Cancer sleep symptom-related phenotypic clustering differs across three cancer specific patient cohorts

Abstract

Specific sleep disorders have been linked to disease progression in different cancers. We hypothesised sleep symptom clusters would differ between cancer types. The aim of this study was to compare sleep symptom clusters in post-treatment melanoma, breast and endometrial cancer patients. Data were collected from 124 breast cancer patients (1 male, 60 ± 15 years, 28.1 ± 6.6 kg/m² ), 82 endometrial cancer patients (64.0 ± 12.5 years, 33.5 ± 10.4 kg/m² ) and 112 melanoma patients (59 male, 65.0 ± 18.0 years, 29.1 ± 6.6 kg/m² ). All patients completed validated questionnaires to assess sleep symptoms, including the Epworth Sleepiness Scale (ESS), Pittsburgh Sleep Quality Index (PSQI), Insomnia Severity Index (ISI), and Functional Outcomes of Sleep Questionnaire-10 (FOSQ-10). Snoring, tiredness, observed apneas, age, BMI, and gender data were also collected. Binary values (PSQI, ISI, FOSQ), or continuous variables for sleepiness (ESS) and perceived sleep quality (PSQI), were created and sleep symptom clusters were identified and compared across cancer cohorts. Four distinct sleep symptom clusters were identified: minimally symptomatic (n = 152, 47.7%); insomnia-predominant (n = 87, 24.9%); very sleepy with upper airway symptoms (n = 51, 16.3%), and severely symptomatic with severe dysfunction (n = 34, 11.1%). Breast cancer patients were significantly more likely to be in the insomnia predominant or severely symptomatic with severe dysfunction clusters, whereas melanoma patients were more likely to be minimally symptomatic or sleepy with upper airway symptoms (p <0.0001). Endometrial cancer patients were equally distributed across symptom clusters. Sleep symptom clusters vary across cancer patients. A more personalised approach to the management of sleep-related symptoms in these patients may improve the long term quality of life and survival.

Keywords: breast cancer; endometrial cancer; melanoma; sleep symptom clusters.

FIGURE 1

Estimated response probabilities of binary symptom‐variable for each symptom‐cluster (minimally symptomatic, cluster 1 = orange; insomnia predominant, cluster 2 = green; very sleepy with upper airway symptoms cluster 3 = purple; and severely symptomatic with severe dysfunction cluster 4 = blue). Sleep quality and sleepiness are not shown as they were not binary variables

FIGURE 2

Estimated proportion of patients within each of the four symptom‐clusters (cluster 1, minimally symptomatic; cluster 2, insomnia predominant; cluster 3, very sleepy with upper airway symptoms; and cluster 4, severely symptomatic with dysfunction) having scores within bands for (a) ESS total scores and (b) PSQI question 6. (a) For the following bands shown for ESS total scores: (0–5, “not sleepy” (yellow); 6–10, “mildly sleepy” (orange); 10–15, “moderately sleepy” (maroon); and > 15, “severely sleepy” (purple)). (b) The following bands for question 6 of PSQI are (“very good”, yellow; “fairly good”, light green; “fairly bad”, dark green; and “very bad”, blue. PSQI, Pittsburgh Sleep Quality Index; ESS, Epworth Sleepiness Scale

FIGURE 3

Frequency distribution for number of patients falling into each of the four symptom‐clusters. Cluster 1, minimally symptomatic; cluster 2, insomnia predominant; cluster 3, very sleepy with upper airway symptoms; and cluster 4, severely symptomatic with dysfunction

FIGURE 4

Heatmap profile of symptom burden within the four symptom‐clusters (cluster 1, minimally symptomatic; cluster 2, insomnia predominant; cluster 3, very sleepy with upper airway symptoms; and cluster 4, severely symptomatic with dysfunction), for symptom‐variables derived from ESS, FOSQ, ISI, PSQI, and SB (STOP‐BANG) questions, respectively. Values within each cell represent response probabilities for each cluster, where cell colour represents relatively low (green) and high (brown) degrees of symptom burden. For ESS total scores (0–5, “not sleepy”; 6–10, “mildly sleepy”; 10–15, “moderately sleepy”; and > 15, “severely sleepy.” PSQI, Pittsburgh Sleep Quality Index; ESS, Epworth Sleepiness Scale; ISI, Insomnia Severity Index; FOSQ, Functional Outcomes of Sleep Questionnaire; SB, STOP‐BANG

FIGURE 5

Proportion and distribution of covariates (cancer group, gender, and age) across the four symptom‐clusters (cluster 1, minimally symptomatic; cluster 2, insomnia predominant; cluster 3, very sleepy with upper airway symptoms; and cluster 4, severely symptomatic with dysfunction). (a) Bar graph describing the proportion of patients with a history of breast cancer (blue), endometrial cancer (yellow), and melanoma (red) within each symptom‐cluster. (b) Bar graph describing proportion of female (light blue) and male (dark blue) genders within each symptom‐cluster. (c) Distribution of age (years) within each symptom‐cluster (cluster 1, orange; cluster 2, green; cluster 3, purple; cluster 4, blue. Greater width indicates higher density probability of a particular age