It is estimated in the United States that there are over one and a quarter million persons that are unaware that they have diabetes. Although diabetes detection drives have been carried out for over 14 years, the number of undiagnosed diabetics remains alarmingly high. Too frequently detection drives consist of simply distributing some method of discovering glycosuria but fail to follow through to see that the person receives a definitive diagnosis and treatment if necessary. Wilkerson and Krall's 1 studies have not only shown the incidence of diabetes, but their complete follow-up studies have demonstrated the high mortality rate of complications from this disease.
It has been known that the fasting blood-sugar determination is unreliable in detecting mild or asymptomatic diabetics. Fox et al., in this issue of The Journal, p. 622, found that the glucose-oxidase tests for determining glycosuria on the same specimen of urine was nearly twice as