A recent study by the Cholesterol Treatment Trialists’ (CTT) Collaboration evaluated whether the wide range of adverse effects listed in statin product labels are causally attributable to statin therapy. Although statins are well established as effective in reducing major vascular events and cardiovascular mortality, concerns about possible harms, many derived from non-randomised or non-blinded evidence, have led to extensive lists of undesirable effects in Summaries of Product Characteristics (SmPCs). Such listings may overstate risk, contribute to patient anxiety, and reduce adherence. The investigators undertook a comprehensive meta-analysis of individual participant data from large, double-blind randomised controlled trials to provide more reliable evidence regarding these potential harms.
The background to the analysis highlights the strong benefits of statin therapy. Over five years, effective statin regimens can prevent approximately 1000 major vascular events per 10,000 patients with established occlusive disease and about 500 events among high-risk individuals without prior events. Established adverse effects include rare myopathy and rhabdomyolysis, a small excess of milder muscle symptoms (mainly during the first year), and a moderate dose-dependent increase in diabetes incidence. However, numerous other outcomes, such as hepatic dysfunction, depression, cognitive impairment, sleep disturbance, renal injury, lung disease, and pancreatitis, have been associated with statins in observational studies that are prone to confounding and reporting bias. The authors argued that only large, blinded trials can reliably assess moderate risks for common outcomes.
Eligible trials were double-blind randomised studies with at least 1000 participants and at least two years of scheduled treatment. Two comparison types were included: statin versus placebo, and more intensive versus less intensive statin regimens. Researchers searched an electronic medicines compendium for five commonly used statins (atorvastatin, fluvastatin, pravastatin, rosuvastatin, and simvastatin) to identify all adverse events listed in SmPCs. After excluding previously assessed muscle and diabetes outcomes, 66 prespecified undesirable outcomes across 15 body systems were analysed.
Nineteen statin-versus-placebo trials contributed 123,940 participants with a median follow-up of 4.5 years. These trials provided extensive and systematically collected adverse event data. Of the 66 outcomes assessed, only four showed statistically significant excess risks after FDR adjustment. Abnormal liver transaminases occurred more frequently with statins (0.30% per annum vs 0.22%). Other liver function test abnormalities were also increased (0.25% vs 0.20%). The combined absolute annual excess for liver biochemistry abnormalities was small at approximately 0.13%. Urinary composition alterations showed a modest excess (0.21% vs 0.18%), as did oedema (1.38% vs 1.31%). All these absolute risk increases were very small, generally less than 0.1% per year.
Four additional trials (30,724 participants) compared more intensive with less intensive statin regimens. These analyses supported a dose-dependent relationship for liver transaminase and other liver function abnormalities, reinforcing a likely causal effect of statins on liver biochemistry. However, no significant excess was observed for urinary composition changes or oedema in these intensity comparisons, suggesting these latter findings might not reflect clear dose-related causality.
No significant excess risks were identified for the remaining 62 prespecified outcomes. These included many conditions commonly cited in product labels and patient information leaflets, such as cognitive impairment, depression, sleep disturbance, peripheral neuropathy, erectile or sexual dysfunction, acute kidney injury, interstitial lung disease, and other hepatobiliary outcomes. The absence of associations in well-powered, blinded trials strongly argues against causal links between statins and these conditions.
The authors emphasise that non-randomised studies are vulnerable to confounding, expectation bias, and differential reporting, especially when patients and clinicians are aware of potential side effects. Such biases may generate spurious associations that become codified in regulatory labels despite weak evidence. In contrast, the double-blind randomised design used here minimises these errors and provides more dependable estimates of harm.
The investigators conclude that, beyond the known risks of muscle effects and diabetes, statins are associated only with small, clinically modest increases in liver biochemical abnormalities and possibly minor changes in urinary composition and oedema. Most other labelled adverse effects lack reliable supporting evidence. Consequently, current product labelling may overstate risks and inadvertently discourage beneficial treatment. The authors recommend that regulatory and informational materials be revised to better reflect high-quality evidence, thereby enabling patients and clinicians to make more informed decisions about statin use.
Source: The Lancet
Image Credit: iStock
