Part 4. Interaction between air pollution and respiratory viruses: time-series study of daily mortality and hospital admissions in Hong Kong.

BACKGROUND

Populations in Asia are not only at risk of harm to their health through environmental degradation as a result of worsening pollution problems but also constantly threatened by recurring and emerging influenza epidemics and. pandemics. Situated in the area with the world's fastest growing economy and close to hypothetical epicenters of influenza transmission, Hong Kong offers a special opportunity for testing environmental management and public health surveillance in the region. In the Public Health and Air Pollution in Asia (PAPA*) project, the Hong Kong research team assessed the health effects of air pollution and influenza as well as the interaction between them. The team also assessed disparities in the health effects of air pollution between relatively deprived and more affluent areas in Hong Kong. The aim was to provide answers to outstanding research questions relating to the short-term effects of air pollution on mortality and hospital admissions; the health effects of influenza with a view to validating different measures of influenza activity according to virologic data; the confounding effects of influenza on estimates of the health effects of air pollution; the modifying effects of influenza on the health effects of air pollution; and the modifying effects of neighborhood social deprivation on the health effects of air pollution.

DATA

Data on mortality and hospital admissions for all natural causes, as well as the subcategories of cardiovascular diseases (CVD) and respiratory diseases (RD), were derived from the Hong Kong Census and Statistics Department and the Hospital Authority. Daily concentrations of nitrogen dioxide (NO2), sulfur dioxide (SO2), particulate matter with an aerodynamic diameter < or = 10 pm (PM10); and ozone (O3) were derived from eight monitoring stations with hourly data that were at least 75% complete during the study period. Three measures of influenza and respiratory syncytial virus (RSV) activity were derived from positive isolates of specimens in the virology laboratory of Queen Mary Hospital (QMH), the main clinical teaching center at The University of Hong Kong and part of the Hong Kong Hospital Authority network of teaching hospitals: influenza intensity (defined as the weekly proportion of positive isolates of influenza in the total number of specimens received for diagnostic tests); the presence of influenza epidemic (defined as a period when the weekly frequency of these positive isolates is > or = 4% of the annual total number of positive isolates [i.e., twice the expected mean value] in two or more consecutive weeks); and influenza predominance (defined as a period of influenza epidemic when the weekly frequency of RSV was less than 2% for two or more consecutive weeks). The weekly proportion of positive isolates of RSV in total specimens was determined in the same way as for influenza intensity. A social deprivation index (SDI) was defined by taking the average of the proportions of households or persons with the following six characteristics in each geographic area using the census statistics: unemployment; household income < U.S. $250 per month; no schooling at all; never-married status; one-person household; and subtenancy. A Poisson regression with quasi-likelihood to account for overdispersion was used to develop core models for daily health outcomes, with a natural spline smoothing function to filter out seasonal patterns and long-term trends in this time-series study of daily mortality and hospital admissions, and with adjustment for days of the week, temperature, and relative humidity (RH). Air pollutant concentration values were entered into the core model to assess the health effects of specific pollutants. The possible confounding effects of influenza were assessed by observing changes in magnitude of the effect estimate when each influenza measurement was entered into the model; and interactions between air pollution and influenza were assessed by entering the terms for the product of the air pollutant concentration and a measurement of influenza activity into the model. A Poisson regression analysis was performed to assess the effects of air pollution in each area belonging to low, middle, or high social deprivation strata according to the tertiles of the SDI. The differences in air pollution effects were tested by a case-only approach. RESULTS The excess risk (ER) estimates for the short-term effects of air pollution on mortality and hospitalization for broad categories of disease were greater in those 65 years and older than in the all-ages group and were consistent with other studies. The biggest health impacts were seen at the extremes of the age range. The three measures employed for influenza activity based on virologic data-one based on a proportion and the other two using frequencies of positive influenza isolates-were found to produce consistent health impact estimates, in terms of statistical significance. In general, we found that adjustment for influenza activity in air pollution health effect estimations took account of relatively small confounding effects. However, we conclude that it is worthwhile to make the adjustment in a sensitivity analysis and to obtain the best possible range of effect estimates from the data, especially for respiratory hospitalization. Interestingly, interaction effects were found between influenza activity and air pollution in the estimated risks for hospitalization for RD, particularly for 03. These results could be explained in terms of the detrimental effects of both influenza viruses and air pollutants, which may be synergistic or competing with each other, though the mechanism is still unknown. The results deserve further study and the attention of both public health policy makers and virologists in considering prevention strategies. IMPLICATIONS In Hong Kong, where air pollution may pose more of a health threat than in North American and Western European cities, the effects of air pollution also interact with influenza and with residence in socially deprived areas, potentially leading to additional harm. Asian governments should be aware of the combined risks to the health of the population when considering environmental protection and management in the context of economic, urban, and infrastructure development. This is the first study in Asia to examine the interactions between air pollution, influenza, and social deprivation from an epidemiologic perspective. The biologic mechanisms are still unclear, and further research is needed.