Poor Diagnostic Power of Functional Bowel Disorders With Colonoscopy

Investigators examined the diagnostic yield of colonoscopies for patients with functional bowel disorders without alarm features.

Colonoscopies had little diagnostic yield for patients with functional bowel disorders (FBD) without alarming features, according to results from a prospective cross-sectional study, which were published in Clinical Gastroenterology and Hepatology.

Investigators recruited patients (N=917) without known gastrointestinal conditions who were recommended for an outpatient colonoscopy at the gastroenterology department of the Sheffield Teaching Hospital in the United Kingdom during 2019. Patients completed the Rome IV diagnostic questionnaire at home before the procedure to record their demographics and symptoms. Clinicians were blinded to reported symptoms.

Among questionnaires, self-reported symptoms met the criteria for functional diarrhea (FD) (n=177), irritable bowel syndrome with predominant diarrhea (IBS-D) (n=160), IBS with mixed symptoms (IBS-M) (n=129), functional constipation (FC) (n=109), and IBS with predominant constipation (IBS-C) (n=71). The remaining 271 participants did not meet the criteria for FBD.

The patients who met the criteria for a FBD were more likely to be women (61% vs 41%; P <.0001) and to be younger (mean age, 55 vs 62 years; P <.0001) compared with patients not meeting the FBD criteria. The 2 groups differed significantly for the alarm features of unintentional weight loss (24% vs 13%; P <.0001), nocturnal symptoms (40% vs 9%; P <.0001), fever (8% vs 2.6%; P =.004), family history of irritable bowel disease (IBD) (6.5% vs 2.6%; P =.02), anemia (22% vs 43%; P <.0001), and presence of any alarm feature (98% vs 100%; P =.03) for participants with and without an FBD, respectively.

Of the alarm features, abnormal gastrointestinal exam (adjusted odds ratio [aOR] 4.3; 95% CI, 1.5-12.2), increased inflammatory markers (aOR 2.4; 95% CI, 1.4-4.2), and rectal bleeding (aOR 1.8; 95% CI, 1.0-3.0) were all independent predictors of organic disease.

The number of alarm features for each patient increased the likelihood of organic disease in a stepwise manner (0 for patients with no features; 10% for 2-3 features; 15% for ³3 features).

Among participants with an FBD compared with no FBD, a higher prevalence of organic disease was observed (12.2% vs 8.1%; P =.07); and significantly more IBD (6.2% vs 2.2%; P =.012) and microscopic colitis (2.9% vs 0; P =.004) diagnoses were made, but fewer colorectal cancers (3.1% vs 5.9%; P =.046) were diagnosed, respectively.

The overall diagnostic yield of colonoscopy for patients with FBD was 12%. Stratified by specific subtype (P =.005), the colonoscopy had the lowest power for IBS-C (~6%), followed by IBS-M (~9%), and IBS-D (~17%).

Limitations of this study were that microscopic colitis may have been underestimated as a quarter of eligible patients did not have a biopsy sampled during colonoscopy; also, celiac disease and idiopathic bile acid diarrhea were not considered in the study for patients with chronic diarrheal symptoms.

The study authors concluded that despite the presence of alarm features among patients, few were diagnosed with organic disease by colonoscopy.

Follow @Gastro_Advisor


Asghar Z, Thoufeeq M, Kurien M, et al. Diagnostic yield of colonoscopy in patients with symptoms compatible with Rome IV functional bowel disorders [published online August 31, 2020]. Clin Gastroenterol Hepatol. doi: 10.1016/j.cgh.2020.08.062