A database search was independently conducted by two reviewers (MR and NT) in MEDLINE (accessed via PubMed), SCOPUS, EMBASE, SCIENCE DIRECT, LILACS and the Cochrane Central Register of Controlled Trials for the period of January 1990 until October 2016. 15,16 The search strategy was specific for each database and included a combination of the Medical Subject Heading and free text terms for ¨Acyclovir¨, “Pityriasis rosea” and “Clinical trial” (Chart 2). For unpublished studies we reviewed grey literature and experts’ opinion on the topic. In addition, we scanned bibliographies of published studies by a manual search of the literature; reviewed virtual libraries of universities and theses; requested authors for papers presented at conferences and congresses of dermatology societies; and contacted the pharmaceutical industry. We also searched at ClinicalTrials.gov database to find additional studies that were not included in the initial search. There are no restrictions of language to select the studies for this review.

We included parallel-clinical trials (experimental and quasi-experimental), with open-labeled or blinded designs; studying the efficacy of acyclovir for PR. Participants included were immunocompetent and nonpregnant females and males, children and adults with PR diagnosed according to the investigator criteria and specific clinical features.

Studies should have two or more study arms, at least one arm on acyclovir (with or without symptomatic treatment), and a control group (on placebo, symptomatic treatment, antibiotics or nothing). Symptomatic treatment consists on oral or topical antihistamines, oral or topical steroids, and topical balsams.9,17

We excluded studies lacking complete data in any arm of study, after exhausting efforts to contact the authors for full data. Overlapping studies were included after discarding the one with the smaller study population.

The comparisons were (1) Acyclovir vs. placebo; (2) Acyclovir vs. symptomatic treatment; (3) Acyclovir vs. antibiotic; (4) Acyclovir vs. observation; (5) Acyclovir plus symptomatic treatment (combined therapy) vs. symptomatic treatment alone.

The primary outcome was the proportion of patients achieving regression of lesions after the treatment of oral acyclovir. Regression of lesions is defined as decrease or disappearance of erythema in all lesions leaving desquamation or pigmentation. Secondary outcomes were (i) The proportion of patients that achieved cessation of new lesions. (ii) The proportion of patients that had decreased systemic symptoms (pruritus). (iii) Time for complete clearance of lesions. Outcomes were measured at several weeks after treatment was initiated.

Two reviewers (MR and NT) independently screened the titles and abstracts of all the retrieved articles, followed by the full texts of the articles considered potentially eligible for the study. They independently collected the data using a standardized data sheet. Both authors confirmed all entries and checked at least twice for completeness and accuracy. Disagreements were resolved by discussion with a third reviewer (HG). Authors were contacted by email for missing data and clarifications, in order to achieve completeness of studies.

The pooled relative risk (RR) was the effect measured of the primary outcome, with 95% confidence intervals (95%CI). As the heterogeneity was considerable, the pooled RR was calculated based on a random effect model, according to the ‘standard approach’.18 Statistical heterogeneity was assessed with the I2 statistic.19

For secondary outcomes, dichotomous data was calculated using RR and continuous data using the standardized mean difference, with 95%CI and a random effect model.

Included trials were characterized with descriptive statistics (percentages and mean). We performed analyses with Stata13® and RevMan 5.1®.

Every study was independently evaluated for ‘risk of bias’ by two reviewers (MR and NT), considering the domains and criteria suggested by the Cochrane Handbook v.5.1.0.12 Disagreements were discussed and resolved with a third reviewer (HG).

Subgroup analysis, publication bias and sensitivity analysis were not possible due to the limited number of trials in the meta-analysis.

324 participants entered to the meta-analysis, 159 were allocated to acyclovir, and 165 to control. Studies failed to report reasons for dropouts.

Comparison: Acyclovir vs. Placebo (two trials)

Acyclovir was superior to placebo after one week of treatment (RR 5.72, 95%CI 2.36-13.88, I2=0%); however, there was no difference at two weeks (RR 6.08, 95%CI 0.94-39.36, I2=85%) (Figure 2).10,23

Figure 2.

Forest plots of meta-analysis for the main outcome (regression of lesions) after one A and two weeks of treatment B. Trials included compared acyclovir vs. placebo

(0.32MB).

Acyclovir was similar to erythromycin after two weeks, but superior after four (RR 8.0 95%CI 1.13-56.33) and eight (RR 2.16 95%CI 1.13-4.15) weeks of treatment.22

Comparison: Acyclovir vs. observation (one trial)

Acyclovir was superior to observation after one (RR 3.02 95%CI 1.13-8.08), two (RR 2.92 95%CI 1.50-5.66) and four (RR 1.51 95%CI 1.10-2.08) weeks of treatment.21

Comparison: Combined therapy vs. symptomatic treatment (two trials)

Combined treatment was not significantly superior after two (one study (23): RR 3.11 95%CI 0.93-10.39), four (two studies (19, 23): RR 1.46 95%CI 0.93-2.29, I2=69%) and eight (one study (23): RR 0.98 95%CI 0.86-1.12) weeks of treatment (Figure 3).24

Figure 3.

Forest plots of meta-analysis for the main outcome (regression of lesions) after four weeks of treatment. Trials included compared combined therapy vs. symptomatic treatment alone. Combined RR was obtained only for comparison at week 4

(0.21MB).

Outcome: Cessation of new lesions

Comparison: Acyclovir vs. Placebo (two trials)

Acyclovir was not significantly superior after one (RR 4.85 95%CI 0.16-145.64, I2=83%) and two (RR 1.42 95%CI 0.73-2.75, I2=92%) weeks of treatment (Figure 4).10,23

Figure 4.

Forest plot of the meta-analysis for secondary outcome cessation of lesions after one A and two B weeks of treatment. Trials included compared acyclovir vs. placebo

(0.38MB).

Comparison: Combined therapy vs. symptomatic treatment (one trial)

Combined therapy was not significantly superior after one week of treatment (RR 1.67, 95%CI 0.85-3.426).20

Comparison: Acyclovir vs. Placebo (one trial)

Acyclovir was not superior to decrease symptoms (e.g. fatigue, headache, sore throat, difficulty in concentrating, irritability, insomnia, and nausea) after one week (RR 12.75, 95%CI 0.78-207.32).10 Nevertheless, it was superior at two (RR 11.79, 95%CI 1.73-80.15) weeks of treatment (Figure 5).

Figure 5.

Comparative diagram for decrease of symptoms analysis after one A and two B weeks of treatment. Trials included compared acyclovir vs. placebo

(0.16MB).

Comparison: Acyclovir vs. antibiotic (one trial)

Acyclovir was not significantly superior to erythromycin to decrease pruritus after two (RR 13.22, 95%CI 0.91-192.02), four (RR 1.47, 95%CI 0.83-2.61) and eight (RR 1.20, 95%CI 0.79-1.83) weeks of treatment.22

Comparison: Combined therapy vs. symptomatic treatment (one trial)

Combined therapy decreased significantly pruritus after one (SMD -1.31 95%CI -2.19, -0.41), two (SMD -1.61 95%CI -2.53, -0.68), three (SMD -1.25 95%CI-2.13, -0.27) and four (SMD -1.21 95%CI -2.10, -0.34) weeks of treatment.20

Comparison: Acyclovir vs. Placebo (two trials)

Acyclovir was not superior (SMD -0.04 95%CI -1.47, 1.38) (Figure 6).10,25

Figure 6.

Forest plot of the meta-analysis for secondary outcome duration of disease. Trials included compared acyclovir vs. placebo

(0.21MB).

The rest of analyses are not presented, as there was not data available.

One study10 demonstrated that acyclovir was superior to placebo for decreasing symptoms at week 1 and 2 (Figure 5); another study showed the significant decrease of itching from week 1 to 4 with combined treatment.20 It is important to underline that other reports had elucidated the anti-pruritic effect of acyclovir in PR.9 Other treatments, erythromycin and phototherapy, have been also proposed for pruritic management in PR with controversial results.31 Most of the guidelines suggest that, as PR is a self-limited illness, most patients do not need to be treated. Other authors propose that patients with recalcitrant symptoms, such as pruritus, are the exception: they may benefit from treatment with macrolides, phototherapy and antivirals.9

When erythromycin is compared to placebo for PR, 3 out of 4 patients on interventional group cleared completely within 2 weeks.32 Nevertheless, this study had methodological deficiencies in randomization, allocation concealment and analysis, with the adverse events inherent to macrolides.33 Other studies suggested immunomodulatory and anti-inflammatory effects of macrolides in PR due to Chlamydia, Legionella and Mycoplasma involvement, but not assertively.34-36 Clinical trials have shown that efficacy of macrolides, such as erythromycin, clarithromycin and azithromycin has not been proven.36-38 One study compared acyclovir vs. erythromycin and showed that the antiviral was significantly superior for regression of lesions, but not significant for decrease of symptoms.22 Therefore, we recommend the use of antiviral over macrolides supported by the possible viral etiology of PR.39

Phototherapy with narrowband UVB and UVA1 has been proposed as methods to control PR severity and pruritus.40,41 Other studies have proven their efficacy in decreasing severity but failed to change itch and course of the disease.42,43 There is inadequate evidence for phototherapy, and sessions are given thrice a week for at least one month.27 Although it might be efficient, it may not be feasible and accessible for all patients.31

Based on the results, we think that symptomatic treatment (oral antihistamines and topical steroids/calamine) may be considered as baseline treatment for PR, and the early addition of acyclovir for pruritus control, especially in extensive, relapsing or persistent disease.11,17,44

Acyclovir was not superior to placebo, neither combined to symptomatic treatment in cessation of new lesions at week 1 and 2 (Figure 4). In addition, acyclovir failed to decrease significantly the duration of PR compared to placebo. However, studies that entered into the meta-analysis had opposed results independently (Figure 6).10,25

Clinical trials included were not enough to distinguish efficacy of children from adult population. Drago et al. have found different characteristics between them.45 Children have less prevalence of systemic symptoms and less average duration of exanthema. Therefore, we believed that acyclovir in children may not be a reasonable alternative for PR, until trials demonstrate efficacy in this population.

Adverse events were reported in both acyclovir and control groups (Table 2). In an observer-blinded trial, 8 patients on acyclovir and antihistamines reported increased sleep, headache, nausea, vomiting and dysgeusia.20 These adverse events, although they are expected in patients on acyclovir, were not reported in the rest of trials.46 A triple-blinded trial reported abdominal pain and diarrhea in placebo group, possibly due to nocebo effect.25,47

Unclear risk of bias was prevalent for allocation concealment (selection bias) and incomplete outcome data (attrition bias) (Figure 7A). The rest of domains had in general low risk of bias. Drago et al,10 did not perform randomization neither allocation concealment. Two trials failed to blind participants from treatment received (Figure 7B).20,24

Figure 7.

Risk of bias analysis. A. Between studies; B. Within studies

(0.61MB).

We faced several limitations. First, studies did not evaluate all outcomes and they differed in temporality. Therefore, meta-analysis was not possible for all outcomes and weeks of the follow-up. Second, some studies had high risk of bias in randomization and allocation concealment. However, we decided to include all trials since they represent interventional studies with similar established outcomes. Finally, there were few trials for each comparison, thus, subgroup analysis was not possible.