都會區大氣臭氧與氣溫對敏感性次族群健康之相關分析

Abstract

近年本國空氣品質不良情況已受控制，唯生成機制複雜之二次污染物-臭氧，未因空氣污染管制措施而改善，反有逐年上升之趨勢。而在全球暖化的情境下，歐美氣候及空氣品質預測模式均指出高溫及相關之臭氧濃度，未來將逐年上升，顯現其對敏感性族群之健康風險將日趨嚴重。本研究首先分析北台灣臭氧與其前驅物的時空變化與長期趨勢；進而評估各項臭氧指標對於台北都會區14歲以下孩童的呼吸道感染與其他上呼吸道疾病就醫的相對風險；後續更細部考量空氣污染物、氣溫及呼吸道流行病毒對上呼吸道感染之就醫風險；最終，分析極端溫度及其連續數日之事件對於台北、台中及高雄都會區65歲以上族群死因別死亡風險分析。 本研究使用1994-2007年間衛生署死亡登記資料〔全死因、心血管死因(ICD9 390-459)及呼吸道疾病(ICD9 460-519)〕、氣象資料(平均溫度、相對濕度及大氣壓力等)及空氣品質資料〔一氧化碳(CO)、非甲烷碳氫化合物(non-methane hydrocarbon，NMHC)、氮氧化物(NOx)、一氧化氮(NO)、二氧化氮(NO2)、懸浮微粒(PM10、PM2.5)、二氧化硫(SO2)、最大臭氧小時值(O3, max)、最大八小時臭氧平均值(O3, 8h)及24小時臭氧平均值(O3, 24h)〕；2000-2007年間全民健保100萬人承保歸人檔中台北都會區14歲以下孩童之急性呼吸道感染(ICD9 CM 460-466)、其他上呼吸道疾病(ICD9 CM 470-478)及上呼吸道感染(ICD9 CM 465.9)；以及1999-2007年間疾病管制局收集之台北都會區呼吸道相關病毒檢測結果。本研究分析上述各因子之時空趨勢及相關性，後續則以廣義線性模式(generalized linear model，GLM)評估風險因子在特定延遲日下對疾病就醫之風險，及遞延分配模式(distributed lag model，DLM)量化各別風險因子累計數日之相對危險(relative risk，RR)及95%信賴區間(confidence interval，CI)。 本研究發現臭氧的濃度受到地形與氣候的影響，北台灣的臭氧濃度逐年上升，即使在臭氧前驅物逐漸下降的的情形下，臭氧的生成效率與影響時間逐漸拉長，都會區中的NO滴定效應明顯且中高濃度的臭氧超標站日數發生頻率有增加的趨勢。在健康風險上，臭氧24小時平均值在九月到隔年一月的秋冬季學期期間對於14歲以下孩童呼吸道就醫風險具有顯著相關，對急性呼吸道感染的RR為1.025 (95% CI: 1.022-1.028)，對其他上呼吸道疾病的RR為1.035 (95% CI: 1.023-1.047)，有明顯的暴露劑量反應關係。考量呼吸道病毒流行情形後，台北都會區內上呼吸道感染就醫受到空氣污染物NO2的影響最為顯著，在冬季時RR為1.19(95% CI: 1.18-1.21)，O3, 24h次之(RR = 1.06，95% CI: 1.04-1.07)；病毒則以腺病毒對就醫之相對風險最高，RR為1.05 (95% CI: 1.02-1.07)；低溫時有較高就醫風險。在極端溫度對年長者反應上，過高(>30 &ordm;C)或過低(<18 &ordm;C)的溫度可能會造成65歲以上年長敏感族群全死因及心血管疾病死亡風險增加，但對呼吸道死因則不明顯，連續三日以上高(低)溫事件所造成死亡風險並不顯著，我們亦發現近年台灣都會區年長族群對於高溫所造成死亡的敏感度似有下降。

Ambient ozone and temperature have been predicted to increase in next decades; consequently the health risks associated with ozone and temperature deserve study. This study aims to: (1) analyze the temporal and spatial patterns for the concentrations of ambient ozone and its precursors in Northern Taiwan; (2) associated the risk of respiratory diseases for population younger than 15 years with various ozone metrics in Taipei metropolitan; (3) evaluate the risk of upper respiratory infection associated with atmospheric indices and circulating respiratory viruses for population younger than 15 years in Taipei metropolitan; and (4) assess the mortality risk related to temperature and its long lasting events for population at 65 years old or older in Taipei, Taichung and Kaohsiung. We used the vital statistics data and the measurements of weather and air pollutants including CO, non-methane hydrocarbon (NMHC), NOx, NO, NO2, PM10, PM2.5,SO2, maximum hourly ozone (O3, max), maximum 8-hour moving average ozone (O3, 8h) and 24-hour average ozone (O3, 24h) in 1994-2007. The insurance reimbursement claims in 2000-2007 from a randomly selected sample containing one million insured individuals were also used. The daily clinic visits for cases on acute respiratory infections (ICD9 CM 460-466, ARI), other diseases of upper respiratory (ICD9 CM 470-478, ODUR) and upper respiratory infection (ICD9 CM 465.9, URI), and daily deaths from all causes (ICD all), circulatory diseases (ICD9 390-459) and respiratory diseases (ICD9 460-519) were retrieved for analyses. The daily virus-specific positive isolation rates were calculated based on the viruses surveillance data in Taipei in 1999-2007. The relative risks (RRs) with 95% confidence intervals (CIs) for mortality and morbidity associated with temperatures, concentrations of air pollutants and viruses positive isolation rates were estimated using generalized linear model (GLM) and distributed lag model (DLM) with Poisson regressions. The results showed that the concentrations of ambient ozone varied with geographic and seasonal conditions. The concentrations of ozone precursors decreased significantly, while all ozone metrics were increasing over the study period. Ozone generated comparatively efficient in afternoon causing a relatively long ozone exposure and a relatively frequent occurrence of the moderate and high ozone concentrations (80-120 ppb) in Northern Taiwan. The hospital admissions for ARI and ODUR for children aged 14 years and younger in Taipei were significantly correlated with 24-hour average ozone (O3, 24h) during September-January. As the positive isolation rates of circulating respiratory viruses were included in the models, the relative risks of URIs for children were significantly associated with the ambient levels of NO2 and O3, 24h, and the isolation rates of influenza type B, parainfluenza viruses and adenoviruses. The excess risk for URIs was higher in the cold season (November-April) than that in the warm season (May-October). The cumulative mortality risks for the elderly (65 years and above) due to all causes and cardiovascular diseases were significantly associated with high and low temperature in metropolitans of Taiwan. However, no significant associations between temperatures and mortality from respiratory diseases were observed. There was no significant excess of mortality due to exposing to extreme low/high temperatures for consecutive three days and above. The decreaseing trends of mortality risk to the high temperature but the cold temperature for the elderly population in the studied metropolitans was noticed.