The Rome V criteria for the diagnosis of irritable bowel syndrome in secondary care: a diagnostic accuracy study

The Rome V criteria for irritable bowel syndrome (IBS) were proposed in May, 2026. Their performance for the diagnosis of IBS in clinical practice is as yet unknown. We aimed to conduct a diagnostic accuracy study to examine this issue.

This diagnostic accuracy study was conducted in the specialist IBS clinic in Leeds Teaching Hospitals National Health Service Trust, Leeds, UK. Consecutive adults aged 16 years and older who were referred with suspected IBS between Sept 22, 2016, and June 18, 2024, provided complete symptom data. We applied standardised investigations (including full blood count, C-reactive protein, and coeliac serology), with assessors masked to symptom status. We assessed performance of the Rome V criteria, compared with the previous iterations, Rome IV and Rome III. The reference standard used to confirm IBS was the presence of lower abdominal pain or discomfort in association with altered stool form or frequency combined with no evidence of organic gastrointestinal disease after standardised investigations. Sensitivity, specificity, and positive and negative likelihood ratios were calculated for each of the diagnostic criteria, with 95% CIs.

Of the 745 eligible patients recruited, 726 patients had complete Rome V symptom data (546 [75%] were female, 180 [25%] were male, and mean age was 36·9 years [SD 14·0]). 417 (57%) patients met Rome V criteria for IBS. The level of agreement between the Rome V and either the Rome III or Rome IV criteria was fair to moderate (κ=0·36-0·55). Of patients with complete Rome V symptom data, 590 had IBS according to the reference standard, of whom 390 met Rome V criteria for IBS. Sensitivity of the Rome V criteria was therefore 66·1% (95% CI 62·1-69·9) and specificity was 80·1% (72·4-86·5). Positive likelihood ratio for the Rome V criteria was 3·33 (95% CI 2·40-4·74) and negative likelihood ratio was 0·42 (0·37-0·49). The Rome III and Rome IV criteria for IBS performed differently, with sensitivities of 87·5% (95% CI 84·4-90·3) and 78·9% (75·4-82·1) and specificities of 75·0% (66·3-82·4) and 81·0% (73·4-87·2). Positive and negative likelihood ratios for the Rome III criteria were 3·50 (2·61-4·84) and 0·17 (0·13-0·21) and for the Rome IV criteria were 4·16 (2·98-5·95) and 0·26 (0·22-0·31).

The Rome V criteria for IBS identified a different group of patients, with lower sensitivity than either Rome III or IV criteria in this single centre secondary care study. The clinical relevance of this is uncertain.

None.