Immunosuppression and Other Risk Factors for Lip Cancer after Kidney Transplantation

Increased rates of lip cancer have been consistently observed after kidney transplantation, as well as in another immunodeficient state, infection with HIV (1). A meta-analysis of registry-based cancer data (1) estimated that lip cancer risk is 30-fold higher among kidney transplant recipients compared with the general population. Although this observation supports its classification as an immunodeficiency-related cancer, the precise etiology of lip cancer in the context of immunosuppression is unknown. The single previous epidemiologic study of lip lesions in kidney transplant recipients (2) included only two cases of squamous cell carcinoma (SCC).

Epidemiologic studies of lip cancer in the general population are also limited. Data on risk factors for lip cancer come from large case series (3-7) and a few case-control studies (8-12). Lip cancer is causally associated with smoking (13), and an association with exposure to solar UV radiation has been suggested, although evidence for causality is insufficient (14). A role for human papillomavirus (HPV) infection is possible but remains unsubstantiated (15). The association with alcohol consumption, which is causal in cancer of the oral cavity (16), is uncertain.

Lip cancer is composed predominantly of SCC and usually arises in the epithelial layer of the vermillion, a transition zone between the inner labial mucosa and the outer skin of the lip. Although cancer of the lip is likely to share certain risk factors with cancer of the skin, the existence of a transitional epithelium, combined with its anatomic positioning, sees the vermillion vulnerable to a unique combination of exposures with established carcinogenic potential. Herein, we report risk factor data from a population-based, nationwide Australian cohort of kidney transplant recipients.

The study cohort and method of cancer data ascertainment have been described in detail previously (17). Briefly, the Australia and New Zealand Dialysis and Transplant (ANZDATA) Registry contains comprehensive information on all patients who commence maintenance dialysis or receive a kidney transplant in Australia and New Zealand (18). For this analysis, the cohort was restricted to patients in Australia registered on the ANZDATA Registry who received their first transplant between January 1, 1982 and September 30, 2003 (n = 8,173). Patients with a diagnosis of lip cancer before transplantation were excluded (n = 11); patients were not excluded based on a history of other types of cancer.

Data on all patients registered on the ANZDATA Registry are actively collected via biannual surveys of their treating dialysis or transplant unit (18). Thus, data on immunosuppression and other potential risk factors for lip cancer were directly obtained, or derived from, data held on the ANZDATA Registry.

Diagnoses of lip cancer within the ANZDATA cohort were ascertained through probabilistic data linkage with the Australian National Cancer Statistics Clearing House (NCSCH). The NCSCH compiles information on all diagnoses of primary invasive cancer, excepting nonmelanoma skin cancer (NMSC), as notified by statute to each of the eight state and territory cancer registries of Australia (19). Data on the NCSCH were available from January 1, 1982 until December 31, 2001, 2002, or 2003, depending on the state or territory of cancer registration.

For each diagnosis of lip cancer, the date of diagnosis, topography, and morphology, according to the International Classification of Diseases, 10th edition (ICD10), and International Classification of Diseases for Oncology, 2nd and 3rd editions, were obtained. All diagnoses within the ICD10 C00 rubric were included; cancers arising in the cutaneous part of the lip (ICD10 C430, melanoma of skin of lip and C440, skin of lip) were excluded. General population cancer incidence rates were also obtained for each subsite of the lip by 5-y age group, sex, calendar year, and state/territory, for each year since 1982.

Ethical approval was obtained from all relevant institutional ethics committees. The requirement for informed consent from patients was waived because the researchers received only deidentified data.

Lip Cancer Incidence by Subsite of Lip. For each patient, person-years of follow-up were accumulated from the date of first transplantation until the date of first lip cancer diagnosis, death, last contact, or the latest date to which cancer data were available, whichever occurred first. Standardized incidence ratios (SIR), comparing the number of observed cancers with that expected based on the application of general population cancer rates, were calculated with 95% confidence intervals (95% CI) for each subsite of the lip, including the upper (ICD10 C000) and lower vermillion (ICD10 C001), inner lip and commissures (ICD10 C003-C006), and lip, unspecified (ICD10 C009).

Lip Cancer Incidence by Transplant Function. SIRs for all lip cancers were calculated for periods of transplant function and for periods of dialysis after transplant failure. Immunosuppressive therapy is often ceased, and always reduced, when dialysis treatment is reinstituted after transplant failure (20). Therefore, periods of transplant function were taken as signifying the current receipt of immunosuppression and periods of dialysis were taken as signifying nonreceipt. For this analysis, follow-up and incident cancers during the first 3 mo of each period were disregarded (n = 0 cases during transplant function and n = 2 during dialysis) for all patients, as cancers diagnosed within this time would almost certainly have developed in the preceding period. For comparison purposes, the SIR was also computed for the period of dialysis before transplantation, based on all patients on the ANZDATA Registry first dialyzed since January 1, 1982 until September 30, 2003 (n = 23,764).

To address possible confounding by differences in the age-sex structure of the dialysis and transplanted populations, risk during periods of transplant function and failure was also assessed using multivariate Poisson regression. Incidence rate ratios (IRR) with 95% CIs were compared for each period relative to that during the first functioning transplant, with adjustment for age and sex. Analyses were also adjusted for the duration of transplant function (years), a surrogate for the cumulative duration of immunosuppression, as patients with transplant failure had, by definition, fewer years of continuous transplant function.

Risk Factors for Lip Cancer during the First Functioning Transplant. Risk factor analyses were confined to the first functioning transplant; that is, follow-up was censored at transplant failure and reinstitution of dialysis. Risk factors examined were exposure to solar UV radiation and those related to patients' immunosuppressive and medical history. UV exposure was assessed through several surrogate measures collected at entry onto ANZDATA, including self-reported race (Caucasoid, non-Caucasoid), country of birth (Australia/New Zealand, Europe, Asia, other/unknown), and residential latitude. Residential latitude was classified into one of three bands (<30°S, 30-35°S, ≥36°S), which broadly demarcate Australia's most populous cities, Brisbane, Sydney, and Melbourne (21). In addition, sex was considered as a surrogate measure of UV exposure; females have lower occupational exposure than do males (21) and are also afforded protection through the use of lipstick (11).

Data were available on the receipt of individual immunosuppressive agents at transplantation and at 1, 2, 3, 6, 12, 24, 36, and 60 mo after transplantation and at 5 yearly intervals thereafter. Using this information, a time-dependent variable was constructed to represent the current receipt of each agent during each of these intervals. These included the calcineurin inhibitors, cyclosporine and tacrolimus, and the antiproliferative agents, azathioprine and mycophenolate. Receipt of the steroid prednisolone was not examined given its near universal (95%) use within the cohort, nor was the receipt of the recently introduced mammalian target of rapamycin inhibitor, sirolimus, due to insufficient follow-up time. Combination therapy with azathioprine and prednisolone, with and without cyclosporine, and other triple therapies, was examined for descriptive purposes.

Other immunosuppression-related factors examined were duration, approximated by time since transplantation (time dependent, <2, 2-4, 5-9, ≥10 y), and exposures reflective of the intensity of immunosuppression. These included the receipt of lymphocyte-depleting agents, specifically, antithymocyte globulin (ATG) and the anti-CD3 agent muromonab-CD3 (time dependent from time of receipt, given for either induction immunosuppression or for treatment of acute rejection); donor source (living related, cadaveric/living unrelated); the number of human leukocyte antigen mismatches between the recipient and the donor (0-1, 2-3, 4-6); and the receipt of a different organ transplant in addition to a kidney, either at the time of or after kidney transplantation (time dependent, yes/no).

Other factors examined were current age (time dependent, single years), a history of nonlip cancer recorded on the NCSCH before transplantation (yes/no, excluding NMSC), smoking history (never, former, current), cause of primary renal disease leading to renal failure (diabetic nephropathy, primary and secondary glomerulonephritis, hypertensive/arteriopathic, other/unknown), the number of years of dialysis before first transplantation (<1, 1-2, ≥2), and diabetes comorbidity (none, type I, type II).

Poisson regression modeling was used to determine IRRs with 95% CIs for each risk factor. All variables with two-sided statistical significance (P < 0.10) in univariate analysis were considered multivariately. The final multivariate model was determined using a forward stepwise approach and included only those covariates with two-sided statistical significance (P < 0.05) after adjustment for each other and for age, sex, time since transplantation, and the current receipt of each immunosuppressive agent, which were included a priori. Statistical interaction was assessed between sex and the current receipt of either azathioprine or cyclosporine, as these immunosuppressive agents can enhance the effects of UV exposure (22-25). Multivariate analyses were conducted separately for each subsite of lip (upper and lower vermillion, inner lip) due to the likelihood of differences in exposures and thus etiology across sites. For all variables, missing data were analyzed as “unknown,” unless otherwise stated. All analyses were done using Stata version 10 (StataCorp LP).

The cohort consisted of 8,162 eligible patients (4,822 males, 3,340 females). As summarized in Table 1, most patients (88%) were of Caucasoid race. At entry onto the ANZDATA Registry, over half of the cohort resided at latitudes between 30°S and 35°S.

Smoking status at ANZDATA entry was known for approximately two thirds of the cohort, of whom 39% reported a history of smoking. A diagnosis of nonlip cancer before transplantation was observed for 202 patients and included, among other types, cancers of the kidney and urinary tract (33%), prostate (8% of male cancers), and breast (15% of female cancers). Glomerulonephritis was the single greatest (36%) cause of primary renal disease. Patients were dialyzed for a median of 1.4 years before transplantation. The median age at first transplantation was 43 years, and the majority (79%) of first transplants were of cadaveric origin. Other organ transplantations, mostly pancreatic, were received by 232 patients either at the time of or during the period of first kidney transplant function.

A small proportion of patients (11%) underwent first transplantation before 1985, half were transplanted during 1985 to 1995, and 39% were transplanted after 1995. Half the patients had at some stage received triple therapy with cyclosporine, azathioprine, and prednisolone, and 22% were administered the lymphocyte-depleting antibodies ATG and muromonab-CD3. Transplant failure with reinstitution of dialysis occurred in 2,054 (25%) patients, and 794 (10%) underwent retransplantation. Three or more transplants were received by 82 (1%) patients.

In total, 203 cases of lip cancer were identified after transplantation, all of which were of squamous cell origin and occurred in patients of Caucasoid race. The median time to diagnosis was 6.1 years after transplantation, and the 5- and 10-year cumulative incidence was 1.2% and 3.7%, respectively. Subsite of the lip was specified for 180 (89%) of cases. The majority of cases (139, 77%) occurred on the vermillion of the lower lip, 22 (12%) on the upper vermillion, and a small number (11, 6%) on the inner aspect of the lip. As shown in Table 2, SIRs were significantly raised for all subsites of the lip for both males and females. In general, SIRs were higher among female than male patients, owing to the lower population rates for lip cancer among females.

Almost all cases of lip cancer arose during periods of transplant function; the great majority (180, 90%) occurred during the first functioning transplant and 20 (10%) occurred during a higher-order transplant (Table 3). The SIR for lip cancer during all periods of dialysis after transplant failure (SIR, 2.16; 95% CI, 0.05-12.05) was significantly lower than during periods of transplant function (SIR, 52.26; 95% CI, 45.27-60.02; P = 0.001) and was comparable with that observed during the period of dialysis before transplantation (SIR, 3.44; 95% CI, 2.23-5.08; P = 0.649). The SIR for lip cancer during higher-order transplantation (SIR, 74.00; 95% CI, 45.20-114.28) was similar to that during the first functioning transplant (SIR, 50.61; 95% CI, 43.48-58.56; P = 0.107). These results were broadly replicated in multivariate Poisson regression modeling, with and without adjustment for age, sex, and the duration of transplantation (Table 3).

Risk factor analysis was conducted for cancers occurring during the first functioning transplant only; a total of 180 lip cancers were observed in 8,161 patients. Separate analyses were conducted for cancers of the lower (n = 121) and upper (n = 18) vermillion; there were insufficient cases (n = 11) on the inner aspect of the lip.

The independent risk factors for cancer of the lower vermillion (Table 4) were increasing year of age (IRR, 1.03; 95% CI, 1.02-1.05; P < 0.001), greater time since transplantation (IRR, 3.96; 95% CI, 1.63-9.64; P = 0.002 for ≥10 years since transplantation), current receipt of either azathioprine (IRR, 2.67; 95% CI, 1.39-5.15; P = 0.003) or cyclosporine (IRR, 1.63; 95% CI, 1.00-2.65; P = 0.049), and a history of smoking (IRR, 2.13; 95% CI, 1.12-4.07; P = 0.022 for current smoker at entry onto ANZDATA). Female sex (IRR, 0.29; 95% CI, 0.18-0.46; P < 0.001), non-Australia/New Zealand country of birth (P = 0.006), and increasing duration of prior dialysis (IRR, 0.86; 95% CI, 0.75-0.98; P = 0.029) were associated with significantly decreased risk. Receipt of lymphocyte-depleting agents also conferred lower risk (IRR, 0.59; 95% CI, 0.35-0.99; P = 0.048); similar results were observed for both ATG and muromonab-CD3, and irrespective of use for induction or for the treatment of rejection (data not shown). Current receipt of mycophenolate was associated with a significant reduction in risk (IRR, 0.28; 95% CI, 0.12-0.69; P = 0.006) in univariate, but not multivariate, analysis. There was no significant interaction between sex and use of either azathioprine (P = 0.739) or cyclosporine (P = 0.893).

A significant association for cancer of the lower vermillion was observed with latitude of residence (P = 0.019), although the relationship was not linear. Compared with patients residing at latitudes between 30°S and 35°S, risk among patients residing at latitudes of less than 30°S was not significantly different (IRR, 0.97; 95% CI, 0.60-1.58; P = 0.908), whereas a significantly increased risk was observed among patients residing at latitudes of 36°S or greater (IRR, 1.70; 95% CI, 1.14-2.55; P = 0.010).

Cancer of the upper vermillion was independently associated with female sex (IRR, 3.30; 95% CI, 1.24-8.78; P = 0.016) and with current use of cyclosporine (IRR, 8.77; 95% CI, 1.14-67.48; P = 0.037). Greater time since transplantation (P = 0.058) and a history of smoking (IRR, 4.63; 95% CI, 0.89-24.19; P = 0.069 for ever smoker at entry onto ANZDATA) were of borderline significance.

Risk for lip cancer in this Australian cohort of kidney transplant recipients was strongly related to the current receipt of immunosuppression, decreasing to pretransplantation level on transplant failure and reinstitution of dialysis when medical immunosuppression is usually withdrawn. During the first functioning transplant, risk for cancer of the lower vermillion was independently positively associated with increasing year of age, a history of smoking, surrogate indices of increased exposure to solar UV radiation, the duration of immunosuppression, and the current receipt of azathioprine or cyclosporine, immunosuppressive agents that can potentiate UV-related cellular damage. These data provide evidence that, in addition to the effects of immunosuppression per se, specific immunosuppressive agents can directly increase the risk for cancer of sun-exposed parts of the lip in this population.

Our finding of an association between lip cancer risk and immunosuppression accords with incidence data in this population and in those with HIV and supports the classification of lip cancer as an immunodeficiency-related cancer (1). Risk was entirely and rapidly reversible on cessation of immunosuppression, strongly suggesting a causal role (26). The reversal of risk has not been previously documented for lip cancer, nor explicitly for NMSC, although it is consistent with data on NMSC occurrence from one case series (27).

Our data suggest that lip cancer in transplant recipients shares a similar etiologic profile to that in immunocompetent patients in terms of smoking history and exposure to solar UV radiation. An association with UV exposure was implicated by the marked predilection for the lower vermillion, the part of the lip that receives the greatest exposure to direct solar UV radiation (28), and greater risk among males than females, reflecting established differences between the sexes in levels of cumulative solar UV exposure (21). Risk was also highest in patients of Caucasoid race, a function of melanin deficiency within the vermillion relative to that of darker-skinned individuals (29), and in patients of Australian or New Zealand birth. This may be due to differences in skin type but also in patients' early life exposure to solar UV radiation, as implicated in studies of migrants in which rates of lip cancer more closely reflect that of their country or region of origin (8, 30).

Latitude, an approximation of solar intensity, is strongly inversely correlated with incidence of cutaneous SCC (31). However, a linear association between lip cancer incidence and latitude was not observed in this study, nor is it observed in the general Australian population (32). One explanation for this apparent incongruence with other indices of exposure to solar UV radiation is that the anatomic positioning of the lips makes them particularly susceptible to irradiation when the sun is low in the sky (33).

In skin, exposure to solar UV radiation reduces local and systemic immune responses and this effect may be amplified in the context of iatrogenic immunosuppression (34). In addition, certain immunosuppressive agents can directly potentiate the damaging effects of UV radiation, independent of their effect on immune cells. In this study, current use of either azathioprine or cyclosporine was associated with increased risk for lip cancer. Azathioprine has been shown to sensitize DNA to UVA radiation, reducing the minimal erythema dose in skin cells of treated patients (22, 25), while cyclosporine inhibits DNA repair and apoptosis in UVB radiation-exposed human keratinocytes (23, 24). The direct contribution by these agents to cancer risk has not been fully established epidemiologically. However, in one retrospective cohort study, a significant association was reported between the presence of cutaneous SCC and receipt of azathioprine, irrespective of the receipt of other agents (35). In addition, a lower incidence of NMSC has been observed in patients randomized to a low-dose relative to normal-dose cyclosporine regimen (36) as well as in those receiving azathioprine and prednisolone compared with azathioprine, cyclosporine, and prednisolone (37, 38).

Use of lymphocyte-depleting antibodies results in the rapid, and in some cases sustained, depletion of CD4 T lymphocytes (39). Although an inverse correlation between CD4 T-lymphocyte count and risk for posttransplant NMSC has been reported (40), the role of lymphocyte-depleting agents is uncertain. Risk of cutaneous SCC was not associated with receipt of ATG or muromonab-CD3 in a large, population-based retrospective cohort study in kidney and heart transplant recipients (38). We observed a slight decrease in lip cancer risk with receipt of these agents. However, this finding may have been influenced by our modeling approach; patients remained classified as antibody exposed after first receipt despite the possibility that lymphocyte count may have returned to a level comparable with those who were not exposed.

HPV infection has been implicated in lip carcinogenesis. Although both cutaneous and mucosal subtypes have been isolated in lip SCC in immunocompetent patients (41, 42), their role is presently unclear. Persistent infection is a necessary step in HPV-associated carcinogenesis (15); the immunosuppressed seem predisposed to persistent infection, as indicated by the increased risk for cancers at all sites for which HPV is causative (1). A role for persistent viral infection in lip carcinogenesis is supported in our data by the increasing risk with increasing cumulative duration of immunosuppression.

An interaction between exposure to solar UV radiation and infection with cutaneous (β) HPV subtypes has recently been associated with increased risk for NMSC in immunocompetent hosts (43, 44). HPV oncogenes have been shown to interfere with cellular responses to UV-related DNA damage (34). This may be exacerbated in the context of immunosuppression due to persistence of infection; the use of immunosuppressive agents that directly potentiate the effects of UV radiation may also play a role. The association in transplant recipients between NMSC and history of cutaneous warts is well documented (45). The finding in immunocompetent patients that risk for lip cancer was highest in those with a sun-sensitive skin phenotype and the presence of common sporadic warts (12) suggests that a similar UV-HPV mechanism may occur in lip carcinogenesis.

This study had several strengths, including the large size of the cohort, which encompassed all kidney transplant recipients in Australia, and the nationwide population-based ascertainment of lip cancer. Therefore, patients were under follow-up unless they emigrated from Australia. The long period of follow-up, which included periods of higher-order transplant function and failure, enabled time-dependent assessment of immunosuppression. Data on the receipt of individual immunosuppressive agents were also sufficiently detailed so as to allow time-dependent examination, which is important given the practice of switching between agents, particularly in the first year after transplantation (46), and changes in treatment regimen over time (47).

There were also some limitations. Owing to diagnostic difficulties in establishing the precise site of origin of lip cancer in some instances, together with ambiguity in the applied definition of lip cancer, some cases will have been misclassified as having occurred on the cutaneous portion of the lip and vice versa. Therefore, although the lip cancer rubric specifically excludes “skin of lip,” it is possible that some misclassification will have occurred.

Cancer ascertainment will also have been affected by the validity of the data linkage technique, and the accuracy and completeness of the ANZDATA and cancer registries. The linkage methodology we used follows closely that used previously in identifying acquired immunodeficiency syndrome–related non–Hodgkin's lymphoma on the New South Wales Cancer Registry, which was 99% sensitive and 100% specific (48). The accuracy and completeness of registration on the ANZDATA Registry is routinely verified by active follow-up, and Australian cancer registration is of high quality (32).

Our risk factor analyses were limited to the information contained on the ANZDATA Registry, and consequently, the role of alcohol consumption and HPV infection could not be examined, and data on country of birth and smoking history were incomplete. In addition, for exposures collected at entry onto the ANZDATA Registry, which may have been several years before transplantation and in some instances when patients were children, misclassification is possible. We could not consider drug dose and thus the extent of immunosuppression, which may have differed between immunosuppressive agents. Nor, as described, could we model definitively the long-term effects of receipt of lymphocyte-depleting antibodies because of uncertainty in the precise duration of effect. In addition, as this was an observational study, we cannot exclude the possibility that unmeasured patient-specific factors taken into account when prescribing drug type and dose may have contributed to the findings. Lastly, our examination of exposure to solar UV radiation relied on the use of surrogate indices. However, the consistency of the association across the broad range of indices suggests that an association with actual UV exposure is highly likely.

To conclude, lip cancer occurs at markedly increased rates after kidney transplantation. Risk is strongly related to the current receipt of immunosuppression and quickly reverts to pretransplantation level on cessation, thereby supporting a role for immunodeficiency in lip carcinogenesis. Exposure to solar UV radiation, potentiated by the use of specific immunosuppressive agents, seems to be causally associated with lip cancer in this setting. This finding may also help inform the high rate of posttransplant NMSC, especially considering the reported association between UV exposure and cutaneous HPV infection in skin carcinogenesis. Our data reinforce the sun avoidance message advocated to transplant recipients and support the extension of sun protection methods to include lip sun screening agents.

No potential conflicts of interest were disclosed.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank the dialysis and transplantation units throughout Australia for their dedication and care in regularly submitting the information on which this analysis has been done, the ANZDATA staff who have created and maintained the database so accurately, the staff of the state and territory cancer registries for the use of their data, and the Australian Institute of Health and Welfare and the Cancer Council Victoria for their assistance in the conduct of this study.