Ertugliflozin and new cases of obstructive sleep apnea in the VERTIS CV trial

Obstructive sleep apnea (OSA) has been independently associated with important health outcomes, including hypertension, cardiovascular (CV) events, stroke, heart failure, chronic kidney disease (CKD), and metabolic dysfunction (1–3). OSA prevalence in adults ranges between 9 and 36% with higher rates in older men [4]. Among individuals with type 2 diabetes (T2D), concomitant OSA is associated with a higher risk of CV disease (CVD), atrial fibrillation, peripheral neuropathy, diabetic foot disease, CKD, and all-cause mortality [3]. Furthermore, in a randomly selected, nationally representative sample of Medicare beneficiaries, patients with untreated OSA had significantly higher mean total annual healthcare costs [5]. Treatment adherence with continuous positive airway pressure (CPAP), the mainstay intervention for OSA, is poor, ranging from 34 to 60% [6]. While improving CPAP adherence is key to improving OSA outcomes, the discovery of effective medical therapies could transform OSA care.

The sodium-glucose cotransporter 2 inhibitors (SGLT2i) were initially developed and approved to treat hyperglycemia in patients with T2D, but potent CV and kidney benefits were subsequently demonstrated in large-scale randomized trials. Their potential efficacy in OSA is just beginning to be explored. In a recent study, 36 patients with T2D and OSA were randomized to either metformin and glimepiride or metformin and the SGLT2i dapagliflozin and treated for 24 weeks. Home sleep testing was performed at the beginning and end of the study. Dapagliflozin led to significantly improved apnea–hypopnea index (AHI), a small rise in oxygen saturation, and better scores on the Epworth Sleepiness Scale [7]. In another small retrospective study of patients with OSA who had refused CPAP therapy, SGLT2i treatment was associated with significantly lower AHI [8]. Results from analyses exploring the effect of empagliflozin on OSA incidence from the EMPA-REG OUTCOME trial found that empagliflozin versus placebo reduced the incidence of OSA in patients with T2D and atherosclerotic CVD (ASCVD) by more than half (4.6/1000 patient-years [1.13%] vs. 2.2/1000 patient-years [0.57%]; hazard ratio [HR], 0.48; 95% confidence interval [CI], 0.27–0.83; P = 0.009) (9). Notably, this effect was largely independent of weight loss.

To further investigate the effect of SGLT2 inhibition on OSA incidence, we performed a post hoc analysis of VERTIS CV, a double-blind, placebo-controlled trial of the efficacy and safety of the SGLT2i ertugliflozin in patients with T2D and ASCVD.

VERTIS CV (NCT01986881↗) consisted of 8246 patients ≥ 40 years of age with T2D and ASCVD randomized to ertugliflozin 5 mg, 15 mg, or placebo once daily, with the primary outcome of composite of major adverse CV events. In these exploratory analyses, we evaluated the impact of ertugliflozin on incident OSA. A pooled analysis of ertugliflozin 5 and 15 mg (measured from the first dose of study medication until treatment end plus an additional 14 days) versus placebo was conducted. Patients with OSA at baseline (549; 6.7%) were excluded. Incident OSA events were based on investigator-reported adverse events during the trial using the MedDRA term sleep apnea syndrome (synonyms at: https://bioportal.bioontology.org/ontologies/MEDDRA?p=classes&conceptid=10040979↗). Kaplan–Meier estimates for time to first OSA event by treatment assignment for participants without OSA at baseline were plotted. A stratified Cox proportional hazards regression model was constructed to compare treatments with adjustment for age, sex, geographic region, baseline body mass index (BMI; < 35, ≥ 35 kg/m²), baseline glycated hemoglobin (HbA1c; < 9.0, ≥ 9.0%), and baseline estimated glomerular filtration rate (eGFR; < 60, ≥ 60 mL/min/1.73 m²) and with stratification by enrollment cohort (before and after protocol amendment).

Small studies have suggested the potential efficacy of SGLT2i versus placebo in OSA (7, 8). An exploratory analysis of the large EMPA-REG OUTCOME trial suggested that empagliflozin may prevent OSA (9). Based on prior findings (9), we hypothesized that ertugliflozin treatment would reduce the risk of new-onset OSA in the VERTIS CV trial. We found a significant 48% adjusted relative risk reduction that was observed with ertugliflozin versus placebo for incident OSA, similar to the magnitude of effect observed with empagliflozin [9].

Several potential mechanisms may underpin SGLT2i efficacy to improve OSA risk. The first is weight loss associated with SGLT2i, as urinary glucose excretion induced by this mechanism results in a daily net caloric loss. However, weight loss associated with SGLT2i is modest, and, in a mediation analysis from the EMPA-REG OUTCOME trial, weight loss was not significantly associated with the OSA effect [10]. It is also possible that the resultant reduction in plasma volume from osmotic diuresis (and accompanying natriuresis) decreases cardiac preload, resulting in less loop gain and downstream beneficial effects in respiratory physiology relevant to both obstructive and central sleep apnea (9).

There are several limitations to the present study. Patients were identified with OSA by investigator reporting of adverse events, which may have under-reported incident sleep apnea since periodic diagnostic sleep studies were not required in the protocol. Adverse event reporting is known to under-report incident OSA in large trials. Nonetheless, the results were highly consistent between the EMPA-REG OUTCOME trial and VERTIS CV, using two different SGLT2i. Furthermore, any misclassification of the outcome would tend to bias toward the null hypothesis and would not explain our observed results. The AHI was also not measured prospectively in this study; therefore, no comment can be made on the magnitude of AHI reduction or the baseline severity and type of sleep apnea in this population.

In VERTIS CV, the SGLT2i ertugliflozin reduced the incidence of OSA in patients with T2D and ASCVD, an effect similar in magnitude to that previously reported with empagliflozin, contributing to the growing body of literature that SGLT2i may serve to prevent (and possibly treat) sleep apnea syndromes. Accordingly, the SGLT2i class deserves additional exploration for this potential new role, including in patients without diabetes, as well as a further inquiry into the mechanisms involved.