Abstract
Background
The ethnic disparity in ischemic stroke between Mexican Americans (MAs) and non-Hispanic whites (NHWs) may be partly attributable to disparities in sleep and its disorders. We therefore assessed whether pre-stroke sleep apnea symptoms (SA risk) and pre-stroke sleep duration differed between MAs and NHWs.
Methods
MA and NHW ischemic stroke survivors in the Brain Attack Surveillance in Corpus Christi (BASIC) project reported sleep duration, and SA symptoms on the validated Berlin questionnaire, both with respect to their pre-stroke baseline. Log binomial and linear regression models were used to test the unadjusted and adjusted (demographics, vascular risk factors) associations of high-risk Berlin scores and sleep duration with ethnicity.
Results
Among 862 subjects, 549 (63.7%) were MA, and 514 (59.6%) had high risk for pre-stroke SA. The MA and NHW subjects showed no ethnic difference, after adjustment for potential confounders, in pre-stroke SA risk (risk ratio [95% CI]: 1.06 [0.93,1.20]) or in pre-stroke sleep duration (on average MAs slept 2.0 fewer minutes than NHWs, 95%CI: -18.8, 14.9 minutes).
Conclusions
Pre-stroke SA symptoms are highly prevalent, but ethnic differences in SA risk and sleep duration appear unlikely to explain ethnic stroke disparities.
Keywords: ethnicity, stroke, sleep apnea, sleep duration, sleep disordered breathing, insufficient sleep, Mexican American, cerebrovascular accident
1. Introduction
Mexican Americans (MAs), the largest subgroup of Hispanics, have a higher risk for first and recurrent stroke than do non-Hispanic whites (NHWs), with nearly double the risk at younger ages.[1] Reasons for this ethnic disparity are uncertain, but could include an increased prevalence of sleep disorders, such as sleep apnea (SA), an independent stroke risk factor.[2] Similarly, self-reports of short (< 7 hours) or long (> 9 hours) sleep duration are associated with an increased risk of stroke.[3] Studies suggest that Hispanics may have a higher prevalence of SA and higher odds of longer sleep compared to NHW.[4, 5]
Information about pre-stroke SA and sleep duration would allow hypotheses to be generated about the contributions of short or long sleep or disordered sleep to elevated stroke risk in MAs compared with NHWs. Our aim was to compare pre-stroke SA and sleep duration in MAs and NHWs included in a population-based stroke study. We hypothesized that MAs would have a higher prevalence of pre-stroke SA and shorter or longer sleep duration than NHWs.
2. Methods
Subjects were recruited between 7/8/2010 - 1/20/2014 from the Brain Attack Surveillance in Corpus Christi (BASIC) Project, an ongoing, population-based stroke surveillance study conducted in the bi-ethnic community of Nueces County, Texas. Our hospital surveillance captures nearly all cases of ischemic stroke and detailed methods have been published previously.[1, 6, 7] Briefly, stroke cases from all 7 acute care hospitals in Corpus Christi are identified through active surveillance of Emergency Department and hospital logs and passive surveillance of ICD-9 codes. Complete case capture is facilitated through the geographic isolation of the community with only sparsely populated surrounding areas. The lack of an academic medical center in the area limits referral bias. Study neurologists validate each case based on review of source documentation. The current analysis is restricted to ischemic stroke cases who agreed to undertake a study interview. The Institutional Review Boards of the University of Michigan and the local hospitals approved this project. Informed consent was obtained from each subject or surrogate.
Demographics and medical history were obtained from the medical record and a baseline interview, with the subject or a proxy if the subject was unable to participate. Ethnicity was defined by chart designation, which we have previously shown has a 96.3% agreement with self-report in this community.[8] During the baseline interview, the Berlin questionnaire, a validated SA screening tool, and a sleep duration question (“How many hours of sleep do you usually get a night (or when you usually sleep)?”) were asked in reference to the pre-stroke state.[9] The Berlin questionnaire has an 89% positive predictive value for detecting sleep apnea (apnea/hypopnea index (AHI)≥5) and 71% negative predictive value (detecting no sleep apnea with AHI<5) in a primary care setting.[9] The Berlin questionnaire was used to screen for sleep apnea in the Sleep in America 2005 poll and other studies.[10-13] The questionnaire consists of three categories that query snoring/apneas (category 1), sleepiness (category 2), and hypertension and body mass index (BMI)>30 (category 3). Each category is scored as low (0) or high risk (1) and an individual is considered high risk if two or more category scores are high risk. Total reported sleep duration was further categorized as short (≤6 hours), long (≥ 9 hours), or normal (7-8 hours), based on typical classifications.[14] With an estimated 50% prevalence for overall high risk pre-stroke SA, we had 82% power to detect a risk ratio of 1.2 for pre-stroke SA in MAs compared with NHWs given our sample size.
Descriptive statistics summarized baseline characteristics, Berlin scores, and sleep duration. Unadjusted ethnic comparisons for categorical and continuous variables, including sleep duration (hours), were made by chi-square and Kruskal-Wallis tests. Log binomial regression with robust standard errors was used to test the unadjusted and adjusted associations between high risk Berlin score overall and Berlin category scores and ethnicity. Two separate models (model 1 and model 2) were used to adjust for pre-specified potential confounders including age (continuous), sex, current smoking status, current alcohol intake (categorical: <1, 1-14, >14 drinks per week), diabetes, high cholesterol, atrial fibrillation, hypertension, and BMI (quintiles). Model 1 tested the adjusted association between high risk Berlin score overall or category 3 and ethnicity. Hypertension and BMI were not included in model 1 since these variables are part of the definitions of category 3 and the overall score. Hypertension and BMI were additionally included in model 2, which tested the adjusted association between category 1 or 2 and ethnicity. Linear regression with robust standard errors tested the unadjusted and adjusted (same pre-specified confounders) difference in sleep duration (≤6 hours, 7 or 8 hours, ≥ 9 hours) by ethnicity. SAS version 9.3 (SAS Institute, Cary, N.C., USA) was used for analyses.
3. Results
During the study period, there were 1,419 validated ischemic strokes among MAs and NHWs. There were 134 patients excluded because greater than 6 months elapsed between stroke and the interview, the interview was pending or they completed the interview, but not the sleep questions. Of 1,285 eligible patients with validated ischemic strokes, 862 (67.1%) completed the interview that included sleep questions; 549 (63.7%) of these subjects were MA. Baseline characteristics are shown in Table 1. Proxies completed the surveys for 272 (31.5%) of subjects, with no ethnic difference (p= 0.30). MAs had a greater prevalence of diabetes, hypertension, and a higher BMI, but a lower prevalence of atrial fibrillation, smoking, and lower use of alcohol compared with NHWs. MAs were also less likely to complete high school or post-high school education than NHWs.
Table 1. Baseline characteristics and pre-stroke sleep duration among stroke subjects in BASIC.
| All n=862 n (%) | Mexican American n=549 n (%) | non-Hispanic Whites n=313 n (%) | p-value | |
|---|---|---|---|---|
| Female | 419 (48.6) | 272 (49.5) | 147 (47.0) | 0.47 |
| Age, median (IQR) | 69 (59, 80) | 67 (58, 79) | 71 (61, 82) | 0.003 |
| Health insurance | ||||
| Nueces County, self pay, or not insured | 125 (14.5) | 83 (15.1) | 42 (13.4) | <0.001 |
| Medicare/Medicaid, VA/Tricare/Champs | 211 (24.5) | 158 (28.8) | 53 (16.9) | |
| HMO,PPO or HMO, PPO and Medicare | 526 (61.0) | 308 (56.1) | 218 (69.6) | |
| Education | ||||
| <High school | 340 (39.5) | 299 (54.7) | 41 (13.1) | <0.001 |
| High school | 231 (26.9) | 134 (24.5) | 97 (31.0) | |
| >High School | 289 (33.6) | 114 (20.8) | 175 (55.9) | |
| Current alcohol use | ||||
| none | 212 (24.6) | 169 (30.8) | 43 (13.7) | <0.01 |
| <1 | 380 (44.1) | 242 (44.1) | 138 (44.1) | |
| 1-14 | 220 (25.5) | 108 (19.7) | 112 (35.8) | |
| >14 | 50 (5.8) | 30 (5.5) | 20 (6.4) | |
| Current smoker | 182 (21.1) | 98 (17.9) | 84 (26.8) | 0.002 |
| Proxy response | 272 (31.5) | 180 (32.8) | 92 (29.4) | 0.3 |
| BMI: n, median (IQR) | 859, 27.44 (24.37, 32.42) | 547, 28.53 (24.96, 33.28) | 312, 26.32 (23.70, 30.27) | <0.001 |
| NIH score, median (IQR) | 4 (2, 9) | 4 (2, 9) | 4 (2, 10) | 0.72 |
| Diabetes | 403 (46.8) | 313 (57.0) | 90 (28.8) | <0.001 |
| Hypertension | 699 (81.1) | 467 (85.1) | 232 (74.1) | <0.01 |
| Atrial fibrillation | 140 (16.2) | 70 (12.8) | 70 (22.4) | <0.01 |
| Congestive heart failure | 94 (10.9) | 60 (10.9) | 34 (10.9) | 0.98 |
| Coronary artery disease | 270 (31.3) | 169 (30.8) | 101 (32.3) | 0.65 |
| Stroke/TIA history | 257 (29.8) | 173 (31.5) | 84 (26.8) | 0.15 |
| Sleep duration | ||||
| Total (hours), median (IQR) | 8 (6, 8) | 8 (6, 8) | 8 (6, 8) | 0.52 |
| Short, ≤6 h | 268 (31.4) | 177 (32.4) | 91 (29.6) | 0.67 |
| Normal, 7-8 h | 424 (49.7) | 269 (49.3) | 155 (50.5) | |
| Long , ≥ 9 h | 161 (18.9) | 100 (18.3) | 61 (19.9) |
Abbreviations: IQR = interquartile range, VA = veterans administration, HMO = health maintenance organization, PPO = preferred provider organization
Overall, 514 (59.6%) subjects had a high risk of pre-stroke SA, which was more common in MAs (61.9%) than NHWs (55.6%), but not significantly (p=0.07). Among the three Berlin categories, only high risk category 3 was more prevalent among MAs, given their higher average BMI and prevalence of hypertension (Table 2). After adjustment, no ethnic differences in SA risk or in the symptom-based categories, category 1 and 2 were found. Approximately half of all subjects reported normal sleep duration (7 or 8 hours), with no difference in sleep duration category distribution by ethnicity (Table 1). No difference in sleep duration was found by ethnicity before or after adjustment (Table 2). Additional post-hoc adjustment for education level did not significantly change the relationship between SA risk or sleep duration and ethnicity.
Table 2. Ethnic differences in pre-stroke Berlin score and sleep duration.
| All n=862 | Mexican Americans n=549 | Non-Hispanic Whites n=313 | Unadjusted RR or mean difference (95%CI) | Model 1 RR or mean difference (95%CI) | Model 2 RR or mean difference (95%CI) | |
|---|---|---|---|---|---|---|
| Sleep apnea risk | ||||||
| High risk overall | 514 (59.6) | 340 (61.9) | 174 (55.6) | 1.11 (0.99, 1.26) | 1.06 (0.93, 1.20) | - |
| High risk category 1 | 470 (54.5) | 311 (56.6) | 159 (50.8) | 1.12 (0.98, 1.27) | 1.09 (0.95, 1.25) | 1.05 (0.91, 1.20) |
| High risk category 2 | 276 (32.0) | 176 (32.1) | 100 (31.9) | 1.01 (0.82, 1.24) | 0.87 (0.7, 1.07) | 0.86 (0.69, 1.07) |
| High risk category 3a | 743 (86.2) | 495 (90.2) | 248 (79.2) | 1.13 (1.06, 1.21) | 1.11 (1.04, 1.18) | - |
| Sleep duration (minutes) | - | - | - | -5.4 (-22.1, 11.3) | -0.9 (-17.7, 15.8) | -2.0 (-18.9, 14.9) |
p<0.001
Model 1 adjustment: age (continuous), sex, current smoking status, current alcohol intake (categorical: <1, 1-14, >14 drinks weekly), diabetes, hyperlipidemia, atrial fibrillation.
Model 2 additionally adjusted: BMI (quintiles) and hypertension. Overall Berlin score and category 3 scores were not adjusted for BMI and hypertension because they are part of the score itself.
4. Discussion
In this population-based stroke study, we found that a high risk of pre-stroke SA is common in stroke patients, but the prevalence of pre-stroke SA and short or long sleep duration was not higher in MAs than NHWs. Given our sample size, we had 82% power to detect a risk ratio of 1.2 for pre-stroke SA in MAs compared to NHWs, a small but clinically meaningful difference. These results suggest that differences between MA and NHW stroke patients in sleep duration and SA risk are unlikely explanations for important disparities observed in their cerebrovascular health. Such disparities could still be explained, in theory, by more severe SA in MAs or a stronger association between SA and stroke among MAs,[2] and an alternative study design that included control subjects might have been more informative. However, the current negative results help to address a compelling question of vital importance that has never been investigated previously in a community sample of similar size.
One prior bi-ethnic study (n=176) did assess pre-stroke SA risk, by the Berlin questionnaire, among 77 Hispanic (mostly Cuban decent) and 21 NHW subjects and found a similar proportion of Hispanics (60%) at high risk for SA, but a lower proportion of high-risk NHWs (33%).[15] With adjustment, Hispanics had higher odds of SA. Differences in our results may relate to our larger sample, baseline differences in characteristics of NHWs, or our population-based design.
The main limitation of this study was the absence of objective measures for SA and sleep duration prior to stroke. Although the Berlin questionnaire may not be a good predictor of SA post-stroke, it has been validated in general populations, and our use was in reference to the pre-stroke state.[16] Therefore, we would anticipate that it would perform as well as in the general population, though using it in reference to the pre-stroke state may introduce recall bias. Because some stroke patients were unable to complete questionnaires due to aphasia or other stroke-related issues, some questionnaires were completed by proxies. However, bed partner responses to the Berlin questionnaire may predict sleep apnea based on polysomnographic criteria better than the patient, there is fair patient-proxy agreement for high risk pre-stroke SA based on Berlin questionnaire, and no ethnic difference in the use of proxies was identified (p=0.30).[17, 18] Self-report of sleep duration is variably correlated with more objective measures of total sleep time, and may better reflect other constructs such as sleep quality or time in bed.
This population-based study provides evidence for a high prevalence of pre-existing SA in stroke patients. However, no ethnic differences were identified in the prevalence of pre-stroke SA or sleep duration.
Highlights.
High risk of pre-stroke SA is common in ischemic stroke patients.
The prevalence of high pre-stroke SA risk was not higher in MAs than NHWs.
Short or long sleep duration prevalence was not higher in MAs than NHWs.
Sleep duration and pre-stroke SA risk does not explain bi-ethnic stroke disparities.
Acknowledgments
This study was funded by R01HL098065 and R01NS070941. Dr. De Lott is supported by T32NS7222.
Footnotes
Conflicts: The ICMJE Uniform Disclosure Form for Potential Conflicts of Interest associated with this article were submitted for each author.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Morgenstern LB, Smith MA, Sanchez BN, Brown DL, Zahuranec DB, Garcia N, et al. Persistent ischemic stroke disparities despite declining incidence in Mexican Americans. Annals of neurology. 2013;74:778–85. doi: 10.1002/ana.23972. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Yaggi HK, Concato J, Kernan WN, Lichtman JH, Brass LM, Mohsenin V. Obstructive sleep apnea as a risk factor for stroke and death. The New England journal of medicine. 2005;353:2034–41. doi: 10.1056/NEJMoa043104. [DOI] [PubMed] [Google Scholar]
- 3.Fang J, Wheaton AG, Ayala C. Sleep duration and history of stroke among adults from the USA. n/a-n/a. Journal of sleep research. 2014 doi: 10.1111/jsr.12160. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Kripke DF, Ancoli-Israel S, Klauber MR, Wingard DL, Mason WJ, Mullaney DJ. Prevalence of sleep-disordered breathing in ages 40-64 years: a population-based survey. Sleep. 1997;20:65–76. doi: 10.1093/sleep/20.1.65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Ramos AR, Wohlgemuth WK, Dong C, Gardener H, Wright CB, Boden-Albala B, et al. Race-ethnic differences of sleep symptoms in an elderly multi-ethnic cohort: the Northern Manhattan Study. Neuroepidemiology. 2011;37:210–5. doi: 10.1159/000334315. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Al-Wabil A, Smith MA, Moye LA, Burgin WS, Morgenstern LB. Improving efficiency of stroke research: the Brain Attack Surveillance in Corpus Christi study. Journal of clinical epidemiology. 2003;56:351–7. doi: 10.1016/s0895-4356(03)00005-2. [DOI] [PubMed] [Google Scholar]
- 7.Piriyawat P, Smajsova M, Smith MA, Pallegar S, Al-Wabil A, Garcia NM, et al. Comparison of active and passive surveillance for cerebrovascular disease: The Brain Attack Surveillance in Corpus Christi (BASIC) Project. American journal of epidemiology. 2002;156:1062–9. doi: 10.1093/aje/kwf152. [DOI] [PubMed] [Google Scholar]
- 8.Caveney AF, Smith MA, Morgenstern LB, Lisabeth LD. Use of death certificates to study ethnic-specific mortality. Public health reports (Washington, DC : 1974) 2006;121:275–81. doi: 10.1177/003335490612100309. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Netzer NC, Stoohs RA, Netzer CM, Clark K, Strohl KP. Using the Berlin Questionnaire to identify patients at risk for the sleep apnea syndrome. Annals of internal medicine. 1999;131:485–91. doi: 10.7326/0003-4819-131-7-199910050-00002. [DOI] [PubMed] [Google Scholar]
- 10.Netzer NC, Hoegel JJ, Loube D, Netzer CM, Hay B, Alvarez-Sala R, et al. Prevalence of symptoms and risk of sleep apnea in primary care. Chest. 2003;124:1406–14. doi: 10.1378/chest.124.4.1406. [DOI] [PubMed] [Google Scholar]
- 11.Hiestand DM, Britz P, Goldman M, Phillips B. Prevalence of symptoms and risk of sleep apnea in the US population: Results from the national sleep foundation sleep in America 2005 poll. Chest. 2006;130:780–6. doi: 10.1378/chest.130.3.780. [DOI] [PubMed] [Google Scholar]
- 12.Chung F, Yegneswaran B, Liao P, Chung SA, Vairavanathan S, Islam S, et al. Validation of the Berlin questionnaire and American Society of Anesthesiologists checklist as screening tools for obstructive sleep apnea in surgical patients. Anesthesiology. 2008;108:822–30. doi: 10.1097/ALN.0b013e31816d91b5. [DOI] [PubMed] [Google Scholar]
- 13.Blondet M, Yapor P, Latalladi-Ortega G, Alicea E, Torres-Palacios A, Rodriguez-Cintron W. Prevalence and risk factors for sleep disordered breathing in a Puerto Rican middle-aged population. Sleep & breathing = Schlaf & Atmung. 2009;13:175–80. doi: 10.1007/s11325-008-0216-4. [DOI] [PubMed] [Google Scholar]
- 14.Cappuccio FP, Cooper D, D'Elia L, Strazzullo P, Miller MA. Sleep duration predicts cardiovascular outcomes: a systematic review and meta-analysis of prospective studies. European heart journal. 2011;32:1484–92. doi: 10.1093/eurheartj/ehr007. [DOI] [PubMed] [Google Scholar]
- 15.Ramos AR, Guilliam D, Dib SI, Koch S. Race/ethnic differences in obstructive sleep apnea risk in patients with acute ischemic strokes in south Florida. Sleep & breathing = Schlaf & Atmung. 2014;18:165–8. doi: 10.1007/s11325-013-0865-9. [DOI] [PubMed] [Google Scholar]
- 16.Kotzian ST, Stanek JK, Pinter MM, Grossmann W, Saletu MT. Subjective evaluation of sleep apnea is not sufficient in stroke rehabilitation. Topics in stroke rehabilitation. 2012;19:45–53. doi: 10.1310/tsr1901-45. [DOI] [PubMed] [Google Scholar]
- 17.Sagaspe P, Leger D, Taillard J, Bayon V, Chaumet G, Philip P. Might the Berlin Sleep Questionnaire applied to bed partners be used to screen sleep apneic patients? Sleep medicine. 2010;11:479–83. doi: 10.1016/j.sleep.2010.01.007. [DOI] [PubMed] [Google Scholar]
- 18.Reeves SL, Brown DL, Chervin RD, Morgenstern LB, Smith MA, Lisabeth LD. Agreement between stroke patients and family members for ascertaining pre-stroke risk for sleep apnea. Sleep medicine. 2014;15:121–4. doi: 10.1016/j.sleep.2013.09.010. [DOI] [PMC free article] [PubMed] [Google Scholar]