Although survival rates of children with cancer have increased substantially due to the introduction of intensive treatment protocols, these treatments themselves also lead to morbidity and mortality.
We aimed to 1) obtain an overview of the various causes of death among children with cancer, and 2) identify factors contributing to treatment-related mortality (TRM).
Eligible participants were children diagnosed with cancer between 2003 and 2015 in the university hospitals of Groningen or Amsterdam (NL), and who had died between 2003 and 07/2017. Data on various demographic and disease/treatment factors were collected by use of a purpose-designed data collection form. Categorization of TRM or not-TRM was according to the classification of Alexander (2015), causes were further specified according to both this classification and the ICD-10.
402 deaths were included in the analysis. Of these, 84 (20.9%) were classified as TRM (causes see table 1), 317 (78.9%) were classified as not-TRM, and for 1 (0.2%) no classification was possible. In univariate logistic regression, several factors were significantly associated with increased occurrence of TRM; diagnosis of hematological malignancy, absence of relapse, allogeneic HSCT, and expected five year survival of >40% (vs. <40%). In multivariate analysis, diagnosis of hematological malignancy and absence of relapse remained significant.
We found that 1 in 5 deaths in children with cancer were treatment-related. In multivariate testing, diagnosis of hematological malignancy was associated with occurrence of TRM. The same was true for absence of relapse, which we think is due to the higher rate of tumor-related deaths in this group.