Study Population and Data Collection

This study includes all deaths reported to CDC among US children with acute hepatitis of unknown etiology with onset of October 1, 2021‒June 6, 2023. A child was reported if they initially met the criteria defined in the CDC health advisory and if the child’s death was related to their hepatitis episode, regardless of the time between hepatitis onset and death. Reported cases underwent further investigation including medical chart abstractions, caregiver interviews, laboratory testing, and, when possible, pathologic evaluation of liver tissue specimens. Health departments or clinicians completed medical chart abstractions using a standardized case report form to collect demographics, medical and illness history, laboratory test results, treatment received, and outcomes. Interviews were conducted with a parent or caregiver at the discretion of the health department to obtain additional information on each patient’s symptoms; healthcare use; ill contacts at home, school, or daycare; and other potential exposures such as diet and travel. In addition, autopsy reports, discharge summaries, death certificates, and other available medical records were requested for further review to better understand the course of illness and elucidate the possible role of adenovirus or other pathogens in development of severe acute hepatitis and death.

Specimen Collection and Testing

Laboratory testing for clinical management was not standardized; testing varied by individual patient and by clinical judgment. Most respiratory and fecal specimens were tested locally by multiplex PCR panel. In addition, testing for adenovirus by nucleic acid amplification testing (e.g., PCR or quantitative PCR) of blood, respiratory, or stool samples was requested at the discretion of the treating clinician and conducted at a clinical or reference laboratory. Available residual adenovirus-positive blood, stool, respiratory, and tissue specimens were sent to select reference laboratories or CDC for adenovirus typing using Sanger sequencing of the 6 hypervariable regions of the hexon gene (15). Formalin-fixed, paraffin-embedded (FFPE) tissues from liver biopsy, explant, or autopsy underwent routine evaluation at the clinical institutions or medical examiner’s office; available specimens were submitted to CDC. CDC’s evaluation of specimens included histopathology, immunohistochemistry (IHC), and conventional PCR and sequencing for adenovirus, as well as IHC and PCR for other infectious etiologies, as indicated based on the pathologic findings and clinical and epidemiologic history.

Record Review

All deaths reported to CDC as part of the hepatitis of unknown etiology investigation through June 6, 2023, are noted in this report. Some children were excluded as cases by the reporting jurisdiction as additional clinical information became available and an etiology was identified. For the remaining children, clinical and epidemiology teams at CDC reviewed all available medical records and related laboratory and pathological evaluations to develop consensus opinions on possible etiologies and confirm each child met the reporting criteria. Two pediatricians (P.A.G. and H.L.K.) independently evaluated reported underlying medical conditions to determine whether those conditions may have contributed to development of or increased risk for acute hepatitis. We excluded children with an etiology for their hepatitis from further investigation and analysis. We describe here in detail the children with no known or suspected etiology after review of all available records.

Statistical Analysis

We used descriptive statistics to summarize the data and results. We analyzed all data in R version 4.2.1 (The R Foundation for Statistical Computing, https://www.r-project.org). CDC reviewed the investigation protocol and confirmed that it was conducted consistent with applicable federal law and CDC policy (45 C.F.R. part 46.102(l)(2), 21 C.F.R. part 56; 42 U.S.C. §241(d); 5 U.S.C. §552a; 44 U.S.C. §3501 et seq.).

Eligibility and Record Availability

During the national investigation, a total of 23 pediatric deaths with hepatitis onset occurring during October 1, 2021–June 6, 2023, were reported to CDC from 13 US jurisdictions: Florida, Georgia, Indiana, Michigan, Missouri, Nebraska, New York City, Pennsylvania, Puerto Rico, South Carolina, Tennessee, Texas, and Wisconsin (Table 1). Fourteen (60.9%) of the 23 children did not meet the reporting criteria defined in the health advisory on the basis of detailed medical record and laboratory testing review by the reporting jurisdiction or CDC. The most common reason for exclusion was identification of an alternative etiology for the hepatitis (n = 12). Alternative etiologies were complications of sepsis from pathogens other than adenovirus (n = 6), congenital conditions (n = 3), and unintentional injury (n = 3) (Table 1). After review of all information provided to CDC, 8 (34.8%) children were confirmed to meet the reporting criteria and are further described in this report. One child’s eligibility could not be determined because records were unavailable.

Record availability differed by child (Table 2). A case report form was completed for all children, whereas an exposure questionnaire was available for 5 (62.5%). An autopsy had been conducted for 4 children; all those reports were available for review. A discharge summary was available for all 8 children and a death certificate was available for 6. Other available records included emergency department visit notes, admission history and physical notes, progress and consultation notes, and additional laboratory and radiologic testing results. FFPE autopsy tissue specimens from 3 children were submitted to CDC for evaluation; of those 3 children, 1 also had native liver explant submitted, and 1 a liver biopsy.

Demographics and Clinical Characteristics at Initial Visit

The median age at acute hepatitis onset was 2.4 (range 0.2–6.3) years (Table 3). Five patients (62.5%) were male and 3 (37.5%) female; 5 (5/8, 62.5%) identified as Hispanic or Latino ethnicity. At initial visit, 2 children were immunocompromised and receiving immunosuppressive therapy (1 for ongoing cancer treatment and 1 after liver transplant), 2 other children had conditions that may have increased their risk of developing hepatitis or complications of adenovirus infection in the right clinical circumstance (1 was small for gestational age; 1 had history of poor weight gain and possible gastrointestinal disorder), and 4 children had no known underlying conditions (Table 3; Appendix Table 1). Children with underlying conditions that may have increased their risk and those with no known underlying conditions are henceforth called previously healthy.

All children manifested hepatitis signs and symptoms (i.e., dark colored urine, pale stool, jaundice, or scleral icterus); most also reported other systemic symptoms (87.5%), gastrointestinal symptoms (87.5%), and respiratory symptoms (75.0%) during their illness course (Table 3; Appendix Table 2). The median duration of symptoms before emergency department or inpatient admission was 7 (range 1–36) days (Figure). Laboratory evaluations showed elevated transaminases; median peak ALT was 4,852 (range 219–5,648) U/L and AST 7,500 (range 786–10,000) U/L. Median peak total bilirubin level was 13.7 (range 1.3–19.4) mg/dL and INR was 8.0 (range 1.4–13.0) (Table 3). Before the hepatitis episode, all but 1 child had received >2 doses of hepatitis B vaccine >3 weeks before their hepatitis episode (62.5% had 3 or 4 doses) and 5 children had received 1 or 2 doses of hepatitis A vaccine >3 months before their hepatitis episode. One child had received COVID-19 vaccine >3 months before the episode.

Adenovirus and Other Pathogen Testing

For every child, >2 specimen types were tested for adenovirus by multiplex PCR panel (using respiratory or stool specimens) or by quantitative PCR (using blood specimens); the most common specimen types tested were blood (n = 7), respiratory (nasopharyngeal, n = 6), FFPE liver tissue (n = 6), and stool/rectal swab (n = 3) (Table 4). Four of the 8 children tested positive for adenovirus; each was positive in both respiratory specimens and whole blood. One child also tested positive in stool and another in both serum and liver tissue. Blood specimens had the highest adenovirus detection rate (4/7, 57.1%), particularly whole blood (4/5, 80%). Adenovirus typing data were available for 2 children.

Of the 4 children who tested positive for adenovirus, 2 (patients 1 and 2) had immunocompromising conditions before hepatitis onset; 1 was positive for adenovirus species B, type 7 (respiratory and stool samples), and 1 was positive for species C, type 1, in whole blood and species C in liver tissue. The third child (patient 3), initially immunocompetent, tested positive for adenovirus (type unknown) after becoming immunocompromised after a liver transplant received during the initial hospital stay; before the liver transplant, the child had tested negative for adenovirus in respiratory, whole blood, and liver specimens. The fourth child (patient 4) who tested positive for adenovirus (type unknown) had no underlying conditions but was <1 year of age. Of the 4 children who were negative for adenovirus (patients 5–8), all were considered previously healthy (Table 4; Figure; Appendix Table 1).

All children tested positive for >1 viral or bacterial pathogen (Table 5). After adenovirus, the most commonly detected pathogens were bacterial pathogens isolated in culture from normally sterile sites (blood or joint fluid, n = 4), followed by Clostridium difficile (stool specimens, n = 2), rhinovirus/enterovirus (respiratory specimens, n = 2), and respiratory syncytial virus (respiratory specimens, n = 2). In addition, 1 child had serologic evidence of acute Epstein-Barr virus infection. Of note, none of the children had evidence of acute SARS-CoV-2 infection; 2 children were reported to have had previous SARS-CoV-2 infection on the basis of serologic results in their medical records.

Liver specimens from 6 children underwent routine pathological examination at the clinical institution or medical examiner’s office; 3 children had FFPE liver tissue specimens submitted to CDC. In 3 children (patients 3, 6, and 8), FFPE liver tissue specimens were negative for adenovirus by IHC at the clinical institution. Of the 3 children for whom FFPE liver tissue specimens were submitted to CDC, 2 (patients 5 and 7) were negative for adenovirus by IHC and PCR at CDC. FFPE liver tissue from 1 child (patient 2) was positive by IHC and PCR at the clinical institution and CDC, and adenovirus species C was detected by PCR at CDC. In addition, adenovirus species C was detected in whole blood and respiratory samples (Table 4). In the same child (patient 2), liver tissue pathology was consistent with classic adenovirus hepatitis including smudge cells. Pathologic findings seen in liver varied in the other 5 children (e.g., acute hepatitis, massive necrosis, interface hepatitis).

Outcomes

The median duration of hospitalization was 14 days (range <1–18 days). Seven (87.5%) children had acute liver failure as noted on the medical abstraction form; 2 of those underwent liver transplantation during acute hospitalization (Table 3; Figure). Seven children had a death certificate or autopsy report available that described the cause of death, including liver-related causes (viral hepatitis, acute cholestatic hepatitis, acute liver failure, hepatic encephalopathy, acute liver transplant rejection), cardiopulmonary causes (acute respiratory distress syndrome, cardiopulmonary arrest), cerebral herniation, or multiorgan failure. Among the 4 children who tested positive for adenovirus, 3 had adenovirus noted in the death certificate or autopsy report. Adenovirus was not mentioned on the death certificate for the fourth adenovirus-positive child (patient 3), who tested positive only after receiving a liver transplant.