Liver transplantation for adenomatosis: European experience

Multiple adenomas (MAs) occur either in normal livers or in cases of underlying liver pathology, such as glycogen storage disease (GSD) or congenital or acquired vascular anomalies. Initially, Flejou et al.1 described a special entity, liver adenomatosis (LA), in cases with more than 10 adenomas in a normal liver in the absence of GSD or vascular disease. Great progress has been made in the diagnosis, imaging, and pathogenicity of adenomas. In particular, the quality of nontumoral parenchyma is important in the management of MAs. So, considering the recent advances in this field, the term of adenomatosis has been extended and refers today to a high number of these tumors in the liver.2 LA is a heterogeneous disease, and imaging often underestimates the number of lesions present. Chiche et al.3 described 2 forms of LA: a massive form with an enlarged liver and typically large and necrotic tumors, and a multifocal form with a normal-sized liver in which 1 or 2 adenomas may predominate over the smallest lesions.

The risks of hemorrhage and malignant transformation (MT) justify tumor resection in many cases of solitary hepatocellular adenoma (HCA).4-10 However, this surgical approach can be difficult or impossible in cases of LA due to the large number of adenomas. Therefore, liver transplantations (LTs) have been reported in the literature.11-17 Nevertheless, the management of LA remains controversial despite some progress in the prediction of the potential risk of complications (which is likely related to the size and biomolecular subtypes of the adenomas), and LA management ranges from simple follow-up to partial or total liver resection.18-21

This report describes a large retrospective European cohort of LT patients treated for adenomatosis and examines the main factors leading to transplantation and the transplantation results, with the goal of identifying the optimal role for LT in LA. This study also provides recommendations for LT decision-making in LA.

Patients and Methods

Prospectively collected data from the European Liver Transplant Registry (ELTR)22 were reviewed after obtaining approval from the board of the European Liver and Intestine Transplant Association. Patients undergoing LT for LA between January 1, 1986, and July 15, 2013, were selected. Patients with GSD type IA who underwent LT for LA were included.

The ELTR has collected prospective data on LT in 145 centers all over Europe since 1968. We selected cases with the diagnostic code “adenomatosis” in the ELTR database. A standard computerized file was provided to the contributing centers with detailed instructions for the collection of accurate and uniform information. However, participation in the ELTR is voluntary and may contain coding errors. Therefore, we directly contacted the centers and, notably, all of the French and Belgium centers to check the validity of the ELTR list. Patients were included only if the primary indication was LA (exclusion of hepatocellular insufficiency).

Results

According to the ELTR, 64 patients were identified as having a LT for LA in 22 centers in 5 European countries during the study period. Several centers did not provide data, others reported more or less cases than the registered number in the ELTR database, and several coding errors were identified.

Additional cases were collected after direct calls to the French centers. Thirteen French centers reported 24 patients, but only 5 of these patients were listed in the ELTR database. Two Belgium centers reported 6 patients; 4 Italian centers reported 7 patients; 2 Spanish centers reported 3 patients; and 1 German center reported 9 patients.

Eventually, a total of 49 patients underwent a LT for histologically confirmed adenomatosis throughout this long period, quite consistently, because 30 underwent transplantation before 2006 and 19 underwent transplantation after (Fig. 1). Table 1 shows the patient characteristics. Thirty-one of the 49 (63%) patients were women, and 18 (37%) patients were men. Information on estrogen intake was not available in 12/31 women, even after advanced research. In the other 19 female patients, only 6 had a history by oral contraceptive intake. The following initial clinical manifestations of the disease were noted: 22 follow-up cases (14 for glycogenosis); 6 cases of abdominal pain with hemorrhagic manifestations; 6 cases of accidental discovery, including trauma in most cases; and 4 cases of hepatomegaly. Twenty-seven of the 49 (55%) patients were younger than 30 years. We obtained information for maturity onset diabetes of the young (MODY) 3 diabetes23-25 for 2 patients, and 2 patients had a family history of adenomatosis, including 2 sisters in this study. Sixteen (33%) patients had GSD type IA in this study.

Discussion

LA is a benign disease, although 2 potential lethal risks may justify aggressive management: cases of the massive form with complications, such as iterative and/or severe bleeding, and cases showing MT.

This ELTR study is the largest cohort of LT in LA reported in the literature. Although we only identified 21 patients who received LT for adenomatosis without GSD in the English language literature, our series of 33 patients without GSD is the largest reported to date (Table 6).

This study first described a population of transplanted patients to analyze the primary factors that lead to transplantation in current practice. Several interesting findings from the data emerged from our analyses. First, we observed a predominance of women (63%) but a lower incidence compared to the usual HCA population. The median age was 28 years (23-41), and more than half of the patients were younger than 30 years old. Notably, we expected to find a large number of massive and highly symptomatic forms, but we found that 75% of the transplanted patients were asymptomatic. Specifically, 43% of the patients presented multifocal disease, and the maximal size of the adenomas in 9 (18%) patients was ≤4 cm in liver explants, but this concerned patients who had undergone previous hepatectomies or patients with liver disease hampering the possibility of partial hepatectomies.

Concerning the indication of LT, because hemorrhage appeared to be the main indication in this disease in the case reports of the literature we expected to collect a higher percentage of patients transplanted for hemorrhage in this study. However, only 5 patients underwent transplantation for this complication: 1 patient had dramatic hemorrhage with a history of several previous partial resections, and 4 patients presented with underlying liver disease that contraindicated resection. Iterative severe or life-threatening hemorrhages are rare in LA. Moreover, we have an expanding armamentarium and experience in controlling bleeding angiographically, so this indication of LT should remain exceptional. It should be restricted to hemorrhages in case of failure or impossibility of embolization and/or partial hepatectomy, because of the location of the bleeding tumor, the number of hemorrhages, or the underlying liver disease.

The concern of malignancy happens to be, in this series, the major indication. Two situations were observed before transplantation: (1) a proven diagnosis of MT of 1 adenoma or (2) the fear of malignancy. In the first case, the transformation of 1 or more HCAs among lots of other nodules was a major concern, which explains the decision to perform LT. In 1 patient, this decision was likely too late because the disease was obviously progressive and HCC in the liver explant was very advanced.

The second situation, the fear of malignancy, is more debatable. Notably, we observed that 14 (75%) patients with GSD underwent transplantation for HCC suspicion or disease progression, although only 4 (25%) patients with GSD had HCC that was histologically confirmed before LT. Twelve (36%) patients without GSD underwent transplantation with histological proof of HCC, and 11 (33%) patients underwent transplantation for HCC suspicion or disease progression without histological proof. The risk of malignancy was clearly overestimated in the group with GSD because the literature considers this underlying disease as a risk factor for the MT of adenomas.8 These data are supported by a high frequency of β-CAT–mutated adenomas5, 7 in this pathology. We identified nearly 37 patients transplanted for MAs and GSD in the literature, but only 6 patients showed evidence of HCC in liver explants. Therefore, 85% of the patients who underwent transplantation for adenoma exhibited no HCC on liver explants, both in our study and in the literature. However, patients with GSD underwent transplantation earlier than other patients (median age was statistically lower), perhaps before MT.

MT was a frequent feature in patients with vascular disease associated with adenomas, as this presentation occurred in 4 of the 7 patients. However, it was not statistically relevant in our analyses of risk factors, likely because of the low number of these cases.

In the 9 patients with normal livers, (1) the median age was higher; (2) the form of LA, either massive or multifocal, was usually evolutive; and (3) the MT was multifocal in liver explants in 5 of the 9 patients.

These data demonstrate that follow-up must be recommended in LA to diagnose this lethal complication. However, implementation of this follow-up approach is not clearly established. Magnetic resonance imaging may be the best method to study increases in adenoma size or intratumoral modifications with apparitions of a portal washout, and biopsy or surgical exploration should also be discussed. The means and timing of follow-up in LA should likely be adaptable in this heterogeneous disease. Follow-up must be frequent and accurate in evolutive forms of adenomas and in patients with vascular disease or a β-CAT mutation, whereas follow-up can be less frequent in nonevolutive forms of the disease and in patients with small adenomas, particularly young patients, using ultrasonography for basic routine examinations. In addition, patient age >30 years was statistically associated with malignancy.

Analyses of biomolecular pathways revealed another interesting finding of this study. Biomolecular studies are not performed in all centers, which is why we did not obtain complete data for our 49 patients. Nevertheless, even when incomplete, these data are informative. The HNF1α subgroup was the most frequent feature in LA without liver disease or GSD, as reported in the literature; however, we also observed that different biomolecular subgroups of adenomas could be present in the same patient41 and that MT was possible across biomolecular subtypes, even in the HNF1α group (2 HCCs with underlying vascular disease of the liver). Therefore, more data are mandatory to better understand the link between HCC and biomolecular subtype, and it is clear that the risk of malignancy, in the case of adenomatosis, should not be stated only on the biopsy of 1 sole adenoma and does not exclusively concern β-CAT–mutated adenomas.

The clinical results of LT in this series were relatively good (41 patients alive and well), although 43% of the patients had undergone previous liver surgery and some had vascular abnormalities. Nevertheless, we must take into account the 3 operative deaths and 5 retransplantations, which is not negligible. Transplantation teams have to assess the potential technical difficulties of LT related to prior vascular or hepatic surgery. This past history may be a factor of operative bleeding and/or morbimortality. So, the performance of preventive LT is debatable, although the opposite risk is delayed transplantation after the MT of several tumors. The patients with MT who recurred after LT died of the disease. These 3 patients had massive LA with multifocal HCC identified on liver explants. One of these 3 patients had vascular abnormalities; 1 patient had GSD and vascular abnormalities; and 1 patient had no underlying liver disease (no GSD, no vascular abnormalities). However, all 3 patients had advanced HCC in liver explants with vascular involvement (obvious for 2 patients on preoperative imaging), which is a well-known very poor prognostic factor. Like in transplantation for HCC, diagnosis of vascular invasion on imaging should indeed be an absolute contraindication to transplantation. In conclusion, MT seems to be a good indication if it is not too advanced. This implies that in the case of suspicion of MT, diagnosis has to be rapidly confirmed in order to discuss LT without delay.

Even with this study, it is difficult or impossible to provide evidence-based recommendations for the indication of LT in LA. It should be discussed with caution because of the benign nature of this disease, the mortality of LT, and liver graft shortage. Some authors have stated that there is no indication for transplantation in LA,18 and this series demonstrates the opposite. Nevertheless, 1 major problem is the access to LT for these patients without cirrhosis and therefore with low Model for End-Stage Liver Disease (MELD) score. Several years ago (approximately before 2006-2007 in Europe), before MELD score was adopted for organ allocation, patients with LA had access to transplant quite easily. Today, with the European allocation's rules based on MELD score, those patients can be transplanted only as MELD exceptions or with living donors (3 in this series). That is why it is important to define the good candidates for transplantation. In most patients, it seems that the indication for LT will not depend on a single factor.

Our results highlight the crucial necessity of precise histological study of several adenomas (pathomolecular characteristics and mandatory proof of MT) and nontumoral livers before any discussion of LT. This study also supports that the restrictive definition of LA that excludes GSD or vascular disease is not justified even if these underlying pathologies influence the management of these patients.

In conclusion, this study demonstrates that LA is an extremely rare indication for LT, representing approximately 0.03% of LTs in Europe, which should be validated in cases of exceptionally aggressive types of this disease. Considering the characteristics of this population (previous liver surgery, liver underlying disease, vascular abnormalities), LT is associated with a significant rate of morbimortality. On the other hand, many complications of this benign disease can be handled by modern radiological or surgical approaches. If not, mostly in case of MT, and after assessment of the risk-benefit ratio, this European experience shows that LT can be justified. However, the access to LT depends on an expert decision because the system of graft allocation is based on the MELD score, and LA has no impact on liver function. Therefore, the identification of several criteria is mandatory to aid the decision-making process.

The strength of this multicenter series was the collection of information on extremely rare diseases. These results should serve as a catalyst to initiate prospective multicenter studies to address the many remaining issues of this rare disease, such as the identification of high-risk patients, prediction of malignancy, and the timing of LT. A prospective worldwide registry of patients transplanted for LA, with extensive study of history, pretransplant modern imaging, and precise pathological examination of the explanted liver with biomolecular characterization would be the best way to progress in that heterogeneous and complex rare disease.