Distinct Sleep Disturbance Profiles in Patients with Gastrointestinal Cancers Receiving Chemotherapy

INTRODUCTION

Sleep disturbance is a frequent and distressing symptom in patients receiving chemotherapy that has detrimental effects on their cognitive and functional status, quality of life (QOL), and disease progression.¹ While sleep disturbance in patients with breast,² lung,³ prostate,⁴ and gynecologic⁵ cancer has been documented, less is known about the occurrence of, risk factors for, and impact of this symptom in patients with gastrointestinal cancers. In previous studies of patients with gastrointestinal cancers, occurrence rates for sleep disturbance ranged from 38%⁶ to 63%⁷ and severity scores were in the moderate range.^{7, 8} In a longitudinal study of 361 patients with colorectal cancer,⁹ sleep disturbance was reported by 56% of patients prior to and by 52% during chemotherapy. These findings suggest that a significant number of patients with gastrointestinal cancers report moderate to high levels of sleep disturbance during chemotherapy.

In terms of risk factors, two studies of patients with colorectal cancer found that higher rates of sleep disturbance were associated with pain, anxiety, fatigue, retirement, and the existence of multiple comorbid conditions.^{10, 11} In another cross-sectional study of 434 patients with colorectal cancer,⁶ the occurrence of sleep disturbance was positively correlated with pain and anxiety. In addition, compared to patients with liver cancer who did not experience sleep disturbance,¹² those who reported this symptom had higher rates and severity of co-occurring symptoms. In terms of patient outcomes, sleep disturbance was associated with a poor treatment response,⁹ worse QOL,¹³ higher risk of earlier death,⁹ and lower overall survival.⁹

While these studies provide important information on sleep disturbance in patients with gastrointestinal cancers, several limitations warrant consideration. First, only one study described changes over time in sleep disturbance in these patients.⁹ Second, only four studies examined risk factors for higher levels of sleep disturbance.^{6, 10–12} In addition, little information is available on specific sleep characteristics (e.g., sleep quality, sleep maintenance) in these patients. Finally, none of these studies used a person-centered analytic approach (e.g., latent variable modeling) to evaluate for distinct sleep disturbance profiles in patients with gastrointestinal cancers. Therefore, the purposes of this study were to identify subgroups of patients with gastrointestinal cancers with distinct sleep disturbance profiles and evaluate for differences among these subgroups in demographic, clinical, and sleep characteristics, as well as co-occurring symptoms and QOL outcomes.

METHODS

RESULTS

DISCUSSION

This study is the first to identify subgroups of patients with gastrointestinal cancers with distinct sleep disturbance profiles. Consistent with previous studies of patients with gastrointestinal cancers,^{6, 7, 9} almost 65% of our patients reported high levels of sleep disturbance across all six assessments (i.e., the mean total GSDS scores were above the clinically meaningful cutoff score of ≥43). Similar to the findings from our previous report of the total sample of patients with heterogenous cancer diagnoses,³² the mean total GSDS score for the patients with gastrointestinal cancers (i.e., 50.2) suggests a high overall level of sleep disturbance. However, in the LPA analysis for the total sample that identified the same three sleep disturbance profiles, compared to the patients with the other types of cancer (i.e., breast, lung, gynecological), a higher percentage of patients with gastrointestinal cancers were classified in the Low class. Our current analysis of only patients with gastrointestinal cancers suggests that within cancer type analyses are warranted to identify the characteristics associated with membership in the higher classes even when a specific type of cancer may have a lower overall symptom burden compared to other cancer types. While the relative proportions of patients in the High and Very High classes were slightly higher in the total sample (i.e., 74.8%) compared to the patients with gastrointestinal cancers (i.e., 64.8%), for both samples, their total GSDS scores were almost identical. It should be noted that total GSDS scores of approximately 60 are indicative of significant sleep disturbance as reported in studies of shift workers³³ and mothers and fathers of newborn infants.³⁴

The GSDS subscale scores provide specific information about differences in a number of sleep characteristics among the patient subgroups. As shown in Table 3, for all of the GSDS subscales, scores increased in a stepwise fashion among the three classes. However, across the three classes, insufficient quantity of sleep and a high number of mid-sleep awakenings were the only two subscales that had scores above the clinically meaningful cutoff, which suggests that the problem occurs on three or more days per week.

In addition, the scores on the GSDS subscales allow for an evaluation of two common problems associated with sleep disturbance, namely: difficulty with the initiation of sleep (i.e., sleep onset latency) and difficulty with the maintenance of sleep (i.e., early awakenings and mid-sleep awakenings). The relatively high mid-sleep awakening (i.e., 3.9) score in the Low class as well as the relatively high mid-sleep awakening (i.e., 5.0) and early awakening (i.e., 3.9) scores in the High class suggest that both groups have problems with sleep maintenance. In contrast, patients in the Very High class had problems with both sleep initiation and maintenance. Consistent with previous studies of patients undergoing radiation therapy³⁵ and treatment for breast cancer,³⁶ the use of sleep medications was extremely low. Taken together, these findings suggest that sleep disturbance is a significant problem in patients with gastrointestinal cancers. Clinicians need to determine the underlying processes for these sleep disturbances and teach patients effective sleep hygiene interventions (e.g., increases in daytime activity, establishment of a regular bedtime).^{37, 38}

One goal of this study was to identify common and distinct risk factors associated with higher levels of sleep disturbance within and across cancer types. To achieve this goal, as summarized in Table 5, we identified characteristics associated with membership in the Very High sleep disturbance class compared to the Low sleep disturbance class in both the current sample of patients with gastrointestinal cancers and the total sample with heterogeneous types of cancer.³² For both samples, the common risk factors associated with membership in the Very High class were: younger age, not being married/partnered, being unemployed, having childcare responsibilities, lack of regular exercise, having a lower performance status, and having a higher comorbidity burden. Findings regarding age differences in sleep disturbance are inconsistent with some finding associations with younger ^{39, 40} and others with older ^{41, 42} age. These inconsistencies may be related to specific types of cancer and/or the timing of the assessments.

The added stress associated with some of the other demographic risk factors may explain their associations with very high levels of sleep disturbance. For example, patients who are not married or partnered may experience higher levels of loneliness which is associated with sleep disturbance.⁴³ The added stress of caring for children while receiving chemotherapy may contribute to a lack of sleep or poorer quality of sleep.⁴⁴ In addition, being unemployed may be associated with an increased financial burden and associated with sleep problems.⁴⁵ While these characteristics are not modifiable, clinicians can provide referrals to social services and/or psychological counseling to help patients manage their stress.

Lack of regular exercise is the only modifiable risk factor associated with membership in the Very High sleep disturbance class. The benefits of exercise to decrease sleep problems are well documented.^{10, 46} For example, a meta-analysis found that patients who participated in an aerobic exercise program (i.e., 4 to 8 week, 80 to 149 minutes per week) experienced significant reductions in sleep disturbance (i.e., medium to large effects).⁴⁷ Therefore, clinicians need to encourage patients to increase aerobic exercise during and following chemotherapy.

Consistent with prior studies,^{3, 36} higher levels of sleep disturbance were associated with a higher comorbidity burden and lower functional status in our patients with gastrointestinal cancer. Of note, the differences in SCQ and KPS scores between the patients in the Low and Very High classes represent not only statistically significant but clinically meaningful differences (i.e., d= 0.8 and 1.3, respectively). It is reasonable to hypothesize and is supported by studies of patients with other chronic conditions (e.g., hypertension, diabetes) that higher comorbidity burden contributes to sleep disturbance.^{48, 49} In addition, findings from this study suggest that a low functional status may contribute to a lack of regular exercise and associated sleep disturbance.

In terms of the distinct factors that were associated with membership in the Very High class, in the patients with gastrointestinal cancers, three factors were identified, namely: having received prior cancer treatments, having a higher total number of prior cancer treatments, and having higher number of comorbidities. While not assessed in the total sample, a higher percentage of patients in the Very High class self-reported diagnoses of back pain and depression. Compared to the 58% of patients in the Low class, 80.6% of the patients with gastrointestinal cancers in the Very High class had received at least one additional treatment prior to chemotherapy. While the timing between the previous and current treatments was not evaluated, this finding supports previous research that suggests that the effects of cancer treatments on sleep disturbance may be cumulative.⁷ The high occurrence rates for self-reported back pain (33.3%) and depression (31.7%) in the Very High class of patients with gastrointestinal cancers provide some insights into the associations between sleep disturbance and comorbidity burden. For example, in the general population both back pain⁵⁰ and depression⁵¹ are associated with sleep disturbance.

In our previous studies,^{52, 53} we found that patients with gastrointestinal cancers experienced an average of 10 to 15 multiple co-occurring symptoms. Similar to the GSDS subscale scores, levels of trait and state anxiety, depressive symptoms, cognitive dysfunction, morning and evening fatigue, and pain intensity and interference were significantly different among the three sleep disturbance classes of patients with gastrointestinal cancers (Table 4). It is interesting to note that for all three sleep disturbance classes, morning energy scores were below the clinically meaningful cutoff score. However, only patients in the Very High class had scores that were well above the clinically meaningful cutoff for all of the other symptoms listed in Table 4.

When comparisons were done between the current sample and the total sample (Table 5), our findings suggest that very high levels of sleep disturbance are associated with multiple co-occurring symptoms. It should be noted that in both samples, the differences in all of the symptom severity scores, between the Low and Very High classes represent not only statistically significant but clinically meaningful differences (i.e., effect sizes ranged from 0.6 [morning energy] to 1.8 [morning fatigue]). While previous studies have reported on positive associations between sleep disturbance and fatigue,⁴⁶ anxiety,⁵⁴ depression,⁵⁵ and pain⁵⁶ in oncology patients, none of these studies assessed all of these co-occurring symptoms in the same sample of patients. The hypothesized mechanisms for the co-occurrence of sleep disturbance and other symptoms include: increases in proinflammatory cytokine activity; dysregulation of the hypothalamic-pituitary-adrenal axis; and misaligned circadian rhythms.^{55, 57} Future studies need to determine the common and distinct mechanisms for the co-occurrence of these symptoms.

In terms of generic and disease-specific QOL outcomes, except for the spiritual well-being subscale of the QOL-PV, significant differences were found among the three sleep disturbance classes. Compared to the Low class, all of the differences across the various domains of QOL in the Very High class represent not only statistically significant but clinically meaningful differences (d = 1.3 to 1.7). Of note, across all three classes, patients with gastrointestinal cancers reported PCS scores of <50 which is lower than the normative score for the general population.²⁷ Consistent with previous studies in patients with lung³ and ovarian⁵⁸ cancer, high levels of sleep disturbance were associated with poorer QOL outcomes.

CONCLUSIONS

This study is the first to identify subgroups of patients with gastrointestinal cancers with distinct sleep disturbance profiles and identify risk factors associated with higher levels of sleep disturbance. Additional research is warranted to explore underlying molecular mechanisms that contribute to the development of sleep disturbance and the co-occurrence of other common symptoms during chemotherapy. Clinicians need to assess for the common risk factors and associated co-occurring symptoms, as well as initiate personalized symptom management interventions and referrals.