Abstract
The present randomized, double-blind, placebo-controlled, multicenter clinical trial was designed to compare the efficacy and tolerability of sorivudine [1-β-d-arabinofuranosyl-E-(2-bromovinyl)uracil] and acyclovir for the treatment of dermatomal herpes zoster in human immunodeficiency virus (HIV)-seropositive patients. A total of 170 HIV-seropositive adults presenting with herpes zoster (confirmed by direct fluorescent-antigen testing and/or viral culture) were enrolled and randomized to receive a 10-day course of orally administered sorivudine (40 mg once daily plus acyclovir placebos) or acyclovir (800 mg five times daily plus sorivudine placebo). Patients were monitored daily to document the events of cutaneous healing, pain, zoster-related complications, and drug-related adverse events. Patients were reassessed on days 21 and 28 and then once monthly for 1 year. The primary efficacy endpoint was time to the cessation of new vesicle formation. Secondary efficacy endpoints included times to other events of cutaneous healing, resolution of pain, and frequency of dissemination and zoster recurrence. In a multivariate analysis, sorivudine was superior to acyclovir for reducing the times to the cessation of new vesicle formation (relative risk [RR] = 1.54, 95% confidence interval [CI] = 1.00 to 2.36; P = 0.049) and total lesion crusting (RR = 1.48, 95% CI = 1.07 to 2.04; P = 0.017). In a univariate analysis, there was a trend favoring sorivudine for the cessation of new vesicle formation (median of 3 versus 4 days; P = 0.07) and a significant advantage for time to total lesion crusting (median of 7 versus 8 days; P = 0.02). The time to the resolution of zoster-associated pain, the frequency of dissemination, and the frequency of zoster recurrence were not different between the two treatment groups. Both drugs were well tolerated. Sorivudine is an effective drug for the treatment of herpes zoster in HIV-infected patients and results in accelerated cutaneous healing when compared with acyclovir therapy.
Herpes zoster is a common opportunistic infection among individuals with human immunodeficiency virus (HIV) infection. The incidence of herpes zoster among HIV-infected persons is about 30 cases per 1,000 person-years, which is approximately 15-fold higher than the incidence in an HIV-seronegative control population (3, 10). Following prompt treatment with a drug active against varicella-zoster virus (VZV), most cases of herpes zoster in HIV-infected persons resolve without sequelae. However, a wide variety of complications, including persistent skin lesions, disseminated VZV, encephalitis, and acute retinal necrosis have been reported and occur with increased frequency in individuals with CD4+-cell counts of <200/mm3 (6). Only limited data are available from controlled prospective trials to guide physicians in the selection of appropriate therapy for herpes zoster in this population.
Sorivudine [1-β-d-arabinofuranosyl-E-5-(2-bromovinyl)uracil; BV-araU], a thymidine analog containing a bromovinyl side chain at the 5 position, is an investigational antiviral drug with extremely potent in vitro activity against VZV. Like acyclovir, sorivudine is monophosphorylated by virally encoded thymidine kinase. Unlike acyclovir, diphosphorylation of sorivudine is also dependent on viral enzymes. Sorivudine triphosphate blocks viral DNA replication by inhibiting DNA polymerase activity, but it is not incorporated into elongating viral DNA (4, 19). For clinical VZV isolates, the geometric mean effective concentration of sorivudine required to produce a 50% reduction in viral plaque formation is 0.001 μg/ml, making sorivudine at least 1,000-fold more active than acyclovir against VZV (7, 13). In addition to its excellent in vitro activity against VZV, sorivudine exhibits good activity against herpes simplex virus (HSV) type 1 (HSV-1) (comparable to the activity of acyclovir) but no activity against HSV-2 (4, 7). Unlike acyclovir, sorivudine is well absorbed after oral administration, with oral bioavailability of >60% (11, 12). Because of its excellent in vitro activity and favorable pharmacokinetic profile, sorivudine has been evaluated as a once-daily oral therapy for herpes zoster and varicella (2, 8, 16).
In this report we describe the results of a multicenter, randomized, double-blind, placebo-controlled clinical trial comparing sorivudine with acyclovir for outpatient therapy of herpes zoster in HIV-infected patients. This study was conducted by the National Institutes of Allergy and Infectious Diseases Collaborative Antiviral Study Group (CASG) in collaboration with the AIDS Clinical Trials Group (ACTG) (ACTG Protocol 169).
MATERIALS AND METHODS
Subject eligibility.
HIV-seropositive patients who were at least 18 years of age and who presented with a localized cutaneous eruption that was of less than 72 h in duration and that was clinically consistent with herpes zoster were evaluated for enrollment. Patients were required to have a Karnofsky Performance Status Index of ≥60 (i.e., “requires occasional assistance but is able to care for most personal needs”). All women of childbearing potential were required to have a negative serum pregnancy test prior to enrollment. Excluded from the study were patients who were receiving topical or systemic therapy with other drugs active against VZV, were unable to take oral medication, had other acute or life-threatening opportunistic infections, had evidence of visceral or cutaneous dissemination of herpes zoster, and had laboratory evidence of significant renal, hepatic, or bone marrow dysfunction. Also excluded were patients receiving treatment with 5-fluorouracil (5-FU) or 5-FU derivatives because of the potential for a serious drug interaction with sorivudine. All study subjects provided written informed consent in compliance with federal and local institutional review board guidelines.
Randomization and dosing.
In this double-blind trial, each patient was assigned to one of two treatment groups according to a computer-generated randomization scheme. At study enrollment, each patient received a box of 10 blister-pack cards, with each card containing 1-day’s supply of study medication. Patients in the acyclovir group took one 800-mg acyclovir tablet by mouth five times daily (7:00 a.m., 11:00 a.m., 3:00 p.m., 7:00 p.m., and 11:00 p.m.) and one sorivudine placebo tablet at 7:00 a.m. for 10 days. Similarly, patients in the sorivudine group took 40 mg of sorivudine by mouth at 7:00 a.m. and acyclovir placebo tablets five times daily. The study medications were identical in appearance to their corresponding placebos.
Clinical protocol.
Prior to the initiation of the study medication, each subject underwent a medical history review and a complete physical examination, and blood was collected for baseline laboratory studies. Patients were classified by using the Karnofsky Performance Status Index (score of 0 to 100) and the modified HIV classification system for HIV infection from the Centers for Disease Control and Prevention. Clinical samples for viral cultures were obtained from cutaneous lesions, and the lesions were scraped to obtain specimens for identification of VZV antigens by direct fluorescent-antigen testing. Patients were evaluated daily until the lesions were completely crusted for the presence of new vesicle formation within the involved dermatome; the percentage of the lesions that were maculopapular, vesicular, pustular, scabbed, or healed; and evidence of cutaneous or visceral dissemination. Patients were questioned regarding the severity of zoster-related pain (scored on a scale of 0 to 4), zoster-related sleep and activity impairment (scale of 0 to 3), and analgesic use (scale of 0 to 5). Pill counts were recorded to document compliance. Patients were discontinued from the study during the treatment phase for documentation of a pathogen other than VZV causing the skin lesions, visceral or cutaneous VZV dissemination, abnormal laboratory values (alanine aminotransferase and aspartate aminotransferase (AST) levels greater than fivefold the upper limits of normal; absolute neutrophil count, <500/mm3; platelet count, <30,000/mm3; creatinine level, >2.5 mg/dl), pregnancy, any serious adverse event possibly related to study medication, the need to administer protocol-prohibited medications, or an inability to comply with the protocol. Following completion of the 10-day drug treatment, patients returned on days 21 and 28 for evaluation of lesion status, pain severity, and zoster-related complications.
For the next 11 months, patients were interviewed monthly in the clinic or by telephone and were questioned regarding zoster-related pain, zoster-related complications, and recurrences of herpes zoster. Patients were discontinued from the study during the long-term follow-up phase in the event of herpes zoster recurrence or death.
Laboratory and safety studies.
Clinical samples for viral culture and a direct fluorescent-antigen assay were obtained on study day 1 prior to the initiation of treatment with the study medication to confirm the diagnosis of herpes zoster. In addition, specimens for acute- and convalescent-phase VZV serology were collected on study days 1 and 28. To assess drug safety, blood and urine were collected on study days 1, 3, 5, and 10 for routine hematologic profile, biochemical studies, and urinalysis. The CD4+ lymphocyte count was determined on day 1.
Statistical analyses.
For the purposes of sample size calculation, the time to the cessation of new vesicle formation was considered the primary study endpoint. We assumed that the proportion of patients in the inferior treatment group who continued to form new vesicles 4 days after enrollment would be 40 to 50% and that an alternative therapy would be considered superior if it reduced this proportion to 15 to 25%. To demonstrate a 25% difference between the inferior and superior treatment groups with a significance level of 5% and a power of 90% on the basis of a two-tailed test, a minimum of 79 evaluable patients were targeted for enrollment in each treatment group.
Clinical findings were recorded on standardized case report forms and were entered into a computerized database. Quality control procedures were instituted to ensure the accuracy of the data collection and entry processes. The distribution of covariates in the two treatment arms was checked to confirm the validity of the randomization process. Data for all 170 patients who received study medication were included in the safety analyses. For efficacy calculations, all except four patients with disease proven to be caused by HSV were included in an intent-to-treat analysis. The primary endpoint was the time to the cessation of new vesicle formation. Secondary endpoints included frequency of dissemination and times to events of cutaneous healing and pain resolution. All endpoints defined by time to event were analyzed by time-to-failure methods for censored observations. A multivariate analysis with the Cox proportional hazard model provided estimates of relative risk (RR), 95% confidence intervals (CIs), and P values. Covariates included in the Cox model were age, gender, duration of lesions at randomization, CD4 count, stage of HIV infection, and level of pain at enrollment.
The multiple testing procedure of O’Brien and Fleming (9) was used to determine the need for early termination of the trial if one treatment group performed markedly better than the other. Interim analyses were performed and efficacy and safety data were assessed by an independent Data Safety and Monitoring Board of the National Institute for Allergy and Infectious Diseases after enrollment of one-third and two-thirds of the projected total study population.
RESULTS
Demographics of study population.
A total of 170 subjects were enrolled at 30 participating centers (median number of patients per site, 5; range, 1 to 17) between September 1991 and June 1994. Virologic studies established that 165 patients had herpes zoster, 4 patients had rashes caused by HSV (type not specified), and 1 patient had a rash of unknown etiology. The demographics of the study population are summarized in Table 1. The median age of the subjects was 36 years, and 90% of the subjects were male. The median CD4+ lymphocyte count was 171 cells/mm3; 34% of the population had CD4+ lymphocytes count of <100 cells/mm3 and 45% had CD4+ lymphocyte counts of >200 cells/mm3. The median Karnofsky score was 90.
TABLE 1.
Demographics of study population
| Characteristic | Acyclovir group (n = 82) | Sorivudine group (n = 84) | Total (n = 166) | P value |
|---|---|---|---|---|
| Age (yr [median]) | 35 | 37 | 36 | 0.358 |
| Sex (no. [%] of subjects) | 0.220 | |||
| Male | 76 (93) | 73 (87) | 149 (90) | |
| Female | 6 (7) | 11 (13) | 17 (10) | |
| Race (no. [%] of subjects) | 0.422 | |||
| White | 56 (68) | 55 (65) | 111 (67) | |
| African American | 13 (15) | 16 (19) | 28 (17) | |
| Hispanic | 13 (16) | 10 (12) | 23 (14) | |
| Other | 1 (1) | 3 (4) | 4 (2) | |
| CD4 count (no. of cells/mm3 [median]) | 185 | 141 | 171 | 0.249 |
| No. (%) of subjects with the following: | ||||
| Prior herpes zoster | 28 (34) | 27 (33) | 55 (33) | 0.869 |
| Prior acyclovir use | 25 (30) | 30 (36) | 55 (33) | 0.474 |
| Anti-retroviral therapy | 50 (61) | 58 (69) | 108 (65) | 0.275 |
The duration of rash at enrollment (mean, 1.7 days) was ≤1 day for 42% of the subjects, 2 days for 37% of the subjects, and 3 days for 21% of the subjects (Table 2). The median number of discrete vesicles at enrollment was 75; 84% of the patients described prodromal pain in the involved dermatome. Acute neuritis at the baseline was reported to be absent or mild for 55% of the patients and moderate to incapacitating for 45% of the patients. Of the 166 subjects evaluated for efficacy, 104 completed 12 months of posttherapy follow-up, 19 were lost to follow-up, and 43 reached other study endpoints (Table 3).
TABLE 2.
Baseline clinical characteristics of the study population
| Characteristic | Acyclovir group (n = 82) | Sorivudine group (n = 84) | Total (n = 166) | P value |
|---|---|---|---|---|
| No. (%) of subjects with the following duration of rash: | 0.743 | |||
| <1 day | 5 (6) | 8 (10) | 13 (8) | |
| 1 day | 30 (37) | 27 (32) | 57 (34) | |
| 2 days | 30 (37) | 31 (37) | 61 (37) | |
| 3 days | 17 (20) | 18 (21) | 35 (21) | |
| No. of vesicles (median) | 96 | 70 | 75 | 0.279 |
| No. (%) of subjects with prodromal pain | 73 (89) | 66 (79) | 139 (84) | 0.068 |
| No. (%) of subjects with the following pain severity: | 0.507 | |||
| None | 24 (29) | 34 (40) | 58 (35) | |
| Mild | 17 (21) | 16 (19) | 33 (20) | |
| Moderate | 24 (29) | 19 (23) | 43 (26) | |
| Severe or incapacitating | 17 (21) | 15 (18) | 32 (19) |
TABLE 3.
Reasons for study termination and deathsa
| Reason for study termination or death | No. of subjects
|
|
|---|---|---|
| Acyclovir group (n = 84) | Sorivudine group (n = 86) | |
| Treatment phase (days 1 to 28) | ||
| HSV disease (not VZV) | 2 | 2 |
| Disseminated zoster | 1 | 1 |
| Possible adverse effects | ||
| Anxiety or confusion | 2 | 0 |
| Rash | 0 | 1 |
| Elevated LFTs | 1 | 1 |
| Ganciclovir for CMV | 0 | 1 |
| Failed to meet entry criteria | ||
| Elevated baseline LFTs | 1 | 0 |
| Other pathogen (unspecified) | 1 | 0 |
| Trauma (automobile accident) | 1 | 0 |
| Recurrent zoster | 4 | 1 |
| Follow-up phase (months 2 to 12) | ||
| Recurrent zoster | 6 | 12 |
| Death | 5 | 4 |
| Lost to follow-up | 8 | 11 |
| Deathb | ||
| HIV disease or wasting | 3 | 5 |
| Malignancy | 2 | 4 |
| Other OI | 2 | 2 |
Abbreviations: LFTs, liver function tests; OI, opportunistic infection; CMV, cytomegalovirus.
Two acyclovir recipients and seven sorivudine recipients were discontinued from the study after reaching defined endpoints (e.g., zoster recurrence) but died within 12 months of administration of study medication. Data for those subjects are included in this tabulation.
Clinical efficacy. (i) Cutaneous healing.
In the time-to-event univariate analysis (Fig. 1), there was a trend toward a more rapid cessation of new vesicle formation in the sorivudine group (median of 3 versus 4 days), but the difference did not reach statistical significance (P = 0.07). The time to total lesion crusting (Fig. 2) was significantly faster with sorivudine therapy than with acyclovir therapy (P = 0.02). The time to complete lesion healing was not different between the two treatment groups (P = 0.13). For the subpopulation of patients who entered the study with a rash duration of ≤1 day, no differences in healing endpoints were evident between the two treatment groups. However, for patients with a rash duration of ≥2 days at the time of presentation, the times to the cessation of new vesicle formation (P = 0.02) and total crusting (P = 0.004) were significantly shorter for the sorivudine-treated patients.
FIG. 1.
New vesicle formation. Kaplan-Meier curves demonstrating time to cessation of new vesicle formation for the sorivudine and acyclovir treatment groups (P = 0.07 by log-rank test).
FIG. 2.
Lesion healing. Kaplan-Meier curves demonstrating time to total lesion crusting for the sorivudine and acyclovir treatment groups (P = 0.02 by log-rank test).
Multifactorial Cox regression analysis (Table 4) demonstrated that sorivudine was superior to acyclovir for the time to the cessation of new vesicle formation (RR = 1.54, 95% CI = 1.00 to 2.36; P = 0.049) and for time to total lesion crusting (RR = 1.48, 95% CI = 1.07 to 2.04; P = 0.017). New vesicle formation occurs over a finite interval; therefore, patients presenting later in the course of illness should have a shorter observed duration of new vesicle formation. This assumption, based on the natural history of herpes zoster, was confirmed in the regression analysis, as indicated in Table 4.
TABLE 4.
Multifactorial Cox regression model
| Significant factor | P value | RR | 95% CI |
|---|---|---|---|
| Time to cessation of new vesicle formation | |||
| Sorivudine treatment | 0.049 | 1.54 | 1.00–2.36 |
| Rash duration of >1 day | 0.003 | 1.96 | 1.26–3.05 |
| CD4 count of <200/mm3 | 0.021 | 0.60 | 0.39–0.93 |
| Time to 100% crusting | |||
| Sorivudine treatment | 0.017 | 1.48 | 1.07–2.04 |
| CD4 count of <200/mm3 | 0.010 | 0.66 | 0.47–0.90 |
The CD4+ lymphocyte count correlated with rates of healing, with faster resolution observed in the group with low CD4 counts. A CD4+ lymphocyte count of <200/mm3 was a significant factor for a more rapid time to the cessation of new vesicle formation (RR = 0.60, 95% CI = 0.39 to 0.93; P = 0.021) and total crusting (RR = 0.66, 95% CI = 0.47 to 0.90; P = 0.010) (Table 4). Furthermore, in the subpopulation of patients with CD4 counts of ≤200 cells/mm3, lesion resolution was significantly more rapid in the sorivudine group (n = 49) than in the acyclovir group (n = 42), with faster times to the cessation of new vesicle formation (P = 0.03) and total crusting (P = 0.008). In the subpopulation of patients with CD4+ cell counts of >200/mm3, no differences in the healing rates were apparent between the acyclovir (n = 39) and sorivudine (n = 35) treatment groups for either new vesicle cessation (P = 0.83) or crusting (P = 0.54). Sorivudine was maximally effective in the subpopulation with low CD4 counts. For all patients treated with sorivudine, faster lesion resolution was evident in the group with CD4 counts of ≤200/mm3 than in the group with CD4 counts of >200/mm3 for both the new vesicle cessation (P = 0.02) and total crusting (P = 0.006) endpoints. No such differences were seen in the acyclovir treatment group.
(ii) Resolution of pain.
At each patient contact, information was recorded regarding the presence and severity of pain in the involved dermatome. The median time to resolution of zoster-associated pain (with no subsequent recurrence of the pain) was 54 days in the sorivudine group and 63 days in the acyclovir group (P = 0.22), as indicated in Fig. 3.
FIG. 3.
Pain. Kaplan-Meier curves demonstrating time to resolution of zoster-associated pain (ZAP) for the sorivudine and acyclovir treatment groups (P = 0.22 by log-rank test).
(iii) Quality of life.
Comparisons were made between the patients’ quality-of-life responses made on day 1 and those made on day 21 (or day 28 if day 21 data were unavailable). No significant differences were noted between the two treatment groups for reduction in pain severity, analgesic use, or sleep interruption or for improvement in level of activity.
(iv) Herpes zoster-related complications.
Of the 166 HIV-seropositive zoster patients enrolled in the study, only 2 developed possible VZV dissemination. One sorivudine recipient with zoster in a T-4 dermatome developed an axillary maculopapular rash on study day 4, but the subject completed the 10-day course of study medication. No clinical samples of the axillary rash were obtained for culture, but the investigator thought that it was consistent with cutaneous VZV dissemination. One patient in the acyclovir group developed lower-extremity paresthesias on study day 8 but completed the treatment phase. He was diagnosed with possible VZV myelitis on study day 13 (although no confirmatory laboratory studies were performed) and was treated with intravenous acyclovir, with resolution of his symptoms.
Three patients developed motor neuropathies within the involved dermatomes. Two patients with herpes zoster of the arm reported swelling of the arm and hand, with one case of swelling possibly being due to bacterial superinfection.
(v) Herpes zoster recurrences.
Recurrent episodes of herpes zoster were reported by 13 sorivudine recipients and 10 acyclovir recipients (P = 0.54). In the acyclovir group, four of the recurrences occurred during the acute (28-day follow-up) phase and six of the recurrences were reported during the 11-month follow-up. In the sorivudine group, 1 recurrence was seen during the acute phase and 12 were seen during the follow-up phase. The median time to zoster recurrence was 134 days in the sorivudine group and 58 days in the acyclovir group (P = 0.10).
Safety and tolerance. (i) Clinical adverse events.
Overall, both acyclovir and sorivudine were well tolerated and not associated with serious clinical adverse events, although 63% of the study participants reported some potential adverse event. The most commonly reported symptoms included nausea or vomiting, dizziness, and headache (Table 5). The adverse event profiles were not different between the acyclovir and the sorivudine recipients. Two patients in the sorivudine group discontinued the study medication because of adverse events potentially related to the study medications (one with rash and one with elevated lactate dehydrogenase levels). In the acyclovir group, patients were discontinued from the study because of anxiety or confusion (2 patients) and elevated hepatic enzyme levels (1 patient) (Table 3). For each of these patients, the relationship of the adverse event to the study medication was considered to be “unknown” or “possible” by the investigator.
TABLE 5.
Most frequently reported clinical adverse events
| Adverse event | No. (%) of subjects with adverse event
|
|
|---|---|---|
| Acyclovir group (n = 82) | Sorivudine group (n = 84)a | |
| Nausea or vomiting | 19 (23) | 11 (13) |
| Headache | 9 (11) | 6 (7) |
| Fever | 3 (4) | 7 (8) |
| Abdominal pain | 5 (6) | 4 (5) |
| Diarrhea | 4 (5) | 4 (5) |
| Dizziness | 5 (6) | 1 (1) |
| Rash | 2 (2) | 4 (5) |
Differences between treatment groups were not significant for any adverse event.
(ii) Laboratory abnormalities.
Laboratory abnormalities were noted in both treatment groups, although a causal relationship with the study drugs was frequently unclear for this population of patients who were receiving multiple other medications. The only laboratory abnormalities occurring during the treatment phase for which there were statistically significant differences between the two drugs were mild leukopenia associated with sorivudine and mild elevations in AST levels associated with acyclovir (Table 6). Among patients with normal peripheral blood leukocyte counts at the time of entry into the study, 7 of 48 in the acyclovir group and 12 of 36 in the sorivudine group developed grade 1 to 2 leukopenia (1,000 to 4,000/mm3) (P = 0.040). Grade 1 to 2 neutropenia (750 to 1,500/mm3) was seen in 14 of 71 acyclovir recipients and 15 of 57 sorivudine recipients who had normal neutrophil counts at enrollment (P = 0.686). One patient in each treatment group developed grade 3 or 4 neutropenia (<750/mm3). Among patients with normal baseline AST levels at enrollment, grade 1 or 2 abnormalities (levels of 1.25 to 5 times the upper normal level) were seen among 14 of 56 patients in the acyclovir group and 5 of 53 patients in the sorivudine group (P = 0.036). Two patients (one from each treatment group) were discontinued from the study because of abnormal liver function tests.
TABLE 6.
Laboratory abnormalities during therapya
| Clinical abnormality and treatment | No. (%) of subjects
|
P value | ||
|---|---|---|---|---|
| Normal at entry | Progression to grade 1 or 2 | Progression to grade 3 or 4 | ||
| Leukopenia | 0.040 | |||
| Acyclovir | 48 (59) | 7 | 0 | |
| Sorivudine | 36 (45) | 12 | 0 | |
| Neutropenia | 0.686 | |||
| Acyclovir | 71 (88) | 14 | 1 | |
| Sorivudine | 57 (72) | 15 | 1 | |
| Elevated AST level | 0.036 | |||
| Acyclovir | 56 (72) | 14 | 0 | |
| Sorivudine | 53 (66) | 5 | 0 | |
Toxicity grading for leukopenia: grade 1, ≥2,500 to <4,000/mm3; grade 2, ≥1,000 to <2,500/mm3. Toxicity grading for neutropenia: grade 1, ≥1,000 to <1,500/mm3; grade 2, ≥750 to <1,000/mm3; grade 3, ≥500 to <750/mm3; grade 4, <500/mm3. Toxicity grading for AST level: grade 1, 1.25 to 2.5 times the upper limit of normal; grade 2, 2.6 to 5 times the upper limit of normal.
Deaths.
Eighteen deaths occurred within the 12 months of study drug administration (7 in the acyclovir group and 11 in the sorivudine group). Nine of the deaths occurred in patients who had previously been discontinued from the study (six terminated from the study because of recurrences of herpes zoster), and nine deaths occurred among patients still in long-term follow-up. All deaths occurred among patients with CD4+ lymphocyte counts of ≤200/mm3 at the time of study entry. In each case, the cause of death was an opportunistic infection, an AIDS-related malignancy, or wasting syndrome (Table 3). No patients died as a result of VZV infection or from adverse events related to a study medication. No patient received 5-FU during the course of the study or developed bone marrow aplasia. The mean date of death was study day 139 for the sorivudine group and study day 165 for the acyclovir group; survival curves were not different between the two treatment groups (P = 0.72).
DISCUSSION
On the basis of clinical experience and studies performed with other immunocompromised patient populations, most clinicians have treated herpes zoster in HIV-infected patients with acyclovir. Unless there is evidence of a complication requiring intravenous antiviral therapy, herpes zoster is usually managed on an outpatient basis with a 1- to 2-week course of orally administered acyclovir at a dosage of 800 mg five times daily, although data validating this approach are limited. We report the results from the largest controlled clinical trial of antiviral therapy for herpes zoster in HIV-infected patients that has been conducted to date. These data confirm that oral administration of antiviral drugs for herpes zoster in patients with AIDS is an effective and appropriate therapeutic strategy and can be safely conducted on an outpatient basis. Even though the study population comprised significantly immunocompromised individuals (median CD4+ lymphocyte count, 171 cells/mm3), the incidence of herpes zoster-related complications was low. Overall, the clinical presentation and response to therapy in this HIV-infected population were relatively similar to those seen in members of the general population with herpes zoster.
Sorivudine is an effective treatment for herpes zoster in HIV-infected individuals and results in accelerated cutaneous healing when compared with standard oral acyclovir therapy. Furthermore, sorivudine is administered as a single capsule once daily, unlike the five-times-daily dosing schedule required for acyclovir. This simplified dosing regimen is more convenient and may result in improved compliance among HIV-infected individuals, who are likely to be taking many other medications. Famciclovir and valacyclovir, dosed three times daily for herpes zoster, have not yet been evaluated extensively for this indication in HIV-infected patients, but they will likely prove to be as effective as acyclovir (14).
Previous studies of herpes zoster in immunocompromised patients have demonstrated that the cessation of new vesicle formation is a reliable clinical indicator of the termination of viral replication and is a critical endpoint for the assessment of antiviral efficacy (1, 17). In the multivariate analysis, new vesicle formation terminated faster (RR = 1.54, 95% CI = 1.00 to 2.36; P = 0.049) in the group treated with sorivudine than in the acyclovir recipients. Similarly, total crusting occurred faster (RR = 1.48, 95% CI = 1.07 to 2.04; P = 0.017) in the sorivudine-treated population than in the group receiving acyclovir. The finding that the beneficial effects of sorivudine treatment on lesion healing were most prominent in the patients with lower CD4 counts was unexpected and not easily explained.
Potential complications of herpes zoster in immunocompromised patients include viral dissemination and the development of prolonged neuralgic pain. The incidence of VZV dissemination (cutaneous or visceral) in HIV-infected patients with herpes zoster is uncertain, since no natural history data for untreated populations are available. Clinical experience suggests that VZV dissemination occurs less frequently in HIV-infected patients than in patients with lymphoproliferative malignancies or organ transplantation (6). In this study, one patient (who was receiving sorivudine) developed possible cutaneous VZV dissemination, but this patient did not require additional therapy. Another patient (who was receiving acyclovir) developed possible myelitis secondary to disseminated VZV, but the patient recovered completely following a course of intravenous acyclovir.
In this study, no differences were seen between the acyclovir and sorivudine treatment groups with regard to the duration of acute or chronic pain. Although 100% of patients reported pain at some point during the treatment phase, 50% of the study population described no discomfort at the 28-day follow-up visit. When contacted at 6 and 12 months, persistent zoster-associated pain was reported by 7 and 4% of the study population, respectively. Many previous studies of herpes zoster have documented that older age is an important risk factor for protracted pain. The relatively young age of the study population (median age, 36 years) was likely an important factor in the relatively small number of patients with severe, prolonged zoster-associated pain.
A prominent feature of the natural history of herpes zoster in HIV-infected patients is the propensity for shingles to recur, which is an unusual event even among other immunocompromised populations. At the time of enrollment, 27% of the subjects had experienced one previous case of herpes zoster and 6% reported two or more previous episodes. Not all of these prior episodes were physician documented, and some may have resulted from etiologies other than VZV infection. During the 12-month follow-up period, 14% of our study subjects experienced a recurrence of herpes zoster. Five patients developed recurrent zoster during the 28-day acute phase. Some of these acute-phase recurrences involved different dermatomes, while others occurred in the same dermatome as the index case episode and may represent relapse rather than true recurrence. In this study, there were no differences in herpes zoster recurrence rates between treatment groups. This differs from the results of a previously reported clinical trial which found a lower frequency of herpes zoster recurrences in sorivudine-treated patients compared with the frequency of recurrences in the acyclovir-treated group (2).
In view of the dramatically superior in vitro activity of sorivudine against VZV, why were bigger differences in clinical outcome not apparent between the two treatment groups? The dose of sorivudine selected for use in this clinical trial was based on careful pharmacokinetic studies which demonstrated that a 40-mg daily dose produces trough levels in serum far in excess of the 50% effective concentration for VZV (12). Even though sorivudine is more than 95% protein bound in serum, even the unbound portion should be sufficient for effective inhibition of VZV replication (11). An alternative explanation is that we are achieving maximal clinical benefit with drugs such as acyclovir and sorivudine. By the time that the patient presents with clinically apparent herpes zoster, VZV replication has already occurred in dorsal root ganglion cells and the virus has traveled along the sensory nerve to the skin, producing an intense inflammatory response along the nerve tract. At that point interventions with a potent antiviral drug can only be expected to limit further viral replication and will not affect the existing pathological changes. The limited superiority of the activity of sorivudine over that of acyclovir demonstrated in this study suggests that attempts to develop new antiviral agents with even more potent in vitro activity may not result in additional clinical benefit.
One of the products of sorivudine metabolism is bromovinyluracil (BVU). In vitro studies conducted long before clinical trials with sorivudine were initiated had identified BVU as a potent inhibitor of dihydropyrimidine dehydrogenase (DPD). Among other enzymatic activities, DPD is involved in the metabolism of 5-FU, a cancer chemotherapeutic agent (5). This creates the potential for a significant drug interaction in patients concomitantly receiving sorivudine and 5-FU. With DPD inhibited by BVU, the activity of 5-FU is sustained, which may result in severe bone marrow suppression. In Japan, where sorivudine was originally synthesized, sorivudine was released for use in the treatment of herpes zoster (at a dosage of 50 mg three times daily) in September 1993. Within a few weeks, multiple instances of cancer patients on 5-FU therapy receiving sorivudine for herpes zoster were reported, despite warnings to physicians (18). Tragically, this inappropriate prescription of medications resulted in 18 deaths due to severe bone marrow toxicity (15). Sorivudine was withdrawn from the Japanese market in October 1993.
The interaction between sorivudine and 5-FU was well recognized when this clinical trial was planned, and 5-FU therapy was a strict exclusion criterion for study participation. Two interim analyses performed by the independent Data Safety and Monitoring Board during the course of this study found no evidence of excess toxicity or unexplained deaths, and the study was permitted to proceed to full enrollment. Following completion of this clinical trial, however, Bristol-Myers Squibb discontinued clinical development of sorivudine due to concerns expressed by regulatory authorities about the possibility of serious drug interactions in patients erroneously prescribed both 5-FU and sorivudine. Therefore, despite sorivudine’s documented efficacy and safety in clinical trials conducted in the United States and other countries, the drug is unlikely to receive licensure in the United States.
ACKNOWLEDGMENTS
This work was supported by funds from the National Institute for Allergy and Infectious Diseases (NIAID) Collaborative Antiviral Study Group (contracts N01-AI-15113 and N01-AI-65306), the Adult AIDS Clinical Trials Group (contract 1-U0L AI-38858), the National Cancer Institute (contract CA RO-1-13148), and the Division of Research Resources (contract RR-032). Additional funding and support plus sorivudine and placebo study medications were provided by Bristol-Myers Squibb. Acyclovir and placebo study medications were provided by Burroughs-Wellcome.
Appendix
Participating investigators. Alabama: L. Austin, G. Cloud, J. Gnann (CASG study chair), L. Sherrill, S.-J. Soong, and R. Whitley (CASG principal investigator) CASG Central Unit, University of Alabama at Birmingham, Birmingham. California: P. Joseph, Infectious Diseases Medical Group, Oakland; J. Lalezari and L. Drew, Mount Zion Medical Center, San Francisco, and M. Wallace, Naval Medical Center, San Diego. Colorado: K. Baum, University of Colorado Health Sciences Center, Denver. Connecticut: M. Beltangady, N. Brennan-Rowe, D. DeHertogh, G. Denisky, J. Harkins, C. Johnson, L. Pacelli, and S. Oshona, Bristol-Meyers Squibb. District of Columbia: D. Parenti, George Washington University, Washington, and C. Gilbert, Veterans Administration Medical Center, Washington. Georgia: C. Newman, Medical College of Georgia, Augusta. Hawaii: M. Health-Chiozzi, University of Hawaii, Honolulu. Illinois: R. Novak, University of Illinois at Chicago, and M. Till, Northwestern Memorial Hospital, Chicago. Massachusetts: M. Sands, Baystate Medical Center, Springfield; C. Crumpacker (ACTG protocol 169 study chair), Beth Israel Hospital, Boston; P. Kazanjian, Brigham and Women’s Hospital, Boston; M. Hirsch, Massachusetts General Hospital, Boston; S. Hammer and A. Karchmer, New England Deaconess Hospital, Boston; and P. Fairchild, University of Massachusetts Medical Center, Worcester. New Mexico: G. Mertz, University of New Mexico, Albuquerque. New York: D. Mildvan, Beth Israel Medical Center, New York, H. Sacks, Mount Sinai Medical Center, New York; M. Grieco, Saint Lukes/Roosevelt Hospital Center, New York; R. Steigbigel, State University of New York, Stony Brook; and R. Reichman, University of Rochester Medical Center, Rochester. North Carolina: G. Davis, Burroughs-Wellcome, Research Triangle Park. Ohio: R. MacArthur, Medical College of Ohio, Toledo, R. Fass, Ohio State University, Columbus, and J. Bernstein, Veterans Administration Medical Center, Dayton. Texas: J. Smith, University of Texas Health Sciences Center, San Antonio; M. Borucki and R. Pollard, University of Texas Medical Branch-Galveston, Galveston; and S. Tyring, University of Texas Medical Branch-Galveston, Clear Lake. Virginia: C. Schleupner, Veterans Administration Medical Center, Salem. Washington: L. Corey and T. Schacker, University of Washington, Seattle.
NIAID program staff. C. Laughlin (CASG project officer), T. Gaither, and B. Alston (Division of AIDS), National Institutes of Health, Bethesda, Md.
REFERENCES
- 1.Balfour H H, Bean B, Laskin O L, Ambinder R F, Meyers J D, Wade J C, Zaia J A, Aeppli D, Kirk L E, Segreti A C, Keeney R E Burroughs Wellcome Collaborative Antiviral Study Group. Acyclovir halts progression of herpes zoster in immunocompromised patients. N Engl J Med. 1983;308:1448–1453. doi: 10.1056/NEJM198306163082404. [DOI] [PubMed] [Google Scholar]
- 2.Bodsworth N J, Boag F, Burdge D, Généreux M, Borleffs J C C, Evans B A, Modai J, Colebunders R, Thomis J the Multinational Sorivudine Study Group. Evaluation of sorivudine (BV-araU) versus acyclovir in the treatment of acute localized herpes zoster in human immunodeficiency virus-infected adults. J Infect Dis. 1997;176:103–111. doi: 10.1086/514011. [DOI] [PubMed] [Google Scholar]
- 3.Buchbinder S P, Kate M H, Hessol N A, Lin J Y, O’Malley P M, Underwood R, Holberg S D. Herpes zoster and human immunodeficiency virus infection. J Infect Dis. 1992;166:1153–1156. doi: 10.1093/infdis/166.5.1153. [DOI] [PubMed] [Google Scholar]
- 4.Descamps J, Sehgal R K, DeClereq E, Allauden S. Inhibitory effect of E-5-(2-bromovinyl)-1-β-d-arabinofuranosyluracil on herpes simplex virus replication and DNA synthesis. J Virol. 1982;43:332–336. doi: 10.1128/jvi.43.1.332-336.1982. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Desgranges C, Razaka G, DeClerq E, Herdewijn P, Balzarini J, Drouillet F, Bricaud H. Effect of E-5-(2-bromovinyl)uracil on the catabolism and antitumor activity of 5-fluorouracil in rats and leukemic mice. Cancer Res. 1986;46:1094–1101. [PubMed] [Google Scholar]
- 6.Glesby M J, Moore R D, Chaisson R E. Clinical spectrum of herpes zoster in adults infected with human immunodeficiency virus. Clin Infect Dis. 1995;21:370–375. doi: 10.1093/clinids/21.2.370. [DOI] [PubMed] [Google Scholar]
- 7.Machida H. Comparison of susceptibilities of varicella-zoster virus and herpes simplex viruses to nucleoside analogs. Antimicrob Agents Chemother. 1986;29:524–526. doi: 10.1128/aac.29.3.524. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Niimura M. A double-blind clinical study in patients with herpes zoster to establish YN-72(Brovavir) dose. In: Lopez C, Mori R, Roizman B, Whitley R J, editors. Immunobiology and prophylaxis of human herpesvirus infections. New York, N.Y: Plenum Press; 1990. pp. 267–275. [DOI] [PubMed] [Google Scholar]
- 9.O’Brien P C, Fleming T R. A multiple testing procedure for clinical trials. Biometrics. 1979;35:549–556. [PubMed] [Google Scholar]
- 10.Rogues A-M, Dupon M, Ladner J, Ragnaud J-M, Pellegrin J-L, Dabis F Groupe D’Epidemiologie Clinque du SIDA en Aquitaine. Herpes zoster and human immunodeficiency virus infection: a cohort study of 101 co-infected patients. J Infect Dis. 1993;168:245. doi: 10.1093/infdis/168.1.245. [DOI] [PubMed] [Google Scholar]
- 11.Sherman J, Devault A, Natarajan C, Harkins J, Hedden B, Stouffer B, Whigan D, Kassalow L, Grasela D, Surgerman A, Reilly K. Program and abstracts of the 30th Interscience Conference on Antimicrobial Agents and Chemotherapy. Washington, D.C: American Society for Microbiology; 1990. SQ 32,756 (BV-araU): characteristics and pharmacokinetics in healthy young and elderly male volunteers, abstr. 1103; p. 270. [Google Scholar]
- 12.Sherman, J., L. Kassalow, B. J. Harkins, B. J. Suites, B. Stougger, D. Whigan, A. A. Shugerman, and S. A. Smith. 1990. SQ32,756 (BV-araU): characteristics and pharmacokinetic evaluation in healthy male volunteers. Antivir. Res. 13(Suppl. 1):100.
- 13.Shigeta S, Yokota T, Iwabuchi T, Baba M, Konno K, Ogata M, DeClercq E. Comparative efficacy of antiherpes drugs against various strains of varicella-zoster virus. J Infect Dis. 1983;147:576–584. doi: 10.1093/infdis/147.3.576. [DOI] [PubMed] [Google Scholar]
- 14.Sullivan M, Skiest D, Signs D, Young C the Famciclovir Herpes Zoster in HIV Study Group. Fourth Conference on Retroviruses and Opportunistic Infections. 1997. Famciclovir in the management of acute herpes zoster (HZ) in the HIV-positive patients, abstr. 704. [Google Scholar]
- 15.Swinbanks D. Deaths bring clinical trials under scrutiny in Japan. Nature. 1994;369:697. doi: 10.1038/369697b0. [DOI] [PubMed] [Google Scholar]
- 16.Wallace M R, Chamberlin C J, Sawyer M H, Arvin A M, Harkins J, LaRocco A, Colopy M W, Bowler W A, Oldfield E C. Treatment of adult varicella with sorivudine: a randomized placebo-controlled trial. J Infect Dis. 1996;174:249–255. doi: 10.1093/infdis/174.2.249. [DOI] [PubMed] [Google Scholar]
- 17.Whitley R J, Gnann J W, Hinthorn D, Liu C, Pollard R B, Hayden F, Mertz G J, Oxman M, Soong S J the NIAID Collaborative Antiviral Study Group. Disseminated herpes zoster in the immunocompromised host: a comparative trial of acyclovir and vidarabine. J Infect Dis. 1992;165:450–455. doi: 10.1093/infdis/165.3.450. [DOI] [PubMed] [Google Scholar]
- 18.Yawata M. Deaths due to drug interaction. Lancet. 1993;342:1166. [Google Scholar]
- 19.Yokata T, Kono K, Mori S, Shigeta S, Kumangai M, Watanabe Y, Machida H. Mechanism of selective inhibition of varicella-zoster virus replication by 1-beta-d-arabinofuranosyl-E-5-(2-bromovinyl)uracil. Mol Pharmacol. 1989;36:312–316. [PubMed] [Google Scholar]