Affiliations: [a]
Department of Neonatology, Kanagawa Children’s Medical Center, Yokohama, Kanagawa, Japan
| [b]
Department of Biostatistics, School of Medicine, Yokohama City University, Yokohama, Kanagawa, Japan
Correspondence:
[*]
Address for correspondence: Junichi Saito, MD, PhD, Department of Neonatology, Kanagawa Children’s Medical Center, 2-138-4 Mutsukawa, Minami-ku, Yokohama, Kanagawa 232-8555, Japan. Tel.: +81(45)711 2351; Fax: +81(45)721 3324; E-mail: [email protected].
Abstract: BACKGROUND:Serum interleukin-6 (IL-6) may predict adverse outcomes of neonatal encephalopathy (NE); however, limited data regarding the predictive utility of IL-6 during neurodevelopmental follow-up are available. We aimed to determine the utility of IL-6 for predicting adverse outcomes at 18 to 22 months of age. METHODS:Eighty-seven patients with NE who received therapeutic hypothermia were enrolled in this study. Serial serum IL-6 levels during the first 3 postnatal days were collected. Patients were classified into three groups: 1) death, 2) survival with moderate to severe neurodevelopmental disability (NDD) at 18–22 months of age, and 3) survival without NDD (favorable outcome). The predictive ability of IL-6 was determined by the area under the receiver-operating characteristic curve (AUC). RESULTS:Serial IL-6 data of 80 patients with NE were available and showed peak levels on postnatal day 1; these levels gradually decreased toward day 3. By 18–22 months of age, 13 and 17 patients died and experienced moderate to severe NDD without death, respectively. Fifty patients experienced favorable outcomes. Higher IL-6 levels on day 1 predicted the composite adverse outcome (including death and survival with NDD; n = 30; AUC, 0.648). Higher IL-6 levels on day 1 predicted death (n = 13; AUC, 0.799), whereas higher IL-6 levels on day 1 predicted survival with NDD (n = 17; AUC, 0.536). CONCLUSIONS:The AUC of IL-6 that predicted survival with NDD was lower than the AUC of IL-6 that predicted death; therefore, IL-6 may have insufficient utility for predicting NDD without death.
