The impact of preexisting or acquired Kaposi sarcoma herpesvirus (KSHV) infection in kidney transplant recipients was evaluated in a prospective study. Serum collected from kidney donors and recipients before transplantation were tested for antibodies against KSHV latent nuclear antigen. Three groups of recipients were defined: group A (KSHV+), group B (KSHV−, KSHV+ donor) and group C (donor and recipient KSHV−). Blood was collected from recipients, every 3 months for 3 years, for KSHV viremia (groups A and B), quantitative (group A) and qualitative serology (group B). Data of group C recipients were extracted from a French database. The prevalence of KSHV antibodies was 1.1% in donors and 3.2% in recipients. There were respectively 161, 64 and 4744 recipients in groups A, B and C.
In group A, 13% developed Kaposi's sarcoma (KS). Age >53.5 years (p = 0.025) and black skin (p = 0.0054) were associated with KS development. In group B, three recipients developed clinical manifestations related to KSHV infection. There was no difference in terms of survival and graft loss between the three groups. In conclusion, although kidney recipients should be aware of the additional risk of KSHV morbidity, KSHV+ recipients should not be systematically excluded from kidney transplantation.
In 2000, the transmission of human herpesvirus type 8 (HHV-8) or Kaposi sarcoma herpesvirus (KSHV) via a renal transplantation was formally documented by analysis of the highly variable sequence of the orf-K1 gene and polymorphisms in the orf-73 gene of the KSHV genome, demonstrating that the KSHV strains isolated from the Kaposi's sarcoma (KS) lesions of a kidney recipient originated from the organ donor (1). However, the frequency of KSHV transmission by kidney transplantation and its clinical consequences are unclear, as is the magnitude of the risk of KSHV reactivation in patients infected before transplantation (2). In several retrospective studies carried out in different countries, the rate of seroconversion after renal transplantation in previously KSHV-seronegative recipients has been estimated to range from 2% to 12.1%, and the rate of KS to vary between 0% and 8% of patients who seroconverted (2–5). In previously reported limited retrospective studies, 23%–28% of recipients, who were KSHV seropositive before transplantation developed KS (6, 7). We therefore conducted a nationwide prospective multicenter study in France on a cohort of kidney transplant recipients to evaluate the consequences of KSHV infection acquired before transplantation or transmitted through the graft.
In this French nationwide multicenter prospective study, we included kidney recipients undergoing transplantation between October 2000 and December 2003. All patients signed an informed consent form before inclusion in the study, which was performed in accordance with the rules of our institute and French legislation. Donors and recipients were tested for the presence of antibodies to the KSHV latent nuclear antigen (LANA) with a latent immunofluorescence test (see below), at the time of transplantation. Recipients were assigned to three groups on the basis of the results of these tests: group A (KSHV seropositive recipients, regardless of the serological status of the donor), group B (KSHV seronegative recipients receiving a kidney from a seropositive donor) and group C (seronegative recipients with a seronegative donor). We retrospectively tested serum samples from all donors for antibodies against a structural viral glycoprotein by ELISA (K8.1 glycoprotein; see below). No reclassification of their recipients was performed on the basis of these test results.
The recipients in groups A and B attended checkups every 3 months for 3 years. At each visit, a dermatological examination was carried out and blood was sampled for quantitative KSHV viremia determination (groups A and B), quantitative KSHV serological assays for antibodies against the K8.1 glycoprotein (group A), qualitative KSHV serological tests based on three serological assays for KSHV: LANA immunofluorescence, lytic antigen immunofluorescence and EIA for K8.1 glycoprotein (group B). Seroconversion was defined as the recent detection of anti-KSHV antibodies in two different assays (latent immunofluorescence, and/or lytic immunofluorescence and/or EIA for K8.1 glycoprotein). The data for group C were extracted from the CRISTAL database of the Agence de Biomédicine, the French organ procurement and transplantation agency. In addition to routine clinical monitoring by transplant teams, the recipients in group C had at least one dermatological consultation per year, as recommended by the skin and organ transplantation group of the French Society of Dermatology and French kidney transplantation recommendations.
Antibodies to the KSHV LANA were detected using an immunofluorescence assay (IFA) on the primary effusion cell line BC-3, which is latently infected with KSHV but not with Epstein–Barr virus. The IFA was performed as previously described, with a serum dilution of 1/100 (8).
The IgG antibodies directed mainly against lytic KSHV antigens were detected by a commercial indirect IFA (Biotrin Human Herpesvirus-8 IgG Immunofluorescent Assay, Biotrin, Lyon, France). This test uses BC-3 cells, in which the lytic replication cycle has been induced, as a source of KSHV and serum at a 1/64 dilution, and was performed according to the manufacturer's instructions.
The K8.1 EIA has been described in detail elsewhere [“http://www.journals.uchicago.edu/doi/full/10.1086/423326”\l“rf18”] (9). In brief, serum samples were diluted 1:20 and added to ELISA plates that had been precoated with recombinant K8.1 protein. After incubation and washes, a conjugate antihuman antibody was added, and, after a further period of incubation and removal of unbound antibody, a substrate solution containing nitrophenyl phosphate (Sigma) was added and the plates were immediately read spectrophotometrically at 405 nm by use of a standard EIA plate reader (Diagnostics Power CP400). To define the cutoff value, the mean optical density readings for the 10 negative controls (blood donors) were determined. Samples were scored as negative if their optical density was <5 SD above the mean of the negative samples. Fluctuations in K8.1 antibody levels in individual patients over time were studied by using the original ELISA OD values for statistical analysis.
Quantification of KSHV viremia
KSHV viral load was determined in each sample by real-time quantitative PCR analysis using primers derived from the ORF 26 gene on an ABI PRISM 7500 instrument (Applied Biosystems, Courtaboeuf, France). The reaction conditions and methodological approach used were as previously described (10,11).
According to the protocol, failure time data were analyzed for the first 36 months after enrolment. Thus, all events, such as deaths, graft loss and KS occurring after 36 months were ignored.
The cumulative incidence function (CIF) of graft loss and KS was estimated in a competing-risk framework, using a nonparametric estimator (with death prior the event defining the competing-risk endpoint). The Gray's test was used to compare CIF (12). The distribution of overall survival was estimated by the Kaplan–Meier method and survival curves were compared using log-rank tests.
In group A, univariate prognostic analyses for KS incidence were based on the Gray's test, implemented in the cmprsk package (http://biowww.dfci.harvard.edu/~gray/cmprsk_2.1--4.tar.gz). Factors significantly associated with outcome at the 5% level were introduced into a multivariate regression model, with stepwise selection (13). The strength of the association was estimated by the hazard ratio (HR) and its 95% confidence interval (95% CI). Changes in antibody levels detected with the K8.1 ELISA were analyzed with a mixed regression model for repeated measurements.
All statistical tests were two-sided, with p-values of 0.05 and below considered significant. All statistical analyses were performed on SAS 9.1 (SAS Inc., NC) and R 2.8.0 (R Development Core Team 2006) software packages.
Description of groups A, B and C
Overall, 94.3% of donors and 71.8% of recipients, transplanted in France between October 2000 and December 2003, were included in the study. The prevalence of antibodies against the KSHV LANA was 1.08% among French donors and 3.24% among French recipients. A higher prevalence was observed in large cities, especially in Paris and Marseille. Among donors without antibodies to KSHV LANA, 1.7% had antibodies to the K8.1 membrane glycoprotein.
In total, 161 kidney recipients were assigned to group A and 64 were assigned to group B (Figure 1). The main clinical characteristics of these patients are shown in Table 1. The donors of seven of the recipients in group A were seropositive for KSHV. Group C consisted of 4744 recipients (seronegative recipients with a seronegative donor).
Table 1. Description of kidney recipients included in groups A and B
Parameters no. (%)
Group A (KSHV + recipients; KSHV ± donor) 161
Group B (KSHV − recipients; KSHV + donors) 64
*In group A: heart (1), pancreas (3) and liver (3); in group B: heart (1), lung and liver (1).
Country of origin
North, Western or Central Europe
Africa, South of Europe, Mediterranean countries and Arabian Peninsula
French overseas departments
Dark complexion (phototype IV–V)
Fair complexion (phototype I–III)
Type of nephropathy
More than one transplanted organ*
Cytomegalovirus (valaciclovir or ganciclovir)
Pneumocystis carinii (sulfamethoxazole–trimethoprim or sulfadoxine–pyrimethamine)
Herpes simplex virus (aciclovir or valaciclovir)
Tuberculosis (rifampicin and/or isoniazid)
Kaposi's sarcoma in group A
Twenty-one (13%) of the 161 patients developed KS and 10 died without developing KS. All patients who developed KS received a kidney from a donor who was seronegative with two KSHV antibody tests (K8.1 EIA and latent IFA). Cumulative KS incidence rates were 4.4% at 6 months after transplantation, 8.8% at 12 months, 12% at 24 months and 13% at 36 months. Table 2 summarizes the results of univariate analyses for the identification of factors associated with KS development. Skin color (p = 0.005), age at transplantation (p = 0.041) and country of birth (p = 0.040) had a significant effect on the cumulative incidence of KS. KS tended to be more frequent after induction treatment with monoclonal anti-Il2-receptor antibody, but this difference was not statistically significant. None of the 17 recipients treated with sirolimus developed KS. In the multivariate analysis, the three prognostic factors for KS selected from Table 2 were jointly considered, namely age at transplantation, country of birth and skin color. In the final model, only age >53.5 years (HR = 3.29 [1.16; 9.34], p = 0.025) and black skin (HR = 4.55 [1.56; 13.51], p = 0.0054) were retained as independent prognostic factors for KS.
Table 2. Group A: Univariate analyses of factors associated with Kaposi's sarcoma
N patients (%)
3-year cumulative incidence (%)
Hazard ratio [95% CI]
p-Value (Gray test)
N KS = number of patients with Kaposi's sarcoma.
2.37 [0.70; 8.03]
Age >53.5 years
2.59 [1.01; 6.64]
0.66 [0.16; 2.67]
0.82 [0.19; 3.55]
Country of origin and skin color
North, Western and Central Europe
1.43 [1.02; 2.01]
Africa, South of Europe, Mediterranean countries and Arabian Peninsula
10.3 [1.36; 78.7]
12.1 [0.65; 223.1]
3.91 [1.39; 11.02]
2.42 [0.55; 10.6]
0.85 [0.36; 1.98]
2.20 [0.94; 5.15]
1.51 [0.64; 3.57]
0.86 [0.33; 2.20]
Cytomegalovirus (valaciclovir or ganciclovir)
1.94 [0.82; 4.59]
Herpes simplex virus (aciclovir or valaciclovir)
2.13 [0.91; 4.98]
Pneumocystis carinii (sulfamethoxazole-trimethoprim or sulfadoxine-pyrimethamine)
1.18 [0.48; 2.90]
Sequential blood samples were not available for 12 recipients, for various reasons (death, graft loss, and distance from the transplantation center). KSHV viremia was detected in 23 recipients, 13 of whom developed KS. In these recipients, KSHV viremia was detected before or at KS diagnosis in five cases and after KS diagnosis in eight cases, with estimated sensitivity of viral load at 23.8%. Other 10 patients with detectable KSHV viremia presented no clinical signs of KSHV infection, with an estimated specificity of 92.2%. K8.1 EIA analysis showed that anti-KSHV antibody levels decreased over time (p < 10−4).
Clinical and virological events in group B
Only three of the 64 (4.6%) recipients in group B developed clinical signs potentially related to KSHV infection. One patient, in whom KSHV viremia was detectable for 2 months, developed fever, hepatosplenomegaly, pharyngeal erythema and pancytopenia 8 months after transplantation (14). Bone marrow aspiration showed normal cellularity, with no signs of hemophagocytosis, but KSHV DNA was detected by PCR. Nine months after grafting, KSHV viremia increased to 18 million copies/106 mononuclear cells and KSHV DNA was also detected in pharyngeal samples. The patient was successfully treated with four rituximab infusions, resulting in the clearance of KSHV viremia 1 month later. Antibodies to KSHV antibodies, detectable by lytic immunofluorescence and K8.1 EIA, were found 24 months after transplantation. Two recipients presented with skin and visceral KS, 17 and 27 months after transplantation with low-level KSHV viremia detected on diagnosis (330 and 100 copies/106 mononuclear cells, respectively). KSHV seroconversion was documented by lytic and latent immunofluorescence tests, 19 and 23 months, respectively, after transplantation. The cumulative incidence rates for HHV-8 viremia and antibody detection at 36 months after transplantation were 5.8% and 31.22%, respectively. None of the recipients with evidence of KSHV seroconversion received kidneys from the same donor.
Survival and graft loss
No patient from group C (4744 patients) had a clinical event that could be related to KSHV infection. The overall survival did not differ between the three groups (p = 0.69), with 3-year survival estimated at 93.7%[95% CI: 90.1; 97.6] in group A, 96.7%[95% CI: 92.4; 100] in group B and 94.7%[95% CI: 94.0; 95.3] in group C.
Similarly, no significant differences in the cumulative incidence of graft loss (p = 0.64) and death with a functional graft (p = 0.83) were observed between groups (Figure 2A and B). At 36 months, the cumulative incidence of graft loss was 11% in groups A and B and 8.5% in group C.
Follow-up after the end of the study
Clinical follow-up of recipients was extended beyond the study, with a median follow-up of 48.9 months after transplantation (95% CI: 46.1–53.7 months). Three recipients of group A developed KS at 37, 44 and 53 months after transplantation. None was treated by sirolimus and had positive KSHV viremia during the course of the study. One patient, originally from Mali, presented with a refractory pleural effusion during treatment with sirolimus for KS and died 8 months after the diagnosis of primary effusion lymphoma.
KSHV seroprevalence varies greatly worldwide. Between 1% and 10% of people are infected in developed countries, whereas up to 80% of the population may be infected in some areas of sub-Saharan and equatorial Africa (15). It is therefore not surprising that a high percentage of group A recipients had a black skin (43.2%) and is originated from countries with a known high prevalence of KSHV infection. In France, the prevalence of KSHV infection among organ donors (1.08%), as determined with latent IFA, is low. In Ile-de-France, the prevalence of this infection (2.1%) was similar to that reported for blood donors, based on the same latent IFA (16). Higher prevalence rates are generally obtained with the lytic IFA (17). Kidney recipients were more frequently found to be KSHV seropositive (3.24%) than donors (1.08%). Previous studies have reported higher KSHV seroprevalence rates in dialysis patients (5). Some patients may have been infected by transfusion, although transmission via blood is thought to be rare and has been estimated at 0.1% per transfused component in the United States (18).
As in previous reports (2,3,5–7), most of the KS cases in our kidney recipients were due to KSHV reactivation, as 91% of the recipients with KS belonged to group A. The prevalence of KS in kidney recipients infected with KSHV before transplantation was found to be 15%, 5 years after transplantation. This percentage is unlikely to change radically in the next few years as the mean time between organ transplantation and KS onset is 20 months, with a range of a few weeks to 18 years (19). The cofactors associated with KS development differed from those identified in retrospective studies (6,7), with age and black skin identified as the only independent risk factors in our study. Age is widely accepted to have an impact on the development of classical KS, and may also affect that of iatrogenic KS (20), although such an effect is not frequently reported for this type of cancer (21). Among KSHV positive recipients, the impact of black skin on KS development has never been reported. In this study, HLA matching was similar in recipients with or without KS (data not shown). A recent report on a South African population described an association between higher HHV-8 load in saliva (presumably reflecting higher reactivation rates) and two HLA A haplotypes typical of black African populations (22). As many different immunosuppressive treatment combinations were used in the transplant units participating in this study, too few patients were treated with identical protocols for a meaningful statistical comparison. However, no patient treated with sirolimus developed KS. Sirolimus has tumor inhibitor properties and has been used successfully for KS treatment, although occasional relapses are observed (23–25). We found that the KSHV viral load and the results of the optical density readings in the K8.1 EIA were not predictive of the development of KS in asymptomatic patients. The 3-month interval between consecutive blood samples may have been too long for detection of the transient KSHV viremia that may precede KS onset. However, determining the KSHV viral load may be useful in patients with clinical signs, to confirm or exclude the presence of replicative KSVH infection.
The transmission of symptomatic KSHV infection to a seronegative recipient through the graft has been reported before (26) and seems to be frequent in some settings (27). In group B, we observed only two cases of KS in 64 patients. In France, clinical events resulting from the transmission of KSHV from donor to recipient were therefore not frequent. In addition, the antiviral treatment (ganciclovir or acyclovir) prescribed to prevent infection with cytomegalovirus or herpesvirus in 68.7% of group B patients, including the two recipients with KS, may have lowered the number of primary clinical infections, although it did not prevent KS in these two cases. In some case reports, primary KSHV infection resulted in a potentially lethal acute syndrome characterized by fever, hemophagocytic syndrome and bone marrow failure (26,28,29). One of the patients in group B presented severe primary HHV-8 infection 8 months after transplantation. KSHV viremia was detected 2 months before the clinical manifestations and KSHV seroconversion was confirmed more than 1 year later. All clinical signs were resolved by treatment with rituximab (14).
The small number of clinical events in group B contrasts with the high seroconversion rate, which had reached 31.22% at 36 months. To date, after a median follow-up of 48.9 months after transplantation, only three of the seroconverting recipients in group B (see above) have had any clinical manifestations potentially related to KSHV infection or have been found to have detectable levels of KSHV DNA in peripheral blood. Preexisting or acquired KSHV infection had no impact on survival and graft loss 3 years after transplantation. This may be partially due to improvements in KS management in recent years (23,25,30).
This study is subject to several limitations. The high specificity but lower sensitivity (31) of the latent IFA used to define groups A to C may have resulted in some patients being assigned to group B or C, although they were actually seropositive at the time of transplantation and should therefore have been assigned to group A. The inclusion of group C patients in group A would have reduced the cumulative frequency of KS development. However, no cases of KS were observed in group C, suggesting that latent IFA at the time of transplantation had correctly identified all patients who developed KS due to prior infection with KSHV (group A) or transmission of the virus from the donor (group B). In addition, given the high specificity of the latent IFA, we are confident that the patients and donors identified with this assay really were infected with KSHV. The cumulative KS incidence rates are therefore valid for patients with KSHV infection diagnosed with a stringent assay. Furthermore, the low number of patients in groups A and B did not allow detecting slight differences of graft survival or death between the three groups.
In conclusion, the observation that 13% of KSHV seropositive transplant recipients developed KS without major impact on survival suggests that KSHV seropositive recipients should not be systematically excluded from kidney transplantation. However, they should be aware of the additional risk of KSHV morbidity. Actually, KSHV seropositive donors cannot be detected prior transplantation due to the lack of reliable KSHV serologic assays for use in emergency conditions. Knowing the KSHV serologic status of the donor may however be useful in order to be able to closely monitor these patients.
We thank all the French hospital coordinators of organ procurement, the Regional support units of the French Biomedicine Agency, all the physicians of the French nephrology, urology, dermatology and virology departments who participated in this study, Nouredine Balegroune, Lilia Ben M’Barek, Mirela Duman, Agnès Dumont, Olivier Lada, Catherine Milliancourt and Anne Marie Roque-Afonso, MD, for data management. This work was supported by the French Society of Dermatology, Agence de Biomédicine, Programme Hospitalier de Recherche Clinique and CNRS grants.