On March 27, 2001, just 19 months after it had been approved, rapacuronium bromide (Raplon, Organon Inc, West Orange, NJ) was voluntarily withdrawn from the market by the manufacturer. The withdrawal from the market of a previously approved drug is not a novel event. Since 1979, the annual rate of drugs withdrawn from the market has been approximately 3%. In the case of rapacuronium bromide, postmarketing surveillance, largely consisting of anecdotal case reports, suggested that the risk of injury from bronchospasm was more than that suggested in premarket clinical trials. The unique and important aspects of the report by Rajchert et al. (1) in this issue of Anesthesia & Analgesia are the methods used. These techniques move beyond the anecdotal to arrive in the quantitative realm of evidence-based medicine. The approaches used afford not only an estimate of the incidence of the adverse event and its predisposing factors but also an approximation of the risk relative to that of alternative muscle relaxants.
Automated anesthesia information management systems are computers that capture perioperative information in a digital format and permit use of this information to enhance patient management. In addition to serving an important function in departmental quality assurance, these systems provide the opportunity to link specific intraoperative events (i.e., bronchospasm) to potential risk factors (i.e., use of rapacuronium). The authors of this report provide a simple and elegant example of the power of information management in the perioperative environment. These kinds of investigations are now increasingly feasible with anesthesia information management systems. Despite the awesome capabilities of automated information management systems, current estimates indicate that they are used by only 1% of all anesthesiology departments in the United States.
The authors queried their computerized anesthesia database to perform two types of studies: a historical cohort study and a case control study. Cohort studies are observational trials in which patients with and without a risk factor or exposure are identified and then followed forward in time to see if they develop the outcome of interest. Cohort studies begin with a risk factor (such as exposure to rapacuronium) and examine the consequences. Risk is defined as the probability of developing a condition over a specified time period. Mathematically, risk is simply the number of individuals who develop the condition divided by the total number of individuals who were candidates to develop the condition. The strength of the association between the risk factor and the outcome is called the relative risk. Relative risk measures the probability of bronchospasm if rapacuronium is used divided by the probability if rapacuronium is not used. A relative risk value more than one implies that the use of rapacuronium increases the risk of bronchospasm.
Unfortunately, the authors did not provide the information necessary to calculate the relative risk for developing bronchospasm when rapacuronium was administered versus an alternative muscle relaxant. However, they did the next best thing: they performed a case control study.
Case control studies are useful when the event under study is rare and it is difficult to assemble a large cohort. Case control studies begin with an outcome (such as bronchospasm) and look backwards over time to detect possible associations or risk factors for the development of the outcome. The previous exposures of both the cases and controls are then analyzed in an attempt to identify an exposure present in the cases but absent in the control patients. The strength of the association between the outcome and the risk factor in a case control study is expressed as an odds ratio. The odds ratio measures the odds of having the risk factor (rapacuronium) if the condition (bronchospasm) is present divided by the odds of having the risk factor if the condition is absent. The odds ratio of 10.1 in the case control study means that the odds of having received rapacuronium are increased 10 times in patients with bronchospasm. The odds ratio closely approximates the relative risk when the event is rare such as bronchospasm in an anesthetized patient.
A limitation common to both case control and retrospective cohort studies is bias. Bias occurs when factors that influence the results or outcome are introduced. The study and control groups may differ from each other by a factor that is likely to affect the results. This factor is often referred to as a confounding variable. The statistical method of multivariate analysis adjusts for confounding variables. Rajchert et al. (1) used univariate analyses that do not allow for adjustment. A multivariate logistic regression model would have provided more power to the analysis. The limited number of events in this study precluded the authors’ ability to perform a multivariate analysis.
Nonetheless, the study by Rajchert et al. (1) provides compelling evidence above and beyond that which has been previously reported. Previous communications of bronchospasm after rapacuronium in children have been case reports. Case reports, although useful in a heuristic sense, provide little information on the frequency of these events or their associations with predisposing factors. Case series data are the weakest link in the hierarchy of medical evidence. Cohort and case control studies offer more power for identifying risk factors for disease, making statistical inferences, and testing hypotheses.
Premarket drug evaluation includes phase I, II, and III studies that typically limit drug exposures to a few thousand patients. These studies seek to determine the safety and efficacy (performance under optimal conditions of the clinical trial) rather than effectiveness (performance in uncontrolled “real-life” situations) of a new therapeutic drug and are not large enough to detect rare but serious complications. Premarket clinical trials of rapacuronium reported a 3.2% incidence of bronchospasm.
Phase IV begins once a new drug application is approved and the drug is on the market. The Food and Drug Administration (FDA) then relies on its Adverse Event Reporting System. This depends upon voluntary reporting by physicians, pharmacists, pharmaceutical companies and patients through facsimile, telephone, or MedWatch reports. These adverse event reports are triaged to safety evaluators in the Office of Post-Marketing Drug Risk Assessment to determine if a signal is present. If there appears to be a significant signal, the matter is referred to the division that originally approved the drug.
This information is then provided to the pharmaceutical company and they are instructed to issue either a “Dear Doctor” letter or to revise the package insert to include a boxed warning. Two recent pediatric examples are the “Dear Doctor” letter concerning propofol for pediatric intensive care unit sedation and the boxed warning on the package insert for succinylcholine regarding its routine use in children. The third alternative is to voluntarily withdraw the drug from the market, as was done with rapacuronium.
The FDA acted swiftly and responsibly, thanks largely to the efforts of those individuals who voluntarily reported their adverse experiences. The FDA has come to recognize the power of computerized patient record systems to augment the voluntary reporting system. It has just announced contracts that afford the agency access to three large commercial databases. The goal is to enable the regulatory body to investigate the public health significance of reported adverse events. Perhaps a consortium of institutions with automated anesthesia record system capability could cooperatively pool their resources to augment this effort. Meanwhile, the search for a fast-acting, short duration, nondepolarizing muscle relaxant continues.