Insulin and gall stones: a population case control study in southern Italy.

BACKGROUND Hyperinsulinaemia has been associated with many common diseases in developed countries, such as ischaemic heart disease and colon cancer. Gall stones are also very prevalent in these countries but little is known about the association between insulin and gall stones. AIMS To study the relationships between insulin and the incidence of gall stones in a sample of the general population. SUBJECTS AND METHODS Between May 1985 and June 1986, systematic sampling from the electoral register of Castellana, a small town in southern Italy, yielded 2472 subjects who had their gall bladder checked for gall stones by ultrasonography. Between May 1992 and June 1993, 1962 of the 2235 subjects without gall stones at the first examination agreed to a further ultrasound examination. A total of 101 subjects with newly diagnosed gall stones and 303 randomly chosen controls entered the study. Serum insulin was determined by radioimmunoassay, and concentrations of cholesterol, cholesterol high density lipoprotein (HDL), glucose, and triglycerides by standard enzymatic colorimetric methods. Unconditional multiple logistic regression was used to study the association between insulin and gall stones, controlling for the most common confounding factors. RESULTS In individuals with no clinical diagnosis of diabetes and serum glucose <7 mmol/l, insulin was associated with gall stones. This association persisted even after controlling for sex, age, body mass index, and serum glucose. The risk of gall stones in the highest quintile of serum insulin was 2.66 (95% confidence interval 1.04-6.72; chi(2) test for trend, p=0.03). The association of insulin with gall stones persisted when total and HDL cholesterol were entered in the logistic regression models, and only slightly decreased when serum triglycerides were included in the model. CONCLUSIONS The results of the study indicate that hyperinsulinaemia may play an important role in the aetiology of gall stones even in individuals without diabetes and with normal serum glucose levels.