Characteristics and Treatment Outcome of Childhood Acute Lymphoblastic Leukemia: Comparison Between Ethnic Populations in Israel

Background and aims: Differences in outcome and disease characteristics of childhood acute lymphoblastic leukemia (ALL) among ethnic groups were reported worldwide. We explored such differences among ethnic populations in Israel.

Methods: Patient and disease characteristics and outcome were compared retrospectively among Jewish, Muslim, Bedouin, Druze, and Christian populations. Dose intensity and adherence to chemotherapy was measured by comparing dose and duration of treatment between the ethnic groups. Treatment duration was measured in days, from the first day of induction until the beginning of maintenance, compared to the scheduled time table according to the protocol, and expressed as percentage of delay. All data were retrieved from the Israeli National Center of Childhood ALL database after approval of the ethics committees of the participating institutions.

Results: Between the years 1984-2011, 1286 patients aged 1-21 years were enrolled into 4 Israeli National Studies (INSs) for childhood ALL. The patients included 963 Jews (75%), 211 Muslims (16%), 54 Bedouins (4%), 28 Druze (2%) and 30 Christians (2%). In the first ALL INS-84 protocol (132 patients) the outcome of Muslims was inferior to that of the Jews in patients with pre-B ALL (10 years event free survival (EFS): 33.3%±13.6 Vs 68.8% ±5.3 (p=0.043)). The other ethnic groups were too small for statistical analysis. In the three subsequent ALL-BFM-based protocols since 1989: ALL-INS 89, 98, and 2003/7 the outcome improved with time for both Jewish and Muslim patients, abrogating the earlier group difference. In the combined latter 3 studies, the Bedouins had a poorer outcome compared to Jews, Muslims, Druze, and Christians with 5-year EFS of 60.3%±7.2%, 80.4%±1.4%, 77.3%±3.2%, 84.0%±7.3%, and 88%+6.8% (p=0.020), respectively. The difference was mainly due to higher relapse rates in high risk (HR) group (5 year cumulative incidence of relapse (CIR) of 85.7%, 30.3%, 29.4% (p=0.003) in the Bedouin, Jew and Muslim respectively) and in T ALL (5 year CIR of 45.8%, 16.9%, 13.9% (p=0.05) respectively). Higher rate of death in remission (DIR) due to infections was observed in the medium risk (MR) group (5 year DIR of 11%, 1.5%, 0.9% (p=0.005) in the Bedouin, Jew and Muslim respectively) and in pre-B ALL (10.9%, 1.3%, 2.1% (p=0.001) respectively). As opposed to the other ethnic groups the current study did not improve outcome of Bedouins. No major differences in clinical and leukemia cell characteristics (age, gender, white blood cell count, phenotype, central nervous system involvement), as well as in early response (day 8 peripheral blood prednisone response, day 15 and 33 bone marrow morphology) and in BFM risk group stratification, could be detected between the Bedouin and the rest. However, the Bedouins had a tendency to higher incidence of t(1;19),and lower incidence of high hyperdiploid karyotype. Adherence to chemotherapy was compared between 27 Bedouins and matched 57 Jews and 43 Muslims, treated with the last ALL-INS 2003/7 protocol. Lower incidence of treatment delay was demonstrated among Jews compared to Bedouins and Muslims (P = 0.01 and 0.03, respectively), but no significant difference was found between Bedouins and Muslims (p=0.46).

Conclusion: The ethnic Bedouin group had inferior outcome with higher rate of relapse and death in remission compared to Jew, Muslim and Druze patients with no improvement over the years. Adherence to chemotherapy was similar to that of the Muslims. Other possible factors such as adherence to supportive treatment in a lower socioeconomic group and/or inherited factors associated with susceptibility to different subtypes of leukemia and to a different response to chemotherapy and toxicity might have contributed to the observed differences. Further studies are needed to investigate whether Bedouin ethnicity may serve as a marker for high risk stratification and whether pharmacogenetic data, yet unknown, contribute to their higher infection related toxicity. These data could aid in optimizing anti-leukemic and supportive treatment.