Response to Bartoli et al.

To the Editor: We would like to thank Bartoli et al. (1) for their supplementary analysis to our recently published article (2) comparing metabolic syndrome in patients with bipolar disorder and schizophrenia assessed within the same study. Meta-analytic data comparing major diagnostic subgroups with each other are indeed currently lacking. We agree with Bartoli et al. that such pooled comparisons are relevant, as they allow investigation of the frequency of metabolic syndrome in patients with different disorders who were assessed with the same methodology. Moreover, such analyses could assist with dissecting risk factors associated with mental disorders from those independent of the underlying illness, at least inasmuch as such data were actually collected and uniformly reported. Comparison and pooling of data across major diagnostic categories also allows investigation into the effects of specific treatments, particularly mood stabilizers and antipsychotics, in psychotic and nonpsychotic conditions. If risk stratification is observed, this information could help guide clinicians in monitoring and treatments decisions.

However, although the complementary findings of Bartoli et al. (1) are relevant, we believe they should still be interpreted with caution. Mitchell et al. (3) found disease stage and medication history to be influential, and these factors are not controlled for in these comparisons. Moreover, both Mitchell et al. (4) in schizophrenia patients and Vancampfort et al. (2) in bipolar patients found that age was also a significant moderating variable for the prevalence of metabolic syndrome in severely mentally ill patients. As in the general population, older age is associated with higher metabolic syndrome prevalence rates. However, in the majority (8/11) of the studies included in the Bartoli et al. analysis (1), patients with bipolar disorder and schizophrenia were not matched for age. Therefore, we suggest that more research with age-matched, medication-equivalent clinical populations that have similar durations of illness is needed before we can conclusively state that patients with bipolar disorder and schizophrenia have comparable risks for metabolic syndrome.

Finally, we agree with Bartoli et al. (1) that further research should assess the relative contribution to metabolic syndrome of not only different psychiatric diagnoses, but also other components such as genetics, medication use, and health behaviors. To allow such analyses, these highly relevant moderating and mediating variables should be assessed more uniformly in studies reporting on metabolic syndrome frequencies in patients with specific mental disorders.