Defining postnatal growth failure among preterm infants in Indonesia

**Background:** Postnatal growth failure (PGF) is common among preterm infants and is associated with poor neurodevelopment. Traditional single-point weight-based indicators (weight-for-age z-score <−1.28 or <−1.5 at discharge) have limitations, including inability to predict adverse outcomes and failure to account for normal postnatal weight loss. A newer indicator—decline in weight-for-age z-score ≥1.2 from birth to discharge—has been proposed but requires validation across different populations.

**Methods:** The Cohort of Indonesian PreTerm infants for long-term Outcomes (CIPTO) study is a prospective cohort study conducted at Cipto Mangunkusumo General Hospital (level III NICU), Jakarta, Indonesia. This report includes 650 preterm infants (<37 weeks gestational age) born between 2020–2021 who survived to discharge. Exclusion criteria included parents unwilling to participate, self-discharge against medical advice, or planned relocation. Anthropometric measurements were performed by trained staff using calibrated tools. PGF was defined using three indicators: (1) weight-for-age z-score <−1.28 (<10th percentile) at discharge, (2) weight-for-age z-score <−1.5 (<7th percentile) at discharge, and (3) decline in weight-for-age z-score ≥1.2 from birth to discharge (indicating moderate-to-severe malnutrition). Weight gain was calculated using three methods: early 1-point, average 2-point average, and exponential 2-point (g/kg/day). Full enteral feeding was defined as >120 kcal/kg/day. Statistical analyses used Mann–Whitney U-test, Kruskal–Wallis test, chi-square test, Cohen’s kappa, and Cox regression.

**Key Results:** Among 650 preterm infants (56.5% male, mean gestational age 33.1 weeks, mean birth weight 1,898 g), the prevalence of PGF varied dramatically by definition: 307 (47.2%) by weight-for-age z-score <−1.28, 270 (41.5%) by weight-for-age z-score <−1.5, and only 51 (7.8%) by decline in weight-for-age z-score ≥1.2. The traditional indicators paradoxically showed that infants classified as having PGF had significantly higher weight gain than those without PGF (e.g., for weight z-score <−1.28: median 8.79 vs. 5.72 g/kg/day by early 1-point method, p<0.001). In contrast, the decline-based indicator correctly identified infants with lower weight gain (median 4.44 vs. 8.04 g/kg/day, p<0.001). The decline-based indicator also identified that a history of invasive ventilation was a significant risk factor for PGF (RR=3.04; 95% CI: 1.41–6.53; p=0.004), while traditional indicators did not. Infants with PGF by the decline-based indicator had longer median duration of total parenteral nutrition (2 vs. 0.5 days) and longer time to achieve full oral feeding (10 vs. 5 days). Combination feeding (breast milk plus formula) was associated with lower risk of PGF compared to exclusive breast milk (RR=0.28; 95% CI: 0.12–0.65; p=0.003) by the decline-based indicator.

**Clinical Implications:** The decline in weight-for-age z-score ≥1.2 appears to be a more clinically meaningful indicator for identifying PGF among Indonesian preterm infants compared to traditional single-point indicators, which may overdiagnose PGF and lead to unnecessary or excessive nutritional interventions. This indicator correctly identifies infants with poor weight gain, prolonged parenteral nutrition dependence, and delayed oral feeding—all clinically relevant outcomes. The finding that invasive ventilation is a risk factor for PGF highlights a vulnerable subgroup requiring targeted nutritional support. The study supports adoption of this growth-pattern-based indicator in Indonesian NICUs, though limitations include single-center design, small numbers of extremely preterm infants, and lack of long-term neurodevelopmental outcome data to calculate sensitivity and specificity.