Reclassification of the Etiology of Infant Mortality With Whole-Genome Sequencing

Key Points Question What proportion of infant mortality is explained by genetic diseases? Findings In this cohort study of 112 infant deaths, single-locus genetic diseases were the most common antecedent of infant mortality (41%). Treatments positively associated with outcomes were available for 30% of these genetic diseases. Meaning The study results suggest that because treatable genetic diseases are associated with considerable infant mortality, strategies for neonatal diagnosis may be associated with decreased infant mortality.


Study Design
This was a retrospective cohort study with 3 observational arms (Figure).It was approved by the institutional review boards of the University of California, San Diego and San Francisco.They did not consider postmortem WGS to be human participants research and provided a waiver of informed consent for the study.All participants underwent standard genetic testing as clinically indicated.
Premortem WGS was performed with informed parental consent either as a clinical diagnostic test or in research protocols (ClinicalTrials.gov:[31][32][33] The indications for infant WGS were those published by Blue Shield-California.Postmortem WGS was performed for all infants with blood sample retains that were archived in the Rady Children's biorepository.Administrative data were obtained from public sources in the San Diego Study of Mothers and Infants from 2015 to 2019.Data from 2020 were not yet available.All data have been deidentified.Results were reported according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines for reporting observational studies.

Race and Ethnicity
Infant race and ethnicity were classified by parents and extracted from the EHR.Race and ethnicity options were defined by the EHR.Race and ethnicity were assessed since to our knowledge there is a paucity of information regarding the diagnostic use of genome sequencing in racial and ethnic minority groups and a dearth of reference genome sequences from racial and ethnic minority groups, from which the racial-specific and ethnicity-specific allele frequencies used in genome interpretation are determined.Infants with a racial and ethnic classification of "other" were those whose parents did not categorize them as American Indian or Alaska Native, Asian, Black or African American, Hispanic, Native Hawaiian or Pacific Islander, non-Hispanic, or White and included multiracial infants.The efficacy of interventions for diseases associated with infant mortality in this study was adjudicated with the Genome-to-Treatment system and a similar online compendium of treatable genetic disorders. 34,35
Postmortem WGS was of singleton proband samples.Whole-genome sequencing was performed with 2 × 101 nt and a depth of more than 30 fold (Illumina).Read alignment to GRCh37 and variant diplotype identification was with DRAGEN (Illumina) and included copy number and structural variant identification.Semiautomated interpretation was performed using MOON (InVitae), GEM, and Enterprise (Fabric Genomics) as described. 37,38Inputs were variant call files, manually curated lists of observed human phenotype ontology terms, and metadata.Reportable diplotypes were identified by filtering and ranking disease phenotype match, variant pathogenicity, and rarity using decision trees, bayesian models, neural networks, and natural language processing and classified according to American College of Medical Genetics and Genomics guidelines by molecular laboratory directors. 37Variants of uncertain significance were only included if located in a gene that was casually associated with a genetic disease whose expected clinical features in infancy clearly overlapped the observed phenotypes in the proband and was known to be associated with infant mortality (variants of uncertain significance suspicious).Whole-genome sequencing was interpreted once.Reanalysis of negative cases was not performed.

Adjudication of Efficacy of Interventions for Genetic Diseases
An expert panel undertook a structured adjudication of the indications, contraindications, efficacy, and evidence of efficacy of 9911 drug, device, dietary, and surgical interventions for 563 severe childhood genetic diseases, as described previously. 37Of these, 421 diseases and 1527 effective interventions (15%) were retained and integrated with 13 genetic disease information resources (Genome-to Treatment: https://gtrx.radygenomiclab.com).For diseases associated with infant mortality, the efficacy of interventions was adjudicated with the Genome-to-Treatment system and a similar online compendium of treatable genetic disorders. 38,39

Statistical Analyses
Groups were compared with χ 2 and Fisher exact tests.Unadjusted P values less than .01were considered significant.[40][41] Of 21 maternal and infant characteristics examined, 5 known risk factors for infant death differed significantly between the 46 infant deaths associated with genetic diseases and 66 without (Table 2).Premature birth, placental abruption, and maternal infection were more common in infant deaths without genetic diseases, and polyhydramnios was more common in genetic diseaseassociated deaths.

Comparison of Genetic Diseases in Infant Deaths and Survivors
To further understand genetic determinants of infant deaths, we compared all 112 infants who died with all 434 surviving infants who underwent diagnostic WGS as inpatients at the same health system during the same period [26][27][28][29][30][31][32][33]36 (Figure). Thee were no differences in sex, race, or ethnicity between infant deaths and survivors with or without underlying genetic diseases (eTables 1 and 2 in Supplement 1).However, single-locus genetic diseases were more common (46   1; eTable 3 in Supplement 1).[38][39][40][41] The genetic diseases identified in infants who died and those who survived differed ( a The candidacy of TAB1 as a cause of infant death will be presented elsewhere.[38][39][40][41][42][43] They had been admitted to the regional neonatal ICU with encephalopathy and   managed with immunoglobulin replacement and antibiotic prophylaxis.[47][48][49] Rapid WGS at neonatal ICU admission may have been followed by CVID12 diagnosis and treatment with intravenous immunoglobulin and promptly administered broad-spectrum antibiotics.[47][48][49] Infant 183 died of cardiac arrest on DOL 123.CACNA1C long QT syndrome 8 (LQT8; MIM#618447) was identified postmortem.Rapid WGS at neonatal ICU admission may have been followed by collection of a specific family history, electrocardiography, and, if abnormal, βadrenoreceptor blockade.50 Infant 212 died following a bradycardic arrest in the emergency department on DOL 42.PPA2 infantile sudden cardiac failure (ISCF; MIM#617222) was identified postmortem.51 There is a frequent association between ISCF and neonatal sudden cardiac death, and it has been treated with cardioverter defibrillator implant and heart transplant.52 Rapid WGS during either of 2 prior pediatric ICU admissions may have been followed by diagnosis and consideration of these interventions. 51,52assification of Cause of Death

JAMA Network Open | Pediatrics
Reclassification of the Etiology of Infant Mortality With Whole-Genome Sequencing Figure.Flow Diagram of the Observational Study of Infant Survivors and Infants Who Died Who Underwent Whole-Genome Sequencing (WGS) for Diagnosis of Genetic Diseases During Care at Rady Children's JAMA Network Open.2023;6(2):e2254069.doi:10.1001/jamanetworkopen.2022.54069(Reprinted)February 9, 2023 2/17 Downloaded From: https://jamanetwork.com/ on 09/17/2023 c The left side of the diagram represents 45 infant deaths who received WGS postmortem and 67 who received rapid WGS for diagnosis of a suspected genetic disease during intensive care unit (ICU) admission.The right side of the diagram represents the control group, comprising infant survivors who received rapid WGS for diagnosis of a suspected genetic disease during ICU admission.a2020 deaths projected from average of 2015 to 2019 total deaths.bStandard genetic tests included chromosomal microarray, gene, and panel sequencing.c

Table 2 .
Demographic and Clinical Characteristics of 112 Infants in San Diego County Who Died Who Received WGS a (continued) a A genetic disease was identified in 46 infant deaths.bTheraceand ethnicity of infants were classified by parents.Race and ethnicity options were defined by the electronic health record.Infants with a racial and ethnic classification of "other" were those whose parents did not categorize them as Asian, American Indian or Alaska Native, Black or African American, Hispanic, Native Hawaiian or Pacific Islander, and non-Hispanic White and included multiracial infants.cCourse classification: neonatal: death occurred at fewer than 27 days of life.Known condition: multiple hospitalizations or extended hospital stay with a known diagnosis.First presentation: first presentation or admission in a previously healthy child.

Table 3 .
Comparison of Death Certificates With WGS Findings in 45 Infant Deaths a (continued) Reclassification of the Etiology of Infant Mortality With Whole-Genome Sequencing a One certificate missing all fields was excluded.JAMA Network Open | PediatricsJAMA Network Open.2023;6(2):e2254069.doi:10.1001/jamanetworkopen.2022.54069(Reprinted) February 9, 2023 10/17 Downloaded From: https://jamanetwork.com/ on 09/17/2023 We classified the cause of death of the 112 infants according to the US Centers for Disease Control Wide-ranging Online Data for Epidemiologic Research guidelines 53 and single-locus genetic disease as a separate, single category (Table4).In the latter, genetic disease was the leading cause of death(46 deaths [41.1%]).The proportion of the 10 leading causes of death that were reclassified as genetic disease varied widely: 24 of 40 deaths (58%) with malformations were reclassified as genetic diseases.Malformations, which had been the leading cause of death, was relegated to third after exclusion of genetic disease.Congenital diaphragmatic hernia was the only common malformation in which genetic diseases were underrepresented (1 of 5).The second leading cause of death was prematurity(21 deaths [19%]), of which only 1 (5%) was reclassified as genetic disease.SIDS was the third leading cause of death (4 deaths [4%]), of which 1 (of infant 170) was reclassified as a genetic disease.Two other infants (183 and 212) died of isolated cardiopulmonary arrest that could be characterized as SIDS.Five of the 10 leading causes of death were unchanged, with inclusion of genetic disease as a category (maternal complications of pregnancy, prematurity, incidental events, complications of placenta, cord and membranes, intrauterine hypoxia/birth asphyxia, maternal pregnancy complications, and incidental events and hemorrhage).Thus, genetic diseases were not associated with infant deaths with known nongenetic risk factors.Eighteen of 30 deaths (57%) classified as "all other" were reclassified as genetic disease.

Table 4 .
Relative Proportions of Leading Causes of Death in 112 Infant Deaths in San Diego County Before and After WGS Guimier A, Achleitner MT, Moreau de Bellaing A, et al.PPA2-associated sudden cardiac death: extending the clinical and allelic spectrum in 20 new families.Genet Med.2022;24(4):967.doi:10.1016/j.gim.2022.02.002 53.US Centers for Disease Control and Prevention.Underlying cause of death, 1999-2020.Accessed January 11, 2023.https://wonder.cdc.gov/wonder/help/ucd.html 54.Manickam K, McClain MR, Demmer LA, et al; ACMG Board of Directors.Exome and genome sequencing for pediatric patients with congenital anomalies or intellectual disability: an evidence-based clinical guideline of the American College of Medical Genetics and Genomics (ACMG).Genet Med.2021;23(11):2029-2037.doi:10.1038/s41436-021-01242-6 55.Barwell J, Snape K, Wedderburn S. The new genomic medicine service and implications for patients.Clin Med (Lond).2019;19(4):273-277.56.Best S, Brown H, Lunke S, et al.Learning from scaling up ultra-rapid genomic testing for critically ill children to a national level.NPJ Genom Med.2021;6(1):5.doi:10.1038/s41525-020-00168-357.Lunke S, Eggers S, Wilson M, et al; Australian Genomics Health Alliance Acute Care Flagship.Feasibility of ultrarapid exome sequencing in critically ill infants and children with suspected monogenic conditions in the Australian public health care system.JAMA.2020;323(24):2503-2511. doi:10.1001/jama.2020.7671Comparison of sex, race, and ethnicity of 112 San Diego County (SD) infant deaths and 434 SD infant survivors who received WGS between 2015 and 2020 eTable 2. Comparison of sex, race, and ethnicity of 46 SD infant deaths with genetic diseases and 114 SD infant survivors with genetic diseases by WGS between 2015 and 2020 eTable 3. 131 single locus (Mendelian) genetic diseases identified by WGS in 114 of 434 SD infants who survived eTable 4. Demographic and clinical characteristics of 112 SD infant deaths who received WGS and 199 SD infant deaths who did not receive WGS between 2015 and 2020 eTable 5. Demographic and clinical characteristics of all SD infant deaths and the subset who had a Rady Children's Hospital (RCH) electronic health record (EHR) between 2015 and 2019 Abbreviations: NA, not applicable; WGS, wholegenome sequencing.aGeneticdiseases were classified as a single category.52.