Reply to Meira e Cruz et al.: A Reappraisal on the Associations between Sleep-disordered Breathing, Insomnia, and Cardiometabolic Risk

From the Authors:

We thank Meira e Cruz and colleagues for their comments on our recent report on the prospective associations between sleep-disordered breathing (SDB) and insomnia with incident hypertension and diabetes (1). We fully agree that it is important to examine the comorbid insomnia and sleep apnea (COMISA) phenotype as it relates to cardiometabolic risk given the potential for each disorder to cause synergistic physiological stressors or for this phenotype to potentially identify a sleep apnea endotype characterized by low arousal threshold and autonomic reactivity. Prior research has shown that SDB and insomnia frequently cooccur, and COMISA has been associated with impairments to daytime functioning and quality of life (2–4). Accordingly, in sensitivity analyses reported in our paper, we explored whether a phenotype described by both insomnia and objectively assessed SDB differed from either SDB or insomnia alone in its association with incident hypertension or diabetes. In our analytic sample, almost 10% of individuals had comorbid SDB and insomnia at baseline. Results showed that the COMISA group had increased odds of both incident hypertension and diabetes compared with those with neither SDB nor insomnia. As reported in Table E3 in the online supplement of Reference 1, we found some evidence for a stronger association for COMISA and incident hypertension in our fully adjusted models (odds ratio [OR], 2.09; 95% confidence interval [CI], 1.45–3.00) compared with SDB alone (OR, 1.55; 95% CI, 1.14–2.09) or insomnia alone (OR, 1.37; 95% CI, 1.04–1.81). However, no statistical interaction was demonstrated between SDB and insomnia for incident hypertension. For incident diabetes, we did not see evidence for a stronger association for the combined phenotype compared with the individual phenotypes, which differed from Meira e Cruz and colleagues’ reported cross-sectional association. There are many reasons for potential study differences, including population and study design differences such as the use of cross-sectional versus longitudinal analyses, covariate adjustments, and use of objective versus questionnaire-based measures of SDB. Taken together, Meira e Cruz and colleagues’ and our findings suggest that COMISA is both cross-sectionally and prospectively associated with increased cardiometabolic risk, although the magnitude of risk of the combined conditions compared with individual conditions is unclear. There is a need to better define subsets of patients with SDB at highest risk for incident cardiometabolic disease. It is biologically plausible that the physiological disturbances resulting from the cooccurrence of insomnia and SDB may amplify the risk for adverse cardiometabolic health. There is also growing evidence that SDB is a heterogeneous disorder (5). In any given patient, disease may be caused by a different combination of pathophysiological mechanisms (low arousal threshold, elevated loop gain, increased pharyngeal collapsibility, and reduced neuromuscular compensation) that results in differences in cardiometabolic risk. Future efforts are warranted to examine whether COMISA represents the cooccurrence of two common disorders and/or a unique endotype that increases cardiometabolic risk. Improved disease definitions may further clarify which patients with SDB (and insomnia) are at most risk and inform mechanistically informed treatment approaches.