Nonmelanoma skin cancer after liver transplantation. Study of risk factors



Nonmelanoma skin cancer (NMSC) is a frequent complication after liver transplantation, but the risk factors of posttransplant NMSC have not been well defined. In a prospectively followed series of 170 liver transplant recipients, we assessed the incidence of NMSC, compared it with the expected incidence in the general population, and investigated which risk factors were related to NMSC. After a median follow-up of 62 months, 27 patients developed 43 NMSC. The relative risk of NMSC was 20.26 (95% confidence interval: 14.66-27.29) as compared with sex- and age-matched population. In univariate analysis, older age, male sex, Child-Turcotte-Pugh A or B at transplantation, treatment with mycophenolate mofetil, skin type, and total pretransplant sun burden were associated to the development of NMSC. In multivariate analysis, only skin type and total sun burden were independently related to NMSC. In conclusion, risk of posttransplant NMSC may be estimated combining skin type and an easy estimation of total sun burden. No individual immunosuppression regimen seems to be related to a higher risk of NMSC. (Liver Transpl 2005;11:1100–1106.)

Nonmelanoma skin cancer (NMSC) is the most frequently diagnosed malignancy after liver and other solid organ transplantations.1 Its incidence may be higher than 80% 20 years after transplantation in countries with high sun exposure.2 In most cases, these tumors are diagnosed sufficiently early to allow curative treatment, but squamous-cell carcinomas appear to be more aggressive in transplant recipients than in the general population. They may recur, metastasize, and be the cause of death.1 So, aggressive preventive measures and frequent dermatologic assessment have been recommended.3

The identification of those patients with a higher risk of development of NMSC may be useful to reinforce preventive measures and develop a tailored schedule of dermatologic follow-up in patients with higher risk of NMSC. Furthermore, if specific immunosuppressive regimens are found to be related to a higher risk of NMSC, these regimens should be avoided in patients with other risk factors of NMSC. Therefore, we have analyzed a cohort of prospectively followed liver transplantation (LT) recipients to determine which risk factors are independently associated with the development of NMSC.


NMSC, Nonmelanoma skin cancer; LT, liver transplantation.

Patients and Methods

Schedule of Dermatology Visits

We analyzed the data from a prospectively followed cohort of adult LT recipients who received their first LT between May 1990 and April 2003. All patients were living in Spain or Portugal (latitude 40°-45° North). At the time they were transplanted, all patients were counseled to avoid sun exposure and to use sun screens. They were followed by a single senior dermatologist 6 and 12 months after LT and every year thereafter. Furthermore, they were evaluated whenever a cutaneous malignancy was suspected by the hepatologist. The frequency of the dermatology visits was increased to 1 every 4 to 6 months when a malignant or premalignant lesion was observed during post-LT follow-up.

In their baseline dermatology visit, the skin type, according to Fitzpatrick4 (Table 1), and their history of sun exposure before transplantation, according to Espana et al.,5 were assessed (Table 2). Total sun burden was assessed by a questionnaire completed by the patients about their usual place of residence, the sun exposure during their holidays, and the weekly time of occupational exposure in four periods of life until LT (<20, 20-40, 40-60 and >60 years).5 To obtain the total score, we summed the points obtained for each of these 3 factors of exposure in each of the 4 periods of life to obtain a theoretical maximum score of 24 (actual maximum score was 19). In order to represent the results, total sun burden was arbitrarily grouped as low (0-5 points), medium (6-10 points), and high (more than 10 points). All the lip and cutaneous malignancies were histologically confirmed. All the patients were followed until their last dermatology visit before the end of July 2004.

Table 1. Skin Sensitivity to Sunlight
Skin TypeDescription
IAlways burn, never tan
IIAlways burn, sometimes tan
IIISometimes burn, sometimes tan
IVSometimes burn, always tan
VNever burn, sometimes tan
VINever burn, always tan
Table 2. Score System to Assess Total Sun Exposure (Total Sun Burden)
 Sun Exposure at Different Ages
<20 Years20–40 Years40–60 Years>60 Years
Usual residence    
Occupational sun exposure   
 <20 hours/week1111
 ≥20 hours/week3333
Sun exposure during vacations    
 <14 days/year1111
 ≥14 days/year2222


Induction immunosuppression protocols have been described previously.6 Until 1996, immunosuppression was based on the combination of cyclosporine, azathioprine, and steroids. From 1997 to 2003, it was based on cyclosporine or tacrolimus, combined with steroids; most patients also received azathioprine or mycophenolate mofetil.

Until 1996, maintenance immunosuppression was based on triple therapy, combining cyclosporine, azathioprine, and prednisone, with the exception of those patients with intolerance to azathioprine or with diabetes mellitus. Cyclosporine monotherapy was not attempted before the end of the second posttransplant year. After 1997, steroid withdrawal was attempted between the third and the sixth posttransplant month. After steroid withdrawal, azathioprine was withdrawn; thus monotherapy with cyclosporine or tacrolimus was usually attempted between the third and the ninth month after transplantation.

Treatment of rejection consisted on an increase of cyclosporine or tacrolimus doses, or the administration of up to 3 1-g boluses of methylprednisolone. Refractory rejection was treated with OKT3.

Statistical Analysis

For the purposes of the study, lip malignancies were considered as NMSC. For comparison with the general population, observed cases of NMSC in the study group were compared with those expected based on age- and sex-specific incidences of neoplasia and cancer-related mortality rates for Navarra (Spain) in the period 1993-1997.7 The 95% confidence limit for the relative risk (observed to expected ratio) was obtained after assuming a Poisson distribution for the development of NMSC.

The cumulative incidence of NMSC was obtained with the Kaplan-Meier method (considering the first NMSC as event). The potential association of the following variables was investigated: age, sex, cigarette smoking (if higher than 10 pack-years), alcohol (if higher than 80 g/day for 10 years), hepatitis C, Child-Turcotte-Pugh status at transplantation (A or B vs. C), hepatocellular carcinoma, major immunosuppressive drugs (cyclosporine vs. tacrolimus), antimetabolite (azathioprine or mycophenolate mofetil) therapy (if longer than 6 months), mycophenolate mofetil therapy (if longer than 6 months), steroid therapy (if longer than 6 months), graft rejection requiring high-dose steroid treatment in the first 6 months after LT, use of OKT3, skin type (I or II vs. III or IV), and total sun burden before transplantation. To investigate the possible influence of these factors on the development of NMSC, every variable was considered in univariate Cox proportional hazards model. All predictors with P < 0.2 in univariate analysis were entered in a multivariate Cox proportional hazard model; P < 0.05 was considered significant. All statistical analyses were performed with the software SPSS 11.0.


During the mentioned period, 231 transplantations were performed among 222 adult patients. Fifty-two patients were excluded because their survival was less than 6 months (17 patients), they refused to attend the Dermatology Clinic follow-up (16 patients, including 3 patients with poor short-term prognosis), or incomplete data were available at the baseline Dermatology Clinic visit (19 patients). The remaining 170 patients were studied, and their general characteristics are shown in Table 3. After a median follow-up of 62 months (cumulative follow-up: 992 patient-years), 27 patients developed 43 NMSC. The incidence of NMSC was 43.3 per 1,000 patient-years. The cumulative actuarial incidence (Kaplan-Meier) is shown in Figure 1.

Table 3. General Characteristics of the 170 Liver Transplant Recipients Included in the Study
  • NOTE: Values expressed as mean (SD).

  • *

    One patient without cirrhosis, not included.

Age (years)55 (9.77)
 23–40 years16 (9%)
 41–60 years92 (54%)
 61–73 years62 (36%)
Sex (male/female)125 (74%)/45 (26%)
Cause of liver disease 
 Hepatitis C49 (29%)
 Alcohol72 (42%)
 Hepatitis B13 (8%)
 Other36 (21%)
Child-Turcotte-Pugh status (A/B/C)*24 (14%)/81 (48%)/64 (38%)
History of smoking55 (32%)
Skin type (I/II/III/IV)2 (1%)/35 (21%)/129 (76%)/4 (2%)
Sun burden (low/medium/high)60 (35%)/79 (46%)/31 (18%)
Rejection36 (21%)
Main immunosuppressive therapy 
 Cyclosporine95 (56%)
 Tacrolimus75 (44%)
Antimetabolite therapy >6 months102 (60%)
Mycophenolate mofetil >6 months21 (12%)
Steroid therapy >6 months96 (56%)
Figure 1.

Cumulative actuarial incidence of cutaneous neoplasia in 170 liver transplantation recipients.

LT recipients had a relative risk of developing NMSC of 20.26 (95% confidence interval: 14.66-27.29), as compared with a sex- and age-matched general population. Sixteen patients developed 22 squamous-cell carcinomas. The cumulative actuarial incidence of squamous-cell carcinoma is shown in Figure 2A, and their locations are shown in Table 4. Fifteen patients developed 21 basal-cell carcinomas. Cumulative actuarial incidence of basal-cell carcinoma is shown in Figure 2B, and their specific locations are shown in Table 4.

Figure 2.

Cumulative actuarial incidence of squamous cell carcinoma (2A) and basal cell carcinoma (2B) in 170 liver transplantation recipients.

Table 4. Location of Squamous-Cell and Basal-Cell Carcinomas in 170 Liver Transplant Recipients
 Squamous CellBasal Cell
Forehead and scalp98
Ear and retroauricular23
Upper limbs30
Lower limbs10

In univariate analysis, older age, male sex, Child-Turcotte-Pugh A or B cirrhosis at transplantation, treatment with mycophenolate mofetil for more than 6 months, skin type I or II, and higher total sun burden were associated with the development of NMSC (Table 5). In multivariate analysis, only skin type and total sun burden were independently associated with the development of NMSC (Table 5).

Table 5. Risk Factors for Nonmelanoma Skin Cancer in 170 Liver Transplant Recipients
VariableUnivariate AnalysisMultivariate Analysis
HR (95% CI)PHR (95% CI)*P
  1. Abbreviations: HR, hazard ratio; CI, confidence interval; HCC, hepatocellular carcinoma; MMF, mycophenolate mofetil.

Age (years)1.07 (1.02–1.12)0.0071.03 (0.98–1.09)0.2
Sex (male)2.52 (0.87–7.31)0.091.36 (0.41–4.51)0.6
Smoking0.98 (0.43–2.24)0.95  
Alcohol0.88 (0.40–1.92)0.74  
Hepatitis C1.16 (0.52–2.58)0.72  
HCC1.46 (0.36–3.29)0.36  
Child-Turcotte-Pugh (A or B)1.99 (0.80–4.94)0.131.77 (0.69–1.77)0.2
Tacrolimus1.01 (0.41–2.51)0.98  
Antimetabolite1.03 (0.44–2.41)0.94  
Steroids >6 mo.1.16 (0.48–2.81)0.75  
Rejection treatment0.74 (0.31–1.80)0.51  
OKT30.04 (0.00–1.10)0.44  
MMF3.33 (0.94–11.83)0.062.45 (0.64–9.37)0.2
Skin type (I or II)2.10 (0.97–4.52)0.063.27 (1.44–7.41)0.005
Sun burden (units)1.22 (1.11–1.34)<0.00011.20 (1.07–1.35)0.002

Figures 3A and 3B show the actuarial incidences of NMSC according to skin type (I and II vs. III and IV) and the total sun burden. Seven patients developed more than 1 NMSC. Characteristics of these neoplasia and general characteristics of these patients are shown in Table 6. All the tumors were resected with wide margins and no patient had tumor recurrence.

Figure 3.

Actuarial incidences of cutaneous malignancy patients with skin types I and II or skin types III and IV stratified according to their pretransplant sun burden. Pretransplant sun burden has been arbitrarily grouped according to total sun burden in low (0-5 points), medium (6-10 points), and high (more than 10 points).

Table 6. General Characteristics of the Patients and of the Tumors in Liver Transplant Recipients With More Than 1 Nonmelanoma Skin Cancer
SexIndication of LTImmunosuppressionRejectionSkin TypeTSBTumor HistologyLocation
  1. Abbreviations: TSB, total sun burden; HCV, hepatitis C virus; HCC, hepatocellular carcinoma.

MaleAlcoholic cirrhosisCyclosporineNoII4Squamous cellLeg
      Basal cellForehead
MaleHCV cirrhosisCyclosporineNoIII15Squamous cellLip
      Basal cellForehead
      Basal cellForehead
      Basal cellFace
MaleAlcoholic cirrhosis + HCCCyclosporineYesII8Squamous cellRetroauricular
      Squamous cellRetroauricular
MaleHemochromatosis + HCCCyclosporineNoIII12Squamous cellScalp
      Squamous cellScalp
      Squamous cellScalp
      Basal cellForehead
      Squamous cellEyelid
      Basal cellScalp
MaleAlcoholic cirrhosisTacrolimusNoIII8Basal cellForehead
      Basal cellForehead
      Basal cellRetroauricular
MaleHemochromatosis + HCCCyclosporineNoII8Basal cellEar
      Basal cellEar
FemaleHCV cirrhosis + HCCTacrolimusNoIII4Squamous cellScalp
      Squamous cellArm
      Basal cellFace
      Squamous cellBack


NMSC is a frequent complication in LT recipients.3 In this series, 16% of patients developed NMSC. This incidence is higher that the incidence reported in most other studies,8–14 but it is comparable to the incidence of NMSC in recipients of other solid organs.1 The difference between our series of LT recipients and others may be explained in part by some differences in the length of follow-up, the higher ages of the patients in our series (which is related to a higher total sun burden), or geographical area. Furthermore, some authors have suggested that the incidence of NMSC in LT may have been underreported in some series.15 In our series, the strict dermatology follow-up of the patients warrants that the number of NMSC reported is correct.

It has been suggested that the incidence of NMSC is lower in LT recipients than in other solid organ transplant recipients, because the requirement of immunosuppression is lower after LT.1 Our data confirm this difference. In our series, the incidence of NMSC in patients undergoing LT was lower to the incidence found in heart transplant recipients from our center with similar lengths of follow-up5 (16% vs. 26%).

The strict follow-up of our patients may have been helpful in diagnosing all the malignancies at an early stage. Squamous-cell carcinomas appear to be more aggressive in transplant recipients than in the general population.1, 3 Its rapid growth, probability of recurrence, or ability to metastasize are not exceptional. In our series, the strict follow-up may have contributed to the good outcome. None of the patients had tumor recurrences or metastases during follow-up.

In our series, the risk of developing NMSC was linked to both skin type and pretransplant total sun burden. Indeed, the combination of a simple quantitative score of sun burden and skin type provided an easy estimate of the risk of neoplasia. In patients with a high risk of NMSC, the usual counseling of avoiding sun exposure3 must be reinforced. Our policy of frequent dermatology visits has allowed the early diagnosis in all the cases. We think that this strict policy of follow-up is advisable, at least for patients at high risk of developing NMSC.

In previous papers,9, 15, 16 age was found to be an independent risk factor for NMSC. In the present study, we also found an association between age and NMSC in univariate analysis, but this association loss its significance in multivariate analysis. Obviously, a higher age at transplantation increases the risk of having a higher sun burden. In fact, we found a highly significant correlation between sun burden score and age (R = 0.288; P < 0.001). To ensure that the score is related to the risk of NMSC, irrespective of the age of the patients, we analyzed if the score predicts the risk of NMSC in patients younger and older than the mean age of our series (55 years). The significant association between NMSC and sun burden persisted in both groups of age: in patients up to 55 years, hazard Rratio for every unit of sun burden was 1.29 (95% confidence interval: 1.12-1.48; P = 0.003), and in patients older than 55 years, hazard ratio was 1.18 (95% confidence interval: 1.04-1.33; P = 0.01).

As previously found in other series,12 our LT recipients had a 20-fold risk of developing cutaneous neoplasia, as compared with the sex- and age-matched general population in the same geographical area. This increased risk of neoplasia seems to be related to immunosuppression, but in this series, no individual immunosuppressive regimen was found to be associated with the development of NMSC. In univariate analysis, NMSC development was associated with treatment with mycophenolate mofetil, but this association disappeared in multivariate analysis. Previous studies have found that there is an association between NMSC and a high level of immunosuppression17 or treatment of rejection episodes.18 In LT recipients, Mithoefer et al. found a higher incidence of NMSC in patients receiving cyclosporine than in those receiving tacrolimus.15 This finding was not confirmed in our series. The use of a less carcinogenic immunosuppression regimen might be useful in patients high risk of NMSC. Thus far, only sirolimus appears unrelated to the development of neoplasia. Indeed, recent data in renal transplant patients suggest that immunosuppression with sirolimus may protect against skin cancer.19

In conclusion, LT recipients have an increased incidence of NMSC. The combination of skin type and an easy estimation of sun burden may predict the risk of developing NMSC. Our aggressive policy of dermatology follow-up allowed for the diagnosis of these tumors at an early stage and avoided local recurrence and NMSC-related mortality.