Hypercapnia in hospitalized children and adolescents with anorexia nervosa as a predictive marker for readmission: a prospective study

PURPOSE: To determine whether hypercapnia is associated with risk of hospital readmission related to anorexia nervosa (AN) in children and adolescents. METHODS: We performed a prospective study of patients ≤ 18 years old admitted due to AN decompensation from November 2018 to October 2019. Both subtypes of AN, restricting subtype (AN-R) and binge-eating/purging subtype (AN-BP), were included. Study participants were evaluated upon admission, at discharge and six months after discharge. T-tests or Mann–Whitney U tests was used to compare means values. Pearson or Spearman correlations were used to measure the association between two variables. Logistic regression models were developed to evaluate the relationship between scoring methods and readmission. RESULTS: Of the 154 persons admitted during the study period, 131 met the inclusion criteria. Median age was 15.1 years. At admission, 71% of participants were malnourished and 33 (25%) had been previously admitted. We observed a marked decrease in venous pH and stable pCO(2) elevation during follow-up period. Hypercapnia at discharge was associated with a twofold increased likelihood of readmission and the odds of readmission increased as discharge pCO(2) rose. These findings did not depend on AN subtype or participant sex. Electrolytes persisted within the normal range. CONCLUSION: Hypercapnia and respiratory acidosis are common alterations in children and adolescents hospitalized due to AN decompensation. Hypercapnia persists for at least 6 months after discharge despite clinical improvement and is associated with higher odds of readmission. This is the first study to identify an abnormal laboratory finding as a potential predictor of readmission in AN. LEVEL OF EVIDENCE: IV: Multiple time series without intervention.