Investigating Fatigue in Adults with Neuroendocrine Neoplasms Using a Biopsychosocial Model

Abstract

Purpose

Neuroendocrine neoplasms (NEN) are rare but increasingly recognized malignancies. Fatigue is common amongst individuals with NEN, but factors associated with fatigue in this population are not well described. Understanding these factors is essential for optimizing supportive care and quality of life. This study aimed to investigate factors associated with fatigue in patients with NEN using a biopsychosocial model.

Methods

Participants enrolled on a rare tumor natural history study (NCT03739827) were included in this analysis if they had NEN, were ≥18 years old, and completed at least one evaluation of patient-reported outcome measures. Fatigue was assessed using the Patient-Reported Outcome Measurement Information System (PROMIS).

Results

Sixty-four adult participants with NEN (mean age=59 years; SD=12.0) completed PROMIS measures. Over half (53%) had PROMIS Fatigue T-scores >55 indicating fatigue, including 31% in the moderate to severe range. Fatigue was highly correlated with pain, functional status, and psychosocial factors including depression, anxiety, and emotional support. There were no significant differences in fatigue based on identified sociodemographic or disease-related factors, although a significant association between fatigue and serotonin, a neuroactive peptide secreted by some NENs, was found in a subset of patients.

Conclusion

Fatigue is common in patients with NEN, with one-third having moderate to severe fatigue impacting daily functioning. Pain, anxiety, and depression were significantly associated with fatigue, suggesting that interventions targeting these symptoms also may improve fatigue in these patients. Further work is needed to investigate additional NEN-specific factors to optimize management of fatigue in individuals living with this long-term disease.

Introduction

Neuroendocrine neoplasms (NENs) are rare tumors most commonly originating in the gastrointestinal tract, pancreas, or bronchopulmonary system, and have several unique features which differentiate them from other tumor types [1]. A hallmark of NENs is their ability to secrete various peptides, some of which, such as serotonin, are hormonally active and cause a constellation of systemic symptoms including flushing, diarrhea, and respiratory distress, commonly known as carcinoid syndrome (CS)[2]. In contrast, non-neuroactive tumor markers of NENs, including chromogranin A, can be useful in diagnosis but do not directly cause symptoms [2]. Management of NENs often involves disease-specific treatments, such as somatostatin analogs and peptide receptor radionuclide therapy, instead of or in addition to more traditional surgery, chemotherapy, and radiation [1]. While prognosis depends on tumor stage and grade, most individuals with NEN live with the disease for many years, with a median overall survival of around 9 years [3]. Most individuals with NEN report symptoms related to their disease, which can vary based on tumor size and location, type of hormone secreted, and disease status. Across all types of NEN, fatigue is one of the most frequent symptoms, reported by 36% to 61% of patients [4–7].

Fatigue is a common and burdensome complaint amongst cancer patients and survivors. The National Comprehensive Cancer Network defines cancer-related fatigue (CRF) as “a distressing, persistent, subjective sense of physical, emotional, and/or cognitive tiredness or exhaustion related to cancer or cancer treatment that is not proportional to recent activity and interferes with usual functioning” [8]. Fatigue is prevalent across cancer types, and is often reported as having an even greater negative impact on daily activities and quality of life than other common cancer-related symptoms such as pain and nausea [9]. The etiology of CRF is thought to be multifactorial, with clinical, sociodemographic, physiologic, and psychologic factors all likely playing a role [10]. Fatigue is highly correlated with other impairments in health-related quality of life (HRQOL), and many studies have demonstrated a correlation between fatigue and depression in various types of cancer [11–14]. Prior work also shows that CRF can be related to level of physical symptom distress, as well as sociodemographic factors such as sex, level of education, and household size [15, 16]. Results are somewhat less consistent for disease- and treatment-related factors, with some studies finding correlations between level of fatigue and receipt of therapies such as radiation or chemotherapy [16], with others finding no such correlation with disease stage or treatments [11, 17].

Despite the high prevalence of fatigue in patients with NEN, there has been limited investigation into factors associated with fatigue in this population. Given the unique characteristics of NEN, including their indolent nature, variety of treatment modalities, and potential for hormone secretion, it is reasonable to suspect that fatigue in these patients could be different than in those with other malignancies. Most prior studies of fatigue in patients with NEN have assessed fatigue as one domain of HRQOL. Several studies have found that patients with high symptom burden, in particular frequent diarrhea and flushing in the setting of CS, have more severe fatigue in addition to other markers of impaired HRQOL [5, 18, 19]. In contrast, a recent study by Hummelshoj et al specifically assessed fatigue and did not find an association with CS [7], suggesting that further investigation into disease-related predictors of fatigue in patients with NEN is needed. In addition, less work has been done in this population to investigate sociodemographic and psychosocial factors that are correlated with fatigue in other cancers. To date, no studies have used a conceptual framework to investigate factors related to fatigue in individuals with NEN. The biopsychosocial model is a framework for understanding health that takes biological, psychological, and sociodemographic factors into account [20, 21]. Using this approach, we examined sociodemographic, biologic/disease-related, and psychosocial factors that may be associated with fatigue in patients with NEN. Better understanding of these associations can aid in identifying patients with NEN who are likely to experience significant fatigue as well as inform interventions to improve fatigue in this population.

The current study aimed to investigate fatigue in a cohort of patients with NEN enrolled on a rare solid tumor natural history study. Specifically, we aimed to 1) characterize the prevalence and severity of fatigue in this population, and 2) investigate factors associated with fatigue using a biopsychosocial approach (Figure 1). Under these aims, we hypothesized that 1) there would be a higher burden of fatigue in this cohort than in the general population, and 2) that sociodemographic factors including sex and marital status, disease-related factors including metastatic disease and time since diagnosis, and psychological factors including anxiety and depression would all be associated with fatigue.

Fig. 1.

A biopsychosocial model of fatigue investigated in individuals with neuroendocrine neoplasms

Study-specific variables were categorized into one of three domains according to the biopsychosocial model. The arrows highlight the complex interactions between the different domains and between the domains and fatigue. Abbreviations: NEN, neuroendocrine neoplasm. ^a Enrollment group refers to participants who participated entirely remotely (field group) versus participants enrolled and seen at the NIH Clinical Center (clinic group); ^b Functional status includes patient-reported cognitive function, mobility, and upper extremity function.

Methods

Participants

Individuals enrolled on the Natural History and Biospecimen Acquisition Study for Children and Adults with Rare Solid Tumors (NHRST; NCT03739827, [22]), a protocol developed as part of the NCI My Pediatric and Adult Rare Tumor (MyPART) network to study the biology and clinical course of pediatric and adult rare solid tumors, with the goal of translating these findings to improve care and treatment. Individuals ≥4 weeks old with a rare solid tumor, defined as less than 15 cases per 100,000 people yearly, were eligible to participate. Participants could enroll remotely (field group), with an option for patients to visit the National Institutes of Health (NIH) Clinical Center (NIHCC) for in-person evaluations (clinic group). Participants were recruited using a variety of mechanisms including self and physician referrals and engagement with advocacy groups. The NHRST was approved by the NIH IRB and participants completed an informed consent process.

Participants who enrolled on the NHRST protocol from January 2019 to February 2024 were included in this analysis if they had a diagnosis of NEN, including patients with both neuroendocrine tumor (NET) and neuroendocrine carcinoma (NEC), and had completed Patient Reported Outcome Measurement Information System (PROMIS) measures at least once to assess fatigue. Pediatric patients (<18yo at time of enrollment) were excluded from statistical analysis due to small numbers.

Study Design & Data Collection

Comprehensive data collection for NHRST included medical history from patient questionnaires and medical records, imaging, pathology, and patient-reported outcome (PRO) measures at enrollment and prospectively. For the clinic group, additional examinations included clinical laboratory studies and imaging that were conducted at the NIHCC. Participants were contacted annually to complete PRO measures and obtain updated medical and family history information.

PRO Measures

Questionnaires were administered electronically to participants via email or on a clinic-provided tablet. PRO measures were built in Scribe, an electronic survey administration platform, to allow encrypted administration of questionnaires via email to be completed remotely with direct data deposit into the Labmatrix^® database.

PROMIS.

The primary outcome of fatigue as well as other psychosocial outcomes were assessed using PROMIS, a set of validated measures that evaluate functioning, well-being, and HRQOL [23]. Participants completed adult self-report PROMIS measures (ages ≥18 years) assessing domains of Fatigue as well as Anxiety, Depression, Mobility, Upper Extremity, Pain Interference, Cognitive Function, and Emotional Support. All measures, with the exception of the Adult Depression-Short Form, were administered as a computerized adaptive test, where questions are pulled from an item bank for each domain and refined based on the individual’s response [24]. Each PROMIS measure provides a T-score with a mean of 50 and a standard deviation of 10, with higher scores indicating more of the outcome being measured. T-scores are generated based on normative data from the general U.S. population, with a PROMIS Fatigue T-score of <55 being within normal limits, 55–60 indicating mild fatigue, 60–70 moderate fatigue, and >70 severe fatigue. PROMIS measures have been validated in patients with cancer in addition to the general population [23, 25], and have shown good reliability (Cronbach’s alpha 0.86 to 0.96) in other cancer populations [26]. A difference in PROMIS T-score of 3 points is generally considered the minimally clinically important difference [27].

Numeric Rating Scale for Pain (NRS-11).

Pain intensity was assessed using the self-report NRS-11, an 11-point numeric scale that rates severity of pain from 0 (no pain) to 10 (worst pain possible) in the last week [28]. The NRS-11 is validated to assess pain intensity in ages 8 years and older [29, 30].

Pain Interference Index (PII).

This self-report measure was used to assess the impact of pain on daily functioning. Participants completed a 12-item PII, with each item rated on a 7-point Likert scale from 0 (Not at all) to 6 (Completely), and averaged to obtain the total score, with higher total scores indicating increased pain interference across activities [22]. A 6-item PII has been validated in pediatric and young adult patients [31, 32].

Laboratory Values

Relevant biochemical markers were collected from a subset of participants in the clinic group and included for analysis if they were collected within six months of baseline PROMIS measure collection. Biochemical markers included hemoglobin (normal range 11.2–15.7 g/dL) a marker of anemia that can cause fatigue, serum chromogranin A (normal range <93 ng/mL), a non-neuroactive tumor marker, and serum serotonin (normal range ≤330 pg/mL), a neuroactive peptide that causes CS.

Statistical Analysis

Descriptive statistics were used to define the characteristics of the sample at baseline (initial assessment). PROMIS Fatigue T-score was used as the primary outcome to assess fatigue, and descriptive statistics were used to determine prevalence and severity of fatigue in the sample. Comparison of fatigue in this NEN sample compared to the general population was conducted using a one-sample t-test. The first Fatigue T-score collected after study enrollment was used as the baseline value to characterize fatigue and to investigate factors associated with fatigue as described below. For participants who completed PROMIS measures at one year follow-up, change in fatigue over time was analyzed using a paired t-test.

To investigate the association between fatigue and biopsychosocial factors, between-group comparisons were conducted using Mann-Whitney U-test or Kruskal-Wallis test to evaluate for differences in Fatigue T-scores based on disease-related and sociodemographic factors. The relationship between Fatigue T-scores and psychosocial variables and laboratory values with continuous outcome measures was investigated using Spearman correlation coefficients (r_s). A p-value of <0.05 was used for all tests of significance. Given the exploratory and hypothesis-generating nature of this study, adjustments were not made for multiple comparisons. Analysis was conducted using IBM SPSS Statistics (Version 30.0).

Results

Participants

Ninety-nine patients with NEN were enrolled on the NHRST study from January 2019 to February 2024, of whom 5 were pediatric and excluded due to age. Sixty four out of 94 adults (67.7%) completed PROMIS measures at least once and were included in this analysis. There was no significant difference between adult NEN participants who completed PROMIS measures and those who did not with respect to sex, race, age at baseline, enrollment group (field versus clinic), primary tumor location, or presence of metastases.

At baseline, the total sample (n=64) had a mean age of 59.0 years (SD 12.0; range 27–79). The majority had gastroenteric or pancreatic NEN and had metastatic disease. Three participants had NECs. Median time from diagnosis was 3.2 years (interquartile range [IQR] 6.8). Participants were mostly white (n=54, 84%) and non-Hispanic (n=59, 92%). The cohort was highly educated, with a mean of 17.1 years of education (standard deviation [SD] 2.9).

Fatigue and Relation to Biopsychosocial Variables

Fatigue was prominent, with 53.1% (n=34) having a Fatigue T-score >55 at baseline, indicating greater than normal fatigue. Of these, 21.9% had mild fatigue, 25% had moderate fatigue, and 6.3% had severe fatigue. The mean baseline Fatigue T-score was 54.0 (SD 10.6), which is significantly higher than the general population mean of 50 (t=3.025, df=63, p=0.004, 95% confidence interval [CI]: 51.4 to 56.6).

With respect to sociodemographic factors, Fatigue T-scores were not associated with age at baseline (r_S=−0.100, p=0.430), and there were no significant differences in fatigue according to sex, enrollment group, or marital status, where unmarried participants include those who are single, divorced, or widowed (Table 1). There were not sufficient non-white participants to examine level of fatigue based on race.

Table 1.

Baseline PROMIS Fatigue T-scores by sociodemographic and disease-related categorical variables

	n (%)	PFTS (median, IQR)	Test statistic^	P-value
Sex			466.0	0.767
Male	25 (39.1%)	53.5 (16.5)
Female	39 (60.9%)	55.5 (15.7)
Cohort			491.5	0.803
Clinic	34 (53.1%)	56.3 (12.8)
Field	30 (46.9%)	52.1 (17.6)
Marital status			215.5	0.167
Unmarried	14 (25.5%)	57.1 (20.2)
Married	41 (74.5%)	50.7 (16.8)
Metastatic disease			367.0	0.475
No	13 (20.6%)	48.5 (22.2)
Yes	50 (79.4%)	55.5 (12.5)
Prior chemotherapy or radiation			468.0	0.961
No	41 (64.1%)	55.2 (16.6)
Yes	23 (35.9%)	55.5 (11.7)
Primary tumor location#			6.745	0.080
Gastroenteric	27 (44.3%)	55.4 (13.1)
Pancreatic	19 (31.1%)	48.5 (13.5)
Bronchopulmonary	8 (13.1%)	68.1 (21.3)
Other	7 (11.5%)	55.4 (18.1)

PFTS, PROMIS Fatigue T-Score; IQR, Interquartile range;

^{^}Mann-Whitney U-test used for two-way comparisons, Kruskal-Wallis test used for comparison of more than two groups (primary tumor location);

^#3 participants with unknown primary tumor location not included in comparison

With respect to disease-related factors, there was no significant association between Fatigue T-scores and time since diagnosis (r_s=0.001, p=0.994) or age at diagnosis (r_s=−0.118, p=0.355). In addition, there was no difference in Fatigue T-scores based on presence of metastatic disease or prior chemotherapy or radiation treatment (Table 1). Regarding tumor location, there was a trend (p=0.080) toward higher Fatigue T-scores in patients with bronchopulmonary NENs (median 68.1, IQR 21.3) compared to other locations (gastroenteric [55.4, 13.1], pancreatic [48.5, 13.5], and other [55.4, 18.1]).

For the subset of patients with biomarker data available, there was no correlation between hemoglobin values (n=21; median=13.0g/dL, IQR=1.4, range=10.5–16.5) and Fatigue T-scores (r_s=0.112, p=0.629). Chromogranin A levels (n=12 with one extreme outlier excluded; median=101.5ng/mL, IQR=105, range=25–442) and Fatigue T-scores also were not correlated (r_s=−0.056, p=0.863). However, serotonin values (n=14, median=729.5pg/mL, IQR=939, range=76–1960) were significantly correlated with Fatigue T-scores (r_s=0.638, p=0.014).

Regarding pain, NRS-11 worst pain scores in the last week ranged from 0 to 9 with a median of 2 (IQR 5) (Table 2). Pain scores were significantly correlated with Fatigue T-score (r_s=0.555, p<0.001). Similarly, pain interference as measured by both PROMIS Pain Interference T-scores (r_s=0.710, p<0.001) and PII scores (r_s=0.667, p<0.001) were highly correlated with Fatigue T-scores.

Table 2.

Correlation between baseline PROMIS Fatigue T-score and other PRO measures

Measure	Median score (IQR)	Correlation with PFTS, rSa	95% CI, rS
Pain in last week, NRS	2.0 (5.0)	0.56	0.34 – 0.71
Pain Interference, PII (n=42)	0.8 (2.0)	0.67	0.43 – 0.82
Pain Interference, PROMIS	52.7 (17.8)	0.71	0.54 – 0.82
Anxiety, PROMIS	52.2 (8.9)	0.46	0.23 – 0.64
Depression, PROMIS	45.5 (16.5)	0.57	0.36 – 0.72
Emotional support, PROMIS (n=42)	50.4 (9.6)	−0.41	−0.11 – −0.64
Cognitive function, PROMIS	49.7 (15.8)	−0.57	−0.36 – −0.72
Mobility, PROMIS	47.0 (11.7)	−0.56	−0.34 – −0.72
Upper extremity, PROMIS	47.6 (18.9)	−0.55	−0.34 – −0.71

PROMIS, Patient-Reported Outcomes Measurement Information System; PRO, patient-reported outcome; IQR, interquartile range; PFTS, PROMIS Fatigue T-score; r_s, Spearman correlation coefficient; CI, confidence interval; PII, Pain interference index;

^aall correlation coefficients significant at p<0.01 level

With respect to psychosocial factors, more severe anxiety and depression were significantly correlated with more severe fatigue (Table 2). Emotional support was negatively correlated with Fatigue T-scores, with less fatigue in patients reporting more emotional support. Functional status, as measured by cognitive function, mobility, and upper extremity function, also was negatively correlated with Fatigue T-scores, with more severe fatigue in participants who reported less functionality in day-to-day activities.

Longitudinal Examination of Fatigue

PROMIS measures were completed by 32 of the 64 participants at the one-year follow up evaluation. Longitudinal data was incomplete due mostly to lack of participant completion of annual PROMIS measures, as well as death before the one-year mark in two participants. There was no significant difference between the baseline (mean 54.9, SD=8.7) and one-year follow up (mean=55.5, SD=7.6) Fatigue T-scores for these participants (p=0.639). Associations between fatigue and other psychosocial factors, such as anxiety (r_s=0.53, p<0.01) and depression (r_s=0.58, p<0.01), as well as mobility (r_s=−0.59, p<0.01), pain (r_s=0.36, p=0.04), and pain interference (r_s=0.54, p<0.01) remained significant at one year follow up.

Discussion

The results from this study demonstrate significant fatigue in adults with NEN, with over half reporting higher than normal levels of fatigue and significantly higher average Fatigue T-scores than the general population, as hypothesized. Nearly a third of participants (31.3%) had moderate to severe fatigue, a level which is generally considered to be clinically significant and impact HRQOL and clinical outcomes [33]. These findings, along with those from other studies [6, 7, 18], underscore that fatigue is a significant concern in the NEN population and warrants close attention, investigation, and ideally treatment and prevention.

Longitudinal analysis was limited, but fatigue appeared relatively stable over time, given the lack of change of Fatigue T-scores between baseline and one year follow up in the 32 patients who completed repeat measures, as well as the lack of association between Fatigue T-scores and time since diagnosis. This finding is particularly notable given the indolent nature of NENs, indicating that a subset of people with NEN live with fatigue for a significant length of time while under treatment or surveillance, suggesting both a high cumulative burden of fatigue and conversely a wide window of opportunity for intervention.

The second hypothesis regarding the association of biopsychosocial factors with fatigue was partially supported. With respect to sociodemographic factors, there did not appear to be significant differences in fatigue between males and females in this cohort, which contrasts with other populations where females are more fatigued [34]. Although there was no significant difference in fatigue by marital status, less fatigue was reported in patients with more emotional support.

With respect to disease-related factors, metastatic disease, which was hypothesized to be associated with more fatigue, did not show a statistically significant difference, nor did prior receipt of chemotherapy or radiation. These results contrast with findings in other studies, where more advanced disease and more intensive therapy are associated with higher prevalence and severity of fatigue [35, 36]. However, the current study results should be interpreted with caution, as the small sample size and small number of participants without metastatic disease may have limited the ability to detect a statistical difference. Similarly, the trend toward higher fatigue in patients with bronchopulmonary NEN compared to other locations also may be due to small sample size and outsized influence of a few outliers. It does highlight, however, that more investigation is warranted into NEN-specific disease-related factors such as hormone-secreting status, symptoms related to hormone secretion (flushing, diarrhea, etc.) and somatostatin analog use. A subset of the cohort in this study had laboratory values available for serotonin, the bioactive peptide that causes CS, and these levels were highly correlated with Fatigue T-scores. In contrast, chromogranin A levels, a non-bioactive marker of tumor burden, were not associated with fatigue. These findings suggest the possibility that symptomatology, perhaps more than tumor burden, is a significant driver of fatigue in patients with NEN. Further investigation is needed in this area, but the current study suggests that fatigue could be a valuable clinical endpoint in studies of therapy for NEN, particularly those targeted at symptomatic improvement.

Regarding psychosocial and HRQOL-related factors, all were highly associated with fatigue. Like in other populations [12–14], depression was highly correlated with fatigue, as was anxiety. Both pain intensity and pain interference had particularly high correlation coefficients with fatigue, suggesting a strong association between pain and fatigue, a relationship that is well-described both in other forms of cancer and in other chronic disease processes [12, 14, 37, 38]. Fatigue also was highly associated with both physical and mental functional status, demonstrating the mind-body connection and the impact fatigue may have on patients’ everyday lives.

While the observational nature of this study does not allow for conclusions regarding directionality of the relationship between fatigue and other factors, potentially modifiable factors strongly associated with fatigue include pain, anxiety, and depression. Optimizing pain management and mental health treatment may help alleviate the fatigue burden in these patients who may live with NEN-associated fatigue for many years. Per the biopsychosocial model, these relationships are likely bidirectional, suggesting that interventions targeting fatigue also may benefit these other factors as well. Substantial evidence in other cancer types suggest that exercise is associated with decreased CRF [39, 40] and can reduce levels of anxiety and depression [41, 42]. Other mind-body interventions, including individualized counseling, support groups, and mindfulness-based practices [43–45] have been beneficial in managing fatigue as well as depression and anxiety in some cancer populations. Further investigation is warranted to investigate the utility of such interventions in the NEN population, who may live for many years with this disease.

There are several limitations of this study. The sample size was small, which limited statistical power as well as the ability to use more complex statistical techniques to control for multiple factors or perform subgroup analyses. The cohort had significant heterogeneity, and future investigations would benefit from separate analyses by tumor location, presence of metastases, and degree of differentiation (NET vs NEC). Longitudinal data was only available for half of participants, reducing sample size and ability to investigate trajectories of fatigue over time. Institutional factors may limit generalizability, as individuals enrolled on a study at the NIH may have more advanced, complex, or atypical disease presentations. Furthermore, the sample was predominately white and highly educated, further limiting generalizability. The inclusion of the field group was an attempt to alleviate these biases, and therefore, multi-center collaborative studies and additional outreach recruitment strategies should be utilized in future research. Finally, data was collected as part of a larger natural history study, and as such, many NEN-specific variables were not collected. Future studies should look to investigate these disease-specific factors. Given these limitations, results of this study should primarily serve as hypothesis-generating for future investigations of fatigue in the NEN population.

In conclusion, this study suggests a substantial burden of fatigue in patients with NEN, a disease that patients typically live with for many years. Their fatigue appears highly associated with pain, functional status, and psychosocial factors, and may be associated with hormone secretion and other symptomatology. This work demonstrates the utility of studying CRF using the biopsychosocial model, which can be applied to studies of CRF in this and other cancers. Further longitudinal research with a larger, more diverse sample is needed to investigate the symptoms and treatments that are specific to NEN to better understand the disease-related factors affecting fatigue, as well as trajectories of fatigue over time. A natural history study of patients with NEN is currently underway and the results will help provide more disease-specific insights on fatigue and other related symptoms. Additional investigation will lead to better understanding of the biopsychosocial factors influencing fatigue in this population, which will help identify patients at risk for clinically significant fatigue, inform the clinical use of tools to monitor fatigue, guide supportive care interventions, and inform the use of fatigue as a valuable clinical endpoint in future work.