Language Outcomes in Moderate-to-Late Preterm Infants at 2 Years of Corrected Age

Article information

Neonatal Med. 2026;33(1):66-73

Publication date (electronic) : 2026 May 31

doi : https://doi.org/10.5385/nm.25016

Sung Hee Lee, MD, Ju Hyun Jin, MD, Shin Won Yoon, MD

Department of Pediatrics, National Health Insurance Service Ilsan Hospital, Goyang, Korea

Correspondence to: Shin Won Yoon, MD Department of Pediatrics, National Health Insurance Service Ilsan Hospital, 100 Ilsan-ro, Ilsandong-gu, Goyang 10444, Korea Tel: +82-31-900-0259 Fax: +82-31-900-0343 E-mail: swyoon@nhimc.or.kr

Received 2025 December 20; Revised 2026 April 20; Accepted 2026 May 9.

Abstract

INTRODUCTION

The global incidence of preterm births has been increasing, partly because of advances in maternal age and the growing use of assisted reproductive technologies. Preterm infants are at an elevated risk of various developmental disorders, including delays in speech and language [1,2]. Language development is critical for communication and has broad implications for cognitive, academic, and social outcomes [3]. Therefore, early identification and intervention for language delays are essential.

The majority of preterm births (approximately 80%) are classified as moderate-to-late preterm (MLPT), defined as births between 32 and 36 weeks of gestational age (GA) [4]. Although MLPT infants have traditionally been considered to have relatively favorable developmental outcomes, accumulating evidence indicates that they are also at an increased risk of neurodevelopmental delays [4,5]. However, research on the long-term developmental outcomes of MLPT infants in Korea remains limited [6,7].

A previous study by our group has reported that Korean MLPT children at school age had a mean full-scale intelligence quotient (FSIQ) of 92.89±11.90, with 24.3% demonstrating borderline intellectual functioning (FSIQ 70–85) [8]. Another nationwide population-based study conducted by our team has determined that MLPT children were significantly more likely to have language disorders than term children (adjusted odds ratio, 1.48) [6].

Similarly, recent international studies have emphasized that MLPT infants, the largest subgroup of preterm births, should not be underestimated in terms of their risk for adverse neurodevelopmental outcomes [4,9]. A population-based study from Germany has demonstrated a persistent decline in language function among MLPT children from 20 months to 8 years of age compared with their term peers, and recommended language assessments around 2 years of corrected age to enable timely intervention [9].

Despite these findings, the long-term language outcomes of infants with MLPT in Korea remain underexplored. The present study aimed to evaluate language development at approximately 2 years of corrected age with a particular focus on infants with MLPT. A better understanding of these outcomes may inform clinical follow-ups and early intervention strategies for this vulnerable population.

MATERIALS AND METHODS

1. Study population

This retrospective study included 119 preterm infants who visited the Growth and Development Clinic at the National Health Insurance Service Ilsan Hospital between January 2020 and June 2023 for follow-up at a corrected age of 18 to 24 months. The clinic provides regular medical evaluations, developmental assessments, and counseling for preterm infants born at <36 weeks of gestation and high-risk full-term infants.

Infants born at ≥36 weeks of gestation (n=9) and those lost to follow-up (n=28) were excluded. The remaining participants were classified into two groups according to their GA: early preterm (<32 weeks GA) and MLPT (32–35 weeks GA) (Figure 1).

Figure 1.

The flow chart of the study population. Abbreviations: NHIS, National Health Insurance Service; GA, gestational age.

RESULTS

1. Characteristics of infants and mothers

A total of 82 preterm infants were included: 29 early preterm infants (mean GA, 29.8±1.3 weeks; mean birth weight, 1,336.2± 234.5 g) and 53 MLPT infants (mean GA, 33.7±0.9 weeks; mean birth weight, 1,984.9±367.1 g) (Table 1).

Table 1.

The Characteristics of Preterm Infants and Their Mothers

Neonatal morbidities, including low Apgar scores, RDS, use of invasive ventilators, duration of ventilator support, duration of NICU admission, and treatment of PDA were significantly more frequent in early preterm infants than in MLPT infants (P<0.05). No statistically significant differences were observed between the two groups in terms of severe maternal depressive symptoms (19% in the early preterm group vs. 14% in the MLPT group, P=0.721).

2. Language development outcomes

The K-DST was completed by 79.3% of early preterm infants and 66.0% of MLPT infants. Among infants who completed the K-DST, 8.7% of early preterm infants and 14.3% of MLPT infants scored below −2 SD in the language domain. When language delay was defined as a score below −1 SD, the prevalence was 30.4% and 31.4% in early preterm and MLPT infants, respectively.

The BSID-III was performed in 82.8% of early preterm infants and 90.6% of MLPT infants. Among infants who underwent the BSID-III, 25.0% of early preterm infants and 10.4% of MLPT infants scored ≤70 in the language domain, indicating significant language delay. Mild language delay (scores 70–84) was observed in 20.8% of early preterm and 25.0% of MLPT infants. Overall, 35.4% of the MLPT infants had language scores <85, indicating a risk of language delay. The proportion of infants requiring speech therapy was 34.5% among early preterm infants and 26.4% among MLPT infants.

SELSI was administered to 55.2% of early preterm and 47.2% of the MLPT infants at a mean age of 24 to 25 months. Among those tested, 68.8% of early preterm and 56.0% of MLPT infants scored below –2 SD, indicating significant language delay.

No statistically significant differences were observed between early preterm and MLPT infants in the prevalence of severe language delay or the need for speech therapy, suggesting that MLPT infants remain at a comparable risk for language developmental difficulties.

The MLPT group demonstrated better outcomes in the cognitive domain. In the cognitive domain, 20.8% of early preterm infants scored below –2 SD, compared with 2.1% of MLPT infants (P=0.014) (Table 2).

Table 2.

Language Assessment Outcomes in Preterm Infants at 2 Years of Corrected Age

Multivariate logistic regression analysis, including sex, SGA status, and NICU morbidity, showed no significant association with language delay.

DISCUSSION

Investigating whether MLPT birth affects language development revealed that infants with MLPT at a corrected age of approximately 2 years are also at risk of language delays and often require speech and language therapy. Although MLPT infants have higher GAs and fewer neonatal complications than early preterm infants, their risk of language delay remains substantial. These results highlight the importance of continuous developmental surveillance and timely intervention in all preterm infants, including those with the MLPT.

Concerns regarding the long-term developmental outcomes of MLPT infants are growing [15-17]. Because MLPT infants typically experience more favorable short-term morbidities and shorter NICU stays than early preterm infants, long-term follow-up is often insufficient. Consequently, subtle but clinically meaningful developmental difficulties, particularly in language, may go unnoticed in early childhood.

Previous studies have reported similar results. An Australian longitudinal cohort study demonstrated that MLPT infants exhibited developmental delays compared with their term-born peers, with the most pronounced deficits observed in the language domain [4]. Similarly, the Bavarian Longitudinal Study has revealed that MLPT infants were at significantly higher risk for poorer language outcomes than term-born children [9]. Another recent study in Spain has reported a significant language developmental delay in the MLPT group compared with their full-term peers [17]. Meanwhile, studies on MLPT infants in Korea remain scarce. To the best of our knowledge, this is the first Korean study to comprehensively evaluate language performance in MLPT infants using both the SELSI and BSID-III, thereby providing valuable insights into their early language development.

In the K-DST language domain, no statistically significant differences in language delay were observed between early preterm and MLPT infants. Early preterm (8.7%) and MLPT (14.3%) infants scored below two SDs, with a slightly higher proportion in the MLPT group. When language delay was defined as a score below –1 SD, the prevalence was 30.4% in early preterm infants and 31.4% in MLPT infants. As the K-DST is a parent-reported questionnaire, these findings may reflect subjective parental perceptions rather than objective developmental differences [11]. In addition, the K-DST was administered to a higher proportion of early preterm infants (79.3%) than MLPT infants (66.0%), raising the possibility of selection bias, whereby MLPT infants who underwent the K-DST assessment may have been perceived to be at higher developmental risk. Both the BSID-III and SELSI, which are clinician-administered assessments, have demonstrated that early preterm and MLPT infants have an increased risk of language delays [12,13].

When examining other developmental domains of BSID-III, the MLPT group demonstrated better cognitive outcomes than the early preterm group. However, in the language domain, no significant differences were observed between the two groups, and both groups showed a high prevalence of language delay (mild language delay: 20.8% vs. 25.0%; significant language delay: 25.0% vs. 10.4%). Thus, language development may remain particularly vulnerable even in MLPT infants, supporting previous studies reporting pronounced developmental delays in the language domain [4,9]. Therefore, heightened awareness of the risk of language delay and timely initiation of speech and language interventions are crucial to support optimal developmental trajectories in this population.

The mechanisms underlying language delay in MLPT infants are likely multifactorial [18]. Although these infants usually avoid the severe complications observed in extremely preterm infants, subtle neurological immaturity, altered brain connectivity, and limited early language exposure may contribute to developmental vulnerability [19]. In addition, maternal psychological factors, including postpartum depression, can affect the quality of parent–infant interaction and consequently influence language development [20-24]. In the present study, maternal depressive symptoms were observed in 47.6% of mothers of early preterm infants and 41.2% of mothers in the MLPT group. Severe depressive symptoms (scores ≥25), which are more strongly associated with a Diagnostic and Statistical Manual of Mental Disorders-based diagnosis of major depressive disorder [24], were observed in 19.0% and 14.0% of the patients in the two groups, respectively. These findings indicate a high prevalence in both groups, with no statistically significant differences between them. Therefore, screening for maternal mental health should be considered an essential component of follow-up care to support optimal developmental outcomes in preterm infants.

As MLPT infants constitute the majority of preterm births, their developmental challenges have significant public health and socioeconomic implications [25]. The potential long-term consequences of language delay in this large population highlight the need for early detection and intervention strategies [26,27]. Language delay in early childhood can extend beyond communication difficulties, affecting later academic performance, social competence, and emotional well-being [6,8,28]. Therefore, recognizing and addressing language delay at an early stage is essential to promote optimal long-term developmental and educational outcomes [29-31].

This study has some limitations. First, it was conducted at a single center with a relatively small sample size, which may limit its generalizability. SELSI was performed in 16 (55.2%) early preterm infants and 25 (47.2%) MLPT infants. SELSI was preferentially administered to high-risk infants, which may have resulted in a relatively high proportion of language delays among the infants assessed (68.8% in early preterm infants and 56.0% in MLPT infants). However, when considering the entire cohort, the overall prevalence of language delay was 37.9% in early preterm infants and 26.4% in MLPT infants. In addition, objective clinician-administered BSID assessments were conducted in 90.6% of MLPT infants, of whom 35.4% scored <85. This finding is consistent with the observed need for language therapy in 26.4% of infants in this group. Second, owing to the small sample size, we were unable to analyze the risk factors for language delay in the MLPT group. Third, because the follow-up clinic at our institution enrolled only infants born at <36 weeks’ gestation, those born at 36 weeks’ gestation were not included in the present study, although they were classified as late preterm infants. Finally, because a full-term control group was not included, a direct comparison with term-born peers was not possible.

Nevertheless, this study has important clinical implications. To the best of our knowledge, this is the first Korean study to comprehensively assess language outcomes in MLPT infants using two standardized tools (SELSI and BSID-III). This dual-assessment approach provides objective evidence-based insights into language development, an area that has been relatively under-recognized in this population. Although a full-term control group was not included for comparison, the developmental performance of these preterm infants can still be interpreted relative to population norms because the BSID-III provides standardized normative data (mean±SD, 100±15), and SELSI results are reported as percentiles. A previous Korean study has demonstrated a significant correlation between the K-BSID-III language scale and SELSI scores, suggesting that the K-BSID-III is a useful tool for predicting language developmental delay, particularly in infants with mild or definite delay [32]. In the context of the limited data available in Korea, our findings add clinically meaningful evidence to support developmental surveillance and early intervention strategies for MLPT infants. Previous Korean studies have reported that 27.5% of infants born at ≥28 weeks of gestation showed language delay when assessed using the BSID-III [33]. Data from the Korean Neonatal Network (KNN) have indicated that 18% of very-low-birth-weight infants required language support by 3 years of age, although the KNN cohort included only infants born at <32 weeks, limiting long-term outcome data to very preterm infants [34]. In our study, speech therapy was administered to 26.4% of MLPT infants, highlighting that even MLPT infants remain at risk for language difficulties.

In conclusion, MLPT infants, who are often considered to be at a lower developmental risk, may experience language delays comparable to those of early preterm infants. These results underscore the importance of regular developmental surveillance, including formal language assessment at approximately 2 years of corrected age and early intervention to improve long-term neurodevelopmental outcomes in MLPT infants.

Article information Continued