Response

Chen et al's recent article^{Reference Chen, Tsai, Chen, Pan, Su and Chen1} in the BJPsych provides valuable insights into the potential benefits of lipid-modifying agents in reducing mortality risks among individuals with schizophrenia. However, several important limitations warrant further discussion.

First, the study's reliance on prescription data as a proxy for medication adherence is a significant limitation. As noted by the authors, actual medication adherence cannot be accurately determined from prescription data alone. This issue is particularly pertinent in the context of schizophrenia, where non-adherence to prescribed medications, including lipid-modifying agents, is known to be common. This non-adherence might have led to an underestimation of the true effects of lipid-modifying agents on mortality risk, resulting in potential exposure misclassification. Future studies should consider incorporating measures of actual medication intake, such as pharmacy refill records or electronic adherence monitoring, to provide a more accurate assessment.

Second, although the study acknowledges the absence of lifestyle data, the potential confounding effects of unmeasured variables such as smoking, alcohol consumption, physical activity and socioeconomic status are substantial. These factors are well-documented contributors to mortality risk and are particularly relevant in the population with severe mental illness.^{Reference Plana-Ripoll, Pedersen, Agerbo, Holtz, Erlangsen and Canudas-Romo2} The lack of adjustment for these variables might have biased the observed associations, either exaggerating or underestimating the true effects of lipid-modifying agents. Future research should strive to include these critical covariates to enhance the robustness of the findings.

Third, the absence of laboratory data precludes the exploration of biological mechanisms underlying the observed associations. Specifically, the lack of data on lipid levels, inflammatory markers and oxidative stress indicators limits the ability to elucidate the pathways through which lipid-modifying agents may confer mortality benefits. Incorporating such biomarkers in future studies would provide a more comprehensive understanding of the therapeutic effects and mechanisms of lipid-modifying agents in schizophrenia.

Fourth, the study does not sufficiently address the issue of potential misclassification of suicide deaths. Misclassification can occur if some deaths classified as accidental or of undetermined cause were, in fact, suicides. This potential misclassification might lead to an underestimation of the suicide mortality risk associated with schizophrenia.^{Reference Correll, Solmi, Croatto, Schneider, Rohani-Montez and Fairley3} Implementing more rigorous methods to ascertain the cause of death, possibly through the inclusion of coroners’ reports or validation against other databases, would mitigate this issue.

Finally, another significant limitation is the potential influence of time-dependent changes in treatment practices and healthcare policies over the nearly two-decade study period. Changes in medication availability, clinical guidelines and healthcare access might have influenced prescribing patterns and mortality outcomes. A more detailed analysis accounting for temporal changes and their potential impact on study results would provide a clearer picture of the long-term effects of lipid-modifying agents.

Addressing these additional limitations in future research will be crucial for substantiating and extending Chen et al's findings. We commend the authors for their important work and hope that our commentary provides constructive insights for further investigations.