The split renal function, obtained using Technetium 99m mercaptoacetyltriglycine renography, is currently considered as a robust, accurate and reproducible parameter, provided that it is calculated at a time when no escape of the tracer from the kidney has occurred. The question arises as to whether the simultaneous administration of furosemide with the tracer (F0 test) might accelerate the escape of the tracer, resulting in an underestimation of the split renal function. From a large database, we selected 36 clinically stable children in whom both F0 and F+20 (administration of furosemide 20 min after the tracer) diuretic renography had been performed. In all cases, the F+20 test preceded the F0 test. The mean interval between the two tests was 17 months. The split renal function was calculated on the basis of the 1-2 min background-corrected renal activity using the integral method, slope method and Rutland-Patlak plot. In order to evaluate the effect on the split renal function of an early escape of the tracer, the patients were analysed according to the T(max) value of the renogram for both the F0 and F+20 tests. For the F+20 test, all T(max) values were more than 3 min. For the F0 test, the T(max) value was more than 3 min in 53 kidneys and less than 3 min in 19 kidneys (unilateral in all 19 patients). For the F0 test, for the kidneys with a T(max) value of less than 3 min, there was a tendency to underestimate the split renal function, taking as reference the split renal function observed in the F+20 test. This observation, however, was only statistically significant for the slope method (P=0.03). There was a tendency for lower values with the Rutland-Patlak plot (P=0.07), but, for the integral method, no difference was observed (P=0.5). In conclusion, the simultaneous administration of furosemide with the tracer induces an early acceleration of renal transit. This should be taken into account when calculating the split renal function, for example by favouring the integral method.