Human Herpesvirus 8 (HHV8) Transmission and Related Morbidity in Organ Recipients


Celeste Lebbe,


The aims of the study were to assess the risk of HHV8 transmission resulting from organ transplantation, and related morbidity in liver, heart and kidney transplant recipients. Donor and recipient serologies were screened between January 1, 2004 and January 1, 2005 using HHV8 indirect immunofluorescence latent assay (latent IFA) and indirect immunofluorescent lytic assay (lytic IFA). Recipients negative for latent IFA with a donor positive for at least one test were sequentially monitored for HHV8 viremia and underwent serological tests over a period of 2 years. The results showed that among 2354 donors, HHV8 seroprevalence was 9.9% (lytic IFA) and 4.4% (latent IFA). A total of 454 organ recipients (281 renal, 116 liver and 57 heart) were monitored over a 2-year period. Seroconversion was observed in 12 patients (cumulative incidence 28%) whose donor had positive latent IFA and in 36 patients (cumulative incidence 29%) whose donors were positive only for lytic IFA, without differences across types of transplants. Positive HHV8 viremia was detected in only 4 out of 89 liver transplant recipients during follow-up and not in recipients of other types of transplant. Two liver transplant recipients and one kidney transplant recipient developed KS. In conclusion, although HHV8 transmission is a frequent event after organ transplantation, HHV8-related morbidity is rather rare but can be life threatening. Donor screening is advisable for monitoring HHV8 seronegative liver transplant recipients.




human herpesvirus-8


immunofluorescence assay


Kaposi's sarcoma






In a first French national prospective study, 31% of seronegative recipients seroconverted for HHV8 after receiving a kidney from a HHV8 seropositive donor (2009). However, HHV8-related morbidity was low, since only 3 out of 64 patients subsequently developed either symptomatic primary infection or KS. In contrast, in our experience of organ recipients who received other types of organ from HHV-8 positive donors, higher HHV8-related morbidity was observed: in 6 out of 12 heart transplant recipients and in 2 out of 6 liver transplant recipients, including five cases of hemophagocytic syndrome with either polyclonal lymphoproliferation or KS, three of which were fatal (unpublished data). These alarming data prompted us to perform another French national prospective study aiming to determine the frequency of HHV8 seroconversion and HHV8-related morbidity according to the type of organ transplanted.

Patients and Methods


This French multicenter nationwide study was conducted from January 1, 2004 to January 1, 2005. All donors and recipients were screened at transplantation for HHV8 serological status, using both indirect immunofluorescence latent assay (latent IFA) and indirect immunofluorescent lytic assay as recommended for epidemiological studies, particularly in low-HHV8 seroprevalence populations (2001). Indeed HHV8 serological methods are not standardized, and most are “homemade” assays. As a rule, latent IFA is considered as highly specific but it has a lower sensitivity than lytic tests. Lytic IFAs are very sensitive but lack specificity. It is generally recommended to combine lytic IFA with a second assay in order to confirm positivity (2001, 2006).

Since our aim was to evaluate the risk of HHV8 transmission and prevent morbidity we decided (1) that donors positive on either latent IFA or lytic IFA were potential transmitters, (2) to exclude latent IFA-positive recipients (since latent IFA is a very specific technique, we considered these recipients as already contaminated by HHV8 before transplantation), (3) to include recipients with negative latent IFA but positive lytic IFA because we could not rule out false positives.

According to donor (D) HHV8 status, transplant recipients(R) could be assigned to one of four groups: D Latent IFA + R Lytic IFA + or −, D only Lytic IFA + R Lytic IFA + or −. Transplant recipients were prospectively monitored for HHV8 viremia and underwent serological tests every 2 weeks for the first 6 months and then monthly for 18 months, in addition to their usual biological and clinical follow-up. Over the whole follow-up period, a median number of five blood samples per patient were analyzed (first to third quartile: 2 to 10, maximum 10). Given that some patients had a follow-up of less than 12 months, this actually corresponded to a median number of seven blood samples over 12 months (first to third quartile: 3 to 10, maximum 18) thus amounting to a regular follow-up, although less frequent than initially planned.

Recipients with confirmed positive HHV8 viremia > 1000 copies/mL had immediate complete clinical and biological monitoring for HHV8-related disorders and rituximab 375 mg/m2 was administered weekly in four infusions.

Characteristics of recipients at baseline were extracted from the CRISTAL database of the Agence de la Biomédecine, the French organ procurement and transplantation Agency. After 2 years of follow-up, the following end-points were collected for all recipients: Kaposi sarcoma, Castleman's disease, HHV8-related lymphoma, transplant failure, death.

The study was approved by the Paris Saint-Louis Ethics Committee and patients gave written informed consent.

Serological assays

Antibodies against HHV8 latent nuclear antigen were detected by immunofluorescent assay (IFA) carried out on the BC-3 cell line. The sensitivity and specificity of this test are respectively 61–72%, and 99–100%. IgG antibodies directed principally against lytic HHV-8 antigens were detected using a commercial indirect IFA (Diasorin, Human Herpesvirus-8 IgGImmunofluorescent Assay, Biotrin, Lyon, France). The sensitivity and specificity of this test are respectively 100% and 94% according to the manufacturer and as previously described (2009, 1996, 2007). Sera were used at 1:100 and 1:64 dilutions, respectively, for latent and lytic IFA.

Quantification of HHV8 viremia

DNA was extracted from whole blood using a DNA Easy Kit (Qiagen France). HHV-8 viral load was determined in each sample by real-time quantitative PCR analysis of 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 (2009). The cut-off for positive HHV8 was 100 copies/mL, while the cut-off for rituximab therapy was 1000 copies/mL.

Statistical analysis

Data were presented as frequencies and percentages or median and range. Probabilities of seroconversion and seroreversion over follow-up time were estimated with their 95% confidence interval (95% CI) using usual methods for censored data. The risk of seroconversion was compared between patients groups using log-rank tests. All tests were two-sided, and p-values of 0.05 or less were regarded as indicating statistical significance.


Characteristics of donors and characteristics of recipients at baseline

Between January 1, 2004 and January 1, 2005, 2354 donors were serologically tested for HHV8 (89.3% of the total number). The seroprevalence was 9.9% using lytic IFA, 4.4% using latent IFA and 12.0% when either test was considered.

A total of 479 HHV8 (latent IFA)-negative patients received an organ from a seropositive donor according to our algorithm, and were initially considered for follow-up. Twenty five were excluded from analysis because of the lack of HHV8 serological follow-up. There remained 454 recipients, including 281 kidney, 116 liver and 57 heart transplant recipients. Baseline characteristics were available for 426 of these 454 recipients as shown in Table 1.

Table 1. Characteristics of patients at baseline
VariableHeart and or lungLiverKidney
No. of patients with data5498274
Mean age  (standard deviation), years46(13.8)45.6(17.4)48.6(13.6)
Gender, n (%)
Type of transplant, n (%)
 Langerhans Islets0(0)0(0)1(<1)
Induction therapy
 Ciclosporin, n (%)44(81)10(10)119(43)
 Tacrolimus, n (%)11(20)80(82)139(51)
 Corticosteroids, n (%)49(91)89(91)267(97)
 Azathioprine, n (%)12(22)3(3)11(4)
 Mycophenolate, n (%)38(70)48(49)246(90)
 Antilymphocyte serum, n (%)40(74)2(2)124(45)

Probability of HHV8 seroconversion and seroreversion

The results were divided up according to the four groups previously defined: D Latent IFA + R Lytic IFA + or −, D only Lytic IFA +R Lytic IFA+or− (Table 2).

Table 2. Rates of HHV8 seroconversion and seroreversion according to donor and recipient serological status at baseline
Donor HHV-8 serological statusLatent IFA +/lytic IFA + or −Latent IFA −/lytic IFA +
Recipient HHV-8 serological statusLatent IFA −/lytic IFA −Latent IFA −/lytic IFA+Latent IFA −/lytic IFA −Latent IFA −/lytic IFA+
  1. No. = number.

  2. −, Negative.

  3. +, Positive.

  4. d, days.

  5. *Not all patients tested on serology had testing for viremia.

Number of recipients (no = 454)804323794
Organ transplanted
No. of HHV-8 serological testing (%)
 Median (interval) if >16.5(2 to 23)8(2–20)7(2–30)7(2–22)
Median follow-up (range), months14(0 to 29)14(0–35)12(0–34)14(0–29)
Latent IFA (no.)1364314
 Probability estimate (95% CI)30% (13 to 44)21% (3–36)32% (21–41)24% (10–36)
 Median time lapse (range) months12(3–18)10(4–17)9(3–21)9(1–21)
 Lytic IFA (no.)42 116 
 Probability estimate (95% CI)89% (59–98)-79% (68–86)-
 Median time lapse (range) months9(1–24)-9(2–20)-
 Both tests (no.)12 36 
 Probability estimate (95% CI)28%(11–41)-29%(18–39)-
 Median time lapse (range) months15(3–18)-10(3–21)-
Latent IFA (no./no. seroconversion)11/134/637/4311/14
 Median time lapse (range) months1(0.5–7)2(21d–7)2(9d–6)2(10d–21)
 Lytic IFA (no./no. seroconversion)27/42-73/116-
 Probability estimate (95% CI)100%-94%(83–98)-
 Median time lapse (range) months3(8d–17)-1.5(6d–18) 
No.Lytic IFA negativation-22-50
 Probability estimate (95% CI)-100%-97%(82–100)
 Median time lapse (range) months-5(1d–18)-6(2–21)
HHV-8 viremia*
No. with viremia>100 during follow-up/1/622/350/1811/72
 no. tested    
No. of patients with Kaposi’ sarcoma1/80 liver1/43 liver0/2371/94
 transplant, HHV8transplant HHV8  
 viremia positiveviremia positive  

After an overall median follow-up of 13 months, the probability of HHV8 seroconversion, defined by the occurrence of at least one positive serological test with the latent IFA technique, ranged from 21% to 32% (Table 2). The median time-lapse ranged from 9 to 12 months, across the four groups and the frequency of seroconversion tended to be higher in the groups of double-negative (latent and lytic IFA) recipients. The probability of lytic seroconversion when lytic IFA was negative at baseline ranged from 79% to 89%. Seroconversion, defined by the occurrence of both latent and lytic IFA-positive tests (only in recipients who were double-negative at baseline) was seen in respectively 28% of patients whose donor had positive latent IFA and 29% of patients whose donors were positive only for lytic IFA, within a median time lapse of 15 and 10 months, respectively.

When heart, liver and kidney transplant recipients were considered separately, the probability of seroconversion according to the two tests ranged from 23% to 33% with no significant difference according to the organ transplanted (p = 0.65) (Figure 1). For this analysis, no subclassification of patients according to donor and recipient initial serological status was performed, given the small numbers of patients in each group.

Figure 1.

Probability of HHV-8 seroconversion using latent IFA, lytic IFA or both tests according to the organ transplanted.

In can be noted that we were able to show that most patients who secondarily seroconverted according to latent or lytic IFA seroreverted during follow-up, whatever the type of organ transplanted (Table 2). Table 3 shows the evolution of recipient HHV8 serological status during follow-up. For both latent IFA and lytic IFA we can distinguish three patterns: (i) patients who maintained seroconversion; (ii) patients evidencing seroconversion then seroreversion; (iii) reseroconversion after seroreversion. Eighteen to 30% of patients with HHV8 seroreversion using latent IFA later registered positive using the same technique.

Table 3. Recipient HHV8 serological status during follow-up
 Latent IFA +/lytic IFA + or −Latent IFA −/lytic IFA +
Donor HHV-8 serological status Recipient HHV-8 serological statusLatent IFA −/ Lytic IFA −Latent IFA −/ lytic IFA+Latent IFA −/ lytic IFA −Latent IFA −/ lytic IFA+
  1. −, Negative.

  2. +, Positive.

  3. d, days.

Number of patients (No)804323794
Latent IFA
No change673719480
Stable seroconversion 2 2  6 3
Transient seroconversion (seroreversion)10 4 31 9
Becomes+, comes back to baseline and changes again 1 0  6 2
Lytic IFA
No change382112144
Stable seroconversion15 0 430
Stable seroreversion 0 5  015
Transient seroconversion (seroreversion)12 0 33 0
Transient seroreversion 0 4  011
Becomes +, comes back to baseline and changes again15 0 40 0
Becomes −, comes back to baseline and changes again 013  024

HHV8 viremia and HHV8-related morbidity

Despite the high frequency of seroconversion, positive HHV8 viremia was infrequent since only four out of 89 liver transplant recipients had positive HHV8 viremia during follow-up, as shown in Figure 2. No patients had positive HHV8 viremia among the 224 kidney transplant recipients, nor among the 37 heart transplant recipients tested. Three donors of the four liver recipients had positive latent IFA while one was only lytic IFA-positive. Patients 1, 2 and 3 were positive for lytic IFA at baseline. Patients 1 and 2 remained asymptomatic. Patients 3 and 4 developed Kaposi's sarcoma and hemophagocytic syndrome. Patient 3 received 4 infusions of rituximab beginning 8 weeks after confirmation of HHV8 viremia (8 months after transplantation). At that time, he was neutropenic but free from symptoms. Despite rituximab therapy the patient died 4 months later from extensive KS associated with polyclonal HHV8-related lymphoproliferation. Patient 4 was living in Italy and his first HHV8 viremia was discovered 5 months after transplantation concomitantly with the appearance of mucosal and lymphadenopathic KS associated with hemophagocytic syndrome. Since rituximab has been shown to be potentially harmful in patients with KS, the calcineurin inhibitor was discontinued and sirolimus substituted, without rituximab use, and this enabled the resolution of KS and the hemophagocytic syndrome within 3 months.

Figure 2.

Descriptive analysis of HHV8 serological tests and HHV8 viremia in four patients with liver transplant presenting positive HHV8 viremia during follow-up.

One kidney transplant recipient developed KS with limited cutaneous lesions and asymptomatic gastric involvement 6 months after transplantation without concomitant or preceding HHV8 positive viremia. This patient was initially seronegative on latent and lytic IFA. The donor was HHV8 positive on lytic IFA and negative on latent IFA. The patient seroconverted on lytic IFA 2 months after transplantation. KS was controlled with minimized immunosuppression and a switch from calcineurin inhibitors to sirolimus.

When all potential HHV8-related diseases are considered, with a median follow-up time of 4.5 years (interquartile range 4.2–4.8), only two out of 116 (1.7%) liver transplant recipients and one out of 281 (0.3%) kidney recipients developed KS. None developed HHV8-related multicentric Castleman's disease or lymphomas. No HHV8-related disease was observed in heart transplant recipients.


This prospective study provides significant new information on HHV8 infection after transplantation.

First of all, we were able to assess HHV8 seroprevalence for a second time among organ donors in France and compare it to data obtained 4 years earlier in the same country (2009). According to these data, HHV8 seroprevalence using latent IFA rose from 1.08% to 4.4% over this period. This difference cannot be attributed to technical variability since the same test was used in the same laboratory for both studies. The influence of donor age, which was 6 years older between the two study periods, could be involved since HHV8 seroprevalence increases with age (2002). An alternative explanation is a trend to increasing HHV8 seroprevalence in France over time, which remains to be confirmed by independent studies.

This study enabled us to extend our original data restricted to kidney recipients to all organ transplant recipients. HHV8 seroconversion, defined by the occurrence of two distinct positive tests, was observed in 23–33% of patients with no statistical differences according to the organ transplanted, although the highest value was observed in the kidney transplant population. It should be noted that we did not measure the life-time risk of seroconversion, but only gave estimations for a 24-month period taking into account that the majority of seroconversions occurred within 12 months, when more than 50% of patients were still in follow-up. These data are quite similar to data obtained in our previous study showing a 31% seroconversion rate in a kidney transplant population using the same latent and lytic IFA tests, as well as a K8.1 EIA test. The interpretation of latent IFA seroconversion among recipients initially positive for lytic IFA is debatable because the lack of specificity of lytic tests precludes differentiation between true previous HHV8 infection and false initial positivity.

In accordance with the earlier study, the prevalence of HHV8 positive viremia was very low, less than 1% (4 patients out of 454 recipients), as opposed to the high probability of HHV8 seroconversion. This could be related to an adapted immune response against HHV8 as demonstrated by Lambert et al. (2006). One possible alternative explanation could be the limited number of HHV8-DNA testing performed during the recipients’ follow-up.

It is noteworthy that the four patients with positive HHV8 viremia were liver transplant recipients, while no heart nor kidney transplant recipients developed positive HHV8 viremia during these 2 years of follow-up. Two patients had high levels of viremia (>1000 copies/mL) and developed KS while the others with HHV-8 viremia levels between 100 and 250 copies/mL remained asymptomatic. Three patients were positive for HHV8 on lytic IFA at baseline and only one of them seroconverted according to latent IFA 2 months after transplantation. One patient who was negative both for latent and lytic IFA seroconverted for both tests. While transient HHV8 reactivation cannot be excluded for the two patients with low levels of HHV8 viremia, virological and clinical data favor a true HHV8 primary infection for the other two. HHV8 reservoir cells mostly consist of B lymphocytes, monocytes, endothelial cells or their progenitors which are widely present in the liver. A higher risk of symptomatic HHV8 transmission by liver transplant is likely, although the small number of events in our study precludes any definitive conclusion. Recently Pietrosi et al. conducted a prospective survey of liver transplant recipients in Italy (2011). A total of 215 recipients and 179 donors were serologically screened for HHV8. Eight recipients were considered HHV8 D+R−, among whom four developed primary HHV8 infection, all of them with positive HHV8 viremia. The discrepancy with our results (high incidence of clinical events in the Pietrosi et al. study) could probably be explained by different methodological designs and techniques and the number of subjects.

In anecdotal case reports, rituximab has been of benefit in HHV8 symptomatic primary infection (2006). Our retrospective personal analysis of French national data in 2000 showed resolution of severe symptomatic primary infection in two liver transplant recipients by early treatment with rituximab (4 months after transplantation and 1 month after initial positive viremia), in contrast to a heart transplant recipient who was treated rather late after KS was diagnosed (12 months after transplant and 8 months after initial positive viremia). In this study, the only patient with symptomatic HHV-8 primary infection who was treated with rituximab had a fatal outcome. So far, the experience of rituximab is too limited in these patients to make recommendations about its use. The availability of mTOR inhibitors, used preferentially because of the KS experience in the management of these patients, will probably be an obstacle to solving this issue.

Finally this study shows the variability of HHV8 serological tests at individual level in a population of transplant recipients. Therefore we believe that HHV8 serologic tests are useful in epidemiological studies, but cannot be considered reliable on an individual basis to confirm HHV8 transmission in organ transplant recipients, and this is a limitation of serological tests in this population. Indeed, immunoreactivity is decreased in patients treated with immunosuppressant agents. Under these regimens, both T and B cell responses are impaired by blocking of cellular proliferation after antigen stimulation, as well as inhibition of the cytokine production necessary for such stimulation, particularly within the 6 months after transplantation. Hypogammaglobulinemia and decreased response to vaccination antigens such as pneumococcus, diphtheria and tetanus are frequent in transplant recipients (2003). Variability in HHV8 serological responses has been already described in the context of HIV infection (2003, 2007). Apart from HIV patients, some cases of HHV8 seroreversion have also been demonstrated in hemodialysis patients (2006). These HHV8 seroreversions observed in immuno-suppressed patients probably result from a defective antibody response, although variability in HHV8 serological tests cannot be totally excluded. Despite these limitations at individual level, we believe that serological tests in transplant recipients remain a valuable epidemiological tool at population level and there are recent examples supporting this, such as the increase in mean antibody level with the intensity of BK virus infection in transplant recipients (2008). In the IMPACT study on the prevention of CMV disease using valganciclovir prophylaxis in D+/R− transplant recipients the proportion of seroconversions (IgM or IgG) was 55–62% (2010). All patients who had CMV disease seroconverted, 92.3% after disease onset and 7.7% at the onset of disease.

In conclusion the low morbidity of HHV8 primary infection, whatever the organ transplanted, and the lack of reliable specific HHV8 serological tests available in emergency argue against routine HHV8 screening of donors in France for the purpose of excluding HHV8-positive donors. However screening of this type aiming to provide results after transplantation using two different HHV8 serological tests could be helpful to monitor recipients at higher risk of being contaminated, such as liver transplant recipients. The management of risk-prone recipients is open to discussion. Systematic monitoring of HHV8 viremias raises problems of cost-effectiveness. However, waiting for manifestations of HHV8 primary infection such as fever, cytopenia, lymphoproliferation of KS, to perform HHV8 viremia may not be sufficient to avoid unfavorable outcome.


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, liver, cardiac and lung transplantation departments who participated in this study, Sophie Minh Muzeaux and Anne Marie Roque-Afonso, MD, for data management. This work was supported by the French Society of Dermatology, and grants from the Agence de Biomédecine and Programme Hospitalier de Recherche Clinique.

Funding Source

Agence de la Biomédecine, PHRC.


The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.