Artigos

Introduction

In Brazil, in 2021 and 2022, almost four million children and adolescents younger than 19
years were hospitalized. 1 Among these, more than one million were younger than one year. In
this group, the main causes for hospitalization include perinatal morbidity (56%), infectious
diseases (12%), and respiratory diseases (12%). Early childhood illnesses are, at least in part,
associated with perinatal conditions. Exposure to adverse perinatal conditions increase the
risk of hospitalizations within the first year of life. 2 Moreover, the risk of hospitalization in
children younger than five years increases with male gender, low-income families, order of
birth, tobacco exposure, living in rural areas, short duration of breastfeeding, malnutrition,
maternal age, low maternal education level, higher household density, low birth weight
(LBW), and prematurity. 2,3
Oliveira & Barbieri (2017) studied the effects of perinatal factors on health outcomes in a
Brazilian birth cohort. They found that, among others, gestational age (GA) and LBW
(< 2,500 g) were associated with greater risks of hospitalizations during the first year of life. 4
In the absence of other Brazilian studies, similar findings were reported in a large German
birth cohort (2007-2013), in which very low birth weight (VLBW, < 1,500 g) and LBW
(1,500-2,500 g) were associated with higher risks of hospitalizations during the first year of
life after the perinatal period. 5 In a Swedish birth cohort (1973-1994), birth weight was
strongly associated with overall health (all-cause and cause-specific hospitalizations and
sickness absences) during infancy, after which it weakens throughout childhood and
adolescence, and increases again in adulthood. 6
However, LBW is often a proxy for prematurity and/or intrauterine growth restriction
(IUGR), not necessarily a cause of morbidity per se. This is important because birth weight
alone does not accurately reflect the newborn body composition and has been criticized
because the cut-off of < 2,500 g is the same across all GAs. 7

5

In this regard, Villar et al studied the accuracy of three anthropometric ratios commonly used
to estimate body composition in newborns: weight/length ratio (W/L, kg/m), body-mass index
(BMI, kg/m 2 ), and ponderal index (PI, kg/m 3 ). They showed that W/L by GA was a better
predictor of fat mass (FM) and fat-free mass (FFM) than BMI or PI, independently of sex,
GA, and timing of measurement. 8 However, to the best of our knowledge, there is no study
investigating the association of low birth W/L and hospitalizations during the first year of life.
Therefore, this study aimed at investigating whether a low W/L at birth (< 3 rd percentile for
sex and GA) is independently associated with hospitalizations during the first year of life in a
Brazilian birth cohort.

Patients and methods

This is a retrospective analysis of prospectively collected data from a cohort study of all births
from two Brazilian cities (Brazilian Ribeirão Preto and São Luís Birth Cohort Studies -
BRISA). 9,10 For this study, only children born at Ribeirão Preto were eligible. The cohort study
was approved by the institutional review board (IRB) (#11157/2008 and #4250/2016) and was
conducted according to the Brazilian regulations on medical research. For this analysis, the
need for signed informed consent was waived by the IRB. This study followed the STROBE
recommendations. 11
Briefly, part of the mothers started being followed-up during pregnancy (prenatal cohort, one
visit with 5 months of gestation), while most entered the study upon delivery (birth cohort).
Data collection took from January 2010 to June 2011. Ribeirão Preto is in the countryside of
São Paulo state, Brazil, and has a human development index of 0.8, higher than most of the
country. Participants were recruited in all maternities, public and private. The mothers were
interviewed by trained researchers a few hours after delivery, given they consented and were
able to answer the questions, which comprised: demographics, general and reproductive

6

health, past conditions, habits, pregnancy course, prenatal care, delivery, and the newborn.
Anthropometric measurements of the newborn were collected from medical charts.
After inclusion and the assessment at birth, all mothers/children were invited to a follow-up
visit after their first birthday, which occurred between 2011 and 2013. Trained researchers
collected information on mental health, habits, contraception, feeding, general health of the
children, among others. The children underwent blood draw, oral health assessment, and
anthropometric measurements.
Variables
The main outcome was the answer to the question: “Was the child admitted to a hospital
anytime since birth up to their first birthday?”. They could answer yes, no, or do not know. A
second question was used to determine if hospitalizations occurred in the neonatal period or
after: “If yes, how old was the child on their first hospitalization?”
The independent variables were maternal factors (age, skin color, years of education, marital
status), gestational factors (prenatal care, GA, delivery type, type of health system, and
smoking, diabetes, and hypertension during pregnancy), and newborn factors (sex, W/L, 5 th -
minute Apgar score < 7).
GA was determined from the date of the last menstrual period before pregnancy and the first
ultrasound exam. GA was estimated primarily from the date of the last menstrual period,
unless it differed by more than 7% from the GA estimated by the ultrasound, in which case
the latter was considered. 12 Prenatal care was analyzed as adequate (at least six appointments)
or inadequate. W/L was categorized as low when below the 3 rd percentile for GA and sex
according to the Intergrowth 21 st reference. 13
Statistical analysis
Bivariate comparisons were made using Student’s t test, Mann-Whitney’s U test, or Fisher’s
exact test, as appropriate. Cox proportional hazards models were used to investigate the

7

association of low W/L and hospital admissions in the first year of life. Crude and adjusted
hazard ratios (HR) were calculated along with their corresponding 95% confidence intervals
(95%CI). The variables used in model adjustment were selected with the help of the DAGitty
tool (http://www.dagitty.net, Supplementary Figure 1 ). Significance was set at 5%. Stata SE
14.0 (StataCorp, USA) was used.

Results

Among the 8,342 births in the database, stillbirths (n=42), children who died within the first
year of life (n=68), and twins (n=188) were not included. The final database contained 8,044
eligible infants. Among these, 4,126 (51.5%) attended to the follow-up visit, but 39 had
missing data. There were 741 children (18.1%) who had been hospitalized at least once during
the first year of life ( Figure 1 ).
The mothers were usually young adults (20-34 years-old), mostly white, married or in stable
relationships, and with 9-11 years of education. The mothers of children who were
hospitalized were significantly younger, more frequently non-white, had a higher prevalence
of hypertension during pregnancy ( Table 1 ). Almost all mothers received adequate prenatal
care, mostly in the Brazilian public health system (SUS), but more than a half underwent a
cesarean delivery ( Table 2 ). The mothers of children who were hospitalized were less likely
to have received adequate prenatal care, and used the public health system more frequently.
Hospitalized children were mostly boys, with lower GA and 1 st and 5 th -minute Apgar scores,
more frequently born preterm and with a low 5 th -minute Apgar scores (< 7), had lower birth
weight, length, head circumference, and W/L ratio ( Table 3 ). The proportion of children with
a low birth W/L was higher in children who were hospitalized (7.7% vs. 2.2%).
For our main objective, Table 4 presents the variables associated with hospitalization during
the first year of life, including or not hospitalizations at birth in a neonatal intensive care unit
(NICU). In the univariate analysis, several factors were associated with increased risk of the

8

outcome: low W/L, inadequate prenatal care, hypertension during pregnancy, black skin
colors, and using the public health system. When we excluded neonatal hospitalizations in a
NICU, the results were similar, adding low socioeconomic level (classes D/E) and less than
12 years of education. However, in the multivariate analysis, after adjusting for the other
independent variables, a low birth W/L was associated with greater risk of all hospitalizations
(aHR 2.67, 95%CI 1.98, 3.60), but this association disappeared when we excluded neonatal
hospitalizations (aHR 1.58, 95%CI 0.98, 2.54). The only variable independently associated
with greater hazard of hospitalization beyond the neonatal period was hypertension during
pregnancy (aHR 1.32, 95%CI 1.05, 1.66).

Discussion

In this study, in a large Brazilian cohort of live births, we showed that a low birth W/L was
associated with an increased risk of hospitalizations during the first year of life, but not
beyond the neonatal period, even after adjustment for prenatal care, maternal hypertension,
skin color, socioeconomic level, educational status, and using the public health system.
Rüdiger et al, in a very large German birth cohort, showed that children born with VLBW
(< 1,500 g) had a 3.9-fold higher risk of hospitalization – with a distinct morbidity pattern –
during their first years of life, compared to normal birth weight infants, even after adjusting
for sex, area of living, and the presence of previous perinatal hospitalization. 5 In that study,
the cumulative hospitalization was 72% for VLBW, 39% for LBW, and 22% for normal birth
weight infants, which is similar to our data (71.4% for VLBW, 29.9% for LBW, and 16.9%
for normal birth weight infants).
In the study by Helgertz et al, in Sweden, the association between LBW and all-cause and
cause-specific hospitalizations and sickness absence was strong during infancy, weakened
during childhood and adolescence, and increased again in adulthood. 6 The authors also

9

demonstrated that the effect of birth weight on hospitalizations is stronger among children of
less educated mothers, which is consistent with our findings.
In the study by Oliveira & Barbieri, which analyzed the same database as we did, the
independent risk factors for hospitalization in the first year of life were maternal hypertension
during pregnancy, use of the public health system, preterm birth (< 37 weeks), male sex,
LBW (< 2,500 g), and 5 th -minute Apgar score < 7. 4 However, similarly to our findings, the
association of LBW and hospitalizations was only significant up to 30 days of life.
In all these previous studies, the proxy for suboptimal fetal growth was LBW, which can be
misleading. The cut-off of 2,500 g was established in the beginning of the 20 th century, when
GA calculation was not routine, and the definition of preterm birth was based on weight,
length, and morphological features of the newborn. 7 Because the same cut-off is
inappropriately applied over all GAs, we hypothesized that a variable that could be adjusted
for GA and sex and that better reflected the newborn body composition, that is, the W/L ratio,
could be more strongly associated with the outcome of hospitalizations during the first year of
life, which, in fact, occurred. The relative risks reported by Oliveira & Barbieri for LBW were
1.58 (95%CI 1.27, 1.96) and 1.46 (95%CI 1.17, 1.82), lower than the HR what we found, 2.67
(95%IC 1.98, 3.60). 4 Another reason to abandon LBW as a predictor for infant morbidity is
the so-called epidemiologic paradox of LBW: countries and regions with higher levels of
socioeconomic development and lower infant mortality rates have been consistently reporting
higher LBW rates. 14 The explanations for the paradox include underreporting of live births in
less developed regions and the availability of perinatal care services over underlying social
conditions.
For older children, WHO adopts BMI over body weight because the former is a better
surrogate for adiposity. 15 However, Villar et al (2017) demonstrated that, for newborns, W/L
for age reflects neonatal body composition, in terms of lean mass, fat mass, and fat

10

percentage, better than BMI for age. 8 The association between a low birth W/L and perinatal
morbidity is not new. In 1997, Williams et al published a study involving newborns ? 34
weeks of GA, in which a low birth W/L was associated with poorer perinatal outcomes,
including death. This association was stronger than that of SGA. 16 In the following year, the
same authors reported that a low birth W/L, as a surrogate for asymmetrical intrauterine
growth, was best associated with poorer neonatal outcomes (cerebral palsy and death), even
after adjusting for preterm and twin birth. 17
The explanation for a greater association of W/L with perinatal morbidity is that LBW alone
is not the causative factor. 5 In fact, it has been replaced by measures of birth weight adequacy:
small (SGA), adequate (AGA), and large (LGA) for GA. SGA newborns have higher
morbidity when compared to AGA newborns, even if they weight > 2,500 g at birth. 7
Therefore, using the adequacy of birth weight for GA seems more reliable than birth weight
alone in predicting risk of morbidity and mortality. 18 Currently, the gold standard for
classification of anthropometric indices at birth are the Intergrowth 21 st standards because
Villar et al showed that W/L better estimates neonatal body composition than BMI. 8 However,
our results showed that being born with a low W/L, that is, being wasted, does not seem to
increase morbidity in the first year of life beyond the neonatal period. This is different from
what was shown in the studies by Rüdiger and Helgertz. 5,6 Possible explanations for this
discrepancy are that the W/L ratio is a more specific measure of wasting, while the LBW
criterion (< 2,500 g at any GA) probably incorporate other prenatal exposures and intrinsic
patient factors, such as preterm birth, IURG, congenital malformations, among others, which
can be more strongly associated with infancy and childhood morbidity than wasting alone.
This study has many limitations. First, not all infants in the birth cohort were included in the
analysis, only those attending to the follow-up visit. However, there were no differences
between those included or not regarding birth characteristics (Supplementary table 1), but

11

small differences in maternal factors. Briefly, non-participant mothers were more frequently
younger than 20 years, had ? 8 or ? 12 years of education, were from social classes A/B or
D/E, were smokers, received less prenatal care, and underwent cesarean deliveries in the
private health system. All these differences, although statistically significant, given the large
sample size, are of little magnitude. Second, the information on hospital admissions was
retrospective and, therefore, can be memory biased.

Conclusion

A low birth W/L for GA and sex was not associated with an increased risk of hospitalizations
during the first year of life beyond the neonatal period in a Brazilian cohort of live births.

References

1. BRAZIL. DATASUS [Internet]. 2019 [citado 6 de agosto de 2019]. Disponível em:
http://datasus.saude.gov.br
2. Ray KN, Lorch SA. Hospitalization of Rural and Urban Infants During the First Year of Life.
Pediatrics. 1 o de dezembro de 2012;130(6):1084–93.
3. Caetano J do R de M, Bordin IAS, Puccini RF, Peres C de A. Fatores associados à internação
hospitalar de crianças menores de cinco anos, São Paulo, SP. Rev Saude Publica. junho de
2002;36(3):285–91.
4. Oliveira KF de, Barbieri MA. The influence of gestational age and perinatal factors on
hospital admissions during the first year of life [Portuguese] [Master’s]. [Ribeirão Preto]:
University of São Paulo; 2017.
5. Rüdiger M, Heinrich L, Arnold K, Druschke D, Reichert J, Schmitt J. Impact of birthweight
on health-care utilization during early childhood - A birth cohort study. BMC Pediatr. 1 o de
março de 2019;19(1).

12

6. Helgertz J, Nilsson A. The effect of birth weight on hospitalizations and sickness absences: a
longitudinal study of Swedish siblings. J Popul Econ. 1 o de janeiro de 2019;32(1):153–78.
7. Hughes MM, Black RE, Katz J. 2500-g Low Birth Weight Cutoff: History and Implications
for Future Research and Policy. Matern Child Health J. 23 de fevereiro de 2017;21(2):283–9.
8. Villar J, Puglia FA, Fenton TR, Cheikh Ismail L, Staines-Urias E, Giuliani F, et al. Body
composition at birth and its relationship with neonatal anthropometric ratios: the newborn
body composition study of the INTERGROWTH-21st project. Pediatr Res. 31 de agosto de
2017;82(2):305–16.
9. Confortin SC, Ribeiro MRC, Barros AJD, Menezes AMB, Horta BL, Victora CG, et al. RPS
Brazilian Birth Cohorts Consortium (Ribeirão Preto, Pelotas and São Luís): history, objectives
and methods. Cad Saude Publica. 2021;37(4).
10. da Silva AAM, Simões VMF, Barbieri MA, Cardoso VC, Alves CMC, Thomaz EBAF, et al.
A protocol to identify non-classical risk factors for preterm births: the Brazilian Ribeirão
Preto and São Luís prenatal cohort (BRISA). Reprod Health. 19 de dezembro de
2014;11(1):79.
11. Christen MC, Altman DG, Egger M, Gøtzsche PC, Pocock SJ, Vandenbroucke JP, et al.
STROBE Statement: STrengthening the Reporting of OBservational studies in Epidemiology
[Internet]. 2023 [citado 23 de janeiro de 2023]. Disponível em: https://www.strobe-
statement.org/
12. Verburg BO, Steegers EAP, De Ridder M, Snijders RJM, Smith E, Hofman A, et al. New
charts for ultrasound dating of pregnancy and assessment of fetal growth: longitudinal data
from a population-based cohort study. Ultrasound in Obstetrics and Gynecology [Internet].
abril de 2008 [citado 23 de março de 2023];31(4):388–96. Disponível em:
https://onlinelibrary.wiley.com/doi/10.1002/uog.5225

13

13. Intergrowth-21st. Intergrowth-21st Standards and Tools [Internet]. 2009 [citado 23 de março
de 2023]. Disponível em: https://intergrowth21.tghn.org/standards-tools/
14. Silva AAM da, Silva LM da, Barbieri MA, Bettiol H, Carvalho LM de, Ribeiro VS, et al. The
epidemiologic paradox of low birth weight in Brazil. Rev Saude Publica [Internet]. outubro de
2010;44(5):767–75. Disponível em:
http://www.scielo.br/scielo.php?script=sci_arttext&pid=S0034-
89102010000500001&lng=en&tlng=en
15. Silveira VMF da, Horta BL. Peso ao nascer e síndrome metabólica em adultos: meta-análise.
Rev Saude Publica [Internet]. fevereiro de 2008 [citado 23 de março de 2023];42(1):10–8.
Disponível em: http://www.scielo.br/scielo.php?script=sci_arttext&pid=S0034-
89102008000100002&lng=pt&tlng=pt
16. Williams MC, O’Brien WF. A comparison of birth weight and weight/length ratio for
gestation as correlates of perinatal morbidity. Journal of Perinatology [Internet]. 1997 [citado
23 de março de 2023];17(5):346–50. Disponível em:
http://www.ncbi.nlm.nih.gov/pubmed/9373837
17. Williams M, O’Brien W. Low Weight/Length Ratio to Assess Risk of Cerebral Palsy and
Perinatal Mortality in Twins. Am J Perinatol [Internet]. 4 de abril de 1998 [citado 23 de março
de 2023];15(04):225–8. Disponível em: http://www.thieme-connect.de/DOI/DOI?10.1055/s-
2007-993931
18. Mayer C, Joseph KS. Fetal growth: a review of terms, concepts and issues relevant to
obstetrics. Ultrasound in Obstetrics & Gynecology [Internet]. 31 de fevereiro de 2013 [citado
23 de março de 2023];41(2):136–45. Disponível em:
https://onlinelibrary.wiley.com/doi/10.1002/uog.11204