Abstract
Background. Statins exert both anti-inflammatory and immunomodulatory effects. There is evidence that statins trigger tumor-specific apoptosis in patients with multiple myeloma (MM). The etiologic role of statins in the risk of MM, however, is not well understood. Methodological challenges of observational studies on statin use include use of self-reported data on drug exposure, cross-sectional exposure assessment, and short duration of follow-up with limited sample size. To rigorously investigate the long-term effects of statins on the incidence of MM, we conducted a pharmacoepidemiologic study using nationwide computerized inpatient/outpatient visits and prescription records from the Taiwan universal health care claims database.
Methods. Our cohort study included 15.5 million individuals (men=7,642,972; women=7,787,542) aged 20-80 years with no prior history of cancer as of 1/1/2001 and no prescription for any cholesterol-lowering drug from 1/1/2000 to 12/31/2000. During the 10-year follow-up period (1/1/2001 to 12/31/2010), 1,550 incident cases of MM (men=853; women=697) were identified. We identified cholesterol-lowering drugs using the WHO Anatomical Therapeutic Chemical classification system and defined statin users as subjects who filled a minimum 2 consecutive prescriptions within 90 days for lipophilic statins (atorvastatin, lovastatin, simvastatin) and hydrophilic statins (pravastatin, fluvastatin, rosuvastatin). Non-statin cholesterol-lowering drugs included fibrates, nicotinic acid and derivatives, and other lipid modifying agents. To specifically evaluate the effect of statins, we did not include subjects who used both statins and non-statin cholesterol-lowering drugs. For each drug, we calculated the defined daily dose (DDD) recommended by the WHO. We examined the association between statin use and MM risk using Cox proportional hazard models.
Results. Cholesterol-lowering drug users were older than never users: 78% and 59% of statins and non-statins cholesterol-lowering drug users, respectively, were older than 50-year-old, compared with only 26% among never users. Compared with never users, use of statins only was associated with a 20% lower risk of MM (Hazard Ratios (HR)=0.80; 95% confidence interval (CI)=0.68-0.94), after adjustment for age, sex, and Charlson comorbidity index. Each 50 unit increase in DDD was associated with a 2% lower risk of MM (HR=0.98; CI=0.96-0.99; p-trend=0.001). The inverse association was limited to lipophilic statins. To control for confounding by indication, we conducted subgroup analysis of individuals with a history of hyperlipidemia. Among subjects in this subgroup, subjects who used statin had a significantly lowered risk of MM (HR=0.67; CI=0.52-0.88) than those never used any cholesterol lowering drugs. Of the 6 individual types of statins evaluated, atorvastatin was the only drug that was inversely associated with MM risk (RR=0.59; CI=0.38-0.91).
Conclusion. In this large population-based cohort study, we found that lipophilic statin use, but not use of hydrophilic statins or non-statin cholesterol-lowering drugs, was inversely associated with risk of MM and the protective effect strengthens with increasing cumulative dose. These findings were not confounded by indication for treatment. Further studies, particularly clinical trials with sufficient duration of follow-up, are needed to clarify the role of statins in MM risk.
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.
