Sawicki Peter T, Bender Ralf, Selke Gisbert W, Klauber Jürgen, Gutschmidt Sven
Institut für Qualität und Wirtschaftlichkeit im Gesundheitswesen (IQWiG), Köln.
Med Klin (Munich). 2006 Mar 15;101(3):191-7. doi: 10.1007/s00063-006-1044-6.
After the recall of rofecoxib (Vioxx), there was repeated national and international discussion on the potential number of patients harmed by causally related cardio- and cerebrovascular events. In individual cases, it cannot be determined whether a myocardial infarction or stroke that occurred during rofecoxib therapy was actually directly caused by this drug. On the basis of the results of the Vioxx Gastrointestinal Outcomes Research (VIGOR) trial and German prescription data provided by the Scientific Institute of the Local Health Care Fund, the authors therefore conservatively estimated the number of patients harmed by rofecoxib in Germany between 2001 and 2004.
Under simplifying assumptions that, as in the VIGOR study, the risk of rofecoxib or naproxen therapy can be described by a Cox model with exponentially distributed event times, it is possible to calculate the daily risk of cardio- and cerebrovascular events in patients treated with these drugs. The estimated number of patients experiencing cardio- and cerebrovascular events under rofecoxib or naproxen therapy can be calculated by multiplying the daily risks by the defined daily doses prescribed in Germany. The difference between these numbers produces the estimated number of patients harmed by rofecoxib.
On the basis of the data pool, a total of 7,092 additional diseased or deceased patients due to rofecoxib therapy were estimated (95% confidence interval: 2,004-15,416). The simplifying assumptions made together with the underreporting of events in the VIGOR trial are more likely to lead to an underestimation than an overestimation of affected patients. When assessing the benefit-harm ratio of rofecoxib, it needs to be considered that its protective gastrointestinal effects were not assessed compared with the optimum long-term therapy. It can be assumed that a comparison of rofecoxib with a combination of nonsteroidal anti-inflammatory drugs (NSAIDs) and gastric mucosal barrier protectors (e. g., misoprostol) would not have shown an advantage in favor of rofecoxib therapy.
The example of rofecoxib and the relatively high number of patients harmed by it in Germany indicate that, before widely prescribing a new drug, a more thorough assessment of the benefit-harm ratio of the drug is required as well as a stronger consideration of therapeutic alternatives and a timely conduct of meaningful clinical studies. The results of these studies should be promptly communicated in full to physicians and patients.
罗非昔布（万络）撤市后，国内外多次就因心血管和脑血管事件受伤害患者的潜在数量展开讨论。在个别案例中，无法确定罗非昔布治疗期间发生的心肌梗死或中风是否确实由该药物直接导致。因此，基于罗非昔布胃肠道转归研究（VIGOR）试验结果以及地方医保基金科学研究所提供的德国处方数据，作者保守估计了2001年至2004年间德国因罗非昔布而受伤害的患者数量。
在简化假设下，如同VIGOR研究那样，罗非昔布或萘普生治疗的风险可用事件时间呈指数分布的Cox模型来描述，由此能够计算接受这些药物治疗患者发生心血管和脑血管事件的每日风险。罗非昔布或萘普生治疗下发生心血管和脑血管事件的估计患者数量，可通过将每日风险乘以德国规定的每日剂量来计算。这两个数字之差即得出因罗非昔布而受伤害的估计患者数量。
基于该数据库，估计因罗非昔布治疗而另外患病或死亡的患者共有7092例（95%置信区间：2004 - 15416）。VIGOR试验中所做的简化假设以及事件报告不足，更有可能导致对受影响患者的低估而非高估。在评估罗非昔布的利弊比时，需要考虑的是，与最佳长期治疗相比，其对胃肠道的保护作用未得到评估。可以假定，罗非昔布与非甾体抗炎药（NSAIDs）和胃黏膜屏障保护剂（如米索前列醇）联合使用的比较不会显示出罗非昔布治疗的优势。
罗非昔布的例子以及德国因该药受伤害患者数量相对较多表明，在广泛开具一种新药之前，需要对药物的利弊比进行更全面的评估，更多地考虑治疗替代方案，并及时开展有意义的临床研究。这些研究结果应及时完整地传达给医生和患者。
