Bayer’s Response to “Potential for Conflict of Interest in the Evaluation of Suspected Adverse Drug Reactions: Use of Cerivastatin and Risk of Rhabdomyolysis”

JAMA has invited this response from Bayer Corporation to the article authored by Psaty et al (the “Article”), which appears on page 2622 in this issue. Although the Article is presented as a scientific review of facts, sections are copied verbatim from a report submitted by the lead author in a US lawsuit in which he served as a paid expert for the plaintiff. Moreover, the Article’s conclusions are based upon selected exhibits from another lawsuit in which another author (Dr Ray) submitted a report for plaintiffs. Bayer won that court case (as well as the only other cerivastatin cases that have gone to trial). At least 2 of the authors continue to participate in lawsuits against Bayer.

This response presents a more complete picture of the information available in the public record. It is submitted on Bayer’s behalf by one of its outside counsel because Bayer views the Article as the publication of a disputed position taken in ongoing litigation.

Cerivastatin was an HMG-CoA reductase inhibitor shown to be highly effective in lowering serum LDL cholesterol. After extensive preclinical and clinical testing,^{1 - 5} the FDA determined that cerivastatin was safe and effective and approved cerivastatin for sale in the United States.⁶ Regulatory authorities in over 80 countries approved cerivastatin as safe and effective.⁷

At the time of cerivastatin’s US launch in January 1998, rhabdomyolysis was a well-known side effect of the statin class. All statin labeling carried rhabdomyolysis warnings. The medical literature was replete with reports of statin-associated rhabdomyolysis.^{8 - 9} Moreover, the scientific community recognized that coadministration of fibrates with statins increased the risk of myotoxicity.^{10 - 12}

Based on this class-wide experience, the original FDA-approved cerivastatin labeling warned of rhabdomyolysis and of a potential fibrate interaction, even though neither the independent clinical investigators nor the FDA reviewers ever characterized any event in the cerivastatin premarketing clinical trials as rhabdomyolysis.¹³ The Article uses a definition of rhabdomyolysis—“combination of CK elevations greater than 10 times the ULN with symptoms”—much broader than the proposed ACC/AHA Advisory definition (which requires renal impairment).¹⁴ But even using that definition, the original cerivastatin labeling specifically disclosed that such cases had occurred in <0.2% of clinical trial patients.¹¹ When higher doses became available, the labeling reflected that 0.4% of clinical trial patients developed this condition.¹⁵ No scientific evidence suggests that the risk of myopathy in cerivastatin patients was higher than that disclosed in the label. Thus, at all times that cerivastatin was marketed, both the nature and the magnitude of the myopathy risk were accurately set forth in the labeling.

Postmarketing, Bayer actively monitored the safety of cerivastatin through phase 4 clinical trials and by collecting and analyzing spontaneous adverse event reports (SAERs). Neither FDA nor the Article criticized Bayer’s diligence and good faith in reporting SAERs. Indeed, FDA has audited Bayer’s compliance with adverse event reporting covering the period of cerivastatin’s marketing and found no deficiencies.¹⁶

After receiving SAERs of rhabdomyolysis, Bayer—without FDA prompting—amended the labeled warnings regarding myopathy 5 times.¹⁷ The December 1999 label change specifically contraindicated coadministration with gemfibrozil in the United States.¹⁸ Another label change in May 2001 emphasized that untitrated use of the highest cerivastatin dose could lead to increased risk of rhabdomyolysis.¹⁹

By August 2001, Bayer and FDA agreed that SAERs of rhabdomyolysis continued, particularly with gemfibrozil coprescription and untitrated use, despite the label changes.²⁰ Accordingly, Bayer voluntarily suspended marketing cerivastatin.²¹

FDA Effectively Monitors the Safety of Medicines Postmarketing. Using cerivastatin as a supposed example, the Article’s central thesis is that FDA lacks the resources to monitor medicines effectively postmarketing.

This Article’s premise, however, is demonstrably untrue. FDA, in fact, did carefully monitor the safety of cerivastatin. In June 2000, FDA completed a detailed comparison of SAER reporting rates for statin drugs with respect to rhabdomyolysis.²² FDA found that cerivastatin had the highest reporting rate for rhabdomyolysis of any of the statins on the market. To understand this observation, FDA considered these data from several different perspectives.²³ FDA noted the high rate of rhabdomyolysis when cerivastatin was combined with gemfibrozil, commenting that “the voluntary upgrade of the labeling by the manufacturer for cerivastatin from a Warning to a Contraindication regarding combination use with gemfibrozil is consistent with our findings in the AERS database.”²³

FDA observed that “reporting of adverse events to FDA’s spontaneous reporting system is highest in the first few years after a new drug is approved.”²³ Given this known bias, “reporting rates were also calculated for only the first two years of marketing for each statin, so as to make comparison with cerivastatin (the newest statin) more balanced.”²³ Although reporting rates for monotherapy were much more comparable to other statins, the reporting rate for combination use remained higher for cerivastatin.²³

FDA acknowledged that “several major limitations . . . must be borne in mind when examining and interpreting reporting rates. . . . ”²³ FDA stated that “perhaps most importantly, the reporting rates presented in this document are NOT incidence rates and are in no way reflective of the actual occurrence of rhabdomyolysis among users of statins and/or fibrates out in the ‘real world’ of clinical practice.”²³

Although the Article apparently disagrees with FDA’s experts on the limitations of reporting rates, the contention that FDA lacked resources to adequately evaluate the postmarketing data on cerivastatin is specious. FDA’s internal memorandum reflects a careful analysis that is consistent with positions of other scientists experienced in interpreting SAER data²⁴ and with Bayer’s conduct.

FDA’s analysis was not limited to fatal rhabdomyolysis cases, as the Article suggests, since no fatal rhabdomyolysis cases were reported in the United States during cerivastatin’s first 2 years of marketing.

The Gemfibrozil Interaction. The Article’s criticism of Bayer for not contraindicating gemfibrozil in the United States until December 1999 and for not undertaking an earlier pharmacokinetic interaction study with gemfibrozil employs hindsight analysis.

Statin-fibrate drug interactions had been known to the scientific community for several years prior to cerivastatin’s launch.^{10 ,25 - 29} Numerous studies attempting to elucidate the interaction between statins and fibrates consistently showed the absence of a pharmacokinetic interaction.^{30 - 31}

Thus Bayer, like other researchers,³² hypothesized a possible pharmacodynamic mechanism and tested that hypothesis in vitro. Ultimately, this testing showed no pharmacodynamic interaction between Baycol and gemfibrozil.⁷

A pharmacokinetic interaction between any statin and gemfibrozil first was demonstrated in August 2000 with simvastatin.³³ Based on this research, Bayer undertook its own clinical interaction study, which similarly revealed a pharmacokinetic interaction when cerivastatin was coadministered with gemfibrozil.³⁴ Later studies in 2001 with lovastatin demonstrated that pharmacokinetic interactions with gemfibrozil were not unique to cerivastatin among the statin class.³⁵

Bayer Responded to the Postmarketing Data. The Article quotes out-of-context an internal company document discussing postmarketing data. That memorandum actually documents the scientific unreliability of such data and recommended that Bayer undertake a reliable epidemiological study.³⁶

Thereafter, Bayer retained the former chief FDA pharmacoepidemiologist and undertook a comprehensive search for a viable epidemiological database.³⁷ A retrospective study was completed that confirmed an increased myopathy risk with gemfibrozil combination therapy but no difference in the monotherapy myopathy rate.³⁸

Bayer Complied With Its Disclosure Obligations. The Article implies that Bayer should have provided physicians with comparative SAER reporting rates. However, FDA regulations prohibit manufacturers from making comparisons to other medicines based on anything other than data from well-controlled clinical trials.³⁹ Indeed, even when placebo-controlled clinical trials are available for each medication individually, FDA practice precludes comparative statements unless such statements are based on head-to-head comparisons, based upon FDA’s long-standing belief that well-controlled head-to-head clinical trials provide the only reliable basis upon which to base drug comparisons.

It always is possible to second guess decision-making after the fact. However, when judged fairly by all the facts in their proper context, Bayer’s conduct in the marketing of cerivastatin from 1997 until its voluntary withdrawal from the market in August 2001 was responsible, appropriate, and consistently motivated by concern about the safety and welfare of patients.