Ajay Gupta1, David Thompson2, Andrew Whitehouse2, Tim Collier3, Bjorn Dahlof4, Neil Poulter5, Rory Collins6, Peter Sever7. 1. National Heart and Lung Institute, Imperial College London, London, UK; Royal London Hospital, Barts Health NHS Trust, Whitechapel, London, UK; William Harvey Research Institute, Queen Mary University of London, London, UK. 2. National Heart and Lung Institute, Imperial College London, London, UK. 3. Department of Medical Statistics, London School of Hygiene & Tropical Medicine, London, UK. 4. Department of Molecular and Clinical Medicine, Institute of Medicine, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden. 5. Imperial Clinical Trials Unit, Imperial College London, London, UK. 6. Clinical Trial Service Unit and Epidemiological Studies Unit, Nuffield Department of Population Health, University of Oxford, Oxford, UK. 7. National Heart and Lung Institute, Imperial College London, London, UK. Electronic address: p.sever@imperial.ac.uk.
Abstract
BACKGROUND: In blinded randomised controlled trials, statin therapy has been associated with few adverse events (AEs). By contrast, in observational studies, larger increases in many different AEs have been reported than in blinded trials. METHODS: In the Lipid-Lowering Arm of the Anglo-Scandinavian Cardiac Outcomes Trial, patients aged 40-79 years with hypertension, at least three other cardiovascular risk factors, and fasting total cholesterol concentrations of 6·5 mmol/L or lower, and who were not taking a statin or fibrate, had no history of myocardial infarction, and were not being treated for angina were randomly assigned to atorvastatin 10 mg daily or matching placebo in a randomised double-blind placebo-controlled phase. In a subsequent non-randomised non-blind extension phase (initiated because of early termination of the trial because efficacy of atorvastatin was shown), all patients were offered atorvastatin 10 mg daily open label. We classified AEs using the Medical Dictionary for Regulatory Activities. We blindly adjudicated all reports of four prespecified AEs of interest-muscle-related, erectile dysfunction, sleep disturbance, and cognitive impairment-and analysed all remaining AEs grouped by system organ class. Rates of AEs are given as percentages per annum. RESULTS: The blinded randomised phase was done between February, 1998, and December, 2002; we included 101 80 patients in this analysis (5101 [50%] in the atorvastatin group and 5079 [50%] in the placebo group), with a median follow-up of 3·3 years (IQR 2·7-3·7). The non-blinded non-randomised phase was done between December, 2002, and June, 2005; we included 9899 patients in this analysis (6409 [65%] atorvastatinusers and 3490 [35%] non-users), with a median follow-up of 2·3 years (2·2-2·4). During the blinded phase, muscle-related AEs (298 [2·03% per annum] vs 283 [2·00% per annum]; hazard ratio 1·03 [95% CI 0·88-1·21]; p=0·72) and erectile dysfunction (272 [1·86% per annum] vs 302 [2·14% per annum]; 0·88 [0·75-1·04]; p=0·13) were reported at a similar rate by participants randomly assigned to atorvastatin or placebo. The rate of reports of sleep disturbance was significantly lower among participants assigned atorvastatin than assigned placebo (149 [1·00% per annum] vs 210 [1·46% per annum]; 0·69 [0·56-0·85]; p=0·0005). Too few cases of cognitive impairment were reported for a statistically reliable analysis (31 [0·20% per annum] vs 32 [0·22% per annum]; 0·94 [0·57-1·54]; p=0·81). We observed no significant differences in the rates of all other reported AEs, with the exception of an excess of renal and urinary AEs among patients assigned atorvastatin (481 [1·87%] per annum vs 392 [1·51%] per annum; 1·23 [1·08-1·41]; p=0·002). By contrast, during the non-blinded non-randomised phase, muscle-related AEs were reported at a significantly higher rate by participants taking statins than by those who were not (161 [1·26% per annum] vs 124 [1·00% per annum]; 1·41 [1·10-1·79]; p=0·006). We noted no significant differences between statin users and non-users in the rates of other AEs, with the exception of musculoskeletal and connective tissue disorders (992 [8·69% per annum] vs 831 [7·45% per annum]; 1·17 [1·06-1·29]; p=0·001) and blood and lymphatic system disorders (114 [0·88% per annum] vs 80 [0·64% per annum]; 1·40 [1·04-1·88]; p=0·03), which were reported more commonly by statin users than by non-users. INTERPRETATION: These analyses illustrate the so-called nocebo effect, with an excess rate of muscle-related AE reports only when patients and their doctors were aware that statin therapy was being used and not when its use was blinded. These results will help assure both physicians and patients that most AEs associated with statins are not causally related to use of the drug and should help counter the adverse effect on public health of exaggerated claims about statin-related side-effects. FUNDING: Pfizer, Servier Research Group, and Leo Laboratories.
RCT Entities:
BACKGROUND: In blinded randomised controlled trials, statin therapy has been associated with few adverse events (AEs). By contrast, in observational studies, larger increases in many different AEs have been reported than in blinded trials. METHODS: In the Lipid-Lowering Arm of the Anglo-Scandinavian Cardiac Outcomes Trial, patients aged 40-79 years with hypertension, at least three other cardiovascular risk factors, and fasting total cholesterol concentrations of 6·5 mmol/L or lower, and who were not taking a statin or fibrate, had no history of myocardial infarction, and were not being treated for angina were randomly assigned to atorvastatin 10 mg daily or matching placebo in a randomised double-blind placebo-controlled phase. In a subsequent non-randomised non-blind extension phase (initiated because of early termination of the trial because efficacy of atorvastatin was shown), all patients were offered atorvastatin 10 mg daily open label. We classified AEs using the Medical Dictionary for Regulatory Activities. We blindly adjudicated all reports of four prespecified AEs of interest-muscle-related, erectile dysfunction, sleep disturbance, and cognitive impairment-and analysed all remaining AEs grouped by system organ class. Rates of AEs are given as percentages per annum. RESULTS: The blinded randomised phase was done between February, 1998, and December, 2002; we included 101 80 patients in this analysis (5101 [50%] in the atorvastatin group and 5079 [50%] in the placebo group), with a median follow-up of 3·3 years (IQR 2·7-3·7). The non-blinded non-randomised phase was done between December, 2002, and June, 2005; we included 9899 patients in this analysis (6409 [65%] atorvastatin users and 3490 [35%] non-users), with a median follow-up of 2·3 years (2·2-2·4). During the blinded phase, muscle-related AEs (298 [2·03% per annum] vs 283 [2·00% per annum]; hazard ratio 1·03 [95% CI 0·88-1·21]; p=0·72) and erectile dysfunction (272 [1·86% per annum] vs 302 [2·14% per annum]; 0·88 [0·75-1·04]; p=0·13) were reported at a similar rate by participants randomly assigned to atorvastatin or placebo. The rate of reports of sleep disturbance was significantly lower among participants assigned atorvastatin than assigned placebo (149 [1·00% per annum] vs 210 [1·46% per annum]; 0·69 [0·56-0·85]; p=0·0005). Too few cases of cognitive impairment were reported for a statistically reliable analysis (31 [0·20% per annum] vs 32 [0·22% per annum]; 0·94 [0·57-1·54]; p=0·81). We observed no significant differences in the rates of all other reported AEs, with the exception of an excess of renal and urinary AEs among patients assigned atorvastatin (481 [1·87%] per annum vs 392 [1·51%] per annum; 1·23 [1·08-1·41]; p=0·002). By contrast, during the non-blinded non-randomised phase, muscle-related AEs were reported at a significantly higher rate by participants taking statins than by those who were not (161 [1·26% per annum] vs 124 [1·00% per annum]; 1·41 [1·10-1·79]; p=0·006). We noted no significant differences between statin users and non-users in the rates of other AEs, with the exception of musculoskeletal and connective tissue disorders (992 [8·69% per annum] vs 831 [7·45% per annum]; 1·17 [1·06-1·29]; p=0·001) and blood and lymphatic system disorders (114 [0·88% per annum] vs 80 [0·64% per annum]; 1·40 [1·04-1·88]; p=0·03), which were reported more commonly by statin users than by non-users. INTERPRETATION: These analyses illustrate the so-called nocebo effect, with an excess rate of muscle-related AE reports only when patients and their doctors were aware that statin therapy was being used and not when its use was blinded. These results will help assure both physicians and patients that most AEs associated with statins are not causally related to use of the drug and should help counter the adverse effect on public health of exaggerated claims about statin-related side-effects. FUNDING: Pfizer, Servier Research Group, and Leo Laboratories.
Authors: Emily Herrett; Elizabeth Williamson; Kieran Brack; Alexander Perkins; Andrew Thayne; Haleema Shakur-Still; Ian Roberts; Danielle Prowse; Danielle Beaumont; Zahra Jamal; Ben Goldacre; Tjeerd van Staa; Thomas M MacDonald; Jane Armitage; Michael Moore; Maurice Hoffman; Liam Smeeth Journal: Health Technol Assess Date: 2021-03 Impact factor: 4.014