In light of the attention that the article on the ADVANTAGE (Assessment of Differences between Vioxx and Naproxen To Ascertain Gastrointestinal Tolerability and Effectiveness) study (1) has received, we thought it would be useful to provide additional information clarifying the specific vascular events that contributed to the aggregate end points reported and the process by which individual cases were assigned to those end points.

Lisse and colleagues (1) reported cardiovascular events in the ADVANTAGE study using 2 aggregate end points: confirmed thrombotic cardiovascular serious adverse experiences and the Antiplatelet Trialists' Collaboration (APTC) combined end point. The former included reports of myocardial infarction, unstable angina, sudden cardiac death, ischemic stroke, transient ischemic attack, peripheral arterial thrombosis, peripheral venous thrombosis, and pulmonary embolism that were confirmed by adjudication according to a prespecified program-wide standard operating procedure. This procedure was initiated in 1998 before the start of the ADVANTAGE trial (2). The APTC combined end point included cardiovascular and hemorrhagic deaths and deaths of unknown cause, nonfatal myocardial ischemia, and nonfatal stroke (3) and was the prespecified end point in the pooled analyses of cardiovascular events with rofecoxib (4, 5).

All investigator reports of potential cardiac, cerebrovascular, or peripheral arterial or venous thrombotic events that used one of a set of prespecified adverse experience terms were prospectively adjudicated by external blinded panels of medical specialists for inclusion in the confirmed thrombotic and APTC end points. In addition, all deaths that did not meet criteria for this external adjudication were prospectively reviewed by a Merck cardiologist blinded to treatment assignment to determine whether the death met the broader APTC end point criteria. Tables 1 and 2 show the confirmed thrombotic cardiovascular serious adverse experiences and APTC combined end point events, respectively, from the ADVANTAGE trial as reported by Lisse and colleagues (1).

Recent media reports have focused on the U.S. Food and Drug Administration's (FDA's) medical review of unblinded data from this study, in which the FDA reviewer commented that the death of 1 patient in the rofecoxib group, which the investigator had listed as from "hypertensive heart disease" on the basis of autopsy findings, was, in the FDA reviewer's opinion, a case of sudden death. Although the term retrospectively proposed by the FDA reviewer would have met criteria as a potential thrombotic event eligible for adjudication if used by the investigator, the term hypertensive heart disease did not trigger adjudication in the existing standard operating procedure. Therefore, this case was not prospectively adjudicated and is not included as a confirmed thrombotic event in Table 1. However, on the basis of internal blinded review, it was determined prospectively that this patient's death met the criteria of the APTC combined end point, and as shown in Table 2, this patient's death was included in the combined APTC end point in the article by Lisse and colleagues (1). It was also included in the pooled analyses of cardiovascular events with rofecoxib published by Konstam and associates (4) and Weir and coworkers (5).