Liver transplantation is the only potentially curative therapy for patients with end-stage liver disease and early-stages hepatocellular carcinoma (HCC) but priority and equity to access to transplantation are a controversial issue.1,2

Among liver transplant recipients (LTR), the highest incidence of mortality occurs within the early postransplant period and it is close to 10% in elective recipients at 90 days.3 Although some baseline clinical features of donors or recipients may justify some cases of mortality,4–5 in most patients the postoperative complications directly derived from transplantation surgery are responsible for early death and usually unpredictable.6–9

Before the upcoming of Milan criteria10 there was a worldwide moratorium on access to transplant of patients with HCC. Currently, liver transplantation due to HCC accounts for 15–30% of all transplants performed in most of the institutions.11,12 The indication for liver transplantation (HCC vs non-HCC) seems to have an impact on early postoperative mortality but literature around this issue is heterogeneous.13,14 Targeted studies examining the difference between mortality rates after whole-graft elective liver transplantation from brain-dead donors without inclusion of recipients with acute liver failure or retransplantation are lacking.

Liver transplant recipients due to HCC (HCC-LTR) could show lower perioperative mortality rates, mainly due to preserved pretransplant liver function, less portal hypertension15 and reduced comorbidities.16 In addition, while prioritization of LTR due to end-stage liver disease (non-HCC-LTR) depends on the MELD score in most centres in Europe and USA, prioritization of HCC-LTR is more arbitrary (ie. MELD exceptions) and may vary among different countries or even among institutions within the same country.17

Therefore, the aims of this study were to analyse the perioperative differences in a consecutive cohort of LTR classified according to indication to transplant and assess their impact on early mortality during 90 day after transplantation.

The present study is a retrospective analysis of a prospectively collected database from a single centre. A consecutive cohort of adult (>16 aged) LTR with brain-dead donors from January 1995 to December 2017 was analysed.

All LTR after liver transplantation were admitted at Intensive Care Unit (ICU) and included in the study. Exclusion criteria were as follows: patients with acute liver failure enlisted for urgent transplantation, multiple organ transplantation, retransplantation, split or reduced donation, living donation and donation after cardiac death. Subsequently, the cohort was ordered according to transplant indication (HCC-LTR vs non-HCC-LTR).

The STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) recommendations18 were followed for the study design and the manuscript preparation.

The study fulfils all ethical requirements of the WHO code (Declaration of Helsinki) and was approved by the Regional Clinical Research and Ethics Committee (Code 0823-N-18, Act No. 278, ref 3948) without requirement of signed informed consent.

According to our regional prioritization system,19 patients with HCC (HCC-LTR cohort) empirically received 15 MELD points at inclusion in the waiting list. Extra-MELD points were granted only to a subgroup of patients with an increased risk of tumour progression (ie. a single tumour≥3cm, multinodular or with alpha fetoprotein>200ng/mL). In such patients one extra-MELD point was added every month within the waiting list up to 20 MELD points. The non-HCC-LTR cohort included LTR with end-stage liver disease and poor liver function irrespective of cirrhosis aetiology. They were prioritized according their actual MELD score which was updated at each patient visit. Patients enlisted for special indications such as refractory ascites or recurrent encephalopathy were also included in the non-HCC-LTR and prioritized as follows. In patients with refractory ascites, MELD-Na was considered after 3 months in the waiting list. In patients with encephalopathy, MELD 15 at inclusion, one extra MELD point every three months up to 18 points, and 1 extra MELD point every other month thereafter. Both indications were limited to MELD 20.

Throughout the study period, three different surgical techniques were performed in our institution. Until 1992, the Standard technique was the standard of care whenever the patient could tolerate the vascular clamping test. Subsequently, Preservation of inferior cava vein technique became the most commonly performed. In 2007, Preservation of inferior cava vein with a temporary portacaval shunt was adopted and since then it has been considered the technique of choice.

All patients included in the study had been admitted to ICU immediately after surgery. Initial postoperative treatment throughout the study period was fairly homogeneous, with no significant clinical changes in terms of supportive treatment, antimicrobial prophylaxis or length of stay in ICU. Doppler ultrasound was routinely performed to rule out vascular complications within the first 24h after transplant, at day 7th and whenever clinically indicated. Immunosuppressive regimens, however, have been changed according to the state of art of each period. Since 1997, tacrolimus was available in its oral formulation and after 2008 it is administered in a triple therapy regimen using initial low doses. Currently, induction therapy with anti-CD25 antibodies is restricted to patients with severe graft dysfunction or pretransplant renal impairment in order to avoid calcineurin inhibitors within the first 5–7 postoperative days.

From January 1995 to December 2017, 892 adult patients were transplanted and all of them were considered for inclusion. Exclusion criteria were met in 115 patients (Fig. 1) and therefore the overall cohort consisted of 777 adult LTR of which 259 (33.3%) formed the HCC-LTR cohort and 518 (66.7%) the non-HCC-LTR cohort. Annual numbers and proportions of both cohorts is shown in Fig. 2.

Figure 1.

Cohort study. Flow-chart with inclusion and exclusion criteria of patients.

(0.16MB).

Figure 2.

Annual numbers and proportions in both cohorts of liver transplant recipients.

(0.4MB).

Outcomes of cohorts

Length of stay in ICU showed no meaningful differences between cohorts in the central tendency values (Mann–Whitney U test) but it did it in variability or extreme values (Moses test, p=0.012).

The overall cumulative mortality during 90 posttransplant days, when patients with urgent retransplant were censored, was 13.0% in non-HCC-LTR vs 5.9% in HCC-LTR (non-adjusted HR=2.12, 95% CI=1.25–3.61, p=0.005) (Fig. 3). When this outcome was adjusted by their independent predictors, the mortality related to transplant indications (non-HCC-LTR vs HCC-LTR as reference) reach no statistical significance (adjusted HR=1.60, 95% CI=0.93–2.76, p=0.086) (Fig. 4).

Figure 3.

Comparison of liver transplant recipients cohorts. Un-adjusted Kaplan–Meier curve of mortality (patients with need of urgent retransplant were censored).

(0.17MB).

Figure 4.

Comparison of liver transplant recipients cohorts. Adjusted Kaplan–Meier curve of mortality (patients with need of urgent retransplant were censored).

(0.17MB).

In liver transplant recipients, early postoperative mortality is conditioned by complications emerged after transplant surgery and initial graft function, which could be associated with certain clinical pretransplant features and the way to access to transplant. In elective liver transplant candidates, waiting list prioritization is an eternally controversial issue since prioritization systems have to be periodically audited and further adapted to demographic changes and novel transplant indications.

The present study adds additional information to the actual body of research in this field. It does not try to demonstrate the homogeneity of the groups to obtain the validity of the results and predictors, but just the opposite, it tries to show that the cohorts HCC-LTR and non-HCC-LTR are not homogeneous and the mortality predictors have no the same risk in both cohorts.

First of all, in our population the non-HCC-LTR who received their first liver transplant from brain-dead donors showed a crude early mortality rates doubled compared to HCC-LTR (non-adjusted HR=2.12, 95% CI=1.25–3.60, p=0.005). The confirmation of a lower early perioperative mortality rate in HCC-LTR in other studies, it would be equivalent to a “possible advantage” in a short term in comparison to non-HCC-LTR. At present once the Milan criteria are universally applied, this early advantage for HCC patients could be balanced in the long term by mortality risk derived from tumour recurrence.14,20 However, taking into account the current guidelines for treatment of patients with HCC,21 these patients could have in future a lower mortality on the waiting list, a higher probability of transplantation, and a greater early postransplantation survival.22

Secondly, independent predictors of early postoperative mortality in LTR are almost exclusively related to postoperative complications: shock at ICU admission, graft vascular thrombosis and MODS occurrence. And the mortality related to transplant indication (non-HCC-LTR vs HCC-LTR as reference) reached no statistical significance when was adjusted by these independent predictors (adjusted HR=1.60, 95% CI=0.93–2.76, p=0.086). Therefore, the crude difference observed between HCC and non-HCC patients could be only justified by these predictors.

Two meaningful predictors of mortality (postoperative shock and MODS) had higher incidence in non-HCC-LTR. The first is related to a higher need for transfusions within surgery and early reinterventions.23 The MODS is the confluence of higher incidence of postoperative shock, renal dysfunction, graft dysfunction, and a higher rate of postoperative infections. All of them cause a longer mechanical ventilation time and ICU length of stay.

Some of predictors analysed in our work have been previously reported but within heterogeneous populations that included patients with acute liver failure or retransplantation.24–31 Other risk factors that are not widely accepted could also be discussed. LTR stature is not universally accepted as a risk factor.5 However, a study demonstrated that coincidence of short stature recipients and donors with larger size and weight was an independent risk factor for early dysfunction and secondary graft lost.32 Pretransplant serum creatinine is a well-known predictor of postransplant mortality33–34 but has collinearity with MELD-Na, which was expected since serum creatinine is contained in the MELD-Na. Also, blood lactic acid levels tested just after transplantation could also be considered a risk factor, although its clearance is more important than the postoperative peak.35–36

There are several aspects for practical application of our results. On the one hand, a prospective study with separated cohorts according to reason for transplant (HCC vs non-HCC) should be taken into account in future studies to analyse the early mortality in liver transplant recipients. This could prove if overall cumulative mortality is due to other predictors as a higher risk of postoperative shock and MODS. On the other hand, there is a need to assess whether this “possible advantage” of short-term mortality in HCC-LTR (to which an exceptions to the MELD are applied) is in fact compensated.37–40 On the contrary, a re-design of transplant access protocols of the “harmed cohort” should be considered.

Our study have several limitations. It is a single-centre involvement which could decrease external validity. It is a very old historic series which contain a certain grade of heterogenity of the patients. It is a retrospective study, however there are several strengths including a large sample of patients with homogeneous characteristics and the prospective database of most of the explanatory variables.

In conclusion, our study report a higher crude postoperative mortality in non-HCC-LTR compared to HCC-LTR due to higher incidence of shock at ICU admission and occurrence of postoperative multiple organ dysfunction syndrome. The indication to transplant (non-HCC compared to HCC as reference) did not show a meaningful statistical difference in early mortality rate during 90 days after transplantation.

Juan Carlos Pozo-Laderas: concept and design of study, data colleted, statistical analysis, interpretation of results and writing of the manuscript. Ipek Guler: Statistical study. Manuel Rodríguez-Perálvarez: concept and design, interpretation of results, writing and critical revision of the manuscript. Juan Carlos Robles, Ana Mula, Pedro López-Cillero and Carmen de la Fuente: critical revision of the manuscript.

The authors declare no conflict of interest.