Air Pollution and Adverse Birth Outcomes in the

Air Pollution and Adverse Birth Outcomes in the South Coast Air Basin, 1989 -2000 Michelle Wilhelm, Ph. D. Beate Ritz, M. D. , Ph. D. Department of Epidemiology UCLA School of Public Health

Why Study Air Pollution and Pregnancy? Ø Developing organism is uniquely sensitive to environmental toxins within a short time window Ø Adverse outcomes are common; in US: Ø ~10% are preterm Ø ~ 5% are low weight Ø Immediate and long term health effects Ø Infant morbidity and mortality Ø Adverse effects on adult health?

Existing Epidemiologic Studies Ø Most completed in last 5 years (35 published studies) Ø China, U. S. , Brazil, Czech Republic, Canada, S. Korea, Taiwan, Lithuania, Croatia, Sweden, Mexico Ø Two recent review articles Ø Maisonet et al. 2004, Environmental Research, May; 95(1): 106 -15 Ø Glinianaia et al. 2003, Epidemiology, 15: 36 -45

Outcome Events Studied Ø Deaths Ø Intrauterine (28 weeks of gestation to birth) Ø neonatal (<28 days) Ø infant (< 1 year) Ø Low birth weight (LBW) Ø Ø Ø Weight at birth <2500 g Born LBW at term vs. preterm Reduction in mean weight Small for gestational age (SGA; <10 th percentile of weight for gestational age) Preterm births (<37 weeks of gestation) or Ø reduction of mean gestational age Ø Malformations (cardiac)

Methodologic differences Ø Outcome categories Ø Ø Term LBW, all LBW, SGA, preterm and/or LBW Fetal, neonatal, infant deaths Ø Air pollutants Ø Area-wide averages (all monitoring stations) Ø Use monitor closest to maternal residency Ø Ø Ø Pollutants measured, mixtures, co-pollutants Scaling of units for pollutants Timing of exposure during most relevant pregnancy periods (varies by pollutant and outcome? ) Ø Covariates included in model

South Coast Air Basin Ø Large number of births (~half of all CA births, most in LA county) Ø Birth certificates are readily available Ø Network of government air monitoring stations Ø Nonattainment for PM 10, PM 2. 5 and O 3

Summary of So. CAB Studies Ø Ø Ø Ritz and Yu (1999) evaluated association between CO exposures during 3 rd trimester and risk of term LBW Infants born 1989 -1993 to women living in So. CAB Exposure based on CO measurements at SCAQMD monitoring stations (n=18) – Focused on mothers who resided within 2 miles of a CO monitoring station (n=125, 573) Ø For each child calculated last trimester average CO concentration using the ambient monitoring station within 2 miles

Summary of So. CAB Studies Ø Logistic regression used to evaluate association between 3 rd trimester CO and term LBW Ø Dichotomous outcome - <2, 500 g vs. 2, 500 g Ø Variables included in models Ø Ø Ø Ø Ø parity sex of the infant maternal age maternal race/ethnicity maternal educational attainment interval since previous live birth <12 months prenatal care gestational age transportation time to work (from census data)

Summary of So. CAB Studies Ø Other risk factors not available on birth certificates Øpre-pregnancy weight, weight gain, and height of mother Øsocial factors (marital status? , occupational exposures to toxins? ) Øbehavioral factors (e. g. smoking, caffeine use, alcohol consumption during pregnancy)

Adjusted Odds Ratios (95%CI) for Term LBW 3 rd trimester ambient CO levels 1989 -1993, 18 monitoring stations in So. CAB All children Higher parity children Young Women case N=2, 809 case N=1, 454 case N=420 non-case N=122, 764 non-case N=73, 687 non-case N=15, 111 ___________________________________ CO level (ppm): <2. 2 1. 0 2. 2 - <5. 5 1. 04 1. 03 1. 02 (0. 96, 1. 13) (0. 92, 1. 15) (0. 83, 1. 26) 1. 22 1. 33 1. 54 (1. 03, 1. 44) (1. 07, 1. 65) (1. 07, 2. 22) 5. 5 1. 0 ___________________________

Adjusted Odds Ratios (95%CI) for Term LBW 3 rd trimester ambient CO levels 1989 -1993, 6 monitoring stations in So. CAB w/ CO, NO 2, O 3, PM 10 All children Higher parity children Young Women case N=1, 100 case N=584 case N=146 non-case N=46, 921 non-case N=28, 313 non-case N=5, 918 ___________________________________ CO level (ppm): <2. 2 1. 0 2. 2 - <5. 5 1. 10 1. 24 1. 30 (0. 91, 1. 32) (0. 96, 1. 60) (0. 77, 2. 20) 1. 38 1. 92 5. 08 (0. 86, 2. 22) (1. 02, 3. 62) (1. 77, 14. 63) 5. 5 1. 0 ___________________________

Summary of So. CAB Studies Ø Ritz et al. (2000) evaluated association between air pollution and preterm birth ØInfants born 1989 -1993 to women living in So. CAB (n=97, 158) Ø Exposure assessment similar to Ritz and Yu (1999) ØFirst month of pregnancy and weeks prior to birth

Summary of So. CAB Studies Logistic regression with dichotomous outcome (<37 wks vs. 37 wks completed gestation) Ø Controlled for: Ø Ømaternal age, race, education; parity, interval since previous live birth, prenatal care, infant gender, previous LBW or preterm birth, smoking reported as “pregnancy complications”, season

Odds ratio (and 95% CI) for Preterm Birth Per 50 ug/m 3 increase in PM 10 and 3 ppm increase in CO 1989 -1993 Crude Two Exposure Other Risk Multipollutant OR Periods Factors+Season Model _____________________________________________ PM 10 – All stations: 6 wks 1. 20 1. 18 1. 15 1. 19 (1. 09 -1. 33) (1. 07 -1. 31) (1. 04 -1. 26) (1. 03 -1. 29) (1. 01 -1. 40) 1 st month 1. 16 (1. 06 -1. 26) 1. 13 (1. 04 -1. 24) 1. 09 (1. 00 -1. 19) 1. 09 (0. 99 -1. 20) 1. 12 (0. 97 -1. 29) CO – All stations: 6 wks 1. 12 (1. 08 -1. 17) 1. 13 (1. 08 -1. 17) 1. 06 (1. 02 -1. 10) 1. 04 (0. 99 -1. 10) 1. 05 (0. 97 -1. 12) 1. 05 (1. 01 -1. 09) 1. 01 (0. 97 -1. 04) 1. 04 (0. 99 -1. 09) 1. 03 (0. 96 -1. 10) 1 st month 1. 04 (1. 01 -1. 09) _____________________________________________

Summary of So. CAB studies Ø Ø Ritz et al. (2002) evaluated association between air pollution and occurrence of birth defects Used California Birth Defect Monitoring Program data to evaluate neonates/fetuses delivered in So. CAB during 1987 -1993 Evaluated 6 different common heart defects Exposure based on SCAQMD monitoring station data during first 3 months of pregnancy for each infant

CO and (isolated) Ventricle Septum Defects (multi-pollutant model)

O 3 and Pulmonary artery and valve defects (multi-pollutant model)

Results on Birth Defects Ø Risk of certain cardiac heart defects was three times greater at high exposure levels Ø Ventricle septum birth defects (CO) Ø Aortic and pulmonary artery and valve defects (O 3) Ø Risks were observed in 2 nd month of pregnancy when heart formation occurs

Exposure Assessment Ø Based on concentrations measured at government air monitoring stations Ø May not account for differential exposure within neighborhoods due to proximity to sources Ø Women residing closer to freeways and major roadways may experience greater exposure Ø Pollutants released directly in motor vehicle exhaust Ø Pollutants formed adjacent to roadways

Exposure Assessment Ø Existing measurement data indicate concentrations of motor vehicle exhaust pollutants (CO, NO 2, primary particles) are: Ø Elevated near roadways and decline rapidly with distance Ø Correlated with traffic counts Ø Elevated inside and outside homes and schools located close to heavy traffic roadways

Source: Ott, 1977

Is CO a marker for traffic related pollution? Y. Zhu and W. Hinds, UCLA Particle center Epidemiologic studies ignore potential spatial heterogeneity of vehicle-related air pollution when using exposure data from government air monitoring stations

Traffic Density How can we estimate traffic-related contributions using existing data for large areas? Ø Self-reported traffic density on street of residence Ø Residential distance to major roads/freeways Ø Measured traffic density on main roads near homes Ø Average traffic density in census block groups, tracts or wards Ø Distance weighted traffic density (DWTD) Ø Air dispersion models (e. g. Caline 4)

Summary of So. CAB Studies Ø Wilhelm and Ritz (2003) evaluated association between residential proximity to traffic and ABOs Ø For 1994 -96, in 112 LA zip codes, we identified all Ø Term low birth weight (LBW) and preterm infants (n=31, 653) and a random sample of controls (~same n) Ø Mapped residential birth addresses using GIS (ESRI Street. Map) Ø ~86% had electronic address data, of those ~91% could be mapped Ø Transferred Caltrans annual average daily traffic (AADT) count data for each year on to ESRI Street. Map

Distance Weighted Traffic Density (DWTD) DWTD value calculated for each subject Ø for all streets within 750 ft. (228. 6 m) radius of home Ø weighted traffic count on each street by distance from home to street (using a Gaussian distribution) Ø summed weighted counts for all streets within radius

Summary of So. CAB Studies Ø Evaluated change in estimates when adjusting for following risk factors (based on birth records) Ø Level of prenatal care, maternal age, maternal race/ethnicity, maternal education, infant sex, parity, interval since last live birth, previous LBW or preterm infant, gestational age (for LBW), season, year of birth Ø Also evaluated change in estimates when adjusting for: Ø Background air pollution concentrations (annual averages) Ø One or more freeways within buffer Ø Census block-group level indicators of SES Ø Household income, per capita income, fraction of children in poverty, age of structure, median home value, median gross rent

Association (RR point estimate, 95% CI) between Residential DWTD and Risk of Preterm Birth for infants born between 1994 -1996 to mothers living in 112 zip codes located in LA County, CA Preterm cases=13, 464 controls=21, 124 DWTD Unadjusted DWTD Adjusted for covariates, background concentrations and one or more freeways w/in 750 ft. buffer Quintile of Distance-Weighted Traffic Density (DWTD) <20 th percentile 20 -<40 th percentile 40 th-<60 th percentile 60 th-<80 th percentile ≥ 80 th percentile 1. 00 (0. 94 -1. 07) 1. 05 (0. 98 -1. 11) 1. 10 (1. 04 -1. 17) 1. 11 (1. 04 -1. 18) 1. 0 0. 99 (0. 93 -1. 05) 1. 02 (0. 96 -1. 09) 1. 07 (1. 01 -1. 13) 1. 08 (1. 01 -1. 15) Fall/Winter Third Trimester (birth month Jan-June) <20 th percentile 20 -<40 th percentile 40 th-<60 th percentile 60 th-<80 th percentile ≥ 80 th percentile 1. 0 0. 99 (0. 91 -1. 09) 1. 05 (0. 97 -1. 15) 1. 11 (1. 02 -1. 21) 1. 18 (1. 08 -1. 28) 1. 0 0. 99 (0. 90 -1. 07) 1. 04 (0. 95 -1. 13) 1. 08 (1. 00 -1. 18) 1. 15 (1. 06 -1. 26) Spring/Summer Third Trimester (birth month July-Dec) <20 th percentile 20 -<40 th percentile 40 th-<60 th percentile 60 th-<80 th percentile ≥ 80 th percentile 1. 01 (0. 92 -1. 10) 1. 04 (0. 95 -1. 13) 1. 10 (1. 00 -1. 20) 1. 04 (0. 96 -1. 14) 1. 0 0. 99 (0. 91 -1. 08) 1. 01 (0. 92 -1. 10) 1. 05 (0. 97 -1. 15) 1. 01 (0. 92 -1. 10) DWTD percentile values are: <1, 537, 1, 537 -5, 338, 5, 339 -11, 722, 11, 723 -24, 711, ≥ 24, 712.

Low inversion layers trap pollutants in colder seasons

Association (OR point estimate, 95% CI) between Residential DWTD and Risk of LBW and Preterm Birth for infants born between 1994 -1996 to mothers living in 112 zip codes located in LA County, CA Preterm and LBW cases=3, 509 controls=21, 124 DWTD Unadjusted DWTD Adjusted for covariates, background concentrations and one or more freeways w/in 750 ft. buffer Quintile of Distance-Weighted Traffic Density (DWTD) <20 th percentile 20 -<40 th percentile 40 th-<60 th percentile 60 th-<80 th percentile ≥ 80 th percentile 1. 02 (0. 91 -1. 15) 1. 05 (0. 93 -1. 17) 1. 16 (1. 04 -1. 30) 1. 13 (1. 01 -1. 26) 1. 01 (0. 90 -1. 14) 1. 05 (0. 93 -1. 18) 1. 14 (1. 01 -1. 28) 1. 12 (0. 99 -1. 26) Fall/Winter Third Trimester (birth month Jan-June) <20 th percentile 20 -<40 th percentile 40 th-<60 th percentile 60 th-<80 th percentile ≥ 80 th percentile 1. 0 0. 96 (0. 82 -1. 14) 0. 99 (0. 84 -1. 17) 1. 14 (0. 97 -1. 34) 1. 24 (1. 06 -1. 45) 1. 0 0. 96 (0. 81 -1. 14) 0. 99 (0. 83 -1. 17) 1. 12 (0. 95 -1. 33) 1. 24 (1. 04 -1. 47) Spring/Summer Third Trimester (birth month July-Dec) <20 th percentile 20 -<40 th percentile 40 th-<60 th percentile 60 th-<80 th percentile ≥ 80 th percentile 1. 08 (0. 92 -1. 27) 1. 10 (0. 94 -1. 29) 1. 17 (1. 00 -1. 38) 1. 03 (0. 87 -1. 20) 1. 06 (0. 90 -1. 25) 1. 11 (0. 94 -1. 31) 1. 16 (0. 98 -1. 36) 1. 01 (0. 85 -1. 20) DWTD percentile values are: <1, 537, 1, 537 -5, 338, 5, 339 -11, 722, 11, 723 -24, 711, ≥ 24, 712.

Association (OR point estimate, 95% CI) between Residential DWTD and Risk of Term LBW for infants born between 1994 -1996 to mothers living in 112 zip codes located in Los Angeles County, CA Term LBW cases=3, 771 controls=26, 351 DWTD Unadjusted DWTD Adjusted for covariates, background concentrations and one or more freeways w/in 750 ft. buffer Quintile of Distance-Weighted Traffic Density (DWTD) <20 th percentile 20 -<40 th percentile 40 th-<60 th percentile 60 th-<80 th percentile ≥ 80 th percentile 1. 0 1. 13 (1. 02 -1. 27) 1. 16 (1. 04 -1. 29) 1. 18 (1. 05 -1. 31) 1. 16 (1. 04 -1. 30) 1. 0 1. 11 (0. 99 - 1. 25) 1. 16 (1. 03 - 1. 30) 1. 15 (1. 02 - 1. 29) 1. 11 (0. 99 - 1. 26) Fall/Winter Third Trimester (birth month Jan-June) <20 th percentile 20 -<40 th percentile 40 th-<60 th percentile 60 th-<80 th percentile ≥ 80 th percentile 1. 0 1. 18 (1. 00 -1. 38) 1. 26 (1. 08 -1. 48) 1. 29 (1. 10 -1. 51) 1. 33 (1. 13 -1. 55) 1. 0 1. 20 (1. 01 - 1. 42) 1. 33 (1. 12 - 1. 58) 1. 33 (1. 12 - 1. 57) 1. 33 (1. 11 - 1. 58) Spring/Summer Third Trimester (birth month July-Dec) <20 th percentile 20 -<40 th percentile 40 th-<60 th percentile 60 th-<80 th percentile ≥ 80 th percentile 1. 0 1. 10 (0. 95 -1. 28) 1. 07 (0. 92 -1. 24) 1. 08 (0. 93 -1. 26) 1. 03 (0. 88 -1. 20) 1. 04 (0. 89 - 1. 22) 1. 02 (0. 87 - 1. 20) 1. 01 (0. 86 - 1. 18) 0. 96 (0. 81 - 1. 13) DWTD percentile values are: <1, 524, 1, 524 -5, 266, 5, 267 -11, 568, 11, 569 -24, 579, ≥ 24, 580.

DWTD Measure Ø Relatively crude exposure assessment approach ØDoes not take meteorology or motor vehicle emission rates into account Can be applied to large population using existing data Ø Experimental support for exponential decay from roadways Ø

EPOS Survey Ø Nested case-control study of LA County births Ø LBW/Preterm children and term normal weight controls (targeted n=1000 case and 1000 controls over 1 year - 2003) Ø Ø Ø Randomly selected from birth certificates in 111 LA zip codes Mothers interviewed by phone or completed a mail survey 3 -7 months after births Collected information on: Ø indoor air pollution sources Ø in-transit exposures Ø time-activity data Ø Individual level risk factors for LBW/preterm birth during pregnancy: Ø Smoking, alcohol, occupation, psychosocial stress

EPOS Response Rates # of Mail Surveys # of Telephone Surveys # of Home Visits Total # of surveys completed Response rates: All selected subjects Subjects w/available and correct telephone no’s and addresses 1020 1491 64 2575 39% ~75%

Data Linkages Ø Most studies based on existing data sources ØBirth and death certificate information, CBDMP data Ø Residential addresses and SSN’s now available from state for birth certificates ØAir pollution measurements from government monitors ØTraffic data from Caltrans

Major Questions Ø Adequacy of air pollution exposure assessment methods? Ø Reliance on existing monitoring networks Ø Potential intracommunity variation in pollutant concentrations Ø Time spent indoors and indoor sources Ø Reliance on birth certificate zip code or address Ø Residential and work mobility patterns for pregnant women Ø Which pollutants are important? Ø Associations reported for CO, NO 2, TSP, SO 2, PM 10, PM 2. 5, PAHs Ø Identifying pollutants of concern and biological mechanisms of action

Major Questions Ø Vulnerable pregnancy periods? Ø Associations reported most consistently for early or late pregnancy Ø Confounding due to unmeasured risk factors? Ø E. g. , smoking, alcohol, diet, pregnancy weight gain Ø Ø Susceptible subpopulations? Importance of early pregnancy losses (unreported)?

Discussion

Adjusted Odds ratios (and 95% CI) for Term LBW Ambient CO levels at South Central LA station only, 1989 -1993 1 st Trim 2 -mile radius 2 nd Trim 2 -mile radius 3 rd Trim 5 -mile radius _____________________________________________________ CO level: <50 th 1. 0 50 -95 th >=95 th 1. 0 0. 87 (0. 73 -1. 03) 1. 02 (0. 85 -1. 20) 1. 06 (0. 89 -1. 26) 1. 07 (0. 99 -1. 16) 0. 82 (0. 54 -1. 24) 0. 97 (0. 65 -1. 44) 1. 24 (0. 87 -1. 77) 1. 24 (1. 06 -1. 45) ____________________________________________________ 2 mile radius: case N=572, non-case N=23, 533; 5 mile radius: case N=2, 805, non-case N=94, 160

u Adjusting for census block group level SES indicators changed association estimates minimally (maximum of 6% but mostly by 1 -2%) u Stratification on median values of SES indicators suggested greater associations in lower SES areas Outcome Association for women residing in census block groups where fraction of children in poverty <28% (median) Association for women residing in census block groups where fraction of children in poverty ≥ 28% Preterm birth 1. 03 (0. 95 -1. 11) 1. 15 (1. 03 -1. 27) Preterm and LBW birth 1. 07 (0. 90 -1. 27) 1. 18 (0. 97 -1. 43) Term LBW 1. 07 (0. 90 -1. 27) 1. 25 (1. 03 -1. 51) *Results for highest vs. lowest DWTD quintile, adjusting for covariates, background pollutant concentrations (continuous) and census block group indicators of SES (continuous).