Patients on renal replacement therapies are at increased risk of cancer compared with the age- and gender-matched general population (1, 2). The magnitude of the increased risk varies with the modality of renal replacement therapy, with transplantation conferring a much greater risk of cancer than dialysis (3). Compared with the age- and gender-matched population, the overall risks of cancer are at least 1.5- and 2.5-fold greater among the dialysis and transplant patients, respectively. Similar to other immunodeficient settings, the increased cancer risk in the transplant population is type specific and is highest among the virus-related malignancies, including human papillomavirus–related cervical cancers (5–10 times increased risk), Epstein-Barr virus–related posttransplantation lymphoproliferative disease (PTLD; 15 times increased risk), and human herpesvirus-8–related Kaposi sarcomas (200 times increased risk) (4–6). Although certain cancers developing in patients on dialysis and after transplantation share similar risk factors (such as hepatitis B for liver cancer), a different pattern of site-specific cancer risk is observed in the dialysis population (1). Most cancers that are more frequent on dialysis compared with the general population are associated with the severity and duration of renal failure, such as renal cancer and acquired cystic disease.
Apart from the effects of impaired immunity and chronic inflammation, the total time on dialysis had been suggested as a potential oncogenic risk factor for patients on dialysis. Previous analyses of registry data from the United States, the United Kingdom, and Australia reported a significant trend of increasing risk of certain cancer types, such as kidney cancer, with increasing time on dialysis. The highest risk of kidney cancer was found in those who had been on dialysis for more than 10 years, declining from a 7-fold excess risk to 3.7- and 3.2-fold excess risks after 5 and 1 years on dialysis, respectively (7). A single-center study also reported an association between dialysis duration and an increased risk of renal cell carcinoma after transplantation, but the effects of dialysis time and other site-specific cancers were not examined (8). We hypothesized that the influences of prolonged exposure to the uremic state associated with chronic dialysis may also contribute to cancer development after transplantation. In this study, we aimed to determine the association between the length of time on maintenance dialysis before the first kidney transplantation and the risk of site-specific and overall incident cancers after transplantation.
Over a median follow-up of 4.4 years (interquartile range, 1.7–7.7 years), a total of 6417 kidney transplant recipients were included between 1997 and 2009 and followed up for 4.4 years (interquartile range, 1.7–7.7 years), with 33,670 person-years. During which time, 454 (7.1%) patients developed cancers, 1370 (21.4%) experienced graft loss, and 785 (12.2%) died. There was a linear relationship between the duration of dialysis and the risk of cancer after transplantation (Ptrend=0.02). Table 1 shows the baseline characteristics of the total recipient population with and without cancers.
Cancer Frequency and Incidence Rate
PTLD was the most common cancer type followed by malignant melanoma, urinary tract (kidney, urothelial, and bladder), cervical, lung, prostate, and breast cancers in our study cohort. Cervical cancer (n=39; 21.1%) was the most common solid organ cancer in women followed by colorectal (n=21; 11%) and breast (n=16; 8.7%) cancers. In men, urinary tract cancer (n=49; 18.2%) was the most frequent cancer followed by PTLD (n=43; 16.0%) and malignant melanoma (n=42; 15.6%).
Effect of Duration of Maintenance Dialysis and Overall Cancer Risk
Table 2 shows the unadjusted and adjusted risk factors for cancers after the first kidney transplantation. Age and smoking status at the time of transplantation were independent risk factors for cancer in our study cohort. After adjustment for age at transplantation and other significant variables from the univariate analyses (smoking status, race, mycophenolate dose at 12 months, and immunologic status), the duration of maintenance dialysis was a risk factor for cancer after transplantation. A significant increase in cancer incidence was observed in patients with increasing duration on maintenance dialysis (Ptrend=0.03; df=3; χ2=10.3). Compared with recipients who received less than 1.5 years of dialysis before the first transplantation, the adjusted hazard ratio (HR) for all incident cancers was 1.42 (95% confidence interval [95% CI], 1.15–1.83; P=0.01) among those who had been on dialysis for more than 4.5 years before receiving their first kidney transplant.
Effects of Duration of Maintenance Dialysis and Site-Specific Cancer Risk
Figure 1 shows the site-specific risk of the eight common cancers after kidney transplantation and the association with the duration of maintenance dialysis before the first transplantation. There was a significant increased risk of cancers of the urinary tract associated with increasing duration of maintenance dialysis before the first kidney transplantation (Ptrend=0.0027; df=3; χ2=11.95). Compared with recipients who received less than 1.5 years of maintenance dialysis, the adjusted HR (95% CI) for urinary tract cancers among recipients who had been exposed to dialysis for more than 4.5 years was 2.57 (1.33–4.95; P=0.005). After adjustment for the effects of age at transplantation, smoking status, and immunosuppression use, the excess risk for lung cancer was also highest among recipients with the longest dialysis duration (adjusted HR [95% CI], 3.32 [1.0–1.14]; P=0.05). For colorectal cancer, the overall risk was greatest among recipients who had been on maintenance dialysis between 2.5 and 4.5 years (adjusted HR [95% CI], 2.57 [1.03–6.41]; P=0.043) compared with those on dialysis for less than 1.5 years before the first kidney transplantation. However, increasing duration on dialysis before transplantation was not associated with an increased risk of colorectal (Ptrend=0.22; df=3; χ2=4.38) and lung (Ptrend=0.2; df=2; χ2=3.86) cancers.
Adjusted Cumulative Incidence Probability of Cancer
Figure 2 shows the cumulative incidence function of all cancers and the duration of maintenance dialysis before the first transplantation after adjusting for competing cardiovascular-related deaths and other risk factors such as age at transplantation, smoking status, and immunosuppression use after transplantation. The adjusted cumulative incidence probability (95% CI) of having any cancers at 12 years after transplantation were 0.23 (0.17–0.30), 0.21 (0.15–0.26), 0.18 (0.10–0.26), and 0.15 (0.12–0.18) among those who had been on dialysis more than 4.5 years, between 2.5 and 4.5 years, from 1.5 to 2.5 years, and less than 1.5 years, respectively. After adjusting for competing risk, there was a significant increase in the overall trend (P=0.02) of increasing time on dialysis and the overall risk of cancer. Compared with those who had been on dialysis for more than 4.5 years, the estimated differences in the cumulative incidence probability (95% CI) of having cancer for recipients who had been on dialysis between 1.5 and 2.5 years and less than 1.5 years were 0.0934 (0.014–0.17; P=0.02) and 0.0875 (0.012–0.16; P=0.02), respectively.
The adjusted cumulative incidence probability (95% CI) of having urinary tract cancer at 12 years after transplantation were 0.051 (0.02–0.08), 0.020 (0.011–0.03), 0.027 (0.002–0.05), and 0.021 (0.005–0.04) among those who had been on dialysis more than 4.5 years, between 2.5 and 4.5 years, from 1.5 to 2.5 years, and less than 1.5 years, respectively. In contrast to the cause-specific model, there were no significant pairwise differences in the cumulative incidence cancers of the urinary tract system observed between those who had been on dialysis for a longer duration and those recipients who had been on dialysis (>4.5 years) for a shorter period of time (<1.5 years).
We have found that recipients exposed to a longer period of maintenance dialysis were at a greater risk of having cancer after transplantation. There was a linear relationship between duration of dialysis and cancer risk after transplantation, with 30% and 60% increased risks of all incident cancers if recipients had been on maintenance dialysis for more than 2.5 and 4.5 years before the first transplantation, respectively. The adjusted cumulative incidence probability (95% CI) of having any cancers (excluding nonmelanocytic skin cancers) at 12 years after transplantation were 0.23 (0.17–0.30), 0.21 (0.15–0.26), 0.18 (0.10–0.26), and 0.15 (0.12–0.18) among those who had been on dialysis more than 4.5 years, between 2.5 and 4.5 years, from 1.5 to 2.5 years, and less than 1.5 years, respectively. The excess risks for lung, urinary tract, and colorectal cancers were also highest among those who had been on dialysis for a longer time period before the first transplantation.
It is well accepted that the long-term immunosuppressive therapy used to maintain host tolerance in kidney transplant recipients contributes to the heighted cancer risk after transplantation. However, the effects on the susceptibility to cancer are limited not only in transplant recipients but also commence even before exposure to chronic immunosuppression. This is supported by robust epidemiologic evidence from registry data, suggesting that cancer risk increases by at least 1.2- and 1.5-fold in the predialysis and dialysis populations, respectively (3, 4, 6, 9). In addition to the elevated cancer risk observed in patients with more advanced stage disease, older men with mild to moderate stage chronic kidney disease (CKD) are also at risk of developing cancer by at least 1.4-fold compared with men without CKD (10). More recently, a population-based, prospective cohort study also reported a direct association between the levels of albumin-creatinine ratios and the increased incidence of cancer, especially bladder and renal cancers, independent of the effects of diabetes (11).
Although the exact etiologic causes for the increased cancer risk in CKD are unclear, several factors that relate to the effects of kidney disease have been proposed as potential oncogenic factors. Besides the uremic-induced endothelial damage leading to the initiation and progression of atherosclerosis in patients with CKD (12), biological studies have suggested that uremia and the consequent effects associated with uremic toxins may also contribute to cancer development. First, the state of uremia is associated with immunodeficiency and, in particular, impairs T-cellular immunity (13). This may explain the higher incidence of the immunodeficiency-related cancers, such as lymphoma, cervical cancer, and melanoma, found in our study and by others. Second, patients undergoing maintenance dialysis are exposed to a variety of environmental factors such as substances within the dialysate (nitrites, chloramines, and other unknown trace elements), which may be carcinogens. Therefore, longer exposure to these toxins on dialysis may further increase the incidence of cancer after transplantation independent of the effects of immunosuppression. This hypothesis is corroborated by our study findings because a dose-dependent relationship between the duration of maintenance dialysis before the first transplantation and the overall cumulative cancer risk after transplantation was observed in our population cohort using cause-specific and competing risk analyses. We also found that this association is independent of the types of dialysis modality used before the first transplantation. Third, uremia is also a state of chronic inflammation. Chronic inflammation has long been recognized as a risk factor for cancer. The presence of chronic inflammation leading to DNA mutations in proliferating cells and the deregulatory release of cytokines and chemokines are thought to be the critical elements for cancer development and progression (14–16).
Our results also show that the effect sizes from the cause-specific and the subdistribution hazards models differ for different cancers after transplantation. In the cause-specific model, urinary tract cancer incidence is increased in excess of 2.6 times greater among those on dialysis for more than 4.5 years compared with those less than 1.5 years. On the contrary, there were no significant differences in the cumulative incidence of cancers of the urinary tract systems with the various times and durations on maintenance dialysis before the first kidney transplantation after adjustment for cardiovascular-related deaths. The Kaplan-Meier methods have been widely used for estimating survival and cumulative incidence function in medical research. However, if there are more than one single event (or failures), and if these events are dependent, the Kaplan-Meier estimates may be biased (17–19). In the context of our study, cardiovascular events and cancer risk are not mutually exclusive. Factors that may have affected the cancer risk (such as uremia) can also influence the overall risk of cardiovascular-related events. Therefore, if the Kaplan-Meier method is used, the cumulative incidence of cancer in transplant recipients would be overestimated.
Our study has a number of strengths. It is based on a large, population-based cohort of patients on renal replacement therapy with very little missing baseline data and 33,670 person-years of follow-up. The prospective nature and the completeness of the data set suggest that selection and ascertainment biases in the exposure and study factors are minimized. To our knowledge, this is also the first study that explicitly assessed the effects of dialysis duration and the cumulative risk of all and site-specific cancers after transplantation.
Our study has a number of potential limitations. The lack of association between time on dialysis and the risk site-specific cancers after transplantation may be due to the limited number of cancers in individual sites within the study cohort and therefore insufficient power to detect any significant differences in the site-specific cancer risk across the varying times and durations of dialysis before the first transplantation. Despite adjustment for all confounding factors, there may be unmeasured residual effects such as the dose and duration of immunosuppression use for the treatment of primary disease, finer details of smoking habits, and alcohol consumption, which may have altered the strength and the magnitude of association between dialysis duration and cancer risk after transplantation. Although we have excluded cancers that caused the primary renal failure in our analyses, exposure factors such as ethyl benzenes and phenacetin that were responsible for causing renal dysfunction may account for cancer development after transplantation. Finally, opportunistic screening with ultrasound may have contributed to the apparent increased risk of urinary tract cancers in all recipients after transplantation. However, the effects are unlikely to be differential among recipients with varying times on dialysis before transplantation.
Implication for Clinical Practice and Future Research
Understanding an individual’s cancer risk is important for long-term clinical management and cancer prevention strategies. Our findings are consistent with emerging literature showing that having end-stage kidney disease and being on dialysis is a risk factor for disease in other organ systems, and this increased risk appears to be associated with the duration and the severity of kidney disease. Unlike cardiovascular disease where receiving a transplant may reverse and improve the cardiovascular risk profile, such as hyperparathyroidism developed during dialysis, chronic immunosuppression use will further increase the overall cancer risk after transplantation. This study therefore has implications for pretransplantation and posttransplantation care, particularly surrounding the issues of cancer screening and prevention in long-term dialysis patients with coexisting comorbidities. There had been much controversy and considerable debate about cancer screening in patients with end-stage kidney disease. Some have argued against routine screening in dialysis patients because of competing risks of cardiovascular deaths and the reduced life expectancy limits screening benefits (20, 21). Others have recommended that screening should be part of the standard clinical management because of the higher incidence of cancer, particularly for those who are waiting on the transplant waiting list by virtue of the expected increased cancer risk associated with long-term immunosuppression after transplantation. Our study findings suggest that early detection and cancer screening is probably warranted among longer-term listed dialysis patients waiting for transplantation and also for those who have received a kidney transplant after waiting for a prolonged period on dialysis. Future studies should further explore whether the mortality benefits observed through screening can be achieved in all patients with CKD regardless of age, comorbidities, and the time on dialysis.
In conclusion, increasing time on dialysis is a significant risk factor for cancers in kidney transplant recipients independent of competing events such as age and cardiovascular deaths. Strategies to improve cancer surveillance among recipients who had been on dialysis for a longer time may be warranted.
MATERIALS AND METHODS
The Australia and New Zealand Dialysis and Transplant Registry (ANZDATA) records clinical data from all patients from all renal units in Australia and New Zealand from their first use of renal replacement therapy. We included all persons within the registry who received their first deceased and live kidney donor transplants between 1 January 1997 and 31 December 2009. We excluded recipients with multiple organ grafts, who received preemptive transplants, whose primary diseases were caused by cancer such as multiple myeloma and renal cancer, and those with a history of cancer before commencement of renal replacement therapy (except for nonmelanocytic skin cancers). We have excluded patients with a prior history of cancer on dialysis because the primary aim of this study was to assess the association of time on dialysis and the risk of incident cancers after transplantation.
Ascertainment of Cancers
The ANZDATA registry also records all new cancers in all patients, except for squamous and basal cell cancers of the skin. Cancers within the registry are coded for sites and cell type using codes adapted from the International Classification of Diseases for Oncology, First Edition. Cancer records by the ANZDATA registry are robust and accurate (22). With the exception of lip cancers and multiple myeloma, previous analyses that compared the records of incident cancer diagnoses in patients on renal replacement therapy with those reported to the New South Wales Cancer Registry (a mandatory reporting requirement for all hospitals and institutions that diagnosed or treated someone with cancers in New South Wales) found good agreement between the two registries (22). We included all cancers, except nonmelanocytic skin cancers, premalignant or in situ lesions, in our analyses.
Baseline characteristics for recipients diagnosed with incident cancers and no cancers were compared with chi-square test, t test, and analysis of variance. Analysis of variance was used to compare the mean of the continuous variables between the groups of patients exposed to the various durations of dialysis before transplantation. For survival analyses, the follow-up period was defined from the time of transplantation to the time of first cancer diagnoses after transplantation. Those who did not develop cancer were censored at the time of death, graft loss, or last contact, whichever occurred first. The proportions free from incident cancers were calculated using the Kaplan-Meier method.
We applied univariate and multivariable Cox proportional hazards models to assess the relationship between incident cancers and the duration of maintenance dialysis before the first transplantation while controlling for the effects of age, sex, ethnicity, body mass index, comorbidities such as diabetes mellitus, lung disease, coronary artery disease, cerebrovascular disease, peripheral vascular disease, smoking status, primary renal disease, dialysis modality before the first transplantation, immunologic status, immunosuppression, and type of induction therapy. All explanatory variables that had an association with cancer after transplantation at P<0.25 in the unadjusted analyses were included in the multivariable-adjusted analyses. The least significant variables were then removed from the base model using a stepwise backward elimination process until variables with P values less than 0.05 remained in the final model. Potential effect modification was tested between the study factor (duration of maintenance before the first transplantation) and also other covariates using two-way interaction terms. The proportional hazards assumption of all Cox models was checked graphically by plotting the Schoenfield residuals, but there were no evidence of departures from proportional hazards for any variable. Site-specific analyses were also conducted to assess the relationship between the duration of maintenance dialysis before transplantation and the risk for common cancers such as cancers of the urinary tract systems, breast, PTLD, prostate, melanomas, lung, and colorectal cancers. Eight separate Cox regression models were developed for each individual cancer sites.
Competing Risk Analyses
As a secondary analysis, we conducted a nonparametric estimation of the adjusted cumulative incidence function of cancer after transplantation, taking into account the informative nature of censoring due to competing risk. The cumulative incidence of cancer is estimated using two main steps. We first considered the event of interest (i.e., cancer incidence) and other competing events (such as cardiovascular-related death) and then calculated the Kaplan-Meier estimate of the overall “events”. Anyone who was not experiencing the “event” (i.e., event free) was considered censored. A stratified proportional subdistribution hazards model was also developed to estimate the covariate effects on the cumulative incidence function (23). All analyses were undertaken using SAS statistical software 9.2 and R statistical software (24).
1. Vajdic CM, van Leeuwen MT, McDonald SP, et al.. Increased incidence of squamous cell carcinoma of eye after kidney transplantation. J Natl Cancer Inst
2007; 99: 1340.
2. Webster AC, Craig JC, Simpson JM, et al.. Identifying high risk groups and quantifying absolute risk of cancer after kidney transplantation: a cohort study of 15,183 recipients. Am J Transplant
2007; 7: 2140.
3. Stewart JH, Vajdic CM, van Leeuwen MT, et al.. The pattern of excess cancer in dialysis and transplantation. Nephrol Dialysis Transplant
2009; 24: 3225.
4. Birkeland SA, Lokkegaard H, Storm HH. Cancer risk in patients on dialysis and after renal transplantation. Lancet
2000; 355: 1886.
5. Grulich AE, van Leeuwen MT, Falster MO, et al.. Incidence of cancers in people with HIV/AIDS compared with immunosuppressed transplant recipients: a meta-analysis [review] [30 refs]. Lancet
2007; 370: 59.
6. Vajdic CM, McDonald SP, McCredie MR, et al.. Cancer incidence before and after kidney transplantation. JAMA
2006; 296: 2823.
7. Stewart JH, Buccianti G, Agodoa L, et al.. Cancers of the kidney and urinary tract in patients on dialysis for end-stage renal disease: analysis of data from the United States, Europe, and Australia and New Zealand. J Am Soc Nephrol
2003; 14: 197.
8. Goh A, Vathsala A. Native renal cysts and dialysis duration are risk factors for renal cell carcinoma in renal transplant recipients. Am J Transplant
2011; 11: 86.
9. Maisonneuve P, Agodoa L, Gellert R, et al.. Cancer in patients on dialysis for end-stage renal disease: an international collaborative study. Lancet
1999; 354: 93.
10. Wong G, Hayen A, Chapman JR, et al.. Association of CKD and cancer risk in older people. J Am Soc Nephrol
2009; 20: 1341.
11. Jorgensen L, Heuch I, Jenssen T, et al.. Association of albuminuria and cancer incidence. J Am Soc Nephrol
2008; 19: 992.
12. Gris JC, Branger B, Vecina F, et al.. Increased cardiovascular risk factors and features of endothelial activation and dysfunction in dialyzed uremic patients. Kidney Int
1994; 46: 807.
13. Wilson WE, Kirkpatrick CH, Talmage DW. Immunologic responsiveness in uremia. Ann Intern Med
1965; 62: 1.
14. Moss SF, Blaser MJ. Mechanisms of disease: inflammation and the origins of cancer [review] [65 refs]. Nat Clin Pract Oncol
2001; 2: 90.
15. Thun MJ, Henley SJ, Gansler T. Inflammation and cancer: an epidemiological perspective [review] [57 refs]. Novartis Found Symp
1949; 256: 6.
16. Schwartsburd PM. Chronic inflammation as inductor of pro-cancer microenvironment: pathogenesis of dysregulated feedback control [review] [53 refs]. Cancer Metastasis Rev
2003; 22: 95.
17. Zhou B, Latouche A, Rocha V, et al.. Competing risks regression for stratified data. Biometrics
2011; 67: 661.
18. Forsblom C, Harjutsalo V, Thorn LM, et al.. Competing-risk analysis of ESRD and death among patients with type 1 diabetes and macroalbuminuria. J Am Soc Nephrol
2011; 22: 537.
19. Kim HT. Cumulative incidence in competing risks data and competing risks regression analysis. Clin Cancer Res
2007; 13: t-65.
20. Holley JL. Screening, diagnosis, and treatment of cancer in long-term dialysis patients [review] [51 refs]. Clin J Am Soc Nephrol
2007; 2: 604.
21. Holley JL. Preventive medical screening is not appropriate for many chronic dialysis patients. Semin Dialysis
2000; 13: 369.
22. Webster AC, Supramaniam R, O’Connell DL, et al.. Validity of registry data: agreement between cancer records in an end-stage kidney disease registry (voluntary reporting) and a cancer register (statutory reporting). Nephrology
2010; 15: 491.
23. Zhang X, Zhang MJ. SAS macros for estimation of direct adjusted cumulative incidence curves under proportional subdistribution hazards models. Comput Methods Programs Biomed
2011; 101: 87.
24. R Foundation for Statistical Computing, Vienna, Austria, 2012.