Indian Pediatrics 1999; 36:669-676 

From the Department of Pediatrics, K.E.M. Hospital, Pune 411011, India.

Reprint requests: Dr. Sudha Chaudhari, Consultant, DivisIon of Neonatology, Department of Pediatrics, K.E.M. Hospital, Pune 411 all, India.

Manuscript received: November 12, 1998; Initial review completed: January 5, 1999; Revision accepted: March 8, 1999.

Abstract:

Objective: To evaluate the intelligence, visuo-motor perception, emotional problems and preschool skills in low birth weight (LBW) infants and the impact of social and environmental factors on their development. Design: A prospective cohort study. Setting: Infants discharged from a Neonatal Special Care Unit of a referral hospital with birth weight less than 2000 g followed up in the High Risk Clinic. Methods: Low birth weight infants were assessed by Stanford Binet Scales of intelligence, Bender Gestalt Testforvisuo-motor perception, Human Figure Drawingfor emotional indicators and occupational therapy assessment. A detailed evaluation of their environment and socio-economic status was done. Hearing and ophthalmic assessment was also done and the school progress report was scrutinized. Results: Two hundred and one LBW and seventy one control children were assessed. The mean IQ of LBW children was within normal limits (94.3), though . significantly lower than controls (101.3). Preterm SGA children had the lowest mean 1Q. Visuo- motor perception and preschool skills and language development was poorer in LBW children. There was no difference in the emotional indicators. Thirteen per cent of LBW children had borderline IQ, as compared to 5.6% in controls (p < 0.05). Mother's education and spaciousness of the house had a positive impact and chronic medical problems had a negative impact on the IQ. Conclusions: The mean IQ of LBW children was within normal limits. The incidence of children with borderline intelligence (IQ 70-85) was significantly higher than controls. Mother's education had a positive impact on the intelligence of the children. A longer follow up is necessary to identify "slow learners".

Key words: Intelligence quotients, Low birth weight, Slow learners.

The improved survival of low birth weight infants has not resulted in an increase in the incidence of major disabilities like cerebral palsy, as feared by many neonatologists. In fact, the incidence of cerebral palsy in very low birth weight infants, has remained quite stable (4.5% - 10%) in the Western countries in the last 20 years (l). However, these positive findings are also accompanied by reports of increasingly higher incidence of mild learning difficulties(2). These children with learning difficulties have borderline intelligence and low cognitive abilities, they function below the average level but not in the range of mental retardation. The cognitive difficulties do not become apparent till the child enters school. It is important to identify this group of children (formerly termed as minimal brain dysfunction) as they may be put under undue pressure by parents and teachers to perform better and 'cope' with normal children in their class, which in turn may lead to emotional and behavioral problems.

This long term follow up study was undertaken in. two phases. In the first phase of the study, major handicaps and neurologic sequelae were identified at the end of 3 years(3). We are now reporting the second phase of the study, where the emphasis was on detection of abnormalities of cognition, visuomotor perception, fine motor coordination, skills which are very important for reading and writing. The influence of social and environ- mental factors on the development of a child cannot be overlooked, so this study also aimed at studying in great details the social milieu in which the child was growing up.

Subjects and Methods

The cohort consisted of infants weighing less than 2000g discharged from a Neonatal Special Care Unit between October 1987 to April 1989 and this study was conducted between 1993-95. The low birth weight infants' intrauterine growth status was classified as AGA or SGA using the criteria of Singh et al.(4).The children were followed up in the High Risk Clinic (HRC) which was held at the IDH Rehabilitation Center twice a week. Full term normal neonates with birth weight more than 2500 g with a normal antenatal and perinatal course, matched for socio-economic class with LBW infants, were enrolled as controls at birth.

The cohort at its inception had 249 LBW and 86 control infants. This study reports the assessment at 6 years of 216 LBW and 71 controls who were available for follow up. The reasons for loss to follow up have been re- ported earlier(3). The children were reminded of the follow up visit, just prior to their sixth birthday, by sending a letter. A home visit was made by the Social Worker in case of defaulters.

Methods of Assessment

The development of the child was assessed using three different methods. The children with cerebral palsy and/or mental retardation could not perform these complicated tests of intelligence due to their motor problems or mental retardation and were hence not included in the final analysis. The normal birth weight control children went through the same protocol of assessment as the LBW group. The following tests were administered to the cohort children by a trained psychologist.

Stanford Binet Intelligence Scales

The intelligence of the child was deter- mined by using the Indian adaptation by Kulshreshtha(5). An IQ of 85 or above was considered as normal, while as an IQ between 70-84 was considered as borderline intelligence.

Bender Gestalt Test

Visuo-motor perception of the child was assessed by using the Bender Gestalt test(6). A set of nine figures was presented to the child, one at a time, and the child was asked to copy them on a piece of paper. A score of 8.4 :t 4.12 was considered normal at 6 years.

Human Figure Drawing

Emotional state of the child was assessed by asking the child to draw a human figure on a prescribed piece of paper. Koppitz(7) has described 30 emotional indicators, which can be interpreted from this drawing. The presence of three or more indicators was considered as abnormal.

Occupational Therapist's Assessment (Ayres and Bobath)(8)

This detailed assessment was done under 7 headings - (i) gross motor (ii) fine motor (iii) perception-body image and visuo-motor (iv) intersensory integration (v) language development (vi) preschool skills and (vii) activities of daily living. It was graded into three categories - (i) normal, (ii) child needs some assistance from the therapist and (iii) abnormal (when the child could not perform the activity even with assistance).

Hearing and Ophthalmic Assessment

Hearing assessment was done using a pure tone auditometer by the Audiologist to detect any hearing loss. Ophthalmic examination was done to detect refractive errors and squints.

Social and Environmental Assessment

The social worker made a home visit and made detailed observation on the living conditions - locality, spaciousness of the house, etc. The socio-economic status (SES) was deter- mined by using the revised Kuppuswamy scale(9). Parental education was also classified according to Kuppuswamy Scale(9).

School Progress Report

The annual school progress report card was read and categorized as good (>60%), fair (35- 60%) and poor (<35%). Descriptive progress reports were categorized by the teachers them- selves in the 'Remarks' section.

Statistical Analysis

The data was translated into SPSS PC+. Student '1' test was used for comparing mean IQs and Bender-Gestalt scores between the study and control groups. ANOV A was used for comparing school performance with mean IQ. Chi-square was used for comparing categorical data such as occupational therapy assessment and IQ distribution. Multiple linear regression was done to study the impact of various socio-environmental factors on IQ.

Results

Baseline Data

Two hundred and sixteen low birth weight infants (birth weight less than 2000 g) were available for the six year assessment. Out of these, 10 children had already been diagnosed as cerebral palsy and/or mental retardation in the previous study(3). One child died at home due to fever and convulsions before all the assessments could be completed. Four children with hearing impairment could not be tested satisfactorily and their assessment was considered as invalid. This analysis is restricted to 201 children, 11] girls and 90 boys. One hundred and nineteen infants were outborn and 82 were inborn. The frequency distribution of their birth weight and gestation is shown in Table I. Additional risk factors at birth were septicemia/meningitis in 35 (17.4%), hyperbilirubinemia in 25 (12.4%), intraventricular hemorrhage in 14 (6.97%), seizures in ]6 (7.9%), apnea in 18 (8.9%), birth asphyxia in 7 and meconium aspiration syndrome in 2. Many of the infants had more than two of these risk factors. Seventy one normal infants with birth weight more than 2500 g formed the control group.

TABLE I

 Frequency distribution of Birth Weight and Gestation in Study Population (n=201)

Birth weight (g)	n(%)	Gestation week	n (%)
<1000	2(1)	≤28	1 (0.5)
1000-1249	26 (12.9)	29-30	4 (2.0)
1250-1499	32 (15.9)	31-32	24 (11.9)
1500-1749	99 (49.3)	33-34	70 (34.8)
1750-1999	42 (20.9)	35-36	67 (33.3)
		≥ 37	35 (17.4)

Full term SGA = 35(17.4%), Preterm AGA=79(39.3%), Preterm SGA - 87 (43.3%)

Cognitive Assessment and Preschool Skills
Stanford Binet intelligence Scales

The distribution of intelligence quotients of the study and the control group is shown in Table II. Seventeen per cent of the LBW children had below normal IQ, as compared to 5.6% in the controls (p <.05). Out of the 7 children with IQs in the retardation category (50-69), four had seizure disorder and were on anticonvulsant therapy.

TABLE II

 Distribution of Intelligence Quotients at 6 years

IQ	Control (n=71)	LBW (n = 201)
	n(%)	n(%)
1. 50-69    Retarded	0(0)	7(3.5)*
2. 70-84    Borderline	4(5.6)	27(13.4)*
3. 85-110    Normal	55(77.5)	141(70.1)
4. > 110    Bright	12(16.9)	26(10)

* p < 0.05

The mean IQ of LBW children was within normal limits, though significantly lower than that of controls (Table Ill). The mean IQs of LBW children categorized by their intrauterine growth status showed that the lowest mean IQs were observed in the preterm SGA group (Table Ill).

TABLE III

 Comparison of Mean Intelligence Quotients of Controls with Study Groups

Group	n	mean IQ SD	p value
1. Control	71	101.38 (10.2)	-
2. All LBW	201	94.30 (13.6)	0.000
3. Full term SGA	35	96.02 (15.4)	0.03
4. Preterm AGA	79	95.85 (14.6)	0.009
5. Preterm SGA	87	92.22 (11.6)	0.000

AGA = Appropriate for gestational age.
SGA = Small for gestational age.

Bender Gestalt Test

The number of children who completed this test was much smaller, because many of the children got bored or lost interest after a while and did not finish the test. One hundred and sixty eight LBW children completed the test with mean score of 10.57 :t 3.5. Sixty one controls completed the test and had a mean score of 8.19 :t 3.1 and this difference was significant (p<0.05). Lower scores indicate better visuo-motor perception.
 
Human Figure Drawing (Koppitz)

Human figure drawing on a specified piece of paper was scored for emotional indicators. One hundred and seventy seven LBW children and 61 controls completed the drawing. There were 31 (17.5%) LBW children and 8 (13.1 %) controls who showed abnormal findings, but the difference was not statistically significant.

Occupational Therapy Assessment (Ayres and Bobath)

The number of LBW children who required assistance for language assessment were 34 (16.9%) as against 4 (5.6%) in the control children (p <0.05). Sixteen LBW children had poor preschool skills as compared to none in the control group (p <0.05).

School Performance

The children were studying either in pre- primary or first grade. The report cards of 117 LBW children and 55 control children were evaluated. Table IV provides a comparison of school performance and mean IQs in different groups. This shows that in controls and LBW children school performance conformed well with the IQs.

TABLE IV

Comparison of School performance with Mean Intelligence Quotients

School Performance	LBW (n=170)	Mean IQ	Controls (n=5)	Mean IQ
Good	101	101.45 (6.7)	37	106.07(6.6)
Fair	45	87.95 (8.8)	17	96.67 (7.8)
Poor	21	80.05 (4.3)	1	53.5(1.8)

p <0.05 (ANOVA for mean IQ).

Social and Environmental Assessment

Thirty per cent of the mothers had some degree of college education, whereas one per cent of the mothers were illiterate. Five per cent of the children lived in independent bungalows and the rest in lower and middle class localities. Most of the children (91 %) came from middle or lower middle class families (SES II or III). Fifty six per cent of the children lived in overcrowded houses.

A multiple linear regression analysis was done using IQ as the outcome variable and infant and socio-environmental information as predictor variable. The predictor. variables included socio-economic status, locality, spaciousness of the house, mother's education, etc., mental and motor age on}3ayley Scales at one year (available from the previous phase of the study), and chronic medical problems like seizures and asthma.

The regression analysis showed that mother's education (β coefficient 0.33373, p <0.001), spaciousness of the house (β coefficient 0.19424, p <0.020), and a mental age at one year (β coefficient 0.22959, p <0.009) had a significant impact on the 6 year IQ. Chronic medical problems had a negative impact on the 6 year IQ (β coefficient - 0.3573, P <0:020). Multiple R of the final module was 0.4354.

Hearing and Ophthalmic Assessment

Four children had sensorineural hearing loss, and one girl had very mild sensorineural hearing loss. Three children had conductive hearing loss. Six children had convergent squint, 3 of whom had associated hyper- metropia. Two children had myopia, 1 had cataract (which was not present in Infancy), and one child was diagnosed to have Coat's disease.

Interventions

Twenty seven children had borderline intelligence and were lagging behind in school. Special methods of teaching and stimulation were explained to the parents so that they could attempt to improve their academic performance. Advice regarding change of medium of instruction from English to Marathi for children with borderline intelligence was given. For children with emotional and behavioral problems, special play therapy sessions were conducted and behavioral modification therapy was explained to parents. Children with problems of fine motor activities, intersensory integration, language development and preschool skills were given intervention, through a home based program. The children whose mothers did not do the exercises regularly at home or those with severe problems were provided intervention at the TDH Rehabilition Center using Montessori equipment.

Discussion

Long term follow ups are difficult and expensive. It is however important to know the later outcome in LBW infants as subtle intellectual impairments and learning problems do not become apparent till much later in child- hood. An area of intense interest for many years has been whether pregnancy and birth complications lead to learning difficulties in later childhood. Sameroff(10) concluded that reproductive complications, in fact, contribute little to these problems by the time children reach school age. The present low birth weight study was undertaken to explore this complex issue. The low birth weight children were assessed at the age of 6 years with a battery of tests designed to give a global picture of their intellectual, visuo-motor, language and cognitive status. Appropriate interventions were given by the psychologist, occupational therapist and audiologist whenever a problem was detected. The mean IQ of the LBW children was within normal limits, though significantly lower than that of controls. Klein(11) reported similar findings in their VLBW study. Amongst the low birth weight children, it was the children with a double risk preterm and small for gestational age, who showed the lowest mean IQ. The LBW children also showed poor visuo-motor perception compared to controls. Poor visuo-motor perception in infants weighing less than 1500 g has been reported by many authors(12,13). Visuo-motor perception is extremely important for learning to write. In children with poor visuo-motor perception, the letters tend to collide, they are formed in strange ways and these children cannot write on a line. They cannot color within a given figure or cut and paste and this causes problems in school.

These findings were confirmed by an independent assessment by the occupational therapist. The low birth weight children had significant problems with preschool skills, compared to controls. They also needed more assistance when language development was assessed. Hunt(l4) also described deficiency in language abilities in VLBW children. As far as emotional indicators are concerned, there was no statistically significant difference between control and low birth weight children.

Thirteen per cent of the low birth weight children had "borderline intelligence", which was significantly more than the incidence in the control group. These children may turn out to be "slow learners" and may need shifting to a special school when the school curriculum becomes more demanding. In Maharashtra, it is a general policy to promote all children in primary schools, unless they are very poor in their studies. So it is only when these children reach secondary school that their inability to cope may become obvious. In the Scottish low birth weight study( 15) of infants weighing less than 1750 grams, the incidence of borderline intelligence was 28.5%.

Unlike the Western studies, none of the low birth weight infants in our study received assisted ventilation, as we did not have those facilities in our Unit. Ross(16) reported an incidence of 22% in VLBW children at 4 years. There are no Indian figures for comparison.

Although low birth weight per se has been recognized as .a significant risk factor, the interaction between low birth weight and adverse environment has not been well established. Hunt( 17) has reported that parental education had a very strong impact on VLBW children when they were followed up at 8 years. We also found that mother's education had a strong impact on the 6 year IQ. Mutch(18) reported a strong impact of socio- economic status and mother's education. Escalona(19) reports lower IQs in socio- economically deprived LBW children, who according to her, are subject to double hazards -biologic and social. We found that spaciousness of the house seemed to influence IQ. The mental age at one year determined on the Bayley Scales strongly influenced the IQ at 6 years. Chronic medical problems had a negative impact on the IQ and the largest group here was seizure disorders.

Although we have been able to gain some insight into the cognitive status of the LBW children at school entry, further follow up is necessary to determine how these children fare with increasingly complex tasks at school. Lastly, in the words of Aylward(20) "environ- mental effects can exacerbate or ameliorate the influence of non-optimal prenantal or perinatal events. Environmental influences have a tendency to obscure biomedical and biological effects with longer follow up". Hence a longer follow up is absolutely essential.

Acknowledgement

The authors gratefully acknowledge the financial support provided by Indian Council of Medical Research, New Delhi, and the guidance for statistical analysis given by Dr Ashish Eavdekar.