Rhabdomyolysis with statins
That muscle problems can occur with statins and other lipid lowering drugs is an accepted problem. While some people seem unable to take statins because of muscle soreness or weakness, the vast majority are unaffected. There is something of a biological progression, from muscle soreness, through more severe muscle problems, to increased levels of creatinine kinase enzymes, to rhabdomyolysis, and even death from rhabdomyolysis (in about 1 in 15 cases).
Most of the spontaneous reports of rhabdomyolysis to the FDA were associated with cerivastatin, now withdrawn [1]. The problem with spontaneous reporting is that while it may identify cases, there is always uncertainty about denominators, so rates of adverse events are imprecise. Randomised trials are poor at finding rare but serious adverse events, because the events do not occur in sufficient numbers. Only 12 cases of rhabdomyolysis were reported in 30 RCTs reviewed [1].
One way of trying to overcome both these problems is by using large cohort studies, where people prescribed a drug, statins in this case, are enrolled in good databases that can identify cases of the adverse events examined. Such a cohort study on statins and rhabdomyolysis [2] is informative.
Study
This was a retrospective cohort study of patients in 11 US health plans providing pharmacy benefits, and with automated claims files covering prescription drugs, outpatient visits, and hospital admissions. Patients with a first prescription of statin or fibrate were entered, as long as there was no such prescription in the previous six months.
Potential cases of hospital admission for rhabdomyolysis were identified from records of members of the cohort using coded discharge diagnoses. Also used were claims for measurement of creatinine kinase within seven days of admission or discharge, or a discharge diagnosis of renal failure plus a creatinine kinase measurement.
Three assessors blind to statin or fibrate exposure status reviewed abstracts of medical records. Rhabdomyolysis was defined as severe muscle injury present at admission, plus a diagnosis of rhabdomyolysis or creatinine kinase more than 10 times the upper limit of normal. Severe rhabdomyolysis was defined as a serum creatinine kinase above 10,000 IU/L or more than 50 times the upper limit of normal.
Results
The cohort included a quarter of a million people with 225,000 person years of monotherapy for statin or fibrate, and 7,300 years of combined therapy of statin plus fibrate. There was little information for fluvastatin or lovastatin, and these drugs were ignored in favour of the bulk of statin information on atorvastatin, pravastatin, and simvastatin. Fibrate information was much lower than that for statins, and included gemfibrozil and fenofibrate. Within the cohort there were 77,000 person years not exposed to lipid-lowering drugs, during which no cases of rhabdomyolysis were reported.
Thirty-one patients met the inclusion criteria for rhabdomyolysis. Seven of these were excluded because rhabdomyolysis occurred during a period when the prescription records showed that they were not exposed to a lipid-lowering drug. In each case their hospital records showed that they were taking a statin at the time of the event.
Of the 24 events remaining, there were 13 events on statin monotherapy, three on fibrate monotherapy, and eight cases with combined therapy with both statin and a fibrate. Various doses of each statin and each fibrate were being used for monotherapy and combined therapy. Three-quarters of the events were defined as severe rhabdomyolysis. Hospital stay was 1-11 days (average 6). Two patients underwent haemodialysis, and one died.
Monotherapy
There was no difference in rate of hospital admission for rhabdomyolysis between atorvastatin, pravastatin, and simvastatin. The combined event rate was 0.44 per 10,000 person years of exposure (Table 1), and the one-year number needed to harm (NNH) was 22,700. For cerivastatin, the rate was about 10 times higher, with a NNH of 1,870. For fibrates the NNH was 3,550. Age over 65 years and having diabetes increased the risk of rhabdomyolysis with statin monotherapy, but duration of use made no difference to the event rate.
Table 1: Rhabdomyolysis cases, rate, and NNH for different lipid-lowering treatments
Lipid-lowering
drug |
Number
|
Person
years
|
Rhabdolyolysis
cases
|
Rate
per 10,000 patient
years (95% CI) |
One-year
NNH
|
| Atorvastatin, pravastatin, simvastatin | |||||
| Cerivastatin | |||||
| Gemfibrozil, fenofibrate |
Combined therapy
Combined use of statin and fibrate increased the risk of rhabdomyolysis. With atorvastatin or simvastatin plus a fibrate the incidence rate was between 17 and 23 per 10,000 person years, about 40 times higher than with statins alone. The NNH for one year of therapy with atorvastatin, pravastatin, or simvastatin plus a fibrate to produce one case of hospital admission for rhabdomyolysis was 1,670. For a patient aged 65 years or older with diabetes treated with both a statin and a fibrate the NNH was 480.
Combined use of cerivastatin plus gemfibrozil produced a rate of about 1000 per 10,000 person years, with an NNH of about 10.
Comment
The problems researchers face over rare but serious adverse events are clear here. Start with a quarter of a million patients and end up with only 24 actual events. Those 24 events include several drugs, at several doses, and given either separately or in combination. And while this is probably the best study of rhabdomyolysis with statins we have, even this has a problem. Seven cases occurred when patients were not on statins according to the prescription analysis, but were on statins according to the hospital records. Fortunately including or excluding the results made no difference.
Despite all these concerns we have some reasonably clear results. Rhabdomyolysis did not occur in the cohort when they were not taking statins or fibrates, as best we can judge. Rhabdomyolysis did occur when they were taking statins or fibrates. We have a reasonable estimate of how frequently the events occurred, both in single and combined use (Figure 1). This allows an estimate of risk of rhabdomyolysis, which is really extremely rare in people taking the most commonly prescribed statins.
Figure 1: NNH for rhabdomyolysis for different treatments, combinations, and conditions
With fibrates, or combined therapy, or people more at risk, or with certain combinations (now impossible because one of the drugs has been withdrawn), the risks are higher, or were even common. What we also know is that the risk of dying was even lower. In the 225,000 person years of therapy, even including the withdrawn cerivastatin, only 31 cases occurred, and only 3 of those died. As 10 of the cases were associated with cerivastatin, it would be reasonable to estimate that risk of death with statin, fibrate or combination with drugs commonly prescribed is of the order of 1 per 100,000 per year, and is probably less common than that.
And finally
For those interested in such things, another study examined the history of cerivastatin and rhabdomyolysis [3]. It points out that information was available to company and FDA associating cerivastatin monotherapy and increased rate of rhabdomyolysis some while before this was disseminated.
References:
- PD Thompson et al. Statin-associated myopathy. JAMA 2003 289: 1681-1690.
- DJ Graham et al. Incidence of hospitalized rhabdomyolysis in patients treated with lipid-lowering drugs. JAMA 2004 292: 2582-2590.
- BM Psaty et al. Potential for conflict of interest in the evaluation of suspected adverse drug reactions. JAMA 2004 292: 2622-2631.