Advances in neonatal care have reduced mortality among very low birth weight (VLBW) infants in Thailand; however, comorbidities and suboptimal postnatal growth remain major challenges [1–3]. Extrauterine growth restriction (EUGR), defined as growth below the expected intrauterine trajectory, has been associated with adverse neurodevelopmental and cognitive outcomes [4–7]. However, its definition remains controversial regarding the cut-off, timing of assessment, and classification. Static EUGR is commonly defined as body weight (BW) < 10th percentile at 36 weeks postmenstrual age (PMA), whereas longitudinal EUGR is defined as a decline in BW z-score > 1 SD from birth to 36 weeks PMA [8–10]. These approaches identify different groups of infants and may provide complementary information on postnatal growth. This study aimed to determine the prevalence of static and longitudinal EUGR and identify independent risk factors for each among VLBW infants.

From a total of 148 eligible infants, 23 died, and 18 were discharged or referred out from the studied hospital before PMA 36 weeks, as shown in Figure 1. The remaining 107 infants were enrolled in the study; 55 (51.4%) had static EUGR and 35 (32.7%) had longitudinal EUGR. The mean maternal age was 30.6 ± 6.47 years, and 81 infants (75.7%) were singletons. Baseline characteristics of the infants are shown in Table 1. Of the 107 infants, 32 (29.9%) were SGA at birth; the remaining 75 non-SGA infants were included in the sensitivity analysis.

Figure 1.

Study flow diagram.

Univariate analysis showed that static EUGR was associated with higher GA and SGA at birth and maternal morbidity. RDS requiring surfactant therapy and BPD were inversely associated with static EUGR. This finding likely reflects confounding by gestational age, as infants with static EUGR had a significantly higher gestational age at birth. In the multivariable analysis, SGA at birth was the only factor independently associated with static EUGR (adjusted OR 9.01, 95% CI 2.08–39.08, p = 0.003). Delayed attainment of full enteral feeding was associated with increased odds of static EUGR, although the association did not reach statistical significance (adjusted OR 2.57, 95% CI 0.93–7.13, p = 0.069) (Table 2).

A sensitivity analysis excluding infants born SGA (n = 75) is presented in Table 3. After exclusion of SGA infants, no factors were independently associated with static EUGR in the multivariable analysis. Delayed attainment of full enteral feeding remained positively associated with static EUGR (OR 2.29, 95% CI 0.78–6.68, p = 0.131), although the association was not statistically significant.

Univariate analysis of the same variables showed that longitudinal EUGR was associated with lower birth weight, lower gestational age, SGA status, NEC, BPD, delayed full enteral feeding, and longer duration of mechanical ventilation. Multivariable analysis identified delayed full enteral feeding as the only statistically significant factor associated with longitudinal EUGR (aOR 3.98, 95% CI 1.48–10.68, p = 0.006) (Table 4). NEC showed a strong univariate association (OR 5.36, 95% CI 1.67–17.23, p = 0.005) but was excluded from the multivariable model due to sparse events. This association remained statistically significant after excluding SGA infants (OR 3.75, 95% CI 1.26–11.16, p = 0.018), with a magnitude of effect consistent with the primary analysis, supporting the SGA-independence of this finding (Table 5).

The two definitions showed only partial overlap, with a substantial proportion of infants classified by one definition but not the other (Table 6). Of the 107 infants, 23 (21.5%) met both definitions of EUGR, 32 (29.9%) met the criteria for static EUGR only, 12 (11.2%) met the criteria for longitudinal EUGR only, and 40 (37.4%) met neither definition. Thus, only 41.8% (23/55) of infants with static EUGR also met the criteria for longitudinal EUGR, whereas 65.7% (23/35) of infants with longitudinal EUGR also met the static definition. Agreement between the two classifications was low, with an observed agreement of 58.9% and a Cohen's kappa coefficient of 0.19 (SE 0.09, p = 0.019), indicating only slight agreement beyond chance.

Extrauterine growth restriction (EUGR) is commonly defined as a weight below the 10th percentile at either hospital discharge or between 36 and 40 weeks of gestation(8). This “static” definition yields the highest prevalence across studies [12], and may be useful for basic nutritional screening. However, static EUGR does not account for body composition, linear or head growth, nor does it adequately reflect genetic growth potential or pre-birth conditions. Most notably, its association with long-term neurodevelopmental outcomes is relatively weak when compared to longitudinal definitions [13–15].

Longitudinal EUGR has been defined as a decline in body weight z-score of >1 SD from birth to 36 weeks PMA. This approach incorporates birth conditions and genetic background and has demonstrated greater predictive value for neurodevelopment, including lower Bayley II Mental and Psychomotor Developmental Index scores [16,17].

In the present study, 51.4% of infants met the static criteria, and 32.7% met the longitudinal criteria. The reported prevalence of EUGR varies substantially depending on the definition and study population. For example, Yapicioglu Yildizdas et al. reported 95.3% prevalence using longitudinal criteria [18], while a study from Thailand identified 64.8% with static EUGR [19]. Agreement between the two definitions was quite weak in the studied cohort (Cohen’s kappa coefficient of 0.186), indicating that the two classifications identify largely different groups of infants and should not be used interchangeably.

Although delayed attainment of full enteral feeding was associated with more than a twofold increase in the adjusted odds of static EUGR, the confidence interval was wide and included the null value. Therefore, the present study cannot exclude either a clinically important association or no association, and larger studies are needed to obtain a more precise estimate.

The association between higher gestational age and static EUGR may appear counterintuitive but is inherent to the static definition, which classifies infants according to weight at 36 weeks' postmenstrual age regardless of their postnatal growth trajectory. Infants born at later gestational ages have less time to achieve catch-up growth before assessment at 36 weeks, whereas more immature infants have a longer interval for nutritional rehabilitation despite experiencing greater neonatal morbidity. Consequently, static EUGR may preferentially identify infants with fetal growth restriction rather than those with postnatal growth failure. The apparent inverse associations of RDS and BPD with static EUGR are likely explained by their close relationship with gestational age rather than a protective effect. More immature infants have higher rates of these morbidities but also a longer period before 36 weeks' postmenstrual age, reducing the likelihood of meeting the static EUGR definition. These findings highlight an important limitation of the static EUGR definition. Because it depends on a single anthropometric measurement at 36 weeks' postmenstrual age, static EUGR may be influenced by gestational age at birth and fetal growth status rather than reflecting true postnatal growth failure. In contrast, longitudinal EUGR, based on the decline in weight z-score over time, may better capture postnatal growth trajectory and the impact of nutritional factors.

In a sensitivity analysis restricted to infants who were not small for gestational age (SGA) at birth (n = 75), no factors were independently associated with static EUGR. Delayed attainment of full enteral feeding remained positively associated with static EUGR (OR 2.29, 95% CI 0.78–6.68), consistent with the direction of the association observed in the full cohort (adjusted OR 2.57, 95% CI 0.93–7.13), although neither association reached statistical significance.

In univariate analysis, longitudinal EUGR was associated with lower birthweight, lower GA, SGA status, NEC, BPD, delayed full enteral feeding, and longer duration of mechanical ventilation. NEC showed a strong univariable association (OR 5.36, 95% CI 1.67–17.23, p = 0.005). However, because of the small number of NEC events and the limited number of longitudinal EUGR outcomes, NEC was not included in the primary multivariable model to avoid unstable parameter estimates. In addition, NEC is closely related to illness severity and enteral feeding practices, making it difficult to disentangle its independent association within a parsimonious model.

In the multivariable model, delayed achievement of full enteral feeding was the most consistently associated factor. In the full cohort (n = 107), delayed full feeding was associated with markedly higher odds of longitudinal EUGR (aOR 3.98, 95% CI 1.48–10.68, p = 0.006), while SGA at birth was not associated with longitudinal EUGR (p > 0.05). This pattern contrasts with the static EUGR analysis (where SGA strongly predicted static EUGR) and supports the conceptual advantage of a longitudinal definition: It is less driven by birth size and more reflective of postnatal growth trajectory (i.e., growth faltering over time rather than being small at birth). Importantly, the association between delayed full feeding and longitudinal EUGR remained in the sensitivity analysis restricted to infants who were not SGA at birth (n = 75; OR 3.75, 95% CI 1.26–11.16, p = 0.018). Clinically, delayed full feeding likely captures a combination of factors—including feeding intolerance, gastrointestinal immaturity, and overall illness severity — that can limit cumulative energy/protein delivery during a critical period for growth. The present findings are consistent with previous studies demonstrating an association between inadequate nutritional intake and postnatal growth failure. Hu et al. reported that every 100-kcal/kg/day deficit was associated with a 0.1 SD decline in weight z-score [20], and Khasawneh et al. demonstrated that early caloric intake < 100 kcal/kg/day by day 15 was strongly associated with EUGR [21]. Collectively, these findings underscore the importance of careful nutritional management during the early neonatal period and suggest that infants with delayed attainment of full enteral feeding should be closely monitored for longitudinal EUGR. Breastfeeding at discharge was inversely associated with longitudinal EUGR in the non-SGA sensitivity cohort (OR 0.18, p = 0.020) but was not included in the multivariable model as it represents a discharge outcome rather than a risk factor preceding EUGR development.

The differing patterns of associated factors suggest that static EUGR is more strongly influenced by fetal growth status at birth, whereas longitudinal EUGR is more closely associated with postnatal clinical course and nutritional management. This distinction is clinically meaningful, as longitudinal EUGR may be the more appropriate measure for assessing adequacy of postnatal nutritional care. The present findings contribute to the growing evidence that nutrition remains an important and potentially modifiable factor associated with postnatal growth outcomes in VLBW infants.

This study has several limitations. First, its single-center retrospective design may limit generalizability to NICUs with different patient populations, nutritional protocols, or resources. Second, registry data lacked several potentially relevant variables, including early-onset sepsis, detailed parenteral nutrition composition, timing of breast milk fortification, and donor human milk use, resulting in potential unmeasured confounding. Third, the relatively small sample size limited the number of variables included in multivariable models, and sparse events precluded adjustment for some factors, notably NEC. Fourth, excluding infants who died, were transferred, or were discharged before 36 weeks PMA may have introduced survivor selection bias; baseline characteristics of these infants were unavailable, limiting applicability to the sickest VLBW infants. Finally, the BPD definition (supplemental oxygen > 28 gt; 28 days) included all severity grades, which may have heterogeneous effects on nutritional tolerance and growth. Severity-stratified BPD definitions may provide more precise estimates in future studies.

In VLBW infants, static EUGR at 36 weeks PMA was strongly associated with SGA at birth, suggesting that it partly reflects pre-existing fetal growth restriction. Longitudinal EUGR identified a partially distinct group and was associated with delayed achievement of enteral intake of 120 mL/kg/day. These associations do not establish causality, as delayed feeding may reflect illness severity and other clinical factors. Prospective studies incorporating detailed nutritional and illness-severity data are needed to identify modifiable determinants of postnatal growth.

During the preparation of this work, the author(s) used Claude (Anthropic, San Francisco, CA, USA) to assist with English language editing and manuscript preparation, including sentence refinement, grammar correction, and structural organization during manuscript revision. After using this tool, the author(s) reviewed and edited the content as needed and take full responsibility for the content of the published article.

No funding was received for this study.

The de-identified datasets generated and/or analyzed during the current study are not publicly available due to patient privacy and ethical restrictions. Data are available from the corresponding author upon reasonable request, subject to approval by an appropriate institutional ethics committee.