Inflammation and sudden cardiac death in a community-based population of older adults: The Cardiovascular Health Study

Author's Accepted Manuscript

Inflammation and Sudden Cardiac Death in a Community-based Population of Older Adults: The Cardiovascular Health Study Ayman A. Hussein MD, John S. Gottdiener MD, Traci M. Bartz MS, Nona Sotoodehnia MD, Christopher DeFilippi MD, Vincent See MD, Rajat Deo MD, David Siscovick MD, Phyllis K. Stein PhD, Donald Lloyd-Jones MD www.elsevier.com/locate/buildenv

PII: DOI: Reference:

S1547-5271(13)00729-7 http://dx.doi.org/10.1016/j.hrthm.2013.07.004 HRTHM5367

To appear in:

Heart Rhythm

Cite this article as: Ayman A. Hussein MD, John S. Gottdiener MD, Traci M. Bartz MS, Nona Sotoodehnia MD, Christopher DeFilippi MD, Vincent See MD, Rajat Deo MD, David Siscovick MD, Phyllis K. Stein PhD, Donald Lloyd-Jones MD, Inflammation and Sudden Cardiac Death in a Community-based Population of Older Adults: The Cardiovascular Health Study, Heart Rhythm, http://dx.doi.org/10.1016/j. hrthm.2013.07.004 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 galley 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.

1

Inflammation and Sudden Cardiac Death in a Community-based Population of Older Adults: The Cardiovascular Health Study Ayman A. Hussein1, MD, John S. Gottdiener1, MD, Traci M. Bartz2, MS, Nona Sotoodehnia2, MD, Christopher DeFilippi1, MD, Vincent See1, MD, Rajat Deo3, MD, David Siscovick2, MD, Phyllis K. Stein4, PhD and Donald Lloyd-Jones5, MD. 1

University of Maryland, Baltimore, MD University of Washington, Seattle, WA 3 University of Pennsylvania, Philadelphia, PA 4 Washington University, St Louis, MO 5 Northwestern University, Evanston, IL 2

Running title: Inflammation and sudden cardiac death Word count (including text, abstract, references, legends and tables): 4949 words Conflict of interest / Relationship with industry: None Funding sources: The research reported in this article was supported by contracts HHSN268201200036C, N01-HC85239, N01-HC-85079 through N01-HC-85086, N01-HC-35129, N01 HC-15103, N01 HC-55222, N01-HC-75150, N01-HC-45133, and grant HL080295 from the National Heart, Lung, and Blood Institute (NHLBI), with additional contribution from the National Institute of Neurological Disorders and Stroke (NINDS). Additional support was provided through AG-023629, AG-15928, AG-20098, and AG-027058 from the National Institute on Aging (NIA). A full list of principal CHS investigators and institutions can be found at http://www.chs-nhlbi.org/pi.htm. Address of correspondence: John S. Gottdiener MD Director, Echocardiography Laboratory Professor of Medicine, University of Maryland Medical Center 110 S. Paca St. Floor 7, Baltimore, MD 21201. Tel: 410 328 6190. Fax 410 328 3530 Abstract Objectives: To study the association between inflammation and sudden cardiac death (SCD) in a community-based population of older adults. Background: Inflammation is linked to adverse cardiovascular events but its association with SCD has been controversial. Older subjects, who are at particular risk for SCD, were underrepresented in previous studies addressing this issue. Methods: In the Cardiovascular Health Study, 5806 and 5382 participants had measurements of C-reactive protein (CRP) and Interleukin-6 (IL6), respectively; and were followed for up to 17 years. SCD risk as a function of baseline IL-6 and CRP was assessed in the overall population and a group of participants without known prevalent cardiac disease. Results: In univariate analyses, both IL-6 [Hazard ratio (HR) 1.79 for 1+ log IL-6, 95% confidence interval (CI) 1.50-2.13; 5th vs. 1st quintile HR 3.36, 95%CI 2.24-5.05] and CRP (HR 1.31 for 1+ log CRP, 95%CI 1.18-1.45; 5th

2

vs. 1st quintile HR 2.00, 95%CI 1.40-2.87) were associated with SCD risk.

In covariate-adjusted analyses,

accounting for baseline risk factors, incident myocardial infarction and heart failure, the association with SCD risk persisted for IL-6 (HR 1.26 for 1+ log IL-6, 95%CI 1.02-1.56, 5th vs. 1st quintile HR 1.63, 95%CI 1.03-2.56), but was significantly attenuated for CRP (HR 1.13 for 1+ log CRP, 95%CI 1.00-1.28, 5th vs. 1st quintile HR 1.34, 95%CI 0.88-2.05). Similar findings were observed in participants without prevalent cardiac disease. Conclusion: Greater burden of inflammation, assessed by IL-6 levels, is associated with SCD risk beyond traditional risk factors, incident myocardial infarction and heart failure. Key words: death, sudden; inflammation

Abbreviations list Sudden cardiac death (SCD) Coronary heart disease (CAD) Cardiovascular Health Study (CHS) C-reactive protein (CRP) Interleukin-6 (IL6) HF (HF) Coefficient of variation (CV) Enzyme-linked immunosorbent assay (ELISA) Myocardial infarction (MI) Left ventricular ejection fraction (LVEF) Transient ischemic attack (TIA) High-density lipoprotein cholesterol (HDL-C) Estimated glomerular filtration rate (eGFR) Hazard ratios (HR) Confidence intervals (CI)

ptrend across quintiles (pt) Introduction Sudden cardiac death (SCD) is a major public health problem. (1, 2) Over the past few decades, primary and secondary prevention measures have led to a significant decline in cardiovascular mortality from coronary heart disease (CAD), but the rates of SCD have declined to a lesser extent (2) which emphasizes the need for a better understanding of SCD epidemiology and risk factors in order to reduce its burden in modern societies. It has been suggested by epidemiological and pathological studies that coronary atherosclerosis, plaque inflammation and acute plaque rupture underlie most cases of SCD. (3, 4) In patients without identifiable

3

macroscopic cause of SCD, focal areas of myocardial inflammation have been reported. (5, 6) Previous observations, therefore, suggest that inflammation may underlie a significant subset of SCD events. In fact, elevated levels of inflammatory biomarkers have been associated with multiple cardiovascular endpoints, (7-9) including SCD. (10) However, the relationship between inflammation and SCD risk is controversial with data suggesting that the association is no longer observed after adjusting for traditional risk factors in multivariable analyses. (11, 12) Furthermore, older subjects, who are at particular risk for SCD, (2, 13, 14) were largely underrepresented in previous studies addressing inflammation and SCD risk. We aimed to study the association between inflammation and SCD risk in a large community-based ambulatory population of older adults. We hypothesized that a greater burden of inflammation is associated with an increased risk of SCD beyond traditional risk factors. Methods Study population, assessment and cardiovascular events The Cardiovascular Health Study (CHS) is a longitudinal community-based study of adults aged 65 years or older at recruitment. The rationale, design and methods of CHS have been previously published. (15, 16) The CHS original (n=5201) and minority (n=687) cohorts were enrolled from 1989 to 1990 and from 1992 to 1993, respectively. The current analysis included 5806 participants who had measurements of inflammatory biomarkers, C-reactive protein (CRP) and interleukin-6 (IL6), from sera collected at enrollment (Figure 1). The CHS was approved by the institutional review boards at the participating sites. All subjects gave written informed consent at time of enrollment. Baseline assessments included a standardized questionnaire that assessed for variable health and behavioral risk factors. For each cardiovascular condition, self-reports were confirmed by components of the baseline examination, review of medical records or surveys of treating physicians. For instance, the baseline examination sought to identify major Q waves or the combination of minor Q waves and ST-T-wave changes on electrocardiograms to confirm self-reported myocardial infarction (MI). Self-reported heart failure (HF) was confirmed by symptoms, physical signs and the use of both diuretics and either digitalis or a vasodilator. Further confirmation of MI or HF, was sought from treating physicians by questionnaire or from hospitals by discharge

4

summaries, as well as by review of medical records. Coronary heart disease was defined as a history of angina, coronary revascularization, or previous MI. All subjects had physical examinations at baseline. (17) After the initial assessment, enrolled subjects were contacted every 6 months for follow-up, alternating between telephone interviews and clinic visits through 19981999 and by telephone interviews only thereafter. New cardiovascular events, including cardiovascular deaths, incident HF and MI, were adjudicated by a centralized events committee. (17) Information on death was obtained from reviews of medical records, physician questionnaires, interviews with next-of-kin, death certificates, autopsy reports and coroners’ reports. The ascertainment of death was 100%. To identify cases of SCD, all cases of fatal cardiovascular death were reviewed and re-adjudicated. SCD was defined as a sudden pulseless condition presumed due to a malignant ventricular arrhythmia in a previously stable individual without evidence of a non-cardiac cause of the arrest. We a priori sought to exclude cases with non-arrhythmic characteristics, including those with evidence of progressive hypotension or advanced congestive HF before death. All SCD events in this analysis occurred out of the hospital or in the emergency room. All deaths that occurred under hospice, nursing home care or in subjects with life-threatening non-cardiac comorbidities were not considered SCD.

Available data from death certificates, informant interviews, physician questionnaires,

coroner reports, and hospital discharge summaries were reviewed, in addition to circumstances surrounding the event, to help classify whether the subject had experienced SCD. For non-witnessed deaths, the participant must have been seen within 24 hours of the arrest in a stable condition and without evidence of a non-cardiac cause of the arrest. Each cardiovascular death identified from the central adjudication process was re-adjudicated by a cardiologist’s record review in order to identify those who had SCD. A blinded second physician review of a random sample of 70 of these death records showed an 88% inter-reviewer agreement and =0.74 for SCD. (18) Inflammatory biomarkers assays Blood sample acquisition as well as analytical and quality-assurance methods in CHS have been previously reported. (19) All measurements of inflammatory biomarkers were performed in a central blood analysis laboratory. Baseline levels of CRP and IL-6 were measured in blood samples collected at baseline for the original CHS cohort (1989-90) and the minority cohort (1992-93). These samples were stored at -80ºC for 5-6 years prior to the measurements and thawed just prior to laboratory assays (maximum of 3 freeze-thaw cycles). CRP was measured

5

by using stored plasma and a high-sensitivity enzyme-linked immunosorbent assay (ELISA) (20) developed and validated in CHS laboratories (analytical coefficient of variation “CV” 5.1%). (21) IL-6 was measured from stored sera using a commercial ELISA kit (Quantikine IL-6, R&D Systems, Minneapolis, USA; CV 6.3%). (22) IL-6 measurements were obtained for fewer subjects (n=5380) than CRP (n=5806) due to insufficient serum volume and higher volume requirements for the IL-6 assay. Technologists performing the assays were blinded to the outcome of interest. Statistical analysis The hypothesis of association between higher levels of inflammatory biomarkers and SCD risk was tested with IL-6 and CRP evaluated as continuous variables (Log transformed values), as well as categorized into quintiles. The risk of SCD as a function of inflammatory biomarkers was assessed in the overall population and a group of participants without known prevalent cardiac disease at enrollment (n=3035 with CRP; 2842 with IL-6) (Figure 1). This secondary analysis excluded subjects with CAD, previous MI or HF, reduced left ventricular ejection fraction (LVEF<55%), valvular disease, atrial fibrillation, ventricular conduction delay or prolonged QT interval and those with missing information regarding any of these variables. The incidence rates of SCD in IL-6 and CRP quintiles were estimated by using the Kaplan-Meier method and compared with the log-rank test. Cox proportional hazards models were employed to evaluate the association between baseline CRP and IL-6 levels and SCD risk. Models were adjusted for demographics as well as baseline factors found to have a statistically significant association with SCD in univariable analyses (p<0.05). These factors included age, race, gender, smoking, history of HF, CAD, MI, stroke or transient ischemic attack (TIA), qualitative LVEF (categorical variable: <45% vs. 45-55% vs. >55%), left ventricular mass by electrocardiography, ventricular conduction delay, abnormal Q waves and prolonged QT intervals, levels of total and high-density lipoprotein cholesterol (HDL-C), serum fasting glucose, glomerular filtration rate (eGFR) by the modified diet in renal disease formula, systolic blood pressure and use of aspirin, digoxin, antiarrhythmic and antihypertensive medications. Additionally, we adjusted the analyses for incident HF and MI to account for the well-established relationship between these conditions and SCD. Participants who had no evidence of MI or HF at baseline, but had an event during follow-up had an updated status for the duration of follow-up. Similar analyses were performed in the group of participants without known cardiac disease and adjusted for the above risk factors, as appropriate. Finally, in a

6

subgroup of participants with serial echocardiograms, multivariable analyses additionally adjusted for change in LVEF over time. Hazard ratios (HR) and their 95% confidence intervals (CI) are reported. Statistical analyses were performed using JMP pro version 9.0 (SAS; NC, USA) and STATA version 12.1(StataCorp; TX, USA). P-values <0.05 were considered statistically significant. Results The demographics and clinical characteristics of the study population (n=5806) are summarized in Table 1. Over a median follow-up of 13.1 years (interquartile range 7.9-16.3), there were 317 (5.5%) and 284 (5.3%) cases of SCD in participants with CRP and IL-6 measures, respectively (Figure 1). Traditional cardiovascular risk factors were more common among SCD cases than participants without SCD (Table 1). In participants without known prevalent cardiac disease at enrollment, there were 101 (3.3%) and 91 (3.2%) SCD cases in subjects with available CRP and IL-6 measures, respectively (Figure 1). In the overall population, there was a significant association between baseline levels of inflammatory biomarkers and SCD risk (Figure 2A,2B). The incidence rate of SCD increased from 3.6 per 1000 person-years in the first CRP quintile to 6.2 per 1000 person-years in the 5th CRP quintile (Figure 2A, ptrend across quintiles “pt”<0.0001) and from 2.6 per 1000 person-years in the first IL-6 quintile to 6.5 per 1000 person-years in the 5th IL-6 quintile (Figure 2B, pt<0.0001). In proportional hazards analyses, the unadjusted HRs per log increase of IL-6 and CRP were 1.79 (95%CI 1.50-2.13) and 1.31 (95%CI 1.18-1.45), respectively. In categorical analyses, there was a significant increase in SCD risk with increasing IL-6 (5th vs. 1st quintile unadjusted HR 3.36, 95%CI 2.24-5.05, pt<0.0001) and CRP (5th vs. 1st quintile unadjusted HR 2.00, 95%CI 1.40-2.87, pt<0.0001) quintiles. In covariateadjusted analyses accounting for demographics and baseline risk factors, the association between inflammatory biomarkers levels and SCD risk persisted for both IL-6 (HR per Log IL-6 increase 1.33, 95%CI 1.08-1.65, 5th vs. 1st quintile HR 1.80, 95%CI 1.14-2.83, pt=0.006) and CRP (HR per log CRP increase 1.16, 95%CI 1.03-1.31; 5th vs. 1st quintile HR 1.46, 95%CI 0.96-2.23, pt=0.008) (Table 2). In time-dependent models which, in addition to baseline risk factors, accounted for incident HF and MI; IL-6 levels remained associated with SCD risk (HR per Log IL-6 increase 1.26, 95%CI 1.02-1.56, 5th vs. 1st quintile HR

1.63, 95%CI 1.03-2.56, pt=0.03) but the

7

association between CRP levels and SCD was significantly attenuated (HR per log CRP increase 1.13, 95%CI 1.001.28; 5th vs. 1st quintile HR 1.34, 95%CI 0.88-2.05, pt=0.02) (Table 2). The association between baseline IL-6 quintiles and SCD was also observed in participants without known cardiac disease at enrollment (Figure 2, Panel D). The incidence rate of SCD increased from 1.4 per 1000 personyears in the first IL-6 quintile to 3.8 per 1000 person-years in the 5th IL-6 quintile (pt=0.01). There was no significant trend in SCD incidence rates across CRP quintiles (Figure 2, Panel D, pt=0.2). In unadjusted proportional hazards analyses, there was a significant association between IL-6 levels and SCD risk (HR per log increase of IL-6 1.62, 95%CI 1.19-2.21; 5th vs. 1st quintile HR 2.81, 95%CI 1.38-5.69, pt=0.002) but only a weak association between CRP levels and SCD (HR per log CRP increase 1.21, 95%CI 1.00-1.47; 5th vs. 1st quintile HR 1.56, 95%CI 0.87-2.81, pt=0.04) (Table 3). Similar findings were observed in covariate-adjusted analyses, accounting for incident HF and MI (Table 3). Finally, 3017 participants had follow-up echocardiograms (180 SCD cases). In multivariable models, accounting for baseline factors, incident HF and MI as well as change in LVEF over time, baseline CRP levels were associated with SCD only when treated as a categorical variable (5th vs. 1st quintile HR 1.27, 95%CI 0.82-1.96, pt=0.047; HR per log increase of CRP 1.11, 95%CI 0.97-1.26, p=0.1). However, in similar analyses, IL-6 levels were associated with SCD regardless of whether they were treated as categorical or continuous variable (5th vs. 1st quintile HR 1.74, 95%CI 1.08-2.81, pt=0.02; HR per log increase of CRP 1.26, 95%CI 1.01-1.57, p=0.04). Discussion In a large community-based population of older adults followed for up to 17 years, there was a significant association between inflammation and long term risk of SCD. The association between inflammatory biomarkers, IL-6 in particular, and SCD risk persisted in covariate-adjusted analyses accounting for a series of traditional risk factors. Furthermore, the association between IL-6 levels and SCD risk was not significantly attenuated after adjusting for change in LVEF over time, incident HF and myocardial infarction, showing no important confounding effect which represents an important and novel finding. In addition, in this group of elderly subjects, the association between IL-6 levels and SCD risk was more robust compared to CRP levels, suggesting that IL-6 may provide useful information regarding SCD risk beyond traditional risk factors. Importantly, the association between IL-6 levels and SCD risk was also observed in a group of participants without known prevalent cardiac disease at

8

enrollment which has significant clinical implications, given the challenging nature of SCD risk assessment in such individuals in the community. Previously published data on the association between inflammatory biomarkers and SCD risk have been controversial. The observation of elevated levels of CRP in CAD patients who died suddenly (23)

raised

researchers’ interest in assessing the relationship between inflammation and SCD risk in various populations. In the Physician's Health Study, elevated CRP levels were associated with SCD. (10) More recent data from the PRIME (11) and Nurses’ Health (12) studies suggested that the association of inflammation, assessed by CRP levels, and SCD risk is mediated by comorbid conditions and was not observed in multivariable analyses accounting for traditional risk factors. In PRIME, (11) however, levels of IL-6 were found to be associated with SCD. Other published reports on inflammation and SCD have addressed high risk populations such as ischemic cardiomyopathy patients receiving implantable cardioverter-defibrillators. In such patients, initial observations have suggested a significant association between CRP levels and malignant ventricular arrhythmias, (24) but a subsequent prospective study did not observe an association between CRP levels and arrhythmic death. (25) Compared to the initial post-mortem data, which provided information on a temporal association between inflammation and SCD events, (29) the current findings suggest a long term association in which elevated levels of inflammatory biomarkers were associated with future SCD events. Furthermore, compared to previous studies that reported a significant association between inflammation and SCD, the current study has the strengths of larger population size, (10, 11) longer duration of follow-up, (10, 11) and the adjustment for incident HF and MI. Moreover, it extends prior observations of a significant association between inflammation and SCD risk to the general population of older adults. This population is of particular interest given that it is a growing segment in modern societies and contributes the majority of SCD cases to the community. (2, 13, 14) The pathophysiological mechanisms underlying the association between inflammation and SCD need further investigation. Potential hypothetical explanations include an association with subclinical atherosclerosis, poor general health state or direct involvement in arrhythmogenesis leading to SCD. Traditionally, inflammation has been suggested to play an important role in the initiation and progression of coronary atherosclerosis as well as the conversion of stable plaques to an unstable phenotype. (7, 8, 26-29) It is possible that the observations in the current study reflect an advanced subclinical atherosclerotic burden (30) in subjects with higher inflammatory

9

burden, (22, 31) and hence the association with SCD. The association between inflammation and acute coronary events is well established, and CAD is the most common pathologic substrate in SCD victims. (2) It is possible that many SCD victims develop myocardial ischemia prior to their collapse, but in contrast to those who present to the emergency rooms with acute coronary syndromes, SCD victims experience a sudden circulatory collapse, most commonly from a life threatening arrhythmia. The association between inflammation and arrhythmogenesis is possible but there is a paucity of published data. In fact, the association between CRP levels and malignant ventricular arrhythmias is still controversial, (24, 25) whereas the association between IL-6 levels and life-threatening arrhythmias has not been investigated to date. The observations from the current study suggest that this issue deserves further investigation. It is possible that a higher burden of inflammation is associated with structural myocardial changes, such as scarring and fibrosis, which in turn may be arrhythmogenic. (5, 6) This process may be even more pertinent in older SCD patients who are much less likely to have acute coronary thrombosis than younger SCD victims. (32) The possibility of direct involvement in SCD pathogenesis is suggested by the stronger association of IL-6, the true indicator of inflammatory burden compared to CRP as its primary inducer in the liver, (33) with SCD risk, and is concordant with previously published reports which suggest that IL-6 is a better predictor than CRP of adverse cardiovascular outcomes. (20, 34) This explanation is, however, hypothetical and deserves further investigation. Study limitations The study has the inherent limitations of observational studies and the findings, therefore, may have been affected by residual confounding and do not establish causality. However, the associations between IL-6 levels and SCD risk persisted after adjusting for an extensive number of risk factors that are usually assessed in clinical practice, as well as incident HF and MI suggesting that inflammation may be involved in a pathophysiological process which may lead to sudden death independent of these conditions. Second, our observations are made in population of older adults and therefore may not necessarily be generalized to younger populations. Third, information about defibrillators was not available and these could not be accounted for in multivariable analyses. However, the analyses adjusted for baseline risk factors and incident conditions that otherwise could indicate the need for defibrillators. Finally, there is no standard operational definition that has consistently been used across

10

epidemiologic studies. Nonetheless, despite differences in populations, study designs, the availability of clinical data and the operational definitions used to classify SCD, the results of prior epidemiologic studies focusing on potential determinants of SCD typically have been consistent. In CHS, SCD was adjudicated from medical records and interviews with next of kin as to circumstances surrounding the event, and most cases did not have an autopsy and heart rhythm monitoring at time of sudden collapse. Misclassification of the outcome is possible, and would tend to underestimate the true association. Importantly, the operational definition used to define SCD in this report has been used in multiple previous studies from CHS and the findings from these analyses have replicated findings from studies with other operational definitions of SCD. Conclusion In a large community-based population of older adults, we found a significant association between inflammation and sudden death risk. This relationship was also observed in a group of participants without cardiac disease and persisted in covariate-adjusted analyses accounting for an extensive numbers of risk factors including incident HF and MI. The findings suggest that a greater burden of inflammation, assessed by IL-6 measures, in older adults is associated with increased sudden death risk beyond traditional cardiovascular risk factors.

References 1. Roger VL, Go AS, Lloyd-Jones DM, et al. Heart disease and stroke statistics--2011 update: A report from the American Heart Association. Circulation. 2011;123:e18-e209.

2. Deo R, Albert CM. Epidemiology and genetics of sudden cardiac death. Circulation. 2012;125:620-37.

3. Manfredini R, Portaluppi F, Grandi E, Fersini C, Gallerani M. Out-of-hospital sudden death referring to an emergency department. J Clin Epidemiol. 1996;49:865-8.

4. Sudha ML, Sundaram S, Purushothaman KR, Kumar PS, Prathiba D. Coronary atherosclerosis in sudden cardiac death: An autopsy study. Indian J Pathol Microbiol. 2009;52:486-9.

5. Lecomte D, Fornes P, Fouret P, Nicolas G. Isolated myocardial fibrosis as a cause of sudden cardiac death and its possible relation to myocarditis. J Forensic Sci. 1993;38:617-21.

11

6. Wang HY, Huang WL, Yang CF, et al. [Morphologic features of sudden cardiac death in yunnan province, with emphasis on myocarditis]. Chung-hua Ping Li Hsueh Tsa Chih. 2007;36:805-9.

7. Ridker PM, Cushman M, Stampfer MJ, Tracy RP, Hennekens CH. Inflammation, aspirin, and the risk of cardiovascular disease in apparently healthy men. N Engl J Med. 1997;336:973-9.

8. Ridker PM, Hennekens CH, Buring JE, Rifai N. C-reactive protein and other markers of inflammation in the prediction of cardiovascular disease in women. N Engl J Med. 2000;342:836-43.

9. Ridker PM, Stampfer MJ, Rifai N. Novel risk factors for systemic atherosclerosis: A comparison of C-reactive protein, fibrinogen, homocysteine, lipoprotein(a), and standard cholesterol screening as predictors of peripheral arterial disease. JAMA. 2001;285:2481-5.

10. Albert CM, Ma J, Rifai N, Stampfer MJ, Ridker PM. Prospective study of C-reactive protein, homocysteine, and plasma lipid levels as predictors of sudden cardiac death. Circulation. 2002;105:2595-9.

11. Empana JP, Jouven X, Canoui-Poitrine F, et al. C-reactive protein, interleukin 6, fibrinogen and risk of sudden death in European middle-aged men: The PRIME study. Arterioscler Thromb Vasc Biol. 2010;30:2047-52.

12. Korngold EC, Januzzi JL,Jr, Gantzer ML, Moorthy MV, Cook NR, Albert CM. Amino-terminal pro-B-type natriuretic peptide and high-sensitivity C-reactive protein as predictors of sudden cardiac death among women. Circulation. 2009;119:2868-76.

13. Kannel WB, Wilson PW, D'Agostino RB, Cobb J. Sudden coronary death in women. Am Heart J. 1998;136:205-12.

14. Albert CM, Chae CU, Grodstein F, et al. Prospective study of sudden cardiac death among women in the United States. Circulation. 2003;107:2096-101.

15. Fried LP, Borhani NO, Enright P, et al. The cardiovascular health study: Design and rationale. Ann Epidemiol. 1991;1:263-76.

12

16. Psaty BM, Kuller LH, Bild D, et al. Methods of assessing prevalent cardiovascular disease in the cardiovascular health study. Ann Epidemiol. 1995;5:270-7.

17. Ives DG, Fitzpatrick AL, Bild DE, et al. Surveillance and ascertainment of cardiovascular events. the cardiovascular health study. Ann Epidemiol. 1995;5:278-85.

18. Sotoodehnia N, Siscovick DS, Vatta M, et al. Beta2-adrenergic receptor genetic variants and risk of sudden cardiac death. Circulation. 2006;113:1842-8.

19. Cushman M, Cornell ES, Howard PR, Bovill EG, Tracy RP. Laboratory methods and quality assurance in the cardiovascular health study. Clin Chem. 1995;41:264-70.

20. Harris TB, Ferrucci L, Tracy RP, et al. Associations of elevated interleukin-6 and C-reactive protein levels with mortality in the elderly. Am J Med. 1999;106:506-12.

21. Macy EM, Hayes TE, Tracy RP. Variability in the measurement of C-reactive protein in healthy subjects: Implications for reference intervals and epidemiological applications. Clin Chem. 1997;43:52-8.

22. Jenny NS, Tracy RP, Ogg MS, et al. In the elderly, interleukin-6 plasma levels and the -174G>C polymorphism are associated with the development of cardiovascular disease. Arterioscler Thromb Vasc Biol. 2002;22:2066-71.

23. Burke AP, Tracy RP, Kolodgie F, et al. Elevated C-reactive protein values and atherosclerosis in sudden coronary death: Association with different pathologies. Circulation. 2002;105:2019-23.

24. Biasucci LM, Giubilato G, Biondi-Zoccai G, et al. C reactive protein is associated with malignant ventricular arrhythmias in patients with ischaemia with implantable cardioverter-defibrillator. Heart. 2006;92:1147-8.

25. Biasucci LM, Bellocci F, Landolina M, et al. Risk stratification of ischaemic patients with implantable cardioverter defibrillators by C-reactive protein and a multi-markers strategy: Results of the CAMI-GUIDE study. Eur Heart J. 2012;33:1344-50.

13

26. Kuller LH, Tracy RP, Shaten J, Meilahn EN. Relation of C-reactive protein and coronary heart disease in the MRFIT nested case-control study. multiple risk factor intervention trial. Am J Epidemiol. 1996;144:537-47.

27. Koenig W, Khuseyinova N, Baumert J, et al. Increased concentrations of C-reactive protein and IL-6 but not IL-18 are independently associated with incident coronary events in middle-aged men and women: Results from the MONICA/KORA Augsburg case-cohort study, 1984-2002. Arterioscler Thromb Vasc Biol. 2006;26:2745-51.

28. Koenig W, Lowel H, Baumert J, Meisinger C. C-reactive protein modulates risk prediction based on the Framingham score: Implications for future risk assessment: Results from a large cohort study in southern Germany. Circulation. 2004;109:1349-53.

29. Luc G, Bard JM, Juhan-Vague I, et al. C-reactive protein, interleukin-6, and fibrinogen as predictors of coronary heart disease: The PRIME study. Arterioscler Thromb Vasc Biol. 2003;23:1255-61.

30. Kuller LH, Shemanski L, Psaty BM, et al. Subclinical disease as an independent risk factor for cardiovascular disease. Circulation. 1995;92:720-6.

31. Cesari M, Penninx BW, Newman AB, et al. Inflammatory markers and cardiovascular disease (the health, aging and body composition study). Am J Cardiol. 2003;92:522-8.

32. Burke AP, Farb A, Pestaner J, et al. Traditional risk factors and the incidence of sudden coronary death with and without coronary thrombosis in blacks. Circulation. 2002;105:419-24.

33. Biasucci LM, Vitelli A, Liuzzo G, et al. Elevated levels of interleukin-6 in unstable angina. Circulation. 1996;94:874-7.

34. Cesari M, Penninx BW, Newman AB, et al. Inflammatory markers and onset of cardiovascular events: Results from the health ABC study. Circulation. 2003;108:2317-22.

14

Figure legends: Figure 1: Enrollment and analysis flow of the study population. CRP: C-reactive protein. IL-6: interleukin-6. CHS: Cardiovascular Health Study. Figure 2: Incidence rates of sudden cardiac death (SCD) in participants of the Cardiovascular Health Study according to quintiles of baseline levels of interleukin-6 (IL-6) and C-reactive protein (CRP).

15

Table 1. Baseline demographics and clinical characteristics. Numbers are percentages unless otherwise specified. * mean±SD. **median (interquartile range)

Characteristics N Age, years* Male gender Race, African American Current smoker Diabetes mellitus Body mass index, Kg/m2* Coronary heart disease History of myocardial infarction History of stroke History of transient ischemic attack Hypertension Left ventricular ejection fraction> 55% Ventricular conduction delay Atrial fibrillation Prolonged QT interval Triglycerides, mg/dl** High-density lipoprotein cholesterol, mg/dl* Total cholesterol, mg/dl* Low-density lipoprotein cholesterol, mg/dl* Fasting glucose, mg/dl* Systolic blood pressure, mmHg* Diastolic blood pressure, mmHg* Aspirin Antihypertensive medications Diuretics Beta blockers Angiotensin converting enzyme inhibitors Statins Antiarrhythmic medications Class Ia Class Ib Class Ic Class III Digitalis

All 5806 72.8±5.6 42.5 15.3 12.0 16.2 26.7±4.7 19.5 9.6 4.2 2.8 44.3 90.1 9.5 2.7 14.2 120 (92-164) 54.2±15.8 211.2±39.2 129.8±35.6 111.2±37.3 136.5±21.8 70.7±11.4 33.8 47.3 29.0 12.8 7.4 2.3

no SCD 5489 72.8±5.6 41.5 15.1 11.9 15.6 26.7±4.7 18.6 8.8 4.0 2.7 44.0 91.7 9.0 2.6 13.7 119 (92-163) 54.5±15.8 211.4±39.1 129.9±35.6 110.7±36.7 136.3±21.7 70.7±11.4 33.4 47.0 28.7 12.6 7.2 2.3

SCD 317 73.4±5.8 58.7 18.6 13.9 26.7 27.0±4.4 36.0 22.4 8.5 4.7 50.2 75.7 18.4 3.5 22.6 128 (97-177) 50.1±14.7 207.8±41.5 128.9±37.0 119.2±45.9 139.5±22.7 71.4±11.9 40.7 54.0 34.4 15.8 9.8 2.2

p value 0.05 <0.0001 0.1 0.29 <0.0001 0.26 <0.0001 <0.0001 0.0005 0.05 0.03 <0.0001 <0.0001 0.37 <0.0001 0.02 <0.0001 0.15 0.65 0.001 0.01 0.29 0.008 0.02 0.03 0.11 0.11 0.95

2.2 1.0 0.1 0.1 8.4

2.2 1.0 0.1 0.1 8.2

2.5 2.2 1.0 0.3 12.9

0.67 0.06 0.006 0.2 0.005

Table 2. Cox analyses of the association between interleukin-6 (IL-6) and C-reactive protein (CRP) levels and sudden cardiac death risk in the Cardiovascular Health Study.

16

HR: hazard ratio. CI: confidence interval. * adjusted for age, race, gender, smoking, history of HF, coronary disease, MI, stroke or transient ischemic attack, left ventricular ejection fraction, left ventricular mass by electrocardiography, ventricular conduction delay, abnormal Q waves and prolonged QT intervals on electrocardiograms, levels of total and high-density lipoprotein cholesterol, serum fasting glucose, glomerular filtration rate estimated by the modified diet in renal disease formula, systolic blood pressure and use of aspirin, antiarrhythmic medications, digoxin, and antihypertensive medications including angiotensin-converting enzyme inhibitors and diuretics. **adjusted for above parameters, incident HF and MI (time-dependent analyses)

HR IL-6 +1 Log increase Quintiles Q1 (reference) Q2 Q3 Q4 Q5 CRP +1 Log increase Quintiles Q1 (reference) Q2 Q3 Q4 Q5

1.79

Univariate 95%CI P value 1.50-2.13

1 1.55 2.42 2.27 3.36

1.00-2.40 1.60-3.65 1.49-3.46 2.24-5.05

1.31

1.18-1.45

1 0.95 1.34 1.90 2.00

0.63-1.43 0.92-1.95 1.34-2.72 1.40-2.87

Covariate-adjusted* HR 95%CI P value

Covariate-adjusted** HR 95%CI P value

<0.001

1.33

0.009

1.26

<0.001

1 1.11 1.71 1.46 1.80

0.006 0.69-1.79 1.10-2.66 0.92-2.32 1.14-2.83

1 1.08 1.64 1.34 1.63

0.67-1.74 1.06-2.55 0.85-2.14 1.03-2.56

<0.001

1.16

1.03-1.31

0.02

1.13

1.00-1.28

<0.001

1 0.91 1.22 1.50 1.46

0.008 0.59-1.41 0.80-1.84 1.00-2.25 0.96-2.23

1 0.91 1.16 1.46 1.34

0.59-1.42 0.76-1.76 0.97-2.18 0.88-2.05

1.08-1.65

1.02-1.56

0.04 0.03

0.049 0.02

Table 3. Cox analyses of the association between interleukin-6 (IL-6) and C-reactive protein (CRP) levels and sudden cardiac death risk in participants of the Cardiovascular Health Study without known cardiac disease at enrollment. HR: hazard ratio. CI: confidence interval. *,** adjusted for factors in Table 2 legend, as appropriate. Univariate

Covariate-adjusted*

Covariate-adjusted**

17

IL-6 +1 Log increase Quintiles Q1 (reference) Q2 Q3 Q4 Q5 CRP +1 Log increase Quintiles Q1 (reference) Q2 Q3 Q4 Q5

HR

95%CI

P value

HR

95%CI

P value

HR

95%CI

P value

1.62

1.19-2.21

0.003

1.46

1.04-2.05

0.03

1.42

1.01-2.00

0.049

0.002

1 1.30 2.05 1.79 2.23

0.02 0.62-2.70 1.01-4.14 0.85-3.75 1.06-4.69

1 1.26 1.90 1.60 2.16

0.61-2.63 0.94-3.84 0.76-3.37 1.03-4.54

0.05

1.18

0.96-1.45

0.1

1.16

0.94-1.43

0.04

1 0.54 1.12 0.98 1.47

0.09

1 0.57 1.09 0.98 1.40

1 1.57 2.46 2.25 2.81

0.77-3.21 1.24-4.88 1.11-4.57 1.38-5.69

1.21

1.00-1.47

1 0.56 1.16 1.10 1.56

0.28-1.13 0.65-2.07 0.60-2.02 0.87-2.81

0.26-1.09 0.62-2.01 0.52-1.85 0.79-2.75

0.03

0.2 0.1

0.28-1.16 0.60-1.96 0.52-1.85 0.75-2.61

Figure 1

Figure 2A

Figure 2B

Figure 2C

Figure 2D