Daneman Nick, Low Donald E, McGeer Alison, Green Karen A, Fisman David N
Institute for Clinical Evaluative Sciences, Sunnybrook Health Sciences Centre, USA.
Clin Infect Dis. 2008 Apr 15;46(8):1131-8. doi: 10.1086/529440.
Community-acquired pneumonia caused by Streptococcus pneumoniae is a major source of morbidity and mortality. Macrolide antibiotics are recommended as empirical first-line therapy for patients with community-acquired pneumonia. Guidelines suggest a 25% rate of high-level macrolide resistance in the community as the threshold beyond which macrolides should not be used. We evaluated the implications of this threshold for clinical failure rates.
We developed a theoretical model linking the prevalence of macrolide resistance to patient outcomes, based on the epidemiological concept of risk difference. We estimated the risk of clinical failure as a function of the likelihood and impact of discordant therapy and of the probability of clinical failure even in the presence of optimal therapy. The model was parameterized on the basis of the best available data derived from the published medical literature, and clinical failures were valued monetarily using an expected net benefit approach.
Under the proposed 25% resistance threshold, the risk difference for such therapy would be 1.2% (95% credible interval, 0.5%-3.1%) for death, 1.6% (95% credible interval, 0.5%-3.2%) for bacteremia, and 3.3% (95% credible interval, 1.1%-5.7%) for prolonged clinical course; excess risks of death were valued at >$10,000 per empirical treatment of community-acquired pneumonia and were further elevated in high-risk populations. Excluding low-level resistance resulted in a 4-fold underestimation of projected risks.
A 25% resistance threshold that fails to consider low-level resistance will result in high excess rates of morbidity and mortality because of discordant therapy. Whether projected failure rates are classified as unacceptable is an important health policy question, because risk of clinical failure needs to be weighed against other considerations.
肺炎链球菌引起的社区获得性肺炎是发病和死亡的主要原因。大环内酯类抗生素被推荐作为社区获得性肺炎患者的经验性一线治疗药物。指南建议将社区中高水平大环内酯类耐药率25%作为不应使用大环内酯类药物的阈值。我们评估了该阈值对临床失败率的影响。
基于风险差异的流行病学概念，我们建立了一个将大环内酯类耐药率与患者预后相关联的理论模型。我们将临床失败风险估计为不恰当治疗的可能性和影响以及即使在最佳治疗情况下临床失败概率的函数。该模型根据已发表医学文献中的最佳可用数据进行参数化，并使用预期净效益方法对临床失败进行货币估值。
在所提议的25%耐药阈值下，此类治疗的死亡风险差异为1.2%（95%可信区间，0.5%-3.1%），菌血症风险差异为1.6%（95%可信区间，0.5%-3.2%），临床病程延长风险差异为3.3%（95%可信区间，1.1%-5.7%）；每例社区获得性肺炎经验性治疗的额外死亡风险估值超过10,000美元，在高危人群中进一步升高。排除低水平耐药会导致对预计风险的低估达4倍。
未考虑低水平耐药的25%耐药阈值将因不恰当治疗导致高发病率和死亡率。预计失败率是否被归类为不可接受是一个重要的卫生政策问题，因为临床失败风险需要与其他因素权衡。
