Objective: Studies have shown that increased co-morbidity is associated with poor pharmacological adherence. We undertook to determine the feasibility of using the modified RxRisk-V co-morbidity index to predict adherence to lipid lowering therapy (LLT). Methods: Using RxAmerica data, patients =18 years and with =18 months of continuous health plan enrollment from 2001-2005 were included in the analysis if they were 'new starts' with any class of LLT, defined as no prior treatment in the class for six months. Adherence ratios (defined as proportions of drug-available days during the follow-up period) were calculated and patients with adherence ratios =0.80 were considered adherent to LLT. Using a modified RxRisk-V, co-morbid conditions (CCs) were identified based on one-year of prescription claims prior to the index LLT prescription. Multivariable logistic regression was used to estimate the age- and sex-adjusted odds for adherence associated with various levels of disease co-morbidity. Results: A total of 19,458 patients were identified as new starts with an LLT class. The mean age of patients was 55 years (SD 12.1), 48% were females, and 43% had =3 CCs. Results of the regression analysis showed that patients with 1-2 CCs were less likely to be adherent (OR: 0.90; CI: 0.83-0.99) compared to patients with no CCs. Patients with =3 CCs were more likely to be adherent (OR: 1.10; CI: 1.01-1.18). The OR for adherence was significantly decreased for individuals with anxiety and tension, pain disorders, and tuberculosis. The OR was significantly increased for patients with cardiovascular diseases, psychiatric disorders, gastric acid disorders, and others. Conclusion: These results show that the relationship between adherence and degree of co-morbidity takes a U-shaped distribution; patients with lower levels of co-morbidity are less adherent compared to patients with no co-morbidity, and patients with higher levels of co-morbidity are more adherent.