Early Hyperglycemia and Risk of Infections in Preterm VLBW Infants in NICU: Data from a Multicenter, Randomized, Placebo-Controlled Trial

Paolo Manzoni; Elena Tavella; Ilaria Stolfi; Lorenza Pugni; Hubert Messner; Milena Maule; Mauro Stronati

doi:10.1055/s-0036-1592389

American Journal of Perinatology, Table of Contents

Early Hyperglycemia and Risk of Infections in Preterm VLBW Infants in NICU: Data from a Multicenter, Randomized, Placebo-Controlled Trial

Paolo Manzoni ¹, Elena Tavella ¹, Ilaria Stolfi ², Lorenza Pugni ³, Hubert Messner ⁴, Milena Maule ⁵, Mauro Stronati ⁶

¹Department of Neonatology and NICU, S. Anna Hospital, Torino, Italy
²Department of Neonatology, Policlinico Umberto I^, Roma
³Department of NICU, IRCCS Mangiagalli Hospital, Milano
⁴Department of NICU, Ospedale Regionale, Bolzano, Italy
⁵Department of Biomedical Sciences and Human Oncology, Cancer Epidemiology Unit, University of Torino, Italy
⁶Department of Patologia Neonatale, IRCCS Policlinico San Matteo, Pavia

Abstract

Background: Hyperglycemia in the early days of life has been occasionally reported as a possible risk factor for development of Candida infections in preterm ELBW neonates. Additional data are needed to assess the relationships between early hyperglycemia and late-onset sepsis in preterm neonates.

Objective: This is a secondary analysis of data from a multicenter RCT in Italy and New Zealand, whose original protocol was previously published. The trial aimed at assessing effectiveness of bovine lactoferrin (LF) supplementation (100 mg/day, alone or in combination with the probiotic LGG [10⁶ CFU/day] versus placebo) in prevention of late-onset sepsis (Manzoni et al, JAMA 2009) and NEC (Manzoni et al, EHD 2014) in VLBW infants. We tested the hypothesis that early (within the 5^th day of life) hyperglycemic (>200 mg/dL) spells can be associated with the occurrence of proven late-onset infections in preterm VLBW neonates.

Design/Methods: We analyzed the data of all infants enrolled in the RCT, and then separately for treatment groups (LF vs. placebo). Per the original protocol, daily surveillance and monitoring of glycaemia levels were performed and recorded. Management of hyperglycemia followed institutional protocols and clinical judgment of attending physicians. Multivariable logistic regression was performed to assess whether hyperglycemia was associated with late-onset culture-proven infections, and separately with infections by gram positives, gram-negatives, and fungal agents.

Results: Among 740 VLBW infants enrolled, 86 featured at least an episode of microbiologically confirmed late-onset infection, and 34 had at least one episode of early hyperglycemic spell recorded. After controlling for all variables significantly associated with infections (i.e., LF exposure, birth weight, gestational age), occurrence of at least one episode of early hyperglycemic spell retained a significant and independent association with the occurrence of infections only by Gram-positives (OR: 5.45; 95% CI: 1.92–15.42; p < 0.001) and fungal agents (OR: 3.37; 95% CI: 1.01–11.97; p = 0.04), but not by gram negatives. Of note, the day of onset of infections occurred significantly earlier in hyperglycemic infants compared with normoglycemic: 13.9 versus 20.1 mean days (p = 0.03), regardless of the pathogen.

Conclusion: Early hyperglycemic spells are significantly predictive of development of LOS by gram-positives and fungal microorganisms in preterm infants. Prophylactic strategies and reinforced monitoring should be addressed to these infants.

On behalf of the Italian Task Force for the study and prevention of Neonatal Fungal Infections; the Italian Society of Neonatology.