Prevalence of constipation: agreement among several criteria and evaluation of the diagnostic accuracy of qualifying symptoms and self-reported definition in a population-based survey in Spain.

The aims of this study were to estimate the prevalence of chronic constipation and to evaluate the diagnostic accuracy of the symptoms and the self-reported definition of constipation. A cross-sectional survey was conducted in the general community in 1999. A questionnaire comprising 21 items was developed and mailed to a random sample of 489 subjects who were aged between 18 and 65 years and who belonged to a Spanish population. In the 349 subjects (71%) responding to the questionnaire, the prevalence of self-reported constipation was 29.5% (95% confidence interval (CI): 24.9, 34.3) versus 19.2% (95% CI: 15.1, 23.3) and 14.0% (95% CI: 10.4, 17.7) based on Rome I and Rome II criteria, respectively. Agreement was good between self-reported and Rome I criteria (kappa: 0.68) and between Rome I and Rome II criteria (kappa: 0.71), and it was moderate between self-reported and Rome II criteria (kappa: 0.55). Female gender was identified to be a risk factor for constipation; fiber intake and physical exercise were found to be protective factors. Likelihood ratios were higher for the presence of anal blockage and straining and for the absence of hard stools. Chronic constipation is a highly prevalent problem, especially in women. Different prevalence estimates of constipation were observed using different criteria, although agreement between them was acceptable. Anal blockage, straining, and hard stools show the greatest accuracy for the diagnosis of constipation.