Episodic nature of symptoms in irritable bowel syndrome.

OBJECTIVES: The objectives of this study were to determine whether the symptoms of diarrhea (defined as loose or watery stools), constipation (hard or lumpy stools), abdominal pain, and bloating occur in episodes rather than sporadically in patients with irritable bowel syndrome (IBS); to identify rules for defining the onset and termination of symptom episodes; and to assess the overlap of these episodes.

METHODS: IBS patients kept a symptom log in which they rated the consistency of each bowel movement (BM) on the Bristol Stool Scale for 3 months. Each night they transferred these data to an internet website and also rated abdominal pain and bloating for that day. Data were analyzed for 124 patients who completed at least 21 consecutive diary days (mean of 73 days) without taking laxative, antidiarrheal, or IBS-specific medications. For each symptom in each patient, we computed the correlation between consecutive observations (autocorrelations) in the diary to determine whether the symptom tended to occur in clusters of several instances, as would happen in episodes vs. happening randomly. Next, we compared different patterns by which diarrhea and nondiarrhea stools alternate to identify episode definitions that captured at least 75% of loose/watery stools. A similar pattern analysis was performed for constipation. Pain and bloating episodes were defined as days with an intensity rating >3 on a 0-10 scale. These patterns were converted into rules for defining the onset and termination of symptom episodes. Last, we used these episode definitions to examine the overlap of pain with episodes of diarrhea, constipation, and pain.

RESULTS: Significant (P<0.05) autocorrelations were found in the Bristol Stool Scale ratings of 69.4% of patients and in the daily abdominal pain and bloating ratings of 52.4% and 68.5% of patients, respectively. Defining a diarrhea episode as two or more loose/watery stools never separated by >1 nonloose/watery stool or by a day without a BM captured 76% of all loose/water stools. Defining constipation episodes as two or more hard/lumpy stools never separated by >1 nonhard/lumpy stool captured 80% of hard/lumpy stools. Sequences of 3 or more days without a BM were also defined as constipation episodes because they were strongly associated with hard stools. Average episode durations were 2.1 days for diarrhea, 4.5 days for constipation, 3.1 days for pain, and 3.5 days for bloating. Overlap analysis showed that only 41.6% of constipation episode days and 67.0% of diarrhea episode days were pain episode days. Bloating and pain coexisted on 59.1% of days on which either type occurred.

CONCLUSIONS: Loose/watery stools and hard/lumpy stools occur in well-defined episodes. Pain and bloating also occur in episodes, but contrary to the Rome criteria more than half of the pain episodes occur outside episodes of abnormal stool consistency.