American Journal of Epidemiology © The Author(s) 2018. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. Vol. 187, No. 11 DOI: 10.1093/aje/kwy116 Advance Access publication: June 12, 2018 Original Contribution Temperature and Term Low Birth Weight in California * Correspondence to Dr. Rupa Basu, California Office of Environmental Health Hazard Assessment, Air and Climate Epidemiology Section, 1515 Clay Street, 16th Floor, Oakland, CA 94612 (e-mail: Rupa.Basu@oehha.ca.gov). Initially submitted December 15, 2017; accepted for publication May 25, 2018. Few investigations have explored temperature and birth outcomes. In a retrospective cohort study, we examined apparent temperature, a combination of temperature and relative humidity, and term low birth weight (LBW) among 43,629 full-term LBW infants and 2,032,601 normal-weight infants in California (1999–2013). The California Department of Public Health provided birth certificate data, while meteorological data came from the California Irrigation Management Information System, US Environmental Protection Agency, and National Centers for Environmental Information. After considering several temperature metrics, we observed the best model fit for term LBW over the full gestation (per 10-degrees-Fahrenheit (°F) increase in apparent temperature, 13.0% change, 95% confidence interval: 4.1, 22.7) above 55°F, and the greatest association was for third-trimester exposure above 60°F (15.8%, 95% confidence interval: 5.0, 27.6). Apparent temperature during the first month of pregnancy exhibited no significant risk, while the first trimester had a significantly negative association, and second trimester, last month, and last 2 weeks had slightly increased risks. Mothers who were black or older, delivered male infants, or gave birth during the warm season had infants at the highest risks. This study provides further evidence for adverse birth outcomes from heat exposure for vulnerable subgroups of pregnant women. adverse birth outcomes; apparent temperature; California; epidemiology; humidity; low birth weight; temperature; term low birth weight Abbreviations: BMI, body mass index; LBW, low birth weight; LMP, last menstrual period; PM2.5, particulate matter with an aerodynamic diameter less than or equal to 2.5 μm; WIC, Special Supplemental Nutrition Program for Women, Infants, and Children; ZCTA, zip code tabulation area. ambient temperature or extreme heat and LBW specifically remains largely unexplored. However, heat exposure causes damage to the placenta, developing cells, and vascular system (8). Our study objective was to evaluate LBW from both longand short-term temperature exposure in California from 1999 to 2013. Because many LBW infants may be premature with other risk factors, we limited our study population to term infants. As with all LBW infants, term LBW infants (defined as both less than 2,500 grams and delivered between 37 and 44 gestational weeks) also present severe public health challenges, affecting almost 2% of births (6), as they have greater risks for infant mortality (9) as well as lifelong consequences, ranging from respiratory difficulties throughout childhood (10) to psychological distress in adulthood (11). We also considered the following maternal factors as effect modifiers to identify high-risk groups: age, educational attainment at the time of giving birth, While the associations between temperature and several health outcomes have been well-documented (1, 2), only recently have epidemiologic investigations considered adverse birth outcomes such as preterm delivery, stillbirth, and low birth weight (LBW; <2,500 grams) from temperature (3). Results from previous studies have varied depending on locations, metrics and windows of exposure, covariables, and outcomes (3). In the United States, the rates of preterm delivery, stillbirth, and low birth weight are 12% (4), <1% (5), and 8% (6), respectively. While the rates for preterm delivery and stillbirth have remained stable (4, 5), the rate for LBW continues to rise steadily (6). Many specific risk factors remain largely unknown for LBW, although air pollution exposure, smoking, alcohol and drug use during pregnancy, maternal body mass index (BMI) and weight gain during pregnancy, and high blood pressure and other chronic health conditions have been reported in previous studies (7). The link between 2306 Am J Epidemiol. 2018;187(11):2306–2314 Downloaded from https://academic.oup.com/aje/article-abstract/187/11/2306/5036486 by guest on 09 June 2020 Rupa Basu*, Reina Rau, Dharshani Pearson, and Brian Malig Temperature and Term Birth Weight in California 2307 race/ethnicity, prepregnancy BMI, enrollment in the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC), poverty level, and smoking. Because prior evidence supports the link between long-term exposure to criteria air pollutants and increased risk of LBW (12), we considered them as potential confounders and effect modifiers. METHODS Study population Meteorological and air pollution data Meteorological data were obtained from 3 different sources: the California Irrigation Management Information System (http://www.cimis.water.ca.gov/), the US Environmental Protection Agency Air Quality System Data Mart (http://www.epa. gov/ttn/airs/aqsdatamart/), and the National Centers for Environmental Information (http://www.ncdc.noaa.gov/). As described previously (14), we calculated mean apparent daily temperature, a combination of temperature and relative humidity, in degrees Fahrenheit (°F). A weather station within 10 km of the maternal zip code tabulation area (ZCTA) in the same climate zone for each birth was chosen to assign exposure based on proximity and coverage of possible pregnancies. The 16 climate zones were previously defined by the California Energy Commission based on weather patterns, energy use, and other factors related to climate (15). Each meteorological monitor had to include enough data to cover 95% of possible pregnancies during the study period. The US Environmental Protection Agency provided air pollution data for the 14-year study period. We evaluated potential confounding and effect modification using maximum daily Am J Epidemiol. 2018;187(11):2306–2314 Statistical analysis Using a retrospective cohort design and logistic regression models, we examined the association between apparent temperature and term LBW, while adjusting for season and year of delivery, maternal age, maternal education, maternal race/ ethnicity, and infant sex. Each of these variables was tested for confounding separately using the base model containing apparent temperature and term LBW. We also abstracted data from the US Census Bureau to examine neighborhood socioeconomic status, specifically percentage of families living below the poverty level (https://www.census.gov/programs-surveys/acs/data.html; https://www.census.gov/census2000/sumfile3.html). To better account for nonlinearity for apparent temperature during the entire year and long-term trends over time, we stratified by a cutoff value for apparent temperature based on dose-response curves for each temperature metric. We added an apparent temperature squared term to the model with each apparent temperature exposure metric to check for linearity and also ensured that there was no remaining nonlinearity after stratification by the lower and higher cutoff values. All results are first presented by climate zone then combined using meta-analyses to produce an overall estimate. They are expressed as a percent change in odds of term LBW per 10-degree increase (°F), along with corresponding 95% confidence intervals. We used SAS, version 9.3 (SAS Institute, Inc., Cary, North Carolina), to perform all analyses. Downloaded from https://academic.oup.com/aje/article-abstract/187/11/2306/5036486 by guest on 09 June 2020 Prior to beginning this study, the California Health and Human Services Agency’s Committee for the Protection of Human Subjects approved the research protocol. We abstracted data from all full-term normal and LBW births occurring in California between January 1, 1999, and December 31, 2013. Using birth certificate data from the Center for Health Statistics and Informatics Birth Data File (13), we included singleton births with data available for the following infant characteristics: birth date (month, day, and year), birth weight (in grams), gestational age (37–44 weeks), and sex (male or female). The difference between the date of birth and the date of the last menstrual period (LMP) defined gestational age. If only the month of LMP was available, LMP was assigned as the 15th day of the month. To ensure that births with shorter gestations were not overrepresented in the last year of data, we limited our study to births with LMP before February 28, 2013. We also abstracted data on the following maternal characteristics: age (in years, categorized as 11–18, 19–24, 25–29, 30–39, and ≥40), race/ethnicity (non-Hispanic white; non-Hispanic black; Hispanic; non-Hispanic Asian; or other non-Hispanic, which included American Indian, Hawaiian, and other Pacific Islander), educational attainment at the time of giving birth (less than high school, some college, or college graduate or beyond), and residential zip code. For data after 2007, we also examined WIC enrollment status, prepregnancy BMI (calculated as weight (kg)/ height (m)2: underweight, <18; normal, 18–24; or overweight/ obese ≥25), and cigarette use 3 months prior to pregnancy (yes or no). 1-hour measurements for carbon monoxide, nitrogen dioxide, and sulfur dioxide, as well as 8-hour averages for ozone. Carbon monoxide and nitrogen dioxide measurement had to be located within a 5-km radius of each mother’s residential ZCTA to be included, while ozone and sulfur dioxide monitors had to be within a 20-km radius. We also considered confounding or effect modification by daily mean particulate matter with an aerodynamic diameter less than or equal to 2.5 μm (PM2.5). Air pollutant monitors had to include at least 90% of possible pregnancies during the study period to be included. For each mother, we calculated apparent temperature exposure for the entire pregnancy, defined as the period between the conception date (estimated as 2 weeks after the LMP date) and the delivery date, as well as during the 3 trimesters of pregnancy, first month, month prior to birth, and 2 weeks prior to birth. First trimester was defined as the period from conception through 13 gestational weeks, the second trimester as gestational weeks 14–26, and the third trimester from the 27th week of gestation to delivery. For each trimester and the entire pregnancy, we used the weekly exposure mean value, calculated as the mean of all of the readings taken during that week, provided there were at least 5 measurements. Monthly averages were combined using weekly data, where at least 3 weeks of data were available. These monthly averages were then used to estimate exposure for the entire pregnancy. Data for at least 75% of the weeks during each trimester and 75% of the months for the entire pregnancy were required in order to be included in the analysis. Monthly, trimester, and entire-pregnancy exposure means for each gaseous pollutant and for PM2.5 were calculated using the same method. However, because PM2.5 was often monitored every third or sixth day, only 1 reading was required to create a weekly exposure. 2308 Basu et al. RESULTS A total of 43,629 full-term LBW infants and 2,032,601 normal-weight infants constituted our final study population. As shown in Table 1, the distributions between the 2 groups were similar according to maternal factors (age, education, and race/ethnicity), region, and season of delivery. However, LBW infants were more likely to be female or younger in gestational age than those who were of normal weight. Table 2 depicts the mean and standard deviation for apparent temperature and criteria air pollutants according to climate zone. Apparent temperatures were generally higher in noncoastal areas than in coastal areas. Pearson correlation coefficients were generally low between apparent temperature and air pollutants (data not shown), with the highest correlation between apparent temperature and ozone (r = 0.32). Generally, annual apparent temperature and LBW were stable throughout the study period, although 2013 had slightly fewer births; December 2013 was the end of the study period, so the number of births tapered off. We used various cutoff values for apparent temperature based on deflection points shown in Figure 1: 55°F (12.8°C) for full gestation, last month, and last 2 weeks of exposure; 60°F (15.6°C) for first-trimester exposure; and 65°F (18.3°C) for first month and third-trimester exposure. Because associations above these cutoff values produced results that were more consistent according to exposure metric, we focused our study on models using apparent temperature values above the cutoff values. Because our results suggested second-trimester exposure was linear, we did not use a cutoff value for apparent temperature for this exposure period. Apparent temperature squared was no longer significant for models stratified by the lower or upper values for cutoff points, and a linear model was now sufficient for all exposure periods considered. Table 3 lists the overall results according to apparent temperature metric above designated cutoff values. The best model fit was found for full gestational exposure. Third-trimester exposure had the greatest association (per 10°F increase in apparent temperature, change of 15.8%, 95% confidence interval 5.0, 27.6), and was largely responsible for the significant associations found during full gestational exposure (13.0% change, 95% confidence interval: 4.1, 22.7). Meanwhile, first month had no association (−0.8% change, 95% confidence interval: −4.9, 3.5), first trimester had a significantly negative association (−7.4% change; 95% confidence interval: −12.5, −2.0), and second trimester, last month, and last 2 weeks Table 1. Summary Characteristics of Term Infants (%) According to Birth Weight, California, 1999–2013 Characteristic Low Birth Weight (n = 43,629) Normal Birth Weight (n = 2,032,601) Gestational week 37 30.7 8.4 38 27.6 19.1 39 20.5 30.0 40 11.7 24.5 41 5.4 11.7 42 2.2 3.6 43 1.2 1.7 44 0.7 0.9 Male 43.3 51.0 Female 56.7 49.0 Infant sex Maternal age, years 11–18 8.1 5.4 19–24 28.0 25.5 25–29 25.1 27.1 30–39 25.0 38.9 3.8 3.2 Less than high school 29.5 26.5 High school 27.3 25.9 ≥40 Maternal educational level Some college 20.8 21.7 College or beyond 22.3 25.8 23.1 30.8 Maternal race/ethnicity Non-Hispanic white Non-Hispanic black 8.4 4.5 Hispanic 49.7 51.7 Non-Hispanic Asian 17.7 12.1 Other non-Hispanic 1.1 1.1 Noncoastal 74.9 71.6 Coastal 25.1 28.4 Cold 48.7 48.0 Warm 51.3 52.0 Region Season of delivery all had slightly positive associations (changes ranging from 1.7% to 2.7%). Figures 2 and 3 illustrate results of possible effect modifiers of the apparent temperature-term LBW association based on full gestational exposure at or above 55°F. Figure 2 depicts variations according to maternal demographics: age, education, and race/ethnicity. As shown, older and more educated mothers generally had greater risk. Although the age group ≥40 years Am J Epidemiol. 2018;187(11):2306–2314 Downloaded from https://academic.oup.com/aje/article-abstract/187/11/2306/5036486 by guest on 09 June 2020 To further analyze effect modification according to maternal factors (age, education, race/ethnicity, prepregnancy BMI, and cigarette use), season of birth, infant sex, and poverty level, we used the model including the apparent temperature exposure metric with the best model fit as determined by Akaike information criterion. Additionally, using the same final model, we adjusted for air pollutants to assess potential confounding by adding each pollutant separately, and considered effect modification by adding an interaction term containing apparent temperature and each pollutant to the final model. Presence of confounding was determined if P < 0.05 for the added variable, while effect modification was based on a significant result (also P < 0.05) for an interaction term. Temperature and Term Birth Weight in California 2309 Table 2. Mean (Standard Deviation) of Exposure Metrics for Full Gestation According to Climate Zone and Region in California, 1999–2013 Climate Region/ Representative City Climate Zone No. Apparent Temperature <55°F ≥55°F PM2.5, μg/m3 Ozone, ppm Nitrogen Dioxide, ppb Carbon Monoxide, ppm 0.39 (0.16) Coastal Eureka 1 48.3 (2.1) San Francisco 3 51.9 (2.0) 56.2 (0.8) Santa Maria 5 52.6 (1.5) 56.3 (1.0) Los Angeles Airport 6 54.0 (0.8) 59.4 (2.5) 10.1 (1.3) 0.027 (0.003) 14.7 (6.7) San Diego 7 54.3 (0.6) 61.0 (3.0) 11.7 (1.8) 0.032 (0.004) 15.2 (3.6) Napa 2 50.9 (2.5) 56.5 (1.0) 0.020 (0.003) 11.0 (2.3) San Jose 4 53.0 (1.4) 58.4 (1.8) 0.025 (0.006) Long Beach 8 54.4 (0.4) 62.2 (3.4) 13.4 (3.2) 0.024 (0.005) 0.69 (0.26) Los Angeles Civic Center 9 54.6 (0.3) 62.0 (3.5) 13.8 (3.0) 0.028 (0.006) 0.73 (0.33) 0.72 (0.29) 0.022 (0.004) 0.032 (0.003) 0.45 (0.12) Riverside 10 53.9 (0.9) 61.4 (3.6) 16.8 (4.5) 0.030 (0.006) Red Bluff 11 52.5 (1.6) 61.7 (3.6) 7.9 (3.8) 0.030 (0.007) Stockton 12 52.6 (1.6) 60.3 (3.1) 10.2 (2.8) 0.027 (0.006) 16.4 (4.2) 0.41 (0.18) 16.9 (4.1) 0.037 (0.007) 18.1 (3.7) 0.43 (0.11) 0.041 (0.007) 21.0 (3.9) 0.19 (0.07) 0.037 (0.009) 12.5 (4.1) 0.29 (0.13) Fresno 13 53.1 (1.3) 62.6 (4.3) Barstow 14 51.9 (1.9) 61.5 (3.7) Brawley 15 71.0 (6.0) Bishop 16 47.6 (5.3) 58.9 (2.3) 8.2 (1.1) 0.032 (0.003) Abbreviation: PM2.5, particulate matter with an aerodynamic diameter less than or equal to 2.5 μm. had a large negative association, this finding was not significant because of a relatively small number of births in this oldest age category. Black mothers, followed by white mothers, had infants at the greatest risks, while Hispanic and Asian mothers also had increased positive associations, although the latter 2 findings were not statistically significant. Figure 2 also includes results for other modifiers (season of birth and infant sex). Warm season of birth had a greater association than cold season, and this difference was more pronounced during the third trimester (data not shown). Male infants had slightly greater risk than did female infants. Figure 3 shows the results for variables available from 2007 through 2013. Cigarette use did not make a difference in our study, but only 3% of women smoked 3 months prior to pregnancy. Not being enrolled in WIC was associated with a greater risk, as was being underweight. Percentage in poverty was a confounder but not an effect modifier. No criteria air pollutant examined confounded the apparent temperature and term LBW association (Figure 4), but ozone was an effect modifier. DISCUSSION We found that long-term apparent temperature exposure, specifically full gestational exposure driven largely by the third trimester, was positively associated with term LBW in California. Recent studies evaluating other outcomes, such as mortality and hospital or emergency visits, have found that acute temperature exposure resulted in adverse health outcomes (14, 16–18). Am J Epidemiol. 2018;187(11):2306–2314 Previous birth outcome studies focusing on preterm delivery and stillbirth in California have also suggested that the week prior to birth represented a critical window of exposure (19, 20). Because air pollutants were found to be associated with adverse birth outcomes from longer-term exposures in previous studies (21–23), we ensured that they were not confounders or effect modifiers here. Few previous studies have addressed temperature or heat index and LBW, as summarized in a recent review in which investigators reported slight decreases or no associations between ambient temperature and LBW (3). However, several of these studies reported associations during the second trimester, where we found only a slight positive association relative to thirdtrimester or full-gestation exposures. The previous studies also did not limit the exposures to warmer temperatures and have other methodological differences. In one prior study, investigators examined term LBW specifically and its relationship with apparent temperature in the United States, and they observed similar results for increases during the third trimester and full gestational exposure (24). Another recent study conducted in Massachusetts also found that greater third-trimester exposure was associated with decreased birth weight and that higher fullgestational exposures increased the association with LBW (25). However, this latter study did not exclude preterm LBW, so there may have been other factors related to preterm delivery that were not adjusted for in the analysis. Results varied according to maternal demographics and other factors. Black mothers were at greatest risk from higher Downloaded from https://academic.oup.com/aje/article-abstract/187/11/2306/5036486 by guest on 09 June 2020 Noncoastal 2310 Basu et al. A) B) –3.2 –3.4 –3.4 Log(Odds) Log(Odds) –3.3 –3.6 –3.8 –3.5 –3.6 –3.7 –3.8 20 40 60 80 100 20 –3.2 Log(Odds) Log(Odds) 100 –3.0 –3.7 –3.8 –3.4 –3.6 –3.8 –3.9 20 40 60 80 100 20 40 60 80 100 Apparent Temperature, °F Apparent Temperature,°F F) –3.0 –3.0 –3.2 –3.2 Log(Odds) Log(Odds) 80 D) –3.5 –3.6 E) 60 –3.4 –3.6 –3.8 –3.4 –3.6 –3.8 20 40 60 80 20 100 Apparent Temperature, °F G) 40 60 80 100 Apparent Temperature, °F –3.0 Log(Odds) –3.2 –3.4 –3.6 –3.8 30 40 50 60 70 80 90 Apparent Temperature, °F Figure 1. Apparent temperature metrics and log(odds) of term low birth weight according to exposure period, in California, 1999–2013. A) First month of pregnancy; B) first trimester; C) second trimester; D) third trimester; E) 1 month prior to birth; F) 2 weeks prior to birth; G) full gestation. All models included apparent temperature and apparent temperature squared (with one exception: no apparent temperature squared term for second trimester) assuming specific baseline values. Covariables included maternal race/ethnicity, maternal age group, maternal education, infant sex, and year and season of birth. Am J Epidemiol. 2018;187(11):2306–2314 Downloaded from https://academic.oup.com/aje/article-abstract/187/11/2306/5036486 by guest on 09 June 2020 C) 40 Apparent Temperature, °F Apparent Temperature, °F Temperature and Term Birth Weight in California 2311 Table 3. Estimated Percent Change in Odds of Term Low Birth Weight per 10°F Increase in Apparent Temperature According to Exposure Period, Californiaa, 1999–2013 Apparent Temperature Cutoff Value, °F No. of Births Above Cutoff Value % Change Above Cutoff Value Full gestation 55 1,553,780 13.0 First month 65 564,493 −0.8 −4.9, 3.5 First trimester 60 831,621 −7.4 −12.5, −2.0 Exposure Period N/Ab Second trimester 95% CI 4.1, 22.7 1.7 0.3, 3.2 65 611,413 15.8 5.0, 27.6 Last month 55 1,241,616 2.7 −0.1, 5.5 Last 2 weeks 55 1,245,687 2.6 0.2, 5.0 Abbreviations: CI, confidence interval; N/A, not applicable. Climate zone 1 was excluded in all analyses; climate zones 1, 2, 3, and 5 were excluded in third-trimester analyses because of insufficient sample sizes. b A cutoff value was not needed for second-trimester apparent temperature because our results suggested that this exposure period was linear. a mothers were mostly white or Asian, which could explain why white mothers are the group at second-highest risk for term LBW. These findings support strong connections among socioeconomic status, race/ethnicity, and educational attainment, as well as maternal age. Finally, because maternal cigarette use based on self-report was so low in our study population, it could not be evaluated as a modifying risk factor for term LBW and did not serve as a valid marker for socioeconomic status. We were able to consider neighborhood poverty level, but we found that it also did not modify our results. Because LBW generally has multiple etiologies, a clear biological mechanism or cause is unknown. The critical window of exposure appears to be the third trimester, particularly for births occurring during the warm season. With heat exposure, decreased uterine blood flow and increased pituitary secretion of antidiuretic hormone and oxytocin may induce labor (29). 50 50 40 40 40 30 30 30 30 30 20 10 0 –10 –20 20 10 0 –10 –20 20 10 0 –10 –20 % Change in Odds 50 40 % Change in Odds 50 40 % Change in Odds 50 % Change in Odds % Change in Odds temperatures, while Hispanic mothers fared the best, despite the 2 groups having similar socioeconomic status, age distributions, and educational attainment at the time of giving birth. This difference was most likely due to Hispanic mothers having strong family support and networking, access to state or federally funded programs, and better nutrition during pregnancy. In California, 75% of WIC enrollees are Hispanic while only 5% are black (26). Mothers who do not qualify for the WIC program, but who are poor, may not be able to afford maternity care. Higher maternal education resulted in greater risks for term LBW following high apparent temperature exposure. Highly educated mothers tend to bear children at older ages, but Hispanic mothers tend to be younger, and while there have been several studies linking younger age with adverse birth outcomes, older maternal age (ages ≥45 years) has also been associated with adverse birth outcomes (27) and greater racial disparities (28). Older and more highly educated 20 10 0 –10 –20 20 10 0 –10 –20 –30 –30 –30 –30 –30 –40 –40 –40 –40 –40 m ar W l oo C al m Fe e al e Educational Level M e or M or ge ge e le ol oll C C l e m hoo So Sc S H h ig an H Th ss Age Group, years Le 0 ≥4 9 –3 30 9 –2 25 4 –2 19 8 –1 11 te hi n ia As nic pa is H k ac Bl W Race/Ethnicity Infant Sex Birth Season Figure 2. Estimated percent change in odds of term low birth weight per 10°F increase in full-term apparent temperature above 55°F, stratified by maternal demographics, infant sex, and birth season, in California, 1999–2013. All models adjusted for maternal race/ethnicity, maternal age group, maternal education, infant sex, and year and season of birth. HS: high school. Am J Epidemiol. 2018;187(11):2306–2314 Downloaded from https://academic.oup.com/aje/article-abstract/187/11/2306/5036486 by guest on 09 June 2020 1,980,970 Third trimester 70 70 60 60 60 60 50 50 50 50 40 30 20 10 40 30 20 10 % Change in Odds 70 % Change in Odds 70 % Change in Odds % Change in Odds 2312 Basu et al. 40 30 20 10 40 30 20 10 0 0 0 –10 –10 –10 –10 –20 –20 –20 –20 s Ye o BMI N o s Ye N t gh ei rw ve O al m or N t gh ei w er nd U After 2007 WIC Cigarette Use Figure 3. Estimated percent change in odds of term low birth weight per 10°F increase in full-term apparent temperature above 55°F, stratified by body mass index (BMI), enrollment in the Special Supplemental Nutrition Program for Women, Infant, and Children (WIC), and cigarette use 3 months prior to pregnancy, in California, 2007–2013. All models adjusted for maternal race/ethnicity, maternal age group, maternal education, infant sex, and year and season of birth. There are several limitations to this study. We relied on ambient monitoring data rather than on individual monitoring to assign exposures. By limiting the study to mothers whose residential ZCTAs were within 10 km of a meteorological monitor, we attempted to reduce exposure misclassification, but there could have been some residual misclassification. We assigned full gestational exposure based on the residential ZCTA at the time of giving birth because we did not have information on previous addresses. A previous study reported that although 15%–30% of pregnant women reportedly changed residences during their pregnancy, the impact on health associations were relatively low (34). Thus, using the ZCTA of the residential address at the time of giving birth should have been sufficient 50 50 40 40 40 30 30 30 30 20 10 0 –10 –20 20 10 0 –10 % Change in Odds 50 40 % Change in Odds 50 % Change in Odds % Change in Odds There is also evidence that pregnant women may not be able to thermoregulate efficiently. When body temperatures rise, the body generally shifts blood flow from the vital organs to the skin’s surface in an effort to cool down (30). Thermoregulation may be inadequate when too much blood is diverted from the vital organs of the mother and the developing fetus, depriving the fetus of adequate nutrition (30). A sudden rise in temperature or extreme heat also causes heat stress (31). Increased prenatal inflammation may be another factor contributing to LBW or other adverse birth outcomes (32). Increased blood viscosity, elevated cholesterol levels associated with higher temperatures, and a higher sweating threshold have also been reported in susceptible subgroups (33). 20 10 0 –10 20 10 0 –10 –20 –20 –20 –30 –30 –30 –30 –40 –40 –40 With Without PM2.5 With Without Ozone –40 With Without NO2 With Without CO Figure 4. Estimated percent change in odds of term low birth weight per 10°F increase in full-term apparent temperature above 55°F, with and without each specified air pollutant, in California, 1999–2013. All models adjusted for maternal race/ethnicity, maternal age group, maternal education, infant sex, and year and season of birth. CO, carbon monoxide; NO2, nitrogen dioxide; PM2.5, particulate matter with an aerodynamic diameter less than or equal to 2.5 μm. Am J Epidemiol. 2018;187(11):2306–2314 Downloaded from https://academic.oup.com/aje/article-abstract/187/11/2306/5036486 by guest on 09 June 2020 0 Temperature and Term Birth Weight in California 2313 ACKNOWLEDGMENTS Author affiliations: Air and Climate Epidemiology Section, Office of Environmental Health Hazard Assessment, California Environmental Protection Agency, Oakland, California (Rupa Basu, Dharshani Pearson, Brian Malig); and Department of Environmental Health Sciences, School of Public Health, University of California, Berkeley, Berkeley, California (Rupa Basu, Reina Rau). The opinions expressed in this article are those of the authors, and do not represent those of the California Environmental Protection Agency or the Office of Environmental Health Hazard Assessment. Conflict of interest: none declared. REFERENCES 1. Basu R. High ambient temperature and mortality: a review of epidemiologic studies from 2001 to 2008. Environ Health. 2009;8:40. 2. Turner LR, Barnett AG, Connell D, et al. Ambient temperature and cardiorespiratory morbidity: a systematic review and metaanalysis. Epidemiology. 2012;23(4):594–606. Am J Epidemiol. 2018;187(11):2306–2314 3. Zhang Y, Yu C, Wang L. Temperature exposure during pregnancy and birth outcomes: an updated systematic review of epidemiological evidence. Environ Pollut. 2017; 225: 700–712. 4. Martin JA, Hamilton BE, Ventura SJ, et al. Births: final data for 2011. Natl Vital Stat Rep. 2013;62(1):1–69, 72. 5. MacDorman MF, Mathews TJ. Understanding racial and ethnic disparities in US infant mortality rates. NCHS Data Brief. 2011; September (74):1–8. 6. Hamilton BE, Martin JA, Osterman MJ, et al. Births: final data for 2014. Natl Vital Stat Rep. 2015;64(12):1–64. 7. Valero de Bernabé J, Soriano T, Albaladejo R, et al. Risk factors for low birth weight: a review. Eur J Obstet Gynecol Reprod Biol. 2004;116(1):3–15. 8. Li DK, Janevic T, Odouli R, et al. Hot tub use during pregnancy and the risk of miscarriage. Am J Epidemiol. 2003; 158(10):931–937. 9. Hauck FR, Tanabe KO, Moon RY. Racial and ethnic disparities in infant mortality. Semin Perinatol. 2011;35(4): 209–220. 10. Caudri D, Wijga A, Gehring U, et al. Respiratory symptoms in the first 7 years of life and birth weight at term: the PIAMA Birth Cohort. Am J Respir Crit Care Med. 2007;175(10): 1078–1085. 11. Wiles NJ, Peters TJ, Leon DA, et al. Birth weight and psychological distress at age 45–51 years: results from the Aberdeen Children of the 1950s cohort study. Br J Psychiatry. 2005;187:21–28. 12. Li X, Huang S, Jiao A, et al. Association between ambient fine particulate matter and preterm birth or term low birth weight: an updated systematic review and meta-analysis. Environ Pollut. 2017;227:596–605. 13. California Department of Health Services. Birth Statistical Master Files, 1999–2013. Sacramento, CA: Office of Vital Statistics; 2013. 14. Basu R, Feng WY, Ostro BD. Characterizing temperature and mortality in nine California counties. Epidemiology. 2008; 19(1):138–145. 15. California Energy Commission. 2008 Building Energy Efficiency Standards for Residential and Nonresidential Buildings. Sacramento, CA: California Energy Commission; 2008. http://www.energy.ca.gov/2008publications/CEC-4002008-001/CEC-400-2008-001-CMF.PDF. 16. Green RS, Basu R, Malig B, et al. The effect of temperature on hospital admissions in nine California counties. Int J Public Health. 2010;55(2):113–121. 17. Basu R, Malig B. High ambient temperature and mortality in California: exploring the roles of age, disease, and mortality displacement. Environ Res. 2011;111(8): 1286–1292. 18. Basu R, Pearson D, Malig B, et al. The effect of high ambient temperature on emergency room visits. Epidemiology. 2012; 23(6):813–820. 19. Basu R, Malig B, Ostro B. High ambient temperature and the risk of preterm delivery. Am J Epidemiol. 2010;172(10): 1108–1117. 20. Basu R, Sarovar V, Malig BJ. Association between high ambient temperature and risk of stillbirth in California. Am J Epidemiol. 2016; 183(10):894–901. 21. Green R, Sarovar V, Malig B, et al. Association of stillbirth with ambient air pollution in a California cohort study. Am J Epidemiol. 2015;181(11):874–882. 22. Basu R, Pearson D, Ebisu K, et al. Association between PM2.5 and PM2.5 constituents and preterm delivery in California, 2000–2006. Paediatr Perinat Epidemiol. 2017; 31(5):424–434. Downloaded from https://academic.oup.com/aje/article-abstract/187/11/2306/5036486 by guest on 09 June 2020 to characterize exposure for most women. Although we had data on several maternal factors, we did not have information on multiple pregnancies for the same mother so we could not consider repeated outcomes. We also did not have data on occupational addresses, air conditioning use, or daily time-activity patterns, so we could not account for them. Maternal smoking data were based on self-report, and because it is not socially desirable to smoke, mothers may have inaccurately reported less or no exposure. Along with prepregnancy BMI and WIC enrollment, maternal cigarette smoking data were available only for birth records beginning in 2007 (7 years in our study), making the estimates less reliable. Adjustments for multiple hypothesis testing were not conducted, because corrections would likely widen our confidence intervals, but the effect estimates would remain similar. The results that were produced were consistent according to each exposure metric, and thus, very unlikely due to chance. This study adds to the growing body of literature that suggests that pregnant women and their fetuses are vulnerable populations following ambient heat exposure. During heat advisory warnings, pregnant women should be included as a vulnerable subgroup for extra precautions, particularly specific subgroups found to be at greatest risk. The results from this study will not only be informative to pregnant women and their health-care professionals, but also other caretakers, family members, and employers. The Intergovernmental Panel on Climate Change has stated that hot days, elevated nighttime temperatures, and heat waves have become more frequent in recent years, and their duration and intensity are likely to increase in the future (35); thus, adverse health outcomes from heat exposure are more likely as well. The present study focused on apparent temperature, although the association during a heat wave is likely to be substantially greater. Given the significant associations for apparent temperature and LBW found in this study, more large-scale studies of temperature and LBW, including modifiers of the association, in other locations are warranted. 2314 Basu et al. 29. Stan CM, Boulvain M, Hirsbrunner-Amagbaly P, et al. Hydration for treatment of preterm labour. Cochrane Database Syst Rev. 2002;(2):CD003096. 30. Bouchama A, Knochel JP. Heat stroke. N Engl J Med. 2002; 346(25):1978–1988. 31. Carolan-Olah M, Frankowska D. High environmental temperature and preterm birth: a review of the evidence. Midwifery. 2014;30(1):50–59. 32. Bastek JA, Weber AL, McShea MA, et al. Prenatal inflammation is associated with adverse neonatal outcomes. Am J Obstet Gynecol. 2014;210(5): 450.e1–450.e10. 33. Astrand P-O, Rodahl K, Dahl HA, et al. Textbook of Work Physiology: Physiological Bases of Exercise. 4th ed. Champaign, IL: Human Kinetics; 2003. 34. Bell ML, Belanger K. Review of research on residential mobility during pregnancy: consequences for assessment of prenatal environmental exposures. J Expo Sci Environ Epidemiol. 2012;22(5):429–438. 35. Field CB, Barros VR, Dokken DJ, et al. eds. Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part A: Global and Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, UK: Cambridge University Press; 2014 Am J Epidemiol. 2018;187(11):2306–2314 Downloaded from https://academic.oup.com/aje/article-abstract/187/11/2306/5036486 by guest on 09 June 2020 23. Ng C, Malig B, Hasheminassab S, et al. Source apportionment of fine particulate matter and risk of term low birth weight in California: exploring modification by region and maternal characteristics. Sci Total Environ. 2017;605–606:647–654. 24. Ha S, Zhu Y, Liu D, et al. Ambient temperature and air quality in relation to small for gestational age and term low birthweight. Environ Res. 2017;155:394–400. 25. Kloog I, Melly SJ, Coull BA, et al. Using satellite-based spatiotemporal resolved air temperature exposure to study the association between ambient air temperature and birth outcomes in Massachusetts. Environ Health Perspect. 2015; 123(10):1053–1058. 26. United States Department of Agriculture Food and Nutrition Service. Special supplemental Nutrition Program—WIC: Table III: composition of ethnic participant enrollment. 2012. https://fns-prod.azureedge.net/sites/default/files/wic/TABLEIII-Composition-of-Ethnic-Participant-Enrollment-2012.pdf. Accessed July 16, 2018. 27. Carolan M. Maternal age ≥45 years and maternal and perinatal outcomes: a review of the evidence. Midwifery. 2013;29(5): 479–489. 28. Buescher PA, Mittal M. Racial disparities in birth outcomes increase with maternal age: recent data from North Carolina. N C Med J. 2006;67(1):16–20.