Poor reproducibility of the 75-g oral glucose tolerance test (75-g OGTT) is a major criticism. To test whether dilution might be a contributor, we studied the effect of administering it at 300, 600, and 900 ml on the coefficient of variation (CV) and overall tolerability. Each dilution was randomly repeated three times by 35 subjects: 11 lean (body fat [BF]: 11.5+/-1.5%, body mass index (BMI): 24.1+/-0.8 kg/m(2), age: 33+/-3 years), 12 normal (BF: 24.3+/-1.3%, BMI: 23.9+/-0.7 kg/m2, age: 33+/-3 years), and 12 obese (BF: 34.1+/-1.5%, BMI: 34.5+/-1.6 kg/m(2), age: 41+/-4 years). The protocol followed, the American Diabetes Association (ADA) guidelines with venous samples drawn at 0, 15, 30, 45, 60, 90, and 120 min. Scales assessing palatability, acceptability, satiety, nausea, and dizziness were also completed. No differences were detected in CV between the three dilutions at the 2 h-diagnostic-endpoint in any group. CV for glycemia, however, was lower for the 300-ml OGTT (10+/-2.1%) than either the 600 ml (17.9+/-2.1%) or 900-ml OGTT (19.9+/-4.4%) at 45 min in the lean group (P<0.05). CV for insulinemia was lower for the 300 ml (30.6+/-4.5%) and 600 ml (30.6+/-4.1%) OGTT than the 900-ml OGTT (53.7+/-7.9%) at 60 min in the obese group (P<0.05). When affective ratings were compared, the 600-ml OGTT scored highest on palatability and acceptability (P<0.05). We concluded that the dilution of the 75-g OGTT improves the overall tolerability but not reproducibility of results. Practitioners may choose to give it at the best-tolerated dilution.