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


Subjects and Clinical Characteristics

The subjects in this report were prospectively recruited from the authors' (GSS and CPG) out-patient tertiary GI practices, as well as from campus-based advertisements from 2009 to 2013. All study participants who agreed to participate completed seven days of sleep actigraphy monitoring and daily log of GI and extraintestinal symptoms, as detailed below. In addition, comprehensive multidimensional symptom, affective state and quality of life questionnaires also were implemented. At the conclusion of the 1-week study period, subjects returned their actigraph watches and daily logs for data analysis. A total of 51 individuals were approached to participate, and only one IBS patient did not complete the study. The IBS group was composed of adult patients (≥18 years old) who were both clinically diagnosed with IBS by a gastroenterologist and met Rome III diagnostic criteria.[19] All IBS patients had organic bowel disease excluded via a comprehensive evaluation performed at the discretion of the treating gastroenterologist. The control group was comprised of healthy individuals over 18 years of age without ROME III criteria for any functional GI disorder, prior gastrointestinal diagnoses or active GI symptoms. Exclusion criteria for both study groups included a history of structural GI illness, prior GI surgery, major medical illness, history of alcohol or substance abuse, or history of a sleep disorder or sleep apnoea. Use of sleep medications (as needed, less than once a week) prior to enrollment was acceptable, but was not permitted during study participation. IBS patients were continued on all previously prescribed medications, including anti-depressants at a stable medication dose without adjustment for at least 4 weeks prior to enrollment. Using pilot data estimate of mean waking episodes during sleep of 10 ± 4 in IBS and 8 ± 4 in controls, a sample size of 50 subjects was calculated to allow a power of 0.80, alpha = 0.05 to detect a statistically significant difference in waking episodes in the study groups. The study protocol was approved by the Human Research Protection Office (Institutional Review Board) at Washington University School of Medicine and Barnes-Jewish Hospital, St. Louis, Missouri.

Study Measures

Demographics and Medical History. Gender, age, race, body mass index and marital status were recorded. Medical history included documented past medical history of IBS, other GI- and non-GI functional disorders, sleep disorders and other GI illnesses, current medications, tobacco and alcohol use.

GI Symptom Measures. The ROME III Research Diagnostic Questionnaire was administered to establish presence of ROME III-defined functional GI disorders.[20] The GI Symptom Rating Scale for IBS (GSRS-IBS) was administered as a validated rating scale, consisting of fifteen items assessing prototypical IBS symptoms.[21] GI symptom burden within the 2 weeks preceding enrollment was used to evaluate symptom Severity, Bother and Frequency. GI symptom Severity and Bother were assessed with 10-cm Visual Analog Scales (VAS), using previously described methods.[22] Symptom Frequency (total number of symptomatic days) within the preceding 2 weeks was quantified (0–14 days).

During the 7-day enrollment period, a daily bowel symptom log was completed before bedtime each evening, which recorded bowel symptom Severity and Bother (10-point Likert scale), accessory bowel symptoms (bloating/distention, gas/flatus, mucus and urgency along a five-point Likert scale where 1 = 'none/mild' and 5 = 'severe'), the number of bowel movements over the preceding 24 h, and a Bristol stool scale for the predominant bowel pattern that day.[23]

Somatic Symptom Measures. At baseline, all participants completed the Patient Health Questionnaire-15 (PHQ-15) as a validated assessment of somatic symptoms, including headache, arthralgia, back pain and fatigue.[24] Furthermore, a daily assessment of common non-GI symptoms were assessed in a symptom log completed each evening (back/hip pain, headache, neck/shoulder pain, achiness, muscle/joint pain, fatigue and sexual dysfunction, where 1 = 'none/mild' and 5 = 'severe').

Psychological Assessments. The Beck Depression Inventory (BDI) and Beck Anxiety Inventory (BAI), validated 21 question multiple-choice self-reported inventory measuring the severity of depression and anxiety over the previous 4 weeks were used to screen for these mood disorders[25,26] The Visceral Sensitivity Index (VSI) was employed as a validated 15-item self-report questionnaire used to measure GI specific anxiety.[27,28]

Quality of Life Measures. The IBS-QOL is a validated measure of IBS-specific HRQOL,[29] with overall scores averaging 63.2 ± 18.5 in IBS samples. The Work Productivity and Activity Impairment questionnaire for IBS (WPAI-IBS) consists of six items and is a validated measure used to quantify the effects of IBS on productivity and daily activities; outcomes are expressed as impairment percentages, with higher numbers indicating greater impairment and less productivity.[30] The short form 36 (SF-36) was used as a validated assessment of health status and its impact on general HRQOL.[31,32] The SF-36 is normalised to a maximal score of 100 (higher scores indicating better HRQOL), and divided into physical and mental domains. This assesses the role that medical conditions and pain have on physical and emotional well-being, and on limitation of day-to-day and pleasurable activities.

Sleep-specific Measures. All subjects also completed a National Sleep Foundation Daily Sleep Log (www.sleepfoundation.org) with questions regarding subjective sleep quality (estimated time to fall asleep, number of awakening at night and sleep duration in hours). The log, completed in the morning to reflect the previous night's sleep, also allowed free-form recording of any perceived issues which may have interrupted sleep (e.g. pain, urination, dreams and environmental factors). The Pittsburgh Sleep Quality Index (PSQI) was completed at the conclusion of the monitoring period as a validated retrospective self-report measure of sleep quality and disturbance. Individuals with sleep problems or poor sleep quality have higher PSQI scores [minimum score = 0 (better); maximum score = 21 (worse)], with scores >5 being regarded as reflecting poorer sleep quality.[33]

Actigraph Recording

Subjects wore a wrist-mounted actigraph monitor (Motionlogger, Ambulatory Monitoring, Ardsley, NY, USA) on their nondominant hand continuously for the duration of the 7-day study period. The uploaded data was analysed on ActionW software version 2.7. Subjects were permitted to remove the monitor to brief periods of time (e.g. bathing), but were instructed to record these events both using the 'Event' button on the actigraph, and in their daily logs. Actigraphy has been validated as an objective measure of sleep when compared to polysomnography and offers the advantage of 'real-world' testing.[34] Actigraphy parameters of particular interest included mean sleep episode duration, waking episodes during sleep, longest undisturbed sleep episode, and total hours of sleep per day (Figure 1). In addition, every subject's daily actigraph recording was manually compared to the patient's daily sleep log by a single, blinded investigator (AP) to ensure accuracy of the automated actigraph interpretation algorithm in classifying sleep status.

Figure 1.

Example of sleep actigraphy from a study subject. Actigraph tracing from a 7-day study period (time along x-axis), and measured activity (y-axis) Sleeping hours (red) are magnified to show detail, with waking episodes during sleep (red arrows) are recorded via patterned movement. Sleep episode duration (purple interval) is defined by the time between waking episodes.

Statistical Analysis

Grouped values are reported as mean, standard error of mean and 95% confidence intervals or medians with associated range, where appropriate. Between-group comparisons were performed using Student's t-tests for continuous variables, or independent samples median and Mann–Whitney U-test for nonparametric measures; Chi-square or Fisher's exact analyses were carried out on binomial data as indicated. In each case, P < 0.05 was required for statistical significance. Pearson correlations were performed to establish: (i) the relationship between the objective sleep actigraphy measures and following-day reports of GI- and somatic symptoms; (ii) to determine the association of subjective reports of sleep quality with the actigraphy data and (iii) to assess the relationship of anxiety measures with sleep quality. To conduct these analyses, each patient-day was regarded as a discrete data point. Univariate linear regression models were developed to assess the actigraphy sleep measures on overall and IBS-specific QOL, as well as reports of recent IBS symptoms. Sobel mediational analyses which included actigraphy sleep measures (waking episodes), generalised and visceral-specific anxiety, and depression measures were performed to explore if effect of sleep on IBS symptoms might be mediated by concomitant mood disturbances. Pre-conditionally, all of the variables to be included in the analysis were required to be significantly correlated using Pearson correlations, and linear regression models were developed in including the independent (e.g. waking episodes during sleep) and dependent variable of interest (IBS symptoms on GSRS-IBS), followed by second order models which included the mediational variables of interest (e.g. VSI, HADS anxiety) to determine the unstandardised regression coefficients and their standard errors. Statistical analysis was carried out using spss 22 software (IBM, Armonk, NY, USA).