气温与湿度的交互作用对呼吸系统疾病的影响

Interaction of temperature and relative humidity on emergency room visits for respiratory diseases

为评价平均气温、相对湿度及其交互作用对呼吸系统疾病急诊就诊人数的影响,采用广义相加模型（GAM）,在控制了时间长期趋势、＂星期几效应＂、节假日效应、空气污染等因素的影响后,分析2009-2011年北京市平均气温、相对湿度及其交互作用对呼吸系统疾病急诊就诊人数影响的暴露反应关系.结果显示,平均气温与呼吸系统疾病急诊就诊人数呈现近似U型的非线性关系,其作用临界点为12℃,当平均气温低于12℃时,气温每升高1℃,呼吸系统疾病急诊就诊人数减少2.26%（95%CI-2.43,-2.09）;当气温高于12℃时,气温每升高1℃,呼吸系统疾病急诊就诊人数增加0.92%（95%CI 0.72,1.11）.相对湿度与呼吸系统疾病的效应也呈现U型的分布特征,作用阈值为51%,当相对湿度≤51%时,相对湿度每增加10%,呼吸系统疾病急诊就诊人数减少3.43%（95%CI-3.47%,-3.38%）;当相对湿度〉51%时,其每增加10%呼吸系统疾病急诊就诊人数增加1.80%（95%CI 1.76%,1.85%）.平均气温对呼吸系统疾病的影响受相对湿度水平的调节.在低温环境下,相对湿度越小,气温对呼吸系统疾病的影响越显著,气温每升高1℃,呼吸系统疾病急诊就诊人数减少2.71%（95%CI-2.88,-2.53）;而高温环境下,当相对湿度较大时气温健康效应较强,即气温每升高1℃,呼吸系统疾病急诊就诊人数增加1.37%（95%CI 1.13,1.61）.

To quantitatively evaluate the effect of ambient temperature（AT）, relative humidity（RH）, and their interaction on emergency room visits（ERVs） for respiratory diseases in Beijing, a generalized additive model（GAM） was used to analyze the exposure-effect relationship between AT, RH and daily respiratory disease（ERVs） from 2009 to 2011 in Beijing, as well as their interaction effect on such visits. The model was considered with some potential confounding factors, such as long time trend, ＂day of week＂ effect, holiday effect, and air pollution. An obvious U-shaped pattern was found between temperature and daily respiratory disease（ERVs） with the optimum temperature threshold at 12℃. Below that optimum temperature threshold, a 1℃ increase was associated with a decrease of 2.26%（95%CI：-2.43,-2.09） for（ERVs）. Above that temperature threshold, a 1℃ increase was associated with an increase of 0.92%（95%CI： 0.72, 1.11）. A U-shaped pattern was also observed between RH and daily respiratory disease（ERVs） with the optimum RH threshold at 51%. Below that RH threshold, the（ERVs） increased by 3.43%（95%CI：-3.47%,-3.38%） for a 10% decrease. Above that RH threshold, the（ERVs） increased by 1.80%（95%CI： 1.76%, 1.85%） for a 10% increase. There was a synergistic effect of temperature and RH on respiratory diseases, which meant that the temperature effect differed by RH level. Below the temperature threshold, the temperature effect was stronger in lower RH levels, and the effect estimate per 1℃decrease in temperature was an 2.71%（95%CI：-2.88,-2.53） increase for respiratory disease（ERVs）. However, above the temperature threshold, the temperature effect was greater in higher humidity levels, and the effect estimate per 1℃increase in temperature was a 1.37%（95%CI： 1.13, 1.61） increase for respiratory disease（ERVs）.