Models calculate risk for post-hepatectomy liver failure in cirrhosis
A pair of preoperative and postoperative models estimated the probabilities for post-hepatectomy liver failure in patients with cirrhosis, according to a recently published study.
While the safety of elective hepatectomies in patients with cirrhosis has increased in the last few decades, mortality due to post-hepatectomy liver failure (PHLF) has been reported as high as 50% along with prolonged hospitalization, increased costs and poor long-term outcomes, according to Mathieu Prodeau, MD, from the Lille University Hospital in France, and colleagues.
“As PHLF explains most of the postoperative morbidity and mortality in the context of cirrhosis, the probability of PHLF occurrence and the estimation of its severity could be important information during patients’ selection process, either for confirming surgery, choosing the surgical technique or determining the need for dedicated perioperative care,” they wrote.
The study comprised 343 patients with cirrhosis. The main indication for liver resection was hepatocellular carcinoma (87.8%) and alcohol abuse was the main cause of underlying cirrhosis.
The researchers ranked PHLF as “grade 0/A” for either absence or first grade of PHLF, grade B, or grade C, based on the International Study Group of Liver Surgery definition.
The variables for the preoperative model included intention-to-treat laparoscopy, which correlated with a lower risk for PHLF (OR = 0.31; 95% CI, 0.18-0.53); and platelet count and remnant total liver volume, which correlated with a higher risk for PHLF.
The postoperative model estimated a lower risk for PHLF with per protocol-laparoscopy (OR = 0.25; 95% CI, 0.12-0.51), platelet count and remnant total liver volume. The model also included increased blood loss as a variable that correlated with a higher risk for PHLF.
The preoperative model discriminated between PHLF grade B/C vs. grade 0/A (AU C= 0.72) and the definite presence of grade C vs. grade 0/A/B (AUC = 0.73). The postoperative model yielded significantly better predictions (P < .001) for both grade B/C vs. 0/A (AUC = 0.77) and grade C vs. 0/A/B (AUC = 0.81).
“We suggest this model to be part of the selection tools for hepatectomy in patients with cirrhosis, in order to balance the risk of surgical resection in regards with alternative treatments,” Prodeau and colleagues wrote .”The probabilities of each PHLF grade can be computed from our online calculator, the use of which may facilitate external validation of those models in external independent cohorts.” – by Talitha Bennett
Disclosures: The authors report no relevant financial disclosures.