Neuropsychological Symptoms and Intrusive Thoughts Are Associated With Worse Trajectories of Chemotherapy-Induced Nausea

INTRODUCTION

While the occurrence and severity of chemotherapy-induced vomiting (CIV) has decreased with anti-emetic prophylaxis, unrelieved chemotherapy-induced nausea (CIN), that affects 30% to 60% of oncology patients, remains a significant clinical problem.^1,2 One of the major challenges with determining specific risk factors for the occurrence and/or severity of CIN is that CIN was not assessed as a unique symptom but rather as a composite symptom that included vomiting (i.e., chemotherapy-induced nausea and vomiting (CINV)).³

In the two studies that investigated risk factors for the occurrence² and severity⁴ of CIN, demographic and clinical characteristics were the foci of these investigations. Across these cross-sectional² and longitudinal⁴ studies, risk factors for CIN included: younger age,⁴ lower level of education,² having childcare responsibilities,^2,4 poorer functional status,^2,4 a higher number of co-morbidities,² increased emetogenicity of the chemotherapy (CTX) regimen,⁴ shorter CTX cycle length,² and an inadequate antiemetic regimen.² While most of these risk factors are not modifiable, they can be used to identify patients at increased risk for CIN. An identification of modifiable risk factors for CIN may allow for the development and testing of novel interventions to decrease this debilitating side effect of CTX.

Emerging evidence suggests that the microbiome-gut-brain-axis (MGBA) plays a role in the development of a number of “neuropsychological” symptoms associated with the administration of CTX^5–7 and reactions to stress.^8,9 In addition to its direct inflammatory effects on the gastrointestinal (GI) tract^10,11 and associated increase in GI symptoms,^4,12 CTX alters the functioning of the gut microbiome which regulates epithelial permeability and host immunity^5,7 and modulates a number of brain functions.¹³ These CTX-induced changes alter the bidirectional communications between the gut microbiome and the brain that are mediated by vagal activation and the serotonergic system. These bidirectional changes in communication within the MGBA are implicated in the development of chronic pain⁶ as well as in changes in mood and cognition.⁹ In addition, alterations in the gut microbiome affect the functioning of the hypothalamic-pituitary-adrenal (HPA) axis which can influence sleep quality, anxiety, depression, and reactions to stress.^9,14

While most of the research on symptoms associated with alterations in the MGBA is preclinical (for reviews see 5–7, 9, 13) or done in patients with other chronic medical conditions,¹⁵ evidence from our research group suggests that compared to patients who do not report the occurrence of CIN, patients with CIN have changes in gene expression associated with perturbations in pathways involved in the MGBA.¹² In addition, we found that in the same sample of patients, over two cycles of CTX, higher CIN severity scores were associated with higher levels of a number of GI symptoms (i.e., vomiting, feeling bloated, lack of appetite, difficulty swallowing, mouth sores, and constipation).⁴ In this later paper,⁴ we hypothesized that these additional GI symptoms were related to CTX-induced changes in the MGBA.

Given the limited evidence on modifiable risk factors associated with increased levels of CIN; the emerging evidence on associations between CTX-induced changes in MGBA and a number neuropsychological symptoms (i.e., pain, fatigue, sleep disturbance, cognitive changes, depression, anxiety)^5–7,13 and stress;⁹ and our own preliminary data;^2,4,12 in this study, we extend our previous findings to examine the relationships among a number of neuropsychological symptoms, as well as stress measures and CIN severity. Specifically, the purposes of this study, in a sample of outpatients with breast, gastrointestinal (GI), gynecological (GYN), or lung cancer who received two cycles of CTX (n=1251), were to evaluate for inter-individual differences in the severity of CIN and to determine which demographic, clinical, symptom, and stress characteristics are associated with higher initial levels as well as with the trajectories of CIN.

METHODS

RESULTS

DISCUSSION

As an extension of our previous study to identify modifiable risk factors for increases in CIN severity,⁴ this study is the first to identify associations among multiple neuropsychological symptoms, as well as cancer-specific intrusive thoughts, and worse CIN trajectories. In brief, higher levels of sleep disturbance, depression, and morning fatigue, lower levels of trait anxiety, as well as higher levels of intrusive thoughts were associated with higher initial levels of CIN. Morning energy was the only symptom that influenced the trajectory of CIN. In addition, lower KPS scores and shorter CTX cycle lengths were associated with higher initial levels of CIN and younger age and higher emetogenicity of the CTX regimen were associated with both higher initial levels as well as a worse trajectories of CIN severity.

In our exploratory analyses (Table 1), all of the neuropsychological symptoms that we assessed (i.e., evening fatigue, morning fatigue, evening energy, morning energy, depression, sleep disturbance, trait anxiety, state anxiety, attentional function, pain) were associated with either initial levels and/or the trajectories of CIN severity. However, only five of these symptoms (i.e., sleep disturbance, depression, trait anxiety, morning fatigue, morning energy) remained significant in the final model.

Consistent with previous reports, higher levels of sleep disturbance^2,36–38 and depression^2,38 were associated with increases in CIN severity. Associations among these three symptoms may be partially explained by shared underlying mechanisms. For example, all three symptoms can occur as a result of CTX-induced alterations in the gut microbiome,^12,39–41 as well as through increases in systemic levels of pro-inflammatory cytokines.^42,43 In addition, as was found in previous studies,^36,38,44 higher average fatigue scores were associated with increases in both CIN occurrence rates and severity scores. While in our previous cross-sectional study,² patients with CIN reported higher severity scores for both morning and evening fatigue, in the current study, only higher levels of morning fatigue were associated with increases in CIN severity. It is interesting to note that in the HLM model, the symptoms of sleep disturbance, depression, and morning fatigue made independent contributions to explain inter-individual variability in CIN severity. While we^45–48 and others^49,40 have evaluated the impact of the symptom cluster of fatigue, sleep disturbance, depression and/or pain on various QOL outcomes, the underlying hypothesis for this symptom cluster was that cancer and its treatments contributed to cytokine-induced sickness behavior.^51,52 The current findings, as well as our previous reports,^4,12 and the work of others^5–7 suggest that CTX-induced alterations in the MGBA may contribute to the development of these often co-occurring symptoms.

While the symptoms of fatigue and lack of energy have been used interchangeably, work from our group^53,54 and others^55–58 has demonstrated that they are distinct but related symptoms. Of note, in our previous cross-sectional study,² patients with CIN reported lower levels of both morning and evening energy. However, in the current study only lower morning energy scores were associated with a worst trajectory of CIN severity during the second cycle of CTX. This association may be linked to the relatively high levels of sleep disturbance reported by this sample.^59,60 This hypothesis warrants confirmation in future studies.

Previous research⁶¹ and our cross-sectional study² found that compared to patients without CIN, those with CIN reported higher trait anxiety scores. However, in the current study a lower level of trait anxiety was associated with an increase in CIN severity. One plausible explanation for this discrepancy is that anxious patients in this study may have been receiving anti-anxiety medications. Future studies need to evaluate for cross level interactions between co-occurring symptoms and pharmacologic and nonpharmacologic interventions and changes in the severity of CIN.

A growing body of evidence suggests that unrelieved stress is associated with disruptions in the MGBA (for reviews see 62–66). While in cancer patients, fear of disease recurrence and/or progression is associated with increased stress,⁶⁷ our previous² and current study were the first to demonstrate an association between a disease-specific measure of stress (i.e., intrusion subscale of the IES-R) and the occurrence and severity of CIN, respectively. In the multivariate analysis from our previous study, for each one-point increase in the intrusion subscale score, there was a 1.35 increased odds of reporting the occurrence of CIN. In the current study, higher intrusion scores were associated with a higher severity of CIN. This subscale of the IES-R assesses disturbing visuals and feelings associated with the patients’ cancer and its treatments.

In our previous longitudinal study, that evaluated associations between a number of GI symptoms and CIN severity,⁴ we provided a detailed discussion of the demographic (i.e., younger age, having child care responsibilities) and clinical (i.e., lower KPS score, higher SCQ score, higher emetogenicity of the CTX regimen) characteristics that were included in the final HLM model. In the current study, only three of these characteristics were retained in the final HLM (i.e., age, KPS score, emetogenicity of the CTX regimen) and a new clinical characteristic was identified (i.e., shorter cycle length). Of note and consistent with the current study, in our previous cross-sectional study,² compared to patients who received CTX on a 14-day cycle, patients on a 21-day cycle had a 42% decrease in odds of belonging to the CIN group (i.e., CIN occurrence). These findings support the idea that a shorter duration between CTX infusions is associated with increases in both the occurrence and severity of CIN.⁶⁸

Several limitations need to be considered. While this study included a large sample of patients who were assessed six times over two cycles of CTX and evaluated for multiple co-occurring symptoms in the same patients, future studies should enroll patients prior to the initiation of CTX and follow them through to the completion of CTX. In addition, future studies should code not only the emetogenicity of the CTX regimen and the types of anti-emetics presecribed but include an evaluation of whether or not the anti-emetic regimen reflected evidenced-based guideline recommendations. Equally important, the relationships between CIN and other GI symptoms are undoubtedly complex, additional research is warranted using analytic techniques like parallel process growth modeling to determine if a particular GI symptom is influencing the severity of one or more GI symptoms including CIN. Finally, future studies need to collect samples to evaluate for associations between/among changes in the severity of these co-occurring symptoms and changes in the oral and gut microbiota.

Despite these limitations, the findings from this study suggest that common symptoms associated with cancer and its treatment are associated with increased severity of CIN. This work, as well as our previous findings,^2,4,12 provide information on risk factors for CIN occurrence and severity that clinicians can assess and manage. In addition, this work provides direction for future studies that can directly examine the associations among multiple co-occurring symptoms and CTX-induced alterations in the MGBA.