Maternal Smoking During Pregnancy, Growth, and Bone Mass in Prepubertal Children

Authors


  • Presented in part at the 19th Annual Meeting of the American Society for Bone and Mineral Research, Cincinnati, Ohio, U.S.A. September 10–14, 1997.

Abstract

There have been no studies of smoking during pregnancy and bone mineralization in children. The objective of this population-based longitudinal study was to determine whether maternal smoking during pregnancy is associated with bone mass and other growth variables in prepubertal children. We studied 330 8-year-old male and female children representing 47% of those who originally took part in a study of risk factors for Sudden Infant Death Syndrome in 1988. The main outcome measures were bone mineral density measured by a Hologic QDR2000 densitometer: birth weight, placental weight, height, and weight. Maternal smoking during pregnancy was associated with deficits in growth with these children having lower height (−1.53 cm, 95% confidence interval [CI] −3.03 to −0.03) and a trend to lower weight (−1.35 kg, 95% CI −2.75 to 0.11) at age 8. Furthermore, there was a disproportionate deficit in bone mass such that those children whose mothers smoked during pregnancy had lower size adjusted bone mass at the lumbar spine (−0.019 g/cm2, 95% CI −0.033 to −0.005) and femoral neck (−0.018 g/cm2, 95%CI −0.034 to −0.002) but not total body (−0.005 g/cm2, 95% CI −0.015 to 0.005). This association was only present for children born at term. Mothers who smoked during pregnancy also had lower placental weight (− 56 g, 95% CI −95 to −17), and further adjustment for placental weight led to nonsignificant results for smoking with both growth and bone parameters, suggesting that these associations may be mediated through placental size and function. Maternal smoking habit in 1996 was not significantly associated with bone mass at any site. In conclusion, this study has demonstrated a long-term negative association between maternal smoking during pregnancy and both growth and bone mass in children born at term, and suggests that the timing of exposure rather than the dose or duration is critical. If these associations are present in other populations and they persist until the attainment of peak bone mass, then our findings suggest that osteoporosis prevention programs should start very early in the life cycle.

INTRODUCTION

Fractures are a major public health problem in both males and females.(1) Bone density is one of the major predictors of these osteoporotic fractures(2) and is the result of the amount of bone gained in early life (i.e., peak bone mass) and subsequent bone loss.(3) Physical activity and, to a lesser extent, diet (particularly calcium intake) during adolescence and early adulthood have been implicated as determinants of peak bone mass.(4,5) However, the vast majority of adult bone mass is attained before age 14,(6) but relatively little is known about the effect of genetics or lifestyle factors on bone acquisition from birth to puberty. A British group has reported that weight at age 1 year could predict peak bone mass(7) and also bone mass in the elderly, suggesting that skeletal growth may be programmed during intrauterine or early postnatal life.(8) One potential candidate for this programming is smoking during pregnancy. Smoking is deleterious to bone in adults particularly in postmenopausal women.(9-13) Passive smoking is unlikely to be directly harmful to bone in children given the low levels of exposure to nicotine (but not necessarily other components of smoke).(14) It may also be associated due to its known association with socioeconomic status.(15) However, smoking by the mother during pregnancy may well be detrimental because it can lead to intrauterine growth retardation,(16) and this effect appears to be more marked later in the pregnancy. (17) Whether there is an effect on bone mineralization is uncertain since there have been no studies. In this study, therefore, we asked the following research questions: Is smoking during pregnancy associated with a deficit in bone mass in 8-year-old male and female children either before or after adjustment for size? Is this association independent of other factors related to socioeconomic status?

MATERIALS AND METHODS

In 1988, there were 6779 live births in Tasmania. Of these, 1380 were identified as being at higher risk of Sudden Infant Death Syndrome by previously published criteria(18) and were invited to take part in a longitudinal study. In Southern Tasmania, there were 735 births that met these criteria. Of these, 696 mothers (95%) agreed to an in-hospital interview, and 581 (80%) agreed to the 1 month follow-up. Comprehensive sociodemographic, obstetric, and perinatal information involving over 450 pieces of information is also available on these subjects from the pre-existing Tasmanian Infant Health Survey(18) database. Data of potential relevance to the current study was available for: maternal smoking during pregnancy (both dichotomous and cigarettes per day in each trimester), breastfeeding (intention and actual), socioeconomic data (marital status, employment status, educational level, household income), birth weight, and placental weight.

The 696 subjects who agreed to the in-hospital interview in 1988 were approached during 1996 to take part in a further study. After 8 years, we were able, through the use of school lists, to definitely identify 551 of these subjects (80%) which closely approximates the number likely to be available in Tasmania utilizing Australian Bureau of Statistics data on annual outward migration rates.

This study was granted approval by the University of Tasmania Ethics Committee (human experimentation). Subjects who provided informed consent to take part underwent a protocol including measurement of bone mass, anthropometrics, questionnaire, assessment of duration of breastfeeding, current maternal and paternal smoking, and socioeconomic factors as well as current diet (Food Frequency Questionnaire), physical activity, and sunlight exposure. Weight was measured by electronic scales (calibrated at the beginning of the study by the manufacturer). Height was measured by a stadiometer. Questions on socioeconomic status related to maternal education (4 point scale), paternal employment status, and marital status (6 point scale).

Bone mass was assessed utilizing the technique of dual-energy X-ray absorptiometry at the total body, lumbar spine, and femoral neck. The machine utilized was a Hologic QDR 2000 (Waltham, MA, U.S.A.) densitometer on array setting. Bone mass was examined in two separate ways: bone mineral content (BMC) and bone mineral density (BMD). Precision estimates in vivo are not available in our subjects for ethical reasons but are 1–2% in adults. The longitudinal coefficient of variation for our machine during 1996 using daily measurements of a spine phantom was 0.54%.

Statistics

A combination of univariate and multivariate linear modeling techniques were utilized. Both BMC and BMD were regarded as dependent variables. The association with smoking was then examined. Three steps were taken. First, a univariate association was assessed at all sites of interest. Second, it was then adjusted for birth weight. The final step was to adjust for gender and current size. For BMC, bone area at that site was utilized. For BMD, current height and weight were both added as independent variables to the model relating smoking to BMD so that the effect of smoking could be examined in a child of the same weight and height. To examine the role of other lifestyle factors, associations were then adjusted for current calcium intake, sports participation, and winter sunlight exposure simultaneously. Placental weight was then added to the model to examine if it could be acting as an intermediate variable. To examine the effect of socioeconomic status, associations were adjusted individually for maternal education, employment status, and marital status as measured were all correlated with each other. A p-value of 0.05 (two-tailed) or a 95% confidence interval (CI) not including the null point was regarded as statistically significant. All statistical calculations were carried out on SPSS version 6.02 for Windows (SPSS Inc., Chicago, IL, U.S.A.).

RESULTS

There were 330 subjects in all, representing a response rate of 60% of those available or 47% of those in the original study. Demographics and study factors of interest are presented in Table 1. Males and females were similar in terms of height and weight. Due to selection factors, males weighed more at birth and were less likely to be premature than females but were similar in terms of other factors, such as proportion breastfed and proportion of maternal smoking during pregnancy.

Table Table 1.. Demographic Details of Subjects
original image

Associations between maternal smoking and bone mass in children born at term are presented in Table 2. Results were identical in terms of effect size and statistical significance for BMC and BMD, so only results for BMD are presented here. Furthermore, coefficients for smoking were similar in both genders, so a combined analysis is presented.

Table Table 2.. Univariate and Multivariate Associations Between Maternal Smoking and Areal Bone Mass in 8-Year-Old Children Born at Term
original image

Smoking by the mother during pregnancy was associated with a deficit in bone mass at age 8 in children born at term (Table 2 and Fig. 1). These children also had gender-adjusted deficits in birth weight (−106 g, 95% CI −246 to 34), current weight (−1.35 kg, 95% CI −2.75 to 0.11), and current height (−1.53 cm, 95% CI −3.03 to −0.03) at 8 years (Fig. 1). The magnitude of the deficit in bone mass was reduced after the addition to the model of current weight and height but remained statistically significant at the femoral neck and lumbar spine. There was no apparent dose effect (Fig. 2). Smoking mothers also had lower placental weight (−56 g, 95% CI −95 to −17), and further adjustment for placental weight led to nonsignificant results for smoking (lumbar spine (−0.012 g/cm2, 95% CI −0.027 to 0.003), femoral neck (−0.014 g/cm2, 95% CI −0.031 to 0.003), total body (−0.001 g/cm2, 95% CI −0.011 to −0.009) as well as growth parameters, suggesting these associations may be mediated through the placenta. This effect was not observed if birth weight rather than placental weight was included in the model (data not shown). Smoking was not associated with size-adjusted deficits in bone mass in children born prior to 37 weeks (lumbar spine [−0.010 g/cm2, 95% CI −0.039 to 0.019], femoral neck [0.011 g/cm2, 95% CI −0.016 to 0.038], total body [0.003 g/cm2, 95% CI −0.017 to 0.023]).

Figure FIG. 1..

Maternal smoking dose during the third trimester of pregnancy, growth parameters, and bone mass in 8-year-old children born at term. These children have deficits in placental weight, birth weight, current weight, current height, and bone mass at the femoral neck, spine, and to a lesser extent at the total body. Differences are greater in magnitude for bone mass and placental weight than size parameters. Results are presented in SDs to allow direct comparisons.

Figure FIG. 2..

Dose of smoking and bone mass in 8-year-old children born at term. There is no clear association between the amount of cigarettes smoked per day and bone mass at the femoral neck or lumbar spine. Results are presented as mean ± SEM.

Maternal smoking habits changed somewhat in the 8 years between assessments, with 16 taking up smoking and 33 ceasing smoking. There was good agreement between smoking in 1988 and smoking in 1996 (κ = 0.69, 95% CI 0.61–0.77). However, the nondifferential changes in smoking resulted in a dilution of the association with bone mass such that current maternal smoking was not significantly associated with size-adjusted deficits in bone mass at any site in these same children (femoral neck −0.009 g/cm2 [95% CI −0.027 to 0.009], lumbar spine −0.014 g/cm2 [95% CI −0.030 to 0.002], total body 0.005 g/cm2 [95% CI −0.006 to 0.017]).

The associations with smoking were not substantially altered by adjustment for current sports participation, dietary calcium intake, or sunlight exposure (data not shown). Furthermore, adjustment for maternal education, marital status, or employment status did not alter the reported associations. Indeed, none of these measures was associated with BMD. Smoking and breastfeeding were negatively associated with each other, resulting in results of borderline significance for both if they were included simultaneously in a model. However, a combined effect was evident when results are presented in a stratified manner. Of those children born at term, those who were breastfed and whose mothers did not smoke during pregnancy had higher gender-adjusted bone mass at all sites at age 8 when compared with those who were not breastfed and whose mothers smoked during pregnancy (femoral neck 0.034 g/cm2 [95% CI 0.007–0.061], lumbar spine 0.040 g/cm2 [95% CI 0.018–0.062), total body 0.023 g/cm2 [95% CI 0.008–0.038]) (Fig. 3).

Figure FIG. 3..

Breastfeeding, maternal smoking, and bone mass in 8-year-old children born at term. Children who were both breastfed and whose mothers did not smoke during pregnancy have bone mass that is 3.3–7.7% higher depending on site than those who were not breastfed and whose mothers smoked (all statistically significant). Children who had one exposure but not both had intermediate bone mass. Results are presented as mean ± SEM.

The availability of pre-existing data allowed assessment of the characteristics of nonresponders. As compared with responders, nonresponders had lower rates of breastfeeding (25 vs. 53%, p < 0.0001), higher rates of smoking (64 vs. 49%, p = 0.001), lower maternal age in 1988 (23.1 vs. 25.1 years, p < 0.0001), lower levels of maternal education (p < 0.0001), higher levels of paternal unemployment (30 vs. 6%, p = 0.008) and a trend to higher birth weight (3.19 v. 3.07 kg, p = 0.06). All other comparisons were not statistically significant.

DISCUSSION

This longitudinal study has shown that smoking during pregnancy is negatively associated with current bone mass in 8-year-old children. This association is apparent for both growth and disproportionately for bone mineralization with up to a 5% site-dependent deficit in bone mass at age 8. If these associations persist into adult life, they suggest that osteoporosis prevention programs may need to start very early in the life cycle.

Smoking has deleterious effects on bone in all age groups reported to date.(9-13) It has been associated with increased postmenopausal bone loss(13) as well as deficits in BMD in men(11) and young women,(12) although the major effect appears to be on postmenopausal bone loss.(19) To date, there have been no studies in children. Biologically, it would appear implausible that passive smoking should have a negative effect on bone in childhood, given putative mechanisms such as lower body weight and sex hormone levels, because the level of exposure is insufficient, with studies estimating 30- to 100-fold lower exposure to nicotine with environmental exposure.(14) This is consistent with the present study, which found no association with current smoking. However, smoking exposure in utero may well be harmful. Previous studies have linked smoking during pregnancy with defects in growth both at birth(16) and up to 5 years of age,(20) although there may be some catch-up phenomenon.(21) In this study, children of smoking mothers had lower birth weight as well as lower height and weight at 8 years of age. Children also had significant deficits in bone mass, which persisted after adjustment for size suggesting a long-term decrease in both growth and bone mineralization. Interestingly, this association was only present in children born at 37 weeks of gestation or later and not in those born prematurely. Studies of growth have suggested that exposure to smoking is more harmful later in the pregnancy(17,22) and our findings are in agreement with this hypothesis, although power considerations limit any conclusions we might draw about premature children with regard to the size of any potential deficit with smoking. Similarly, insufficient women changed their smoking habits during pregnancy for us to comment on whether it is exposure during the third trimester or the duration of exposure during pregnancy that is most important.

We did not find evidence of a dose response. This may in part be due to misclassification of exposure. A recent report indicated a much stronger relationship between a biological measure of exposure (cotinine) and growth when compared with self-report.(14) Alternatively, there may be a threshold effect, whereby exposure to small levels of smoking in utero is all that is required for a detrimental effect and our data support this concept with little incremental change at higher doses.

There are a number of potential mechanisms for the detrimental effect of smoking either separately or in combination. These include impaired placental size and function,(22-26) lower maternal glucose levels,(27) deficient maternal diet,(28) and lower volume of maternal breast milk.(29) It would appear most likely that smoking primarily interferes with placental function. Recent studies have suggested defects in both placental size(22,24) and function with reported alterations in endothelial function,(23,24) epidermal growth factor,(26) and cytotrophoblast differentiation.(27) In support of this hypothesis, we found a decrease in placental weight in smoking mothers, and adjustment for this decrease led to nonsignificant results for smoking, suggesting that placental weight is an intermediate variable for the negative effect of smoking. Furthermore, our results indicate that the deficit is bone mass is partially explicable by low birth weight, but more so by subsequent growth deficits, suggesting that growth, at least during the first 8 years, may be programmed in utero.

The combined effect of benefit from not smoking and breastfeeding is substantial, with bone mass being on the order of 3–7% higher at age 8 depending on the site. This equates to ∼0.5 SD at all sites after adjustment for size and is similar in magnitude to the difference in bone mass between fracture cases and controls which, according to a recent overview, varies between 0.5 and 1 SD.(2) If this benefit persists until the time of peak bone mass or even later in life, modification of these behaviors could have a substantial impact on fracture risk in later life. It remains speculative at present as to whether this is the case, but there is some evidence to support the concept that early life exposures may have an impact in the elderly. Cooper et al. reported similar correlations between weight at 1 year and bone mass(7) in the elderly.(8)

This study has a number of potential limitations. The children in this study are not representative of Southern Tasmanian children. Selection factors (which did not include smoking) resulted in a higher proportion of males, lower proportion of breastfeeding, and higher proportion of smoking mothers than in the Tasmanian population as a whole. The effect of this has been to increase the efficiency of the study because the prevalence of smoking was ∼50%. Mietinen states that for exposure outcome associations to be generalizable to other populations, three key criteria need to be met regarding definition of eligibility, sample size, and adequate distribution of study factors, all of which are met by this study.(30) The availability of pre-existing data allowed us to examine nonresponse. Subjects who were lost to follow-up or did not participate were under-represented in terms of breastfeeding and over-represented in terms of smoking and younger mothers, suggesting possible lower bone mass in nonparticipants. Smoking has been linked with socioeconomic status,(15) thus it is possible that any reported association may potentially be mediated by associated defects in lifestyle. However, adjustment for current diet, physical activity, and sunlight exposure did not modify these associations, Furthermore, more generic measures of socioeconomic status were not associated with bone mass in this sample, suggesting that the associations with smoking are biological. Similarly, other investigators have reported that measures of socioeconomic status do not alter the association between maternal smoking and intrauterine growth retardation.(31) This hypothesis is further supported by the evidence for periods of critical exposure for a harmful effect of smoking in that there is only evidence of a harmful effect for those children born at term but not with current maternal or paternal smoking or premature children exposed in utero. The above statements also suggest that nonresponse bias evident in this study will not affect the apparent size of the smoking effect because it is not modified or mediated by socioeconomic factors. A further potential weakness of this study is the use of array setting on our densitometer. Recent work has indicated that this can introduce magnification error in both BMC and bone area.(32) However, this paper concentrates on BMD, which is not affected, and the smoking associations with both BMD and BMC are remarkably consistent, suggesting that this problem may be less serious in our sample due to the relative homogeneity in body size introduced by having children of the same age.

In conclusion, this study has demonstrated associations between maternal smoking during pregnancy and bone mass in 8-year-old male and female children born at term. It would appear plausible to suggest that this association is biological and that the timing of exposure rather than the dose or duration appears critical. If these associations can be documented in other populations and they persist until the timing of peak bone mass, then our findings suggest that osteoporosis prevention programs should start very early in the life cycle.

Acknowledgements

Special thanks also to Carole Goff (Research Coordinator), Jenny Cochrane (Data Manager), Denise Kaye, and the staff of the Medical Imaging Department at Royal Hobart Hospital and Jack Allan. This work was supported by the National Health and Medical Research Council of Australia, Royal Hobart Hospital Acute Care Program, Lions Club of Australia, Coca-Cola Amatil, Tasmanian Dairy Authority, and Talays.

Ancillary