Notice: Wiley Online Library will be unavailable on Saturday 30th July 2016 from 08:00-11:00 BST / 03:00-06:00 EST / 15:00-18:00 SGT for essential maintenance. Apologies for the inconvenience.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Transplantation and subsequent continuous immunosuppression have been long associated with an increased risk for malignancies.1–6 Several studies, mainly on renal transplant recipients, have reported increased cancer incidence rates in comparison with the general population.6–17 In liver transplant recipients, posttransplant malignancies have been shown to occur in 5%-16% of patients,18–41 depending on the length of follow-up, and posttransplant malignancies are a major cause of death in these patients.18, 42–46
Although numerous studies have reported incident cancers after liver transplantation, only a few studies have compared the incidence with an age-matched and gender-matched control population. Results from single-center studies may be noncomparable because of ethnical and demographical variations, dissimilar inclusion criteria for different cancer types, and, importantly, variations in the immunosuppression regimen. Furthermore, there exist controversies regarding the risk of some cancer types in these studies. Because of these reasons, more studies are needed to obtain reliable data on cancer risk patterns in an attempt to reach consensus on optimal monitoring of immunosuppression, cancer surveillance programs, and strategies to minimize cancer risk and thus ultimately improve patient survival and quality of life.
The aim of this study was to describe the cancer risk pattern in Finnish liver transplant patients. The hypothesis was that the incidence of some cancer types is increased among liver transplant patients. Possible risk factors for the development of cancer and the impact of different forms of cancer on outcome were also studied.
All liver transplant patients from the Helsinki University Central Hospital (Finland) transplanted between 1982 and 2005 were included in the study. Clinical data, consisting of age, gender, indication for transplantation, initial immunosuppressive agents used, acute rejection episodes, treatment with monoclonal or polyclonal antibodies, and cytomegalovirus status, were collected from the Finnish liver transplant registry and patient records.
The cohort was compared with the national population register, and the correct personal identification number and data on vital status were achieved for every cohort member. All residents of Finland since January 1, 1967 have a unique personal identification code that is used in all main registers in Finland. Follow-up for cancer through the files of the population-based Finnish Cancer Registry was done automatically with the personal identifier used as a key. Follow-up for cancer started at the date of first transplant and ended at death or on December 31, 2005, whichever was first. Further division was made by the time that elapsed since liver transplantation. In the subanalysis among patients with acute rejection, follow-up for cancer started at the date of rejection.
The numbers of observed cases and person years at risk were counted, by 5-year age groups, separately for 4 calendar periods (1982-1987, 1988-1993, 1994-1999, and 2000-2005). The expected numbers of cases for all cancers combined and for specific cancer types were calculated by the multiplication of the number of person years in each gender and age group by the corresponding cancer incidence rate in all of Finland during the period of observation. The specific cancer types a priori selected for the analysis included cancer sites with known or suspected exceptional risk in earlier studies and other common cancer types to give the whole picture of the cancer situation among Finnish liver transplant patients.
All patients received calcineurin inhibitor–based initial immunosuppression. The majority received cyclosporine-based therapy combined with azathioprine and methylprednisolone. Only some patients, participating in controlled clinical trials, had tacrolimus-based initial immunosuppression. Furthermore, some immunologically unstable patients, particularly patients presenting with a recurrent early acute rejection episode after steroid-treated acute rejection and patients with late acute rejection episodes, were converted from cyclosporine to tacrolimus. Mycophenolate mofetil was added to some patients with calcineurin inhibitor–induced renal toxicity, and calcineurin inhibitor doses were then reduced or withdrawn. The initial target of cyclosporine concentration was 200-250 ng/mL, decreasing with time to maintain a level of 70-150 ng/mL; for tacrolimus, these values were 15-20 and 5-10 ng/mL, respectively.
All acute rejection episodes were histologically confirmed. Acute rejection episodes were initially treated with steroids (3 mg/kg/day) for 5 days. Steroid-resistant rejections were treated with OKT3 monoclonal antibodies or in some cases with anti-thymocyte globulin polyclonal antibodies. These antibodies were also in a few cases used as induction therapy when calcineurin inhibitor–based initial immunosuppression was considered too risky.
The standardized incidence ratio (SIR) was calculated by the division of the observed number of cases by the expected number. The 95% confidence intervals (CIs) for the SIR were based on the assumption that the number of observed cases followed a Poisson distribution. Potential risk factors (transplant indication, initial calcineurin inhibitor agent, presence of acute rejection, antibody therapy, retransplantation, and cytomegalovirus status) for the development of cancer were considered in a Cox proportional hazards model adjusted for age, gender, and time since transplantation. The same analysis was also done separately for the development of lymphoma, skin cancer, and other cancer. Survival functions and cumulative incidence rates were calculated with the Kaplan-Meier method. P values < 0.05 were considered statistically significant. Statistical analyses were performed with SPSS statistical software version 14.0 (SPSS, Inc., Chicago, IL).
The cohort consisted of 540 liver transplant patients, and this produced 3222 person years of follow-up. The mean age of patients was 43 years (standard deviation 18 years), and the mean length of follow-up was 6.3 years (range 0-24 years). Demographic and clinical features of the cohort are presented in Table 1. The 1-, 5-, 10-, and 20-year patient survival rates were 88%, 80%, 70%, and 59%, respectively, and 402 patients were alive at the closing date of follow-up (December 31, 2005).
Table 1. Pretransplant and Posttransplant Demographic Characteristics for All Patients and Separately for Patients Who Developed Posttransplant De Novo Cancer
All Patients (n = 540)
Patients with Posttransplant Cancer (n = 36)
In 2 patients, a liver tumor was also detected.
Eight patients did not have either cyclosporine or tacrolimus.
All rejections occurred before detection of cancer.
A total of 39 posttransplant de novo cancers were found in 36 patients. In addition, there were 11 cases of basal cell carcinoma of the skin that were not included in the total number of cancers. The mean time from transplantation to detection of cancer in these patients was 61 months (range 4-172 months). The cumulative incidence of posttransplant de novo cancer at 1, 5, 10, and 20 years was 3%, 5%, 13%, and 16%, respectively (Fig. 1). The respective cumulative de novo cancer–related mortality rates were 0%, 1%, 2%, and 2%. Eighty percent of the posttransplant cancers were detected at an age of 45-74 years.
Four of 8 cases of non-Hodgkin lymphoma were reviewed as posttransplant lymphoproliferative disorder. The mean time from transplantation to detection of posttransplant lymphoproliferative disorder (48 months, range 10-88 months) was similar to that for other non-Hodgkin lymphomas (31 months, range 10-76 months).
There were 11 patients with diagnosed nonliver cancer before liver transplantation. These included 1 lymphoma, 1 ventricle carcinoid, 3 colon cancers, 1 rectal cancer, 1 pancreatic cancer, 1 spinocellular skin cancer, 1 medullar meningeoma, 1 case of rectal carcinoid with metastasis to the liver, and 1 case of thymoma with later hepatocellular carcinoma. None of these 11 patients developed de novo cancer post-transplant. Cancer recurrences after transplantation were not accepted as outcome events in this study.
The overall SIR, in comparison with the general population, was 2.59 (95% CI 1.84-3.53; Table 2). Besides the cancer types selected for the SIR analysis, there were 4 additional cases of cancer: 1 in the urinary bladder, 1 in the ovaries, 1 in the body of the uterus, and 1 fibrosarcoma. Non-Hodgkin lymphoma, nonmelanoma skin cancer, and basal cell carcinoma were the only types to show significantly elevated SIRs (Table 2). Nonmelanoma skin cancer included Kaposi sarcoma and squamous cell carcinoma. Basal cell carcinoma, which was not included as cancer in the overall SIR, also showed a significant excess.
Table 2. Observed and Expected Numbers of Cancer Cases and Standardized Incidence Ratios with 95% Confidence Intervals Among the Finnish Liver Transplant Patients in 1982-2005 by Primary Site
The overall SIR was higher for males (SIR 4.16, 95% CI 2.61-6.30) than for females (SIR 1.74, 95% CI 1.01-2.78). Six of 8 observed non-Hodgkin lymphoma cases occurred in males. The SIR for non-Hodgkin lymphoma was also higher for males (SIR 26.7, 95% CI 9.79-58.1) than for females (SIR 5.72, 95% CI 0.69-20.7). Nonmelanoma skin cancer showed similar SIRs for both males (SIR 36.2, 95% CI 9.87-92.7) and females (SIR 40.2, 95% CI 14.8-87.5).
The SIR was more elevated in children (SIR 18.1, 95% CI 2.19-65.5) than adults (SIR 5.77, 95% CI 1.87-13.5 for ages of 17-39 years and SIR 2.27, 95% CI 1.55-3.20 for ages ≥ 40 years). The SIR for non-Hodgkin lymphoma was 123 (95% CI 3.12-686) for ages < 17 years, 55.7 (95% CI 6.74-201) for ages of 17-39 years, and 9.42 (95% CI 3.06-22.0) for ages ≥ 40 years. Nine of 10 observed nonmelanoma skin cancer cases occurred among the oldest transplant patients (≥40 years at transplantation); the respective SIR for this group was 36.1 (95% CI 16.5-68.5).
SIRs were further calculated according to the posttransplant follow-up period, namely, <2 years, 2-9 years, and ≥10 years. As shown in Table 3, the SIR value was higher in the earlier periods. All detected non-Hodgkin lymphomas and nervous system cancers (all meningeomas) occurred before 10 years post-transplant, whereas nonmelanoma skin cancers occurred quite uniformly during all follow-up periods.
Table 3. Standardized Incidence Ratios According to the Posttransplant Follow-Up Period for All Sites and for Cancer Sites with At Least 2 Observed Cases During the Whole Follow-Up
Time Since Liver Transplantation
Data are presented as standardized incidence ratio (95% confidence interval).
Patients with acute liver failure (113 patients) had higher SIR (3.35, 95% CI 1.61-6.16) than patients with chronic liver disease (381 patients, SIR 2.39, 95% CI 1.58-3.48) or a liver tumor (46 patients, SIR 2.47, 95% CI 0.30-8.92).
Cyclosporine was used in 82% of patients (441 patients) as the initial calcineurin inhibitor agent (Table 1). Similarly, 92% of cancer cases (36/39 cases) were detected in this group of patients. Also, the cyclosporine group produced 2938 patient years of follow-up versus 276 person years in the tacrolimus group. The overall SIR, in comparison with the general population, for patients with cyclosporine as the initial agent was 2.61 (95% CI 1.83-3.61), and cancer sites that showed statistically significantly increased SIRs in this group were non-Hodgkin lymphoma (SIR 13.36, 95% CI 5.37-27.5), nonmelanoma skin cancer (SIR 41.79, 95% CI 20.04-76.84), and basal cell carcinoma (SIR 4.06, 95% CI 2.03-7.26). The overall SIR for patients with tacrolimus as the initial agent was 2.32 (95% CI 0.48-6.79) in comparison with the general population. No cancer site showed significantly increased SIRs in the tacrolimus group.
The groups of patients with acute rejection (276 patients) had lower SIR (1.77, 95% CI 0.95-3.02) than the group of patients without rejection (261 patients, SIR 3.52, 95% CI 2.39-5.27). Of patients with acute rejection, 17% (47/276 patients) had been treated with monoclonal or polyclonal antibodies. SIR values for this group (1.71, 95% CI 0.35-4.98) did not markedly differ from that of the group which had not been treated with antibodies (SIR 1.79, 95% CI 0.86-3.29).
The Cox proportional hazards model was used to make subcategories comparable in terms of gender, age, and length of follow-up (Table 4). A history of acute rejection was associated with a relative cancer risk estimate < 0.5 in comparison with those without rejection; the hazard ratio for non–antibody-treated rejection episodes was statistically significant (P = 0.02). Antibody therapy for acute rejections did not alter the cancer risk significantly. Antibody induction therapy, on the other hand, increased the relative risk of cancer 4.32-fold (P = 0.004) and the relative risk of skin cancer 5.98-fold (P = 0.01) in comparison with patients that had not been treated with antibodies (Table 4).
Table 4. HRs for the Development of Cancer After Liver Transplantation According to Cox Proportional Hazard Analysis Adjusted for Age, Gender, and Time Since Transplantation
Twenty-five patients developed posttransplant noncutaneous cancer. In a retrospective examination of their patient records, cancer was detected during routine follow-up in only 3 of these cases. The case of breast cancer was detected during routine mammography, the case of rectal cancer was detected during protocol colonoscopy due to the patient's colitis ulcerosa, and the case of fibrosarcoma was detected during a routine visit to the physician. The last patient had increased follow-up due to unclear hypersedimentation and hematuria. In the remaining 22 patients, cancer was detected after investigations due to a symptom.
Of the 36 patients that developed posttransplant cancer, 14 had died by the end of follow-up. Of these 14, 7 patients had their respective de novo malignancy as the cause of death. Five patients died of a solid organ tumor [1 ovarian, 1 pancreatic, 1 rectal, 1 brain (lymphoma), and 1 prostate cancer], and 2 died of posttransplant lymphoproliferative disorder. The remaining 7 patients had a cause of death not directly related to their cancer. One patient died of fungal infection after being diagnosed with Kaposi sarcoma 2 months earlier. One patient died of cerebral infarction, 2 died of kidney failure, 1 died of myocardial infarction, 1 died of noncompliance to immunosuppressive therapy, and 1 died of coronary disease.
This study included all liver transplant patients from Finland, with up to 24 years from transplantation and no patient lost to follow-up. The identification of patients and follow-up for deaths and emigration are complete for the period of this study. The cancer registration system in Finland is virtually complete,47 and the computerized record linkage procedure is precise.48 Therefore, technical incompleteness does not cause bias in the results.
On the basis of our data, 1 of 6 liver transplant patients is estimated to develop some form of cancer by 20 years after transplantation (Fig. 1). This study also shows that liver transplant patients have a 2.6-fold cancer incidence in comparison with the normal population. The immunosuppressive agents and doses used in Finland have not markedly changed over the years,49–50 but the SIR will decrease along with higher proportions of patients with longer follow-up.
The overall SIR value in our study (2.6) is somewhat lower than that reported in most similar studies. Haagsma et al.21 reported an overall SIR of 4.3, Sheiner et al.18 reported a SIR of 3.2, and Herrero et al.51 reported a SIR of 3.2 for nonskin cancers and 16.9 for skin cancers. Only Oo et al.41 reported a lower SIR (2.1) in their study from England and Wales in comparison with the present study. Variations in these results may be explained by several factors. First, the immunosuppression regimens differ between transplant centers. Second, some studies have not included children,21 who have the highest relative cancer risk. Third, indications for liver transplantation vary. For instance, hepatitis C has been associated with a higher risk of posttransplant lymphoproliferative disorder,52–54 and at our center, fewer than 3% were transplanted for viral hepatitis. The extremely low number of patients with chronic viral hepatitis (<3%) and alcoholic liver disease (10%) in our transplant population also increases the relative number of acute patients (21%) to proportions higher than those seen in most transplant populations. Finally, the length of follow-up and number of patients included differ widely between studies,18, 21, 41, 51 and the relative cancer risk decreases along with longer follow-up times. Relative risk estimates derived from different studies are also difficult to compare because of differences in cancer incidence in the general populations used as reference rates. The higher the incidence rates in a reference population are, the lower the SIR estimate corresponding to a same given absolute incidence of posttransplant cancer will be.
Kyllönen et al.9 from our center studied cancer incidence after kidney transplantation in Finland, using population incidence rates as references similar to those used in our study. In their study, 22% of patients developed cancer by 20 years, and the overall SIR was 3.3. Cancer types with the highest SIRs, namely skin (SIR 39) and lip cancer (SIR 23), were similar to those in our study. However, Kyllönen et al. reported a higher incidence for cancer in the small bowel (SIR 12) and thyroid gland (SIR 8.1) but a rather low SIR for lymphomas (SIR for nodal lymphomas 4.8). The last observation corresponds to previous studies,6, 46, 55 which have reported a risk of non-Hodgkin lymphoma up to 8.4 times higher in non–kidney transplant patients versus kidney transplant patients.6 This fact has been explained by the larger percentage of pediatric patients in the liver transplant population.55 It must be pointed out that the study by Kyllönen et al. included 20,817 patient years of follow-up, which is more than 5 times that of our cohort. Also, kidney transplants have been performed during a longer time in Finland (since 1967), and the immunosuppression regimen has varied through time and is different from that used for liver transplant patients.
As in other studies on liver transplant patients,18, 19, 21, 26, 28, 29, 31, 41, 42, 56, the most common cancer types in our cohort were lymphoma and skin cancer. Non-Hodgkin lymphoma, which included 4 cases of posttransplant lymphoproliferative disorder, occurred more frequently in males, in patients transplanted at a younger age, and soon after transplantation. Some studies have reported that the use of an alternative immunosuppressive regimen, such as mycophenolate mofetil, is associated with a lower risk for leukemia and lymphoma.57–61 In light of our results, such alternative immunosuppressive strategies might be most suitably considered for the susceptible group of patients mentioned previously.
Nonmelanoma skin cancer, on the other hand, occurred more often at an older age, and the risk did not vary between genders or with the time after transplantation. Induction therapy with antibodies was associated with a 6-fold risk of skin cancers in comparison with patients that had not received antibodies. Routine check-up of the skin among both men and women at all times after transplantation, especially at older ages and among patients with a history of antibody induction therapy, is therefore important. Transplant patients are informed about the higher risk of skin cancers associated with organ transplantation, and the skin is routinely examined at follow-up, but it is unclear to what extent this results in detection of malignant lesions at earlier stages.
Besides lymphoma and skin cancer, other cancer types were scarce in the number of cases observed in this study. Thus, it is understandable that data on rare types of cancers could not be reliably studied. It seems notable, however, that meningeoma seemed to occur predominantly in the early posttransplant period, perhaps as a result of an accelerated growth of a preexisting, asymptomatic tumor caused by immunosuppression. Also noteworthy is the finding that transplantation from a cytomegalovirus-positive donor to a cytomegalovirus-negative recipient was associated with a 1.8-fold increase in total cancer risk, although statistical significance was not achieved in this study.
An interesting finding was the less increased cancer incidence among patients with a history of acute rejections, which most strongly correlated with lymphomas. This association seemed to be especially true for non–antibody-treated acute rejections. On the basis of the study setting used here, an explanation for this association is difficult to provide. Sanchez et al.29 reported that excessive immunosuppression to the point of no rejection episodes is expected to increase the incidence of de novo tumors. Most acute rejection episodes at our center occur within 3 weeks after transplantation, and therefore differences in posttransplant immunosuppression levels during such a short period are unlikely to explain the differences in cancer risk. What can be concluded at least is that heavy short-term immunosuppressive therapy with steroids and even antibodies, such as for acute rejection episodes, does not seem to increase the risk of cancer in these patients.
Nonskin cancer was detected during routine follow-up in only 12% (3/25) of cases in this study. Herrero et al.51 reported correspondingly that 3 of 28 patients that developed nonskin cancer were detected in surveillance studies, despite an aggressive surveillance protocol at their center. Haagsma et al.21 reported that half of all cancer cases (including skin cancers) were detected during yearly visits to the transplant center but did not reveal whether these patients presented symptoms or not. The protocol for posttransplant cancer surveillance at our center includes routine colonoscopies every 2-3 years among patients transplanted for primary sclerosing cholangitis. A regular check-up at the gynecologist is advised for all female patients. Tumor markers are studied when survival time exceeds 15 years, and intense investigations are initiated if some symptom develops at any time during follow-up. It is unclear whether a more intense cancer surveillance strategy would have detected any of the cancer cases at an earlier stage.
In conclusion, cancer incidence among liver transplant patients is higher than that in the general population. The most common cancer types are nonmelanoma skin cancer and non-Hodgkin lymphoma. Male gender, young age, and the immediate posttransplant period present as risk factors for non-Hodgkin lymphoma, whereas old age and antibody induction therapy increase the risk for skin cancers.
This study points out the importance of cancer surveillance after liver transplantation as well as the need for innovative immunosuppression strategies associated with less cancer risk.