Skip to content

Diagnosis and the Rome criteria

TL;DR — IBS has no structural, serological or physiological marker; it is defined entirely by a symptom rule, and that rule has been rewritten five times since 1978. Each rewrite changes who has the disease: switching from Rome III to Rome IV halved community prevalence (9.0% → 4.6% in a 5,931-person Anglophone survey) almost entirely because the minimum pain frequency rose from 2–3 days/month to 1 day/week (Palsson 2020, PMID 31917991). Rome V criteria were published in May 2026 (Drossman 2026, PMID 42031435); the first accuracy study found they identify a different group again — sensitivity 66.1% (95% CI 62.1–69.9), specificity 80.1% (72.4–86.5), agreement with Rome III/IV only fair-to-moderate (κ 0.36–0.55) (Staller 2026, PMID 42392123) — while the Rome V global survey found overall DGBI prevalence essentially unchanged from Rome IV (40.9% vs 40.5%) with IBS at 8.5% (Sperber 2026, PMID 42613194). Every criteria set, from Manning onward, has performed only modestly against a clinical reference standard (Rome III positive LR 3.35, negative LR 0.39; Sood 2015, PMID 26076071), and a positive diagnosis with limited testing is non-inferior to exhaustive exclusion over five years (Begtrup 2013, PMID 23357491; Engsbro 2021, PMID 33029843). Subtyping by stool form is the most stable classification available and is still only moderately so (κ 0.60 over 12 months; Khasawneh 2026, PMID 41447016), and self-reported Bristol scores correlate weakly with measured stool water content (r = 0.36; Nordin 2022, PMID 36087104). When reading any effect estimate in this knowledge base, check which criteria defined the population — the populations are not interchangeable.

Why the definition is the subject, not the preamble

In most conditions the diagnostic criteria are an administrative layer over a disease that exists independently. In IBS the criteria are constitutive: there is no reference standard other than "characteristic symptoms plus absence of organic disease after investigation," which is the definition used as the comparator in the accuracy studies themselves (Staller 2026, PMID 42392123; Goodoory 2025, PMID 39466700). Consequences that propagate through the rest of this knowledge base:

  1. Prevalence is a property of the rule. See epidemiology-and-burden.
  2. Trial populations are not directly comparable across criteria eras. Rifaximin's registration trials used Rome II (Pimentel 2011, PMID 21208106); ATLANTIS used Rome IV (Ford 2023, PMID 37858323); most low-FODMAP trials used Rome III. Pooling them assumes criteria-invariance of treatment effect; targeted PubMed and ClinicalTrials.gov searches on 2026-09-02 retrieved no within-population treatment analysis across Rome III, IV and V — see OPEN-QUESTIONS.md OQ-1.
  3. "Sensitivity" of a criteria set is partly circular, because the reference standard is itself symptom-based; Whitehead and Drossman argued this explicitly, noting that studies discriminating IBS from structural disease "cannot provide meaningful information on the sensitivity or positive predictive value because IBS is defined only by exclusion of structural disease" (Whitehead 2010, PMID 20179688).

The criteria, in sequence

Set Year Core requirement Key change Reference
Manning 1978 ≥2 of 6 symptoms (distension; pain relieved by defecation; looser and more frequent stools with pain onset; mucus; incomplete evacuation) First attempt at a positive symptom diagnosis rather than exclusion; derived in 109 patients, 32 with IBS and 33 with organic disease at 17–26 months Manning 1978, PMID 698649
Kruis 1984 Weighted score from history, examination, ESR, blood count Explicitly incorporates tests into the score; at sensitivity 64% specificity was 99%, at sensitivity 83% specificity 97% (n=479 outpatients, 108 IBS, 209 organic) Kruis 1984, PMID 6724251
Rome I 1992/1994 Symptom cluster over ≥3 months First consensus-committee criteria (process described in Drossman 1999, PMID 10457038)
Rome II 1999 Abdominal discomfort/pain ≥12 weeks in 12 months + 2 of 3 features Duration threshold formalised Thompson 1999, PMID 10457044
Rome III 2006 Recurrent abdominal pain or discomfort ≥3 days/month for 3 months + 2 of 3 (relief with defecation; onset with change in stool frequency; onset with change in stool form) Subtyping moved to stool form (Bristol scale) rather than "predominance" Longstreth 2006, PMID 16678561
Rome IV 2016 Recurrent abdominal pain only ≥1 day/week for 3 months + 2 of 3 "Discomfort" dropped (untranslatable, ambiguous); frequency quadrupled; "functional" de-emphasised in favour of disorders of gut–brain interaction Drossman 2016, PMID 27144617; Palsson 2016, PMID 27144634
Rome V 2026 Published May 2026 as part of the Rome V process Updated definitions, diet and microenvironment content, updated algorithms, two new DGBI (abdominal migraine; inability to belch syndrome) Drossman 2026, PMID 42031435; Sperber 2026, PMID 42613194

The Rome IV rationale for dropping "discomfort" was that the term is ambiguous and has no equivalent in some languages; the frequency threshold rose from 2–3 days/month to ≥1 day/week (Goodoory 2024, PMID 38423348).

What each rewrite did to the denominator

  • Rome III → Rome IV halved community IBS. In 5,931 adults across the USA, Canada and UK, Rome IV IBS prevalence was 4.4–4.8% by country versus 9.0% by Rome III overall (4.6% vs 9.0%), "primarily due to higher Rome IV minimum pain frequency." Functional constipation and functional diarrhoea became more prevalent by Rome IV, i.e. patients were reclassified rather than lost (Palsson 2020, PMID 31917991).
  • In people who self-identify as having IBS, 89% of those meeting Rome III but not Rome IV failed Rome IV specifically on the pain-frequency change (Goodoory 2024, PMID 38423348).
  • Rome IV → Rome V left the aggregate almost unchanged. In a 28,771-adult internet survey across 15 countries, ≥1 DGBI was present in 40.9% under Rome V versus 40.5% under Rome IV; IBS was 8.5%, behind unclassified bowel disorder (9.6%) and chronic constipation (8.9%). The two new categories were abdominal migraine (5.1%) and inability to belch syndrome (1.4%) (Sperber 2026, PMID 42613194).
  • But aggregate stability is not individual stability. In secondary care, agreement between Rome V and either predecessor was only fair-to-moderate (κ 0.36–0.55) — the same headline number is being generated from partly different people (Staller 2026, PMID 42392123).

Measured accuracy of each criteria set

All figures below are against a symptom-plus-negative-investigation reference standard, in secondary care unless stated.

Criteria Population Sensitivity Specificity LR+ LR− Source
Manning (≥3) 361 Mayo outpatients vs organic GI disease 58% 74% — — Talley 1990, PMID 2318433
Manning (≥3) vs all non-IBS GI disease 42% 85% — — Talley 1990, PMID 2318433
Rome III (pooled) 22 studies, 7,106 patients — — 3.35 (2.97–3.79) 0.39 (0.34–0.46) Sood 2015, PMID 26076071
Rome III Leeds, n=726 87.5% (84.4–90.3) 75.0% (66.3–82.4) 3.50 (2.61–4.84) 0.17 (0.13–0.21) Staller 2026, PMID 42392123
Rome IV Leeds, n=726 78.9% (75.4–82.1) 81.0% (73.4–87.2) 4.16 (2.98–5.95) 0.26 (0.22–0.31) Staller 2026, PMID 42392123
Rome IV Leeds, n=170 82.1% 85.1% 5.51 (2.95–11.3) 0.21 (0.14–0.31) Goodoory 2025, PMID 39466700
Rome IV, pain frequency relaxed to 3 days/month Leeds, n=170 90.2% 85.1% 6.06 (3.25–12.2) 0.11 (0.07–0.19) Goodoory 2025, PMID 39466700
Rome V Leeds, n=726 66.1% (62.1–69.9) 80.1% (72.4–86.5) 3.33 (2.40–4.74) 0.42 (0.37–0.49) Staller 2026, PMID 42392123

Two things are visible in that table. First, no set has ever achieved a positive likelihood ratio above ~6 — the criteria shift probability, they do not settle it. Second, the newest set performs worse on sensitivity than either predecessor in the only study to have tested it, and its authors state plainly that "the clinical relevance of this is uncertain" (Staller 2026, PMID 42392123). A simple modification — keeping pain-only but relaxing frequency back to 3 days/month — outperformed unmodified Rome IV on both sensitivity and likelihood ratios (Goodoory 2025, PMID 39466700), which suggests the Rome IV frequency threshold was set tighter than the data support.

Adding biomarkers has not rescued accuracy: across 22 studies, eleven candidate biomarkers "performed no better than symptom-based criteria," and the best combinations (faecal calprotectin + intestinal permeability + Rome I: LR+ 26.4, 11.4–61.9; serum biomarkers + psychological markers: LR− 0.18, 0.12–0.25) have not been replicated into practice (Sood 2015, PMID 26076071). In primary care, symptom criteria alone cannot exclude organic disease — performance across 25 primary diagnostic studies was "highly variable" — but the low pre-test probability in that setting "argues against exhaustive diagnostic evaluation" (Jellema 2009, PMID 19575763).

Positive diagnosis versus diagnosis of exclusion

The one randomised test of the strategy question compared a positive diagnostic strategy (limited blood tests) against exclusion (extensive blood tests, faecal parasites, sigmoidoscopy with biopsies) in 302 primary-care patients aged 18–50 meeting Rome III criteria without alarm signals. The positive strategy was non-inferior at one year (Begtrup 2013, PMID 23357491). At five-year registry follow-up: no coeliac disease and no gastrointestinal or gynaecological cancers in either arm, negligible and comparable IBD rates, and the positive-diagnosis arm underwent more lower endoscopies during years 1–5 (23 vs 13 patients) yet still saved endoscopies overall (Engsbro 2021, PMID 33029843). This is the empirical basis for the "positive diagnosis" recommendation that now appears in ACG, BSG and AGA documents (guidelines) — and it is one modestly sized Danish trial in patients under 50. The corresponding yield data are on differential-diagnosis-and-exclusion.

Subtyping: stool form, and how badly it holds

Rome III moved subtyping to stool form via the Bristol Stool Form Scale, calling subtyping "controversial" even as it did so (Longstreth 2006, PMID 16678561). Three independent problems have accumulated since:

  • Criteria-set dependence. Rome II and Rome III subtyping agreed in only 46% of 249 patients (κ 0.19). Direct constipation↔diarrhoea switches were rare (8%); nearly all disagreement involved the alternating/mixed/unsubtyped categories (Ersryd 2007, PMID 17767480).
  • Instrument validity. Self-reported BSFS correlates with measured stool water content at only r = 0.36 (p<0.0001), with heavy overlap between adjacent categories: of stools scored BSFS 6–7 ("loose"), only 52% had water content ≥78% (Nordin 2022, PMID 36087104). Physician- and patient-assigned BSFS agree hardly at all in a head-to-head comparison (κ = −0.01 across 7 points, κ = 0.04 collapsed to 3 categories; concordant in 28% of 64 cases) (Engel 2026, PMID 42289094).
  • Temporal instability. Over 12 months in 352 UK registry participants, stool-form subtyping was the most stable of four systems (κ 0.60; 83% of IBS-D stable), followed by a psychological-burden classification (κ 0.54), most-troublesome-symptom (κ 0.47) and a seven-cluster latent-class model (κ 0.37). Only 73.6% still met Rome IV criteria at all (Khasawneh 2026, PMID 41447016).

Subtype distribution also depends on which criteria are applied: pooling 96 studies across 52 countries, IBS-D was commoner under Rome III (26.2%) and IBS-C under Rome IV (34.2%) (Arif 2025, PMID 40359286).

The competing classification: latent classes, not bowel habit

A Leeds programme has argued that bowel habit is the wrong axis and that IBS is better cut by joint gastrointestinal-severity and psychological burden:

Study Population Finding
Black 2022, PMID 35531932 1,375 community adults self-identifying as IBS; 558 met Rome IV criteria for a non-IBS functional bowel disorder Five clusters found, correlating poorly with the Rome IV diagnostic labels; 75% classified as "mild IBS" by the IBS-derived model; one-third fluctuated into IBS within 12 months. Conclusion: the functional bowel disorders may be "a spectrum of IBS rather than separate disorders"
Black 2024, PMID 36858142 752 Rome IV IBS Seven-cluster model; the four highest-psychological-burden clusters had worse quality of life, lower earnings, more drugs, and >£1,000/person/year IBS-related costs in cluster 6
Black 2025, PMID 38876193 2,195 Rome IV IBS in the Rome Foundation Global Epidemiology Study All seven clusters reproduced globally; a de novo model produced 10 clusters, ≥2 apparently duplicates, almost all mapping onto the original seven
Goodoory 2026, PMID 41714311 379 secondary/tertiary-care referrals (249 Rome IV positive) Higher-psychological-burden clusters had more severe IBS-SSS, worse QoL, lower discharge rates, and more IBS drugs prescribed over 12 months

Independent factor analysis of the Rome IV criteria across 26 countries supports the construct's global coherence at the item level (Hreinsson 2023, PMID 36889555). The unresolved tension is that the latent-class system predicts burden and utilisation better than stool form does, while stool form is more stable over time and is the axis on which every drug is licensed (Khasawneh 2026, PMID 41447016).

What the criteria are used for, and where that breaks

Use Criteria dependence Consequence
Population prevalence Very high Rome III → IV halved measured IBS (Palsson 2020, PMID 31917991)
Trial eligibility Very high Rome IV populations are smaller, more pain-dominant; Rome III trial estimates may not transfer
Drug labelling Subtype-dependent Agents are licensed for IBS-C or IBS-D, a split that is only κ 0.60 stable at 12 months (Khasawneh 2026, PMID 41447016)
Clinical diagnosis Moderate Clinicians frequently diagnose IBS in patients failing formal criteria; in the Leeds accuracy study 590/726 met the clinical reference standard but only 417 met Rome V (Staller 2026, PMID 42392123)
Excluding organic disease Low Criteria alone cannot do this (Jellema 2009, PMID 19575763); see differential-diagnosis-and-exclusion

Open questions

  • Are treatment effects criteria-dependent? Targeted PubMed and ClinicalTrials.gov searches on 2026-09-02 retrieved no trial stratifying an IBS drug or diet effect by Rome III versus Rome IV versus Rome V status within the same population, despite the criteria selecting demonstrably different people (Staller 2026, PMID 42392123; Palsson 2020, PMID 31917991).
  • Should Rome IV's pain-frequency threshold be relaxed? A single-centre modification study says yes on accuracy grounds (sensitivity 90.2% vs 82.1% at identical specificity; Goodoory 2025, PMID 39466700) and Rome V did not adopt it.
  • Why did Rome V's sensitivity fall to 66.1% in its first accuracy test, and does that replicate outside a single UK specialist clinic (Staller 2026, PMID 42392123)?
  • Is the latent-class/psychological-burden classification clinically actionable, or only prognostic? It predicts cost and utilisation (Black 2024, PMID 36858142; Goodoory 2026, PMID 41714311); a targeted PubMed search on 2026-09-02 retrieved no trial randomising treatment by cluster.
  • Should the non-IBS functional bowel disorders be abolished into an IBS spectrum, as latent-class data suggest (Black 2022, PMID 35531932)? Rome V retained them (Sperber 2026, PMID 42613194).
  • Can any subtyping instrument be made objective? BSFS self-report tracks stool water content at r = 0.36 and physician report barely at all (Nordin 2022, PMID 36087104; Engel 2026, PMID 42289094).

References

  1. Manning AP, Thompson WG, Heaton KW, Morris AF. Towards positive diagnosis of the irritable bowel. Br Med J. 1978;2(6138):653-4. PMID 698649
  2. Kruis W, Thieme C, Weinzierl M, Schüssler P, Holl J, Paulus W. A diagnostic score for the irritable bowel syndrome. Its value in the exclusion of organic disease. Gastroenterology. 1984;87(1):1-7. PMID 6724251
  3. Drossman DA. The functional gastrointestinal disorders and the Rome II process. Gut. 1999;45 Suppl 2:II1-5. PMID 10457038
  4. Thompson WG, et al. Functional bowel disorders and functional abdominal pain. Gut. 1999;45 Suppl 2:II43-7. PMID 10457044
  5. Longstreth GF, Thompson WG, Chey WD, Houghton LA, Mearin F, Spiller RC. Functional bowel disorders. Gastroenterology. 2006;130(5):1480-91. PMID 16678561
  6. Drossman DA. Functional Gastrointestinal Disorders: History, Pathophysiology, Clinical Features and Rome IV. Gastroenterology. 2016. PMID 27144617
  7. Palsson OS, et al. Rome IV Diagnostic Questionnaires and Tables for Investigators and Clinicians. Gastroenterology. 2016. PMID 27144634
  8. Drossman DA, Chang L, Tack J. Disorders of Gut-Brain Interaction and the Rome V Process. Gastroenterology. 2026;170(6):1083-1098. PMID 42031435
  9. Sperber AD, et al. Rome V global epidemiology and validation survey. Gut. 2026 Aug 18 (online ahead of print). PMID 42613194
  10. Staller K, Goodoory VC, Khasawneh M, Black CJ, Ford AC. The Rome V criteria for the diagnosis of irritable bowel syndrome in secondary care: a diagnostic accuracy study. Lancet Gastroenterol Hepatol. 2026;11(9):812-820. PMID 42392123
  11. Palsson OS, Whitehead W, Törnblom H, Sperber AD, Simren M. Prevalence of Rome IV Functional Bowel Disorders Among Adults in the United States, Canada, and the United Kingdom. Gastroenterology. 2020;158(5):1262-1273.e3. PMID 31917991
  12. Goodoory VC, Khasawneh M, Black CJ, Ford AC. Assessing Diagnostic Performance of Modifications to the Rome IV Criteria for Irritable Bowel Syndrome. Clin Gastroenterol Hepatol. 2024;22(9):1942-1943. PMID 38423348
  13. Goodoory VC, Riggott C, Khasawneh M, Black CJ, Ford AC. Validating Simple Modifications to the Rome IV Criteria for the Diagnosis of Irritable Bowel Syndrome in Secondary Care. Aliment Pharmacol Ther. 2025;61(2):354-362. PMID 39466700
  14. Sood R, Gracie DJ, Law GR, Ford AC. Systematic review with meta-analysis: the accuracy of diagnosing irritable bowel syndrome with symptoms, biomarkers and/or psychological markers. Aliment Pharmacol Ther. 2015;42(5):491-503. PMID 26076071
  15. Jellema P, van der Windt DA, Schellevis FG, van der Horst HE. Systematic review: accuracy of symptom-based criteria for diagnosis of irritable bowel syndrome in primary care. Aliment Pharmacol Ther. 2009;30(7):695-706. PMID 19575763
  16. Whitehead WE, Drossman DA. Validation of symptom-based diagnostic criteria for irritable bowel syndrome: a critical review. Am J Gastroenterol. 2010;105(4):814-20. PMID 20179688
  17. Talley NJ, Phillips SF, Melton LJ, Mulvihill C, Wiltgen C, Zinsmeister AR. Diagnostic value of the Manning criteria in irritable bowel syndrome. Gut. 1990;31(1):77-81. PMID 2318433
  18. Begtrup LM, et al. A positive diagnostic strategy is noninferior to a strategy of exclusion for patients with irritable bowel syndrome. Clin Gastroenterol Hepatol. 2013;11(8):956-62.e1. PMID 23357491
  19. Engsbro AL, et al. A positive diagnostic strategy is safe and saves endoscopies in patients with irritable bowel syndrome: A five-year follow-up of a randomized controlled trial. Neurogastroenterol Motil. 2021;33(3):e14004. PMID 33029843
  20. Ersryd A, Posserud I, Abrahamsson H, Simrén M. Subtyping the irritable bowel syndrome by predominant bowel habit: Rome II versus Rome III. Aliment Pharmacol Ther. 2007;26(6):953-61. PMID 17767480
  21. Nordin E, Hellström PM, Brunius C, Landberg R. Modest Conformity Between Self-Reporting of Bristol Stool Form and Fecal Consistency Measured by Stool Water Content in Irritable Bowel Syndrome and a FODMAP and Gluten Trial. Am J Gastroenterol. 2022;117(10):1668-1674. PMID 36087104
  22. Engel T, et al. Physician-Patient Discordance in Bristol Stool Scale Classification and Overall IBS Burden Assessment. Am J Gastroenterol. 2026 Jun 9 (online ahead of print). PMID 42289094
  23. Khasawneh M, Goodoory VC, Ng CE, Ford AC, Black CJ. Stability of Classification Systems for Irritable Bowel Syndrome. Aliment Pharmacol Ther. 2026;63(8):1132-1139. PMID 41447016
  24. Black CJ, Houghton LA, Ford AC. Latent class analysis does not support the existence of Rome IV functional bowel disorders as discrete entities. Neurogastroenterol Motil. 2022;34(11):e14391. PMID 35531932
  25. Black CJ, Ng CE, Goodoory VC, Ford AC. Novel Symptom Subgroups in Individuals With Irritable Bowel Syndrome Predict Disease Impact and Burden. Clin Gastroenterol Hepatol. 2024;22(2):386-396.e10. PMID 36858142
  26. Black CJ, et al. Novel Irritable Bowel Syndrome Subgroups Are Reproducible in the Global Adult Population. Clin Gastroenterol Hepatol. 2025;23(6):1039-1048.e7. PMID 38876193
  27. Goodoory VC, et al. Novel Symptom Subgroups in Patients With Irritable Bowel Syndrome Are Associated With Healthcare Utilisation in Secondary and Tertiary Care. Aliment Pharmacol Ther. 2026;63(11):1474-1486. PMID 41714311
  28. Hreinsson JP, et al. Factor Analysis of the Rome IV Criteria for Major Disorders of Gut-Brain Interaction (DGBI) Globally and Across Geographical, Sex, and Age Groups. Gastroenterology. 2023;164(7):1211-1222. PMID 36889555
  29. Arif TB, et al. Global prevalence and risk factors of irritable bowel syndrome from 2006 to 2024 using the Rome III and IV criteria: a meta-analysis. Eur J Gastroenterol Hepatol. 2025;37(12):1314-1325. PMID 40359286
  30. Ford AC, et al. Amitriptyline at Low-Dose and Titrated for Irritable Bowel Syndrome as Second-Line Treatment in primary care (ATLANTIS). Lancet. 2023;402(10414):1773-1785. PMID 37858323
  31. Pimentel M, et al. Rifaximin therapy for patients with irritable bowel syndrome without constipation. N Engl J Med. 2011;364(1):22-32. PMID 21208106