Safety and Immunogenicity of Zoster Vaccine Live in Human Immunodeficiency Virus–Infected Adults With CD4⁺ Cell Counts >200 Cells/mL Virologically Suppressed on Antiretroviral Therapy

Live attenuated herpes zoster vaccine administered to HIV-infected adults suppressed on antiretroviral therapy with CD4⁺ counts ≥200 cells/µL was generally safe and immunogenic. Antibody responses were similar to those observed in older adults without HIV infection who received the same vaccine.

Abstract

Background

Herpes zoster (HZ) risk is increased in human immunodeficiency virus (HIV)–infected persons. Live attenuated zoster vaccine (ZV) reduces HZ incidence and severity in adults; safety and immunogenicity data in HIV-infected adults are limited.

Methods

We conducted a randomized, double-blind, placebo-controlled trial in HIV-infected adults virally suppressed on antiretroviral therapy (ART). Participants, stratified by CD4⁺ count (200–349 or ≥350 cells/µL), were randomized 3:1 to receive ZV or placebo on day 0 and week 6. The primary endpoint was serious adverse event or grade 3/4 signs/symptoms within 6 weeks after each dose. Immunogenicity (varicella zoster virus [VZV]–specific glycoprotein enzyme-linked immunosorbent assay and interferon-γ enzyme-linked immunospot assay responses) was assessed at 6 and 12 weeks postvaccination.

Results

Of 395 participants (296 ZV vs 99 placebo), 84% were male, 47% white, 29% black, and 22% Hispanic; median age was 49 years. Safety endpoints occurred in 15 ZV and 2 placebo recipients (5.1% [95% confidence interval {CI}, 2.9%–8.2%] vs 2.1% [95% CI, .3%–7.3%]; P = .26). Injection site reactions occurred in 42% of ZV (95% CI, 36.3%–47.9%) vs 12.4% of placebo recipients (95% CI, 6.6%–20.6%) (P < .001). Week 12 median natural log VZV antibody titer was higher for ZV (6.30 [Q1, Q3, 5.64, 6.96]) vs placebo (5.48 [Q1, Q3, 4.63, 6.44]; P < .001) overall and in the high CD4⁺ stratum (P = .003). VZV antibody titers were similar after 1 or 2 ZV doses. Polymerase chain reaction–confirmed HZ occurred in 2 participants (1 ZV; 1 placebo); none was vaccine strain related.

Conclusions

Two doses of ZV in HIV-infected adults suppressed on ART with CD4⁺ counts ≥200 cells/µL were safe and immunogenic.

Clinical Trials Registration

NCT00851786.

Herpes zoster (HZ) occurs with increased frequency and severity in elderly or immunocompromised persons [1–5]. While the incidence has decreased in human immunodeficiency virus (HIV)–infected adults virologically suppressed on antiretroviral therapy (ART) [6], HZ continues to occur, particularly in the setting of immune reconstitution inflammatory syndrome following initiation of ART [6–9]. Severe postherpetic neuralgia or other complications are also more common in HIV-infected persons with low CD4⁺ T-cell counts, and may not be prevented with ART [1–5]. Vaccine-induced immunity to varicella zoster virus (VZV) may reduce the incidence and severity of HZ in this population [10–17].

The efficacy of live attenuated Oka/Merck VZV vaccine (ZOSTAVAX; zoster vaccine live [ZV]) was established in the Shingles Prevention Study (SPS), a large, randomized, double-blind, placebo-controlled trial in adults aged ≥60 years; the burden of HZ illness was reduced by 61.1%, the incidence of postherpetic neuralgia by 66.5%, and the incidence of HZ by 51.3% vs placebo [11]. In a second study, the ZOSTAVAX Efficacy and Safety Trial (ZEST) in adults aged 50–59 years, ZV also reduced the incidence of HZ by 69.8% [12]. These outcomes were correlated with postvaccination increases in VZV glycoprotein enzyme-linked immunosorbent antibody (gpELISA) and the geometric mean fold rise (GMFR) in VZV antibody titer measured at 6 weeks postvaccination. No serious adverse events (SAEs), or clinical, laboratory, or body system abnormalities aside from injection site reactions (ISRs) and headache occurred more commonly in ZV than placebo recipients. More recently, data from studies of a recombinant adjuvanted HZ subunit vaccine (HZ/su) showed that it was 96.6% effective in reducing the incidence of HZ in adults ≥50 years of age, 91.3% effective in those aged ≥70 years, 91.2% and 88.8% effective, respectively, in reducing postherpetic neuralgia, and was generally safe [16, 17]. Immunocompromised persons and people with HIV infection were excluded from all of these vaccine efficacy trials.

Previous studies demonstrated that heat-inactivated VZV vaccine increased cell-mediated immune responses to VZV, reduced the incidence and severity of HZ, and was safe and well tolerated in immunocompromised VZV-seropositive bone marrow and hematopoietic stem cell transplant recipients [14, 15]. Two doses of live attenuated VZV vaccine administered 12 weeks apart to HIV-infected, VZV-seronegative children was safe and immunogenic, although the study was not powered to assess efficacy [15]. The safety, immunogenicity, and efficacy of HZ vaccines have not been fully characterized in HIV-infected adults, including those virologically suppressed on ART. This study was conducted to address the hypothesis that 2 doses of ZV would have comparable safety and immunogenicity to that observed in the SPS when given to HIV-infected adults virologically suppressed on ART with CD4⁺ T-cell counts of ≥200 cells/µL. A 2-dose schedule was selected to address the concern that lower vaccine immunogenicity might be observed in HIV-infected persons compared with healthy adults (Supplementary Data).

METHODS

Study Design

We conducted a randomized, double-blind, placebo-controlled trial at 43 AIDS Clinical Trials Group Network sites to evaluate safety and immunogenicity of 2 doses of ZV or placebo in HIV-infected adults ≥18 years of age. The study was reviewed and approved by local institutional review boards at all participating sites. All participants provided written informed consent. An independent data and safety study monitoring committee reviewed the safety data and interim results.

Study Population

Eligible participants had CD4⁺ T-cell counts of ≥200 cells/µL on a stable ART regimen (no change within 90 days prior to entry), undetectable plasma HIV RNA levels (lower limit of detection <75 copies/mL) within 90–180 days of entry, and no detectable plasma HIV RNA within 30 days prior to enrollment and were either VZV seropositive or had a history of varicella or HZ >1 year prior to entry. Randomization was stratified by screening CD4⁺ T-cell count: ≥200–349 cells/µL (low CD4⁺ stratum) or ≥350 cells/µL (high CD4⁺ stratum). Participants were excluded if they had a nadir CD4⁺ count <100 cells/µL by history, prior VZV vaccines, or other immunosuppressive disorders; were pregnant or breastfeeding; or were taking chronic antiherpesvirus medications.

Intervention

Participants received 1 subcutaneous injection of 0.65 mL of ZV or placebo on day 0 and at week 6.

Follow-up

Study visits occurred at weeks 2, 6, 8, and 12. If the second vaccination was delayed, evaluations were performed at 2 and 6 weeks, respectively, after the second vaccination. Participants reporting signs or symptoms suggestive of HZ, varicella, or vaccine-related adverse events (AEs) of severity greater than Division of AIDS (DAIDS) grade 1 or mild [18] were scheduled for an interim clinic visit. All participants were followed for 42 days after each vaccination and were contacted by telephone after each vaccination and at week 24 to collect HZ or vaccination-related AEs.

Safety Evaluations

All AEs of DAIDS grade 2 or higher [18], any signs or symptoms that led to a change in treatment, and all signs and symptoms suggestive of HZ or varicella occurring within 42 days after each vaccination were recorded. Thereafter, only serious AEs attributed to vaccination, signs and symptoms suggestive of HZ or varicella, and deaths were recorded. Vaccination report cards completed by participants recorded daily temperatures, systemic AEs, rash (HZ- or varicella-like), nonstudy vaccines or medications received during 42 days after each vaccination, and ISRs for 5 days after each vaccination, and were reviewed at each study visit.

Identification, Evaluation, and Confirmation of Suspected Cases of Herpes Zoster

Participants experiencing varicella-like, HZ-like, or generalized rashes were asked to notify the site and return for an interim clinic visit within 72 hours of onset. Study personnel recorded rash characteristics, HZ or varicella exposure history, complications, and any medications used; performed a targeted physical examination; and, from those with rashes that were clinically judged to be varicella-like or HZ-like, obtained 2 swabs of lesions for polymerase chain reaction (PCR) testing to confirm VZV and differentiate vaccine strain from wild type. Participants were examined every third day until no new lesions appeared and all older lesions were healing, and were offered supportive treatment at the discretion of the investigator.

All lesion specimens were shipped to PPD Vaccines and Biologics, LLC, in Wayne, Pennsylvania, for PCR detection of VZV and herpes simplex virus (HSV) DNA, as previously described [11]. If VZV DNA was detected, the suspected case was classified as confirmed; if VZV DNA was not detected, a case was not considered confirmed. “Clinical HZ” was defined as signs and symptoms compatible with HZ without laboratory confirmation.

Immunogenicity Evaluations

Immunogenicity evaluations were performed at 6 weeks after each dose of ZV or placebo. VZV-specific gpELISA antibody responses were measured for all participants (PPD Vaccines and Biologics). Interferon-gamma (IFN-y) enzyme-linked immunospot (ELISPOT) cell-mediated immune responses (University of Pittsburgh Immunology Specialty Laboratory, Pennsylvania) were measured in peripheral blood mononuclear cells obtained on the first 40 participants entering stage 2 in each CD4 stratum.

Primary and Immunogenicity Endpoints

The primary composite safety endpoint was the occurrence of International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH)–defined SAEs or DAIDS grade 3 and 4 signs and symptoms, excluding SAEs related to trauma, during the 6-week period after receipt of any dose of ZV or placebo [18, 19]. Immunogenicity endpoints were VZV-specific gpELISA antibody and IFN-γ ELISPOT responses 6 weeks after 1 and 2 doses of ZV or placebo.

Statistical Analysis

Target accrual for this study was 400 participants randomized 3:1 to ZV or placebo to accrue 285 participants who received ZV and had evaluable safety data. A decision rule for stage 1 interim safety was if >3 of the first 36 ZV recipients experienced a composite safety endpoint, then a decision about whether to continue, modify, or halt the study would be made. Per the protocol, the end-of-study decision rule for the ZV arm reflected the observed safety endpoint rate in the placebo arm: ZV would be considered to have acceptable safety if ≤18 of 285 ZV recipients with evaluable safety data experienced a study-defined primary endpoint. For further details of the derivation of the sample size and decision rules, see Supplementary Data (event rates based in part on [20]). Analyses of safety endpoints employed a 2-sided α = .05 level.

The immunogenicity endpoint was defined as the within-arm means of the natural log-transformed VZV-specific gpELISA antibody titers at 6 weeks after each vaccination. To control the overall type I error on immunogenicity at a 1-sided .025 level, sequential testing was used such that antibody titers at week 6 were only tested at the 1-sided .025 level if week 12 titers were significantly different at the 1-sided .025 level. A second immunogenicity endpoint was defined as the within-arm, per-protocol maximum GMFR in VZV-specific IFN-γ ELISPOT response in a subset of 40 participants in each CD4 stratum with prespecified 1-sided .05-level test.

RESULTS

Study Participant Characteristics

A total of 395 participants (296 ZV and 99 placebo) were enrolled from April 2009 through June 2011 (Figure 1); demographics are summarized in Table 1. The majority were on ART regimens composed of nucleoside reverse transcriptase inhibitors plus either nonnucleoside reverse transcriptase inhibitors or boosted protease inhibitors.

Figure 1.

Flowchart summarizing participant randomization, cohort assignment, dosing completion, and participant disposition, by treatment arm.

Table 1.

Demographic and Baseline Characteristics of Study Participants

Characteristic	Total (N = 395)	ZOSTAVAX (n = 296)	Placebo (n = 99)
Male sex	333 (84)	252 (85)	81 (82)
Race/ethnicity
Black/African American	116 (29)	90 (30)	26 (26)
White	185 (47)	141 (48)	44 (44)
Hispanic/Latino, regardless of race	87 (22)	60 (20)	27 (27)
Other/unknown	7 (2)	5 (2)	2 (2)
Age, y, median (Q1, Q3)	49 (44, 56)	49 (44, 56)	49 (44, 55)
Nadir CD4+, cells/µL, median (Q1, Q3)	188 (108, 287)	197 (109, 287)	178 (104, 285)
Entry CD4+, cells/µL, median (Q1, Q3)	394 (281, 605)	399 (282, 607)	362 (276, 579)
Prior AIDS diagnosis	385 (97)	288 (97)	97 (98)
Varicella >5 y prior to entry	296 (75)	219 (74)	77 (78)
HZ >1 year prior to entry	131 (33)	101 (34)	30 (30)

Data are presented as No. (%) unless otherwise indicated.

Abbreviations: HZ, herpes zoster; Q1, first quartile; Q3, third quartile.

Vaccine Safety

Tables 2 and 3 summarize vaccine safety data. Overall, 17 protocol-defined primary endpoints were observed within 6 weeks of the first or second vaccine dose, 15 (5.1%) in ZV and 2 (2.1%) in placebo recipients. Eight participants (2.7%) in the ZV group and 2 (2.0%) in the placebo group experienced ICH-defined SAEs while grade 3 and 4 AEs not associated with an SAE occurred in 7 (2.4%) participants in the ZV arm.

Table 2.

Primary Safety Endpoints and Vaccine Adverse Events

Endpoint/Adverse Event	ZOSTAVAX (n = 295) Estimate, % (95% CI)	Placebo (n = 97) Estimate, % (95% CI)	P Valuea
Primary safety endpointsb	n = 15 5.1 (2.9–8.2)	n = 2 2.1 (.3–7.3)	.261
Injection site reactions	n = 124 42.0 (36.3–47.9)	n = 12 12.4 (6.6–20.6)	<.001
Rashes (generalized, varicella-like, or HZ-like)	n = 15 5.1 (2.9–8.2)	n = 4 4.1 (1.1–10.2)	1.00
Fever >38.3°C	n = 12 4.1 (2.1–7.0)	n = 6 6.2 (2.3–13.0)	.405

Abbreviations: CI, confidence interval; HZ, herpes zoster.

^a P values for the primary endpoint between arms are based on Fisher exact tests.

^bInternational Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use–defined serious adverse events (SAEs) and Division of AIDS grade 3 and 4 non-SAE-associated signs or symptoms during 6 weeks following each vaccination.

Table 3.

Characteristics and Timing of Primary Safety Endpoints

Primary Safety Endpointa	ZOSTAVAX			Placebo
	Event	Week	CD4+ Stratum	Event	Week	CD4+ Stratum
SAEa	Cutaneous VZVb	1	High	Avascular hip necrosis	8	Low
	Pneumonia	6	Low	Malaise, pneumonia	12	Low
	Drug-induced delirium	6	High
	Complex migraine	6	High
	Bulimia	8	Low
	Depression	8	Low
	UTI	8	High
	Cellulitis	9	Low
Grade ≥3 non-SAEa	Fever	0	High
	Acute pain	2	Low
	Hypertension	4	Low
	Fever	5	Low
	Headache	9	High
	Anorexia	12	Low
	Fever	12	High

Abbreviations: SAE, serious adverse event; UTI, urinary tract infection; VZV, varicella zoster virus.

^aInternational Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use–defined SAEs and Division of AIDS grade 3 and 4 non-SAE-associated signs or symptoms occurring during the 6 weeks following each vaccination.

^bCutaneous VZV event was originally reported as possible disseminated VZV, later updated to the adverse event term “vaccine-related rash,” deemed probably related to study therapy.

ISRs, characterized as pain, tenderness, edema, erythema, or rash, were reported more frequently in participants who received at least 1 dose of ZV compared with placebo recipients (42% and 12.4%, respectively). The majority of ISRs were reported as grade 1 or 2. The number of participants with and severity of ISRs were similar across CD4⁺ strata. With the exception of ISRs, there were no statistically significant differences in AEs between arms or across CD4⁺ strata.

Nineteen participants reported varicella- or HZ-like rashes (Table 2). Twelve had specimens tested by PCR; 2 were PCR-positive for wild-type VZV, 2 were PCR positive for HSV, and 8 were PCR negative.

Suspected or Confirmed Herpes Zoster Cases

Four participants had clinically suspected HZ (Table 4); 2 were confirmed HZ based on PCR testing, which was positive for wild-type VZV (1 ZV, 1 placebo recipient).

Table 4.

Suspected Cases of Herpes Zoster

Study Arm	CD4+ Stratum	Week	No. of Dermatome(s)	PCR Result
Placebo	Low	19	Single	Wild-type VZV positive
ZOSTAVAX	Low	7	Single	Wild-type VZV positive
	Low	8	Single	HSV positive
	High	1	Single → Disseminateda	Negative

Abbreviations: HSV, herpes simplex virus; PCR, polymerase chain reaction; VZV, varicella zoster virus.

^aParticipant experienced prolonged, disseminated, cutaneous VZV-like rash between weeks 1 and 6. Four specimens were tested during this time and all were PCR-negative for VZV and HSV. This event was originally reported by the investigator as possible disseminated VZV but subsequently updated to the adverse event term as “vaccine-related rash,” deemed probably related to study therapy.

Vaccine Immunogenicity

Tables 5 and 6 summarize VZV gpELISA results by treatment arms pooled across CD4⁺ strata. The median of log_n VZV antibody titer was significantly higher in the ZV arm than for placebo at both week 6 and 12 (P < .001 for both). The mean changes from baseline for log_n VZV antibody titer were higher in the vaccine arm at weeks 6 and 12 compared with placebo. VZV titers were higher in the ZV arm in both CD4⁺ strata compared with placebo (Supplementary Data).

Table 5.

Varicella Zoster Virus Glycoprotein Enzyme-Linked Immunosorbent Assay Antibody Log_n Titer by Treatment Arm, Pooled Across CD4 Strata

Timepoint	ZOSTAVAX (n = 296)	Placebo (n = 99)	P Valuea
Baseline
No. missing	5	2
Median (Q1, Q3)	5.59 (4.96, 6.50)	5.47 (4.71, 6.51)	.312
Week 6
No. missing	10	3
Median (Q1, Q3)	6.41 (5.62, 6.98)	5.58 (4.73, 6.49)	<.001
Week 12
No. missing	23	8
Median (Q1, Q3)	6.30 (5.64, 6.96)	5.48 (4.63, 6.44)	<.001
Change from baseline to week 6
No. missing	14	3
Median (Q1, Q3)	0.39 (0.08, 0.93)	0.00 (–0.10, 0.20)	<.001
Change from baseline to week 12
No. missing	26	8
Median (Q1, Q3)	0.46 (0.11, 0.93)	0.02 (–0.19, 0.19)	<.001

Abbreviations: Q1, first quartile; Q3, third quartile.

^a P value based on Wilcoxon tests.

Table 6.

Varicella Zoster Virus Antibody Geometric Mean Titer and Geometric Mean Fold Rise, by Treatment Arm Pooled Across CD4 Strata

GMT/GMFR	ZOSTAVAX (n = 296)	Placebo (n = 99)	P Valuea
No. missing	12	3
Week 6 GMT, mean (95% CI)	534.4 (480.0–594.9)	263.7 (204.0–340.8)	<.001
Week 6 GMFR, mean (95% CI)	1.78 (1.64–1.92)	1.05 (.98–1.12)
No. missing	23	9
Week 12 GMT, mean (95% CI)	530.3 (477.8–588.6)	250.3 (191.7–326.8)	<.001
Week 12 GMFR, mean (95% CI)	1.80 (1.66–1.95)	1.04 (.96–1.13)

Abbreviations: CI, confidence interval; GMFR, geometric mean fold rise; GMT, geometric mean titer.

^a P values for geometric means between arms based on t tests of log-transformed data.

The log_n gpELISA VZV antibody titer was modeled longitudinally using linear mixed-effects methods including treatment, CD4⁺ count, baseline titer, and age. ZV recipients had significantly higher antibody titers (P < .001) than placebo, and participants in the higher CD4⁺ stratum had significantly higher antibody titers than those in the lower CD4⁺ stratum (P = .003) (data not shown). VZV-specific antibody geometric mean titers (GMTs) and GMFR were also significantly higher in ZV recipients compared with placebo at both week 6 and week 12 (P < .001; Table 6).

Table 7 summarizes VZV-specific IFN-γ ELISPOT responses by treatment arms pooled across CD4⁺ strata. The maximum GMFR in the vaccine arms was compared with placebo and the GMFR at week 6 and week 12 is included to illustrate trends. With the prespecified 10% type I error for this analysis, the maximal GMFR was greater in the vaccine arms than in the placebo arm (P = .07). The actual GMFR was significantly greater at week 6 in the vaccine arm than in the placebo arm, although at week 12 the difference was not significant. Additional analyses of the interactions between GMFR and nadir CD4⁺ cell count, baseline CD4⁺ cell stratum, and age were not significant (data not shown).

Table 7.

Geometric Mean Fold Rise in Varicella Zoster Virus Enzyme-Linked Immunospot Responses, by Treatment Arm Pooled Across CD4 Strata

Treatment Arm	Week 6 GMFR (90% CI)	Week 12 GMFR (90% CI)	Maximum GMFR (90% CI)
ZOSTAVAX	1.8 (.8–3.7) (n = 56)	3.8 (1.8–8.0) (n = 54)	6.7 (3.4–13.3)
Placebo	0.4 (.2–1.1) (n = 18)	1.1 (.3–3.4) (n = 16)	1.6 (.7–3.4)
P valuea	.09	.17	.07

Abbreviations: CI, confidence interval; GMFR, geometric mean fold rise.

^aStatistically significant result was defined as a P value <.10 based on analysis of covariance.

DISCUSSION

The incidence of HZ in HIV-infected persons is estimated at 9.3 per 1000 person-years in those on ART, and HZ continues to cause substantial morbidity [9, 21, 22]. In 1 study, 28% of persons with HIV and HZ developed complications, including 12% with postherpetic neuralgia [22]. The risk is greatest in the 90 days after starting ART and in those with detectable plasma HIV RNA and CD4⁺ counts <350 cells/μL [22]; thus, effective vaccine prevention in still needed in this population.

The SPS and ZEST demonstrated that ZV was safe and effective in reducing the incidence of HZ, morbidity, and postherpetic neuralgia compared with placebo in older adults [11, 12]. We demonstrated that ZV is safe and immunogenic in HIV-infected adults with CD4⁺ counts of ≥200 cells/µL who were virologically suppressed on ART. ISRs were more frequent in ZV than placebo recipients, but there were no statistically significant differences in the rates of other vaccination-related SAEs or grade 3 or 4 AEs between the ZV and placebo arms.

In both the SPS and ZEST, gpELISA antibody responses were significantly higher for ZV than placebo [11, 12, 23–25]. In the SPS, the estimated fold differences in both GMT and GMFR from baseline for ZV vs placebo were 1.7 each (95% confidence interval, 1.6–1.8). The antibody response explained 45.9% of the ZV effect on HZ outcomes. In ZEST, the GMFR in gpELISA antibody was a better correlate of protection than the absolute levels of antibody at week 6 postvaccination. Both the VZV gpELISA titer and the GMFR in VZV antibody titer postvaccination were significantly higher for ZV than placebo recipients in our study, and baseline CD4⁺ count was significantly related to VZV antibody response after controlling for baseline titer and age; that is, VZV antibody responses were more robust in ZV recipients in the higher than the lower CD4⁺ stratum. Week 6 and 12 postvaccination VZV gpELISA GMTs and GMFR in our study were comparable to those reported in the SPS, although somewhat lower than those observed in ZEST. VZV-specific cell-mediated immune responses in a subset of participants in the SPS indicated that responder cell frequency and VZV-specific IFN-γ ELISPOT assay results correlated with vaccine protection [24]. Similarly, in a subset of participants from our study, maximal IFN-γ ELISPOT responses were significantly greater in ZV recipients than for placebo recipients.

Data evaluating the safety, immunogenicity, and efficacy of either live attenuated ZV or recombinant HZ/su in HIV-infected adults, including those virologically suppressed on ART, are limited. In a study of heat-inactivated vaccine, 80 HIV-infected adults with CD4⁺ counts of <200 cells/µL who were on ART for at least 30 days were randomized 3:1 to 4 doses of vaccine or placebo [26]. IFN-γ ELISPOT and gpELISA antibody 30 days after the fourth dose were generally higher in vaccine than placebo recipients; AE rates were similar. However, immune responses were less robust than in our study, although all participants in that study had CD4⁺ cell counts <200 cells/µL; plasma HIV RNA was not reported. Another study enrolled 67 HIV-infected adults with CD4⁺ cell counts of ≥400 cells/µL and plasma HIV RNA levels <1000 copies/mL who received 2 doses of live-attenuated varicella vaccine administered 12 weeks apart [27]. Although the vaccine was safe, VZV responder cell frequency, lymphocyte proliferation, and IFN-γ ELISPOT responses were only modestly higher in vaccine than placebo recipients. The lower immunogenicity observed in that study may have resulted from lower vaccine potency compared to that of ZV in our study, and not all participants were on ART or virologically suppressed. Last, in 94 persons with HIV virologically suppressed on ART with CD4⁺ cell counts ≥200 cells/µL, similar to our study population, HZ/su was safe and immunogenic, as measured by VZV CD4⁺ T-cell responses and gpELISA antibody levels [28]. Although generally similar results were observed for those with CD4⁺ counts of 50–199 cells/µL or those not on ART but with CD4⁺ counts >500 cells/µL, numbers were too small in these subsets to draw firm conclusions [28].

Our study has several limitations. It was not powered to fully evaluate efficacy of ZV in preventing HZ or HZ-related complications. Confirmed HZ due to wild-type VZV occurred in only 1 vaccine and 1 placebo recipient. Neither developed postherpetic neuralgia or other complications. Although the study was not powered for efficacy, the magnitude of the VZV gpELISA antibody titers in the higher CD4⁺ cell stratum and the GMFR in VZV antibody titers overall in ZV recipients were similar to responses associated with significant reductions in the incidence of HZ and postherpetic neuralgia in HIV-uninfected adults in SPS [11]. Similar to the experience in healthy adults, we did not observe a further increase in antibody response after a second ZV dose, suggesting that only 1 ZV dose is necessary to induce an immune response that may be protective [29]. We did not follow participants beyond 24 weeks, so longer-term durability of the ZV immune response is unknown. Last, our sample size was not large enough to detect differences in rare AEs in the ZV and placebo arms. Of note, SAEs, sometimes fatal, have been described after inappropriate ZV administration [30–32].

In a recent update, the Advisory Committee on Immunization Practices preferentially recommends the use of adjuvanted recombinant HZ/su vaccine over live attenuated ZV to prevent HZ and related complications in adults aged ≥50 years; however, a recommendation for immunocompromised persons, including those with HIV, is lacking, as these were excluded from efficacy studies [33]. We demonstrated that live attenuated ZV, in the largest study to date of HZ vaccines in HIV-infected adults with CD4⁺ counts ≥200 cells/µL who are virologically suppressed on ART, is safe and associated with immunogenicity similar to that in HIV-uninfected older adults in the SPS. While the benefit-risk profile supports the use of ZV in this population to potentially reduce the incidence of and morbidity associated with HZ, data remain insufficient to preferentially recommend HZ/su or ZV.

Supplementary Data

Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Acknowledgments. We thank all of the site investigators, research personnel, and especially the study participants. A5247 participating sites and personnel are summarized in Supplementary Appendix I.

Author contributions. C. A. B., J. W. A., J. H. J., A. F. R., P. A., and J. L. L.: study concept/design, data analysis/interpretation, and manuscript preparation. C. A. B., M. C. K., M. G., P. T., and J. L. L.: patient enrollment, data collection, data interpretation, and manuscript preparation. S. R., L. P., and D. B.: study and pharmacy regulatory oversight, protocol preparation, manuscript preparation, and review. D. R.: data collection, data management, manuscript preparation. B. J. C. M., C. R. R., R. M., and C. J.: laboratory assay data collection and analysis, data interpretation. J. C. M. and Z. P.: data analysis/interpretation and manuscript preparation.

Disclaimer. The opinions expressed in this manuscript are those of the authors and may not necessarily reflect those of Merck & Co, Inc, or the National Institutes of Health (NIH). All coauthors approved the final version of the manuscript.

Financial support. This work was supported by the National Institute of Allergy and Infectious Diseases (NIAID) of the NIH (award numbers UM1 AI068634, UM1 AI068636, UM1 AI106701, UM1 AI069432, UM1 AI069494, UO1 AI068634, UM1 AI069471, UM1 AI069419, UM1 AI069511, UM1 AI069467, UM1 AI069534, and UM1 AI069418). Merck & Co, Inc, collaborated on this study by providing ZOSTAVAX and financial support for varicella zoster virus-specific glycoprotein enzyme-linked immunosorbent antibody and polymerase chain reaction testing.

Potential conflicts of interest. C. A. B. has received grants from the NIH/NIAID, personal fees from GlaxoSmithKline, and grants/contracts from Gilead Sciences to her institution. J. W. A. and M. C. K. have received grants from NIH/NIAID. B. J. C. M. has received grants from NIH. M. S. has received grants from Gilead Sciences and personal fees from UpToDate and Clinical Care Options. P. T. has received personal fees from Merck, ViiV, and Gilead. J. L. L. has received grants from Bristol-Myers Squibb and ViiV. A. F. R., P.A., J. M., and Z. P. are employees of Merck & Co, Inc, and may hold stock and/or stock options in the company or receive other support from the company outside the work. All other authors report no potential conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.