Transient glucose intolerance (TGI) is an important cause of WHO G-ORS (Glucose-Oral Rehydration Solution) treatment failure and hospitalization in dehydrated children during acute diarrhoea. This retrospective case-control study was designed to determine some risk and predictive factors for developing TGI among moderately dehydrated patients with acute diarrhoea while under G-ORS therapy. Among moderately dehydrated patients, files of 22 patients with TGI and 66 other dehydrated patients without intolerance were reviewed. Patients with TGI were younger (9.7 +/- 10.5 mo and 11.6 +/- 7.8 mo, respectively, p < 0.05), the median age being 6 mo in the TGI group and 10 mo in non-TGI group. There was no difference between groups for sex, admission season, history of fever or vomiting, frequency of vomiting, presence of blood, mucous or leukocyte in stool, presence of associated disease and duration of diarrhoea on admission. The admission haemoglobin, white blood cell, blood pH, sodium and potassium levels were similar in both groups. The mean serum chloride level (116.8 +/- 6.9, 109.6 +/- 7.9 mEq/l, respectively, p < 0.05) was higher in the TGI group and the bicarbonate level was lower (12.9 +/- 3.8, 15.3 +/- 6.0, respectively, p < 0.05). Stool frequency was also higher in patients with TGI (11.2 +/- 5.3/24 h, 5.9 +/- 4.4/24 h, respectively, p < 0.05). No difference was found between the nutritional status of children in both groups. Mon children were breastfed in the group without TGI (34/56, 61%, 6/18, 33% respectively, p = 0.079, OR = 0.32, 95% CI [0.09-1.11]). It was concluded that patients with TGI are younger and have high stool frequency. Although statistical significance could not be shown, breastfeeding seems to protect children from TGI, as it protects from diarrhoea.