Trends and Resource Use for Kernicterus Hospitalizations in the United States

We conducted a retrospective serial cross-sectional analysis of the Healthcare Cost and Utilization Projects’ (HCUP) Kid’s Inpatient Database (KID) and identified neonatal hospitalizations with kernicterus between 2006 to 2016 in the United States.⁷  The KID is a sample of pediatric discharges from all community, nonrehabilitation hospitals in the United States participating in HCUP. KID is sampled at 10% of uncomplicated in-hospital births and 80% of other pediatric cases. The 2016 KID sample has 47 participating member states, 4200 hospitals with pediatric discharges, and almost 6 million weighted pediatric discharges.⁸  We used the sampling weights provided by the HCUP to generate national estimates.

KID was analyzed from 2006 to 2016, and all neonatal (age ≤28 days) hospitalizations were extracted. Neonatal hospitalizations were extracted using the ‘Neomat’ variable which identifies discharges with neonatal and/or maternal diagnoses and procedures. To prevent double counting, all admissions ≤28 days old that were transferred out to other facilities were excluded. A similar methodology has been applied before in other studies.⁶  Neonatal hospitalizations with kernicterus were identified using the International Classification of Diseases, 9th/10th Revision, Clinical Modification (ICD-9/10-CM) codes in the primary or secondary fields (Supplemental Table 3). To increase the reliability of our case definition of kernicterus, cases were required to have a diagnosis code for kernicterus and a procedure code for phototherapy or exchange transfusion.⁶  To reduce errors associated with polycythemia, we excluded those newborns who had a diagnosis code for polycythemia and a procedure code for exchange transfusion while analyzing rates for exchange transfusion.⁶ 

We studied baseline characteristics of the study population for potential confounding assessment. Patient-level characteristics such as gestational age, sex, race, median household income according to zip code (<$36 000, $36 000 to $44 999, >$45 000), primary payer (Medicare/Medicaid, private insurance, self-pay, or no charge) and hospital-level characteristics such as hospital location (urban or rural), hospital bed size (small, medium, and large), region (Northeast, Midwest, or North Central, South, and West), and teaching status were studied. Race was described on the basis of the categorization within the KID. The cost of hospitalization was estimated using cost-to-charge ratios available from HCUP.⁹  The cost was adjusted for inflation using consumer price index data.¹⁰ 

Continuous variables were reported as medians and IQRs, and categorical variables were reported as proportions with SE. Outcomes of neonatal hospitalizations with kernicterus were reported in terms of phototherapy use, exchange transfusion use, mortality, and length of stay. The incidence rate of kernicterus was calculated by dividing the number that met the case definition for kernicterus (numerator) by the corresponding total number of neonatal hospitalizations (denominator) and expressed as per 100 000 neonatal hospitalizations. Variables with cell counts of <10 were not reported in keeping with the HCUP guidelines.¹¹  Survey procedures were used to account for clustering. For trend analysis, the Cochran Armitage test for dichotomous dependent variables was used.¹²  Statistical analyses were performed by using (SAS) 9.4 (SAS Institute Inc. Cary, NC). Two-sided P value <.05 was considered significant.

A total of 16 094 653 weighted neonatal hospitalizations were identified between 2006 to 2016 within the KID, of which 20.5% had a diagnosis of jaundice. During the study period, 81 neonatal hospitalizations met the case definition for kernicterus, yielding an overall incidence of 0.5 per 100 000. Table 1 shows the demographic characteristics of neonatal hospitalizations with kernicterus. The rate of kernicterus (per 100 000) differed according to demographic and hospital characteristics and was higher among males (0.59), Asian or Pacific Islanders (1.04), preterm infants (0.58), the Northeast census region (1.2), and urban teaching hospitals (0.72). Between 2006 and 2016, the incidence of kernicterus decreased from 0.7 to 0.2 per 100 000 (P-trend .03), which is shown in Fig 1. Phototherapy, exchange transfusion, and the combination of both were performed in 92.5%, 29.8%, and 100% of neonates with kernicterus, respectively. The most common diagnoses associated with kernicterus was isoimmization (22%). Other associated diagnoses were glucose-6-phosphate-dehydrogenase deficiency, birth trauma, sepsis, and dehydration.

Table 2 shows that the overall median length of stay for kernicterus was 5 days (IQR, 3–8 days). The overall median inflation-adjusted cost of hospitalization was $5470 (IQR, $1609–$19 989). No mortality was noted in this cohort during the study period.

Using a large nationally representative database, we found a downward trend in the rate of hospitalizations for neonatal kernicterus. The current study addresses an important gap in our knowledge regarding the national level incidence rate of kernicterus in the United States in contemporary times. We found the overall incidence rate of kernicterus to be 0.5 per 100 000 neonatal hospitalizations. This is lower than the reported rate of 0.8 per 100 000 in California, and other developed countries such as the United Kingdom (0.9), Denmark (1.2), Sweden (1.3), and Canada (2).^4,13–16  This could be attributed to differences in the racial composition of the various populations, breastfeeding rates, case ascertainment, and different time periods studied.¹⁷ 

Burke et al previously showed that the rate of kernicterus decreased from 5.8 to 1.6 per 100 000 neonates between 1988 to 2005. The current study extends that of Burke et al, and we demonstrate the continued decrease in the rate of kernicterus from 0.7 to 0.2 per 100 000 neonatal hospitalizations between 2006 and 2016. This is very reassuring, and some of the possible reasons for this include the widespread adoption of universal predischarge bilirubin screening, increase surveillance for hyperbilirubinemia in the outpatient setting, institution of system-level measures to screen for hyperbilirubinemia and aggressively manage infants with high bilirubin levels, and increased phototherapy use. Several large studies have shown that universal bilirubin screening led to a reduction in the incidence of bilirubin levels >25 mg/dL.^16,18,19  The potential reasons for the occurrence of kernicterus are well described,²⁰  but its continued occurrence in the United States and other developed countries with advanced medical systems suggests that current strategies need to be reviewed and strengthened.

We also found that newborns of Asian or Pacific Islander and Black race or ethnicity had higher rates of kernicterus. This could be because of higher incidence of jaundice among Asians,²¹  genetic polymorphisms of the enzymes involved in bilirubin metabolism on the basis of ancestral origin, racial and ethnic differences in the occurrence of hemolytic disease of the newborn,^22,23  misperception of decreased risk of hyperbilirubinemia among Black infants,²⁴  health inequities resulting from social, economic, structural disparities that lead to low health literacy among mothers of minority races,²⁵  parental underrecognition of the importance of follow-up,²⁶  and possible disparities in care received during hospitalization and after discharge.^27,28 

Our study is limited given the use of retrospective administrative data. Clinical information such as serum bilirubin levels and the signs and symptoms of bilirubin encephalopathy, etc. which would have helped to ascertain cases were not available. We identified kernicterus-related hospitalizations using ICD-9/10 CM codes. These codes are not validated and do not make a distinction between acute bilirubin encephalopathy versus chronic bilirubin encephalopathy or Kernicterus or bilirubin-induced neurologic dysfunction (BIND). We limited our study to the neonatal population and, therefore, may not capture those cases which may have developed at a later stage or deaths related to kernicterus which occurred after the neonatal period. Given the rarity and limited data available on kernicterus, the use of administrative databases such as the KID is suitable for tracking the incidence of kernicterus over time.

Although kernicterus has been called a “never event,” identified on a list of “serious, egregious, preventable adverse events” since 2002,²⁹  its continued occurrence despite the decreasing trend in the United States is worrying. Equally worrying is the disproportionately higher rate in Asian and Black infants. A multipronged approach including designating kernicterus as a reportable event, strengthening newborn hyperbilirubinemia care practices and bilirubin surveillance, parental empowerment, and removing barriers to care can potentially decrease the rate of kernicterus further.

We acknowledge the Healthcare Cost and Utilization Project (HCUP) sponsored by the Agency for Healthcare Research and Quality (Rockville, MD) and its partner organizations that provide data to the HCUP. We also thank Paula Umscheid for proofreading the manuscript.