Frequent Human Herpesvirus-6 Viremia but Low Incidence of Encephalitis in Double-Unit Cord Blood Recipients Transplanted Without Anti-thymocyte Globulin

Abstract

Cord blood transplantation (CBT) is a known risk factor for human herpesvirus-6 (HHV-6) infection. We analyzed the nature of HHV-6 infections in 125 double-unit CBT recipients (median age 42 years) transplanted for hematologic malignancies with calcineurin-inhibitor/ mycophenolate mofetil prophylaxis and no anti-thymocyte globulin (ATG). One-hundred and seventeen (94%) patients reactivated HHV-6 by quantitative plasma PCR (median peak 7,600 copies/mL, range 100-160,000) at a median of 20 days (range 10-59) after transplantation. HHV-6 encephalitis occurred in 2 patients (1.6%), of whom one died and the other recovered with therapy. There was no association between high level HHV-6 viremia (≥ 10,000 or ≥ 25,000 copies/mL) and age, diagnosis, conditioning intensity, or dominant unit characteristics, or between high level viremia and transplant outcomes (engraftment, cytomegalovirus reactivation, day 100 grade II-IV acute graft-versus-host disease, day 100 transplant-related mortality, or 1-year disease-free survival). HHV-6 therapy delayed the onset of cytomegalovirus reactivation. Interestingly, HHV-6 resolution was observed in untreated patients, and resolution of viremia correlated with absolute lymphocyte count recovery. We observed a low incidence of encephalitis and no association with CBT outcomes. Our data suggests therapy in uncomplicated viremia may not be warranted. However, further investigation of the risk-benefit of HHV-6 viremia treatment and standardization of PCR testing is required.

Introduction

Human herpesvirus-6 (HHV-6) is a member of the Roseolovirus genus in the Betaherpesvirus subfamily of human herpesviruses. Type A and B variants have been identified, exhibiting different biological characteristics and disease associations. HHV-6B is highly prevalent in the human population, affecting over 90% of healthy individuals during childhood(1). Primary HHV-6 infection is recognized as the cause of exanthema subitum and fever(2). Like other herpesviruses, HHV-6 remains latent in host cells and very few cases of recurrent infection ever occur in immunocompetent adults(3). Early HHV-6 reactivation, however, occurs in about half of allogeneic hematopoietic stem cell transplantation (HSCT) recipients(4, 5) and end-organ disease is an important complication following allogeneic HSCT(6-8). HHV-6 reactivation can cause life threatening hepatitis(9), interstitial pneumonia(10), and encephalitis(7, 11-15). Reactivation has also been associated with subtle cognitive dysfunction(16), fever, skin rash(5, 6), and delayed neutrophil and platelet engraftment(1). HHV-6 can infect hematopoietic progenitor cells and resultant reduction in colony formation in vitro has been described(17). This could explain the delayed engraftment reported in patients who reactivate HHV-6 after allografting(18, 19). HHV-6 may also have immunomodulating properties that could enhance the reactivation of cytomegalovirus (CMV)(11).

Reported risk factors for HHV-6 reactivation in HSCT recipients include myeloablative conditioning(20, 21), one-antigen human leukocyte antigen (HLA)-mismatch(11, 22) CMV reactivation(5, 11, 23), total body irradiation(24), unrelated donor grafts(25), bone marrow grafts(26), gender-mismatched grafts(8) and the use of cord blood (CB) as a HSC source(13, 15, 27, 28). In a recently published study of 235 patients who underwent allogeneic HSCT with matched related sibling or unrelated donors including 34 recipients of CB, HHV-6 reactivation reduced day 180 overall survival(29). Studies in children have correlated blood levels > 1,000 viral copies/mL with increased non-relapse mortality(30). Studies in CB transplantation (CBT) recipients, in particular, indicate an increased incidence and severity of HHV-6 associated disease(13, 15). The goal of the current study was to analyze the incidence and severity of HHV-6 viremia and end-organ disease, and the natural history of HHV-6 viremia in the first 100 days following CBT in patients transplanted at Memorial-Sloan Kettering Cancer Center (MSKCC).

Methods

Patient and Graft Characteristics

Eligible patients for this retrospective analysis included all consecutive adult and pediatric first allograft recipients transplanted with CB for the treatment of hematologic malignancies at MSKCC. All CBT recipients during the study period received double-unit grafts. Of 141 consecutive double-unit CBT recipients transplanted from 2/28/2006 through 3/21/2012, 125 were evaluable for HHV-6 reactivation. A total of 16 patients were not, 14 due to insufficient testing (one or fewer tests performed due to physician preference but not due to early death) and 2 due to death within the first 3 weeks after CBT unrelated to HHV-6. All patients provided written informed consent for transplantation and outcome analysis and the study was approved by the MSKCC Institutional Review/Privacy Board.

CB units were selected on the basis of 4-6/6 HLA-A,-B antigen, -DRB1 allele match to the recipient, a cryopreserved total nucleated cell (TNC) dose of at least 1.5 × 10⁷/kilogram (kg)/unit (increased to 2.0 × 10⁷/kg/unit in 2011), and the bank of origin as previously described(31). Unit-unit HLA-match was not considered in CB unit selection.

Conditioning Regimens and GVHD Prophylaxis

All patients were hospitalized in high-efficiency particulate air filtered rooms and received similar supportive care. Pre-transplant conditioning varied according to patient's age, diagnosis, remission status, extent of prior therapy, and co-morbidities, and consisted of high dose myeloablative, reduced intensity, or non-myeloablative regimens. All patients received granulocyte-colony-stimulating factor (G-CSF) after CBT until neutrophil recovery. All patients received a calcineurin-inhibitor and mycophenolate mofetil (MMF) for graft-versus-host disease (GVHD) prophylaxis starting day -3, and none received anti-thymocyte globulin (ATG)(32, 33).

Viral Prophylaxis, Monitoring, and Anti-viral Treatment

All patients received acyclovir anti-viral prophylaxis at 250 mg/m² intravenously every 8 hours followed by 400 mg orally twice daily upon discharge (or 200 mg every 8 hours in children). CMV seropositive recipients did not receive CMV-specific prophylaxis. Post-transplant surveillance for HHV-6 was performed at Viracor-IBT Laboratories using quantitative PCR of HHV-6 DNA from plasma. The lower limit of detection was 100 DNA copies/ml(34). Patients were monitored post-transplant with testing once to twice weekly from day 14 (or earlier if clinically indicated) to approximately day 60, and thereafter at the treating physicians discretion. Over the course of the study, CMV reactivation was monitored by antigenemia until 6/2010 after which PCR was substituted. Patients treated for HHV-6 received intravenous foscarnet at either induction (90 mg/kg IV every 12 hours) or maintenance (90 mg/kg IV every 24 hours) doses. Induction versus maintenance dosing was chosen per physician preference according to the patient's renal function. Therapy guidelines were not standardized during the study period, however, and institution of therapy depended upon the patient's clinical status (including symptoms such as fever and renal function), the viremia level, and the treating physician's preference.

Study Definitions

HHV-6 reactivation was defined as ≥ 1 positive HHV-6 PCR value of ≥ 100 copies/mL. The duration of HHV-6 viremia was measured as days from the first to last positive test. A single positive test was, therefore, recorded as 1 day of viremia. HHV-6 encephalitis was defined as clinical encephalitis (confusion, memory loss, seizures) combined with magnetic resonance imaging (MRI) findings consistent with limbic encephalitis in the setting of simultaneous HHV-6 viremia(35). Isolation of HHV-6 from cerebro-spinal fluid (CSF) was not mandatory for the diagnosis of encephalitis. HHV-6 pneumonia was defined as pulmonary symptoms with infiltrates and concurrent HHV-6 isolation from the bronchial alveolar lavage (BAL) in the absence of other causative organisms as previously described(36).

Statistical Analysis

Fisher exact test compared categorical variables, and Wilcoxon rank sum test compared continuous variables, in patients with high and low peak HHV-6 viremia levels. Kaplan-Meier methodology estimated one-year disease-free survival (DFS). The cumulative incidence of neutrophil and platelet engraftment was evaluated using competing risk analysis with death as the competing event. The incidence of acute GVHD was estimated treating death and relapse as competing events. For the incidence of relapse, transplant-related mortality (TRM) was the competing event, and for the TRM calculation relapse was the competing event. Time-dependent Cox regression assessed the association between high HHV-6 viremia and CBT outcomes. For DFS, the hazard ratio (HR) and 95% confidence interval (CI) was estimated, and cause-specific HRs were estimated for other transplant outcomes. Time-dependent Cox regression was used to analyze the incidence of CMV reactivation in relation to high versus low level HHV-6 viremia. The association between weekly median values of HHV-6 and absolute lymphocyte counts was evaluated using Spearman's correlation coefficient, and tested considering repeated measures over time using generalized estimating equations (GEE) method with an unstructured correlation structure and robust covariance matrix. The Wilcoxon rank sum test was used to compare HHV-6 viral loads in patients who were observed and those who were treated. A day 45 landmark analysis of 1-year DFS was performed to examine the effect of treatment in patients who had HHV-6 viremia. In all analyses statistical significance was defined as P < 0.05 based on a two-sided test. Statistical analyses were performed in software packages SAS 9.2 (SAS Institute Inc., Cary, NC, USA), and R version 2.13 (The R Foundation for Statistical Computing).

Results

Characteristics of Patients and Grafts

Table 1 provides the patient and graft demographics for the entire cohort (n = 125). The median age was 42 years. The majority of patients had acute leukemia and received myeloablative conditioning.

Table 1. Demographics of 125 patients and their grafts.

Characteristic	Value
Median age	42 years (range 0.9-69)
Median weight	67 kilogram (range 8-125)
Diagnosis, N (%)
AML	43 (34%)
ALL	24 (19%)
Other leukemia, MDS, CML	12 (10%)
Lymphoma, CLL	46 (37%)
Conditioning, N (%)
Myeloablative	94 (75%)
Non-myeloablative	31 (25%)
Median unit-recipient HLA-match
10 allele	6/10 (range 2-9/10)
Median Infused TNC Dose × 107/kg
Larger unit	2.7 (range 1.4-11.3)
Smaller unit	2.0 (range 0.9-7.1)

N indicates number; AML, acute myelogenous leukemia; ALL, acute lymphoblastic leukemia; MDS, myelodysplastic syndrome; CML, chronic myelogenous leukemia; CLL, chronic lymphocytic leukemia; HLA, human leukocyte antigen; TNC: total nucleated cell dose.

Incidence, Severity and Duration of HHV-6 Viremia

Figure 1 depicts the incidence and severity of HHV-6 viremia in the first 100 days post-CBT. One-hundred seventeen of 125 patients (94%) reactivated HHV-6. The median day of HHV-6 onset was 20 days post-CBT (range 10-59 days), the median peak was 7,600 copies/mL (range 100-160,000), and the median time to peak viremia was 23 days (range 12-62 days). The median duration of viremia was 10 days (range 1-60 days).

Figure 1. The incidence and severity of HHV-6 viremia in the first 100 days after CBT (n = 125).

Nearly all patients reactivated, 77% ≥ 1,000, but only 6% > 100,000 and none ≥ 1,000,000.

Patients were classified as having low level (peak viral load < 10,000 copies/mL) or high level (peak viral load ≥ 10, 000 copies/mL) HHV-6 viremia during the first 100 days. This cut-off was arbitrary based on the distribution of viral load in our patient population and previous reports of this level representing a clinically relevant threshold of viral load for increased HHV-6 end-organ disease risk(37). Of patients with high level viremia ≥ 10,000 copies/ml, 33/51 (65%) had fever at viremia onset. This compared to 22/66 (33%) patients with low level viremia (p < 0.01).

HHV-6 End Organ Disease

Two patients developed HHV-6 encephalitis. The first was a heavily pre-treated 62 year-old with lymphoma who underwent non-myeloablative CBT. The patient reactivated HHV-6 on day 23 to 13,100 copies/mL. Foscarnet was started on day 27 and continued for 15 days with resolution of viremia. On day 43 the patient developed confusion, MRI showed encephalitis, and CSF was positive for HHV-6. Foscarnet was resumed, but the patient died of HHV-6 encephalitis. The second patient was a 16 year-old myeloablative CBT recipient with acute lymphoblastic leukemia who reactivated HHV-6 on day 18 to 21,000 copies/mL. On day 21, HHV-6 was 118,000 copies/mL. On day 24 the patient developed delirium and MRI showed encephalitis. The patient was treated with foscarnet induction and made a full recovery.

Four additional patients had HHV-6 isolated from a bronchoalveolar lavage (BAL) in the setting of viremia. At the time of the lavage, HHV-6 was not detected in the blood in one patient already on foscarnet therapy, and blood PCR was 9,000, 22,400, and 65,500 copies/mL in the other 3 patients. None of these patients fulfilled criteria for HHV-6 pneumonia, although 2 of the 4 patients received foscarnet therapy for viremia.

Recipient Risk Factors for High Level Viremia

A comparison was performed between patient and graft characteristics in those with either no viremia or a peak viremia of < 10,000 copies/mL (n = 74) versus ≥ 10,000 copies/mL (n = 51, Table 2). As the definition of high level viremia remains controversial, a second analysis used ≥ 25,000 copies/mL (n = 31) as the cut-off value for high-level viremia(38). In both analyses we found no association between high level viremia and recipient age, diagnosis, conditioning regimen, or graft characteristics.

Table 2. Comparison of recipient and graft demographics in 125 patients with no or low HHV-6 viremia versus high level HHV-6 viremia.

Characteristics	No or Low Level Viremia (< 10,000 copies/mL, n = 74)	High Level Viremia (≥ 10,000 copies/mL, n = 51)	P value
Median age	42 years (range 1-69)	43 years (range 1-64)	0.89
Median weight	70 kg (range 8-104)	65 kg (range 10-125)	0.13
Diagnosis, N (%)
Acute leukemia, MDS, CML (n = 79)	41 (52%)	38 (48%)	0.05
Lymphoma inc. CLL (n = 46)	33 (72%)	13 (28%)
Conditioning, N (%)
TBI-based myeloablative (n = 78)	44 (56%)	34 (44%)	0.26
Chemo-based myeloablative (n = 16)	8 (50%)	8 (50%)
Non-myeloablative (n = 31)	22 (71%)	9 (29%)
Median unit-recipient HLA-match*
A, -B antigen, -DRB1 allele	5/6 (range 4-6/6)	5/6 (range 4-6/6)	0.55
10 allele	6/10 (range 2-9/10)	6/10 (range 3-9/10)	0.13
Median dose
Infused TNC × 107/kg*	2.0 (range 1.2-11.3)	2.2 (range 1.4-10.7)	0.08
Infused CD34+ × 105/kg*	1.0 (range 0.1-4.8)	0.9 (range 0.2-3.7)	0.57

^*Unit dominant in engraftment.

N indicates number; kg, kilogram; AML, acute myelogenous leukemia; ALL, acute lymphoblastic leukemia; MDS, myelodysplastic syndrome; CML, chronic myelogenous leukemia; CLL, chronic lymphocytic leukemia; chemo, chemotherapy; HLA, human leukocyte antigen; TNC, total nucleated cell dose.

Transplant Outcomes and the Association with Peak HHV-6 Viremia

CBT outcomes for the entire 125 patient cohort were analyzed with a median follow-up among survivors of 31.8 months (range 4.4-77.2 months). Day 100 neutrophil and platelet engraftment were 98% (95%CI: 94-100) and 80% (95%CI: 73-87), respectively. The median unit-recipient HLA-match of the engrafting unit was 6/10 alleles (range 2-9/10), and the median infused TNC of the engrafting unit was 2.2 × 10⁷/kg (range 1.2-11.3). The incidence of day 100 grade II-IV acute GVHD was 51% (95%CI: 42-60). Day 100 TRM was 11% (95%CI: 5-17). One-year relapse incidence was 11% (95%CI: 6-17) and the 1-year DFS was 65% (95%CI: 57-74).

Cox regression analysis tested the association between level of viremia within the first 100 days and CBT outcomes (Table 3). There was no difference in the speed and success of neutrophil and platelet engraftment, acute GVHD, or TRM by 100 days after CBT, or DFS 1-year post-transplant in patients with high level (≥ 10,000 copies/mL) viremia versus no or low level (< 10,000 copies/mL) viremia. Specifically, among engrafting patients the median time to neutrophil engraftment was 22 days (range 7-38) versus 21 days (range 7-43 days) in the high and no/low viremia groups, respectively. Additionally, the median time to platelet engraftment was 48 days (range 32-99 days) versus 46 days (range 21-90 days) in high and no/low viremia groups, respectively. In a second analysis using a peak viremia cut-off of ≥ 25,000 copies/mL (n = 31) within the first 100 days, we also found no differences in engraftment, acute GVHD, TRM, or DFS between groups (data not shown).

Table 3. Cox regression analysis of no or low (< 10,000 copies/mL, n = 74) versus high (≥ 10,000 copies/mL, n = 51) peak viremia within the first 100 days and transplant outcomes (n = 125).

	Sustained Engraftment* at 100 Days		Day 100 II-IV aGVHD	Day 100 TRM	1-Yr DFS
	Neutrophils	Platelets
HR for ≥ 10,000 Viremia (95% CI)	0.98 (0.68-1.41)	1.01 (0.68-1.50)	1.41 (0.86-2.31)	1.04 (0.51-2.13)	0.71 (0.39-1.30)
P value	0.92	0.95	0.17	0.90	0.27

aGVHD indicates acute graft-versus-host disease; TRM, transplant related mortality; DFS, disease free survival; HR, hazard ratio; CI, confidence interval.

^*Estimate incorporates speed and success of neutrophil and platelet engraftment (speed and success).

Relationship between HHV-6 and CMV Viremia

We compared rates of HHV-6 viremia according to recipient CMV serostatus. There were no differences in the rates of HHV-6 viremia in CMV seropositive and CMV seronegative patients with 28/70 (40%) and 23/55 (42%) of each group developing high level HHV-6 viremia, respectively. In addition, using time-dependent Cox regression, the incidence of CMV reactivation in CMV seropositive patients was not different in patients who developed high versus low level HHV-6 viremia (HR = 1.04, 95%CI: 0.60-1.80, p = 0.897).

Relationship between HHV-6 Viremia and Absolute Lymphocyte Recovery

To evaluate for a potential association between HHV-6 viremia and immune reconstitution, we analyzed viral load over time and the recovery of absolute lymphocyte counts. We found these were negatively correlated with a Spearman's correlation coefficient of ρ = -0.339 (p = 0.009) (Figure 2).

Figure 2. Relationship between HHV-6 viral load and absolute lymphocyte recovery after CBT.

There was a significant association between absolute lymphocyte recovery (measured in K/mcL) and resolution of HHV-6 viremia. The HHV-6 viral load in the first 60 days post-CBT is shown in black circles and ALC recovery is shown in red stars (all values for all patients). The solid lines are splines over time for HHV-6 viremia (black) and ALC recovery (red), respectively.

Treatment of HHV-6 Viremia

One-hundred and ten patients either reactivated HHV-6 and never reactivated CMV, or reactivated HHV-6 prior to CMV. In this 110 patient cohort, HHV-6 viremia kinetics were compared in those patients whose HHV-6 was treated with foscarnet (n = 31) and those whose HHV-6 was observed (n = 79, Table 4). Patients treated for HHV-6 received foscarnet promptly after viremia reactivation. Treated patients were more likely to be febrile at HHV-6 onset: 22/31 (71%) were febrile (median 38.7°, range 38-40.5) compared with 30/79 (38%) of observed patients (median 38.8°, range 38-39.6, p < 0.01). Moreover, while the onset of HHV-6 viremia and time to peak viremia were similar between groups, treated patients had a significantly higher peak HHV-6 viremia (p < 0.01). Notably, the duration of HHV-6 viremia (days to resolution) was similar in the 2 groups.

Table 4. Comparison of HHV-6 viremia kinetics in 110 patients in whom HHV6 viremia was either observed or treated with foscarnet.

HHV-6 Characteristics	Observed Patients (n = 79) (Median, range)	Treated Patients (n = 31) (Median, range)
Onset	20 days (10-40)	19 days (10-56)
Peak (copies/ml)	3,800 (100-148,000)	24,100 (2,000-160,000)
Day of Peak	23 days (12-42)	21 days (14-57)
Foscarnet Onset	-	24 days (17-83)
Foscarnet Duration	-	9 days (1-22)
Viremia Duration*	9 days (1-49)	8 days (1-60)

^*Reflects number of days from first to last positive.

In CMV sero-positive patients who initially received foscarnet therapy for HHV-6, subsequent CMV reactivation was delayed compared to those whose HHV-6 viremia was observed: 59 days (range 46-81) versus 39.5 days (range 19-274) after CBT (p = 0.004). In addition, 2 patients whose HHV-6 was observed (but none of the HHV-6 treated patients) later developed CMV pneumonia. Finally, 7 additional patients of the study cohort were not included in the above n = 110 analysis as they reactivated CMV before HHV-6. In these, foscarnet therapy for CMV abrogated subsequent HHV-6 viremia (data not shown).

A day 45 landmark analysis was performed in patients who reactivated HHV-6 by day 45 to evaluate the effect of therapy on survival. The 1-yr DFS survival of foscarnet treated patients (n = 27) was 62% (95%CI: 41-78). This compared to 72% (95%CI: 61-81) in observed patients (n = 78) (log-rank test p = 0.843).

Discussion

CB has been successfully used as an alternative HSC source for the treatment of hematologic malignancies in patients who do not have suitably HLA-matched adult donors. Progression-free survival after single-unit CBT in children(39) or double-unit CBT in adults(32, 40, 41) is comparable to hematopoietic allografts from matched unrelated adult donors. CBT carries a high risk of infectious morbidity and mortality(42, 43), however, and the incidence and severity of HHV-6 associated disease may be increased in this population(13, 15). The naïve T-cells in CB, and hence the absence of HHV-6 primed T-cells, may contribute to the higher rate of HHV-6 reactivation following CBT. End-organ disease attributable to HHV-6, including limbic encephalitis and pneumonia, has been reported in CBT recipients when the viral copy number exceeds 10,000 copies/mL(37, 44). Furthermore, Hill et al reported that the incidence of HHV-6 associated post-transplant acute limbic encephalitis was more frequent (9.9%) and associated with higher mortality after CBT than adult donor HSCT(13).

In 125 CBT recipients from a single center, we found that nearly all patients (94%) reactivated HHV-6 early after transplantation, with 41% of patients reactivating to levels ≥ 10,000 copies/mL. No patient or graft characteristic correlated with high level viremia. This viremia incidence is higher than the 69% incidence in 54 CBT recipients reported by Betts et al(38), the 76% reactivation rate reported in 21 CBT recipients by Zerr et al(45), and the 72% incidence in 68 CBT recipients reported by Hill et al(13). However, we observed a lower median peak viremia of 7,600 copies/mL than that reported by Hill et al (median peak viral load in that study was 34,000 copies/mL with the upper limit of the range being > 1 million)(13). No patient in our analysis had viremia greater than 1 million copies/mL. Studies reporting high viral loads may have included some patients with chromosomal integration as 1 viral copy may exist per white blood cell (1 × 10⁷ HHV-6 genomes/ml of whole blood)(46). A factor that may be even more important in interpretation of HHV-6 studies is that differences in incidences of viremia and viral loads may be explained by varying sensitivity of different PCR assays as has previously been described for both Epstein-Barr virus and CMV infections (47, 48). The high incidence of viremia in this study could potentially be explained by an overly sensitive PCR at our institution. This considerably complicates therapy guidelines as a widely applicable viremia threshold for treatment cannot be currently recommended.

While we observed a higher incidence of viremia than prior reports(13, 38, 45), our series is notable for a very low incidence of HHV-6 encephalitis (less than 2% as compared, for example, with the 9.9% reported by Hill et al(13)). In addition, while subtle neurocognitive deficits described by others(16) may not have been detected in our retrospective analysis, no other types of end-organ disease were diagnosed in our population, although 2 of the 4 patients with HHV-6 isolated from the bronchoalveolar lavage received foscarnet therapy for viremia which potentially could have abrogated pulmonary disease. The low number of encephalitis patients precluded any evaluation of risk factors for end-organ disease, although both patients with encephalitis had viral loads greater than 10,000 copies/mL. We cannot rule out that our intervention with foscarnet therapy mitigated subsequent end-organ disease in the 31 treated patients. While the 79 patients who were observed (Table 4) did not develop disease, their median peak viremia levels were lower than treated patients. This could have contributed to the relatively low level of disease in our study. Only a large randomized controlled trial would be able to further investigate this question.

There was no association between the development of high level viremia and CBT outcomes including sustained donor engraftment, aGVHD, day 100 TRM, and DFS. DFS was also comparable in a landmark analysis when foscarnet therapy was taken into account. While consistent with the findings of Betts et al from the standpoint of survival(38), this is in contrast with Dulery et al who reported that high level viremia adversely affected overall survival(29), and the findings of Zerr et al who reported HHV-6 reactivation was associated with higher rates of aGVHD and mortality in 315 allograft recipients (including 21 CBTs)(45). It would also suggest that routine ex vivo generation of cytotoxic T-lymphocytes for HHV-6 for adoptive immunotherapy(49) may not be warranted.

Because the clinical implications of HHV-6 viremia are unclear, clinical management has been inconsistent, especially given the substantial concerns regarding the toxicity of foscarnet therapy(50-52). Nevertheless, recognizing the limitations of a retrospective review of non-randomized patients, this varying clinical practice permitted the comparison of the natural history of viremia in patients who were observed with those who were treated. Treated patients were more likely to be febrile, had higher viral loads, and promptly responded to therapy. As a group, patients who did not receive foscarnet had lower levels of viremia and, notably, had resolution of viremia without therapy. The mechanism of viral clearance is of great interest given this is observed in the first 6 weeks after transplantation. We postulate that the omission of ATG from the pre-transplant conditioning could have facilitated the low incidence of encephalitis and the ability of many patients to resolve without therapy. The significant association between diminishing viral load and lymphocyte recovery (Figure 2) merits further investigation into the role of innate natural killer cell immunity and T-cell repertoire recovery in HHV-6 clearance.

It is evident from this study that it remains unclear which patients require and/or benefit from therapy for HHV-6 viremia after CBT. Moreover, even the need for routine viral monitoring is open to question, as is the value of therapy in asymptomatic patients. This is practically relevant as many patients may be able to be safely observed. At this time, however, we recommend HHV-6 monitoring in CBT recipients as our understanding of HHV-6 is incomplete, and as this population as transplanted at MSKCC remain at a low but finite risk of end-organ disease. Furthermore, detection of viremia, especially at high levels, will enhance clinician vigilance for early signs of end-organ disease, such as subtle neuro-cognitive deficits that could otherwise be overlooked in patients with incipient encephalitis, or other symptoms such as pneumonitis that has been reported in other studies of allograft recipients.

At this time we strongly consider therapy in patients with symptoms like fever or with very high levels of viremia > 25,000 copies/ml at MSKCC, although the risk-benefit of this approach in those without evidence of end-organ disease is not established and would require a large randomized study. The CMV status of the patient can also be taken into account in this decision making given Foscarnet treatment of HHV-6 viremia delays subsequent CMV reactivation. This is at least one practical benefit of HHV-6 therapy in CMV seropositive CBT recipients given their high CMV reactivation rate and attendant morbidity in CMV seropositive CBT recipients(53). Additionally, patients who are treated for HHV-6 viremia without evidence of end-organ disease are given lower doses of Foscarnet (the equivalent of maintenance doses for CMV viremia) to mitigate the nephrotoxicity of Foscarnet therapy. Ultimately, clinicians will require greatly enhanced knowledge of HHV-6 biology to improve management of CBT patients who are at risk for both HHV-6 end-organ disease and the toxicity of pre-emptive therapy. Standardization of PCR assays would also greatly assist in the understanding of HHV-6 biology in allograft recipients and ultimately aid in therapy decisions.