Abstract
目的
分析不同季节不同度量方式的臭氧(ozone, O3)浓度对儿童哮喘门诊就诊人次的影响,探讨该影响在不同类型人群中的差异。
方法
收集兰州市三所三级甲等综合医院2014年1月1日至2017年12月31日儿童哮喘门诊的就诊资料,收集同期空气污染数据和气象数据,考虑到O3浓度与儿童哮喘门诊就诊人次及气象因素之间的非线性关系,利用广义相加时间序列模型分析O3浓度变化与儿童哮喘门诊就诊人次的短期关联性,同时进一步对性别、年龄进行分层分析,考虑到O3在1 d内浓度变化情况,同时采用每日1 h最大值浓度(O3max1h)、每日8 h最大值浓度(O38h)以及每日24 h均值浓度(O324h)这三种O3暴露的不同度量方式作为O3短期暴露指标,并分别开展模型分析。
结果
夏季时,O3短期暴露水平的升高对于儿童哮喘门诊就诊人次的增加有显著影响。在滞后当天(lag0)的情况下,大气O3max1h浓度每增加10 μg/m3,儿童哮喘门诊就诊人次增加3.351% (95%CI:1.231%~5.516%);O38h浓度每增加10 μg/m3,儿童哮喘门诊就诊人次增加3.320% (95%CI:0.197%~3.829%);O324h浓度每增加10 μg/m3,夏季儿童哮喘门诊就诊人次增加6.600% (95%CI:0.914%~12.607%);此外,O3max1h暴露增加也会引起0~5岁儿童及男性哮喘门诊就诊人次的显著增加。
结论
兰州市夏季时O3短期暴露水平的升高对于儿童哮喘门诊就诊风险的增加有显著影响,O3max1h与儿童哮喘门诊就诊人次增加的关联更密切。
Keywords: 臭氧, 空气污染物, 儿童, 哮喘, 门诊病人
Abstract
Objective
To analyze the effects of ozone (O3) concentrations measured with different approaches across different seasons on the total number of childhood asthma-related clinical visits, as well as the differentiation of such effects across different groups of patients.
Methods
The outpatient data of three grade A tertiary hospitals in Lanzhou City spanning from 1 January 2014 to 31 December 2017, as well as air pollution and meteorological data during the same period were collected. Considering the nonlinear relationship between O3 concentrations and the total number of childhood asthma-related clinical visits and meteorological factors, a generalized additive temporal sequence model was employed to analyze the short-term association between changes in O3 concentrations and the total number of childhood asthma-related clinical visits. Taking into account of the variations in O3 concentrations within 1 day, this study adopted different measurement approaches to address the three types of O3 exposures, namely, the maximum 1 h daily concentration (O3max1h), the maximum 8 h daily concentration (O38h) and the mean 24 h daily concentration (O324h) as the short term exposure indicators to O3, followed by a model-based analysis.
Results
The increase in short-term exposure levels to O3 in summer had a significant effect on the increase in the total number of childhood asthma-related clinical visits. With lag0 for the current day, every 10 μg/m3 increase in atmospheric concentration of O3max1h was associated with an increase in the total number of childhood asthma-related clinical visits by 3.351% (95%CI: 1.231%-5.516%); for every 10 μg/m3 increase in O38h concentration, the total number of childhood asthma-related clinical visits increased by 3.320% (95%CI: 0.197%-3.829%); for every 10 μg/m3increase in O324h concentration, the total number of childhood asthma-related clinical visits in summer increased by 6.600% (95%CI: 0.914%-12.607%); moreover, an increase in exposure to O3max1h also led to a significant rise in the total number of childhood asthma-related clinical visits among the males.
Conclusion
The increase in short-term exposure levels to O3 in summer in Lanzhou City has a significant effect on the increase in the total number of childhood asthma-related clinical visits; O3max1h is more closely correlated with the increase in the total number of childhood asthma-related clinical visits.
Keywords: Ozone, Air pollutants, Child, Asthma, Outpatients
由于燃煤使用量、机动车数量等不断增加,我国的空气污染模式已由传统的煤烟型污染向煤烟型和机动车尾气型混合污染转变。臭氧(ozone, O3)作为一种二次污染物,随着全球气候变暖和人类活动排放的氮氧化合物增多,已经成为我国部分城市夏季的首要污染物[1],其与人群健康的关系也逐渐受到公众关注。多项研究发现,O3短期暴露可增加哮喘、支气管炎等呼吸系统疾病的发病风险。儿童的每千克体重氧耗量(oxygen consumption per kg body weight)高于成人,且呼吸系统和免疫系统发育尚不完善,使其更易遭受空气污染侵害[2],易出现咳嗽、喘息、哮喘、支气管炎等呼吸系统症状和疾患。目前,我国关于O3暴露与儿童哮喘门诊量的定量分析研究相对较少,仅限于合肥[3]、重庆[4]、香港特别行政区[5]等少数城市,且已有的研究大多采用单独度量方式的O3浓度,但是,考虑到O3浓度的日变化和季节变化特征以及O3前体物释放的变异性,不同度量形式的O3浓度可能受到地区、季节、天气、城市化、人口密度等因素影响,单独采用某一度量方式的O3浓度指标而不考虑其他度量方式下的浓度变化,很难全面获取O3对人群健康影响的准确结果,故在实际研究中应尽可能给出多种度量标准下的效应值。O3浓度监测一般采取3种度量方式,包括O3每日1 h最大值浓度(O3max1h)、每日8 h最大值浓度(O38h)以及每日24 h均值浓度(O324h)。近年来,兰州市高度重视大气污染防治工作,空气质量明显改善,但O3污染却日益凸显,成为兰州市夏秋季节的主要污染物之一,因此本研究采用时间序列广义相加模型及人群分层的分析方法,收集兰州市三所医院儿童哮喘门诊的就诊资料,采用多种度量方式的O3浓度指标,分析兰州市O3浓度暴露对儿童哮喘门诊就诊人次的影响。
1. 资料与方法
1.1. 数据来源
儿童每日呼吸系统疾病门诊病历资料来源于兰州市三所三级甲等综合医院的电子就诊记录,资料选取的时间段为2014年1月1日至2017年12月31日,由医院信息系统导出,依据患儿的住址筛选兰州市本地门诊就诊患者作为研究对象,根据第10版《国际疾病分类》(International Classification of Disease Version 10th,ICD-10)的编码筛选出哮喘。
同期兰州市大气污染物监测资料来源于全国城市空气质量实时发布平台(http://106.37.208.233:20035/),选取兰州市逐日监测数据,包括细颗粒物(PM2.5)、可吸入颗粒物(PM10)、二氧化硫(SO2)、二氧化氮(NO2)和一氧化碳(CO)。为获取与对比不同O3度量方式的健康效应,我们同时从全国城市空气质量实时发布平台收集2014—2017年的O3每小时浓度数据。O3的每小时浓度值是兰州市区4个空气质量监测点的平均浓度,反映了兰州市区的空气污染整体情况,具有较好的代表性。
采用O3监测值作为O3暴露浓度的流行病学研究中,使用多种度量方式评估其健康效应,包括O3max1h、O38h (包括滚动8 h和固定8 h)和O324h[6-7]。兰州市O3浓度从11:00开始呈现出逐渐上升趋势,午后15:00~16:00达到峰值,16:00后随着光照和紫外线辐射减弱,O3浓度逐渐降低[8]。我们参考以往相关研究报道[9-10],同时考虑兰州市O3污染特点,O38h采用固定时间段(11:00~18:59)O3每小时平均浓度8个值的平均值。同期的每日气象数据由甘肃省气象局提供,包括气温、相对湿度。
1.2. 统计学分析
哮喘发病是小概率事件,每日哮喘门诊就诊人次近似服从Poisson分布。采用基于Poisson回归的广义相加模型(generalized additive models,GAM)分析O3暴露与哮喘门诊量的关联,在模型中采用自然样条平滑函数控制时间序列的长期趋势、气温、湿度、星期几效应、节假日效应等混杂因素的影响,并根据赤池信息量准则(Akaikes information criterion,AIC)确定混杂因素的自由度[11]。为了评估不同季节时O3的差异,本研究根据兰州市气候特征,同时结合已有文献对四季的划分标准[12],将全年分为春季(3—5月)、夏季(6—8月)、秋季(9—11月)、冬季(12—2月),采用分层分析的方法分别将其纳入模型进行统计分析,基本模型表达为:Log(E(Yt))=α+βZt+ns(Timet, df)+ns(Temt, df)+ns(RHt, df)+as.factor(dow)+as.factor(holiday),式中,Yt为第t日哮喘门诊就诊人次,E(Yt)为第t日的门诊就诊人次期望值,α为该模型的截距,Zt为第t日的O3污染物浓度,β为回归系数,ns为自然样条平滑函数,Timet为时间变量,Temt为第t日平均温度,RHt为第t日平均湿度,dow为星期几效应,holiday为节假日效应。根据赤池信息量准则,最终纳入模型的时间、温度、湿度的自由度分别7、3、3。同时,计算O3浓度每增加10 μg/m3所致儿童哮喘门诊就诊人次的超额危险度(excess risk,ER),公式为:ER=[EXP(β×10)-1]×100%。考虑大气污染作用的滞后效应,参考现有相关研究[3-5],纳入污染当天,滞后1、2、3、4、5、6、7 d(lag0、lagl、lag2、lag3、lag4、lag5、lag6、lag7),计算O3的单日滞后效应,以及1~7 d多日移动平均滞后效应(lag01、lag02、lag03、lag04、lag05、lag06、lag07)。
此外,针对O3max1h、O38h和O324h暴露的健康效应进行敏感性分析,采用对应季节O3max1h、O38h和O324h的最显著效应值,通过构建双污染物和多污染物模型和改变时间变量自由度(df=6~10)来评价模型结果的稳定性。采用R3.4.4软件mgcv软件包进行分析,检验水准为0.05。
2. 结果
2.1. 描述性分析
儿童哮喘日门诊就诊人次的描述性分析结果见表 1。2014年1月1日至2017年12月31日,兰州市三所三级甲等医院平均每天儿童哮喘门诊就诊人次为6例(1~31例),男女性别比接近2 ∶1,幼童(年龄0~5岁)占56.24%,6~14岁儿童占43.76%,春、夏、秋、冬4个季节哮喘门诊就诊人次分别为4、6、5、3例。
表 1.
2014—2017年兰州市儿童哮喘日门诊就诊人次的描述性分析
Descriptive statistics on daily hospital outpatient visits for childhood asthma in Lanzhou during 2014 to 2017
| Category | x±s | Minimum | Percentile | Maximum | ||
| P 25 | P 50 | P 75 | ||||
| 0-14 years | 6±3 | 1 | 3 | 5 | 7 | 31 |
| Male | 4±2 | 1 | 2 | 3 | 5 | 20 |
| Female | 2±1 | 1 | 1 | 2 | 3 | 12 |
| 0-5 years | 4±2 | 1 | 2 | 3 | 5 | 15 |
| 6-14 years | 2±1 | 1 | 1 | 2 | 4 | 17 |
| Spring | 4±2 | 1 | 2 | 4 | 6 | 18 |
| Summer | 6±4 | 1 | 3 | 5 | 8 | 22 |
| Autumn | 5±3 | 1 | 3 | 5 | 8 | 31 |
| Winter | 3±1 | 1 | 2 | 4 | 6 | 14 |
气象因素和其他大气污染物基本情况见表 2。日平均温度和相对湿度分别为11.28 ℃和50.83%。
表 2.
2014—2017年兰州市气象因素各指标和空气污染物的描述性分析
Descriptive statistics for meteorological variable and air pollutant levels in Lanzhou during 2014 to 2017
| Category | x±s | Minimum | Percentile | Maximum | ||
| P 25 | P 50 | P 75 | ||||
| PM, particulate matter. | ||||||
| Meteorological factors | ||||||
| Temperature/℃ | 11.28±9.64 | -12.30 | 2.45 | 12.90 | 19.80 | 29.90 |
| Relative humidity/% | 50.83±15.15 | 18.00 | 39.00 | 50.02 | 61.74 | 96.09 |
| Air pollutants | ||||||
| PM2.5/(μg/m3) | 53.30±27.19 | 12.73 | 35.40 | 45.69 | 63.59 | 269.43 |
| PM10/(μg/m3) | 122.90±78.25 | 18.98 | 81.00 | 107.94 | 145.25 | 1484.54 |
| SO2/(μg/m3) | 22.39±14.47 | 3.54 | 10.88 | 18.29 | 30.42 | 81.87 |
| NO2/(μg/m3) | 45.98±15.74 | 7.80 | 36.40 | 45.36 | 52.90 | 146.60 |
| CO/(mg/m3) | 1.34±0.73 | 0.34 | 0.83 | 1.09 | 1.62 | 4.65 |
O3max1h、O38h和O324h的描述性分析结果见表 3。O3的暴露浓度均值O3max1h为(77.59±33.44) μg/m3,O38h为(84.03±36.53) μg/m3,O324h为(38.70±19.97) μg/m3。
表 3.
2014—2017年兰州市臭氧浓度不同度量方式的描述性分析
Descriptive statistics for different ozone metrics in Lanzhou during 2014 to 2017
| Category | x±s | Minimum | Percentile | Maximum | ||
| P 25 | P 25 | P 25 | ||||
| O3max1h, the maximum 1 h daily concentration; O38h, the maximum 8 h daily concentration; O324h, the mean 24 h daily concentration. | ||||||
| O3max1h/(μg/m3) | 77.59±33.44 | 15.00 | 57.20 | 69.40 | 82.40 | 252.60 |
| Spring | 83.38±33.80 | 36.50 | 62.00 | 72.60 | 86.90 | 197.80 |
| Summer | 87.28±41.93 | 28.40 | 63.20 | 71.00 | 87.50 | 252.60 |
| Autumn | 73.15±28.77 | 15.00 | 54.40 | 68.00 | 82.80 | 195.20 |
| Winter | 66.29±21.59 | 15.80 | 52.20 | 62.00 | 77.20 | 158.60 |
| O38h/(μg/m3) | 84.03±36.53 | 8.00 | 56.00 | 78.00 | 106.00 | 221.00 |
| Spring | 99.23±32.82 | 20.00 | 78.00 | 98.00 | 118.00 | 221.00 |
| Summer | 104.43±39.93 | 10.00 | 76.00 | 103.00 | 134.00 | 197.00 |
| Autumn | 72.02±29.93 | 8.00 | 51.00 | 67.00 | 87.00 | 182.00 |
| Winter | 59.94±19.88 | 15.00 | 46.00 | 58.00 | 71.00 | 140.00 |
| O324h/(μg/m3) | 38.70±19.97 | 4.94 | 24.52 | 33.96 | 48.31 | 126.11 |
| Spring | 44.81±20.36 | 14.30 | 29.18 | 40.60 | 56.59 | 116.26 |
| Summer | 48.36±22.78 | 17.23 | 32.22 | 40.61 | 56.21 | 126.11 |
| Autumn | 31.42±14.5 | 8.03 | 20.75 | 28.51 | 38.01 | 84.56 |
| Winter | 29.91±13.74 | 4.94 | 19.17 | 26.96 | 38.25 | 78.58 |
2.2. O3急性健康效应分析结果
控制了时间长期趋势、日均温度、湿度、星期几效应和节假日效应后,单污染模型运行结果显示,O3max1h (lag0)对儿童哮喘门诊就诊人次有影响(P < 0.05),O3max1h浓度每增加10 μg/m3,春季儿童哮喘门诊就诊人次增加1.358% (95%CI:0.126%~2.606%),夏季增加3.351% (95%CI:1.231%~5.516%),秋季增加2.359% (95%CI:1.052%~3.683%),冬季增加1.972% (95%CI:0.484%~3.482%),见表 4。
表 4.
不同滞后天数O3max1h浓度每升高10 μg/m3儿童哮喘门诊就诊人次的超额危险度
Excess risk of childhood asthma hospital outpatient visits associated with a 10 μg/m3increase in O3max1h concentrations with different lag days %
| Lag days | ER (95%CI) | |||
| Spring | Summer | Autumn | Winter | |
| * P < 0.05. ER, excess risk. lag0 to lag7: single-day lagged model, lag0 refers to air pollution exposure of the current day, lag7 is exposure of the previous seventh day; lag01 to lag07: multi-day lagged model, lag01 represents the rolling two day average concentration of the current day and the day before, lag07 is the rolling eight-day average concentration of the current day and the previous seven days. | ||||
| Lag0 | 1.358 (0.126 to 2.606)* | 3.351 (1.231 to 5.516)* | 2.359 (1.052 to 3.683)* | 1.972 (0.484 to 3.482)* |
| Lag1 | 0.233 (-0.978 to 1.458) | -1.312 (-3.078 to 0.486) | -1.868 (-3.180 to 0.539) | -0.878 (-2.268 to 0.531) |
| Lag2 | -0.122 (-1.377 to 1.149) | -0.450 (-1.961 to 1.084) | -1.431 (-2.776 to 0.068) | 0.229 (-1.152 to 1.630) |
| Lag3 | -1.070 (-2.429 to 0.308) | 0.136 (-1.504 to 1.803) | -0.311 (-1.606 to 1.002) | -1.055 (-2.613 to 0.529) |
| Lag4 | -0.024 (-1.257 to 1.225) | -0.662 (-2.318 to 1.022) | 0.035 (-1.287 to 1.375) | 0.626 (-0.837 to 2.110) |
| Lag5 | 0.542 (-0.744 to 1.844) | 1.419 (-0.164 to 3.027) | -0.191 (-1.537 to 1.175) | -0.136 (-1.543 to 1.291) |
| Lag6 | 0.301 (-0.906 to 1.523) | 0.491 (-1.216 to 2.227) | -0.131 (-1.496 to 1.253) | -0.932 (-2.349 to 0.506) |
| Lag7 | 0.474 (-0.832 to 1.797) | -2.552 (-4.981 to 0.062) | -0.740 (-2.037 to 0.575) | 0.506 (-1.054 to 2.090) |
| Lag01 | 1.536 (-0.163 to 3.264) | 0.929 (-1.510 to 3.428) | -2.687 (-4.375 to 0.969) | 0.814 (-1.030 to 2.694) |
| Lag02 | 1.961 (-0.216 to 4.186) | 0.304 (-2.679 to 3.379) | -2.884 (-4.943 to 0.779) | 0.884 (-1.221 to 3.033) |
| Lag03 | 1.419 (-1.192 to 4.100) | 0.528 (-3.179 to 4.377) | -3.018 (-5.423 to 0.551) | 0.270 (-2.221 to 2.824) |
| Lag04 | 1.347 (-1.671 to 4.457) | -0.656 (-5.24 to 4.149) | -1.678 (-4.610 to 1.344) | 1.293 (-2.044 to 4.743) |
| Lag05 | 2.298 (-1.314 to 6.043) | 1.651 (-4.132 to 7.782) | -2.068 (-5.560 to 1.552) | 1.399 (-2.970 to 5.964) |
| Lag06 | 1.769 (-2.055 to 5.743) | 3.614 (-3.852 to 11.659) | -2.254 (-6.414 to 2.092) | -0.008 (-5.804 to 6.145) |
| Lag07 | 0.982 (-2.616 to 4.712) | 1.127 (-7.630 to 10.714) | -3.721 (-8.703 to 1.533) | 0.765 (-6.056 to 8.082) |
控制了时间长期趋势、日均温度、湿度、星期几效应和节假日效应的情况下,单污染模型运行结果显示,夏季时,O38h (lag0)对儿童哮喘门诊就诊人次有影响(P < 0.05),此时O38h浓度每增加10 μg/m3,夏季儿童哮喘门诊就诊人次增加3.320% (95%CI:0.197%~3.829%),春季、秋季和冬季时O38h对儿童哮喘门诊就诊人次的ER差异均无统计学意义(表 5)。夏季时,O324h (lag0)对儿童哮喘门诊就诊人次有影响(P < 0.05),此时O324h浓度每增加10 μg/m3,夏季儿童哮喘门诊就诊人次增加6.600% (95%CI:0.914%~12.607%),春季、秋季和冬季时O324h对儿童哮喘门诊就诊人次的ER差异均无统计学意义(表 6)。O3max1h (lag0)、O38h (lag0)、O324h (lag0)的效应值最显著,本研究选用O3滞后当天(lag0)的效应值进一步进行亚组分析和敏感性分析。
表 5.
不同滞后天数O38h浓度每升高10 μg/m3儿童哮喘门诊就诊人次的超额危险度
Excess risk of childhood asthma hospital outpatient visits associated with a 10 μg/m3increase in O38h concentrations with different lag days %
| Lag days | ER (95%CI) | |||
| Spring | Summer | Autumn | Winter | |
| * P < 0.05. Abbreviations as in Table 4. | ||||
| Lag0 | -0.383 (-1.579 to 0.828) | 3.320 (0.197 to 3.829)* | 0.552 (-1.149 to 2.283) | -2.203 (-3.787 to -0.593) |
| Lag1 | 0.360 (-0.855 to 1.590) | -1.689 (-3.761 to 0.428) | 0.046 (-1.53 to 1.648) | -0.232 (-1.869 to 1.432) |
| Lag2 | -1.449 (-2.869 to -0.008) | -1.106 (-3.083 to 0.912) | 0.272 (-1.324 to 1.895) | 1.171 (-0.359 to 2.724) |
| Lag3 | 0.590 (-0.805 to 2.005) | 1.074 (-1.636 to 3.860) | -0.511 (-2.148 to 1.154) | 0.166 (-1.401 to 1.757) |
| Lag4 | 1.367 (-0.001 to 2.753) | 2.747 (-1.046 to 6.685) | 0.299 (-1.281 to 1.903) | 0.762 (-0.767 to 2.315) |
| Lag5 | -0.400 (-1.672 to 0.888) | -3.538 (-7.437 to 0.525) | 0.261 (-1.279 to 1.825) | -0.841 (-2.353 to 0.695) |
| Lag6 | 0.313 (-0.91 to 1.552) | -0.361 (-4.213 to 3.646) | 0.273 (-1.242 to 1.811) | -0.362 (-1.941 to 1.243) |
| Lag7 | 0.060 (-1.497 to 1.641) | -0.597 (-4.584 to 3.557) | 1.038 (-0.566 to 2.667) | -0.895 (-2.404 to 0.637) |
| Lag01 | 0.251 (-1.248 to 1.772) | 0.542 (-2.186 to 3.346) | 0.550 (-1.652 to 2.801) | -2.429 (-4.610 to 0.197) |
| Lag02 | 0.575 (-1.169 to 2.349) | -0.559 (-4.371 to 3.406) | 0.813 (-1.801 to 3.496) | -1.481 (-4.234 to 1.35) |
| Lag03 | 0.052 (-2.056 to 2.206) | 1.880 (-3.507 to 7.568) | 0.822 (-2.172 to 3.907) | -2.040 (-5.409 to 1.448) |
| Lag04 | 0.140 (-2.718 to 3.081) | 2.916 (-2.582 to 8.724) | 1.124 (-2.367 to 4.74) | -2.219 (-6.268 to 2.004) |
| Lag05 | 0.864 (-2.969 to 4.848) | 0.377 (-5.828 to 6.991) | 1.251 (-2.773 to 5.441) | -4.524 (-9.092 to 0.274) |
| Lag06 | -2.218 (-6.916 to 2.716) | -1.961 (-8.973 to 5.591) | 1.300 (-3.317 to 6.137) | -5.045 (-10.445 to 0.680) |
| Lag07 | -2.271 (-7.518 to 3.273) | -3.019 (-11.132 to 5.835) | 2.307 (-3.077 to 7.990) | -6.712 (-8.969 to 0.050) |
表 6.
不同滞后天数O324h浓度每升高10 μg/m3儿童哮喘门诊就诊人次的超额危险度
Excess risk of childhood asthma hospital outpatient visits associated with a 10 μg/m3increase in O324h concentrations with different lag days %
| Lag days | ER (95%CI) | |||
| Spring | Summer | Autumn | Winter | |
| * P < 0.05. Abbreviations as in Table 4. | ||||
| Lag0 | -1.055 (-2.897 to 0.821) | 6.600 (0.914 to 12.607)* | 1.752 (-0.647 to 4.209) | -2.261 (-6.110 to 1.745) |
| Lag1 | -0.498 (-2.373 to 1.413) | 3.552 (-0.338 to 7.593) | -2.468 (-5.130 to 0.269) | -4.455 (-8.325 to 0.421) |
| Lag2 | -1.088 (-3.358 to 1.236) | -2.162 (-5.842 to 1.662) | -0.634 (-3.473 to 2.289) | -0.220 (-3.861 to 3.559) |
| Lag3 | -0.751 (-2.996 to 1.546) | -0.539 (-4.299 to 3.369) | 2.593 (-1.779 to 7.160) | 1.874 (-1.838 to 5.726) |
| Lag4 | -0.196 (-2.382 to 2.038) | 1.187 (-2.568 to 5.087) | 3.030 (-2.183 to8.520) | 2.030 (-1.794 to 6.004) |
| Lag5 | -0.033 (-2.051 to 2.027) | -2.107 (-5.866 to 1.802) | 3.273 (-2.055 to 8.891) | -0.651 (-4.319 to 3.156) |
| Lag6 | 1.091 (-0.798 to 3.016) | 0.511 (-3.342 to 4.518) | 1.103 (-4.161 to 6.656) | -2.145 (-5.963 to 1.829) |
| Lag7 | 1.067 (-1.232 to 3.418) | 1.505 (-2.498 to 5.672) | -1.963 (-6.748 to 3.067) | -0.133 (-3.848 to 3.726) |
| Lag01 | -0.509 (-2.745 to 1.779) | 8.115 (-2.446 to 14.098) | -0.117 (-3.645 to 3.542) | -6.442 (-11.553 to 1.036) |
| Lag02 | -0.801 (-3.425 to 1.895) | 6.567 (-0.357 to 13.972) | -2.218 (-7.136 to 2.960) | -6.086 (-12.403 to 0.686) |
| Lag03 | -1.475 (-4.674 to 1.831) | 6.014 (-1.878 to 14.54) | -1.225 (-7.883 to 5.915) | -4.232 (-11.895 to 4.098) |
| Lag04 | -1.773 (-5.757 to 2.38) | 6.749 (-2.012 to 16.294) | -0.938 (-9.439 to 8.360) | -2.028 (-11.291 to 8.204) |
| Lag05 | -2.023 (-6.667 to 2.851) | 3.623 (-6.055 to 14.299) | 0.186 (-10.06 to 11.600) | -2.631 (-13.478 to 9.575) |
| Lag06 | -1.232 (-6.561 to 4.400) | 4.382 (-6.527 to 16.564) | 2.964 (-10.017 to 17.817) | -6.308 (-12.75 to 8.038) |
| Lag07 | 0.816 (-5.114 to 7.117) | 4.637 (-7.882 to 18.858) | -1.652 (-14.032 to 12.511) | -6.538 (-11.281 to 9.966) |
2.3. O3浓度对不同性别和年龄儿童哮喘门诊就诊人次的影响
O3max1h (lag0)、O38h (lag0)、O324h (lag0)的效应值最显著,选用O3滞后当天(lag0)的效应值进一步进行亚组分析(表 7~9)。
表 7.
O3max1h浓度每升高10 μg/m3时不同性别和年龄儿童哮喘门诊就诊人次的超额危险度
Excess risk of childhood asthma hospital outpatient visits associated with a 10 μg/m3 increase in concentrations of O3max1h by gender and age %
| Variables | ER(95%CI) | |||
| Spring | Summer | Autumn | Winter | |
| * P < 0.05. ER, excess risk. | ||||
| Male | 0.382 (-1.478 to 2.277) | 1.997 (0.197 to 3.829)* | 4.877 (0.818 to 9.099)* | 2.030 (1.794 to 6.004)* |
| Female | 0.639 (-0.556 to 1.849) | 1.752 (-0.647 to 4.209) | 0.552 (-1.149 to 2.283) | 0.762 (-0.767 to 2.315) |
| 0-5 years | 1.551 (-0.314 to 3.451) | 3.351 (1.231 to 5.516)* | 5.805 (0.686 to 11.185)* | 2.162 (0.556 to 3.794)* |
| 6-14 years | -0.250 (-2.683 to 2.244) | 2.945 (-0.033 to 6.012) | 0.939 (-0.578 to 2.480) | 1.367 (-3.392 to6.361) |
表 9.
O324h浓度每升高10 μg/m3时不同性别和年龄儿童哮喘门诊就诊人次的超额危险度
Excess risk of childhood asthma hospital outpatient visits associated with a 10 μg/m3 increase in concentrations of O324h by gender and age %
| Variables | ER(95%CI) | |||
| Spring | Summer | Autumn | Winter | |
| ER, excess risk. | ||||
| Male | 1.576 (-0.050 to 3.126) | 5.805 (-0.686 to 11.185) | 1.310 (-0.733 to 3.395) | 1.603 (-2.120 to 5.469) |
| Female | 1.948 (-0.588 to 4.549) | 4.342 (-4.905 to 14.489) | 0.579 (-1.416 to 2.613) | 0.759 (-1.557 to 3.130) |
| 0-5 years | -0.595 (-4.263 to 3.213) | 7.106 (-4.832 to 20.542) | 0.379 (-5.174 to 6.257) | 2.610 (-1.229 to 6.599) |
| 6-14 years | 0.262 (-1.938 to 2.512) | 3.499 (-4.179 to 11.791) | -0.280 (-2.530 to 2.023) | -0.102 (-4.990 to 5.038) |
表 7结果显示,控制了时间长期趋势、日均温度、湿度、星期几效应和节假日效应的情况下,按照性别分层,O3对男性的影响明显大于女性,夏季和秋季时,O3max1h浓度每增加10 μg/m3,男童哮喘门诊就诊人次分别增加1.997% (95%CI:0.197%~3.829%)和4.877% (95%CI:0.818%~9.099%);O3对女性的效应差异均无统计学意义(P>0.05)。
按照年龄分层,0~5岁组儿童在夏季、秋季和冬季时O3max1h对儿童哮喘门诊就诊人次的ER差异有统计学意义,在秋季时危害效应最强;O3max1h对6~14岁组儿童在各个季节的危害效应差异均无统计学意义。
表 8结果显示,控制了时间长期趋势、日均温度、湿度、星期几效应和节假日效应的情况下,按照性别分层,O38h对男性和女性在各个季节的影响均不明显(P>0.05)。按照年龄分层,0~5岁组儿童在夏季、秋季,O38h对儿童哮喘门诊就诊人次的ER差异有统计学意义,在夏季时危害效应最强;O38h对6~14岁组儿童在各个季节的危害效应差异均无统计学意义。
表 8.
O38h浓度每升高10 μg/m3时不同性别和年龄儿童哮喘门诊就诊人次的超额危险度
Excess risk of childhood asthma hospital outpatient visits associated with a 10 μg/m3 increase in concentrations of O38h by gender and age %
| Variables | ER(95%CI) | |||
| Spring | Summer | Autumn | Winter | |
| * P < 0.05. ER, excess risk. | ||||
| Male | 1.713 (-0.626 to 4.108) | 1.973 (-1.581 to 5.656) | 1.987 (-0.092 to 4.108) | 1.215 (-0.825 to 3.297) |
| Female | 0.486 (-2.410 to 3.468) | -1.249 (-5.536 to 3.232) | 1.713 (-4.501 to 8.331) | -0.727 (-7.147 to 6.137) |
| 0-5 years | 0.226 (-1.662 to 2.150) | 3.521 (0.806 to 6.309)* | 2.926 (1.250 to 4.629)* | 1.394 (-0.378 to 3.198) |
| 6-14 years | 1.669 (-0.135 to 3.506) | -0.606 (-3.688 to 2.574) | 0.331 (-2.276 to 3.007) | 0.973 (-1.884 to 3.913) |
表 9结果显示,控制了时间长期趋势、日均温度、湿度、星期几效应和节假日效应的情况下,按照性别分层,O324h对男性和女性在各个季节的影响均不明显(P>0.05);按照年龄分层,O324h对不同年龄组儿童在各个季节的危害效应差异均无统计学意义。
2.4. 敏感性分析结果
O3max1h (lag0)、O38h (lag0)、O324h (lag0)的效应值最显著,选用O3滞后当天(lag0)的效应值进一步进行敏感性分析(表 10、11)。
表 10.
双污染物模型O3浓度每升高10 μg/m3时儿童哮喘门诊就诊人次的超额危险度
Excess risk of childhood asthma hospital outpatient visits associated with a 10 μg/m3 increase in O3 concentrations in two-pollutant models %
| Variables | ER(95%CI) | ||
| O3max1h | O38h | O324h | |
| O3max1h, the maximum 1 h daily concentration; O38h, the maximum 8 h daily concentration; O324h, the mean 24 h daily concentration. PM, particulate matter; ER, excess risk. | |||
| Single model | 3.351 (1.231 to 5.516) | 3.320 (0.197 to 3.829) | 6.600 (0.914 to 12.607) |
| O3+PM2.5 | 3.127 (1.513 to 5.195) | 3.193 (0.124 to 3.768) | 6.456 (0.637 to 11.208) |
| O3+PM10 | 3.191 (1.581 to 5.255) | 3.055 (0.148 to 3.794) | 6.573 (0.733 to 10.137) |
| O3+NO2 | 3.477 (1.501 to 5.310) | 3.446 (0.136 to 3.789) | 6.671 (0.724 to 12.428) |
| O3+ SO2 | 3.085 (1.239 to 4.985) | 3.342 (0.092 to 3.609) | 6.774 (1.028 to 12.834) |
| O3+ CO | 3.338 (1.644 to 5.484) | 3.415 (0.216 to 3.846) | 6.929 (1.386 to 13.202) |
| O3+SO2+NO2 | 3.329 (1.167 to 5.321) | 3.120 (0.169 to 3.629) | 6.687 (1.014 to 11.307) |
| O3+SO2+PM10 | 3.253 (1.118 to 5.476) | 3.097 (0.117 to 3.466) | 6.471 (0.971 to 11.671) |
| O3+SO2+PM2.5 | 3.355 (1.138 to 5.383) | 3.155 (0.127 to 3.491) | 6.565 (0.833 to 10.782) |
| O3+SO2+CO | 3.402 (1.136 to 5.289) | 3.421 (0.135 to 3.742) | 6.524 (0.901 to 11.763) |
| O3+NO2+PM10 | 3.338 (1.192 to 5.409) | 3.351 (0.102 to 3.573) | 6.684 (1.012 to 12.213) |
| O3+NO2+PM2.5 | 3.385 (1.216 to 5.446) | 3.285 (0.187 to 3.446) | 6.529 (1.206 to 12.102) |
| O3+NO2+CO | 3.379 (1.197 to 5.229) | 3.460 (0.201 to 3.733) | 6.721 (1.314 to 12.786) |
| O3+PM10+PM2.5 | 3.293 (1.124 to 5.268) | 3.172 (0.144 to 3.776) | 6.456 (0.637 to 10.208) |
| O3+PM10+CO | 3.275 (1.148 to 5.294) | 3.155 (0.141 to 3.664) | 6.473 (0.704 to 10.375) |
| O3+PM2.5+CO | 3.401 (1.136 to 5.789) | 3.246 (0.151 to 3.704) | 6.571 (0.745 to 11.123) |
表 11.
不同时间自由度O3浓度每升高10 μg/m3时儿童哮喘门诊就诊人次的超额危险度的敏感性分析
Sensitivity analysis of the effect of excess risk for every 10 μg/m3 increase in O3 concentration on outpatients with childhood asthma at different degrees of freedom %
| Items | ER(95%CI) | ||
| O3max1h | O38h | O324h | |
| O3max1h, the maximum 1 h daily concentration; O38h, the maximum 8 h daily concentration; O324h, the mean 24 h daily concentration; ER, excess risk. | |||
| df=6 | 3.330 (1.345 to 5.354) | 3.463 (0.267 to 3.924) | 6.676 (0.868 to 11.371) |
| df=7 | 3.351 (1.231 to 5.516) | 3.320 (0.197 to 3.829) | 6.600 (0.914 to 12.607) |
| df=8 | 3.502 (1.449 to 5.597) | 3.464 (0.247 to 3.955) | 6.562 (0.843 to 11.426) |
| df=9 | 3.238 (1.435 to 5.385) | 3.249 (0.104 to 3.832) | 6.452 (1.007 to 11.833) |
| df=10 | 3.279 (1.258 to 5.488) | 3.399 (0.158 to 3.886) | 6.712 (0.945 to 12.030) |
双污染物和多污染物敏感性模型分析结果如表 10所示,控制了时间长期趋势、日均温度、湿度、星期几效应和节假日效应的情况下,O3max1h、O38h、O324h与儿童哮喘门诊就诊风险的双污染物及多污染物模型拟合结果提示,在单污染物模型基础上分别引入PM2.5、PM10、NO2、SO2及同时引入两种污染物后,三种O3暴露浓度的ER值本身无明显变化,显示本研究模型稳定性较好。
比较时间自由度的变化对儿童哮喘的影响,由表 11可见,控制了时间长期趋势、日均温度、湿度、星期几效应和节假日效应的情况下,当自由度在6~10之间变化时,三种O3暴露浓度的ER及其95%CI无明显变化,表明模型相对稳定。
3. 讨论
本研究运用广义相加模型的分析方法控制多种影响因素后发现,兰州市夏季O3短期暴露水平升高对于儿童哮喘门诊就诊人次的增加有显著影响。在滞后当天(lag0)的情况下,当O38h浓度上升10 μg/m3时,儿童哮喘门诊就诊人次增加3.320% (95%CI:0.197%~3.829%)。国内有研究显示,O38h每升高10 μg/m3会引起重庆市和合肥市儿童哮喘门诊就诊人次分别增加0.200% (95%CI:-0.100%~0.400%)[3]和1.661% (95%CI:0.371%~3.065%)[4]。香港特别行政区的一项研究显示,O38h每升高10 μg/m3会引起儿童哮喘就诊人次增加3.900% (95%CI:2.800%~5.100%)[9]。韩国和澳大利亚的研究亦发现,O38h每升高10 μg/m3会引起儿童哮喘就诊人次增加5.385% (95%CI:2.564%~8.718%)[13]和3.047% (95%CI:0.714%~5.476%)[14]。本研究结果与已有研究结果的数值范围基本一致。
本研究采用的三个不同度量方式的O3暴露指标每升高10 μg/m3时,儿童哮喘就诊风险均呈不同程度的增加,三者对儿童哮喘的影响基本是一致的,即O3暴露对儿童哮喘的影响均在滞后当天出现危害效应。在本研究时段内比较三个不同O3暴露指标后显示,O324h (lag0)增加10 μg/m3时的危害效应最大,能够引起儿童哮喘就诊风险增加6.600% (95%CI:0.914%~12.607%);O3max1h (lag0)增加10 μg/m3时的危害效应次之,可引起儿童哮喘就诊风险增加3.351% (95%CI:1.231%~5.516%);O38h (lag0)增加10 μg/m3时的危害效应最小,引起儿童哮喘就诊风险增加3.320% (95%CI:0.197%~3.829%)。因此,虽然目前O38h是公认的能够更好代表O3急性健康效应的指标,但是在本研究中,与其他两个O3浓度指标相比,O38h暴露度量方式并没有显示出与儿童哮喘更加显著的关联性,这与以往在香港特别行政区、广州、苏州发现的O38h的健康效应更为显著的研究结论并不完全一致[9, 15-16]。本研究结果与以往结果不同的原因可能有两点:一方面大气O3污染水平的地区差异较大,当地的大气O3污染特征不同可能会影响研究结果的可比性,不同区域的气候特征和人群生活习惯亦存在差异,有可能会造成O3的健康效应在不同地区的结果不一致;另一方面,模型参数(如自由度的选择)、数据指标形式(如使用不同度量方式的O3指标)的不同也可能影响研究结果的比较。
季节分层结果显示,O3max1h在不同的季节中对不同性别和年龄儿童哮喘的影响有明显差别,O3max1h在夏季、秋季和冬季对男童和0~5岁组的影响较大,其中,秋季时男童和0~5岁组哮喘患者就诊风险增加的百分比受到O3max1h影响最为显著。从临床上来说,不同季节诱发哮喘发生的原因是不同的,夏季时,兰州O3浓度在日出前最低,之后逐渐增加,在15:00~16:00达到一天中的最高值,此时草木生长比较茂盛,空气中花粉等过敏源也比较多,儿童在午后外出活动的时间更多,儿童哮喘发作的概率较高。秋季和冬季,兰州的气温在午后12:00回升,太阳辐射逐步增加,O3浓度呈现出逐渐上升趋势,至15:00达到峰值,此时儿童外出活动更多[8, 17],同时因秋冬季节温差较大,平均气温低,气候比较干燥,儿童不易咳出呼吸道分泌物,空气污染和气温变化可引起儿童哮喘的发作,且影响更大[17-18]。
季节分层结果显示,O38h在不同的季节中对不同性别和年龄儿童哮喘的影响有明显差别,O38h在夏季和秋季对0~5岁组儿童的影响较大,O3浓度在白天午后12:00~17:00较高[8],0~5岁组儿童户外活动时间较长,同时从临床角度分析,夏季和秋季草木生长比较茂盛,花粉过敏导致的过敏性哮喘增加;夏季和秋季也是霉菌、螨虫等的繁殖高峰期,对霉菌、螨虫过敏者就会发病,另外空气闷湿也会导致患者呼吸不畅,也可能引起哮喘发作[19]。
季节分层结果显示,O324h在不同的季节中对不同性别和年龄儿童哮喘有影响,但差异无统计学意义,分析可能的原因是O3形成的过程依赖于温度,其浓度在炎热的夏季较高、在寒冷的冬季较低,故夏季午后的污染最为严重[8]。O324h浓度是通过计算O3的每日每小时平均浓度的平均值,与O3max1h和O38h相比,不能很好地反映O3浓度变化对儿童哮喘的影响。
多污染物模型中,当额外纳入一种或两种污染物,O3的效应与单污染物模型相比略有波动,但本研究的结果仍然较稳定,与Li等[15]在广州的研究结果一致。可能是因为在单污染物模型中,不考虑污染物之间的相互影响,只考虑一种污染物的作用,因此污染物效应相对稳定,但在多污染物模型中,由于污染物间的相关系数较高,不同污染物间存在共线性,可能使污染物效应发生细微变化。此外,不同地区之间,污染物的不同来源可能导致组分间的相互影响,从而导致这种不一致性的产生。
本研究选取了我国西北内陆干旱地区的城市兰州为研究地点,采用广义相加时间序列讨论了不同度量方式的O3暴露水平及其对儿童哮喘门诊就诊风险的影响,基于O3浓度的日变化特征采用多个度量方式计算O3暴露与儿童哮喘门诊就诊风险的相关性,研究结果提示兰州市开展O3污染治理及减缓与防护人群急性健康效应的必要性,为制定科学有效的大气质量控制政策提供了定量研究的数据支撑。
本研究同时存在一定的局限性,研究采用的O3数据是地面监测数据,并不能完全代表研究对象的实际O3暴露量,本研究为生态学研究,不一定能外推于其他情况。
综上所述,夏季时兰州市O3暴露水平的度量指标O3max1h、O38h和O324h均对儿童哮喘就诊风险的增加有显著影响,其中男性和0~5岁儿童为敏感人群。民众应加强对O3污染的关注,建议政府部门合理分配卫生资源,开展相关健康教育,减少O3暴露,降低其对居民健康的危害。
Funding Statement
甘肃省人民医院院内科研基金(18GSSY4-24)
Supported by the Research Fund Project of Gansu Provincial Hospital (18GSSY4-24)
References
- 1.徐 甜, 王 志远, 张 兵, et al. 中国环境保护重点城市空气质量指数时空变化特征. 中国公共卫生. 2016;32(8):1027–1031. doi: 10.11847/zgggws2016-32-08-06. [DOI] [Google Scholar]
- 2.Ferrante G, Antona R, Malizia V, et al. Asthma and air pollution. Ital J Pediatr. 2014;40(S1):A75. doi: 10.1186/1824-7288-40-S1-A75. [DOI] [Google Scholar]
- 3.Zhang YW, Ni H, Bai LJ, et al. The short-term association between air pollution and childhood asthma hospital admissions in urban areas of Hefei City in China: A time-series study. Environ Pollut. 2019;169:510–516. doi: 10.1016/j.envres.2018.11.043. [DOI] [PubMed] [Google Scholar]
- 4.Ding L, Zhu DJ, Peng DH, et al. Air pollution and asthma attacks in children: A case-crossover analysis in the city of Chongqing, China. Environ Pollut. 2017;220(PtA):348–353. doi: 10.1016/j.envpol.2016.09.070. [DOI] [PubMed] [Google Scholar]
- 5.Dai YR, Qiu H, Sun SZ, et al. Age-dependent effect of ambient ozone on emergency asthma hospitalizations in Hong Kong. J Allergy Clin Immun. 2018;141(4):1532–1534. doi: 10.1016/j.jaci.2018.01.006. [DOI] [PubMed] [Google Scholar]
- 6.杨春雪. 细颗粒物和臭氧对我国居民死亡影响的急性效应研究[D]. 上海: 复旦大学, 2012.
- 7.闫美霖. 广州市越秀区臭氧短期暴露与人群死亡风险的时间序列研究及健康风险评估[D]. 北京: 北京大学, 2013.
- 8.尹伊. 基于CMAQ模式、MEIC清单对兰州市主城区臭氧污染特征的初步研究[D]. 兰州: 兰州大学, 2019.
- 9.Ko FWS, Tam W, Wong TW, et al. Effects of air pollution on asthma hospitalization rates in different age groups in Hong Kong. Clin Exp Allergy. 2007;37(9):1312–1319. doi: 10.1111/j.1365-2222.2007.02791.x. [DOI] [PubMed] [Google Scholar]
- 10.班 婕, 李 湉湉. 北京市不同度量方式下臭氧短期暴露人群急性健康效应研究. 环境与健康杂志. 2016;33(4):287–291. doi: 10.16241/j.cnki.1001-5914.2016.04.002. [DOI] [Google Scholar]
- 11.Capraz O, Deniz A, Dogan N, et al. Effects of air pollution on respiratory hospital admissions in Istanbul, Turkey, 2013 to 2015. Chemosphere. 2017;181:544–550. doi: 10.1016/j.chemosphere.2017.04.105. [DOI] [PubMed] [Google Scholar]
- 12.Chang Q, Liu S, Chen ZJ, et al. Association between air pollutants and outpatient and emergency hospital visits for childhood asthma in Shenyang City of China. Int J Biometeorol. 2020;64(9):1539–1548. doi: 10.1007/s00484-020-01934-9. [DOI] [PubMed] [Google Scholar]
- 13.Lee JT, Cho YS, Son JY. Relationship between ambient ozone concentrations and daily hospital admissions for childhood asthma/atopic dermatitis in two cities of Korea during 2004-2005. Int J Environ Health Res. 2010;20(1):1–11. doi: 10.1080/09603120903254033. [DOI] [PubMed] [Google Scholar]
- 14.Petroeschevsky A, Simpson RW, Thalib L, et al. Associations between outdoor air pollution and hospital admissions in Brisbane, Australia. Arch Environ Health. 2001;56(1):37–52. doi: 10.1080/00039890109604053. [DOI] [PubMed] [Google Scholar]
- 15.Li TT, Yan ML, Ma WJ, et al. Short-term effects of multiple ozone metrics on daily mortality in a megacity of China. Environ Sci Pollut Res Int. 2015;22(11):8738–8746. doi: 10.1007/s11356-014-4055-5. [DOI] [PubMed] [Google Scholar]
- 16.Yang CX, Yang HB, Guo S, et al. Alternative ozone metrics and daily mortality in Suzhou: The China Air Pollution and Health Effects Study (CAPES) Sci Total Environ. 2012;426:83–89. doi: 10.1016/j.scitotenv.2012.03.036. [DOI] [PubMed] [Google Scholar]
- 17.刘金悦. 济南市大气污染与儿童呼吸系统疾病住院量的相关性分析[D]. 济南: 济南大学, 2019.
- 18.陈浪. 石家庄市空气污染对儿童呼吸系统疾病门诊影响及温度的修饰效应[D]. 石家庄: 华北理工大学, 2019.
- 19.马依拉·尔肯. 乌鲁木齐市大气污染物, 气象因素与儿童哮喘住院人数的相关研究[D]. 乌鲁木齐: 新疆医科大学, 2014.
