SEARCH

SEARCH BY CITATION

Keywords:

  • acute lymphocytic leukemia;
  • mortality;
  • trends;
  • children;
  • socioeconomic factors

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

BACKGROUND.

Mortality from childhood leukemia has declined substantially in developed countries but less markedly in the developing world. This study was designed to describe mortality trends in childhood leukemia and the impact of social inequalities on these trends in Brazil from 1980 to 2002.

METHODS.

Cancer mortality data by cause and estimates of resident population stratified by age and sex were obtained from the Brazilian Mortality Information System (SIM) for the years 1980 to 2002. Age-standardized (ages 0-19 years) mortality rates were calculated by the direct method using the 1960 world standard population. Trends were modeled using linear regression with 3-year moving average rates as the dependent variable and with the midpoint of the calendar year interval (1991) as the independent variable. The Index of Social Exclusion was used to classify the 27 Brazilian states. Pearson correlation was used to describe the correlation between social exclusion and variations in mortality in each state.

RESULTS.

Age-standardized mortality rates for boys decreased from 2.05 per 100,000 habitants in 1984 to 1.44 100,000 habitants in 1995, whereas the observed corresponding decline among girls was from 1.60 per 100,000 habitants in 1986 to 1.14 per 100,000 habitants in 1995. Statistically significant declining trends in mortality rates were observed for boys (adjusted correlation coefficient [r2] = 0.68; P < .001) and girls (adjusted r2 = 0.62; P < .001). Significant negative correlations between social inequality and changes in mortality were noted for boys (r = −0.66; P = .001) and for girls (r = −0.78; P < .001).

CONCLUSIONS.

A consistent decrease in mortality rates from childhood leukemia was noted in Brazil. Higher decreases in mortality were observed in more developed states, possibly reflecting better health care. Cancer 2007. © 2007 American Cancer Society.

Leukemias are the most frequent childhood neoplasm worldwide, corresponding to approximately 33% of all malignancies in the group ages birth to 14 years. Acute lymphocytic leukemia (ALL) is the most common subtype.1 Low incidence rates usually are observed in developing countries; although, in Brazil, the rates are similar to those reported in developed regions (4.1 per 100,000 boys and 3.6 per 100,000 girls).2

In the developed world, mortality from childhood had a decline >60% from the early 1960s onward,3 whereas smaller favorable trends were noted later in less developed regions, including South America.4 Advances in therapy have improved the prognosis of childhood leukemia dramatically in the last 30 years. In the U.S. and Europe, the current 5-year survival rates for children with ALL and acute myeloid leukemia (AML) are approximately 80% and 40%, respectively,5, 6 compared with mortality rates of 61% and 23%, respectively, observed from 1975 to 1984.5 Paradoxically, improvements in treatment have widened the gap of inequality between children in resource-rich countries and children in poor nations. The most important factors are availability of drugs at affordable cost and development of centers or groups of excellence to ensure the efficacy and safety of chemotherapy.7

Socioeconomic characteristics have been associated with discrepancies in health and with disease incidence and mortality in many developed countries.8–10 The relation between social inequalities and cancer has been studied well for adults,11 but there are few reports regarding the influence of those inequalities for childhood leukemia incidence, mortality, and survival.12–18

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Cancer mortality data and estimates of the resident population at ages birth to 4 years, 5 to 9 years, 10 to 14 years, and 15 to 19 years from 1980 to 2002 were derived from the Ministry of Health database (DATASUS). Leukemia deaths were coded according to the International Classification of Diseases (ICD). The ninth ICD revision was used from 1980 to 1995 (ICD codes 204–208), and the 10th revision was used from 1996 to 2002 (ICD codes C91-C95). Age-specific leukemia death rates for the groups ages birth to 4years, 5 to 9 years, 10 to 14 years, and 15 to 19 years were computed according to sex. Age-standardized mortality rates for children aged <19 years were obtained by the direct method using the 1960 world standard population. Trends in mortality were modeled using standard linear regression methods, with the age-standardized mortality coefficient (3-year centered moving average) as the dependent variable and the calendar year as the independent variable. A centralized variable (ie, X − 1991 [in which 1991 is the midpoint of the series]), was used to minimize colinearity of terms of the equation. Therefore, the model is represented by the following equation:

  • equation image

We also estimated the annual percent change (APC) for each age group by fitting a straight-line regression to the natural logarithm of the rates, with calendar year used as a regressor variable in a joinpoint regression analysis19 using the Joinpoint Regression Program (version 3.0). The 95% confidence intervals (95% CIs) were calculated for each estimated APC (EAPC). If these intervals excluded zero, then the EAPCs were statistically significant (P < .05). EAPC analyses could not be performed when an observation contained a zero mortality rate.

The Social Exclusion Index (SEI), which ranges from 0 (worst) to 1 (best), was used to classify the 27 Brazilian states. This index was built on the basis of 3 dimensions: suitable life conditions, knowledge, and youth vulnerability. The first dimension includes a poverty indicator (percentage of family heads with insufficient income in each city), an employment indicator (percentage of individuals aged >10 years working legally), and an inequality indicator (percentage of family heads with monthly income [in U.S. dollars] >$1000 and ≤$1000). The second dimension comprises 2 indicators: literacy (percentage of the population aged ≥5 years that is able to read and to write) and years of study (mean number of years of study for family head), and the last dimension includes a youth indicator (percentage of the population aged ≤19 years) and a violence indicator (homicides rate per 100,000 habitants). All indicators were obtained from the 2000 Demographic Census. We adapted these indicators to provide indexes by using a methodology similar to that applied for the Human Development Index,20 represented by the formula below.

  • equation image

The SEI, which is a synthesis of all 3 dimensions of social exclusion, was obtained through the sum of the 7 partial scores, which were weighted as follows: poverty (17.0), employment (17.0), literacy (5.7), education (11.3), youth (17.0), and violence (15.0). The lowest SEI was observed in the state of Maranhao (0.197), and the highest SEI was observed in the Federal District (0.850) (Fig. 1).

thumbnail image

Figure 1. Map of social exclusion, Brazil, 2000. AC, indicates Acre; AL, Alagoas; AP, Amapa; AM, Amazonas; BA, Bahia; CE, Ceara; DF, Distrito Federal; ES, Espirito Santo; GO, Goias; MA, Maranhao; MT, Mato Grosso; MS, Mato Grosso do Sul; MG, Minas Gerais; PR, Parana; PB, Paraiba; PA, Para; PE, Pernambuco; PI, Piaui; RJ, Rio de Janeiro; RN, Rio Grande do Norte; RS, Rio Grande do Sul; RO, Rondonia; RR, Roraima; SC, Santa Catarina; SP, São Paulo; SE, Sergipe; TO, Tocantins.

Download figure to PowerPoint

Pearson correlation coefficients were used to describe the correlation between social exclusion and variations in mortality in each Brazilian state. For all statistical tests, a Type 1 error (α) of 5% was adopted.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

From 1980 to 2002, 22,189 deaths from leukemia were registered among Brazilian children. Among boys, the highest age-adjusted mortality rate was observed in 1984 (2.05 per 100,000), and the lowest was noted in 1995 (1.44 per 100,000); whereas, among girls, the highest rate was 1.60 per 100,000 in 1986, and the lowest rate was 1.14 per 100,000 in 1995 (Table 1). The highest and lowest mean mortality coefficients for boys and girls were observed in the states of Rio Grande do Sul (boys, 2.42; girls, 1.87) and Maranhao (boys, 0.61; girls, 0.43) (Table 2).

Table 1. Age-specific and Age-standardized Childhood Leukemia Mortality Rates According to Year and Sex: Brazil, 1980–2002
YearMortality rate (per 100,000)
Age groupAge-adjusted
0–4 years5–9 years10–14 years15–19 years0–19 years
BoysGirlsBoysGirlsBoysGirlsBoysGirlsBoysGirls
19801.941.611.911.601.651.342.161.171.921.45
19811.981.821.781.251.331.282.011.221.791.42
19822.191.821.711.281.631.192.121.301.931.43
19832.151.781.981.261.711.301.721.561.911.49
19842.231.692.061.681.711.202.151.182.051.46
19852.031.631.571.311.511.041.94.991.781.27
19862.022.082.001.601.641.241.571.331.831.60
19872.001.281.801.361.721.151.641.121.811.24
19881.681.481.351.491.331.171.831.281.551.37
19891.781.551.581.031.371.061.821.131.651.22
19901.431.481.831.181.681.141.641.441.631.32
19911.661.411.631.251.571.221.921.141.691.27
19921.641.491.661.051.261.171.571.011.541.20
19931.261.241.661.061.471.151.931.121.561.15
19941.421.341.431.091.341.301.941.101.521.22
19951.251.201.631.031.211.011.711.301.441.14
19961.471.441.431.071.381.081.771.141.511.20
19971.571.461.651.161.671.001.841.181.681.22
19981.771.581.681.031.481.031.871.161.711.22
19991.551.541.521.071.351.061.681.241.531.25
20001.411.701.421.031.631.251.941.291.581.34
20011.551.521.421.231.350.981.871.091.541.23
20021.451.351.751.181.631.111.751.111.631.20
Table 2. Mean Childhood Leukemia Mortality Coefficient, Average Annual Variation, and Annual Percent Change in Mortality Rates, According to Sex and State of Residence: Brazil, 1980–2002
StateMean mortality coefficient (β0)Average annual increment (β1)EAPC
BoysGirlsBoysGirlsBoysGirls
  • EAPC indicates estimated annual percent change; NC, no change.

  • *

    Statistically significant.

  • Comprises 1991–2000.

Acre1.381.450.0600.005NCNC
Alagoas1.110.920.0350.0222.94*1.71
Amapa2.061.55−0.0070.079NCNC
Amazonas1.401.100.0040.0280.343.17*
Bahia0.980.76−0.0130.002−0.680.45
Ceara1.170.970.0390.0222.89*2.12
Distrito Federal2.231.65−0.053−0.061−2.53−4.25*
Espirito Santo1.871.35−0.028−0.037−1.67−2.94*
Goias1.991.36−0.062−0.034−2.54*−2.18
Maranhao0.610.430.0370.0326.38*6.77*
Mato Grosso1.311.230.0050.0180.33NC
Mato Grosso do Sul2.091.500.0000.0040.890.09
Minas Gerais1.801.39−0.048−0.043−2.51*−2.92*
Para1.270.920.0050.0050.440.58
Paraiba1.170.880.0010.0061.461.54
Parana2.171.62−0.054−0.019−2.58*−1.43
Pernambuco1.351.00−0.0000.006−0.080.71
Piaui0.730.630.0300.0314.27*NC
Rio de Janeiro2.351.81−0.042−0.044−1.98*−2.37*
Rio Grande do Norte1.331.110.0530.0423.36*3.63*
Rio Grande do Sul2.421.87−0.056−0.047−2.44*−2.46*
Rondonia1.220.75−0.0100.0350.36NC
Roraima1.151.25−0.1000.013NCNC
Santa Catarina2.111.84−0.012−0.028−0.12−1.40
São Paulo2.071.60−0.039−0.027−1.85*−1.62*
Sergipe1.090.920.0450.032NCNC
Tocantins0.950.670.0170.034NCNC

From 1980 to 2002, childhood leukemia mortality decreased significantly in both sexes (APC, −0.98; 95% CI, −1.36 to −0.61). A similar decrease was observed for girls (APC, −0.92; 95% CI, −1.36 to −0.48) and boys (APC, −1.04; 95% CI, −1.46 to −0.62). Statistically significant downward trends in mortality were noted in all age groups. The highest decrease in mortality was noted for children in the group ages birth to 4 years (APC, −1.47; 95% CI −0.79 to −2.15), followed by the subsequent age groups: ages 5 to 9 years (APC, −1.19; 95% CI, −0.65 to −1.72), ages 10 to 14 years (APC, −0.60; 95% CI, −0.11 to −1.09), and ages 15 to 19 years (APC, −0.42; 95% CI, −0.02 to −0.81). Among boys, the highest decrease in mortality was noted for children in the group ages birth to 4 years (APC, −1.95; 95% CI, −2.69 to −1.20), followed children in the group ages 5 to 9 years (APC, −0.99; 95% CI, −1.63 to −0.34). However, no statistically significant changes were noted for older boys ages 10 to 14 years (APC, −0.46; 95% CI, −1.17 to +0.25) and ages 15 to 19 years (APC, −0.42; 95% CI, −1.01 to +0.17). Among girls, the highest decrease in mortality was noted in the group ages 5 to 9 years (APC, −1.48; 95% CI, −0.71 to −2.24), followed by the groups ages birth to 4 years (APC, −0.93; 95% CI, −1.70 to −0.15), and ages 10 to 14 years (APC, −0.78; 95% CI, −1.28 to −0.28). No statistically significant decreasing trend was observed for girls in the group ages 15 to 19 years (APC, −0.36; 95% CI, −1.05 to +0.32).

The decreasing trends in mortality rates were most notable in the states of Distrito Federal (girls: EAPC, −4.25), Espírito Santo (girls: EAPC, −2.94), Goias (boys: EAPC, −2.54), Minas Gerais (boys: EAPC, −2.51; girls: EAPC, −2.92), Parana (boys: EAPC, −2.58), Rio de Janeiro (boys: EAPC, −1.98; girls: EAPC, −2.37), Rio Grande do Sul (boys: EAPC, −2.44; girls: EAPC, −2.46), and São Paulo (boys: EAPC, −1.85; girls: EAPC, −1.62) (see Table 2). Significant increasing trends were observed in the states of Alagoas (boys: EAPC, +2.94), Ceara (boys: EAPC, +2.89), Maranhao (boys: EAPC, +6.38; girls: EAPC, +6.77), Piaui (boys: EAPC, +4.58), Amazonas (girls: EAPC, +3.17), and Rio Grande do Norte (boys: EAPC, +3.36; girls: EAPC, +3.63) (see Table 2).

Linear regression models indicated a statistically significant downward trend in childhood leukemia mortality for both sexes (boys: β0 = 1.68; β1 = −0.019; adjusted correlation coefficient [r2] = 0.68; P < .001; girls: β0 = 1.30; β1 = −0.013; adjusted r2 = 0.62; P < .001) (Figs. 2, 3).

thumbnail image

Figure 2. Age-adjusted mortality rates and linear trend for girls in Brazil ages 0 to 19 years from 1980 to 2002.

Download figure to PowerPoint

thumbnail image

Figure 3. Age-adjusted mortality rates and linear trend for boys in Brazil aged 0 to 19 years from 1980 to 2002.

Download figure to PowerPoint

There was a statistically significant correlation between SEI and the change in mortality rates in each state for both boys (r = −0.66; P < .001) and girls (r = −0.78; P < .001) (ie, decreases in mortality were more prominent in the Brazilian states that had better socioeconomic conditions) (Figs. 4 and 5).

thumbnail image

Figure 4. Correlation between the Social Exclusion Index and the mean annual increment in childhood leukemia mortality rates for girls in Brazil from 1980 to 2002. AC, indicates Acre; AL, Alagoas; AP, Amapa; AM, Amazonas; BA, Bahia; CE, Ceara; DF, Distrito Federal; ES, Espirito Santo; GO, Goias; MA, Maranhao; MT, Mato Grosso; MS, Mato Grosso do Sul; MG, Minas Gerais; PR, Parana; PB, Paraiba; PA, Para; PE, Pernambuco; PI, Piaui; RJ, Rio de Janeiro; RN, Rio Grande do Norte; RS, Rio Grande do Sul; RO, Rondonia; RR, Roraima; SC, Santa Catarina; SP, São Paulo; SE, Sergipe; TO, Tocantins.

Download figure to PowerPoint

thumbnail image

Figure 5. Correlation between the Social Exclusion Index and the mean annual increment in childhood leukemia mortality rates among boys in Brazil from 1980 to 2002. AC, indicates Acre; AL, Alagoas; AP, Amapa; AM, Amazonas; BA, Bahia; CE, Ceara; DF, Distrito Federal; ES, Espirito Santo; GO, Goias; MA, Maranhao; MT, Mato Grosso; MS, Mato Grosso do Sul; MG, Minas Gerais; PR, Parana; PB, Paraiba; PA, Para; PE, Pernambuco; PI, Piaui; RJ, Rio de Janeiro; RN, Rio Grande do Norte; RS, Rio Grande do Sul; RO, Rondonia; RR, Roraima; SC, Santa Catarina; SP, São Paulo; SE, Sergipe; TO, Tocantins.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

To our knowledge, this is the first report analyzing time trends in childhood leukemia mortality on a national basis in Brazil. The assessment of mortality trends is a useful tool for monitoring changes in the epidemiologic profile of a population. In Brazil, important changes in age- and disease-specific mortality have occurred in the last 20 years. In 1979, infectious diseases were the main cause of death among individuals aged <19 years, and malignant neoplasms were responsible for only 1.1% of all deaths among individuals in this age group (ninth in the rank). Two decades later, these rates have changed dramatically: today, infectious diseases are the fourth most important cause of mortality in this age group, accounting for 6.8% of all deaths, and are surpassed by perinatal conditions (33.1%), external causes (22.6%), and congenital malformations (9.9%). By contrast, by 2004, neoplasms represented the sixth major cause of death in this age group (3.3%).21

The mean childhood leukemia mortality rate in the period was 1.68 per 100,000 for boys and 1.30 per 100,000 for girls. Among girls, this rate was lower than the rates observed in many countries, such as Italy, Bulgaria, Portugal, Argentina, Mexico, Costa Rica, the Russian Federation, and Romania.22 The mortality rate for girls described in the current study was similar to the rates observed in Australia, France, and Japan but still was higher than the rates observed in Finland, Germany, the U.K., the U.S., and Canada.22

From 1980 to 2002, childhood leukemia mortality in girls decreased by 17.2% (percent change), with an APC of −0.92%. This decline was lower than the downward trends noted in the U.S. (percent change, −60.4%; APC, −3.2 from 1975 to 2003).23

Boys in Brazil had a lower mean mortality rate than the rates reported among boys in Bulgaria, Hungary, Argentina, Costa Rica, Italy, Mexico, Portugal, Russian Federation, Spain, and Japan.22 The mortality rate for boys described in the current study was similar to the rates observed in U.K., France, Australia, and Hong-Kong but still was higher than the rates observed in Austria, Finland, Germany, Norway, Sweden, the U.S., and Canada.22

In this study, the observed reduction in mortality was slightly higher for boys than for girls. This finding is compatible with results reported from the U.S., where boys (ages birth-19 years) have shown the more favorable trends in childhood leukemia mortality (percent change, −58.0%; APC, −3.5 from 1975 to 2003).23 In Argentina and in many European countries, such as Bulgaria, Denmark, France, Hungary, and Portugal, this advantage for boys in mortality reduction also was reported; however, it should be noted that rates for boys consistently are higher to begin with.

In China, from 1987 to 1999, a sharp decline in childhood leukemia mortality was observed in the youngest age group (ages birth-4 years: boys-rural, −4.3; girls-rural, −5.9; girls-urban, −3.7), but the decline was less pronounced for boys living in urban areas (EAPC, −2.3); whereas, for the children ages 5 to 14 years, no significant changes were noted.24 This may be one of the only populations (ie, children in rural China) where rates are higher in girls than in boys. In Western European countries, significant reductions in mortality from childhood leukemia were observed over the last 40 years.25

Mortality data are very useful to evaluate the impact of treatment advances, particularly in those countries where incidence and survival estimates from cancer registries still are not broadly available.26 However, it is important to emphasize that the interpretation of mortality trends should be cautious, because these data are subject to many other influences. One important aspect is the quality of diagnostic data. A range of conditions may mimic acute leukemia and related disorders, particularly in the young child. There is a need of good-quality diagnostic material to permit morphologic analysis, immunophenotyping, and cytogenetics. An equivocal choice of treatment for patients with acute leukemia reduces significantly the possibility of successful treatment.27 Factors that influence the stability or even the increasing mortality observed in some Brazilian states (Alagoas, Ceara, Maranhao, Piaui, Amazonas, and Rio Grande do Norte) must be evaluated carefully. It is important to highlight that most of these states are situated at the northeast region of Brazil, where there is a reduced number of medical facilities that specialize in pediatric oncology. In addition, the mean number of physicians per 1000 habitants in Brazil is 1.42, but this number ranges from 0.21 physicians per 1000 habitants in Rondonia (north region) to 2.90 physicians per 1000 habitants in the Federal District.

Another possibility is that the observed increase in mortality may have been caused by an improvement in the quality of death certification. In fact, in several states that have shown an increase or stabilization in childhood leukemia mortality rates (Acre, Tocantins, Rio Grande do Norte, Amazonas, Amapa, Sergipe, Ceara, and Alagoas), we observed a concomitant, significant decrease in mortality rates from ill-defined causes. The only exception was the state of Maranhao, where the quality of death certification apparently worsened over the study period (data not shown).

An evaluation of trends in childhood leukemia incidence was not performed in this study, but data from the population-based cancer registry of the city of São Paulo demonstrated that, for the interval from 1969 to 1998, incidence rates tended to stabilize for both sexes.28 In Goiania, incidence rates for boys and girls during the intervals from 1988 to 1989, from 1990 to 1993, and from 1995 to 1998 were 2.5, 3.3, and 4.7, and 1.8, 2.7, and 2.7, respectively.29 Therefore, it is possible that the decline in mortality could be attributed to enhancements in survival rates.

The Brazilian Cooperative Group for Treatment of Childhood Acute Lymphocytic Leukemia started clinical trials in 1980. Three consecutive multicenter studies in 994 children with newly diagnosed ALL have been completed. In this population, event-free survival has improved from 50% in the period from 1980 to 1982 to 70% for children treated from 1985 to 1988.30 According to the Brazilian Society of Pediatric Oncology, there are 6 well-established cooperative groups with ongoing protocols for the treatment of leukemia, retinoblastoma, osteosarcoma, Ewing sarcoma, Wilms tumor, and germ cell tumors; and these protocols potentially may have an impact on survival and mortality rates. According to de Camargo (2004),31 the 5-year overall survival rate for children with ALL treated at the A. C. Camargo Hospital in São Paulo, Brazil, has improved dramatically, from a low of 13% during the 1970s to 55% at the end of the 1990s. It is noteworthy that that series included all children, even those who were admitted at the hospital with relapsed leukemia. It has been demonstrated that the results from treatment in children with cancer are much better when specialized care is given at a pediatric cancer center.32, 33 The importance of multidisciplinary treatment in improving patient outcome has been well documented for solid and hematologic malignancies, and this approach is available only at a pediatric cancer center.32–34 The roles of specialized, multidisciplinary teams and access to progressively more complex equipment and services are crucial for success in diagnosis, treatment, and long-term follow-up.

de Camargo also described a statistically significant improvement in the 5-year overall survival rate for childhood AML at the A. C. Camargo Hospital in São Paulo, Brazil, from 0% for children who were admitted between 1975 and 1979 to 31% for children who were admitted between 1995 and 1999.31 This improvement has been credited to an enhancement of support therapy as well as the introduction of more aggressive chemotherapy regimens, similar to what has been noted in other countries. In the U.S., increases in the 5-year survival rates for both childhood ALL and childhood AML were observed when comparing the 2 10-year periods from 1974 to 1983 and from 1984 to 1993.35

Our finding that trends in mortality were correlated to socioeconomic status corroborates the results from previous studies, which reported a statistically significant association between socioeconomic status and cancer mortality.10, 36–38 Over the past decade, interest has increased in social inequalities in health and life expectancy in industrialized countries.39 Socioeconomic status plays a key role in health, not only for those in poverty but at all levels. A socioeconomic status-health gradient has been reported.9

Several authors have described ethnic and racial differences in childhood ALL outcomes, and most studies have described poorer results for black children.40–43 In fact, these differences in survival emerge from distinct levels of socioeconomic status, but few studies have explored the role of nutrition14, 44, 45 and socioeconomic status as prognostic factors in childhood ALL.46, 47

The International Society of Pediatric Oncology and the Latin-American Society of Pediatric Oncology48 have produced statements concerning the value of cooperative programs between developed and developing countries as one of the strategies to reduce the gap between the therapeutic-diagnostic opportunities and the access to them. Since then, several successful partnerships between Latin America, Europe, and the U.S. have been established.49, 50 In Recife, which is located at the northeast region of Brazil, the 5-year event-free survival rate has increased from 32% to 63%, and the authors suggest that an effort of putting resources into pediatric cancer units and the establishment of twinning programs could overcome the impact of socioeconomic status.51, 52 Such strategies have a noticeable role in the progress of pediatric oncology in developing countries, but other initiatives still are required. It is improbable that a single method would be successful in making effective cancer treatment for children more widely available.53 Health in children results from many different factors, such as race, language, culture, geography, and socioeconomic status; and, unfortunately, these factors cannot be addressed fully only by changing access to health care.54

A recent study that included 127 children who were eligible for clinical trials indicated that parent's race and socioeconomic status were powerful predictors of clinician-parent communication during the informed consent process for pediatric leukemia trials. Minority parents and those from lower socioeconomic status received less information, had an incomplete understanding, and had a lower participation rate during the informed consent conference.55 It also is known that these factors may lead to nonadherence and treatment abandonment. Moreover, although it is presumed that adherence has an impact on treatment outcomes, the reverse causality may play a role (ie, a good outcome may encourage subsequent adherence and, under the same hypothesis, a bad initial result could discourage parents from continuing, particularly during a long period of treatment).56

A limitation of our study is pertinent to design. “Ecologic fallacy” cannot be ruled out, because individual assessment of socioeconomic status was not performed in our study, and the smaller unit analyzed (state) was too large to represent a neighborhood effect. A possible solution for this problem would be the use of smaller units (cities) to make the groups more homogeneous, but this approach may lead to other problems, such as greater migration between groups and less accurate estimation of disease rates.57 Which measurements are used in particular countries is dependent on the type of socioeconomic information commonly available in those countries.39 The main reason for using household or area-based measures of social class is that individual-level data are unavailable. Area-based measures of socioeconomic position, usually based on census data, have been used in several countries as a proxy for individual or household social class.39

Recent studies have indicated that approximately 30% of the poor population in Brazil is living in the metropolitan areas, which are distinguished by a strong social heterogeneity.20 In Brazil, there still is a broad manifestation of social exclusion. The old exclusion continues to be the signature of less developed regions, expressed by low schooling, poverty, and income inequalities. Conversely, the new exclusion rapidly is reaching the more developed areas through generalized unemployment, youth isolation, and deprivation in monoparental families.20 Despite the decrease in childhood leukemia mortality observed in Brazil in the last 20 years, we observed great heterogeneity among states that was related closely to socioeconomic development. This finding points to the need of interventions to improve the outcomes of children with leukemia who live in poor regions.

Acute leukemia is a progressive clonal disorder determined by mutations. The etiology of childhood leukemia remains undefined, but research studies have established a causal association with ionizing radiation, Down syndrome, and exposure to chemotherapeutic agents.58 Other factors, such as parental occupation, magnetic fields, pesticides, breastfeeding, dietary factors, parental smoking, chemical exposures, and genetic susceptibility, also have been described as risk factors or protective factors.58 To our knowledge, there is no nationwide study in Brazil regarding the prevalence of these environmental risk factors and their impact on childhood leukemia incidence and mortality.

The burden of childhood cancer and leukemia in developing countries will increase in the coming years,59 and such an increase will require a strong effort to eliminate disparities in survival and mortality caused by a complex interplay of poverty, culture, environment, and social injustice.60 The use of public health surveillance data can help in the identification of these inequalities and also in the implementation of a set of strategies to eliminate them, such as raising awareness, education and training, improved access to health care, changing of institutional practices, and other public policies.61 It is essential to eliminate the gap between discovery (diagnosis) and delivery (treatment) as well as between knowledge and action60 to warrant early diagnosis, treatment, and cure for all children with cancer.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

We thank Dr. Patricia Buffler, Dr. Carlos Rodriguez-Galindo, and Dr. Ricardo Brentani for their valuable suggestions and careful revisions of this article.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES