Printer Friendly

Birth weight, ethnicity, and exposure to trihalomethanes and haloacetic acids in drinking water during pregnancy in the born in Bradford cohort.

Introduction

During drinking-water treatment, disinfectants such as chlorine react with natural organic matter and bromide ions present in the water, producing disinfection by-products (DBPs) (IPCS 2000). There are > 600 different species of DBPs including trihalomethanes (THMs), haloacetic acids (HAAs), haloacetonitriles, halofuranones, and nitrosamines (Richardson et al. 2007). THMs [comprising chloroform, bromodichloromethane (BDCM), dibromochloromethane (DBCM), and bromoform] are the most prevalent class in drinking waters, followed by HAAs (comprising bromo-, chloro-, dibromo-, dichloro-, tribromo-, trichloro-, bromochloro-, bromodichloro-, and dibromochloro-acetic acids) (Krasner et al. 1989). THMs are regulated at 100 [micro]g/L per sample in the United Kingdom (DWI 2010), whereas HAAs are currently unregulated.

Exposure to volatile THMs via showering and bathing plays an influential role in determining total blood dose via dermal exposure and inhalation, whereas ingestion plays a relatively minor role (Backer et al. 2000; Nuckols et al. 2005). Inhalation and dermal routes are expected to contribute little to uptake of nonvolatile HAAs (Xu et al. 2002; Xu and Weisel 2003), making ingestion the dominant route for HAA exposure.

The relationship between DBPs and fetal growth outcomes has been investigated in > 30 epidemiological studies since the early 1990s, but the evidence for an association with THMs and also with HAAs is inconsistent. However, some studies suggest that brominated species may be most relevant (Costet et al. 2012; Rivera-Nunez and Wright 2013).

Epidemiological studies have typically examined THMs and HAAs in isolation, but these are not necessarily good proxies for each other, nor for other DBP classes. In reality, DBPs exist as complex mixtures in drinking-water supplies, which may be geographically specific, thus contributing to inconsistent findings in this research field. One recent epidemiological study has examined THM-HAA mixtures using area-level exposure data (Rivera-Nunez and Wright 2013); however, to our knowledge, there are no studies to date incorporating individual water use into DBP exposure assessment that have addressed THM-HAA mixture effects.

In this study we aimed to investigate, for the first time, potential effects of joint THM-HAA exposures, incorporating individual estimates of uptake, on birth weight and ethnic differences in the Born in Bradford (BiB) prospective birth cohort.

Methods

Study population. Born in Bradford (BiB) is a longitudinal multi-ethnic birth cohort study aiming to examine the impact of environmental, psychological, and genetic factors on maternal and child health and well-being (Wright et al. 2013). Bradford is a city in the north of England with high levels of socioeconomic deprivation and ethnic diversity. Approximately half of the births in the city are to mothers of South Asian origin. Women were recruited while waiting for their glucose tolerance test, a routine procedure offered to all pregnant women registered at the Bradford Royal Infirmary, at 26-28 weeks gestation. The only exclusion criterion was if a woman planned to move away from Bradford before the birth (Raynor and Born in Bradford Collaborative Group 2008). For those providing informed consent to participate in BiB, a baseline questionnaire was completed via an interview with a study administrator. The baseline questionnaire was transliterated into Urdu and Mirpuri, and questionnaires in Mirpuri were administered by trained bilingual interviewers because it has no written form (Raynor and Born in Bradford Collaborative Group 2008). Information on birth weight, gestational age, and health during pregnancy was obtained from clinical records.

The full BiB cohort recruited 12,453 women during 13,776 pregnancies between 2007 and 2010. The BiB cohort represents 64% of all pregnancies at Bradford Royal Infirmary and is broadly characteristic of the city's maternal population (Wright et al. 2013). Ethical approval for the data collection was granted by Bradford Research Ethics Committee (Ref 07/H1302/112).

Our initial data set contained 13,525 babies for whom clinical record data were available. We restricted the data set to 13,199 singleton births. For mothers with repeated pregnancies in BiB, we randomly selected one of these pregnancies (excluding 1,271 babies). We excluded 2,100 due to missing water use data, 98 for whom THM exposure could not be calculated for all trimesters, 1 record with missing birth weight, and 531 preterm births. This left 9,198 babies eligible for the analysis. After taking into account missing covariate data, final models included up to 7,438 babies.

Exposure assessment. Water use. The baseline questionnaire ascertained typical daily consumption of tap water, bottled water, tea, coffee, and squash (concentrated diluting juice, including any other drinks made up with tap water) at home, work/ study, or elsewhere; water filtering at home and work; and frequency and duration of showering, bathing, and swimming. We calculated daily consumption (liters) of cold tap water (sum of tap water and squash), hot beverages made from tap water (sum of tea and coffee), and total tap water (sum of tap water, squash, tea, coffee) and minutes spent showering, bathing, and swimming per week.

THM and HAA concentrations in drinking water. Routine monitoring data on THMs and other parameters were provided by Yorkshire Water, for the eight water supply zones (WSZ) covering the study area from January 2006 to March 2011. Each WSZ was sampled nine times per year on average, giving 374 data points in total. Additional samples for HAA analysis were taken quarterly from these eight WSZs, by Yorkshire Water staff alongside their routine sampling regime, between June 2007 and November 2010. The additional samples were picked up by study staff, and analyzed for HAAs. HAAs were analyzed using a modified form of U.S. Environmental Protection Agency (EPA) Method 552.3 (Tung et al. 2006; U.S. EPA 2003). The derivatized HAAs (methyl esters) were measured using gas chromatography with microelectron capture detection (Agilent 6890; Santa Clara, CA, USA).

Predictive modeling of both THMs and HAAs was undertaken to obtain estimates of DBP concentrations for times and places for which data were sparse, in line with previous studies (Nieuwenhuijsen et al. 2008; Toledano et al. 2005). Log-transformed THMs were modeled using linear regression, with a spline in month, a factor for year, and a factor for WSZ in order to provide monthly WSZ-specific concentrations. THM samples below the limit of detection (LOD) were assigned a value equal to half the LOD, per Malliarou et al. (2005). Bromoform was not modeled individually because so many data points were below the LOD; instead total brominated THMs (THMBr) were modeled.

Only three HAAs had sufficient detectable data points to be modeled: 158 points each for dichloroacetic acid (DCAA) and for trichloroacetic acid (TCAA) and 143 data points for bromodichloroacetic acid (BDCAA). Model selection for each HAA was first performed in a frequentist framework in R 2.15.2 (R Core Team 2012). Square root-transformed DCAA and TCAA and log-transformed BDCAA were modeled, with all models including a factor for WSZ and a spline on time. Models selected additionally included the following variables: conductivity (all HAA models), temperature (DCAA and TCAA models only), TOC (TCAA model only), and total chlorine (BDCAA model only). To account for parameter uncertainty and impute missing covariate data, the best model for each HAA was then run in a Bayesian framework in WinBUGS 1.4.3 (Lunn et al. 2000) to predict mean DCAA, TCAA, and BDCAA concentration levels by WSZ and year-quarter. All HAA data points for quarter 2 of 2009 were anomalously low compared with data for all other sampling quarters, suggesting possible unreliability of the laboratory analysis data for this quarter. Therefore, data from quarter 2 of 2009 were excluded, and the Bayesian model was allowed to predict HAA values for this quarter. In addition, the Bayesian model was allowed to predict HAA values for the first quarters of 2007 and 2011 (i.e., extrapolating to one additional quarter either side of the HAA sampling window), to cover as many women's pregnancies as possible, but without overstretching the model.

Time-weighted average DBP concentrations. Each woman was assigned modeled THM and HAA concentrations (micrograms per liter) for the WSZ encompassing her residence postcode at time of recruitment, and additionally for her workplace postcode if applicable and possible, within a GIS (geographic information system) via WSZ-postcode link. Workplace exposure could not be assigned to employed women if workplace address/postcode data was missing or insufficient to allow geocoding (n = 1,157) or if a woman's workplace was outside the WSZs included in the exposure assessment (n = 514). Time-weighted average concentrations were calculated for each woman. The first trimester was defined as days 1-93, the second trimester as days 94-186, and the third as day 187 to the day preceding delivery. The time weighting was based on the proportion of the whole pregnancy or trimester falling into each month (THMs) or quarter (HAAs). Each mother was thus assigned concentrations (micrograms per liter) for total trihalomethane (TTHM), chloroform, BDCM, DBCM, THMBr, DCAA, TCAA, and BDCAA for each trimester/whole pregnancy. TTHM, DCAA, TCAA, and BDCAA were summed to give DBP7 concentration, a joint THMHAA exposure metric for use in regression models. For 15% of women, time-weighted average HAA concentrations could not be calculated for their pregnancy because the pregnancy commenced before January 2007 (the earliest extent of the HAA models). For those women assigned workplace concentrations (n = 2,353; 58.4% of 4,024 employed women), weightings of 40/112 and 72/112 were applied to work and residence, respectively, based on 8 hr at work 5 days per week, out of a total of 112 waking hours per week.

Integrated uptake of THMs in blood via multiple activities. Time-weighted average THM concentrations (micrograms per liter) were multiplied by individual water use and THM uptake factors [micrograms/(micrograms per liter)liter for consumption or micrograms/(micrograms per liter) min for showering/bathing] to estimate whole pregnancy and trimester-specific average THM uptake into blood (micrograms/day) for each activity (water consumption, showering, bathing). THM uptake factors were calculated from biomonitoring studies (Aggazzotti et al. 1995; Backer et al. 2000; Lynberg et al. 2001) or based on uptake factors previously used in the literature (Villanueva et al. 2007). Further details are given in Supplemental Material, "THM uptake factors," and uptake factor values are shown in Supplemental Material, Table S1. Uptake calculations for swimming used swimming pool THM concentrations (Chu and Nieuwenhuijsen 2002). Uptakes for each activity were summed to give the exposure metric of integrated uptake--total blood dose, for chloroform, BDCM, DBCM, THMBr, and TTHM. We also calculated integrated TTHM uptake via showering, bathing, and swimming as a separate exposure metric so that this could be included in a model with DBP7 (sum of TTHM and HAA3) ingestion via water consumption.

Ingestion of THMs and HAAs via drinking water consumption. Time-weighted average TTHM and HAA concentrations (micrograms per liter) were multiplied by water consumption (liters/day) to calculate the exposure metrics of whole pregnancy and trimester-specific ingestion via water consumption (micrograms/day) of TTHM, DCAA, TCAA, BDCAA, HAA3 (sum of DCAA, TCAA, BDCAA), and DBP7, the latter being an individual-level metric of joint THM-HAA exposure.

For both integrated THM uptakes and THM-HAA ingestion metrics, boiling adjustment factors were applied to beverages made with hot water (THMs: -92%; DCAA +43.5%; TCAA -36.9%; BDCAA -56.5%) and filtering reduction factors applied to cold filtered tap water (THMs -90%; DCAA -61.8%; TCAA -67.4%; and BDCAA -78.5%) (Edwards 2014; Smith 2011). All other drinks (including squash) were treated as unfiltered.

Statistical analysis. We analyzed continuous birth weight because this had greatest statistical power compared with term low birth weight (LBW) and small for gestational age (SGA) and because the latter outcomes do not necessarily identify pathologically small infants comparably for white British and Pakistani-origin subpopulations because of different underlying birth weight distributions (Moser et al. 2008). Multiple linear regressions examining the relationship between continuous birth weight and categorical exposure metrics (using tertiles of THM or HAA exposure metrics and using individual water use activities were categorized as follows: cold tap water consumption: 0.0-0.4 L/day, 0.6-1.0 L/day, 1.2-1.4 L/day, [greater than or equal to] 1.6L/day; showering: 0 min/week, 1-60 min/week, 61-105 min/week, > 105 min/week; bathing: 0 min/week, 1-60 min/week, 61-120 min/ week, > 120 min/week; swimming: yes vs. no) were run in STATA 12.1 (StataCorp, College Station, TX, USA).

To examine joint THM-HAA exposure, we included two exposure terms [DBP7 ingestion via water consumption and integrated TTHM uptake via showering, bathing, and swimming (both as tertiles)] in the model. We chose to examine joint exposure in this way because, to our knowledge, there is no information available from which HAA uptake factors can be calculated, and thus integrated DBP7 uptake across all activities could not be estimated. To incorporate THM and HAA exposures on a comparable scale within models, a joint THM-HAA exposure term was estimated as DBP7 ingestion via water consumption (because ingestion is the dominant route for HAA exposure), with the remainder of exposure included as a separate term for TTHM uptake via showering/ bathing/swimming (because HAA exposure via these activities is negligible).

We also analyzed the relationship between time-weighted area-level DBP7 exposure (as tertiles) and continuous birth weight, to investigate whether any relationship could be observed between birth weight and DBP levels in tap water, that is, without taking into account individual-level water consumption/ activities. This model was included to aid interpretation of models examining individuallevel DBP exposures, to help rule out potential confounding by unmeasured factors related to water consumption or water-related activities.

All models were adjusted a priori for 10 maternal factors [caffeine intake ([greater than or equal to] 200 mg/day vs. < 200 mg/day), education (highest educational qualification categorized as no qualifications, school, further education, higher education, other), fasting and postload glucose from oral glucose tolerance test (continuous), ethnicity (categorized as white British, Pakistani origin, other), smoking (categorized as never smoker, ever smoker, current smoker), parity (categorized as 0, 1, [greater than or equal to] 2), age (continuous), body mass index (BMI) at recruitment as quartiles (because prepregnancy BMI was not available), Index of Multiple Deprivation (IMD) 2010 quintiles of deprivation (McLennan et al. 2011)] and two infant factors (gestational age at delivery as linear and quadratic terms, sex).

Spearman's correlations between exposure metrics were calculated.

Models were run for total population and subgroup analyses stratifying by the two largest ethnic groups white British and Pakistaniorigin. The ethnic category "other" comprised a number of ethnicities and was not considered meaningful to examine because of its small size and lack of cultural cohesiveness.

We consideredp < 0.05 statistically significant. The F-test was used to a) test whether the categorical exposure term, as a whole, was significant within regression models (giving a p-value for overall significance), and b) to test whether exposure-ethnicity interaction terms were, as a whole, significant within regression models (giving a p-value for interaction). p-Values for a linear trend across exposure categories were derived by including the exposure tertiles (coded as 0, 1, 2) as a continuous variable in the model.

Because HAA ingestion is driven by water consumption (which may change during pregnancy) and may be sensitive to filtering assumptions, we conducted a sensitivity analysis for trimester 2 only (because water consumption data were collected at 26-28 weeks, and may be more reflective of behavior during the second trimester), applying stricter filtering criteria to calculation of HAA ingestion. Sensitivity analyses additionally adjusting models for environmental tobacco smoke and language of questionnaire completion were conducted because these variables showed the greatest difference between Pakistani-origin women in the highest THM uptake tertile versus those in the lower tertiles. Overall pattern and direction of relationship, and magnitude of any change in point estimates, were examined to determine if conclusions were robust or sensitive to changes in the model.

Results

White British (40%) and Pakistani-origin (44%) women are the two predominant ethnic groups in the population, and there are differences in birth weight outcomes, employment status, deprivation, parity, diabetes, smoking, and alcohol and caffeine consumption between the two groups (Table 1). Compared with white British women, women of Pakistani origin drink less water from all sources combined, spend less time bathing but more time showering, and very few go swimming (2% vs. 14% white British). There were few or no missing data on maternal age, employment, education, IMD quintile, or smoking, but approximately 3-4% of the population were missing data on parity, BMI, fasting and post-load glucose levels, and 8.9% were missing data on caffeine consumption, which reduced numbers in final regression models. Time-weighted average DBP concentrations in tap water to which mothers were exposed differed by ethnicity, but only marginally and not in any consistent direction (Table 2). However, estimates of THM uptakes and HAA ingestion are lower on average for Pakistani-origin women compared with white British women, reflecting differences in water use between the two groups. Individual-level exposure metrics within the THM class and within the HAA class were highly correlated (see Supplemental Material, Table S2).

Table 3 presents adjusted model estimates for mean difference in birth weight associated with integrated TTHM and THMBr uptake tertiles for whole pregnancy, and trimesterspecific results are shown in Supplemental Material, Table S3. Compared with bivariate models, adjusted coefficients changed considerably, indicating that the relationship between THM uptake and birth weight is confounded. For white British infants, there is no evidence of an association between birth weight and THM uptake. For Pakistaniorigin infants, there are statistically significant reductions in mean birth weight of approximately 50 g for the highest exposure tertile compared with the lowest tertile [-53.7 g; 95% confidence interval (CI): -89.9, -17.5 for [greater than or equal to] 1.82 vs. < 1.05 [micro]g/day of total THM; and -56.4 g; 95% CI: -93.1, -19.6 for [greater than or equal to] 0.26 vs. < 0.14 [micro]g/day of THMBr] and significant trends (p < 0.01) across increasing exposure tertiles. Findings were similar for individual trimesters (see Supplemental Material, Table S3). p-Values for interaction by ethnicity were significant for TTHM uptake in whole pregnancy, and all THM uptakes in trimesters 2 and 3. Results for chloroform and BDCM uptake were similar (see Supplemental Material, Table S4).

We find no evidence of association between birth weight and HAA ingestion (via drinking water consumption) for either the total population or ethnic subgroups for whole pregnancy (Table 4) or for separate trimesters (data not shown).

In joint THM and HAA exposure models (Table 5; see also Supplemental Material, Table S5), we find no evidence of association between birth weight and DBP7 ingestion via water consumption. However, for TTHM uptake via showering, bathing, and swimming we observe statistically significant trends across tertiles, and reductions in mean birth weight for the highest exposure tertile for Pakistaniorigin infants for whole-pregnancy (-67.4 g; 95% CI: -106.1, -28.6) and trimester-specific exposures. Interactions between ethnicity and TTHM uptake via showering, bathing, and swimming were significant except for trimester 1. We found statistically significant trends across time-weighted average DBP7 concentration tertiles for whole pregnancy (mean birth weight difference for highest tertile: -60.2 g; 95% CI: -117.4, -3.1) (Table 5) for Pakistani-origin infants, and for white British infants for trimester 1 (trimester-specific data not shown).

Longer bathing duration is associated with birth weight reductions for Pakistani-origin but not white British infants (Pakistani-origin: -55.9 g; 95% CI: -102.2, -9.7 and white British: -17.6 g; 95% CI: -59.3, 24.2 for > 120 min/week vs. 0 min/week bathing). Cold tap water consumption is associated with increased birth weight for Pakistaniorigin infants but not for white British infants (Pakistani-origin: 71.4 g; 95% CI: 23.4, 119.4 and white British: 23.7 g; 95% CI: -16.6, 63.9 for > 1.6 L/day vs. 0-0.4 L/day cold tap water consumption).

We examined differences between Pakistani-origin women in the highest THM uptake tertile versus those in the lower tertiles. Those in the highest tertile were slightly younger (highest tertile: mean age, 27.0 years; 95% CI: 26.7, 27.4 vs. lower tertiles: mean age, 27.7 years; 95% CI: 27.6, 27.8), 22% versus 42% completed the questionnaire in Urdu or Mirpuri, 28% versus 21% were employed, 8% versus 12% had gestational diabetes. In terms of lifestyle, 6% versus 2% were current smokers, 33% versus 23% were exposed to environmental tobacco smoke, and 7% versus 5% consumed > 200 mg caffeine per day. In sensitivity analyses additionally adjusting models for environmental tobacco smoke and language of questionnaire completion, neither variable altered the relationship between birth weight and THM uptake for Pakistani-origin infants (data not shown).

Discussion

To our knowledge, this is by far the largest study to date on DBPs and birth weight with data on individual-level water use. Moreover, it is the first study in DBP epidemiology to incorporate detailed individual-level estimates of joint THM-HAA exposures. Our results suggest that maternal exposure to THMs during pregnancy (measured as both integrated THM uptake and exposure to THMs via showering, bathing, and swimming) is inversely associated with term birth weight for infants of Pakistani-origin. No associations were observed for white British infants. We found no evidence of association between birth weight and ingestion of HAAs alone, or THMs and HAAs combined, via drinking water consumption.

Strengths of our exposure assessment were modeling of DBP data to give estimates both temporally and spatially, even for times and places for which data were sparse; comprehensive exposure assessment combining individual-level water usage with modeled area-level DBP concentrations; and accounting for exposure modifiers (boiling/filtering).

Given the size of our study, it has greater statistical power than similar studies that have preceded it. It also benefits from many detailed individual-level covariates to address potential confounding and the homogeneity of the Pakistani-origin ethnic classification, because we did not lump all South Asian ethnicities together.

It is a limitation that we could not account for residential mobility during pregnancy, and only partially account for mobility between WSZs due to workplace location. However, given the low spatial variability in DBPs between WSZs in the study area, we expect the impact to be small in terms of exposure misclassification.

Compared with previous studies that had similar exposure assessment (i.e., integrated THM uptake), the 50-g birth weight reduction in Pakistani-origin infants we estimated for TTHM uptake > 1.82 [micro]g/day (compared with reference < 1.05 [micro]g/day, representing an exposure contrast of at least 0.77 [micro]g/day) is consistent with 45- to 50-g birth weight reduction associated with a 1-[micro]g/day increase in TTHM internal dose reported by Grazuleviciene et al. (2011). In contrast, two previous studies report no evidence for an association between birth weight and personal THM exposure, either as integrated THM uptake or via separate consumption or showering/bathing pathways (Hoffman et al. 2008; Villanueva et al. 2011).

We found no evidence of association between HAA exposure and birth weight, consistent with three other studies (Hoffman et al. 2008; Horton et al. 2011; Wright et al. 2004). In contrast, other studies have reported birth weight deficits associated with HAA5 exposure (Rivera-Nunez and Wright 2013) and high levels of urinary TCAA (Zhou et al. 2012).

We found evidence of association between time-weighted average DBP7 concentration and reduced birth weight, similar to the results of Rivera-Nunez and Wright (2013), who reported birth weight deficits of 39-45 g for DBP9 (sum of TTHM and HAA5) concentration (e.g., -39 g; 95% CI: -62, -16 for > 97 [micro]g/L vs. 0 [micro]g/L of DBP9). However, we found no evidence of association for individual-level DBP7 ingestion. This may reflect the positive health benefits of higher water consumption, and that women who drink lots of water tend to be healthier, both of which could mask possible adverse effects of DBP exposure via this route; and we did observe associations between cold tap water consumption and increased birth weight. Behavior in pregnancy may be influenced by the health of the pregnancy, and we cannot rule out reverse causality in the water consumption or bathing associations we observe. That we observed a relationship between birth weight and DBP7 levels in tap water suggests that the relationship between integrated THM exposure and birth weight observed for Pakistani-origin infants cannot be explained by residual confounding due to unmeasured variables related to showering and bathing (e.g., use of personal care products, which could differ by ethnicity).

We observed that uptake of TTHM via showering, bathing, and swimming--predominantly via inhalation and dermal absorption routes--was associated with reduced birth weight in a dose-response manner for Pakistani-origin infants, when adjusting for exposure to THM and HAA ingestion via water consumption. This suggests that birth weight reductions observed in integrated THM uptake models are driven by exposure via showering, bathing, and swimming (rather than consumption). A previous study also found THM exposure via showering/bathing, but not other routes, to be associated with fetal growth restriction (Costet et al. 2012).

To explain inconsistent findings in DBP epidemiology, it has been hypothesized that THMs may be a proxy for an unmeasured putative agent, for example, another DBP/ pollutant. The present study suggests that THMs are not acting as a proxy for HAAs, or vice versa. It is possible that the association we observed between birth weight and THM uptake could reflect other DBPs (e.g., emerging DBPs), for which THMs could act as a proxy. However, of 24 previous studies examining quantitative measures of THM exposure and fetal growth outcomes, 18 report evidence for associations between exposure and outcome (Bove et al. 1995; Costet et al. 2012; Danileviciute et al. 2012; Gallagher et al. 1998; Grazuleviciene et al. 2011; Hinckley et al. 2005; Hoffman et al. 2008; Iszatt et al. 2014; Kramer et al. 1992; Kumar et al. 2014; Levallois et al. 2012; Lewis et al. 2006; Porter et al. 2005; Rivera-Nunez and Wright 2013; Summerhayes et al. 2012; Toledano et al. 2005; Wright et al. 2003, 2004), whereas 6 do not (Dodds et al. 1999; Horton et al. 2011; Patelarou et al. 2011; Savitz et al. 1995; Villanueva et al. 2011; Yang et al. 2007). It is thus difficult to dismiss THMs as a proxy for other DBPs, because it would have to be the same proxy in all these studies. This is unlikely, given that DBP mixture composition and corresponding toxicity has been shown to vary substantially in different study locations (Jeong et al. 2012).

To our knowledge, there is no existing literature on the relationship between birth weight and DBP exposure specifically for infants of Pakistani-origin with which we can compare our main findings. However, ethnic differences (Caucasian vs. non-Caucasian) in term LBW risk associated with high TTHM exposure have previously been observed (Lewis et al. 2006).

Smoking, drinking alcohol, and caffeine consumption are known risk factors for adverse fetal growth outcomes, which may also be related to water consumption/activities and thus potentially confound the relationship between DBP exposure and birth weight. In the present study, very few Pakistani-origin women smoked, drank alcohol, or consumed [greater than or equal to] 200 mg caffeine per day, but prevalence of these activities was much higher among white British women. With higher prevalence, these risk factors may have greater impact in the white British subpopulation--the exposure-outcome relationship could be more confounded, or the adverse effects of these risk factors may dwarf potential deleterious effects of DBPs, making it more difficult to tease out effects of DBPs in this population. The differences in effect between ethnic groups could be explained by residual confounding because other factors such as diet and stress, which may differ by ethnicity, were not taken into account in this study.

Alternatively, there may be potential differences in DBP uptake or metabolism by ethnicity, resulting from differences in body composition, hepatic function, or genetic variation between ethnic groups. South Asian adults (including those of Pakistani origin) have greater relative fat mass than European-origin adults (McKeigue et al. 1991). Additionally, in this population Pakistani-origin infants, although lighter in weight, are relatively more adipose than white British infants (West et al. 2013). THMs take longer to partition out of physiological compartments such as adipose tissue, compared to blood (Blount et al. 2011). Greater fat mass could result in greater relative uptake of THMs, which are lipophilic, into adipose tissue for the same maternal exposure, potentially resulting in longer-lived exposure for those with greater fat mass.

It is also plausible that ethnic variation in genes related to metabolic enzymes could result in different patterns of metabolism by ethnicity. Specific polymorphisms in CYP2D6 and GSTT1 genes have been associated with significant differences in blood THM concentrations following showering (Backer et al. 2008). Other chemical compounds demonstrate different patterns of metabolism by ethnicity--for example, arsenic (Brima et al. 2006) and various pharmaceutical drugs (Yasuda et al. 2008). Within BiB, associations between residential greenness and birth weight differ by ethnicity, with a positive association for white British infants, but not for those of Pakistani-origin (Dadvand et al. 2014).

Conclusions

This large study, which is the first to incorporate individual-level ingestion estimates of DBP mixtures, substantially strengthens a considerable body of evidence suggesting that exposure to THMs during pregnancy is associated with adverse fetal growth outcomes, including reduced birth weight. Our observation of a birth weight reduction of approximately 50 g associated with high THM exposure for Pakistani-origin infants could have a proportionally greater health impact on these infants now and throughout life, because they already weigh considerably less on average than white British infants at birth. This study makes a valuable contribution to the more limited evidence base for HAAs by indicating that HAA exposures are not associated with term birth weight in our study population.

There is virtually no epidemiological literature on THM-HAA mixtures and fetal growth, and this study advances the field in this respect by showing that associations with THMs were not the result of confounding by HAA exposures in our study population. Any future studies of DBPs and fetal growth should examine as wide a range of DBP classes/species as possible, and assess mixture effects. Further studies examining only single classes of DBPs are not warranted because they are unlikely to advance this field.

http://dx.doi.org/10.1289/ehp.1409480.

References

Aggazzotti G, Fantuzzi G, Righi E, Predieri G. 1995. Environmental and biological monitoring of chloroform in indoor swimming pools. J Chromatogr A 710:181-190.

Backer LC, Ashley DL, Bonin MA, Cardinali FL, Kieszak SM, Wooten JV. 2000. Household exposures to drinking water disinfection by-products: whole blood trihalomethane levels. J Expo Anal Environ Epidemiol 10:321-326.

Backer LC, Lan Q, Blount BC, Nuckols JR, Branch R, Lyu CW, et al. 2008. Exogenous and endogenous determinants of blood trihalomethane levels after showering. Environ Health Perspect 116:57-63; doi:10.1289/ehp.10049.

Blount BC, Backer LC, Aylward LL, Hays SM, LaKind JS. 2011. Human exposure assessment for DBPs: factors influencing blood trihalomethane levels. In: Encyclopedia of Environmental Health, (Nriagu JO, ed). Amsterdam:Elsevier, 100-107.

Bove FJ, Fulcomer MC, Klotz JB, Esmart J, Dufficy EM, Savrin JE. 1995. Public drinking water contamination and birth outcomes. Am J Epidemiol 141:850-862.

Brima EI, Haris PI, Jenkins RO, Polya DA, Gault AG, Harrington CF. 2006. Understanding arsenic metabolism through a comparative study of arsenic levels in the urine, hair and fingernails of healthy volunteers from three unexposed ethnic groups in the United Kingdom. Toxicol Appl Pharmacol 216:122-130.

Chu H, Nieuwenhuijsen MJ. 2002. Distribution and determinants of trihalomethane concentrations in indoor swimming pools. Occup Environ Med 59:243-247.

Costet N, Garlantezec R, Monfort C, Rouget F, Gagniere B, Chevrier C, et al. 2012. Environmental and urinary markers of prenatal exposure to drinking water disinfection by-products, fetal growth, and duration of gestation in the PELAGIE birth cohort (Brittany, France, 2002-2006). Am J Epidemiol 175:263-275.

Dadvand P, Wright J, Martinez D, Basagana X, McEachan RR, Cirach M, et al. 2014. Inequality, green spaces, and pregnant women: roles of ethnicity and individual and neighbourhood socioeconomic status. Environ Int 71:101-108.

Danileviciute A, Grazuleviciene R, Vencloviene J, Paulauskas A, Nieuwenhuijsen MJ. 2012. Exposure to drinking water trihalomethanes and their association with low birth weight and small for gestational age in genetically susceptible women. Int J Environ Res Public Health 9:4470-4485.

Dodds L, King W, Woolcott C, Pole J. 1999. Trihalomethanes in public water supplies and adverse birth outcomes. Epidemiology 10:233-237.

DWI (Drinking Water Inspectorate). 2010. 2010 No. 911. Water, England and Wales. The Water Supply Regulations 2010. Available: http://dwi.defra.gov. uk/stakeholders/legislation/wsr2010.pdf [accessed 29 January 2015].

Edwards SC. 2014. Haloacetic Acids in Public Drinking Water and Risk of Adverse Birth Outcomes in the Born in Bradford Cohort [PhD Dissertation]. London:Imperial College London.

Gallagher MD, Nuckols JR, Stallones L, Savitz DA. 1998. Exposure to trihalomethanes and adverse pregnancy outcomes. Epidemiology 9:484-489.

Grazuleviciene R, Nieuwenhuijsen MJ, Vencloviene J, Kostopoulou-Karadanelli M, Krasner SW, Danileviciute A, et al. 2011. Individual exposures to drinking water trihalomethanes, low birth weight and small for gestational age risk: a prospective Kaunas cohort study. Environ Health 10:32; doi:10.1186/1476-069X-10-32.

Hinckley AF, Bachand AM, Reif JS. 2005. Late pregnancy exposures to disinfection by-products and growth-related birth outcomes. Environ Health Perspect 113:1808-1813; doi:10.1289/ehp.8282.

Hoffman CS, Mendola P, Savitz DA, Herring AH, Loomis D, Hartmann KE, et al. 2008. Drinking water disinfection by-product exposure and fetal growth. Epidemiology 19:729-737.

Horton BJ, Luben TJ, Herring AH, Savitz DA, Singer PC, Weinberg HS, et al. 2011. The effect of water disinfection by-products on pregnancy outcomes in two southeastern US communities. J Occup Environ Med 53:1172-1178.

IPCS (International Programme on Chemical Safety). 2000. Environmental Health Criteria 216: Disinfectants and Disinfectant By-Products. Geneva:World Health Organization. Available: http://apps.who.int/iris/bitstream/10665/42274/1/WHO_EHC_216.pdf [accessed 11 November 2014].

Iszatt N, Nieuwenhuijsen MJ, Bennett JE, Toledano MB. 2014. Trihalomethanes in public drinking water and stillbirth and low birth weight rates: an intervention study. Environ Int 73:434-439.

Jeong CH, Wagner ED, Siebert VR, Anduri S, Richardson SD, Daiber EJ, et al. 2012. Occurrence and toxicity of disinfection byproducts in European drinking waters in relation with the HIWATE epidemiology study. Environ Sci Technol 46:12120-12128.

Kramer MD, Lynch CF, Isacson P, Hanson JW. 1992. The association of waterborne chloroform with intrauterine growth retardation. Epidemiology 3:407-413.

Krasner SW, McGuire MJ, Jacangelo JG, Patania NL, Reagan KM, Aieta EM. 1989. The occurrence of disinfection by-products in US drinking water. J Am Water Works Assoc 81:41-53.

Kumar S, Forand S, Babcock G, Richter W, Hart T, Hwang SA. 2014. Total trihalomethanes in public drinking water supply and birth outcomes: a crosssectional study. Matern Child Health J 18:996-1006.

Levallois P, Gingras S, Marcoux S, Legay C, Catto C, Rodriguez M, et al. 2012. Maternal exposure to drinking-water chlorination by-products and small-for-gestational-age neonates. Epidemiology 23:267-276.

Lewis C, Suffet IH, Ritz B. 2006. Estimated effects of disinfection by-products on birth weight in a population served by a single water utility. Am J Epidemiol 163:38-47.

Lunn DJ, Thomas A, Best N, Spiegelhalter D. 2000. WinBUGS--a Bayesian modelling framework: concepts, structure, and extensibility. Stat Comput 10:325-337.

Lynberg M, Nuckols JR, Langlois P, Ashley D, Singer P, Mendola P, et al. 2001. Assessing exposure to disinfection by-products in women of reproductive age living in Corpus Christi, Texas, and Cobb County, Georgia: descriptive results and methods. Environ Health Perspect 109:597-604.

Malliarou E, Collins C, Graham N, Nieuwenhuijsen MJ. 2005. Haloacetic acids in drinking water in the United Kingdom. Water Res 39:2722-2730.

McKeigue PM, Shah B, Marmot MG. 1991. Relation of central obesity and insulin resistance with high diabetes prevalence and cardiovascular risk in South Asians. Lancet 337:382-386.

McLennan D, Barnes H, Noble M, Davies J, Garratt E, Dibben C. 2011. The English Indices of Deprivation 2010. London:Department for Communities and Local Government. Available: https://www.gov.uk/ government/uploads/system/uploads/attachment_ data/file/6320/1870718.pdf [accessed 8 February 2015].

Moser K, Stanfield KM, Leon DA. 2008. Birthweight and gestational age by ethnic group, England and Wales 2005: introducing new data on births. Health Stat Q 39:22-31,34-55.

Nieuwenhuijsen MJ, Toledano MB, Bennett J, Best N, Hambly P, de Hoogh C, et al. 2008. Chlorination disinfection by-products and risk of congenital anomalies in England and Wales. Environ Health Perspect 116:216-222; doi:10.1289/ehp.10636.

Nuckols JR, Ashley DL, Lyu C, Gordon SM, Hinckley AF, Singer P. 2005. Influence of tap water quality and household water use activities on indoor air and internal dose levels of trihalomethanes. Environ Health Perspect 113:863-870; doi:10.1289/ehp.7141.

Patelarou E, Kargaki S, Stephanou EG, Nieuwenhuijsen M, Sourtzi P, Gracia E, et al. 2011. Exposure to brominated trihalomethanes in drinking water and reproductive outcomes. Occup Environ Med 68:438-445.

Porter CK, Putnam SD, Hunting KL, Riddle MR. 2005. The effect of trihalomethane and haloacetic acid exposure on fetal growth in a Maryland county. Am J Epidemiol 162:334-344.

R Core Team. 2012. R: A Language and Environment for Statistical Computing. Vienna, Austria:R Foundation for Statistical Computing. Available: http://www.R-project.org.

Raynor P, Born in Bradford Collaborative Group. 2008. Born in Bradford, a cohort study of babies born in Bradford, and their parents: protocol for the recruitment phase. BMC Public Health 8:327; doi:10.1186/1471-2458-8-327.

Richardson SD, Plewa MJ, Wagner ED, Schoeny R, DeMarini DM. 2007. Occurrence, genotoxicity, and carcinogenicity of regulated and emerging disinfection by-products in drinking water: a review and roadmap for research. Mutat Res 636:178-242.

Rivera-Nunez Z, Wright JM. 2013. Association of brominated trihalomethane and haloacetic acid exposure with fetal growth and preterm delivery in Massachusetts. J Occup Environ Med 55:1125-1134.

Savitz DA, Andrews KW, Pastore LM. 1995. Drinking water and pregnancy outcome in central North Carolina: source, amount, and trihalomethane levels. Environ Health Perspect 103:592-596.

Smith RB. 2011. Assessment and Validation of Exposure to Disinfection By-products during Pregnancy, in an Epidemiological Study examining Associated Risk of Adverse Fetal Growth Outcomes (PhD Dissertation). London:Imperial College London.

Summerhayes RJ, Morgan GG, Edwards HP, Lincoln D, Earnest A, Rahman B, et al. 2012. Exposure to trihalomethanes in drinking water and small-for-gestational-age births. Epidemiology 23:15-22.

Toledano MB, Nieuwenhuijsen MJ, Best N, Whitaker H, Hambly P, de Hoogh C, et al. 2005. Relation of trihalomethane concentrations in public water supplies to stillbirth and birth weight in three water regions in England. Environ Health Perspect 113:225-232; doi:10.1289/ehp.7111.

Tung HH, Unz RF, Xie YF. 2006. HAA removal by GAC adsorption. J Am Water Works Assoc 98:107-112. U.S. EPA (U.S. Environmental Protection Agency). 2003. Method 552.3: Determination of Haloacetic Acids and Dalapon in Drinking Water by Liquid-Liquid Microextraction, Derivatization, and Gas Chromatography with Electron Capture Detection. Revision 1.0. EPA 815-B-03-002. Cincinnati, OH:U.S. EPA Office of Ground Water and Drinking Water. Available: http://nepis.epa.gov/Exe/ZyPURL. cgi?Dockey=901V0400.txt [accessed 11 November 2014].

Villanueva CM, Gagniere B, Monfort C, Nieuwenhuijsen MJ, Cordier S. 2007. Sources of variability in levels and exposure to trihalomethanes. Environ Res 103:211-220.

Villanueva CM, Gracia-Lavedan E, Ibarluzea J, Santa Marina L, Ballester F, Llop S, et al. 2011. Exposure to trihalomethanes through different water uses and birth weight, small for gestational age, and preterm delivery in Spain. Environ Health Perspect 119:1824-1830; doi:10.1289/ehp.1002425.

West J, Lawlor DA, Fairley L, Bhopal R, Cameron N, McKinney PA, et al. 2013. UK-born Pakistani-origin infants are relatively more adipose than white British infants: findings from 8704 mother-offspring pairs in the Born-in-Bradford prospective birth cohort. J Epidemiol Community Health 67:544-551.

Wright JM, Schwartz J, Dockery DW. 2003. Effect of trihalomethane exposure on fetal development. Occup Environ Med 60:173-180.

Wright JM, Schwartz J, Dockery DW. 2004. The effect of disinfection by-products and mutagenic activity on birth weight and gestational duration. Environ Health Perspect 112:920-925; doi:10.1289/ehp.6779.

Wright J, Small N, Raynor P, Tuffnell D, Bhopal R, Cameron N, et al. 2013. Cohort profile: the Born in Bradford multi-ethnic family cohort study. Int J Epidemiol 42:978-991.

Xu X, Mariano TM, Laskin JD, Weisel CP. 2002. Percutaneous absorption of trihalomethanes, haloacetic acids, and haloketones. Toxicol Appl Pharmacol 184:19-26.

Xu X, Weisel CP. 2003. Inhalation exposure to haloacetic acids and haloketones during showering. Environ Sci Technol 37:569-576.

Yang CY, Xiao ZP, Ho SC, Wu TN, Tsai SS. 2007. Association between trihalomethane concentrations in drinking water and adverse pregnancy outcome in Taiwan. Environ Res 104:390-395.

Yasuda SU, Zhang L, Huang SM. 2008. The role of ethnicity in variability in response to drugs: focus on clinical pharmacology studies. Clin Pharmacol Ther 84:417-423.

Zhou WS, Xu L, Xie SH, Li YL, Li L, Zeng Q, et al. 2012. Decreased birth weight in relation to maternal urinary trichloroacetic acid levels. Sci Total Environ 416:105-110.

Rachel B. Smith, (1) * Susan C. Edwards, (1) * Nicky Best, (1) John Wright, (2) Mark J. Nieuwenhuijsen, (1,3) and Mireille B. Toledano (1)

(1) MRC-PHE (Medical Research Council-Public Health England) Centre for Environment and Health, Department of Epidemiology and Biostatistics, School of Public Health, Imperial College London, St Mary's Campus, Norfolk Place, London, United Kingdom; (2) Bradford Institute for Health Research, Bradford Royal Infirmary, Bradford, United Kingdom; (3) Center for Research in Environmental Epidemiology (CREAL), IMIM (Hospital del Mar Medical Research Institute), CIBERESP (Centro de Investigation Biomedica en Red de Epidemiologia y Salud Publica), Barcelona, Spain

* These authors contributed equally to this work.

Address correspondence to M.B. Toledano, Department of Epidemiology and Biostatistics, Imperial College London, St Mary's Campus, Norfolk Place, London W2 1PG, UK. Telephone: 44 20 7594 3298. E-mail: m.toledano@imperial.ac.uk

Supplemental Material is available online (http:// dx.doi.org/10.1289/ehp.1409480).

Born in Bradford (BiB) is only possible because of the enthusiasm and commitment of the Children and Parents in BiB. We are grateful to all the participants, health professionals, and researchers who have made BiB happen. We thank Yorkshire Water, particularly C. Hamilton, for provision of routine sampling data and assistance with add-on water sampling. We also thank J. Bennett for providing a trihalomethane model for use in this study, M. KostopoulouKaradaneli (University of the Aegean) and E. Goslan (Cranfield University) for analysis of haloacetic acid levels in water samples, E. Petherick for BiB data extraction, H. Wright for preliminary analyses of trihalomethanes and birth outcomes, and A. Lavigne for preliminary statistical advice.

This research was funded by HiWATE (Health Impacts of Long-Term Exposure to Disinfection By-products in Drinking Water in Europe) (EU 6th Framework Programme contract no. Food-CT-2006-036224), the Joint Environment & Human Health Programme (NERC grant NE/ E008844/1), an Economic and Social Research Council studentship (PTA-031-2006-00544 to R.B.S.), an MRC Capacity Building Studentship 2010-2013 to S.C.E., and a Research Training support stipend (S.C.E.). The MRC-PHE Centre for Environment and Health is funded by the UK Medical Research Council and Public Health England.

The authors declare they have no actual or potential competing financial interests.

Received: 14 November 2014; Accepted: 31 August 2015; Advance Publication: 4 September 2015; Final Publication: 1 May 2016.
Table 1. Birth outcomes and maternal characteristics.

                                      Total (n = 9,198)

                                               %
                                            missing
                            % (a) or         data     % undertaking
Variable                mean [+ or -] SD      (b)       activity

Birth outcomes
Term birth             3,295 [+ or -] 484      0
weight (g)
Term LBW (d)                   4               0
SGA                            12              0
Gestational age        39.8 [+ or -] 1.2       0
(weeks)
Maternal
characteristics
Maternal age                                   0
(years)
  < 25                         33
  25-29                        32
  30-34                        23
  [greater than or             13
  equal to] 35
Employed (yes)                 44             0.1
Highest educational                           0.2
qualification
  No qualifications            21
  School                       31
  Further education            15
  Higher education             25
  Other                        8
IMD quintile (e)                              0.2
  1 (most deprived)            66
  2                            18
  3                            11
  4                            3
  5 (least deprived)           2
Parity                                        4.0
  0                            40
  1                            27
  > 2                          29
BMI at recruitment     28.3 [+ or -] 5.5      3.6
(kg/[m.sup.2])
Gestational                    8              3.8
diabetes
Fasting glucose         4.5 [+ or -] 0.5      4.1
(mmol/L)
2 hr post-load          5.7 [+ or -] 1.5      4.1
glucose (mmol/L)
Current smoker                 14            0.04
ETS during                     32             0.4
pregnancy
Alcohol                        16             3.1
consumption (f)
Caffeine [greater              17             8.9
than or equal to]
200 mg/day (g)
Water use
Consumption (L/day)
  Cold tap water       1.14 [+ or -] 0.81      0           94
  Hot tap water        0.50 [+ or -] 0.58      0           81
  Total tap water      1.64 [+ or -] 0.95      0           99
  Bottled water        0.29 [+ or -] 0.54      0           35
  All water            1.92 [+ or -] 1.03      0           100
Showering                93 [+ or -] 74        0           72
(min/week) (h)
Bathing                 123 [+ or -] 119       0           68
(min/week) (h)
Combined                151 [+ or -] 126       0           100
showering/bathing
(min/week)
Swimming                 73 [+ or -] 47        0            7
(min/week) (h)

                               White British (n = 3,665, 40%)

                                               %
                                            missing
                            % (a) or         data     % undertaking
Variable                mean [+ or -] SD      (b)       activity

Birth outcomes
Term birth             3,424 [+ or -] 482      0
weight (g)
Term LBW (d)                   2               0
SGA                            8               0
Gestational age        40.0 [+ or -] 1.2       0
(weeks)
Maternal
characteristics
Maternal age                                   0
(years)
  < 25                         28
  25-29                        38
  30-34                        20
  [greater than or             13
  equal to] 35
Employed (yes)                 64            0.03
Highest educational                           0.1
qualification
  No qualifications            20
  School                       34
  Further education            17
  Higher education             19
  Other                        10
IMD quintile (e)                              0.2
  1 (most deprived)            50
  2                            22
  3                            18
  4                            6
  5 (least deprived)           3
Parity                                        3.5
  0                            48
  1                            29
  > 2                          19
BMI at recruitment     28.9 [+ or -] 5.8      3.0
(kg/[m.sup.2])
Gestational                    5              3.9
diabetes
Fasting glucose         4.4 [+ or -] 0.4      4.3
(mmol/L)
2 hr post-load          5.4 [+ or -] 1.3      4.3
glucose (mmol/L)
Current smoker                 28             0.1
ETS during                     43             0.3
pregnancy
Alcohol                        34             6.8
consumption (f)
Caffeine [greater              31             7.4
than or equal to]
200 mg/day (g)
Water use
Consumption (L/day)
  Cold tap water       1.11 [+ or -] 0.94      0           91
  Hot tap water        0.72 [+ or -] 0.75      0           81
  Total tap water      1.84 [+ or -] 1.14      0           99
  Bottled water        0.40 [+ or -] 0.61      0           47
  All water            2.23 [+ or -] 1.20      0           100
Showering                86 [+ or -] 64        0           71
(min/week) (h)
Bathing                 151 [+ or -] 142       0           76
(min/week) (h)
Combined                175 [+ or -] 139       0           100
showering/bathing
(min/week)
Swimming                 72 [+ or -] 49        0            4
(min/week) (h)

                               Pakistani origin
                               (n = 4,098, 45%)
                                               %
                                            missing
                            % (a) or         data
Variable                mean [+ or -] SD      (b)

Birth outcomes
Term birth             3,186 [+ or -] 462      0
weight (g)
Term LBW (d)                   6               0
SGA                            16              0
Gestational age        39.7 [+ or -] 1.2       0
(weeks)
Maternal
characteristics
Maternal age                                   0
(years)
  < 25                         36
  25-29                        28
  30-34                        24
  [greater than or             13
  equal to] 35
Employed (yes)                 23             0.1
Highest educational                           0.2
qualification
  No qualifications            26
  School                       31
  Further education            13
  Higher education             26
  Other                        4
IMD quintile (e)                              0.2
  1 (most deprived)            79
  2                            14
  3                            6
  4                            0
  5 (least deprived)           0
Parity                                        4.3
  0                            31
  1                            24
  > 2                          40
BMI at recruitment     27.9 [+ or -] 5.2      4.1
(kg/[m.sup.2])
Gestational                    11             3.8
diabetes
Fasting glucose         4.6 [+ or -] 0.6      3.8
(mmol/L)
2 hr post-load          5.9 [+ or -] 1.6      3.8
glucose (mmol/L)
Current smoker                 3             0.05
ETS during                     25             0.5
pregnancy
Alcohol                       0.2             0.1
consumption (f)
Caffeine [greater              6              9.6
than or equal to]
200 mg/day (g)
Water use
Consumption (L/day)
  Cold tap water       1.18 [+ or -] 0.65      0
  Hot tap water        0.33 [+ or -] 0.31      0
  Total tap water      1.51 [+ or -] 0.71      0
  Bottled water        0.15 [+ or -] 0.38      0
  All water            1.66 [+ or -] 0.77      0
Showering                94 [+ or -] 78        0
(min/week) (h)
Bathing                  96 [+ or -] 81        0
(min/week) (h)
Combined                128 [+ or -] 107       0
showering/bathing
(min/week)
Swimming                 75 [+ or -] 38        0
(min/week) (h)

                       Pakistani origin
                       (n = 4,098, 45%)

                        % undertaking      Ethnicity
Variable                   activity       p-value (c)

Birth outcomes
Term birth                                  < 0.001
weight (g)
Term LBW (d)                                < 0.001
SGA                                         < 0.001
Gestational age                             < 0.001
(weeks)
Maternal
characteristics
Maternal age                                < 0.001
(years)
  < 25
  25-29
  30-34
  [greater than or
  equal to] 35
Employed (yes)                              < 0.001
Highest educational                         < 0.001
qualification
  No qualifications
  School
  Further education
  Higher education
  Other
IMD quintile (e)                            < 0.001
  1 (most deprived)
  2
  3
  4
  5 (least deprived)
Parity                                      < 0.001
  0
  1
  > 2
BMI at recruitment                          < 0.001
(kg/[m.sup.2])
Gestational                                 < 0.001
diabetes
Fasting glucose                             < 0.001
(mmol/L)
2 hr post-load                              < 0.001
glucose (mmol/L)
Current smoker                              < 0.001
ETS during                                  < 0.001
pregnancy
Alcohol                                     < 0.001
consumption (f)
Caffeine [greater                           < 0.001
than or equal to]
200 mg/day (g)
Water use
Consumption (L/day)
  Cold tap water              98            < 0.001
  Hot tap water               81            < 0.001
  Total tap water            100            < 0.001
  Bottled water               21            < 0.001
  All water                  100            < 0.001
Showering                     69            < 0.001
(min/week) (h)
Bathing                       66            < 0.001
(min/week) (h)
Combined                     100            < 0.001
showering/bathing
(min/week)
Swimming                      2              0.62
(min/week) (h)

Abbreviations: ETS, environmental tobacco smoke;
IMD, Index of Multiple Deprivation; LBW, low birth
weight; SGA, small for gestational age.

(a) Percentage in category. (b) Percentage missing out
of total column n. (c) p-Value for f-test and chi-square
test for continuous and categorical variables respectively,
comparing white British vs. Pakistani-origin. (d) Birth weight
< 2,500 g and gestation [greater than or equal to] 37 weeks.
(e) Index of Multiple Deprivation (IMD) quintiles-quintiles
of relative deprivation at Lower Super Output Area (LSOA)
level across England. (f) only for those undertaking the activity.

Table 2. Maternal DBP exposures.

                                                 Total
                                              (n = 9,198)

                                                        % missing
  Variable                          Mean [+ or -] SD    data (b)

DBP exposures (whole pregnancy)
Time-weighted average
concentration ([micro]g/L)
  TTHM                             45.6 [+ or -] 4.0        0
  Chloroform                       37.8 [+ or -] 3.8        0
  BDCM                              6.6 [+ or -] 0.6        0
  DBCM                              0.9 [+ or -] 0.2        0
  THMBr                             7.7 [+ or -] 0.8        0
  DCAA                              8.9 [+ or -] 2.1       15
  TCAA                             12.5 [+ or -] 2.6       15
  BDCAA                             1.3 [+ or -] 0.5       15
Integrated THM uptake
([micro]g/day)
  TTHM                             1.86 [+ or -] 1.66       0
  Chloroform                       1.61 [+ or -] 1.46       0
  BDCM                             0.20 [+ or -] 0.16       0
  DBCM                             0.03 [+ or -] 0.03       0
  THMBr                            0.25 [+ or -] 0.21       0
Ingestion of HAA via
drinking water consumption
([micro]g/day)
  DCAA                             15.7 [+ or -] 10.4      15
  TCAA                             17.2 [+ or -] 11.0      15
  BDCAA                             1.6 [+ or -] 1.2       15
  HAA3                             34.5 [+ or -] 21.4      15

                                             White British
                                            (n = 3,665, 40%)

                                                        % missing
  Variable                          Mean [+ or -] SD    data (b)

DBP exposures (whole pregnancy)
Time-weighted average
concentration ([micro]g/L)
  TTHM                             45.6 [+ or -] 4.1        0
  Chloroform                       37.6 [+ or -] 3.8        0
  BDCM                              6.7 [+ or -] 0.7        0
  DBCM                              0.9 [+ or -] 0.3        0
  THMBr                             7.8 [+ or -] 0.9        0
  DCAA                              8.5 [+ or -] 2.2       15
  TCAA                             12.5 [+ or -] 2.8       15
  BDCAA                             1.3 [+ or -] 0.5       15
Integrated THM uptake
([micro]g/day)
  TTHM                             2.27 [+ or -] 1.98       0
  Chloroform                       1.96 [+ or -] 1.76       0
  BDCM                             0.24 [+ or -] 0.19       0
  DBCM                             0.04 [+ or -] 0.04       0
  THMBr                            0.30 [+ or -] 0.24       0
Ingestion of HAA via
drinking water consumption
([micro]g/day)
  DCAA                             17.5 [+ or -] 12.4      15
  TCAA                             18.5 [+ or -] 12.5      15
  BDCAA                             1.7 [+ or -] 1.3       15
  HAA3                             37.7 [+ or -] 24.7      15

                                           Pakistani origin
                                            (n = 4,098, 45%)

                                                        % missing
  Variable                          Mean [+ or -] SD    data (b)

DBP exposures (whole pregnancy)
Time-weighted average
concentration ([micro]g/L)
  TTHM                             45.6 [+ or -] 4.0        0
  Chloroform                       38.0 [+ or -] 3.7        0
  BDCM                              6.5 [+ or -] 0.5        0
  DBCM                              0.9 [+ or -] 0.2        0
  THMBr                             7.6 [+ or -] 0.6        0
  DCAA                              9.2 [+ or -] 2.0       15
  TCAA                             12.6 [+ or -] 2.4       15
  BDCAA                             1.3 [+ or -] 0.5       15
Integrated THM uptake
([micro]g/day)
  TTHM                             1.49 [+ or -] 1.20       0
  Chloroform                       1.29 [+ or -] 1.05       0
  BDCM                             0.16 [+ or -] 0.13       0
  DBCM                             0.02 [+ or -] 0.02       0
  THMBr                            0.20 [+ or -] 0.16       0
Ingestion of HAA via
drinking water consumption
([micro]g/day)
  DCAA                             14.7 [+ or -] 8.3       15
  TCAA                             16.6 [+ or -] 9.2       15
  BDCAA                             1.7 [+ or -] 1.1       15
  HAA3                             32.9 [+ or -] 17.5      15

                                    Ethnicity
  Variable                         p-value (b)

DBP exposures (whole pregnancy)
Time-weighted average
concentration ([micro]g/L)
  TTHM                                0.80
  Chloroform                         < 0.001
  BDCM                               < 0.001
  DBCM                               < 0.001
  THMBr                              < 0.001
  DCAA                               < 0.001
  TCAA                                0.17
  BDCAA                              < 0.001
Integrated THM uptake
([micro]g/day)
  TTHM                               < 0.001
  Chloroform                         < 0.001
  BDCM                               < 0.001
  DBCM                               < 0.001
  THMBr                              < 0.001
Ingestion of HAA via
drinking water consumption
([micro]g/day)
  DCAA                               < 0.001
  TCAA                               < 0.001
  BDCAA                               0.18
  HAA3                               < 0.001

Abbreviations: BDCAA, bromodichloroacetic acid;
BDCM, bromodichloromethane; DCAA, dichloroacetic acid;
DBCM, dibromochloromethane; HAA3, the sum of DCAA, TCAA,
and BDCAA; TCAA, trichloroacetic acid; THMBr,
total brominated THMs; TTHM, total trihalomethanes.

(a) For 15% of women, time-weighted average
HAA concentrations for their pregnancy could not be
calculated because their pregnancy commenced before
January 2007 (the earliest extent of the HAA models).
(b) p-Value for f-test and chi-square test for continuous
and categorical variables respectively, comparing white
British vs. Pakistani-origin.

Table 3. Relationship between term birth weight
and whole pregnancy THM exposure.

                                          Total (n = 7,438)

                                              Crude mean
                                            difference in
Integrated THM uptake                     term birth weight
([micro]g/day)                      n        (g) (95% CI)

TTHM
  < 1.05                          2,538       Reference
  [greater than or equal to]      2,447   0.3 (-26.5, 27.0)
  1.05 to < 1.82
  [greater than or equal to]      2,453   15.4 (-11.4, 42.2)
  1.82
  p-Value for trend (d)                         0.262
  p-Value for significance (e)                  0.438
THMBr
  < 0.14                          2,516       Reference
  [greater than or equal to]      2,496   20.1 (-6.6, 46.8)
  0.14 to < 0.26
  [greater than or equal to]      2,426    27.7 (0.8, 54.6)
  0.26
  p-Value for trend (d)                         0.043
  p-Value for significance (e)                  0.112

                                       Total (n = 7,438)

                                   Adjusted (a) mean
                                     difference in
Integrated THM uptake              term birth weight
([micro]g/day)                       (g) (95% CI)         n

TTHM
  < 1.05                               Reference         750
  [greater than or equal to]      -20.1 (-43.0, 2.7)     980
  1.05 to < 1.82
  [greater than or equal to]      -21.6 (-45.3, 2.1)    1,344
  1.82
  p-Value for trend (d)                  0.072
  p-Value for significance (e)           0.125
THMBr
  < 0.14                               Reference         678
  [greater than or equal to]      -11.0 (-33.9, 11.8)   1,000
  0.14 to < 0.26
  [greater than or equal to]      -20.0 (-44.0, 3.9)    1,366
  0.26
  p-Value for trend (d)                  0.101
  p-Value for significance (e)           0.256

                                   White British (n = 3,044)

                                   Adjusted (b) mean
                                  difference in term
Integrated THM uptake                birth weight
([micro]g/day)                       (g) (95% CI)         n

TTHM
  < 1.05                               Reference        1,432
  [greater than or equal to]      -13.6 (-53.4, 26.1)   1,098
  1.05 to < 1.82
  [greater than or equal to]      10.8 (-26.9, 48.4)     768
  1.82
  p-Value for trend (d)                  0.442
  p-Value for significance (e)           0.371
THMBr
  < 0.14                               Reference        1,491
  [greater than or equal to]       1.6 (-38.8, 42.0)    1,086
  0.14 to < 0.26
  [greater than or equal to]       6.5 (-31.9, 45.0)     721
  0.26
  p-Value for trend (d)                  0.717
  p-Value for significance (e)           0.070

                                   Pakistani origin
                                      (n = 3,298)

                                   Adjusted (b) mean
                                     difference in         p-Value
Integrated THM uptake              term birth weight     interaction
([micro]g/day)                        (g) (95% CI)           (c)

TTHM
  < 1.05                               Reference            0.011
  [greater than or equal to]       0.6 (-31.0, 32.1)
  1.05 to < 1.82
  [greater than or equal to]      -53.7 (-89.9, -17.5)
  1.82
  p-Value for trend (d)                  0.009
  p-Value for significance (e)           0.006
THMBr
  < 0.14                               Reference            0.074
  [greater than or equal to]       -6.5 (-38.0, 25.0)
  0.14 to < 0.26
  [greater than or equal to]      -56.4 (-93.1, -19.6)
  0.26
  p-Value for trend (d)                  0.006
  p-Value for significance (e)           0.007

Abbreviations: BDCM, bromodichloromethane; THMBr, total brominated
THMs; TTHM, total trihalomethanes.

(a) Adjusted for 10 maternal factors (caffeine intake, IMD,
education, fasting glucose, post-load glucose, ethnicity, smoking,
parity, age, BMI) and 2 infant factors (gestational age at delivery
as linear and quadratic terms, sex). (b) Ethnic subgroup analyses
excluded ethnicity covariate. (c) p-Value for significance of
exposure-ethnicity interaction term, as a whole, from F-test. (d)
p-Value for linear trend across tertiles, derived by including the
exposure term (coded as 0, 1, 2) as continuous variable in the
model. (e) p-Value for significance of categorical exposure term,
as a whole, within the model, from F-test.

Table 4. Relationship between term birth weight
and whole pregnancy HAA ingestion.

                                          Total (n = 6,529)

                                              Crude mean
                                          difference in term
                                           birth weight (g)
Ingestion ([micro]g/day)            n          (95% CI)

BDCAA
  < 1.01                          2,247        Reference
  [greater than or equal to]      2,145   -7.0 (-35.6, 21.5)
  1.01 to < 1.86
  [greater than or equal to]      2,137   -13.4 (-42.0, 15.1)
  1.86
  p-Value for trend (d)                          0.357
  p-Value for significance (e)                   0.654
HAA3
  < 23.82                         2,132        Reference
  [greater than or equal to]      2,174   -34.9 (-63.7, -6.1)
  23.82 to < 38.83
  [greater than or equal to]      2,223   -11.6 (-40.3, 17.0)
  38.83
  p-Value for trend (d)                          0.443
  p-Value for significance (e)                   0.053

                                  Total (n = 6,529)    White British
                                                        (n = 2,651)
                                  Adjusted (a) mean
                                  difference in term
                                   birth weight (g)
Ingestion ([micro]g/day)               (95% CI)              n

BDCAA
  < 1.01                              Reference             930
  [greater than or equal to]      0.7 (-23.3, 24.8)         835
  1.01 to < 1.86
  [greater than or equal to]      1.6 (-22.9, 26.0)         886
  1.86
  p-Value for trend (d)                 0.901
  p-Value for significance (e)          0.992
HAA3
  < 23.82                             Reference             843
  [greater than or equal to]      -18.9 (-43.3, 5.6)        771
  23.82 to < 38.83
  [greater than or equal to]      -0.6 (-25.5, 24.4)       1,037
  38.83
  p-Value for trend (d)                 0.973
  p-Value for significance (e)          0.222

                                    White British       Pakistani
                                     (n = 2,651)         origin
                                                       (n = 2,916)

                                  Adjusted (b) mean
                                  difference in term
                                   birth weight (g)
Ingestion ([micro]g/day)               (95% CI)             n

BDCAA
  < 1.01                              Reference            899
  [greater than or equal to]      4.3 (-34.0, 42.6)       1,030
  1.01 to < 1.86
  [greater than or equal to]      19.2 (-19.5, 57.9)       987
  1.86
  p-Value for trend (d)                 0.331
  p-Value for significance (e)          0.590
HAA3
  < 23.82                             Reference            895
  [greater than or equal to]      -9.5 (-50.4, 31.3)      1,120
  23.82 to < 38.83
  [greater than or equal to]      11.7 (-27.5, 50.9)       901
  38.83
  p-Value for trend (d)                 0.525
  p-Value for significance (e)          0.558

                                   Pakistani origin
                                     (n = 2,916)

                                  Adjusted (b) mean
                                  difference in term     p-Value
                                   birth weight (g)    interaction
Ingestion ([micro]g/day)               (95% CI)            (c)

BDCAA
  < 1.01                              Reference           0.876
  [greater than or equal to]      14.2 (-21.3, 49.7)
  1.01 to < 1.86
  [greater than or equal to]      19.0 (-17.5, 5.5)
  1.86
  p-Value for trend (d)                 0.310
  p-Value for significance (e)          0.568
HAA3
  < 23.82                             Reference           0.971
  [greater than or equal to]      0.1 (-34.9, 35.1)
  23.82 to < 38.83
  [greater than or equal to]      13.1 (-24.8, 51.0)
  38.83
  p-Value for trend (d)                 0.498
  p-Value for significance (e)          0.721

Abbreviations: BDCAA, bromodichloroacetic acid; DCAA,
dichloroacetic acid; HAA, haloacetic acid; HAA3, sum of DCAA, TCAA,
and BDCAA; TCAA, trichloroacetic acid.

(a) Adjusted for 10 maternal factors (caffeine intake, IMD,
education, fasting glucose, post-load glucose, ethnicity, smoking,
parity, age, BMI) and 2 infant factors (gestational age at delivery
as linear and quadratic terms, sex). (b) Ethnic subgroup analyses
excluded ethnicity covariate. (c) p-Value for significance of
exposure-ethnicity interaction term, as a whole, from F-test (d)
p-Value for linear trend across tertiles, derived by including the
exposure term (coded as 0, 1, 2) as continuous variable in the
model. (e) p-Value for significance of categorical exposure term,
as a whole, within the model, from F-test

Table 5. Relationship between term birth weight and whole
pregnancy joint THM and HAA exposures (DBP7).

                                             Total (n = 6,529)

                                                Crude mean
                                            difference in term
                                             birth weight (g)
Exposure                             n           (95% CI)

Individual-level exposure
model including
2 exposure terms (d)
  DBP7 via drinking water
    consumption
    ([micro]g/day) (e)
    < 58.6                         2,162         Reference
    [greater than or equal to]     2,173   -39.4 (-68.2, -10.7)
    58.6-97.0
    [greater than or equal to]     2,194    -14.5 (-43.2, 14.1)
    97.0
    p-Value for trend (f)                          0.326
    p-Value for significance (g)                   0.025
  Uptake of TTHM via showering,
  bathing, swimming
  ([micro]g/day) (h)
    < 0.85                         2,228         Reference
    [greater than or equal to]     2,148     3.8 (-24.8, 32.4)
    0.85-1.63
    [greater than or equal to]     2,153    11.2 (-17.4, 39.8)
    1.63
    p-Value for trend (f)                          0.371
    p-Value for significance (g)                   0.739
Time-weighted average DBP7
concentration ([micro]g/L) (i)
    < 65                           1,426         Reference
    [greater than or equal to]     4,410   -43.3 (-72.0, -14.5)
    65-75
    [greater than or equal to]      693    -99.8 (-143.5, -56.1)
    75
    p-Value for trend (f)                         < 0.001
    p-Value for significance (g)                  < 0.001

                                    Total (n = 6,529)       White
                                                           British
                                                         (n = 2,651)

                                    Adjusted (a) mean
                                   difference in term
                                    birth weight (g)
Exposure                                (95% CI)              n

Individual-level exposure
model including
2 exposure terms (d)
  DBP7 via drinking water
    consumption
    ([micro]g/day) (e)
    < 58.6                              Reference            957
    [greater than or equal to]     -23.6 (-48.0, 0.8)        795
    58.6-97.0
    [greater than or equal to]      0.4 (-24.0, 24.8)        899
    97.0
    p-Value for trend (f)                 0.955
    p-Value for significance (g)          0.085
  Uptake of TTHM via showering,
  bathing, swimming
  ([micro]g/day) (h)
    < 0.85                              Reference            626
    [greater than or equal to]     -23.9 (-48.2, 0.5)        844
    0.85-1.63
    [greater than or equal to]     -31.7 (-57.0, -6.4)      1,181
    1.63
    p-Value for trend (f)                 0.017
    p-Value for significance (g)          0.036
Time-weighted average DBP7
concentration ([micro]g/L) (i)
    < 65                                Reference            649
    [greater than or equal to]     -30.7 (-54.9, -6.5)      1,729
    65-75
    [greater than or equal to]     -44.8 (-82.3, -7.4)       273
    75
    p-Value for trend (f)                 0.007
    p-Value for significance (g)          0.020

                                      White British       Pakistani
                                       (n = 2,651)         origin
                                                         (n = 2,916)

                                    Adjusted (b) mean
                                   difference in term
                                    birth weight (g)
Exposure                                (95% CI)              n

Individual-level exposure
model including
2 exposure terms (d)
  DBP7 via drinking water
    consumption
    ([micro]g/day) (e)
    < 58.6                              Reference            817
    [greater than or equal to]     -16.3 (-55.3, 22.7)      1,090
    58.6-97.0
    [greater than or equal to]     21.3 (-16.4, 58.9)       1,009
    97.0
    p-Value for trend (f)                 0.278
    p-Value for significance (g)          0.171
  Uptake of TTHM via showering,
  bathing, swimming
  ([micro]g/day) (h)
    < 0.85                              Reference           1,308
    [greater than or equal to]     -28.9 (-71.9, 14.2)       938
    0.85-1.63
    [greater than or equal to]     -8.2 (-48.5, 32.1)        670
    1.63
    p-Value for trend (f)                 0.912
    p-Value for significance (g)          0.361
Time-weighted average DBP7
concentration ([micro]g/L) (i)
    < 65                                Reference            555
    [greater than or equal to]     -30.7 (-66.6, 5.3)       2,041
    65-75
    [greater than or equal to]     -20.1 (-78.6, 38.3)       320
    75
    p-Value for trend (f)                 0.251
    p-Value for significance (g)          0.277

                                     Pakistani origin
                                        (n = 2,916)

                                     Adjusted (b) mean     p-Value
                                    difference in term     inter-
                                     birth weight (g)      action
Exposure                                 (95% CI)            (c)

Individual-level exposure
model including
2 exposure terms (d)
  DBP7 via drinking water
    consumption
    ([micro]g/day) (e)
    < 58.6                               Reference          0.640
    [greater than or equal to]       4.8 (-31.8, 41.5)
    58.6-97.0
    [greater than or equal to]      18.1 (-19.5, 55.8)
    97.0
    p-Value for trend (f)                  0.335
    p-Value for significance (g)           0.595
  Uptake of TTHM via showering,
  bathing, swimming
  ([micro]g/day) (h)
    < 0.85                               Reference          0.032
    [greater than or equal to]      -14.9 (-48.7, 18.9)
    0.85-1.63
    [greater than or equal to]     -67.4 (-106.1, -28.6)
    1.63
    p-Value for trend (f)                  0.001
    p-Value for significance (g)           0.002
Time-weighted average DBP7
concentration ([micro]g/L) (i)
    < 65                                 Reference          0.589
    [greater than or equal to]       -35.2 (-73.1,2.7)
    65-75
    [greater than or equal to]     -60.2 (-117.4, -3.1)
    75
    p-Value for trend (f)                  0.027
    p-Value for significance (g)           0.073

Abbreviations: DBP7, sum of TTHM, DCAA, TCAA, and BDCAA; HAA,
haloacetic acid; THM, trihalomethanes; TTHM, total trihalomethanes.
(a) Adjusted for 10 maternal factors (caffeine intake, IMD,
education, fasting glucose, post-load glucose, ethnicity, smoking,
parity, age, BMI) and 2 infant factors (gestational age at delivery
as linear and quadratic terms, sex). (b) Ethnic subgroup analyses
excluded ethnicity covariate. (c) p-Value for significance of
exposure-ethnicity interaction term, as a whole, from F-test. (d)
This model includes both exposure terms: DBP7 via drinking water
consumption; and uptake of TTHM via showering, bathing, swimming.
(e) Consumption via drinking water ([micro]g/day) of sum of TTHM,
DCAA, TCAA, and BDCAA. (f) p-Value for linear trend across
tertiles, derived by including the exposure term (coded as 0, 1, 2)
as continuous variable in the model. (g) p-Value for significance
of categorical exposure term, as a whole, within the model, from
F-test. (h) [micro]g/day uptake into blood via these activities.
(i) Sum of time-weighted average concentrations of TTHM, DCAA,
TCAA, and BDCAA.
COPYRIGHT 2016 National Institute of Environmental Health Sciences
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2016 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Research: Children's Health
Author:Smith, Rachel B.; Edwards, Susan C.; Best, Nicky; Wright, John; Nieuwenhuijsen, Mark J.; Toledano, M
Publication:Environmental Health Perspectives
Article Type:Report
Geographic Code:4EUUK
Date:May 1, 2016
Words:11102
Previous Article:Organophosphate insecticide metabolites in prenatal and childhood urine samples and intelligence scores at 6 years of age: results from the...
Next Article:Developmental effects of the ToxCast[TM] phase I and phase II chemicals in Caenorhabditis elegans and corresponding responses in zebrafish, rats, and...
Topics:

Terms of use | Privacy policy | Copyright © 2022 Farlex, Inc. | Feedback | For webmasters |