We conducted a population-based cohort study to estimate the risk of myopathy associated with use of lipid-lowering drugs. Using data from general practices in the United Kingdom in 1991 through 1997, we identified three cohorts of individuals 40 to 74 years of age. One cohort comprised 17,219 persons who had received at least one prescription for lipid-lowering drugs in the period; a second cohort consisted of patients with a hyperlipidemia diagnosis who had not been prescribed lipid-lowering drugs (N = 28,974); and a third cohort comprised 50,000 individuals from the general population with no diagnosis of hyperlipidemia. The incidence rate of myopathy in the cohort of users of lipid-lowering drugs was 2.3 per 10,000 person-years [95% confidence interval (95% CI) = 1.2–4.4], which exceeded the incidence rates observed in the nontreated hyperlipidemia cohort [0 per 10,000 person-years (95% CI = 0.0–0.4)] and the general population [0.2 per 10,000 person-years (95% CI = 0.1–0.4)]. The relative risks of myopathy in current users of fibrates and statins compared with nonusers were 42.4 (95% CI = 11.6–170.5) and 7.6 (95% CI = 1.4–41.3), respectively. Potential risk factors other than drug use could not explain our findings in the nested case-control analysis. We conclude that use of lipid-lowering drugs is associated with a substantially greater risk of myopathy, which is most pronounced for fibrates. The absolute risk of myopathy in users of lipid-lowering drugs is, however, small.
From the Departments of 1Epidemiology
3Clinical Pharmacology, Institute of Public Health, University of Southern Denmark, Odense University, Denmark;
2Centro Español de Investigación Farmacoepidemiológica, Madrid, Spain;
4Department of Neurology, Odense University Hospital, Denmark.
Address correspondence to: David Gaist, Epidemiology, Institute of Public Health, University of Southern Denmark, Odense University, Sdr Boulevard 23A, 5000 Odense C, Denmark.
Submitted June 26, 2000; final version accepted December 4, 2000.
Lipid-lowering drugs have occasionally been associated with myopathy. The clinical presentation varies, but it typically includes muscle pain, weakness and tenderness, and moderate to markedly elevated serum creatinine kinase activity. 1,2 Fibrates (clofibrate, fenofibrate, bezafibrate, etofibrate, and beclofibrate), statins (lovastatin, simvastatin, and pravastatin), and nicotinic acid derivatives have been suggested to cause myopathy in reports comprising single cases and case series. 1,3–7 Rapid remission of symptoms after discontinuation of therapy has frequently been reported. 1 Use of combined lipid-lowering drugs may increase the risk of myopathy and result in a more severe clinical presentation occasionally leading to rhabdomyolysis. 2 The pathogenetic mechanism underlying this form of drug-induced myopathy is unknown. 1,2
Although rare, cases of myopathy have also been reported in the large clinical trials of statins conducted in more recent years. In the Expanded Clinical Evaluation of Lovastatin Study, which included 8,245 patients traced for 48 weeks, myopathy (defined as muscle symptoms with a creatine kinase elevation greater than ten times the upper limit of normal) was identified in five patients receiving lovastatin. 8 In the Scandinavian Simvastatin Survival Study, which offered a median follow-up of 5.4 years in 4,444 patients, a single patient on 20 mg of simvastatin developed myalgia with a creatine kinase level 100 times above the upper limit, and elevated creatine kinase without myalgia was reported in another six patients. 9
Epidemiologic studies of myopathy related to the use of lipid-lowering drugs in a population setting have not, to the best of our knowledge, previously been performed. We therefore carried out a large follow-up study based on data from British general practitioners. Our study provides estimates of the absolute and relative risks of myopathy in patients who received lipid-lowering drugs.
Subjects and Methods
The General Practice Research Database contains computerized medical information entered by general practitioners in the United Kingdom. 10 Data derived from practices covering approximately three million individuals in the United Kingdom are systematically recorded and sent anonymously to the Medicine Control Agency, which collects and organizes this information for research projects. The general practitioners generate prescriptions directly from the computer, and these are recorded into the patient’s computerized file. The information recorded includes demographic data, clinical diagnoses from outpatient visits, consultant referrals, and hospitalizations. A previous validation study has documented that more than 90% of all referrals present in the manual records in the general practitioners’ offices are entered into the general practitioners’ computers with a code that reflects the clinical diagnosis. 11 An additional requirement of this data resource is that the indication for any new course of therapy be entered in the computer. In addition, the general practitioner may record laboratory results and other medical data in a free text comment field. A modification of the Oxford Medical Information System (OXMIS) classification is used to code specific diagnoses, and a drug dictionary based on the Prescription Pricing Authority drug dictionary is used to record drugs. To date, more than 50 pharmacoepidemiologic studies have been published using the General Practice Research Database as the original source of information. 11
The study period started on January 1991 and ended on September 1997. The source population comprised patients 40–74 years of age with a current registration status of permanent or died. We identified three cohorts of patients free of disorders listed in Table 1 (potential study outcome diagnoses) and of exclusion criteria (alcohol abuse, diabetes, cancer, acquired immune deficiency syndrome, thyroid disease, connective tissue diseases, familial neuropathy diseases, and myasthenia gravis) before the start date. The first cohort included all subjects with a prescription for lipid-lowering drugs within the study period (N = 17,219). We also sampled from the source population a cohort of 50,000 patients who had no diagnosis of hyperlipidemia or use of lipid-lowering drugs at any time. This general population cohort was matched by age and sex to the cohort of lipid-lowering drug users. Finally, a third cohort with a diagnosis of hyperlipidemia but no use of lipid-lowering drugs was identified (N = 28,974). The three cohorts were followed up until the earliest of the following: any of the diagnoses in Table 1 (potential study outcome), any exclusion criteria, death, age of 75 years, or end of study period. We evaluated the risk of myopathy in the three cohorts.
Case Ascertainment and Validation
Myopathy is a rare disease. So that we would not overlook potential cases, we chose a sensitive screening strategy, which also included a large number of nonspecific codes, such as “weak muscle,” “leg weakness,” etc (Table 1). We identified 533 patients with a code included in Table 1 in the three cohorts. After manual review of their computerized profile, we could safely exclude a case of myopathy according to our eligibility criteria in 480 patients. In most cases these patients experienced localized muscular pains resulting from musculoskeletal (N = 208), vascular (N = 64), neurologic (N = 49), or various other (N = 6) disorders. The diagnosis could not be confirmed in a further 135 patients, because only one of our criteria was fulfilled in those cases. Finally, 18 patients were excluded because they fulfilled our exclusion criteria, either before (N = 5) or within a few months (N = 13) of follow-up. After this review, we thus identified in the three cohorts 53 potential cases of myopathy. To confirm the diagnosis, we requested from the general practitioners clinical records for all potential cases of myopathy. We received information on 46 cases (87%). A person was defined as a case of myopathy when he or she presented at least two of the following criteria: (1) clinical diagnosis of myopathy confirmed by the general practitioner; (2) muscle weakness, muscle pain, or muscle tenderness (two of these symptoms); and (3) creatine kinase concentration above the reference limit. After review of all available information by our group, which included an experienced consultant in neurology, 33 did not meet our criteria of idiopathic myopathy. The reasons for exclusion were: diagnosis not confirmed (28), vascular pathology (1), mechanical pathology (1), other disease (1), and prior history (2). All patient personal identifiers and drug use information were suppressed before review to maintain confidentiality and avoid information bias.
We defined two time windows of exposure for lipid-lowering drugs in the cohort of hyperlipidemia treated: current use and past use. Current use was use corresponding to the prescription supply plus an additional 30 days. Past use was all person-time after the end of current use. Duration of therapy was defined as the period corresponding to consecutive prescriptions. The reference group was the general population.
We calculated the crude incidence rates and 95% confidence intervals (95% CIs) of myopathy. We calculated incidence rates by dividing the number of cases of myopathy by the corresponding amount of person-time in that cohort. We also performed a nested case-control analysis to examine the effect of age, sex, smoking status, body mass index, and calendar year. All 13 incident cases of myopathy identified in the three cohorts were used as cases. To select controls a date during the study period was generated at random for all members of the three study cohorts. If the random date of a study member was included in his or her eligible person-time, we used his or her random date as the index date and marked that person as an eligible control. All inclusion and exclusion criteria applied during the selection of cases were applied to the control selection procedure. A random series of 1,000 controls was sampled. We computed estimates of relative risk and 95% CI of myopathy associated with current use of individual lipid-lowering drugs compared with non-use, with unconditional logistic regression.
The study cohort of hyperlipidemia treated included 17,219 persons who received 255,350 prescriptions, a mean of 15 prescriptions per user. In all, 10,003 patients were treated with simvastatin, and the second most favored drug, bezafibrate, was used by 4,163 patients (Table 2). The age and sex distribution of the three study cohorts is presented in Table 3. During 258,405 person-years of follow-up, we identified four cases of myopathy in the general population cohort, corresponding to an incidence rate of 0.2 per 10,000 person-years (95% CI = 0.1–0.4) (Table 4). We found no case of myopathy in the hyperlipidemia nontreated group. In all, nine cases of myopathy were identified in the hyperlipidemia treated cohort, corresponding to an incidence rate of 2.3 per 10,000 person-years (95% CI = 1.2–4.4). The relative risk of myopathy associated with current use of lipid-lowering drugs was substantially raised for statins (7.6; 95% CI = 1.4–41.3), and even more so for fibrates (42.4; 95% CI = 11.6–170.5). Current use of other lipid-lowering drugs conveyed no measurable effect, although this finding could be due to the limited number of person-years of observation in this group.
We examined the influence of other potential confounders in a nested case-control analysis. Owing to the small number of cases, we only examined the risk factors one at a time (Table 5). Persons 65 or more years of age, smokers, and individuals with a body mass index exceeding 24 kg/m2 were at increased risk of myopathy, although the relative risks conferred by these factors were smaller than those associated with current use of lipid-lowering drugs.
Current use of fenofibrate carried the highest (164.0; 95% CI = 8.7–3106.0) and simvastatin the lowest (6.1; 95% CI = 0.7–56.2) relative risk of myopathy (Table 6).
We found that use of lipid-lowering drugs greatly increased the risk of myopathy. Compared with the general population, current users of statins were at an eightfold increased risk of myopathy and users of fibrates were at a 42-fold increased risk of this disorder. The absence of any myopathy cases in a large cohort of patients with hyperlipidemia, who were not treated pharmacologically, indicates that hyperlipidemia per se is not associated with myopathy.
We made efforts to control for potential confounders. Patients with conditions that are known to predispose for or to be associated with myopathy were excluded from the cohorts. We went through the medication histories of the cases for drugs that have previously been associated with myopathy, for example, cyclosporin, penicillamine, and amiodarone, 1 but identified only a single case of concurrent use of enalapril in a current user of bezafibrate. The patient, however, continued to use enalapril uneventfully after cessation of fibrate treatment.
Among the 533 potential cases identified through our initial screening procedure, we had to exclude five patients because manual review of the computerized profile revealed that they fulfilled the exclusion criteria before the start of follow-up. This finding indicates a small weakness in our procedure for identifying patients fulfilling the exclusion criteria. This misclassification of eligibility leads to a slight overestimation of the available person-time of a magnitude that should have little bearing on our results.
Myopathy associated with use of lipid-lowering drugs has been reported to occur more frequently in subjects with impaired renal function. 1,2 We therefore scanned the records of all cases for diagnoses compatible with impaired renal function but found none. Finally, in a nested case-control analysis, we examined the effect of various potential confounders, such as age, sex, and smoking status. Although the small number of cases only allowed a univariate approach, the rate ratio estimates for these confounders were generally small. We therefore find it unlikely that any of the measured confounders are responsible for the strong association between lipid-lowering drugs and myopathy observed in our study. Owing to the purely observational nature of our study, we cannot rule out that at least part of the observed effects were due to unmeasured or inadequately controlled confounding.
The results of this first epidemiologic study to examine the risk of myopathy in users of lipid-lowering drugs confirm the suspicion raised by numerous case histories and case series. 1,3–7 From a public health perspective, we find it reassuring that treatment with lipid-lowering drugs is rarely associated with myopathy. According to our study, treating 10,000 patients with statins for a year would give rise to a single case of myopathy. If the same number of patients were treated with fibrates, roughly six cases of myopathy would occur. Although study design differences make straightforward comparisons difficult, our results seem to confirm the findings of the large clinical trials concerning the relatively small absolute magnitude of the problem. The risk of myopathy in users of simvastatins was thus recently estimated to be one case per 10,000 person-years on the basis of data from two megatrials, which included more than 12,000 patients randomized to simvastatin. 12
We conclude that the use of lipid-lowering drugs is associated with a marked increase in the risk of myopathy. The absolute risk of myopathy associated with these treatments, however, is small.
We thank the Boston Collaborative Drug Surveillance Program for providing access to the dataset and the general practitioners for their excellent collaboration.
1. Le Quintrec JS, Le Quintrec JL. Drug-induced myopathies. Baillieres Clin Rheumatol 1991; 5: 21–38.
2. Pierce LR, Wysowski DK, Gross TP. Myopathy and rhabdomyolysis associated with lovastatin-gemfibrozil combination therapy. JAMA 1990; 264: 71–75.
3. Giraud P, Cassou M, Paul R, Guidet M. [Muscular toxicity due to fenofibrate. Apropos of a case (Letter)]. Rev Rhum Mal Osteoartic 1982;49:162 [in French].
4. Rush P, Baron M, Kapusta M. Clofibrate myopathy: a case report and a review of the literature. Semin Arthritis Rheum 1986; 15: 226–229.
5. Magarian GJ, Lucas LM, Colley C. Gemfibrozil-induced myopathy. Arch Intern Med 1991; 151: 1873–1874.
6. Litin SC, Anderson DF. Nicotinic acid-associated myopathy: a report of three cases. Am J Med 1989; 86: 481–483.
7. Maher VM, Pappu A, Illingworth DR, Thompson GR. Plasma mevalonate response in lovastatin-related myopathy (Letter). Lancet 1989; 2: 1098.
8. Bradford RH, Shear CL, Chremos AN, Dujovne C, Franklin FA, Gould AL, Hesney M, Higgins J, Hurley DP. Expanded Clinical Evaluation of Lovastatin (EXCEL) study results. I. Efficacy in modifying plasma lipoproteins and adverse event profile in 8245 patients with moderate hypercholesterolemia. Arch Intern Med 1991; 151: 43–49.
9. Pedersen TR, Berg K, Cook TJ, Faergeman O, Haghfelt T, Kjekshus J, Miettinen T, Musliner TA, Olsson AG, Pyorala K, Thorgeisson G, Tobert JA, Wedel H, Wilhelmssen L. Safety and tolerability of cholesterol lowering with simvastatin during 5 years in the Scandinavian Simvastatin Survival Study. Arch Intern Med 1996; 156: 2085–2092.
10. Jick H, Jick SS, Derby LE. Validation of information recorded on general practitioner based computerised data resource in the United Kingdom. BMJ 1991; 302: 766–768.
11. García Rodríguez LA, Peréz Gutthann S. Use of the UK General Practice Research Database for pharmacoepidemiology. Br J Clin Pharmacol 1998; 45: 419–425.
12. Gruer PJK, Vega JM, Mercuri MF, Dobrinska MR, Tobert JA. Concomitant use of cytochrome P450 3A4 inhibitors and simvastatin. Am J Cardiol 1999; 84: 811–815.