Maternal alcohol use during pregnancy and offspring trajectories of height and weight: A prospective cohort study

Maternal alcohol use during pregnancy and offspring trajectories of height and weight: A prospective cohort study

Drug and Alcohol Dependence 153 (2015) 323–329 Contents lists available at ScienceDirect Drug and Alcohol Dependence journal homepage: www.elsevier...

489KB Sizes 0 Downloads 26 Views

Drug and Alcohol Dependence 153 (2015) 323–329

Contents lists available at ScienceDirect

Drug and Alcohol Dependence journal homepage: www.elsevier.com/locate/drugalcdep

Maternal alcohol use during pregnancy and offspring trajectories of height and weight: A prospective cohort study夽 Linda M. O’Keeffe a,b,∗ , Patricia M. Kearney b , Richard A. Greene a , Luisa Zuccolo c,d , Kate Tilling d , Debbie A. Lawlor c,d , Laura D. Howe c,d a

National Perinatal Epidemiology Centre, Department of Obstetrics and Gynecology, Cork University Maternity Hospital, Cork, Ireland Department of Epidemiology and Public Health, University College Cork, Cork, Ireland c UK Medical Research Council Integrative Epidemiology Unit at the University of Bristol, Bristol BS8 2BN, UK d School of Social and Community Medicine, University of Bristol, Bristol BS8 2BN, UK b

a r t i c l e

i n f o

Article history: Received 21 September 2014 Received in revised form 6 February 2015 Accepted 28 February 2015 Available online 12 March 2015 Keywords: Alcohol Pregnancy Growth

a b s t r a c t Background: Previous studies have examined associations between alcohol use in pregnancy and offspring birth size but evidence on whether associations persist during childhood is limited. Methods: We examined the association between maternal drinking during pregnancy and trajectories of offspring weight and height from 0 to 10 years in 7597 mother–child pairs in the Avon Longitudinal Study of Parents and Children. To strengthen the inference, we compared the maternal alcohol-offspring growth association with the partner alcohol-offspring growth association, to partially control for unmeasured confounding. We also performed sensitivity analyses restricting our analysis to women of white ethnicity and participants with three or more growth measures. Results: Maternal occasional or light daily drinking during pregnancy was not associated with reduced birth weight, birth length or offspring growth trajectories up to age 10 years. The infants of heavy drinking mothers were born 0.78 cm shorter (95% CI −1.34, −0.22) and 0.22 kg lighter (95% CI −0.34, −0.09) than infants of pregnancy abstainers but by age 10, offspring of heavy drinking mothers were of comparable height (mean difference 0.59 cm, 95% CI −0.93, 2.11) and weight (mean difference 0.41 kg, 95% CI −0.70, 1.52). These associations were not observed for heavy partner drinking and were not altered in sensitivity analyses. Conclusion: Maternal occasional or light daily drinking is not associated with birth weight, birth length or postnatal growth, but residual confounding may persist. Maternal heavy drinking may have an intrauterine association with reduced birth weight and length but this association is overcome during childhood. © 2015 Published by Elsevier Ireland Ltd.

1. Introduction Upwards of 35% of women in contemporary European birth cohorts and cross sectional studies report alcohol use during pregnancy (Andersen et al., 2012; Bakker et al., 2010; Colvin et al., 2007; Kelly et al., 2009; McCarthy et al., 2013; Mullally et al., 2011; O’Keeffe et al., 2013, 2015). Public health guidelines are polarized on the existence of a safe alcohol consumption threshold during pregnancy. Guidelines from Canada (Butt et al., 2011),

夽 Supplementary material can be found by accessing the online version of this paper at http://dx.doi.org and by entering http://dx.doi.org/10.1016/j.drugalcdep. 2015.02.035. ∗ Corresponding author at: National Perinatal Epidemiology Centre, Cork University Maternity Hospital, Cork, Ireland. E-mail addresses: [email protected], l.okeeffe[email protected] (L.M. O’Keeffe). http://dx.doi.org/10.1016/j.drugalcdep.2015.02.035 0376-8716/© 2015 Published by Elsevier Ireland Ltd.

the United States (Department of Health and Human Services, 2005), Ireland (Health Services Executive Ireland (HSE), 2009) and New Zealand (Ministry of Health, 2010) advise complete abstinence from alcohol during pregnancy while the National Institute of Health and Care Excellence (NICE) (2015) and Royal College of Obstetricians and Gynecologists (2006) in the United Kingdom advise an upper limit of 1–2 unit [28 g] not more than once or twice per week. A recent systematic review demonstrated that heavy alcohol consumption in pregnancy was associated with preterm birth, lower birth weight and size for gestational age, but found little evidence of harm at moderate levels (Patra et al., 2011). However, whether the smaller size at birth associated with heavy alcohol consumption persists into childhood or is overcome with age is not known. Moreover, the safety of lower consumption levels set out in the NICE guidelines is not established in relation to postnatal growth trajectories.

324

L.M. O’Keeffe et al. / Drug and Alcohol Dependence 153 (2015) 323–329

Given evidence of strong social patterning of alcohol use in pregnancy, there is the potential for residual confounding in its association with offspring outcomes because simple adjustment for one or two measures of socioeconomic position (as is often done in studies that have explored its relationship to perinatal and later outcomes) is likely to be inadequate to fully capture this confounding (Howe et al., 2013a). Specifically, studies that have shown modest consumption to be unrelated to outcomes might be masking a detrimental effect of even low levels of consumption since this pattern of exposure is common amongst those from more affluent and middle-class backgrounds in whom adverse outcome are likely to be low (Lewis et al., 2012). By contrast any influence of heavy consumption might be exaggerated as this pattern of consumption is more common amongst those from lower socioeconomic positions. Comparisons of maternal and partner associations can be used to investigate the likelihood that residual confounding might explain associations of maternal pregnancy exposures with offspring outcomes (Alati et al., 2006; Howe et al., 2012). These comparisons assume that the confounding structure for maternal and partner drinking in pregnancy is the same and that the association of maternal drinking and partner drinking with later offspring anthropometry (partly through later drinking and parenting styles) are similar. With respect to the question of intrauterine effects of maternal alcohol consumption in pregnancy, if associations with offspring growth trajectories are similar for maternal and partner alcohol consumption, this may be indicative of residual confounding by shared familial environmental or genetic factors. A stronger maternal, compared with partner, alcohol consumption—offspring growth trajectory association could be explained by maternal intrauterine effects, which are not plausible for partner consumption. Previous work in the Avon Longitudinal Study of Parents and Children (ALSPAC) cohort has shown that infants born to women who reported drinking one to two drinks daily with at least one binge, or three or more drinks daily with or without binges had a mean birth weight approximately 150 g less than infants whose mothers drank prior to pregnancy but reported abstinence during pregnancy (Passaro et al., 1996). Building on this work, we describe the association between maternal alcohol use during pregnancy and individual trajectories of height and weight between birth and 10 years of age. Second, we investigate potential residual confounding of estimates by comparing the associations of maternal drinking during pregnancy with that of her partner’s drinking during pregnancy.

2. Methods ALSPAC is a prospective birth cohort study in Southwest England. Pregnant women resident in one of the three Bristolbased health districts with an expected delivery date between April 1, 1991 and December 31, 1992 were invited to participate and the study has been described elsewhere in detail (Boyd et al., 2012; Fraser et al., 2012). ALSPAC initially enrolled a cohort of 14,451 pregnancies of which 13,867 live births occurred representing 13,761 unique women. Follow-up has included parent and child completed questionnaires, links to routine data and clinic attendance. Ethical approval was obtained from the ALSPAC Law and Ethics Committee and the local research ethics committees and written consent for participation in ALSPAC was obtained for all participants. Please note that the study website contains details of all the data that is available through a fully searchable data dictionary (University of Bristol, 2013).

2.1. Exposure classification Maternal alcohol consumption: Postal questionnaires at 18 weeks gestation collected information on the mother’s and partner’s alcohol consumption before pregnancy and during the first trimester. Response categories were never, less than once a week, at least once a week, one to two glasses every day, three to nine glasses every day, or 10 glasses a day. Examples were provided to specify that one glass was equivalent to one unit (8 g) of alcohol. Questions also asked about the number of binging days occurring during the preceding month which corresponded to approximately the fourth month of pregnancy. A binge was defined as consumption of the equivalent of two pints of beer, four glasses of wine or four pub measures of spirit on a single day (approximately four standard sized alcohol drinks or 40–45 g of absolute alcohol). In line with the summary index of maternal drinking created previously in the ALSPAC cohort (Passaro et al., 1996), we combined the information on binging behavior in the second trimester and drinking patterns in early pregnancy. The final categories for women and their partners were: (i) Pregnancy abstainer: drank alcohol pre-pregnancy, abstained in early pregnancy and did not binge mid pregnancy [women only]; (ii) General non-drinkers: reported not drinking three months prior to conception, not drinking in early pregnancy and not binging mid pregnancy; (iii) Occasional pregnancy drinker: drank less than daily in early pregnancy and did not binge mid pregnancy; (iv) Occasional binge drinkers during pregnancy: drank less than daily in early pregnancy and binged at least once in mid pregnancy; (v) Light daily pregnancy drinker: drank one or two drinks per day in early pregnancy, did not binge in mid pregnancy; and (vi) Heavy pregnancy drinker: drank one or two drinks per day in early pregnancy and binged at least once in mid pregnancy, or drank at least three drinks per day in early pregnancy with or without any reported binged in mid pregnancy. Due to evidence of poorer birth outcomes in non-drinkers compared with alcohol consumers (Mullally et al., 2011; Whitehead and Lipscomb, 2003), for all analyses of maternal alcohol consumption pregnancy abstainers were the reference group and for partner drinking, occasional drinkers served as the reference group. Only participants who provided information on their drinking habits both before and in early pregnancy (i.e., data available for both mother and her partner), with infants born after 23 weeks gestation, valid data on birth weight and birth length and complete data on all confounders were included, leading to a total sample of 7957 women–partner–offspring trios. Height and weight were available for ALSPAC offspring from research clinics, routine healthcare records, and parent-reports (full details of measurements in Supporting material S11 ). Height and weight at birth, 3 months, 1, 3, 7 and 10 years was predicted for all individuals from multilevel models, details of which are provided in the online supplement and have been published previously (Howe et al., 2013b). We included maternal and partner alcohol in the multilevel models, and estimated at each age the mean height/weight differences for each category of alcohol consumption compared with the reference category. Analyses were carried

1 Supplementary material can be found by accessing the online version of this paper at http://dx.doi.org and by entering http://dx.doi.org/10.1016/j.drugalcdep. 2015.02.035.

L.M. O’Keeffe et al. / Drug and Alcohol Dependence 153 (2015) 323–329

325

Table 1 Characteristics of women by their pregnancy alcohol consumption, n = 7957.

Partner alcohol consumption Occasional pregnancy drinker [reference] General non-drinker Occasional binge drinker Light daily pregnancy drinker Heavy pregnancy drinker Maternal education Less than O level O level A level Degree or above Parity 0 1 2+ Maternal smoking [1st trimester] No <10 per day 10–19 per day >19 per day Maternal age (years) Maternal height (meters) Maternal BMI Gestational age at birth (weeks)

Pregnancy abstainer, [reference] n = 2919 n (%)

General non-drinker, n = 504 n (%)

Occasional pregnancy drinker, n = 3253 n (%)

Occasional binge drinker, n = 1172 n (%)

488 (16.7%) 116 (4.0%) 1818 (62.3%) 8 (0.3%) 490 (16.8%)

97 (19.3%) 138 (27.4%) 217 (43.1%) 2 (0.4%) 50 (10.0%)

422 (13.0%) 68 (2.1%) 2034 (62.5%) 10 (0.3%) 722 (22.1%)

63 (5.4%) 12 (1.0%) 785 (67.0%) 1 (0.1%) 311 (26.5%)

1 (3.0%) 0 6 (18.2%) 1 (3.0%) 25 (75.8%)

2 (2.6%) 2 (2.6%) 27 (35.5%) 0 45 (59.2%)

p < 0.001

712 (24.4%) 1066 (36.5%) 724 (24.8%) 417 (14.3%)

204 (40.5%) 179 (35.5%) 77 (15.3%) 44 (8.7%)

661 (20.3%) 1118 (34.4%) 840 (25.8%) 634 (19.5%)

384 (32.8%) 423 (36.1%) 268 (22.9%) 97 (8.3%)

7 (21.2%) 7 (21.2%) 6 (18.2%) 13 (39.4%)

21 (27.6%) 27 (35.5%) 17 (22.4%) 11 (14.5%)

p < 0.001

1541 (52.8%) 982 (33.6%) 396 (13.6%)

193 (38.3%) 178 (35.3%) 133 (26.4%)

1466 (45.1%) 1201 (36.9%) 586 (18.0%)

475 (40.5%) 404 (34.5%) 293 (25.0%)

18 (54.6%) 9 (27.3%) 6 (18.2%)

43 (56.6%) 15 (19.7%) 18 (23.7%)

p < 0.001

2481 (85.0%) 226 (7.7%) 155 (5.3%) 57 (2.0%)

376 (74.6%) 53 (10.5%) 47 (9.3%) 28 (5.6%)

2675 (82.2%) 309 (9.5%) 215 (6.6%) 56 (1.7%)

784 (66.9%) 166 (14.2%) 173 (14.8%) 49 (4.2%)

23 (69.7%) 2 (6.1%) 7 (21.2%) 1 (3.0%)

30 (39.5%) 16 (21.1%) 22 (29.0%) 8 (10.5%)

p < 0.001

Mean (SD) 28.2 (4.6) 164 (6.6) 23.0(4.0) 39.5 (1.8)

Mean SD) 27.6 (4.8) 163.2 (6.5) 23.0 (4.2) 39.2 (2.1)

Mean (SD) 29.2 (4.6) 164.4 (6.7) 22.8 (3.6) 39.5 (1.8)

Mean (SD) 28.4 (4.7) 164.1 (6.7) 23.3 (3.8) 39.5 (1.8)

Mean (SD) 32.1 (4.4) 165.6 (5.8) 22.8 (2.8) 39.5 (1.4)

Mean (SD) 30.8 (5.5) 165.3 (7.1) 23.5 (3.7) 39.2 (2.1)

Light daily pregnancy drinker, n = 33 n (%)

Heavy pregnancy drinker, n = 76 n (%)

p Value for comparison*

p < 0.001 0.001 0.005 0.06

* p Value from chi squared difference in proportion between alcohol consumption groups for categorical variables and mean difference from student t test for continuous variables.

out using the statistical packages Stata12 (StataCorp,LP, Texas), MLwiN v2.27 (Centre for Multilevel Modelling, Bristol, UK, Centre for Multilevel Modelling, 2013) and the Stata command ‘runmlwin’ (Leckie and Charlton, 2013). We considered the following as potential confounders: maternal education, parity, maternal smoking, maternal age, maternal height, maternal BMI and partner drinking (all measured by mother- or partner-completed questionnaires; details in Supporting material S1,1 and included interactions between each confounder and the intercept and linear slopes in the multilevel models. We examined whether associations differed between females and males by looking at stratified results and computing a test for interaction between infant gender and maternal/partner alcohol consumption. In order to assess the robustness of our results to the exclusion of offspring with missing exposure, outcome or confounder data we examined the association between maternal and partner alcohol consumption and birth weight and length among excluded participants. We performed sensitivity analyses restricting our analysis to (i) women of white ethnicity, and (ii) participants with three or more growth measures, and we also evaluated whether adjusting for the following variables altered our findings: gestational age at birth (a potential mediator), pregnancy intention, partner education, partner smoking, maternal hospital admission, marital status, household social class, and partner height (potential confounders, not included in main analyses due to considerable amounts of missing data; measurement of variables and descriptive statistics of these confounders detailed in Supporting material S1 and Table S11 ). 3. Results Amongst the 7957 trios included (approximately 55% of the cohort recruited at birth), 57% of women and 96% of their partners consumed alcohol during the woman’s pregnancy [Supporting material S1 and Table S21 ]. Participants included in our analysis

were less likely to smoke and tended to be of higher socioeconomic position than those excluded due to missing data, but there was no clear pattern of difference in terms of maternal or partner alcohol consumption. Women classified as general non-drinkers and occasional binge drinkers during pregnancy tended to have lower educational attainment than pregnancy abstainers. Light daily pregnancy drinkers were more likely to have a degree than all other categories of drinker while heavy pregnancy drinkers and pregnancy abstainers had similar educational distributions (Table 1). Maternal heavy pregnancy drinking or light daily pregnancy drinking was correlated with partner heavy or daily use during pregnancy. Among women who drank during pregnancy, increasing severity of consumption was correlated with the severity of smoking. Associations of maternal and partner drinking with maternal characteristics (Table 2) are broadly similar, although maternal smoking is more strongly associated with maternal than partner drinking. Maternal general non-drinking is associated with a 0.22 cm lower birth length compared with pregnancy abstainers (95% CI −0.47 to 0.04); this height differential persists across childhood, with the difference at age 10 being −0.56 cm (95% CI −1.23 to 0.10) (Table 3). No clear differences, however, were seen for the maternal general non-drinking category in terms of weight at birth or later in childhood. Maternal occasional drinking, occasional binge drinking during pregnancy and light daily pregnancy drinking are not strongly associated with lower birth weight or birth length or height or weight later in childhood compared with pregnancy abstainers after adjustment for confounders and mutual adjustment for partner drinking (Table 3). Heavy maternal drinking during pregnancy is associated with 0.78 cm lower birth length (95% CI −1.34, −0.22) and 0.26 kg lower birth weight (95% CI −0.34, −0.09) (Table 3). By age 2 years, however, infants of heavy drinkers have similar heights and weights to the infants of abstainers (mean difference in height 0.11 cm,

326

L.M. O’Keeffe et al. / Drug and Alcohol Dependence 153 (2015) 323–329

Table 2 Characteristics of women’s partner’s level of drinking during the woman’s pregnancy, n = 7597.

Maternal education* Less than O level O level A level Degree or above Parity 0 1 2+ Smoking [1st trimester] No <10 per day 10–19 per day >19 per day Maternal age (years) Maternal height (meters) Maternal BMI Gestational age at birth (weeks)

Occasional drinker, [reference] n = 1073 n (%)

General non-drinker, n = 336 n (%)

Occasional pregnancy drinker, n = 4887 n (%)

Light daily pregnancy drinker, n = 22 n (%)

Heavy pregnancy drinker, n = 1639 n (%)

p Value for comparison**

271 (25.3%) 342 (31.9%) 283 (26.4%) 177 (16.5%)

136 (40.5%) 119 (35.4%) 49 (14.6%) 32 (9.5%)

1277 (26.1%) 1821 (37.3%) 1129 (23.1%) 660 (13.5%)

1 (4.6%) 7 (31.8%) 9 (40.9%) 5 (22.7%)

306 (18.6%) 531 (32.4%) 462 (28.2%) 342 (20.9%)

p < 0.001

420 (39.1%) 405 (37.8%) 248 (23.1%)

152 (45.2%) 102 (30.4%) 83 (24.4%)

2353 (48.2%) 1706 (34.9%) 828 (16.9%)

4 (18.2%) 9 (40.9%) 9 (40.9%)

807 (49.2%) 567 (34.6%) 265 (16.2%)

p < 0.001

920 (85.7%) 62 (5.8%) 69 (6.4%) 22 (2.1%)

242 (72.0%) 39 (11.6%) 35 (10.4%) 20 (6.0%)

3888 (79.6%) 491 (10.1%) 385 (7.9%) 123 (2.5%)

19 (86.4%) 1 (4.6%) 1 (4.6%) 1 (4.6%)

1298 (79.2%) 179 (10.9%) 129 (7.9%) 33 (2.0%)

p < 0.001

Mean (SD) 28.8 (5.0) 164 (6.6) 23.0 (4.1) 39.5 (1.7)

Mean (SD) 27 (5.5) 163.5 (6.8) 23.0 (4.6) 39.3 (2.2)

Mean (SD) 28.4 (4.5) 164 (6.7) 23 (3.8) 39.4 (1.8)

Mean (SD) 32.4 (3.8) 163.1 (9.0) 22.3 (2.1) 39.3 (3.0)

Mean (SD) 29.7 (4.7) 164.7 (6.5) 22.7 (3.5) 39.5 (1.8)

p < 0.001 0.003 0.14 0.27

* O-Level. (Ordinary Level; exams taken in different subjects usually at age 15–16 at the completion of legally required school attendance, equivalent to today’s UK General Certificate of Secondary Education), O-Level only, A-Level (Advanced-Level; exams taken in different subjects usually at age 18), or university degree or above. ** p Value from chi squared difference in proportion between alcohol consumption groups for categorical variables and mean difference from student t test for continuous variables.

Table 3 The association between maternal alcohol consumption in pregnancy and offspring length/height and weight at birth and ages 2 and 10 years. Mean (SD) values in pregnancy abstainers (the reference category)

Mean difference (95% confidence interval) compared with pregnancy abstainers

Pregnancy abstainer

General non-drinker

Occasional pregnancy drinker

Occasional binge drinker

Light daily pregnancy drinker

Heavy pregnancy drinker

Height (cm) Birth 2 years 10 years

50.22 (1.46) 85.95 (2.41) 141.03 (5.20)

−0.22 (−0.47, 0.04) −0.27 (−0.56, 0.02) −0.56 (−1.23, 0.10)

0.06 (−0.06, 0.19) −0.14 (−0.29, 0.00) −0.25 (−0.57, 0.08)

0.04 (−0.14, 0.21) −0.16 (−0.36, 0.04) −0.19 (−0.65, 0.27)

0.42 (−0.49, 1.33) −0.25 (−1.27, 0.76) 0.31 (−1.84, 2.46)

−0.78 (−1.34, −0.22) 0.11 (−0.56, 0.78) 0.59 (−0.93, 2.11)

Weight (kg) Birth 2 years 10 years

3.49 (0.42) 12.56 (1.17) 23.23 (6.02)

−0.03 (−0.09, 0.02) 0.01 (−0.13, 0.14) −0.08 (−0.89, 0.73)

0.00 (−0.03, 0.03) −0.04 (−0.11, 0.03) −0.05 (−0.71, 0.61)

0.01 (−0.03, 0.05) −0.02 (−0.11, 0.07) 0.05 (−0.64, 0.73)

−0.01 (−0.20, 0.18) −0.09 (−0.56, 0.37) 0.36 (−1.03, 1.76)

−0.22 (−0.34, −0.09) −0.09 (−0.41, 0.22) 0.41 (−0.70, 1.52)

Offspring height and weight values are predicted from the multi-level models, based on models adjusted for gender, confounders and partner’s alcohol consumption. Values are for females and the baseline category of all covariates. No evidence of gender interaction for the association between alcohol consumption and height/weight trajectories was observed. Height/lengths are in centimeters, weight is in kilograms.

95% CI −0.56, 0.78) and (mean difference in weight −0.09 kg, 95% CI −0.41, 0.22) (Table 3). By age 10, any weight or height differences have been entirely attenuated such that infants of heavy drinkers are of comparable height and weight to the offspring of pregnancy abstainers (mean difference in height 0.59 cm, 95% CI −0.93, 2.11) and (mean difference in weight 0.41 kg, 95% CI −0.70, 1.52) although these estimates also have wide confidence intervals that span positive as well as negative values. This increased growth in the offspring of maternal heavy drinkers tends to occur after the age of 3 months for height growth (Table S31 ] and after age 1 year for weight (Table S41 ) For all analyses of maternal alcohol use, unadjusted and confounder-adjusted results were similar (Table S51 ). Although, the reference groups for the maternal and partner alcohol comparisons differ, in contrast to the observed associations for maternal drinking, there is very little evidence of an association between heavy partner alcohol use during pregnancy and smaller length or weight at birth, but the offspring of partners who drink heavily tend to be shorter and lighter at 2 and 10 years (Table 4, and Tables S6 and S71 ).

3.1. Sensitivity analyses Our results for the associations of maternal drinking during pregnancy were similar in sensitivity analyses using observed height and weight at 2 and 10 years rather than the values predicted by the multilevel model, providing reassurance that our modeling strategy was appropriate (Table S81 ). Results were also similar when restricting analysis to offspring with three or more growth measures or women of white ethnicity (Tables S9 and S101 ). Findings were also robust to further adjustment for gestational age at birth, pregnancy intention, partner education, partner smoking, maternal hospital admission, marital status, household social class, and partner height (Tables S11–S181 ). Despite the participants included in our analysis being more socioeconomically advantaged than those excluded due to missing data, our analysis of the association between maternal alcohol consumption and birth weight and birth length (Table S191 ) showed that the observed associations did not differ between included and excluded participants, providing reassurance that missing data is unlikely to have biased our findings.

L.M. O’Keeffe et al. / Drug and Alcohol Dependence 153 (2015) 323–329

327

Table 4 The association between partner alcohol consumption in pregnancy and offspring length/height and weight at birth and ages 2 and 10 years. Mean (SD) values in pregnancy abstainers (the reference category)

Mean difference (95% confidence interval) compared with pregnancy abstainers

Occasional pregnancy drinker

General non-drinker

Occasional binge drinker

Light daily pregnancy drinker

Heavy pregnancy drinker

Height (cm) Birth 2 years 10 years

50.24 (1.43) 86.05 (2.55) 141.29 (5.32)

−0.25 (−0.57, 0.07) −0.35 (−0.74, 0.01) −0.85 (−1.70, 0.00)

−0.03 (−0.20, 0.14) −0.01 (−0.20, 0.19) −0.42 (−0.86, 0.01)

0.43 (−0.75, 1.61) −0.56 (−0.75, 1.86) −0.83 (−1.94, 3.61)

−0.07 (−0.27, 0.13) −0.25 (−0.48, −0.02) −0.93(−1.44, −0.42)

Weight (kg) Birth 2 years 10 years

3.48 (0.42) 12.57 (1.15) 24.21 (5.90)

−0.07 (−0.14, 0.00) −0.30 (−0.48, 0.13) −0.55 (−1.60, 0.50)

−0.02 (−0.05, 0.02) 0.03 (−0.07, 0.12) −0.12 (−0.65, 0.41)

0.10 (−0.14, 0.34) −0.16 (−0.76, 0.43) 0.16 (−3.11, 3.43)

−0.02 (−0.06, 0.03) −0.05 (−0.15, 0.06) −0.61 (−1.23, 0.00)

Values are predicted from the multi-level models, based on models adjusted for gender, confounders and mother’s alcohol consumption. Values are for females (baseline category) and no evidence of gender interaction for the association between alcohol consumption and height/weight trajectories was observed. Height/lengths are in centimeters, weight is in kilograms.

4. Discussion In this large prospective birth cohort in the South West of England, we found evidence that heavy alcohol consumption during pregnancy was associated with lower birth length and birth weight, but that these associations were subsequently overcome with age. By age ten, infants of heavy maternal drinkers are of comparable height and weight to infants of women who abstained from alcohol during pregnancy. The associations were not attenuated by adjustment for confounders or partner drinking. We did not find any evidence that heavy alcohol use in partners was associated with lower offspring birth weight or length, and in contrast to the maternal drinking results, heavy partner drinking was associated with shorter height and lighter weight later in childhood. Although the reference groups for maternal and partner exposure were different, making maternal, compared with partner, alcohol consumption—offspring growth trajectory associations not fully compatible, the overall discordance between the maternal and partner associations could imply that the detrimental effect of maternal heavy drinking during pregnancy on birth weight and length occurs through intra-uterine mechanisms. However, it should be noted that maternal smoking was more strongly associated with maternal alcohol use than with partner alcohol use, so we cannot exclude residual confounding by maternal smoking since our measure of smoking may not fully capture all aspects of smoking behavior. For all other categories of gestational drinkers we did not find any evidence of birth length or weight differences or height and weight growth differences up to 10 years. Our findings demonstrating no evidence of adverse associations between occasional or light daily drinking and birth length and weight are consistent with evidence on the observational associations of moderate alcohol use with head circumference, abdominal circumference, femur length in utero (Bakker et al., 2010) and birth weight (McCarthy et al., 2013; Patra et al., 2011). However, evidence on the impact of occasional or light daily consumption on growth outcomes postnatally is sparse and less consistent; some studies have also shown no association between more moderate drinking levels and postnatal growth/body size (O’Callaghan et al., 2003), but some previous studies have demonstrated height and weight deficits persisting into childhood (Barr et al., 1984; Day et al., 1991, 2002, 1990, 1999). Reasons for conflicting results include dissimilarities in the measurement and classification of exposure to alcohol during pregnancy, follow-up time and analysis differences and population differences where most studies showing an adverse association between light or moderate drinking and growth have been carried out in far more disadvantaged populations than the relatively socially advantaged ALSPAC cohort.

Our findings on birth weight and length differences among heavy drinkers are consistent with a number of large population based prospective cohort studies and results from meta-analysis showing adverse associations with growth at birth among heavy alcohol users during pregnancy (Patra et al., 2011). Similarly, our data illustrating catch up growth among heavy drinkers is compatible with evidence of associations with alcohol that are not evident at age 14 in a cohort of middle class white, married women in Seattle, a group similar to ALSPAC participants (Sampson et al., 1994). However, our results conflict with evidence demonstrating a deleterious effect of heavy alcohol consumption on growth during childhood among more disadvantaged populations including African-American women (Covington et al., 2002), teenagers in Pittsburgh (Cornelius et al., 2002), socioeconomically disadvantaged women in South Africa (Carter et al., 2012) and women of lower socioeconomic position in Northern France (Larroque and Kaminski, 1998). Genetic, nutritional, ethnic and socioeconomic differences and dissimilarities in the measurement and classification of exposure relating to the range and burden of the distribution of alcohol in each category may explain discordance between studies. Although we did not observe associations between our moderate drinker categories [occasional/light daily drinkers] in pregnancy and offspring growth trajectories, other studies have shown associations with birth outcomes even at lower levels (Nykjaer et al., 2014) and thus residual confounding even after adjustment for socio-demographic and other characteristics of women can obscure true associations between alcohol and growth. We have used maternal–partner exposure comparisons in this study to explore whether residual confounding is likely. An underlying assumption of this approach is that confounders are similarity related to the maternal and partner exposure in question. This approach works well for examining potential confounding due to shared postnatal familial socioeconomic, lifestyle, environmental or genetic characteristics but may not work well for confounding by maternal characteristics that might also have intrauterine effects. In our case, maternal smoking was more strongly associated with maternal than partner alcohol use. Given the known associations between maternal smoking and birth size and offspring growth (Oken et al., 2007; Howe et al., 2012) and difficulty in capturing smoking status fully and accurately in a questionnaire, the observed associations between maternal alcohol use and growth might reflect an effect of maternal smoking, or at least in part be explained by this, rather than fully due to their alcohol consumption. This study of postnatal growth trajectories among alcohol exposed offspring represents one of the largest and most robust studies of gestational alcohol and off-spring postnatal growth to

328

L.M. O’Keeffe et al. / Drug and Alcohol Dependence 153 (2015) 323–329

date, with a sample size close to 8000 women with complete data on exposure, outcomes and confounders, prospective follow up and multiple repeated measures of growth. However, due to loss to follow-up and missing data our analysis represents less than 60% of the ALSPAC cohort originally recruited at birth, thus potentially undermining generalizability due to attrition bias. In relation to broader generalizability, our cohort is predominantly of white ethnicity and relatively affluent thus potentially reducing generalizability to other more diverse populations. We relied on self-reported measures of alcohol consumption, but ascertainment of exposure has been done concurrently before women knew the outcomes of their pregnancy, reducing the risk of recall bias. Moreover, as data was collected in 1991 and 1992 under-reporting may be less of a concern due to much less awareness of the potential fetal effects of alcohol use during pregnancy. However, for the light daily drinking category, our associations may be underpowered due to small numbers (n = 33) and thus these results should be interpreted carefully. Habitual daily alcohol intake as well as binge consumption is combined in a summary index to reduce the risk of misclassification of exposure. However, because of the definition of ‘non-drinker’ (none in the 3 months prior to pregnancy or in early pregnancy), misclassification may occur in this category due to an inability to differentiate women abstaining from alcohol prior to pregnancy due to intention to get pregnant versus those who never drink alcohol regardless of pregnancy intention. We have also performed multiple sensitivity analyses which support the chosen analytical approach and strengthen the validity of our findings. Maternal heavy drinking is associated with lower birth weight and birth length but these associations are overcome with increasing offspring age. Our data do not demonstrate an adverse association between occasional or light daily drinking and birth weight, birth length or postnatal growth, weight and height up to age 10. However, residual confounding cannot be discounted and abstinence from alcohol during pregnancy should continue to be emphasized in routine antenatal care. Author disclosures Role of funding source Health Research Board (HRB) in Ireland under Grant No. PHD/2007/16 (to L.M.O.K); UK Medical Research Council Population Health Scientist Fellowship (G1002375 to L.D.H); The UK Medical Research Council, the Wellcome Trust and the University of Bristol provide core funding for ALPSAC (Grant reference 102215/2/13/2); The UK MRC and the University of Bristol provide core funding for the MRC Integrative Epidemiology Unit at the University of Bristol (MC UU 12013/5 and MC UU 12013/9). None of the funding sources had any further role in the study design; in the collection, analysis and interpretation of data; in writing of the report; or in the decision to submit the paper for publication. Contributors Linda O’Keeffe had the original idea for the study, wrote the first draft of the article, applied critical revisions to the article based on other co-author recommendations and approved the final version for publication. Patricia M Kearney contributed to the design of the study with other co-authors, contributed to critical revisions of the article and approved the final version for publication. Richard Greene conceptualized the design of the study with other co-authors, contributed to critical revisions of the article and approved the final version for publication.

Luisa Zuccolo conceptualized the design of the study with other co-authors, assisted in interpretation of the data, reviewed and revised the manuscript and approved the final version for publication. Kate Tilling conceptualized the design of the study including original modeling of growth trajectories, assisted to the interpretation of the data, contributed to critical revisions of the article and approved the final version for publication. Debbie Lawlor conceptualized the design of the study, contributed to critical revisions of the article and approved the final version for publication. Laura Howe supervised the analysis and write up of the manuscript, contributed to critical revisions of the article and approved the final version for publication. Conflict of interest No conflict declared. Acknowledgements We are extremely grateful to all the families that took part in this study, the midwives for their help in recruiting them and the whole ALSPAC team, which includes interviewers, computer and laboratory technicians, clerical workers, research scientists, volunteers, managers, research scientists, volunteers, managers, receptionists and nurses. Appendix A. Supplementary data Supplementary material related to this article can be found, in the online version, at http://dx.doi.org/10.1016/j.drugalcdep.2015. 02.035 References Alati, R., Al Mamun, A., Williams, G.M., O’Callaghan, M., Najman, J.M., Bor, W., 2006. In utero alcohol exposure and prediction of alcohol disorders in early adulthood—a birth cohort study. Arch. Gen. Psychiatry 63, 1009–1016. Andersen, A.M.N., Andersen, P.K., Olsen, J., Grønbæk, M., Strandberg-Larsen, K., 2012. Moderate alcohol intake during pregnancy and risk of fetal death. Int. J. Epidemiol. 41, 405–413. Bakker, R., Pluimgraaff, L.E., Steegers, E.A.P., Raat, H., Tiemeier, H., Hofman, A., Jaddoe, V.W.V., 2010. Associations of light and moderate maternal alcohol consumption with fetal growth characteristics in different periods of pregnancy: the Generation R Study. Int. J. Epidemiol. 39, 777–789. Barr, H.M., Streissguth, A.P., Martin, D.C., Herman, C.S., 1984. Infant size at 8 months of age: relationship to maternal use of alcohol, nicotine, and caffeine during pregnancy. Pediatrics 74, 336–341. Boyd, A., Golding, J., Macleod, J., Lawlor, D.A., Fraser, A., Henderson, J., Molloy, L., Ness, A., Ring, S., Smith, G.D., 2012. Cohort Profile: the ‘Children of the 90s’—the index offspring of the Avon Longitudinal Study of Parents and Children. Int. J. Epidemiol. 42, 111–127. Butt, P., Beirness, D., Gliksman, L., Paradis, C., Stockwell, T., 2011. Alcohol and Health in Canada: A Summary of Evidence and Guidelines for Low Risk Drinking. Canadian Centre on Substance Abuse, Ottawa. Carter, R.C., Jacobson, J.L., Molteno, C.D., Jiang, H., Meintjes, E.M., Jacobson, S.W., Duggan, C., 2012. effects of heavy prenatal alcohol exposure and iron deficiency anemia on child growth and body composition through age 9 years. Alcohol. Clin. Exp. Res. 36, 1973–1982. Centre for Multilevel Modelling, 2013. Mlwin: A Software Package for Fitting Multilevel Models, http://www.bristol.ac.uk/cmm/software/mlwin/ (accessed on 26/05/2013). Colvin, L., Payne, J., Parsons, D., Kurinczuk, J.J., Bower, C., 2007. Alcohol consumption during pregnancy in nonindigenous west Australian women. Alcohol. Clin. Exp. Res. 31, 276–284. Cornelius, M.D., Goldschmidt, L., Day, N.L., Larkby, C., 2002. Alcohol, tobacco and marijuana use among pregnant teenagers: 6-Year follow-up of offspring growth effects. Neurotoxicol. Teratol. 24, 703–710. Covington, C.Y., Nordstrom-Klee, B., Ager, J., Sokol, R., Delaney-Black, V., 2002. Birth to age 7 growth of children prenatally exposed to drugs: a prospective cohort study. Neurotoxicol. Teratol. 24, 489–496. Day, N.L., Goldschmidt, L., Robles, N., Richardson, G., Cornelius, M., Taylor, P., Geva, D., Stoffer, D., 1991. Prenatal alcohol exposure and offspring growth at 18 months

L.M. O’Keeffe et al. / Drug and Alcohol Dependence 153 (2015) 323–329 of age: the predictive validity of two measures of drinking. Alcohol. Clin. Exp. Res. 15, 914–918. Day, N.L., Leech, S.L., Richardson, G.A., Cornelius, M.D., Robles, N., Larkby, C., 2002. Prenatal alcohol exposure predicts continued deficits in offspring size at 14 years of age. Alcohol. Clin. Exp. Res. 26, 1584–1591. Day, N.L., Richardson, G., Robles, N., Sambamoorthi, U., Taylor, P., Scher, M., Stoffer, D., Jasperse, D., Cornelius, M., 1990. Effect of prenatal alcohol exposure on growth and morphology of offspring at 8 months of age. Pediatrics 85, 748–752. Day, N.L., Zuo, Y.L., Richardson, G.A., Goldschmidt, L., Larkby, C.A., Cornelius, M.D., 1999. Prenatal alcohol use and offspring size at 10 years of age. Alcohol. Clin. Exp. Res. 23, 863–869. Department of Health and Human Services, 2005. U.S. Surgeon General Releases Advisory on Alcohol Use in Pregnancy. Department of Health and Human Services, http://www.surgeongeneral.gov (accessed on 20/04/2013). Fraser, A., Macdonald-Wallis, C., Tilling, K., Boyd, A., Golding, J., Smith, G.D., Henderson, J., Macleod, J., Molloy, L., Ness, A., 2012. Cohort profile: the Avon Longitudinal Study of Parents and Children: ALSPAC mothers cohort. Int. J. Epidemiol. 42, 97–110. Health Services Executive Ireland (HSE), 2009. Alcohol and Pregnancy. Health Services Executive Ireland (HSE), http://www.yourdrinking.ie/alcohol-andpregnancy (accessed on 25/04/2013). Howe, L.D., Matijasevich, A., Tilling, K., Brion, M.J., Leary, S.D., Smith, G.D., Lawlor, D.A., 2012. Maternal smoking during pregnancy and offspring trajectories of height and adiposity: comparing maternal and paternal associations. Int. J. Epidemiol. 41, 722–732. Howe, L.D., Tilling, K., Galobardes, B., Lawlor, D.A., 2013a. Loss to follow-up in cohort studies: bias in estimates of socioeconomic inequalities. Epidemiology 24, 1–9. Howe, L.D., Tilling, K., Matijasevich, A., Petherick, E.S., Santos, A.C., Fairley, L., Wright, J., Santos, I.S., Barros, A.J.D., Martin, R.M., Kramer, M.S., Bogdanovich, N., Matush, L., Barros, L.H., Lawlor, D.A., 2013b. Linear spline multilevel models for summarising childhood growth trajectories: a guide to their application using examples from five birth cohorts. Stat. Methods Med. Res. (in press) pii:0962280213503925. Kelly, Y., Sacker, A., Gray, R., Kelly, J., Wolke, D., Quigley, M.A., 2009. Light drinking in pregnancy, a risk for behavioural problems and cognitive deficits at 3 years of age? Int. J. Epidemiol. 38, 129–140. Larroque, B., Kaminski, M., 1998. Prenatal alcohol exposure and development at preschool age: main results of a French study. Alcohol. Clin. Exp. Res. 22, 295–303. Leckie, G., Charlton, C., 2013. runmlwin: a Program to run the MLwiN multilevel modeling software from within Stata. J. Stat. Softw. 52, 1–40. Lewis, S.J., Zuccolo, L., Smith, G.D., Macleod, J., Rodriguez, S., Draper, E.S., Barrow, M., Alati, R., Sayal, K., Ring, S., 2012. Fetal alcohol exposure and IQ at age 8: evidence from a population-based birth-cohort study. PLoS ONE 7, e49407. McCarthy, F.P., O’Keeffe, L.M., Khashan, A.S., North, R.A., Poston, L., McCowan, L.M., Baker, P.N., Dekker, G.A., Roberts, C.T., Walker, J.T., Kenny, L.C., 2013.

329

Association between maternal alcohol consumption in early pregnancy and pregnancy outcomes. Obstet. Gynecol. 122, 830–837. Ministry of Health, 2010. Alcohol and Pregnancy: A Practical Guide for Health Professionals. Ministry of Health, Wellington. Mullally, A., Cleary, B.J., Barry, J., Fahey, T.P., Murphy, D.J., 2011. Prevalence, predictors and perinatal outcomes of peri-conceptional alcohol exposure—retrospective cohort study in an urban obstetric population in Ireland. BMC Pregnancy Childbirth 11, 27. National Institute of Health and Care Excellence (NICE), 2015. Antenatal care: routine care for healthy pregnant women. In: NICE Clinical Guidelines 62. Nykjaer, C., Alwan, N.A., Greenwood, D.C., Simpson, N.A., Hay, A.W., White, K.L., Cade, J.E., 2014. Maternal alcohol intake prior to and during pregnancy and risk of adverse birth outcomes: evidence from a British cohort. J. Epidemiol. Comm. Health 68, 542–549. O’Callaghan, F.V., O’Callaghan, M., Najman, J.M., Williams, G.M., Bor, W., 2003. Maternal alcohol consumption during pregnancy and physical outcomes up to 5 years of age: a longitudinal study. Early Hum. Dev. 71, 137–148. O’Keeffe, L.M., Kearney, P.M., Greene, R.A., 2013. Surveillance during pregnancy: methods and response rates to a hospital based cross sectional study of the Pregnancy Risk Assessment Monitoring System in Ireland. In: BMC Pregnancy Childbirth 13. O’Keeffe, L.M., Kearney, P.M., McCarthy, F.P., Khashan, A.S., Greene, R.A., North, R.A., Poston, L., McGowan, L.M., Baker, P.N., Dekker, G.A., Walker, J.J., Taylor, R., Kenny, L.C., 2015. Prevalence and predictors of alcohol use during pregnancy: findings from international multicentre cohort studies. BMJ Open, in press. Oken, E., Levitan, E., Gillman, M., 2007. Maternal smoking during pregnancy and child overweight: systematic review and meta-analysis. Int. J. Obes. 32, 201–210. Passaro, K.T., Little, R.E., Savitz, D.A., Noss, J., 1996. The effect of maternal drinking before conception and in early pregnancy on infant birthweight. Epidemiology 7, 377–383. Patra, J., Bakker, R., Irving, H., Jaddoe, V.W., Malini, S., Rehm, J., 2011. Dose-response relationship between alcohol consumption before and during pregnancy and the risks of low birthweight, preterm birth and small for gestational age (SGA)—a systematic review and meta-analyses. BJOG 118, 1411–1421. Royal College of Obstetricians and Gynecologists, 2006. Alcohol consumption and the outcomes of pregnancy. In: RCOG Statement No 5. Royal College of Obstetricians and Gynecologists. Sampson, P.D., Bookstein, F.L., Barr, H.M., Streissguth, A.P., 1994. Prenatal alcohol exposure, birthweight, and measures of child size from birth to age 14 years. Am. J. Public Health 84, 1421–1428. University of Bristol, 2013. Avon Longitundinal Study of Parents and Children Data Dictionary. University of Bristol, Bristol, http://www.bris.ac.uk/alspac/ researchers/data-access/data-dictionary/ (accessed on 17/07/2013). Whitehead, N., Lipscomb, L., 2003. Patterns of alcohol use before and during pregnancy and the risk of small-for-gestational-age birth. Am. J. Epidemiol. 158, 654–662.