Background In irritable bowel syndrome (IBS) subtyping is used in research and clinical practice. Knowledge of subtype stability is needed for proper design of trials and treatment strategies. Aims To evaluate the stability of Rome III IBS subtypes over time and to determine the optimal time period for prospective, diary-based subtyping. Methods Rome III IBS patients aged 18-70 years enrolled in two identical, randomised, placebo-controlled trials of probiotics, were included. No difference was found on stool pattern, thus patients were analysed as one group. Patients scored defaecations according to Bristol Stool Form Scale for 10 weeks. IBS subtypes were determined for all 1- and 2-week periods. Subtype distribution and stool pattern over time were determined. The proportions of patients having the same subtype all weeks (stable patients) or having a predominant subtype (same subtype ≥60% of time) were determined. Results A total of 126 patients, mean age 46 ± 15 years, 72% women were included. Subtype distribution was similar over time with IBS with constipation, IBS with diarrhoea and IBS unsubtyped constituting one-third of the population each. Even though only 18-35% had the same subtype all weeks, the majority of patients had the same subtype for ≥60% of time (82-98%). Sixty-nine per cent had the same predominant and baseline subtypes. Two-week data increased the proportion of stable patients, of patients with a predominant subtype, and of patients who had similar baseline and predominant subtype. Conclusions Most IBS patients change subtype over time. However, an underlying stool pattern stability was demonstrated in the majority of patients. To increase stability, we recommend 2-week data for IBS subtyping.