Near-infrared spectroscopy as a diagnostic tool for necrotizing enterocolitis in preterm infants.

Abstract

We aimed to investigate whether splanchnic tissue oxygen saturation (rsSO2) measured by near-infrared spectroscopy (NIRS) could contribute to the early diagnosis of necrotizing enterocolitis (NEC). We retrospectively included infants with suspected NEC, gestational age <32 weeks and/or birth weight <1200 g in the first 3 weeks after birth. We calculated mean rsSO2, cerebral tissue oxygen saturation (rcSO2), variability of rsSO2 (coefficients of variation [rsCoVAR] = SD/mean), and splanchnic-cerebral oxygenation ratio ([SCOR] = rsSO2/rcSO2) in the period around the abdominal radiograph to confirm or reject NEC. Of the 75 infants, 21 (28%) had NEC (Bell's stage ≥2). Characteristics of infants with and without NEC differed only on mechanical ventilation and nil-per-os status. RsSO2 tended to be higher and rcSO2 lower in infants with NEC. RsCoVAR (median [range]) was lower (0.11 [0.03-0.34]) vs. 0.20 [0.01-0.52], P = 0.002) and SCOR higher (0.64 [0.37-1.36]) vs. 0.47 [0.16-1.09], P = 0.004) in NEC infants. Adjusted for postnatal age, mechanical ventilation, and nil-per-os status, a 0.1 higher rsCoVAR decreased the likelihood of NEC diagnosis with likelihood ratio (LR) 0.38 (95% CI 0.18-0.78) and a 0.1 higher SCOR increased it with LR 1.28 (1.02-1.61). Using NIRS, high SCOR may confirm NEC and high variability of rsSO2 may rule out NEC, when suspicion arises. Near-infrared spectroscopy may contribute to the diagnosis of necrotizing enterocolitis. When clinical signs are present a high splanchnic-cerebral oxygenation may indicate necrotizing enterocolitis. A low splanchnic-cerebral oxygenation ratio and high variability of splanchnic tissue oxygen saturation may rule out necrotizing enterocolitis. Whether a bedside real-time availability of the splanchnic-cerebral oxygenation ratio and variability of splanchnic tissue oxygen saturation improves NEC diagnosis needs to be further investigated.