Air Pollution and Daily Hospital Admissions in Metropolitan Los Angeles.

We used daily time-series analysis to evaluate associations between ambient carbon monoxide, nitrogen dioxide, particulate matter [is greater than or equal to] 10 [micro]m in aerodynamic diameter ([PM.sub.10]), or ozone concentrations, and hospital admissions for cardiopulmonary illnesses in metropolitan Los Angeles during 1992-1995. We performed Poisson regressions for the entire patient population and for subgroups defined by season, region, or personal characteristics, allowing for effects of temporal variation, weather, and autocorrelation. CO showed the most consistently significant (p [is less than] 0.05) relationships to cardiovascular admissions. A wintertime 25th-75th percentile increase in CO (1.1-2.2 ppm) predicted an increase of 4% in cardiovascular admissions. [NO.sub.2], and, to a lesser extent, [PM.sub.10] tracked CO and showed similar associations with cardiovascular disease, but [O.sub.3] was negatively or nonsignificantly associated. No significant demographic differences were found, although increased cardiovascular effects were suggested in diabetics, in whites and blacks (relative to Hispanics and Asians), and in persons older than 65 years of age. Pulmonary disease admissions associated more with [NO.sub.2] and [PM.sub.10] than with CO. Pulmonary effects were generally smaller than cardiovascular effects and were more sensitive to the choice of model. We conclude that in Los Angeles, atmospheric stagnation with high primary (CO/[NO.sub.2]/[PM.sub.10]) pollution, most common in autumn/winter, increases the risk of hospitalization for cardiopulmonary illness. Summer photochemical pollution (high [O.sub.3]) apparently presents less risk. Key words: air pollutants, carbon monoxide, epidemiology, Los Angeles, morbidity, nitrogen dioxide, ozone, particulate matter. Environ Health Perspect 108:427-434 (2000). [Online 27 March 2000] http://ehpnet1.niehs.nih.gov/docs/2000/108p427-434linn/abstract.html

Time-series analyses of daily mortality or morbidity have shown statistical associations with air pollution in cities throughout the world. Physiologic/toxicologic mechanisms of these phenomena remain unknown, and time-series analyses have not clearly linked specific pollutants with specific health outcomes (1-3); thus, their application to pollution-control policy decisions remains controversial (4). Combustion-related particulate matter, the only pollutant common to virtually all locations of time-series studies, has been the focus of scientific and regulatory attention (1,2,4-6). However, recent studies in a number of North American cities also associate cardiovascular and/or pulmonary disease incidence with pollutant gases such as carbon monoxide, nitrogen dioxide, and/or ozone (7-12). Where they are not highly correlated, gas and particulate pollutants appear to have separate statistically and medically significant influences on cardiopulmonary morbidity (9,11,12).

The Los Angeles metropolitan area has been studied relatively little by time-series analysis, but is a good candidate for study because of its large diverse population ([approximately equals] 14 million); detailed monitoring of air quality and hospital admissions; mild climate, which should limit confounding of pollution effects by weather stresses; and severe but widely variable air pollution (with maximum levels of primary pollutant gases, secondary photochemical oxidant gases, and particulate pollution occurring at somewhat different times and places). Powerful tests of pollution effects should be possible in the entire metropolitan population and in subpopulations defined geographically, demographically, or clinically. We hypothesized that regional and/or seasonal differences in time-series analysis results in the general population and/or in particular subgroups, would allow us to distinguish effects associated with primary pollutants (CO or [NO.sub.2]), photochemical oxidants ([O.sub.3]), or particulate matter more clearly than has been possible elsewhere. If so, we could rank these categories of pollution in terms of their public health impact, and thus provide useful guidance for regulatory policymaking and for future research on mechanisms. To test this hypothesis, we analyzed daily admission data for 1992-1995 from the South Coast Air Basin (Los Angeles, Riverside, San Bernardino, and Orange Counties in California, excluding mountain and desert regions of the first three counties) in relation to daily levels of CO, [NO.sub.2], [O.sub.3], and particulate matter [is less than or equal to] 10 [micro]m in aerodynamic diameter ([PM.sub.10]).

Methods

Data acquisition and management. After its institutional review board verified confidentiality protection, the California Office of Statewide Health Planning and Development (OSHPD) (Sacramento, CA) provided records of hospital admissions in the metropolitan counties for 1992-1995 (the only years with adequately comparable [PM.sub.10] data). The records included hospital identifier, date, principal and additional diagnoses as International Classification of Diseases (ICD; World Health Organization, Geneva) codes, All-Patient-Refined Diagnosis-Related Group (APR-DRG; 3M Inc., Murray, UT)--a broader classification based on Medicare diagnosis-related groups, sex, age, ethnic group, and residence zip code. Daily counts after 21 December 1995 were excluded from analysis because the records for numerous patients not discharged until 1996 were missing, and all 1995 data were excluded from ethnic-group analyses because of changes in OSHPD ethnic classifications. Broad principal-diagnosis categories used in analyses were cardiovascular (APR-DRG 103-144); cerebrovascular (APR-DRG 14-17 and 22); pulmonary (APR-DRG 75-101); and abdominal--a negative control category thought to be unrelated to pollution (APR-DRG 146-207). More-specific principal diagnoses, thought likely to associate with air pollution on the basis of previous epidemiologic or toxicologic evidence, were also analyzed: congestive heart failure (CHF) (APR-DRG 127); myocardial infarction (APR-DRG 111, 115, and 121); cardiac arrhythmia (APR-DRG 138); occlusive stroke (APR-DRG 14); asthma (ICD 493); and chronic obstructive pulmonary disease (COPD) (APR-DRG 88). Analyses excluded patients younger than 30 years of age (with exceptions noted in "Results") and prescheduled admissions.

We obtained air pollution and meteorologic data from the South Coast Air Quality Management District (SCAQMD; Diamond Bar, CA) and from the National Weather Service. These data included hourly [PM.sub.10] from six SCAQMD stations with continuous monitors; hourly CO, [O.sub.3], [NO.sub.2], temperature, and relative humidity from those stations plus others; 24-hr-average [PM.sub.10] measured every sixth day by high-volume samplers at or near each continuous [PM.sub.10] station; and barometric pressure and rainfall at the Los Angeles International Airport. Figure 1 shows monitor locations. Analyses related daily admission counts with 24-hr averages of environmental variables. For [O.sub.3], maximum hourly concentrations were also analyzed; they correlated highly with 24-hr averages in all seasons (r [is greater than or equal to] 0.79) and showed similar relationships to daily morbidity ("Results"). We did not analyze relative humidity because many data were missing or out of range. Stations differed in their continuous [PM.sub.10] monitoring techniques and their relationships of continuous to high-volume sampler data. On the assumption that high-volume data were more comparable throughout the basin, we used season- and station-specific linear regressions to adjust continuous data to conform with high-volume data. We defined seasons to begin 1 January (winter), 1 April (spring), 1 July (summer), and 1 October (autumn).

[Figure 1 ILLUSTRATION OMITTED]

Geographic differences were investigated across six regions defined by continuous [PM.sub.10] monitoring stations (Figure 1). A region consisted of all zip codes that had a majority of their area closest to its station, except that some western coastal zip codes, which were separated from their closest station (region 1) by mountains, were assigned to region 2 to better represent their air quality. Admitted patients were assigned to regions by their residence zip codes. We excluded the 6.7% with zip codes missing or outside the South Coast Air Basin from regional analyses. We determined pollutant gas concentrations and temperatures for each region by averaging across all monitoring stations within it. Missing air monitoring data (4.4% for [PM.sub.10], smaller percentages for other variables) were replaced using analysis of variance with maximum likelihood estimation of missing values. For [PM.sub.10], data from all days in the same season and all stations in the basin were used in estimation; for other variables, only stations in the same region were used.

Statistical analyses. We used BMDP software (SPSS Inc., Chicago) and SAS software (SAS Institute, Inc., Cary, NC) for statistical analyses. Descriptive statistics and correlation patterns were examined regionally and seasonally for admission counts and atmospheric variables. Further descriptive analyses were performed to contrast weather and pollution characteristics between days with unusually high and unusually low observed morbidity relative to predicted values accounting for cyclical and secular trends. Predicted values were from regressions with indicator variables for the day of the week and for weekday holidays, with longer term variation modeled by fitting cubic splines to successive 28-day intervals of data. We then compared weather and pollution statistics between days with high admissions (residual [is greater than] 85th percentile) and low admissions (residual [is less than] 15th percentile), as well as the immediately preceding days.

Time-series analytical approaches included a) ordinary least squares (OLS) regression with admission count and atmospheric data filtered by the Shumway 19-day weighted moving average procedure (13), with or without an autoregressive component; b) regression of log-transformed daily admission counts using polynomial distributed lag models (14,15); and c) Poisson regression with allowance for overdispersion and autocorrelation, adapted from the analytical strategy of the Particle Epidemiology Evaluation Project (16) with modifications. In principle, daily counts are Poisson distributed and require approach c; however, given the generally large counts with filtering or smoothing, distributions were reasonably near normal so that other approaches were also feasible (17). The different approaches yielded similar conclusions when considering cardiovascular diseases. Polynomial-distributed lag models showed the largest significant effects consistently at lag 0, and effects beyond lag 1 were nearly always nonsignificant. Therefore, we adopted Poisson regression as the primary analytical tool. Predictors of daily admission counts included basis variables of a cubic-spline smooth on time (which accounted for secular trends and seasonal variation); indicator variables for the day of the week and for weekday holidays; indicator variables for hot days (maximum temperature [is greater than] 85th percentile for entire study period), cold days (minimum temperature [is less than] 15th percentile), and rain days ([is greater than] 0.01 inches at the Los Angeles International Airport); continuous atmospheric variables (one or more pollutant concentrations, barometric pressure, and mean temperature); and an autoregressive term--the residual admission count at lag 1, determined in a preliminary regression including all other predictors. Seasonal variation was more complex for pulmonary diseases than for others, probably because the timing and intensity of winter infectious disease outbreaks varied from year to year. Thus, cubic splines were determined at 28-day intervals when smoothing pulmonary disease counts, and at 4-month intervals otherwise.

Results

Seasonal air quality and hospital admissions. Table 1 presents seasonal pollution, weather, and hospital admission statistics for the entire basin for 1992-1995. Overall means [+ or -] SDs were 1.5 [+ or -] 0.8 ppm for CO, 3.4 [+ or -] 1.3 parts per hundred million (pphm) for [NO.sub.2], 45 [+ or -] 18 [micro]g/[m.sup.3] for [PM.sub.10], and 2.4 [+ or -] 1.2 pphm for [O.sub.3]. We determined basinwide means from the six regional values by weighting each region according to its proportion of cardiovascular plus pulmonary admissions (considered to reflect its proportion of the population at risk). These levels reflect a [is greater than] 80% reduction of CO since the 1960s (18), and more modest reductions in the other pollutants. Year-round means and SDs of daily admissions were 428 [+ or -] 76 for cardiovascular, 207 [+ or -] 54 for pulmonary, 74 [+ or -] 14 for cerebrovascular, and 244 [+ or -] 39 for abdominal diseases. Seasonal means for abdominal diseases (not tabulated) varied [is less than] 3%. All disease categories showed marked variation by day of the week, consistent across seasons. Relative to Monday admissions, Sunday admissions averaged 64% for cardiovascular, 70% for pulmonary, 76% for cerebrovascular, and 67% for abdominal diseases.

Table 1. Air(a) and hospital admission statistics for the entire South Coast Air Basin, by season.

Variable
(units)                Season    Mean [+ or -] SD    Min    Max

CO                     Winter    1.7 [+ or -] 0.8    0.5    5.3
(ppm)                  Spring    1.0 [+ or -] 0.3    0.4    2.2
                       Summer    1.2 [+ or -] 0.4    0.3    2.7
                       Autumn    2.1 [+ or -] 0.8    0.6    4.3

[NO.sub.2]             Winter    3.4 [+ or -] 1.3    1.1    9.1
(pphm)                 Spring    2.8 [+ or -] 0.9    1.1    6.1
                       Summer    3.4 [+ or -] 1.0    0.7    6.7
                       Autumn    4.1 [+ or -] 1.4    1.6    8.4

[PM.sub.10]            Winter     37 [+ or -] 19       5    115
([micro]g/[m.sup.3])   Spring     42 [+ or -] 12      14     83
                       Summer     49 [+ or -] 10      14     78
                       Autumn     54 [+ or -] 22      15    132

[O.sub.3]              Winter    1.4 [+ or -] 0.7    0.2    4.4
(pphm)                 Spring    3.2 [+ or -] 1.0    0.9    7.0
                       Summer    3.3 [+ or -] 0.8    0.4    6.3
                       Autumn    1.5 [+ or -] 0.9    0.1    4.7

Temperature            Winter   14.8 [+ or -] 2.7    8.4   23.2
(mean, [degrees] C)    Spring   19.2 [+ or -] 3.0    9.5   29.2
                       Summer   23.9 [+ or -] 2.3   19.0   31.2
                       Autumn   16.4 [+ or -] 3.7    8.7   26.7

Rain                   Winter     28
(% of days)            Spring      4
                       Summer      1
                       Autumn     10

Cardiovascular         Winter    450 [+ or -] 77     300    607
adm/day                Spring    428 [+ or -] 76     277    586
                       Summer    406 [+ or -] 70     239    559
                       Autumn    428 [+ or -] 76     273    610

Pulmonary              Winter    241 [+ or -] 54     117    574
adm/day                Spring    196 [+ or -] 34     118    329
                       Summer    172 [+ or -] 27     107    256
                       Autumn    220 [+ or -] 63     115    595

Cerebrovascular        Winter     77 [+ or -] 14      44    126
adm/day                Spring     74 [+ or -] 14      43    116
                       Summer     72 [+ or -] 14      41    114
                       Autumn     75 [+ or -] 14      45    117

Abbreviations: adm, admissions; max, maximum; min, minimum; pphm, parts per hundred million.

(a) Pollutant concentrations are averaged across six regions, each weighted according to its proportion of cardiovascular plus pulmonary admissions.

Table 2 shows pairwise correlations of basinwide average daily pollutant concentrations, mean temperature, and barometric pressure within each season. [NO.sub.2] showed high positive correlations (r [is greater than or equal to] 0.8) with CO in all seasons, and correlations nearly as high with [PM.sub.10]. [O.sub.3] was positively correlated with all three other pollutants only in the spring, and most strongly with [PM.sub.10]. [O.sub.3] showed a weaker positive relationship to [PM.sub.10] in the summer and a negative relationship to CO and [NO.sub.2] in the winter. Higher mean temperatures were associated with higher pollutant levels in all seasons, with the exception of CO in the autumn. Barometric pressure showed varying relationships with pollutants. Expressing the data as residuals from cubic-spline smoothing brought about no marked changes in these correlations, except that in autumn the positive relationship between [O.sub.3] and temperature became nonsignificant. Regional measurements and basinwide averages correlated strongly for every pollutant in every season (r [is greater than] 0.7), except for [PM.sub.10] in the summer (r = 0.5-0.6 for some regions). Different regions' measurements of a given pollutant also correlated positively in all seasons. Southern coastal region 4 and eastern inland region 6 contrasted most sharply, with r-values between 0.3 and 0.7. In light of this generally similar behavior of air quality in different regions, time-series analyses focused on the entire basin, and regional comparisons were limited to regions 4 and 6.

Table 2. Pairwise correlation coefficients (r) for atmospheric factors expressed as basinwide 24-hr averages,(a) by season.(*)

Factor         Season   [NO.sub.2]   [PM.sub.10]   [O.sub.3]

CO             Winter      0.89         0.78        -0.43
               Spring      0.92         0.54         0.29
               Summer      0.94         0.72         0.03
               Autumn      0.84         0.58        -0.36

[NO.sub.2]     Winter      -            0.88        -0.23
               Spring      -            0.67         0.35
               Summer      -            0.80         0.11
               Autumn      -            0.80        -0.00

[PM.sub.10]    Winter      -            -           -0.01
               Spring      -            -            0.63
               Summer      -            -            0.40
               Autumn      -            -            0.28

[O.sub.3]      Winter      -            -            -
               Spring      -            -            -
               Summer      -            -            -
               Autumn      -            -            -

[T.sub.mean]   Winter      -            -            -
               Spring      -            -            -
               Summer      -            -            -
               Autumn      -            -            -

Factor         Season   [T.sub.mean]   BP(a)

CO             Winter       0.22        0.43
               Spring       0.38        0.15
               Summer       0.51       -0.17
               Autumn      -0.08        0.38

[NO.sub.2]     Winter       0.38        0.38
               Spring       0.53        0.03
               Summer       0.51       -0.21
               Autumn       0.28        0.12

[PM.sub.10]    Winter       0.37        0.39
               Spring       0.64       -0.18
               Summer       0.44       -0.30
               Autumn       0.40       -0.03

[O.sub.3]      Winter       0.33       -0.11
               Spring       0.57       -0.23
               Summer       0.18       -0.05
               Autumn       0.62       -0.42

[T.sub.mean]   Winter       -           0.13
               Spring       -          -0.34
               Summer       -          -0.20
               Autumn       -          -0.39

Abbreviations: BP, barometric pressure; T, temperature.

(a) BP measured only at the Los Angeles International Airport; other variables averaged from measurements in all six regions, weighted according to regions' proportions of total cardiovascular plus pulmonary admissions.

(*) p < 0.05 for r > 0.10; p < 0.01 for r > 0.13.

Figure 2 shows average concentrations by region and season for CO, [PM.sub.10], and [O.sub.3]. The mean and variance of CO decreased markedly in the spring and summer in all regions. Autumn and winter CO were highest in the southern coastal region 4. [PM.sub.10] was highest in the summer and autumn, particularly in the eastern inland region 6, but seasonal variation was less for [PM.sub.10] than for CO. [O.sub.3] was highest in the spring and summer, particularly in inland regions 3 and 6.

[Figure 2 ILLUSTRATION OMITTED]

Contrast of atmospheric conditions between days with high and low admission counts. Table 3 summarizes significant (p [is less than] 0.05) differences in basinwide weather and pollution statistics between days with unusually high and unusually low admission counts (residuals from cubic-spline smoothing) in a particular broad disease category. High-admission days (and/or immediately preceding days) tended to have relatively warm dry weather. Primary pollution (CO and [NO.sub.2]) was significantly elevated on winter, spring, or autumn days with high cardiovascular admissions; spring and summer days with high pulmonary admissions; and spring and autumn days with high cerebrovascular admissions. Elevated [PM.sub.10] tended to accompany elevated primary pollutants on days with high cardiovascular or pulmonary admissions; [PM.sub.10] also was associated with high abdominal admissions in the spring. [O.sub.3] was increased (along with the other pollutants) on days with high pulmonary admissions in spring and summer, the seasons of the highest mean [O.sub.3] concentrations (Table 1). By contrast, [O.sub.3] was decreased on days with high cardiovascular admissions in the winter, when 03 was generally low and negatively correlated with the other pollutants.

Table 3. Atmospheric variables showing significant (p < 0.05) differences between days with high and low hospital admissions in the entire basin, by season.(a)

Disease
category          Season   CO    [NO.sub.2]   [PM.sub.10]

Cardiovascular    Winter    +        +             +
                  Spring    +        +
                  Summer
                  Autumn    +        +             +

Pulmonary(c)      Spring    +        +             +
                  Summer    +        +             +
                  Autumn

Cerebrovascular   Winter
                  Spring    +        +
                  Autumn   (b)       +

Abdominal(d)      Spring                           +

Disease
category          Season   [O.sub.3]   [T.sub.high]

Cardiovascular    Winter       -            +
                  Spring                    +
                  Summer
                  Autumn

Pulmonary(c)      Spring       +            +
                  Summer       +            +
                  Autumn                    +

Cerebrovascular   Winter                   (b)
                  Spring                   (b)
                  Autumn                    +

Abdominal(d)      Spring

Disease
category          Season   [T.sub.low]   BP   Rain

Cardiovascular    Winter                 +     -
                  Spring       (b)             -
                  Summer                 +
                  Autumn

Pulmonary(c)      Spring                       -
                  Summer        +
                  Autumn

Cerebrovascular   Winter
                  Spring        +
                  Autumn                       -

Abdominal(d)      Spring                       -

Abbreviations: BP, barometric pressure; T, temperature.

(a) High- and low-admission days (> 85th and < 15th percentiles, respectively) are determined by residuals from regressions accounting for temporal effects (see text). A significant increase in atmospheric variables on high-admission days relative to low-admission days (and/or the immediately preceding days) is indicated by +; a significant decrease by -.

(b) The change in atmospheric measurement from the preceding day was significantly more positive on high- than low-admission days, although values measured on high and low days were not significantly different.

(c) For pulmonary diseases in winter, high-admission days' increases in [NO.sub.2], [PM.sub.10], and high temperature approached significance (p < 0.10).

(d) For abdominal diseases in winter, high-admission days' increases in CO approached significance (p < 0.10).

Table 4 shows mean weather and pollution conditions on days of high and low cardiovascular admissions in the winter and summer for the contrasting southern coastal region 4 and eastern inland region 6. In the summer, pollution (except for CO) and heat were markedly greater in region 6, but there were no clear pollution or temperature differences between high- and low-admission days in either region. In the winter, CO was markedly higher in region 4, and other regional differences were modest. In region 4, winter high-admission days had significantly higher temperature, barometric pressure, CO, [NO.sub.2], and [PM.sub.10], and significantly lower probability of rain, than low-admission days. In region 6, these tendencies were less obvious, but CO and [NO.sub.2] were significantly elevated on the days preceding high-admission days. In similar analyses of pulmonary diseases (not tabulated), we found only a few significant associations with high admissions: high same-day [PM.sub.10] in region 4 in the winter, high previous-day CO and low previous-day [O.sub.3] in region 6 in the winter, and high previous-day [NO.sub.2] and [O.sub.3] in region 6 in the summer.

Table 4. Atmospheric differences between days of high and low cardiovascular admissions(a) in most contrasting regions/seasons.

                                     Region 4 (Long Beach)
Variable               Admissions
(units)                   (n)         Winter        Summer

CO (ppm)                  Low           2.1          1.0
                          High          2.9(*)       1.1

[NO.sub.2] (pphm)         Low           3.4          2.8
                          High          4.4(*)       2.9

[PM.sub.10]               Low          27.6         38.0
([micro]g/[m.sup.3])      High         42.8(*)      36.9

[O.sub.3] (pphm)          Low           1.4          2.8
                          High          1.4          2.9

[T.sub.high]              Low          17.9         25.6
([degrees] C)             High         20.1(*)      25.6

[T.sub.low]               Low          10.6         17.7
([degrees] C)             High         10.4         17.2

BP(b) (mbar)              Low       1,016.2      1,012.9
                          High      1,018.5(*)   1,013.0

Rain(b)                   Low          38            0
(% of days)               High         12(*)         0

                          Region 6 (Riverside)
Variable
(units)                   Winter       Summer

CO (ppm)                   1.1           1.1
                           1.3(**)       1.2

[NO.sub.2] (pphm)          2.5           3.8
                           2.9(**)       4.0

[PM.sub.10]               39.1          84.6
([micro]g/[m.sup.3])      48.5          83.5

[O.sub.3] (pphm)           1.9           4.4
                           1.8           4.5

[T.sub.high]              19.2          33.7
([degrees] C)             19.6          33.9

[T.sub.low]                8.9          17.1
([degrees] C)              9.2          17.0

BP(b) (mbar)           1,017.3       1,012.5
                       1,016.4       1,013.5(*)

Rain(b)                   26             0
(% of days)               22             2

(a) See Table 3 for definition of high-and low-admission days.

(b) Measured at the Los Angeles International Airport, closer to region 4 than region 6. Other measurements made within the indicated region.

(*) Significant (p < 0.05) differences.

(**) Although this difference did not reach significance, the difference between days immediately preceding high- and low-admission days was significant (p < 0.05).

Analyses of admission counts in broad disease categories. Table 5 presents the results from single-pollutant autoregressive Poisson models, including all of the time and weather predictors mentioned in "Methods," relating daily average concentrations with same-day hospital admission counts over the entire basin for each broad disease category. Admissions of patients under 30 years of age are not included. To a good approximation, the coefficients represent proportionate increases in admission counts expected from unit increases in pollutant concentrations.

Table 5. Poisson regression coefficients (SEs): hospital admissions in broad disease categories in the entire metropolitan area, versus same-day pollution levels.(a)

Pollutant,
units         Season      Cardiovascular

CO            All         0.032 (0.003)(*)
(ppm)         Winter      0.038 (0.006)(*)
              Spring      0.010 (0.015)
              Summer      0.035 (0.014)(*)
              Autumn      0.027 (0.006)(*)

[NO.sub.2]    All         0.014 (0.002)(*)
(pphm)        Winter      0.016 (0.004)(*)
              Spring      0.001 (0.006)
              Summer      0.011 (0.005)(*)
              Autumn      0.014 (0.003)(*)

[PM.sub.10]   All       0.00064 (0.00012)(*)
([micro]g/    Winter    0.00095 (0.00024)(*)
[m.sup.3])    Spring   -0.00031 (0.00037)
              Summer    0.00039 (0.00041)
              Autumn    0.00065 (0.00020)(*)

[O.sub.3]     All        -0.007 (0.003)(**)
(pphm)        Winter     -0.021 (0.008)(**)
              Spring      0.003 (0.005)
              Summer      0.001 (0.005)
              Autumn     -0.003 (0.007)

Pollutant,
units         Season        Pulmonary           Cerebrovascular

CO            All        0.007 (0.005)           0.009 (0.007)
(ppm)         Winter     0.016 (0.009)          -0.008 (0.014)
              Spring     0.014 (0.024)           0.107 (0.033)(*)
              Summer     0.020 (0.021)           0.030 (0.033)
              Autumn     0.020 (0.008)(*)        0.008 (0.012)

[NO.sub.2]    All        0.007 (0.003)(*)        0.004 (0.004)
(pphm)        Winter     0.011 (0.005)(*)       -0.013 (0.007)
              Spring     0.007 (0.010)           0.042 (0.012)(*)
              Summer     0.004 (0.008)           0.009 (0.012)
              Autumn     0.012 (0.004)(*)        0.007 (0.006)

[PM.sub.10]   All      0.00057 (0.00018)(*)    0.00006 (0.00025)
([micro]g/    Winter   0.00081 (0.00032)(*)   -0.00021 (0.00052)
[m.sup.3])    Spring   0.00010 (0.00061)       0.00126 (0.00083)
              Summer   0.00061 (0.00061)       0.00127 (0.00096)
              Autumn   0.00078 (0.00029)      -0.00004 (0.00039)

[O.sub.3]     All        0.008 (0.004)           0.003 (0.005)
(pphm)        Winter    -0.006 (0.010)           0.028 (0.016)
              Spring     0.011 (0.008)           0.011 (0.011)
              Summer     0.006 (0.007)           0.007 (0.011)
              Autumn     0.009 (0.011)          -0.003 (0.014)

Pollutant,
units         Season       Abdominal

CO            All        0.003 (0.004)
(ppm)         Winter     0.006 (0.008)
              Spring    -0.007 (0.019)
              Summer     0.021 (0.018)
              Autumn     0.006 (0.007)

[NO.sub.2]    All        0.004 (0.002)(*)
(pphm)        Winter     0.002 (0.005)
              Spring    -0.004 (0.007)
              Summer     0.008 (0.006)
              Autumn     0.007 (0.004)

[PM.sub.10]   All      0.00017 (0.00014)
([micro]g/    Winter   0.00013 (0.00030)
[m.sup.3])    Spring   0.00039 (0.00047)
              Summer   0.00068 (0.00052)
              Autumn   0.00008 (0.00022)

[O.sub.3]     All        0.003 (0.003)
(pphm)        Winter    -0.012 (0.010)
              Spring     0.000 (0.006)
              Summer     0.011 (0.006)
              Autumn     0.000 (0.008)

(a) Regression analyses used 24-hr average measurements of pollutants and same-day admission counts for patients [is greater than or equal to] 30 years of age throughout the South Coast Air Basin. Example interpretation: the coefficient 0.038 relating winter cardiovascular admissions to CO indicates that admissions increase by a factor of [e.sup.0.038], i.e., by 3.9%, with a 1-ppm increase in CO concentration, after allowing for the effects of time and weather on admission rates.

(*) Significant in expected direction, p < 0.05.

(**) Significant in the "wrong" direction, p < 0.05.

Primary pollution, as represented by CO and [NO.sub.2], showed the most consistent associations with cardiovascular-disease admissions; they were significantly related in year-round analyses and in single-season analyses except for the spring. [PM.sub.10] showed a similar pattern of relationships, but was nonsignificant in the summer as well as the spring. The cardiovascular disease/primary pollution relationship was not very sensitive to inclusion or exclusion of weather or other pollutant variables in the models. In the winter, the interquartile range of CO concentrations was 1.1 to 2.2 ppm, and the corresponding predicted increase in cardiovascular admissions was 4.2%, which represented approximately 20 extra admissions/day.

Pulmonary-disease admissions were significantly related to [NO.sub.2] or [PM.sub.10] in year-round and winter analyses, and also to [NO.sub.2] or CO in autumn. Cerebrovascular-disease admissions were significantly related only to CO or [NO.sub.2] only in the spring. Abdominal-disease admissions were significantly related only to [NO.sub.2], and only in the year-round analysis.

[O.sub.3] showed either negative or nonsignificant positive relationships with cardiovascular, pulmonary, cerebrovascular, and abdominal disease admissions in year-round and single-season analyses. The same was true in OLS regressions of Shumway-filtered data (not tabulated). Alternative analyses intended to give additional weight to high-[O.sub.3] conditions, by expressing exposure in terms of daily maxima or in exceedances of an assumed threshold, or by restricting the analysis to the three high-[O.sub.3] inland regions, still showed no significant positive associations. In Poisson models excluding mean temperature and barometric pressure, daily mean [O.sub.3] showed significant positive associations with pulmonary diseases in the spring and year-round.

Table 6 illustrates the sensitivity of results to the choice of regression procedure and model for the cardiovascular disease/CO relationship in the winter and the pulmonary disease/[O.sub.3] relationship in the spring. Across a broad range of models with and without weather and other pollutants as predictors, the estimated winter CO effect was always significant and was reasonably consistent in size. None of the other pollutants' effects was significant when included in a model with CO. The spring [O.sub.3] effect on pulmonary admissions was significant when [O.sub.3] was the only atmospheric factor in the model, predicting a 1.5% increase in admissions for a 1-pphm increase in daily mean [O.sub.3] concentration. However, the effect was nonsignificant if the model included weather and/or other pollutant variables. Interpretation of these findings is complicated by collinearity and by possibly different characteristics of exposure measurement error for different pollutants (19). Nevertheless, it seems clear that in the winter, CO was the analyzed atmospheric factor that was most closely linked with excess cardiovascular morbidity. In the spring, [O.sub.3] was the pollutant most closely associated with excess pulmonary morbidity; however, morbidity was still more closely associated with warm temperatures, and all four pollutants tended to rise with temperature, making interpretation difficult.

Table 6. Sensitivity of key results to choice of regression procedure and model.

Procedure (additional predictors)                     Slope (SE)

Cardiovascular admissions versus CO, winter

Poisson autoregressive (time)                      0.044 (0.005)(*)
Poisson autoregressive (time, weather)             0.038 (0.006)(*)
Poisson autoregressive (time, weather,
  [O.sub.3])                                       0.036 (0.007)(*)
Poisson autoregressive (all above + [PM.sub.10])   0.033 (0.011)(*)
Poisson autoregressive (all above + [NO.sub.2])    0.047 (0.013)(*)
PDL(b) (time, weather)                             0.044 (0.008)(*)
OLS autoregressive (time, weather)                  19.6 (3.1)

Pulmonary admissions versus [O.sub.3], spring

Poisson autoregressive (time)                      0.012 (0.004)(*)
Poisson autoregressive (time, weather)             0.003 (0.005)
Poisson autoregressive (time, weather, CO)         0.002 (0.005)
Poisson autoregressive (all above + [PM.sub.10])   0.008 (0.006)
Poisson autoregressive (all above + [NO.sub.2])    0.008 (0.006)
PDL(b) (time, weather)                             0.003 (0.011)
OLS autoregressive (time, weather)                  1.58 (1.79)

                                                   Relative
Procedure (additional predictors)                  risk(a)

Cardiovascular admissions versus CO, winter

Poisson autoregressive (time)                       1.050
Poisson autoregressive (time, weather)              1.043
Poisson autoregressive (time, weather,
  [O.sub.3])                                        1.040
Poisson autoregressive (all above + [PM.sub.10])    1.037
Poisson autoregressive (all above + [NO.sub.2])     1.053
PDL(b) (time, weather)                              1.050
OLS autoregressive (time, weather)                  1.048

Pulmonary admissions versus [O.sub.3], spring

Poisson autoregressive (time)                       1.015
Poisson autoregressive (time, weather)              1.003
Poisson autoregressive (time, weather, CO)          1.002
Poisson autoregressive (all above + [PM.sub.10])    1.009
Poisson autoregressive (all above + [NO.sub.2])     1.009
PDL(b) (time, weather)                              1.004
OLS autoregressive (time, weather)                  1.010

(a) Predicted relative risk at the 75th percentile concentration of pollutant, versus the 25th percentile.

(b) Cubic polynomial distributed lag model, lag 0 slope given, lags 1-3 slopes nonsignificant.

(*) Significant in expected direction, p < 0.05.

Because diabetes mellitus is an important risk factor for cardiovascular disease, we reanalyzed cardiovascular admissions separately for diabetics (all of those with ICD code 250 entered among four additional diagnoses in the record; approximately 20% of all patients) and for others, using the autoregressive Poisson model. In year-round analyses, the slope [+ or -] SE was 0.039 [+ or -] 0.006 for diabetics as compared to 0.031 [+ or -] 0.004 for others. Year-round analyses of [NO.sub.2] and [PM.sub.10] effects showed similar modest slope increases for diabetics, as did single-season analyses. None of the slope differences between diabetics and others was statistically significant.

Analyses of cardiovascular disease admission counts by age, sex, and ethnicity. Table 7 presents results from single-pollutant autoregressive Poisson models applied to cardiovascular admission counts in three age strata (30-64, 65-74, and [is greater than or equal to] 75 years of age) separately for men and women. Results for [O.sub.3] (not tabulated) were never statistically significant. CO effects were near-significant for women 30-64 years of age and sig-nificant in all other age-sex groups in year-round analyses and in one or more seasonal analyses. Effect sizes increased with age similarly in both sexes, but age-related differences were not significant. [NO.sub.2] effects showed a similar pattern of significance, but with less suggestion of age dependence. [PM.sub.10] effects were also significant year-round and/or in one season for all groups except men [is greater than or equal to] 75 years of age.

Table 7. Poisson regression coefficients (SEs): cardiovascular disease admissions in the entire metropolitan area versus same-day pollution levels, by age and sex.(a)

Sex, age
in years                              Year-round
[count](b)         Pollutant(c)      coefficient

Male 30-64         CO              0.014 (0.007)(*)
[99 [+ or -] 21]   [NO.sub.2]      0.007 (0.004)
                   [PM.sub.10]    0.0003 (0.0003)

Male 65-74         CO              0.037 (0.009)(*)
[61 [+ or -] 14]   [NO.sub.2]      0.014 (0.005)(*)
                   [PM.sub.10]    0.0008 (0.0003)(*)

Male 75+           CO              0.040 (0.009)(*)
[59 [+ or -] 13]   [NO.sub.2]      0.013 (0.005)(*)
                   [PM.sub.10]    0.0003 (0.0003)

Female 30-64       CO              0.017 (0.009)
[68 [+ or -] 15]   [NO.sub.2]      0.015 (0.004)(*)
                   [PM.sub.10]    0.0007 (0.0003)(*)

Female 65-74       CO              0.033 (0.009)(*)
[56 [+ or -] 12]   [NO.sub.2]      0.014 (0.005)(*)
                   [PM.sub.10]    0.0002 (0.0003)

Female 75+         CO              0.040 (0.007)(*)
[88 [+ or -] 17]   [NO.sub.2]      0.014 (0.004)(*)
                   [PM.sub.10]    0.0005 (0.0003)

Sex, age
in years                          Significant (p < 0.05) positive
[count](b)         Pollutant(c)   coefficients for separate seasons

Male 30-64         CO             Winter 0.040 (0.013),
[99 [+ or -] 21]                    autumn 0.025 (0.013)
                   [NO.sub.2]     Winter 0.017 (0.008),
                                    autumn 0.016 (0.007)
                   [PM.sub.10]    Winter 0.0016 (0.0005)

Male 65-74         CO             Autumn 0.045 (0.016)
[61 [+ or -] 14]   [NO.sub.2]     Autumn 0.024 (0.009)
                   [PM.sub.10]    Autumn 0.0013 (0.0006)

Male 75+           CO             Summer 0.068 (0.034),
[59 [+ or -] 13]                    autumn 0.042 (0.015)
                   [NO.sub.2]     Autumn 0.020 (0.008)
                   [PM.sub.10]    None

Female 30-64       CO             None
[68 [+ or -] 15]   [NO.sub.2]     Winter 0.018 (0.008)
                   [PM.sub.10]    None

Female 65-74       CO             Winter 0.043 (0.015)
[56 [+ or -] 12]   [NO.sub.2]     Winter 0.017 (0.008)
                   [PM.sub.10]    Winter 0.0012 (0.0006)

Female 75+         CO             Winter 0.047 (0.013),
[88 [+ or -] 17]                    autumn 0.025 (0.011)
                   [NO.sub.2]     Winter 0.021 (0.008)
                   [PM.sub.10]    Winter 0.0012 (0.0005)

(a) These results are for 1992-1994 only. See footnote to Table 5 for explanation of regression procedure and coefficients.

(b) Annual mean daily admission count [+ or -] SD, for patients [is greater than or equal to] 30 years of age.

(c) [O.sub.3] results not tabulated; none was significantly positive.

(*) Year-round relationship significant, p < 0.05.

Table 8 presents the results from single-pollutant autoregressive Poisson models applied to cardiovascular admission counts for adults [is greater than or equal to] 30 years of age in four ethnic categories--white (non-Hispanic), black, Hispanic, and other. The other category includes people of Asian-Pacific ancestry (the large majority), Native Americans, and others not classifiable in the first three groups. [O.sub.3] effects (not tabulated) were never significant. Regression coefficients, though not significantly different, suggested meaningful ethnic differences in exposure--response relationships. CO, [NO.sub.2], and [PM.sub.10] effects were significant in whites in year-round, winter, and autumn analyses. In blacks, CO and [NO.sub.2] effects were significant year-round (also in the winter for CO) and were similar to these effects in whites. CO and [NO.sub.2] effects in Hispanics were significant year-round but were smaller than these effects in whites and blacks. The remaining (other) category, with a relatively small number of admissions, showed consistently small and nonsignificant regression slopes.

Table 8. Poisson regression coefficients (SEs): cardiovascular disease admissions in the entire metropolitan area versus same-day pollution levels, by ethnic category.(a)

Category                               Year-round
[count](b)          Pollutant(c)       coefficient

White               CO               0.034 (0.005)(*)
[290 [+ or -] 53]   [NO.sub.2]       0.014 (0.003)(*)
                    [PM.sub.10]     0.0006 (0.0002)(*)

Black               CO               0.031 (0.010)(*)
[49 [+ or -] 11]    [NO.sub.2]       0.014 (0.006)(*)
                    [PM.sub.10]     0.0003 (0.0004)

Hispanic            CO               0.019 (0.009)(*)
[63 [+ or -] 13]    [NO.sub.2]       0.010 (0.005)(*)
                    [PM.sub.10]     0.0005 (0.0003)

Other               CO               0.001 (0.013)
[29 [+ or -] 8]     [NO.sub.2]       0.002 (0.007)
                    [PM.sub.10]    -0.0004 (0.0005)

                                   Significant (p < 0.05)
Category                           positive coefficients
[count](b)          Pollutant(c)   for separate seasons

White               CO             Winter 0.038 (0.008),
[290 [+ or -] 53]                    autumn 0.036 (0.008)
                    [NO.sub.2]     Winter 0.018 (0.005),
                                     autumn 0.017 (0.005)
                    [PM.sub.10]    Winter 0.0011 (0.0003),
                                     autumn 0.0007 (0.0003)

Black               CO             Winter 0.042 (0.017)
[49 [+ or -] 11]    [NO.sub.2]     None
                    [PM.sub.10]    None

Hispanic            CO             None
[63 [+ or -] 13]    [NO.sub.2]     None
                    [PM.sub.10]    None

Other               CO             None
[29 [+ or -] 8]     [NO.sub.2]     None
                    [PM.sub.10]    None

(a) These results are for 1992-1994 only. See footnote to Table 5 for explanation of regression procedure and coefficients.

(b) Annual mean daily admission count [+ or -] SD, for patients [is greater than or equal to] 30 years of age.

(c) [O.sub.3] results not tabulated; none was significantly positive.

(*) Year-round relationship significant, p < 0.05.

Analyses of admission counts for more specific diagnoses. Table 9 presents results in adults [is greater than to equal to] 30 years of age from single-pollutant autoregressive Poisson models relating basinwide daily average pollutant concentrations with same-day admission counts. Occlusive strokes showed the most consistent positive relationships to pollution: significant associations with [O.sub.3] in the summer only; and with CO, [NO.sub.2], and [PM.sub.10] year-round and in at least two single-season analyses. Asthma, COPD, and CHF were significantly associated with CO and [NO.sub.2] year-round and in one or more single-season analyses. Myocardial infarction was associated with CO and [NO.sub.2], and arrhythmias with CO, in year-round analyses only.

Table 9. Poisson regression coefficients (SEs): hospital admissions in more specific disease categories in the entire metropolitan area, versus same-day pollution levels.(a)

Disease                              Year-round
[count](b)         Pollutant        coefficient

Myocardial         CO             0.040 (0.009)(*)
infarction         [NO.sub.2]     0.011 (0.005)(*)
[47 [+ or -] 11]   [PM.sub.10]   0.0006 (0.0003)
                   [O.sub.3]     -0.007 (0.007)

Congestive heart   CO             0.025 (0.009)(*)
failure            [NO.sub.2]     0.010 (0.005)(*)
[49 [+ or -] 11]   [PM.sub.10]   0.0004 (0.0003)
                   [O.sub.3]     -0.001 (0.007)

Cardiac            CO             0.023 (0.009)(*)
arrhythmia         [NO.sub.2]     0.006 (0.005)
[50 [+ or -] 10]   [PM.sub.10]   0.0002 (0.0003)
                   [O.sub.3]     -0.001 (0.007)

Occlusive stroke   CO             0.044 (0.009)(*)
[45 [+ or -] 10]   [NO.sub.2]     0.020 (0.005)(*)
                   [PM.sub.10]   0.0013 (0.0003)(*)
                   [O.sub.3]      0.007 (0.007)

Asthma(c)          CO             0.028 (0.010)(*)
[38 [+ or -] 9]    [NO.sub.2]     0.014 (0.005)(*)
                   [PM.sub.10]   0.0003 (0.0004)
                   [O.sub.3]     -0.001 (0.008)

COPD               CO             0.019 (0.007)(*)
[89 [+ or -] 19]   [NO.sub.2]     0.008 (0.004)(*)
                   [PM.sub.10]   0.0003 (0.0002)
                   [O.sub.3]     -0.007 (0.005)

Disease                          Significant (p < 0.05) positive
[count](b)         Pollutant     coefficients for separate seasons

Myocardial         CO            None
infarction         [NO.sub.2]    None
[47 [+ or -] 11]   [PM.sub.10]   None
                   [O.sub.3]     None

Congestive heart   CO            Summer 0.074 (0.038)
failure            [NO.sub.2]    Winter 0.019 (0.009)
[49 [+ or -] 11]   [PM.sub.10]   None
                   [O.sub.3]     None

Cardiac            CO            None
arrhythmia         [NO.sub.2]    None
[50 [+ or -] 10]   [PM.sub.10]   None
                   [O.sub.3]     None

Occlusive stroke   CO            Winter 0.036 (0.017),
[45 [+ or -] 10]                   summer 0.091 (0.039),
                                   autumn 0.032 (0.015)
                   [NO.sub.2]    Winter 0.027 (0.010),
                                   autumn 0.021 (.008)
                   [PM.sub.10]   Winter 0.0024 (0.0006),
                                   autumn 0.0012 (0.0005)
                   [O.sub.3]     Summer 0.025 (0.012)

Asthma(c)          CO            Winter 0.045 (0.017),
[38 [+ or -] 9]                    autumn 0.039 (0.016)
                   [NO.sub.2]    Winter 0.028 (0.010),
                                   autumn 0.019 (0.008)
                   [PM.sub.10]   None
                   [O.sub.3]     None

COPD               CO            Winter 0.035 (0.012),
[89 [+ or -] 19]                   autumn 0.029 (0.011)
                   [NO.sub.2]    Autumn 0.016 (0.006)
                   [PM.sub.10]   None
                   [O.sub.3]     None

(a) See footnote to Table 5 for explanation of regression procedure and coefficients.

(b) Annual mean daily admission count [+ or -] SD for patients [is greater than or equal to] 30 years of age.

(c) See text concerning asthma in patients < 30 years of age.

(*) Year-round relationship significant, p < 0.05.

We also analyzed asthma admissions for patients 0-29 years of age. In year-round analyses, slopes [+ or -] SEs were 0.036 [+ or -] 0.016/ppm CO, 0.024 [+ or -] 0.008/pphm [NO.sub.2], and 0.0011 [+ or -] 0.0006/[micro]g/[m.sup.3] [PM.sub.10]--all significant (p [is less than] 0.05) and appreciably larger than the slopes in adults [is greater than] 30 years of age. [O.sub.3] effects were nonsignificant. Most of the admitted patients in this youngest group were children: the mean age was 7.

Relationships of cardiovascular disease admissions to CO or [PM.sub.10] in the two most contrasting regions. Table 10 presents comparative statistics for cardiovascular admissions, CO, and [PM.sub.10] in southern coastal region 4 and eastern inland region 6. The two regions showed reasonably similar daily admission counts and similar positive correlations of daily CO and [PM.sub.10] levels, but markedly different concentration ranges (Figure 2). By OLS regression allowing for autocorrelation, a wintertime 1-ppm rise in CO predicted a 9-[micro]g/[m.sup.3] rise in [PM.sub.10] in region 4, but a 25 [micro]g/[m.sup.3] rise in region 6. In single-pollutant autoregressive Poisson models, region 4 showed highly significant relationships between [PM.sub.10] and admissions, year-round and in the winter, despite its low [PM.sub.10]. In region 6, despite its high [PM.sub.10], regression slopes were significantly lower than in region 4, and were not significantly different from zero. Admissions showed a more plausible relationship with CO across the two regions, with highly significant positive slopes in region 4 and modestly lower non-significant slopes in region 6, consistent with its generally lower and less variable CO concentrations.

Table 10. Daily cardiovascular (CV) admissions, CO, and [PM.sub.10]: statistics for the two most contrasting regions.

                                      Region 4
Statistic               Season      (Long Beach)

CV admission count      All      48-67
interquartile range     Winter   52-71

CO interquartile        All      0.93-2.40
range (ppm)             Winter   1.33-3.17

[PM.sub.10]             All      28-45
interquartile range     Winter   20-43
([micro]g/[m.sup.3])

Correlation,            All      0.56(*)
CO vs. [PM.sub.10]      Winter   0.70(*)

Correlation,            All      0.63(*)
CO vs. [PM.sub.10]      Winter   0.72(*)
residuals(a)

Regression slope,       All      0.022 (0.005)(*)
CV vs. CO(b)            Winter   0.027 (0.008)(*)

Regression slope,       All      0.0012 (0.0003)(*)
CV vs. [PM.sub.10](b)   Winter   0.0018 (0.0006)(*)

                                    Region 6
Statistic               Season     (Riverside)

CV admission count      All      43-57
interquartile range     Winter   44-59

CO interquartile        All      0.81-1.58
range (ppm)             Winter   0.74-1.70

[PM.sub.10]             All      44-86
interquartile range     Winter   24-61
([micro]g/[m.sup.3])

Correlation,            All      0.55(*)
CO vs. [PM.sub.10]      Winter   0.75(*)

Correlation,            All      0.69(*)
CO vs. [PM.sub.10]      Winter   0.76(*)
residuals(a)

Regression slope,       All      0.012 (0.008)
CV vs. CO(b)            Winter   0.017 (0.015)

Regression slope,       All      0.0001 (0.0002)
CV vs. [PM.sub.10](b)   Winter   0.0003 (0.0004)

(a) Observed values minus values predicted by day of the week and cubic-spline smooth of longer term trends.

(b) From single-pollutant autoregressive Poisson models including temporal and weather effects.

(*) Significant, p < 0.05.

Discussion

Limitations; recommendations for future research. Problems with this and other time--series studies include exposure misclassification, response misclassification, and model misspecification. Exposure misclassification occurs when the monitored environmental factors are not the ones responsible for health effects, when monitoring errors are appreciable and differ by time and location, when monitoring station data poorly represent background air near patients' homes, when personal microenvironments differ from background, or when exposures that precipitate hospital admissions occur away from home. Future expansion of the monitoring program, at least for particulate pollution, should provide better background concentration estimates, allowing more powerful tests for regional differences in effects. New personal monitoring studies, designed to elucidate longitudinal relationships between background and personal exposures, might help to disentangle the effects of particulate pollutants and covarying gases (e.g., CO and [NO.sub.2]). Small panel studies have suggested that personal particulate exposures track background concentrations closely in healthy older adults and children in The Netherlands (20,21), but not in older adults with COPD in Los Angeles (22). To our knowledge, no longitudinal studies of personal CO exposure have been reported.

Response misclassification can result from errors in diagnosis or in medical record coding. Reviews suggest that 15-20% of assigned ICD codes are inaccurate (23,24). Inaccuracies should increase random errors in specific disease counts and reduce the statistical significance of disease/pollution relationships, but should not introduce bias unless coding inaccuracies covary with pollution. Misdiagnoses are hard to evaluate, but are undoubtedly important, given the complexities of disease processes and the fuzzy boundaries between diagnoses. Wrong diagnoses or codes would likely shift patients to different specific disease counts within the same broad category, and thus should have relatively little effect on broad-category analyses. In any event, we have had only limited success in finding specific pollutant-disease relationships with mechanistic or public-health implications. Future studies focusing on precise diagnoses and accounting for other risk factors (e.g., additional diagnoses and particular demographic characteristics) might be more successful.

Limitations of our primary analytical model include the use of only one pollutant at a time and possibly incomplete accounting for weather and temporal influences. Because estimated CO effects were similar in various single- and multipollutant models that accounted for seasonal and weekly cycles, more complete modeling of weather or temporal effects should not change the conclusions concerning CO. By contrast, estimates of [O.sub.3] effects were highly sensitive to inclusion or exclusion of weather and other pollutant variables. If pollutants not in the model affected admissions, the likely result would be to overestimate effects of the modeled pollutant and underestimate total effects of pollution (12). Thus, effects we associated with CO might be at least partly due to covarying gases (e.g., oxides of nitrogen) or to particulate substances. Similarly, incomplete accounting for lagged effects would likely result in overestimated effects of very recent exposure, but underestimated cumulative effects of recent and earlier exposures (14).

Conclusion

In general, our results from metropolitan Los Angeles appear consistent with reports from elsewhere (7-12) that day-to-day increases in urban CO and/or [PM.sub.10] and/or [NO.sub.2] are associated with meaningful increases in cardiovascular illnesses. We found only a few equivocally positive relationships between cardiopulmonary morbidity and [O.sub.3], in situations when other pollutants and heat stress increased along with [O.sub.3]. This is surprising, in light of severe [O.sub.3] pollution in Los Angeles, obvious acute respiratory effects of [O.sub.3] in animal and human exposure studies (2), and recent observations of [O.sub.3]-related hospital admissions in Toronto, Canada, where [O.sub.3] levels are lower than in Los Angeles (12). [O.sub.3] has been linked to mortality in Los Angeles (13), although [PM.sub.10] might explain that association (25). On the other hand, a recent time--series study of asthma admissions in central and western Los Angeles (26) generally supports our findings, showing associations with [PM.sub.10] but not with [O.sub.3]. The tendency of [O.sub.3] concentrations to decrease indoors, where most people spend most of their time (27), might attenuate morbidity/[O.sub.3] relationships, but would not likely do so in Los Angeles more than in Toronto. In any event, our results suggest that the excess risk of hospitalization in Los Angeles is greater on high-primary-pollution days than on high-[O.sub.3] days. The greatest risk of pollution-related hospital admissions apparently occurs on autumn/winter days with weak Santa Ana weather conditions, when air incursion from the desert approximately counterbalances that from the ocean, resulting in maximal atmospheric stagnation.

We could not distinguish clearly among CO-, [NO.sub.2]- and particle-associated effects. CO showed the strongest statistical relationships with most indices of morbidity even in the regions and seasons with the highest and widest ranging [PM.sub.10]. [NO.sub.2] tracked CO closely enough that CO-associated effects might reasonably be attributed to [NO.sub.2] and/or another oxide of nitrogen. Weaker statistical relationships of illness to [PM.sub.10] might have resulted from less accurate exposure assessment even if [PM.sub.10] were inherently more toxic (19). Too little is known about the relationships between the ambient background and personal exposures to judge which pollutants are most subject to exposure misclassification. Even if [PM.sub.10]/morbidity associations were entirely explainable by CO/morbidity associations, some particulate species closely associated with CO might be the active agent(s). Alternatively, our findings might reflect separate effects of CO and some component of [PM.sub.10], as suggested by the Schwartz (9) Tucson, Arizona, study. One argument against CO effects per se is that typical ambient background CO concentrations are below normal bloodstream concentrations of metabolically produced CO (3). Even on most high-CO days in Los Angeles, inhaling the background concentration should reduce the blood's oxygen-carrying capacity by [is less than] 1%. However, CO concentrations near sources (e.g., heavy traffic) exceed background levels and may cause appreciable cardiovascular stress (9,10). If both are driven by atmospheric stagnation, these higher microenvironmental concentrations should track background levels. Thus, a low range of monitored background CO does not necessarily rule out an effect of CO on cardiovascular morbidity.

The observed association of all pulmonary diseases with [PM.sub.10] or [NO.sub.2] more than CO, and of all cardiovascular diseases with CO more than [PM.sub.10] (Table 5), appears consistent with the well-known properties of CO as a circulatory toxicant without direct effects on the lungs, and of some particulate species as respiratory irritants. The association of occlusive strokes with all four tested pollutants appears consistent with the hypothesis of Seaton et al. (28) that urban pollution provokes alveolar inflammation, releasing mediators which increase blood coagulability. A previous finding of increased plasma viscosity during a primary pollution episode in Germany (29) also supports that hypothesis. By our data, we cannot test Seaton et al.'s (28) attribution of the inflammatory effect to ultrafine particles.

We found possibly meaningful demographic differences in morbidity/pollution relationships, although none of them attained statistical significance. Persons [is greater than or equal to] 65 years of age and diabetics showed somewhat increased cardiovascular disease effects as compared to others without those risk factors, but men did not appear to be more at risk than women of similar age. Persons younger than 30 years of age showed the largest pollution-related effects on asthma. Although air pollution health risks are believed to fall disproportionately on ethnic minorities (30), whites usually showed the largest pollution-related effects on cardiovascular disease. Effect sizes in blacks, the minority group generally at greatest risk for cardiovascular disease, were similar to those in whites, whereas effect sizes were generally smaller in Hispanics and undetectable in the other (predominantly Asian) ethnic category. Definitive interpretation would require evaluation of ethnic differences in exposure, susceptibility, and access to hospitals. On average, [O.sub.3] exposures in the basin appear higher for whites than for blacks or Asian/Pacific Islanders (31). Differences in other exposures apparently have not been studied formally, but the high-primary-pollution regions 2 and 4 have high proportions of minority residents. Thus, smaller effects in some minorities (if real) probably are not explained by less exposure.

The relatively nonspecific pattern of diagnoses suggests that excess patients admitted to hospitals on high-pollution days in metropolitan Los Angeles are individuals with preexisting problems which make them highly vulnerable to any extra stresses on their oxygen delivery systems, including unfavorable changes in the air environment. If so, generalized efforts to preserve cardiopulmonary health should help to prevent (or at least to postpone) pollution-associated illnesses. Our findings suggest that control of primary pollutants is more important to public health than control of [O.sub.3], which in any case depends on control or primary pollutants.

REFERENCES AND NOTES

(1.) Lipfert FW. Air Pollution and Community Health. New York:Van Nostrand Reinhold, 1994;347-413, 516-536.

(2.) Committee of the Environmental and Occupational Health Assembly of the American Thoracic Society. State of the art: health effects of outdoor air pollution, part I. Am J Respir Crit Care Mad 153:3-50 (1996).

(3.) Committee of the Environmental and Occupational Health Assembly of the American Thoracic Society. State of the art: health effects of outdoor air pollution, part II. Am J Respir Crit Care Med 153:477-498 (1996).

(4.) Vedal S. Ambient particles and health: lines that divide. J Air Waste Manag Assoc 47:551-581 (1997).

(5.) Dockery DW, Pope CA. Acute respiratory effects of particulate air pollution. Annu Rev Public Health 15:107-132 (1994).

(6.) Pope CA III, Bates DV, Raizenne ME. Health effects of particulate air pollution: time for reassessment? Environ Health Perspect 103:472-480 (1995).

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

(8.) Burnett RT, Dales RE, Brook JR, Raizenne ME, Krewski D. Association between ambient carbon monoxide levels and hospitalizations for congestive heart failure in the elderly in 10 Canadian cities. Epidemiology 8:162-167 (1997).

(9.) Schwartz J. Air pollution and hospital admissions for cardiovascular disease in Tucson. Epidemiology 8:371-377 (1997).

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

(11.) Schwartz J. Air pollution and hospital admissions for heart disease in eight US counties. Epidemiology 10:17-22 (1999).

(12.) Burnett RT, Smith-Doiron M, Stieb D, Cakmak S, Brook JR. Effects of particulate and gaseous air pollution on cardiorespiratory hospitalizations. Arch Environ Health 54:130-139 (1999).

(13.) Kinney PL, Ozkaynak H. Associations of daily mortality and air pollution in Los Angeles County. Environ Res 54:99-120 (1991).

(14.) Pope CA, Schwartz J. Time series for the analysis of pulmonary health data. Am J Respir Crit Care Med 154:S229-S233 (1996).

(15.) Gold DR, Damokosh Al, Pope CA, Dockery DW, McDonnell WF, Serrano P, Retama A Castillejos M. Particulate and ozone effects on the respiratory function of children in southwest Mexico City. Epidemiology 10:8-16 (1999).

(16.) Samet JM, Zeger SL, Berhane K. The association of mortality and particulate air pollution. In: Particulate Air Pollution and Daily Mortality: Replication and Validation of Selected Studies. Cambridge, MA:Health Effects Institute, 1995;1-104.

(17.) Thurston GD, Kinney PL. Air pollution epidemiology: considerations in time-series modeling. Inhal Toxicol 7:71-83 (1995).

(18.) California Air Resources Board. Ten-Year Summary of California Air Quality Data, 1963-1972. Sacramento, CA:State of California Resources Agency, 1974.

(19.) Lipfert FW, Wyzga RE. Air pollution and mortality: the implications of uncertainties in regression modeling and exposure measurement. J Air Waste Manag Assoc 47:517-523 (1997).

(20.) Janssen NAH, Hoek G, Brunekreef B, Harssema H, Mensink I, Zuidhof A. Personal sampling of [PM.sub.10] in adults: relation between personal, indoor, and outdoor concentrations. Am J Epidemiol 147:537-547 (1998).

(21.) Janssen NAH, Hoek G, Harssema H, Brunekreef B. Personal exposure to fine particles in children correlates closely with ambient fine particles. Arch Environ Health 54:95-101 (1999).

(22.) Linn WS, Gong H, Clark KW, Anderson KR. Day-to-day particulate exposures and health changes in Los Angeles area residents with severe lung disease. J Air Waste Manag Assoc 49:PM108-PM115 (1999).

(23.) Hsia DC, Ahern CA, Ritchie BP, Moscoe LM, Krushat WM. Medicare reimbursement accuracy under the prospective payment system, 1985 to 1988. J Am Med Assoc 268:896-899 (1992).

(24.) Fisher ES, Whaley FS, Krushat WM, Malenka DJ, Fleming C, Baron JA, Hsia DC. The accuracy of Medicare's hospital claims data. Am J Public Health 82:243-248 (1992).

(25.) Kinney PL, Ito K, Thurston GO. A sensitivity analysis of mortality/[PM.sub.10] associations in Los Angeles. Inhal Toxicol 7:59-69 (1995).

(26.) Nauenberg E, Basu K. Effect of insurance coverage on the relationship between asthma hospitalizations and exposure to air pollution. Public Health Rep 114:135-148 (1999).

(27.) Hayes SR. Use of an indoor air quality model (IAQM) to estimate indoor ozone levels. J Air Waste Manag Assoc 41:161-170 (1991).

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

(29.) Peters A, Doring A, Wichmann HE, Koenig W. Increased plasma viscosity during an air pollution episode: a link to mortality? Lancet 349:1582-1587 (1997).

(30.) Sexton K, Gong H, Bailar JC, Ford JG, Gold DR, Lambert WE, Utell MJ. Air pollution health risks: do class and race matter? Toxicol Ind Health 9:843-878 (1993).

(31.) Korc ME. A socioeconomic assessment of human exposure to ozone in the South Coast Air Basin of California. J Air Waste Manag Assoc 46:547-557 (1996).

Address correspondence to W.S. Linn, 51 Medical Science Building, Rancho Los Amigos National Rehabilitation Center, 7601 Imperial Highway, Downey, CA 90242 USA. Telephone: [phone]. Fax: [phone]. E-mail: [e-mail]

We thank the following for donating data and for help and advice: G. Cox, J. White, H. Lomas, B-M. Kim, M. Zeldin, S. Prasad, and F. Lurmann. We thank J. Solomon for programming and data quality assurance; C. Linn for consultation on hospital admission data interpretation; S. Terrell for computer artwork; and K. Clark, V. Valdez, and M. Avila for administrative assistance. We thank M.P. Jones, M. Lipsett, and J. Schwartz for advice on data handling and statistical analysis.

Major support for this project was provided by the American Heart Association, Western States Affiliate (grant 1134-GI1). The Southern California Environmental Health Sciences Center, funded by the National Institute of Environmental Health Sciences (grant 5P30 ES07048-03), provided additional support.

Received 13 July 1999; accepted 1 December 1999.

William S. Linn,(1)(2) Yaga Szlachcic,(1)(3) Henry Gong, Jr.,(1)(2)(3) Patrick L. Kinney,(4) and Kiros T. Berhane(2)

(1) Rancho Los Amigos National Rehabilitation Center, Downey, California, USA

(2) Department of Preventive Medicine

(3) Department of Medicine, Keck School of Medicine, University of Southern California, Los Angeles, California, USA

(4) Division of Environmental Sciences, Joseph A. Mailman School of Public Health, Columbia University, New York, New York, USA