Reply to “Relevance of Postoperative Peak Transaminase After Elective Hepatectomy”
We are very grateful to Müller et al for their careful reading of our manuscript on the influence of postoperative peak serum transaminase (PST) on the postoperative course of an elective hepatectomy,1 and their constructive comments.
The first refers to the definition of the extent of liver resection, which was based in this study, as in the vast majority of published studies, on the number of segments removed using the classical minor/major dichotomy (up to 2 vs 3 or more). We agree that this distinction is somewhat arbitrary and does not totally reflect the technical challenges and hazards of liver resections. In this study, however, the Receiver Operating Charasteristic (ROC) curves representing the sensitivity and the sensibility of the number of resected segments as a determinant of postoperative severe morbidity identified 3 resected segments as the most relevant cut-off. Although we must admit that the area under the ROC curve was relatively low at 0.663, this is precisely the minimal number of resected segments that defines a major hepatectomy. In addition, when using the number of segments removed as a continuous rather than a categorical variable in a linear regression model, we also failed to show a significant impact of the extent of resection on PST- Aspartate Amino Transferase (AST) (P = 0.498) or PST- Alanine Amino Transferase (ALT) (P = 0.304).
Müller et al elegantly suggest that the classification of the extent of liver resection should not simply take into account the number of resected segments or the remnant liver volume, but also distinguish the different liver segments removed.2 To address this, we changed our classification of hepatectomies as follows: nonanatomical, mono-segmentectomy, left hepatectomy, and bisegmentectomy involving segments 2 to 6 were classified as “minor resections,” whereas right hepatectomy, trisectionectomy, and bisegmentectomy or trisegmentectomy involving segments 7 and 8 were classified as “major resections.” In a linear regression model including this classification, there was a trend for “major resection” to be independently correlated with PST-AST (P = 0.066), but not with PST-ALT (P = 0.181). The other determinant of PST that we had identified (concomitant in situ ablation, red blood cell transfusion, the use of inflow occlusion, and the duration of surgery) remained independently correlated with PST, whereas duration of inflow occlusion remained nonsignificantly correlated with either PST-AST (P = 0.873) or PST-ALT (P = 0.792).
Remnant liver volume was measured in 346 patients (out of 651, 53.1%), when the number of segments planned to be resected, the underlying liver parenchyma, or a history of previous liver resection dictated it. In these patients, some of whom had undergone a portal vein embolization, the mean remnant liver volume-to-body weight ratio (RLBWR) at the time of surgery was 0.88% (interquartile range 0.72–1.15), with only 9 patients having a RLBWR below 0.5%. RLBWR was not correlated with either PST-AST (Spearman ρ = −0.022; P = 0.683) or PST-ALT (Spearman ρ = 0.003; P = 0.957). The median RLBWR was also not significantly different in patients with and without morbidity (0.87% vs 0.91%; P = 0.611), severe morbidity (0.84% vs 0.91%; P = 0.362), or 90-day mortality (0.80% vs 0.89%; P = 0.346).
Another issue raised is the technique used for parenchyma dissection. An ultrasonic dissector was most frequently used (564 patients, 86.6%), whereas bipolar tissue sealer (158 patients, 24.3%), Kelly-clasy (60 patients, 9.2%), and harmonic scalpel (33 patients, 5.1%) were used less frequently (in some patients, 2 or more of these techniques were used). Radiofrequency-assisted resection was never used in this series. The use of a bipolar tissue sealer, a Kelly-clasy technique, or a harmonic scalpel did not influence PST. By contrast, the use of an ultrasonic dissector was correlated with increased PST-AST (median 344 vs 274; P < 0.001) and PST-ALT (median 348 vs 263; P = 0.