Comparison of New Topical Treatments for Herpes Labialis: Efficacy of Penciclovir Cream, Acyclovir Cream, and n-Docosanol Cream Against Experimental Cutaneous Herpes Simplex Virus Type 1 Infection

Background  There are 3 new topical treatments for herpes labialis that have either been approved by the US Food and Drug Administration (penciclovir cream [Denavir] and n-docosanol cream [Abreva]) or recently undergone extensive clinical evaluation (acyclovir cream). The relative efficacy of these products is unknown.

Objective  To compare the efficacy of penciclovir cream, acyclovir cream, n-docosanol cream, and acyclovir ointment in an experimental animal model of cutaneous herpes simplex virus type 1 (HSV-1) disease.

Design  The backs of guinea pigs were infected with HSV-1 using a vaccination instrument. Active treatments and corresponding vehicle controls were applied for 3 to 5 days beginning 24 hours after inoculation.

Main Outcome Measures  After completion of treatment, the animals were killed and the severity of the infection assessed from the number of lesions, the total lesion area, and the lesion virus titer.

Results  Penciclovir cream effected modest reductions in lesion number (19%), area (38%), and virus titer (88%) compared with its vehicle control, and each of these differences was significantly greater (P<.05) than the reductions effected by acyclovir ointment (0%, 21%, and 75%, respectively). The acyclovir cream effect (reductions of 4%, 28%, and 77%, respectively) was less than that of penciclovir cream, and this difference was confirmed by 2 additional head-to-head experiments. Two experiments with n-docosanol cream failed to show statistically significant differences by any parameter between n-docasonol cream and vehicle control–treated sites or between n-docosanol and untreated infection sites.

Conclusions  In this model, the efficacy of penciclovir cream was greater than acyclovir cream, acyclovir cream was greater than or equal to acyclovir ointment, and acyclovir ointment was greater than n-docosanol cream. Since our model was designed to evaluate compounds that function primarily through antiviral activity, the negative findings with n-docosanol in these studies do not exclude that it might work clinically through other mechanisms.

FOLLOWING a period of uncertainty regarding efficacy,^1,2 antiviral treatments for herpes labialis have been approved by regulatory agencies in the United States and other countries. In the United States, acyclovir ointment (Zovirax Ointment) is approved for use in immunocompromised patients and penciclovir cream (Denavir) for use in otherwise healthy adults.^3,4 In other countries, 1 or more ethical drug treatments, including penciclovir cream, acyclovir cream,⁵ and idoxuridine in dimethyl sulfoxide,⁶ are generally available. There is experimental evidence that high-dose peroral nucleoside analogue therapy may be highly effective, possibly because of delivery of high concentrations of drug to the site of the infection.⁷ A multitude of treatments are available over the counter, with claims based on the potential of the ingredients to reduce lesion pain or promote wound healing through occlusion.

Because of the large number of patients and the expense required to conduct a therapeutic trial in herpes labialis, it is unlikely that the relative clinical efficacy of these various treatments will ever be determined. The dorsal cutaneous guinea pig model is a well-standardized animal model of cutaneous herpes disease and permits an experimental approach to this question. Drug efficacy in the model correlates with antiviral activity and skin penetration.^8,9 Guinea pig skin is generally more permeable to nucleoside drugs than human skin, but the relative order of permeation of different compounds appears to be the same.^10,11

The present report describes 2 sets of experiments in the guinea pig model. In the first experiment, we compared penciclovir cream, acyclovir cream, and acyclovir ointment with vehicle controls. In the second group of experiments, we examined the efficacy of a new over-the-counter agent, n-docosanol cream (Abreva). The results are correlated with the efficacy of other treatments in the model and the outcome of clinical trials.

Both 10% and 12% n-docosanol cream and 10% and 12% stearic acid–supplemented vehicle controls were provided by Lidak Pharmaceuticals (La Jolla, Calif). The 10% n-docosanol cream is now marketed over the counter as Abreva by GlaxoSmithKline (Research Triangle Park, NC). The 5% acyclovir cream (Zovirax Cream [Europe]), 5% acyclovir ointment (Zovirax Ointment), and their corresponding vehicles were provided by GlaxoWellcome (Research Triangle Park, NC). The 1% penciclovir cream (Denavir [United States], Vectavir [Europe]) and its vehicle were provided by SmithKline Beecham (King of Prussia, Pa).

The virus was the laboratory strain HSV-1 E115, originally obtained from Andre Nahmias, MD, MPH (Emory, Ga), and used in our model since 1980.^11,12 Virus stock for inoculation of guinea pigs contained 1 × 10⁸ plaque-forming unit (pfu)/mL and was prepared in mink lung cells (MV-1-Lu, American Type Culture Collection).

Hartley strain, outbred, female, albino guinea pigs, weighing approximately 500 g each, were obtained from Charles River Breeding Laboratories Inc, Wilmington, Mass. During treatment, animals were housed in individual cages. Before inoculation and again immediately before experimental virus lesion assessment, the hair was removed from the animal dorsum with a chemical depilatory.

Test compounds were evaluated for irritation on the depilated, uninfected dorsum of guinea pigs. Test compounds and their vehicles were each applied at 3 different sites once per day, 2 times per day, and 4 times per day for 3 days and evaluated for irritation on the morning of the fourth day using an irritation score of 0 to 4. A score of 4 indicated severe erythema with punctate bleeding. A mean score of 2.5 or greater indicated that the compound and/or its vehicle at that dosing frequency caused sufficient irritation to compromise assessment of the viral infection and potentially affect the outcome of the experiment.

Guinea pigs were inoculated with HSV-1 E115 (day 0) in 4 different areas on the depilated dorsum, right and left midback, and right and left rump by multiple shallow punctures with a vaccination instrument as originally performed by Hubler et al.¹³ Between 40 and 60 discrete lesions developed at each of the 4 infection sites. Test treatments were begun 24 hours after inoculation (day 1) and given 1 to 4 times per day at 8 AM, noon, 4 PM, and 8 PM, depending on dermal irritation, for a total of 3 days. The day after completion of treatment (day 4), the dorsum of each animal was again depilated, and the severity of infection at each site was assessed from the number of lesions, the total lesion area, and titer of virus in the excised infection site. The procedures have been described in detail elsewhere.¹²

Paired data (drug and drug vehicle) were evaluated by the Wilcoxon signed rank test. To compare 2 drugs with one another, efficacies were expressed as the percent difference in a lesion severity measure (lesion number, area, or virus titer) between the active formulation and its control and tested by a Mann-Whitney rank sum procedure. All probability determinations were 2-tailed, and P≤.05 was considered significant.

Penciclovir and acyclovir creams were examined for efficacy in relationship to their corresponding vehicles, which were applied at contralateral infection sites. Acyclovir ointment and the ointment vehicle were studied concurrently as a control. The different drug formulations were compared with one another using the percentage of improvement in lesion severity measures from the drug vs drug vehicle assessments. Each treatment was evaluated at 12 different infection sites on the guinea pig dorsum as described in the "Materials and Methods" section. Treatments were only applied once a day because both cream formulations were irritating to guinea pig skin. The results are shown in Table 1.

Penciclovir cream had the greatest effect. Reductions in lesion number (19%), area (38%), and virus titer (88%) were demonstrated in comparison to the vehicle, and each of these differences was significantly greater (P<.05) than the reductions effected by acyclovir ointment (0%, 21%, and 75%, respectively). The acyclovir cream effect on lesion parameters (reductions of 4%, 28%, and 77%, respectively) was less than that of penciclovir cream and similar in magnitude and not significantly different from those of acyclovir ointment. The differences between acyclovir and penciclovir cream reached statistical significance for virus titer (P<.05).

To further examine the potential differences between penciclovir and acyclovir cream, the 2 treatments were compared directly with one another in 2 additional experiments. Treatments were only applied once a day because both formulations were irritating to guinea pig skin. The results are shown in Figure 1. In both experiments, penciclovir cream–treated sites had fewer mean numbers of lesions than at acyclovir cream–treated sites (46-49 vs 52-56, respectively) and a smaller mean lesion area (197-219 vs 258-280 mm², respectively). Three of these 4 differences were statistically significant (P<.05). There was no difference between the 2 treatments in the mean lesion virus titer.

Two separate experiments were done to evaluate the potential efficacy of n-docosanol cream. Because the formulation lost its consistency when n-docosanol was removed, the sponsor provided a vehicle control in which stearic acid was added, which gave the preparation a similar turgor and appearance to the active formulation. The vehicle with stearic acid proved to be highly irritating and could only be applied once per day, whereas n-docosanol cream was tolerated at a maximum dosing frequency of 3 times per day. After consultation with the sponsor, the first experiment was performed in which 10% n-docosanol cream was applied 3 times per day for 3 days and compared with untreated infection sites. Acyclovir ointment and ointment control were applied 4 times per day for 3 days, our standard comparator regimen in this model. The results are shown in Table 2. Acyclovir ointment effected statistically significant reductions in all lesion parameters compared with the ointment vehicle control, whereas n-docosanol had no effect. The amount of reduction in the mean lesion area by acyclovir ointment was 22%.

To further explore the potential activity of n-docosanol, a second experiment was performed in which the concentration of n-docosanol was increased to 12%, the duration of treatment was increased from 3 to 5 days, and efficacy parameters were assessed on 3 days (days 4, 5, and 6) instead of just on day 4. The 12% stearic acid–supplemented cream was used as the vehicle control. Acyclovir ointment and ointment control were again included as comparators. All treatments were applied 3 times per day for 5 days. The results are shown in Figure 2. Acyclovir ointment was effective on all 3 days, whereas n-docosanol showed no benefit. The amount of reduction in the mean lesion area by acyclovir ointment on days 4, 5, and 6 was 23%, 31%, and 28%, respectively.

The relative efficacy of the treatments tested in this study was penciclovir cream greater than acyclovir cream greater than or equal to acyclovir ointment greater than n-docosanol cream. Total lesion area, the major clinical parameter in the guinea pig model,⁹ was reduced 38%, 28%, 21%, and 0%, respectively, by these treatments. Prior studies¹⁴ with 15% idoxuridine in dimethyl sulfoxide in this model have shown a reduction in lesion area of 80%. Each of these products has also been evaluated in clinical trials as therapy for herpes labialis with lesion healing time as the primary efficacy variable.^1,4,6,15-17 The efficacy values for these 5 treatments in the guinea pig model and in clinical trials against herpes labialis in normal hosts are compared in Table 3.

As shown in Table 3, the rank order of results with different compounds in our model generally parallels the rank order of results in clinical trials. Infection in the model is roughly twice as sensitive to the antiviral treatments as is herpes labialis and so exaggerates the degree of efficacy compared with the outcome in the human condition. This is, in part, because dorsal cutaneous HSV guinea pig disease is a primary infection, whereas herpes labialis is a recrudescence. In the latter instance, the secondary immune response truncates the course of the illness and abbreviates the window of opportunity for chemotherapy. Awan et al²¹ have convincingly demonstrated the negative impact of the immune response on drug efficacy by studies of topical anti-HSV therapy in mice with and without the adoptive transfer of HSV-sensitized mononuclear cells.

It is possible that the dosing regimen of acyclovir cream was not optimal in our animal experiments. Specifically, more frequent dosing of acyclovir cream might have been associated with greater efficacy. The potential impact of dosing frequency on comparative studies of penciclovir and acyclovir in animal models has been previously reported²² and may be related to differences in the stability of the intracellular triphosphate metabolites of the 2 drugs. In the present experiments, skin irritation limited treatment applications to once a day.

Evidence from the experimental model and in vitro studies of drug penetration through animal and human skin samples^11,23 showed that better drug delivery significantly improved efficacy and that we were in the early portion of a dose-response curve with our current therapies for cutaneous HSV infection. To some extent, this is apparent in the clinical trials of topical formulations (Table 3). To further substantiate this principle in herpes labialis, we compared 3 different doses of peroral famciclovir in the treatment of UV radiation–induced recurrences. There was a significant dose-response effect, indicating that higher peroral doses of antiviral agents or topical formulations providing improved intraepidermal drug penetration are strategies that should be explored in the treatment of herpes labialis.⁷

There has been considerable controversy recently concerning the potential effect of the vehicle comparator arm on the apparent efficacy of topical antiherpetic compounds in clinical trials.^18,24 Treatment advocates have claimed that the vehicles have efficacy, therefore reducing the overall benefit of the formulation compared with an untreated control arm. Potential mechanisms of vehicle efficacy include antiviral activity, acceleration of wound healing by occlusion, and the eschar-dissolving potential of vehicle solvents such as propylene glycol. It is also possible that drug vehicles could have an adverse effect on lesion severity, for example, by skin irritation, such that the efficacy of a treatment would be exaggerated by drug-drug vehicle comparisons. These issues are best examined experimentally, because the placebo effect in human studies makes it difficult to interpret a trial in which one group of patients has not received treatment.

The issue of vehicle effects is of importance in the case of n-docosanol cream. The n-docosanol cream formulation loses its consistency when n-docosanol is omitted. To conduct clinical trials, it was necessary to have a vehicle that matched the active formulation in appearance, and this was accomplished by adding stearic acid. A small pilot trial (n = 63) that used a stearic acid–containing vehicle control showed efficacy,²⁵ but a subsequent much larger study (n = 846) found no benefit.¹⁷ Because of a concern that the vehicle may have had a beneficial effect on herpes labialis and masked an effect by n-docosanol, a polyethylene glycol control formulation was prepared. The clinical trials were repeated using polyethylene glycol as the control, and a 15% reduction in lesion healing time in the n-docosanol arm recently was reported.¹⁸ The differing trial results and the complicated issue of vehicle effects as a potential explanation for the differences make it hard to evaluate this product.

In the present studies, n-docosanol cream was examined in several ways. When compared against untreated lesion sites, no clinical or antiviral activity was seen (Table 2). When compared against a stearic acid–containing vehicle on multiple days, again no clinical or antiviral activity was found. In the latter experiment, acyclovir ointment and ointment control (polyethylene glycol) were run as comparators. There were no significant differences in any measure between the stearic acid–containing n-docosanol vehicle and polyethylene glycol ointment, indicating that stearic acid was inactive in this model. In summary, despite reports of n-docosanol antiviral activity in vitro,²⁶ we found that both n-docosanol and stearic acid were inactive as antiviral agents against experimental cutaneous HSV lesions in the dorsal cutaneous guinea pig model. Since our model was designed to evaluate compounds that function primarily through antiviral activity, our studies do not exclude the possibility that n-docosanol cream might have beneficial effects on herpes labialis through other mechanisms, such as stimulation of wound healing or anti-inflammatory properties.²⁷

Accepted for publication February 20, 2001.

Studies of n-docosanol cream were funded by a grant from Avanir Pharmaceuticals (formerly Lidak Pharmaceuticals), San Diego, Calif. The remaining support was provided by the University of Utah.

Presented in part at the Eighth International Conference on Antiviral Research, Sante Fe, NM, April 25, 1995, and the 11th International Conference on Antiviral Research, San Diego, Calif, April 7, 1998.

Corresponding author and reprints: Spotwood L. Spruance, MD, Division of Infectious Diseases, School of Medicine, Room 4B319, University of Utah, 50 North Medical Dr, Salt Lake City, UT 84132.