Objectives. To compare the physical growth, current health status, and utilization of health care resources by extremely low birth weight (ELBW) and control (C) adolescents and to look at changes over time.
Methods. A longitudinal regional cohort study was conducted. Growth measures were converted to z scores on the National Center for Health Statistics growth curves. Information regarding current health status/health care utilization was obtained by parental interviews.
Results. A total of 154 (91%) of 169 ELBW survivors between 12 and 16 years and 125 (86%) of 145 controls participated. Neurosensory impairments were present in 28% of ELBW survivors and 2% of control participants. Mean z scores for both height and weight were below 0 for ELBW survivors (weight: −0.35; height: −0.55) compared with control participants (weight: 0.40; height: 0.28). However, among ELBW survivors, significant catch-up growth occurred in both parameters between age 8 and adolescence but remained stable among control participants. ELBW survivors had a higher prevalence of visual problems (57% vs 21%), seizures (7% vs 1%), developmental delay (26% vs 1%), learning disabilities (34% vs 10%), and hyperactivity (9% vs 2%) and used more specialists and community resources than did control participants.
Conclusions. Although physical growth continues to be compromised and substantial morbidity remains among ELBW survivors at adolescence, there seems to be some catch-up growth, a reduction in the prevalence of acute health problems, and a decrease in the utilization of medical resources.
- extremely low birth weight
- functional status
- chronic problems
- health care resources
Extremely low birth weight (ELBW) and very low birth weight (VLBW) children are reported to have a significantly higher prevalence of poor physical growth,1–3 ill health, and rehospitalizations during midchildhood compared with their peers.4–9 Until recently, there was little information on the growth trajectory and functional limitations of these children to adolescence. The newer studies suggest that in addition to the well-described neurologic and cognitive deficits,10,,11these children continue to have significant health-related morbidity12 and lower attainment of growth parameters.13–17 Because only a few studies are population based9,,12 or have a large enough sample of ELBW survivors, it is important to conduct additional longitudinal research to obtain a true picture of the overall burden of morbidity and to determine the special health care needs of this vulnerable population.
The primary focus of this article is to describe and compare the physical growth, current health status, and utilization of health care resources at adolescence of a regional cohort of infants who were ELBW and matched term controls. A secondary objective is to make longitudinal comparisons between 8 years of age and adolescence to determine whether there are changes in growth and key health variables over time. We hypothesized that the ELBW cohort would continue to be more compromised in physical growth and use more health care resources than their same-age peers at adolescence.
ELBW survivors who were 501 to 1000 g birth weight and born between 1977 and 1982 to residents of a geographically defined region in central-west Ontario were followed longitudinally from birth. The neurodevelopmental outcomes of this cohort were reported previously at age 3,18 5,19 8,20 and adolescence.11 Term control participants were recruited at 8 years of age from a random list obtained through the directors of 2 Hamilton school boards and matched for gender, age, and socioeconomic status (SES) to each individual child (1977–1981 births).20 At the time of assessment, the children ranged in age from 12 to 16 years unadjusted age.
The key demographic variables collected were gender, family status (single parent/2-parent family), birth weight and gestational age, SES, and maternal education. Gestational age was determined by the best estimate obtained through a combination of the mother's last menstrual period, physical examination, and early ultrasound, if available. Small for gestational age (SGA) was defined as birth weight less than the third centile when plotted on the Usher and McLean growth curves.21 SES was defined by the Hollingshead 4-factor index.22 Neurosensory impairments were reported previously in both groups and include cerebral palsy (CP), mental retardation, blindness, deafness, microcephaly, and hydrocephalus.11,,18,20
Anthropometric measurement, including standing height, weight, and head circumference at both 8 years of age and adolescence (unadjusted age), were obtained by standard techniques (balance beam weight scale with height rod). Occipitofrontal head circumference was measured with a disposable paper tape. The z scores were calculated for height and weight measurements with ANTHRO software package (1.01 Edition, Centers for Disease Control and Prevention, Atlanta, GA), using the age- and gender-specific reference population data provided by the National Center for Health Statistics (NCHS) growth charts (for children from birth to 18 years).23 The NCHS growth curves do not provide data on head circumference measures beyond 36 months. We therefore used the gender-specific centile data from the Nelhaus head circumference growth chart.24 Body mass index (BMI) was measured by the relation of body weight to height squared. Formal assessment of sexual maturity was not performed, except for inquiry regarding the age at menarche in girls. The height and weight measurements of the biological parents were obtained by self-report. Ford et al17 reported no significant differences between measured versus estimated height among parents in their study.
Health Status and Health Care Utilization
Information on health status was obtained through parent interview on the same day as the assessments of the adolescents.11 The questionnaire was devised by the authors using items from the Ontario Child Health Study Questionnaire,25 National Health Interview Survey,26 Survey of Disabled Children,27 and other relevant health variables from the literature.6,,7The questionnaire was administered by a nurse clinician and was not validated before use.
Extra Health Care Expenses
Parents were asked whether they had incurred any out-of-pocket expenses as a result of the health of their child, beyond those covered by universal health care in Canada. The recall period for 1-time major expenses was the child's lifetime. For ongoing care and expenses, the time period was the past year. Descriptive information was collected regarding mobility, vision, hearing and speech aids, special furniture, equipment and gadgets, adaptations to home and car, drugs, respite care, equipment maintenance, and expenses related to health care visits (transportation, parking, accommodation). These items were extracted from several sources26,,27 and by consultation from experts in the field.
This study was approved by the Ethics Committee of the Hamilton Health Sciences Corporation, and written informed consent was obtained from the parents of all participants.
Descriptive statistics such as frequency distributions, means, standard deviations, and range were used. The ELBW and control participants were compared as a group with respect to their growth and health outcome variables using Student's t tests and 95% confidence intervals (CIs) to determine differences between means. Bivariate analyses relied on χ2 analyses with odds ratio and 95% CIs when appropriate.28 The McNemar χ2 test was used to determine change over time in the 8-year and adolescence comparisons. Paired sample tests were done to compare growth z scores at age 8 and adolescence.P < .05 was considered significant.
Linear regression analyses using SPSS (SPSS, Inc, Chicago, IL) were performed to identify whether prematurity was a factor in determining growth at adolescence after controlling for child's age, gender, SES, and maternal and paternal height and weight. This analysis was repeated in the ELBW group to determine whether gestational age and/or birth weight predicted growth. Because of multiple testing, the significance level was set at P < .01.
Between 1977 and 1982, inclusive, 179 ELBW children survived to hospital discharge (survival rate: 48%); 10 children subsequently died (4 of whom were severely impaired), leaving 169 survivors. Of these, 8 were lost and 5 refused (4 of 13 of these had neurosensory impairments), and 2 lived too far away to return for assessment. Thus, 154 (91%) of 169 were assessed. Neurosensory impairments were noted in 43 participants (28%): CP (9), isolated bilateral blindness (5), unilateral blindness (6), deafness (1), hydrocephalus (2), profound mental retardation (3), autism (5), and multiple impairments (12).
Among 145 control participants, none died, 10 were lost, 8 refused, and 2 lived too far away (including 1 child with mild impairment). Of 125 (86%) of 145 assessed , 2% had neurosensory impairment (1 had mild CP, and 1 had unilateral deafness).
Except for a higher prevalence of neurosensory impairments among the ELBW cohort, there were no differences between groups in gender, 2-parent families, SES, or maternal education. Mean birth weight of ELBW teens was 835 g (standard deviation [SD]: 124); mean gestation was 27 weeks (SD: 2); 22% of ELBW infants were SGA.21 Mean birth weight of controls was 3401 g (SD: 481), and all were born at term (see Table 1).
As a group, ELBW adolescents were 5.8 cm shorter than term control participants (P < .0001). Height-for-age meanz scores also were significantly lower in the ELBW cohort (P < .0001). Although the overall mean zscores for the ELBW cohort are within 1 SD of the normative means, a small but significantly higher proportion of ELBW adolescents were ≥2 SD below the mean (8.3% vs 0.8%; P < .01) (see Table 2).
The mean body weight measurements of the ELBW cohort were lower than that of control participants by 5.8 kg (P < .0001). Weight-for-age z scores for the ELBW cohort also were significantly lower than that for term control participants (P < .0001). Again, a significantly higher proportion of the ELBW cohort had mean weight-for-age z scores ≥2 SD below the mean (6.2% vs 0.8%; P = .02).
Z scores were unavailable through the NCHS program. The mean head circumference of the ELBW cohort was significantly smaller than that of control participants by 1.8 cm (P < .0001). In addition, a significantly higher proportion of the ELBW cohort had head circumference 2 SD below the mean on the Nelhaus growth curves (15.5% vs 1.6%; P < .0002). There were significant differences between ELBW and control adolescents by gender: the mean head circumference of ELBW adolescents was 1.6 cm lower than in male control participants and 1.9 cm lower than in female control participants (P < .0001).
Sexual maturity was evaluated in girls only. There were no differences between groups in the proportion who had achieved menarche (90% vs 91%), and there were no differences in the mean age at onset of menarche (ELBW: 12 years [SD: 1.1]; control: 12.2 years [SD: 1.1]). ELBW boys and girls were significantly lower than control participants in both height-for-age z scores (differences in means: boys, −0.80 [95% CI: −1.28 to −0.31; P < .001]; girls, −0.85 [95% CI: −1.16 to −0.55; P < .001]) and weight-for-age z scores (differences in means: boys, −0.70 [95% CI: −1.19 to −0.22; P < .01]; girls, −0.64 [95% CI: −0.95 to −0.33; P < .001]). However, a slightly higher proportion of ELBW boys but not girls had height-for-age z scores ≥2 SD below mean compared with their term peers (boys: 12% vs 2% [P = .08]; girls: 5% vs 2% [P = .5]). ELBW boys also had a higher proportion with weight-for-age z scores ≥2 SD below mean, whereas there were no differences in the corresponding figures between ELBW and control girls (boys: 12% vs 2% [P = .08]; girls: 1% vs 0% [P = 1.0]).
SGA and Neurosensory Impairments
There were no significant differences in the mean heights and weights between ELBW adolescents who were SGA or appropriate for gestational age (AGA). Nor were there any significant differences between those with and without neurologic impairments. However, growth data were not available for 6 adolescents with severe impairments.
The mean BMI of the ELBW cohort at 8 years of age was 14.9 (SD: 1.7) compared with 16.8 (SD: 2.4) for control participants (mean difference: −1.9; 95% CI: −2.4 to −1.3; P < .001). At adolescence, the mean BMI was 19.7 (SD: 3.9) for the ELBW cohort and 21.2 (SD: 3.6) for control participants (mean difference: −1.5; 95% CI: −2.5 to −0.6; P = .001). There was a significant increase in BMI between the 2 ages for both groups (P< .001), which is consistent with the increase in body fatness as children grow. The mean BMI change over time was not different between the ELBW cohort and control participants (0.5). Overall the BMI percentiles for the ELBW cohort and control participants were within the normal range at both ages.
When gender differences for BMI were explored, the BMI for both genders of ELBW children were significantly lower than that of control participants (all P < .05) at both age 8 and adolescence. The change in mean BMI between the 2 ages was not statistically different between ELBW and control boys (mean difference: −0.2; 95% CI: −1.3–0.8; P = .7). However, mean BMI change over time was significantly higher for ELBW girls than for control girls (mean difference: 0.7; 95% CI: −0.3–1.7;P = .02).
Parental measurements were obtained from 141 mothers and 140 fathers in the ELBW cohort and 118 mothers and 115 fathers in the control cohort. There were no statistically significant differences in mean maternal or paternal heights or weights between the ELBW and control cohorts (Table 2). The mean height and weight of both parent groups were within 2 cm of the 50th centile for height and 2 kg of the 50th centile for weight based on the 1983 Metropolitan Life Insurance Company Height and Weight tables.
Determinants of Growth
Because some data for parental height and weight were missing, the height/weight regression analysis was performed on a reduced sample of adolescents: for height, sample size = 245 (ELBW = 132, control = 112); for weight, sample size = 236 (ELBW = 128, control = 108).
Both maternal (P < .001) and paternal height (P < .01) were significant independent predictors of height at adolescence. However, parental weight was not found to predict weight at adolescence. When a child's age, gender, SES, and group status, in addition to parental variables, were considered, all variables except for SES were significant predictors of height at adolescence. Significant predictors of weight at adolescence were child's age and group status (P < .001). After controlling for explanatory variables (child's age, gender, SES, and parental height/weight), prematurity was associated with being 6 cm shorter (95% CI: 3.8–8.0 cm) and 7.2 kg lighter than control participants (95% CI: 2.8–7.2 kg).
In the subanalysis of the ELBW cohort only, birth weight and gestational age were not found to be significant predictors of height or weight. As observed in the analysis of both groups, child's age and parental height predicted height at adolescence, and only child's age predicted weight at adolescence.
ELBW adolescents had similar current prevalence of chronic health conditions such as allergies, asthma, or cardiac or renal problems as the control participants. The prevalence of asthma and recurrent bronchitis, which was significantly higher than control participants in the past, was less problematic at adolescence. However, ELBW adolescents continued to experience a higher prevalence of seizures (P = .03). A significantly higher proportion of the ELBW cohort had multiple (≥3) health problems (35% vs 7%;P < .0001). There were no differences in the proportion who required surgery in the past 2 years. Overall, a much higher proportion of ELBW children had surgical procedures in the past (68% vs 36%; P < .00001). The control group had surgery mainly for tonsillectomies and ear tubes, whereas the reasons for surgery among the ELBW cohort were more varied and, in addition to the above, included hernia repairs, eye surgery, cardiac surgery, and shunts for hydrocephalus (Table 3).
Current Functional Limitations
ELBW adolescents had a significantly higher prevalence of functional limitations by parent report in most domains. Developmental delay, clumsiness, emotional problems, learning difficulties, and visual problems were much more prevalent than in control participants. In addition, a third of the ELBW adolescents had limitations in their ability to participate in normal activities at school, extracurricular activities, and/or other activities that are normal for age compared with 9% of control participants (P < .0002). However, only 5% had reduced self-care abilities. The majority of limitations were attributable to chronic conditions, mainly neurosensory impairments. In contrast, control adolescents were limited by short-term problems, mainly musculoskeletal injuries. The proportion with any functional limitation was 81% for the ELBW cohort and 42% for control participants (P < .0001). The mean number of functional limitations per individual child also was higher among the ELBW cohort (ELBW: mean, 2.0 [SD: 1.8]; control: 0.6 [SD: 0.8];P < .0001). School absenteeism was no different between the 2 cohorts (see Table 4).
Utilization of Health Care and Other Resources in the Past 2 Years
A significantly higher proportion of ELBW adolescents were seen by pediatricians, ophthalmologists, ear/nose/throat specialists, occupational therapists, and speech therapists. There were no differences between the cohorts in the proportion of adolescents who visited emergency rooms or who were hospitalized in the past 2 years. There was a similar incidence of hospitalizations in both cohorts at adolescence (7% vs 7%). The mean number of hospitalizations did not differ (ELBW: 2.0 [SD: 1.2]; control: 2.0 [SD: 1.7]). The mean number of hospital days also were similar for both groups (ELBW: 6.0 [SD: 5.0]; control: 6.0 [SD: 6.0]), after exclusion of 3 outliers from the control group who had prolonged hospitalizations (suicide attempt , gastrointestinal problems , and anorexia ) (see Table 5).
In the past 2 years, there were no differences between groups in the utilization of social services such as children's aid, court officials, or counselors for emotional and/or behavioral problems. However, ELBW children used a significantly higher proportion of remedial educational resources than did control participants (P < .0001, odds ratio = 8.0).
More than a third of ELBW adolescents were wearing prescription glasses compared with 10% of control participants (P < .0001). One ELBW and 1 control adolescent wore hearing aids. There were no statistically significant differences in the use of prescription medications between the 2 groups. These medications included antibiotics, Ritalin, anticonvulsants, inhalers, and pain medications. Only 3% of the ELBW adolescents were using mechanical aids and wheelchairs for mobility.
A significantly higher proportion of parents of ELBW adolescents indicated that they had incurred out-of-pocket expenses as a result of their child's health (ELBW: 10%; control: 1%; P < .001). This was primarily for mobility aids for 3 adolescents with multiple impairments (prosthesis , pogo chair , stander , bath board , bathtub railing , wheelchair , braces , special shoes , ramps , van ), hearing aid , visual aids for 6 blind adolescents (special computer , Braille equipment and accessories , tapes/talking books , optical reader , talking calculators , special watches , liquid indicator , white cane , artificial tears ), and other special needs (diapers, pants, mattress covers, and extra laundry costs for 2 adolescents with incontinence) respite care (4). Except for 1 hearing aid, there were no similar expenses among control adolescents.
Comparison Between Age 8 and Adolescence: Growth, Health Status, and Resource Utilization
Growth z scores were compared for the 2 time periods, using paired data (ELBW: 136; control: 121). The ELBW cohort improved significantly in height and more remarkably in weight (mean height-for-age z score at 8 years: −0.69; at adolescence: −0.50 [mean difference: −0.19; 95% CI: −0.30 to −0.06;P < .01]; weight-for-age z score at 8 years: −0.79; at adolescence: −0.30 [mean difference: −0.48; 95% CI: 0.62 to −0.35; P < .001]; Fig 1). The z scores for control participants did not change significantly between age 8 and adolescence for either height (P < .4) or weight (P = 1.0). Although the mean increase in height from age 8 to adolescence did not differ between groups (P = 1.0), weight gain on the ELBW cohort during this period was significantly greater than in control participants (P < .001). Thus, the disproportion of shorter height for weight among the ELBW cohort became more apparent at adolescence (Fig 1).
Health variables from our longitudinal database also were examined for comparability between age 8 and adolescence. Only variables that matched exactly in concept and semantics in the questionnaires were compared. By parent report, the incidence of seizures, asthma, cardiac problems, clumsiness, and developmental delay did not change significantly between the 2 time periods in either group. Despite this stability, a higher proportion of ELBW children were reported to have limitations in everyday activities at adolescence compared with at 8 years of age (31% vs 8%; P < .0001). In the control group, there also was some increase in functional limitations reported for these variables over time (9% vs 1%; P = .002). In both cohorts, there was an increase in the proportion of children who were taking prescription medications at adolescence compared with at 8 years of age (ELBW: 27% vs 16% [P = .001]; control: 18% vs 6% [P = .001]). There also was an increase in both cohorts in fractures at adolescence (ELBW: 16% vs 9% [P = .01]; control: 24% vs 12% [P= .001]). There was a lower incidence of hospitalizations in both cohorts at adolescence compared with at 8 years of age (ELBW:P < .0001; control: P < .0001).
The results of this study confirm our hypothesis that at adolescence, ELBW survivors have lower growth attainment on all parameters; have a higher prevalence of functional limitations in most domains; and use more educational, health care, and other resources than term control participants. These data are derived from one of the largest and oldest population-based cohorts of ELBW survivors who have been followed longitudinally from birth. In addition, information on the same variables was obtained from a sociodemographically matched comparison group, and the attrition rate was low.
Although ELBW survivors continue to be lighter and shorter than their same-age peers, the majority of infants (>90%) were within 2 SD of the mean at adolescence. We demonstrated a catch-up in their growth parameters between age 8 and adolescence, with a significantly greater catch-up in weight than in height. It seems that the catch-up in weight was more remarkable for ELBW girls than for ELBW boys, and this is consistent with the significantly higher change in BMI over time for ELBW girls. This rapid catch-up in weight is not of concern, as the BMI values do not suggest a tendency for obesity and its resultant health consequences.29 A similar increase in height and weightz scores, between age 8 and 14 years, was reported by Ford et al17 for both ELBW and VLBW children. Other investigators also reported lower mean z scores on all anthropometric measurements in ELBW16,,17 and VLBW13,,17 adolescents in comparison with control participants. Premature infants who were SGA had even lower meanz scores than their AGA counterparts.16
At adolescence, the mean head circumference of the ELBW cohort in our study was significantly smaller than same-age control participants, and a higher proportion were below 2 SD of the mean. Similar reduction in head growth attainment has been reported by other authors,13,,16,17 with a greater deficit in girls than in boys.13 However, Ford et al17 observed a much larger increase in the head circumference z scores in girls than in boys between age 8 and adolescence. VLBW children who were SGA had lower head circumference than the AGA group.16Subnormal head circumference has been reported to be associated with poor cognitive function and academic achievement.3,,13
We determined that neither the proportion who had reached menarche nor the mean age at onset of menarche was different between the ELBW and control girls. However, a formal assessment of sexual maturity was not performed. Other investigators reported no differences in sexual maturation rates by gender in infants who were ELBW16,,17and VLBW13 in comparison to term control participants. Powls et al13 reported that the bone age of VLBW adolescents was advanced with reference to their chronological age and speculated that this may contribute to shorter height at adulthood in the premature cohort. Peralta-Carcelen et al16 also reported advanced bone age in ELBW adolescents compared with control participants, after adjusting for covariates such as height, sexual maturity, gender, and race.
Similar to the study by Peralta-Carcelen et al,16 we found no differences in parental height and weight in our study participants. Of all of the variables, parental height and group status were significant predictors of height at adolescence. After controlling for explanatory variables, prematurity was associated with being significantly shorter and lighter than control participants. Ford et al17 reported lower mid-parental height zscores among their VLBW cohorts. By adolescence, the VLBW cohorts had exceeded their parents' height z scores.17Powls et al13 reported lower maternal but not paternal height z scores. Multiple regression analyses revealed that although some of the differences in stature were accounted for by the lower maternal height, the significant differences observed between the groups remained.13
Only 1 previous study addressed the functional limitations and special health care needs of children who were <750 g birth weight compared with larger VLBW and term control participants at adolescence.12 Consistent with this study, we found that as a group, a higher proportion of the ELBW cohort than control participants had functional limitations such as visual, emotional, and behavioral problems; developmental lag; learning disabilities; reduced self-care abilities; and limitation in school and everyday activities. In both cohorts, the proportion with functional limitations as reported by parents was higher at adolescence than at 8 years of age, despite a reduction in acute health problems and relative stability in chronic health conditions. Although we can only speculate, this may be a reflection of greater expectations on the part of the parents for independent functioning of their children at the older age.
On the positive side, adolescents in our study demonstrate a remarkable resilience, and a significant majority rate their quality of life as quite acceptable, even when they have disabilities.30Except for a much higher proportion who wore prescription glasses, only a small number of ELBW adolescents required compensatory aids. Nevertheless, despite universal health care in Canada, significant out-of-pocket expenses were incurred by some parents of ELBW children toward visual aids and equipment for blind children and for special adaptations for those with multiple impairments. Thus, in addition to the emotional toll and family stress caused by chronic childhood illness,31 parents of ELBW children face a further economic burden of expenses related to special health care and other support services.
Several investigators reported that at school age, ELBW children use a disproportionate amount of resources for remedial education, as well as for behavioral and psychological therapy for hyperactivity and emotional problems.6–8,10,11 It is likely that the ELBW cohort will have reduced economic productivity as a result of lower educational attainment, which translates into a greater dependence on the family, health care system, and government resources. It seems from our study and that of Hack et al12 that the higher utilization of medical and social services continues as late as the adolescent years. These data provide useful information and should be incorporated in estimating the overall lifetime economic costs as a consequence of neonatal intensive care.32,,33
Although physical growth continues to be compromised and substantial morbidity remains, this study provides some reassuring information on catch-up growth, reduction in the prevalence of acute health problems, and decrease in the utilization of medical resources by the ELBW cohort at adolescence. It should be noted that the racial composition of our study participants was predominantly white, and they had the additional benefit of access to universal health care. This might limit the generalizability of our findings to other populations. It also is possible that the growth deficits reported may not be applicable to the more recent survivors of neonatal intensive care, as there is now greater awareness to ensure optimal nutrition, both in the neonatal intensive care unit and in the early years after discharge. However, because of the marked immaturity of the current survivors and the continuing high rate of disabilities,34,,35 it is likely that the functional limitations and the high utilization of health and educational resources may not be any different from those reported by us and others.12
Neonatologists, educators, and therapists should be made aware of the high prevalence of long-term morbidity among ELBW survivors. This information is vital for counseling parents and for planning resources for children with special needs. We believe that the approach to improve the outcome of future extremely premature infants should be focused not only on intervention strategies to prevent the complications of prematurity36 but also on comprehensive medical and social support programs after graduation from neonatal intensive care.37,,38
This study was supported by Grant 04447 from the Ontario Ministry of Health (Toronto, Ontario, Canada) and Grant No. ROI HS-08385 from the Agency for Health Care Research and Quality (Rockville, MD).
We thank the adolescents who were born prematurely and those who were born at term and to their parents for taking the time and effort to participate in our ongoing studies. We appreciate the advice of Dr Peter Rosenbaum and John Horsman regarding health care utilization and Drs Stephanie Atkinson and Janice Randall Simpson regarding the growth analyses. We thank Lorraine Hoult for her assistance in conducting the study and Diane Turcotte for typing the manuscript. The support of the Department of Pediatrics, McMaster University, and the Children's Hospital at the Hamilton Health Sciences Corporation is gratefully acknowledged.
- Received December 5, 2000.
- Accepted February 15, 2001.
Reprint requests to (S.S.) McMaster University, 1200 Main St, West, Hamilton, ON L8N 3Z5, Canada. E-mail:
- ELBW =
- extremely low birth weight •
- VLBW =
- very low birth weight •
- SES =
- socioeconomic status •
- SGA =
- small for gestational age •
- CP =
- cerebral palsy •
- NCHS =
- National Center for Health Statistics •
- BMI =
- body mass index •
- CI =
- confidence interval •
- SD =
- standard deviation •
- AGA =
- appropriate for gestational age
- Saigal S,
- Hoult LA,
- Streiner DL,
- Stoskopf BL,
- Rosenbaum PL
- Hack M,
- Taylor HG,
- Klein N,
- Mercuri-Minich N
- Ericson A,
- Källén B
- Saigal S,
- Rosenbaum P,
- Hattersley B,
- Milner R
- ↵Hollingshead AS. Four Factor Index of Social Status. New Haven, CT: Yale University; 1975
- ↵National Center for Health Statistics. NCHS Growth Curves for Children, Birth–18 years, United States. Vital and Health Statistics Series II, No. 65. Hyattsville, MD: Public Health Service; 1977:78–165
- Nelhaus G
- ↵National Center for Health Statistics. National Health Interview Survey. Hyattsville, MD: Vital and Health Statistics; 1981. Series 1041
- ↵Survey of Disabled Children. UK: Office of Population Censuses and Surveys; 1989
- Gardner MJ,
- Altman DG,
- Hill JO,
- Trowbridge FL
- Vohr BR,
- Wright LL,
- Dusick AM,
- et al.
- Ment LR,
- Vohr B,
- Allan W,
- et al.
- Copyright © 2001 American Academy of Pediatrics