Unfavorable outcomes associated with preterm birth are well described [4], however most studies assess preterm birth as a dichotomous variable based on the cutoff point of 37 weeks, without considering the differences that may exist between a 35-week and a 37-week infant. Few studies have addressed the consequences of late preterm in comparison with term births. Although health problems are not as evident in infants born 34-36 weeks as those born before 32 weeks, late preterm children are still at greater risk of health problems than those born at term [5].

In the first 24 hours after birth, children were examined by trained field workers supervised by a pediatrician. Gestational age was estimated using an algorithm proposed by the National Center for Health Statistics (NCHS) [7] based on the last menstrual period [8]. If the birthweight, length and head circumference were inconsistent with the normal curves for the gestational age calculated, or if the date of the last menstrual period was unknown, then gestational age was determined using the Dubowitz method [9] which was performed on almost all newborns. Preterm birth was defined as birth at <37 weeks of gestational age. Preterm births of between 34/0 and 36/6 weeks of gestational age were defined as late preterm births. Births with unknown gestational age accounted for 0.3% (n = 13) of all births. Intra-uterine growth restriction (IUGR) at birth was defined according to Williams' curves for birth weight for gestational age and sex [10]. Newborns with weight for age below the 10^th percentile of the Williams curve were classified as having IUGR and were excluded from the current analyses.

Family income in the month prior to delivery was collected as a continuous variable and quintiles were calculated. Women who were single, widowed, divorced, or lived without a partner were classified as single mothers. Maternal skin colour was classified as white or black/mixed according to the interviewer's observation. The mother's formal education was categorized as 0, 1-4, 5-8 and ≥ 9 completed school years. Maternal age in years at delivery was categorized as 12-15, 16-19, 20-34 and ≥ 35 years. Smoking habits during pregnancy were based on maternal self-report. Women who smoked at least one cigarette per day in any trimester of pregnancy were classified as smokers during pregnancy. Type of delivery was classified as vaginal or caesarian section.

Children whose mothers lived in the urban area of Pelotas were visited at home at twelve months and two years of age. Mothers were interviewed by trained interviewers and anthropometric measurements were performed both in the mother and her child. Information on the child's breastfeeding and frequency of symptoms and diseases was collected. Losses to follow-up and refusals to participate in the study were 5.7% and 6.5% for the twelve month and two year old follow-ups, respectively.

Maternal and child weight, as well as maternal height and child length were measured after completing the interview. Children were weighed undressed together with their mothers after the mother had been weighed alone, and the child weight was calculated as the difference between the two measures. Tanita electronic scales (Tanita, Tokyo, Japan) with a 150 kg maximum and 100 g precision were used in the study. Scales were calibrated on a weekly basis using standard weights. Child's length was measured in supine position using an AHRTAG portable infantometer with 1 mm precision, that was custom built from wood for the purpose of the study. Interviewers underwent training and standardization in anthropometric techniques and 5-10% of interviews were repeated by supervisors. Growth indicators (weight-for-age, length-for-age and weight-for-length z-scores) were calculated based on the growth curves published by the World Health Organization using Anthro 2005 software [11]. Children with z-scores below -2 for any of the above indicators were considered to have growth failure. Children with weight-for-age z-scores below -2 were classified as underweight; those with length-for-age z-scores below -2, as stunted; and those with weight-for-length z-scores below -2, as wasted. Weight and length gains between birth and each follow-up visit were calculated by subtracting the initial weight/length from the last weight/length values.

Information about child health problems was collected at the twelve month and two year follow-ups. Any episode of wheezing, medical diagnosis of pneumonia, diarrhea in the last fifteen days before the interview was performed, and hospital admissions (defined as a readmission after being discharged in the post-delivery period) were reported by the mothers.

All analyses were restricted to newborns from singleton pregnancies, with adequate weight for age at birth, that were successfully followed-up either at twelve months or at two years of age. Comparison categories were late preterm (34/0-36/6 weeks of gestational age) and term children (37/0-42/6). Crude associations between late preterm birth and each growth indicator (considering each of the growth indicators as individual outcomes) were explored using the chi-squared test. Multivariable analyses were carried out using logistic regression for growth failure and linear regression for weight and length gains. Potential confounders of the association between late preterm birth and each outcome were entered in the adjusted analysis. An operational definition of confounding was used, that is, variables that were associated with both the outcome and the predictor of interest, and were not part of the causal chain. The standard significance level of 5% was used, but variables with a p-value ≤ 0.2 were retained in the model in order to control for residual confounding.

Mothers of late preterm children reported shorter breastfeeding duration at the 12 and 24 month interviews than mothers of term children (Table ​(Table2).2). Prevalence of wheezing and hospitalizations from birth up to 12 months of age were higher among late preterm children than among children born at term. From birth to 24 months pneumonia and hospitalization, but not wheezing, were reported more frequently by mothers of late preterm children.

At birth, late preterm children weighed on average 2806.83 (SD: 402.23) grams and measured 46.69 (SD: 2.13) cm in length. The corresponding values for term newborns were 3365.05 (SD: 387.93) grams and 48.98 (SD: 1.91) cm. At 12 and 24 months late preterm children weighted on average 9.40 (SD: 1.22) and 12.00 (SD: 1.58) kilograms, respectively, and measured 73.47 (SD: 2.76) and 85.70 (SD: 3.54) cm in length, respectively. The corresponding values for term children were 9.90 (SD: 1.27) and 12.67 (SD: 1.71) kilograms and 74.75 (SD: 2.90) and 87.11 (SD: 3.48) cm at 12 and 24 months, respectively.

From birth to 12 months, late preterm children gained on average 61.66 (SE: 66.72) grams more in weight and 1.04 (SE: 0.15) cm more in length than term infants. From birth to 24 months late preterm children gained on average 121.31 (SE: 90.81) grams less than term children, whereas their mean gain in length was 0.88 (SE: 0.18) cm higher. Contrary to what was observed with regard to differences in weight gain, differences in length gain were statistically significant in both follow-ups.

Prevalence of growth failure at 12 and 24 months was greater among late preterm children than among term children (Table ​(Table2).2). At 12 months the crude odds of being underweighted, stunted and wasted were respectively, 3.39, 2.72 and 4.61 higher among late preterm than among term children (Table ​(Table3).3). At 24 months, the corresponding odds were twice as high as for term children (3.98, 2.61 and 2.34) (Table ​(Table44).

Adjusted analyses for growth failure, when comparing late preterm with term children at 12 months, accounted for maternal (family income, maternal skin color, maternal schooling, maternal age and height, and smoking during pregnancy) and child characteristics (breastfeeding duration and history of wheezing, pneumonia and hospitalization). Adjusted odds of being underweight, stunted and wasted at 12 months of age among late preterm children were more than twice as high as for term children as indicated by the OR (2.57, 2.35 and 3.98, respectively) (Table ​(Table3).3). Adjusted linear regression analyses confirmed what was observed in the crude analyses, with late preterm children gaining 1.19 cm more in length than term children and with no difference in weight gain (Table ​(Table33).

A key limitation of the study is that most of the confounders studied were self-reported by mothers. Furthermore, in the analyses of late preterm birth and wasting at 12 and 24 months of age, the low number of children born late preterm that were wasted resulted in reduced precision.

Some methodological issues of this study are worthy of being discussed. First, some of the increase in morbidity among late preterm children may be attributable to observation and detection bias, because mothers and medical doctors may be more attentive to monitor symptoms and signs of medical complications and diseases in preterm than in term infants. However, higher hospitalization rates indicate that late preterm children are at increased risk of developing more severe illnesses than term children [12]. A small number of studies have investigated short-term rehospitalization rates (2-4 weeks of discharge from the birth hospitalization) among late preterm children [12-14]. These studies all reported higher rehospitalization rates among late preterm children than in term infants. We were able to locate only two published studies about rehospitalization among late preterm children after the neonatal period [15,16]. Escobar et al found that gestational age of 36 weeks was a predictor of elevated risk for hospitalization within 15 to 182 days after discharge [15]. McLaurin et al showed that late-preterm infants, whether discharged early (<4 days from the date of birth) or late after birth, were almost twice as likely as term infants to have been rehospitalized at some time during their first year of life [16]. Severity of illness leading to hospitalization is also reflected in the increased risk of mortality among preterm as compared to term infants. Previous analyses of our cohort showed that adjusted relative risks for neonatal and infant mortality, respectively, were 5.1 (1.7- 14.9) and 2.1 (1.0, 4.6) times higher for late-preterm children than for term children [17].

Second, in the study of McLaurin et al [16], bronchiolitis due to respiratory syncytial virus and pneumonia were the most common causes of rehospitalization during the first year among both late-preterm and term infants. Although immaturity of the respiratory system may put late preterm children at higher risk of more severe community-acquired pulmonary infections, part of the hospitalizations may be due to Berkson bias since late preterm births were more prevalent among low income families.

Third, it has already been shown that preterm infants and children born small for gestational age exhibit different rates of growth in early childhood [18]. Prematurity and IUGR frequently present together in the same newborn. The exclusion of small-for-gestational age newborns from the current study aimed to prevent this potential confounding effect.

The importance of analyzing the effect of late preterm births on growth indicators is supported by the marked increase in preterm births observed across the three population-based perinatal studies carried out in Pelotas in 1982, 1993 and 2004. Preterm births increased from 6.3% in 1982 to 11.4% in 1993, and 14.7% in 2004 [19]. Nearly two out of three preterm births in 2004 occurred at late-preterm gestations. The present study showed that the nutritional risks expressed in growth failure linked to late prematurity during infancy are extended to early childhood, at least until the age of two years.

Despite growing faster, late preterm children were at increased risk of underweight and stunting during the first two years of life. The strength of the association between late preterm birth and stunting was similar for the first and the second years with an almost two-fold increase in risk among late preterm children as compared to children born at term. Wasting was associated with late preterm birth only in the first year of life, although the small number of wasted children in both follow-ups may be responsible for the lack of association in the second year of life. In fact, only 4 (1.1%) late preterm and 7 (0.3%) term children showed signs of wasting at 12 months. At 24 months, there were 3 (0.8%) and 10 (0.4%) late preterm and term children, respectively, with signs of wasting. We found no studies that have evaluated the long-term nutritional status of late preterm children. Thus, we know neither the prevalence rates nor risk estimates for wasting and stunting in children from other populations. Previous studies suggest that the gastrointestinal tract is likely to be less developed in late preterm children [3,20], which may lead to difficulties in sucking and swallowing, delay in successful breastfeeding, poor weight gain, and dehydration during early postnatal weeks [21-23].

Late preterm children and term children gained on average 9 kg in weight from birth to 24 months of age, with a statistically non-significant advantage for late preterm children (who gained on average 13.23 g more than term children). Mean child length at 24 months was of 85.70 cm for late preterm children and 87.11 cm for term children, indicating that those born at term had on average a higher stature at the age of two years. However, linear growth from birth to 24 months of age was higher among late preterm children than among term children: although still shorter than term children at 24 months, late preterm children grew on average 1.26 cm more than term children.

Late preterm children gained as much as weight as their term counterparts and they grew more linearly than the term children in both follow-ups. Their growth, however, was not enough to prevent underweight and stunting in the first and second year of life. Analyses from the 1982 Pelotas Birth Cohort [18] showed that despite their earlier disadvantage, preterm children gradually caught up with their appropriate birth weight, term counterparts. This catch-up occurred primarily between mean ages 23 and 47 months. As a result, it is possible that the growth failure observed in late preterm children from the 2004 cohort will be reversed in the coming two years.