- Top of page
Previous reports of cancer after kidney transplantation have been limited by small numbers of patients in single-center studies and incomplete ascertainment of cases in large registries.
We examined rates of malignancies among first-time recipients of deceased or living donor kidney transplantations in 1995–2001 (n = 35 765) using Medicare billing claims.
For most common tumors, e.g. colon, lung, prostate, stomach, esophagus, pancreas, ovary and breast, cancer rates were roughly twofold higher after kidney transplantation compared with the general population. Melanoma, leukemia, hepatobiliary tumors, cervical and vulvovaginal tumors were each approximately fivefold more common. Testicular and bladder cancers were increased approximately threefold, while kidney cancer was approximately 15-fold more common. Kaposi's sarcoma, non-Hogkin's lymphomas, and nonmelanoma skin cancers were more than 20-fold increased than in the general population. Compared with patients on the waiting list, several tumors were more common after transplantation (p < 0.01): nonmelanoma skin cancers (2.6-fold), melanoma (2.2-fold), Kaposi's sarcoma (9.0-fold), non-Hodgkin's lymphoma (3.3-fold), cancer of the mouth (2.2-fold), and cancer of the kidney (39% higher).
The rates for most malignancies are higher after kidney transplantation compared with the general population. Cancer should continue to be a major focus of prevention in kidney transplantation.
- Top of page
Long-term complications of kidney transplantation have assumed increasing importance as short-term patient and graft survival have improved. There are reasons to believe that cancer should be more common after kidney transplantation than in the general population, and more common than in comparable patients on dialysis. Immunosuppressive agents may cause DNA damage and interfere with normal DNA repair mechanisms. In addition, a number of cancers have been linked to viral infections that are common after transplantation. Finally, immune surveillance, which ordinarily prevents the growth and development of malignancies, may be impaired by immunosuppressive medications.
It is difficult to precisely ascertain the incidence of most tumors, and to compare their rates of occurrence with those in the general population, using data from small, single-center studies. There have been few reports on the incidence of cancer from transplant registries (1–5). Reporting of cancer to registries is often incomplete, and the extent that registry data underestimate the true incidence of cancer is difficult to determine.
We linked data from the United States Renal Data System (USRDS) to Medicare billing claims to detect the occurrence of cancers after kidney transplantation. In the USRDS 2003 Annual Data Report, we reported the 3-year incidences of most major post-transplant malignancies, some of the clinical correlates to cancer after transplantation, and the associations between cancer and outcomes (6). In the present report, we compare sex-specific rates of cancer after transplantation with those in the general US population, adjusting for age. We also compare rates of cancer in kidney transplant recipients with rates among comparable patients on the deceased donor waiting list.
- Top of page
Studies in the general population have led to the development of guidelines for cancer screening and prevention (9–12). Whether or not these guidelines are applicable to kidney transplantation recipients depends on the life-expectancy of the individual patient and the risk of developing cancer (13). The present results suggest that the incidences of the most common cancers (colon, lung, prostate and breast) are roughly twofold higher in the first 3 years after kidney transplantation than in the general population. This suggests that measures to reduce the risk of these malignancies, e.g. screening for colon cancer, smoking cessation, screening for prostate and breast cancer, might also be appropriate for transplant recipients whose life expectancy is not several-fold less than individuals from the general population.
Single-center studies are generally too small to compare the rates of cancer after transplantation with those in the general population, and few registry studies have compared the incidence of cancer in kidney transplant recipients with the general population (1–5). The results of these registry studies are in general agreement with those of the current study (Table 6). However, in the current study the increase in the incidence of uterine cancer, compared with the general population, was much higher than that in other studies, and it is possible that some of these were coded as cervical cancer in other studies. Why the increased incidence of bone cancer is higher in the current study compared with other studies is not clear, although in some cases metastatic disease could have been miscoded as bone cancer.
We were unable to find any correlations (by intention-to-treat analysis) between the types of initial immunosuppressive agents used and nonskin malignancies (Table 2). However, we did not examine the effects of these agents on individual tumors, and others have reported that antibody induction was associated with post-transplant lymphomas (14,15). Although there are theoretical reasons for believing that sirolimus might be associated with a reduced incidence of cancer (16), neither sirolimus, nor any other agent, were associated with nonskin cancers. However, only 6% of the study population were treated with sirolimus (Table 7), and the maximum duration of follow up in this study was only 3 years. We did find associations between the types of initial immunosuppressive agents and the incidence of nonmelanoma skin cancer (Table 2). Specifically, tacrolimus was associated with a 35% lower incidence of skin cancer, while azathioprine was associated with a 17% higher incidence (p = 0.04). Antibody induction was associated with a 17% lower risk of subsequent skin cancers (Table 2). These results should be interpreted with caution, as the selection of immunosuppressive agents is often tailored to individual patients based on risk, and it is impossible to completely remove this bias in the multivariate statistical analysis. Only randomized controlled trials can determine the effects of immunosuppressive agents on outcomes.
Table 7. Initial immunosuppression
| Microemulsion cyclosporine||16 985||47|
| Sirolimus||2241|| 6|
| Mycophenolate mofetil||22 052||62|
| IL–2 receptor antibody||6448||18|
| Other antibody||8424||24|
Cancer has been reported to be more common among hemodialysis patients compared with the general population (17). Although dialysis patients do not generally receive immunosuppressive mediations, uremia is associated with abnormalities in the immune system. For this and other reasons, it is plausible that the rate of cancer after kidney transplantation may not be different than the rate of cancer in patients with end-stage kidney disease who have not undergone transplantation. Our analysis suggests that the rates of several cancers, which are increased after transplantation compared with the general population, are similar vs. comparable patients on the waiting list. However, several other tumors (e.g. skin, kidney and lymphomas) were more common after transplantation than on the waiting list (Table 5).
There is increasing evidence that viruses may play an important role in the pathogenesis of skin cancers (18,19), cervical carcinomas (20,21), Kaposi's sarcomas (22,23), hepatobiliary tumors (24,25), and lymphomas (26,27). This may help to explain why these malignancies are particularly common after kidney transplantation. The high prevalence of viral hepatitis in the dialysis population may explain why hepatobiliary tumors are higher in transplant recipients compared with the general population, but not higher compared with the waiting list. The pathogenesis of renal cell carcinoma is unclear, but it is well known that kidney cancer often occurs in end-stage kidney disease (28). The present results suggest that kidney cancer is even more common after transplantation compared with patients on the waiting list.
The incidence of nonskin cancer is higher among transplant recipients who are older, male, White (compared with Black recipients, Asians, and other ethnic groups), non-Hispanic (compared with Hispanic), and lower among patients who had diabetes as a cause of kidney disease (Table 2). These findings are in general agreement with those of another, recent study (15). It is not clear why diabetes was associated with a lower risk, although we could speculate that a lower rate of cigarette smoking among diabetics could reduce the overall risk of cancer (29). We also found that the adjusted incidence of cancer was not different in patients transplanted in 1999–2001 compared with 1995–1998 (Table 2). In another recent report the incidence of nonskin cancers increased over time (15). This analysis used data reported to the United Network for Organ Sharing, and did not adjust for the type of immunosuppressive agents used (15). It is possible that these differences, as well as differences in the populations analyzed (the current analysis included only Medicare beneficiaries), explain these different results.
A major limitation of this study (and other registry studies) is the relatively short duration of follow up (3 years). An inherent limitation of registry data comes from the fact that centers may inconsistently report events that occur late after transplantation, when patients have returned home and the transplant center can no longer follow patients closely. In the current study, we did not rely on reporting from centers to detect cancers, but rather relied on physicians, hospitals and clinics to bill for services related to the diagnosis and treatment of cancer. Nevertheless, reliance on Medicare claims limited our analysis to the first 3 years after transplantation.
Another limitation of this analysis is that the population of patients who have Medicare as their primary provider is not a random sample of the whole kidney transplant population. Thus, the results of this analysis may not be entirely applicable to the whole transplant population. Nevertheless, the Medicare primary provider population is a sizeable portion of the whole transplant population (47% in this study), and the populations analyzed in this study were similar compared with those excluded from the analysis (Table 8).
Table 8. Comparison of transplantation recipients included or excluded from the present analysis
|Patients' characteristic||Included (n = 35 765)||Excluded (n = 40 702)||p-value (χ2)|
| 18–34||19%||20%|| 0.0003|
| 50–64||34%||33%|| 0.0032|
| 65 +||12%||4%||<0.0001|
| Asian||5%||5%|| 0.1210|
| Other||3%||3%|| 0.7143|
| Female||40%||41%|| 0.0235|
| Male||60%||59%|| 0.0235|
|Primary cause of kidney failure|
| Glomerulonephritis||24%||24%|| 0.7898|
| Cystic kidney disease||7%||10%||<0.0001|
|Pre–transplantation dialysis time|
| <1 years||13%||31%||<0.0001|
| 1 ≤ 2 years||20%||22%||<0.0001|
| 2 ≤ 3 years||21%||10%||<0.0001|
| 3+ years||43%||10%||<0.0001|
|Education (highest level attained)|
| Unable to work||39%||24%||<0.0001|
| Able to work||45%||60%||<0.0001|
| Unknown||16%||16%|| 0.5736|
| Microemulsion cyclosporine||47%||48%|| 0.0562|
| Sirolimus||6%||6%|| 0.4306|
| Mycophenolate mofetil||62%||62%|| 0.3428|
| IL-2 receptor antibody||18%||20%||<0.0001|
In summary, this analysis suggests that most tumors occur more frequently after kidney transplantation than in the general population, and some cancers occur more frequently than in comparable patients on the transplantation waiting list. Altogether, these results suggest cancer should continue to be a major focus of prevention in kidney transplantation.