LAUSR

With great interest we read the article by Wu et al, published online ahead of print by JMRI. Wu et al studied the renal effects of x-ray contrast media in Sprague–Dawley (SD) rats by means of magnetic resonance imaging (MRI). They applied dynamic contrast-enhanced (DCE)-MRI followed by data processing according to the Toft twocompartment pharmacokinetic model and a deconvolution algorithm to obtain quantitative data on glomerular filtration rate (GFR), renal blood flow (RBF), and renal blood volume (RBV). These data, however, deviate strikingly from data obtained by “gold standard” methods, and are hard to reconcile with established paradigms of renal hemodynamics. Baseline data (pre-x-ray contrast) reported by Wu et al in their Table 2 were: mean values for RBF: 1.68 mL/min per g of kidney weight (kw); for GFR: 0.60 mL/min per g kw; and for RBV: 0.57 mL per g kw. Data obtained by gold standard methods such as ultrasound transit-time difference flowmeters (Transonic) and clearance techniques (clearance of inulin, creatinine, paraaminohippuric acid, etc.) greatly differ from these data. Thus, values for RBF in SD rats of comparable body weight and kidney weight are at least twice as high as reported by Wu et al and mostly in the range of 7–9 mL/min per g kw. Even though the GFR data given by Wu et al are rather low as compared to those obtained by clearance techniques, the filtration fraction (FF 5 GFR/effective renal plasma flow 5 GFR/RBF * (1 – hematocrit); reported hematocrit: 0.42) derived from their data would be about 62%. FF as determined by gold standard methods is 15– 30% in SD rats. If FF were indeed 62%, then the hematocrit of the blood in efferent arterioles would amount to 0.65. At such an excessive hematocrit the viscosity of the blood increases to an extent that would jeopardize its flow. What also casts doubt on the reliability of the data by Wu et al is that their RBF values (as depicted by color-coded pixel-wise maps in fig. 7a,a’) are quite homogenous across the renal layers. This is contrary to the long-established fact that cortical blood flow per tissue volume is much higher than medullary blood flow. It is just as unlikely that the data by Wu et al on RBV (0.57 mL/g), ie, an RBV fraction of 57% (at a density of about 1 g/mL are correct: the notion that more than half of the renal tissue is occupied by blood is obviously erroneous. Studies that utilized various techniques report RBV fractions of 15–30%. While we feel that the erroneous data at baseline also question Wu et al’s results following x-ray contrast media, as renal physiologists, our main concern is another: if uncontested, these published quantitative data on key variables of renal hemodynamics might be taken as reference by colleagues, eg, of the imaging community, who are not experts for kidney hemodynamics.