Objectives: To evaluate the relationship between levels of serum insulin, the homeostasis model assessment (HOMA) and IGF-binding protein-1 (IGFBP-1) as factors related to myocardial infarction (MI) risk, and their interaction with lifestyle-related risk factors. Design: The Stockholm epidemiology programme (SHEEP), a case-control study, consisting of 749 first-time MI cases (510 men, 239 women) and 1101 healthy controls (705 men, 396 women) was used. Methods: The risk of developing MI was assessed by calculating odds ratios (OR) and synergistic interactions (SI) between serum insulin, IGFBP-1, HOMA and other variables related to MI risk (including smoking) in men and women. Results: Subjects with elevated levels of insulin and HOMA (. 75th percentile) had increased MI risks when compared with individuals with low levels. ORs for elevated insulin and HOMA (adjusted for age and residential area) for men: insulin 1.6 (95% confidence interval (CI) 1.3-2.1) and HOMA 1.5 (95% CI 1.1-1.9) and for women: insulin 2.1 (95% CI 1.5-2.9) and HOMA 1.9 (95% CI 1.3 -2.8). Women with low levels of IGFBP-1 (,10th percentile) showed a tendency towards elevated MI risk even if this was not statistically significant (OR 1.5 (95% CI 0.9-2.6)). Smokers with high levels of serum insulin had greatly increased MI risk (OR for men: 4.7 (95% CI 3.0-7.2) and OR for women: 8.1 (95% CI 4.5 -14.8)). SI scores based upon these interactions were statistically significant. Conclusions: These results might have preventive cardiovascular implications as they clearly suggest that subjects with insulin resistance are particularly susceptible to the hazards of smoking.

Objectives. Genes with a possible role for the development of the insulin resistance syndrome (IRS) were scanned for novel single-nucleotide polymorphisms (SNPs) using bioinformatics. Methods. GenBank mRNA sequences were compared to the human EST database using gapped BLAST BLAST, software that is available on the internet. Mismatches between the search and the EST sequences indicated potential SNPs. Thirty-two SNPs in 13 genes were randomly chosen for experimental veri®cation. PCR and direct sequencing were used to determine the`true' SNPs. A random sample of 30 Swedish men with slightly elevated diastolic blood pressure (85±94 mmHg) obtained from a population-based study was selected for the sequencing.After completion of these stages, the potential SNPs were checked against the large and rapidly expanding SNP databases HGBASE and NCBI. Results. EST searches of 146 genes revealed 106 potential SNPs in 44 genes. Experimental analysis of 32 of these potential SNPs veri®ed two SNPs; endothelin receptor A 1471 G/C (3¢ UTR) and PAI-1 Trp514Arg from a T/C exchange. These two SNPs were also identi®ed in the NCBI and HGBASE databases together with two polymorphisms that were not experimentally identi®ed in our homogenous Swedish population. Overall, the HGBASE and NCBI databases contained entries of 22% (23 out of 106) of the SNPs identi®ed through our EST searches. Conclusions. In the search for genetic variations causing complex diseases like IRS in homogenous populations (such as the Swedish one used here), important information can be obtained through bioinformatic searches of human genome databases and experimental veri®cation.