Date Published: October 21, 2008
Publisher: Public Library of Science
Author(s): Vlado Perkovic, Christine Verdon, Toshiharu Ninomiya, Federica Barzi, Alan Cass, Anushka Patel, Meg Jardine, Martin Gallagher, Fiona Turnbull, John Chalmers, Jonathan Craig, Rachel Huxley, Giuseppe Remuzzi
Abstract: BackgroundMarkers of kidney dysfunction such as proteinuria or albuminuria have been reported to be associated with coronary heart disease, but the consistency and strength of any such relationship has not been clearly defined. This lack of clarity has led to great uncertainty as to how proteinuria should be treated in the assessment and management of cardiovascular risk. We therefore undertook a systematic review of published cohort studies aiming to provide a reliable estimate of the strength of association between proteinuria and coronary heart disease.Methods and FindingsA meta-analysis of cohort studies was conducted to obtain a summary estimate of the association between measures of proteinuria and coronary risk. MEDLINE and EMBASE were searched for studies reporting an age- or multivariate-adjusted estimate and standard error of the association between proteinuria and coronary heart disease. Studies were excluded if the majority of the study population had known glomerular disease or were the recipients of renal transplants. Two independent researchers extracted the estimates of association between proteinuria (total urinary protein >300 mg/d), microalbuminuria (urinary albumin 30–300 mg/d), macroalbuminuria (urinary albumin >300 mg/d), and risk of coronary disease from individual studies. These estimates were combined using a random-effects model. Sensitivity analyses were conducted to examine possible sources of heterogeneity in effect size. A total of 26 cohort studies were identified involving 169,949 individuals and 7,117 coronary events (27% fatal). The presence of proteinuria was associated with an approximate 50% increase in coronary risk (risk ratio 1.47, 95% confidence interval [CI] 1.23–1.74) after adjustment for known risk factors. For albuminuria, there was evidence of a dose–response relationship: individuals with microalbuminuria were at 50% greater risk of coronary heart disease (risk ratio 1.47, 95% CI 1.30–1.66) than those without; in those with macroalbuminuria the risk was more than doubled (risk ratio 2.17, 1.87–2.52). Sensitivity analysis indicated no important differences in prespecified subgroups.ConclusionThese data confirm a strong and continuous association between proteinuria and subsequent risk of coronary heart disease, and suggest that proteinuria should be incorporated into the assessment of an individual’s cardiovascular risk.
Partial Text: Over recent decades, substantial progress has been made in understanding the role that “classical risk factors”—namely blood pressure, smoking, cholesterol, diabetes, and obesity—play in the aetiology of coronary heart disease (CHD) [1–6]. Subsequent studies and meta-analyses have identified other factors that may provide additional important predictive information regarding the risk of CHD, including inflammatory markers , haemostatic factors [8,9], left ventricular hypertrophy, and markers of kidney dysfunction .
This overview of the epidemiological evidence suggests that proteinuria is independently associated with increased risk of subsequent CHD. The results from this meta-analysis of 26 cohort studies, including information on over 7,000 CHD events among almost 170,000 individuals, suggest that people with proteinuria have a risk of CHD that is at least 50% greater than those without. Moreover, there was some evidence to indicate a dose–response relationship such that the strength of the association was substantially higher among individuals with macroalbuminuria compared with those with microalbuminuria. Furthermore, the relationship was consistent across diverse population subgroups including individuals with and without diabetes. The included studies were largely population-based suggesting that these findings are broadly generalisable and less likely to have been affected by interventions (e.g., blood pressure lowering) than data from clinical trials.