Methods

Patients

A retrospective analysis was performed on 284 consecutive patients who underwent an allogeneic sibling/haploidentical or unrelated donor stem cell transplant at University College Hospital between January 1998 and December 2002. All patients undergo regular long-term follow-up at UCH, and follow-up data were available from the medical records on 247 of these patients, and the following variables were analysed: diagnosis, donor source, conditioning regimen, pretransplantation serology for VZV, time to discontinuation of aciclovir, time to reactivation of VZV, severity of zoster infection, association between active GVHD requiring immunosuppression and reactivation, and association between VZV reactivation and disease relapse.

VZV reactivation

The diagnosis of VZV reactivation was based on clinical findings, with confirmation by detection of VZV DNA and/or viral culture in equivocal cases. Treatment was with aciclovir 30 mg/kg/day for 5 days intravenously (i.v.), in the majority of patients, followed, in some, by a 7-day course of valaciclovir 3 g/day orally. A proportion of patients received outpatient treatment only, with aciclovir 4 g/day orally for 7 days. Following treatment for VZV reactivation, patients recommenced aciclovir at 200 mg p.o. twice daily for 3–6 months.

Conditioning regimens

Patients received three main categories of conditioning regimen as follows:

1. Full intensity with T-cell depletion:
   • Total body irradiation (TBI) in combination with cyclophosphamide and, in some cases, thiotepa or fludarabine, followed by the infusion of G-CSF-mobilised peripheral blood stem cells or bone marrow, which had undergone T-cell depletion by either the addition of Campath-1G or -1H, or by CD34⁺ selection. A proportion of these patients also received in vivo Campath. Additional GVHD prophylaxis was given to those patients who received only in vitro Campath, in the form of cyclosporine.
   • Patients undergoing transplants from haploidentical donors received antithymocyte globulin, thiotepa, fludarabine and TBI, followed by the infusion of G-CSF-mobilised peripheral blood stem cells or bone marrow, which had undergone T-cell depletion by CD34⁺ selection.
2. Full intensity without T-cell depletion:
   • TBI in combination with either cyclophosphamide, melphalan or etoposide followed by the infusion of unmanipulated G-CSF-mobilised peripheral blood stem cells or bone marrow, with cyclosporine and methotrexate as GVHD prophylaxis.
   • Yttrium90-labelled CD45, busulphan and cyclophosphamide followed by the infusion of unmanipulated G-CSF-mobilised peripheral blood stem cells or bone marrow, with cyclosporine and methotrexate as GVHD prophylaxis.
3. Reduced intensity:
   • fludarabine, melphalan and Campath-1H followed by the infusion of unmanipulated G-CSF-mobilised peripheral blood stem cells or bone marrow. GVHD prophylaxis was with cyclosporine.

Prophylaxis

All patients received anti-viral prophylaxis using aciclovir 600 mg/day p.o. in divided doses from the first day of conditioning to day -1, 750 mg/m²/day i.v. in divided doses from day 0 until engraftment, and then 400 mg/day (200 mg p.o. twice daily) until the CD4⁺ cell count exceeded 200/mm³ and immunosuppressive therapy was discontinued. Aciclovir was recommenced if immunosuppression was restarted subsequently for the treatment of GVHD.

Surveillance for CMV reactivation was performed by PCR, and pre-emptive therapy commenced, as described previously.⁹

Statistical analysis

Patients who did not reactivate VZV were censored at the time of their last follow-up or death, and the cumulative incidence of VZV reactivation was estimated using Kaplan–Meier analysis.

Results

Incidence of VZV reactivation

The median follow-up for the entire cohort of 247 patients was 479 days (range 2–2026 days).

In all, 27 patients developed VZV reactivation at a median of 579 days following transplantation (range 259–1783 days). The actuarial incidence at 1 year was 2%, at 3 years 21 and at 5 years 34% (Figure 1).

Figure 1.

Cumulative incidence of VZV reactivation post transplant in the entire cohort of 247 patients.

Full figure and legend (37K)

Only one of the 247 patients had a breakthrough zoster infection while on aciclovir prophylaxis.

The other 26 in whom reactivation occurred had discontinued aciclovir prophylaxis prior to the reactivation.

A total of 64 patients was available for analysis following the planned discontinuation of aciclovir, with a median follow-up from the time of aciclovir discontinuation of 697 days (range 45–1614 days)

The median time on aciclovir for those patients who subsequently reactivated VZV (n=26) was 418 days (range 164–987 days), and for those who did not reactivate (n=38) was 398 days (range 116–959 days). The median time from stopping aciclovir to zoster infection was 135 days (range 16–1364 days) in the 26 patients in whom reactivation occurred, and the actuarial incidence of VZV reactivation following discontinuation of prophylaxis was 39% at 1 year and 44% at 3 years (Figure 2).

Figure 2.

Cumulative incidence of VZV reactivation following discontinuation of aciclovir prophylaxis, in 64 patients in whom aciclovir had been stopped.

Full figure and legend (37K)

Clinical features

Of the 27 VZV reactivations, 25 (93%) occurred in a dermatomal distribution. In one patient, cutaneous dissemination and presumed VZV pneumonitis occurred, and in one other case, there was cutaneous dissemination with gastrointestinal involvement (gastric ulceration). In this patient, aciclovir had been discontinued despite ongoing immunosuppression, in a protocol violation. In one case, a Ramsay Hunt syndrome developed, with no resolution of the resulting facial palsy.

Post-herpetic neuralgia of varying persistence occurred in seven patients (26%).

In the single case of clinical VZV reactivation, despite ongoing aciclovir prophylaxis, the infection was localised and responded promptly to high-dose oral aciclovir. No information is available concerning compliance with the prophylactic regimen.

In all, 11 patients were treated as outpatients, 13 as in-patients, two were managed at referring centres (details not available) and one patient declined any treatment. Of the 27 patients, 26 (96%) responded to higher doses of aciclovir, with one case of dermatomal zoster failing to respond to 10 days of i.v. aciclovir 30 mg/kg/day, as shown by the continued appearance of new vesicular lesions. Therapy was switched to i.v. foscarnet, to which a complete response was made.

Risk factors

Pre-transplant recipient VZV serology was available in 174 patients, of whom 163 (94%) were seropositive for VZV IgG. All reactivations occurred in seropositive recipients. Donor serology was available in 50 patients, of whom 47 (94%) were VZV seropositive.

The incidence of reactivation was similar in all of the main transplant groups, with 10/100 (10%) reactivating following reduced-intensity conditioning transplants, and 17/147 (12%) following full-intensity conditioning. Overall, conditioning regimens that involved T-cell depletion resulted in 18/164 (11%) reactivation, and those that were T-cell replete, 9/83 (11%) reactivation. Furthermore, of the 64 patients who were followed up beyond discontinuation of aciclovir, the incidence of subsequent reactivation was not influenced by whether the conditioning regimen contained T-cell depletion (41% zoster in T deplete group, 43% in T replete group). No differences in zoster incidence were noted when patients were subdivided according to age, donor type or underlying diagnosis (Table 1).

Table 1 - Rate of VZV reactivation according to subgroupings.

Full table

No association between active GVHD and VZV reactivation was seen, with only one case of limited chronic GVHD and one of extensive chronic GVHD at the time of zoster infection. In addition, no GVHD was triggered by infection. In two patients, disease relapse occurred within 6 months following VZV reactivation.

Discussion

The reactivation of VZV is a significant cause of morbidity in recipients of haemopoietic stem cell transplants (HSCT). The prompt use of anti-viral therapy has reduced the mortality historically associated with early infection, although cutaneous dissemination and visceral involvement still occur, as discussed above.

The introduction of anti-viral prophylaxis has therefore been investigated as a means of preventing clinical reactivation until immune reconstitution is adequate either to permit only a contained localised infection or, potentially, to abrogate such infection completely.

Previous studies have reported the use of differing doses of aciclovir for 6 months post-allogeneic stem cell transplant. In all of these studies, VZV reactivation was totally prevented during the period of prophylaxis, but infection occurred in a substantial proportion following discontinuation. Ljungman et al¹⁰ (1200 mg/day) described reactivation rates of 31% at 6 months after discontinuation of prophylaxis, Selby et al¹¹ (3200 mg/day) 45% at 6 months and Steer et al¹² (600 mg/day) 33% at 12 months.

The only published data on ultra-low-dose prophylaxis comes from Kanda et al,¹³ who gave aciclovir 400 mg/day until immunosuppression was discontinued. Data on 86 patients were analysed, of whom 45 had received the above prophylactic regimen. There were no cases of breakthrough VZV reactivation. After cessation of aciclovir at a median of 152 days, zoster infection occurred in five of 17 patients, giving an actuarial incidence of reactivation at 1 year after stopping aciclovir of 29%. All five of the patients in whom reactivation had occurred, however, had restarted immunosuppression prior to the time of reactivation, an observation that is consistent with the one reactivation that occurred in the current study when aciclovir was discontinued while immunosuppression was ongoing.

This led to the conclusion that not only was ultra-low-dose aciclovir effective in preventing zoster infection early post transplant, but that prolonging prophylaxis until the cessation of immunosuppression, rather than for a set period (eg 6 months) post transplant, might lead to a reduction in the incidence of zoster infection once prophylaxis is discontinued. This was interpreted as supporting the theory that the use of ultra-low-dose aciclovir could permit a level of subclinical reactivation that results in the reconstitution of sufficient VZV-specific T-cell immunity to maintain the virus long term in a clinically inactive state, and so prevent zoster infection.¹⁴

The data in the current study does not support this hypothesis. A larger group of patients have been followed up for a significantly longer period, with 64 patients stopping aciclovir prophylaxis, and data being available on a median follow-up of 697 days following cessation. A cumulative reactivation incidence off aciclovir of 39% at 12 months was demonstrated, a figure not dissimilar to previous findings in studies of higher dose/shorter prophylaxis. It would appear, therefore, that reducing the dose of prophylactic aciclovir to 400 mg/day does not prevent zoster infection occurring once prophylaxis is discontinued.

This study would, however, support the suggestion that use of aciclovir prophylaxis makes an impact on the clinical presentation and associated morbidity of zoster infection, with extremely low levels of cutaneous and visceral dissemination, and no deaths. It is possible that this reflects the contribution of at least partially reconstituted virus-specific immunity in ameliorating the severity of the clinical syndrome, and provides an opportunity to manage these more localised infections in the outpatient setting, with the beneficial impact on patient quality of life and on the cost of therapy that this would entail.

In addition, this study confirms the findings of previous reports of aciclovir prophylaxis in providing no evidence to suggest that acquired resistance to aciclovir is a consistent problem. We describe only one case of probable aciclovir resistance, although no tests were performed to confirm this in vitro.

With respect to how recipient immunity could be optimised in order to prevent reactivation, alternative methods are clearly required, as the use of pharmacological prophylaxis does not assist, despite evidence that the clinical illness, once aciclovir is stopped and reactivation occurs, is less severe.

While it is theoretically possible that prolonging the period of prophylaxis could result in a sufficiently robust immune system such that zoster would not then occur, the absence of even a trend to this with the use of prophylaxis would suggest that clinical infection is required to re-establish a protective antigen-specific response. One possibility would be to continue low-dose prophylaxis long term. This would be relatively nontoxic and easily administered, but there is a theoretical concern that viral resistance may emerge over lengthy periods of time, and, on a practical level, patients are often understandably keen to minimise long-term drug use, particularly as they have a considerable chance of avoiding viral reactivation.

Of note, once reactivations have occurred, further episodes are unusual, with only 2% reported in a large series of transplant patients who were not given prophylaxis, confirming the role of the antigen-stimulated immune response in maintaining the virus in an inactive state.⁵

A different approach to zoster reduction has been the use of heat-inactivated VZV vaccine, in an attempt to enhance the VZV-specific CD4⁺ T-cell population. It is postulated that the expansion of the virus-specific CD4⁺ population provides help for VZV-specific CD8⁺ T cells, as well as generating interferon- and other Th1 cytokines, and thus contributing to the prevention of clinical reactivation.^{1, 15} Vaccination at 1, 2 and 3 months post transplant resulted in no reduction in zoster incidence, but a major decrease in the severity of the infections documented.¹⁶ In the allogeneic transplant cohort, infection in the vaccinated group resulted in a mild cutaneous eruption, with <20 lesions and no post-herpetic neuralgia, in marked contrast to the severity of infection in the unvaccinated group. As proposed, enhanced recall responses to VZV antigen were demonstrated by T-cell proliferation in vitro, in vaccinated individuals, with no difference in the humoral response, as measured by VZV IgG titres. The same group have subsequently reported in a randomised trial¹⁷ the efficacy of this vaccine schedule with an additional pre-transplant dose, when given to recipients of autologous transplants. The vaccinated cohort had a significantly reduced incidence of zoster post-procedure (13 vs 33%), with enhanced in vitro CD4⁺ T-cell proliferation to VZV antigen. Although there are fundamental differences between autologous and allogeneic transplantation, this study provides the first evidence that clinical VZV reactivation can be suppressed by vaccination in a setting where otherwise a high incidence of zoster infection would be expected.

In allogeneic transplantation, it is possible that the administration of vaccine later in the post transplant period, when partial T-cell reconstitution has occurred, may result in sufficient enhancement of virus-specific immunity to prevent viral reactivation, and the high-risk pre-vaccination period could be covered by the use of propylactic aciclovir.

In conclusion, the regimen of ultra-low-dose aciclovir described here provides an effective, nontoxic low-cost method of minimising the risk of early VZV reactivation. Once prophylaxis has been discontinued, reactivation is not prevented by this regimen, but infection is of a localised dermatomal pattern in 93% of those who reactivate and responds well to treatment doses of aciclovir. This suggests that the use of higher doses of aciclovir or of better absorbed but more expensive alternatives such as prophylaxis is not necessary, although no comment can be made regarding the dose of aciclovir when used to protect against cytomegalovirus in addition to VZV, as in some transplant centres. Finally, if the re-establishment of latency without clinical infection is the aim in this group of patients, alternative/additional strategies will be required.

References

References