The reproducibility of the oral glucose tolerance test over long (5 years) and short periods (1 week)

Summary Three oral glucose tolerance tests (oGTT) have been performed in 312 non-diabetic relatives of diabetics over a period of 10 years. In a second study 6 identical oGTT's have been performed at weekly intervals in 55 individuals. In this study the variance, calculated from the logarithmic values, increased in the following order: fasting (0.026), 1 h (0.035), 2 h (0.044) and 3 h values (0.047). The sum of the 1 h and 2 h values showed the lowest variance (0.024). No significant difference of the variances was found in the 43 individuals in whom both the long-term and the short-term studies have been performed. Thus, a great proportion of the total variance of glucose observed over longer periods only represents a random variation. This random variation is much higher than most other factors which might influence the result of an oGTT. A diagnosis based on a single oGTT is of only limited value.Supported by a grant from Deutsche Forschungsgemeinschaft (Ko 457/8)