Statin Use, Clinical Fracture, and Bone Density in Postmenopausal Women: Results from the Women's Health Initiative Observational Study

Context Some observational studies have shown fewer fractures in patients receiving statins, but other studies have shown no effect. The studies have been small and had limited ability to adjust for potential confounders. Contribution In this subanalysis of the Women's Health Initiative, postmenopausal women had similar rates of hip, lower arm or wrist, and other fractures whether or not they used statins. The authors adjusted for many potential confounders, and the estimates of fracture rates were very precise. Implications Statins do not seem to prevent fractures in postmenopausal women. The Editors The long-term efficacy and safety of 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors (statins) have been established in large multicenter trials of cholesterol-lowering for preventing coronary events in both sexes (1-4). Recent laboratory studies have shown that statins stimulate bone formation in cultured osteoblasts, neonatal murine calvaria, and the cortical bone of mice (5). Statins administered orally increased the trabecular bone volume of female rats by 90%. These findings raise the possibility that statin treatment might prevent both coronary and fracture events, two major causes of morbidity in older women, later in life. Early epidemiologic studies examining the association of statin use with risk for hip fracture produced encouraging results (6-9). These studies were limited, however, by either small numbers of fractures (6) or lack of data on important potential confounders (7-9). More recent studies had mixed results (10-12), with some showing no association (11, 12). Few studies have examined associations with both fracture rates and bone density in the same study group. Therefore, we examined the association of statin use with levels of bone density and rates of hip, lower arm or wrist, and other clinical fractures in the Women's Health Initiative (WHI) Observational Study cohort of postmenopausal women. Methods Study Group The study group for this paper is the WHI Observational Study, a prospective cohort study that enrolled 93 716 women ages 50 to 79 years from 1994 to 1998 at 40 clinical centers throughout the United States. Study methods have been described in detail elsewhere (13). Briefly, women were eligible if they were postmenopausal, were unlikely to relocate or die within 3 years, were not enrolled in the WHI Clinical Trial, and were not participating in any other clinical trial. At baseline, women completed screening and enrollment questionnaires by interview and self-report, physical examination, and blood specimen collection. Human subjects review committees at each participating institution reviewed and approved the study. Follow-up and Outcome Ascertainment Women are sent questionnaires annually to report any hospitalization and a wide variety of outcomes, including clinical fractures of any type. Follow-up time ranged from 2 to 6 years per participant as of February 2001 (median duration, 3.9 years). At that time, 2.8% of participants (n = 2632) had withdrawn or were lost to follow-up [2.7% of statin users and 2.8% of nonusers]. Hip fractures are confirmed by central review of radiology reports. Other fractures are counted on the basis of self-report. Nonetheless, in the WHI Clinical Trial, in which all fractures are adjudicated, 81% of self-reported nonhip clinical fractures are confirmed by physician review of medical records, suggesting that the self-report of such fractures is reasonably accurate. For this report, we classified fractures into three mutually exclusive categories: 1) hip fractures, 2) lower arm or wrist fractures, and 3) other clinical fractures. Clinically recognized vertebral fractures were classified as other clinical fractures. Bone mineral density at the total hip, posterioranterior spine, and total body was measured at baseline in three clinical centers among 6442 women (97% of participants enrolled in Pittsburgh, Pennsylvania; Birmingham, Alabama; and Phoenix and Tucson, Arizona) with dual-energy x-ray absorptiometry using a Hologic QDR densitometer (Hologic, Inc., Waltham, Massachusetts). Standard protocols for positioning and analysis were used by technicians who were trained and certified by the University of California, San Francisco, Bone Density Coordinating Center, San Francisco, California. The ongoing quality assurance program includes monitoring spine and hip phantom scans; reviewing a random sample of all scans and flagging scans with specific problems; hardware or software change control, including in vitro and in vivo cross-calibration; and scanning calibration phantoms across instruments and clinical sites. Statin Exposure and Potential Confounders Participants were asked to bring all current prescription medications to their first screening interview. Clinic interviewers entered each medication name directly from the containers into the WHI database, which assigned drug codes using Medispan software (First DataBank, Inc., San Bruno, California). Women reported duration of use for each current medication. Information on dose was not recorded. Current medication use was ascertained by using identical methods at the year 3 clinic visit. Current statin medication use was defined as use of any HMG-CoA reductase inhibitor. Duration of use was examined in three categories (<1 year, 1 to 3 years, or >3 years). Statin medications were further categorized into three groups according to their demonstrated potency for lipid-lowering on the basis of a dose-efficacy trial (14): low potency (fluvastatin and lovastatin), medium potency (pravastatin), and high potency (atorvastatin, simvastatin). Other lipid-lowering medications were fibrate, colestipol, probucol, cholestyramine, niacin, or nicotinic acid. All covariates were ascertained at baseline. Current use of thiazide diuretics, alendronate, corticosteroid, and sedative or hypnotic medications was recorded by using the same procedures described. Current and previous use of hormone replacement therapy was ascertained by interview using a detailed questionnaire that measured type, route of administration, number of pills per day or week, and duration for each hormonal preparation ever taken. For the purposes of this report, hormone replacement therapy was defined as current use of any estrogen with or without progestin. Dietary supplements, including calcium preparations, taken at least twice weekly for the past 2 weeks were also entered into the database. Dietary intake of calcium was measured by using a semi-quantitative food-frequency questionnaire (15). Total calcium intake was defined as the sum of calcium from diet and supplements. Baseline questionnaires ascertained information on race or ethnicity, history of fracture or coronary heart disease (history of myocardial infarction or angina), current and past smoking, coffee consumption (cups per day), and time spent walking outside the home for more than 10 minutes without stopping (minutes per week). Alcohol consumption was estimated from the food-frequency questionnaire. Physical function was measured by using the 10-item Medical Outcomes Study scale (16). Weight was measured to the nearest 0.1 kg on a balance-beam scale with the participant dressed in indoor clothing without shoes. Height was measured to the nearest 0.1 cm by using a wall-mounted stadiometer. Body mass index was calculated as weight in kg/height in m2. Statistical Analysis The characteristics of women taking a statin medication at baseline were compared with those of women not taking statin medication by using chi-square tests to determine the statistical significance of the differences. Age-adjusted incidence rates of hip, lower arm or wrist, and other clinical fractures per 1000 person-years were calculated according to duration of statin use by the direct method, using the age distribution of the full cohort as the standard population. Women contributed follow-up time until the occurrence of fracture, death, or the end of follow-up, whichever came first. The a priori analysis plan specified selected stratified analyses to determine whether associations between statin use and fracture were apparent in key subgroups of women. For these analyses, women were stratified according to age group (50 to 64 years versus 65 years), current hormone replacement therapy use, body mass index (<25 kg/m2 versus 25 kg/m2, the standard threshold for overweight) (17), history of fracture, and history of coronary disease. Hazard ratios adjusted for age and corresponding 95% CIs were calculated by using Cox proportional-hazards survival models for each fracture category using PHREG in SAS software, version 8.2 (SAS Institute, Inc., Cary, North Carolina). To control for potential confounding factors, hazard ratios and corresponding 95% CIs for categories of duration and potency of statin use were calculated by using multivariate Cox proportional-hazards survival models, using the forced entry approach for variable selection. The multivariate models adjusted for age; race or ethnicity; body mass index; previous fracture; previous coronary disease; current and past hormone replacement therapy use; thiazide, alendronate, corticosteroid, or psychoactive drug use; calcium intake; walking; current and past smoking; coffee intake; and physical function. The multivariate models included individual clinical center as a stratification (or blocking) variable to allow the underlying hazard to be estimated separately for each clinical center. Interaction by clinical center was evaluated by comparing log likelihood statistics (chi-squares) for a model with interaction terms for each clinical center and a reduced model without these terms. Tests for the proportional hazards assumption were conducted by examining plots of the baseline hazard by statin duration categories and by testing interaction terms of statin use by time. Multivariate models were based on 81 896 individuals after exclusion