Antidepressant use in US adults increased 3-fold from 2.5% in 1988-94 to 8.1% in 1999-2002, based on National Health and Nutrition Examination Surveys. As the use of antidepressants increases, a comprehensive understanding of the potential health risks that may be associated with their use becomes increasingly important. This study evaluated the effects of paroxetine and sertraline on low-density lipoprotein cholesterol (LDL-C). An observational cohort study (1997-2004) of adults who had taken paroxetine or sertraline for at least 60 continuous days and had > or =2 LDL-C values measured during the study period, one while taking and one while not taking paroxetine or sertraline. A total of 13 634 LDL-C values clustered within 2682 patients were studied. We conducted mixed model regression analyses to quantify the relationship between antidepressant use and LDL-C values. The number of days taking paroxetine (beta = 0.0045; 95% CI 0.0018, 0.0073) and sertraline (beta = 0.0074; 95% CI 0.0054, 0.0093) prior to the LDL-C test were related to higher LDL-C values, after accounting for age, sex, year LDL-C was tested, co-morbidity, depression and lipid medication. The number of days that had passed since exposure to paroxetine (beta = -0.0013; 95% CI -0.0020, -0.00061) or sertraline (beta = -0.00093; 95% CI -0.016, -0.00022) were related to lower LDL-C values. The significant interaction between exposure to an antidepressant and taking a lipid medication demonstrates that the increase in LDL-C values associated with antidepressant use is ameliorated among patients who were taking a lipid medication when LDL-C was measured. Our study showed that long-term use of paroxetine or sertraline may have a measurable adverse impact on cardiovascular risk in adults. Clinical strategies should be used to address cardiovascular risk while maintaining effective treatment of major depression. In light of these findings, attention to LDL-C values should accompany antidepressant use.