Introduction

Globally, approximately 20 million infants are born with low birth weight (LBW < 2500 g) annually[1]. Approximately 95% of LBW infants are born in developing countries. The two main reasons for LBW are preterm birth (

< 37 weeks) and intrauterine growth restriction (IUGR). Low birth weight is considered to be one of the most significant factors associated with neonatal, post-natal and infant mortality as well as the increased risk of diseases in adulthood[2]. Determinants of LBW may differ between populations depending on whether LBW is due to growth restriction or prematurity. Sixty-four percent of the worldwide deaths of children under the age of five in 2010 were attributable to infectious causes, and preterm birth complications were the second-leading cause of death[3]. Preterm birth complications (14.1%), intrapartum-related complications (9.4%), and sepsis or meningitis (5.2%) were the leading causes of neonatal death. In older children, pneumonia (14.1%), diarrhea (9.9%), and malaria (7.4%) claimed the most lives[3]. Brazil is among the 10 countries with the highest number of preterm births[4].

One of the most important determinants of fetal growth and birth weight is maternal nutritional status[5-7]. Studies have found that short maternal stature increases the risk of LBW[8] and offspring mortality, underweight and stunting[9]. A recent systematic review and meta-analyses[10] found that short-statured women have a higher unadjusted risk of preterm birth and LBW, and tall women have approximately one half of the unadjusted risk of LBW compared to women of the reference height.

In addition, body mass index [BMI = weight (kg)/height (m)²] plays a critical role in the determination of gestational outcome, as maternal overweight and underweight are both strong predictors of birth weight[11,12]. A meta-analysis of 78 studies involving more than one million women in developed and developing countries revealed that maternal undernutrition was strongly associated with a greater risk of LBW[13]. On the other hand, a meta-analysis of 84 studies consisting of a similar number of women indicated that maternal overweight or obesity might reduce the risk of LBW and that the protective influence appeared to be stronger in developing countries[14], although the authors identified this association as spurious and due to bias in the analysis. Based on the above research, it is evident that maternal stature and BMI both contribute significantly to the birth weight of children, but no information is available regarding the impact of the combination of short stature and BMI on the risk of LBW.

The Brazilian state of Alagoas has the lowest human development index (0.677) in the country[15]. The population of Alagoas is estimated to be around three million[16], and 56.6% of the population live in absolute poverty; 32.2% live in extreme poverty[17]. Of the irregular settlements in the state (e.g., shantytowns and illegally occupied urban areas), 94.6% are located in the capital city of Maceió[18]. Previous studies conducted by our research group in the shantytowns of Maceió have revealed a high prevalence of individuals presenting with a combination of short stature and obesity, particularly women[19,20].

The aim of the present study was to examine the hypothesis that obese mothers of short stature are at increased risk of bearing children with LBW, as has been described previously for undernourished mothers of short stature. For this purpose, the association between LBW and short maternal stature, coupled with maternal current nutritional status, has been investigated in a sample population drawn from poor communities in Maceió-AL, Brazil.

Methods

The study was approved by the Ethics Committees in Research of the Federal University of Alagoas (Maceió, AL, Brazil; protocol number 20090132001-6) and of the Federal University of São Paulo (São Paulo, SP, Brazil; protocol number 1731/11). The aims and objectives of the study were explained to all potential participants, and it was emphasized that participation was voluntary. Written informed consent was obtained from the mothers for themselves and on behalf of the children participants prior to the commencement of the study. None of the invited participants refused to participate in the study.

Results

In the present study, 2226 mother-child pairs were included from 23 shantytowns; 51% of the children were girls (1413), and 49% were boys (1349). The population was very homogeneous with respect to socioeconomic and educational levels (Table 1). The average monthly family income was Brazilian reals (R$) 452 ± 322, equivalent to US$ 224 ± 160, at an exchange rate of US$ 2.02, in September 2012. The majority of families (70%) reported a monthly income of less than one minimum Brazilian salary (R$ 622 or US$ 308), while a considerable number (25.5%) reported an even lower household income (

< ½ minimum salary). Fifty percent of the mothers had attended school for less than 4 years. Forty-six percent of the children were overweight or obese, and 5% were underweight. The mean birth weight of the children was 3230 ± 575 g (Table 1). Short stature was observed in 25% of the women. Slightly under half (47.5%) of the short-statured mothers were either overweight or obese. Ten percent of the children surveyed had LBW. The frequencies of LBW were 11.6% in girls and 8.1% in boys (Pearson x²-test, P = 0.003, OR: 1.460 CI 95%:1.13 - 1.88).

The analysis of the sociodemographic characteristics of the mothers associated with LBW is shown in Table 2. Younger mothers (less than 20 years of age) had more than twice the risk of having LBW children, and lower family income (≤ ½ Brazilian minimum salary) was also associated with an increased risk (34%) of LBW. Regarding maternal nutritional status, underweight mothers were almost 3 times more likely to deliver children with LBW (P<0.0001), while a 32% lower chance was observed for overweight mothers compared to mothers with a normal body mass index.

To maximize the possibility of detecting the influence of maternal stature in LBW, we contrasted the first (stature less or equal to 152 cm) and fourth (above 160.4 cm) height quartiles (Table 2). Mothers in the first quartile of stature had a tendency of increased risk (42%) of having LBW children compared with those in the 4^th stature quartile (p=0.078).

Considering the relation between child birth weight and maternal stature, children of mothers in the 1^st quartile of stature weighed an average of 125 g less than those of mothers in the 4^th quartile (p= 0.002) (Table 3). The mean birth weights of children of mothers in the 1^st, 2^nd and 3^rd quartiles of stature were not significantly different.

Because we detected a modification of the effect of maternal BMI by stature in relation to LBW children, we stratified the analysis of factors influencing birth weight by maternal stature. Table 4 and 5 show the results from the univariate and multivariate analysis of factors associated with low birth weight stratified by maternal stature. Tables 4 and 5 correspond to the 1st and 4^th quartiles of maternal stature, respectively.

In mothers of short stature (1^st quartile), univariate analysis showed that younger age at childbirth and underweight were significantly associated with LBW (OR: 3.36 and 3.68, respectively), but overweight decreased the risk of giving birth to a child with LBW compared with normal BMI (OR: 0.36, 95% CI: 0.15-0.87) (Table 4). Multivariate analysis confirmed these results, as LBW remained associated with maternal age below 20 years (OR= 3.05, 95% CI: 1.44 - 6.47), although the effect of underweight was attenuated (OR: 2.26, 95% CI: 0.92-5.95). Overweight mothers had a significantly lower risk of LBW (OR: 0.38), and the same trend of decreased risk was observed in obese mothers, although this effect was not statistically significant.

The results for the mothers in the 4th stature quartile are presented in Table 5. The univariate analysis revealed that lower family income and being underweight were related to LBW (OR: 1.81 and 2.41, respectively), and obese, but not overweight, mothers showed a tendency towards increased risk of LBW. Age, family income, BMI, underweight, overweight and obesity were then included in the multivariate analysis. The results confirmed a significant association between both lower family income and underweight and LBW (OR: 1.88 and 2.85, respectively). The data from obese mothers demonstrated a similar trend, although not statistically significant, for increased risk of LBW (OR: 1.66, 95% CI: 0.84-3.25).

Discussion

The causes of LBW are different in developing and developed countries[1]. In developed countries, preterm birth is major cause of LBW, whereas in developing nations, most LBW newborns are born at term but are small for gestational age (SGA)[26]. Because it is relatively difficult to obtain accurate gestational age estimates in very low income communities, we used birth weight to compare the nutritional status of the mother to the birth conditions of the child. The prevalence of LBW (10%) in this study population from shantytowns around the state capital Maceió was 17.5% higher than that reported for the state of Alagoas (7.6%) and 15% higher than the average in Brazil (8.4%)[27]. The prevalence of LBW in Brazil has remained stable (approximately 8%) since 2000 despite an increase in the number of preterm births. One possible explanation for this stability is the reduction in the frequency of intrauterine growth retardation, which counterbalances the negative effects of increased premature births[28]. The prevalence of LBW within Brazil is, however, somewhat perplexing in that the richer and more developed areas exhibit a higher prevalence of LBW than the less privileged regions. It has been suggested that this apparent paradox may result from inadequate recordkeeping and the lower survival rate of premature babies in the poorest municipalities. In isolated areas of the country, data such as gestational age, birth weight, number of home births and number of live births may be recorded incorrectly or not at all. Indeed, it is estimated that the frequency of underreporting of birth events in Brazil is approximately 7.2%[29,30].

Various studies have shown that poor families in different parts of the world have a greater risk for children with LBW than their more affluent counterparts[31-34]. A recent report from the United States confirmed that low income and income inequality exert a negative effect on birth weight[35]. The present study corroborates these findings because family income was found to be inversely associated with LBW throughout the entire sample and for normal stature mothers in the 4^th quartile (OR: 1.34, 95% CI: 0.99-1.81, P = 0.06 and adjusted OR: 1.88, 95% IC: 1.07-3.29, P = 0.027, respectively), although other factors were more relevant as determinants of LBW in short-statured mothers. In addition to being an extremely poor population, the average height of all mothers was low, 156 cm, which corresponds to the 15th percentile of the WHO reference for 19-year-old girls [21]. For this reason, we divided the population into quartiles of stature and compared the lowest quartile to the highest quartile. A significant negative effect on the birth weight of children was observed when comparing mothers in the first and second quartiles of stature to those in the fourth quartile of stature.

No association between length of schooling and LBW could be established in the present study, most likely because the sample population was very homogeneous with respect to this variable. The mean length of schooling was 2.9 years, and more than 40% of the mothers had less than 4 years of education. However, the association between poor maternal education and LBW has been observed in developed and developing countries and is well described in the literature[8,36]. A study of 78,582 children from Rio de Janeiro, Brazil, concluded that both level of education and maternal age were strongly associated with LBW[37].

Along with the factors mentioned above, gender can also influence the fetal weight[38,39]. The results presented herein revealed that the risk of LBW was 46% higher in girls, a finding that agrees with those from previous studies conducted in the southern Brazilian state of Paraná[40] and in Indonesia[41].

The results of the present study confirmed that young mothers (

< 20 years old) were 2.2 times more likely to give birth to children with LBW than their older counterparts when the entire sample was considered. This association became insignificant when mothers in the fourth quartile of stature were assessed, while younger short-statured mothers were 3.1 times more likely to give birth to LBW infants. This finding demonstrates a stronger likelihood of giving birth at an early age among short-statured mothers.

Maternal BMI is known to be a determining factor for fetal development in that both excess weight and weight deficiency can lead to adverse gestational outcomes[12,13,42]. A cohort study demonstrated that maternal underweight not only increased the risk of LBW but also the risk of premature births and neonatal mortality[43]. Additionally, a study of more than 2 million children and 751,912 low- and medium-income mothers from 54 emerging countries[9] found a direct relationship between short maternal stature and increased LBW, perinatal mortality and stature deficit in their children.

The present study confirmed a strong association between LBW and maternal underweight, regardless of stature, as underweight mothers were 2.43-fold more likely to have LBW offspring than the group as a whole. On the other hand, a marginal protective effect of overweight in mothers was found (OR; 0.72, 95% CI: 0.51-1.01, P = 0.060). As the effect of BMI appeared to be different in mothers of short and normal stature, we carried out an analysis stratified according to the 1^st (short) and 4^th (normal) stature quartiles. In short-statured mothers, age and underweight were associated with LBW; overweight mothers were protected (adjusted OR: 0.38, 95% CI: 0.16-0.95, P = 0.037) against having LBW children, and this tendency was also observed among obese mothers. Among mothers in the 4^th stature quartile, income and underweight were the factors associated with LBW in the multivariate analysis; however, in this subgroup, being overweight or obese was not significantly associated with protection against LBW. Obese mothers showed a trend towards increased risk of LBW (OR: 1.66, 95% CI: 0.84-3.25, P = 0.14). These results show that, in addition to the association between LBW, maternal short stature and undernutrition that has been described in the literature, mothers who were overweight at the time of the study were less likely to have babies with LBW. We can speculate that in extreme poverty, overweight acts as a protective factor and ensures better health for the children of mothers with short stature. This finding is the opposite of those described for developed countries such as England and the United States, where overweight (BMI > 25 kg/m²) non-short mothers were observed to have a higher risk of adverse fetal and perinatal complications and infant death [44,45]. Among mothers with normal stature, the results of the present study, although not significant, indicated a higher risk of LBW in those with obesity at the time of the study. This finding is in line with the literature results concerning the harmful consequences of maternal obesity on fetuses and infants in developed countries[44,45] .

It is well known that adult short stature reflects health processes throughout life and is associated with nutritional stress during the early stages of life[8,9,46]. The biomechanical characteristics of mothers of short-stature must also be considered because this condition is associated with a narrower pelvis, which increases the likelihood of cephalopelvic disproportion and obstructed labor[9]. Moreover, a short-statured mother may not be able to supply sufficient amounts of nutrients to the fetus, a situation that results in greater susceptibility to a negative gestational outcome. Within this context, overweight short-statured mothers have a surplus of energy and/or nutrients that may protect the fetuses.

The present study reports the effects of combined maternal short stature and overweight or obesity on the birth weight of children for the first time . The coexistence of stature deficit and excess weight is being observed at an increasingly early age in developing countries. A study conducted in 2001 established that the prevalence of overweight and obese women of short stature in poor communities of Alagoas was 30%[20], while a more recent investigation revealed that the phenomenon of stunting in combination with obesity had become prevalent in underprivileged children from the same area[47]. In the present study, we observed that the frequency of overweight and obesity was 20% higher when combined with short stature (47.6%) than when combined with normal stature (39.5%).

Our study has several limitations: the lack of an assessment of other risk factors for low birth weight, such as smoking status and alcohol intake, parity, and other maternal and obstetric conditions; the inability to discriminate between at-term and premature children; and the cross-sectional study design that prevented us from establishing a clearer causal relationship between the mothers’ BMI and LBW. Mothers’ BMI was determined an average of 30 months following childbirth. We believe that it is unlikely that maternal BMI changed dramatically after childbirth, as the establishment of obesity, overweight or even undernutrition in adult, low-income women usually requires several years and is directly related to their long-term living conditions.

In conclusion, the most important finding of this study was that overweight short-statured mothers showed a significantly lower risk of LBW. Thus, overweight seemed to be a protective factor against LBW in extreme poverty conditions. In addition, younger mothers had a higher risk of LBW and the effect was stronger among those with short stature, as the effect was not significant among normal-stature mothers when analyzed alone. Income had a negative impact on birth weight only in the normal-statured mothers, while in short-statured mothers, the negative effects of low age and the protective effects of overweight remained the most important factors in the multivariate analysis. Our other results are agree with the literature, as maternal underweight, regardless of stature, was associated with LBW, and taller women (4^th quartile) gave birth to significantly heavier babies compared those in the first and second quartile. Because this is a cross-sectional study, prospective studies are necessary to confirm the present findings.