Smoking is associated with schizophrenia, but not with mood disorders, within a population with low smoking rates: a matched case-control study in Bucaramanga, Colombia.

When comparing current smoking in schizophrenia patients versus the general population, the average odds ratio (OR) was 5.3 in a meta-analysis of 42 studies from 20 nations. Limited tobacco access can eliminate this strong association in some nations. Out of the 42 ORs, 37 were significantly higher than 1. Of the 5 non-significant ORs, three came from Colombian studies comparing current smoking prevalences in schizophrenia versus those in the general population (18%). However, the 3 ORs were not adjusted for confounders. We hypothesized that the association between schizophrenia and smoking is so strong that it can be detected in populations with limited access to smoking after carefully controlling for confounders. Of the three Colombian studies, one included 73 schizophrenia patients (DSM-IV schizophrenia or schizoaffective disorder) and 111 patients with mood disorders (DSM-IV bipolar or major depressive disorders). The current study matched each of these patients with 2 controls from the general population and performed more sophisticated statistical analyses. Prevalences of current smoking were 26% for schizophrenia patients and 10% for their matched controls (adjusted Mantel-Haenszel OR=3.1, 95% CI, 1.4-6.8), and 7% for patients with mood disorders and 12% for their matched controls (adjusted Mantel-Haenszel OR=0.62; CI, 0.28-1.4). The previously observed lack of association between schizophrenia and current smoking was due to lack of control of important confounding variables because of the absence of a control group. This re-analysis, which used a careful matching that controlled for confounders, suggests that the association between schizophrenia and smoking behaviors can stand in populations with low monetary income and low smoking rates. This study also suggests that the association between severe mood disorders (bipolar and major depression) and smoking is not as strong as that observed between schizophrenia and smoking, and may not be observable in countries where people have limited economic resources.