2. Article Overview Title Prospective Study of Obstructive Sleep Apnea and Incident Coronary Heart Disease and Heart Failure: The Sleep Heart Health Study Study began in 1994 and ended in 2006 Submitted August 2009 Accepted May 2010 Published in Circulation July 2010
3. Sources Publication dates ranged from 1975-2009 Selection of Journals and Studies American Journal of Cardiology American Journal of Medicine Circulation Framingham Heart Study Journal of the American Medical Association Journal of Applied Physiology New England Journal of Medicine Sleep
4. Definitions Obstructive Sleep Apnea (OSA) Continuous Positive Airway Pressure Therapy Community Based Prospective Cohort Study Polysomnography Electroencephalogram Electrooculogram Chin electromyogram Inductance plethysmography Apnea Hypopnea Apnea-hypopneaindex (AHI) Incident CHD Incident Heart Failure Cox proportional hazards regression modeling Linear regression splines LOWESS Adjusted hazard ratio
5. Purpose of the Study “To assess the independent contribution of OSA to cardiovascular disease, the Sleep Heart Health Study (SHHS) was initiated in 1994 as a multicenter, prospective cohort study of the cardiovascular consequences of OSA.” “The present report details the incidence of CHD and heart failure in SHHS participants free of these conditions at the baseline examination.”
6. Study Design Adults 40 years of age and older were recruited from existing population-based studies of cardiovascular and pulmonary disease Participants completed questionnaires on sleep habits and general health Height, weight, and blood pressure were measured Underwent overnight polysomnography Parent cohorts provided data Ongoing surveillance for cardiovascular disease through April 2006
7. Study Sample 10,737 cohort participants invited to participate 6,441 were enrolled in study 760 participants from NYU-Cornell site excluded 783 participants excluded due to prevalent heart disease or heart failure at baseline 21 participants lacked follow-up data and were excluded 455 participants were missing baseline measures and excluded 4,422 subjects remained for participation 2,495 females 1,927 males
8. Methods In-home polysomnography Polysomnograms scored centrally and AHI determined Ongoing surveillance for CHD and heart failure by parent-cohorts Follow-up procedures performed by cohorts Median follow-up ranged from 8.3-9.2 years All CHD and heart failure outcomes extracted from hospital and physician’s records by trained abstractors ~5 years post-baseline, survey relative to diagnosis of and treatment for OSA completed by 3,794 participants
9. Covariates Baseline prevalent CHD or heart failure Smoking status Medication Diabetes Blood pressure Weight Race/ethnicity Height Total cholesterol HDL cholesterol
10. Statistical Analysis Performed with SAS version 9.2 Descriptive statistics presented by category of OSA severity based on the AHI Testing for improved model fit Cubic or quadratic functions Linear regression splines LOWESS Association between sex and AHI Models created for males and females
11. Statistical Analysis cont. Models using AHI as continuous dependent variable adjusted for the following independent variables: (1) Age, race, BMI, and smoking status (2) Variables from model (1) plus total and HDL cholesterol and diabetes (3) Variables from model (2) plus hypertension Models repeated using parent cohort instead of race Models constructed to stratify age >70 years old ≤ 70 years old
12. Results Increasing severity with OSA was associated with male sex, higher BMI, higher systolic blood pressure, lower HDL cholesterol, and higher prevalence of hypertension and diabetes Median AHI in women was 2.7 (interquartile range 0.8 to 7.5) and 6.2 in men (interquartile range 2.3 to 14.3) During follow-up period there were 473 incident CHD events 76 CHD deaths 185 MI’s 212 revascularization procedures
13. Results cont. During follow-up there were 308 incidents of heart failure; 144 also had incident CHD Rate of events increased with severity of OSA in men, but not as clearly in women When adjusted for age, race, BMI, and smoking status there was a significant association between AHI with incident CHD in men
14. Discussion This study found an association between incident CHD and OSA in men that was considerably weaker than previous clinic-based studies Screening of non-clinic-based population identified many asymptomatic individuals with OSA SHHS may underestimate the true cardiovascular risk associated with OSA due to median age of 62 Study demonstrates a 58% higher adjusted risk for incident heart failure for men with OSA
15. Strengths Community-based recruiting limited referral bias Detailed covariate data Exclusion criteria provided Few received treatment for OSA allowing for assessment of natural history of untreated OSA Includes both men and women Ethnically diverse
16. Weaknesses Older age of the cohort Echocardiograms were not routinely performed Varying protocols at different facilities Varying frequency of follow-ups Authors do not appear to be individuals who extracted data from medical records BMI is imperfect way to collect anthropometric data Diet and exercise not recorded
17. Significance Potentially significant because it does suggest an increased risk of incident heart failure in those with OSA Those with OSA tend to have all other risk factors predisposing them to heart disease, so it is hard to demonstrate if OSA has significant impact independent of those risk factors
