Higher Lifetime Stress and Symptom Burden Contribute to the Occurrence of Shortness of Breath

Patients and Settings

This study is part of a larger, longitudinal study of the symptom experience of oncology outpatients receiving chemotherapy.⁴⁹ Eligible patients were ≥18 years of age; had a diagnosis of breast, gastrointestinal, gynecological, or lung cancer; had received chemotherapy within the preceding 4 weeks; were scheduled to receive at least two additional cycles of chemotherapy; were able to read, write, and understand English; and gave written informed consent. Patients were recruited from two Comprehensive Cancer Centers, one Veterans Affairs hospital, and four community-based oncology programs during their first or second cycle of chemotherapy. The major reason for refusal was being overwhelmed with their cancer treatment.

Study Procedures

Study was approved by the Committee on Human Research at the University of California, San Francisco and by the institutional review board at each of the study sites. Written informed consent was obtained from all patients. Patients completed the shortness of breath measure, a total of six times over two cycles of chemotherapy. Stress and symptom measures were completed at enrollment (ie, prior to the second or third cycle of chemotherapy). Medical records were reviewed for disease and treatment information.

Instruments

Data Analysis

Descriptive statistics and frequency distributions were generated for sample characteristics at enrollment using the Statistical Package for the Social Sciences (SPSS) version 28 (IBM Corporation, Armonk, NY). As was done previously,⁷⁵ unconditional LCA was used to identify shortness of breath profiles that characterized unobserved subgroups of patients (ie, latent classes) over the six assessments. Before performing the LCA, patients who reported the occurrence of shortness of breath for ≤1 of the six assessments were identified and labeled as the “None” class (n = 943, 70.5%). Then, the LCA was performed on the remaining 395 patients using MPlus Version 8.4.⁷⁶

Estimation was carried out with full information maximum likelihood with standard error and a χ² test that are robust to non-normality and non-independence of observations (“estimator=MLR”). Model fit was evaluated to identify the solution that best characterized the observed latent class structure with the Bayesian Information Criterion (BIC), Vuong-Lo-Mendell-Rubin likelihood ratio test (VLMR), entropy, and latent class percentages that were large enough to be reliable.⁷⁷ Missing data were accommodated for with the use of the Expectation-Maximization (EM) algorithm.⁷⁸

Differences among the latent classes in demographic, clinical, and symptom characteristics, as well as stress and resilience measures, were evaluated using parametric and nonparametric tests. A value of P < .05 was considered statistically significant. Bonferroni corrected value of P < .008 was considered statistically significant for the pairwise contrasts.

Latent Class Solution

As noted in our previous publication,²⁶ the 943 patients (70.5%) who had ≤1 occurrence of shortness of breath over the six assessments were classified as the None class. For the remaining 395 patients whose data were entered into the LCA, a three-class solution was selected because the BIC for that solution was lower than the BIC for the 2-class solution. In addition, the VLMR was significant for the 3-class solution, indicating that three classes fit the data better than two classes. Shortness of breath classes were labeled based on their distinct trajectories for the occurrence: Decreasing (8.2%), Increasing (7.8%), and High (13.5%) (Supplemental Figure 1).

Demographic and Clinical Characteristics

As noted in our previous publication,²⁶ significant differences were found among the latent classes for many of the demographic and clinical characteristics (Supplemental Table 1). In brief, compared to the None class, the High class was more likely to live alone, less likely to be employed, and more likely to report a previous or current history of smoking. In addition, they were more likely to have multiple cancer treatments, more likely to have lung metastasis, more likely to be receiving chemotherapy on 21- or 28-day cycles, and more likely to self-report a diagnosis of osteoarthritis, back pain, and rheumatoid arthritis. Compared to the None class, the Decreasing and High classes had lower KPS scores, a higher number of comorbidities, higher SCQ scores, and were more likely to self-report a diagnosis of depression. Compared to the None and Decreasing classes, the High class was more likely to be older and more likely to have lung cancer.

Stress and Resilience Scores

Compared to the None class, the Decreasing and High classes reported higher PSS total, IES-R total, IES-R intrusion, IES-R hyper-arousal, and LSC-R affected sum scores. Compared to the None class, the High class reported higher IES-R avoidance subscale and LSC-R total scores. Compared to the None and Increasing classes, the High class reported higher LSC-R PTSD sum scores and lower CDRS scores (Table 1).

Occurrence of SLEs

Compared to the None class, the High class reported higher occurrence rates for physical neglect, sexual harassment, being forced to touch and have sex before the age of 16, and being separated or divorced. Compared to the None class, the Increasing and High classes reported higher occurrence rates for being forced to touch and have sex at ≥ 16 years. Compared to the None class, the Decreasing class reported a higher occurrence rate for having parents separated or divorced (Table 2).

Symptom Severity Scores

Compared to the None class, the other three classes reported higher levels of morning fatigue and sleep disturbance. Compared to the None class, the Decreasing and High classes reported higher levels of depressive symptoms, trait and state anxiety, cognitive dysfunction, and pain interference and decrements in morning energy. Compared to the None class, the Decreasing and High classes were more likely to have both cancer and non-cancer pain. Compared to the None class, the Increasing and High classes reported higher levels of evening fatigue. Compared to the None class, the High class reported a higher worst pain intensity score (Table 3).

Stress Measures

Compared to the None class, patients in both the Decreasing and High classes reported higher levels of global stress. Similarly, in previous studies of palliative care patients³² and adolescent patients with asthma,⁷⁹ the severity of shortness of breath was positively correlated with global stress. One potential explanation for this relationship, that warrants additional investigation, is that both disruptions in the HPA axis and increases in systemic inflammation occur in oncology patients receiving chemotherapy that contribute to the occurrence of dyspnea.^32,80

In terms of cancer-related stress, compared to the None class, both the Decreasing and High classes reported higher IES-R total, intrusion, and hyperarousal scores. Of note, in the Decreasing and High classes, 23.1% and 14.0% met the criteria for partial PTSD and 14.4% and 23.4% met the criteria for PTSD, respectively. In addition, fears of recurrence and declines in physical function in oncology patients may contribute to the severity of PTSD symptoms.³⁴ These findings suggest that both cancer-related stress and relatively high levels of PTSD result in decreases in the threshold of dyspnea perception. This hypothesis is supported by a study that found that individuals who had PTSD following the World Trades Center disaster were twice as likely to experience shortness of breath than individuals without PTSD.⁸¹

While both the Decreasing and High classes showed similar trends in the severity of global and cancer-specific stress at enrollment, no explanation is readily apparent for why this pattern was not observed in the Increasing class. The distinct trajectories of shortness of breath for the Increasing class compared to the Decreasing class may be explained by differences in the provision of timely symptom management inventions (eg, steroids, oxygen, opioids, thoracentesis); different etiologies for shortness of breath (eg, anemia, pleural effusion); and/or the presence of a variety of triggers (eg, smoking, air pollution). If these trajectories are replicated in an independent sample, additional phenotypic characteristics and interventions warrant evaluation to determine the specific risk factors for membership in the Increasing class.

In terms of the overall number of SLEs, as well as the occurrence of specific SLEs, the primary differences were between the None and the High classes. Specifically, the High class reported higher occurrence rates for physical neglect, sexual harassment, forced to touch at <16 years, forced sex at <16 years of age, and being separated or divorced. In addition, compared to both the None and Increasing classes, the High class reported a higher LSC-R PTSD sum score. Our findings are consistent with previous reports that found that individuals who experienced interpersonal violence, abuse, and neglect had higher levels of PTSD symptoms, compared to those with other types of trauma (eg, unexpected death of a loved one).⁸²

These findings suggest that for the High class in particular, the cumulative impact of various types of stress, particularly adverse childhood experiences (ACEs) may contribute to the occurrence of shortness of breath in patients receiving chemotherapy. Accumulating evidence suggests that exposure to early psychological stress contributes to the lifelong responsiveness of the HPA axis to stress.⁸³ Specifically, repeated activation of the HPA axis may result in blunted cortisol responses to a variety of stressors (eg, airway inflammation).³¹ Over time, reduced inhibitory feedback associated with stress contributes to airway sensitization and chronic/refractory dyspnea.³¹ This hypothesis is supported by a study of the general population⁸⁴ that found that exposure to a higher number of traumatic events and the occurrence of PTSD were associated with increased airflow limitations.⁸⁵ Given the paucity of research on associations between dyspnea and SLEs in oncology patients, additional mechanistic studies are warranted.

Co-occurring Symptoms

Anxiety is hypothesized to play a crucial role in dyspnea perception.⁸⁶ Specifically, anxiety is known to amplify shortness of breath by increasing anxiety sensitivity (ie, the fear of anxiety symptoms⁸⁷) and activating the limbic system.^88,89 Therefore, it is not surprising that across our four distinct shortness of breath profiles, both trait and state anxiety scores exceeded the clinically meaningful cutoffs. Our findings are supported by a study of healthy volunteers,⁹⁰ that found that higher trait anxiety was associated with dyspnea unpleasantness, as well as higher state anxiety levels. Additional investigations are warranted on the common and distinct roles of trait and state anxiety in the affective dimension of shortness of breath.

Depressive symptoms scores approached or exceeded cutoff scores only for the Decreasing and High classes. Evidence suggests that depression decreases the threshold of shortness of breath by altering the perception of respiratory sensations.⁸⁹ Of note, in studies of patients with COPD,^91,92 depression scores were positively correlated with the frequency of shortness of breath; functional impairment related to shortness of breath; emotional and cognitive responses to shortness of breath; and catastrophic thinking. In addition, in a study of patients with lung cancer that used latent variable modeling to create subgroups of patients with distinct functional status profiles,³⁷ patients with the Severe Disability profile reported higher levels of shortness of breath and depressive symptoms. In terms of interventions for the co-occurrence of dyspnea and depression, in a study of patients with advanced cancer,⁹³ the administration of sertraline resulted in decreases over time in the severity of both depression and shortness of breath. Additional research is warranted on the efficacy of antidepressants to decrease one or both of these symptoms.

Consistent with previous findings,^{15,16,39,42,44} morning fatigue scores exceeded clinically meaningful cutoffs in the Decreasing, Increasing, and High classes. This finding is not surprising, given that compared to the None class, these patients reported higher levels of sleep disturbance. In terms of evening fatigue, while all three classes of patients with shortness of breath had scores above the clinically meaningful cutoff, only the Decreasing and High classes had higher scores than the None class. Additional research is warranted to understand the specific causes of fatigue associated with shortness of breath in patients undergoing chemotherapy (eg, respiratory muscle weakness, hypoxia).

Consistent with four studies,^11,21,24,42 patients in the High class reported higher worst pain scores. In addition, patients in both the Decreasing and High classes were more likely to have both cancer and noncancer pain. This finding is aligned with a longitudinal study of a large cohort of Medicare recipients that reported that compared to individuals without dyspnea, individuals with the symptom had higher pain prevalence rates (ie, 18% versus 64%).⁹⁴ In addition, individuals with chest, back, or arthritis pain were substantially more likely to report dyspnea. Of note, the relative risk of dyspnea resolving was greatly increased if the pain problem resolved. One potential explanation for this association is that long-term physical inactivity associated with chronic pain results in physical deconditioning and secondary shortness of breath.⁹⁴ Another potential explanation for the inter-connectedness between the two symptoms is that the perceptions of pain and shortness of breath activate similar brain cortical regions and shares common neural mechanisms.^95,96

While patients in all four of our shortness of breath profiles had GSDS scores above the clinically meaningful cutoff, consistent with two studies of patients with advanced cancer,^15,39 compared to the None class, the other three classes reported higher levels of sleep disturbance. Of note, normal human sleep causes a rapid and shallow breathing pattern; an unpredictable depth of breathing; a significant decrease in tidal volume; and decrements in ventilation and gas exchange during rapid eye movement (REM) sleep.⁹⁷ Given the interconnectedness between sleep and respiration, shortness of breath during the night may contribute to decrements in sleep duration and sleep quality and increases in daytime tiredness (ie, morning fatigue). In terms of potential interventions, in a study of patients with COPD,⁹⁸ progressive relaxation exercises improved dyspnea, fatigue, and sleep disturbance. The improvements in all three symptoms may be the result of increases in lung function and skeletal muscle relaxation, and/or decreases in stress.

Consistent with a previous finding in patients with COPD, asthma, or heart failure,⁴⁸ higher occurrence rates of shortness of breath were associated with more severe decrements in cognitive function on our sample. In addition, in a study of healthy volunteers,⁹⁹ experimentally induced dyspnea interfered with cognitive function. In other studies of community-dwelling older adults¹⁰⁰ and patients with COPD,¹⁰¹ the co-occurrence of depression, anxiety, and sleep disturbance was associated with decrements in cognitive function.

Resilience

All four of our classes had CDRS scores below the normative data for adults in the United States.⁶³ Specifically, compared to both the None and Increasing classes, the High class had significantly lower resilience scores. Given that the High class reported a higher comorbidity burden, a lower functional status, a lower level of social support, and poorer quality of life,²⁶ this finding is not surprising. Likewise, in patients with pulmonary disease,¹⁰² lower levels of resilience were correlated with higher levels of anxiety and depression and lower quality of life scores. Repetitive episodes of shortness of breath may cause fear that leads to the avoidance of daily activity or social interactions, neuropsychological symptoms, and decreases in resilience.^15,103,104 However, because social buffering may modulate HPA reactivity to stressors,¹⁰⁵ resilience training may decrease shortness of breath in oncology patients.¹⁰⁶

Limitations

Several limitations warrant consideration. Given that our sample was relatively homogeneous in terms of gender, race, and ethnicity, our findings may not generalize to more diverse racial and ethnic groups. Given that the major reason for refusal to participate was being overwhelmed with cancer and its treatments, stress levels in these patients may represent underestimates. While this study used a valid and reliable measure to assess the subjective experience of shortness of breath, future studies need to evaluate for correlations with objective measures of pulmonary function and/or neuroimaging. In addition, detailed information is needed on the etiology of shortness of breath and the use of pharmacologic and nonpharmaco-logic treatments. Finally, a more detailed evaluation of demographic and clinical characteristics associated with membership in the Increasing class warrants investigation.