C-Reactive Protein, Cardiovascular Risk Factors, and Mortality in a Prospective Study in the Elderly
Abstract—Serum C-reactive protein (CRP) reflects inflammation and predicts cardiovascular disease in middle-aged individuals. We investigated CRP, risk factors, and 10-year mortality in 3 elderly cohorts (aged 75, 80, and 85 years; n=455) of the population-based Helsinki Ageing Study. Clinical and laboratory examinations were performed at baseline, and in 1998, CRP was measured by a sensitive method (sensitivity 0.3 mg/L) from frozen serum samples. Mortality data were retrieved from national registers. Serum CRP ranged from 0.18 to 170.0 mg/L (interquartile range 0.68 to 4.10 mg/L, median 1.60 mg/L). CRP correlated significantly with body mass index and plasma insulin and was associated with smoking at baseline. An inverse correlation was found with albumin and total and HDL cholesterol. CRP was not associated with diabetes or cardiovascular disease but was significantly (P=0.015) higher in persons with (n=70) than without (n=385) dementia. During the 10-year follow-up, 61% (n=278) of the cohort died; half of the deaths were due to cardiovascular diseases. Mean CRP in survivors and nonsurvivors was 3.16 and 5.22 mg/L (P=0.017), respectively. After controlling for age and sex, baseline CRP (per 10 mg/L) significantly predicted the 10-year total mortality (risk ratio 1.20, 95% CI 1.08 to 1.32) and cardiovascular mortality (risk ratio 1.22, 95% CI 1.10 to 1.35). Predictive value was found in the 75-year-old cohort, but it was clearly attenuated in the 80- and 85-year-old cohorts. The results indicate that CRP is associated with several cardiovascular risk factors in the elderly. CRP alone predicts overall and cardiovascular mortality, but the prediction was significant in only the 75-year-old cohort.
Reprint requests to Timo Strandberg, MD, Department of Medicine, University of Helsinki, PO Box 340, FIN-00029 HYKS, Finland.
Presented in part in abstract form at the 39th Annual Conference on Cardiovascular Disease Epidemiology and Prevention, American Heart Association, Orlando, Fla, March 24–27, 1999.
- Received October 11, 1999.
- Accepted December 7, 1999.
Cardiovascular diseases (CVDs) are the main causes of mortality in elderly populations. The basic process of most CVD, atherosclerosis, is now considered to be partly an inflammatory disorder,1 and this may be 1 of several possible explanations for the findings that even slightly elevated concentrations of an acute-phase reactant, C-reactive protein (CRP), have predicted coronary events in middle-aged men and women in several studies.2 3 4 5 6 CRP has also been associated with CVD events7 and mortality8 in healthy individuals aged ≥65 years. However, there are scant data about the epidemiology and predictive value of CRP, specifically in the elderly (aged >75 years), at the population level. Therefore, we measured baseline CRP in population-based cohorts aged 75, 80, and 85 years and related it to total and CVD mortality during a 10-year follow-up.
The design of the Helsinki Ageing Study, a prospective birth cohort study in the City of Helsinki, Finland, has been described in detail.9 In 1989, 651 random persons of 3 birth cohorts, born in 1904, 1909, and 1914, were examined. Baseline frozen serum samples were available from 455 unselected persons in 1998, and these samples were tested for CRP.
The baseline evaluation included postal questionnaires for the subject and for a close informant, a structured interview conducted by a community nurse, an examination by a general practitioner, and laboratory examinations. Hypertension was defined as a past diagnosis with medication prescribed or current blood pressure >160/95 mm Hg. Diabetes was defined as a past diagnosis or current use of antidiabetic medication. Angina pectoris was diagnosed according to the criteria of Rose.10 The presence of other CVDs was based either on data from hospital records or on clinical examinations, including echocardiography.9 CVDs were also graded by the examining physician according to the New York Heart Association (NYHA) classification. Dementia at baseline was diagnosed as described previously.11 The individuals were classified as healthy elderly if their subjective or objective (according to the physician) health was good or moderate, if they did not suffer from any significant clinical disease, and if they had a normal reported exercise tolerance.
Blood samples for routine laboratory analyses (including blood lipids and albumin) were drawn after an overnight fast. ApoE phenotype was determined as described previously.11 Dehydroepiandrosterone (DHEA) was measured with a commercial kit (Coat-A-Count DHEA-soy, Diagnostics Products).12 CRP was measured from baseline frozen (−20°C) serum samples in 1998 with a sensitive immunoenzymometric assay that made use of 2 monoclonal antibodies (sensitivity 0.3 mg/L, Medix Diacor). The standard of this assay was the World Health Organization 1st International Reference Standard for CRP Immunoassay 85/506.12A The between-run coefficient of variation was 9.4% for level 1.4 mg/L and 9.8% for level 14.4 mg/L. The range for this assay was 0.3 to 30 mg/L.
The census status (100%) was determined annually until December 1998. During this time, 278 persons with known CRP died. The death certificates were evaluated for cause of death according to the 9th revision of the International Classification of Diseases (ICD-9) by one of the authors (R.S.T.) before determination of CRP.
Data were analyzed with Biomedical Data Processing (BMDP) software.13 The differences in laboratory variables were tested by ANOVA. For these calculations, logarithmic transformations of CRP were used. Stepwise logistic regression analysis was used for testing the risk of death, and the life-table method and the Cox proportional hazards model were used for survival analyses. Age and sex were forced into the analyses as covariates.
Clinical characteristics of the population are shown in Table 1⇓. Because it was a population-based study, only part of the individuals were designated as “healthy” by our criteria. The range of serum CRP was 0.18 to 170.0 mg/L (median value 1.60 mg/L, interquartile range 0.68 to 4.10 mg/L). Forty individuals (8.6%) had CRP values >10 mg/L, which is the cut point generally considered to signify clinically significant inflammation, and only 4 (0.4%) had CRP values >50 mg/L. CRP was not associated with the age group, but the logarithmically transformed value had a significantly positive correlation with plasma insulin (r=0.135, P=0.007) and a correlation of borderline significance with body mass index (r=0.092, P=0.050). Inverse associations were found with serum albumin (r=−0.302, P<0.001), total cholesterol (r=−0.146, P=0.002), and HDL cholesterol (r=−0.271, P<0.001). Unadjusted associations of CRP levels with the covariables, computed as group means (±SE) of categorized variables, are shown in Table 2⇓. The P value is the significance level obtained for testing the simultaneous equality of all group means (logarithmic value) in a 1-way ANOVA. CRP was not associated with diabetes or history of myocardial infarction or heart failure but increased significantly with increasing NYHA class and was significantly (P=0.015) higher in persons with (n=70) than without (n=385) prevalent dementia at baseline. In addition, no significant associations were found between CRP and serum DHEA or apoE phenotype (data not shown).
During the 10-year follow-up, 278 persons (61%) died; 53% of the deaths were attributed to CVDs. Mean CRP levels in survivors and nonsurvivors were 3.16±0.43 and 5.22±0.78 mg/L, respectively (P=0.017). The corresponding CRP value of those who died of cardiovascular causes was 5.15±1.34 mg/L (P=0.076 versus survivors). The excess mortality of the elderly with slightly elevated CRP (>5 mg/L, which corresponds to the middle of the clinically “nonsignificant ” range) was already evident during the first year of follow-up but did not increase thereafter (Figure⇓). Although the curves differed in all birth cohorts, the prognostic value became weaker with increasing age (Table 3⇓) and was significant in only the 75-year-old cohort (P=0.045). When adjusted for age and sex, baseline CRP (per 10 mg/L) significantly predicted 10-year total mortality (risk ratio 1.20, 95% CI 1.08 to 1.32) and cardiovascular mortality (risk ratio 1.22, 95% CI 1.10 to 1.35).
To our knowledge, this is the first population-based study that reports cross-sectional and prospective data on CRP specifically in persons aged ≥75 years. Even in old age, CRP is linked to several cardiovascular risk factors, including smoking and low HDL cholesterol. CRP is also associated with total and cardiovascular mortality during follow-up. However, the prognostic value is more pronounced in the short term and in 75-year-old than in older individuals.
The potential limitation of the present study is that the results are based on analyses from 1 measurement of CRP in serum samples collected 10 years earlier. At the individual level, a casual CRP measurement may not represent “true” basal CRP, because intervening infections easily increase CRP. Thus, the impact of CRP on mortality might be explained simply by infections. However, in our population, 91% had CRP levels under the “clinically significant” level of 10 mg/L. Moreover, the mortality curves were similar for cardiovascular mortality, conventionally not considered an infectious disease.
Although some high values increased the mean, the distribution of CRP in this population was otherwise remarkably similar to that in the middle-aged men population,14 with the interquartile ranges being almost identical (0.68 to 4.10 mg/L versus 0.69 to 3.95 mg/L). The geometric mean of CRP (1.60 mg/L) was also very similar to that of middle-aged men (1.72 mg/L)14 and of healthy elderly subjects (1.57 mg/L) in another study of the elderly.8 Characteristic associations of CRP with some, but not all, risk factors and disease states that have been observed in younger persons were also found in our elderly cohort. To facilitate comparisons in Table 2⇑, we partly used categorization of variables presented earlier in a middle-aged male population.6 Associations of CRP with BMI, lipids, and smoking are similar to those associations in middle-aged men6 14 and women,15 whereas chronic conditions, such as diabetes, hypertension, or CVD, were not associated with CRP in our aged population. This may reflect the overriding effect of these conditions on survival in old age, but then the fact that NYHA class was associated with CRP is somewhat surprising. The total impact of disability, physical condition with its sequelae (BMI, lipids, and insulin), and smoking on CRP could be an explanation, because functional disability has earlier been shown to be associated with interleukin-6 levels in plasma.16 An interesting association was found between CRP and prevalent dementia (54% being either probable or uncertain Alzheimer-type dementia).11 Although there are reports suggesting “acute-phase response” in the pathology of Alzheimer’s disease,17 18 the most likely explanation for higher CRP in patients with dementia is their proneness to infections and minor trauma leading to systemic acute-phase response.
The present results have several implications. Although the results show that CRP predicts mortality in the elderly, the age cohort design of the present study demonstrates that the prediction is gradually diluted after 75 years. However, the results suggest that in the basal state, CRP is remarkably stable with age in humans. The similarity of the present results (with only 1 CRP measurement) to other studies also reflects technical reliability of current sensitive assays. It was somewhat unexpected that the impact of age on CRP was minor in our elderly cohorts, because interleukin-6, which stimulates CRP production in the liver, has been reported to increase with age.19 Because CRP has prognostic value, it can be used as an indicator of cardiovascular risk,20 keeping in mind that this effect seems to vanish in individuals aged 75 to 80 years. Finally, there is emerging evidence that part of the clinical effect of aspirin3 and statins21 22 23 may be due to anti-inflammatory properties. These actions may also be relevant in elderly patients.
This study was supported by Ragnar Ekberg’s Foundation. The authors thank Liisa Melamies, Clinical Biochemist, at Medix Diacor for CRP determinations.
Kuller LH, Tracy RP, Shaten J, Meilahn EN. Relation of C-reactive protein and coronary heart disease in the MRFIT nested case-control study. Am J Epidemiol. 1996;144:537–547.
Ridker PM, Buring JE, Shih J, Matias M, Hennekens CH. Prospective study of C-reactive protein and the risk of future cardiovascular events among apparently healthy women. Circulation. 1998;98:731–733.
Lowe GDO, Rumely A, Sweetnam PM, Yarnell JWG. D-dimer, C-reactive protein and risk of ischemic heart disease: the Speedwell study. Blood Coagul Fibrinolysis. 1999;10(suppl 1):S92–S93.
Koenig W, Sund M, Fröhlich M, Fischer H-G, Löwel H, Döring A, Hutchinson WL, Pepys MB. C-reactive protein, sensitive marker of inflammation, predicts future risk of coronary heart disease in initially healthy middle-aged men. Circulation. 1999;99:237–242.
Tracy RP, Lemaitre RN, Psaty BM, Ives DG, Evans RW, Cushman M, Meilahn EN, Kuller LH. Relationship of C-reactive protein to risk of cardiovascular disease in the elderly. Arterioscler Thromb Vasc Biol. 1997;17:1121–1127.
Rose GA. Chest pain questionnaire. Milbank Q. 1965;43:32–39.
National Biological Standard Board. First international standard for human C-reactive protein established 1986, code 85/506. National Institute for Biological Standards and Controls.
Dixon WJ, ed. BMDP: Statistical Software. London, UK: University of California Press; 1988.
Mendall MA, Patel P, Ballam L, Strachan D, Northfield TC. C reactive protein and its relation to cardiovascular risk factors: a population based cross sectional study. BMJ. 1996;312:1061–1065.
Hak AE, Stehouwer CDA, Bots ML, Polderman KH, Schalkwijk CG, Westendorp ICD, Hofman A, Witteman JCM. Associations of C-reactive protein with measures of obesity, insulin resistance, and subclinical atherosclerosis in healthy, middle-aged women. Arterioscler Thromb Vasc Biol. 1999;19:1986–1991.
Cohen Pieper, Harris TB, Rao KMK, Currie MS. The association of plasma IL-6 levels with functional disability in community-dwelling elderly. J Gerontol Med Sci. 1997;52A:M201–M208.
Wood JA, Wood PL, Ryan R, Graff-Radford NR, Pilapil C, Robitaille Y, Quirion R. Cytokine indices in Alzheimer’s temporal cortex: no changes in mature IL-1β or IL-1RA but increases in the associated acute phase proteins IL-6, α2-macroglobulin and C-reactive protein. Brain Res. 1993;629:245–252.
Ridker PM, Glynn RJ, Hennekens CH. C-reactive protein adds to the predictive value of total and HDL cholesterol in determining risk of first myocardial infarction. Circulation.. 1998;97:2007–2011.
Ridker PM, Rifai N, Pfeffer MA, Sacks FM, Moye LA, Goldman S, Flaker GC, Braunwald E, for the CARE investigators. Inflammation, pravastatin and the risk of coronary events after myocardial infarction in patients with average cholesterol levels. Circulation. 1998;98:839–844.
Ridker PM, Rifai N, Pfeffer MA, Sacks F, Braunwald E. Long-term effects of pravastatin on plasma concentration of C-reactive protein. Circulation. 1999;100:230–235.