Pain as an Adverse Event During Chemotherapy in Children with Acute Lymphoblastic Leukemia: A Cross-Sectional Nationwide Study

Abstract

Introduction

Pain is a common adverse event of chemotherapy in pediatric oncology patients. Understanding pain’s impact on these patients is essential for improving their quality of life and treatment outcomes. This short communication aims to describe the characteristics of pain in children with Acute Lymphoblastic Leukemia (ALL) and to explore its interactions with hospitalization duration and relapse, considering variables such as age, gender, and risk level.

Methods

This retrospective study analyzed data from 610 observations of pediatric patients with ALL treated under the Brazilian GBTLI-99 protocol. Pain was recorded according to CTCAE criteria. Statistical analyses, including Chi-square, Mann-Whitney, Poisson, and logistic regression models, were applied to explore associations between pain, hospitalization length, relapse, and other variables.

Results

Our results indicated that patients who presented with pain experienced longer hospitalization lengths. No significant difference in the presence of pain in female or male pediatric patients was observed.

Conclusion

This study is a first step towards treating pain as a limiting event that compromises the quality of life and response to treatment in Brazilian pediatric cancer patients.

Introduction

In Brazil, Acute Lymphoblastic Leukemia (ALL) accounts for 25% of pediatric cancer cases and 78.6% of pediatric leukemia cases. ¹ Only a few studies have explored the epidemiological characteristics of ALL in Brazilian children; what is known refers to a slight predominance in boys aged between 2 and 9 years.^1,2 The main treatment for ALL in childhood is chemotherapy. Brazilian referral cancer centers use well-established international protocols and a national protocol created by the Brazilian Childhood Cooperative Group for ALL Treatment (GBTLI). ALL subtypes differ according to biological, cellular, and molecular characteristics, response to therapy, and risk of relapse, all resulting in different outcomes. The stratification into a Low-Risk or High-Risk Relapse group determines treatment and reflects a range of clinical, biological, and genetic features, such as age and gender, white blood cell (WBC) count at diagnosis, and immunophenotypic, cytogenetic and molecular characteristics, as well as early medullar response to induction therapy. ³

Pain is considered one of the most frequent and uncomfortable Adverse Events (AEs) during pediatric cancer treatment, followed by nausea, vomiting, fatigue, decreased appetite, and sleep disturbances. ² It can compromise the quality of life of these patients and increase the length and the costs of hospitalization.^1,3

Unmanaged pain and its associated physical discomfort contribute to emotional distress, sleep disturbances, restricted social interactions, and behavioral challenges, which affect the psychosocial well-being of children and their families. ⁴ Prolonged hospitalization due to pain management further imposes a considerable economic burden on healthcare systems, underscoring the importance of early pain assessment and intervention.^5-7

Cancer-related pain in children is considered complex and challenging, partly due to patient diversities (gender, age, developmental characteristics), differences in treatment trajectories, the variety of oncologic diagnoses, and the complexity of its assessment and management in the pediatric population. The pain may result from the disease itself, be related to frequent intrusive procedures, appear as an adverse effect of the treatment, or as a consequence of generalized psychological and spiritual distress. ³ In this study, we considered treatment-related pain, which can be caused by mucositis, gastritis, and other gastrointestinal tract disorders, infections, and peripheral neuropathy, and is thus considered an AE related to treatment. ⁷

Pain experiences may vary with age, as older children and adolescents tend to report higher pain levels compared to younger children,^7,8 likely due to better self-reporting abilities. While some studies suggest females may report higher pain intensity, evidence in pediatric populations remains inconclusive.9.10 These developmental and gender-related variations highlight the need for tailored pain management strategies that address individual patient needs.

Despite the significance of addressing pediatric cancer pain, there is limited evidence specifically exploring how treatment-related pain interacts with other clinical outcomes, such as hospitalization length and relapse, particularly in pediatric populations undergoing chemotherapy for ALL. The following research question guided this study: How does treatment-related pain in children with Acute Lymphoblastic Leukemia (ALL) interact with hospitalization length and relapse, considering variables such as age, gender, and relapse risk?

Hypothesis

Building on existing research, the following hypotheses were formulated to guide the investigation: (1) children who experience treatment-related pain during chemotherapy have more extended hospitalization periods compared to those without pain; (2) the occurrence of treatment-related pain during chemotherapy is more likely among older children and those classified as high-risk, with no significant differences expected between genders; (3) children with documented treatment-related pain during chemotherapy show no significant difference in relapse rates compared to those without documented pain.

Purpose

This short communication describes the characteristics of pain in children with ALL and explores its interactions with hospitalization duration and relapse, considering variables such as age, gender, and risk level.

Methods

Design

This retrospective study analyzed data from the Brazilian Childhood Cooperative Group for ALL Treatment (GBTLI) ALL-99. Patients considered eligible for treatment with GBTLI-99 belonged to both Low-Risk Relapse and High-Risk Relapse Groups. The protocol was closed with a remission rate of 95.3%, overall survival of 88.1%, and event-free survival of 80.2%. ⁸

This manuscript was prepared in accordance with the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines for reporting observational studies. ⁹

Sample and Data Documentation

Data Source and Extraction

We extracted data referring to the presence of pain from the toxicities and adverse events databank of the general GBTLI-99 database. Pain presence was recorded by the attending physician during the routine evaluations of patients according to the criteria of the Common Terminology Criteria for Adverse Events (CTCAE) version 2.0. ¹⁰ Data from 610 observations were obtained; for the purposes of our analysis, the total observations were categorized into two groups of patients: children and adolescents that presented at least one episode of pain (n = 106), and a group of children and adolescents that did not present with pain during the clinical trial period.

Documentation of Pain Characteristics

Physicians documented treatment-related pain using the Common Terminology Criteria for Adverse Events (CTCAE) version 2.0. According to the instrument’s version used during the data report, pain was classified on a scale from 0, indicating the absence of pain, to 4, indicating incapacitating pain. ¹⁰ Additionally, physicians recorded the site of the pain and treatment phase.

Data Preparation

Only variables relevant to this study were selected from the dataset and transferred to a Microsoft Excel® worksheet. These data were subsequently analyzed using R software (version 3.5.0).

Ethical Considerations

The Institutional Review Board (IRB) of the Centro de Investigações Hematológicas Boldrini reviewed and approved the research protocol (Process Number 1.761.593/2016). Prior to enrollment, all participants and their legal guardians provided informed consent to participate in the original clinical trial, which generated the dataset used in this study. This consent included permission to use anonymized data in secondary research studies.

Data Analysis

We assessed the homogeneity of variables between the groups (with and without pain) using the Chi-square and Mann-Whitney tests. The analysis of variables influencing the length of hospitalization was conducted using adjusted Poisson regressions with a robust variance; this same analysis, considering the influence of relapse, was conducted using adjusted logistic regressions. The Almost-Likelihood method was used to estimate the model for the length of hospitalization, thus allowing for the estimation of robust variances in these phenomena. Logistic regression was used to model the chance of occurrence of events.

Results

This brief report presents the most relevant results of the hypothesized interactions.

Descriptive Characteristics of the Sample

This section presents the descriptive characteristics of the sample, including demographic and clinical data. A total of 610 children and adolescents from 5 to 18 years were enrolled in the clinical trial, and 106 had their pain documented. The majority of children in the study (55.9%) were male. The children’s mean age was 9.7 ± 3.6 years; the minimum and maximum ages were 5 and 18 years, respectively (Table 1).

Table 1.

Descriptive and Clinical Variables.

	Patients with reports of pain		Patients with no reports of pain		P-value
	N	%	N	%
Gender
Female	43	40.6	226	45.0	.485 a
Male	63	59.4	278	55.0
Relapse risk
Low risk	35	33.0	190	37.7	.426 a
High Risk	71	67.0	314	62.3
Age (years)
Mean (S.D)	10.5 (3.9)		9.5 (3.5)		.011 b

^aQui-Squared Test.

^bMann-Whitney Test.

According to the clinical reports that specified the site of pain (n = 66, 62%), the most prevalent locations were lower limbs (n = 43, 65.2%), bones, muscles, and articulations (n = 9, 13.6%) (Table 2).

Table 2.

Pain Reports, According to the Site.

Site of pain	N	%
Lower limbs	43	65.2
Upper limbs	4	6.1
Bones, Muscles and Articulations	9	13.6
Chest	5	7.6
Abdomen	3	4.5
Neck	1	1.5
Jaw	1	1.5
TOTAL	66	100.0

All the reports indicated the intensity of the reported pain. According to the CTCAE criteria, most patients (n = 95, 89.6%) were assessed as having Grade 1 pain. Only three reports (2.8%) indicated high-intensity pain (Grade 3), and no patients scored Grade 4 (Table 3).

Table 3.

Pain Reports, According to the Health Care Professionals’ Evaluation.

Pain score	N	%
Grade 1	95	89.6
Grade 2	8	7.6
Grade 3	3	2.8
Grade 4	-	-
TOTAL	106	100.0

The pain records vary according to the patient’s treatment phase. The majority of pain reports (n = 65, 32.8%) occurred in the maintenance phase, followed by late consolidation (n = 46, 23.2) and intensification (n = 35, 17.8%) (Table 4).

Table 4.

Distribution of Pain Reports According to the Chemotherapy Treatment Phase.

Treatment phase	N	%
Induction	32	16.2
Block A	7	3.5
Block B	2	1.0
Block C	7	3.5
Block D	4	2.0
Intensification	35	17.8
Late consolidation	46	23.2
Maintenance	65	32.8
Total	198	100.0

Analysis of Pain and Hospitalization Length (Hypothesis 1)

The Poisson regression model analyzed the influence of pain on hospitalization length, comparing patients with and without recorded pain while controlling for gender, age, and risk level. Length of hospitalization significantly differed between the two groups (P = 0.037), with patients with recorded pain experiencing days of hospitalization 24% higher [CI 95%: 1% to 51%], than the group that did not present with pain (Table 5).

Table 5.

Ratios of Mean Values and 95% Confidence Intervals (95% CI) for the Relationship Between Pain and Length of Hospital Stay, Controlled for Age, Sex, and Risk, Using the Poisson Regression Model.

Group/Variables	Model 1			Model 2
	Ratios of mean	CI 95%	P-value	Ratios of mean	CI 95%	P-value
Group
No pain	1.00	-	-	1.00	-	-
Pain	1.24	[1.01;1.51]	.037	1.47	[0.74; 2.90]	.272
Age (years)	1.00	[0.98; 1.03]	.747	1.00	[0.97; 1.03]	.844
Gender
Female	1.00	-	-	1.00	-	-
Male	0.83	[0.70; 0.97]	.020	0.86	[0.71; 1.03]	.094
Risk
Low	1.00	-	-	1.00	-	-
High	2.14	[1.68; 2.72]	<.001	2.25	[1.73; 2.92]	<.001
Interaction effect
Pain*Age	-	-	-	1.01	[0.94; 1.09]	.770
Pain*Gender	-	-	-	0.85	[0.57; 1.28]	.444
Pain*Risk	-	-	-	0.77	[0.40; 1.49]	.438

Demographic and Clinical Predictors of Pain (Hypothesis 2)

Children who experienced treatment-related pain during chemotherapy were significantly older than those who did not report pain (mean age 10.5 vs 9.5 years; P = .011). Among the 106 patients with documented pain, 63.2% had only one episode, while 36.8% experienced two or more episodes during the treatment period. No significant difference in the occurrence of pain was observed between female and male patients (P = .485) nor between patients categorized as low-risk vs high-risk for relapse (P = .426) (see Table 1).

Pain and Relapse (Hypothesis 3)

The Logistic Regression model verified the influence of both groups in the control variables (age, gender, and risk). The relationship between pain and relapse did not vary by the measured control variables. There was no significant difference between both groups regarding the chance of relapse (P = .943). However, we found a significant association between age and relapse (P = .005). Each increase of one year in age, decreased by 10% the expected chance of relapse (95% CI: a reduction from 3% to 16 %). No evidence of a significant association between gender and relapse (P = .159) was observed. However, as anticipated, the association between risk and relapse (P = .016) was confirmed (95% CI: 1.13 to 3.28). The results of the analysis of the reference model are presented in Table 6.

Table 6.

Odds Ratio and Confidence Intervals (IC 95%) for the Relationship Between Pain and Relapse, Controlled by Age, Gender, and Risk, According to the Logistic Regression Model.

Group/Variables	Model 1			Model 2
	Odds ratio	CI 95%	P-value	Odds ratio	CI 95%	P-value
Group
No pain	1.00	-	-	1.00	-	-
Pain	1.02	[0.59; 1.76]	0.943	2.04	[0.33; 12.53]	0.443
Age (years)	0.90	[0.84; 0.97]	0.005	0.90	[0.83; 0.98]	0.010
Gender
Female	1.00	-	-	1.00	-	-
Male	1.35	[0.89; 2.04]	0.159	1.53	[0.97; 2.41]	0.069
Risk
Low	1.00	-	-	1.00	-	-
High	1.92	[1.13; 3.28]	0.016	2.11	[1.19; 3.73]	0.011
Interaction effect
Pain*Age	-	-	-	1.01	[0.81; 1.25]	0.964
Pain*Gender	-	-	-	0.50	[0.16; 1.52]	0.222
Pain*Risk	-	-	-	0.61	[0.11; 3.23]	0.558

Discussion

In this study, patients who presented with pain were older than those who did not. Although children as young as 4 can identify and report pain, and those older than 5 can classify its intensity, studies often focus on older children and adolescents due to their ability to express themselves more effectively,^1,11-13 studies with older children or adolescents dominate the literature because these groups can express themselves with greater ease.^2,3 Whether from self-reporting or clinical observation, older pediatric oncology patients are more prone to indicate that they are experiencing pain. ³ Our results align with previous studies that emphasize older children’s ability to communicate pain, highlighting the need for tools to assess pain in younger children more effectively.^7,11,13

According to our results, patients who experienced pain had longer hospitalization durations. Prolonged hospital stays due to chemotherapy-related adverse events not only impact patients clinically and emotionally but also heighten family stress and place additional financial burdens on the healthcare system.^5,6 Studies suggest that early pain assessment using appropriate tools and effective pain management strategies can significantly improve supportive care for these patients.^7,13

High-risk patients experienced longer hospitalizations and higher pain scores, as seen in our study. This aligns with the understanding that these patients often undergo more intensive chemotherapy regimens and, as a consequence, present more treatment-related toxicities, including pain. ⁷ Our results pointed out a higher prevalence of pain during the maintenance phase. However, this is not supported by the literature; studies indicated that higher pain for children is associated with drugs such as vincristine and methotrexate, especially combined. However, it does not link the pain explicitly to any particular treatment phase​.^6,14

As in previous studies,^13,15 our results did not indicate a significant difference between the presence of pain in female or male patients. Although gender and pain differences are better established in adults, it has been documented that from the age of eight, females tend to report higher pain scores than males. These higher scores in females may be influenced by their better capacity to recall the pain experience and their biologically different pain experience, which can be more intense when compared to males.^13,16 This gender difference in pain reports may not have been noted in our study, considering that children and adolescents did not report pain using a validated self-assessment tool. Future studies could explore these gender differences using validated tools and patient-reported outcomes, which increasingly give children and adolescents a voice in reporting their symptoms, potentially offering valuable new insights.

This study has several limitations that must be acknowledged. While the sample size was adequate for a rare disease, the retrospective design relied on clinical judgment, which may have introduced variability in the standardization of pain scores, especially given the developmental differences among patients. Additionally, the lack of pain documentation for many patients limited the distinction between the true absence of pain and the absence of its recording by professionals. This missing data introduces the potential for systematic bias, particularly underestimation of pain prevalence and distortion of its associations with clinical outcomes.

Future prospective studies using standardized tools and patient-reported outcomes are recommended to improve pain assessment, reduce discrepancies, and deepen our understanding of the interactions between these variables, pain, and clinical outcomes.

Conclusion and Implications for Practice and Research

The setting of this pioneering investigation into the presence of pain episodes as an adverse event (AE) during the treatment of ALL in Brazilian children and adolescents is to strengthen the case that pain deserves very accurate reporting in clinical trials so that it can be recognized as a critical AE. The information and gaps in this brief report reveal a great need for constant evaluation and recording of pain in clinical trials as an important aspect that affects patient outcomes.

Our findings demonstrate that, despite the limitations imposed by the retrospective design, accurate pain assessment remains clinically and research paramount. This study marks an important beginning towards tackling pain as an AE of treatment and a limiting factor to the quality of life of pediatric cancer patients. Moving forward, more prospective studies reporting patient-reported outcomes (PRO) measures and other standardized tools will contribute to further understanding and, hence, the management of chemotherapy-induced pain.

ORCID iDs

Fernanda Machado Silva-Rodrigues https://orcid.org/0000-0002-8412-2333

Carlos Alberto Scrideli https://orcid.org/0000-0001-6618-789X

Lucila Castanheira Nascimento https://orcid.org/0000-0002-7900-7111

Ethical Statement

Ethical Approval

The Institutional Review Board (IRB) of the Centro de Investigações Hematológicas Boldrini reviewed and approved the research protocol (Process Number 1.761.593/2016). Prior to enrollment, all participants and their legal guardians provided informed consent to participate in the original clinical trial, which generated the dataset used in this study. This consent included permission to use anonymized data in secondary research studies.