Sleep-Disordered Breathing Is Associated With Recurrent Ischemic Stroke

Brown, Devin L. MD; Shafie-Khorassani, Fatema MPH; Kim, Sehee PhD; Chervin, Ronald D. MD, MS; Case, Erin BA; Morgenstern, Lewis B. MD; Yadollahi, Azadeh PhD; Tower, Susan MD; Lisabeth, Lynda D. PhD


Stroke. 2019;50(3):571-576. 

In This Article


These population-based, longitudinal data with adjustment for potential confounders demonstrate that severity of SDB measured after one ischemic stroke predicts increased risk for another. In contrast, SDB severity as reflected by the REI is not associated with the isolated outcome of all-cause mortality. As SDB may represent a modifiable risk factor for poor stroke outcomes, these findings, including the continuous, graded association, have important implications for secondary stroke prevention strategies and prioritization of clinical trials to test whether treatment of SDB reduces the risk of stroke recurrence.

The association between SDB and recurrent ischemic stroke has received limited prior investigation for comparison to the current results. A single hospital-based sample found that SDB was associated with recurrent ischemic stroke/transient ischemic attack in an unadjusted analysis.[9] A second single-center European study demonstrated that an REI score ≥20, as opposed to lower values, was associated with development of cardiovascular ischemic events including stroke. This relationship persisted after adjustment for multiple potential confounders (13 confounders and 16 recurrent strokes among 138 subjects with untreated obstructive sleep apnea).[25] Yet, a third small single-center sample (n=99) found no association between SDB and combined stroke recurrence and death (3 recurrent strokes, 15 deaths), although this may have been because of limited power.[26] Our results extend the literature through consideration of a much larger sample that supports multivariable adjustment for potential confounders, a population-based design that results in more representative findings, use of a more ethnically diverse sample, identification of a linear association between REI and stroke recurrence, and application of competing risks analysis that can help differentiate associations with individual competing outcomes, such as recurrent stroke and mortality.

Prior single-hospital studies in Europe found that SDB was associated with poststroke mortality after accounting for confounders.[4–6,27] The reasons for the lack of association between SDB and mortality in the current study, and a possible protective effect among NHWs, are not clear. Differences in analytic strategy, with application of competing risks methods in the current analysis, could have contributed. If the effect of SDB on mortality differs by age, the lower age of our study population compared with most of the other studies on this topic could also have contributed. One study in which the mean age of subjects was 56 also did not find an independent association between SDB and mortality.[28]

Given the higher risk of both recurrent ischemic stroke[11] and poststroke SDB in MAs compared with NHWs,[13] we hypothesized that SDB may contribute to the higher recurrence risk in MAs. Alternatively, SDB could contribute to this disparity through a stronger association between SDB and stroke recurrence among MAs. As MA ethnicity was not associated with recurrence in the current, more contemporary data, perhaps as a result of the reduction in the ethnic disparity through time or differences in analytic approaches, we could not formally test this hypothesis. Unexpectedly, our results may suggest a stronger association between REI and stroke recurrence in NHWs than MAs. Although MA ethnicity was associated with higher mortality than NHW ethnicity, this does not appear to be because of SDB, given the lack of confounding of the ethnicity-mortality relationship by REI and lack of an association between REI and mortality in MAs. We are not aware of other studies that have addressed how SDB may contribute to ethnic disparities in stroke recurrence and mortality.

Limitations to the work include the use of a home sleep apnea test to provide objective data on SDB rather the gold standard, laboratory-based polysomnography. However, the ApneaLink Plus has been well-validated; it is better tolerated in the poststroke population;[29] and it makes a sample of the present magnitude feasible. Although we did not collect information about continuous positive airway pressure use, treatment of SDB with continuous positive airway pressure is rare in this community among stroke patients,[30] and any use of continuous positive airway pressure would have attenuated our observed association.