Authors reply: Adherence to guideline-recommended care of late-onset hypertension in females versus males: A population-based cohort study

Abstract

Dear Editor,

We thank Dr. Huang et al. for their thoughtful response to our published manuscript [1, 2]. Their first concern addresses our interpretation of the adjusted hazard ratio (aHR 0.98 [95% CI 0.96, 0.99]) for antihypertensive medication prescription by sex. Although statistically significant, its clinical significance remains uncertain, though we acknowledge its relevance at the population level in the discussion section of our manuscript. To explore potential effect modifiers, we tested a priori interactions between sex and several plausible variables—age, diabetes status, era of hypertension diagnosis, and preexisting cardiovascular disease—by incorporating these multiplicative terms into our model. We then reported aHRs of each stratum from stratified analyses for those variables that were statistically significant on the multiplicative scale. However, we did not report measures of relative excess due to interaction [3]. It is certainly possible that other covariates may have been effect modifiers of the association between sex and prescription of antihypertensive medication.

Second, misclassification bias is a recognized issue in observational studies utilizing administrative data. Nonetheless, the case definition for hypertension used in our study has been validated, showing a sensitivity of 75%, specificity of 94%, positive predictive value of 81%, and negative predictive value of 92%, as detailed in our methods section [4]. As the authors correctly note, administrative data do not capture nuances related to gender and patient preferences that can influence hypertension treatment, which may introduce residual confounding [5].

Third, although the aHR for the prescription of guideline-recommended antihypertensive medication is statistically significant (aHR 0.995, 95% CI [0.994, 0.997]), determining its clinical significance is challenging. We acknowledge that this effect may be clinically relevant at the population level, as discussed in our manuscript. Evaluating antihypertensive medication prescriptions over time, alongside data on actual blood pressure management and drug intolerance, would offer additional insights beyond our current findings.

Lastly, we aimed to address the observation that better cardiovascular outcomes in females compared to males do not appear to be linked to the completion of hypertension-related investigations or the prescription of antihypertensive medication [6]. We proposed that females might benefit more from antihypertensive treatment compared to males, but we recognize that we cannot draw definitive conclusions due to limitations such as insufficient data on treatment adherence, gender-specific issues, and potential residual confounding. More detailed data would enhance our study and support the development of targeted implementation strategies for older populations.

The authors declare no conflicts of interest.