Printer Friendly

Temperature, air pollution, and hospitalization for cardiovascular diseases among elderly people in Denver.

Daily measures of maximum temperature, particulate matter [less than or equal to] 10 pm in aerodynamic diameter (P[M.sub.10]), and gaseous pollution (ozone; nitrogen dioxide, sulfur dioxide, and carbon monoxide) were collected in Denver, Colorado, in July and August between 1993 and 1997. We then compared these exposures with concurrent data on the number of daily hospital admissions for cardiovascular diseases in men and women > 65 years of age. Generalized linear models, assuming a Poisson error structure for the selected cardiovascular disease hospital admissions, were constructed to evaluate the associations with air pollution and temperature. After adjusting the admission data for yearly trends, day-of-week effects, ambient maximum temperature, and dew point temperature, we studied the associations of the pollutants in single-pollutant models with lag times of 0-4 days. The results suggest that [O.sub.3] is associated with an increase in the risk of hospitalization for acute myocardial infarction, coronary atherosderosis, and pulmonary heart disease. S[O.sub.2] appears to be related to increased hospital stays for cardiac dysthythmias, and CO is significantly associated with congestive heart failure. No association was found between particulate matter or N[O.sub.2] and any of the health outcomes. Males tend to have higher numbers of hospital admissions than do females for all of the selected cardiovascular diseases, except for congestive heart failure. Higher temperatures appear to be an important factor in increasing the frequency of hospitalization for acute myocardial infarction and congestive heart failure, and are associated with a decrease in the frequency of visits for coronary atherosclerosis and pulmonary heart disease. Key words: acute myocardial infarction, air pollution, cardiac dysrhythmias, cardiovascular diseases, CO, congestive heart failure, coronary atherosderosis, generalized estimating equations, N[O.sub.2], [O.sub.3], P[M.sub.10], Poisson regression, pulmonary heart disease, S[O.sub.2], temperature.

**********

There is substantial epidemiologic literature indicating a link between air pollution and cardiovascular morbidity and mortality. This includes not only studies of episodic pollution such as occurred during 1930 in the Meuse Valley (Firket 1931), 1948 in Pennsylvania (Shrenk et al. 1949), and 1952 in London (Ministry of Health 1954) but also of the generally low concentrations found in urban areas. Studies have been carried out in North America (Burnett et al. 1997a, 1997b; Morris et al. 1995; Morris and Naumova 1998; Schwartz 1999; Schwartz and Morris 1995; Zanobetti et al. 2000), in Western Europe (Atkinson et al. 1999; Ballester et al. 2001; Diaz et al. 2001; Hoek et al. 2001; Prescott et al. 1998), in Tokyo, Japan (Piver et al. 1999), and in Hong Kong, China (Wong et al. 1999). Most of these studies showed a predominant effect of particulates and carbon monoxide on cardiovascular admissions (Ballester et al. 2001; Burnett et al. 1997b; Morris et al. 1995; Morris and Naumova 1998; Schwartz 1999; Schwartz and Morris 1995; Wong et al. 1999). However, the Tokyo study (Piver et al. 1999) suggests an independent effect of nitrogen dioxide. Furthermore, the studies in Canada (Burnett et al. 1997a) and Spain (Diaz et al. 2001) demonstrated consistent effects for sulfur dioxide and/or ozone on cardiovascular hospital admissions. Several reports have addressed the issue of weather and mortality; extreme temperatures have been associated with increased daily mortality in numerous regions of the world (Braga et al. 2001; Kunst et al. 1993). Mortality has also been observed to increase during periods of 3 or more days of unusual temperatures during summer or winter, showing that temperature variability is an important determinant of human health effects (Braga et al. 2001, 2002; Saez et al. 1995).

It has been suggested that weather and temperature may modify the effects of air pollution on health both at high temperatures (Katsouyanni et al. 1993) and low temperatures (Morris and Naumova 1998). The Intergovernmental Panel on Climate Change (IPCC; 1996) has projected that atmospheric concentrations of carbon dioxide could double in the next 50-100 years. A doubling of atmospheric concentrations of C[O.sub.2]2 could result in an increase in average global surface air temperatures of 1-3[degrees]C because of the greenhouse effect. In addition, because approximately 65% of atmospheric C[O.sub.2] comes from combustion of fossil fuels, increasing concentrations of C[O.sub.2] could also be accompanied by increasing concentrations of other air pollutants, particularly in large urban areas. An increase in surface air temperatures could accompany a greater frequency and duration of heat waves. According to the IPCC (1996), the frequency of extremely hot days in temperate climates approximately doubles for every 2-3[degrees]C increase in temperature during the average summer. Because heat waves often occur in large metropolitan areas during warm summer months, these large cities could experience an increase in the incidence of heat-related morbidity and mortality (McMichael et al. 1996). To address this specific issue, we focused our research on the months of the year when the frequency of exposure to high daily maximum temperatures ([T.sub.max]) and high air pollutant concentrations would be the greatest. The present study focuses on the months of July and August (1993-1997) in Denver, Colorado.

Studies have documented associations between daily variations in air pollution and cardiovascular deaths and hospital admissions (Dockery et al. 2000). Mechanisms for the effects of air pollution on cardiovascular mortality and morbidity may include changes in blood coagulability (Seaton et al. 1995) and changes in the nervous system control of the heart, possibly leading to arrhythmias (Peters et al. 2000). Mechanisms linking temperature to cardiovascular mortality and morbidity have also been postulated. The blood viscosity and cholesterol levels have been found to increase with high temperatures (Keatinge et al. 1986), whereas blood pressure and fibrinogen levels increase during winter, although outdoor temperature does not seem to determine the seasonal variation in fibrinogen (van der Bom et al. 1997). In many epidemiology studies with respiratory and cardiovascular diseases, regression models include adjustments for seasonal climate variability to isolate the contributions of air pollution on the daily number of deaths or hospital admissions for these diseases. In contrast to using seasonal adjustments for the entire year, we focused on the months of July and August, partially to avoid potential bias in the type of seasonal adjustment to be used and partially to focus on short time-frame changes in temperature.

Materials and Methods

Setting. Denver once was among the most polluted cities in the United States, with a brown cloud as a constant reminder of its air pollution woes. The city regularly exceeded the federal air pollution standards for 200 days/year, every year in the late 1970s. Air pollution in Denver mainly derives from motor vehicle exhaust emissions, with industrial pollution playing a minor role. Denver's air pollution levels are also exacerbated by temperature inversions resulting from Denver's geographic location. During a temperature inversion, a warm, less-dense inversion layer of air overlies colder, denser air, forming a "lid" that traps pollutants below it. In Denver the trapping effect is even more pronounced because long periods of calm weather and light winds trap pollutants against nearby mountains. The local government has undertaken sizable efforts to reduce air pollution; the city has now complied with Clean Air Act regulations for particulates, CO, and [O.sub.3] (U.S. EPA 2001) and has cut its pollution dramatically. Nonetheless, the remaining levels of pollutants may continue to have health effects, even at their improved levels. In this study, we defined Denver as all of Denver County.

Hospital admission data. The numbers of daily hospital admissions to all 11 hospitals in Denver County were obtained for the months of July and August during 1993-1997. July and August are of particular interest because these are the months with highest temperatures, combined with relatively high levels of outdoor activity during which exposure would be maximized. The data were provided by the Agency for Healthcare Research and Quality (AHRQ; Rockville, MD). The AHRQ maintains state-specific hospital discharge databases as part of the Healthcare Cost and Utilization Project (HCUP 2001). Data elements included admission date, age, sex, and discharge diagnoses. The study population was restricted to men and women > 65 years of age admitted to the hospitals with a primary discharge diagnosis of the following causes [coded using International Classification of Diseases, 9th Revision-CM (ICD-9-CM); Centers for Medicare & Medicaid Services 2003]: acute myocardial infarction (ICD-9-CM 410.00-410.92), coronary atherosclerosis (ICD-9-CM 414.00-414.05), pulmonary heart disease (ICD-9-CM 416.0-416.9), cardiac dysrhythmias (ICD-9-CM 427.0-427.9), and congestive heart failure (ICD-9-CM 428.0). These particular diseases were chosen because they include a mixture of acute and chronic cardiovascular diseases. Admissions resulting from viral infections (e.g., influenza) and chronic pulmonary dysfunctions (e.g., embolisms, hypertension, thrombosis, aneurysms) were excluded. Observations were available for all diseases on all the days of the study period (310 total days). The unit of analysis was the daily number of admissions for males and females > 65 years of age per 10,000 residents. The numbers of residents in Denver per sex and age for the years 1993-1997 were derived from the Population Estimates Program of the U.S. Census Bureau (2000). From 1993 to 1997, the total population of males and females > 65 years of age has declined from about 64,000 to about 60,000. The percentage of males and females in the > 65 age group was 38% and 62%, respectively.

Environmental data. Air pollution data were extracted from the U.S. Environmental Protection Agency (EPA) Aerometric Information Retrieval System (AIRS) (U.S. EPA 2002). Daily concentrations of the U.S. EPA's criteria air pollutants were obtained from all of the monitoring stations in Denver County. These pollutants included N[O.sub.2], S[O.sub.2], [O.sub.3], CO, and particulate matter [less than or equal to] 10 [micro]m in aerodynamic diameter (P[M.sub.10]). The P[M.sub.10] level was collected over a 24-hr period, whereas gaseous air pollutants were measured hourly. For each day, the 24-hr arithmetic average reading was determined for each station, and then an arithmetic average was calculated over all stations. During July and August of 1993-1997, three monitoring stations measured P[M.sub.10] in Denver. P[M.sub.10] measurements were available for 298 (96%) of the possible days during the study periods (July-August 1993-1997). [O.sub.3] was monitored at two stations during the study periods, and data were available for 310 (100%) of the days. N[O.sub.2] and S[O.sub.2] were measured at one site in all 5 years, with 303 (98%) and 310 (100%) of the days available, respectively. CO was monitored at three stations from 1993 to 1995 and four stations in 1996 and 1997; data were available on all days during the study periods.

Daily 24-hr meteorologic measurements such as [T.sub.max] and dew point temperature (DPT) were provided by the National Climate Data Center (NCDC 2002). During July and August of 1993-1997, three measurement stations were located in Denver County. These meteorologic measurements were available on all days during the study period and were averaged over all stations.

Modeling approach. We used general linear models (GLMs; McCullagh and Nelder 1989) to analyze the frequency of hospital admissions for a given day as a function of environmental exposures. Poisson error structure was assumed because hospital admissions are generally rare events. In a classic Poisson regression, the model assumes

[1] log [E([Y.sub.i])] = [beta][X.sub.i],

where [X.sub.i] is the matrix of covariates on day i, [Y.sub.i], is the morbidity count on day i, and E denotes the expected value. Unlike the normal distribution, which is completely determined by its two parameters, the mean [micro] and the variance [[sigma].sup.2], the relationship between the mean and variance for a Poisson variate is a fixed relationship. Given a Poisson distribution with parameter [lambda], this relationship is [mu] = [[sigma].sup.2] = [lambda].

In the present study, explicit adjustments for seasonal patterns were not necessary because we used only data for July and August from each year. In addition, the daily number of hospital admissions for each disease and the daily levels of each environmental variable in July and August within any given year were assumed to be independent of daily disease counts and environmental levels from the previous years. Although the hospital admission and environmental data are treated as independent from year to year, within a 2-month period in a given year the number of hospital admissions and the levels of the environmental variables for a given day could be correlated with the data for each of these variables from previous days. To account for potential delays in disease incidence after important exposures, lag times of 1-4 days for each of the environmental variables were included as additional model covariates. Model covariates were also centered using their means. Besides an intercept term and indicator variables for sex (0 = female, 1 = male), day of the week, and year of study, all GLM models included an offset or normalizing factor to account for population changes during 1993-1997. The application of an offset is a computational procedure to treat the population size (PS) as a reference value and does not require an additional parameter in the Poisson regression:

[2] log [E([Y.sub.i)] = log(PS) [beta][X.sub.i].

Parameters in the Poisson regression models were estimated using the generalized estimating equations (GEEs) of Liang and Zeger (1986) to account for the possibility of overdispersion and serial correlation (Lipsitz et al. 1994; Schwartz 1993). Overdispersion refers to the increase in the variance of the distribution of the disease counts that can occur when the underlying population is not homogeneous with respect to risk of morbidity. Serial correlation refers to a situation where the daily number of hospital admissions on 2 days close to each other in time is correlated. The GEE for estimating [beta] is an extension of the independence estimating equation for correlated data. The covariance matrix of [Y.sub.i] is modeled as

[3] V = [phi][A.sup.1/2]R([alpha])[A.sup.1/2),

where A is an n x n diagonal matrix with the variance of [Y.sub.i] as the diagonal element, R([alpha]) is the n x n working correlation matrix that is fully specified by the vector of parameters [alpha], and [phi] is the overdispersion parameter (Liang and Zeger 1986; SAS/STAT Software 1997). The working correlation structure for daily number of hospital admissions within a year was assumed to be autoregressive with the following structure:

[4] corr([Y.sub.i],[Y.sub.i'] = [[alpha].sup.|i-i'|].

[alpha] is estimated empirically from the data, and [phi] is estimated from the residual [chi square] using the method of McCullagh and Nelder (1989). Missing values were treated as being missing completely at random and were dropped from the analyses, but appropriate care was taken to ensure that autoregressive parameters were estimated with the proper lags.

Before adding the air pollution variables to the models, the effects of the [T.sub.max] and DPT were investigated. For each cardiovascular disease, the daily admissions per 10,000 population were regressed against the [T.sub.max] and DPT with various lag times from 0 to 4 days. The lags for [T.sub.max] and DPT that minimized the deviance were kept 'fixed for the subsequent analyses incorporating the effect of the air pollutants. Residuals of each model were examined to check whether there were discernible patterns and autocorrelation by means of residual plots and partial autocorrelation function plots, respectively (Schwartz et al. 1996). Once we determined the optimal model for temperature-related effects on hospital admissions (core model), we examined the association between exposure to one air pollutant and daily numbers of hospital admissions per 10,000 population. Each air pollutant was entered linearly (with a log link function) into the regression, and lags between 0 and 4 days were examined. Because of the large number of pollutants and lags tested, emphasis was placed on the consistency of associations across lags and measures, and only the more highly significant findings are discussed. To examine the effect of a period of time with consistently above-median temperatures, the same analyses were carried out containing a variable describing the number of days in a row when maximum daily temperatures were greater than or equal to its 50th percentile. Additional analyses were conducted for each sex separately but resulted in insufficient sample sizes and are not reported here. Other strategies for model choice (e.g., forward, backward elimination) and autocorrelation formulations (exchangeable, unstructured) did not significantly alter the results and are not presented here.

We used PROC GENMOD (SAS System for Windows, Release 8.00; SAS Institute, Cary, NC) in the analysis.

Results

Table 1 shows the mean and variance of the daily numbers of hospital admissions for the cardiovascular diseases per 10,000 population in Denver during the study period. The highest mean daily rate of hospital admissions was observed in males for coronary atherosclerosis (0.74/10,000) and in females for congestive heart failure (0.52/10,000).

Figure 1 shows the annual daily average number of hospital admissions for cardiovascular diseases per 10,000 population for males and females. In general, there appears to be a greater incidence for males than for females, and for most of the diseases a slight increase in daily average number of hospital admissions per 10,000 population was observed during 1993-1997. As a result, the models for each disease include classification variables for annual trends and for sex.

[FIGURE 1 OMITTED]

Table 2 contains the summary statistics of the environmental data in Denver during the study period. Levels of pollutants were generally low if examined in light of U.S. EPA National Ambient Air Quality Standards (U.S. EPA 2001). All of the average pollutant levels would have easily met the standards.

Table 3 shows the correlations among the air pollution and weather variables. [O.sub.3] presented positive correlation with [T.sub.max], negative correlations with other environmental variables, and almost no correlation with P[M.sub.10](0.03). It showed the strongest correlation with [T.sub.max] (0.41), and CO (-0.40). N[O.sub.2] presented high positive correlations with CO (0.73), P[M.sub.10] (0.56), and S[O.sub.2] (0.46). The correlations observed for [O.sub.3] can be explained by the fact that [O.sub.3] is a summertime pollutant caused by the reaction of volatile organic compounds, CO, and N[O.sub.2] in the presence of sunlight. The correlations seen in Table 3 between P[M.sub.10], N[O.sub.2], and S[O.sub.2] probably exist because secondary particles are formed from gaseous precursors, including the S[O.sub.x] and N[O.sub.x]. No consistent trends over the 5 years of this study were seen in the average daily concentrations of the environmental variables.

The single-pollutant GLMs were evaluated based on sequentially identifying factors that significantly reduce the residual error. Table 4 summarizes the results of the single-pollutant analysis for the different cardiovascular diseases in the > 65-year-old population of Denver for up to 4 lag days. The most statistically significant association from the single-pollutant measures investigated, whether positive or negative, is presented. The coefficients from Table 4 can be interpreted in terms of increased daily hospital admissions. For example, the model for coronary atherosclerosis results in a statistically significant 2-day lag effect of [O.sub.3] on daily coronary atherosclerosis hospital admissions. An increase in the daily average level of [O.sub.3] from 20 ppb (25th percentile) to 29.7 ppb (75th percentile; Table 2) is statistically associated with an increase in daily hospital admissions for coronary atherosderosis of ([e.sup.0.012(29.7-20.0])-- 1) x 100% = 12.3% with a 95% confidence interval (CI) of 3.96-21.4%. The percentage change in risk and the 95% CI associated with a 25-75th percentile increase in [T.sub.max], DPT, and/or pollutant measure are reported together with the p-value in Table 4. The estimates for the overdispersion factor for the individual models (not shown) were nearly equal to 1.0, so adjustment of the standard errors was not necessary.

The results indicate that [T.sub.max] at either lag 0 or 1 has a significant association with all cardiovascular diseases except for cardiac dysrhythmias. [T.sub.max] is positively associated (increasing [T.sub.max] leads to increasing hospital admissions) with both acute myocardial infarction (p < 0.01) and congestive heart failure (p = 0.02). In contrast, the [T.sub.max] is negatively associated with the observed number of hospital admissions for coronary atherosclerosis (p < 0.01) and pulmonary heart disease (p < 0.01). A 25-75th percentile increase in [T.sub.max] (from 83 to 93.5[degrees]F) is associated with an 18, 13, 28, and 13% change in risk of hospitalization for acute myocardial infarction, coronary atherosclerosis, pulmonary heart disease, and congestive heart failure, respectively. [T.sub.max] has no significant association with cardiac dysrhythmia hospital admissions (p > 0.05). A 1-day lag DPT association was observed for coronary atherosclerosis (negative, p < 0.01) and for congestive heart failure (positive, p < 0.01). A 25-75th percentile increase in DPT (from 44.2 to 53.3[degrees]F) is associated with a 9 and 16% change in risk of hospitalization for coronary atherosclerosis and congestive heart failure, respectively.

Levels of [O.sub.3] are positively associated with the number of coronary atherosclerosis (lag 2, p < 0.01) and pulmonary heart disease (lag 1, p < 0.01) hospital admissions and negatively associated with the number of acute myocardial infarction hospital admissions (lag 0, p < 0.01). A 25-75th percentile increase in average daily 03 concentrations (from 20 to 29.7 ppb) is associated with an 18, 12, and 21% change in risk of hospitalization for acute myocardial infarction, coronary atherosclerosis, and pulmonary heart disease, respectively. Marginal positive associations are seen for current-day S[O.sub.2] concentrations with hospital admissions for cardiac dysrhythmias (p = 0.055) and for the level of CO with congestive heart failure (lag 3, p = 0.046). A 25-75th percentile increase in average daily S[O.sub.2] concentrations (from 3.8 to 7.2 ppb) is associated with a 9% change in risk of hospitalization for cardiac dysrhythmias. An increase from 0.8 to 1.1 ppm in average daily CO is associated with an 11% change in risk of hospitalization for congestive heart failure. No significant associations were found between particulate matter or N[O.sub.2] for any lag day and any of the health outcomes.

Sex is an important risk factor in all of the selected cardiovascular diseases, except for congestive heart failure. Males had higher numbers of hospital admissions than did females (p = 0.01). When exposed to both higher temperatures and [O.sub.3], males are associated with a 68, 82, and 17% increased risk of hospitalization for acute myocardial infarction, coronary atherosclerosis, and pulmonary heart disease, respectively. Also, males have a 23% increased risk of hospitalization for cardiac dysrhythmias compared with women.

Inclusion of a variable describing the number of days in a row when maximum daily temperatures were greater than or equal to its 50th percentile (88.5[degrees]F; Table 2) had no effect on the model parameter estimates.

Discussion and Conclusions

The purpose of this study was to determine if exposures to higher temperatures and air pollutant concentrations were significantly associated with hospital admissions for cardiovascular diseases. Hospital admissions represent morbidity data, rather than mortality data, and represent a more immediate response to higher temperatures and/or air pollutant concentrations. Studies of morbidity effects can strengthen the consistency of findings regarding the biologic plausibility of mortality effects of temperature and air pollution. This study was conducted during July--August, a period of time when outdoor levels of air pollution likely are a better reflection of actual exposure, given that individuals spend more time outdoors in summer than in winter, and the time when temperatures are usually the highest of the year.

From the results, it is clear that there is heterogeneity in the association of individual components of air pollution and temperature with the disease end points. No association was found between particulate matter or N[O.sub.2] and any of the health outcomes. A large number of daily time-series studies have reported associations between particulate matter and hospitalizations. Most of these studies have focused on respiratory hospital admissions rather than cardiovascular diseases and do not address an issue similar to the one studied in this study. Some recent studies have reported associations between particulate matter and hospitalizations for cardiovascular diseases (Schwartz 1999; Schwartz and Morris 1995), and others have not (Burnett et al. 1997a). Burnett et al. (1997a) could not identify particulate mass as an independent risk factor beyond that attributable to climate and gaseous air pollution. All of these studies have analyzed full-year data, whereas the present study focused on July--August only. A seasonal particulate matter effect might exist, as was seen in a study of Cook County, Illinois (Styer et al. 1995) in which a significant effect of particulate matter on cardiovascular mortality numbers was found in spring and fall, but no significant effect was found in winter and summer.

The significant associations between July--August concentrations of CO and specific cardiovascular diseases are confirmed in a large number of year-round studies (Ballester et al. 2001; Burnett et al. 1997b; Morris et al. 1995; Morris and Naumova 1998; Schwartz 1999; Schwartz and Morris 1995). Morris et al. (1995) found that CO showed a consistent, significant association with hospital admissions for congestive heart failure. Burnett et al. (1997b) demonstrated a similar effect in investigation of 10 Canadian cities. The limited data from laboratory studies on the combined effects of air pollution with either heat stress or cold stress seem to support the possibility of enhanced toxicity of CO associated with extreme temperatures (Yang et al. 1988). Studies in Chicago, Illinois (Morris and Naumova 1998), and Tucson, Arizona (Schwartz 1997), failed to demonstrate an increase in the effect of CO at high temperatures. It is possible that the cut point for the high temperature range was not sufficiently high in the Chicago study, whereas the summer CO levels were not high enough in the Tucson study.

We found a significant negative association of [O.sub.3] with hospital admissions for acute myocardial infarction and a positive association for coronary atherosclerosis and pulmonary heart disease. Only three studies that we reviewed showed evidence of a consistent association between [O.sub.3] and cardiovascular admissions. Prescott et al. (1998) found that cardiovascular admissions for patients [greater than or equal to] 65 years of age had a significant (negative) association with [O.sub.3]. Burnett et al. (1997a) showed that [O.sub.3] was most strongly associated with hospitalization for cardiac diseases in Toronto in the summer. In Hong Kong, Wong et al. (1999) found a significant relation of [O.sub.3] with hospital admissions for cardiovascular diseases. However, most of the reviewed studies do not suggest an impact of [O.sub.3] on the number of hospitalizations attributable to cardiovascular disease (Ballester et al. 2001; Morris et al. 1995; Morris and Naumova 1998; Schwartz 1997; Schwartz and Morris 1995).

It is noteworthy to mention the association between S[O.sub.2] and cardiac dysrhythmias. The physiopathologic mechanisms of S[O.sub.2] causing cardiovascular damage are insufficiently understood, and more research is needed in this area. Only two studies .that we reviewed showed S[O.sub.2] as being positively associated with daily hospital admissions for cardiac diseases (Burnett et al. 1997a; Diaz et al. 2001). It is possible that, in this case, S[O.sub.2] is serving as a general indicator of air pollution exposure rather than a direct causal initiator of cardiovascular damage.

Sex appeared to. be an important risk factor in all of the selected cardiovascular diseases except for congestive heart failure. Males tend to have higher numbers of hospital admissions than do females.

Higher temperatures a) appear to be an important factor in increasing the frequency of hospitalization for acute myocardial infarction and congestive heart failure; b) are associated with a decrease in the frequency of visits for coronary atherosclerosis and pulmonary heart disease; and c) have no significant effect on visits for cardiac dysrhythmias. It is possible that patients with chronic cardiovascular conditions (coronary atherosclerosis and pulmonary heart disease) avoid outdoor exposures during periods of peak heat, resulting in the negative association. In contrast, exposures to higher air pollutant concentrations (except for particulate matter and N[O.sub.2] appear to have a positive effect on the number of hospital admissions for cardiovascular diseases. According to Bernard et al. (2001), climate change may affect exposures to air pollutants by a) affecting weather and thereby local and regional pollution concentrations; b) affecting anthropogenic emissions, including adaptive responses involving increased fuel combustion for power generation; c) affecting natural sources of air pollutant emissions; and d) changing the distribution and types of aeroallergens. According to Bernard et al. (2001), it is expected that, in the future, the pollutants of concern are likely to remain, as now, particulate matter--both from its primary sources and from the secondarily formed particles--and photochemical pollution with [O.sub.3]. Emissions of S[O.sub.x] and N[O.sub.x]due to power generation might increase, but control technologies are available and should be able to control such emissions.

The present study is limited because it uses environmental monitoring data to represent ambient concentrations. The fact that different pollutants have not been measured at the same monitoring stations (e.g., N[O.sub.2] was measured only at one site, and other pollutants at three or four sites) can affect the estimated correlations between the pollutants. Besides this, environmental monitoring data do not necessarily represent individual exposures. This could result in associations between exposures and hospital admissions that are spurious. Also, measurement errors resulting from differences between the population-average exposure and ambient levels cannot be avoided. However, this kind of measurement error is likely to cause a bias toward the null and underestimates the pollutant effects (Zeger et al. 2000). Many individual risk factors, such as smoking, diet, cholesterol level, and obesity, either do not vary significantly over time or vary slowly. In addition, there is no reason to believe that daily variations in the individual risk factors are correlated with daily changes in air pollution; therefore, they are unlikely to be confounding factors in this study (Schwartz 1997). It is possible that with accurate, individual exposures, the major associations seen in this study may not be all of the associations that are important.

Care should be taken with drawing conclusions about the impact of temperature on morbidity because the estimated adverse health effects from exposures to higher temperatures and air pollutants are small. Uncertainties inherent in the analyses go far beyond the sampling variability that is captured by the standard error of the confidence interval. These uncertainties arise from model misspecification, omitted covariates, and errors in the measurement of covariates considered in the models. Care should also be taken in comparing different studies because discrepancies may reflect regional differences in pollution mixture, climate, flora and fauna, period of data collection, measurement methods of air pollutants, or medical practice patterns. Time series do not give precise information on the relationship between air pollution and health for a particular individual. However, they provide valuable information on the overall health impact of air pollution as measured at fixed-site monitoring stations.

In summary, the results of this study in Denver suggest that [O.sub.3] increases the risk of hospitalization for acute myocardial infarction, coronary atherosclerosis, and pulmonary heart disease. S[O.sub.2] appears to be related to increased hospital stays for cardiac dysrhythmias, and CO is significantly associated with congestive heart failure hospitalization. No association was found between particulate matter or N[O.sub.2] and hospitalization for any condition. Males had more hospital admissions for cardiovascular diseases than did females. Higher temperatures appear to be an important factor in increasing the frequency of hospitalization for acute myocardial infarction and congestive heart failure. However, higher temperatures are associated with a decrease in the frequency of visits for coronary atherosclerosis and pulmonary heart disease and have no significant effect on admissions for cardiac dysrhythmias. In contrast, exposures to higher air pollutant concentrations (except for particulate matter and N[O.sub.2]), even at levels that meet federal air quality standards, appear to have an effect of increasing the number of hospital admissions for cardiovascular diseases as a whole.
Table 1. Means and variances of the daily numbers of hospital
admissions for males and females > 65 years of age per 10,000
population in Denver, July-August 1993-1997.

 Males Females

Disease Mean Variance Mean Variance

Acute myocardial infarction 0.51 0.20 0.31 0.08
Coronary atherosclerosis 0.74 0.36 0.40 0.10
Pulmonary heart disease 0.08 0.03 0.06 0.02
Cardiac dysrhythmias 0.36 0.14 0.29 0.08
Congestive heart failure 0.48 0.19 0.52 0.15

Table 2. Summary statistics for daily levels of environmental
variables in Denver, July-August 1993-1997.

 Percentile

Variable Source No. Mean

N[O.sub.2] (ppb) AIRS 303 32.7
S[O.sub.2] (ppb) AIRS 310 5.7
[O.sub.3] (ppb) AIRS 310 25.0
CO (ppm) AIRS 310 0.9
P[M.sub.10] ([micro]g/[m.sup.3]) AIRS 298 24.2
[T.sub.max] ([degrees]F) NCDC 310 87.4
DPT ([degrees]F) NCDC 310 48.6

 Percentile

Variable Minimum 25th 50th

N[O.sub.2] (ppb) 13.0 28.0 33.2
S[O.sub.2] (ppb) 0.4 3.8 5.3
[O.sub.3] (ppb) 5.4 20.0 25.2
CO (ppm) 0.3 0.8 0.9
P[M.sub.10] ([micro]g/[m.sup.3]) 7.0 20.0 24.0
[T.sub.max] ([degrees]F) 62.0 83.0 88.5
DPT ([degrees]F) 26.8 44.2 49.3

 Percentile

Variable 75th Maximum SD

N[O.sub.2] (ppb) 37.9 51.3 7.24
S[O.sub.2] (ppb) 7.2 18.9 2.94
[O.sub.3] (ppb) 29.7 40.2 6.61
CO (ppm) 1.1 1.6 0.27
P[M.sub.10] ([micro]g/[m.sup.3]) 28.0 51.6 6.25
[T.sub.max] ([degrees]F) 92.5 99.0 7.37
DPT ([degrees]F) 53.3 62.4 6.23

Data from U.S. EPA (2002) and NCDC (2002).

Table 3. Pearson correlation coefficients of environmental variables
for Denver, July-August 1993-1997.

 [T.sub.max] DPT N[O.sub.2] S[O.sub.2]

[T.sub.max] 1.00 -0.42 0.34 0.26
DPT -- 1.00 -0.12 -0.15
N[O.sub.2] -- -- 1.00 0.46
S[O.sub.2] -- -- -- 1.00
[O.sub.3] -- -- -- --
CO -- -- -- --
P[M.sub.10] -- -- -- --

 [O.sub.3] CO P[M.sub.10]

[T.sub.max] 0.41 0.11 0.38
DPT -0.26 0.01 * -0.24
N[O.sub.2] -0.30 0.73 0.56
S[O.sub.2] -0.10 0.21 0.36
[O.sub.3] 1.00 -0.40 0.03 *
CO -- 1.00 0.25
P[M.sub.10] -- -- 1.00

* p > 0.05.

Table 4. Summary of results for single-pollutant analysis of hospital
admissions for cardiovascular diseases for males and females > 65
years of age in Denver, July-August 1993-1997.

 Model
Disease covariate Single-day lag [beta] (SE)

Acute myocardial [T.sub.max] 0 0.017 (0.007)
infarction [O.sub.3] 0 -0.020 (0.006)
 Male sex -- 0.517 (0.061)
Coronary [T.sub.max] 1 -0.014 (0.004)
athergsclerosis DPT 1 -0.010 (0.003)
 [O.sub.3] 2 0.012 (0.004)
 Male sex -- 0.601 (0.038)
Pulmonary heart [T.sub.max] 0 -0.035 (0.008)
disease [O.sub.3] 1 0.020 (0.008)
 Male sex -- 0.160 (0.062)
Cardiac S[O.sub.2] 0 0.025 (0.013)
dysrhythmias Male sex -- 0.203 (0.028)
Congestive heart [T.sub.max] 1 0.013 (0.005)
failure DPT 1 0.016 (0.004)
 CO 3 0.332 (0.166)

 Percent change in hospital
Disease admissions (95% CI) p-Value

Acute myocardial 17.5 (2.9-34.3) 0.009
infarction -17.6 (-26.7-7.5) 0.001
 67.7 (48.4-89.5) < 0.0001
Coronary -12.5 (-18.9-5.5 0.001
athergsclerosis -8.7 (-13.6-3.6) 0.003
 12.3 (4.0-21.4) 0.004
 82.4 (69.0-96.8) < 0.0001
Pulmonary heart -28.3 (-38.4-16.5) < 0.0001
disease 21.4 (4.0-41.8) 0.009
 17.4 (3.7-32.8) 0.010
Cardiac 8.9 (-0.34-18.93) 0.055
dysrhythmias 22.5 (15.8-29.6) < 0.0001
Congestive heart 13.2 (2.9-24.4) 0.016
failure 15.7 (7.6-24.4) < 0.0001
 10.5 (0.1-22.0) 0.046

CI, confidence interval. The percentage change in the number of
admissions for a 25-75th percentile change in each pollutant (or for
males compared with females) is presented at its most statistically
significant single lag.


REFERENCES

Atkinson RW, Bremner SA, Anderson HR, Strachan DP, Bland JM, de Leon AP. 1999. Short-term associations between emergency hospital admissions for respiratory and cardiovascular disease and outdoor air pollution in London. Arch Environ Health 54:398-411.

Ballester F, Tenias JM, Perez-Hoyos S. 2001. Air pollution and emergency hospital admissions for cardiovascular diseases in Valencia, Spain. J Epidemiol Community Health 55:57-65.

Bernard SM, Samet JM, Grambsch A, Ebi KL, Romieu I. 2001. The potential impacts of climate variability and change on air pollution-related health effects in the United States. Environ Health Perspect 109:199-209.

Braga ALF, Zanobetti A, Schwartz J. 2001. The time course of weather related deaths. Epidemiology 12:662-667. --. 2002. The effect of weather on respiratory and cardiovascular deaths in 12 U.S. cities. Environ Health Perspect 110:859-863.

Burnett RT, Cakmak S, Brook JR, Krewski D. 1997a. The role of particulate size and chemistry in the association between summertime ambient air pollution and hospitalization for cardiorespiratory diseases. Environ Health Perspect 105:614-620.

Burnett RT, Dales RE, Brook JR, Raizenne ME, Krewski D. 1997b. Associations between ambient carbon monoxide levels and hospitalizations for congestive heart failure in the elderly in 10 Canadian cities. Epidemiology 8:162-167. Centers for Medicare & Medicaid Services. 2003. ICD-9-CM Coordination and Maintenance Committee: Available: http://cms.hhs.gov/paymentsystems/icd9 [accessed 23 June 2003].

Diaz J, Alberdi JC, Pajares MS, Lopez C, Lopez R, Lage MB, et al. 2001. A model for forecasting emergency hospital admissions: effect of environmental variables. J Environ Health 54:9-15.

Dockery DW, Pope CA, Speizer FE. 2000. Effects of particulate air pollution exposure. In: Particle-Lung Interactions (Gehr P, Heyder J, eds). New York:Marcel Dekker, 671-703.

Firket M. 1931. The cause of the symptoms found in the Meuse Valley during the fog of December, 1930. Bull Acad Roy Med Belg 11:683-741.

HCUP 2001. Health Care Cost and Utilization Project 1988-2000. A Federal-State-Industry Partnership in Health Data. Rockville, MD:Agency for Healthcare Research and Quality. Available: http://www.ahrq.gov/data/hcup/hcuppkt.htm [accessed 28 August 2002].

Hoek G, Brunekreef B, Fischer P, van Wijnen J. 2001. The association between air pollution and heart failure, arrhythmia, embolism, thrombosis, and other cardiovascular causes of death in a time series study. Epidemiology 12:355-357.

IPCC. 1996. Climate Change 1995. The Science of Climate Change. Cambridge, UK:Cambridge University Press. Katsouyanni K, Pantazopoulou A, Touloumi G, Tselepidaki I, Moustris K, Asimakopoulos D, et al. 1993. Evidence for interaction between air-pollution and high temperature in the causation of excess mortality. Arch Environ Health 48:235-242.

Keatinge WR, Coleshaw SR, Easton JC, Cotter F, Mattock MB, Chelliah R. 1986. Increased platelet and red cell counts, blood viscosity, and plasma cholesterol levels during heat stress, and mortality from coronary and cerebral thrombosis. Am J Med 81:795-800.

Kunst AE, Looman CW, Mackenbach JP. 1993. Outdoor air temperature and mortality in The Netherlands: a time-series analysis. Am J Epidemiol 137:331-341.

Liang KY, Zeger SL. 1986. Longitudinal data-analysis using generalized linear-models. Biometrika 73:13-22.

Lipsitz SR, Fitzmaurice GM, Orav EJ, Laird NM. 1994. Performance of generalized estimating equations in practical situations. Biometrics 50:270-278.

McCullagh P, Nelder JA. 1989. Generalized Linear Models. London:Chapman and Hall.

McMichael AJ, Ando M, Carcavallo R, Epstein P, Haines A, Jendritsky G, et al. 1996. Climate Change and Human Health. Geneva:World Health Organization.

Ministry of Health. 1954. Mortality and Morbidity during the London Fog of December 1952. Reports on Public Health and Medical Subjects no. 95. London:HMSO.

Morris RD, Naumova EN. 1998. Carbon monoxide and hospital admissions for congestive heart failure: evidence of an increased effect at low temperatures. Environ Health Perspect 106:649-653.

Morris RD, Naumova EN, Munasinghe RL. 1995. Ambient air pollution hospitalization for congestive heart failure among elderly people in seven United States cities. Am J Public Health 85:1361-1365.

NCDC 2002. Online Climate Data. Asheville, NC:National Climate Data Center. Available: http://www.ncdc.noaa.gov/ol/ climate/climatedata.html [accessed 28 August 2002].

Peters A, Liu E, Verrier RL, Schwartz J, Gold DR, Mittleman M, et al. 2000. Air pollution and incidence of cardiac arhythmia. Epidemiology 11:11-17.

Piver WT, Ando M, Ye F, Pottier CJ. 1999. Temperature and air pollution as risk factors for heat stroke in Tokyo, July and August 1980-1995. Environ Health Perspect 107:911-916.

Prescott GJ, Cohen GR, Elton RA, Fowkes FG, Agius RM. 1998. Urban air pollution and cardiopulmonary ill health: a 14.5 year time series study. Occup Environ Med 55:697-704.

Saez M, Sunyer J, Castellsague J, Murillo C, Anto JM. 1995. Relationship between weather temperature and mortality: a time series analysis approach in Barcelona. Int J Epidemiol:24:576-582.

SAS/STAT Software. 1997. The GENMOD procedure. In: Changes and Enhancements through Release 6.12. Cary, NC:SAS Institute Inc, 247-348.

Schwartz J. 1993. Air-pollution and daily mortality in Birmingham, Alabama. Am J Epidemiol 137:1136-1147. --. 1997. Air pollution and hospital admissions for cardiovascular disease in Tucson. Epidemiology 8:371-377. --. 1999. Air pollution and hospital admissions for heart disease in eight U.S. counties. Epidemiology 10:17-22.

Schwartz J, Morris R. 1995. Air-pollution and hospital admissions for cardiovascular disease in Detroit, Michigan. Am J Epidemiol 142:23-35.

Schwartz J, Spix C, Touloumi G, Bacharova L, Barumamdzadeh T, leTertre A, et al. 1996. Methodological issues in studies of air pollution and daily counts of deaths or hospital admissions. J Epidemiol Community Health 50:S3-S11.

Seaton A, MacNee W, Donaldson K, Godden D. 1995. Particulate air pollution and acute health effects. Lancet 345:176-178.

Shrenk HM, Heiman H, Clayton GD. 1949. Air Pollution in Donara, PA: Epidemiology of the Unusual Smog Episode of October 1948. Preliminary Report no. 306. Washington, DC:U.S. Public Health Service.

Styer P, McMillan N, Gao F, Davis J, Sacks J. 1995. Effect of outdoor airborne particulate matter on daily death counts. Environ Health Perspect 103:490-497.

U.S. Census Bureau 2000. Population Estimates for Counties by Age and Sex: Annual Time Series July 1, 1990 to July 1, 1999. Washington, DC:Population Estimates Program, Population Division. Available: http://www.census.gov/ population/estimates/county/cas/cas08.txt [accessed 28 August 2002].

U.S. EPA. 2001. National Ambient Air Quality Standards (NAAQS). Research Triangle Park, NC:U.S. Environmental Protection Agency. Available: http://www.epa.gov/airs/ criteria.html [accessed 28 August 2002].

--. 2002. Aerometric Information Retrieval System. Research Triangle Park, NC:U.S. Environmental Protection Agency. Available: http://www.epa.gov/air/data/index.html [accessed 28 August 2002].

van der Bom JG, de Maat MPM, Bots ML, Hofman A, Kluft C, Grobbee DE. 1997. Seasonal variation in fibrinogen in the Rotterdam study. Thromb Haemost 78:1059-1062.

Wong TW, Lau TS, Yu TS, Neller A, Wong SL, Tam W, et al. 1999. Air pollution and hospital admissions for cardiovascular diseases in Hong Kong. Occup Environ Med 56:679-683.

Yang L, Zhang W, He HZ, Zhang GG. 1988. Experimental studies of combined effects of high temperature and carbon monoxide. J Tongji Med Univ 8:60-65.

Zanobetti A, Schwartz J, Dockery DW. 2000. Airborne particles are a risk factor for hospital admissions for heart and lung disease. Environ Health Perspect 108:1071-1077.

Zeger SL., Thomas D, Dominici F, Samet JM, Schwartz J, Dockery D, et al. 2000. Exposure measurement error in time-series studies of air pollution: concepts and consequence. Environ Health Perspect 108:419-426.

Petra J. M. Koken, (1) Warren T. Piver, (1) Frank Ye, (1) Anne Elixhauser, (2) Lola M. Olsen, (3) and Christopher J. Portier (1)

(1) National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina, USA; (2) Agency for Healthcare Research and Quality, Rockville, Maryland, USA; (3) National Aeronautics and Space Administration, Greenbelt, Maryland, USA

Address correspondence to P.J.M. Koken, National Institute of Environmental Health Sciences, MD A3-06, PO Box 12233, Research Triangle Park, NC 27709 USA. Telephone: (919) 541-2271. Fax: (919) 541-1479. E-mail: koken@niehs.nih.gov

We are grateful for financial support for this project from NASA, Order S-10152-X.

The authors declare they have no conflict of interest. Received 27 August 2002; accepted 19 May 2003.
COPYRIGHT 2003 National Institute of Environmental Health Sciences
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2003, Gale Group. All rights reserved. Gale Group is a Thomson Corporation Company.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Article
Author:Portier, Christopher J.
Publication:Environmental Health Perspectives
Date:Aug 1, 2003
Words:7580
Previous Article:Assessment of dietary exposure to some persistent organic pollutants in the Republic of Karakalpakstan of Uzbekistan.
Next Article:Assessment of potential risk levels associated with U.S. Environmental Protection Agency reference values.


Related Articles
Air Pollution and Daily Hospital Admissions in Metropolitan Los Angeles.
Are There Sensitive Subgroups for the Effects of Airborne Particles?
Effects of Temperature and Air Pollutants on Cardiovascular and Respiratory Diseases for Males and Females Older than 65 Years of Age in Tokyo, July...
The influence of ambient coarse particulate matter on asthma hospitalization in children: case-crossover and time-series analyses. (Children's Health...
Air pollution and hospital admissions for ischemic heart disease in persons with congestive heart failure or arrhythmia.
Cause-specific mortality and the extended effects of particulate pollution and temperature exposure.
Air pollution is a serious cardiovascular risk.
Association of ambient air pollution with respiratory hospitalization in a government-designated "area of concern": the case of Windsor, Ontario.
The effect of particulate air pollution on emergency admissions for myocardial infarction: a multicity case-crossover analysis.
The effects of air pollution on hospitalizations for cardiovascular disease in elderly people in Australian and New Zealand cities.

Terms of use | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters