Greater severity of obstructive sleep apnea (OSA) is associated with a higher risk of contracting COVID-19, and positive airway pressure (PAP) treatment may counter that risk, according to a retrospective analysis from the records of Kaiser Permanente Southern California.
Dr. Dennis Hwang
OSA patients often worry that PAP therapy might increase risk of severe COVID-19, said Dennis Hwang, MD, who presented the study at the American Thoracic Society’s virtual international conference (Abstract A1108). But the findings should be reassuring. “If you have obstructive sleep apnea, and you’re supposed to be using PAP, we recommend that you continue using PAP. It’s good for your overall wellness and reducing the risk of cardiovascular disease, but as it relates to COVID-19, it’s possible that it could protect. And there doesn’t appear to be any risk of increased severity of illness (with use of PAP),” Dr. Hwang said in an interview. He is medical director of sleep medicine for Kaiser Permanente San Bernardino County and cochair of sleep medicine for Kaiser Southern California.
He noted that the retrospective nature of the study makes it difficult to pin down whether PAP therapy is truly protective, “but I think there’s enough that we’ve been able conceptually to understand, to suggest that a direct causative relationship is possible,” said Dr. Hwang.
The results may imply that OSA patients should pay special attention to their OSA when there’s concern about exposure to an infectious agent like SARS-CoV-2. “The intermittent hypoxia at night, which can linger over to the day as increased sympathetic activity, increased heart rate. All of these are stresses to the body. So if you’re going to get infected, you want to start at a healthier level. You want to eliminate your sleep apnea to help reduce your risk of morbidity,” said Esra Tasali, MD, who was asked to comment on the study. Dr. Tasali is associate professor of medicine at the University of Chicago, and director of the Sleep Research Center there.
During the Q&A session after the talk, audience members asked about the timing of PAP use during COVID-19 infection, for example how often it was used during the asymptomatic phase of infection and if PAP has a positive effect. The data were not available, but “I think that the way to go is to understand this chronology,” said Dr. Tasali.
The researchers examined records between 2015 and 2020, using sleep study data, remotely collected daily PAP data, and electronic health records, all from Kaiser Permanente Southern California. Included subjects were adults who had enrolled before Feb. 1, 2020, and had sleep diagnostic or PAP data on record by March 1, 2020. The researchers analyzed PAP adherence between March 1, 2020, and the time of COVID-19 diagnosis, or until the study ended on July 31, 2020.
Patients were defined as being untreated (< 2 hours/night PAP), moderately treated (2-3.9 hours/night), or well treated (4 or more hours/night). Apnea hypopnea index (AHI) was used to determine severity. The analysis included 81,932 patients (39.8% were women, mean age was 54.0 years, 9.9% were Black, and 34.5% were Hispanic). A total of 1.7% of subjects without OSA experienced COVID-19 infection, compared to 1.8% with OSA; 0.3% with OSA were hospitalized and 0.07% underwent intensive care or died.
There were some differences between the two groups. The non-USA population was younger (mean age 47.0 vs. 54.5 years), was less likely to be men (44% vs. 60.3%), had a lower mean body mass index (30.4 vs. 34.3), had fewer comorbidities according to the Charleston Comorbidity Index (1.3 vs. 2.0), and were less likely to have hypertension (5.6% vs. 12.4%; P < .0001 for all).
Infection rates were higher in patients with more severe OSA. The rates in untreated mild, moderate, and severe OSA were 2%, 2%, and 2.4% respectively. The rate among all treated patients was 1.4% (P < .0001). Infection rates also dropped among patients with better treatment: untreated, 2.1%; moderately treated, 1.7%; and well treated, 1.3% (P < .0001).
Not having OSA was associated with a lower infection risk than was having OSA (odds ratio [OR], 0.82; 95% confidence interval, 0.70-0.96). Compared to untreated patients, there was lower infection risk in the moderately treated (OR, 0.82; 95% CI, 0.65-1.03) and well treated (OR, 0.68; 95% CI, 0.59-0.79) groups. Higher infection rates were associated with obesity, higher Charlson Comorbidity score (> 2; OR, 1.29; 95% CI, 1.09-1.53), Black (OR, 1.51; 95% CI, 1.24-1.84) and Hispanic ethnicities (OR, 2.23; 95% CI, 1.96-2.54), and Medicaid enrollment. Increasing age was associated with lower risk of infection, with each 5-year increment linked to reduced risk (OR, 0.88; 95% CI, 0.86-0.90). Dr. Hwang suggested that the age association may be because older individuals were more likely to follow social distancing and other precautions.
A multivariate analysis found that OSA was associated with infection risk according to OSA severity, including mild (OR, 1.21; 95% CI, 1.01-1.44), and moderate to severe (OR, 1.27; 95% CI, 1.07-1.51). There was no association between hospitalization rate or ICU admission/death and presence of OSA or PAP adherence in the data presented, but Dr. Hwang said that an updated analysis suggests that OSA may be associated with a risk of greater COVID-19 severity.
The control group was composed of individuals who had undergone sleep testing, but found to not have OSA. Still, they aren’t necessarily representative of the general population, since symptoms likely drove them to testing. A high percentage were also obese, and the average BMI was 30. “It’s certainly not a ‘normal population,’ but the advantage of what we did in terms of using this control group is that they underwent sleep testing, so they were proven to have no obstructive sleep apnea, whereas if we used a general population, we just don’t know,” said Dr. Hwang.
The study received technical and data support from Somnoware, and was funded by Kaiser Permanente. Dr. Tasali has no relevant financial disclosures.