Boston, MA - Adverse events associated with rosuvastatin (Crestor, AstraZeneca) appear to be higher than with the other available statins, according to a postmarketing analysis of adverse events reported to the FDA [1]. However, an accompanying editorial cautions that studies based on adverse-event reports are unreliable and that even if rosuvastatin does increase relative risk, the absolute number of events remains very low. The study, by Dr Alawi A Alsheikh-Ali (Tufts-New England Medical Center, Boston, MA) and colleagues, and the accompanying editorial, by Dr Scott Grundy (University of Texas Southwestern Medical Center, Dallas), were published online on May 23, 2005 in Circulation [2].

In the primary analysis of the study, for the period in which data were available for all the agents (October 1, 2003 through September 30, 2004), the composite rate of adverse events, which included rhabdomyolysis, proteinuria/nephropathy, or renal failure, was significantly higher (p<0.001) for rosuvastatin than the rates for simvastatin (Zocor, Merck), pravastatin (Pravachol, Bristol-Myers Squibb), and atorvastatin (Lipitor, Pfizer).

In a secondary analysis, the rate of adverse events during the first year of marketing for each drug was also found to be significantly higher for rosuvastatin than for pravastatin or atorvastatin (p=0.001), although the difference with simvastatin did not achieve significance (p=0.02). However, the rate of adverse events for rosuvastatin in this analysis was significantly lower than the rate for cerivastatin (Baycol, Bayer) (p=0.001), which was subsequently withdrawn from the market.

The Boston team note in their report that the adverse events associated with rosuvastatin generally occurred within 12 weeks, which was earlier in the course of events than the adverse events associated with the other statins. Although the majority of events were not fatal, most required hospitalization. Most of the adverse events in the rosuvastatin group occurred at a dose of 10 mg or less.

The senior author of the study, Dr Richard H Karas (Tufts-New England Medical Center), notes in an American Heart Association press release that although the relative risk of adverse events with rosuvastatin appears to be elevated, the absolute rate remains low, with 145 events out of 5.2 million prescriptions, or 1 event per 35 862 prescriptions.

In their report, the Boston investigators note that the advantages of a postmarketing analysis include a "real-life" population sample that is large enough to detect safety issues that can be missed during premarket studies. On the other hand, they note that analyzing adverse events has "intrinsic limitations." For instance, publicity about potential adverse events can influence the reporting rate. In the first years in which statins were on the market, as an example, there were few reports of rhabdomyolysis. It was only when this effect gained recognition that the adverse-event reports began to accumulate. Alsheikh-Ali and colleagues also note, "It remains unclear whether the observed rate of rosuvastatin-associated [adverse-event reports] is due in part to its greater LDL-C-lowering effect compared with the other statins."

In his accompanying editorial, Grundy notes that the problems associated with adverse-event reports was "exhaustively reexamined" in a March 2005 report from the FDA that considered a petition by Dr Sidney Wolfe (Public Citizen, Washington, DC) to remove rosuvastatin from the market. He quotes the conclusion of the FDA report: "All of the available evidence (including preclinical data, premarketing clinical studies, phase 4 clinical studies, and postmarketing adverse-event reports) indicates that Crestor (rosuvastatin) does not pose a risk of muscle toxicity greater than that of other approved statins. With respect to renal toxicity, there is no convincing evidence that Crestor poses a risk of serious renal injury."

In his editorial, Grundy questions whether adverse-event data should play an important role in the choice of statins. "Adverse-event reports can be useful for identifying signals of drug toxicity. They are much less useful for quantifying relative risk of different drugs of the same class," he writes.

In the AHA press statement, Grundy states that "it would be unfortunate for patients to quit taking statins for fear of side effects because that would increase their risk of heart attack." In the same press statement, Dr Alice Jacobs, president of the AHA, said that "the absolute risk with this statin is low. The overwhelming majority of people who are taking it will have no problem at all."

In an interview with heartwire, Dr Roger Blumenthal (Johns Hopkins University, Baltimore, MD) agreed that reports based on adverse events are inherently unreliable but recommended that "especially with rosuvastatin it makes sense to not go to the highest dose unless there's a clear indication."

In response to a question at an AHA press conference, Grundy said he would "be glad to use rosuvastatin" as a first-line agent in a setting where it was the only agent available on the formulary. Grundy said the choice of individual statins was far less important compared with the much more important task of reaching target cholesterol goals.

Just because you choose not to look doesn't mean you can conclude a drug is safe or effective.

A different view is offered by Dr Jonathan Sackner-Bernstein (North Shore University Hospital, Manhasset, NY), who has questioned the safety of nesiritide, a drug used in acute decompensated heart failure. In an interview with heartwire, he said that the new data on rosuvastatin demonstrate that "just because you choose not to look doesn't mean you can conclude a drug is safe or effective." He goes on to ask: "Should our professional organizations be the champions of the importance of proving safety and efficacy of this drug, as they should with all
others? . . . [O]ur obligation as practicing physicians is to resist prescribing unproven medicines when safe and effective alternatives exist."