Drugs used in various treatment protocols used to treat acute lymphoblastic leukemia

The data collected were entered in Microsoft Excel 2007, Version 12.0 (Redmond, WA: Microsoft Corp.), and further analysis was done in IBM SPSS Statistics for Windows, Version 20.0 (Armonk, NY, USA: IBM Corp.). Age, WBC count, platelet count, and hemoglobin were logarithmically transformed and treated as continuous variables. Socioeconomic classification was based on Kuppuswamy’s socioeconomic status scale revision for 2011.[10],[11] Height for age, weight for age, and weight for height (for under five) were computed using revised IAP growth charts 2015 (weight for height of children >5 years had a different categorization as per IAP charts 2015 and was therefore excluded).[12]

Frequencies of all variables were studied to understand the characteristics of the disease in the population. Associations between variables and CR and between variables and induction deaths were also studied. Event-free survival at 5 years was defined as the absence of events (induction failure, deaths during induction, relapse, or death during remission) for the entire 5-year period studied. Variables were also studied for their association with event-free survival at 5 years. All associations were examined on cross tabs using Pearson Chi-square test and Fischer's exact test. To make study of associations more meaningful, height for age, weight for age, and weight for height were transformed into two classes (two standard deviations below normal and above). Weight for height and immunophenotype were censored to exclude “no data available” categories. Variables which were found statistically significant in univariate analysis were further analyzed using logistic binary regression for multivariate analysis.

Results

Of the seventy-five individuals studied, 82.7% had age between 1 and 9 years, the median age of presentation being 4 years [Table 2]. Males contributed to two-third of the study population. About 11% of the children were stunted and >21% were underweight. Sixteen percent of the under-five children had acute malnutrition with four among them having severe acute malnutrition. Massive hepatosplenomegaly and lymphadenopathy were present in 56 (74.7%) and 58 (77.3%) children, respectively. Mediastinal mass was found in eight percent and CNS involvement in four percent of the cases. Fifty-nine percent belonged to FAB class ALL L1, forty percent to ALL L2, and one percent to ALL L3 [Table 3].

Clinical characteristics of 75 previously untreated patients at presentation

Laboratorial characteristics of 75 previously untreated patients at presentation

At the end of first induction cycle, 53 (70.7%) patients achieved CR [Table 4]. Thirteen children died during induction due to hemorrhage and/or infections and one child died due to hepatitis. Among the 29 (38.7%) patients who relapsed, one had testicular relapse, four had CNS relapse, and rest had bone marrow relapse. Fifteen children died during relapse, the single most important cause being infection. To be also noted are the 17 children who died during remission; infections alone lead to death in ten of these children, 3 died due to bleeding and infections, while 4 died due to other causes (hemorrhage – 1, hepatitis – 1, hemorrhage and congestive cardiac failure (CCF) – 1, and infections and CCF – 1). A total of 30 (40%) children lived for >5 years after diagnosis and 21 (28%) of them had event-free 5 years.

Frequency of various events among the 75 children studied

Univariate analysis

None of the parameters studied had a statistically significant association with the CR [Table 5]. Weight for height for under-five children (P = 0.029) and total count (P = 0.019) were found to be significantly associated with deaths during induction [Table 6]. Children with acute malnutrition and a higher WBC count were more likely to die during induction. Age, sex, socioeconomic class, height for age, weight for age, clinical, and laboratorial characteristics other than total count did not have a significant effect on induction deaths.

Relationship of various characteristics to complete remission

Relationship of various characteristics to induction deaths

Weight for age (P = 0.024), weight for height of under-five children (P = 0.009), and lymphadenopathy (P = 0.049) had a strong evidence of association with 5-year event-free survival [Table 7]. Underweight children, malnourished children, and children with lymphadenopathy had significantly poor chances of surviving 5 years' event free. Age, sex, socioeconomic class, height for age, clinical characteristics other than lymphadenopathy, and laboratorial characteristics did not have a significant impact on the 5-year event-free survival.

Relationship of various characteristics to event-free survival

Multivariate analysis

When the significant factors from univariate analysis were studied using multivariate model, only weight for height among under five remained significantly associated with induction deaths (P = 0.021). Similarly, only absence of lymphadenopathy was found to be significantly associated with event-free 5-year survival (P = 0.042).

Discussion

Leukemia treatment in developing countries is peculiar for number of reasons. Clinical variances, adverse socioeconomic, cultural, and political influences make establishment of resources, procurement of drugs, and timely intervention difficult. Even when the most advanced centers of cancer care have survival >90%,[1],[2] they seem distant to the poor resource settings in developing countries. However, recently, specialized centers of cancer care within India have started showing promising results.[13],[14] Recently, specialized many refuse referrals to these centers as they cannot afford to move to such locations and that it would mean not being able to simultaneously support family back home.

The multitude of factors previously described often leads to a late diagnosis of the disease. In addition, potential genetic differences and biology of ALL also contribute to the clinical picture at presentation. This study shows that males are more affected than females with a ratio of 1.7:1. The median age of presentation was found to be 4 years, a year higher than that described in western literatures.[1] However, contradictory to the popular teachings, age was not directly associated with 5-year event-free survival. This is however in line with drawings from Indian studies,[8] opening the question whether age should be considered as factor for risk stratification in our population. Age and sex also did not show any associations with treatment response and deaths during induction.

Kuppuswamy classification was the tool used to categorize socioeconomic class. All children belonged either to the lower/upper lower, middle/lower middle, and upper middle classes of Kuppuswamy classification. Lower/upper lower made 55% of the individuals. The data show that even for the upper middle class (15%), ours was the only treatment available/accessible. However, socioeconomic status did not have any significant effect on the outcomes.

Nutritional factors and concurrent infections are among the important factors adversely affecting survival in developing countries. They also influence toxicity of therapy due to effects on drug metabolism and excretion.[15],[16],[17],[18] A significant effort was made to study this relationship in our series. Results suggest that under-five children with acute malnutrition were more likely to die during induction and have poor chances of surviving 5 years' event free. Children who were underweight also showed similar association with event-free 5-year survival.

Almost 80% of the children had lymphadenopathy and about 75% of them had massive hepatosplenomegaly. Hepatomegaly had long been included as a risk factor in BFM protocols; however, lymphadenopathy, being less common, had not been studied.[3] Our study shows significant association between absence of lymphadenopathy and event-free 5-year survival. Whether this lymphadenopathy is the result of delayed diagnosis or due to the genetic, environmental, or biologic variations in the population remains unclear. Mediastinal mass and CNS involvement were found to be less frequent in our setting. Mediastinal mass, CNS involvement, and massive hepatosplenomegaly were not found to have a significant effect on outcomes.

The proportion of patients with total count >50,000 remained comparable to other studies.[3] Although there was a trend that total count >50,000 was associated with decreased event-free 5-year survival, it was not found to be significant (P = 0.061). However, a high total count at presentation was significantly associated with deaths during induction. Such trend was previously described, in the study by Rubnitz et al. at St. Jude hospital.[19] Platelet count, hemoglobin, and FAB class were not found to have significant effect on treatment response, induction deaths, or survival.

Immunophenotyping results could be obtained only for 9 patients during the period studied. Among the available nine, four had T-cell ALL. This suggests higher proportion of T-cell ALL in Kerala, like in other parts of the country.[4],[8],[20],[21]

Three treatment regimens were simultaneously used during the 5 years studied. None superior in terms of response to treatment or event-free 5-year survival. Each one contributed roughly the same to induction deaths.

The multitude of factors impeding the treatment had resulted in decreased survival rates compared to advanced centers within India and the world. Patients who achieved remission were considerably less (70%) and induction deaths were substantially high (17%). Study of the characteristics of leukemia shows that the population is much different from their western counter parts. And thus, foreign protocols need to be tailored considering the differences in our population, to increase survival and decrease toxicity. Further studies on the clinical and laboratorial characteristics of the disease and their associations with survival and toxicity are essential for formulating such modifications. Amid the adverse factors, an overall survival of 40% at 5 years, in a poor resource setting, is inspirational to further provide quality cancer care. By making specialized healthcare more accessible and giving financial support to needy, we can improve the life of many more. Twinning has been described as one other strategy to improve cancer care in developing countries.[5],[22],[23] This can also be tried within the country to make cancer care more accessible.

Conflict of Interest

There are no conflicts of interest.

References