A potential association between obstructive sleep apnea (OSA) and diabetes is intriguing, as acknowledged in the commentary by Tahrani (1) on our article (2). OSA is prevalent among patients with diabetes and may contribute to several important aspects, including glycemic control, hypertension, cardiovascular disease, and microvascular complications (3,4). In our observational study in a clinical setting, patients with type 2 diabetes and untreated OSA developed higher HbA1c levels (and blood pressure) compared with best-matched patients treated with continuous positive airway pressure (CPAP), despite greater prescription of antidiabetes medications (2). Although data were unavailable, it is likely that the patients studied had significant symptomatic OSA given their BMI and the usual symptom-driven presentation to primary care.
Tahrani (1) suggested that the degree and rapidity (occurring within a year) of the HbA1c difference between the groups was unusual. Several mechanisms for the potential impact of OSA on glycemic control have been proposed, including intermittent hypoxemia, hypercapnia, increased sympathetic drive, sleep disturbance, and short sleep duration (2). For example, even a few nights of partial sleep deprivation can have significant effects on glycemic control (5). CPAP effects on the above mechanisms could explain the rapid and significant difference in glycemia between the patient groups. The degree of difference in HbA1c between the groups is also comparable to other studies. One study of diabetic patients observed that the adjusted mean HbA1c was 3.7% higher in severe OSA patients compared with those without OSA (6). There are, of course, many other potential factors that could also explain our findings, including poorer medication adherence that may accompany a group of patients who may not accept, or have been offered, CPAP treatment. Another possibility is potential cognitive decline with untreated OSA, which might have affected medication adherence. However, we believe these explanations are less likely as a nonadherent or cognitively impaired group would engage less with health care services and perhaps demonstrate early deterioration in other parameters such as blood pressure, but a difference in blood pressure between the groups was not observed until the fifth year after diagnosis.
Given the observational nature of our study, we agree with Tahrani (1) that it is difficult to make confident statements regarding any causal relationship between OSA treatment, glycemic control, and hypertension, as discussed in our article. Current evidence examining the relationship between diabetes and OSA suffers from several drawbacks, including mainly cross-sectional designs, selection bias, variation in diagnostic measures of OSA, lack of consideration for confounding factors, and short CPAP treatment duration (3). There are insufficient data from randomized controlled trials (RCTs) regarding the long-term impact of CPAP treatment on diabetes and its complications (3). While Tahrani proposes RCTs, it should be noted that longer-term placebo-controlled RCTs of OSA treatment in symptomatic patients are generally difficult to conduct and are hampered by CPAP treatment adherence problems. Even RCTs may show an experimental effect that is not due to direct effects of OSA relief but secondary effects such as improved cognition and exercise.
In our study, as noted by Tahrani (1), we did not have sufficient data to examine optimal CPAP usage, as compliance data were very limited. It is currently unclear what optimal CPAP usage is, in particular for a potential impact on metabolism. Another issue is the timing of CPAP intervention in patients with diabetes; early intervention may be more effective. Given that current glycemic management approaches are insufficient, there is a need to carefully examine whether OSA treatment can improve glycemic control in the long term in well-designed studies.
Article Information
Duality of Interest. No potential conflicts of interest relevant to this article were reported.
