The clinical value of the hepatic venous pressure gradient in patients undergoing hepatic resection for hepatocellular carcinoma with or without liver cirrhosis

Abstract The role of hepatic venous pressure gradient (HVPG) measurement in risk stratification before liver resection is an ongoing area of debate. This study examines the impact of preoperative HVPG levels on overall survival (OS)/time to recurrence (TTR) and postoperative complications after hepatic resection of hepatocellular carcinoma (HCC). Thirty-eight HCC patients undergoing HVPG measurement before liver resection at Cambridge University Hospitals NHS Foundation Trust between January 2014 and April 2022 were retrospectively analysed. Statistical analysis comprised univariable/multivariable Cox/logistic regression to identify risk factors of reduced OS/TTR or 90-day post-resection complications and Kaplan–Meier estimator, log-rank, chi-squared, Fisher's exact, and Mann–Whitney U test, or Student's t-test for survival/subgroup analysis. The median HPVG was 6 (range: 0–14) mmHg. The HVPG was an independent risk factor for poorer TTR in the overall cohort (cut-off: ≥7.5 mmHg (17.18/43.81 months; P = 0.009)). In the subgroup analysis of cirrhotic patients (N = 29 (76%)), HVPG was additionally an independent risk factor for lower OS (cut-off: ≥8.5 mmHg [44.39/76.84 months; P = 0.012]). The HVPG had no impact on OS/TTR in non-cirrhotic patients (N = 9 (24%)), nor was it associated with postoperative complications in any cohort. In conclusion, preoperative HVPG levels are useful predictors for TTR and OS in cirrhotic HCC patients undergoing hepatic resection.


Introduction
Advanced chronic liver disease usually leads to increased portal venous pressure due to fibrotic progression of the liver parenchyma, which is associated with life-threatening complications, such as variceal haemorrhage, hepatorenal syndrome, or hepatic encephalopathy [1,2].The assessment of portal venous pressure is known to be a useful tool for monitoring disease progress and treatment response in patients with chronic liver disease [3].
The hepatic venous pressure gradient (HVPG) measurement is an invasive although safe method for a precise indirect estimation of the portal venous pressure and has therefore become the standard measurement method in recent decades [3,4].HVPG values of 5-9 mmHg are considered mild portal hypertension and, e.g., are associated with a higher risk of HCV recurrence after liver transplantation [5,6].Values ≥10 mmHg are categorised as clinically significant portal hypertension (CSPH) and have been linked to the development of hepatocellular carcinoma (HCC), a higher risk of ascites, variceal bleeding, or hepatic encephalopathy in patients with compensated cirrhosis [7][8][9].A meta-analysis in 2015 of eleven studies with a total of 1,737 included subjects concluded a decreased 3-and 5-year overall survival (OS) in HCC patients with compensated cirrhosis and pre-surgical CSPH as well as an increased risk of liver decompensation in a sub-analysis after liver resection of HCC [10].However, in this pooled analysis, patients with CSPH had significantly higher bilirubin levels and a higher international normalised ratio and were more often Child-Pugh class B at baseline than non-CSPH patients [11].In a propensity score matching study, patients with portal hypertension (indirectly assessed by reduced platelet count and splenomegaly or the presence of gastroesophageal varices) with the same model for end-stage liver disease (MELD) score and extent of hepatectomy planning showed no inferior survival or a higher number of postoperative complications [12].A recent study showed safe postoperative courses and promising 5-year OS and disease-free survival rates of 55 and 43%, respectively, after liver resection of HCC in Child-Pugh A patients with CSPH [13].In addition, it has been reported that one-quarter of HCC patients who would benefit from surgery are excluded when CSPH is considered a contraindication [14].
Hence, there is an ongoing discourse about the impact of increased preoperative HVPG levels, especially when amounting to CSPH, on postoperative complications and the long-term outcome in HCC patients undergoing liver resection.Moreover, the current literature is still based on small sample sizes, and international guidelines do not recommend the HVPG measurement before resection per se due to the minimally invasive procedure, cost-intensiveness, and the requirement of interventional expertise that is usually only available in large tertiary centres [13,[15][16][17].Finally, no study has separately investigated the impact of HVPG levels on post-resection complications or the long-term outcome for non-cirrhotic patients with HCC.
Therefore, the purpose of this study is to investigate the value of HVPG measurements prior to liver resection in HCC patients with or without cirrhosis, determining the impact of the pre-resection HVPG level on OS, time to recurrence (TTR), 90-day post-resection complications, and hospital length of stay after liver resection of HCC with curative intent.

Methods
Institutional approval was obtained for this retrospective study (PRN10518).All subjects gave informed consent prior to HVPG measurement.

Inclusion and exclusion criteria
This study included patients ≥18 years of age with histologically or radiologically (Liver Imaging Reporting And Data System 5) diagnosed HCC who underwent HVPG measurement prior to liver resection with curative intent at Cambridge University Hospitals NHS Foundation Trust, UK, between January 2014 and April 2022 [18].Patients with a previous liver resection at the same lesion site or those ultimately not undergoing liver resection after HVPG assessment were excluded from the analysis.

Study endpoints and parameters
The primary endpoint was to evaluate the impact of the pre-resection HVPG level on OS, TTR, the development of 90-day post-resection complications, and post-resection hospital stay length in HCC patients undergoing liver resection with or without liver cirrhosis.In a subgroup analysis, we compared the same endpoints between patients with and without preoperative CSPH (HVPG values ≥10 mmHg) independently of present cirrhosis.The presence of liver cirrhosis was determined from the pathological reports of liver resection.OS was defined as the time from curative resection to death.TTR was defined as the time from resection to progression of intra-or extrahepatic HCC disease, censoring death without recurrence at the time of death.
In addition, all other obtained basic and disease-specific patient characteristics at baseline were examined on their effect on OS, TTR, 90-day post-resection complications, and hospital stay length (see Table 1).Ascites was graded according to the Consensus Conference of the International Ascites Club [19].Splenomegaly was defined as a major diameter >10 cm on axial CT or MRI [20].Additionally included surgical parameters were resection approach, resection extent (minor resection of up to two or major resection of more than two liver segments [21]), and anatomical/non-anatomical resection.
Post-resection complications were analysed for the time of hospitalisation or up to 90 days after resection.Analysed liver complications included liver dysfunction, new or worsening ascites, biliary leakage, hepatic encephalopathy, and post-hepatectomy haemorrhage.Liver dysfunction was defined by 50-50 criteria positive on two or more consecutive days after liver resection, i.e., a prothrombin time of <50% and a serum bilirubin level of >50 µmol/L, considering the blood count on the third, fifth, The role of the hepatic venous pressure gradient  3 and seventh days after resection [22,23].Post-hepatectomy haemorrhage was defined as bleeding requiring transfusion of packed red blood cells for a significant drop in haemoglobin, radiological intervention, or relook laparotomy [24].
Non-liver complications included acute kidney injury, mechanic ileus, pneumonia, type 1 respiratory failure, cardiovascular complications, or sepsis.Prolonged hospitalisation was considered if the hospital stay length exceeded the median of the overall cohort.(1) A 4 Fr Cobra catheter.The free hepatic venous pressure (FHVP) was measured while the catheter floated freely in the hepatic vein.The wedged hepatic venous pressure (WHVP) was obtained with the catheter wedged into the most distal aspect of the vein until no blood could be aspirated.A wedged venography was performed to confirm the position.(2) An occlusion balloon catheter (Berenstein, Boston Scientific, Marlborough, Massachusetts, USA).The WHVP was obtained while the inflated balloon occluded the vein.
The measured values were retrieved by connecting the catheter to a pressure transducer and monitor.A minimum of three readings were obtained for FHVP and WHVP, respectively.The HVPG was calculated by the difference between the mean WHVP and FHVP of all subsequent measurements.In addition, pressures in the right atrium and inferior vena cava were documented.Patients were monitored and required bed rest for 2 h after the procedure.3 Results
Size of largest lesion, macroscopic vascular invasion, and platelet count were significantly higher in patients with non-cirrhotic HCC, while the total bilirubin level was significantly higher in patients with cirrhosis at baseline.
Three (8%) patients died within 90 days post-resection after a median time of 35 (range: 20-45) days.Two (5%) patients with HPVG values <5 mmHg died of postoperative pneumonia-associated complications.The third (3%) patient with an HPVG of 14 mmHg died of cardiovascular complications after acute kidney failure and acute coronary syndrome.Liver dysfunction was noted in all three patients who died within 90 days after surgery and in five patients (13%) in total (median HVPG: 11.17, range: 0-13.5 mmHg).hospital admission did not differ between patients with and without cirrhosis or patients with and without CSPH (see Table 2).

Independent risk factors for reduced 90-day survival, OS, and TTR
Cox regression results for 90-day survival, OS, and TTR are shown in Table 3.
Bold values indicate a P value < 0.05.

Independent risk factors for postoperative complications and prolonged hospital stay
The logistic regression results for 90-day post-resection complications and prolonged hospital stay are summarised in Table 4.The HVPG level was not associated with postresection complications in the overall cohort or sub-cohorts.The only significant independent risk factor was total bilirubin at baseline for acute kidney injury in the overall cohort (odds ratio (OR), 1.37 [95% CI, 1.02-1.83];P = 0.036).Patient age was the only other significant parameter in univariable analysis predicting new/worsening ascites (OR, 1.16 [95% CI, 1.02-1.31];P = 0.028).Cut-off values in ROC analysis were a total bilirubin level of 15.5 μmol/L for predicting acute kidney injury (AUC, 0.77 [95% CI, 0.57-0.97];P = 0.007) and a patient age of 62 years (AUC, 0.71 [95% CI, 0.55-0.87];P = 0.011) for new/worsening ascites after resection.In the subcohort analysis, no significant independent risk factors for post-resection complications were found.The median hospital stay length in the overall cohort was nine (range: 4-58) days.The ECOG status was an independent risk for a prolonged hospital stay in the overall cohort (OR, 10.5 [95% CI, 1.01-108.76];P = 0.049) and in patients with cirrhosis (OR, 16 [95% CI, 1.38-185.41];P = 0.027).

Discussion
In this study, the preoperative HVPG level was a significant independent risk factor for reduced TTR but not OS when considering all HCC patients who received liver resection with curative intent.In the subgroup analysis of patients with cirrhosis, the HVPG was also an independent risk factor for poorer OS.Cut-off values for reduced TTR were 7.5 mmHg in the overall cohort and 8.5 mmHg for poorer OS in cirrhotic patients.When considering only non-cirrhotic patients, HVPG was no risk factor in survival analysis.Furthermore, the HVPG level was no risk factor for 90-day survival, post-resection complications or prolonged hospital stay in any cohort.
Several studies have investigated the impact of preoperative HVPG values on postoperative complications in The role of the hepatic venous pressure gradient  9 cirrhotic HCC patients.The reported 90-day mortality after HCC resection in patients with HVPG values of ≥10 mmHg ranges from 0 to 28%, which is in line with the 10% of patients (one out of ten) with an initial HVPG ≥10 mmHg in our study [13,14,17,[26][27][28][29].While some authors found significant differences in 90-day mortality between patients with and without CSPH [17], the majority did not [14,[26][27][28][29], including our study.However, it is difficult to compare these results in detail due to differences in baseline characteristics, resection techniques, and study designs.The association of HVPG levels with postoperative ascites, haemorrhage, encephalopathy, or prolonged hospitalisation is also variable in the literature, with cut-off values ranging from >5 to ≥10 mmHg and varying significant findings [13,14,17,26,27,30].We could not identify the HVPG level as an independent risk factor for any observed postoperative complication in patients with or without cirrhosis and found no differences between patients with and without CSPH.Some studies additionally reported that HVPG values of ≥10 mmHg are associated with postoperative liver dysfunction, decompensation in patients with cirrhosis, or prolonged hospitalisation, which was not the case in our study [13,14,17,26,28,29].This might result from the small number of patients with CSPH and cirrhosis (N = 9 (24%)) in our cohort and, therefore, a low incidence of complications.
Several studies have examined the impact of HVPG measurements on long-term outcomes after HCC resection.Azoulay et al. found 1-/3-year OS rates of 89%/73% and 1-/3year recurrence-free survival rates of 82%/62% in 79 cirrhotic HCC patients who underwent resection with HVPG values ≥10 mmHg [13].These results were more favourable compared to our patients with CSPH (1-/3-year OS rates: 90%/58.3%;1-/3-year recurrence-free rates: 60%/0%).While median MELD scores (8 versus 7 in our study) and Child-Pugh scores (N = 78 (99%) versus N = 36 (95%) with Child-Pugh A in our study) were comparable, significantly more patients with microvascular (N = 25 (32%) versus N = 27 (71%) in our study) and macrovascular tumour invasion (N = 3 (4%) versus N = 10 (26%) in our study) at time of resection might have led to a worse overall outcome in our analysis.Llovet et al. found a significantly lower median OS in 43 cirrhotic patients with CSPH undergoing HCC resection after HVPG measurement, with 69 months in patients with an HVPG ≥10 mmHg versus 80 months in patients with <10 mmHg [31].These results are comparable to the cutoff value of 8.5 mmHg found in our study to predict a worse OS in cirrhotic patients (≥8.5 mmHg: 44.39 months; <8.5 mmHg: 76.84 months).In contrast, other authors reported no significant associations between the HVPG and OS.Cucchetti et al. analysed 70 HCC patients with and without cirrhosis after HCC resection and found non-significant differences in 1-/3-year OS rates with 100%/100% in patients with a baseline HVPG <10 mmHg and 97.1%/79.4% in patients with HVPG levels ≥10 mmHg [14].A possible explanation could be the combined analysis of patients with (N = 46 (66%)) and without cirrhosis (N = 24 (34%)) without comparing the subgroups.While patients with CSPH (of which nine (90%) out of ten patients had cirrhosis) in our cohort had lower averaged 1-/3-year OS rates of 90%/58.3%,patients without CSPH at baseline had comparable OS rates of 92.9%/82.1%.However, a detailed comparison of the cohort characteristics is not possible due to imbalances in sample size and the lack of pathological information provided.
Furthermore, Silva et al. showed that the HVPG level was no risk factor for poorer OS in univariable or multivariable analysis of 22 Child-Pugh A cirrhosis patients with resected HCC of a single nodule up to 5 cm [32].In contrast, our study found a significant impact of HVPG on OS in cirrhotic patients, of which 27 (93%) had a Child-Pugh A score, although resection was performed in the presence of up to four lesions and a tumour size of up to 110 mm.Other results favouring the HCC resection in Child-Pugh A patients despite CSPH were described by Cortese and Tellado analysing 42 matched HCC patients with Child-Pugh A cirrhosis undergoing hepatic resection with or without CSPH, respectively, reporting no significant differences in 1-/3-year OS rates (CSPH: 85.7%/64.0%;non-CSPH: 92.9%/70.1%;P = 0.604) or disease-free survival rates (CSPH: 61.3%/44.4%;non-CSPH: 59.5%/29.5%;P = 0.296) [29].Although 1-/3-year OS rates (CSPH: 90%/58.3%;non-CSPH: 92.9%/82.1%;P = 0.146) were comparable in our study, 1-/3-year recurrencefree rates (CSPH: 60%/0%; non-CSPH: 72.1%/42.2%;P = 0.048) were significantly reduced in patients with CSPH.Given that the patient characteristics between our studiesalthough again limited comparabledo not considerably differ, the analysis of a smaller cohort of ten (26%) CSPH patients may have resulted in sample bias, contributing to the fact that no CSPH patient remained disease-free after three years postsurgery in our study.
In summary, varying associations between pre-resection HVPG levels and long-term outcomes are reported in the literature, making it difficult to recommend pre-resection HVPG measurement in patients with HCC per se.However, in HCC patients with suspected liver cirrhosis or patients who are borderline surgical candidates, HVPG measurement could be performed simultaneously with a transjugular liver biopsy as an additional marker of disease extent to assist in the decision for or against liver resection.Moreover, combining HVPG measurement with liver biopsy could be particularly useful, as our study intends that abnormal HVPG values should not be considered as an absolute contraindication for liver resection, and a comprehensive assessment of other risk factors such as underlying cirrhosis is essential for an adequate evaluation of potential surgical candidates.Larger, prospective studies are needed to further validate the benefit of HVPG measurement before liver resection.

Limitations
Our study has limitations.The relatively small sample size might have led to sample bias, especially in the smaller non-cirrhotic or CSPH group.Transplantation remains the preferred option for cirrhotic patients within the criteria.Moreover, our current practice favours ablative therapies over liver resection for borderline surgical candidates.Since we only perform HVPG measurements when liver cirrhosis is suspected clinically or by imaging, and the majority were subsequently classified as cirrhotic, thus ultimately limiting the statistical power of our analysis.Due to the retrospective nature of this study, we could not assess for missing information in the medical reports.Therefore, we could not identify the specific catheters used in 31 out of 38 patients and were unable to compare HVPG levels and outcomes between patients with balloon occlusion and conventional catheter measurements.However, both techniques have been recognised in the literature, were performed with at least three repeated measurements, and were assessed in combination with the inferior vena cava and right atrial pressures as internal reference [33,34].Notably, we have adopted the use of balloon occlusion catheters only to standardise the HVPG measurement at our centre since the completion of this study, which is now also the recommended approach in the most recently published Baveno VII Consensus Workshop to reduce random errors in WHVP measurement [35].Finally, it should be noted that the FHVP level is affected by the catheter tip position and vein morphology, which may result in varying results, especially among different operators [36].However, all investigators in our centre followed the same standard operating procedure.

Conclusions
Our results indicate that the preoperative HVPG level in cirrhotic patients with HCC undergoing liver resection has a significant clinical impact on the long-term outcome with poorer OS and TTR from HVPG levels of 8.5 and 7.5 mmHg, respectively, while the occurrence of postoperative complications appears to be independent of the HVPG.

The role of the hepatic venous pressure gradient  5 Table 2 :
Treatment characteristics observed 90-day complications, and postoperative outcome of the overall cohort and the comparison of patients with/without cirrhosis and patients with/without clinically significant portal hypertension (≥10 mmHg)

Figure 1 :
Figure 1: Kaplan-Meier curves for the determined significant cut-off values for time to recurrence in the overall cohort (a) or overall survival (b) and time to recurrence (c) in patients with cirrhosis after liver resection.

Table 1 :
Preoperative basic and disease characteristics of the overall cohort and the comparison of patients with/without cirrhosis and patients with/ without clinically significant portal hypertension (≥10 mmHg)

Table 1 :
ContinuedPotential candidates for HCC resection with clinical or imaging signs suggestive of cirrhosis were referred for HVPG measurement and, if applicable, simultaneous transjugular liver biopsy to assess the severity of advanced chronic liver disease and evaluate the feasibility and safety of liver resection.Measurements were performed by seven consultant interventional radiologists with 2-17 years of experience.Under aseptic technique and local anaesthesia, a 19-G needle was used to puncture the right internal jugular vein (left if the right side is not suitable) under ultrasound guidance.A 7 Fr Super Arrow-Flex® sheath (Teleflex Medical Europe Ltd, Athlon, Leinster, Ireland) was introduced using Seldinger technique, after which a 4 Fr Cobra Catheter Abbreviations: BCLC, Barcelona Clinic Liver Cancer; COPD, chronic obstructive pulmonary disease; ECOG, Eastern Cooperative Oncology Group; HCC, hepatocellular carcinoma; INR, international normalised ratio; HVPG, hepatic venous pressure gradient; MELD, model for end-stage liver disease; UICC, Union for International Cancer Control.*Normally distributed variable displayed with mean ± standard deviation.Bold values indicate a P value < 0.05.2.3 HVPG measurement (Cordis Corp., Miami Lakes, Florida, USA) was used to cannulate the right or middle hepatic veins.Measurements were performed either with: The mean OS for cirrhotic patients (62.27 [95% CI, 50.98-73.56]months) did not differ significantly compared to non-cirrhotic patients (39.4 [95% CI, 28.22-50.58]months; P = 0.439).

Table 3 :
Results of univariable and multivariable Cox regression analysis for 90-day survival, overall survival, and time to recurrence in the overall cohort and patients with or without cirrhosis

Table 4 :
Results of univariable and multivariable logistic regression analysis for 90-day post-resection complications and prolonged hospital stay in the overall cohort and patients with or without cirrhosis