Methods

This study was conducted from January to September 2008 in a community-based setting in the Philadelphia, PA area (trial registration at clinicaltrials.gov , identifier NCT00639223 ). The institutional review boards at the University of Pennsylvania and Chestnut Hill Hospitals approved the study. All participants provided written informed consent. The subjects were recruited from preventive cardiology clinics at University of Pennsylvania and Thomas Jefferson University Hospitals, and 2 suburban Philadelphia cardiology practices. The subjects were eligible if they had had previous, documented SAM leading to discontinuation of at least one statin other than pravastatin, with resolution of myalgia after discontinuation. The exclusion criteria included statin or red yeast rice use during the month before randomization, a history of statin-associated myositis or rhabdomyolysis, a history of generalized chronic pain, the use of medications that inhibit cytochrome P450 CYP3A4, the use of dietary supplements that could mitigate SAM or lower lipids, abnormal baseline laboratory values (creatine phosphokinase >500 U/L, triglycerides ≥400 mg/dl, aspartate aminotransferase or alanine aminotransferase >2.5 times normal, serum creatinine >2 mg/dl, thyroid-stimulating hormone >4.5 μU/ml), and pregnancy.

Eligible participants were randomized to receive red yeast rice 4,800 mg daily (four 600-mg capsules twice daily; Sylvan Bioproducts, Kittanning, Pennsylvania) or pravastatin 40 mg/day (1 overencapsulated 20-mg tablet to appear identical to the red yeast rice capsules and 3 identical-appearing placebo capsules twice daily) for 12 weeks. A sample of red yeast rice was independently analyzed for chemical composition (Eurofins Scientific, Petaluma, California; Table 1 ). An investigational new drug application for using red yeast rice in the present trial was approved by the Food and Drug Administration. Adherence to the study medication was determined by pill counts of the returned study medication every 4 weeks. The mean adherence to pravastatin and red yeast rice (excluding dropouts and subjects who withdrew from medication) was 97% and 93%, respectively (p = 0.10). To ensure that both groups received identical lifestyle education, all participants attended weekly 3.5-hour sessions of a therapeutic lifestyle change program (see Appendix [on-line only]). Attendance at these meetings averaged 83%, with no significant difference in attendance between the 2 groups.

The primary outcome of the present study was the rate of withdrawal from treatment because of intolerable muscle symptoms. The co-primary outcome was the daily pain severity score measured using one question adapted from the Brief Pain Inventory regarding the average pain during the past 24 hours (on a 0 to 10 scale). Participants rated both nonmyalgic and myalgic pain using this scale. The same blinded physician reviewed these pain scales weekly. If the subjects reported intolerable myalgia, their study treatment was discontinued, but all planned measurements were obtained. Isometric hip flexor muscle strength was determined by the same blinded physical therapist at baseline and every 4 weeks using a standard protocol with a hand-held dynamometer (model 01163, Lafayette Industries, Lafayette, Indiana). The hand-held dynamometer has a high correlation (0.91) with isokinetic muscle testing and is useful in comparing serial muscle strength tests.

A fasting blood sample was obtained at baseline and week 12 to determine the low-density lipoprotein (LDL) cholesterol, total cholesterol, high-density lipoprotein (HDL) cholesterol, triglycerides, creatine phosphokinase, aspartate aminotransferase, alanine aminotransferase, and thyroid-stimulating hormone (baseline only) levels. Analyses were performed by the Hospital of the University of Pennsylvania.

Adherence to the dietary recommendations was assessed by collecting two 24-hour diet recalls (1 weekend day and 1 weekday) at baseline and week 12. Blinded, trained staff collected the dietary survey by telephone interview, with data collection centralized at the Penn State Department of Nutritional Sciences (University Park, Pennsylvania). The dietary intake data were analyzed using Nutrition Data System, version 2007 (Nutrition Coordinating Center, Minneapolis, Minnesota). Physical activity was assessed using the Paffenbarger Physical Activity Questionnaire at baseline and week 12. This validated, reliable, self-administered questionnaire quantifies the number of kilocalories subjects expend per week in sports, leisure, and recreational activities.

The subjects were randomized to the study medication in blocks of 4. Although all participants had a history of myalgia with at least one statin drug, many had previously been challenged with other statins. Therefore, the randomization was stratified into 3 groups: those who had not developed myalgia on challenge with a different statin, those who had developed intolerable myalgia with all previous statin challenges, and those who had never been challenged with another statin drug. The randomization sequence was computer-generated using a randomization program available on the Internet ( www.randomization.com ) with the fixed block option. All subjects and study team members were kept unaware of treatment allocation throughout the 12-week study. To assess blinding, the participants guessed their treatment allocation at the end of the study; 37% of those taking pravastatin and 67% of those taking red yeast rice guessed their treatment assignment correctly.

Sample size calculations were performed on the basis of the 7% rate of intolerable myalgias seen with red yeast rice in our previous study and rates of approximately 50% reported in the published data with a statin rechallenge, assuming a 20% dropout rate. According to these rates, a sample size of 20 to 22 subjects per group would provide 80% power to detect a difference of 40% between the 2 groups for intolerable myalgic symptoms, with α = 0.05. All primary analyses were conducted using the intention-to-treat approach. We compared the incidence of withdrawal from the study medication because of intolerable myalgia between the 2 groups using Fisher’s exact test. A linear regression model was used to compare the difference between the treatment groups in the mean Brief Pain Inventory pain severity score, defined as the maximum reported score for myalgic and nonmyalgic pain during each day. The robust variance estimator was used to adjust the standard error estimates for correlation due to repeated measurements. Missing pain scores were treated as missing, because only 2 subjects (dropouts) had missing data (<5%). A linear mixed-effects model was used to analyze hip flexor strength at baseline and weeks 4, 8, and 12. The covariates in the models for pain and muscle strength included stratification assignment, age, baseline thyroid-stimulating hormone level, and baseline physical activity. A secondary analysis was performed to assess the differences between groups in the occurrence of myalgic pain during the intervention period. For participants who reported myalgic pain, we created dichotomous outcomes, defined a priori, indicating whether the subjects had persistent myalgia starting after >2 weeks of study treatment and of ≥1 week’s duration. Comparisons between treatment groups were made using Fisher’s exact tests.

To evaluate treatment efficacy, descriptive statistics were computed for LDL cholesterol, total cholesterol, HDL cholesterol, and triglycerides at baseline and week 12. Linear regression models quantified the differences in the mean LDL cholesterol, total cholesterol, HDL cholesterol, and triglycerides at week 12 across the treatment groups. Each model was adjusted for its respective baseline lipoprotein measure and baseline body mass index. To assess the adherence to the therapeutic lifestyle change program, we compared the mean percentage of change from baseline to week 12 across the treatment groups for the following outcomes using 2-sample t tests: total energy intake, total fat, saturated fat, percentage of energy from fat, fiber, weight, and total kilocalories expended per week. Safety parameters, including liver-associated enzymes and creatine phosphokinase, were analyzed for differences between the 2 groups using 2-sample t tests. The differences in the adverse event rates were compared between treatment groups using Fisher’s exact tests. Statistical analyses were performed using Stata version 9.2 (StataCorp, College Station, Texas). All comparisons were 2-tailed, and the level of significance was set at p = 0.05.