Thirty studies described in 25 articles met the inclusion criteria; 12,507 children were treated with anthracyclines. Nineteen studies recruited children at the start of anthracycline therapy and 11 studies included survivors.
The inter-observer agreement for selecting the studies for inclusion was 95%.
All studies were found to have methodological limitations (median quality score 3; range: 0 to 4).
The reported frequency of A-CHF varied between 0 and 16%.
The RR of possible risk factors of A-CHF were assessed in 10 studies. The following factors were found to be statistically significant: a higher cumulative dose (4/5 studies, i.e. 4 of the 5 studies), radiation therapy in the heart region (3/4 studies), age less than 4 years (1/1 study), children (1/2 studies), a higher maximal dose in 1 week (2/2 studies), daunorubicin (1/1 study), amsacrine (1/1 study), black race (1/1 study), female (1/1 study) and trisomy 21 (1/1 study).
The multivariate regression analysis showed that anthracycline type and the maximal dose in 1 week explained a considerable part (66%) of the variation in A-CHF frequency. The predictive frequency of A-CHF for patients treated with doxorubicin was 3.1% higher than for those treated with daunorubicin (95% CI: 0.6, 11.2). The predictive frequency of A-CHF for patients treated with a maximal dose of greater than 45 mg/m2 within 1 week was 5.8% higher than for those treated with a dose of less than 45 mg/m2 (95% CI: 1.7, 14.1).