The goal of the study was to determine the incidence and variables associated with post-liver transplantation (LT) de novo internal neoplasms development, excluding skin tumors and hepatocellular carcinoma. Medical records were reviewed for recipient/donor demographics, viral serology, cause of liver disease, interval from LT to tumor diagnosis, predisposing factors, immunosuppression and survival. Forty-one neoplasms (31 solid and 10 hematologic) developed in 772 recipients (5.3%) transplanted between 1991 and 2001. Time to tumor diagnosis was longer in patients transplanted before 1995 than in those transplanted afterwards (58 vs. 22 months; p < 0.05). Hematologic neoplasms (HN) appeared earlier than solid (2 vs. 21 months; p < 0.001), were more prevalent in those transplanted after 1995 than before (32% vs. 12.5%), and had lower survival than solid (2 vs. 21 months, p < 0.001). While HCV was the most frequent indication in HN (70%), alcohol was that of solid tumors (71%). Overall, risk factors for de novo neoplasms included alcohol and immunosuppression (p < 0.01). In patients undergoing LT in recent years, there is a higher incidence of HN with de novo internal neoplasms developing at earlier time-points than in those transplanted years ago. Risk factors for tumor development include alcohol, HCV and possibly strong immunosuppression.
Impairment of immunovigilance can result in uncontrolled proliferation and malignant transformation of cells. Indeed, primary or acquired immunodeficiency states are associated with an increased risk for developing malignancy, including lymphomas or Kaposi's sarcoma. Organ transplantation represents an induced state of immunosuppression and a subsequent high incidence of neoplasms has been described in transplant recipients. For instance, the incidence of de novo neoplasms in liver transplant recipients ranges form 3% to 16% (1–6), a percentage significantly higher than that observed in the general population (7,8).
This complication is a major cause of late death in liver transplant recipients, responsible for 25% of the deaths occurring in patients who have survived more than 3 years post-transplantation (9).
In recent years there have been significant changes in the field of liver transplantation (LT), namely the increase of candidates undergoing transplantation at advanced stages of hepatic insufficiency and the development of newer and more potent immunosuppressive drugs. Based on these data, we hypothesized that the risk of developing de novo neoplasms would be higher in patients undergoing LT in recent years as opposed to those transplanted in earlier cohorts.
The main goals of this study were thus, to determine the incidence of de novo neoplasms (excluding skin tumours and recurrent hepatocarcinoma) in patients undergoing LT at one single center, and to identify variables associated with their development.
Patients and Methods
We retrospectively reviewed the medical records of 830 liver transplantations performed in 772 patients from January 1991 to June 2002. Of these, 714 patients received only one transplant, 55 patients received two LT and 3 patients underwent 3 LT. The median follow-up was 51 months (range 2–136 month.) There were 522 males (68%) and 250 females (32%). The median age at time of transplantation was 53 years (range: 19–67 years).
The indications for LT were the following: viral cirrhosis in 381 (46%), alcoholic cirrhosis in 191 (23%), mixed causes (hepatitis C virus plus alcohol) in 79 (9.5%), primary biliary cirrhosis in 25 (3%), cryptogenic cirrhosis in 24 (3%), fulminant hepatic failure in 17 (2%), hepatic tumors in 18 (2%), autoimmune hepatitis in 9 (1%), haemochromatosis in 8 (1%), re-LT in 58 (7%) and other causes in 20 (2.7%).
Induction immunosuppression changed over the years. Overall, the triple standard immunosuppressive regimen was used from 1991 to 1995 (cyclosporine, azathioprine and steroids). After 1995 significant changes were introduced, consisting primarily in the use of more potent immunosuppressive agents [mofetil mycophenolate (MMF), tacrolimus, sirolimus, anti-IL2 receptor antibodies] and the earlier discontinuation of second line immunosuppressive drugs, such as steroids.
In this study, maintenance immunosuppression with one drug alone (calcineurin inhibitor) was achieved in 95% of patients at 1 year of LT (except in those in whom the indication for transplantation was an autoimmune disease and in those in whom the calcineurin inhibitor was switched to another drug due to toxicity). Dosages of cyclosporine and tacrolimus were individually modified according to plasma levels and depending on the patient condition, particularly the presence of comorbidities.
Rejection episodes were histologically proven and were treated, if moderate or severe, with three boluses of 1 g of methyl-prednisolone. Persistent rejection episodes, defined as those with a lack of histological response to six boluses of methyl-prednisolone, were treated with OKT3 or tacrolimus (the latter since 1995).
Follow-up after LT
During the first 2–3 months, recipients were seen in the outpatient clinic twice a month after their hospital discharge. Thereafter the visits were performed every 2–4 months or at more frequent intervals if needed (particularly in patients with recurrent hepatitis C). At 1 year, the interval between visits was switched to 4–6 months whenever possible. An ultrasound examination of the liver graft and chest X-ray were performed at 1 and 3 months, and once a year thereafter. Additional tests were performed when clinically indicated, depending on the clinical status, personal history and LT indication.
Skin malignancies were excluded from this study due to the high prevalence of these tumors in our geographical area and the high likelihood of their underestimation as most are discovered and treated in other centers.
The following variables were analysed in every case: age (donor and recipient), gender, date and indication for LT, previous history of tobacco use or alcohol abuse, previous family cancer, serological status (donor and recipient) of Epstein–Barr virus (EBV), cytomegalovirus (CMV), hepatitis B virus (HBV) and hepatitis C virus (HCV). We also recorded the type of initial and maintenance immunosuppressive regimen, and the incidence and treatment of acute rejection episodes. The type, location, date of tumor diagnosis and management were included in the data analysis. Hematologic malignancies were defined as a heterogeneous group of diseases which included acute and chronic leukemias, B and T cell lymphomas, Hodgkin lymphomas and plasma cell disorders (multiple myeloma). Post-transplant lymphoproliferative disorders were included in this group of malignancies.
Results were expressed as mean or median for quantitative variables and proportions for qualitative variables. The distribution of categorical variables was compared using the Chi-square test or Fisher's exact test as appropriate. Differences between means were accordingly assessed by the Student's t-test or the Mann–Whitney U-test. Kaplan–Meier survival analysis was performed to estimate the risk of developing malignant disease by transplantation year and to determine survival of patients. Multivariate logistic regression analysis was used to identify risk factors independently associated with tumor development. All predictors in the univariate analysis with p < 0.1 were entered into logistic models. p < 0.05 was considered significant. All statistical analyses were performed with SPSS 10.0 (SPSS Inc., Chicago, IL, USA).
Incidence and characteristics of de novo internal neoplasms.
Incidence. There were 41 de novo internal neoplasms in 772 patients, with an overall incidence of 5.3%, none having more than one malignancy. The median age at transplantation of patients developing de novo malignancy was 50 years. The median age at tumor diagnosis was 54.5 years (range: 28–67) with a gender distribution of 24 men (59%) and 17 women (41%).
The indications for LT in these patients are listed in Table 1. The major indication was alcoholic cirrhosis (18 patients), while viral cirrhosis was the cause in 14 patients (13 patients with HCV-related cirrhosis and one patient with HBV-related cirrhosis), coexisting with alcohol abuse in four cases.
Table 1. Comparative features between patients with de novo internal neoplasms and patients without
Baseline features of de novo neoplasms. Histologically, epidermoid tumors were the most frequent neoplasms (20 cases), followed in order of frequency by lymphomas (n = 9) and adenocarcinomas (n = 9). One patient developed a lower extremity sarcoma, another a cerebral astrocytoma and a third patient developed a myeloma. Location and type of malignancies are reflected in Table 2. Histopathologically, most of the B-lymphomas cases (n = 5) were featured as polymorphic variants. The vast majority of tumors were diagnosed in advanced stages. For solid tumors, the TNM classification (Tumor, Node, Metastases) of the International Union Against Cancer was used: stage I, n = 1; stage II, n = 6; stage III, n = 4; stage IV, n = 14; not available, n = 6. For hematologic malignancies, the most frequently used classification was that of Ann-Harbor. Using this classification, all hematologic malignancies were considered to be in advanced stages of disease.
Table 2. Type and location of de novo tumors developing in 41 liver transplant recipients
10 hematological tumors
31 internal solid tumors
7 non-Hodgkin lymphoma
9 head and neck tumors (5 tonsils,
2 tongue, 1 oral cavity, 1 laryngeal cancer)
1 Hodgkin lymphoma
8 lung cancers
1 Burkitt lymphoma
3 bladder cancers
2 breast cancers
2 gastrointestinal cancers
7 other (prostate, vulva, ovary, brain, pancreas, liver and leg)
The number of transplants per year and number of internal de novo neoplasms diagnosed according to the year of transplantation is shown in Figure 1. This latter number was considerably greater in patients who underwent LT in years previous to 1995 (10% of patients vs. 4% of those transplanted after 1995; p < 0.008).
The incidence of tumors during the first 2 years post-transplantation was slightly higher in those transplanted after than before 1995 (2.2% vs. 1.9%). In contrast, this difference was the opposite after the second year of follow-up (Table 3).
Table 3. Differences in incident rates of ‘de novo’ tumors between patients transplanted before and after 1995
Patients transplanted before 1995 (n = 156)
Patients transplanted 1995 and after (n = 616)
Number of tumors diagnosed
1st year of LT
(n = 10)
2nd year of LT
(n = 7)
3rd year of LT
(n = 2)
4th year of LT
(n = 8)
5th year of LT
(n = 6)
≥6th year of LT
(n = 8)
The median time from LT to tumor diagnosis was 40 months (2–94.5). Interestingly, the median time for the development of de novo tumors was shorter in patients undergoing LT in recent years (mainly after 1995) than in those transplanted in earlier cohorts (22 vs. 58 months, p < 0.05) (Figure 2 and Table 4).
Table 4. Characteristics and risk factors for the development of denovo internal neoplasms in patients transplanted before and after 1995
Before 1995 (n = 156)
After 1995 (n = 616)
LT: liver transplantation; NS: nonsignificant, HCV: hepatitis C virus.
Type of tumor
Solid, n = 14 (87.5%);
Solid, n = 17 (68%);
Hematologic, n = 2 (12.5%)
Hematologic, n = 8 (32%)
Lymphoma, n = 1
Lymphoma, n = 8
Adenocarcinoma, n = 5
Adenocarcinoma, n = 4
Epidermoid, n = 8
Epidermoid, n = 12
Others, n = 2
Others, n = 1
Head and neck, n = 7
Head and neck, n = 3
Thorax, n = 7
Thorax, n = 6
Others, n = 2
Abdomen, n = 13
Others, n = 3
Time from LT to tumor diagnosis (median, range)
58 month (2–94)
22 month (4–67)
Patient age at tumor diagnosis (median, range)
Variables associated with tumor development
With tumor: 14 (87.5%)
With tumor: 5 (20%)
Without tumor: 66 (49%)
Without tumor: 78 (17%)
(p = 0.003)
(p = NS)
With tumor: 10 (62.5%)
With tumor: 14(56%)
Without tumor: 45(32%)
Without tumor: 203 (34%)
(p = 0.02)
(p = 0.01)
With tumor: 2(12.5%)
With tumor: 11(44%)
Without tumor: 71(50%)
Without tumor: 313(53%)
(p = 0.004)
(p = NS)
Induction with azathioprine
With tumor: 16(100%)
With tumor: 17(68%)
Without tumor: 138(98%)
Without tumor: 221(38%)
(p = NS)
(p = 0.003)
Treatment and outcome. Nine patients received palliative treatment, while 12 patients underwent surgical intervention with or without radiotherapy or chemotherapy. Overall, 20 recipients were treated with chemotherapy and/or radiotherapy. In all cases, doses of immunosuppressive drugs were reduced.
At the end of the study period, 23 patients had died. The main cause of death was malignant neoplasia (n = 21; 91%). The median survival since tumor diagnosis was 16 months (range: 3–64 months).
Risk factors. With regards to risk factors known to impact the development of de novo tumors: 34 patients had a positive EBV-serology, with no data on the donor; 21 patients (51%) reported a history of tobacco-use before LT and 6 patients had a family history of solid neoplasms. Thirty patients (73%) were administered cyclosporine, azathioprine and steroids as induction immunosuppression. The maintenance immunosuppression regimen was cyclosporine in most cases (35 patients; 85%). Three patients received tacrolimus while the remaining three were administered different protocols as maintenance immunosuppression. Nineteen patients (46%) had developed acute rejection episodes (one episode in 16 patients, two episodes in two patients and three episodes in one patient) with 13 of these being treated with bolus doses of steroids. Four patients had received monoclonal antibodies (OKT3 or basiliximab) and three antithymocyte immunoglobulin (ATGAM) as induction immunosuppression. Therefore, overall there were seven patients in the group of de novo tumors who had been previously treated with monoclonal or policlonal antibodies. These patients developed six solid tumors (two lung cancers, one sarcoma, one prostate cancer, one tongue cancer and one tonsil cancer) and one lymphoma.
There were no cases of ulcerative colitis or primary schlerosing cholangitis.
Comparative analysis of de novo hematologic neoplasms vs. de novo solid internal neoplasms
A comparative analysis between hematologic and solid tumors was performed. The indication for LT was significantly different between both groups: 70% of the patients from the former group received a LT due to HCV-related cirrhosis, while this indication was only present in 19% of the latter group (p < 0.001, Chi-square test). In contrast the number of patients who underwent LT for alcoholic cirrhosis was significantly higher in patients who developed an internal solid tumor than in the hematologic neoplasm group 71% vs. 20%, respectively, (p < 0.01, Chi-square test). Overall, the incidence of tumors was higher in patients undergoing LT for alcoholic cirrhosis than in those undergoing LT for other causes (9% vs. 3%).
Hematologic tumors appeared earlier than solid tumors 14 months vs. 44 months, p = 0.08). In addition, the proportion of solid tumors was greater in patients who had undergone LT before 1995 (67% of solid tumors) than in those transplanted afterwards 33% of solid tumors). In contrast, the proportion of hematologic neoplasms was relatively higher in patients who had undergone transplantation after 1995) 60% of hematologic tumors were diagnosed in patients undergoing LT after 1995).
No significant differences in age, sex, immunosuppressive regimen, rejection episodes, post-transplant infections, HBV, EBV and CMV serological status and familial history of neoplasms were found between both groups.
Eighty percent of the patients with hematologic tumors (8 out of 10) and 48% of those with internal solid tumor (15 out of 31) died after a median of 6.4 months since tumor diagnosis (range: 0.1–81 month.). Survival since tumor diagnosis was significantly higher in patients with solid tumors (21 months, 95% CI: 0–56 month.) than in those with hematologic malignancies (2 months; 95% CI: 0.3–3.5 months) (Figure 3).
Comparative analysis of patients with de novo neoplasms compared with those without
There were no significant differences in age at LT, gender distribution, HBV, EBV and CMV serologies, treatment of acute rejection episodes and Child–Pugh classification at time of LT between the two groups. A higher frequency of alcoholic cirrhosis as an indication for LT was observed in patients who developed tumors as compared with those without (44% vs. 23%, p < 0.01; Chi-square test). The immunosuppressive induction regimen was also different (Table 1): 88% of the patients in the tumor group had received cyclosporine and 81% azathioprine, vs. 73% and 50%, respectively, in the nontumor group.
In addition, the number of rejection episodes was significantly higher among patients who developed de novo tumors than in those without (46% vs. 24%; p < 0.01).
Using multivariate analysis, three variables remained in the model: rejection, history of alcohol abuse and azatioprine treatment (Table 5).
Table 5. Risk factors associated with tumor development (multivariate logistic regression model)
RR: relative risk; CI: confidence interval.
Tumor features and risk factors in different cohorts
Because type and time-to-tumor development were found to change over time (prior and after 1995), we analyzed the characteristics of the tumors (Table 4), the risk factors associated with this complication (Table 4), and the change of the different variables in the two cohorts (Table 6). Four variables were found to be associated with the development of de novo malignancy in the two groups: alcohol (in both cohorts), HCV (<1995), rejection episodes (<1995) and induction with azathioprine (>1995) (Table 4). Immunosuppressive therapy was one of the major changes to occur overtime, with a higher proportion of patients receiving potent immunosuppression in the last cohort as compared with the group of patients transplanted before 1995 (36% vs. 6%, respectively) (Table 6).
Table 6. Evolution of several variables over time in patients with de novo internal neoplasms
Before 1995 (n = 16)
After 1995 (n = 25)
NA: not applicable; NS: nonsignificant
*Potent immunosuppression was defined by the use of tacrolimus (vs. cyclosporine), mofetil mycophenolate (vs. azathioprine) and use of antilympocytic preparations.
Induction with cyclosporine
Induction with tacrolimus
Induction with azatioprine
Induction with MMF or rapamycin
Induction with antilymphocitic preparations
Methyl-prednisolone bolus administration
EBV-positive serology in receptor
Gender (% men)
Donor age (median, range)
Age at LT (median, range)
Child–Pugh stage (median, range)
Follow-up (months) (median, range)
Age at tumor diagnosis (median, range)
Prednisone duration (months) (median, range)
Cyclosporin serum levels (ng/mL) (median, range)
Tacrolimus serum levels (ng/mL) (median, range)
Prednisone doses (mg) (median, range)
Azathioprine doses (mg) (median range)
Malignancies represent a serious complication developing in organ transplant recipients. Its real incidence and factors increasing the risk are not completely defined. This information is relevant for patient management and screening policies in liver transplant units.
In our study, based on a large cohort of liver transplant recipients, the incidence of de novo internal neoplasms was 5.3%. This frequency is higher than that observed in the general population matched for age (1–1.7%) (10), and is similar to that reported in previous studies on LT recipients where skin cancers were also excluded (2–17%) (1–3,7,8,11–22) (Table 7).
Table 7. Incidence of de novo internal neoplasms (after exclusion of skin malignancies) in the literature
Number of patients with LT
De novo internal neoplasms diagnosed
Mean follow-up (months)
LT: liver transplantation.
Catalina MV et al.
Sánchez E et al.
Jiménez C et al.
Bellamy C et al.
Xiol X et al.
Haagsma EB et al.
Catena F et al.
Jain A et al.
Sheiner PA et al.
Jaurrieta E et al.
Galve ML et al.
Peyrègne V et al.
Berenguer M et al.
Bessa X et al.
Jonas S et al.
Levy M et al.
Immunosuppression appears to be a major risk factor associated with the development of this complication. Based on the recent introduction of newer and more potent immunosuppressive drugs in the LT arena, we hypothesized, first, that the number of de novo tumors would increase in recent years; and second, that these tumors would follow a more aggressive behaviour. Our findings partially confirm the two hypotheses. Indeed, we have shown that: (i) despite a shorter follow-up, de novo hematologic neoplasms are more frequent in patients undergoing LT after 1995 than in those transplanted before; (ii) the incident rates per year show a slightly higher incidence of tumor development during the first years post-LT in patients transplanted after 1995 compared with those transplanted before, despite a shorter follow-up at risk; and (iii) the development of both hematologic and solid tumors occurs at a shorter interval after transplantation in patients transplanted in recent years compared with those transplanted in earlier cohorts (Figure 2). Indeed, in patients in whom LT was performed before 1995, the median time of appearance of de novo tumors was 58 months (range: 2–94), while this time was only 22 months (range: 4–67) in those undergoing transplantation more recently. In addition, although no significant differences were found in immunosuppression regimen, type and duration or anti-rejection therapy between patients with hematologic tumors and those with solid neoplasms, and nor between patients with and without tumor (except for azathioprine use), the recent introduction and increased use of more potent first line immunosuppressive drugs, such as tacrolimus or interleukin-2 receptor antibodies (Table 6) may, in part, explain the earlier development of malignancies in recent years. The significant reduction in rejection episodes and subsequent treatment with methyl-prednisolone boluses observed in recent years is probably a reflection of a more profound immunosuppression.
Of note, azathioprine therapy was found to be independently associated with tumor development, a well-known risk factor for tumor development (23).
It is also known that the time of exposure to immunosuppressive drugs is a risk factor for post-LT neoplasms (3–5,24). Immunosuppression allows oncogenic virus activation and proliferation. A direct mutagenic effect caused by immunosuppressive drugs is probably also involved. In our report, a significant proportion of patients with de novo tumors had undergone liver transplantation before 1995 (37%) and as such, had been exposed to immunosuppressive drugs during prolonged periods of time. In fact, the finding that 88% of patients with internal de novo tumors were receiving cyclosporine probably reflects longer immunosuppression exposure (tacrolimus was only introduced in 1995) as opposed to more potent immunosuppression.
Additional important findings from this study may be summarized as follows:
(i) Alcoholic cirrhosis is a frequent indication of LT in patients who develop de novo tumors (44%). Indeed, the incidence of internal de novo malignancies was 9% in patients undergoing transplantation for alcoholic liver disease. This incidence, which is in accordance with the results previously reported by other groups (5,13,17,19,24–26), is higher than that observed in patients undergoing LT for nonalcoholic liver disease (3%). Most tumors in this group of patients are those defined as oropharyngeal tumors (27,28), developing in the oral cavity and larynx. In fact, in our study, 8 out of the 10 patients with oropharyngeal cancer had undergone LT for alcoholic cirrhosis. Based on this finding, we recommend a strict patient follow-up with yearly screening for this type of tumors in patients with a history of alcohol abuse, regardless of abstinence.
(ii) The most common tumors among LT recipients (excluding skin malignancies) are head/neck cancers and lymphomas, the latter appearing earlier than the former. These findings are in accordance with previous studies which have shown that lymphoproliferative disorders usually appear during the first or second year after LT (11). In addition, we also found that hematologic malignancies (mainly non-Hodgkin lymphoma) are more aggressive than other type of tumors, also confirming data from earlier studies (1,4,7–9,20,21,26,29).
(iii) Known risk factors associated with the development of post-transplant lymphoproliferative disorder are Epstein–Barr virus, age and the use of antilymphocytic preparations (29,30). We could not confirm these findings, which may be related to the low number of patients treated with antilymphocytic preparations and the low reliability of serology in the early diagnosis and monitoring of EBV infection. We found, however, an association between HCV infection and the development of these disorders, association which has been recently suggested in several series (29,31), and which probably reflects the chronic stimulation of immune cells, and more specifically of B cells, by the HCV.
(iv) Most tumors were diagnosed at advanced stages. The low specificity of clinical symptoms, great variations in tumor localization, and aggressive behaviour in an immunosuppressed environment may explain this finding.
In conclusion, the majority of de novo internal neoplasms are diagnosed at advanced stages, with hematologic malignancies appearing earlier than solid tumors. Azathioprine use, alcohol, HCV and possibly, prolonged exposure to immunosuppression are risk factors for the development of these neoplasms. In addition, it is likely that new immunosuppressive agents predispose to a higher aggressivity of these tumors. Given the earlier appearance of de novo tumors observed in recent years, the immunologic benefit of intense immunosuppression schemes must be weighed against the risk of development of de novo tumors, particularly in patients undergoing liver transplantation for alcoholic and HCV-related cirrhosis.
This work was supported in part by a Grant from the Instituto de Salud Carlos III (C03/02).