The improved survival of "high risk" infants has generated intense interest in the quality of life of the survivors. However, the emphasis is on early detection of severe impairment and handicapping conditions. Lesser morbidities like postneonatal illness, rehospitalizations, learning problems in children who appear normal in early child-hood, are rarely reported(1). These problems have a major impact on families as far as stress, finance and educational services are concerned. We have already reported major neurologic morbidity in a cohort of high risk infants in our previous publications(2-4). In this communication, we are reporting the less obvious, but equally important morbidities we encountered during a six year follow up.

There is a tendency amongst neo-natologists to be satisfied if a very high risk or a very low birth weight neonate survives and goes home. However, postneonatal mortality continues to be high(5). Unless close follow up is maintained, the baby may die after going home and all the intensive care given to the baby may be of no consequence. This article also reports the mortality in these infants, during a six year follow up.

Infants discharged from a Neonatal Special Care Unit were enrolled in this study using predetermined high risk criteria: (i) gestation <37 weeks, (ii) birthweight <2000 g, (iii) Apgar score <5 at 5 minutes, (iv) septicemia/meningitis, (v) hyperbilirubinemia, (vi) apnea, (vii) seizures, (viii) intraventricular hemorr-hage and (ix) respiratory distress. Enrollment was done during an 18 month period from October 1987 to April 1989 and followup for a period of six years. They were given a specially designed High Risk Card, with visits to the High Risk Clinic (HRC), scheduled one week after discharge, one month of age and monthly thereafter till the age of one year; and six monthly till the final assessment at two and a half years. They were encouraged to come without a prior appointment if there was an illness. A specially designed form for recordng intercurrent illness was given to the parents. All illnesses and their treatment, including hospitalizations were recorded on this form.

The Social Worker made home visits to assess the socio-economic status using the Kuppuswamy Scale(6). She also made home visits whenever appointments were missed. If a baby made one or two initial visits and then stopped coming inspite of sending letters and making a home visit, he or she was termed as a "defaulter". In case of death of the baby, documents regarding treatment and hospitalization were copied and brought back to the hospital. In case of a home death, parents were questioned regarding the signs and symptoms and a sort of ‘verbal autospy’ was conducted, and this information was recorded in the file.

Developmental assessment was done by using the Indian adaptation of Bayley Scales of Infant Development(7) upto 30 months, and these results have been previously published(3). The children were recalled at 6 years for assessing the Intelligence Quotient by Kulshreshta’s adaptation of the Stanford - Binet Test(8) by a trained psychologist. Children with a neurologic impairment, who could not do this complicated test were excluded. An IQ above 84 was considered as normal, and IQ between 70-84 was termed as ‘borderline intelligence’.

For analysis of the data, the cohort was stratified by birth weight, gestation and intrauterine growth status(9). Normal full terms infants with birth weight more than 2500 g with a normal antenatal, natal and postnatal course, matched for socio-economic class, were enrolled as controls.

The data was translated in SPSS PC+. Chi-square test was used for comparing categorical data such as mortality, rehospitalization and borderline intelligence.

Four hundred and twenty five high risk infants were enrolled in the study. Twenty one infants, whose addresses could not be traced, were lost to follow-up; 5 weighed less than 1500 g, 7 between 1500-1999 g, and 9 weighed ³ 2000 g. This follow up is restricted to four hundred and four high risk infants and 86 controls.

The birth weight, gestational age and intrauterine growth status is shown in Table I. There were 246 males and 158 females. Most babies had more than one risk factor. Birth asphyxia was present in 56 (14%), hyperbilirubinemia in 72 (17.8%), septicemia/meningitis in 81 (20%), seizures in 60 (14%); apneic spells in 48 (11%), intraventricular hemorr hage in 28 (6%) and respiratory distress in 48 (11%). There were 38 low birth weight babies who had no other additional risk factors.

There were 40 deaths during the six year period, out of which 38 deaths occurred during the first year of life, 22 in the hospital and 16 at home (Table II). Sixty per cent of these deaths were in infants less than 3 months of age.

The mortality in the VLBW group was significantly higher (Table III) than the rest of the group (p < 0.05). Out of the 38 infants who died in first year, 20 were small for gestational age and 18 were appropriate for gestational age.

The influence of socio-economic status (SES) on mortality was analyzed. There was a significant trend of higher mortality in the lower socio-economic groups (p <0.001). There were no deaths in SES I, 2.2% in SES II and 15.3% in SES III and IV. The mortality amongst the group who attended the HRC regularly (6.4%) was significantly lower (p <0.001) than that of the defaulters (27.6%).

Intercurrent illnesses, especially the incidence of diarrhea and respiratory illnesses in the first year, were comparable in different birth weight strata (Table III).

A total number of 95 children had 144 admissions to the hospital. Eighty seven (91.5%) of these children were admitted in the first year of life. Thirty five infants were admitted for an infection (gastroenteritis - 24, pneumonitis - 6, septicemia - 5), 20 for seizures, 2 for shunt surgery for hydro-cephalus, 8 for protracted diarrhea and 8 with failure to thrive. Fourteen infants were admitted for a top-up transfusion, 12 (85.7%) of these being pre-terms. A significantly larger proportion (p <0.05) of high risk infants (21.3%) were admitted to the hospital compared to control infants (2.3%).

Three hundred and thirty six children who were available for the final 30 month testing on the Bayley Scales of Infant Development, were recalled at 6 years for an IQ assessment. Only 286 children could be tested at six years as 19 had migrated, 2 were dead and 29 were neurologically impaired.

The mean intelligence quotient of the high risk group was 95.21 ± 6.2 and that of the control group was 101.38 ± 10.2 (p<0.05). A significantly larger proportion (p <0.05) of high risk children (14.7%) had borderline intelligence (IQ 70-84) as compared to the control group (5.6%). When the proportion of children with ‘borderline intelligence’ was analyzed according to gestation, birthweight and intrauterine growth status, it was observed that the trend for borderline intelligence increased significantly (p <0.05) with decreas-ing gestation and was significantly more in infants weighing less than 2000 g (Table IV).

Low birth weight infants are known to suffer from high postneonatal mortality, even in advanced countries(10). Indian figures have varied from 3.8% to 15% depending on the length of the study and strength of follow-up(11-12). The mortality in our closely monitored cohort was 9.4%. Out of the 40 deaths, 38 occurred in the first year of life. The importance of very close monitoring in the first year of life needs emphasis. The difference in the mortality rate in the group of children who attended the HR clinic regularly as compared to those who did not, again emphasizes the issue, especially since sixty three per cent of the deaths were in the VLBW group. The mortality rate was related to birth weight and not to gestational age and also did not depend on intrauterine growth status. Three fourths of the hospital deaths were due to an overwhelming infection. This differs from leading causes of death in the Western countries like SIDS(13) and probably reflects the poor environment in their homes and immune incompetence(14) of our VLBW babies. Most of our mothers belonged to a poor socio-economic class with less than seventh standard education. Sharpiro(15) described a much higher incidence of post-neonatal mortality in mothers with less than high school education. We need to pay special attention to counselling of the mothers and intensify maternal education.

The high risk infant remains vulnerable to a wide spectrum of morbidity in the form of postneonatal illness and rehospitaliza-tions(16). Our study also shows a high incidence of rehospitalization in high risk infants (22.7%) compared to normal control group (2.3%). Most of the hospitalizations (90%) occurred in the first year of life, again emphasizing the importance of monitoring and parental support in the first year of life. Hack(17) reports that after the first year, the rate of rehospitalization diminishes rapidly.

Fourteen per cent of children who appeared to have normal development in their third year, showed ‘borderline intelligence’ at the age of 6 years and might turn out to be ‘slow learners’. It is interesting that the incidence of potential slow learners seems to be related more to gestational age than birth weight. Thus, parents and children may need professional help and support and intervention for the first time in much older ages than we have previously supposed. This highlights the need for monitoring the high risk infant well into childhood.

The after care of this ever increasing group of new citizens - graduates of a NSCU - is assuming as much importance as the acute intensive care received by the newborn. The goal now should be to match the improve-ments in technological care with opportunities for enhancing the quality of their future lives.

Contributors: SC designed the study, conducted the High Risk Follow up Clinic and drafted the paper and will act as guarantor for the paper. SK collected the data. AP supervised the project. SD made home visits and collected mortality data.

Funding: Indian Council of Medical Research, New Delhi.
Competing interests: None stated.