Ear& Human Development, 24 (1990) 65-17 Elsevier Scientific Publishers Ireland Ltd.
65
EHD 01094
Language development of very low birth weight infants and fullterm controls at 12 months of age Oscar G. Casiroa, Diane M. Moddemanna, Richard S. Stanwickayb Vinnie K. Panikkar-Thiessen”, Heather Cowanc and Mary S. Cheangb Lkpartments of “Pediatrics and ‘Community Health Sciences,University of Manitoba and ‘Childrens Hospital of Winnipeg, Manitoba (Canada) (Received 12 February 1990; revision received 4 June 1990; accepted 16 July 1990)
Summary Twenty-eight very low birth weight (VLBW) and 32 full term infants were prospectively assessed at one year of age for hearing, language development and neurological status. The prevalence of conductive hearing deficits was the same in both groups. Language scores in VLBW infants were significantly lower than in fullterm controls and 39% had significant language delays. VLBW infants exhibited a shorter attention span and were less likely to understand simple questions, to recognize objects or body parts when named, to initiate speechgesture games, to follow simple commands and to imitate or use words consistently. Language quotients were directly associated with gestational age and five minute Apgar scores and inversely associated with severity of intraventricular hemorrhage, bronchopulmonary dysplasia and length of hospital stay. VLBW small for gestational age infants exhibited more advanced language skills than VLBW appropriate for gestational age infants. Language delays were more prevalent among, but not limited to, infants with mild to moderate neurological abnormalities. The influence of prematurity and VLBW on language development is complex and multifactorial and research is continuing to determine the predictive validity and long term significance of the early language delays described in this study. Correspondence to: Dr. Oscar Casiro, Child Development brook Street Winnipeg, Manitoba R3A 1Sl Canada.
Clinic, Children’s Hospital,
0378-3782/90/$03.50 0 1990 Elsevier Scientific Publishers Ireland Ltd. Published and Printed in Ireland
840 Sher-
66 prekture
infant; month follow up.
language
development;
neurological
outcome;
hearing;
12
Introduction Advances in neonatal and obstetrical care over the last two decades have dramatically decreased the mortality rates for very low birth weight infants [29]. These children, however, are at risk of sustaining physical and developmental deficits of varying severity [28]. The most profound handicaps, such as cerebral palsy, mental retardation, hearing loss and blindness, are usually detected during the first year of life. Very low birth weight infants also have been identified later in childhood to be at risk for language, behavioral and learning problems [9,10,11,14]. The delays in language development of premature infants that have been previously reported [1,4-6,8,15,19,23,33] involved relatively larger, more mature infants, with the assessments performed late in the preschool years. The data available on the early language abilities of VLBW infants is not only scarce but also contradictory. Largo et al. [20] followed a group of premature infants that weighed more than 1,000 g at birth (range 1050-3460 g) until age five. They found that low birth weight and early gestational age were negatively correlated with language development. Also, Kenworthy et al. [16] studied a population of infants weighing less than 1200 g and found that one third exhibited a hearing impairment or speech and language problems between four and eleven years of age. More recently Vohr et al. [32] reported that 28% of a group of VLBW infants exhibited language delays at two years of age. However, Greenberg et al. [12] found no language delays at two years of age in a group of infants weighing less than 1800 g at birth. The purpose of this project was to determine if language delays in infants weighing less than 1500 g at birth could be detected by the end of the first year of life. If communication deficits can be detected earlier than the previously reported two years of age [32], this could facilitate early intervention, and ameliorate or avert the accumulative effects of language deficiencies. This paper describes the language performance of a group of very low birth weight infants at 12 months corrected age (CA) compared to a control group of 12-month-old normal full term infants. Subjects The experimental group was comprised of all newborn babies with a birth weight less than or equal to 1500 g and a gestational age of less than 37 weeks who were consecutively discharged from the NICU of the Winnipeg Children’s Hospital, Winnipeg, Manitoba, Canada between June 1984 and June 1985. Both inborn and outborn infants were included.
67
The control group consisted of healthy full term infants born at the Women’s Centre, Winnipeg, Canada between June 1984 and June 1985. They were selected from the hospital’s ‘computerized registry of all births and transfers using a table of random numbers. Inclusion criteria required an uncomplicated pregnancy and delivery, a gestational age > 37 weeks, a birth weight > 2500 g and an absence of any illness affecting the central nervous system in the first year of life. Socioeconomic status was assessed using the Hollingshead Scale, a six point rating of parental occupation [13]. Scores for paternal and maternal education and occupation were computed for each family. Methods
Gestational age was assessed by early obstetrical ultrasound examination, or calculated from the date of the last menstrual period. Infants whose birthweight was below the third percentile for gestation using the Montreal Intrauterine Growth Charts [30] were classified as small for gestational age @GA). Corrected age (CA) was calculated from the expected date of delivery. VLBW infants were evaluated at 12 months + 2 weeks CA. The control population was evaluated at 12 months of age + 2 weeks. The corrected age was used to calculate all scores in the group of VLBW infants. Hearing Hearing was assessed by an audiologist using soundfield, behavioral, observational and visual reinforcement audiometry as well as acoustic impedance measures. The evaluation techniques and scoring methods used were in accordance with the recommendations of the American Speech and Hearing Association [2] and those published by Northern and Downs [22]. Communication skills Speech and language performance was tested by a speech and language pathologist who had no previous knowledge of the child’s history or group assignment . The assessments were videotaped and scored when replayed. Communication was evaluated using the Reynell Developmental Language Scales [25] and the Receptive-Expressive Emergent Language Scales (REEL) [3]. Raw scores for verbal comprehension and verbal expressive skills were obtained after administration of the Reynell Scales. In the original validation of these scales, only 17 children at one year of age were included. The mean verbal comprehension score was reported to be 6.7 & 4.5 and the mean verbal expressive score 8.5 + 4.6. In our control population of 40 normal fullterm infants, the mean verbal comprehension score was 4.3 f 1.8 and the mean verbal expressive score was 6.5 f 1.8. A receptive or expressive score of more than one standard deviation below that of our control group mean was considered abnormal. The REEL Scales were administered by the speech pathologist as a questionnaire to the parents, utilizing a standard pass/fail scoring system. Receptive, expressive
68
and combined language ages obtained after administration of the REEL Scales were divided by the infant’s age in order to determine the corresponding language quotients. A language delay was considered present if an infant had an average REEL language quotient of less than 85 (a delay of more than two months) and a receptive Reynell raw score of less than 2.5 or an expressive Reynell raw score of less than 4.7. Neurodevelopment All subjects were assessed by one of two develop_mental pediatricians using the revised Gesell Developmental Scales [18] and a standard neurological assessment. Adaptive, motor and language developmental quotients (DQ) were obtained. The degree of neurological abnormality (modified from Nelson and Ellenberg [21]), was defined as: (i) mild - a definite abnormality in physical findings with no interference with normal function (i.e., an infant with brisk reflexes and slightly increased tone in the lower limbs but able to sit, crawl, stand and walk with support); (ii) moderate - a definite abnormality in physical findings with modest interference with normal function (i.e., a patient with significant hypertonia and decreased range of motion of the ankle joints, delayed motor development but likely to walk before age two and who may need splints or surgery in the future); (iii) severe - a definite abnormality in physical findings with significant interference with normal function that cannot be improved substantially (i.e., an infant with truncal hypotonia, severe hype/ hypertonicity of the extremities, a significant delay in gross motor development and unlikely to walk). Ethics Parental written informed consent was obtained prior to enrolement into the study. The project was approved by our institution’s Committee on the Use of Human Subjects for Research. Statistical analysis The data obtained was uploaded to an Amdahl mainframe computer and analyzed using the SAS .program. Two-sample t-tests and Chi-square analysis were used to compare the VLBW with the full term infants. The array of variables tested included not only birth weight but also sex, socioeconomic status and birth order, as well as developmental and language scores. Pearson’s correlation coefficients were calculated between the language quotients and neonatal variables to test for associations. Multiple stepwise regression analyses were performed to study the structural combination of variables that could explain the variability of the language scores. Variables explored as explanatory were taken from the univariate analysis as well as from those historically believed to affect language development, such as sex, birth order, socioeconomic status (SES) and neonatal complications. Partial R2 were used to estimate the relative magnitude of the contributions of each explanatory variables, while controlling for variables already in the model.
69
Results Forty VLBW and 40 full term control infants fulfilled the initial inclusion criteria and were assessed as previously described. Two VLBW infants with severe neurological handicaps and developmental quotients < 70 were excluded, as they would be expected to have delayed language development related to factors other than prematurity. Six VLBW infants were North American Indians. Although we identified suitable full term Native controls, we were either unable to recruit them for the study, or when they were enrolled they did not keep their appointments. In addition, since there are no standards for language development in this population, these six VLBW infants had to be excluded. Hearing Of the 32 VLBW infants included in the study, nine (289’0) had bilateral abnormal impedance tympanometry while only one patient (3070) had a unilateral abnormality. In four (12.5%) VLBW infants abnormal impedance tests were associated with abnormal bilateral soundfield audiometry responses (pure tone average threshold between 40 dB and 70 dB), while none had unilateral deficits. Of the 40 full term control infants, nine (23%) had bilateral abnormal impedance tympanometry and three (8070) had a unilateral abnormality. In six (15%) control infants, abnormal impedance tests were associated with abnormal
TABLE I Clinical and demographic data on very low birth weight infants (VLBW) and controls.
Birth weight(g) (range) Gestational age’ (range) Sex (M:F ratio) SGA BPD IVH I-II III-IV Birth order I II + SES 1,2,3 4,596 a = mean f S.D. b NA = P> 0.05.
VLBW (n = 28)
Controls (n = 33)
1174 f 264 (585-1500) 29 f 2 (25-34) 1.5:l 32% 29%
3320 f 340 (2640-4000) 39 f 1 (38-42) 0.9: 1 -
36% 21%
-
67% 33%
43% 57%
NS NS
46% 54%
60% 40%
NS NS
NY
70
bilateral soundfield audiometry responses (pure tone average threshold between 40 dB and 70 dB), while one (3t70) had a unilateral deficit. Controls with an abnormal audiometry exhibited significantly lower language scores than those with normal audiometry (Expressive REEL 100 & 20 vs. 118 f 27, P < 0.03; Expressive Reynell 4.2 f 1.8 vs. 6.5 + 1.8, P < O.OOl), although this was not the case with VLBW infants (Expressive REEL 108 & 8 vs. 95 f 26, P < 0.9; Expressive Reynell 6.0 f 2.6 vs. 5.1 + 2.5, P < 0.6). To remove the potential negative effects of even mild conductive hearing losses on language development, all patients who exhibited abnormal soundfield audiometry responses were excluded from further analyses. The rest of this paper reports on the remaining 28 VLBW infants and 33 controls and Table I shows the clinical and demographic characteristics of these infants. There were no statistically significant differences in sex, birth order, SES or language spoken at home between the two groups. Language In the VLBW infants language quotients were directly associated with gestational age (r = 0.37; P < 0.06) and 5 min Apgar scores (r = 0.35; P < 0.07) and inversely associated with severity of intraventricular hemorrhage (IVH) (r = 0.39; P < 0.05) and length of hospital stay (r = 0.38; P < 0.05). The nine small for gestational age (SGA) infants had significantly higher quotients than the 19 appropriate for gestational age (AGA) patients (111 f 17 vs. 89 f 24, P < 0.05). The eight patients with bronchopulmonary dysplasia (BPD) had sig-
125
25 --
Rid
RacqAii
RaJ EXprC#aiW
Real Avemga
RayIM4ll Raeeptive
Rayrlall
EXpWJdVO
Fig. 1. Language test scores in VLBW infants with suboptimal (DQ Q 85) and normal (DQ > 85) Gessell adaptive developmental quotients, compared to controls.
71
nificantly lower language quotients compared to the 20 without this condition (79 f 27 vs. 102 f 23, P < 0.05). Within group birth weight and gender differences did not influence language performance. The prevalence of suboptimal development (adjusted adaptive DQ > 70 and < 85) was 18% (n = 5) in the VLBW group, while all the controls had DQs > 85. The 23 VLBW infants with normal development had a mean DQ of 102 it 8, compared to 105 f 8 (NS) for the controls. As shown in Fig. 1, VLBW infants with normal DQs scored better than those with suboptimal DQs, but significantly lower than controls, in all language tests. When language and development were treated as continuous variables, VLBW infants had mean Gesell developmental and language quotients within the normal range (Adaptive DQ 97 f 12, Expressive Language DQ 91 + 18, Receptive Language DQ 92 f 18, Average REEL 95 + 26). Their scores, however, were significantly lower than those of full term controls in all measures. The means and standard deviations of developmental and REEL quotients and of Reynell raw scores for cases and controls are presented in Table II. The close correlation of test scores for the three measures (REEL, Reynell and Gessell language DQ) are shown in Table III. Stepwise multiple regression analysis showed that language differences persisted even after adjusting for the developmental quotient (DQ) and sex. Moreover, the proportion of variability of language scores accounted for by VLBW was higher than the variability
TABLE II Developmental and language scores in VLBW infants and controls at one year of age.
Adaptive DQ Expressive language DQ Receptive language DQ Receptive REEL Expressive REEL Average REEL Receptive Reynell Expressive Reynell
VLBW mean f S.D. (range)
Controls mean + S.D. (range)
P
97 f 12 (74-125) 91 It 18 (53-128) 92k 18 (53-128) 95 + 29 (58-133) 95 + 26 (50-133) 95 +: 26 (29-l 33) 3.2 + 1.7 (O-7) 5.1 f 2.5 (O-10)
1052 8 (85-125) 103 f 16 (76-161) 105 f 11 (85-140) 123 f 18 (75-185) 118 k 27 (67-23 1) 122 + 21 (75-208) 4.3 2 1.8 (O-10) 6.5 + 1.8 (3-10)
**
‘DQ = Gesell Developmental Quotient. 0.05; **p-c 0.01; ***p< 0.001.
‘P
65
EHD 01094
Language development of very low birth weight infants and fullterm controls at 12 months of age Oscar G. Casiroa, Diane M. Moddemanna, Richard S. Stanwickayb Vinnie K. Panikkar-Thiessen”, Heather Cowanc and Mary S. Cheangb Lkpartments of “Pediatrics and ‘Community Health Sciences,University of Manitoba and ‘Childrens Hospital of Winnipeg, Manitoba (Canada) (Received 12 February 1990; revision received 4 June 1990; accepted 16 July 1990)
Summary Twenty-eight very low birth weight (VLBW) and 32 full term infants were prospectively assessed at one year of age for hearing, language development and neurological status. The prevalence of conductive hearing deficits was the same in both groups. Language scores in VLBW infants were significantly lower than in fullterm controls and 39% had significant language delays. VLBW infants exhibited a shorter attention span and were less likely to understand simple questions, to recognize objects or body parts when named, to initiate speechgesture games, to follow simple commands and to imitate or use words consistently. Language quotients were directly associated with gestational age and five minute Apgar scores and inversely associated with severity of intraventricular hemorrhage, bronchopulmonary dysplasia and length of hospital stay. VLBW small for gestational age infants exhibited more advanced language skills than VLBW appropriate for gestational age infants. Language delays were more prevalent among, but not limited to, infants with mild to moderate neurological abnormalities. The influence of prematurity and VLBW on language development is complex and multifactorial and research is continuing to determine the predictive validity and long term significance of the early language delays described in this study. Correspondence to: Dr. Oscar Casiro, Child Development brook Street Winnipeg, Manitoba R3A 1Sl Canada.
Clinic, Children’s Hospital,
0378-3782/90/$03.50 0 1990 Elsevier Scientific Publishers Ireland Ltd. Published and Printed in Ireland
840 Sher-
66 prekture
infant; month follow up.
language
development;
neurological
outcome;
hearing;
12
Introduction Advances in neonatal and obstetrical care over the last two decades have dramatically decreased the mortality rates for very low birth weight infants [29]. These children, however, are at risk of sustaining physical and developmental deficits of varying severity [28]. The most profound handicaps, such as cerebral palsy, mental retardation, hearing loss and blindness, are usually detected during the first year of life. Very low birth weight infants also have been identified later in childhood to be at risk for language, behavioral and learning problems [9,10,11,14]. The delays in language development of premature infants that have been previously reported [1,4-6,8,15,19,23,33] involved relatively larger, more mature infants, with the assessments performed late in the preschool years. The data available on the early language abilities of VLBW infants is not only scarce but also contradictory. Largo et al. [20] followed a group of premature infants that weighed more than 1,000 g at birth (range 1050-3460 g) until age five. They found that low birth weight and early gestational age were negatively correlated with language development. Also, Kenworthy et al. [16] studied a population of infants weighing less than 1200 g and found that one third exhibited a hearing impairment or speech and language problems between four and eleven years of age. More recently Vohr et al. [32] reported that 28% of a group of VLBW infants exhibited language delays at two years of age. However, Greenberg et al. [12] found no language delays at two years of age in a group of infants weighing less than 1800 g at birth. The purpose of this project was to determine if language delays in infants weighing less than 1500 g at birth could be detected by the end of the first year of life. If communication deficits can be detected earlier than the previously reported two years of age [32], this could facilitate early intervention, and ameliorate or avert the accumulative effects of language deficiencies. This paper describes the language performance of a group of very low birth weight infants at 12 months corrected age (CA) compared to a control group of 12-month-old normal full term infants. Subjects The experimental group was comprised of all newborn babies with a birth weight less than or equal to 1500 g and a gestational age of less than 37 weeks who were consecutively discharged from the NICU of the Winnipeg Children’s Hospital, Winnipeg, Manitoba, Canada between June 1984 and June 1985. Both inborn and outborn infants were included.
67
The control group consisted of healthy full term infants born at the Women’s Centre, Winnipeg, Canada between June 1984 and June 1985. They were selected from the hospital’s ‘computerized registry of all births and transfers using a table of random numbers. Inclusion criteria required an uncomplicated pregnancy and delivery, a gestational age > 37 weeks, a birth weight > 2500 g and an absence of any illness affecting the central nervous system in the first year of life. Socioeconomic status was assessed using the Hollingshead Scale, a six point rating of parental occupation [13]. Scores for paternal and maternal education and occupation were computed for each family. Methods
Gestational age was assessed by early obstetrical ultrasound examination, or calculated from the date of the last menstrual period. Infants whose birthweight was below the third percentile for gestation using the Montreal Intrauterine Growth Charts [30] were classified as small for gestational age @GA). Corrected age (CA) was calculated from the expected date of delivery. VLBW infants were evaluated at 12 months + 2 weeks CA. The control population was evaluated at 12 months of age + 2 weeks. The corrected age was used to calculate all scores in the group of VLBW infants. Hearing Hearing was assessed by an audiologist using soundfield, behavioral, observational and visual reinforcement audiometry as well as acoustic impedance measures. The evaluation techniques and scoring methods used were in accordance with the recommendations of the American Speech and Hearing Association [2] and those published by Northern and Downs [22]. Communication skills Speech and language performance was tested by a speech and language pathologist who had no previous knowledge of the child’s history or group assignment . The assessments were videotaped and scored when replayed. Communication was evaluated using the Reynell Developmental Language Scales [25] and the Receptive-Expressive Emergent Language Scales (REEL) [3]. Raw scores for verbal comprehension and verbal expressive skills were obtained after administration of the Reynell Scales. In the original validation of these scales, only 17 children at one year of age were included. The mean verbal comprehension score was reported to be 6.7 & 4.5 and the mean verbal expressive score 8.5 + 4.6. In our control population of 40 normal fullterm infants, the mean verbal comprehension score was 4.3 f 1.8 and the mean verbal expressive score was 6.5 f 1.8. A receptive or expressive score of more than one standard deviation below that of our control group mean was considered abnormal. The REEL Scales were administered by the speech pathologist as a questionnaire to the parents, utilizing a standard pass/fail scoring system. Receptive, expressive
68
and combined language ages obtained after administration of the REEL Scales were divided by the infant’s age in order to determine the corresponding language quotients. A language delay was considered present if an infant had an average REEL language quotient of less than 85 (a delay of more than two months) and a receptive Reynell raw score of less than 2.5 or an expressive Reynell raw score of less than 4.7. Neurodevelopment All subjects were assessed by one of two develop_mental pediatricians using the revised Gesell Developmental Scales [18] and a standard neurological assessment. Adaptive, motor and language developmental quotients (DQ) were obtained. The degree of neurological abnormality (modified from Nelson and Ellenberg [21]), was defined as: (i) mild - a definite abnormality in physical findings with no interference with normal function (i.e., an infant with brisk reflexes and slightly increased tone in the lower limbs but able to sit, crawl, stand and walk with support); (ii) moderate - a definite abnormality in physical findings with modest interference with normal function (i.e., a patient with significant hypertonia and decreased range of motion of the ankle joints, delayed motor development but likely to walk before age two and who may need splints or surgery in the future); (iii) severe - a definite abnormality in physical findings with significant interference with normal function that cannot be improved substantially (i.e., an infant with truncal hypotonia, severe hype/ hypertonicity of the extremities, a significant delay in gross motor development and unlikely to walk). Ethics Parental written informed consent was obtained prior to enrolement into the study. The project was approved by our institution’s Committee on the Use of Human Subjects for Research. Statistical analysis The data obtained was uploaded to an Amdahl mainframe computer and analyzed using the SAS .program. Two-sample t-tests and Chi-square analysis were used to compare the VLBW with the full term infants. The array of variables tested included not only birth weight but also sex, socioeconomic status and birth order, as well as developmental and language scores. Pearson’s correlation coefficients were calculated between the language quotients and neonatal variables to test for associations. Multiple stepwise regression analyses were performed to study the structural combination of variables that could explain the variability of the language scores. Variables explored as explanatory were taken from the univariate analysis as well as from those historically believed to affect language development, such as sex, birth order, socioeconomic status (SES) and neonatal complications. Partial R2 were used to estimate the relative magnitude of the contributions of each explanatory variables, while controlling for variables already in the model.
69
Results Forty VLBW and 40 full term control infants fulfilled the initial inclusion criteria and were assessed as previously described. Two VLBW infants with severe neurological handicaps and developmental quotients < 70 were excluded, as they would be expected to have delayed language development related to factors other than prematurity. Six VLBW infants were North American Indians. Although we identified suitable full term Native controls, we were either unable to recruit them for the study, or when they were enrolled they did not keep their appointments. In addition, since there are no standards for language development in this population, these six VLBW infants had to be excluded. Hearing Of the 32 VLBW infants included in the study, nine (289’0) had bilateral abnormal impedance tympanometry while only one patient (3070) had a unilateral abnormality. In four (12.5%) VLBW infants abnormal impedance tests were associated with abnormal bilateral soundfield audiometry responses (pure tone average threshold between 40 dB and 70 dB), while none had unilateral deficits. Of the 40 full term control infants, nine (23%) had bilateral abnormal impedance tympanometry and three (8070) had a unilateral abnormality. In six (15%) control infants, abnormal impedance tests were associated with abnormal
TABLE I Clinical and demographic data on very low birth weight infants (VLBW) and controls.
Birth weight(g) (range) Gestational age’ (range) Sex (M:F ratio) SGA BPD IVH I-II III-IV Birth order I II + SES 1,2,3 4,596 a = mean f S.D. b NA = P> 0.05.
VLBW (n = 28)
Controls (n = 33)
1174 f 264 (585-1500) 29 f 2 (25-34) 1.5:l 32% 29%
3320 f 340 (2640-4000) 39 f 1 (38-42) 0.9: 1 -
36% 21%
-
67% 33%
43% 57%
NS NS
46% 54%
60% 40%
NS NS
NY
70
bilateral soundfield audiometry responses (pure tone average threshold between 40 dB and 70 dB), while one (3t70) had a unilateral deficit. Controls with an abnormal audiometry exhibited significantly lower language scores than those with normal audiometry (Expressive REEL 100 & 20 vs. 118 f 27, P < 0.03; Expressive Reynell 4.2 f 1.8 vs. 6.5 + 1.8, P < O.OOl), although this was not the case with VLBW infants (Expressive REEL 108 & 8 vs. 95 f 26, P < 0.9; Expressive Reynell 6.0 f 2.6 vs. 5.1 + 2.5, P < 0.6). To remove the potential negative effects of even mild conductive hearing losses on language development, all patients who exhibited abnormal soundfield audiometry responses were excluded from further analyses. The rest of this paper reports on the remaining 28 VLBW infants and 33 controls and Table I shows the clinical and demographic characteristics of these infants. There were no statistically significant differences in sex, birth order, SES or language spoken at home between the two groups. Language In the VLBW infants language quotients were directly associated with gestational age (r = 0.37; P < 0.06) and 5 min Apgar scores (r = 0.35; P < 0.07) and inversely associated with severity of intraventricular hemorrhage (IVH) (r = 0.39; P < 0.05) and length of hospital stay (r = 0.38; P < 0.05). The nine small for gestational age (SGA) infants had significantly higher quotients than the 19 appropriate for gestational age (AGA) patients (111 f 17 vs. 89 f 24, P < 0.05). The eight patients with bronchopulmonary dysplasia (BPD) had sig-
125
25 --
Rid
RacqAii
RaJ EXprC#aiW
Real Avemga
RayIM4ll Raeeptive
Rayrlall
EXpWJdVO
Fig. 1. Language test scores in VLBW infants with suboptimal (DQ Q 85) and normal (DQ > 85) Gessell adaptive developmental quotients, compared to controls.
71
nificantly lower language quotients compared to the 20 without this condition (79 f 27 vs. 102 f 23, P < 0.05). Within group birth weight and gender differences did not influence language performance. The prevalence of suboptimal development (adjusted adaptive DQ > 70 and < 85) was 18% (n = 5) in the VLBW group, while all the controls had DQs > 85. The 23 VLBW infants with normal development had a mean DQ of 102 it 8, compared to 105 f 8 (NS) for the controls. As shown in Fig. 1, VLBW infants with normal DQs scored better than those with suboptimal DQs, but significantly lower than controls, in all language tests. When language and development were treated as continuous variables, VLBW infants had mean Gesell developmental and language quotients within the normal range (Adaptive DQ 97 f 12, Expressive Language DQ 91 + 18, Receptive Language DQ 92 f 18, Average REEL 95 + 26). Their scores, however, were significantly lower than those of full term controls in all measures. The means and standard deviations of developmental and REEL quotients and of Reynell raw scores for cases and controls are presented in Table II. The close correlation of test scores for the three measures (REEL, Reynell and Gessell language DQ) are shown in Table III. Stepwise multiple regression analysis showed that language differences persisted even after adjusting for the developmental quotient (DQ) and sex. Moreover, the proportion of variability of language scores accounted for by VLBW was higher than the variability
TABLE II Developmental and language scores in VLBW infants and controls at one year of age.
Adaptive DQ Expressive language DQ Receptive language DQ Receptive REEL Expressive REEL Average REEL Receptive Reynell Expressive Reynell
VLBW mean f S.D. (range)
Controls mean + S.D. (range)
P
97 f 12 (74-125) 91 It 18 (53-128) 92k 18 (53-128) 95 + 29 (58-133) 95 + 26 (50-133) 95 +: 26 (29-l 33) 3.2 + 1.7 (O-7) 5.1 f 2.5 (O-10)
1052 8 (85-125) 103 f 16 (76-161) 105 f 11 (85-140) 123 f 18 (75-185) 118 k 27 (67-23 1) 122 + 21 (75-208) 4.3 2 1.8 (O-10) 6.5 + 1.8 (3-10)
**
‘DQ = Gesell Developmental Quotient. 0.05; **p-c 0.01; ***p< 0.001.
‘P
