Non-linear structural equation models with correlated continuous and discrete data.

Structural equation models (SEMs) have been widely applied to examine interrelationships among latent and observed variables in social and psychological research. Motivated by the fact that correlated discrete variables are frequently encountered in practical applications, a non-linear SEM that accommodates covariates, and mixed continuous, ordered, and unordered categorical variables is proposed. Maximum likelihood methods for estimation and model comparison are discussed. One real-life data set about cardiovascular disease is used to illustrate the methodologies.

Association between smoking, pancreatic insulin secretion and insulin resistance in Chinese subjects with or without glucose intolerance.

BACKGROUND: There are studies suggesting smoking may increase the risk of type 2 diabetes. Effects of smoking on insulin secretion and insulin resistance (IR) are, however, controversial. METHODS: This is a cross-sectional study. Since there were very few smokers among Hong Kong Chinese women, only men (n = 1068) were analyzed in this report. Fasting and 2-hour plasma glucose and insulin were measured. Insulinogenic index as well as beta-cell function and IR based on homeostatic model assessment (HOMA) by computer model (HOMA Calculator v2.2) were calculated. RESULTS: Of the 1068 men, 147 had newly diagnosed diabetes, 131 newly diagnosed impaired glucose tolerance (IGT) and 790 were non-diabetic normal controls. Smokers had similar fasting and 2-hour insulin levels, insulinogenic index and HOMA derived beta-cell function as compared to non-smokers in the groups with diabetes, IGT or normal oral glucose tolerance test (OGTT). IR was also similar between smokers, ex-smokers and non-smokers in those with normal OGTT. In men with IGT or diabetes, after adjustment for age and body mass index, smokers were more insulin resistant as compared to non-smokers (IR, IGT: 1.59 +/- 1.07 vs 1.03 +/- 0.54, P < 0.05; diabetes: 1.96 +/- 1.36 vs 1.06 +/- 0.45, P < 0.01). With Logistic regression analysis, comparing smokers and non-smokers, IR was independently associated with smoking (odds ratio (95% CI), IGT: 2.23 (1.05, 4.71); diabetes: 3.92 (1.22, 12.58)). None of the other insulin parameters enter into the model among those with normal OGTT or comparing ex-smokers and non-smoker or smokers and ex-smokers. CONCLUSIONS: In Chinese men, smoking did not show any direct association with insulin levels and pancreatic insulin secretion. Smoking men with IGT or diabetes appeared more insulin resistant than their non-smoking counterparts.