In their recent study published in HEPATOLOGY, Bañares et al.1 reported that patients with cirrhosis and a transjugular portosystemic shunt (TIPS) may have an increased risk of hepatocellular carcinoma (HCC) development. As emphasized by the authors, several potential biases are evident in their study, including a selection bias due to the retrospective nature of the study.
At our institution, we have prospectively performed thin and uniform sectioning and pathological staging of all explanted livers with cirrhosis since January 2000 in order to be able to correlate preneoplastic lesions and HCCs with clinical, radiological, and other histopathological variables.2 We selected all patients who underwent transplantation because of end-stage cirrhosis and who were without pretransplant evidence of HCC.
Seven of these 76 patients (9%) had a small, well, or moderately differentiated HCC (diameter between 7 and 17 mm). Given known doubling times, it is very likely that these incidental HCCs developed in the months immediately preceding transplantation.3 The underlying etiology of cirrhosis in these 7 patients was viral (n = 4), alcoholic (n =1), metabolic (n = 1) and a combination of alcoholic and metabolic (n = 1). Several patients also had high-grade and low-grade dysplastic nodules (HGDNs and LGDNs) in their explanted liver. Eleven patients (14%) had a patent TIPS in their explanted liver that was placed at least 3 months before transplantation (mean 8 months, range 3-21 months). TIPS was not significantly associated with any of the known factors involved in hepatocarcinogenesis (gender, age, etiology), although TIPS tended to be placed more frequently in patients with postalcoholic cirrhosis. Assuming that TIPS significantly stimulates hepatocarcinogenesis, the presence of early, incidental HCC and the presence and number of preneoplastic DNs in the explanted livers with cirrhosis should be positively correlated with the presence of a TIPS.
However, neither the presence of HCC nor the presence and number of DNs was correlated with the presence of a TIPS (Table 1). Although the number of patients is much smaller than that in the study of Bañares et al.,1 our series have the advantage that they are prospective and therefore there are considerably less potential biases. One potential bias in our study is that the possible presence of HCCs and DNs before TIPS placement may give an overestimation of the role of TIPS in hepatocarcinogenesis. In view of this, it is striking that all 7 patients with HCC had no TIPS and that the presence and number of the different types of DN was about 2 to 2.5 times lower in patients with a TIPS (Table 1).
|Number (%) or Mean ± SD||P*|
|No TIPS (n = 65)||TIPS ≥3 Months (n = 11)|
|Age||55 ± 10||54 ± 8||.6641|
|Etiology of cirrhosis||.0738|
|Viral||20 (31)||2 (18)|
|Alcohol||22 (34)||8 (73)|
|Biliary||9 (14)||0 (0)|
|Other or combination||14 (21)||1 (9)|
|Incidental HCC**||7 (11)||0 (0)||.5844|
|DN||21 (32)||2 (18)||.4876|
|HGDN||12 (18)||1 (9)||.6777|
|LGDN||15 (23)||1 (9)||.4410|
|Number of DNs||1.3 ± 2.7||0.4 ± 0.9||.2265|
Our findings are in agreement with the two other prospective studies on this subject,4, 5 leading to the conclusion that there is no strong evidence that the creation of a portal systemic shunt represents a risk factor for HCC development. Consequently, we do not agree with their recommendation that patients with TIPS should undergo a more strict HCC-surveillance protocol. Bañares et al.1 further suggest that the indication for a TIPS should be changed: one should be reluctant to place a TIPS in a patient who is not expected to undergo short- or medium-term liver transplantation because of the acceleration of hepatocarcinogenesis. In view of our findings in explanted livers with cirrhosis and because even their own retrospective study shows only a 1.5 times increased risk, we strongly feel that the benefits of TIPS in patients with cirrhosis outweigh the minimal and disputable increase in risk of HCC development.