The anthropometry of children and adolescents may be influenced by the prenatal smoking habits of their grandmothers: a longitudinal cohort study.

OBJECTIVES: Previously, in the Avon Longitudinal Study of Parents and Children (ALSPAC), we have shown different sex-specific birth anthropometric measurements contingent upon whether or not prenatal smoking was undertaken by paternal grandmother (PGM±), maternal grandmother (MGM±), and the study mother (M±). The findings raised the question as to whether there were long-term associations on the growth of the study children over time.
METHODS: Measures of weight, height, body mass index, waist circumference, lean mass, and fat mass of children in the ALSPAC study from 7 to 17 years of age were used. We compared growth in four categories at each age: PGM+M- with PGM-M-; MGM+M- with MGM-M-; PGM+M+ with PGM-M+; MGM+M+ with MGM-M+; and adjusted for housing tenure, maternal education, parity, and paternal smoking at the start of the study pregnancy.
RESULTS: We found that if the PGM had, but the study mother had not, smoked in pregnancy, the girls were taller and both genders had greater bone and lean mass. However, if the MGM had smoked prenatally but the mother had not (MGM+M-), the boys became heavier than expected with increasing age-an association that was particularly due to lean rather than fat mass, reflected in increased strength and fitness. When both the maternal grandmother and the mother had smoked (MGM+M+) girls had reduced height, weight, and fat/lean/bone mass when compared with girls born to smoking mothers whose own mothers had not smoked (MGM-M+).
CONCLUSIONS: This study indicates that smoking in humans can have sex-specific transgenerational effects.

Our program of research into transgenerational effects of cigarette smoking (Miller et al., 2014; Northstone et al., 2014) was instigated as a result of studies from Sweden. These were based on samples of individuals born close to the Arctic Circle in the town of Överkalix. Their longevity and other health outcomes were linked to detailed historical records of harvests experienced by their ancestors (Bygren et al., 2001). Using three independent birth cohorts in the years 1890, 1905, and 1920, Kaati et al. (2002) showed that the paternal grandfathers' plentiful food supply in mid-childhood was associated with a fourfold increased chance of diabetes on the grandchild's death certificate [95% CI 1.3, 12.9]. Their study also showed that cardiovascular mortality in the study individuals was reduced when there had been poor food supply in the father's mid-childhood. Subsequently, sex-specific analysis of the data showed that the mortality rate of the men born in the target years was linked to their paternal grandfather's food supply in mid-childhood, whereas the mortality rate of the women studied was associated solely with their paternal grandmother's food supply (Pembrey et al., 2006). This association was shown in two of three independent cohorts. Exposure sensitive periods involved both paternal grandparents' mid-childhood but also the fetal/infant period for the paternal grandmothers.

In the UK, since the Second World War, there have been no particular years of starvation or glut. In the search for an environmental feature that we could time in regard to the age of exposure at which it occurred, we have chosen the smoking habits of the individual parents. It is well recognized that smoking has strong effects on various physiological systems, and results in a loss of appetite and general reduction in weight compared with nonsmokers (Chiolero et al., 2008). Previously, using the Avon Longitudinal Study of Parents and Children (ALSPAC), we have shown that fathers who started smoking regularly between the ages of 8 and 11 had boys (but not girls) with increased body mass index (BMI), waist circumference, and body fat mass as teenagers (Northstone et al., 2014). We have also shown that nonsmoking mothers exposed prenatally to their own mothers' smoking delivered children who were larger at birth (Miller et al., 2014). After adjustment, the average birth weight, birth length, and BMI measurements of the boys (but not the girls) were greater if the maternal grandmother smoked prenatally: birth weight = +61 [95% CI +30, +92] g; birth length = +0·19 [95% CI +0·02, +0·35] cm; birth BMI = +1·6 [95% CI +0·6, +2·6] g/m2. In a parallel paper (Pembrey et al., 2014), we have shown that exposure of the father to his mother's smoking resulted in a reduction in birth head circumference of his sons if the study mother also smoked in pregnancy, and that this was reflected in reduced IQ in this group. Here, we examine the growth of these children from ages 7 to 17 to determine whether prenatal smoke exposure of either parent is associated with the growth of the offspring, including body composition, and whether it is sex-specific and/or depends on whether the study mother smoked in pregnancy.

MATERIALS AND METHODS

Study samples

The data used in these analyses were collected as part of the ALSPAC, which was designed to assess the ways in which the environment interacts with the genotype to influence health and development (Golding, 2004). Pregnant women resident in the study area in south-west England with an expected date of delivery between 1st April, 1991 and 31st December, 1992 were invited to take part. About 80% of the eligible population did so (Boyd et al., 2013). The initial ALSPAC sample consisted of 14,541 pregnancies; of these, 14,472 had known birth outcomes: 14,062 were live births and 13,988 were alive at 1 year.

Information collected from the parents during their study pregnancy included details of the maternal and paternal grandparents. The two pathways of possible influence of parental prenatal exposure to cigarette smoke on the study child that we will investigate in this article, concern (a) via the maternal grandmother (MGM) to the mother (M) in utero to her study fetus, and (b) via the paternal grandmother (PGM) to the study father (F) while he was in utero and thence to the study conceptus (Fig. 1).

Figure 1

Developmental pedigrees illustrating the prenatal smoke exposures studied. (*all analyses adjusted for paternal smoking; (*) smoking of other grandmother adjusted for in sensitivity analysis)

The exposures

The pregnant study mothers and their partners were sent six questionnaires during pregnancy (ALSPAC, 2014). These elicited information on their current smoking habits and those of their parents (i.e., the study grandparents). If they reported that their mothers had smoked, they were asked whether their mothers had smoked when expecting them—and, if so, were given the responses yes/no/do not know from which to select. Thus, the parents who replied “do not know,” had a mother who smoked but the parent was unsure whether she had smoked during her pregnancy. We have analyzed these data assuming that these women did smoke during pregnancy.

Possible confounders

Potential confounders included in the analyses were the study mother's parity (as ascertained from the maternal report of previous pregnancies resulting in either a live- or still-birth, and coded as 0; 1+); mother's partner smoking during the pregnancy (primarily reported by partner, but maternal report was used if partner report was missing: yes; no); housing tenure as a measure of socioeconomic background (owned or mortgaged; rented public housing; all other), and maternal education (highest level of educational attainment—in five levels of increasing achievement).

Outcomes

Children were measured using standardized methods by the ALSPAC study team in a clinic setting from the age of 7 and every other year thereafter until the age of 17.

Height was measured using the Harpenden Stadiometer (Crymych, UK): shoes were removed, the study child stood with feet flat, so that the under-side of the heels was in contact with the ground. The heels were placed together, so that the medial malleoli were touching (unless the child had knock knees). The child stood straight so that heels, calves, buttocks, and shoulders were in contact with the vertical backboard of the Stadiometer. Shoulders were relaxed and sloping forward in a natural position, hands and arms were loose and relaxed with palms facing medially. The headboard was slid down the backboard until it touched the study child's head. To ensure that the head stayed in contact with the headboard and to minimize the effect of hair thickness, a 1 kg weight was placed on the headboard. The height was recorded to the last completed millimeter.

Waist circumference was measured as the minimum circumference of the abdomen between the iliac crests and the lowest ribs, with the tape perpendicular to the long axis of the body touching the skin but not compressing the tissue. It was measured to the last complete millimeter.

Weight was measured using Tanita scales Body Fat Analyzer model TBF 305 (Arlington Heights, IL). The child was encouraged to pass urine and undress to their underclothes. BMI was calculated as weight (kg) (height (m))2. Total body fat, lean, and bone mass were measured bi-annually from the age of 9 using total-body dual-energy X-ray absorptiometry scans, performed using a Lunar Prodigy dual-energy X-ray absorptiometer (GE Medical Systems Lunar, Madison, WI) (Toschke et al., 2007). Bone mass, lean mass, and fat mass were estimated at each age.

Grip strength was assessed at age 11 using the Jamar hand dynamometer, which measures isometric strength in kilograms. The child sat in a chair with arms and back support and was asked to rest his/her forearms on the arms of the chair with their wrist just over the end of the arm of the chair. The wrist was placed in a neutral position with the thumb facing upwards. The tester demonstrated how to use the dynamometer to the child showing how gripping very tightly registered the best score. The child was given a practice squeeze of the dynamometer to ensure that it felt comfortable. Starting with the right hand, the hand was positioned so that the thumb was round one side of the handle and the four fingers were around the other side. It was important that the instrument felt comfortable for the child and the position of the handle was altered if necessary. The measurer rested the base of the dynamometer on the palm of the child's hand in order to support the weight of the dynamometer, whilst ensuring that the movement of the machine was not restricted. The child was encourage to squeeze as long and as tightly as possible or until the needle stopped rising: the higher the reading, the stronger the grip. The grip strength was measured twice in each hand and the mean of the 4 measurements was used.

Cardio-respiratory fitness

Physical work capacity (Watts) was assessed at a heart rate of 170 bpm (PWC170). This was estimated using standard regression methods from parameters measured using an electronically braked cycle ergometer (Lawlor et al., 2008).

Statistical analyses

Multivariable linear regression models assessed the grandchildren's mean height, weight, BMI, waist circumference, fat mass, lean mass, and bone mass in regard to the parental prenatal smoking exposures. All models were adjusted for parity, maternal education, paternal smoking at the start of pregnancy, and housing tenure. Because maternal prenatal smoking itself is associated with overweight in the offspring (Oken et al., 2007; Ino, 2010), we have analyzed separately the children whose mothers themselves smoked during pregnancy. In line with the evidence in the literature that various effects of cigarette smoking are sex specific (e.g., Zaren et al., 2000), together with the results from our earlier studies (Miller et al., 2014; Northstone et al., 2014; Pembrey et al., 2006), we have analyzed the male and female offspring separately and, where appropriate, have tested for interactions with sex.

RESULTS

Response

The numbers of study children attending for examination at each of the time-points are shown in Table1. It can be seen that there is a steady decline in attendance from 8,290 at age 7 to 5,217 at age 17. However, there was no bias over time in the proportion of the children attending for whom data were available on grandmaternal prenatal smoking—this varied from 88.2% to 89.5% for the MGM history, and from 72.3% to 74.3% for that of the PGM.

Table 1

Attendance at the clinics at which anthropometric measurements were made, and proportions with data on the smoking of the maternal and paternal grandmothers in utero

Age at focus clinic	No. attending	No.(%) with information on MGM	No. (%) with information on PGM
7 Years	8,290	7,352 (88.7%)	5,994 (72.3%)
9 Years	7,722	6,869 (89.0%)	5,602 (72.5%)
11 Years	7,153	6,395 (89.4%)	5,221 (73.0%)
13 Years	6,147	5,504 (89.5%)	4,544 (73.9%)
15 Years	5,515	4,931 (89.4%)	4,096 (74.3%)
17 Years	5,217	4,601 (88.2%)	3,795 (72.7%)

MGM, maternal grandmother; PGM, paternal grandmother.

Anthropometric measures

The results of comparing the anthropometric measures between the children whose grandmothers had smoked while one of their parents was in utero are shown in Supporting Information Tables 1–4 and summarized below.

Child's height

There were no apparent differences in height associated with the mother's prenatal exposure, unless she smoked herself. In the latter scenario [MGM+M+ vs. MGM−M+], her girls were consistently of lower height than expected, ranging from 0.9 to 1.8 cm lower (Supporting Information Table 4).

In contrast, there was a consistency in regard to the father's prenatal exposure—the study children, especially the girls, were taller than expected provided their own mother did not smoke [PGM+M−]. The excess adjusted height varied from 0.2 to 0.7 cm for boys and 0.4 to 0.7 cm for girls (Supporting Information Table 1).

Child's weight, BMI, and waist circumference

There were interesting differences in weight in the children of the nonsmoking mother according to whether she or her partner was exposed in utero. The adjusted differences are shown in Figure 2. The increased child weight with the maternal grandmother smoking [MGM+M− vs. MGM−M−] was apparent for just the boys, whereas that with the paternal grandmother smoking when pregnant [PGM+M− vs. PGM−M] resulted in increased weight in both boys and girls during adolescence (Supporting Information Tables 1 and 2).

Figure 2

Weight of offspring of nonsmoking women showing the difference (kg) between those whose grandmothers smoked prenatally compared with those who did not (MGM, maternal grandmother; PGM, paternal grandmother; M, mother; +, smoked prenatally; −, did not smoke prenatally).

Similarly among nonsmoking women, there were positive associations with BMI which increased with age among both genders for the paternal grandmother smoking in pregnancy but similar effects were much stronger among the boys rather than the girls when the maternal grandmother had smoked. For waist circumference, there were increases in offspring of nonsmoking women if either grandmother had smoked, but the effects were slightly stronger in boys when the MGM had smoked and in girls when the PGM had smoked.

If the mother herself had smoked prenatally, there was no discernible effect of the paternal grandmother smoking prenatally [PGM+M+ vs. PGM−M+] on the weight, BMI, or waist circumference of the child. However, if the maternal grandmother had smoked in pregnancy [MGM+M+ vs. MGM−M+], the study girls [but not boys] tended to weigh less [ranging from 0.9 to 2.2 kg] have slightly lower BMIs and reduced waist circumference [ranging from 0 to 1.7 cm].

Child's components of body composition

For children of mothers who did not smoke prenatally, the effect of the paternal grandmother smoking in pregnancy [PGM+M− vs. PGM−M−] indicated a slightly increased fat mass in the girls [ranging from 0.30 to 0.75 kg], increases in bone mass in both sexes, and strong effects on lean mass that increased with age for the boys, but was less striking for the grand-daughters after 13 years of age. By age 17, the difference between the sexes was significant [interaction P = 0.012]. If the maternal grandmother had smoked prenatally and her daughter had not [MGM+M− vs. MGM−M−], there was little effect on the child's fat mass or bone mass, but there was a strong positive association with lean mass in the study boys, but not the girls (P for interaction at age 17 = 0.006; Fig. 3).

Figure 3

Lean mass of offspring of nonsmoking women showing the difference between those whose grandmothers smoked prenatally compared with those who did not (MGM, maternal grandmother; PGM, paternal grandmother; M, mother; +, smoked prenatally; −, did not smoke prenatally).

If the study mother had smoked prenatally, there were no consistent associations between the child's body composition with the history of the paternal grandmother's smoking [PGM+M+ vs. PGM−M+], but some indication that if the maternal grandmother had smoked prenatally [MGM+M+ vs. MGM−M+] the girls had slightly lower fat, lean, and bone mass than expected from the factors taken into account. The difference between the genders was significant for both lean mass and bone mass at age 9 (P = 0.003 and 0.045, respectively).

Strength and fitness

Given the unexpected associations with lean mass, we carried out further analyses to determine whether the increase in lean mass was reflected in an increase in strength and/or fitness. We therefore looked at the mean levels of these outcomes using the same comparisons and confounders as for the anthropometry measures. The results are shown in Supporting Information Table 5. In brief, there was an association if the maternal grandmother had smoked prenatally but the study mother had not [MGM+M− vs. MGM−M−], with an increase in grip strength in the boys +0.52 [95% CI +0.11, +0.92; P = 0.012], but not girls −0.18 [95% CI −0.56, +0.20]; the test for interaction between the sexes gave P = 0.10. For fitness, there was also a positive effect for this group of boys +2.08 [95% CI +0.91, +3.26; P = 0.001] but not girls: +0.38 [95% CI −0.74, +1.50; P = 0.503]; test for interaction P = 0.061.

Multiple testing

This set of analyses has been designed to look at ways in which the grandmothers' prenatal smoking has influenced the growth of the study child. It is hypothesis generating. There are no other studies to our knowledge that can be used to attempt to replicate our results at this point in time. We are therefore reluctant to be too astringent in rejecting results that do not reach either a Bonferroni or other test for multiple testing. We therefore deliberately take a basic approach, and assess the numbers of results with P values <0.10 or less. These are shown in Table2 for each anthropometric measure.

Table 2

The number of adjusted associations with P < 0.10 (P < 0.05) for each measurement, exposure category, and sex of the study child

	PGM+M−		PGM+M+		MGM+M−		MGM+M+
Boy	Girl	Boy	Girl	Boy	Girl	Boy	Girl
Height	0	5 (3)	0	0	0	0	0	2 (1)
Weight	3 (1)	6 (4)	0	0	4 (3)	0	0	3 (2)
BMI	1 (1)	3 (1)	0	0	5 (5)	0	0	1 (0)
Waist circumference	0	4 (3)	0	1 (1)	4 (3)	1 (0)	0	1 (1)
Fat mass	5 (0)	3 (1)	0	0	1 (0)	0	0	1 (0)
Lean mass	5 (3)	3 (3)	0	0	5 (2)	0	0	1 (1)
Bone mass	5 (3)	4 (1)	0	0	1 (1)	0	0	0
Column total	19 (8)	28 (16)	0	1 (1)	20 (14)	1 (0)	0	9 (5)

MGM, maternal grandmother; PGM, paternal grandmother; M, mother; +, smoked prenatally; −, did not smoke prenatally. Expected numbers in each cell = 0.6 (0.3) for height, weight, and BMI; 0.5 (0.25) for the other measures; expected column totals = 3.8 (1.9).

A clear pattern appears—for each of the eight groups being compared there are 38 sets of analyses, and we would therefore expect 3.8 of these to have P < 0.10 and 1.9 with P < 0.05 by definition. There were just four of the eight groups that clearly showed associations in excess of this: the comparisons of PGM+M− with PGM−M− for both girls and boys; MGM+M− with MGM−M− for boys only, and MGM+M+ with MGM−M+ for girls (Table2). Not all associations are mutually exclusive but many are—e.g., MGM+M+ and MGM+M−.

DISCUSSION

This study was designed to determine whether prenatal smoking by either grandmother had discernible effects on the growth of her child. We have compared seven different anthropometric measures at six time points, distinguishing between the sexes, and comparing four different groups: PGM+M− with PGM−M−; MGM+M− with MGM−M−; PGM+M+ with PGM−M+ and MGM+M+ with MGM−M+. The results are summarized in Table3, which also include birth measurements from our previous study (Miller et al., 2014; Pembrey et al., 2014). Epidemiological strategies often include the search for patterns (Wilson, 1994) and this is the strategy we have used in this set of analyses.

Table 3

Pattern of associations at birth and in childhood for each measurement, sex, and prenatal smoke exposure category

	PGM+M−		PGM+M+		MGM+M−		MGM+M+
Boy	Girl	Boy	Girl	Boy	Girl	Boy	Girl
Birth
Weight	·	·	·	·	↑	·	·	·
Length	·	·	·	·	↑	·	·	·
BMI	·	·	·	·	↑	·	·	·
Head circumference	·	·	↓	·	·	·	·	·
Childhood
Height	·	↑	·	·	·	·	·	↓
Weight	↑	↑	·	·	↑	·	·	↓
BMI	↑	↑	·	·	↑	·	·	·
Waist circumference	·	↑	·	·	↑	(↑)	·	↓
Fat mass	·	↑	·	·	·	·	·	↓
Lean mass	↑↑	↑	·	·	↑↑	·	·	↓
Bone mass	↑	↑	·	·	·	·	·	↓

MGM, maternal grandmother; PGM, paternal grandmother; M, mother; +, smoked prenatally; −, did not smoke prenatally; ↑, positive association; ·, no consistent association; ↓, negative association.

Additional Supporting Information may be found in the online version of this article.

Supplementary Information