Department of Epidemiology and Biostatistics, The George Washington University School of Public Health and Health Services, Washington, District of Columbia
Correspondence to: Paul H. Levine, Department of Epidemiology and Biostatistics, The George Washington University School of Public Health and Health Services, 2100 West Pennsylvania Avenue, NW, 8th Floor, Washington, DC 20037.
Conflict of interest: In 2010 Dr. Levine was a consultant for Novartis Oncology. Novartis did not participate in the analysis or review the manuscript.
Chronic myeloid leukemia (CML) is a rare disease in children and represents approximately 2% of all childhood leukemia. This results in difficulty creating large cohorts of patients for pediatric CML research. The Glivec International Patient Assistance Program (GIPAP) is a patient-access program sponsored by Novartis Oncology and administered by The Max Foundation (MAX) that provides imatinib free of charge to patients in resource-restricted countries who are not able to afford this treatment.
GIPAP highlights a cohort of children (n = 3,188) with CML that provides novel insight into international trends in diagnosis, treatment, and survival. These trends can be compared to outcomes in developed nations to crudely assess the impact of an extended access program for CML treatment such as GIPAP.
Overall survival values for children treated for CML within the GIPAP (89%) suggest that imatinib is very effective in middle and low-income countries.
Chronic myeloid leukemia (CML) is a rare disease in children and represents approximately 2% of all childhood leukemia . CML results from the translocation of chromosomes 9 and 22, t(9;22), the Philadelphia (Ph) chromosome in myeloid precursors, resulting in the aberrant juxtaposition of the BCR and ABL genes and the BCR–ABL fusion protein. This leads to uncontrolled tyrosine kinase activity and the eventual malignant transformation to CML.
At the time of treatment in this cohort, imatinib was the standard therapy for CML, although, other tyrosine kinase inhibitors such as nilotinib and dasatinib have been used to treat CML. Imatinib was approved by the United States (U.S.) Food and Drug Administration in May of 2001 for treatment of CML. Before imatinib, 5-year survival rates for CML were only 33% . Today, 87% of newly diagnosed patients with chronic phase CML achieve complete cytogenetic remission [3, 4]. However, this treatment option is costly, as a 1-year supply of imatinib in the U.S. is $92,000.00 .
The Glivec International Patient Assistance Program (GIPAP) is a patient-access program sponsored by Novartis Oncology and administered by The Max Foundation (MAX) to provide imatinib free of charge to patients in resource-restricted countries who are not able to afford this treatment. In 2010, 90 countries were participating in GIPAP and 33,000 patients have benefited worldwide . Per GIPAP protocol, qualified physicians must diagnose CML patients that are Ph/BCR–ABL positive and are required to be involved throughout all stages of the patient's treatment . MAX determines eligibility based on program criteria and independently approves enrollment.
Prior research suggests that patients participating in GIPAP have a markedly younger age at diagnosis as compared to the United States [6, 7]. GIPAP provides an opportunity to systematically evaluate such findings with a larger number of childhood CML cases than previously reported in the literature . Little is known about CML in pediatric patients from developing countries and the GIPAP cohort allows for novel description of patient characteristics as well as outcomes. The aim of this study is to provide a large-scale descriptive epidemiological review of pediatric CML in the resource-restricted countries participating in GIPAP. A secondary aim is to compare these findings to the pediatric CML experience in the U.S.
Raw data from GIPAP (2002–2010) were collected and received by MAX. Parents provided informed consent and patients, of appropriate age, provided assent to be evaluated in the program. This process was maintained by the individual clinics according to their standard operating procedures. Age groups were divided into 0–5, 6–10, 11–14, and 15–20 years. The age categories were identical to those used in the Surveillance, Epidemiology, and End-Results (SEER) for purposes of comparison. Countries in the analysis were classified according to the United Nations World Bank macro regions: Europe, Africa, Asia, Latin America, and Oceania . Oceania was removed from the survival analysis due to a small number of cases. The European countries in GIPAP consisted of Albania, Moldova, and Russia. Data collected included: gender, date of diagnosis, date of birth, gender, date of first dose of imatinib (considered date of enrollment), country of diagnosis, dose of imatinib, phase of disease at time of enrollment, Ph/BCR–ABL status, and outcomes such as death date, adverse event date, or date of last contact (surrogate for last day that a dose was administered).
An analysis of variance was utilized to assess if age at diagnosis varied by macro region. Kaplan–Meier 3-year overall survival (OS) probabilities were estimated from the time of first dose of imatinib. Survival data, including date of death, were explicitly recorded only during 2002–2007. Over this time period, Cox proportional hazard models were used to evaluate prognostic factors associated with the risk of death. The hazard ratios were calculated to compare the hazard of death between various factors such as gender, age at diagnosis, time from diagnosis to enrollment, original dose, phase of disease at time of enrollment, and geograhic region. All analyses were performed using SAS version 9.2. Additional data on pediatric CML patients in the U.S. were collected from SEER 18 registries dated 2002–2009 . Institutional review board approval was granted for this retrospective review of the GIPAP data.
A total of 3,188 pediatric patients (0–20 years) made up approximately 10% of the overall GIPAP cohort, as the remaining 90% were adults. Table I displays the total number of participants and percentages by key variables. The proportion of patients increased by age-group. Seventy-eight (2,496/3,188) percent of patients were from Asia while 62.2% (1,933/3,188) were male. The mean age at diagnosis among this subset was 14.6 years (95% CI, 14.5–14.8 years) with Europe having a significantly higher mean age at diagnosis [16.1 years (95% CI, 15.1–17.1 years) P = 0.0076] compared to the other regions. Africa had the highest proportion of patients <10 years (15%) while patients from Europe had the highest proportion of patients between 15 and 20 years of age (79%).
Table I. Number and Percentage of Pediatric GIPAP Patients (0–20 Years) by Age at Diagnosis, Sex, Phase of Disease at Enrollment, Time From Diagnosis to Enrollment, and Imatinib History
Patients per macro region
Age at diagnosis (years)
Phase of disease at enrollment
Time from diagnosis to enrollment (days)
Taken imatinib previously
The majority (77.0%) of patients were enrolled in the chronic phase with variation by region; 36% of patients from Europe (Russia, Moldavia, and Albania) were enrolled in the accelerated phase as compared to 8% from the other regions (P < 0.0001). Figure 1 illustrates the Kaplan–Meier survival plot for the entire pediatric cohort (A) and then by phase of disease at enrollment (B), gender (C), age of diagnosis (D), macro region (E), and time from diagnosis to enrollment (F). Overall, pediatric GIPAP patients had a 3-year OS of 89.0% (95% CI 87.4–90.5). The 3-year OS for GIPAP patients did not significantly differ by age but ranged from 87.8% (95% CI 85.6–89.7) for diagnoses between the ages of 15 and 20 years to 92.5% (95% CI 86.5–95.9) for diagnoses between 6 and 9 years (P = 0.1812). Probability of OS in GIPAP had higher rates among all age groups, as compared to SEER, except in the 15–20 year old age group: GIPAP 87.8% versus SEER 91.3%. SEER data from the U.S. found that survival for 0–5 year olds was 74.6% (95% CI 44.7–89.9), 6–9 year olds was 84.4% (95% CI 50.4–95.9), 10–14 year olds was 79.3% (95% CI 63.5–88.9), and 15–20 year olds was 91.3% (95% CI 83.8–95.4).
Significant differences were found in the OS by phase of disease at enrollment (P < 0.0001). Figure 1(B) illustrates that more advanced phase of disease was associated with worsened survival. Those enrolled in the blast crisis phase had an increased risk of death 3.89 times (95% CI 2.51–6.02) higher when compared to those pediatric patients who were enrolled in either the chronic or accelerated phases. Significant differences were found in the OS by time from diagnosis to first treatment in GIPAP (P < 0.0001). Figure 1(F) illustrates that longer time from diagnosis to first treatment was associated with worsened survival. Overall, 32.2% (1,028/3,188) of participants had more than 365 days between the time of diagnosis and first treatment. The risk of death among those patients who commenced therapy with imatinib ≥365 days from diagnosis increased by 2.49 times (95% CI: 1.43–4.33) as compared to those who commenced imatinib therapy within 30 days from diagnosis (P < 0.0001). Europe (Russia, Moldova, and Albania) had the greatest percentage (79.3%, 46/58) of patients treated more than 365 days after diagnosis. The only prognostic factor remaining in multivariate analysis was phase of disease at enrollment. However, no significant differences were found between macro regions when considering survival.
GIPAP highlights a cohort of children with CML that provides novel insight into international trends in diagnosis, treatment, and survival. These trends can be compared to outcomes in developed nations to crudely assess the impact of an extended access program for CML treatment such as GIPAP. The descriptive epidemiology of childhood CML from this cohort can also provide insight into variations in outcome at the international level. Countries such as India were reported to have 95% of CML patients enrolled in GIPAP  but this may not be representative of the general CML population in all countries. An analysis of GIPAP data versus population-based data in the same countries is in progress.
Several findings in the GIPAP cohort were similar to those from the U.S. experience with childhood CML: (1) overall survival around 90%, (2) a more advanced phase of CML at time of enrollment was associated with decreased survival (accelerated or blast crisis), (3) an increase in time from diagnosis to enrollment was associated with decreased survival (especially for those enrolled greater than 1 year from diagnosis), and (4) gender was not associated with a change in survival. These findings are noteworthy as the GIPAP population comes from resource-restricted nations. If such similarities between regions with developed and underdeveloped healthcare systems are sustained, more support may be warranted for partnerships with pharmaceutical companies for future treatment interventions, similar to GIPAP, at the international level. This may streamline or expedite screening procedures and eligibility determinations to shorten the time from diagnosis to treatment. It is reassuring, however, that over time the delay in starting imatinib has decreased in the GIPAP cohort .
Survival comparisons were not formally assessed as the interpretation would be limited, but we propose several hypotheses for why the OS values were lower in SEER than in GIPAP for all age groups but the older adolescent (age 15–19 years) group. In the SEER population, it is possible that imatinib was not the primary chemotherapy used; although, by 2002 when the SEER data were collected, imatinib was the standard approach for CML treatment. In addition, therapies such as blood or bone marrow transplantation may have been used more often in the SEER population. The transplant-related mortality could then potentially contribute to the OS values being lower in SEER than in GIPAP for patients under age 15 years. The survival findings for patients under the age of 10 years may have also been affected by sample size. Only 12% (382/3,188) of the GIPAP pediatric cohort was under the age of 10 years, leading to a much smaller sample size when compared to children age 10–20 years (2,806/3,188, 88%). The survival findings for Europe also may have been affected by smaller sample size as it only included 31 patients. Furthermore, the lower survival values for the American population may be associated with demographic characteristics that result in financial, access, adherence or compliance issues [12, 13]. Not only is imatinib a relatively expensive treatment option, but it must be accessible to children as quickly as possible. There are also age-related challenges with ensuring appropriate adherence to oral medications in adolescents. This is not unique to imatinib but may vary from country to country depending on the interactions of societal norms and individual adolescent maturity level.
There is a lack of longitudinal follow-up of pediatric patients with CML treated with imatinib. In the GIPAP cohort, the duration of therapy with imatinib was not standardized and there was no analysis of outcome based on duration of treatment. Such information would be critical to assess the potential long-term complications and benefits of using this therapy in childhood CML, as the current standard of care is to remain on tyrosine-kinase inhibitor therapy indefinitely. A paucity of data exists on the true long-term complications of imatinib use in childhood cancer. Some small studies have begun to study this including a single institutional report of growth retardation after imatinib therapy . Therefore, any added insight into this area of much needed research would be beneficial. Clinical research through U.S. collaborative groups in childhood cancer has adequately examined imatinib and CML in the context of a Phase 1 trial . There was no need for additional pediatric pharmacokinetics or efficacy studies as this information could be extrapolated from adults to children given the rarity of the disease and its identical etiology and biology in children and adults. Other multi-center trials have successfully used imatinib as adjunct therapy for Ph positive childhood acute lymphoblastic leukemia . But as more molecular pathways are identified as treatment targets, it is likely that combined therapies including imatinib may be forthcoming. In such cases, additional information on the long-term side-effect profile of imatinib will be critical. There is also the possibility of a shift in care within the United States towards using second generation tyrosine-kinase inhibitors such as dasatinib for front-line therapy . This may also impact the need for future research on imatinib. Another consideration is the comparison of outcomes from GIPAP to other developed countries, such as France where a recent report found a 3-year overall survival for childhood CML of 98% . The medical infrastructure created through GIPAP may allow for novel longitudinal evaluation of large numbers of pediatric patients on long-standing courses of imatinib.
This novel cohort of pediatric CML patients provides the clinical platform for the development of unique cancer research opportunities. Most developing nations have lower cancer survival rates when compared to the U.S. , which surprisingly was not the case for this pediatric population with CML. The novel nature of this distinction leaves investigators with ample opportunity to pursue impactful research in the context of a global setting where cancer rates are predicted to increase and poor and developing countries will account for more than half of the global burden . The WHO estimates that 60% of all new cancer cases will be diagnosed in the least developed nations , leaving a patient access program such as GIPAP ripe for novel research leveraged to impact the treatment and outcomes of diseases such as childhood CML.