Subtyping the Irritable Bowel Syndrome by Predominant Bowel Habit: Rome II versus Rome III

A. Ersryd; I. Posserud; H. Abrahamsson; M. Simrén


Aliment Pharmacol Ther. 2007;26(6):953-961. 

In This Article


In the present study we have demonstrated poor agreement between dividing IBS patients into subgroups based on recommendations in the Rome II versus the newly presented Rome III criteria. This finding is of great importance when comparing clinical trials that in the past have included patients based on retrospective recall of supporting symptoms, including stool form and frequency as well as straining and urgency, as proposed by Rome II committee,[6,11] with forthcoming trials that will use the Rome III recommendations, i.e. prospectively assessing the stool form, subdividing patients based on this information.[1,19]

One plausible explanation for the discrepancy between Rome II and Rome III subgrouping might be the way the information was collected from the patients, i.e. in a retrospective way when determining Rome II subgroups versus prospectively using diary cards for Rome III subgroups. There are now several lines of evidence supporting the superiority of diary cards over assessment of bowel habits and GI symptoms retrospectively,[21] especially when evaluating symptoms in IBS patients.[18,22,23] Recall bias is a well known problem in medical research and it might be particularly problematic in IBS patients who have been proposed to have a peculiar confirmatory bias for negative material.[24] This might lead to bias when grading their symptoms retrospectively supported by the high frequency of positive symptom reports on the Rome II Modular Questionnaire in our study versus the much lower frequency on the diary cards. For instance loose, mushy or watery stools more than one-fourth of the time was reported by 77% of the patients, whereas a clearly smaller proportion of subjects had this stool consistency on the BSF diary cards. The same holds true when comparing reports of hard stools, as well as the retrospective versus prospective assessment of abnormal stool frequency. Therefore, subgrouping based on prospective diary cards seems to be more reliable than retrospective symptom recall.

Another possible explanation for our divergent results between Rome II and Rome III subgrouping might relate to the fact that the patients only completed BSF diary cards during 1 week, as compared with the retrospective symptom recall for the previous 3 months. It is well-known that colonic function varies widely, not only in IBS patients,[15,25,26] but also in healthy subjects.[27] One might argue that completing the BSF diary cards for only 1 week is too short, but the optimal time period for this still needs to be defined. The Rome III committee only proposes researchers to use the BSF Scale for subgrouping, but provides no guidance on the optimal time period.[1] In previous studies using diary cards in IBS patients, time periods used ranged from 1 to 6 weeks.[13,15,17,18,22,23,25] In one of these studies, two different IBS samples completed diary cards for 1 and 6 weeks, respectively, with similar subgroups appearing in both samples.[18] Moreover, another recent study came to similar figures as we did when evaluating the agreement between Rome II subgrouping based on retrospective symptom recall and Rome III subgroups based on two-weeks BSF diary card.[25] These findings, to some extent, speak against the short time period used for the completion of the diary cards being the explanation for the discrepancy between Rome II and Rome III subgroups in our study.

Previous studies,[18,22,23] as well as our present study, have demonstrated that the majority of IBS patients have a normal frequency of bowel movements. Moreover, as reports of straining, urgency and incomplete evacuation appear across the spectrum of stool forms in IBS patients,[15,22,23] the Rome III criteria proposed to leave these defecatory symptoms and stool frequency out when subgrouping IBS patients.[1] On the other hand, stool consistency, which reflects colonic transit time,[19] has according to some recent studies been suggested to be the most specific way to subgroup IBS patients,[14,22,23] which was adopted by the Rome III committee.[1] It seems logical that this shift in the definition of the subgroups is the main explanation between the poor agreement between Rome II and Rome III subgroups. A general tendency of IBS patients to report several symptoms[28] probably influences the subgrouping and especially when reporting them retrospectively as stated previously.

The majority of previous studies have found a-IBS to be the most common subgroup,[29] especially when the investigation is primary care office-based.[26] We and others have demonstrated that the presence of diarrhoea is an explanation for referral from primary care to gastroenterologists,[30,31] which might explain higher proportion of d-IBS in some studies from gastroenterology offices.[32,33] In line with the majority of previous studies we found a-IBS to be the most common Rome II subgroup, whereas this was different with Rome III, where IBS-C and IBS-D were the most common subgroups. Again, this has important implications when reading and planning studies using Rome II versus Rome III subgroups.

As in other studies female sex was associated with constipation and the male sex with diarrhoea when using Rome II for subtyping.[33,34,35] With the Rome III criteria on the other hand, there was no significant gender difference between the subtypes. This was a surprising finding and at this stage we do not have any clear explanation for this. A general tendency in females to report retrospective symptoms differently from males might be an explanation, but to the best of our knowledge no evidence for this exists in the literature. The female IBS patients in our study tended to have lower frequency of bowel movements than men, and harder stools, more bloating and flatulence, but these differences were not major. Further studies are needed to elucidate the relationship between gender and symptom pattern in IBS using the Rome III criteria and prospective versus retrospective symptom assessment.

As expected there was a difference in frequency and stool form between the subtypes using Rome II. There was also a significant difference in bloating between IBS-C and IBS-D. In line with previous studies, bloating was more common in constipated patients.[33,36] We also demonstrated more severe discomfort in the patients with Rome III IBS-M, in line with a study by Tillisch et al., where the alternating type reported greater severity of pain.[14] However, besides obvious and expected differences in stool form and frequency, the differences in other symptoms must be considered to be minor, and the similarities between the subgroups were clearer than the differences, which has been reported before.[13]

The interest in subtyping IBS patients has grown stronger as new drugs have been introduced on the market. Their beneficial effects are expected only in different subgroups. Alosetron, a selective serotonin 5-HT3 receptor antagonist, can decrease pain, urgency and stool frequency in patients with IBS and diarrhoea and has been studied in Rome II d-IBS.[8,9] Tegaserod a partial 5-HT4 receptor agonist, has improved stool frequency and form, abdominal pain/discomfort and bloating in women with IBS and constipation, and has been studied in Rome II c-IBS,[10] but has unfortunately recently been withdrawn from the US market due to cardiac side effects.[37] Whether these results can be translated to Rome III subgroups remain to be proven and should be evaluated in future studies. From our investigation it is obvious that a somewhat different patient population would have been included in these studies if Rome III subgroups based on stool form diaries had been used instead of Rome II subgroups based on retrospective symptom reporting.

One potential limitation with our study is that it is not population-based. The patients were recruited either by advertisement in the newspaper or from the gastroenterology outpatient clinic at Sahlgrenska University Hospital. This could have an impact on the subtypes that were presented and how they rated their symptoms. On the other hand, the main purpose of this study was not to assess prevalence figures, but agreement between the different criteria in the same patient. Therefore, for this purpose we do not have any reason to believe that this would have affected our main results.

To conclude, subgrouping of IBS patients based on recommendations from the Rome II and Rome III committee differ substantially. Probably, this is due to the exclusion of defecatory symptoms and stool frequency in Rome III, as well as the difference how the symptom evaluation is performed, i.e. retrospective symptom assessment with Rome II and the use of prospective stool form diary card with Rome III. It is important to take these discrepancies into consideration when comparing studies conducted before versus after the creation of the Rome III criteria.

CLICK HERE for subscription information about this journal.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: