黑碳对呼吸系统发病急性影响及气温修正效应研究

为研究黑碳(BC)对呼吸系统急性发病的影响及气温的修正效应,收集北京市2009~2012年264075例呼吸系统急诊病例与同期空气污染物(BC、PM_2.5、SO₂、NO₂)及气象数据,在划分呼吸道感染部位(上、下呼吸道)与人群年龄的基础上,采用分布-滞后非线性模型与广义相加模型进行建模.首先分别研究BC、气温与发病的(滞后)关联,继而构造二元交互模型探索气温-BC的协同关系,再分层量化BC在不同气温水平的健康影响;并同时纳入气态污染物验证BC结果的稳健性.结果表明,对总呼吸系统、上感、下感而言,气温-发病风险的暴露-响应曲线均近似“V”型,阈值温度分别为24℃、26℃和24℃,且低温的滞后累积影响强于高温.主效应模型揭示BC诱发即时性风险,影响在3d内消失;BC浓度每升高四分位数(IQR),总呼吸系统、上感、下感的超额发病风险(ER)分别为1.97%、2.64%和1.34%.少儿(£14岁)超额发病风险最高(总呼吸系统,3.40%),而老年组(³60岁)结果不甚显著.双污染物模型显示,BC与SO₂共存会放大BC关联风险,尤以上感响应明显;而BC与NO₂共存会适度增强下感风险.BC-气温的非参数二元模型显示,BC升高使发病风险类似对数函数上升,且高温会显著增强BC的健康影响.分层模型得到,每IQR BC在气温高于阈值时导致的下感风险显著高于上感,分别为5.55%、1.27%(P>0.05);而低于阈值时BC所致上、下感风险相当,均在0.55%左右.BC对呼吸系统发病的急性影响与感染部位和气温水平紧密相关,不同年龄段间也体现差异化特征.

Acute effects of black carbon on respiratory diseases and the modulating effect of air temperature

The distributed-lag nonlinear models (DLNMs) and generalized additive models (GAMs) were adopted to reveal the adverse health effects of black carbon (BC) on respiratory emergency room (ER) visits, as well as the potential modifying effect of air temperature. Daily ER visits of respiratory diseases during 2009~2012 in Beijing, daily air pollutants (including BC, PM_2.5, SO₂, and NO₂), and meteorological data covering that period were collected. Moreover, medical data was divided into three sub-groups, including the total respiratory diseases, upper and lower respiratory infections (URI and LRI), respectively. Firstly, we explored the (lag) associations between BC, air temperature and morbidity separately. Secondly, binary interaction models were constructed to explore the synergistic relationship between temperature and BC. Afterwards the effects of BC at varied temperature levels were quantified by hierarchical models. In addition, certain gaseous pollutant (SO₂ or NO₂) was also included to verify the robustness of basic models. The exposure-response relationships between air temperature and respiratory diseases exhibited an overall "V" type, the thresholds were 24℃, 26℃, and 24℃, respectively for total respiratory diseases, URI, and LRI. The cumulative effects of low temperature were stronger than that of high temperature. BC effect was immediately occurred and lagged for a short-term (within 3days). For each interquartile range (IQR) increase in BC concentration, the corresponding excess risk (ER) of total respiratory disease, URI, and LRI were 1.97%, 2.64%, and 1.34%, respectively. Children (£14years) had the highest excess risk (e.g., total respiratory system, ER=3.40%), while the results were not significant in the elderly group (³60years).Further, the coexistence of BC and SO₂ may amplify BC effect on respiratory diseases, especially for URI, while the coexistence with NO₂ mightincrease LRI risk. The nonparametric bivariate-response models showed that, the respiratory risk related to BC increased in a logarithmic manner, and high temperature enhanced BC effect strongly. When temperature exceeding its thresholds, an IQR increments in BC associated with5.55% and 1.27% (P>0.05) increase of URI and LRI morbidity,respectively.Whereas the ERs were both approximated 0.55% for both URI and LRI under temperatures lower than the thresholds. In a word, the acute effects of BC on respiratory disease were closely related to infection sites and air temperature levels, differentiations among agegroups were also revealed.