Epidemiology

Original Article Maternal smoking during pregnancy and rapid weight gain from birth to early infancy Tomosa Mine a, Taichiro Tanaka a,*, Tadashi Nakasone b, Toru Itokazu c, Zentaro Yamagata d, Yuji Nishiwaki a aDepartment of Environmental and Occupational Health, School of Medicine, Toho University, Tokyo, JapanbDepartment of Public Health and Medical Care, Hokubu Regional Public Health Center, Okinawa Prefectural Government, Naha, JapancDepartment of Public Health and Medical Care, Health and Longevity Division, Okinawa Prefectural Government, Naha, JapandDepartment of Health Sciences, Basic Science for Clinical Medicine, Division of Medicine, Graduate School Department of Interdisciplinary Research, University of Yamanashi, Chuo, Yamanashi, Japan article info Article history:

Received 24 December 2015 Accepted 19 April 2016 Available online 3 December 2016 Keywords:

Rapid weight gain Pregnancy Smoking abstract Background:Although several studies have focused on the association between maternal smoking during pregnancy and rapid weight gain (RWG) during infancy, the dose-response relationship has not yet been conﬁrmed, and very few studies have included Asian populations. Using a record-linkage method, we examined the association between maternal smoking during pregnancy and RWG in in- fants at around 4 months of age to clarify the dose-response relationship.

Methods:Two databases were used: maternal check-ups during pregnancy and early infancy check-ups (between April 1, 2013 and March 31, 2014 in Okinawa, Japan) were linked via IDs and provided to us after unlinkable anonymizing. For 10,433 subjects (5229 boys and 5204 girls), we calculated the change in infants' weight z-score by subtracting the z-score of their birth weight from their weight at early infancy check-ups. Smoking exposure was categorized intoﬁve groups. We used Poisson regression to examine the association of maternal smoking during pregnancy with RWG in early infancy.

Results:Overall, 1524 (14.6%) were ex-smoker and 511 (4.9%) were current smoker. Compared with the reference category of non-smokers, the adjusted risk ratio of RWG was 1.18 (95% conﬁdence interval [CI], 1. 0 6e1.32) for ex-smokers, 1.18 (95% CI, 0.93e1.50) for those who smoked 1e5 cigarettes per day, 1.57 (95% CI, 1.24e2.00) for those who smoked 6e10 cigarettes per day, and 2.13 (95% CI, 1.51e3.01) for those who smoked 11 cigarettes per day. There was a clear dose-response relationship.

Conclusion:Our study suggests that maternal smoking during pregnancy is associated in a dose- dependent manner with increased risk of RWG in early infancy.

©2016 The Authors. Publishing services by Elsevier B.V. on behalf of The Japan Epidemiological Association. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/). 1. Introduction Maternal smoking during pregnancy is known to restrict intra- uterine growth, leading to low birth weight. Studies have also indicated that maternal smoking during pregnancy can lead to overweight or obesity in infancy and childhood, 1e4 and that rapid weight gain (RWG) in early infancy can increase the risk of car- diovascular disease and type 2 diabetes in early adulthood. 5,6 Onget al. have shown an association between RWG during theﬁrst 2 years of life and subsequent obesity, suggesting that this 2-year postnatal period is an important opportunity to prevent the later development of obesity and metabolic disease. 7 Several studies have focused on the association of maternal smoking during pregnancy and RWG during infancy. 8e10 However, the dose-response relationship, which is important for inferring a causal relationship, has not yet been conﬁrmed. Second, very few recent studies have included Asian populations. Japanese women generally have lower BMI, lower gestational weight gain, and offspring with lower birth weight than Western women 11e14 ; they also tend to smoke less. The association between maternal smoking during pregnancy and RWG would be strengthened if it were also *Corresponding author. Department of Environmental and Occupational Health, School of Medicine, Toho University, 5-21-16, Omori-Nishi, Ota-ku, Tokyo, 143- 8540, Japan.

E-mail address:[email protected](T. Tanaka).

Peer review under responsibility of the Japan Epidemiological Association. Contents lists available atScienceDirect Journal of Epidemiology journal homepage:http://www.journals.elsevier.com/journal-of-epidemiology/ http://dx.doi.org/10.1016/j.je.2016.10.005 0917-5040/©2016 The Authors. Publishing services by Elsevier B.V. on behalf of The Japan Epidemiological Association. This is an open access article under the CCBY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Journal of Epidemiology 27 (2017) 112e116 observed in a Japanese population. Moreover, low birth weight is known to be an independent risk factor for RWG within theﬁrst 2 years of life. 7Therefore, our study tried to assess the direct effect of maternal smoking on RWG in early infancy in a population without low birth weight.

Health check-ups are provided free of charge in Japan for infants at around 4 months of age, and almost all Japanese infants receive them. We can use the unique resulting dataset to investigate the association between maternal smoking and RWG in early infancy in a community setting via cross-referencing with data collected on the mothers at check-ups during pregnancy. Obtaining data during early infancy also helps minimize the effects of lifestyle factors, such as patterns of movement (e.g., crawling and standing), diet, and sleep habits.

The purpose of this study was to use a record-linkage method to examine the association of maternal smoking during pregnancy with RWG in infants at around 4 months of age and to clarify the dose-response relationship.

2. Methods 2.1. Study subjects All 41 municipalities in Okinawa prefecture provide to free four check-ups for pregnant women and infants (at early and late in- fancy, 18 months, and 36 months) using the same standardized questionnaire. Pregnant women received maternal health check- ups at a medical institute, with questionnaires distributed at mu- nicipalities' health centers. Infants received infant check-ups, with questionnaire mailed to each family from municipalities' health centers.

Two databases were used for this study, one including data collected on mothers at check-ups during pregnancy (maternal check-ups) and the other including information on infants receiving check-ups at around 4 months of age (early infancy check- ups). Okinawa Prefecture collected data of both check-ups from 41 municipalities and linked data cross-referenced via IDs in mater- nity record books. Data were provided to us after unlinkable ano- nymizing. Between April 1, 2013 and March 31, 2014, 15,781 (91.6%) of the 17,229 infants entitled to early infancy check-ups in Okinawa Prefecture received them. After the linkage to the maternal check- up database, 12,373 subjects (6162 boys and 6211 girls) were singleton and had complete data at theﬁrst (around 8e11 weeks of gestation) and the last (around 34e37 weeks of gestation) maternal check-ups. Further, we excluded 1940 subjects falling into one or more of the following categories: preterm gestational age<37 weeks (n¼443), birth weight<2500 g (n¼695) or 4000 g (n¼79), no information about maternal smoking status during pregnancy or anthropometric data (n¼303), and age at check-ups <3 months or>5 months (n¼420). Analyses were carried out on the remaining 10,433 subjects (5229 boys and 5204 girls).

2.2. Information on maternal smoking during pregnancy The mothers were asked about maternal smoking status during pregnancy using a self-reported questionnaire at theirﬁrst maternal check-ups. This questionnaire included questions about their smoking status before and during pregnancy (current or not) and the number of cigarettes they smoked per day. Those mothers who had smoked before but not during pregnancy were classiﬁed as ex-smokers. Current smokers during pregnancy were classiﬁed according to the number of cigarettes smoked daily, in multiples of ﬁve: 1e5, 6e10, and 11 cigarettes per day (there were only three groups because few subjects smoked 16 cigarettes per day).2.3. Rapid weight gain Data on the subjects' weight in early infancy (age 3e5 months) were collected at the early infancy check-ups. Data on the subjects' birthweight and gestational age were transcribed from maternal record books by nurses or public health nurses at the same check- ups. We calculated weight z-score (observed valueemean/stan- dard deviation [SD]) at birth. Calculating weight z-score at birth, Ong et al. adjusted for gestational age and sex, 7and Project Koshu adjusted only for sex, without adjusting for gestational age. 15We used the latter method, though both gave us a similar results in this study. We calculated weight z-score at early infancy according to sex and month of age. The mean and SD were based on the available data on all of the infants receiving early infancy check-ups in the study period (n¼15,781), according to Project Koshu. 15The weight distribution in our data was similar to that found in Japan's 2000 survey on growth of infants and preschool children 16and was not skewed, so we calculated direct z-scores. RWG in early infancy was assessed on the basis of changes in weight z-score from birth to early infancy, with a change exceedingþ0.67 deﬁned as RWG.

According to Ong et al., a z-score of 0.67 represent the width of each percentile band on standard growth charts (2nd, 10th, 25th, 50th, 75th, 90th, and 98th percentile lines); thus, a weight gain exceedingþ0.67 SD indicates an increase across at least one of these percentile bands. 7 2.4. Covariates Information was collected at the maternal check-ups on age, pre-pregnancy body mass index (BMI; weight [kg]/length [m] 2), and weight gain during pregnancy (subtracting the pre-pregnancy weight from the weight at the last maternal check-up). At the early infancy check-ups, information was collected on sex, gestational age at birth, birth order, birth weight, and paternal smoking during pregnancy. We also collected information on how the infants had been fed between birth and the early infancy check-ups and divided them into three feeding categories: exclusive breast-feeding, mixed feeding, and exclusive bottle-feeding.

2.5. Statistical analysis We used Stata 12 for Windows (STATA Corporation, College Station, TX, USA) software for statistical analysis, and we used Poisson regression analysis to estimate crude and adjusted risk ratios (RRs) and 95% conﬁdence intervals (CIs) for RWG during early infancy according to maternal smoking status during pregnancy.

We found no statistical interaction according to infant sex (p¼0.44), so we constructed combined models for both sexes. Our multivariable model included potential confounders identiﬁed in previous studies: infants' sex, birth order (ﬁrst, second, or later), maternal age ( 26 years [25th percentile], 27e33 years [25th-75th percentile], or 34 years [75th percentile]), paternal smoking during pregnancy (current or not), gestational age (continuous), and maternal pre-pregnancy BMI (<18.5, 18.5e22.9, 23e24.9, or 25 kg/m 2). The inclusion of age in days from birth to infancy check-ups in the multivariable model did not alter the results substantially, so we did not adopt this factor as a covariate. This multivariable model was further adjusted individually for type of feeding between birth and the early infancy check-ups (exclusive breast-feeding, mixed feeding, or exclusive bottle-feeding), maternal weight gain (<5, 5e6.9, 7e8.9, 9e10.9, 11e12.9, or 13 kg), and birth weight (continuous). A previous study had found interaction between maternal smoking and type of feeding be- tween birth and early infancy check-ups in the risk of RWG, 17but we found no interaction (p¼0.39). Therefore, we treated feeding as T. Mine et al. / Journal of Epidemiology 27 (2017) 112e116113 a covariate. For trend analysis, Poisson regression analysis was repeated, with smoking categories given integer values.

2.6. Ethical considerations All data provided by Okinawa Prefecture were anonymized in an unlinkable fashion, and the researchers were blinded to personal information about the mothers and infants. This study was approved by the Ethics Committee of Toho University (approval number: 25124).

3. Results Table 1shows the maternal smoking status during pregnancy of the study subjects. In total, 8398 (80.5%) were non-smokers, 1524 (14.6%) were ex-smokers, and 511 (4.9%) reported smoking during pregnancy. A majority (52.8%) of the current smokers smoked 1e5 cigarettes per day. The current smokers tended to be younger, to be overweight, to gain less weight during pregnancy, and to have husbands who smoked during pregnancy. The infants whose mothers smoked were likely to be second or later children and to have received mixed feeding.

Table 2shows the anthropometric data of the infants by sex and maternal smoking status. In both sexes, average birth weight and z- score at birth tended to decrease as the number of cigarettes smoked by their mothers increased. In contrast, average weight and z-score during early infancy tended to increase in line with the number of cigarettes smoked by the mothers. Thus, changes in weight z-score and the incidence of RWG in early infancy increased in line with increased maternal smoking: around 20% of the infantswith non-smoking mothers experienced RWG, whereas the pro- portion was almost 40% among infants whose mothers smoked 11 cigarettes per day.

Table 3shows the dose-response relationship between the smoking category (based on the number of cigarettes smoked by mothers during pregnancy) and the risk of RWG in early infancy, as determined by Poisson regression analysis. Compared with non- smokers, the multivariable-adjusted RRs for RWG were 1.18 (95% CI, 1.06e1.32) for ex-smokers, 1.18 (95% CI, 0.93e1.50) for those who smoked 1e5 cigarettes per day, 1.57 (95% CI, 1.24e2.00) for those who smoked 6e10 cigarettes per day, and 2.13 (95% CI, 1.51e3.01) for those who smoked 11 cigarettes per day. Further adjustment for type of feeding between birth and early infancy, and for maternal weight gain during pregnancy, did not change the RRs substantially (Table 3). In contrast, adjustment for birth weight attenuated the RRs, although a statistically signiﬁcant association between maternal smoking status and RWG remained. When smoking categories were treated as integer values, upward trends across smoking statuses were observed (p for trend<0.05 for each model).

In order to minimize the inﬂuence of the difference in days between birth and early infancy check-ups, we limited to infants who received check-ups at 4 months of age and repeated the analysis, with similar results (eTable 1).

4. Discussion This study showed an increased risk of RWG in early infancy for infants exposed to maternal smoking during pregnancy, and a clear dose-response relationship was observed between the number of Table 1 Characteristics of the study population by maternal smoking during pregnancy.

Current-smoker (number of cigarettes per day) Non-smoker (n¼8398) n (%) a Ex-smoker (n¼1524) n (%) a 1-5 cigarettes/day (n¼270) n (%) a 6e10 cigarettes/day (n¼189) n (%) a 11 cigarettes/day (n¼52) n (%) a p-value b Parental Maternal age, years<0.001 26 (<25th percentile) 1820 (21.7) 645 (42.3) 122 (45.2) 64 (33.9) 12 (23.1) 27-33 (25-75th percentile) 3843 (45.8) 556 (36.5) 103 (38.2) 82 (43.4) 32 (61.5) 34 (>75th percentile) 2731 (32.5) 323 (21.2) 45 (16.7) 43 (22.8) 8 (15.4) Maternal BMI<0.001 <18.5 kg/m 1240 (14.8) 239 (15.7) 53 (19.6) 24 (12.8) 7 (13.5) 18.5e22.9 kg/m 5266 (62.8) 865 (57.0) 139 (51.5) 102 (54.3) 22 (42.3) 23e24.9 kg/m 929 (11.1) 187 (12.3) 30 (11.1) 23 (12.2) 5 (9.6) 25 kg/m 947 (11.3) 228 (15.1) 48 (17.8) 29 (20.7) 18 (34.6) Weight gain during pregnancy<0.001 4.9 kg 913 (10.9) 124 (8.1) 25 (9.2) 32 (14.9) 14 (26.7) 5e6.9 kg 1133 (13.5) 133 (8.7) 19 (7.0) 18 (9.5) 8 (15.4) 7e8.9 kg 2032 (24.2) 254 (16.7) 53 (19.6) 46 (24.3) 12 (23.1) 9e10.9 kg 2106 (25.1) 283 (18.6) 56 (20.7) 31 (16.4) 8 (15.4) 11e12.9 kg 1310 (15.6) 279 (18.3) 58 (21.5) 30 (15.9) 5 (9.6) 13 kg 904 (10.8) 451 (29.6) 59 (21.9) 32 (16.9) 5 (9.7) Paternal smoking during pregnancy 2551 (33.1) 945 (64.2) 210 (80.2) 151 (83.0) 37 (74.0)<0.001 Infancy Sex (boy) 4223 (50.3) 738 (48.4) 145 (53.7) 90 (47.6) 33 (64.5) 0.12 Gestational age c 39.0 (1.1) 39.1 (1.1) 39.1 (1.1) 39.0 (1.1) 38.8 (1.2) 0.18 Birth order<0.001 First 2913 (34.9) 720 (47.5) 119 (44.4) 52 (28.8) 15 (28.9) Second or later 5430 (65.1) 796 (52.5) 149 (55.6) 131 (71.25) 37 (71.2) Feeding between birth and early infancy checkups<0.001 Exclusive breast-feeding 3011 (36.1) 392 (25.9) 67 (25.0) 34 (18.2) 16 (31.4) Mixed-feeding 2923 (35.0) 705 (46.7) 141 (52.6) 115 (61.5) 27 (53.0) Exclusive bottled feeding 2419 (29.0) 414 (27.4) 60 (22.4) 38 (20.3) 8 (15.7) BMI, body mass index.

Due to missing values, the totals for the stratiﬁed subgroups are not equal. aColumn %.bp-value for c2test or Analysis of variance or Fisher's exact test. cMean (standard deviation). T. Mine et al. / Journal of Epidemiology 27 (2017) 112e11 6 114 cigarettes smoked and RWG risk. The association remained after we adjusted for the type of feeding, maternal weight gain during pregnancy, and birth weight, so our results suggest that the asso- ciation of maternal smoking status during pregnancy with RWG in early infancy is independent of these three factors.

Previous studies on the association between maternal smoking during pregnancy and weight gain in infancy have shown dispa- rate results, perhaps largely because of differences in the study periods and outcome indicators. 8,9,18 However, most studies focusing on weight gain in early infancy have shown a positive association between maternal smoking during pregnancy and weight gain. In a cohort study of neonates carried out in Norway in 1992 and 1993, Nafstad et al. used weight gain during theﬁrst year as an outcome indicator and found that infants born to smoking mothers gained more weight, in a dose-dependent manner, than those born to non-smoking mothers. 9They also investigated breast-feeding during theﬁrst 6 months and sug- gested that the observed weight gain could be partially explained by the shorter duration of breast-feeding by smoking mothers. In Canada, Dubois observed infants born in 1998 from birth to 5months of age and found that those born to smoking mothers gained more weight than those born to non-smoking mothers. 18 Harrod et al. used air-displacement plethysmography to measure changes in the body composition of infants born in the United States between July 2010 and November 2013 from birth to 5 months. 19They found that fat-free mass at 5 months was signiﬁ- cantly greater in infants exposed to prenatal smoking than in those not exposed, but that the mean change in weight-for-length z-score of offspring between birth and 5 months was not signiﬁ- cantly different.

Two other studies 20,21 investigated the association between maternal smoking during pregnancy and RWG in infancy using the same deﬁnition of RWG by Ong et al. used in our study. Wijlaars et al. analyzed 2402 infant/parent pairs in England and reported a positive association between maternal smoking during pregnancy and RWG at 3 months of age.

20With non-smokers as the reference category, the odds ratio (OR) of RWG in infants of mothers who smoked during pregnancy was 1.34 (95% CI, 1.00e1.78). However, they did not investigate the dose-response relationship. Lyte et al.

studied 11,134 children and parents in Ireland to investigate the Table 2 Anthropometric data by sex and maternal smoking during pregnancy.

Current smoker Non-smoker Mean (SD)Ex-smoker Mean (SD)1-5 cigarettes/day Mean (SD)6-10 cigarettes/day Mean (SD) 11 cigarettes/day Mean (SD) Total (n¼10,433)8398 1524 270 189 52 Birth weight 3083.1 (317.4) 3076.1 (312.5) 3053.4 (298.6) 3014.9 (306.6) 2947.4 (251.3) z-score at birth 0.21 (0.76) 0.20 (0.75) 0.14 (0.71) 0.05 (0.73) 0.11 (0.60) Weight at early infancy 6960.7 (802.0) 7052.0 (823.9) 7048.1 (818.3) 7165.8 (921.8) 7147.8 (683.9) z-score at early infancy 0.04 (0.94) 0.15 (0.96) 0.12 (0.97) 0.23 (1.07) 0.26 (0.84) Change in weight z-score from birth to early infancy 0.17 (0.94) 0.05 (0.93) 0.02 (1.00) 0.18 (0.99) 0.38 (0.94) Rapid weight gain (%) 1527 (18.8) 335 (22.0) 57 (21.1) 54 (28.6) 20 (38.5) Boys (n¼5229)4223 738 145 90 33 Birth weight 3114.7 (315.2) 3109.3 (311.2) 3082.4 (316.5) 3067.0 (299.1) 2923.6 (207.8) z-score at birth 0.19 (0.75) 0.18 (0.73) 0.12 (0.75) 0.09 (0.71) 0.26 (0.49) Weight at early infancy 7208.8 (780.8) 7296.1 (801.3) 7213.8 (769.7) 7418.2 (882.6) 7276.5 (770.9) z-score at early infancy 0.03 (0.94) 0.14 (0.97) 0.01 (0.95) 0.23 (1.06) 0.14 (0.93) Change in weight z-score from birth to early infancy 0.16 (0.94) 0.04 (0.93) 0.12 (0.98) 0.14 (1.00) 0.41 (0.84) Rapid weight gain (%) 783 (18.5) 166 (22.5) 25 (17.2) 24 (26.7) 13 (39.4) Girls (n¼5204)4175 786 125 99 19 Birth weight 3051.2 (316.4) 3044.8 (310.6) 3019.9 (273.8) 2964.8 (306.3) 2988.6 (315.1) z-score at birth 0.24 (0.77) 0.22 (0.76) 0.16 (0.66) 0.28 (0.75) 0.09 (0.76) Weight at early infancy 6709.7 (743.0) 6822.9 (778.1) 6855.8 (833.9) 6936.3 (900.4) 6924.5 (431.0) z-score at early infancy 0.06 (0.94) 0.20 (0.97) 0.25 (1.05) 0.29 (1.10) 0.26 (0.61) Change in weight z-score from birth to early infancy 0.18 (0.95) 0.03 (1.09) 0.09 (1.09) 0.26 (1.03) 0.18 (1.03) Rapid weight gain (%) 744 (17.8) 169 (21.5) 32 (25.6) 30 (30.3) 7 (36.8) SD, standard deviation.

Table 3 Risk ratios of rapid weight gain at early infancy by maternal smoking during pregnancy, With calculated weight z-scores at birth adjusted for sex and gestational age.

Further adjusted by n/N (%)Crude risk ratio (95% CI)Multivariable risk ratio a (95% CI)Feeding between birth and early infancy checkups (95% CI)Maternal weight gain during pregnancy (95% CI)Birth weight (95% CI) Non-smoker 1527/8398 (18.2) ref ref ref ref ref Ex-smoker 335/1524 (22.0) 1.21 (1.09e1.34) 1.18 (1.06e1.32) 1.18 (1.05e1.32) 1.19 (1.06e1.34) 1.17 (1.05e1.32) Current-smoker 1e5 cigarettes/day 57/270 (21.1) 1.16 (0.91e1.47) 1.18 (0.93e1.50) 1.17 (0.92e1.48) 1.18 (0.93e1.51) 1.14 (0.89e1.44) 6e10 cigarettes/day 54/189 (28.6) 1.57 (1.25e1.98) 1.57 (1.24e2.00) 1.55 (1.22e1.98) 1.58 (1.24e2.01) 1.41 (1.10e1.80) 11 cigarettes/day 20/52 (38.5) 2.12 (1.50e2.99) 2.13 (1.51e3.01) 2.05 (1.43e2.93) 2.13 (1.51e3.01) 1.80 (1.30e2.50) Risk ratio for one category increase 1.17 (1.11e1.23) 1.15 (1.10e1.21) 1.16 (1.10e1.23) 1.16 (1.09e1.22) 1.13 (1.07e1.19) CI, conﬁdence interval. aAdjusted for infant' sex, birth order, maternal age, paternal smoking during pregnancy, gestational age, and pre-pregnant BMI. T. Mine et al. / Journal of Epidemiology 27 (2017) 112e116115 risk of RWG at 9 months, 21and their results suggested a positive association between prenatal exposure to smoking and RWG, with a dose-response relationship. The OR of RWG at 9 months of age for the children of mothers who smoked heavily ( 11cigarettes daily), after adjusting for birth weight and gestation, was 1.85 (95% CI not reported). In our study, the relative risk of RWG, after adjusting for birth weight, in infants whose mothers smoked 11 cigarettes per day was 1.80, which was close to Lyte'sﬁgure.

The mechanisms of the link between maternal smoking during pregnancy and RWG have not been fully clariﬁed, but several hy- potheses can be put forward. First, nicotine, which is transported across the placenta, may explain the physiologic effects of smoking during pregnancy. 22In an animal study, rats prenatally exposed to low doses of nicotine were not smaller at birth but had higher fat levels. 23It has also been suggested that maternal smoking during pregnancy may be associated with lower cord blood leptin, 22and that a low concentration of this hormone at birth may provide a signal for catch-up growth by inhibiting satiety. 7Thus, it is conceivable that the mechanisms linking prenatal smoking with RWG bypass birth weight and type of feeding.

The main strength of our study is that we had access to a large population-based dataset rather than a hospital-based dataset.

Second, we collected information about maternal smoking status before and during pregnancy, and we had information on number of cigarettes smoked per day during pregnancy. Thus, we could observe increased risk of RWG in infants born to ex-smokers and light smokers (1e5 cigarettes per day), which allowed us to assess the dose-response relationship. Also, because we were able to examine the infants in early infancy (at 3e5 months of age), the inﬂuence of lifestyle factors was minimized.

The main limitation of our study is that maternal smoking status was assessed using self-report. Because of the self-report format of our study, the subjects in the ex-smoker group may actually have included some current smokers. Tong et al. reported that 10% of woman who claimed to have stopped smoking showed biochemical evidence of continued smoking. 24 It is possible that we over- estimated the risk of RWG in the ex-smoker group, and that smoking cessation during pregnancy may have a greater effect on RWG in early infancy than the study results indicate. Next, a recent study on the relationship between the socioeconomic status of mothers and RWG in their infants is worthy of attention. 25We did not have access to information on socio-economic status, such as maternal education, household income, and occupation. Further studies that include information on socio-economic status will be necessary. Last, our subjects were from Okinawa Prefecture, which has a distinct culture from the rest of Japan, including food and lifestyle. Therefore, our subjects might not be representative of pregnant Japanese woman.

In conclusion, our results from a large population-based dataset clearly show that maternal smoking during pregnancy results in dose-dependent increases in the risk of RWG in their offspring in early infancy. Thisﬁnding is signiﬁcant because previous studies of maternal smoking status and RWG in infants have not included Asian subjects, who differ physically from Western subjects.

Conﬂicts of interest None declared.

Acknowledgements This work was supported by the Ministry of Health, Labour and Welfare of Japan, Jisedai-Ippan (H25-002,http://www.mhlw.go.jp/seisakunitsuite/bunya/hokabunya/kenkyujigyou/hojokin-koubo- h25/gaiyo/08.html)toZY.

Appendix A. Supplementary data Supplementary data related to this article can be found athttp:// dx.doi.org/10.1016/j.je.2016.10.005.

References 1.Harris HR, Willett WC, Michels KB. Parental smoking during pregnancy and risk of overweight and obesity in the daughter.Int J Obes (Lond). 2013;37:

1356e1363.

2.Raum E, Kupper-Nybelen J, Lamerz A, Hebebrand J, Herpertz-Dahlmann B, Brenner H. Tobacco smoke exposure before, during, and after pregnancy and risk of overweight at age 6.Obes (Silver Spring). 2011;19:2411e2417.

3.Florath I, Kohler M, Weck MN, et al. Association of pre- and post-natal parental smoking with offspring body mass index: an 8-year follow-up of a birth cohort.

Pediatr Obes. 2014;9:121e134.

4.Riedel C, Schonberger K, Yang S, et al. Parental smoking and childhood obesity:

higher effect estimates for maternal smoking in pregnancy compared with paternal smokingea meta-analysis.Int J Epidemiol. 2014;43:1593e1606.

5.Leunissen RW, Kerkhof GF, Stijnen T, Hokken-Koelega A. Timing and tempo of ﬁrst-year rapid growth in relation to cardiovascular and metabolic risk proﬁle in early adulthood.JAMA. 2009;301:2234e2242.

6.Ekelund U, Ong KK, Linne Y, et al. Association of weight gain in infancy and early childhood with metabolic risk in young adults.J Clin Endocrinol Metab.

2007;92:98e103.

7.Ong KK, Ahmed ML, Emmett PM, Preece MA, Dunger DB. Association between postnatal catch-up growth and obesity in childhood: prospective cohort study.

BMJ. 2000;320:967e971.

8.Conter V, Cortinovis I, Rogari P, Riva L. Weight growth in infants born to mothers who smoked during pregnancy.BMJ. 1995;310:768e771.

9.Nafstad P, Jaakkola JJ, Hagen JA, et al. Weight gain during theﬁrst year of life in relation to maternal smoking and breast feeding in Norway.J Epidemiol Com- munity Health. 1997;51:261e265.

10.Suzuki K, Kondo N, Sato M, Tanaka T, Ando D, Yamagata Z. Maternal smoking during pregnancy and childhood growth trajectory: a random effects regres- sion analysis.J Epidemiol. 2012;22:175e178.

11.WHO Expert Consultation. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies.Lancet. 2004;363:

157e163.

12.Harita N, Kariya M, Hayashi T, et al. Gestational bodyweight gain among un- derweight Japanese women related to small-for-gestational-age birth.J Obstet Gynaecol Res. 2012;38:1137e1144.

13.Itabashi K, Miura F, Uehara R, Nakamura Y. New Japanese neonatal anthro- pometric charts for gestational age at birth.Pediatr Int. 2014;56:702e708.

14.Kozuki N, Katz J, Christian P, et al. Comparison of US birth weight references and the International Fetal and Newborn Growth Consortium for the 21st Century Standard.JAMA Pediatr. 2015;169:e151438.

15.Zheng W, Suzuki K, Shinohara R, Sato M, Yokomichi H, Yamagata Z. Maternal smoking during pregnancy and growth in infancy: a covariance structure analysis.J Epidemiol. 2015;25:44e49.

16. National growth survey on preschool children [Home page on the internet].

Tokyo: Ministry of Health, Labour and Welfere. Available from::http://www.

mhlw.go.jp/houdou/0110/h1024-4.html(in Japanese).

17.Wen X, Shenassa ED, Paradis AD. Maternal smoking, breastfeeding, and risk of childhood overweight:ﬁndings from a national cohort.Matern Child Health J.

2013;17:746e755.

18.Dubois L, Girard M. Early determinants of overweight at 4.5 years in a population-based longitudinal study.Int J Obes (Lond). 2006;30:610e617.

19.Harrod CS, Fingerlin TE, Chasan-Taber L, Reynolds RM, Glueck DH, Dabelea D.

Exposure to prenatal smoking and early-life body composition: the healthy start study.Obes (Silver Spring). 2015;23:234e241.

20.Wijlaars LP, Johnson L, van Jaarsveld CH, Wardle J. Socioeconomic status and weight gain in early infancy.Int J Obes (Lond). 2011;35:963e970.

21.Layte R, Bennett A, McCrory C, Kearney J. Social class variation in the predictors of rapid growth in infancy and obesity at age 3 years.Int J Obes (Lond). 2014;38:

82e90.

22.Oken E, Levitan EB, Gillman MW. Maternal smoking during pregnancy and child overweight: systematic review and meta-analysis.Int J Obes (Lond).

2008;32:201e210.

23.Chen WJ, Kelly RB. Effect of prenatal or perinatal nicotine exposure on neonatal thyroid status and offspring growth in rats.Life Sci. 2005;76:1249e1258.

24.Tong VT, Althabe F, Aleman A, et al. Accuracy of self-reported smoking cessa- tion during pregnancy.Acta Obstet Gynecol Scand. 2015;94:106e111.

25.Grifﬁths LJ, Hawkins SS, Cole TJ, Dezateux C. Millennium Cohort Study Child Health Group. Risk factors for rapid weight gain in preschool children:ﬁndings from a UK-wide prospective study.Int J Obes (Lond). 2010;34:624e632. T. Mine et al. / Journal of Epidemiology 27 (2017) 112e11 6 116