Objective. To describe the long-term outcome of very low birth weight infants growing up in poor socioeconomic conditions in Soweto, South Africa.
Methods. A stratified sample of infants weighing <1500 g surviving to hospital discharge was enrolled. Group 1 consisted of 49 infants 1000 to 1499 g who required mechanical ventilation; group 2 consisted of 39 infants 1000 to 1499 g who did not require mechanical ventilation; and group 3 consisted of 25 infants <1000 g (such infants are not routinely ventilated). Growth and neurological status were recorded at follow-up visits at 3, 6, and 12 months' corrected age and the infants were evaluated further using the Bayley scales of infant development between 12 and 18 months.
Results. Fifteen infants died during the period between hospital discharge and 1 year corrected age, and 12 others were lost to follow-up. Although some catch-up growth was noted in the early months, all group means for weight and length were below the 25th percentile at 1 year. Cerebral palsy was diagnosed in nine infants (8 from group 1). Periventricular leukomalacia and/or porencephaly was diagnosed in eight of the nine infants during their initial hospital stay and was also the strongest negative predictor of the Bayley scores. Higher maternal education and better intrauterine growth were associated with higher Bayley scores.
Conclusions. Mortality after hospital discharge in this study cohort was extremely high. However, despite marked differences in socioeconomic conditions and tertiary care facilities, the handicap rates were comparable with recent studies from developed countries, and some of the predictors of handicap, eg, periventricular leukomalacia and porencephaly, were also similar.
Postnatal growth and development of very low birth weight (VLBW) infants (birth weight <1500 g) has been the subject of many studies since the advent of modern neonatal intensive care. However, most of these studies have been conducted in developed countries and little information exists regarding surviving VLBW infants in developing countries. A review of the outcome of VLBW infants, which included all English-language publications from 1959 to 1988, revealed that all published studies from which morbidity rates could be calculated had been done in developed countries.1More recently, Thompson et al2 reported on the outcome of VLBW infants from Cape Town, South Africa, but most of these infants were of mixed race and there is still a paucity of data on the longer-term outcome of surviving black African VLBW infants.
The data from developed countries cannot readily be extrapolated due to a number of factors prevailing in developing countries which may have significant effects on later growth and development. These include: (1) intensive care facilities are often poorly developed, resulting in a different spectrum of survival of VLBW infants; (2) intrauterine growth retardation is much more frequent; and (3) socioeconomic conditions, which are likely to affect both prenatal and postnatal growth and development, are often extremely poor.
The only public sector maternity and neonatal services for Soweto and surrounding areas, which form part of the greater Johannesburg area, are provided by Baragwanath Hospital and a network of seven maternity clinics. With an estimated population of 2 to 3 million, many living in substandard housing or temporary structures, this urban area in recent years has experienced a large influx of rural people. Between 30,000 and 35,000 infants, almost all of whom are black, are born annually within the delivery service; about two-thirds are delivered in the hospital and one-third in the clinics. The number of home deliveries is small, and virtually all VLBW infants born at home who survive are brought to the hospital soon after birth. Thus, apart from a very small number of VLBW infants born in private hospitals, the VLBW infants managed at Baragwanath Hospital can be regarded as representative of the region rather than as a hospital-based sample.
The Baragwanath Hospital neonatal unit admits 600 to 700 VLBW infants annually with an overall survival of just over 60%, the highest mortality being in those <1000 g because, due to the limitations of ventilator facilities, infants weighing <1000 g at birth are not offered mechanical ventilation except under exceptional circumstances. However, infants with birth weight >999 g are ventilated should they require it although, due to continual pressure on ventilator facilities, infants may experience undesirable delays before admission to the neonatal intensive care unit.3,4 Although the in-hospital mortality and morbidity of VLBW infants at Baragwanath Hospital has been well documented, longer-term outcome has not previously been systematically studied because the hospital-based follow-up program, offered to all VLBW “graduates” from the neonatal service, has a high dropout rate.
We therefore initiated a study to examine the growth and development of surviving VLBW infants discharged from the hospital. We describe their outcome up to 18 months' corrected age.
This study was originally planned in conjunction with a much larger study, the Birth to Ten (BTT) cohort study, which is following all infants born in the Johannesburg-Soweto area over a 7-week period during 1990 from birth to age 10 years.5,6 It was clear, however, that insufficient numbers of VLBW infants at highest risk for handicap, ie, those requiring mechanical ventilation and those with birth weight <1000 g, would be enrolled over this short period. Thus, a stratified sample of infants with birth weight <1500 g surviving to hospital discharge representing three predefined subgroups was selected for long-term follow-up. These subgroups are explained below.
Infants 1000 to 1499 g at birth who required mechanical ventilation were consecutively enrolled over a 6-month period.
Infants <1000 g at birth who were not ventilated were consecutively enrolled over a 12-month period.
An extensive questionnaire was administered to each infant's mother. Details of past medical, obstetric, and family history were obtained, together with information regarding the index pregnancy, smoking, alcohol, and drug use. Details of socioeconomic status were obtained that included the highest educational level obtained by the parents, parental occupation or that of the head of the household, family income, the type of housing, and the presence of running water, electricity, and other amenities in the home. Socioeconomic status was further classified according to a locally validated coding system.7
Details of any complications during labor and delivery, Apgar scores, need for resuscitation at birth, the presence and severity of respiratory distress, infections, and subsequent postnatal complications such as patent ductus arteriosus and necrotizing enterocolitis were recorded. In the case of infants requiring mechanical ventilation, additional data such as the oxygenation index at 24 hours of age,8 mean airway pressure, development of pneumothorax, need for pressor support, duration of ventilation, and failed extubations were all recorded. Artificial surfactant was not available in our institution at the time of this study. Infants in groups 1 and 3 had a transfontanelle real-time ultrasound examination of the head during the first 72 hours of life and all infants had ultrasound examinations at 7 to 14 days of age using a sector scanner with a 7.5 MHz transducer as previously described.4 When intraventricular hemorrhage (IVH) or severe periventricular echodensities were noted, ultrasound scans were performed every 1 to 2 weeks until hospital discharge. Wherever possible, a further ultrasound examination was done on all infants at 4 to 6 weeks of age. The most severe grade of IVH was recorded using Papile's classification,9 and later scans recorded the development of post hemorrhagic hydrocephalus and periventricular leukomalacia (PVL).
Birth weight was measured by an electronic scale and head circumference was measured within 72 hours of birth by one of the researchers. Because maternal dates were often not accurately known, gestation was assessed by Ballard scores performed on all infants within 72 hours of birth.10 For each infant a birth weight ratio was calculated and expressed as a percentage of the 50th percentile of Lubchenco's norms for that gestation,11,12 and infants were also given a categorical weight classification according to whether their birthweight was <10th percentile, 10th to 24th percentile, 25th to 50th percentile, or >50th percentile. Weight was subsequently measured twice weekly and head circumference once weekly until hospital discharge. During their hospital stay, VLBW infants received expressed own-mother's milk whenever possible and >90% of mothers provided milk for their infants. This was supplemented with one of the commercially available premature infant formulas. Feeds were adjusted to provide infants with 120 kcal/kg/day.
Infants were scheduled for a follow-up appointment 6 weeks after hospital discharge and then at ages (corrected for prematurity) 3, 6, and 12 months when growth and developmental status were recorded. Infants who did not return for follow-up visits were actively traced by a research nurse. At the 12-month visit, a full clinical and neurodevelopmental examination was performed, and weight, length, and head circumference were measured. Weight- and length-for-age percentiles were calculated using the Nutritional Anthropometry section of the Epi Info program, which is based on World Health Organization international growth reference curves.13 Each infant was assessed using the Bayley scales of infant development between 12 and 18 months' corrected age.14 A combined Bayley score of <70 was regarded as indicating handicap and a score of 70 to 84 was regarded as suspect. Hearing screening was performed at the follow-up clinic by a speech and hearing therapist. Infants in whom a squint was detected or where there was concern regarding vision were referred for a full ophthalmological evaluation.
Comparisons between subgroups were done using analysis of variance if the data were normally distributed or, if not normally distributed, using the Kruskal-Wallis test. A P value of <.05 was regarded as significant, and the Bonferroni correction was used when multiple intergroup comparisons were performed. The χ2and Fisher exact tests were used when appropriate. To determine predictors of weight and length at 1 year, a multiple regression analysis was performed using weight- and length-for-age percentiles at the 1-year visit as the dependent variables and anthropometric and clinical data collected during the initial hospital stay as independent variables. A similar regression analysis was performed to examine the predictors of the Bayley scores using sociodemographic, delivery, and postnatal clinical data as independent variables. Both total Bayley scores and the Bayley mental developmental index (MDI) were tested in this way.
This study was approved by the Committee for Research on Human Subjects of the University of the Witwatersrand, and consent for the study was obtained from the parents of each infant enrolled.
In total, 677 infants <1500 g were admitted to the unit during the year in which enrollment into the study was conducted. Table1 shows the total number of infants in the three categories who were admitted during the year and the numbers enrolled in each group. Forty-nine ventilated VLBW infants were consecutively enrolled in group 1, 39 nonventilated infants 1000 to 1499 g at birth in group 2, and 25 infants with birth weight <1000 g (not ventilated) in group 3. Over the periods of enrollment, only three infants who would have qualified were not enrolled: one mother refused consent for long-term follow-up (group 3) and two infants were missed by the researchers so that important early neonatal data (Ballard scores, head sonars, etc.) were not available (one in group 1 and one in group 3). Of the 113 infants enrolled 90 (79.6%) were born in the hospital, 5 (4.4%) at a clinic, and 18 (15.9%) were born at home but were admitted to the hospital within 12 hours of delivery. Only 38 (33.6%) of the mothers had attended antenatal clinics.
The mean gestational age, birth weight, and head circumference measurements of the three groups can also be seen in Table 1. Although there were no differences in mean birth weight and head circumference between groups 1 and 2, group 1 infants had a significantly lower gestational age than those in group 2. As expected, group 3 infants had a significantly lower mean gestational age, birth weight, and head circumference, and the majority of infants in this group were small for gestational age (SGA) (Table 1). Infants in group 1 required mechanical ventilation for a mean of 11.7 days (range 1–60 days), and total hospital stay for the study infants averaged 50.3 days (range 21 to 158 days) for those with a birth weight of 1000 to 1499 g and 72.1 days (range 54 to 107 days) for those <1000 g. Only eight infants were discharged from hospital later than term corrected age (four in group 1, one in group 2, and three in group 3) and there were no intergroup differences at the age of discharge.
A total of 58% of the families lived in an apartment or house and the remainder lived in a shack or some other form of informal accommodation. Electricity was present in 70% of the homes; only 17% of homes had running hot and cold water although an additional 79% had access only to running cold water either inside or outside the home; 29% had a toilet inside the home, and the same proportion had a telephone in the home. The median household income was R600 (1 US$ = approximately 3 SA Rand in 1990) and median parental education was 7 years of schooling for the mother and 8 years for the father. Only 4% of mothers said they had ever smoked, 89% said they never took alcohol and only 4% admitted to taking alcohol once a week or more. None of the mothers admitted to taking recreational drugs. All infants in this study were black except one, who was of mixed race.
Fifteen infants died during the first year of follow-up and 12 others were lost to follow-up. No intergroup differences were found in the numbers of infants who died or who were lost to follow-up (Table2). Mortality was not affected by whether infants had been born in the hospital or at home. Of the infants who died, six deaths were associated with pneumonia and/or septicemia on subsequent admissions to hospital. Two other infants were admitted to the hospital and died from complications associated with severe protein energy malnutrition and four died at home with a history compatible with a diagnosis of sudden infant death syndrome. Death was associated with severe growth failure and developmental delay due to fetal alcohol syndrome in one infant and the cause of death was unknown in two additional infants. Of those who died, three had PVL/porencephaly identified on cranial ultrasound before hospital discharge (two in group 1 and one in group 3). Of the infants lost to follow-up, none had PVL/porencephaly identified on cranial ultrasound examinations and only one had an IVH larger than grade 2 (grade 3), which resolved without sequelae on follow-up sonars performed in the hospital.
Because a number of infants did not keep scheduled appointments and needed to be traced, the mean ages at which they were seen corresponded to 0.9, 3.4, 7.6, and 13.1 months' corrected age, and not all infants were seen at each time period. The mean weight- and length-for-age percentiles over this period can be seen in Table3. No differences were noted between groups 1 and 2 for weight or length at any age, but group 3 infants were significantly lighter than group 2 infants at a month of age and significantly shorter up to 8 months when compared with the other two groups. There were no longer any significant differences at 13 months of age. Head circumference measurements of group 3 infants were significantly lower than those of the other groups at one month of age but did not differ significantly thereafter (Table 3). Of note was that length measurements (expressed as length-for-age percentiles) were significantly lower on the percentiles than weight measurements for all groups (expressed as weight-for-age percentiles) at every time period up to 13 months of age.
In the multiple linear regression analysis for predictors of both weight- and length-for-age at 13 months, the only significant variable was the rate of increase in head circumference from the time that birth weight was regained until a weight of 1900 g was reached. Weight increase was also significant in the univariate analysis, but was no longer significant in the multivariate analysis. Other variables such as birth weight, gestation, birth weight ratio, time to regain birth weight, etc. showed no association. There was also no association with the length of time that breastfeeding was continued after hospital discharge or the age at which foods other than milk were introduced.
Cerebral palsy (CP) was diagnosed in 9 infants (8 from group 1) although severe truncal and lower limb hypotonia was noted in two other infants (Table 4). Both of the latter infants had been abandoned and were institutionalized. One additional infant with birth weight 700 g, who also had features of fetal alcohol syndrome, developed severe retinopathy of prematurity and was blind. No other infants were diagnosed as having retinopathy of prematurity. Severe hearing loss was not diagnosed in any infant, but testing in some of the infants was inconclusive, especially when it was done at a relatively young age.
Of the nine infants with CP, six also had Bayley scores <70, although both severely hypotonic infants and the infant with retinopathy also had scores <70. A further three infants without signs of CP, all from group 1, had Bayley scores <70. These data are summarized in Table 4. The mean total Bayley, Bayley MDI and psychomotor development index scores for the three groups are shown in Table 5. Group 1 infants had the lowest mean values, but the differences were not significant for the comparison of the total Bayley and Bayley MDI scores. Although analysis of variance showed a significant difference for the Bayley psychomotor development index scores (P = 0.032), the individual comparisons of group 1 with group 2 (P = 0.019) and group 1 with group 3 (P = 0.046) were not significantly different once the Bonferroni correction was applied.
Predictors of Neurodevelopmental Outcome
Of the nine infants diagnosed with CP eight had PVL and/or porencephaly diagnosed on serial ultrasound examinations during their initial hospital stay. Of these, three infants had grade 4 IVH with subsequent posthemorrhagic hydrocephalus and large areas of porencephaly, one of whom required a ventriculo-peritoneal shunt; two infants had grade 3 IVH with subsequent posthemorrhagic hydrocephalus and PVL and both required ventriculo-peritoneal shunts, and three infants had grade 2 IVH but subsequently developed severe PVL. The remaining infant had a small grade 2 IVH, which resolved, and no PVL was detected at the time of the last ultrasound examination at 29 days of age. Of those infants without CP at follow-up, only two had an IVH larger than grade 2 (both grade 3), which resolved without sequelae, although one other infant had a grade 2 IVH and later developed a single anterior 1-cm cyst in the periventricular white matter. No other infants without CP at the 1-year follow-up demonstrated PVL and/or porencephaly on ultrasound. Thus, the finding of PVL/porencephaly on cranial ultrasound had a sensitivity of 89% and a specificity of 99% for predicting CP as diagnosed at 1 year of age.
Using multiple linear regression analysis, the strongest negative predictor of developmental status as measured by the total Bayley scores or the Bayley MDI was the development of cystic PVL or porencephaly during the initial hospitalization of this group of VLBW infants. Other factors significantly associated with a higher Bayley score for the entire group were higher maternal educational level, the adequacy of intrauterine growth as measured by either a higher birth weight ratio or a higher categorical weight classification and fewer total days in supplemental oxygen. None of the many other factors that were analyzed, such as sociodemographic markers, pregnancy-related complications, the many clinical complications in the neonatal period, or growth in hospital (weight gain and increase in head circumference), showed any significant association with Bayley scores. Measurements of weight, length, and head circumference after hospital discharge were also not significantly associated with the Bayley scores.
During the year of study, 66% of infants in our institution with birth weight 1000 to 1499 g survived, although only 24% of those <1000 g at birth survived to hospital discharge. In the United States, neonatal survival for singleton white infants improved from 42% to 85% between 1960 and 1983 for those weighing 1000 to 1499 g at birth, and from 7% to 46% for those <1000 g.15 Similar trends have been seen in other developed countries, although more recent figures show that survival has improved even further in the past decade. Our survival figures for VLBW infants correspond broadly to those experienced in the early 1970's in developed countries.
The infant mortality rate for blacks in the Johannesburg-Soweto area has previously been estimated to be <30/1000,16 although informal settlements and squatter areas were excluded in this survey. The true figure is thus somewhat higher. The Province of Gauteng, which includes Johannesburg, was estimated to have a black infant mortality rate of 43 in 1990, although the estimate for the country is 52/1000 live births.17,18 This compares with 7/1000 for industrialized countries.18 Of the original 113 infants enrolled in this study, 15 were known to have died after hospital discharge, giving a mortality between hospital discharge and 1 year of age of >130/1000. VLBW infants in our area, in addition to having high neonatal mortality, are thus at much greater risk of dying between hospital discharge and 1 year of age when compared with infants of normal birth weight.
The main causes of death related to serious infections in the early months after hospital discharge and probable Sudden Infant Death Syndrome, although, in the latter case the diagnosis was made on history alone. Because no postmortem examinations were done in these cases, other factors may have been responsible for the deaths. No deaths were associated with severe bronchopulmonary dysplasia and PVL was diagnosed by cranial ultrasound in hospital in the case of 3 of the 15 infants who died, not significantly different compared with 8 of the 86 survivors. In addition deaths occurred in the same proportion in all three subgroups followed up. Thus the high mortality could not be related directly to complications suffered in the neonatal period. However, the numbers of infants with bronchopulmonary dysplasia and/or PVL may have been too small to show such an association.
The BTT study, which was taking place concurrent with this study, suggested that the families of VLBW infants came from poorer socioeconomic conditions. For example, 42% lived in informal housing compared with 23% in the overall BTT study, average education of parents of VLBW infants was 1 year less, and family income in the case of VLBW infants was approximately 60% of the average income found in the overall BTT study (De Wet T, unpublished data). Thus, the high mortality after hospital discharge in this group reflects, at least in part, the extremely poor socioeconomic conditions of these families.
The mean birth weight ratios in this study ranged from 87.3% in group 1 to 71.5% in group 3, reflecting a high incidence of suboptimal intrauterine growth. This high rate of intrauterine growth retardation has previously been described in our population19 and is very different from some reported studies from developed countries where the mean birth weight ratios of low birth weight infants are close to 100%.12,20 The highest birth weight ratio was seen in group 1 infants who also had the lowest number of SGA infants. We have shown that better intrauterine growth is associated with a higher incidence of respiratory distress syndrome in our population and it would thus be expected that VLBW infants requiring ventilation would have a higher birth weight ratio.21 Because infants <1000 g at our institution are not usually ventilated, it is not surprising that the majority of those who survived were SGA.
There appeared to be some catch-up growth in the first 3 to 4 months after hospital discharge, but mean weights subsequently declined on the percentiles so that, at 13 months of age, the mean weights were between the 10th and 25th percentiles. Infants in group 3, the majority of whom were SGA, did appear to demonstrate some catch-up growth in relation to the other two groups in that, at 1 year of age, there were no longer any significant anthropometric differences between the groups. Head circumference measurements were close to the 50th percentile for all groups from 3 months' corrected age onward.
The only significant factor associated with weight and length at 1 year was the rate of increase in head circumference during the initial in-hospital stay, which in turn was probably the earliest reflection of the growth potential of these infants in an extrauterine environment. The fact that, throughout the first year of life, the weight of the study infants was significantly higher on the percentiles than the length suggests that postnatal nutrition was not a limiting factor for growth of these infants as a group, although this may have been a factor in isolated cases.
Thus it would seem that this group of infants, coming from extremely poor socioeconomic circumstances and generally suboptimally grown at birth, remained relatively small up to 1 year, especially in terms of length. However, the pattern of postnatal growth demonstrated by this group of VLBW infants is analogous to recent studies reported from developed countries. These studies have generally shown VLBW survivors to be smaller as a group than infants with normal birth weight, especially in those VLBW infants who were SGA at birth.22-25
In this study, 9 (10.5%) of the 86 infants followed were diagnosed as having CP at 1 year. If the rates found in the three groups in our cohort are representative of all infants <1500 g discharged during the year, the majority of whom were 1000 to 1499 g and did not require ventilation, the expected rate of CP would be 5.9% for those surviving to 1 year of age. This compares with the meta-analysis of Escobar et al1 who showed a median CP rate for VLBW survivors of 7.7% (95% confidence interval 5.3 to 9.0), and is also similar to rates shown in other recent studies from developed countries.26-29 It is of note that handicap rates in these studies were similar for those with birth weight 1000 to 1499 g and those <1000 g, and handicap rates also did not differ over time. The high sensitivity and specificity of PVL/porencephaly for the prediction of later CP is in agreement with other studies showing PVL/porencephaly diagnosed on cranial ultrasound to be the best predictor of later CP.30-33 Of the 12 infants lost to follow-up, none had PVL/porencephaly demonstrated on cranial ultrasound, which suggests that the CP rates as calculated above are unlikely to have been higher and may in fact have been lower had there been no losses to follow-up. PVL/porencephaly diagnosed on cranial ultrasound was also the strongest predictor of the Bayley scores at 12 to 18 months of age.
We were unable to show an association between outcome and socioeconomic status using occupational class, income, or any other measures such as quality of housing, conveniences in the home, etc. It is probable that, in a situation such as ours in a typically third-world urban community, the spectrum of socioeconomic status is too narrow to show this association. Maternal education in our study, however, was a predictor of later developmental outcome and is probably a proxy measurement of some of the factors determining socioeconomic status in developed countries.
We have shown previously that a lower birth weight ratio at our institution is associated with a lower incidence of respiratory distress syndrome, but that overall survival was no better than those with a higher birth weight ratio due to the development of other complications, especially infections.21 This study shows that poorer intrauterine growth was associated with lower developmental scores at 12 to 18 months of age. Although most infants with major handicap were from group 1, who had the highest birth weight ratio, this association was nevertheless significant, suggesting that the long-term effects of suboptimal intrauterine growth are more subtle. Thus, although suboptimal intrauterine growth may be associated with a short-term advantage in our population of VLBW infants in the form of less respiratory distress syndrome, the longer-term disadvantages would appear to outweigh the advantages and optimal intrauterine growth should be the goal.
For infants 1000 to 1499 g at birth, the need for mechanical ventilation was associated with an increased risk of handicap (CP alone or combined with Bayley score <70), although nonventilated infants generally had a very good outcome. The association of Bayley scores and the total time in oxygen for the whole cohort is further evidence that later outcome is related to severity of illness in the early neonatal period, especially when assisted ventilation is required. Only 1 of 20 infants <1000 g at birth followed to 1 year had CP although an additional infant had severe retinopathy of prematurity and had a Bayley score <70. The high mortality of infants in this weight category at our institution appears to be associated with handicap rates that are relatively low.
It should be emphasized that infants diagnosed as having CP at 1 year may “outgrow” this condition although other milder cases of CP may be missed at 1 year.34 Nevertheless, developmental assessment at 1 year has been shown to correlate well with later outcome.35 Follow-up of this cohort continues and longer-term outcome will be reported. Despite many differences in population characteristics, tertiary care facilities, and survival figures when compared with VLBW infants in developed countries, the rates of handicap and general outcome measures in this cohort of VLBW infants seem broadly comparable with recent figures from developed countries. However, one-third of infants 1000 to 1500 g who were ventilated and successfully followed up, were handicapped. Ventilating smaller infants <1000 g would be expected to result in even higher handicap rates and could not be recommended in our situation on the basis of this study.
The good follow-up obtained in this study was largely due to the efforts of Z. M. Sefuba, RN, who spent many days tracing some of the families. This study was supported in part by grants from the South African Medical Research Council and the Nestlé Nutrition Research Grant Program, and also formed part of the Birth to Ten Study. Dr. Jackie Galpin of the Department of Statistics and Actuarial Science of the University of the Witwatersrand provided advice on the statistical analysis.
- Received December 6, 1995.
- Accepted April 22, 1996.
Reprint requests to (P.A.C.) Department of Paediatrics, Johannesburg Hospital, Private Bag X39, Johannesburg 2000, South Africa.
- VLBW =
- very low birth weight •
- BTT =
- Birth to Ten (study) •
- IVH =
- intraventricular hemorrhage •
- PVL =
- periventricular leukomalacia •
- MDI =
- mental developmental index •
- SGA =
- small for gestational age •
- CP =
- cerebral palsy
- Escobar GJ,
- Littenberg B,
- Pettiti DB
- Thompson CM,
- Buccimazza SS,
- Webster J,
- Malan AF,
- Molteno CD
- ↵Schlemmer L, Stopforth P. A Guide to the Coding of Occupations in South Africa. Durban, South Africa: Centre of Applied Social Sciences, University of Natal; 1979
- Lubchenco LO,
- Hansman C,
- Dressler M,
- Boyd E
- Morley R,
- Brooke OG,
- Cole TJ,
- Powell R,
- Lucas A
- ↵Sullivan K, Goostein J. Programs for nutritional anthropometry. In: Dean AG, Dean JA, Burton AH, Dicker RC, eds. Epi Info, Version 5: A Word Processing, Database, and Statistics Program for Epidemiology on Microcomputers. Stone Mountain, GA: UDS, Inc; 1990
- ↵Bayley N. Bayley Scales of Infant Development. New York, NY: Psychological Corp; 1979
- Kleinman JC,
- Fowler MG,
- Kessel SS
- ↵Chimere-Dan O. Demographic patterns. In: South African Health Review 1995. Durban, South Africa: Health Systems Trust and the Henry J. Kaiser Family Foundation; 1995:3–15
- ↵Bellamy C. The State of the World's Children 1996. New York, NY: Oxford University Press; 1996
- Stein H,
- Ellis U
- Brownlee KG,
- Ng PC,
- Roussounis SH,
- Dear PRF
- Cooke RWI
- Robertson C,
- Sauve RS,
- Christianson HE
- Pinto-Martin JA,
- Riolo S,
- Cnaan A,
- Holzman C,
- Susser MW,
- Paneth N
- Nelson KB,
- Ellenberg JH
- Ross G,
- Lipper E,
- Auld PAM
- Copyright © 1997 American Academy of Pediatrics