Increased health care utilization by survivors of childhood lymphoblastic leukemia is confined to those treated with cranial or total body irradiation: a case cohort study

Background: Previous studies have indicated that survivors of childhood acute lymphoblastic leukemia (ALL) have an increased morbidity measured in terms of health care utilization. However, earlier studies have several potentially important limitations. To overcome some of these, we investigated hospital contact rates, and predictors thereof, among 5-year survivors of ALL in a population-based setting, and compared them to a control cohort regarding outcome measures from a comprehensive nation-wide health register. Methods: All individuals diagnosed with ALL before the age of 18 in Southern Sweden during 1970-1999 and alive January 2007 (n = 213; male = 107) were identified through the Swedish Cancer Register. Each subject was matched to fifty controls, identified in the Swedish Population Register. All study subjects were linked to the National Hospital Register and detailed information was obtained on all hospital contacts (hospital admissions and outpatients visits) starting five years after cancer diagnosis, and the corresponding date for the controls, until 2009. Results: The median follow-up among the 5-year survivors of ALL was 16 years (range 5-33), accruing a total of 3,527 person-years. Of the 213 5-year survivors, 105 (49.3%) had at least one hospital contact compared to 3,634 (34.1%) of the controls (p < 0.001). Survivors had more hospital contacts (3 [1-6] vs. 2 [1-4] contacts, p < 0.001) and more total days in hospital (6 [2-18] vs. 3 [1-7] days, p < 0.001) than the controls during the study period. Logistic regression analysis showed that survivors treated with cranial irradiation and/or total body irradiation (45% and 7%, respectively) had an increased risk of at least one hospital contact (OR 2.3, 95%CI; 1.5-3.6 and OR 11.0, 95%CI; 3.2-50.7, respectively), while there was no significant difference between the non-irradiated survivors and controls. Conclusions: We show that irradiated survivors of childhood ALL have an increased morbidity measured in terms of hospital contacts, in comparison to non-irradiated survivors and controls, while non-irradiated survivors have not. These findings are encouraging regarding the future morbidity of children currently treated for ALL, as radiotherapy is necessary only for a minority of these.