McDonnell W F, Muller K E, Bromberg P A, Shy C M
Clinical Research Branch, U.S. Environmental Protection Agency, Research Triangle Park, NC 27711.
Am Rev Respir Dis. 1993 Apr;147(4):818-25. doi: 10.1164/ajrccm/147.4.818.
The purpose of this study was to identify personal characteristics that predict individual differences in acute FEV1 response to ozone exposure. Response and predictor data were collected on 290 white male volunteers 18 to 32 yr of age who were each exposed to one of six concentrations of ozone between 0.0 and 0.40 part per million. The sample was divided into an exploratory sample of 96 and a confirmatory sample of 194 subjects. Exploratory analysis indicated that ozone, age, and several other variables explained a significant proportion of the variance in response. In the confirmatory sample, only age and ozone concentration predicted FEV1 decrement. For the combined sample ozone explained 31% of the variance, with age accounting for an additional 4%. The model predicted a decreasing response with increasing age for all nonzero ozone concentrations. For exposure to 0.40 ppm, the model predicts decrements in FEV1 of 1.07 and 0.47 L for 18- and 30-yr-old subjects, respectively. We concluded that for white male subjects age was a significant predictor of response, with older subjects being less responsive to ozone. Furthermore, we demonstrated that exploratory analysis without control of type I statistical error rates may result in apparent findings that cannot be replicated.
本研究的目的是确定能够预测个体在急性暴露于臭氧后第一秒用力呼气量（FEV1）反应差异的个人特征。对290名年龄在18至32岁之间的白人男性志愿者收集了反应数据和预测数据，他们每人暴露于百万分之0.0至0.40的六种臭氧浓度之一。样本被分为一个由96名受试者组成的探索性样本和一个由194名受试者组成的验证性样本。探索性分析表明，臭氧、年龄和其他几个变量解释了反应中很大一部分的方差。在验证性样本中，只有年龄和臭氧浓度能够预测FEV1的下降。对于合并样本，臭氧解释了31%的方差，年龄额外解释了4%。该模型预测，对于所有非零臭氧浓度，随着年龄的增加反应会降低。对于暴露于0.40 ppm的情况，该模型预测18岁和30岁受试者的FEV1分别下降1.07 L和0.47 L。我们得出结论，对于白人男性受试者，年龄是反应的一个重要预测因素，年龄较大的受试者对臭氧的反应较小。此外，我们证明了在不控制I型统计错误率的情况下进行探索性分析可能会导致无法重复的明显结果。
