This cross-sectional study was conducted in neonatal follow-up clinic from April 2017 to February 2018. All VLBW preterm inborn children were consecutively enrolled at the age between 6 months to 3 years of corrected age after written informed consent. The Institute research ethics committee approved the study. Children diagnosed with hearing loss either by clinical assessment or Brainstem evoked response auditory (BERA) were excluded. Baseline demographic details including socioeconomic status and neonatal morbidities were recorded.

Denver Development Screening Tool II and Amiel-Tison scale were used for developmental screening and neurological examination, respectively. Language was assessed by LEST 0-3 scale [4], and hearing assessment was done by clinical examination and BERA, if feasible. LEST administration and interpretation was done as per standard protocol and considered delayed/abnormal if either suspect/questionable or delay was found on LEST 0-3 scale.

Statistical analysis: Assuming speech delay prevalence as 13% [6] with 5% variability, the calculated sample size was 181. Allowing a 5% margin for child’s non-cooperation, a final sample size of 200 was planned. Chi-square test or Fisher’s exact test and students t-test or Mann-Whitney U test were used as applicable. Multivariate logistic regression was used for the analysis of predictors. SPSS version 20 software was used.

There were 1440 VLBW infants discharged between July 2014 till June 2017, of which 490 attended follow-up clinic; we enrolled consecutive 200 children. One child had abnormal BERA and was excluded. Baseline characteristics are described in Table I. Sixty-four (32%) children had language delay, of which 37 (18.5%) were questionable, 18 (9%) were suspect and 9 (4.5%) were delayed. Forty (62%) children had isolated language delay, and rest had it as a part of global development delay. Thirty-three (16.5%) children had suspect development by Denver II and 18 (9%) had either suspect or abnormal neurological examination. All had clinically normal hearing; BERA could be done in 45 (22.5%) children, and was normal.

TABLE I	Baseline Characteristics of the Study Participants (n=200)

The multivariate analysis of predictors for language development showed that late onset neonatal sepsis, patent ductus arteriosus (PDA) and poor socioeconomic status were independent predictors for language delay (Table II). Abnormal neurological examination, and suspect development on the Denver II scale were also significantly associated with language delay.

TABLE II	Univariate and Multivariate Analysis of the Predictors of Language Delay in VLBW Children

In this study, one-third VLBW children with clinically normal hearing had language delay when assessed between 6-36 months corrected age. Late onset neonatal sepsis, PDA, and low socioeconomic status were significant predictors of language delay. We corrected age upto 3 years as most of the children were extreme/early preterm and cannot be directly equated to term children for first 2-3 years [7].

Limitations of the study were inability to do BERA in all cases, and low (34%) follow-up rate in the clinic. Also, we did not screen these children for autism.

Previous studies in apparently healthy Indian children reported prevalence of language delay ranging from 4%-25% [1,5]. VLBW children have higher prevalence ranging from 16-27% [1,8,9]. Mondal, et al. [10] reported the similar prevalence of language delay in term infants discharged from intensive care unit, using LEST 0-3 scale. They found negative home environment and family history of language disorders as risk factors for language delay. They did not find any association between language delay and socioeconomic status or sepsis. However, these results cannot be compared to ours as most of those children were term, normal birth weight with very few neonatal morbidities. Socioeconomic status is an important predictor of language development and previous studies have shown the adverse effects of poverty and low socioeconomic status on language and executive function areas of brain, leading to language delay [10,11], similar to our study. There is a strong adverse association between neonatal sepsis and white matter injury, brain structure metrics, and diffusion in preterms [12]. Other studies also reported neonatal sepsis as an independent predictor of language delay [12,13]. A significant independent association between PDA and language delay noted in the present study was also observed by Singer, et al. [14], who reported that presence of PDA accounted for 13 point decrement in the language score. Unlike previous studies [15], no significant relationship between gestational age and language development, was noted in the present study.

The prevalence of language delay in VLBW preterm infants is high, and therefore, infants with significant morbidities should be routinely screened for early identification and intervention. Structured language assessment and speech stimulation should be routinely performed in high risk follow-up clinics.

Acknowledgement: Mrs. Rita Puri, technician, pediatric neurology lab for conducting BERA.

Contributors: PD: conceived the study, enrolled patients, collected the data; JK: enrolled patients, collected and analyzed the data and drafted the manuscript; KM: conceptualized and designed the study, supervised data collection and analysis, and critically revised the manuscript. All the authors approved the final version of the manuscript and are accountable for all aspects of the work.

Funding: None; Competing interest: None stated.