Obstructive sleep apnea (OSA) is associated with a decreased risk for mortality and in-hospital complications in patients who undergo thrombectomy, new research suggests.
In a cross-sectional study of patients who were treated with endovascular mechanical thrombectomy (MT) for an acute ischemic stroke (AIS), the patients with OSA were at significantly lower risk for death than the patients without OSA. The rate of intracranial hemorrhage was also lower among the OSA group than the non-OSA group.
“Our goal was to determine whether the use of MT as a definitive reperfusion therapy could reduce the disparity in outcomes often observed in OSA vs non-OSA AIS patients,” lead author Justin Lepow, a medical student at New York Medical College, in Valhalla, New York, told meeting attendees.
He presented the findings at the XXV World Congress of Neurology (WCN), which was held online.
Previous studies showed that patients with OSA and AIS had poorer outcomes and that mortality was increased in comparison with patients with stroke but without OSA, Lepow told meeting attendees.
The investigators used the National Inpatient Sample (NIS) to conduct a cross-sectional analysis of patients with AIS treated with MT from 2010 to 2018. During that period, MT was accepted as an effective therapy for AIS but had not been evaluated for patients with OSA.
The NIS is the largest inpatient database in the United States. It represents 20% of all US community hospital discharges annually.
The researchers used International Classification of Diseases-9-Clinical Modification (ICD-9-CM) and ICD-10-CM codes to identify patients treated with MT after AIS.
The investigators employed a battery of stroke severity indices to evaluate baseline characteristics and used other measures to rate severity of illness. Outcomes included in-hospital complications and discharge disposition.
In addition, a multivariable model was constructed to evaluate the independent association of OSA status and mortality after adjusting for baseline covariates.
Of the total cohort of 101,093 patients with AIS who were evaluated, 6412 (6.3%) had OSA. At baseline, the OSA group was younger than the non-OSA group (mean age, 65.6 years, vs 68.5 years). There were fewer women in the OSA group (33.5% vs 50.5%), and there were fewer non-White patients (22.3% vs 29.7%; all comparisons, P < .001).
Also, more of the patients in the OSA group had strokes from 2016 to 2018, which was after the MT clinical trial era (68.8% vs 63.8%; P = .001).
There were also differences in the rate of comorbid conditions at baseline. The percentage of patients with obesity was higher in the OSA group than in the non-OSA group (41.4% vs 10.5%), as was the percentage with atrial fibrillation (47.1% vs 42.2%), hypertension (87.4% vs 78.5%), and diabetes mellitus (41.2% vs 26.9%; all comparisons, P ≤ .001).
Lower Mortality Risk, Fewer Complications
There was no difference between the groups in incidence of patent foramen ovale. There was no significant difference in the composite stroke severity score or the illness severity score, and there was no significant difference in the use of IV thrombolytics.
The rate of in-hospital complications was lower for the OSA group. “We surprisingly found that intracranial hemorrhage was lower in the OSA group” than in the non-OSA group (9.1% vs 21.8%; P = .017), Lepow reported.
Other in-hospital complications favoring the OSA group were mortality (9.7% vs 13.5%, P < .001) and external ventricular drain or ventriculoperitoneal shunt placement for hydrocephalus (0.3% vs 1.1%, P = .009).
No other neurologic or medical complications that were assessed differed between the groups, and the hospital length of stay was the same, at 8.19 days.
In the multivariable analysis, “OSA was found to be independently associated with lower in-hospital mortality after treatment,” Lepow said.
Mortality risk was 22% lower for the OSA patients than the non-OSA patients (odds ratio, 0.78; 95% CI, 0.63 – 0.96; P = .02).
The association of lower mortality with OSA persisted after controlling for obesity, atrial fibrillation, hypertension, diabetes, age, gender, severity of illness, and stroke severity.
“The Jury’s Still Out”
Commenting on the findings for Medscape Medical News, Louise McCullough, MD, PhD, chair of the Department of Neurology at the University of Texas Medical Center, Houston, Texas, called the study interesting and well done.
However, one of the problems with the NIS dataset is that it does not include the National Institutes of Health stroke scale, said McCullough, who was not involved with the research. Still, “they have done a pretty good job trying to control for stroke severity,” she added.
McCullough said she found the lower mortality rate among the OSA group somewhat surprising and paradoxical and offered some possible explanations. She noted that the investigators tried to control for several variables, including age and sex.
“But these might be very important factors” she said. “The patients with OSA were younger and were more often male.”
Another possible cause of the lower mortality in the OSA group is cerebral preconditioning. The bouts of hypoxia during sleep may have induced preconditioning, which experimental models have shown to be protective.
“So possibly, the fact that they have intermittent apnea may protect them when they have a stroke,” McCullough hypothesized. Preconditioning may induce the production of growth factors, such as hypoxia inducible factor, angiogenic factors, or others in response to hypoxia, leading to formation of robust collateral vessels and smaller strokes, she said.
Studies have been conducted on the use of such factors acutely for neuroprotection, but McCullough warned that “the jury’s still out on whether OSA itself is protective. Certainly, we know OSA, especially if it’s untreated, is detrimental.”
She also cautioned that use of the NIS has its limitations. In this study, many members of the non-OSA group could have had undiagnosed OSA, skewing the results.
“So I’m sure in that 100,000 patient population, there probably were patients that had OSA that weren’t coded that way,” she said. The patients themselves may not have known they had OSA, she added.
McCullough also cautioned that the study could not control for every possible variable that may affect mortality, such as infections, deep vein thromboses, or the possibility that some patients may have had many more comorbidities.
Lepow and McCullough have reported no relevant financial relationships
XXV World Congress of Neurology (WCN 2021): Presented October 3–7, 2021.