factORs assOciated with lOw biRth weight in a histORical seRies Of deliveRies in caMpinas , bRazil

692 Rev Assoc Med Bras 2009; 55(6): 692-9 *Correspondência: R. Alexander Fleming, no 101 Campinas SP CEP: 13083-881 Telefone: (19) 3521-9482 Fax: (19) 3521-9304 E-mail: cecatti@unicamp.br Summary Objective. To identify the risk factors associated with low birth weight (<2500 grams). MethOds. Unmatched case-control study performed in a tertiary maternity hospital in Campinas, Brazil, involving 43,499 liveborn infants delivered in the institution between 1986 and 2004. Analysis of the database containing information on deliveries of women who gave birth to infants with low (6,477 cases) and normal (37,467) birth weight were performed. Factors associated with low birth weight were identified according to the odds ratio (OR) and 95% confidence interval (95%CI) in the bivariate analysis and according to the adjusted OR in the multivariate analysis. Sociodemographic characteristics, reproductive history, previous morbidity and factors related to current prenatal care were studied. Results. Extremes of reproductive age, poor education, low maternal weight, smoking beyond the fourth month of pregnancy, previous cesarean section, interdelivery interval ≤24 months and ≥37 months, maternal history of hypertension, cardiopathy and premature delivery, few (≤5) prenatal visits and beginning prenatal care late in pregnancy (after the 3rd month), premature rupture of membranes, increased blood pressure, infectious diseases and hemorrhages during current pregnancy were all associated with low birth weight. Maternal obesity and being a primipara were found to be protective factors. cOnclusiOn. These results confirm the findings of previous studies. The detection and prenatal management of modifiable factors, counseling, pre-conception intervention, adequate prenatal care and the implementation of primary and secondary prevention of maternal morbidity should be a target for all obstetrician as a potential source for reducing the incidence of low birth weight.


intROductiOn
According to the World Health Organization, every newborn infant of less than 2,500 grams at birth is classified as low weight.The incidence is estimated at more than 20 million births per year.Of these, 95.6% occur in developing countries, where 16.5% of newborn infants are of low birth weight, more than twice the percentage found in developed countries (7%). 1 Low birth weight is the result of two obstetrical situations: premature delivery 1,2 and/or a pregnancy in which the fetus has intrauterine growth restriction 3 attributed to poor uteroplacental perfusion 4 .Preterm and low birth weight are the main determining factors of adverse outcomes in childhood and are associated with poorer indicators of child morbidity 1 and mortality 5 .The reduction in the incidence of low birth weight by one-third between 2000 and 2010 with the objective of reducing child mortality is one of the Millennium Development Goals established by the World Health Organization and proposed in 2005 in their "Declaration and Recommendations for Action" 1 .
Over the last 30 years, there has been an improvement in the survival rates of low birth weight infants in developed nations 3 .Nonetheless, in global terms, children born with low birth weight are 20-fold more likely to die prematurely compared to infants of normal birth weight 6 .Likewise, these children experience more morbidity, both in the short and long term.Among these, the principal morbidities include respiratory distress syndrome, necrotizing enterocolitis, amaurosis, deafness, hydrocephaly, mental retardation and cerebral palsy 1 .Children with a birth weight below 2,500 grams have a greater chance of developing attention deficit hyperactive disorder compared to children of normal birth weight 7 .There is also a greater vulnerability to infectious diseases due to a poorer immune response 8 .Even certain conditions in adulthood may be associated with low birth weight, such as Rev Assoc Med Bras 2009; 55 (6): 692-9 coronary disease, cerebral vascular accident, hypertension, type 2 diabetes 9 , insulin resistance 10 and depression 11 .
Therefore, it is clear that low birth weight represents a heavy burden for healthcare services worldwide.In poorer countries where fewer resources are destined for healthcare, the direct, indirect, immediate and later repercussions of low birth weight constitute a relevant public health issue 1 .However, the great majority of scientific evidence accumulated on this subject is generated in developed countries.Therefore, this study was carried out in a referral maternity hospital in a developing country with the objective of identifying risk factors associated with low birth weight.

MethOds
This is an unmatched case-control study carried out at the Center for Woman's Integral Healthcare (CAISM), a tertiary, referral maternity hospital that covers a region of 3 million inhabitants of the city of Campinas in the state of São Paulo, Brazil.A specific database is systematically and prospectively fed with information on obstetrical hospitalizations occurring in the institution, including demographic data and maternal history, reproductive history, prenatal care, delivery, the puerperium and information on the perinatal period.These data are originally collected on forms filled out by the attending physicians from the admission of the patient until her release from hospital.The data are then checked and revised as necessary by a medical supervisor and then entered into the database by technical staff.
The 52,136 records of all deliveries carried out in the institution between January 1986 and December 2004 were originally available for this study.Based on these records, a second database was built in which only those records with information on birth weight were included, excluding the records of infants delivered with a weight of less than 500 grams or more than 4,000 grams at birth, those who were stillborn and twin births.A total of 8,192 records were excluded.The macrosomic newborn infants (those with a birth weight of over 4,000 grams) were excluded because they have specific profiles of maternal and perinatal morbidity and mortality that are different from those of newborn infants with a birth weight that is considered normal 12 .This selection resulted in a database of 43,944 records.
The records were then divided into two groups: 1) Newborn infants with low birth weight (cases), consisting of those whose weight at birth was less than 2,500 grams (6,477 cases; 14.7% of the total population studied and 12.8% of the liveborn infants); and 2) Newborn infants of normal weight (controls), those whose weight at birth ranged from 2,500 to 3,999 grams (37,467 cases; 85.2% of the total population studied and 87.2% of the total number of liveborn infants).Since it was possible to study all these cases, thereby avoiding any selection bias either for cases or controls, this entire sample was evaluated.With respect to the selection of controls, the proportion of almost 6 controls for each case eliminated the need to perform matching or counterbalancing procedures.Due to the incorrectness of available data regarding estimation of gestational age it was not possible to differentiate low birth weight cases between preterm or small for gestational age babies.
The following variables were considered independent: the sociodemographic characteristics of the mother/lifestyle habits (age, marital status, education level, pre-gestational body mass index, smoking habits), obstetrical history (number of pregnancies including the current one, parity, number of cesarean sections, abortions, stillbirths, living children, interval since previous delivery); history of maternal morbidity (cardiopathy, urinary tract infection, chronic hypertension and others, including tuberculosis, syphilis, diabetes, bone sequelae in the pelvis/lower limbs, abdominal surgery), obstetrical history (preeclampsia, preterm delivery and others, including hemorrhage during pregnancy, postpartum hemorrhage, prolonged pregnancy, congenital malformations, puerperal infection, polyhydramnios), factors related to current prenatal care (total number of visits, time when prenatal care was initiated) and complications during current pregnancy (course of blood pressure, infectious diseases, hemorrhages, surgeries during pregnancy, premature rupture of membranes).
Of all the eligible records available in the database, a proportion was excluded from the statistical analysis for each variable due to missing information.These proportions were different for each variable and cannot be considered a selection bias since they did not constitute a subsample with different characteristics from those of the population studied.
Initially, a bivariate analysis was carried out using the Epi Info software program, version 6.04b, in which each independent variable was crossed with the variable represented by the weight of the newborn infant.The possible risk factors were comparatively evaluated between cases (low birth weight) and controls (normal birth weight) and the odds ratio (OR) and 95% confidence interval (95%CI) were calculated.For ordinal variables with a tendency to increase or decrease linearly in risk, the x 2 for linear trends was calculated.To identify the factors and complications independently associated with low birth weight, multiple regression analysis was performed using the SAS software package, version 9.02.For this analysis, the stepwise method for regression models was used, considering low birth weight as the dependent variable and all the other predictor variables used as potential confounding factors for the adjustment of the respective odds ratios.The protocol of this study was approved by the Institutional Review Board prior to initiation.

Results
With respect to the mean age of the mothers in this study, there was no statistically significant difference between cases (25.6 ± 7 years) and controls (25.4 ± 6.5 years).Bivariate analysis of the demographic variables found that young women of 10-19 years of age and adults over 30 years of age, as well as women without a partner, with lower education levels, with low maternal weight prior to pregnancy and smoking beyond the fourth month of pregnancy had a greater risk of giving birth to low birth weight infants.The risk of low birth weight is higher for women who are pregnant for the first time, with no living children, with an interdelivery interval ≤ 24 months or ≥ 37 months or who have not previously delivered and also for those who have had several pregnancies, with previous abortions and stillbirths (Table 1).
Women with a history of chronic hypertension, cardiopathies, preeclampsia or preterm delivery were found to have an increased risk of delivering an infant with low birth weight, as well as those who had received no prenatal care or who had attended fewer than five prenatal visits, who had hypertension (either chronic or gestational), infections, hemorrhages, premature rupture of membranes, or who were submitted to conization, cerclage or other surgeries during pregnancy (Table 2).According to the multiple regression analysis, age, low educational level, low maternal weight and smoking beyond the fourth month of pregnancy were risk factors for low birth weight.Maternal obesity was found to be a protective factor.Parity and marital status were not independently associated with this outcome.Unlike the bivariate analysis, logistic regression found that there is a protective effect in primiparas and that women with a history of cesarean sections have an increased risk of giving birth to infants with a birth weight of less than 2,500 grams.Confirming the preliminary analyses, women with no living children also have an increased risk of having an infant with low birth weight.With respect to the interval between deliveries, both extremes (≤ 24months and ≥ 37 months) were associated with the occurrence of low birth weight.History of previous abortions or stillbirths was not identified as independent factors associated with low birth weight in this population.Maternal history of chronic hypertension, cardiopathy and preterm delivery were risk factors, whereas a history of preeclampsia was not.Having attended fewer than 5 prenatal visits was found to be a risk factor independently associated with the birth of an infant with low birth weight, as well as beginning prenatal care late (after the third month of pregnancy).Finally, the occurrence of premature rupture of membranes, increased blood pressure, infectious diseases and hemorrhages during the current pregnancy were found to be factors significantly associated with low birth weight (Table 3).

discussiOn
In this study, various factors associated with low birth weight were identified in a developing country population with an elevated prevalence of high risk pregnancies, and these results are in agreement with data published in the literature with respect to various risk factors for low birth weight.
Extremes of reproductive age 13,14 , women without a partner 14 , low educational level and low pregestational BMI 1,15,16 were findings shown to be risk factors for low birth weight in this population, and these findings are in agreement with data reported by various other investigators.On the other hand, the present study also showed obesity to be a protective factor, which may reflect the fact that obese women have a greater risk of developing hyperglycemic states or diabetes (gestational or type II), commonly associated with a greater gain in fetal weight.Smoking beyond the 4 th month of pregnancy was shown to be a risk factor irrespective of the number of cigarettes smoked per day, and this finding is in agreement with reports from previous studies [17][18][19][20] .
There is evidence that pregnant women who suffer from chronic psychosocial stress 21 , who have ambivalent feelings with respect to the pregnancy 22 , or who have suffered some form of psychological, physical and/or sexual violence during pregnancy 23 have a greater chance of giving birth to an infant with a birth weight of less than 2,500 grams.Although these factors were not evaluated in this population of pregnant women, systematic incorporation of these forms of evaluation in routine prenatal care represents a challenge, particularly in less developed regions where these factors may contribute significantly towards the occurrence of low birth weight.
Parity was not found to be an independent risk factor for low birth weight, despite the fact that the bivariate analysis showed it to be so, and this is in agreement with data published in the literature.These results show that women with a history of abortions and stillbirths have a greater chance of having infants with low birth weight, this risk increasing as the number of these previous obstetrical outcomes also increases.One possible explanation for this may be the association between these outcomes and morbidities that affect placental vasculature, such as collagenosis and antiphospholipid syndrome, which are also associated with low birth weight.Nevertheless, none of these characteristics remained significantly associated with low birth weight in the multivariate analysis, possibly because they are simultaneously associated with some of the other characteristics identified.As in previous studies 24,25 , short or long interdelivery intervals were associated with low birth weight.In addition, previous cesarean sections were associated with a 20% increase in the occurrence of low birth weight.As well as the tendency of pregnancies with lower gestational ages to result in repeat cesarean sections, it is probable that some of the situations of risk for low birth weight also contributed towards increasing indications for cesarean sections.
In agreement with the findings of Graham et al 26 , the results of the current study show that chronic maternal morbidities such as hypertension and cardiopathy were positively associated with low birth weight.Moreover, they show that morbidities characteristic of pregnancy (hypertension, preeclampsia, history of preterm, infectious diseases, premature rupture of membranes, and hemorrhages during pregnancy) are also risk factors for low birth weight 1,27 .In fact, infections during pregnancy are also considered risk factors, including asymptomatic bacteriuria 28 , bacterial vaginosis 29 , trichomoniasis 30 , malaria 31 and maternal periodontal disease 32,33 .
This study has some limitations that must be taken into consideration, one being the fact that the period of eligibility, both for cases and for controls, was 18 years.Within this period, the way in which a number of factors are managed that in this study were shown to be associated with low birth weight evolved considerably.It is probable that differences in the way of recognizing and treating these conditions (for example, preterm labor or maternal morbidities) may affect the magnitude and the significance of the associations reported here.Nevertheless, this is an effect that is very difficult to control when long historical series are being evaluated.In addition, as said previously, the incorrectness of available data on gestational age did not allow for separating cases of low birth weight between preterm and growth restricted newborns.This is a common problem basically for the majority of historical databases of deliveries where an early ultrasound exam is not systematically available for all women.
Another possible limitation of this study refers to the under-notification of some conditions of interest in the records, the non-uniform classification of the degree of severity, and the grouping together of some conditions into larger, heterogeneous categories.Despite these limitations, it is believed that the detection and prenatal management of changeable risk factors such as low maternal BMI, smoking and the interdelivery interval may contribute towards reducing the occurrence of low birth weight.Likewise, adequate prenatal care was found to reduce the occurrence of adverse perinatal events, including low birth weight 14,34 .Therefore, it is feasible to assume that counseling and pre-gestational intervention could possibly contribute similarly to an improvement in perinatal outcome.The same may be said with respect to primary and secondary prevention of maternal morbidity, whether gestational or pre-gestational.

cOnclusiOn
Low birth weight represents a condition that results in significant repercussions in healthcare systems, particularly in developing countries.These repercussions, as well as representing a significant neonatal problem, reach far beyond the healthcare field in that they represent high economical costs and affect future generations.Various factors have been identified as being associated with low birth weight; many of them are avoidable.A strategy must now be implemented to alter this outcome and combat what is one more silent epidemic.One possibility for the time being is trying to alert obstetricians for identification of these risk factors in order to possibly take specific interventions that could decrease the occurrence of low birth weight.The demonstration that this really helps and happens would be a good objective for a new ambitious intervention study to be performed in a near future.

Table 1 -Crude estimates of risk (Or-95%CI) of low birth weight according to socio-demographic characteristics and obstetrical history
* x2 linear trend p<0.00001

Table 2 -Crude estimates of risk (Or-95%CI) of low birth weight according to history of morbidity and obstetrical pathology and characteristics of the current pregnancy
* Systemic arterial hypertension/ Preeclampsia