Objective: To study the influence of parboiling and the severity of the process on glycaemic and insulinaemic responses to rice in type 2 diabetes. Moreover, to examine changes in starch structure related to parboiling, which may affect the metabolic responses and digestibility. Design: Nine type 2 diabetic subjects ingested four test meals: white bread (WB) and three meals of cooked polished rice of the same variety being non-parboiled (NP), mildly traditionally parboiled (TP) and severely pressure parboiled (PP). The participants ingested the test meals (50 g available carbohydrates) on separate occasions after an overnight fast. Setting: Outpatient clinic, Dept. Endocrinology and Metabolism, Aarhus University Hospital, Denmark. Results: All three rice samples elicited lower postprandial plasma glucose response (NP: 335±43; TP: 274±53; PP: 231±37 mmol/1*180 min.; means±s.e.m.) than white bread (626±80; P<0.001), within rice samples PP tended to be lower than NP (P=0.07). The glycaemic indices were: NP: 55±5, TP: 46±8 and PP: 39±6, and lower for PP than NP (P<0.05). The insulin responses were similar for the three rice meals, which were all lower than that to white bread (P<0.001). Differential scanning calorimetry showed the presence of amylose–lipid complexes in all rice samples and of retrograded amylopectin in PP. Amylose retrogradation was not detected in any of the rice samples. Conclusions: All rice test meals were low-glycaemic in type 2 diabetic subjects. There was no effect of TP on glycaemic index, whereas PP reduced the glycaemic index by almost 30% compared to NP. Sponsorship: The Royal Veterinary and Agricultural University, Aarhus University Hospital, Danish International Development Assistance (DANIDA), Ministry of Foreign Affairs and the ‘Konsul Johannes Fogh-Nielsens og Fru Ella Fogh-Nielsens Legat’ foundation. European Journal of Clinical Nurtition (2000) 54, 380–385