Acta Diabetologica

, Volume 52, Issue 1, pp 183–186

Association between HOMA-IR, fasting insulin and fasting glucose with coronary heart disease mortality in nondiabetic men: a 20-year observational study


    • Institute of Public Health, School of Public Health and Clinical NutritionUniversity of Eastern Finland
    • Department of Internal MedicineLapland Central Hospital
  • Francesco Zaccardi
    • Internal Medicine and Diabetes Care Unit, Policlinico Gemelli HospitalCatholic University of the Sacred Heart
  • Vivian N. Onaemo
    • Institute of Public Health, School of Public Health and Clinical NutritionUniversity of Eastern Finland
  • Sae Young Jae
    • Department of Sports Informatics, College of Arts and Physical EducationUniversity of Seoul
  • Jussi Kauhanen
    • Institute of Public Health, School of Public Health and Clinical NutritionUniversity of Eastern Finland
  • Kimmo Ronkainen
    • Institute of Public Health, School of Public Health and Clinical NutritionUniversity of Eastern Finland
  • Jari A. Laukkanen
    • Institute of Public Health, School of Public Health and Clinical NutritionUniversity of Eastern Finland
    • Department of Internal MedicineLapland Central Hospital
Short Communication

DOI: 10.1007/s00592-014-0615-x

Cite this article as:
Kurl, S., Zaccardi, F., Onaemo, V.N. et al. Acta Diabetol (2015) 52: 183. doi:10.1007/s00592-014-0615-x


Whether glucose and insulin are differently associated with the risk of coronary heart disease (CHD) mortality is unclear. We aimed to estimate the association between insulin resistance (estimated by the homeostasis model assessment for insulin resistance, HOMA-IR), fasting serum insulin (FI) and fasting plasma glucose (FPG) with incident CHD mortality in a prospective study including middle-aged nondiabetic Finnish men. During an average follow-up of 20 years, 273 (11 %) CHD deaths occurred. In a multivariable Cox regression analysis adjusted for age, body mass index, systolic blood pressure, serum LDL-cholesterol, cigarette smoking, history of CHD, alcohol consumption, blood leukocytes and plasma fibrinogen, the hazard ratios (HRs) for CHD mortality comparing top versus bottom quartiles were as follows: 1.69 (95 % CI: 1.15–2.48; p = 0.008) for HOMA-IR; 1.59 (1.09–2.32; p = 0.016) for FI; and 1.26 (0.90–1.76; p = 0.173) for FPG. These findings suggest that IR and FI, but not FPG, are independent risk factors for CHD mortality. Further studies could help clarify these results in terms of screening and risk stratification, causality of the associations, and therapeutical implications.


Insulin resistanceCoronary heart diseaseMortalityMenRisk factors


Type 2 diabetes mellitus (T2DM) and prediabetic states are characterized by an abnormal glucose/insulin homeostasis and an increased risk of cardiovascular disease (CVD) [1].

While the association between fasting plasma glucose (FPG) and CVD has been extensively investigated [2], the role of insulin resistance (IR) as a cardiovascular risk factor is still unclear. IR has been linked to CV and all-cause mortality [3], yet the relationship with coronary heart disease (CHD) events is less clear. In particular, it is not well defined the specific association with CHD mortality, as previously studies tended to combine fatal and nonfatal CHD events (with some including also diabetic subjects at baseline); moreover, whether this association would differ from that of FPG and fasting insulin (FI) has not been investigated.

The homeostasis model assessment for insulin resistance (HOMA-IR) is a well-known method to estimate insulin resistance from FPG and FI. It is considered particularly appropriate for large population-based studies and has been extensively used in previous epidemiological research.

The present study was conducted to investigate if HOMA-IR estimated IR, FPG and FI were associated with the specific risk of CHD mortality among nondiabetic Finnish men over an extended period of follow-up.


Study population

The study sample consisted of participants of an ongoing prospective population-based cohort study designed to investigate the risk factors for cardiovascular diseases and other related outcomes in Eastern Finland [Kuopio Ischaemic Heart Disease Risk Factor Study (KIHD)]. Ethical approval was obtained from the Research Ethics Committee of the University of Kuopio, and the participants gave their written informed consent.

Of 3,235 eligible men, 2,682 (82.9 %) were recruited. The present analysis excluded men with T2DM at baseline, defined as subject with FPG level ≥7.0 mmol/L or a clinical diagnosis of diabetes with dietary, oral antidiabetic or insulin treatment. Complete data on FPG and FI were available in 2,429 subjects.

Laboratory methods and outcome assessments are reported in the Supplementary Material.

Statistical analysis

The associations between HOMA-IR, FPG and FI with the risk of CHD mortality were estimated using a Cox proportional hazards model, simultaneously adjusted for previously established CHD risk factors, such as age, body mass index (BMI), systolic blood pressure (SBP), serum LDL-cholesterol, smoking (Yes/No), history of CHD (Yes/No), alcohol consumption (continuous), blood leukocytes and plasma fibrinogen, and with HOMA-IR, FPG and FI entered as dummy categorical variables (quartiles) for each analysis. The proportional hazards assumption was examined using log plots of survival versus time. Statistical significance was defined as p values <0.05, and all statistical tests were two-sided. Analyses were performed using the SPSS 20.0 IBM software.


Baseline characteristics

At baseline, FI concentration was 11.25 ± 6.75 mU/L (mean ± SD), FPG 4.60 ± 0.50 mmol/L and HOMA-IR 2.59 ± 1.60. Other baseline characteristics of the study population are shown in Supplementary Table 1.

HOMA-IR, fasting insulin, fasting blood glucose and CHD mortality

Among 2,429 men, 273 (11 %) CHD mortality events occurred during the follow-up period (mean, 20.0 years).

As reported in Table 1, the HR of CHD mortality comparing subjects in the top versus bottom (reference) quartile of HOMA-IR was 1.69 (95 % confidence interval (CI) 1.15–2.48; p = 0.008). A similar association was found for FI (HR 1.59, 95 % CI 1.09–2.32; p = 0.016), but not for FPG (HR 1.26, 95 % CI 0.90–1.76; p = 0.173) (Table 2).
Table 1

Multivariate hazard ratios for the variables included in the model with HOMA-IR quartiles


HR (95 % CI)

p value

HOMA-IR (top vs bottom quartile)

1.69 (1.15–2.48)


Age (years)

1.09 (1.06–1.13)


Alcohol/week (g/week)

1.00 (0.99–1.00)


Body mass index (kg/m2)

0.99 (0.95–1.03)


Smoking (Yes)

1.71 (1.29–2.26)


Prevalent coronary heart disease (Yes)

2.36 (1.84–3.01)


Systolic blood pressure (mmHg)

1.01 (1.00–1.02)


Serum LDL-cholesterol (mmol/L)

1.17 (1.04–1.30)


Plasma fibrinogen (g/L)

1.52 (1.23–1.88)


Blood leukocyte count (×109/L)

1.11 (1.03–1.20)


HOMA-IR homeostasis model assessment for insulin resistance, LDL low-density lipoprotein, HR hazard ratio, 95 % CI 95 % confidence interval

Table 2

Multivariate hazard ratios of CHD mortality according to quartiles of HOMA-IR, fasting plasma glucose and fasting serum insulin


HR (95 %CI)

p value

Quartiles of HOMA-IR

 Q1 (0.58–1.64)



 Q2 (1.64–2.19)

0.92 (0.62–1.35)


 Q3 (2.19–3.07)

0.92 (0.72–1.34)


 Q4 (3.07–15.25)

1.69 (1.15–2.48)


Quartiles of fasting plasma glucose (mmol/L)


 Q1 (3.2–4.2)



 Q2 (4.2–4.5)

0.88 (0.62–1.26)


 Q3 (4.5–4.8)

0.76 (0.52–1.11)


 Q4 (4.8–6.2)

1.26 (0.90–1.76)


Quartiles of fasting serum insulin (mU/L)


 Q1 (2.9–7.3)



 Q2 (7.3–9.7)

0.84 (0.57–1.25)


 Q3 (9.7–13.1)

1.04 (0.72–1.50)


 Q4 (13.1–55.7)

1.59 (1.09–2.32)


Multivariate model adjusted for age, prevalent coronary heart disease, cigarette smoking, body mass index, systolic blood pressure, serum LDL-cholesterol, plasma fibrinogen, blood leukocytes and alcohol consumption

HOMA-IR homeostasis model assessment for insulin resistance, HR hazard ratio, 95 % CI 95 % confidence interval, Q quartile

Given the possibility of classification bias in assessing cases of death from death certificates, in a sensitivity analysis, we estimated the association between insulin resistance and all-cause mortality: comparing top versus bottom quartile of HOMA-IR, the HR for all-cause mortality resulted 1.26 (95 % CI 1.04–1.52; p = 0.016). Moreover, comparing quartiles of HOMA-IR after adjustment for incident diabetes as time-varying covariate, the HR for CHD mortality was 1.89 (95 % CI 1.32–2.69; p < 0.001). Lastly, in a similar comparison, HRs for CHD mortality resulted 1.61 (95 % CI 1.14–2.28; p = 0.007) for the subgroup of subjects with baseline CHD and 1.21 (95 % CI 0.76–1.91; p = 0.409) for those without prevalent CHD.


HOMA-IR has been shown to be associated with all-cause mortality as well as with CHD events among nondiabetic subjects [3]. However, the association between IR and the specific risk of fatal CHD events in nondiabetic subjects is less clear. We therefore assessed whether HOMA-IR was associated with CHD mortality in a long follow-up prospective studies including nondiabetic Finnish men. Our results suggest HOMA-IR as an independent risk factor for CHD mortality after adjustment for other potential risk factors and mediators. A similar finding was also evident for FI, but not FPG. A previous individual participant data meta-analysis [2] has indicated that FPG is nonlinearly related to the incidence of combined fatal and nonfatal CHD events, with a risk starting only after 5.6 mmol/L: as 96.2 % of total subjects in our cohort had baseline level of FPG lower than such threshold, this could explain the lack of association between FPG and CHD mortality and would indicate the presence of a nonlinearity also specifically for fatal CHD events.

The pathogenic mechanism of IR as a cardiovascular risk factor is considered to be through the direct atherogenic action of insulin on vessel wall cells and/or indirect through obesity, blood pressure, lipids and metabolic homeostasis [4]. IR has been indeed associated with higher levels of triglycerides, lower levels of HDL-cholesterol, SBP and T2DM [4], all established CHD risk factors. However, our results evidenced an independent effect from other risk factors and potential mediators.

Our findings might have several implications: firstly, they would suggest the use of HOMA-IR (or FI) to screen and stratify the risk of CHD in subjects with FPG value under the (to date) established risk threshold of 5.6 mmol/L. Secondly, they can support the hypothesis that the benefit of glucose reduction in diabetic subjects could be counterbalanced by the atherogenic effect of hyperinsulinaemia in patients treated with intensive glucose control [5]. Thirdly, the results could also advocate a different approach to reduce cardiovascular complications in phenotypically diverse type 2 diabetic subjects, when the primary pathophysiological abnormality is IR/hyperinsulinaemia or reduced insulin secretion.

Some limits of the study should be acknowledged. Our findings are based only on middle-aged nondiabetic Finnish men, and caution is needed in interpretation and inference of the results on women and younger or elderly subjects. Moreover, the analyses related only to baseline values of FPG and FI: as they physiologically fluctuate over time within subjects, this can result in a biased association with the outcomes (regression dilution bias). Finally, these results using HOMA-IR need to be interpreted carefully, given the inherent issues relating to reproducibility (both within subject and between subject).

In conclusion, in this prospective population-based study, HOMA-IR and FI, but not FPG, were associated with long-term CHD mortality among nondiabetic Finnish men. Further studies are warranted to extend our findings in other populations and to define their significance in terms of screening and risk stratification, causality of the associations and therapeutical implications.

Conflict of interest

Sudhir Kurl, Francesco Zaccardi, Vivian N. Onaemo, Sae Young Jae, Jussi Kauhanen, Kimmo Ronkainen and Jari A. Laukkanen declare that they have no conflict of interest.

Human and Animal Rights

All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1975, as revised in 2008.

Informed consent

Informed consent was obtained from all patients for being included in the study.

Supplementary material

592_2014_615_MOESM1_ESM.docx (20 kb)
Supplementary material 1 (DOCX 19 kb)

Copyright information

© Springer-Verlag Italia 2014