To the Editor:
Salerno et al. compared the efficacy of transjugular intrahepatic portosystemic stent shunt (TIPS) versus large-volume paracentesis for the treatment of refractory ascites in the September 2004 issue of HEPATOLOGY.1 In contrast to two other recent trials,2, 3 they reported a significant improvement in survival in patients who received TIPS compared with paracentesis.
Eleven of 33 subjects were crossed over from paracentesis to TIPS because of failure of treatment. We believe that in a small intent-to-treat study, such a crossover can skew the data. As part of their sensitivity analysis, it would be helpful for the authors to confirm that mortality resulting from patients who crossed over from paracentesis to TIPS did not confound their results. Also, the time interval between initial randomization and crossover was not reported. Early crossover would limit the subjects' exposure to paracentesis, which would effectively make them TIPS patients.
It is not clear whether the subjects truly had refractory ascites or whether the fluid overload was a result of dietary noncompliance. We would like to know if urine electrolytes were applied to confirm diet compliance before concluding failure of treatment. The frequency of hepatic encephalopathy was 61% in TIPS patients compared with 39% in paracentesis patients. This appears to be a large difference clinically, even though it is not statistically significant. Finally, although the mean follow-up time was longer in the TIPS group than in the paracentesis group, twice as many TIPS patients were censored, most during the initial 2 years.
Before we seriously consider using a more invasive procedure such as TIPS in patients with ascites, clarification of these issues is important. Not only can TIPS be more expensive, it can potentially add the risks of technical complications, liver failure, and hepatic encephaolopathy.3