Nutrition of Children With Cancer in Brazil: A Systematic Review

Abstract

PURPOSE

The objective of this systematic review was to describe nutrition-related publications on children and adolescents diagnosed with cancer in Brazil.

METHODS

The methodology followed that of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses. Medline, LILACS (the Latin American & Caribbean Health Sciences Literature), and Embase were searched in April 2019, and data extraction and rating of methodologic study quality (according to the National Institutes of Health quality score assessment) were performed independently by reviewers.

RESULTS

Twenty-seven studies met the inclusion criteria, reporting on 3,509 patients from 1994 to 2018. Most of the studies (74%) were of poor quality in methodology and reporting. Different cancer diagnoses were included in 52% of studies, whereas acute leukemia was the exclusive focus in 41%. The majority of the articles (70%) were from institutions in the Southeast Region of Brazil, mainly the state of São Paulo (74%); no publications were from the North Region of the country. Twelve studies addressed nutritional status and body composition, reporting an abundance of malnourished patients in the Brazilian population of children and adolescents with cancer. Six studies on micronutrients pointed to possible deficiencies in this population, with a yet unclear but promising role for supplementation during treatment.

CONCLUSION

Evidence indicates that there is great interest in the impact of nutrition on childhood cancer treatment and clinical outcomes in Brazil. However, there is a need to focus on high-quality research, particularly with multicentric/national studies. This will help establish research priorities and better planned clinical interventions, adapted to each region of the country.

INTRODUCTION

Of the estimated 400,000 annual incident cases of cancer in children 0 to 14 years of age worldwide,¹ more than 80% of those diagnosed live in low- and middle-income countries (LMICs).² In Brazil, an upper-middle-income country, there are approximately 12,500 new patients each year among 1- to 19-year-olds, and cancer currently represents the leading cause of death by disease in this age group.³ With an increasing pediatric population undergoing treatment for cancer, the need for timely supportive care, including nutritional interventions, is of considerable importance.

The impact of poor nutritional status, both undernutrition and overnutrition, on clinical outcomes (including survival) is well recognized. Such conditions can increase treatment-related morbidities and mortality, as well as health care costs and abandonment of therapy.^4-9 Moreover, studies in children with acute lymphoblastic leukemia (ALL) have demonstrated that the remediation of compromised nutritional status during treatment can improve survival, which supports the hypothesis that nutritional status is a modifiable risk factor for adverse outcomes.^10,11 However, there is a significant gap in the literature regarding high-quality research on the relationship between nutrition and cancer in children in LMICs, including in Brazil, which creates obstacles to determining the optimal standards of nutritional care for this patient population and remains a barrier to setting educational and research priorities, thereby preventing the field from advancing. The purpose of this review was to describe published nutrition-related data in children and adolescents diagnosed with cancer in Brazil.

METHODS

The systematic review methodology was guided by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA).¹² The electronic databases of Medline, LILACS (the Latin American & Caribbean Health Sciences Literature), and Embase were searched systematically from inception through April 2019. The inclusion criteria were (1) children ages 0 to 18 years with a cancer diagnosis, (2) nutrition-related investigations, and (3) studies conducted in Brazil. There was no language restriction. Publications that were not indexed, case reports or reviews were excluded. Search strategies are outlined in Appendix Table A1.

CONTEXT

• Key ObjectiveTo describe nutrition-related publications in children and adolescents diagnosed with cancer in Brazil.

• Knowledge Generated Seventy-four percent of included studies were of poor quality. The majority of articles were frominstitutions in the Southeast Region of Brazil, with no publications from the North Region of the country.

• Relevance There is a need to focus on high-quality research, particularly with multicentric studies representative of all regions of the country, in order to establish research priorities and better planned clinical interventions.

Titles and abstracts of references were reviewed by two authors (K.V. and R.D.B.) independently. Lack of consensus was resolved by consultation with a third author (E.J.L.). Potentially relevant publications were retrieved for full-text review.

Data were extracted by two authors (K.V. and L.A.) and verified independently by one author (E.J.L.). Attempts were made to contact authors for additional information when appropriate. Extracted data comprised state of origin, year, study design, sample size, diagnoses, treatment consortium and regimen, definition and categorization of anthropometric data, treatment-related toxicity, and survival data. Because of heterogeneity of the data and the small number of clinical trials evaluating a single diagnosis, statistical analysis was not feasible.

RESULTS

Figure 1 is a flow diagram of the search strategy results on the basis of the PRISMA template. After removal of duplicates, 794 articles met initial search criteria. A full-text review was conducted on 40 articles, 27 of which met eligibility criteria and were included in the analysis. Included studies reported on a total of 3,509 patients between 1994 and 2018. Study characteristics are detailed in Tables 1 and 2.

FIG 1.

Results of the search strategy.

TABLE 1.

Characteristics of Studies on Nutritional Status and Body Composition Included in the Systematic Review

TABLE 2.

Characteristics of Studies on Energy Needs and Clinical Intervention, Micronutrients, Survivors, and Risk Factors

Study Quality

National Institutes of Health quality score assessment¹³ for all included studies was calculated independently by two authors (K.V. and L.A.) and presented in Appendix Table A2. There were 15 initial disagreements, narrowed down to 12. After discussing with 2 more authors (E.J.L. and R.D.B.), all disagreements were resolved. Most studies (74%) were of poor quality (scoring < 8/14) in design and reporting, whereas the others were of fair and good quality.^8,13

Study Characteristics

The majority of studies (52%) included patients with different cancer diagnoses, followed by studies focused on patients with acute leukemia (41%). Five articles (19%) were published in Brazilian Portuguese, whereas the rest were in the English language. Most studies were cross-sectional (70%), mainly retrospective chart reviews, and there were three clinical trials. The majority of the articles (63%) were published in the last decade and represented institutions located in the Southeast Region of Brazil (70%), mainly the state of São Paulo (74%), whereas no publications were from the North Region.

Results by Primary Outcome

Nutritional status and body composition.

Thirteen studies addressed nutritional status and body composition (Table 1).^14-26 Overall, the prevalence of underweight patients at diagnosis varied from 6% to 25% and during treatment from 10.5% to 52.4%. In patients with ALL, the most studied diagnosis, the prevalence of underweight varied from 6% to 21.2% at diagnosis,^19,24-26 increasing to 52.6% in one study that assessed patients after at least one chemotherapy cycle.¹⁶ Stunting prevalence ranged from 8.9% to 42.9%,^{14-16,22,24,25} and one good-quality study reported growth deficit during and several years after treatment, with a negative impact on the final height, particularly in children younger than 4 years of age and in those who received cranial irradiation.²⁶ Overweight and obesity, however, had a reported prevalence of 4.1%-35% at diagnosis,^19,20,23,26 and 7.2%-57.9% during or after treatment.^14-17,23 In one study of patients with craniopharyngioma, 68.4% reported increased waist circumference, with a 44% prevalence of metabolic syndrome. This study also found that obesity at diagnosis was predictive of metabolic syndrome (OR, 6.1; P = .031).²³ Another study found overweight patients to have lower lean mass percentage on average (67.9%) than well-nourished (87.3%) or undernourished (84.1%) patients.¹⁷

Three studies compared different anthropometric measurements to assess nutritional status in patients with several cancer diagnoses. One found strong to moderate correlations between triceps skinfold thickness (TSFT) and body mass index (BMI; r = 0.8), as well as fat mass (r = 0.74) and fat-free mass (r = 0.76) measured by bioelectric impedance analysis.¹⁷ Another study concluded that TSFT and mid–upper arm circumference (MUAC) detected more patients with malnutrition than BMI or weight-for-height (W/H) z scores (P < .05).¹⁸ The third study found no differences in nutritional status according to BMI, MUAC, and TSFT (P > .05).²⁰

Regarding the impact of undernutrition on outcomes, Pedrosa et al²² found no relationship between nutritional status at diagnosis and survival (height-for-age [H/A] P = .229; weight-for-age [W/A] P = .996; W/H P = .887), whereas good-quality studies from Viana et al^24,25 found W/A at diagnosis to be unfavorable for prognosis (P = .0005). Moreover, stunting at diagnosis increased the risk of relapse 8.2-fold (95% CI, 3.1 to 21.9-fold) in one study²⁴ and 2.2-fold (95% CI, 1.2 to 4.1-fold; P = .015) in another.²⁵ Additionally, children with H/A z score above −1.28 at diagnosis had a 50% 8-year disease-free survival, compared with 0% for those with a z score lower then -1.28 (P = .000).²⁵ In contrast, Gelelete et al¹⁹ found 5-year event-free survival of patients who were overweight/obese at diagnosis to be lower (67% v 81%; P = .03) than in nonoverweight/nonobese patients, reporting an overweight/obesity hazard ratio of 1.92 (95% CI, 1.42 to 2.60, P = .031). Last, one study investigated different ALL immunologic subtypes and was unable to demonstrate an association with W/A or H/A z score at diagnosis.²¹

The interpretation of these findings is challenging because of the heterogeneity of methodologies, as well as the classification of nutritional status. However, it seems clear that malnourished patients, both undernourished and overnourished, are widely prevalent in the Brazilian population of children and adolescents with cancer.

Energy needs and clinical intervention.

Four studies addressed energy needs or tested clinical interventions (Table 2).^27-30 One article reported no evidence of increased energy expenditure or change in lipid and carbohydrate metabolism in children with cancer compared with a healthy control group.²⁷ Another study evaluated the use of parenteral nutrition (PN), reporting that most patients (82%) did not meet the recommended energy requirements despite receiving PN.²⁸ In contrast, in a good-quality study evaluating the use of enteral nutrition via feeding tube in patients undergoing hematopoietic stem cell transplantation (HSCT), 47.2% met an indication for this intervention, and no major complications were observed.²⁹ A fourth study evaluated the impact of oral supplementation on nutritional status and found improvement in TSFT from week 0 to week 12 of the intervention (P < .02).³⁰ The available studies are difficult to interpret because of the heterogeneity of objectives and designs, and reinforce the need to further understand nutritional therapy and its impact on Brazilian children with cancer.

Micronutrients.

Six studies explored micronutrient status among children with cancer (Table 2).^31-36 One reported that 70% of children were classified with vitamin C deficiency.³¹ A study of good quality examined vitamin D (25-hydroxyvitamin D) and found that 32% were deficient before HSCT and became more deficient after HSCT (P = .01).³² Another good-quality study supplemented participants with selenium (27 to 100 µg/day, according to age), which led to higher levels of immunoglobulin (Ig) A and IgG in patients with solid tumors (P = .0051 and P = .0055) and fewer neutropenia episodes during chemotherapy in patients with leukemia/lymphoma and solid tumors.³⁴ Three studies focused on zinc in patients with acute leukemia; the first reported lower serum levels than in controls (P < .05) and dietary intakes below the recommended values.³⁵ The second observed that serum zinc levels remained unaltered during treatment,³⁶ whereas the third was a good-quality study that observed no increase in plasma zinc concentrations (P = .217) with supplementation (2 mg/kg/day; maximum, 60 mg/day). Supplementation was associated with weight gain (P = .032) and reduction in the number of episodes of infection (P = .02).³³ Two studies reported on copper. One study found higher serum levels in patients with ALL compared with controls (P < .05) and intake below recommended values,³⁵ whereas another observed that copper levels decreased at the beginning of treatment.³⁶ Taken together, these studies seem to point to possible micronutrient deficiencies in children with cancer in Brazil, with an unclear but promising role for supplementation during treatment.

Survivors.

Four studies examined the impact of treatment on the nutrition of survivors (Table 2).^21,25,37-40 Two studies examined ALL survivors, the first reporting higher serum leptin levels (P < .05) and larger waist circumference (P < .05) in cases compared with controls, with no differences in hip circumference, waist-to-hip ratio, and BMI-for-age z score (P > .05).³⁸ The second study reported 23.4% prevalence of vitamin D insufficiency, 22.8% prevalence of overweight, and 15.8% prevalence of obesity. There was no correlation between nutritional status and exposure to radiation (P = .28). The authors also found that patients at risk for low bone mineral density had lower values of lean body mass measured by dual-energy x-ray absorptiometry (P = .03).³⁷ Another study analyzed medulloblastoma survivors and showed increased waist circumference (P = .024), waist-to-hip ratio (P = .001), and waist-to-height ratio (P = .018) compared with controls.³⁹ By comparison, a study that evaluated the dietary intake of survivors of several types of cancer reported that 25.5% were overweight/obese and had low consumption of fiber, polyunsaturated and monounsaturated fats, and some micronutrients, such as vitamin E, calcium, and selenium.⁴⁰ Overall, these studies indicate that survivors might be at a higher risk for metabolic disease because of elevated anthropometric indicators and that there is an evident need for nutritional education on healthy eating habits.

DISCUSSION

To the authors’ knowledge, this is the first systematic review of nutrition in children with cancer in an LMIC setting, applying rigorous criteria. A small number of Brazilian studies were identified, approximately 1.1 indexed publications per year, increasing to approximately 1.7 per year in the last decade. This is an insufficient amount in such an essential and growing field of research. Moreover, some of those publications were in the Brazilian Portuguese language, inaccessible to most of the global research community.

Most studies were of poor methodologic quality, based on retrospective chart reviews, thereby limiting our conclusions. Additionally, many studies included patients with different tumor types, with resulting difficulty in comparisons deriving from different disease burdens and treatments. Of the studies that focused on specific diagnoses, the majority were hematologic tumors, highlighting the need to also study patients with CNS and solid tumors.

A great discrepancy was noted in the number of publications among different states and regions of Brazil, with the majority of them from the Southeast Region, especially São Paulo. This reinforces the requirement for more representative data from the country, especially from regions such as the north, which has the highest prevalence of stunting in children younger than 9 years old, and the northeast, with the highest prevalence of underweight children and adolescents.⁴¹ Furthermore, although not surprisingly, these under-represented regions are the ones with the greatest prevalence of food insecurity in Brazil.⁴² The authors also observed the nonexistence of national or even multicentric studies on nutrition in pediatric oncology, studies that could bridge the gap in representation of the population of a vast and economically diverse country.

The variations in the prevalences found in nutritional status studies reflect a global lack of standardization on methods to clearly define undernutrition and overnutrition, diversity in the studied populations (different ages and diagnoses), and the country’s socioeconomic inequality. The reported prevalence of underweight patients at diagnosis was approximately 1.5 to 7 times higher than that of the Brazilian population data for children (4.1%) and adolescents (3.4%),⁴¹ and 4.5 to 19 times higher than data representative of South America (1.3%).⁴³ The prevalence of stunting was 1.2 to 6 times higher than the national and South American reported prevalences (7.1%).⁴³ Conversely, the prevalence of overweight or obese patients at diagnosis was compatible with Brazilian population data (overweight, 20.5%-33.5%; obese, 4.9%-14.3%); however, our study suggests a higher prevalence during and after treatment (1.7 times the national prevalence).^41,43 Endemic rates of stunting were found in either older publications when measured at diagnosis or in studies that assessed patients during treatment. In the first case, this might be a reflection of the progressive reduction of height deficit prevalence over the decades in Brazil, which has been associated with the increase of family income, partly attributed to social welfare programs, such as Bolsa Família.^41,44 In the latter, treatment may be accountable for such results, because stagnant growth during cancer treatment has been previously reported for children with ALL in both LMICs and high-income countries.^26,45

The prevalence of vitamin D deficiency reported by Campos et al³² is lower than previously stated in the literature for children with cancer,^46,47 but aligned with national data.⁴⁸ Patients were from different regions of Brazil, with varying incidences of solar radiation in latitudes between −30° to −5°, which might explain this observation. The remaining studies that evaluated micronutrient levels were limited because of their reliance on serum micronutrient values reflecting recent dietary intake rather than nutrient deficiencies. Therefore, the results should be interpreted with caution for the studies reporting on selenium, vitamin C, and zinc.

In comparison with published studies from the United States, the results of our review found the same or lower prevalences of overweight and obesity among survivors of childhood cancer.^37,38,49,50 Concurring with a meta-analysis by Zhang et al,⁴⁹ our review confirmed no correlation of overweight/obesity with previous exposure to cranial radiation. One study reported on the diet of childhood cancer survivors⁴⁰ and found low consumption of fiber, vitamin E, and calcium, as was found also by studies on ALL survivors,^51,52 studies during treatment of ALL (calcium, fiber, zinc, and vitamins D and E),⁵³ and studies in childhood cancer survivors in general (calcium).⁵² The results suggest similar suboptimal dietary intakes among Brazilian children with cancer compared with children in the United States; however, variations in nutrient databases limit direct study comparisons.

In this review we were unable to stratify nutritional outcomes according to the advance in access to medical cancer care in Brazil. The Brazilian Society of Pediatric Oncology was established in 1981, enabling treatment protocols for all cancer diagnoses and improving care, as well as access to care all over the country, with an increasing number of registered treatment centers, resulting in a positive impact on survival. Figure 2 reflects the improvement of outcomes in a single major Brazilian institution. However, many patients in Brazil are treated per protocol, not in a clinical trial, reflecting a deficit in trial enrollments.

FIG 2.

Long-term survival of children and adolescents with cancer.

Other limitations of this systematic review include a publication bias, because we only included indexed journals, and there may have been some Brazilian studies published in this field in nonindexed periodicals. Furthermore, the heterogeneity of themes led to the impossibility of statistical analysis of the data. Despite these limitations, the data suggest that there is great interest in the role of nutritional status and micronutrients in childhood cancer treatment and clinical outcomes in Brazil. Nutritional intervention studies and publications focused on survivors are scarce.

In conclusion, the authors identified a requirement in Brazil to focus on and invest in high-quality research in the field of nutrition in pediatric oncology. There is a particular need to engage in multicentric/national studies that will help establish research priorities and better planned clinical interventions, as well as educational strategies, tailored and adapted to each region of the country. In 2017, a partnership was established between the Brazilian Society of Pediatric Oncology and International Society of Paediatric Oncology through a nutrition task force, aiming to bridge that gap.⁵⁴ The support from such organizations is pivotal in improving nutrition research in pediatric oncology in Brazil, as well as implementing nutritional evaluations nationally at systematic time points. Some suggestions for specific research that could be undertaken are:

1. Prospective cohorts of patients with a single diagnosis or diagnoses that make clinical sense to group, especially solid tumors, to better understand how nutritional variables behave during treatment;

2. Prospective national diet quality and nutrient intake study, representative of different states and cultures;

3. Longitudinal multicentric trial on early nutrition therapy intervention in patients with selected diagnoses, such as medulloblastoma and osteosarcoma; and

4. Clinical trials focusing on specific nutrition interventions, such as protein intake, and their impact on body composition and outcomes.

Overall, it is critical to increase the representation from the under-represented regions in the literature. This may be the way to advance studies of nutrition in children with cancer, not only in Brazil but also in other LMICs.

APPENDIX

TABLE A1.

Search Strategies

TABLE A2.

Quality Criteria for Assessment

AUTHOR CONTRIBUTIONS

Conception and design: Karina Viani, Ronald D. Barr, Elena J. Ladas

Collection and assembly of data: All authors

Data analysis and interpretation: All authors

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO’s conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/go/site/misc/authors.html.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).

No potential conflicts of interest were reported.