Final Report: Health Effects of long-Term Exposure to Particles and Other Air Pollutants in Elderly Nonsmoking California Residents

EPA Grant Number: R827998
Title: Health Effects of long-Term Exposure to Particles and Other Air Pollutants in Elderly Nonsmoking California Residents
Investigators: Knutsen, Synnove F. , Abbey, David E. , Beeson, Larry
Institution: Loma Linda University
EPA Project Officer: Chung, Serena
Project Period: March 1, 2000 through February 28, 2003 (Extended to February 28, 2005)
Project Amount: $763,910
RFA: Airborne Particulate Matter Health Effects (1999) RFA Text |  Recipients Lists
Research Category: Particulate Matter , Air , Health Effects

Objective:

The primary objective of this research project was to determine the risk of exposure to long-term ambient concentrations of fine particulate matter (PM2.5) on all-cause mortality and nonmalignant respiratory mortality as well as lung cancer, cardiovascular disease incidence, and mortality using data from a cohort of 6,338 nonsmoking California Seventh-day Adventists that have been followed prospectively since 1977.

The secondary research objectives were to: (1) assess the health effects of long-term concentrations of other air pollutants (PM10, PM10-2.5, SO4, SO2, O3, NO2) on the same outcomes as mentioned for the primary objective; (2) assess the health effects of mixed pollutants; (3) assess the health effects of long-term concentrations of all the mentioned air pollutants using multipollutant analysis; and (4) evaluate the health effects of air pollutants in sensitive subgroups.

Summary/Accomplishments (Outputs/Outcomes):

This grant made it possible for us to update air pollution estimates for each of the participants in the Adventist Health Study on the Health Effects of Smog (AHSMOG) cohort through 1998. We also followed up mortality through 1998 and linked with the California cancer tumor registry through 1998.

Our main findings were a strong association between fatal coronary heart disease (CHD) and PM in women but not in men. We found similar risks for fatal lung cancer with increased levels of PM in both males and females.

When looking at sensitive subgroups, we found that persons with diabetes and chronic obstructive pulmonary disease (COPD) have a higher risk of fatal CHD in association with increasing levels of PM.

Tracing

As of May 11, 2001, the tracing of those alive was completed, with a questionnaire response rate of 87 percent. For the deceased, surrogate interviews have been completed with a close friend or relative, with a response rate of 76 percent of all known dead (770 of 1,014).

Estimates of Ambient Air Pollution

PM2.5, visibility data as well as directly monitored data have been received from the whole state of California, and we have calculated the individual exposure measures using the same methodology as before.

Other Pollutants. Sonoma Technologies processed the information collected by the California Air Resources Board (CARB) and provided us with monthly ambient levels of air pollutants (PM10, O3, SO2, NO2, SO4) by ZIP Code through March 2000. Since 1992, several air monitoring stations have been discontinued, and others have been started. This created some challenges when developing new algorithms for assessing ZIP Code-specific ambient air pollution levels. Floyd Peterson has done quality checks with data from Sonoma Technologies to make sure the new algorithms produce comparable ambient air pollutant levels as previous algorithms by comparing 3 overlapping years (1990-1992); the consistency was good for all pollutants. The ZIP Code specific ambient air pollution levels for all the pollutants have thus been linked to each individual in our cohort.

Assessment of Outcome

Mortality. All death certificates have been coded by a certified nosologist, and all mortality outcomes have thus been updated through 1998. Since the start of followup in 1977, there were a total of 2,462 deaths, of which 2,393 were all natural cause (ANC) deaths (ICD-9 < 800). Table 1 gives cause specific mortality in males and females separately.

Table 1. Causes of Death in the AHSMOG Study, 1977-1998, According to Age at Time of Death.

Females only
 

Age at time of death

Underlying cause of death

25-49

50-59

60-69

70-79

80+

Total

     All Natural Causes (ICD: 001 – 799)

9

29

115

314

1010

1477

 

Cardiopulmonary mortality

     Ischemic Heart Disease (ICD-9: 410 – 414)

0

3

19

80

296

398

     Sudden Death (ICD-9: 798.1 – 798.2)

0

0

0

0

0

0

     Nonmalignant Respiratory (ICD-9: 460-519)

0

2

5

18

101

126

     Congestive Heart Disease (ICD-9: 428)

0

1

2

5

25

33

     Pneumonia / Influenza (ICD-9: 480 – 487)

0

1

4

10

74

89

     COPD (ICD-9: 490 – 496)

0

1

1

5

15

22

 

Total cancers

     Lung Cancer (ICD-9: 162)

0

0

2

8

9

19

     Respiratory Cancer (ICD-9: 160 – 165)

0

0

3

8

10

21

 

     All Causes (ICD-9: 001 – 999)

11

30

120

321

1023

1505

 
Males only
     All Natural Causes (ICD: 001 – 799)

3

25

114

266

508

916

 

Cardiopulmonary mortality

     Ischemic Heart Disease (ICD-9: 410 – 414)

0

6

38

77

125

246

     Sudden Death (ICD-9: 798.1 – 798.2)

0

0

0

0

0

0

     Nonmalignant Respiratory (ICD-9: 460 – 519)

0

2

5

18

73

98

     Congestive Heart Disease (ICD-9: 428)

0

0

3

6

17

26

     Pneumonia / Influenza (ICD-9: 480 – 487)

0

1

3

11

42

57

     COPD (ICD-9: 490 – 496)

0

1

0

5

16

22

 

Total cancers

     Lung Cancer (ICD-9: 162)

0

0

5

8

8

21

     Respiratory Cancer (ICD-9: 160 – 165)

0

0

5

9

9

23

 

     All Causes (ICD-9: 001 – 999)

7

32

120

274

524

957

Cancer Incidence

From 1977 through 2000, a total of 1,169 incident cancers (ICD-9: 140-208) occurred in the cohort. Table 2 below gives site-specific incidence by age and gender.

Table 2. Incident Cancer in AHSMOG Cohort, 1977-2000. Incidence from 1992 based on CA tumor registry data only.

Females only
 

Age at time of diagnoses

 
 

Un-known

25-44

45-54

55-64

65-74

75-84

85+

Total
(% of total)

Total cancer ICD-9: 140 – 208

15

18

45

107

178

204

111

678

Breast cancer ICD-9: 174

6

4

17

50

55

71

26

229 (33.8)

Colo-rectal ICD-9: 153 –154

1

1

3

10

31

43

25

114 (16.8)

Corpus uteri ICD-9: 182

2

3

6

29

22

17

5

84 (12.4)

Lung cancer ICD-9: 162

0

0

0

2

7

7

9

25 (3.69)

Urogenital ICD-9: 188 – 9

1

0

2

0

9

6

5

23 (3.37)

Hematopoietic ICD-9: 200 – 8

0

0

0

2

10

5

4

21 (3.10)

Lymph node ICD-9: 196

0

0

0

3

2

7

6

18 (2.65)

 

Males only

Total cancer ICD-9: 140 – 208

10

3

17

72

168

154

67

491

Prostate cancer ICD-9: 185

7

0

1

26

79

73

28

214 (43.6)

Colo-rectal ICD-9: 153 – 154

0

0

2

11

23

22

8

66 (13.4)

Urogenital ICD-9: 188 – 9

0

0

2

6

7

16

5

37 (7.5)

Lymph node ICD-9: 196

0

0

3

3

7

7

5

26 (5.3)

Hematopoietic ICD-9: 200 –8

0

0

0

2

12

7

3

24 (4.9)

Lung cancer ICD-9: 162

2

0

0

3

4

7

4

20 (4.1)

The following sections deal with associations between ambient air pollutants and disease.

All Natural Cause Mortality

Associations with PM10–the AHSMOG Cohort. Using the same model as was used by Abbey, et al. (1999), with no exclusion of baseline diseases, we find virtually the same risks as before [RR=1.04 (95% CI: 1.00-1.09) in males vs. 0.99 (95% CI: 0.95-1.02) in females] with each increment of 10µg/m3 of PM10.

However, for the new analyses, we excluded subjects with CHD, stroke, diabetes, or cancer at baseline. This changed the estimates of risk associated with increases in PM10, and they became nonsignificant and more similar in males and females (RR = 1.02, 95% CI: 0.97-1.08 in males vs. 0.99, 95% CI: 0.95-1.04 in females). When adjusting for O3 in two-pollutant models, the relative risk increased to 1.05 in both males and females, whereas relative risk stayed unchanged when the other gaseous pollutants were added to the model.

Associations With the Gaseous Pollutantsthe AHSMOG Cohort. In single-pollutant as well as two-pollutant models with PM10, O3 was associated with decreased ANC mortality in females (RR = 0.83, 95% CI: 0.73-0.96). There also was a nonsignificant positive association (RR = 1.05, 95% CI: 0.98-1.12) with NO2 in males, which remained virtually unchanged in the two-pollutant model with PM10.

Associations With PM2.5–the Airport Cohort. No significant associations between ANC and PM2.5 were found in single-pollutant analyses. In two-pollutant analyses with O3, however, an increase of 10µg/m3 in PM2.5 was associated with an RR of 1.16 (95% CI: 1.01-1.33) in females, whereas the relationship with males was weaker and not significant (RR = 1.07, 95% CI: 0.90-1.27

Cardiopulmonary Mortality (ICD-9 Codes: 401-440, 460-519)

In our previous report on cardiopulmonary (CP) mortality, Abbey, et al. (1999) found no association between CP mortality and NO2 or SO2. A positive, although nonsignificant, association was found with PM10 and O3 in males and a negative, nonsignificant association in females. The association with PM2.5 was not studied. Using the same model as Abbey, et al., and with longer followup through 1998, we find virtually the same associations as before (males RR = 1.03, 95% CI: 0.98-1.09 and females: RR = 0.98, 95% CI: 0.94-1.03).

In our new analyses, we excluded subjects with CHD, stroke, or diabetes at baseline.

PM10 and Gaseous Pollutants–the AHSMOG Cohort. No significant associations were found between CP mortality and PM10, O3, or NO2 in either single or two-pollutant models in any of the genders. In two-pollutant models with O3, the RR associated with a 10µg/m3 increment in PM10 was 0.97 (95% CI: 0.88-1.08) in males and 1.00 (95% CI: 0.93-1.08) in females. However, SO2 showed a significant and negative association with CP mortality in both single- and two-pollutant models in females (RR = 0.94, 95% CI: 0.91-0.97).

PM2.5–The Airport Cohort. The association between PM2.5 and CP mortality was weak in both genders, but positive in females and negative in males. In two-pollutant models with O3, however, the relationship in females became significant (RR = 1.20, 95% CI: 1.01-1.43, whereas O3 was negative (RR = 0.75, 95% CI: 0.56-0.95).

There was a nonsignificant positive association with NO2 in both males and females that was strengthened in two-pollutant models with all three PM fractions. However, none of these associations reached statistical significance.

Nonmalignant Respiratory Mortality (ICD-9 Codes: 460-519)

Using the same modeling approach as in Abbey, et al. (1999), but with longer followup, we find very similar associations (PM10 with O3: RR = 1.08, 95% CI:0.99-1.17 in males and 1.06, 95% CI: 0.98-1.14 in females). However, for NO2 and SO2 we find a positive relationship in males and slightly negative in females, whereas Abbey, et al. (1999) found negative associations both in males and females.

In our newer modeling approach, we excluded prevalent subjects with COPD at baseline.

Associations with PM10 and Gaseous Pollutants–The AHSMOG Cohort. No significant associations were found in the single-pollutant models. In two-pollutant models, however, there was a positive and significant association with PM10 both with O3 and with NO2 in females (RR = 1.12, 95% CI: 1.00-1.25 and RR = 1.12, 95% CI: 1.02-1.22, respectively). The association with PM10 (+O3) also was elevated in males but did not reach statistical significance (RR = 1.09, 95% CI: 0.96-1.24). Associations with NO2 and SO2 varied around the null value in both single- and two-pollutant models.

Associations with PM2.5–the Airport Cohort. In females, the association with PM2.5was elevated in two-pollutant models with all three gaseous pollutants, but highest with O3 (RR = 1.24, 95% CI: 0.78-1.97). The highest association was found for the coarse thoracic size fraction (PM10-2.5) with O3 (RR = 1.84, 95% CI: 1.01-3.38). No association was found in males.

A positive and significant association was found with NO2 in both single and two-pollutant models in females, respectively (RR = 1.47, 95% CI: 1.06-2.05 and RR = 1.45, 95% CI: 1.03-2.03).

Total Cancer Mortality (ICD-9 Code: 140-172 and 174-209)

The study populations for these analyses were those who were free of cancer at baseline in 1977.

Associations with PM10 and Gaseous Pollutants -AHSMOG Cohort. No significant associations were found in the single-pollutant models. In the two-pollutant models, however, there was a positive association with PM10 in both genders both with O3 and with SO2 (Males: RR = 1.15, 95% CI: 0.98-1.34 and RR = 1.07, 95% CI 0.94-1.21 and Females: RR = 1.07, 95% CI: 0.93-1.22 and RR = 1.07, 95% CI: 0.96-1.20). None of the gaseous pollutants showed a relationship with total cancer mortality.

Associations with PM2.5–the Airport Cohort. The strongest association was found for PM2.5 in two-pollutant models with O3 (RR = 1.15, 95% CI: 0.84-1.57 in males and RR = 1.28, 95% CI: 0.94-1.74 in females).

Fatal Coronary Heart Disease (ICD-9: 410-414)

Single Pollutant Models. The association between PM10 and fatal CHD was different between males and females with no association in males but a significant association in females (1.22, 95% CI: 1.07-1.40) versus males (0.94, 95% CI: 0.80-1.11). For PM2.5, the association was stronger (Females: RR = 1.42, 95% CI: 1.06-1.90 vs. Males: RR = 0.90, 95% CI: 0.76-1.05). NO2 showed a positive nonsignificant association in both males and females of 16 and 17 percent increased risk, respectively.

Two-Pollutant Models. The association between PM2.5 and fatal CHD was strengthened in two-pollutant models with the gaseous pollutants, especially with O3 in females. Each 10 µg/m3 increase in PM2.5 was associated with a RR of 2.00 (95% CI: 1.51-2.64). The other two PM fractions also were strengthened when O3 was added to the model. No significant effect was found in males.

Mortality in Sensitive Subgroups

The following sensitive subgroups have been identified:

  • Older age (> 64 yrs and > 74 yrs)
  • Prevalent CHD
  • Prevalent CHD, stroke, or diabetes
  • Past smokers
  • Exposure to environmental tobacco smoke (ETS)
  • Prevalent COPD.

For ANC mortality, no significant associations were found with PM10 in any of the sensitive subgroups.

For cardiopulmonary mortality, males with prevalent CHD, stroke or diabetes showed a positive, association with PM10 in both single- (RR = 1.18, 95% CI: 1.07-1.30) and two-pollutant models with SO2 (RR = 1.22, 95% CI: 1.091.37). No associations were seen in other subgroups and no associations were found in females.

For fatal CHD, the pattern is quite the opposite, with females in sensitive subgroups being at higher risk. Both in elderly females and those with prevalent COPD, PM was associated with significantly higher risks of fatal CHD with RRs ranging from 1.23 (95% CI: 1.07-1.41) (for PM10) to 1.92 (95% CI: 1.32-2.81) (for PM2.5+O3) in the elderly group and from 1.30 (95% CI: 1.03-1.63) (for PM10) to 1.72 (95% CI: 0.98-3.03) (for PM2.5+O3) in those with prevalent COPD. Strangely this trend was reversed for the sensitive group with prevalent CHD, stroke, or diabetes, where the association between fatal CHD and PM was greater in males, although not significant, with PM10 (RR = 1.12, 95% CI: 0.93-1.35) and PM2.5 (RR = 1.21, 95% CI: 0.87-1.68), whereas no association was found in females.

Lung Cancer Incidence and Mortality (ICD-9 code: 162)

We previously have found an association between ambient air pollutants and risk of fatal lung cancer (Abbey, et al., 1999). In single pollutant models, for some pollutants (PM10, O3) the effect was stronger in males than females and for others (SO2 and NO2) the risk estimate was stronger in females. We have also reported an association between O3, PM10, and SO2 and risk of incident lung cancer (Beeson, et al., 1998).

Mean Concentrations. For incident lung cancer we find a 22 percent increased risk (95% CI: 0.91-1.64) for an increase of 10 µg/m3 in PM10 in the single pollutant, which increased to RR = 1.65 (95% CI: 1.06-2.58) in two-pollutant models with O3. No effect was found in females.

The risk of fatal lung cancer was similar in the two genders. Males: RR = 1.11 (95% CI: 0.83-1.48) increasing to 1.32 (95% CI: 0.86-2.03) in two-pollutant models with O3 and females: RR = 1.14 (95% CI: 0.83-1.58) increasing to 1.44 (95% CI: 0.89-2.32) in two-pollutant models with O3.

No independent effect of O3 was found for incident or fatal lung cancer, either in single-pollutant or two-pollutant models with PM10 or the other gaseous pollutants (NO2 and SO2).

Exceedance Frequencies. In males, a monotonic increase in risk of incident lung cancer was found with increasing levels of PM10 both in single- and two-pollutant models with O3. For each 30 days/year of PM10 levels above 100 µg/m3, the risk of incident lung cancer increased by 48 percent (RR = 1.48, 95% CI: 1.04-2.11) and by 81 percent in two-pollutant models with O3 (RR = 1.81, 95 % CI: 1.11-2.94). The findings in females were nonsignificant and weaker (RR = 1.17, 95% CI: 0.70-1.96).

For fatal lung cancer, similar findings were seen in both males and females. In men the risk associated with each 30 days/year of PM10 greater than 100 µg/m3, was increased by 19 percent (95% CI: 0.93-1.71) in single-pollutant models and by 49 percent (95% CI: 0.92-2.42) in the two-pollutant model with O3. For females, the corresponding increases were 16 (95% CI: 0.78-1.76) and 51 percent (95% CI: 0.87-2.60).

For O3 and for incident lung cancer, each 100 hours/year in excess of 150 ppb was associated with a 25 percent increase in males, but this relationship disappeared in the two-pollutant model with PM10. In the two-pollutant model for females, a 34 percent nonsignificant relationship was found.

For fatal lung cancer, in males, each 100 hours/year in excess of 150 ppb of O3 was associated with an increase of 22 percent (95% CI: 0.87-1.71), which increased to 48 percent (95% CI: 0.89-2.46) in the two-pollutant model with PM10. Similar increases were found in females with 24 and 36 percent, respectively.

Incidence of Non-Hodgkin Lymphoma

This paper has been resubmitted after being rejected by “Cancer, Causes and Control.” A 30 percent increased risk of non-Hodgkin lymphoma was found for each increment of 10 µg/m3 of PM10, which remained unchanged in two-pollutant models, with each of the gaseous pollutants. For the gaseous pollutants, there were positive, although not significant, associations with risk of non-Hodgkin lymphoma.

Hospitalizations for Cardiovascular Diseases (1972-1982)

For each increment of 10 µg/m3 of PM10 the risk of being hospitalized for cardiovascular disease was increased by 6 percent (95% CI: 1.001-1.12) in females. No significant association was found in males (RR = 1.02, 95% CI: 0.95-1.09). For PM2.5, the risk associated with the same increment was stronger, although they did not reach statistical significance (Females: RR = 1.16, 95% CI: 0.90-1.50 and Males: RR = 1.03, 95% CI: 0.74-1.45).

Conclusions:

This study assessed the association between PM and risk of mortality and cancer incidence in single- and two-pollutant models.

The association between ANC mortality was weak for both PM and gaseous pollutants. For cardiopulmonary mortality, we found a significant association with PM2.5 in females in two-pollutant models with ozone (RR = 1.20). This is probably a reflection of the strong association we find between an increment of 10 μg/m3 PM2.5 and fatal CHD in females (RR = 2.00) in two-pollutant models with ozone, whereas no significant association was found in males. We also found a 10-12 percent increased risk of nonmalignant respiratory mortality with each increment of 10 μg/m3 of PM10 in two-pollutant models in males and females. The association with PM was strongest for the coarse thoracic size fraction in females with RR = 1.84.

For cancer, we found no significant association between any of the PM fractions and total cancer mortality. However, we did find a similar, but nonsignificant association between PM10 and lung cancer mortality in both males and females. An effect of PM10 on lung cancer incidence was only found in males in two-pollutant models with ozone (RR = 1.65).

We also found an association between PM10 and non-Hodgkin lymphoma with a 30 percent increased risk for each 10 μg/m3 increment of PM10.

When studying the effect of PM in sensitive subgroups (elderly, past smokers, those exposed to ETS, and those with prevalent CHD, stroke, diabetes, or COPD), we found similar risks as found in the cohort as a whole. However, males with prevalent CHD, stroke, or diabetes showed a significant increased risk cardiopulmonary mortality.

Based on our findings, there seems to be an association between long-term ambient levels of PM and risk of cardiac mortality as well as lung cancer and non-Hodgkin lymphoma. Further studies in larger cohorts that are free of the major confounding of smoking and alcohol are needed to study other cause-specific mortality as well as incidence.

Assessment of Accomplishments

The primary objectives have been met for this study.

Several of the secondary objectives are still being addressed by graduate students. The hospitalization data is being analyzed by M.P.H. students and will be completed as they complete their research requirements for their degrees.


Journal Articles on this Report : 1 Displayed | Download in RIS Format

Other project views: All 19 publications 2 publications in selected types All 1 journal articles
Type Citation Project Document Sources
Journal Article Chen LH, Knutsen SF, Shavlik D, Beeson WL, Petersen F, Ghamsary M, Abbey D. The association between fatal coronary heart disease and ambient particulate air pollution:are females at greater risk? Environmental Health Perspectives 2005;113(12):1723-1729. R827998 (Final)
R830547 (2005)
R830547 (2006)
R830547 (2007)
R830547 (Final)
  • Full-text from PubMed
  • Abstract from PubMed
  • Associated PubMed link
  • Full-text: EHP-Full Text HTML
  • Other: EHP-Full Text PDF
  • Supplemental Keywords:

    ambient air, ozone, exposure, risk, risk assessment, health effects, human health, sensitive populations, carcinogen, population, elderly, cumulative effects, susceptibility, epidemiology, modeling, monitoring, analytical, southwest, California, CA,, RFA, Scientific Discipline, Health, Air, ENVIRONMENTAL MANAGEMENT, Geographic Area, particulate matter, Health Risk Assessment, air toxics, Epidemiology, State, Risk Assessments, Susceptibility/Sensitive Population/Genetic Susceptibility, Biochemistry, Atmospheric Sciences, indoor air, tropospheric ozone, genetic susceptability, Biology, Risk Assessment, ambient air quality, elderly adults, PM10, sulfates, Nitrogen dioxide, sensitive populations, PM 2.5, long term exposure, exposure and effects, stratospheric ozone, acute lung injury, ambient air, exposure, air pollution, lung cancer, Sulfur dioxide, particulate exposure, chronic health effects, sensitive subjects, human exposure, Acute health effects, epidemiological studies, elderly, PM, mortality, tobacco smoke, California (CA), indoor air quality, age dependent response, cumulative effects, respiratory, exposure assessment, genetic susceptibility, environmental hazard exposures, toxics

    Relevant Websites:

    http://www.llu.edu/llu/health/ahsmog.htm Exit

    Progress and Final Reports:

    Original Abstract
  • 2000 Progress Report
  • 2001 Progress Report
  • 2002 Progress Report
  • 2003 Progress Report