Kaplan-Meier curves for mortality stratified by tertiles of adiponectin levels in men with no physician-diagnosed prevalent cardiovascular disease (CVD) or heart failure (A), men with CVD but no heart failure (B), and men with physician-diagnosed heart failure (C). Tertile 1 indicates less than 5.27 μg/mL; tertile 2, 5.27 to 9.40 μg/mL; and tertile 3, 9.41 μg/mL or greater.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Wannamethee SG, Whincup PH, Lennon L, Sattar N. Circulating Adiponectin Levels and Mortality in Elderly Men With and Without Cardiovascular Disease and Heart Failure. Arch Intern Med. 2007;167(14):1510–1517. doi:10.1001/archinte.167.14.1510
Copyright 2007 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2007
High adiponectin levels have been associated with reduced cardiovascular risk but have been shown to predict mortality in those at high risk for vascular disease. We examined the relationship between adiponectin levels and mortality in older men with and without diagnosed cardiovascular disease (CVD) and heart failure.
Prospective study of 4046 men aged 60 to 79 years drawn from general practices in 24 British towns and followed up for a mean of 6 years, during which 734 deaths occurred. The men were divided into the following groups according to the presence of physician-diagnosed CVD and heart failure: (1) those with no CVD or heart failure; (2) those with CVD but without heart failure; and (3) those with heart failure (with or without CVD).
After adjustment for a wide range of baseline characteristics, adiponectin levels were positively associated with significantly increased all-cause and CVD mortality in men with no diagnosed CVD or heart failure (top third vs bottom third adjusted relative risk, 1.55 [95% confidence interval (CI), 1.19-2.02; P = .002 for trend] vs 1.53 [95% CI, 1.03-2.27; P = .02 for trend]), as well as in men with diagnosed heart failure ([adjusted relative risk, 2.37 [95% CI, 0.64-8.79; P = .04 for trend] vs 3.43 [95% CI, 0.54-21.70; P = .008 for trend]). No association was seen in those with diagnosed CVD without heart failure. Adjustment for weight loss and renal function made minor differences to these relationships.
In older men, high adiponectin levels are associated with increased all-cause and CVD mortality in those with heart failure and those free of CVD. Such observations suggest that adiponectin levels may reflect a balance of both protective and harmful factors.
Levels of adiponectin, in contrast to other adipokines, are reduced in obese patients and in those with type 2 diabetes mellitus. Adiponectin has been proposed to have insulin-sensitizing, anti-inflammatory, and antiatherogenic properties.1-3 High levels of circulating adiponectin have been associated with reduced risks of cardiovascular disease (CVD) in some population studies,4-8 although some,9-11 including a recent nested case-control study from our group in the original British Regional Heart Study,11 have shown no association or a weak one. However, in recent prospective studies, high levels of plasma adiponectin were shown to be a predictor of mortality in patients with chronic heart failure,12,13 in patients presenting with coronary artery disease,14,15 in patients with chronic kidney disease,16 and most recently in older community-dwelling men and women.17 These conflicting findings raise the possibilities that adiponectin may have different prognostic implications in older subjects or in populations at high risk of vascular disease. Furthermore, despite the increase in visceral fat and insulin resistance with normal aging,18,19 adiponectin levels are known to increase with age.20 It has been recently suggested that the age-related decline of renal function results in a reduction in renal adiponectin clearance and may contribute to the increase in adiponectin levels in the elderly.20 However, little is known about the association between adiponectin levels and mortality in the general older population or its relation to declining renal function, which is itself associated with increased mortality in this group.21 We have therefore examined the relationship between adiponectin levels and all-cause mortality in elderly men (aged 60-79 years). We did so separately in those with and without diagnosed CVD (including coronary heart disease or stroke) and in those with diagnosed heart failure (including those with and without CVD). Our aim was to determine whether adiponectin levels were associated with increased CVD mortality in only those with existing vascular disease or heart failure or whether this extended to all individuals.
The British Regional Heart Study is a prospective study of CVD involving 7735 men aged 40 to 59 years selected from the age and sex registers of a general practice in each of 24 British towns, who were screened during the period between January 1, 1978, and June 30, 1980.22 From February 1, 1998, to March 31, 2000, all surviving men (n = 5547), now aged 60 to 79 years, were invited for a follow-up examination after a 20-year interval. Ethics approval was provided by all relevant local research ethics committees. All men provided informed written consent to the investigation, which was carried out in accordance with the Declaration of Helsinki. All men completed a questionnaire at reexamination, providing details of their medical history and lifestyle behaviors. The men were asked to fast for a minimum of 6 hours, during which they were instructed to drink only water, and to attend for measurements at a specified time between 8 AM and 6 PM. A fasting blood sample was collected using a commercially available system (Sarstedt Monovette; Sarstedt, Numbrecht, Germany), and samples were stored at −70°C for subsequent analyses; 4252 men (76.7% of survivors) attended for examination. The samples were stored at −20°C on the day of collection and transferred in batches for storage at −70°C until analysis, performed after no more than 1 freeze-thaw cycle. Twelve-lead electrocardiography was performed using a commercially available recorder (Sicard 460; Siemens, Erlangen, Germany).
Details of measurement and classification methods for smoking status; physical activity; body mass index (BMI) (calculated as weight in kilograms divided by height in meters squared); social class; alcohol intake; blood pressure; levels of blood lipids, blood glucose, and insulin; and measures of lung function (forced expiratory volume in 1 second) in this cohort have been described previously.22-26 The men were also asked whether they had lost weight in the 3 years prior to the 1998-2000 reexamination and whether the weight loss was intentional or unintentional. Blood pressure measured using an automated blood pressure monitor (Dinamap 1846; Critikon Services, Berkshire, England) was adjusted for observer variation.27 Blood glucose and insulin concentrations were adjusted for the effects of fasting duration and time of day.26 Insulin resistance was estimated according to the homeostasis model assessment (the product of the fasting glucose level [in milligrams per deciliter] and the insulin level [in micro–international units per milliliter] divided by the constant 405).28 (To convert fasting glucose level to millimoles per liter, multiply milligrams per deciliter by 0.0555; to convert insulin level to picomoles per liter, multiply micro–international units per milliliter by 6.945.) Prevalent diabetes included men with a diagnosis of diabetes or men with a fasting blood glucose level of at least 126 mg/dL. Levels of C-reactive protein were assayed by means of ultrasensitive nephelometry (Dade Behring, Milton Keynes, England). Definite and possible left ventricular hypertrophy was defined from the electrocardiogram in accordance with Minnesota coding criteria (Minnesota codes 3.1 and 3.3).29 Estimated glomerular filtration rate (eGFR), estimated from the serum creatinine level using the Modification of Diet in Renal Disease equation developed by Levey et al,30-32 was used as a measure of renal function.
Plasma adiponectin concentrations were determined using an enzyme-linked immunosorbent assay (R&D Systems Europe Ltd, Abingdon, England). The intra-assay and the interassay coefficients of variability were each 7.5%. We have previously shown this method to correlate well with a radioimmunoassay method for adiponectin measurement.11 Adiponectin concentrations were not available in 206 men. There is no evidence that the adipokine levels measured in the present study were influenced by prolonged storage or repeated free-thawing of samples.
The men were asked whether a physician had ever told them that they had angina or myocardial infarction (MI) (ie, heart attack or coronary thrombosis), heart failure, stroke, diabetes mellitus, or any of a number of other CVD conditions. Patient recall of a physician diagnosis of CVD has been shown to be a valid measure of recording diseases in this study population.33,34 The κ statistics comparing the medical record review with the patient's recall of coronary heart disease was 0.82.33 The men were also asked about regular treatment.
All men have been followed up from the initial 1978-1980 examination for cardiovascular morbidity, and follow-up has been achieved for 99% of the cohort.35 In the present analyses, all-cause mortality is based on follow-up from the 1998-2000 rescreening (when patients were aged 60 to 79 years) to December 31, 2005, a mean follow-up of 6 years (range, 5-7 years). Information on death was collected through the established tagging procedures provided by the National Health Service registers. A nonfatal MI was diagnosed according to World Health Organization criteria.36 Cardiovascular deaths include all those with International Classification of Diseases, Ninth Revision, codes 401 to 459. Evidence regarding nonfatal MI and heart failure was obtained by ongoing reports from general practitioners, by biennial reviews of the patients' medical records (including hospital and clinic correspondence) through the end of the study period, and from repeated personal questionnaires completed by surviving subjects after the initial examination.
The results relating adiponectin to coronary heart disease events recently published by our group11 concerned samples taken at the initial 1978-1980 screening, and therefore the analyses in the present article do not overlap with those of the previous study.
According to their recall of a physician diagnosis of CVD (MI, angina, or stroke) or heart failure at the 1998-2000 examination and regular surveillance of general practitioners' medical records of major nonfatal MI, stroke, or heart failure occurring before that point, the men were divided into the following 3 groups on the basis of their cardiovascular and heart failure status: (1) those with no prevalent physician-diagnosed CVD (MI, angina, or stroke) or heart failure (n = 3099); (2) those with prevalent physician-diagnosed CVD but no diagnosed heart failure (n = 830); and (3) those with prevalent physician-diagnosed heart failure (with or without CVD) (n = 117).
The men were divided into 3 groups on the basis of the tertile distribution of adiponectin levels in all men. Thus, similar cutoff points were used for the adiponectin groups in men without diagnosed CVD or heart failure, in those with CVD (no heart failure), and in those with heart failure. Kaplan-Meier curves were used to construct cumulative mortality curves across the 3 adiponectin groups. We used the Cox proportional hazards model to assess the multivariate-adjusted relative risk (RR) for the adiponectin groups. In the adjustment, factors known to be associated with mortality were included. In the multivariate analysis, smoking (never, long-term ex-smokers [>15 years], recent ex-smokers [<15 years], and current smokers), social class (manual vs nonmanual work), physical activity (none/occasional, light, moderate, and moderately vigorous/vigorous),24 alcohol intake (none/occasional, light, moderate, and heavy),24 BMI (<18.5, 18.5-24.9, 25.0-30.0, and ≥30.0), eGFR (<60, 60-69, and ≥70 mL/min/1.73 m2), diabetes mellitus (yes or no), treatment of hypertension (yes or no), use of statins (yes or no), use of β-blockers (yes or no), and weight loss during the preceding 3 years (yes or no) were fitted as categorical variables. Forced expiratory volume in 1 second, homeostasis model assessment product, and levels of C-reactive protein, albumin, and high-density lipoprotein cholesterol were fitted as continuous variables. Tests for trend were performed fitting adiponectin level in its original continuous form. In Table 1 and Table 2, tests for trends were performed across the groups.
During the mean follow-up period of 6 years in the 4046 men with adiponectin data, there were 734 deaths from all causes, of which 309 (42%) were due to CVD. Men with no CVD or heart failure and those with CVD only had lower adiponectin levels compared with men with heart failure (P < .001). The mean adiponectin levels in the 3 groups were 6.77 (interquartile range, 4.38-10.85), 6.63 (interquartile range, 4.10-11.38), and 9.23 (interquartile range, 6.04-15.17) μg/mL for groups 1, 2, and 3, respectively. The cumulative mortality in the 3 groups was 15.0%, 26.1%, and 44.4%, respectively.
The Figure shows the cumulative mortality according to the tertile cutoff distribution of adiponectin levels in the 3 groups. Elevated adiponectin levels were associated with the worst prognosis in men with no diagnosed CVD or heart failure and in men with heart failure (P < .001) and to a weaker extent in men with CVD without heart failure (P = .11).
Table 1 shows the baseline characteristics by the tertiles of adiponectin level in men with and without CVD, excluding those with heart failure. In both groups, elevated adiponectin levels were associated with older age, unintentional weight loss, lower BMI, lower forced expiratory volume in 1 second, and lower albumin level, but were also associated with lower rates of diabetes mellitus, lower use of β-blockers, higher high-density lipoprotein cholesterol levels, and lower insulin resistance (homeostasis model assessment) in both groups. Elevated adiponectin levels were associated with lower C-reactive protein levels in the men with no CVD or heart failure. In the small group of men with heart failure only (Table 2), similar findings were seen.
Table 3 shows the age-adjusted RRs and 95% confidence intervals (CIs) for all-cause mortality and CVD mortality by the tertile groups of adiponectin levels according to CVD and heart failure status. Adiponectin levels were significantly and positively associated with all-cause mortality and CVD mortality in those with no diagnosed CVD or heart failure as well as in those with heart failure. This association was strengthened after further adjustment for alcohol, smoking, BMI, social class, treated hypertension, left ventricular hypertrophy, use of β-blockers, use of statins, diabetes, lung function, levels of high-density lipoprotein cholesterol, albumin, and C-reactive protein, and homeostasis model assessment (Table 3). No significant association was seen between adiponectin levels and total or CVD mortality in those with CVD without heart failure.
Further adjustment for weight loss and low eGFR made minor differences to the relationships seen (Table 3). The increased risk of mortality associated with high adiponectin levels in those with no CVD was seen even after exclusion of men with low eGFR (<60 mL/min/1.73 m2) (adjusted RR, 1.50 [95% CI, 1.13-1.99]) after adjustment for factors in model 2 (Table 3) or exclusion of leaner men (BMI, <25.0) (adjusted RR, 1.89 [95% CI, 1.36-2.63]).
The positive association between adiponectin levels and mortality in those with no CVD or heart failure was seen in those who were never smokers (n = 959) (adjusted RR, 1.90 [95% CI, 0.99-3.62]) and in men who reported good or excellent health (adjusted RR, 1.42 [95% CI, 1.03-1.96]).
In men with no diagnosed CVD or heart failure, elevated adiponectin levels were also associated with increased risk of mortality from noncardiovascular causes (adjusted RR, 1.46 [95% CI, 1.05-2.04; P = .04 for trend]).
In this study of elderly men aged 60 to 79 years, mean adiponectin levels were significantly higher in the older groups and significantly raised in those with physician-diagnosed heart failure. We confirm the results of previous studies that high adiponectin levels are associated with significantly increased mortality in those with heart failure.12,13 Our findings extend the link between high adiponectin levels and mortality to the general older population of men without diagnosed CVD or heart failure. This increased mortality was seen for cardiovascular and noncardiovascular causes. In men with established CVD without heart failure, high adiponectin levels showed weak positive associations with total and CVD mortality, although the findings were not statistically significant.
The elevated mortality risk associated with high adiponectin levels has been observed in several other selected populations with coronary artery disease or renal dysfunction.14-17 It has been suggested that accumulation of adiponectin in those with diseases such as heart failure or chronic kidney disease, unlike the general population, may reflect the wasting and malnutrition that characterize these disease states and is thus a marker of poor prognosis.12,16 This is consistent with the high prevalence of unintentional weight loss and low albumin levels (proxy for malnutrition or cachectic state) observed in those with high adiponectin levels in the heart failure group. In the general population, however, some studies have shown high adiponectin levels to be cardioprotective with regard to CVD events,4,5 although this has not been consistent across population studies.9-11 We have shown that high adiponectin levels were associated with both total and CVD mortality in older men without diagnosed CVD or heart failure, which was not explained by traditional risk factors, weight loss, or renal function in multivariate analyses. The positive association was seen in never smokers, in those who reported good or excellent health, and even after exclusion of lean men (BMI, <25.0). Thus, the increased mortality does not just reflect ill health or low BMI, which has been linked to increased mortality in elderly individuals.37 These findings are in keeping with the most recent findings from the Rancho Bernardo Study, which reported high adiponectin levels to be associated with increased CVD and all-cause mortality in community-dwelling men and women.17
Adiponectin levels increase with age, and it is suggested that the decline in renal function with age may contribute to this increase.20 However, the increased risk in those with no CVD was seen even after exclusion of men with chronic kidney disease (eGFR, <60 mL/min per 1.73 m2).30
Several other possibilities could explain our findings in those with no diagnosed CVD or heart failure. Aging is associated with weight loss and loss in skeletal muscle mass and strength (sarcopenia),38,39 which are significant predictors of mortality in the elderly population.40,41 Thus, high adiponectin levels in elderly persons may be a consequence of weight loss and sarcopenia with aging. Alternatively, it may be directly responsible for increasing energy expenditure and induce weight loss through a direct effect on the brain,42 accelerating sarcopenia in elderly persons and thereby increasing mortality risk. Although the association between adiponectin and mortality remained after adjustment for weight loss, we did not have measures of skeletal muscle mass or of loss of muscle mass in general. An increased adiponectin level might be a marker for early cardiac dysfunction resulting from aging or asymptomatic CVD, increasing the risk of CVD mortality.
The strengths of this study include its representative sampling of men aged 60 to 79 years with a high follow-up rate, its good characterization of CVD status, and the availability of a wide range of risk factors and biological markers, including renal function. The eGFR estimated from serum creatinine level may not be as accurate a measure of renal function as direct measures of GFR. However, the Modification of Diet in Renal Disease Study Group equation to estimate GFR, which takes into account age, has been validated in large populations,31,32 and the eGFR significantly predicted CVD mortality in this cohort.43 The availability of data on intentional and unintentional weight loss in the 3 years prior to sampling was also a major advantage. Our study was performed in a predominantly white male population; we cannot generalize our findings to older women in whom mean adiponectin levels are about 50% higher44 or to other ethnic groups.
In conclusion, despite the favorable cardiovascular risk profile associated with high adiponectin levels, high adiponectin levels were associated with increased cardiovascular and all-cause mortality in this population of older men. The association was particularly strong in those with heart failure but, critically, was present also in those with no diagnosed heart failure or CVD. These data further stress the need to reevaluate the direction of the relationship between adiponectin concentrations and vascular morbidity and mortality in older men and suggest the need for further studies to disentangle relationships, which may help better determine any clinical role for adiponectin.
Correspondence: S. Goya Wannamethee, PhD, Department of Primary Care and Population Sciences, Royal Free and University College Medical School, Rowland Hill Street, London NW3 2PF, England (firstname.lastname@example.org).
Accepted for Publication: March 20, 2007.
Author Contributions: Dr Wannamethee had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Wannamethee and Sattar. Acquisition of data: Whincup, Lennon, and Sattar. Analysis and interpretation of data: Wannamethee. Drafting of the manuscript: Wannamethee and Lennon. Critical revision of the manuscript for important intellectual content: Whincup and Sattar. Statistical analysis: Wannamethee. Obtained funding: Wannamethee, Whincup, and Sattar. Administrative, technical, and material support: Lennon.
Financial Disclosure: None reported.
Funding/Support: This study was supported by the Department of Health (England) (British Regional Heart Study, a British Heart Foundation Research Group) and a British Heart Foundation Project grant for the adiponectin measurements and laboratory analyses reported herein (Drs Wannamethee, Whincup, and Sattar).
Disclaimer: The views expressed in this article are those of the authors and not necessarily those of the Department of Health of England.
Create a personal account or sign in to: