Postchemoembolisation syndrome--tumour necrosis or hepatocyte injury?

Abstract

Transarterial chemoembolisation of liver tumours is typically followed by elevated body temperature and liver transaminase enzymes. This has often been considered to indicate successful embolisation. The present study questions whether this syndrome reflects damage to tumour cells or to the normal hepatic tissue. The responses to 256 embolisations undertaken in 145 patients subdivided into those with hepatocyte-derived (primary hepatocellular carcinoma) and nonhepatocyte-derived tumours (secondary metastases) were analysed to assess the relative effects of tumour necrosis and damage to normal hepatocytes in each group. Cytolysis, measured by elevated alanine aminotransferase, was detected in 85% of patients, and there was no difference in the abnormalities in liver function tests measured between the two groups. Furthermore, cytolysis was associated with a higher rate of postprocedure symptoms and side effects, and elevated temperature was associated with a worse survival on univariate analysis. Multivariate analysis demonstrated that there was no benefit in terms of survival from having elevated temperature or cytolysis following embolisation. Cytolysis after chemoembolisation is probably due to damage to normal hepatocytes. Temperature changes may reflect tumour necrosis or necrosis of the healthy tissue. There is no evidence that either a postchemoembolisation fever or cytolysis is associated with an enhanced tumour response or improved long-term survival in patients with primary or secondary liver cancer.

Figure 1

Survival curves of patients depending on (A) the presence or absence of postchemoembolisation syndrome (temperature >38.5°C, log rank=7.16, d.f.=1, P=0.0076) and (B) cytolysis (increase of serum AST of >100 U and more than 2 × prechemoembolisation value, log rank=0.2, d.f.=1, P=0.65).

Figure 2

Survival of patients with liver cancer, based on their response to first treatment with TACE assessed by CT scan 6–8 weeks following the procedure (log rank=137.3, d.f.=3, P<0.0001).