Dalla Lana School of Public Health, University of Toronto, Ontario, Canada.
Am J Epidemiol. 2013 Jul 1;178(1):118-25. doi: 10.1093/aje/kws463. Epub 2013 May 9.
Seasonal variations in the incidence of pneumonia and influenza are associated with nosocomial Clostridium difficile infection (CDI) incidence, but the reasons why remain unclear. Our objective was to consider the impact of pneumonia and influenza timing and severity on CDI incidence. We conducted a retrospective cohort study using the US National Hospital Discharge Survey sample. Hospitalized patients with a diagnosis of CDI or pneumonia and influenza between 1993 and 2008 were identified from the National Hospital Discharge Survey data set. Poisson regression models of monthly CDI incidence were used to measure 1) the time lag between the annual pneumonia and influenza prevalence peak and the annual CDI incidence peak and 2) the lagged effect of pneumonia and influenza prevalence on CDI incidence. CDI was identified in 18,465 discharges (8.52 per 1,000 discharges). Peak pneumonia prevalence preceded peak CDI incidence by 9.14 weeks (95% confidence interval: 4.61, 13.67). A 1% increase in pneumonia prevalence was associated with a cumulative effect of 11.3% over a 6-month lag period (relative risk = 1.113, 95% confidence interval: 1.073, 1.153). Future research could seek to understand which mediating pathways, including changes in broad-spectrum antibiotic prescribing and hospital crowding, are most responsible for the associated changes in incidence.
肺炎和流感发病率的季节性变化与医院获得性艰难梭菌感染（CDI）的发病率有关，但原因尚不清楚。我们的目的是考虑肺炎和流感的时间和严重程度对 CDI 发病率的影响。我们使用美国国家住院调查样本进行了回顾性队列研究。从国家住院调查数据集确定了 1993 年至 2008 年期间患有 CDI 或肺炎和流感的住院患者。使用每月 CDI 发病率的泊松回归模型来测量 1）肺炎和流感流行高峰与 CDI 发病率高峰之间的时间滞后，以及 2）肺炎和流感流行对 CDI 发病率的滞后影响。在 18465 次出院中发现了 CDI（每 1000 次出院中有 8.52 次）。肺炎流行高峰先于 CDI 发病率高峰 9.14 周（95%置信区间：4.61，13.67）。肺炎流行率增加 1%，在 6 个月的滞后期内累积效应为 11.3%（相对风险=1.113，95%置信区间：1.073，1.153）。未来的研究可以寻求了解哪些中介途径，包括广谱抗生素处方和医院拥挤的变化，是导致发病率相关变化的主要原因。
