Insulin resistance and risk of incident cardiovascular events in adults without diabetes: meta-analysis

Abstract

Background: Glucose, insulin and Homeostasis Model Assessment Insulin Resistance (HOMA-IR) are markers of insulin resistance. The objective of this study is to compare fasting glucose, fasting insulin concentrations and HOMA-IR in strength of association with incident cardiovascular disease.

Methods: We searched the PubMed, MEDLINE, EMBASE, Web of Science, ScienceDirect and Cochrane Library databases from inception to March, 2011, and screened reference lists. Cohort studies or nested case-control studies that investigated the association between fasting glucose, fasting insulin or HOMA-IR and incident cardiovascular disease, were eligible. Two investigators independently performed the article selection, data extraction and risk of bias assessment. Cardiovascular endpoints were coronary heart disease (CHD), stroke or combined cardiovascular disease. We used fixed and random-effect meta-analyses to calculate the pooled relative risk for CHD, stroke and combined cardiovascular disease, comparing high to low concentrations of glucose, insulin or HOMA-IR. Study heterogeneity was calculated with the I(2) statistic. To enable a comparison between cardiovascular disease risks for glucose, insulin and HOMA-IR, we calculated pooled relative risks per increase of one standard deviation.

Results: We included 65 studies (involving 516,325 participants) in this meta-analysis. In a random-effect meta-analysis the pooled relative risk of CHD (95% CI; I(2)) comparing high to low concentrations was 1.52 (1.31, 1.76; 62.4%) for glucose, 1.12 (0.92, 1.37; 41.0%) for insulin and 1.64 (1.35, 2.00; 0%) for HOMA-IR. The pooled relative risk of CHD per one standard deviation increase was 1.21 (1.13, 1.30; 64.9%) for glucose, 1.04 (0.96, 1.12; 43.0%) for insulin and 1.46 (1.26, 1.69; 0.0%) for HOMA-IR.

Conclusions: The relative risk of cardiovascular disease was higher for an increase of one standard deviation in HOMA-IR compared to an increase of one standard deviation in fasting glucose or fasting insulin concentration. It may be useful to add HOMA-IR to a cardiovascular risk prediction model.

Figure 1. Summary of search results.

^aOne publication consisted of two studies. HOMA-IR, Homeostasis Model Assessment insulin resistance.

Figure 2. Random-effect meta-analyses of coronary heart disease risk for the highest category of glucose, insulin or HOMA-IR compared to the lowest category.

^aOr known diabetes was used to define the highest category. ^bParis Prospective Study. ^cHelsinki Policemen Study. ^dMen. ^eWomen. ^fGlomerular Filtration Rate ≥60 ml/min/1.73 m^{2. g}Glomerular Filtration Rate <60 ml/min/1.73 m^2. References are listed in References S1. 95% CI, 95% confidence interval; vs, versus; I-squared, measure of heterogeneity; HOMA-IR, Homeostasis Model Assessment Insulin Resistance.

Figure 3. Results of random-effect meta-analyses comparing cardiovascular disease risk for an increase of one standard deviation.

^a1 study did not specify sex-specific numbers. SD, standard deviation; 95% CI, 95% confidence interval; I², measure of heterogeneity; CHD, coronary heart disease and is defined as fatal or non-fatal myocardial infarction or angina pectoris; CVD, cardiovascular disease and is defined as myocardial infarction, angina pectoris, hemorrhagic stroke, ischemic stroke, arrhythmias, congestive heart failure or sudden cardiac death; HOMA-IR, Homeostasis Model Assessment Insulin Resistance.