Kaplan-Meier analysis of risk of skin cancer in heart transplant recipients after transplantation. BCC indicates basal cell carcinoma; NMSC, nonmelanoma skin cancer; and SCC, squamous cell carcinoma.
Cumulative azathioprine (A), cyclosporine (B), and prednisone (C) doses in patients with basal cell carcinoma (BCC), squamous cell carcinoma (SCC), and no nonmelanoma skin cancer (NMSC). Brackets indicate 95% confidence intervals for the mean values.
Fortina AB, Piaserico S, Caforio ALP, Abeni D, Alaibac M, Angelini A, Iliceto S, Peserico A. Immunosuppressive Level and Other Risk Factors for Basal Cell Carcinoma and Squamous Cell Carcinoma in Heart Transplant Recipients. Arch Dermatol. 2004;140(9):1079-1085. doi:10.1001/archderm.140.9.1079
To examine risk factors for the development of squamous cell carcinoma (SCC) and basal cell carcinoma (BCC) in a cohort of heart transplant (HT) recipients and, in particular, to evaluate the role of the cumulative doses of different immunosuppressive drugs.
Prospective nonconcurrent study.
A dermatology clinic at a university hospital.
A total of 230 HT recipients 18 years or older at the time of transplantation with at least 3 years of follow-up.
Main Outcome Measures
The risk of SCC and BCC in HT recipients and the relationship between development of SCC and BCC and cumulative doses of different immunosuppressive agents, controlling for other potential risk factors (age, sex, sunlight exposure, skin type, and presence of warts).
The cumulative immunosuppressive drug dose 3 years after transplantation (calculated by a weighted linear combination of azathioprine, cyclosporine, and corticosteroid cumulative doses [WLC]) was independently associated with an increased risk of developing SCC but not BCC. On multivariate analysis, patients receiving a WLC higher than the 75th percentile 3 years after HT had a 4 times higher risk of SCC than recipients of a WLC lower than the 50th percentile 3 years after HT (95% confidence interval, 1.4-11.4; P = .008). Other significant risk factors for SCC development were older age at transplantation and a greater occupational sunlight exposure. The risk of developing BCC was only associated with older age at transplantation and skin type II.
The risk of SCC but not of BCC in HT recipients was related to the level of global immunosuppression rather than to 1 specific drug. The level of immunosuppression should be kept as low as possible consistent with survival and function of the transplanted organ.
Nonmelanoma skin cancers (NMSCs) are the most frequent malignancies in organ transplant recipients.1- 3 A reversal in the squamous cell carcinoma(SCC)–basal cell carcinoma (BCC) ratio usually observed in the general population has been described.4- 6 Clearly, immunosuppression is the critical feature differentiating organ transplant recipients from the general population.
Investigations conducted to identify the association between NMSC and specific immunosuppressive drugs have yielded conflicting results.1,7- 11 Recent studies suggest that the risk of NMSC in organ transplant recipients may be related to the cumulative immunosuppressive load rather than to a specific agent.11- 14 However, scarce data are available on the relationship between incidence of NMSC and cumulative dose of different immunosuppressive agents. In fact, many studies are registry based, and the cumulative dose is often impossible to estimate accurately because each patient changes drug intake frequently, particularly during the first months.
Given that the risk of skin cancer has been reported to increase after 3 years after transplantation1,7 and to be uncommon in patients younger than 18 years,2,15 we assessed the risk of SCC and BCC in a cohort of adult heart transplant (HT) recipients observed at our institution for at least 3 years, and we evaluated the role of potential risk factors. In particular, we investigated the influence of the cumulative dose of immunosuppressive therapy on the development of SCC and BCC.
We studied 230 HT recipients (198 men [86.1%]; mean ± SD age at transplantation, 50.2 + 11.0 years) with at least 3 years of follow-up (mean ± SD follow-up, 9.1 + 3 years; range, 3.0-15.6 years) and aged 18 years or older at transplantation. Quarterly, each patient underwent regular dermatologic follow-up visits by the same experienced dermatologist who, on each occasion, performed a thorough total-body examination.
Excision and skin biopsy specimens were examined in the pathology department. The histopathologic diagnoses of BCC, SCC, Bowen disease, and solar keratosis were established in accordance with accepted criteria.16
A standard questionnaire and examination sheet were completed for each patient indicating baseline demographic features, date of transplantation, skin type according to the Fitzpatrick standard criteria,17 history of occupational sunlight exposure, and number of summer holiday weeks spent in the sun during childhood, adulthood, or lifetime. For statistical purposes, patients with skin type III to V were grouped together. There were no patients with skin type I or VI. Each patient's cumulative lifetime sun exposure was calculated in hours.
Heart transplant recipients were treated with cyclosporine and azathioprine (double therapy; n = 37), or with cyclosporine, azathioprine, and oral prednisone (triple therapy; n = 193). Oral prednisone was added to the double therapy regimen of cyclosporine and azathioprine in cases of repeated or persistent rejection or of cyclosporine nephrotoxic effects.
Graft rejection was monitored by endomyocardial biopsy following established protocols.8 Endomyocardial biopsy specimens were obtained via the right internal jugular vein using the Caves-Schultz bioptome and graded according to the International Society for Heart and Lung Transplantation (ISHLT) standardized grading system.18 Acute rejection episodes, defined as ISHLT grade higher than 2, were treated with intravenous administration of methylprednisolone in 189 patients (82.2%). Acute rejection episodes with presence of symptoms were additionally treated with antilymphocyte globulin or antithymocyte globulin in 56 patients (24.3%). Daily drug doses adjusted for body weight were recorded for each patient. Cumulative azathioprine, cyclosporine, and oral prednisone doses at 3 and 6 months and at 1, 2, and 3 years after transplantation were then calculated in milligrams per kilogram of body weight. Also, cumulative intravenous methylprednisolone at 1 year and cumulative total steroid load at 1 year were obtained after conversion of each methylprednisolone dose to an equivalent oral prednisone dose (4 mg of methylprednisolone = 5 mg of oral prednisone).
To evaluate the global immunosuppressive load, we calculated the weighted linear combination of the cumulative azathioprine, cyclosporine, and oral prednisone dose (WLC) at each time point, where WLC = a × the cumulative azathioprine dose + b × the cumulative cyclosporine dose + c × the cumulative oral prednisone dose.19 The values of weights a, b, and c are between 0 and 1, and their sum is 1. The constants a, b, and c were assigned to azathioprine, cyclosporine, and oral prednisone, respectively, to get the same weighted mean dose for each drug (a × the mean azathioprine cumulative dose = b × the mean cyclosporine cumulative dose = c × the mean oral prednisone cumulative dose) at each time point, assuming that the relevance on the risk of cancer occurrence of these different drugs is similar. As an example, after 3 months, the mean azathioprine cumulative dose in our patients was 140.5 mg/kg; the mean cyclosporine cumulative dose was 577.7 mg/kg; and the mean oral prednisone cumulative dose was 18.1 mg/kg. To obtain the same weighted mean dose for the 3 different drugs, we calculated as follows: a = 0.111; b = 0.027; and c = 0.862. The WLC of a patient with an azathioprine, cyclosporine, and oral prednisone 3-month cumulative dose of 266 mg/kg, 895 mg/kg, and 8 mg/kg, respectively, would therefore be 266 × 0.111 + 895 × 0.027 + 8 × 0.862 = 60.591 mg/kg.
Statistical analyses were performed using the statistical package SPSS/PC+ version 11.0 for Windows.20 Results are expressed as median and 25th to 75th percentile unless otherwise specified. Comparisons across groups were made using the Mann-Whitney test. The ordinal data were analyzed using the Pearson χ2 test and, in situations with a low cell count, the Fisher exact test. The 2-tailed Kendall rank test was used to correlate quantitative data.
The survival functions and the probabilities of having developed NMSC at 5, 10, and 15 years after transplantation were computed separately for SCC and BCC for all levels of all variables of interest using the nonparametric Kaplan-Meier product-limit survival estimates.21 Differences between Kaplan-Meier survival curves were analyzed using the Mantel-Haenszel log-rank test. Before applying the Cox proportional hazards method to estimate the possible independent role of the different factors of interest, we checked the necessary assumptions of proportionality of risks. We produced graphs of the log(−logS(t)] vs time, and we verified that the curves so obtained were approximately parallel.22 Results are expressed with hazard ratios (ie, antilog of the coefficient of Cox regression) and their associated 95% confidence intervals (CIs).
Among the 230 HT recipients entered into the study, 48 (20.8%) developed a total of 120 NMSCs after transplantation: 26 patients had invasive or in situ SCCs, 13 had BCCs, and 9 had both tumors. Of the 120 NMSCs, 83 were SCCs, and 37 were BCCs. The SCC/BCC ratio was 2.2:1; most lesions occurred on the head and neck (n = 91; 69.5%); the remaining lesions were on the trunk (n = 18; 13.7%), the upper limbs (n = 20; 15.3%), and the lower limbs (n = 2; 1.5%). A higher percentage of SCCs (88.8%) than BCCs (80%) developed on sun-exposed areas of the body, although this difference did not reach statistical significance. No metastatic or fatal tumors occurred within the study group.
The cumulative incidence of NMSCs in HT recipients by Kaplan-Meier method increased from 11% (95% CI, 7%-15%) after 5 years to 27.5% (95% CI, 20.7%-34.3%) after 10 years. In particular, the cumulative incidence was 6.5% (95% CI, 3.2%-9.7%) and 19.7% (95% CI, 13.3%-26.1%), respectively, for SCCs and 6% (95% CI, 2.9%-9.1%) and 12% (7%-17%), respectively, for BCCs after 5 and 10 years, as shown in Figure 1.
Among patients who developed NMSC, the median time between transplantation and appearance of the first NMSC was 5.2 years (3.5-7.8 years, 25th-75th percentile) and tended to be lower for BCCs than SCCs (4.5 vs 5.9; P = .10).
The cumulative incidence of SCCs and BCCs was higher in men than in women, although this difference did not reach statistical significance (Table 1). The cumulative risk of SCCs and BCCs increased steeply with increasing age at transplantation: 10 years after transplantation, it ranged from 2.1% and 2%, respectively, in patients 43 years or younger at transplantation to 47.9% and 22.7%, respectively, in patients 59 years or older. Moreover, age at transplantation was significantly correlated with the interval from transplantation to the diagnosis of the first NMSC (P<.001; τ = 0.437): older patients showed a shorter induction period for NMSCs.
The development of BCCs was strongly related to skin type II, whereas this association was only slightly significant for SCCs (Table 1). In contrast, the risk of SCCs but not of BCCs significantly increased in patients with higher occupational sunlight exposure (Table 1).
No difference in the cumulative incidence of SCC and BCC was found between patients grouped according to recreational sunlight exposure (low, moderate, and high) or to number of weeks of summer holidays in the sun spent in adulthood or over a lifetime (data not shown). Paradoxically, a significantly higher incidence of SCC occurred in patients who spent fewer weeks of summer holidays in the sun during childhood (P = .01).
The occurrence of both SCCs and BCCs was strongly associated with the presence of solar keratoses (Table 1). The crude odds ratio was 5.3 (95% CI, 2-13.7; P = .001) for SCC and 5.6 (95% CI, 1.7-19.2; P = .005) for BCC development. No association was found between the presence of warts and the occurrence of either SCCs or BCCs (Table 1).
The median cumulative doses of azathioprine, cyclosporine, and oral prednisone (maintenance immuno suppressive therapy) in patients with SCCs, BCCs, or without NMSCs are shown in Figure 2. No significant association was found between the cumulative doses of each drug and the occurrence of SCCs or BCCs. Furthermore, there were no significant differences in SCC or BCC incidence on the basis of cumulative dose of intravenous methylprednisolone (administered to treat acute rejection episodes during the first year after transplantation), use of OKT3 anti-CD3 monoclonal antibody or antilymphocyte globulin/antithymocyte globulin (either as induction or rejection therapy), or number of acute rejection episodes requiring therapy in the first year after transplantation.
The WLC of cumulative doses of azathioprine, cyclosporine, and oral prednisone at 3 and 6 months and at 1, 2, and 3 years were consistently higher in patients with SCCs and lower in patients with BCCs than in patients without NMSCs (data not shown). However, this difference did not reach statistical significance. The cumulative incidence of SCC in patients stratified by age at transplantation (≥54 years or <54 years) was consistently higher in those who exhibited greater WLC. Particularly striking was the difference in the 10 year-posttransplantation cumulative incidence of SCC in patients 54 years or older at transplantation (Table 2).
Risk factors for SCC identified by multivariate Cox proportional hazards model are listed in Table 3. Age at transplantation was the most important prognostic factor: patients 59 years or older at transplantation had a 36.2 relative risk for SCC occurrence vs patients 43 years or younger at transplantation (95% CI, 4.1-314.9; P < .001). Occupational sunlight exposure was an additional risk factor for SCC, with a relative risk rising from 1.1 for patients exposed for less than 13 000 hours (95% CI, 0.3-4.0, P = .60) to 3.4 (95% CI, 1.2-9.9; P = .02) and 4.1 (95% CI, 1.3-13.4; P = .02) for patients exposed for between 13 000 and 29 999 hours and for 30 000 hours or more, respectively.
The WLC at 3 years after transplantation but not at shorter time points was independently associated with an increased risk of developing SCC. Patients with a WLC at 3 years higher than the 75th percentile had a 4 times higher risk of SCC than HT recipients with a WLC at 3 years lower than the 50th percentile (95% CI, 1.5-11.4; P = .008).
The risk of developing BCC was associated only with age at transplantation and skin type II. Patients who received their HT when they were 59 years or older showed an 8.5 times higher risk of BCC than patients 43 years or younger at transplantation (95% CI, 1.1-73.7; P = .04). Patients with skin type II experienced a 5.7 times higher risk than patients with skin types III through V (95% CI, 2.0-16.6; P=.001) (Table 3).
Organ transplant recipients are at increased risk of cutaneous malignancies and have a greater tendency to develop SCCs than BCCs.1,5 Our study highlights the striking risk of skin cancer after organ transplantation in a cohort of adult HT recipients with a long-term follow-up. The steady increase of risk of cutaneous malignancies with time after transplantation is an alarming figure because the number of organ allograft recipients who live for many years after transplantion is rapidly growing. Time-to-event analysis showed that 1 of 5 HT recipients will develop SCCs, and 1 of 8 will develop BCCs within 10 years after transplantation. The excess of SCCs over BCCs leads to an observed SCC/BCC ratio (2:1) that is reversed from what is usually reported in the general population (1:4). This observation, confirmed by several previous studies,4- 6 suggests a different behavior in relation to immunosuppression between the 2 types of NMSCs. However, sparse data are available on the actual relationship between incidence of SCCs and BCCs and the type or the cumulative doses of immunosuppressive drugs.10- 13,23
In our study, we found no association between cumulative doses of each single immunosuppressive drug, whether they were used for immunosuppressive maintenance therapy (azathioprine, cyclosporine, and oral prednisone) or for acute rejection episodes (intravenous methylprednisolone, OKT3, or antilymphocyte globulin/antithymocyte globulin), and the risk of developing SCCs or BCCs. Patients with BCCs showed median cumulative doses of azathioprine, cyclosporine, and oral prednisone even lower than patients without skin cancer. However, on multivariate analysis, patients with a higher WLC after 3 years showed a greater than 4-fold risk of SCCs than patients with a lower WLC at 3 years. The same result was not observed for BCCs. This suggests that the risk of SCC in organ transplant recipients might be associated with the overall level of immunosuppression rather than with a specific immunosuppressive drug.
Our findings support the view of a different relationship between the 2 types of NMSCs and the immune system. The immunogenicity of SCCs and the effective immune response directed against the tumor cells may lead to at least partial control of these tumors in the immunocompetent host.24- 27 Obviously, SCCs could therefore "opportunistically" develop in the presence of a defective immune system. Conversely, BCCs, showing a less immunogenic activity, may not take advantage of a state of immunodeficiency.28,29 In agreement with this hypothesis, the rate of HLA mismatching, likely through an indirect effect on the level of immunosuppression, has been found to be associated in organ transplant recipients with an increased risk of developing SCCs but not BCCs.10
The risk of SCC but not of BCC was increased by a high occupational sunlight exposure. Conversely, the risk of BCC but not of SCC increased substantially in subjects with skin type II. Cumulative sunlight exposure has been shown to be associated with an increased risk of SCC but not of BCC in kidney transplant10,30 and HT recipients31 and in the general population.32,33 In contrast, the risk of BCC in the general population appears to be strongly associated with the tendency to sunburn, but this relationship has not been clearly demonstrated for SCC.34,35 In agreement with this hypothesis, BCC is increasingly observed on non–sun-exposed sites, mainly on the trunk, where sunburns are more frequent.32 In our study, 20% of BCCs developed on nonexposed areas of the body, whereas slightly more than 10% of SCCs occurred on these sites.
This study does not confirm the relationship between the presence of warts and risk of SCC or BCC reported in recent studies.9,31 The association between human papillomavirus and anogenital cancer is well established, but the role of human papillomavirus in cutaneous oncogenesis remains controversial, and further studies are warranted.35
In conclusion, our study suggests that the risk of SCC but not of BCC in HT recipients is related to the level of global immunosuppression after 3 years rather than to a specific immunosuppressive drug. This might explain the reversal of the SCC/BCC ratio in organ transplant recipients compared with the general population, on the basis of a more effective escape from the immunosurveillance mechanism for the highly immunogenic SCC compared with the less immunogenic BCC. The increasing number of organ transplantations performed each year in addition to the increasing long-term survival of transplant recipients points out the need to devote more resources to skin cancer detection. Patients at high risk should maintain as low as possible the level of immunosuppression and receive regular quarterly dermatologic follow-up visits, whereas annual examinations could suffice for the others. All transplant recipients should be advised to avoid sunlight exposure, to use high-factor sunscreens, to wear protective clothing, and to see their physicians immediately when they observe suspect skin changes.
Correspondence: Stefano Piaserico, MD, Unit of Dermatology, Department of Medical and Surgical Sciences, University of Padua, Via Cesare Battisti, 206, 35128 Padua, Italy (firstname.lastname@example.org).
Accepted for publication March 24, 2004.