Changes in serum alpha fetoprotein in patients with recurrent hepatocellular carcinoma following hepatectomy

Background and Aim To study the change in serum alpha fetoprotein ( AFP ) of patients with recurrent hepatocellular carcinoma ( HCC ) after curative resection and to analyze its effect on the survival. Methods We prospectively collected 981 consecutive patients with post‐resectional recurrent HCC between 2005 and 2010 at the E astern H epatobiliary S urgery H ospital. According to the change of AFP from the initial stage to recurrent stage, the patients were divided into stable‐ L (20 ng/m L to 20 ng/m L , n  = 296), stable‐ M (20–400 ng/m L to 20–400 ng/m L , n  = 102), stable‐ H (400 ng/m L to 400 ng/m L , n  = 212), decreasing ( n  = 287), and increasing ( n  = 84) groups. The overall survival ( OS ) and recurrence to death survival ( RTDS ) were analyzed using K aplan– M eier method. Multivariate analysis was performed by C ox proportional hazards regression. Results The stable‐ H /increasing and stable‐ L /decreasing groups had the lowest and highest 5‐year OS and RTDS rates (10.8%/18.8% vs 56.3%/55.0%; 3.4%/5.1% vs 37.7%/33.2%; both P  < 0.001), while the stable‐ M group had the lower rates, which were 29.8% and 23.6% (for OS and RTDS : vs stable‐ L , P  < 0.001 and 0.002; vs deceasing, P  = 0.001 and 0.012; vs increasing, P  = 0.113 and 0.011; vs stable‐ H , both P  < 0.001). Cox regression analysis showed that AFP inconsistency was an independent factor affecting RTDS (decreasing vs stable‐ L , hazard ratio: 1.10, 95% confidence interval: 0.79–1.54, P  = 0.575; increasing vs stable‐ L , 2.93, 2.06–4.16, P  < 0.001). Conclusions The AFP inconsistency was an important prognostic factor for recurrent HCC .