Objective: Our objective was to clarify risk factors associated with the development of severe hepatotoxicity during antituberculosis chemotherapy in Japanese children.
Methods: In a retrospective analysis in a 350-bed referral children's hospital in a metropolitan area, the medical charts of all pediatric patients who received antituberculosis chemotherapy between January 1995 and November 1999 were surveyed. Univariate and multivariate analyses were performed to find any demographic parameters (ie, sex, age, height, and body weight), clinical characteristics (ie, nutritional or developmental status, co-infection with hepatitis viruses [B or C] or human immunodeficiency virus, presence of extrapulmonary tuberculosis, or medical history of liver disease), or individual antituberculosis agents used that would be associated with the likelihood of development of severe hepatotoxicity during antituberculous chemotherapy. Severe hepatotoxicity (defined as an elevation of ALT and AST levels that were greater than 5 times the respective reference values) was attributed to the chemotherapy when it developed in children with normal pretreatment values for these parameters and no other potential causes were identified. Those who had abnormal ALT or AST values before treatment were excluded from analysis.
Results: Among the 117 patients surveyed (58 males and 59 females; age range, 0 to 16 years), 18 were excluded from the analysis because of abnormal baseline ALT and AST values. Severe hepatotoxicity developed in 8 of the 99 eligible children, and all 8 of those children were younger than 5 years old. The univariate analysis revealed that the children in whom hepatotoxicity developed were significantly (P <.05) younger, were predominantly male, had extrapulmonary tuberculosis, and were given pyrazinamide more often than those who had no hepatotoxicity. However, the multivariate logistic regression analysis revealed that only age and the administration of pyrazinamide would have a significant contribution (P <.05) to the development of severe hepatotoxicity, with odds ratios of 143 (95% confidence interval, 4.2 to 4934.9) and 0.60 (95% confidence interval, 0.39 to 0.90), respectively: the estimated probability of development of hepatotoxicity in a typical pediatric patient at 1, 5, and 10 years receiving pyrazinamide with rifampin (INN, rifampicin) and isoniazid would be 0.95, 0.72, and 0.16, respectively.
Conclusions: This study indicated that intensive monitoring of hepatotoxicity should be performed for younger children (<5 years) receiving pyrazinamide for antituberculosis chemotherapy.