We have read with interest the paper by Guevara et al., published in a recent issue of HEPATOLOGY. 1 It is an elegant study that looked at the changes of renal and urinary parameters in 31 patients after contrast medium administration (protocol 1) and assessed the role of contrast medium as an etiology of exacerbating renal failure in another cohort comprised of 60 patients with cirrhosis (protocol 2). Contrary to general belief and our recent findings, 2,3 the authors concluded that contrast medium is not a risk factor for renal failure in patients with cirrhosis. It would be interesting to examine the methodology of this paper and the potential limitation of the authors' conclusions.
We have retrospectively investigated the occurrence of acute renal failure (ARF) in 235 patients with hepatocellular carcinoma undergoing transarterial chemoembolization, a major angiographic procedure. 2 The incidence of ARF was 6.6% in 843 consecutive treatment sessions, and the occurrence of ARF was closely associated with the severity of underlying cirrhosis and number of treatment sessions. These findings were confirmed in a prospective survey. 3 Importantly, we have shown that the risk of ARF after contrast medium administration is cumulative; the proportional increased risk was 30% to 65% for each additional embolization procedure in the same patient. 2,3 Compared with the study by Guevara et al., only 1 out of 60 patients with cirrhosis who had exacerbated renal failure had possible association with the use of contrast medium, whereas the other 2 patients who received contrast medium did not develop significant renal function deterioration. It should be noted that only 3 patients received contrast medium, and all of them underwent a diagnostic computed tomography scan. The amount of contrast medium given in this kind of procedure was quite small, and it is well known that the risk of renal failure is directly related to the amount of contrast medium. In addition, although the use of contrast medium contributed only to a minority of clinical scenarios in patients with cirrhosis who subsequently developed renal dysfunction, this does not necessarily support the hypothesis that contrast medium is not nephrotoxic in the setting of cirrhosis.
Another interesting finding but possible inappropriate interpretation by the authors is that of the patients in protocol 1, 2 had declined glomerular filtration rate (GFR) of more than 20% as well as decreased renal plasma flow (RPF) but no change in their serum creatinine levels. The overall difference for GFR and RPF between baseline and after contrast medium administration may not reach statistical significance because of a relatively low incidence of progressing renal failure in the study patients. Moreover, since GFR and RPF may well better represent true renal function, and serum creatinine level could be spuriously low in patients with advanced cirrhosis, these findings well reflect the fact that contrast medium may induce decreased renal perfusion.
In summary, the results from protocol 2 suggest that contrast mediumassociated renal dysfunction is an infrequent event in patients with cirrhosis, but the results do not support the finding that contrast medium does not predispose to nephropathy, whereas from protocol 1, decreased renal perfusion may indeed occur after contrast medium administration. A larger number of patients should be included to further justify the association between contrast medium and renal dysfunction in patients with cirrhosis.