• No results found

Growth in 10- to 12-Year-Old Children Born at 23 to 25 Weeks' Gestation in the 1990s: A Swedish National Prospective Follow-up Study

N/A
N/A
Protected

Academic year: 2020

Share "Growth in 10- to 12-Year-Old Children Born at 23 to 25 Weeks' Gestation in the 1990s: A Swedish National Prospective Follow-up Study"

Copied!
16
0
0

Loading.... (view fulltext now)

Full text

(1)

ARTICLE

Growth in 10- to 12-Year-Old Children Born at 23 to

25 Weeks’ Gestation in the 1990s: A Swedish National

Prospective Follow-up Study

Aijaz Farooqi, MDa, Bruno Ha¨gglo¨f, MD, PhDb, Gunnar Sedin, MD, PhDc, Leif Gothefors, MD, PhDa, Fredrik Serenius, MD, PhDa

aDepartment of Pediatrics, Institute of Clinical Sciences, University Hospital, Umeå, Sweden;bDepartment of Child and Adolescent Psychiatry, Institute of Clinical Sciences, University Hospital Umeå, andcDepartment of Women’s and Children’s Health, Section for Pediatrics, Uppsala University, Uppsala, Sweden

The authors have indicated they have no financial relationships relevant to this article to disclose.

ABSTRACT

BACKGROUND.Knowledge of long-term growth of extremely preterm infants in rela-tion to gestarela-tional age is incomplete, and there are concerns regarding their poor growth in early childhood. As part of a longitudinal study of a national cohort of infants born at⬍26 weeks’ gestation (extremely immature), growth development from birth to the age of 11 years was examined, and correlates of growth attain-ment were analyzed.

METHODS.Two hundred forty-seven extremely immature children were born alive from April 1990 through March 1992 in the whole of Sweden, and 89 (36%) survived. Growth and neurosensory outcomes of all extremely immature survi-vors were evaluated at 36 months of age. Eighty-six (97%) extremely immature children were identified and assessed at 11 years of age. In this growth study, 83 extremely immature infants (mean [SD]: birth weight, 772 g [110 g]; gestational age, 24.6 weeks [0.6 weeks]) without severe motor disability were followed up prospectively from birth to 11 years old and compared with a matched group of 83 children born at term.zscores for weight, height, head circumference, and BMI were computed for all children. We also examined gender-specific longitudinal growth measures. Predictors of 11-year growth were studied by multivariate analyses.

RESULTS.Extremely immature children were significantly smaller in all 3 growth parameters than the controls at 11 years. Extremely immature children showed a sharp decline in weight and heightzscores up to 3 months’ corrected age, followed by catch-up growth in both weight and height up to 11 years. In contrast to weight and height, extremely immature children did not exhibit catch-up growth in head circumference after the first 6 months of life. The mean BMI zscores increased significantly from 1 to 11 years in both groups. The mean BMI change between 1 and 11 years of age was significantly larger in extremely immature than in control participants. Extremely immature girls showed a faster weight increase than

www.pediatrics.org/cgi/doi/10.1542/ peds.2006-1069

doi:10.1542/peds.2006-1069

Key Words

growth, extremely immature

Abbreviations

EI— extremely immature VLBW—very low birth weight ELBW— extremely low birth weight CVD— cardiovascular disease NSI—neurosensory impairment HC— head circumference EDD— expected date of delivery SES—socioeconomic status SGA—small for gestational age CI— confidence interval PNS—postnatal steroids

Accepted for publication Jun 8, 2006

Address correspondence to Aijaz Farooqi, MD, Department of Pediatrics, University Hospital, SE-901 85 Umeå, Sweden. E-mail: aijaz. farooqi@pediatri.umu.se

(2)

extremely immature boys, whereas catch-up growth in height and head circumference was similar in these groups. Multiple-regression analyses revealed that pre-term birth and parental height were significant predic-tors of 11-year height, and group status (prematurity) correlated strongly with head circumference.

CONCLUSIONS.Children born at the limit of viability attain poor growth in early childhood, followed by catch-up growth to age 11 years, but remain smaller than their term-born peers. Strategies that improve early growth might improve the outcome.

A

DVANCES IN PERINATOLOGY and neonatology in the1990s have led to a dramatic increase in the survival of extremely immature (EI;⬍26 weeks’ gesta-tion) infants born at the threshold of viability.1–3 Once the survival is more assured, the concern becomes fo-cused on growth and development of these infants. Very low birth weight (VLBW; ⬍1500 g) or extremely low birth weight (ELBW;⬍1000 g) children experience sig-nificant growth failure in their early childhood.4–8 Growth outcomes of adolescent ELBW children have recently begun to appear in the literature. All of these reports document compensatory catch-up in growth pa-rameters up to adolescence.9–12However, ELBW adoles-cents remained significantly shorter and lighter than their control peers in the majority of these studies. In most reports, growth outcomes are presented in relation to the birth weight rather than to gestational age. This raises a possibility of introducing bias, because more mature children with fetal growth restriction are includ-ed.13Incomplete cohorts at later follow-up ages, discrep-ancies between reference populations, and definitions of growth failure make it difficult to interpret the results.

There are concerns that disturbances of growth in intrauterine and postnatal life of preterm infants may have long-term implications for their adult health. It has been hypothesized that adaptations made by the fetus or infant when undernourished induce alterations in me-tabolism and hormonal output, possibly increasing the risk for diabetes and cardiovascular disease (CVD) later in life.14Furthermore, there is accumulating evidence of a risk for the development of CVD later in life in growth-restricted infants who exhibit an accelerated weight gain in childhood.15,16 With the exception of 1 population-based study from the United Kingdom,17 information is lacking on growth outcomes of extremely premature infants born in the 1990s. As part of a longitudinal follow-up study in a national Swedish cohort of EI chil-dren (born at ⬍26 weeks’ gestation), we examined the growth development from birth to the age of 11 years and analyzed correlates of growth attainment at 11 years.

POPULATION AND METHODS

The study participants comprised survivors of a national cohort of 247 consecutive live-born EI (⬍26 weeks’ gestation) infants who were born during April 1990 through March 1992 in the whole of Sweden. Of these 247 EI infants, 89 (36%) survived in the neonatal pe-riod, and all were alive at the time of this study (mean age: 11 years). All of the 89 EI survivors were assessed in their neonatal period and at 36 months’ corrected age when they were enrolled in the “1000-g” national Swedish cohort.18,19 The identification of EI subjects at the present assessment, recruitment of control partici-pants, and other methodologic details have been de-scribed elsewhere20and will be briefly repeated here. Of the 89 EI children, 3 were not assessed (2 refused, 1 was abroad). Thus, 86 EI children (97% of all survivors) were followed up to a mean age of 11 years. Three children with severe motor disability were excluded from this study, which thus comprised 83 children (94% of all survivors). Another 3 were on growth hormone treat-ment and were included in the study only until the start of that treatment (at 4 years 10 months, 3 years 8 months, and 4 years 9 months). Thus, 83 (94% of all survivors) EI children were assessed up to 4 years of age, and 80 (90%) were assessed from age 4 up to 11 years. The control participants in our assessment (mean age: 11 years) were recruited by selecting from the national birth register a healthy term child with normal birth weight born at the same hospital, of the same gender, and nearest in birth date (⫾7 days) to the EI child. Three controls were initially selected for every EI child. Be-cause we aimed to have 1 control for every index child, we initially contacted the first of the control families. If the family did not respond or refused to participate, we then approached the second family and, if necessary, the third. Neurosensory impairment (NSI) at child age of 11 to 12 years was identified by reviewing pediatric case records and records from other specialist health servic-es.20Information about the child’s current health status was obtained by the Nordic Child and Family Health Questionnaire,21which included responses to the ques-tions regarding chronic medical and psychiatric condi-tions diagnosed by a medical specialist or child psychol-ogist. Chronic conditions at 11 years of age were defined to include NSI and medical or psychiatric illness with duration of ⱖ12 months. The Nordic Child and Family Health Questionnaire also provided information about the sociodemographic and socioeconomic backgrounds of the study participants.

Data Collection

(3)

mea-sured the weight at birth of all EI children. Thereafter, weight was measured at repeated intervals up to dis-charge from the hospital. Measurement of head circum-ference (HC) and length of the EI children was post-poned at birth on many occasions, but between the second and fourth weeks of postnatal age, all surviving EI children had their length and HC measurements per-formed; subsequently, the 3 growth parameters (weight, length, and HC) were measured at regular intervals up to discharge from the hospital. Data concerning the child’s condition at birth, birth weight, birth length, and HC were recorded on a special form that was sent to the child care center at which the child was seen at regular intervals up to the age of 6 years. All children were registered at the child health care center. After the hos-pital period in the EI children and from birth in the control participants, the length, weight, and HC were measured during health checks either at the hospital or the child health care center up to the age of 6 years. After that age, measurements of length and weight were taken at school by a school nurse at regular intervals (once or more every year). Growth data were collected from the child health care units, hospital records, and school health services until the children reached 11 to 12 years of age. At this assessment (11–12 years), the weight, length, and HC of all the EI and control partic-ipants were measured by a trained nurse in a standard-ized way. The ages at the time of measurement were corrected for gestational age up to the age of 3 years.

Height was measured as supine length until the child could stand up by himself or herself, which is generally at

⬃2 years of age. After the age of 2 years, height was measured with a stadiometer attached to the wall. All height measure-ments were made to the last completed 0.5 cm.

Weight in the first 2 years of life was measured to the last completed 0.1 kg while the child was naked on a balance scale or an electronic scale. Children from ages 2 to 11 years were weighed while wearing minimal cloth-ing on a digital scale with an accuracy of 0.1 kg.

HC was measured in the maximum fronto-occipital plane using nonextensible plastic-coated tape. After 4 years of age, HC is not usually measured at routine health checkups; therefore, HC records were not avail-able from 4 years of age to the time of this assessment at 11 years. The average numbers of measurements of weight and height from birth to 11 years in EI and control partic-ipants were 34 (range: 15– 45) and 18 (range: 10 –36), respectively. The corresponding values of HC measure-ments from birth to 3 years in the 2 groups were 17 (range: 10 –32) and 6 (range: 4 –9), respectively. Data on parental anthropometry (height and weight), obtained by self-re-ports, were available for 97% of the study population.

Data Analysis

The growth data for each child were examined before they were combined with the data sets of the 2 previous

studies for analysis.18,19Before the statistical analysis was performed, the raw growth data were inspected for re-cording errors. This was performed in 2 stages. First, nonpositive age increments were listed and scrutinized. Second, tentative SD scores were established to look for extreme values in either direction. For determining weight/height at exact predefined ages, fourth-degree polynomial-regression22curves were fitted for each child and variable. Interpolated values were then calculated for the predefined ages in EI and control children.

z scores for weight, height, BMI (kg/m2), and HC were computed relative to Swedish population norms being used in Sweden at the time.23,24Swedish reference values for the calculation of HC zscores were unavail-able from 4 years old onward; therefore, we used the United Kingdom growth chart for computing HCzscores at 11 years of age.25 In EI children, z scores for the growth variables from birth to expected date of delivery (EDD) corresponding to 40 weeks’ gestation were com-puted relative to the Swedish reference data based on estimated fetal weight.26 A z score was calculated by subtracting the expected value for the measurement (weight, height, BMI, or HC) from the child’s actual measurement and dividing by the SD for the measure-ment. Azscore of 0 equals the median (50th percentile), a score of⫹2 SDs approximates the 98th percentile, and a score of ⫺2 SDs approximates the 2nd percentile in normally distributed populations. Subnormal growth was defined as azscore of⬎2 SDs below the mean for the growth variables. The cutoffs for overweight at dif-ferent ages in boys and girls were defined as proposed by Cole et al.27When data were available for both parents, zscores were averaged to obtain a midparental heightz score. Otherwise, the measurement from the sole parent was assumed to represent midparental height.

(4)

repeated in the EI children only to determine if growth was predicted by major neonatal complications (intra-ventricular hemorrhage grade 3 or 4 or peri(intra-ventricular leukomalacia, retinopathy of prematurity stage ⱖ3, bronchopulmonary dysplasia, or necrotizing enterocoli-tis), birth weight, or gestational age after controlling for other explanatory variables such as SES, parental an-thropometric measures, and having a chronic condition at this assessment. The study was approved by the re-gional ethical committee at Umeå University.

RESULTS

Infant data and sociodemographic characteristics are shown in Table 1. Ninety-two percent of the EI children were born at perinatal tertiary care centers. In the 83 EI subjects, the mean (SD) gestational age and birth weight were 24.6 (0.7) weeks and 765 (110) g, respectively. Seventeen percent of the EI children were from multiple pregnancies. The mean (SD) birth weightzscore in the EI infants (⫺0.57 [0.98]) was significantly below the reference mean and 6 (7%) were small for gestational age (SGA, birth weight less than⫺2 SDs).

Sociodemographic characteristics, including mater-nal education and single-parent families, were similar in the 2 groups. Twelve percent of the children lived in a single-parent family. At 11 years of age, 45% of the EI children versus 22% of the controls (P⬍.005; odds ratio: 2.6; 95% confidence interval [CI]: 1.3 to 5.2) re-ported chronic conditions that implied an NSI, a medical or a psychiatric condition lasting forⱖ12 months. The chil-dren in the control group were 8 months older, on average, than the index children because of the slower recruitment process for control participants.

Growth From Birth to 11 Years

Raw data on weight, length, HC, and BMI at various ages are shown in Table 2, and data on parental anthropo-metric measures are shown in Table 3.

Weight

Weight z scores are shown in Table 4. EI children showed a marked drop in weightzscores in the neonatal period, and the scores continued to decline up to 3 months’ corrected age. After 3 months’ corrected age,

TABLE 1 Infant Birth, Neonatal Data, and Chronic Conditions at 11 Years and Sociodemographic Characteristics

EI (n⫽83) Control (n⫽83) Birth and neonatal data

Gestational age, mean (SD), wk 24.6 (0.7) 39.2 (1.6)a

23–24 wk,n 27 NA

25 wk,n 56 NA

Birth weight, mean (SD), g 765 (110) 3523 (606)a

SGA (less than⫺2 SDs),n(%)b 6 (7) NA

Female,n(%) 45 (53) 45 (53)

Multiple births,n(%) 14 (17) NA

In vitro fertilization,n(%) 7 (8) NA

Born at tertiary care center,n(%) 77 (92) 77 (92)

Pregnancy-induced hypertension,n(%) 3 (4) NA

Antenatal corticosteroids,n(%) 26 (31) NA

Surfactant treatment,n(%) 20 (24) NA

Neonatal hospital stay, mean (SD), d 106 (24.1) NA

Mechanical ventilation,n(%) 80 (95) NA

Bronchopulmonary dysplasia,n(%)c 31 (37) NA

Intraventricular hemorrhage (grade 3 or 4)/ periventricular leukomalacia,n(%)

9 (11) NA

Retinopathy of prematurityⱖstage 3,n(%) 22 (26) NA

Necrotizing enterocolitis,n(%) 2 (2) NA

Chronic condition at 11 y (⬎1),n(%)d 35 (45) 18 (22)e

NSI,n(%) 10 (12) 1 (1)e

Medical or psychiatric illness,n(%) 30 (36) 17 (20)e

Sociodemographic data

Single-parent family,n(%) 10 (12) 10 (12)

Maternal education⬍9 y,n(%) 9 (11) 8 (10)

Family income, mean (monthly), US dollars 3730 (1545) 4067 (1580) Age at present assessment, mean (SD), y 10.9 ( 0.7) 11.7 (0.8)a

NA indicates not applicable.

aP.005.

bDerived from Swedish reference population.26

(5)

the z scores in the EI children began to increase and continued to do so, reaching the mean of the reference at ⬃11 years of age. The mean difference in weight z scores between the EI and control participants was sig-nificant at all ages (Table 4); however, it decreased from

⫺2.32 at 3 months’ corrected age to⫺0.39 at 11 years of age. The proportion of EI children with subnormal weight (⬍2 SDs below the mean) increased from 7% at birth to 60% at 3 months’ corrected age, after which there was a reduction in this proportion at later ages (Fig 1A). At 11 years, none of the children in the EI cohort or control participants had subnormal weight. Compared with their male controls, EI boys had significantly lower mean weightzscores at all ages from birth to 11 years, whereas between the girls of the 2 groups, this difference disappeared from 7 years of age onward (Fig 2 A and B). At 11 years of age, the EI boys were 5 kg lighter in weight than their control participants (difference in means: EI boys,⫺4.9 [95% CI⫺8.2 to⫺1.6];P⫽.003).

Length/Height

There were not enough length z scores at birth in EI children to make statistical comparisons worthwhile. Compared with their controls and the reference popu-lation, the EI children had significantly lower height z scores at all ages (Table 5). At EDD the mean (SD) height

z scores in EI children were ⫺1.79 (0.85), declining sharply to the lowest values in the next few months (Table 5). Like the scores for weight, the heightzscores increased after 3 months’ corrected age. EI children showed a significant increase in heightzscores between the ages of 3 months (corrected for prematurity) and 3 years (mean increase: 1.44 [95% CI: 1.18 to 1.71]) and between ages 7 and 11 years (mean increase: 0.28 [95% CI: 0.20 to 0.36]). Between the ages of 3 and 7 years, the height z scores did not change in the EI children but remained fairly constant and significantly below zero. In the control participants, the height z scores did not change significantly from 0 between any consecutive ages. At this assessment, the mean (SD) heightzscores were significantly lower in the EI children than in the control participants (⫺0.53 [1.08] vs 0.10 [0.93], respec-tively; P ⬍ .001). The proportion of EI children with subnormal height increased from 37% at EDD to 62% at 3 months’ corrected age and was subsequently reduced at later ages (Fig 1B). At 11 years, a small and nonsig-nificant proportion of EI children (EI, 6%; control, 1%; P⫽.2) remained subnormal in height. A similar pattern of catch-up growth in heightzscores was observed in EI boys and girls (Fig 2 C and D). At 11 years of age, the EI girls were 3.1 cm and the EI boys 5.7 cm shorter than their contemporary control participants (difference in means: girls, ⫺3.1 [95% CI: ⫺6.2 to ⫺0.09]; P⫽ .04; boys,⫺5.7 [95% CI:⫺8.71 to⫺2.7];P⬍.001).

BMI

Compared with their controls the EI children had signif-icantly lower BMIzscores up to 5 years of age, but by 7 years this difference had disappeared (Table 6). There was a significant increase in the mean BMI z scores between ages 1 and 11 years in both groups (mean increase [SD]: EI, 1.5 [1.15]; control, 0.85 [1.4]). The TABLE 2 Raw Growth Data at Each Age

Age (n) Weight, kg Height, cm HC, cm BMI, kg/m2

EI Control EI Control EI Control EI Control

Birth (EI, 83)a 0.765 (0.11)

EDD/birth (EI, 83; C, 83)b 2.7 (0.4) 3.52 (0.6) 46.7 (1.8) 50 (2.6) 32.9 (1.3) 34.9 (1.4) 12.4 (1.2) 14.0 (1.3) 3 mo (EI, 83; C, 83) 4.7 (0.8) 6.0 (0.8) 56 (2.9) 60.8 (2.5) 39.1 (1.7) 40.5 (1.3) 14.8 (1.5) 16.2 (1.4) 6 mo (EI, 83; C, 83) 6.4 (0.9) 7.7 (0.9) 63.2 (2.8) 66.9 (2.4) 42.2 (1.5) 43.3 (1.4) 16.0 (1.4) 17.2 (1.4) 9 mo (EI, 83; C, 83) 7.6 (0.98) 8.9 (1.1) 68.3 (2.8) 71.5 (2.7) 44.2 (1.5) 45.3 (1.4) 16.3 (1.5) 17.4 (1.3) 12 mo (EI, 83; C,83) 8.6 (1) 9.9 (1.2) 72.7 (2.9) 75.3 (2.6) 45.5 (1.4) 46.6 (1.4) 16.2 (1.3) 17.4 (1.4) 2 y (EI, 83; C, 83) 11.3 (1.5) 12.7 (1.6) 84.8 (3.6) 87.3 (3.2) 48.0 (1.6) 49.2 (1.5) 15.7 (1.5) 16.5 (1.2) 3 y (EI, 83; C, 83) 13.5 (2) 15.0 (1.9) 93.6 (4.1) 96.2 (3.9) 49.2 (1.7) 50.6 (1.3) 15.4 (1.6) 16.1 (1.3) 4 y ( EI, 80; C, 80) 15.3 (2.6) 17.2 (2.4) 100.2 (4.9) 103.9 (4.3) — — 15.2 (1.8) 15.8 (1.3) 5 y ( EI, 80; C, 80) 17.6 (3.1) 19.6 (2.9) 107.2 (5.2) 111.0 (4.5) — — 15.3 (2.0) 15.8 (1.5) 7 y (EI, 80; C, 80) 22.7 (4.7) 24.6 (4.1) 119.9 (6.0) 123.7 (5.4) — — 15.7 (2.4) 16.0 (1.9) 9 y (EI, 80; C, 80) 28.9 (6.6) 31.3 (6.1) 131.2 (6.5) 135.5 (5.8) — — 16.6 (2.8) 16.9 (2.5) 11 y ( EI, 80; C, 80) 36.4 (8.8) 39 (7.6) 142.5 (7.5) 146.9 (6.3) 52.6 (1.5) 54.2 (1.2) 17.8 (3.2) 18.0 (2.8)

Data provided are mean (SD). Age is corrected for preterm birth until 3 years of age. C indicates control; —, not available.

aIndicates birth weight in EI infants.

bIndicates raw data at EDD in EI children and at birth in control participants.

TABLE 3 Parental Anthropometric Measurements

EI Control

n Mean (SD) n Mean (SD)

(6)

mean gain in BMI between 1 and 11 years of age was significantly greater in the EI cohort than in the control participants (difference in mean gain: 0.64 [95% CI: 0.25 to 1.03];P⫽.002). At 11 years of age, EI children as a group were relatively heavy for their height, but their mean BMIzscore was not significantly different from 0 (mean difference: 0.29 [95% CI: ⫺0.005 to 0.59]) or from the controls (mean difference: ⫺0.09 [95% CI:

⫺0.5 to 0.31]). Fifteen percent vs 17% of the EI and control participants, respectively, were overweight ac-cording to the cutoffs for BMI proposed by Cole et al.27In an analysis for gender differences in BMI, the mean BMI zscores for EI girls was found to be comparable to those for their controls from ages 2 to 11 years, whereas the mean BMIzscores for EI boys were significantly lower than those for their control participants up to 7 years of age, reaching the mean of the reference at⬃9 years (Fig 2 E and F). Compared with EI boys, there was a trend toward an increase in the proportion of overweight EI girls at various ages (at 5 years: EI girls vs boys, 24% vs 3%;P⫽.005; at 7 years: EI girls vs boys, 20% vs 5%;P

⫽.1; at 11 years: EI girls vs boys, 20% vs 8%;P⫽.1). In the control participants, no such gender differences were found.

HC

There were not enough HC measurements at birth in EI children to make comparisons. At EDD, the mean HCz scores in EI children were significantly lower than those in their control participants (mean difference:⫺1.1 [95% CI:⫺1.33 to⫺0.87];P⬍.0001). These remained signifi-cantly lower than those in the control participants and in the reference mean at all ages at which comparison was possible (Table 7). The control participants did not change theirzscores from the reference mean between any con-secutive ages. A significantly higher proportion of the EI cohort compared with controls had subnormal HC (⬍2 SDs below the mean) at 11 years of age (EI, 22%; control, 1%; P ⬍ .001) (Fig 1C). In addition, there were significant differences between EI and control children by gender at 11 years of age: the mean HC of EI boys was 2 cm lower than that of their control participants, and it was 1.2 cm lower in EI girls than in their controls (P⬍.001). Unlike the increases in height and weight, the EI children did not show any catch-up growth in HC after the age of 6 months (Fig 2 G and H).

SGA, Growth Deficiency, and Chronic Conditions

At this assessment, the meanzscores of weight, height, and HC for 6 EI children who were born SGA were TABLE 4 WeightzScores at Each Age

Age EI Controls Mean Difference (95% CI)/P

Between the Groupsa

n Mean (SD),

zScore

n Mean (SD),

zScore

Birth (EI infants)b 830.57 (0.98)c NA

EDD/birthd 831.62 (0.76)c 83 0.02 (1.15)1.65 (1.94 to1.35)/.001

⌬Birth–EDDb ⫺1.08 (1.01)

3 mo 83 ⫺2.27 (1.38)c 83 0.05 (1.21)2.33 (2.73 to1.93)/.001

⌬EDD/birth–3 mod ⫺0.65 (0.81) 0.03 (0.92)

6 mo 83 ⫺2.03 (1.24)c 830.06 (1.28)1.97 (2.36 to1.58)/.001

⌬3–6 mo 0.25 (0.75) ⫺0.11 (0.62)

9 mo 83 ⫺1.83 (1.17)c 830.25 (1.22)1.58 (1.95 to1.21)/.001

⌬6–9 mo 0.20 (0.56) ⫺0.19 (0.49)

1 y 83 ⫺1.66 (1.04)c 830.34 (1.13)1.31 (1.65 to0.98)/.001

⌬9 mo–1 y 0.17 (0.49) ⫺0.09 (0.41)

2 y 83 ⫺1.24 (1.02)c 830.34 (1.0)0.91 (1.22 to0.60)/.001

⌬1–2 y 0.41 (0.49) 0.01 (0.55)

3 y 83 ⫺0.98 (1.11)c 830.17 (1.02)0.81 (1.14 to0.48)/.001

⌬2–3 y 0.27 (0.37) 0.17 (0.33)

4 y 83 ⫺0.97 (1.26)c 830.06 (1.14)0.90 (1.27 to0.53)/0.001

⌬3–4 y 0.01 (0.39) 0.11 (0.33)

5 y 80 ⫺0.77(1.35)c 80 0.06 (1.22)0.83 (1.23 to0.43)/.001

⌬4–5 y 0.16 (0.34) 0.12 (0.31)

7 y 80 ⫺0.51 (1.38)c 80 0.07 (1.20)0.58 (0.98 to0.18)/.01

⌬5–7 y 0.27 (0.52) 0.01 (0.43)

9 y 80 ⫺0.27 (1.26) 80 0.23 (1.17) ⫺0.50 (⫺0.88 to⫺0.12)/⬍0.05

⌬7–9 y 0.24 (0.44) 0.17 (0.35)

11 y 80 ⫺0.15 ( 1.22) 80 0.24 (1.10) ⫺0.39 (⫺0.75 to⫺0.03)/⬍.05

⌬9–11 y 0.12 (0.36) 0.01 (0.46)

Age was corrected for preterm birth until 3 years of age. NA indicates not applicable;⌬, change inzscores between ages.

aMean difference (95% CI) by unpaired Student’sttest. bIndicates birth weightzscores in EI infants.

(7)

⫺0.48,⫺0.96, and⫺2.6, respectively. The correspond-ing values in the EI children who were born appropriate for gestational age (n ⫽ 74) were ⫺0.12, ⫺0.5, and

⫺0.99, respectively. Of the 6 EI children who were born SGA, at 11 years of age none had weight zscores less than⫺2 SDs, 1 child had a height zscore less than⫺2 SDs, and 5 had HCzscores less than⫺2 SDs. The mean (SD) zscores in height, weight, and HC in EI children with (n⫽33) and without (n⫽47) any chronic condi-tion did not differ significantly at 11 years of age (height z scores: ⫺0.47 [1.03] vs ⫺0.6 [1.1]; weight z scores:

⫺0.13 [1.1] vs⫺0.17 [1.3]; HCzscores:⫺1.36 [1.0] vs

⫺1.1 [1.1], respectively). Three children had been inves-tigated earlier in childhood for short stature and had been started on growth hormone therapy at 4 to 5 years of age, although none of the children had growth hor-mone deficiency. Before starting growth horhor-mone ther-apy these children had a mean height z score of⫺3.1 (SD: 0.6) at 4 years of age.

Parental Anthropometry

Midparental heightzscores were available for parents of 81 children (98%) in each group (ie, in the EI cohort and control participants). There were no differences in ma-FIGURE 1

Percentage of EI (⬍26 week’s gestation) children with sub-normal weight for age (A; weight below2 SDs), height for age (B; height below⫺2 SDs), and HC for age (C; HC below

⫺2 SDs) at different ages23–26: a comparison with controls

(number of children assessed from birth to 4 years: EI, 83; control, 83; 5–11 years: EI, 80; control: 80). Age is corrected for preterm birth up to 3 years of age.aIndicates proportion of EI

infants with subnormal weight at birth;bIndicates

(8)

FIGURE 2

A–H, Graphs illustrating meanzscores of weight, height, BMI, and HC at each age in EI (⬍26 weeks’ gestation) boys (number of boys assessed, birth to 4 years: EI, 38; control, 38; 5–11 years: EI, 36; control, 36) and girls (number of girls assessed, birth to 4 years: EI, 45; control, 45; 5–11 years: EI, 44; control, 44): comparison with control participants. Age is corrected for preterm birth up to 3 years of age.azscores at birth in EI children;bzscores at EDD in EI children and birthzscores in control participants. The null line is thezscore for the reference

(9)

ternal or paternal height or weight between the 2 groups (Table 3).The difference in the mean height z scores between the EI children and their parents (midparental

height zscore) was significantly below 0 at 11 years old (paired t test, t ⫽ ⫺4.99;P ⬍ .001; mean difference in height zscores: ⫺0.56 [95% CI ⫺0.79 to ⫺0.34]). The TABLE 5 HeightzScores at Each Age

Age EI Control Mean Difference (95% CI)/P

Between the Groupsa

n Mean (SD),

zScore

n Mean (SD),

zScore

EDD/birthb 831.79 (0.85)c 830.21 (1.23)1.58 (1.91 to1.25)/.001 3 mo 83 ⫺2.24 (1.32)c 83 0.02 (1.09)2.26 (2.63 to1.89)/.001

⌬EDD/birth–3 mob ⫺0.45 (0.68) 0.24 (0.97)

6 mo 83 ⫺1.84 (1.18)c 830.13 (1.0)1.71(2.04 to1.37)/.001

⌬3–6 mo 0.40 (0.78) ⫺0.15 (0.64)

9 mo 83 ⫺1.5 (1.15)c 830.1 (1.05)1.38 (1.73 to1.04)/.001

⌬6–9 mo 0.36 (0.56)c ⫺0.04 (0.43)

1 y 83 ⫺1.1 (1.11)c 830.02 (1.0)1.1 (1.38 to0.73)/.001

⌬9 mo–1 y 0.41 (0.50) 0.08 (0.37)

2 y 83 ⫺0.87 (1.12)c 830.09 (0.97)0.78 (1.11 to0.46)/.001

⌬1–2 y 0.20 (0.53) ⫺0.07 (0.50)

3 y 83 ⫺0.80 (1.12)c 830.07 (1.03)0.73 (1.06 to0.39)/.001

⌬2–3 y 0.08 (0.42) 0.02 (0.34)

4 y 83 ⫺0.89 (1.18)c 83 0.001 (1.03)0.92 (1.27 to0.58)/.001

⌬3–4 y ⫺0.13 (0.43) 0.07 (0.30)

5 y 80 ⫺0.86 (1.18)c 80 0.00 (1.02)0.86 (1.20 to0.52)/.001

⌬4–5 y 0.03 (0.30) ⫺0.001 (0.25)

7 y 80 ⫺0.81 (1.16)c 800.07 (1.04)0.74 (1.08 to0.41)/.001

⌬5–7 y 0.05 (0.35) ⫺0.07 (0.26)

9 y 80 ⫺0.74 (1.07)c 800.04 (0.95)0.71 (1.02 to0.39)/.001

⌬7–9 y 0.07 (0.28) 0.04 (0.25)

11 y 80 ⫺0.53 (1.08)c 80 0.10 (0.93)0.63 (0.94 to0.32)/.001

⌬9–11 y 0.21 (0.28) 0.13 (0.28)

Age was corrected for preterm birth up to 3 years of age.⌬indicates change inzscores between ages.

aMean difference (95% CI) by unpaired Student’sttest.

bIndicates heightzscores at EDD in EI children and birth heightzscores in control participants.

cStatistically significant difference from 0 (ie, from thezscore of the reference population) by 1-samplettest.

TABLE 6 BMIzScores at Each Age

Age EI Control Mean Difference (95% CI)/P

Between the Groupsa

n Mean (SD),

zScore

n Mean (SD),

zScore

EDD/birthb 831.04 (0.92)c 83 0.18 (1.08)1.22 (1.53 to0.91)/.0001 1 y 83 ⫺1.22 (0.97)c 830.43 (0.94)c ⫺0.79 (1.08 to0.49)/0.0001

⌬EDD/birth–1 yb ⫺0.18 (1.07)0.61 (1.13)

2 y 83 ⫺1.0 (1.1)c 830.37 (0.86)c ⫺0.63 (0.93 to0.33)/.0001

⌬1–2 y 0.22 (0.67) 0.06 (0.65)

3 y 83 ⫺0.72 (1.23)c 830.18 (0.94)0.54 (0.88 to0.21)/.005

⌬2–3 y 0.28 (0.61) 0.19 (0.48)

4 y 83 ⫺0.52 (1.38)c 830.04 (1.0)0.48 (0.85 to0.11)/.05

⌬3–4 y 0.20 (0.55) 0.14 (0.51)

5 y 80 ⫺0.28 (1.46) 80 0.16 (1.11) ⫺0.43 (⫺0.83 to⫺0.03)/⬍.05

⌬4–5 y 0.23 (0.44) 0.20 (0.44)

7 y 80 0.03 (1.55) 80 0.28 (1.23)c ⫺0.26 (0.69 to 0.17)/NS

⌬5–7 y 0.30 (0.76) 0.12 (0.59)

9 y 80 0.26 (1.47) 80 0.47 (1.36)c ⫺0.21 (0.65 to 0.17)/NS

⌬7–9 y 0.23 (0.60) 0.19 (0.47)

11 y 80 0.30 (1.38) 80 0.43 (1.30)c ⫺0.13 (0.50 to 0.31)/NS

⌬9–11 y 0.03 (0.52) ⫺0.04 (0.63)

Age was corrected for preterm birth up to 3 years of age.⌬indicates change inzscores between ages; NS, not significant.

aMean difference (95% CI) by unpaired Student’sttest.

bIndicates BMIzscores at EDD in EI children and at birth in control participants.

(10)

corresponding values in the control participants were not different from 0 (mean difference:⫺0.09 [95% CI⫺0.28 to 0.09]). A majority of the EI children (94%) were within 2 SDs of their mean midparental heightzscores.

Correlates of Growth

Stepwise regression analyses revealed that the zscores for height and HC at 11 years of age continued to be significantly lower in the EI children than in the control participants after adjustment for midparental height, gender, SES, and having a chronic condition. Midparen-tal height (z score) was a significant predictor of the children’s height, explaining 22% of the variance (␤⫽ 0.545;P⬍.0001), followed by the group status (prema-turity), which explained 8% of the variance (␤⫽0.528; P ⬍.001). Group status was the major determinant of HC (zscore) (R2⫽0.24;␤⫽1.0;P.001), and mid-parental height (zscore) showed a weak but significant correlation with HC, explaining 3% of the variance (␤⫽ 0.25; P⫽ .01). Maternal weight was the only variable that correlated with weight or BMIzscores (for weightz score:R2⫽0.26;␤⫽0.042;P.001; for BMIzscore:R2

⫽0.17;␤⫽0.047;P⬍.001). In the separate analyses of boys and girls, preterm birth was a significant predictor of weight in boys but not in girls. In the subanalysis of the EI cohort only, midparental height and birth weight zscores correlated with height at 11 years of age (Table 8). As observed in the analyses of both groups, the only variable that correlated with weight and BMI z scores was mother’s weight. Birth weight z scores predicted head size. Gender, gestational age, SES, and presence of a chronic condition did not correlate with any of the anthropometric measurements in this assessment.

DISCUSSION

This is the first study (based on gestational age) of long-term growth outcomes in children born EI in the 1990s. We have presented our data on the basis of gestational age, thereby avoiding the confounding bias associated with reporting outcomes in terms of birth weight group-ings.13The data are presented inzscores, which provide a more sensitive estimate of deviation of growth than the use of percentiles or cutoff levels for growth deficien-cy.28,29 In EI children, growth data from birth to EDD were compared with Swedish gender-specific intrauter-ine growth standards.26These latter values are in agree-ment with reference data currently being developed for Swedish infants born preterm.30One of the strengths of our study is the large number of measurements at var-ious time points throughout the study period in both EI and control participants. Because the measurements were not made at exact ages, we performed a polyno-mial-regression analysis to reduce the fluctuation in sample size over various ages,22 and the interpolated values could be used for predefined ages. Other notable strengths of the study are its nationwide composition, prospective follow-up, and high follow-up rate. Further-more, the racial composition was homogeneous (97% Swedish or Nordic), and the SES was similar in EI chil-dren and controls.

Our results reveal that EI children have growth fail-ure in early extrauterine life in comparison with those of normal intrauterine growth during the third trimester. The relative decline in growth parameters continued up to 3 months’ corrected age. Growth data are available from a large population-based prospective national fol-low-up study of 283 infants born at⬍26 weeks’ gesta-TABLE 7 HCzScores at Each Age

Age EI Control Mean Difference (95% CI)

Between the Groupsa

n Mean (SD),

zScore

n Mean (SD),

zScore

EDD/birthb 831.30 (0.71)c 830.20 (0.76)1.1 (1.33 to0.87)/0.001 3 mo 83 ⫺0.97 (1.32)c 83 0.12 (0.93)1.09 (1.44 to0.74)/.001

⌬EDD/birth–3 mob 0.33 (0.83) 0.37 (0.76)

6 mo 83 ⫺1.02 (1.20)c 830.09 (1.01)0.93 (1.28 to0.59)/.001

⌬3–6 mo ⫺0.06 (0.54) ⫺0.21 (0.44)

9 mo 83 ⫺0.96 (1.11)c 830.10 (1.0)0.86 (1.19 to0.53)/.001

⌬6–9 mo 0.06 (0.46) ⫺0.01 (0.39)

1 y 83 ⫺0.95 (1.04)c 830.12 (0.91) 0.83 (1.13 to0.53)/.001

⌬9 mo–1 y 0.01 (0.35) ⫺0.02 (0.28)

2 y 83 ⫺0.92 (1.18)c 830.03 (0.92)0.9 (1.22 to0.57)/.001

⌬1–2 y 0.03 (0.54) 0.10 (0.34)

3 y 83 ⫺1.01 (1.23)c 83 0.06 (0.84)1.07 (1.40 to0.74)/0.001

⌬2–3 y ⫺0.08 (0.33) 0.01 (0.31)

11 y 80 ⫺1.11 (1.06)c 800.03 (0.80)1.09 (1.38 to0.79)/.001

⌬3–11 y ⫺0.10 (0.32) ⫺0.10 (0.31)

Age was corrected for preterm birth up to 3 years of age. HCzscores were derived from Swedish reference values23up to 3 years of age. At 11

years, HCzscores were derived from British growth charts.25indicates change inzscores between ages. amean difference (95% CI) by unpaired Student’sttest.

bindicates HCzscores at EDD in EI children and at birth in control participants.

(11)

tion in 1995 in the United Kingdom, the EPICure study.17At EDD, their meanzscores for weight, height, and HC were⫺1.72,⫺2.49, and⫺0.86, respectively, com-pared with ⫺1.6, ⫺1.8, and⫺1.3 at EDD in our study. These data reveal the similar distributions of growth failure in weight, height, and HC in early extrauterine life. High rates of growth failure were also observed in a recent retrospective study from the United States31 concerning growth outcomes of 24 317 infants at discharge who were born between 23 and 34 weeks’ gestation in 1997–2001 (reported from 124 NICUs). At discharge home, of their subgroup of infants born at ⱕ25 weeks’ gestation (n ⫽ 880), 64%, 84%, and 44% had weight, height, and HC parametersⱕ10th percentile, respectively.

At 24 months’ corrected age, the proportions of in-fants with subnormal growth parameters regarding weight, height, and HC in our study were 20%, 13% and 17%, respectively, compared with 13%, 25%, and 38% in the EPICure study at 30 months’ corrected age.17 Thus, the findings were similar for weight, but the pro-portions of infants with subnormal length and HC were considerably higher in the EPICure study. It is possible that the lower disability rate in our cohort of EI children contributed to the better results for growth of HC and height. However, the relatively small sample size in our study limits the comparison of data. Other investigators have also reported increased rates of growth deficien-cies in early childhood among VLBW or ELBW chil-dren.4–7,10,12

At 11 years of age, the EI children were lighter and shorter than the control participants, and the difference was more marked in boys than in girls. However, a majority of these children (⬎90%) were within 2 SDs of the reference mean for age. Our EI cohort displayed a fast catch-up growth in weight and length from 3 months’ corrected age to 3 years, after which there was

a period of late catch-up growth that continued up to this assessment. Catch-up growth in weight was more rapid than the increase in heightzscores. Others have reported late catch-up growth in weight and height of ELBW or VLBW children in their early teens and in adolescence.7–9,11,12 In all of these studies,7–9,11,12preterm children had significantly lower zscores for all anthro-pometric measures during adolescence compared with their normal birth weight peers, and in some of these cohorts7,12 in which outcomes have been reported at a young adult age, the differences in growth parameters have remained significant.32,33Saigal et al33recently re-ported growth outcomes at a young adult age in 147 (89%) of 166 children with birth weight⬍1001 g (mean gestational age: 27 weeks) who were assessed in a pro-spective longitudinal population-based study. Young ELBW adults (males and females) had significantly lower anthropometric measurements than their gender-specific controls.

At 11 years of age, although the EI children had higher mean weightzscores than mean heightzscores, with a discrepancy of 0.4 SD, their mean BMIzscore was not significantly different from that of the control par-ticipants or the reference population. At this assessment, 15% of our EI cohort would be considered overweight according to age- and gender-specific cutoffs for BMI.27 There was a significant increase in BMI over time in both groups, which is consistent with an increase in body fatness as a child grows. However, there were signifi-cantly larger childhood gains of BMI in EI children than in the control participants. There are reports that low birth weight in combination with accelerated weight gain during childhood is associated with an increased risk of CVD in adult life.15,16,34However, these studies did not specifically address children born extremely pre-term, a group with a dramatic increase in survival in the TABLE 8 Multiple-Regression Analysis of Correlates of Weight and Height, BMI, and HC at 11 Years in EI Children (<26 Weeks’ Gestation;

n80)

Independent Variable Weight,␤coefficient (95% CI)a

Height,␤coefficient (95% CI)a

BMI,␤coefficient (95% CI)a

HC,␤coefficient (95% CI)a

Birth weightzscores 0.09 (⫺0.18 to 0.36) 0.26 (0.003 to 0.52)b 0.03 (0.33 to 0.27) 0.37 (0.12 to 0.62)c Gender (male vs female) ⫺0.14 (⫺0.66 to 0.37) 0.07 (⫺0.39 to 0.53) ⫺0.20 (⫺0.79 to 0.38) ⫺0.19 (⫺0.67 to 0.29) Gestational aged 0.41 (1.04 to 0.22) 0.04 (0.60 to 0.51) 0.54 (1.25 to 0.17) 0.13 (0.71 to 0.46) Major neonatal complications (any vs none)e 0.06 (0.51 to 0.62) 0.18 (0.68 to 0.32) 0.26 (0.38 to 0.90) 0.21 (0.73 to 0.32) Chronic condition (any vs none)f 0.08 (0.44 to 0.59) 0.18 (0.27 to 0.64) 0.03 (0.61 to 0.55) 0.06 (0.53 to 0.42) Midparental height (zscores) 0.24 (⫺0.13 to 0.6) 0.60 (0.27 to 0.93)c 0.12 (0.54 to 0.30) 0.23 (0.12 to 0.57) Maternal weight (kg) 0.04 (0.02 to 0.06)c 0.02 (0.001 to 0.034) 0.05 (0.03 to 0.07)c 0.1 (0.01 to 0.02) Social risk (any vs none)gh 0.48 (0.05 to 1.02) 0.16 (0.31 to 0.64) 0.66 (0.01 to 1.32) 0.01 (0.51 to 0.48)

R2 0.29 0.26 0.29 0.17

aRefers to unstandardized coefficients of the independent variables in the linear-regression model with weight, height, BMI, and HCzscores as the dependent variables at 11 years of age. bP.05.

cP.005.

dGestational age was categorized into 2 groups (23–24 vs 25 weeks’ gestation, with 25 weeks’ gestation group as the reference).

eIntraventricular hemorrhage (grade 3 or 4), periventricular leukomalacia, retinopathy of prematurity (stage3), bronchopulmonary dysplasia, or necrotizing enterocolitis. fAt 11 years includes neurosensory, medical, or psychiatric illness lasting12 months.20

gSocial risk factors recorded at the present assessment (ie, at the child’s 11th year).

(12)

past decade,1–3who exhibit substantial growth failure in early infancy.17,31In a recent retrospective study35it was found that the risk of CVD was more strongly related to the speed of childhood gain in BMI than to the BMI attained at any particular age. Thus, it may reasonably be speculated that the pronounced growth restriction in the postnatal period and in early infancy, with accelerated catch-up growth in childhood, may put these immature infants at risk of metabolic and cardiovascular morbidity in later life, a concern that is shared by others.32,33

EI girls showed more pronounced catch-up growth in weight than EI boys, which is compatible with the greater change in BMI in EI girls than in the EI boys. Furthermore, we observed a trend toward an increased prevalence of overweight in the EI girls compared with the EI boys at different ages. A greater increase in BMI in preterm girls at adolescence12and in young adult VLBW girls32has also been reported. As speculated by others,32 the gender differences in growth probably have multi-factorial causes. The greater susceptibility of VLBW boys to neonatal complications has been well described.36In our population the rates of major neonatal complications (intraventricular hemorrhage grade 3– 4 or periventricular leukomalacia, retinopathy of prematurity stageⱖ3, bron-chopulmonary dysplasia, and necrotizing enterocolitis) and the length of neonatal hospital stay were similar in the EI boys and EI girls. However, EI girls had lower rates of chronic conditions than EI boys (33% vs 51%;P⫽.11), but the difference was not significant.

At this assessment, the mean HC zscores in EI boys and girls were 1.3 and 0.89 SDs lower than their con-temporary control participants, respectively. These re-ductions amount to 2 and 1.3 cm in EI boys and girls, respectively. In contrast to the increase in weight and height, our EI cohort did not show any catch-up growth in HC after the first year of life. Our findings are in agreement with others,37–39 which strengthens the evi-dence that the catch-up growth mostly occurs during the first year of life. Furthermore, 21% of the EI children had a subnormal HC at 3 years’ corrected age, and in a similar proportion (22%) HC remained subnormal at 11 years of age. Similar reductions in head-growth attain-ment have been reported in studies of adolescent growth outcomes in VLBW or ELBW children.8,11,12,40As in the studies by Peralta-Carcelen et al11 and Peterson et al,41 we found that the HC was significantly lower in EI children who were born SGA than in those who were appropriate for gestational age. Subnormal head size has been associated with poor developmental outcomes in preterm children.39–43

There were no differences in parental height and weight between our EI and control groups, which is in agreement with reports from other investigators.11,12 Al-though the EI children had significantly lower heightz scores compared with their midparental height by age 11 (mean difference: 0.57 SD), a majority of them (⬎90%)

were within 2 SDs of their mean midparental height. The proportion of children with subnormal height de-creased from 15% to 6% within a period of 2 years (ie, from 9 to 11 years of age). It is likely that some of our EI children had entered puberty. However, we did not col-lect data on pubertal development, and bone ages were not measured. Some investigators have found no differ-ences in sexual maturation rates by gender in children who had ELBW8,11 or VLBW,40 in comparison to term control children. A few studies have reported an ad-vanced bone age in VLBW40 and ELBW adolescents11 with reference to their chronological age and speculated that this may contribute to shorter height in adulthood in preterm children. The final adult stature of extremely preterm infants born in the modern era of perinatal care remains to be determined.

Within the multivariate models, differences in height and head size persisted between EI and control partici-pants when controlling was performed for other explan-atory variables. Of all the tested variables, midparental height and group status predicted height at 11 years old, whereas the major correlate of HC was the group status. In the subanalysis of EI children only, parental height was the major determinant of height, emphasizing the strong genetic influence on the growth. In the EI cohort, birth weight for age correlated with height, but the association was weak. Head size was influenced by the birth weight scores in EI children. In fact, in the majority of the EI children (83%) who were born SGA, the head size remained subnormal at 11 years of age. However, our study does not provide useful data on long-term growth after intrauterine growth impairment, because only 7% of the EI children were below 2 SDs for the mean weight at birth. A number of studies have identi-fied significant correlates of growth and catch-up growth among VLBW or ELBW infants in childhood and ado-lescence. Intrauterine growth restriction has a negative effect on the growth and catch-up growth during child-hood and into adolescence.7,11,44,45 Neonatal complica-tions have been shown to bear a negative relation to growth during early infancy and childhood,17,46,47 and parental size is positively related to growth parameters in childhood and adolescence.7,10–12,44,48

(13)

3-year follow-up of our EI cohort,18,19it was brought to our knowledge that PNS treatment of chronic lung dis-ease in preterm infants was not a common practice in Sweden in the beginning of the 1990s. However, we do not possess the exact information in this regard.

As suggested by Karlberg,50growth from birth to ma-turity has been described as occurring in 3 additive phas-es: infancy, childhood, and puberty (ie, the “ICP model”). The infancy component, the mechanism of which is not well understood, seems to onset at approx-imately midgestation and continues with a decelerating influence up to ⬃3 to 4 years of age. The childhood component starts during the first year of life, has a continuously slow decelerating course, and does not disappear until adult size is attained. It may reasonably be speculated, as suggested by others,48 that preterm birth disrupts the normal regulation of the infancy growth period, resulting in poor growth in the first years, whereas the childhood growth period remains well preserved. Early growth failure might be the result of disruption from the intrauterine physiologic environ-ment and complications in the neonatal period. Interac-tion of the nutriInterac-tional and endocrine factors that govern early growth in infants born prematurely is not perfectly understood. Differences in levels of many circulating hormones between preterm and term infants have been reported,51–54linking them to the early growth failure in premature infants.54 We need to address the specific issues of whether postnatal growth failure is related to nutritional factors and how fast these infants should grow (including catch-up growth). This might enable us to understand whether the postnatal growth failure and its associated poor developmental performance can be moderated by dietary means.

CONCLUSIONS

Our group of preterm children who were born at the limit of viability attained poor growth in their postnatal period and early childhood. This was followed by catch-up growth up to the age of 11 years; nevertheless, the children remained smaller than their term-born peers. It is possible that the growth outcomes reported for these children born in the early 1990s may not be relevant for current survivors in view of the significant advances in the intensive care of extremely preterm infants in the past 15 years, which include greater awareness to ensure optimal nutrition in the neonatal period and during infancy. It is not clear whether chil-dren born extremely preterm are expected to follow growth trajectories similar to those of their term peers, but the severe growth failure exhibited by the children in our EI cohort in their early postnatal life is unequiv-ocal. We believe that by optimizing nutrition in the neonatal period and in infancy, health and growth out-comes may be improved.

ACKNOWLEDGMENTS

This Study was financially supported by the Os-karfonden Foundation, the Sven-Jerrings Fond Founda-tion, and the Kempe-Carlgren’s Fund.

We thank research nurse Margareta Backma¨n (Umeå) and project assistant Nighat Farooqi (Umeå) for assistance in collecting data and establishing an invalu-able contact with the families. We also thank Dr Hans Stenlund in Umeå for statistical advice. We are indebted to the children and their families for cooperation.

REFERENCES

1. Lorenz J. The outcome of extreme prematurity.Semin Perinatol.

2001;25:348 –359

2. Serenius F, Ewald O, Farooqi A, Holmgren PÅ, Håkansson S, Sedin G. Short-term outcome after active perinatal manage-ment at 23–25 weeks gestation: a study from two Swedish tertiary care centers. Part 2: infant survival.Acta Paediatr.2004; 93:1081–1089

3. El-Metwally, Vohr B, Tucker R. Survival and neonatal morbid-ity at the limits of viabilmorbid-ity in the mid 90s: 22–25 weeks.

J Pediatr.2000;137:616 – 622

4. Daily DK, Killbride HW, Wheeler R, Hassanein R. Growth patterns of infants weighing less than 801 g at birth to 3 years of age.J Perinatol.1994;14:454 – 460

5. O’Callaghan MJ, Burns Y, Gray P, et al. Extremely low birth weight infants and controls: a comparison of intellectual abil-ities, motor performance, growth and health.Early Hum Dev.

1995;40:115–125

6. Kitchen WH, Doyle LW, Ford GW, Callanan C. Very low birth weight and growth at 8 years. I: weight and height.Am J Dis Child.1992;146:40 – 46

7. Hack M, Weissman B, Borawski-Clark E. Catch-up growth during childhood among very low-birth-weight children.Arch Pediatr Adolesc Med.1996;150:1122–1129

8. Ford GW, Doyle LW, Davis NM, Callanan C. Very low birth weight and growth into adolescence.Arch Pediatr Adolesc Med.

2000;154:778 –784

9. Hirata T, Bosque E. When they grow up: the growth of ex-tremely low birth weight (⬍or⫽1000 g) infants at adoles-cence.J Pediatr.1998;132:1033–1035

10. Doyle LW. Growth and respiratory health in adolescence of the extremely low-birth weight survivor. Clin Perinatol. 2000;27: 421– 432

11. Peralta-Carcelen M, Jackson DS, Goran MI, Royal SA, Mayo MS, Nelson KG. Growth of adolescents who were born at extremely low birth weight without major disability.J Pediatr.

2000;136:633– 640

12. Saigal S, Stoskopf BL, Streiner DL, Burrows E. Physical growth and current health status of infants who were of extremely low birth weight and controls at adolescence.Pediatrics.2001;108: 407– 415

13. Arnold CC, Kramer MS, Hobbs CA, McLean FH, Usher RH. Very low birth weight: a problematic cohort for epidemiolog-ical studies of very small or immature neonates.Am J Epidemiol.

1991;134:604 – 613

14. Barker DJP.Mothers, Babies and Health in Later Life. Edinburgh, Scotland: Churchill Livingstone; 1998

15. Eriksson JG, Forsen T, Tuomilehto J, Winter PD, Osmond C, Barker DJP. Catch-up growth in childhood and death from cor-onary heart disease: longitudinal study.BMJ.1999;318:427– 431 16. Lucas LA, Fewtrell MS, Cole TJ. Fetal origins of adult disease:

hypothesis revisited.BMJ.1999;319:245–249

(14)

growth and associated problems in children born at 25 weeks of gestational age or less. Arch Dis Child Fetal Neonatal Ed.

2003;88:F492–F500

18. Finnstro¨m O, Otterblad-Olausson P, Sedin G, et al. The Swed-ish national prospective study on extremely low birth weight (ELBW) infants: incidence, mortality, morbidity and survival in relation to level of care.Acta Paediatr.1997;86:503–511 19. Finnstro¨m O, Otterblad-Olausson P, Sedin G, et al.

Neurosen-sory outcome and growth at three years in extremely low birthweight infants: follow-up results from the Swedish na-tional prospective study.Acta Paediatr.1998;87:1055–1060 20. Farooqi A, Ha¨gglo¨f B, Gothefors L, Sedin G, Serenius F.

Chronic conditions, functional limitations, and special health care needs in 10- to 12-year-old children born at 23 to 25 weeks’ gestation in the 1990s: a Swedish national prospective follow-up study. Pediatrics. 2006;118(5). Available at: www.pediatrics.org/cgi/content/full/118/5/e1466

21. Berntsson L, Ko¨hler L.Health and Wellbeing of Children in Five Nordic Countries in 1984 and 1996[doctorial thesis]. Gothenburg, Sweden: Nordic School of Public Health; 2000

22. Royston P. Constructing time-specific reference ranges. Stat Med.1991;10:675– 690

23. Albertsson-Wikland K, Luo ZC, Niklasson A, Karlberg J. Swed-ish population based longitudinal reference values from birth to 18 years of age for height, weight and head circumference.

Acta Paediatr.2002;91:739 –754

24. Karlberg J, Luo ZC, Albertsson-Wikland K. Body mass index reference values for Swedish children.Acta Paediatr. 2003;92: 648 – 652

25. The Child Growth Foundation.British 1990 Growth Reference for Height, Weight, BMI and Head Circumference. London, United Kingdom: Child Growth Foundation; 1996

26. Marsal K, Persson PH, Larsen T, Lilja H, Selbing A, Sultan B. Intrauterine growth curves based on ultrasonically estimated foetal weights.Acta Paediatr.1996;85:843– 848

27. Cole TJ, Mary C, Bellizzi CM, Flegal MK, Dietz WH. Establish-ing a standard definition for child overweight and obesity worldwide: international survey.BMJ.2000;320:1240 –1243 28. Physical status: the use and interpretation of anthropometry.

Report of a WHO expert committee:World Health Organ Tech Rep Ser. 1995;854:1– 452

29. Shann F. Nutritional indices: Z, centile or percent? Lancet.

1993;341:526 –527

30. Niklasson A, Karlberg P. New reference for Swedish infants’ size at birth.Horm Res.1999;51(suppl 2):1–153

31. Clark RH, Thomas P, Peabody J. Extrauterine growth restric-tion remains a serious problem in prematurely born preterm neonates.Pediatrics.2003;111:986 –990

32. Hack M, Schluchter M, Carter L, Rahman M, Cuttler L, Bo-rawski E. Growth of very low birth weight infants to age 20 years.Pediatrics.2003;112(1). Available at: www.pediatrics.org/ cgi/content/full/112/1/e30

33. Saigal S, Pinelli J, Stoskopf B, et al. Comparison of growth of ELBW survivors and NBW from birth to young adulthood.

Pediatr Res.2005;105:57A

34. Ivring RJ, Belton NR, Elton RA, Walker BR. Adult cardiovas-cular risk factors in premature babies [published correction appears inLancet. 2000;356:514].Lancet.2000;355:2135–2136 35. Barker DJP, Osmond C, Forsen TJ, Kajantie E, Eriksson JG. Trajectory of growth among children who have coronary events as adults.N Engl J Med.2005;353:1802–1809

36. Stevenson DK, Verter J, Fanaroff AA, et al. Sex differences in

outcomes of very-low birthweight infants: the newborn male disadvantage.Arch Dis Child Fetal Neonatal Ed.2000;83:F182–F185 37. Hack M, Breslau N. Very low birth weight infants: effect of brain growth during infancy on intelligent quotient at 3 years of age.Pediatrics.1986;77:196 –202

38. Hack M, Breslau N, Fanaroff AA. Differential effects of intra-uterine and postnatal brain growth failure in infants of very low birth weight.Am J Dis Child.1989;143:63– 68

39. Hack M, Breslau N, Weissman B, Aram D, Klein N, Borawski E. Effect of low birth weight and subnormal head size on cogni-tive abilities at school age.N Engl J Med.1991;325:231–237 40. Powls A, Botting N, Cook RWI, Pilling D, Marlow N. Growth

impairment in very low birthweight children at 12 years: cor-relation with perinatal and outcome variables.Arch Dis Child Fetal Neonatal Ed.1996;75:F152–F157

41. Peterson J, Taylor HG, Minich N, Klein N, Hack M. Subnormal head circumference in very low birth weight children: neona-tal correlates and school age consequences. Early Hum Dev.

2006;82:325–334

42. Teplin SW, Burchinal M, Johnson-Martin N, Humphry RA, Kraybill EN. Neurodevelopmental, health, and growth status at age 6 years of children with birth weights less than 1001 grams.

J Pediatr.1991;118(5):768 –777

43. Kitchen WH, Doyle LW, Ford G, Callanan C, Rickards AL, Kelly E. Very low birth weight and growth at 8 years. II: Head dimensions and intelligence.Am J Dis Child.1992;146:46 –50 44. Fewtrell MS, Cole TJ, Bishop JN, Lucas A. Neonatal factors

predicting childhood height in preterm infants: evidence for a persisting effect of early metabolic bone disease.J Pediatr.2000; 137:668 – 673

45. Qvigstad E, Verloove-Vanhorick SP, Ens-Dokkum MH, et al. Pre-diction of height achievement at five years of age in children born very preterm or with very low birth weight: continuation of catch-up growth after two years of age.Acta Paediatr.1993;82: 444 – 461

46. Ehrencrantz RA, Younes N, Lemons JA, et al. Longitudinal growth of hospitalized very low birth weight infants.Pediatrics.

1999;104:280 –289

47. Hack M, Merkatz IR, McGrath SK, Jones PK, Fanaroff AA. Catch-up growth in very-low-birth-weight infants: clinical cor-relates.Am J Dis Child.1984;138:370 –375

48. Niklasson A, Engstro¨m E, Hård AL, Albertsson-Wikland K, Hellstro¨m A. Growth in very preterm children: a longitudinal study.Pediatr Res.2003;54:899 –905

49. Gibson AT, Pearse RG, Wales JK. Growth retardation after dexamethasone administration: assessment of knemometry.

Arch Dis Child.1993;69:505–509

50. Karlberg J. On the modeling of human growth.Stat Med.1987; 6:185–192

51. Rajaram S, Carlson SE, Koo WW, Rangachari A, Kelly DP. Insu-lin-like growth factor (IGF)-1 and IGF-binding protein 3 during the first year in term and preterm infants.Pediatr Res.1995;37: 581–585

52. Wollmann HA. Growth hormone and growth factors during perinatal life.Horm Res.2002;53:50 –54

53. Yeung MY, Smyth JP. Hormonal factors in the morbidities associated with extreme prematurity and the potential benefits of hormonal supplementation.Biol Neonate.2002;81:1–15 54. Cavazzoni E, Gill M, Wales JKH, Clayton PE, Gibson AT. The

(15)

DOI: 10.1542/peds.2006-1069

2006;118;e1452

Pediatrics

Aijaz Farooqi, Bruno Hägglöf, Gunnar Sedin, Leif Gothefors and Fredrik Serenius

1990s: A Swedish National Prospective Follow-up Study

Growth in 10- to 12-Year-Old Children Born at 23 to 25 Weeks' Gestation in the

Services

Updated Information &

http://pediatrics.aappublications.org/content/118/5/e1452

including high resolution figures, can be found at:

References

http://pediatrics.aappublications.org/content/118/5/e1452#BIBL

This article cites 48 articles, 11 of which you can access for free at:

Subspecialty Collections

sub

http://www.aappublications.org/cgi/collection/fetus:newborn_infant_ Fetus/Newborn Infant

following collection(s):

This article, along with others on similar topics, appears in the

Permissions & Licensing

http://www.aappublications.org/site/misc/Permissions.xhtml

in its entirety can be found online at:

Information about reproducing this article in parts (figures, tables) or

Reprints

http://www.aappublications.org/site/misc/reprints.xhtml

(16)

DOI: 10.1542/peds.2006-1069

2006;118;e1452

Pediatrics

Aijaz Farooqi, Bruno Hägglöf, Gunnar Sedin, Leif Gothefors and Fredrik Serenius

1990s: A Swedish National Prospective Follow-up Study

Growth in 10- to 12-Year-Old Children Born at 23 to 25 Weeks' Gestation in the

http://pediatrics.aappublications.org/content/118/5/e1452

located on the World Wide Web at:

The online version of this article, along with updated information and services, is

by the American Academy of Pediatrics. All rights reserved. Print ISSN: 1073-0397.

Figure

TABLE 1Infant Birth, Neonatal Data, and Chronic Conditions at 11 Years and SociodemographicCharacteristics
TABLE 2Raw Growth Data at Each Age
TABLE 4Weight z Scores at Each Age
FIGURE 1Percentage of EI (�26 week’s gestation) children with sub-
+5

References

Related documents

Poor sleepers experienced more nocturnal distur- bances, including disturbed sleep (difficulties falling asleep and staying asleep) and motor symptoms at night (restlessness of

Patients were evaluated using the 21-item Hamilton Depression (HAM-D) rating scale and the Symptom Check List-90-Revised (SCL-90R) before and after REAC brain

Clinical bene fi t from afatinib in an advanced squamous cell lung carcinoma patient harboring HER2 S310Y mutation: a case report. Onco

decreased the levels of Bcl-2, p-Akt and p-ERK, and increased the levels of Bax and active caspase 3 in SW480/OXA cells and HT29/OXA cells, which were further enhanced in the

The survival rate of LAPTM4B knockdown radioresistant cells was decreased, and apoptosis was increased after irradiation, suggesting that LAPTM4B may affect

Consistently, the results of trans-well experiments demonstrated that miR-4324 mimic signi fi cantly impaired the migration activity and invasion potential in both TE-1 and Eca109

Afatinib versus erlotinib for second-line treatment of Chinese patients with advanced squamous cell carcinoma of the lung: a subgroup analysis of the phase 3 LUX-Lung 8 trial..

Furkan Dincer et.al [5] tells about a new type of Metamaterial absorber is designed, characterized and analyzed for better absorption rate in the visible range and