertension or prehypertension. In addition, there is a need for further research on how gender differences influence alcohol-related hypertension risk, especially considering the small sample size of female drinkers in the study.