Sleep Apnea May Accelerate Diabetic Retinopathy

STOCKHOLM — Nonproliferative diabetic retinopathy (NPDR) had a significantly greater risk of progression and complications in patients who also had obstructive sleep apnea (OSA), a large propensity-matched cohort study showed.

In patients with existing NPDR, OSA increased the likelihood of progression to proliferative DR by 75% at 1 year as compared with patients who had NPDR and no OSA. A significant difference persisted out to 5 years. Similarly, the likelihood of new-onset diabetic macular edema (DME) and ocular interventions increased significantly within a year and continued through 5 years in patients with OSA.

Consistent with other studies, the risk of systemic events — stroke, myocardial infarction (MI), and premature death — also were significantly higher in patients with NPDR and OSA, said Ehsan Rahimy, MD, of Stanford University in California, at the American Society of Retina Specialists meeting.

“I won’t belabor … the limitations of some of these registrational database-type studies,” said Rahimy. “They kind of give us a good overview. We can see trends that are worthy of follow-up studies to kind of fill in the knowledge gaps.”

“As a parting thought, many of us see patients with diabetes more frequently than they see their own primary care doctors on an annual basis,” he added. “Give that some consideration. It behooves us all to just be aware of some basic screening questions, such as the STOP-Bang questionnaire. If patients meet even three of the criteria [for OSA], they should be referred to get a sleep study and higher level of care.”

Chicken/Egg Problem

The association between OSA and progression of diabetes “is like a chicken/egg problem,” said Tongalp Tezel, MD, of Columbia University in New York City, during a discussion after the presentation. “Diabetics have a higher rate of obstructive sleep apnea because they have autonomic nerve dysfunction, which affects the respiratory system, but autonomic nerve dysfunction also is a cause of progression of diabetes. How do you isolate obstructive sleep apnea?”

Rahimy said conditions such as diabetes and OSA “have a lot of phenotypic overlap of these conditions, with obesity with sleep apnea, with diabetic retinopathy. It gets back to the construction of the propensity score matching cohorts. There are only so many variables you can control before you start kind of hurting your analyses. We try to be as comprehensive as possible, but there’s no doubt that sleep apnea is just a very difficult thing to completely isolate.”

Session co-moderator David Boyer, MD, of Weill Cornell School of Medicine in New York City, noted that patients continued to progress after diagnosis of OSA.

“Does that imply that the [continuous positive airway pressure or CPAP] mask wasn’t fitting or that there is really no treatment that’s going to stop this?” Boyer asked. “It’s really the group you want to know before they get a mask, how fast did they progress.”

“How do we isolate these patients for the sake of the study?” asked Rahimy. “The index point was once they were diagnosed with sleep apnea with confirmatory CPAP, we required ongoing CPAP assessments, but we don’t know that they were compliant with their CPAP.”

A recent study showed that the GLP-1 receptor agonist tirzepatide (Zepbound) significantly improved OSA symptoms in obese patients, irrespective of CPAP. Alluding to that study, Rahimy said, “Patients are losing weight, and it’s potentially improving their sleep apnea. We potentially are going to have therapeutic options for sleep apnea besides CPAP. There’s a lot of interesting overlap.”

An unidentified ophthalmologist from Switzerland asked whether he should begin referring all of his patients for a sleep test. He said he regularly questions all of his patients with retinal vein occlusion and diabetes about sleep apnea.

“I’m much more conscientious [about sleep apnea],” said Rahimy. “The STOP-Bang is automatically running through my head. Sleep apnea is severely underdiagnosed in the diabetic population. We need to take that into account. If you identify it early and get patients referred for therapy, you could potentially make a positive impact for them, so I do.”

Background, Key Findings

With respect to the retina, OSA and diabetic retinopathy overlap pathophysiologically. Rahimy and colleagues sought to gain some clarity about the relationship between OSA and diabetic retinopathy. For the analysis, they identified patients with NPDR in the TriNetX network of healthcare data and then stratified them by presence of OSA.

The analysis began with 12,321 patients with NPDR and OSA and 128,961 patients with NPDR and no OSA. Using propensity score matching, investigators constructed two cohorts of 9,475 patients each. The primary outcomes were the relative risks for vision-threatening complications (PDR and DME), need for ocular intervention, and systemic events (MI, stroke, death).

Propensity score matching is “a little bit of a balancing act,” said Rahimy. “We can control for various baseline factors … realizing that if you account for too few variables, you’re ignoring potential confounders. If you put too many [factors] in there, then you’re potentially overfitting your model and restricting the availability of potentially otherwise good matches. “

The analysis showed that the patients with OSA had significantly higher (P<0.001) rates of progression to PDR at 1 year (6.1% vs 4.6%), 3 years (9.2% vs 7.0%), and 5 years (10.4% vs 8.1%), as well as new-onset DME (7.0% vs 5.3%, 10.6% vs 8.9%, and 12.5% vs 10.6%). Significantly more patients with OSA required ocular interventions, including intravitreal injections (6.5% vs 4.2% at 5 years) and panretinal photocoagulation (3.4% vs 2.2% at 5 years).

Systemic events also occurred significantly more often in the OSA cohort: stroke (RR 1.29 at 1 year, RR 1.15 at 3 years, 1.11 at 5 years, P<0.0001 to P=0.0051), MI (RR 1.55, 1.37, 1.28, P<0.0001), and death (RR 3.75, 3.61, 1.63, P<0.0001).

Rahimy cited three areas for potential investigation in the future: explore associations between OSA severity and retinopathy outcomes, assess the impact of CPAP on modifying risks, and the impact of GLP-1 agonists on OSA and diabetic retinopathy progression.

Please enable JavaScript to view the

comments powered by Disqus.