Symptom criteria for constipation subtypes in a U.S. national epidemiological study of constipation

Abstract

Introduction. Existing criteria for constipation are based on expert clinical consensus. We conducted a large US national study to evaluate the validity of these symptom-based clinical consensus criteria for functional constipation, IBS constipation and outlet delay. Methods. We contacted a nationwide sample of households by telephone. Among eligible individuals contacted, 69% agreed to participate. A total of 10,018 interviews were completed. Individuals were asked detailed questions about 15 constipation-related symptoms. Interview data were used to identify individuals who met a priori symptom criteria for functional constipation (e.g., Rome criteria), IBS (e.g., Rome), and outlet delay (e.g., Talley etal). Principal components factor analysis was used to determine if the 15 different symptom variables clustered in a manner consistent with a priori criteria. Results. Three factors (i.e., symptom clusters) emerged that overlapped to some degree but were distinct from a priori definitions for functional constipation, IBS, and outlet delay. A fourth factor emerged for self-reported constipation. While number of bowel movements per week (BM/week) is often used to define constipation, it was not strongly correlated with (i.e., did not load with) any of the factors. In contrast, number of completed (i.e., a sense of complete emptying) BM/week was an important item strongly correlated with functional constipation. The factors and loading or weight for each item (i.e., correlation of the item with the factor and visa versa) are summarized below. For self-reported constipation the items (and weights) were self report (0.74), laxatives (0.48), straining (-0.59) and hard stools (-0.67).