Abstract 1103
Poster Session IV, Tuesday, 5/4 (poster 358)
Neonatal jaundice results from the transient imbalance between bilirubin (B) production and elimination. End-tidal carbon monoxide (CO) corrected for inhaled CO (ETCOc) measurements reflect B production, while serum total B (STB) and transcutaneous B (TcB) measurements reflect the balance between the two processes. It has been suggested that because the levels of cord blood STB reflect erythrocyte turnover and the levels of serum alpha-fetoprotein (AFP), and possibly albumin (ALB), reflect liver maturity, these parameters may be used to predict the occurrence of hyperbilirubinemia after birth. We evaluated the usefulness of these parameters for identifying infants at high risk for hyperbilirubinemia [STB ≥ 95th percentile (Bhutani VK et al, Pediatrics 1999;103)]. 154 term- and near-term (≥ 35 wk) healthy Chinese newborns were studied. Cord blood was analyzed for AFP (AFP0), ALB (ALB0), and STB (STB0). Serial STB and TcB (using the Model 102 Minolta/Airshields Jaundice Meter) were performed at 30h (STB30, TcB30), 48h (STB48, TcB48), 96h, and later until STB reached ≥ 95th percentile or STB < 40th percentile or age of baby > 168h. ETCOc (using the CO-Stat™ End Tidal Breath Analyzer, Natus Medical Inc., San Carlos, CA) was measured at 30h (ETCO30) and 48h (ETCOc48). Receiver-operator characteristic (ROC) curves were constructed for each parameter to select optimal cutoff values. The sensitivity, specificity, positive, and negative predictive values for each parameter or combination of parameters were then determined. The predictive value of each parameter, as reflected by the area under curve (AUC), was ranked in the following order: STB48 (0.89), STB30 (0.85), ETCOc48 (0.73), STB0 (0.71), and ETCOc30 (0.64). AFP0 (0.54) and ALB0 (0.40) showed no discriminatory utility. Using any individual parameter alone yielded no predictive value. The combination of STB30 (cutoff ≥7 mg/dL) and ETCOc30 (cutoff ≥ 1.7 ppm) was unsatisfactory with sensitivity of only 88% and specificity of only 48%. Though the combination of STB48 (cutoff ≥10 mg/dL) and ETCOc48 (cutoff ≥ 2.0 ppm) gave the best sensitivity (100%) and specificity (70%), it only identified risk at 48h and thus the treatment of early onset hyperbilirubinemia may be delayed . We found that the most useful combination was STB30 (cutoff ≥ 7 mg/dL) and ETCOc48 (cutoff ≥ 1.9 ppm), which represented a sensitivity of 100% and a specificity of 64%. As in previous studies, we found a very strong correlation between TcB and STB (r=0.86 at forehead, r=0.93 at sternum) but the AUC and thus the discriminatory performance were less. When combining TcB30 with ETCO48, the sensitivity was 100% and specificity was decreased to 50-56%. We therefore conclude that hyperbilirubinemia in healthy Chinese term- and near-term infants can be reliably predicted by the combination of STB30 and ETCOc48 measurements. Furthermore, the combination of TcB30 with ETCOc48 is an attractive noninvasive alternative.
