Effects of Disturbed Sleep on Gastrointestinal and Somatic Pain Symptoms in Irritable Bowel Syndrome

A. Patel; S. Hasak; B. Cassell; M. A. Ciorba; E. E. Vivio; M. Kumar; C. Prakash Gyawali; G. S. Sayuk


Aliment Pharmacol Ther. 2016;44(3):246-258. 

In This Article


Patient Characteristics

A total of 50 subjects (mean 38.6 ± 1.0 years, 44 female) were enrolled. Baseline demographic and clinical characteristics are shown in Table 1. The median age and gender distribution between IBS patients and healthy controls were nearly identical, though systematic matching was not employed. There was no statistically significant difference between the two groups with regard to BMI, race, marital status, or tobacco use. Healthy controls were slightly more educated, and reported greater rates of employment. IBS patients had significantly higher rates of depression, general anxiety and visceral-specific anxiety symptoms as compared to the controls (P < 0.001). The IBS group also reported significantly more severe, bothersome and frequent bowel symptoms than their healthy counterparts (P < 0.001 for all), with more bowel symptoms across all GSRS domains (P < 0.05 for all except constipation), poorer IBS-QOL (P < 0.001) and general HRQOL (SF-36; P = 0.004), as anticipated. Tricyclic medication use was higher in the IBS group (P = 0.03), as expected given its common use in IBS management. This IBS study population did not statistically differ from our overall IBS clinic population with regard to demographics, baseline GI symptom Severity/Frequency, HRQOL or psychological measures.

Subjective Sleep Data

Fewer IBS patients reported waking up feeling 'refreshed' or 'somewhat refreshed' (72.6% of controls vs. 55.8% of IBS patients, P = 0.01). The most commonly cited reason for disturbed sleep was the sleep environment (Figure 2). Only 2.4% of IBS patients attributed their sleep disturbances to GI symptoms, with more patients ascribing sleep disturbances to nocturia and non-GI pain such as back pain, headaches and arthralgias. IBS patients perceived similar sleep duration per night as healthy individuals (approximately 7 h), but reported having significantly increased number of awakenings per night when compared to controls (2.3 ± 0.2 vs. 1.7 ± 0.1, P < 0.001; Table 2). Using the validated PSQI questionnaire, there were significant differences in subjective sleep quality between the two groups (P = 0.025).

Figure 2.

Subjective aetiologies of sleep disturbances by study group.

Actigraphy Sleep Data

In contrast to subjective daily log reports, objective actigraph data revealed IBS patients spent significantly greater amount of time sleeping each day compared to controls (7.7 ± 0.2 vs. 7.1 ± 0.1 respectively, P = 0.008). IBS patients also had a greater number of waking episodes during sleep than controls (12.1 ± 0.6 vs. 9.3 ± 0.5, P < 0.001). As a result, the IBS group had significantly shorter mean sleep episode duration and longest undisturbed sleep episode (P < 0.001 and P = 0.003 respectively; Figure 3). Actigraph measures of disturbed sleep correlated well with total Pittsburgh Sleep Quality (r = 0.43, P < 0.001) in addition to several PSQI subdomains; however, actigraphy did not correlate with self-reported sleep on daily logs (hours of sleep or sleep awakenings, r < 0.12, P > 0.15 for each).

Figure 3.

Actigraphy parameters by study group. Actigraphy parameters of particular interest included mean sleep episode duration (hours) waking episodes during sleep, longest undisturbed sleep episode (hours) and total hours of sleep per day.

Given the known potential influences of tricyclic anti-depressants (TCAs) on sleep, a post hoc analysis of IBS patients on and off tricyclic medications was performed. Thus, we first compared the sleep parameters among IBS patients on TCA (n = 6) compared to those not on TCA (n = 18). No significant differences were found in the IBS subgroups on- or off TCAs in terms of sleep quality (number of waking episodes, sleep duration, or longest undisturbed sleep; P > 0.30 for each). However, significant differences in actigraphy sleep measures persisted when IBS patients not taking TCAs (n = 18) were compared to healthy controls (P < 0.03 for each).

Relationship of Disturbed Sleep and GI Symptom Data in IBS

Number of waking episodes during sleep ('waking episodes') was utilised as a representative measure of sleep disturbance (correlations between waking episodes and longest- and mean sleep episodes were r = −0.50, r = −0.62 respectively, P < 0.001 for each). Significant relationships were found between waking episodes and abdominal pain ratings (VAS, GSRS and daily symptom log all (r ≥ 0.25, P ≤ 0.01 for each) as well as GI Bother (r = 0.43, P < 0.001) and daily abdominal distress (r = 0.37, P < 0.001) as shown in Figure 4. Greater number of symptomatic days also were related to disturbed sleep (P < 0.001). Sleep disturbances did not, however, have significant relationships with bowel pattern (constipation or diarrhoea on the GSRS; P > 0.18 for each) or with other bowel symptoms such as daily reports of bloating, mucus, gas or urgency (P > 0.3 for each). In contrast to that seen with IBS patients, healthy controls had no significant relationship with sleep disturbances, any of the measures of abdominal pain or bowel pattern (data not shown). Self-reported sleep on the Pittsburgh Sleep Quality Index (PSQI), including the total PSQI score, sleep disturbance and sleep latency domains correlated well with self-reported GI pain ratings, including bloating and daily rating of abdominal distress. As with PSQI data, self-reported sleep disturbances correlated with abdominal pain on the GSRS (r = 0.30, P = 0.003), but demonstrated much weaker associations with other GI symptom domains, including bowel symptom Bother and Frequency as well as GSRS nonpain symptoms (P > 0.05 for each).

Figure 4.

Correlation matrix between GI symptoms and actigraphy, subjective sleep report in IBS patients. Correlations are colour coded as indicated by the legend, with positive correlations shown in red, and negative correlations in blue. The colour intensity reflects the strength of the correlation. GSRS, Gastrointestinal Symptom Rating Scale.

Relationship of Disturbed Sleep on Health-related Quality of Life (HRQOL) in IBS

Actigraphic measured waking episodes demonstrated robust relationships with patient reports of both general- and IBS-specific HRQOL in the IBS cohort (r = −0.58 and −0.52 respectively, P < 0.001 for each) (Figure 5). Similarly, several Pittsburg Sleep Quality domains correlated well with HRQOL, including total PSQI, sleep disturbances and days of dysfunction due to sleep (SF-36: r = −0.58 to −0.65, P < 0.001; IBS-QOL: r = −0.53 to −0.56, P < 0.001). Daily sleep log reports of sleep quality were not significantly associated with HRQOL measures.

Figure 5.

Correlation matrix between non-GI symptoms, quality of life, psychatric symptoms and actigraphy, subjective sleep report in IBS patients. Correlations are colour coded as indicated by the legend, with positive correlations shown in red, and negative correlations in blue. The colour intensity reflects the strength of the correlation.

Relationship of Disturbed Sleep and non-GI Symptoms and Psychiatric Features in IBS

Greater numbers of waking episodes detected on actigraphy significantly correlated with several non-GI pain symptoms (back pain, headache, neck pain; r ≥ 0.28, P ≤ 0.01 for each) and somatoform complaints (PHQ-12: r = 0.17, P = 0.036). Total PSQI also correlated well with these specific pain symptoms and somatoform complaints (r = 0.27–0.55, P ≤ 0.05 for each) as well as for the more general symptoms of achiness, muscle pain and fatigue (≥0.35, P < 0.005 for each). Daily sleep logs did not exhibit statistical relationships with these non-GI symptoms.

Mood measures, including depression, anxiety and GI specific anxiety all correlated significantly with waking episodes on actigraphy (r ≥ 0.34, P < 0.001 for each), and several PSQI measures (total score, sleep disturbance and days dysfunction due to sleep; r ≥ 0.28, P ≤ 0.009 for each, the exception being VSI, r = 0.14, P = 0.2). Linear regression models established that mood measures influenced following-day abdominal pain independently of the effects of sleep (waking episodes) in the IBS patients (B ≥ 0.097, P ≥ 0.03 for each). Mediation analysis exploring the relationship of sleep disturbances and mood suggested that sleep effects on abdominal pain symptoms may be partially mediated by the influence of disordered sleep on mood, particularly depression and visceral-specific anxiety (Figure S1 ).