Obstructive sleep apnea (OSA) is a prevalent sleep disorder caused by intermittent partial or complete collapse of the upper airway during sleep resulting in intermittent hypoxemia and hypercapnia, sleep fragmentation, increased oxidative stress, and systemic inflammation. The prevalence of moderate-to-severe OSA in the general population, estimated at 23.4% in women and 49.7% in men, has increased over the past few decades (1,2), due in part to the obesity epidemic (3). Multiple observational studies have reported an association between OSA, insulin resistance, and increased risk of incident type 2 diabetes (4,5). In a meta-analysis, the presence of OSA was independently associated with a 1.35-fold increase in the risk of developing type 2 diabetes (4). In addition to being a risk factor for developing type 2 diabetes, OSA is highly prevalent in adults with type 2 diabetes, with prevalence estimates between 58% and 86%, with moderate-to-severe OSA being present in 25–59% of individuals (4).
Although increasing severity of OSA has been associated with worsened glycemic control (6), continuous positive airway pressure (CPAP), the gold standard treatment of OSA, has not been consistently shown to improve glycemic control or reduce insulin resistance in clinical trials in which CPAP was used for 1–6 months (4,7–10). Multiple factors may contribute to the heterogeneity of the results, including different levels of CPAP adherence, various durations of CPAP therapy, background glycemic control, and the use of antihyperglycemic agents, the number of which has been ever increasing. In well-controlled laboratory experiments, however, there is evidence that short-term CPAP use with ensured compliance of 8 h nightly for 1–2 weeks results in improved glucose metabolism, including reduction in 24-h mean glucose levels in adults with type 2 diabetes (11,12), and reduction in glucose response to glucose challenge and improved insulin sensitivity in those with prediabetes (13). This night-long use of CPAP could be a significant factor, as rapid eye movement–related obstructive apneas and hypopneas, generally occurring in the later part of the night, were found to be associated to glycemic control in type 2 diabetes (14). While 34 million people in the U.S. were estimated to have diabetes in 2018, another 88 million were estimated to have prediabetes, a condition predisposing individuals to future diabetes (15). Despite the evidence that OSA is linked to incident type 2 diabetes and dysglycemia, no long-term data exist as to whether CPAP treatment improves glucose metabolism in people with OSA and type 2 diabetes or whether it lowers conversion rate from prediabetes to type 2 diabetes.
In this issue of Diabetes Care, Loffler et al. (16) report their findings in a substudy of 888 participants from the Sleep Apnea Cardiovascular Endpoints (SAVE) trial, a randomized clinical trial in which 2,687 participants aged 45–75 years with OSA and stable cardiovascular disease were randomized to CPAP treatment plus usual care versus usual care alone. In this substudy, the SAVE investigators aimed to determine the long-term effect of CPAP on glycemic control and type 2 diabetes risk with a median follow-up of 4.3 years, making it the largest clinical trial with the longest duration of follow-up to date examining the impact of CPAP on glycemic control and the risk of incident type 2 diabetes in people with OSA (16). At baseline, 274 (30.8%) participants had known type 2 diabetes, 452 (50.9%) had prediabetes, and 162 (18.2%) had neither. OSA was defined as having an oxygen desaturation index of ≥12 events/h using ≥4% oxygen desaturation criteria during a home screening test. People with severe daytime sleepiness and severe nocturnal hypoxemia were excluded. The mean CPAP usage for the duration of the trial was 3.5 ± 2.3 h/night. After a median follow up of 4.3 years, there was no significant difference between CPAP and usual care in serum glucose, hemoglobin A1c (HbA1c), or antidiabetic medication use in those with known type 2 diabetes. Moreover, no significant differences were found in those with prediabetes or in new diabetes diagnoses. However, possible benefit was seen in women with type 2 diabetes assigned to CPAP, such that glucose levels in those receiving CPAP were lower than those in the usual-care group by an estimate of 1.037 mmol/L (18.6 mg/dL).
This study has several strengths. The researchers are commended on their enormous effort to conduct the largest and the longest randomized clinical trial in people with OSA and established cardiovascular disease. The results give an important insight into the causal relationship between OSA and glucose metabolism and provide information for clinical practice. However, even in the best environment of a clinical trial, CPAP adherence was still low despite implementing a run-in period of 1 week with sham CPAP and excluding those who used CPAP <3 h/night during the run-in period. One question that arises is whether the results were negative because OSA does not have clinically significant adverse effect on glycemic control or in reducing the risk of developing type 2 diabetes—and therefore any treatment of OSA would be ineffective in reducing diabetes risk—or because the participants did not use CPAP for a long enough duration each night to derive metabolic benefits. We suspect it is the former, given that there was no relationship between various metrics of OSA severity or levels of CPAP adherence and glucose levels or HbA1c in this study. However, despite being the largest clinical trial to date exploring the impact of CPAP on glycemic control and diabetes prevention, the study lacked enough statistical power. In those with prediabetes, the incidence of type 2 diabetes in this study was relatively low, with about 15.9% of participants over an average 4.3 years of follow-up, compared with the rate of 11.0 cases per 100 person-years in the Diabetes Prevention Program (17), a study of more than 3,000 participants with both elevated fasting and postload glucose levels. Furthermore, attempts to adjust for antidiabetic medications might not be able to capture the full effects of glucose-lowering agents on HbA1c, thus confounding the effects of CPAP on HbA1c in subjects with known type 2 diabetes. Different criteria of prediabetes used in this study (fasting glucose or HbA1c levels without an oral glucose tolerance test) might explain the findings as fasting glucose, postload glucose, and HbA1c levels may confer different future diabetes risks (18). Thus, the study may not have adequate power to truly detect diabetes conversion rate. Nevertheless, glycemic parameters in participants with prediabetes did not differ between CPAP and usual-care groups.
This study excluded subjects with excessive daytime sleepiness as well as those with severe nocturnal hypoxemia. The exclusion of such individuals could have significant implications in interpreting the results, given that the enrolled subjects may not be representative of patients who are seen in clinical practice. Moreover, a few studies have shown that OSA severity as measured by hypoxemic burden during sleep is associated with insulin resistance and type 2 diabetes only in sleepy subjects (19,20). The issue of sex differences is also of interest; a previous study in patients with type 2 diabetes revealed that the relationship between OSA severity and HbA1c was not seen in women (21), yet the current study suggested benefits of CPAP in women only.
The work by Loffler et al. (16) marks an important step in the area of OSA and glucose metabolism and paves the way to future directions in this field. With an alarming 88 million adults with prediabetes in the U.S., with many of them having concomitant OSA, it is imperative to carry out well-designed clinical trials examining the role of CPAP therapy to slow the progression from prediabetes to type 2 diabetes. While clinicians eagerly await the results of future studies, patients with OSA who derive symptomatic benefit from CPAP therapy should be encouraged to use it.
See accompanying article, p. 1859.
Article Information
Duality of Interest. S.R. receives honoraria from Becton, Dickinson and Company, outside of this work. No other potential conflicts of interest relevant to this article were reported.
