The University of Texas M.D. Anderson Cancer Center Houston, Texas

In 1972 WaIters et al reported that African-American children treated for acute lymphoid leukemia (ALL) had fewer and shorter remissions than European-American children.! Although extent of disease was greater and socio-economic status less in the African-American children, these factors did not entirely explain the difference in response to treatment. Since then several other studies have demonstrated differential therapeutic outcome in native American, Polynesian and Mexican-American children with ALL.2,3,4 The purpose of this presentation is to review the results of these studies and to discuss their possible biological and clinical implications.

African-American Children

The 1972 description of differences in remIssIon rates and survival between African-American and European-American children with ALL was notable because the children were all treated in the same hospital by the same physicians using the same protocols.' The authors reported that African-American children had a 74% remission rate, a median remission duration of 4.9 months and a median survival of 14 months. In contrast, European-American children had a 92% remission rate, a median remission duration of 14 months and a median survival of 23 months. Explanations for this difference were sought with partial success. Severe anemia, marked leucocytosis and massive splenomegaly were mo)re frequent in the African American children and socio-economic status tended to be less. However, after these unfavorable prognostic factors were accounted for, there was still a difference in outcome, suggesting a difference in the biology of the host and/or of ALL between the two ethnic groups. A 1985 report from the same hospital confirmed this difference in treatment outcome and provided more data.5 For 899 children diagnosed with ALL in the years 1972 to 1983 African-American children hald al median remission duration of 2.1 yealrs and a 5 year complete remission rate of 23% , one half of the values for European American children. Clinical and biological features were examined in a subgroup of 423 children who were diagnosed with ALL between 1979 and 1983 and trealted accolding to one protocol. By multivariate analysis, ages less than 2 years and over 10 years, initial white blood cell count greater than 50,000/µ I, F AB L2 morphology, pseudodiploidy and African-American ethnicity were each independent unfavorable prognostic factors. These findings supported the earlier conclusion that European vs. African ancestry in the southern United States influenced treatment outcome for children with ALL in addition to and independent from other clinical and biologic features. The Children's Cancer Study Group described remission and survival experience of 4,361 children enrolled in their ALL treatment studies between 1972 and 1982.6 Both remission duration and survival times were significantly superior in European-American children than in African-American children. They attributed this to differences in disease presentation. Fewer comparative data are av!ailable about differences between African children and European-African children. MacDougal1 compared 41 "Black" children and 89 "White" children with ALL treated between 1974 and 1983 in Johannesburg, South Africa.7 Median survival time was 25 months for Black children and greater than 60 months for White children. Multivarialte analysis of their clinical and hematologic features indicated that Black ethnicity and age were the most significant independent factors associated with short survival. In contrast, Black and White children with acute non lymphoid leukemia had no significant differences In remlsslon rate and survival.

Polynesian Children

A report from New Zealand described the survival of children with ALL diagnosed between 1981 and 1989 and treated at the same hospital according to the same protocols.3 The event-free survival rate for 21 Polynesian children was 42% and for 66 non-Polynesian children 68%. The difference, however, was accounted for by the 20% survival rate of the 10 Pacific Island Polynesian children; the survival rate of New Zealand Maori children was similar to that of' non-Polynesian children. The less favorable prognostic factors of the Polynesian children were thought to partly explain the difference in outcome but an additional undetermined factor was suggested. The disease-free survival rates of children treated for malignant solid tumor and f-or acute non-lymphoid leukemia at the same hospital were similar for Polynesian and non-Polynesian children, indicating that the apparent influence of ethnicity on treatment outcome was peculiar to ALL.

Native American Children

A recent report from New Mexico describes the survival of 28 native American children treated for ALL between 1969 and 1988 and compares it to survival of "Hispanic" or "non-Hispanic white" childlren.² All were treated in the same institutions according to the same protocols. Compliance with treatment was a major problem with the native American children, both during remission induction and continuation therapy. The median survival times were only 8 months for hoys and 37 months for girls as compared to 36 months for "non-Hispanic white" hoys and 140 months for "non-Hispanic white" girls. Multivariate analysis of prognostic factors indicated that ethnicity was a "possible hut confounded" independent prognostic. variable

Mexican-American Children

Although often referred to as "Hispanic", Mexican-American children in the southwestern region of the United States are largely of hoth Spanish and native American ancestry, as reflected in facial features and habitus. In the New Mexico study described above, there were no significant differences in survival between Mexican-American children --described as "Hispanic whites" -and "non-Hispanic white" children treated for ALL. However, a difference has been found in Texas. Culbert et al compared outcomes of 47 Mexican-American children and 39 European-American children residing in Texas who were diagnosed with B-precursor ALL between 1981 and 1985.4 All were treated at M .D. Anderson Cancer Center using the same Pediatric Oncology Group protocol. By standard prognostic criteria they had no significant difference in anticipated risk of relapse. However, with 4 to 8 year follow-up o)n all patients, 51% of Mexican-American children .were in continuo)us complete remission as compared to) 69 % of European-Americans. In a more recent study, awaiting publication, 80 Mexican-American children and 90 European-American children with ALL diagnosed from 1974 to 1985 were compared. All were Texas residents treated at M .D. Anderson Cancer Center according to a series of Pediatric Oncology Group protocols; follow-up was 6 to 17 years. The actuarial 8 year event-free survival rate was 39% for Mexican American children and 52% for the European-American children. The difference was greater for boys than for girls. On multivariate analysis of their clinical features and socio-economic status, Mexican American ethnicity was independently associated with reduced survival.

Discussion :

The influence of ethnicity on survival of children with ALL is subject to question because socio-economic factors are linked with ethnicity. Early diagnosis, ready access to quality health care, ability to purchase drugs and trave1 to treatment centers, sufficient time and energy to maintain compliance with treatment, and conviction of the value of established medicine over folk remedies are important eIements in surviving leukemia. They are less likely within less privileged communities either in developing or developed nations. Since social and economic privileges are closely linked with ethnicity throughout the world, the lower survival rates described for Africans, African-Americans, native Americans and Mexican-American children with ALL might be simply reflections of socio-economic factors . Several observations suggest the contrary. In the Memphis, Houston, Johannesburg, New Mexico and Auckland reports all the children were treated in the same manner in the same hospitals. Multivariate analysis of clinical features, including age, sex and extent of disease, and in two) reports socio-economic status, indicated that ethnicity was independently associated with survival. In the Auckland study no ethnic difference was found in survival rates of children with malignant solid tumors, and both in Auckland and in Johannesburg, none was noted for children with acute non-lymphoid leulkemia. This suggests that the independent influence of ethnicity on response to treatment may be peculiar to) ALL. Curative treatment of ALL is largely composed of. the administration of methootrexate (Mtx) and mercaptopurine (MP), antimetabolites that are less used in treatment of children with malignant solid tumors and '- AN LL. The speculation of Kalwinsky et al that genetic differences in metabolism ,and biovailability of these drugs might explain the ethnic influence on treatment outcome has some support. Genetic differences in metabolism of M P and Mtx have been demonstrated, although not linked with ethnicity. For example, Mtx effects are greater in children with trisomy 21.8 MP is more active in patients with gel1etic deficiency of thiopurine methyl transferase, an enzyme involved in its detoxificatiol.9 The medical challenges in overcoming the ethnic differences in survival of children with ALL are clear. First is to establish ready access to quality health care for all children in order too reduce the negative influences of socio-economic disadvantage and more advanced disease at diagnosis. Second is to investigate the possibility of ethnic differences in drug metabolism and biovailability of the agents commonly used in ALL so that their administration can be modified if. needed. Finally, ,as pointed out two decades ago, since it influences treatment results ethnicity of the patients should be considered in the design, ,analysis, and reporting of clinical trials of ALL therapy.

(1) Walters TR, Bushore M, Simone J. Cancer," 1972; 29:210-214.

(2) Foucar K, Duncan M H , Stidley CA, Wiggins CL, Hunt WC, Key CR. Cancer 1991; 67:2125-2130.

(3) Riugway D, Ske~l1 JE, Mauger DC, Becrott DM. Cancer 1991; 68:451-454.

(4) Culhert S, Frankel L, Hord M, Pinkel D. Proceeding of the American Association for Cancer Research 1990; 31: 201.

(5) Kalwinsky DK, Rivera G, Dahl GV, Roberson P, George S, Murphy SB, et al. Leuk Res 1985; 9: 817-823.

(6) Sather H, Honour R, Sposto, Level C, Hammond D.Pathogenesis of Leukemia and Lymphomas : Environment Influences. Magrath IT, O'Conor GT, Ramot B, ~ditors. New York: Raven Press, 1984:179-187.

(7) MacDougall LG. Leuk Res. 1985;9:765-767.

(8) Garre ML, Relling MV, Kalwinsky D, Dodge R, Crom WR, Abromowitch M. J Pediatr 1987; 111: 606-612.

(9) Lennard L, Lilleyman JS, Van Loon J, Weinshilboum RM Lancet 1990; 336:225-229.