The list of possible side effects attributed to statins is long: muscle pain, memory problems, depression, sleep disturbances, diabetes, and liver damage. However, a meta-analysis by the Cholesterol Treatment Trialists’ (CTT) Collaboration in The Lancet calls many of these associations into question. At the same time, the study confirmed that certain effects on muscles, the liver, and glucose metabolism can indeed occur. What matters, however, is their magnitude — and that is usually significantly lower than what the public criticism of these drugs would suggest.
Surprisingly Few Confirmed Side Effects
The new meta-analysis includes 19 double-blind, randomized, placebo-controlled trials involving 123,940 participants with a median follow-up period of 4.5 years. In addition, the authors evaluated four trials involving 30,724 participants in which more intensive statin therapy was compared with less intensive statin therapy.
They examined 66 adverse events listed in the package inserts for atorvastatin, fluvastatin, pravastatin, rosuvastatin, and simvastatin. Aside from the known effects on muscle function and glucose metabolism, only four of the 66 events remained statistically significant after adjusting for various confounding factors:
- Elevated transaminase levels
- Other liver function test values
- Change in urine composition
- Edema
Elevated transaminase levels occurred at a rate of 0.30% per year with statins and 0.22% per year with placebo. The relative risk (RR) was 1.41. Other abnormal liver function test values were observed in 0.25% of cases per year with statins vs 0.20% of cases per year with placebo. Changes in urine composition were observed in 0.21% of cases per year with statins vs 0.18% of cases per year with placebo, and edema was observed in 1.38% of cases per year with statins vs 1.31% of cases per year with placebo.
For the remaining 62 endpoints, there was no significant evidence of harm after statistical adjustment for potential confounders.
Among the most common concerns expressed on social media is the fear of impaired memory and cognitive function. At first glance, this concern appears biologically plausible, given that cholesterol is essential for neuronal membranes and myelin. However, the brain’s cholesterol metabolism operates largely independently of peripheral low-density lipoprotein (LDL) metabolism.
Clinically, there is no convincing evidence of harm. The recent CTT analysis showed no robust association between statin therapy and cognitive impairment.
Another recent meta-analysis included 42 randomized trials involving 150,405 participants. For lipid-lowering therapies overall, the RR of neurocognitive events was 0.99 (95% CI, 0.88-1.12). For statins alone, the RR was 0.94 (95% CI, 0.72-1.25). No relevant negative effect was found in five specifically examined cognitive domains, namely attention, processing speed, executive function, working memory, and memory.
Depression and Sleep Disorders
Depression and sleep disorders are also often attributed to statins. However, both conditions occur frequently in the age group of patients typically prescribed statins. What matters, therefore, is not whether they occur during therapy but whether they are more frequent than without statins.
The new CTT meta-analysis found no evidence to support this. Neither depression nor sleep disturbances were significantly more common among statin users than among those taking a placebo. In the case of depression, observational data even pointed in the opposite direction.
A meta-analysis from 2025 included 15 studies from 10 countries with 5,403,692 participants. Statin users had a lower risk for depression, with a pooled odds ratio of 0.84 (95% CI, 0.74-0.96). However, the heterogeneity of the data was considerable. Therefore, no antidepressant effect can be inferred from these predominantly observational studies. However, they do not suggest a significant depression-promoting effect.
Muscle Pain
The situation is more nuanced when it comes to muscle pain. There is indeed a causal effect here, but the absolute magnitude is small.
The CTT Collaboration analyzed individual-level data from 19 placebo-controlled trials involving 123,940 participants. During a median follow-up of 4.3 years, 16,835 of 62,028 patients on statins (27.1%) reported muscle pain or muscle weakness. In the placebo group, the figure was 16,446 out of 61,912 participants (26.6%). The RR was 1.03 (95% CI, 1.01-1.06).
The difference was primarily concentrated in the first year of treatment. During this period, statins increased the RR by 7% (RR, 1.07; 95% CI, 1.04-1.10). In absolute terms, this corresponded to 11 additional muscle-related events per 1000 person-years. After the first year, no significant excess was detectable (RR, 0.99; 95% CI, 0.96-1.02).
The authors’ calculation is particularly illustrative: Of 15 muscle complaints reported by patients in the first year of statin therapy, statistically, only one was actually attributable to the medication. In the randomized trials, more than 90% of the muscle complaints reported by patients who were prescribed statins were therefore, mathematically speaking, not attributable to the statin.
'When Placebos Cause Almost as Many Complaints'
The SAMSON study demonstrated just how difficult the question of causality can be. The study included 60 patients who had previously discontinued statins due to side effects. They underwent 12 1-month phases: four with 20 mg of atorvastatin, four with a placebo and four without any medication at all. A total of 49 patients completed the entire study program.
Participants rated the intensity of their symptoms daily via an app on this scale:
- 1 = no or minimal symptoms
- 100 = maximum conceivable symptom intensity
The mean symptom score was 8.0 points during the months without medication. It increased to 16.3 points during periods when participants were taking atorvastatin, but it rose nearly as much — to 15.4 points — during placebo treatment. There was no significant difference between atorvastatin and placebo (P = .39). In other words, about 90% of the added symptom burden seen with atorvastatin, compared with that during the medication-free months, was also seen with placebo.
The larger StatinWISE study yielded similar results. Of 200 patients who had discontinued a statin due to muscle symptoms or intended to do so, 151 were included in the primary analysis. The difference in muscle symptom scores between atorvastatin and placebo was only -0.11 points on a 0-10 point scale (95% CI, -0.36 to 0.14; P = .40). Due to intolerable muscle symptoms, 9% discontinued treatment during the statin period and 7% discontinued treatment during the placebo period.
After the research team communicated the individual study results, 74 of 113 patients (65.5%) reported that they had already resumed statin therapy or intended to do so. After 15 months, 58 of 113 (51.3%) were indeed prescribed a statin again.
Why Statins May Affect Muscles
There is a biological explanation for why a small proportion of muscle symptoms are actually caused by statins. Statins inhibit 3-hydroxy-3-methylglutaryl coenzyme A reductase, thereby blocking the mevalonate pathway. This not only reduces cholesterol synthesis but also affects isoprenoid synthesis, protein prenylation, and coenzyme Q10 formation.
Discussions focus on changes in mitochondrial energy production, intracellular calcium homeostasis, and various signaling pathways in muscle cells. Mechanistic studies also point to possible effects on mitochondrial enzyme complexes and calcium ATPases.
However, the clinical data put these mechanisms into perspective: If 27.1% of patients on statins and 26.6% of patients on placebo report muscle symptoms, the impact on the mevalonate pathway cannot account for the majority of the observed complaints.
The same applies to coenzyme Q10. A decrease in its levels is biologically plausible. However, this does not mean that a Q10 deficiency is the most common cause of statin-associated muscle pain.
Diabetes
The evidence is clearer regarding glucose metabolism. Statins increase the risk for newly diagnosed diabetes, and the effect is dose dependent.
The CTT Collaboration analyzed 19 placebo-controlled trials involving 123,940 participants, as well as four dose-intensity studies involving 30,724 participants. Among patients on low- or moderate-intensity statins, 2420 out of 39,179 developed new-onset diabetes, compared with 2214 out of 39,266 in the placebo group. The annual rates were 1.3% vs 1.2%, corresponding to an RR of 1.10 (95% CI, 1.04-1.16).
With high-intensity statin therapy, the relative effect was greater: 1221 of 9935 patients on high-intensity statin therapy received a new diagnosis of diabetes compared with 905 of 9859 patients on placebo. The annual event rates were 4.8% vs 3.5%, with an RR of 1.36 (95% CI, 1.25-1.48). The metabolic shift itself was minor. Among participants without diabetes, mean glucose levels rose by only 0.04 mmol/L under low- or moderate-intensity statin therapy. The A1c level increased by an average of 0.06 percentage points and by 0.08 percentage points under high-intensity therapy.
Of note, approximately 62% of newly diagnosed cases of diabetes occurred in patients whose baseline glycemia was already in the top quartile. This suggests that the small increase in glycemia becomes clinically apparent primarily in patients whose baseline values are already close to the diagnostic threshold. For patients with high cardiovascular risk, however, the risk-benefit balance remains favorable.
Liver Enzymes
The liver is also an obvious target organ for side effects. This is where statins exert a large part of their effect: By inhibiting cholesterol synthesis, they increase the expression of hepatic LDL receptors, resulting in more LDL being removed from the blood.
It is well documented that laboratory values change. In the current Lancet analysis, 783 patients on statins showed elevated transaminase levels compared with 556 patients on placebo. The annual rate was 0.30% vs 0.22%, with an RR of 1.41 (95% CI, 1.26-1.57).
Other abnormal liver function tests were found in 651 vs 518 participants, corresponding to a rate of 0.25% vs 0.20% per year and an RR of 1.26. Taken together, this corresponded to an absolute annual excess of 0.13%.
The dose-response studies revealed a dose-response effect, further supporting causality. However, clinical interpretation remains crucial: An isolated elevation in transaminase levels is not equivalent to severe hepatotoxicity or liver failure.
Edema and Urinary Changes
Two lesser-known signals from The Lancet analysis are edema and changes in urinary composition:
- Regarding changes in urine composition, 556 events were documented in the statin group compared with 472 in the placebo group. The annual rates were 0.21% vs 0.18%, corresponding to an RR of 1.18 (95% CI, 1.04-1.33).
- Edema was significantly more common, but the difference between the groups was very small: 3495 patients on statins vs 3299 patients on placebo, corresponding to annual rates of 1.38% vs 1.31%. The RR was 1.07 (95% CI, 1.02-1.12).
Treatment Adherence Is Declining
The underlying problem: Both actual and perceived side effects are clinically relevant, because along with other factors, they can contribute to patients taking their statins irregularly or discontinuing their therapy altogether. Researchers distinguish between two terms:
- Adherence describes the extent to which actual medication use aligns with the prescribed dosing regimen. In many studies, a proportion of days covered or medication possession ratio of at least 80% is considered good adherence.
- Persistence, on the other hand, describes how long a patient continues therapy without discontinuing it or interrupting it for a predefined period.
There is significant room for improvement in both areas. A meta-analysis of 76 studies involving 5,898,141 patients found that only 62.4% (95% CI, 58.3%-66.5%) demonstrated good statin adherence. In primary prevention, the rate was 57.5%, and in secondary prevention, the rate was 64.4%.
The problem becomes even more apparent when it comes to persistence. In a German real-world analysis involving 865,732 patients, by day 300, approximately 71% had discontinued their statin therapy — as defined in the study — or had developed a treatment gap of the corresponding duration. After 36 months, only 20.6% remained persistent.
If symptoms are prematurely attributed to the statin, this can contribute to dose reduction or treatment interruption or discontinuation. Inadequate statin therapy, in turn, is associated with less favorable cardiovascular outcomes in observational studies. A recent systematic review on statin adherence and cardiovascular events confirms the link between higher adherence and lower cardiovascular risk. Thus, even supposed side effects can indirectly have clinical consequences.
Conclusion: What Does This Mean for Clinical Practice?
The current evidence paints a much more sobering picture of the side effect profile of statins than the long lists of possible complaints might suggest:
- Muscle symptoms may be causally related, but the absolute excess in the first year is approximately 11 additional events per 1000 person-years. In randomized trials, more than 90% of the muscle complaints reported by patients prescribed statins were, statistically speaking, not caused by the statin.
- Diabetes is a real, dose-dependent side effect. The RR of a new diagnosis increases by about 10% with low- to moderate-intensity therapy and by 36% with high-intensity therapy.
- Changes in liver function tests are also causal and dose dependent. However, the annual excess risk for abnormal liver function tests is only about 0.13%.
- In contrast, large, blinded randomized trials show no reliable increase in risk for cognitive impairment, depression, or sleep disturbances.
There is no question that statins can have side effects. Clinically, however, what matters is how likely it is that a specific symptom was actually caused by the statin — and how this risk compares with the cardiovascular benefits.
For this very reason, symptoms should neither be downplayed nor be hastily attributed to the statins. The practical challenge lies in recognizing genuine side effects, systematically investigating suspected side effects, and finding a lipid-lowering therapy that a patient will not only start but also continue long term.
https://www.medscape.com/viewarticle/statin-side-effects-what-evidence-shows-2026a100112s