Corticosteroids for preventing postherpetic neuralgia

  • Review
  • Intervention

Authors


Abstract

Background

Postherpetic neuralgia is a common, serious painful complication of herpes zoster. Corticosteroids are anti-inflammatory and might be beneficial. This is an update of a review first published in 2008 and previously updated in 2010.

Objectives

To examine the efficacy of corticosteroids in preventing postherpetic neuralgia.

Search methods

We updated the searches for randomised controlled trials (RCTs) of corticosteroids for preventing postherpetic neuralgia in the Cochrane Neuromuscular Disease Group Specialized Register (16 April 2012), CENTRAL (2012, Issue 3), MEDLINE (January 1966 to April 2012), EMBASE (January 1980 to April 2012), LILACS (January 1982 to April 2012), and the Chinese Biomedical Retrieval System (1978 to 2012). We also reviewed the bibliographies of identified trials, contacted authors and approached pharmaceutical companies to identify additional published or unpublished data.

Selection criteria

We included all RCTs involving corticosteroids given by oral, intramuscular, or intravenous routes for people of all ages with herpes zoster of all degrees of severity within seven days after onset, compared with no treatment or placebo but not with other treatments. We did not include quasi-RCTs (trials in which a systematic method of randomisation such as alternation or hospital number was used).

Data collection and analysis

Two authors identified potential articles, extracted data, and independently assessed the risk of bias of each trial. Disagreement was resolved by discussion among the co-authors.

Main results

Five trials were included with 787 participants in total. All were randomised, double-blind, placebo-controlled parallel-group studies. We conducted a meta-analysis of two trials (114 participants) and the results gave moderate quality evidence that oral corticosteroids did not prevent postherpetic neuralgia six months after the onset of herpes (RR 0.95, 95% CI 0.45 to 1.99). One of these trials was at high risk of bias because of incomplete outcome data, the other was at low risk of bias overall. The three other trials that fulfilled our inclusion criteria were not included in the meta-analysis because the outcomes were reported at less than one month or not in sufficient detail to add to the meta-analysis. These three trials were generally at low risk of bias. Adverse events during or within two weeks after stopping treatment were reported in all five included trials. There were no significant differences in serious or non-serious adverse events between the corticosteroid and placebo groups. There was also no significant difference between the treatment groups and placebo groups in other secondary outcome analyses and subgroup analyses. The review was first published in 2008 and no new RCTs were identified for inclusion in subsequent updates in 2010 and 2012.

Authors' conclusions

There is moderate quality evidence that corticosteroids given acutely during zoster infection are ineffective in preventing postherpetic neuralgia. In people with acute herpes zoster the risks of administration of corticosteroids do not appear to be greater than with placebo, based on moderate quality evidence. Corticosteroids have been recommended to relieve the zoster-associated pain in the acute phase of disease. If further research is designed to evaluate the efficacy of corticosteroids for herpes zoster, long-term follow-up should be included to observe their effect on the transition from acute pain to postherpetic neuralgia. Future trials should include measurements of function and quality of life.

Résumé scientifique

Corticostéroïdes pour la prévention de la névralgie post-herpétique

Contexte

La névralgie post-herpétique est une complication fréquente, grave et douloureuse de l'herpès zoster. Les corticostéroïdes sont des anti-inflammatoires qui pourraient s'avérer bénéfiques. Cette revue est une mise à jour d'une revue publiée pour la première fois en 2008 et précédemment mise à jour en 2010.

Objectifs

Examiner l'efficacité des corticostéroïdes dans la prévention de la névralgie post-herpétique.

Stratégie de recherche documentaire

Nous avons mis à jour les recherches d'essais contrôlés randomisés (ECR) sur les corticostéroïdes dans la prévention de la névralgie post-herpétique dans le registre spécialisé du groupe Cochrane sur les affections neuromusculaires (16 avril 2012), CENTRAL (2012, numéro 3), MEDLINE (de janvier 1966 à avril 2012), EMBASE (de janvier 1980 à avril 2012), LILACS (de janvier 1982 à avril 2012), et le Chinese Biomedical Retrieval System (de 1978 à 2012). Nous avons également passé en revue les bibliographies des essais identifiés, contacté les auteurs et contacté les laboratoires pharmaceutiques pour identifier des données supplémentaires, publiées ou non.

Critères de sélection

Nous avons inclus tous les ECR portant sur des corticostéroïdes administrés par voie orale, intramusculaire ou intraveineuse à des personnes de tout âge atteintes d'herpès zoster, quel qu'en soit le degré de gravité, dans les sept jours après son apparition, par rapport à l'absence de traitement ou à un placebo mais sans aucun autre traitement. Nous avons exclu les essais contrôlés quasi-randomisés (essais dans lesquels une méthode de randomisation systématique telle qu'une alternance ou un numéro d'hospitalisation était utilisé).

Recueil et analyse des données

Deux auteurs ont identifié les articles potentiels, extrait les données, et évalué de façon indépendante le risque de biais de chaque essai. Les désaccords ont été résolus par des discussions entre les co-auteurs.

Résultats principaux

Cinq essais ont été inclus, portant sur un total de 787 participants. Ces cinq essais étaient des études randomisées en double aveugle, contrôlées par placebo, en groupes parallèles. Nous avons réalisé une méta-analyse de deux essais (114 participants) et les résultats ont fourni des preuves de qualité moyenne attestant que les corticostéroïdes oraux ne préviennent pas la névralgie post-herpétique six mois après l'apparition de l'herpès (RR 0,95, IC à 95 % 0,45 à 1,99). L'un de ces essais présentait un risque de biais élevé en raison de données de résultats incomplètes, l'autre présentait un faible risque de biais global. Les trois autres essais qui remplissaient nos critères d'inclusion n'ont pas été inclus dans la méta-analyse puisque les résultats étaient présentés à moins d'un mois ou de manière insuffisamment détaillée pour entrer dans la méta-analyse. Ces trois essais présentaient généralement un faible risque de biais. Les événements indésirables pendant le traitement ou dans les deux semaines suivant l'arrêt du traitement étaient présentés dans les cinq essais inclus. Il n'y avait aucune différence significative en termes d'événements indésirables graves ou sans gravité entre les groupes recevant le corticostéroïde et les groupes recevant un placebo. Il n'y avait pas non plus de différence significative entre les groupes de traitement et les groupes recevant un placebo pour l'analyse des autres résultats secondaires et les analyses en sous-groupes. Cette revue a été publiée pour la première fois en 2008 et aucun nouvel ECR à inclure n'a été identifié dans les mises à jour suivantes en 2010 et en 2012.

Conclusions des auteurs

Il existe des preuves de qualité moyenne attestant que les corticostéroïdes administrés ponctuellement lors d'une infection à herpès zoster sont inefficaces pour prévenir la névralgie post-herpétique. Chez les personnes souffrant d'une infection aiguë à herpès zoster, les risques de l'administration de corticostéroïdes ne semblent pas plus élevés qu'avec un placebo, d'après des données de qualité moyenne. Les corticostéroïdes ont été recommandés pour soulager la douleur associée à l'herpès zoster dans la phase aiguë de la maladie. Si l'on conçoit des recherches supplémentaires pour évaluer l'efficacité des corticostéroïdes dans le traitement de l'herpès zoster, un suivi à long terme doit être inclus pour observer leur effet sur la transition d'une douleur aiguë à une névralgie post-herpétique. Les futurs essais devront inclure des mesures de la fonction et de la qualité de vie.

Plain language summary

Corticosteroids for preventing postherpetic neuralgia

Postherpetic neuralgia is a painful condition that is one of the most common complications of an acute herpes zoster infection. Herpes zoster presents as a localised rash resembling localised chicken pox, often called 'shingles'. Postherpetic neuralgia may persist lifelong once it occurs and has major implications for quality of life and use of healthcare resources. Corticosteroids have a potent anti-inflammatory action, which it has been suggested might minimise nerve damage and thereby relieve or prevent the pain experienced by people suffering from this condition. Five trials were identified from a systematic search of the literature which were of high enough quality to be included in the review. These trials involved 787 participants in total. We were able to combine the results from two trials (114 participants) and there was no significant difference between the corticosteroid and control groups in the presence of postherpetic neuralgia six months after the onset of the acute herpetic rash. Two of the three other included trials reported results at less than one month, so these participants did not fulfil the current criteria for a diagnosis of postherpetic neuralgia. The last trial reported results in a format unsuitable for meta-analysis. There were no significant differences in serious or non-serious adverse events between the corticosteroids and placebo groups. There was also no significant difference between the treatment groups and placebo groups in other secondary outcome analyses and subgroup analyses. It can be concluded that, based on moderate quality evidence, corticosteroids are not effective in preventing postherpetic neuralgia.

Résumé simplifié

Corticostéroïdes pour la prévention de la névralgie post-herpétique

La névralgie post-herpétique est une affection douloureuse qui est l'une des complications les plus fréquentes de l'infection aiguë à herpès zoster. L'herpès zoster se présente sous la forme d'une éruption cutanée localisée qui ressemble à une varicelle localisée et est souvent appelé « zona ». La névralgie post-herpétique peut durer toute la vie une fois qu'elle s'est manifestée et elle a des incidences majeures sur la qualité de vie et l'utilisation des ressources des soins de santé. Les corticostéroïdes ont un effet anti-inflammatoire puissant, qui pourrait minimiser les lésions nerveuses et ainsi soulager ou prévenir les douleurs ressenties par les personnes qui souffrent de cette affection. Cinq essais identifiés suite à une recherche systématique de la littérature scientifique étaient de qualité suffisante pour être inclus dans la revue. Ces essais portaient sur un total de 787 participants. Nous avons été en mesure de combiner les résultats de deux essais (114 participants) et il n'y avait pas de différence significative entre les groupes recevant les corticostéroïdes et les groupes témoins en ce qui concerne la présence de névralgie post-herpétique six mois après l'apparition de l'éruption herpétique aiguë. Deux des trois autres essais inclus présentaient les résultats à moins d'un mois, et par conséquent ces participants ne remplissaient pas les critères actuels nécessaires au diagnostic de névralgie post-herpétique. Le dernier essai présentait les résultats sous un format ne convenant pas à une méta-analyse. Il n'y avait aucune différence significative en termes d'événements indésirables graves ou sans gravité entre les groupes recevant les corticostéroïdes et les groupes recevant un placebo. Il n'y avait pas non plus de différence significative entre les groupes de traitement et les groupes recevant un placebo pour l'analyse des autres résultats secondaires et les analyses en sous-groupes. D'après ces données de qualité moyenne, on peut conclure que les corticostéroïdes ne sont pas efficaces pour prévenir la névralgie post-herpétique.

Notes de traduction

Traduit par: French Cochrane Centre 22nd March, 2013
Traduction financée par: Instituts de Recherche en Sant� du Canada, Minist�re de la Sant� et des Services Sociaux du Qu�bec, Fonds de recherche du Qu�bec-Sant� et Institut National d'Excellence en Sant� et en Services Sociaux pour la France: Minist�re en charge de la Sant�

Summary of findings(Explanation)

Summary of findings for the main comparison. Corticosteroids for acute herpes zoster to prevent postherpetic neuralgia
  1. 1 There is a high risk of bias due to inadequately addressed incomplete outcome data of the Esmann 1987 trial, in which six participants were withdrawn, but the reasons and assigned groups of five cases were not specified.

Corticosteroids for acute herpes zoster to prevent postherpetic neuralgia
Patient or population: patients with acute herpes zoster to prevent postherpetic neuralgia
Settings: hospitals and clinics
Intervention: corticosteroids
OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of Participants
(studies)
Quality of the evidence
(GRADE)
Comments
Assumed riskCorresponding risk
ControlCorticosteroids
Presence of PHN six months after the onset of the acute herpetic rash
Clinical manifestation
Follow-up: 6-23 months
193 per 1000183 per 1000
(87 to 384)
RR 0.95
(0.45 to 1.99)
114
(2 studies)
⊕⊕⊕⊝
moderate 1
 
Serious adverse events
Clinical manifestation and laboratory examination
Follow-up: 6-23 months
8 per 100013 per 1000
(3 to 42)
RR 1.65
(0.38 to 5.29)
755
(5 studies)
⊕⊕⊕⊝
moderate 1
 
Non-serious adverse events
Clinical manifestation and laboratory examination
Follow-up: 6-23 months
113 per 1000147 per 1000
(102 to 211)
RR 1.30
(0.9 to 1.87)
755
(5 studies)
⊕⊕⊕⊝
moderate 1
 
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: confidence interval; RR: risk ratio
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Background

Postherpetic neuralgia (PHN) is a painful condition that occurs in people following an acute herpes zoster infection (commonly referred to as 'shingles'). Shingles is an acute vesicular eruption involving one or two adjacent dermatomes, with pain often preceding the eruption by days to weeks (Kost 1996). Herpes zoster results from reactivation of the varicella-zoster virus acquired during chicken pox, the primary varicella infection. Reactivation of latent varicella-zoster virus from dorsal root ganglia is responsible for the classic dermatomal rash and pain that occurs with herpes zoster (Kost 1996).

Herpes zoster is a sporadic disease with an estimated lifetime incidence of 10% to 20%. Its incidence increases sharply with advancing age, roughly doubling in each decade past the age of 50 years. Herpes zoster is uncommon in people less than 15 years old. The normal age-related decrease in cell-mediated immunity is thought to account for the increased incidence in older age (Stankus 2000). People with disease states that affect cell-mediated immunity, such as human immunodeficiency virus (HIV) infection and certain malignancies, are at increased risk. Chronic corticosteroid use, chemotherapy and radiation therapy may increase the risk of developing herpes zoster (Fillet 2002). Ethnic background may influence susceptibility to herpes zoster. Black people are one fourth less likely than white people to develop it. Although herpes zoster is not as contagious as primary varicella infection, people with reactivated infection can transmit varicella-zoster virus to non-immune contacts. There is no seasonal incidence and the areas of the body affected tend to be the chest and abdomen and the territory of the ophthalmic division of the trigeminal nerve.

PHN is one of the most common complications of herpes zoster. It may persist until death and has major implications for quality of life and use of healthcare resources. Although PHN has been defined in different ways, recent data support the distinction between acute (within 30 days of rash onset), subacute (30 to 120 days after rash onset) and PHN (defined as pain lasting at least 120 days from rash onset) (Desmond 2002; Dworkin 1994).

Although age, acute pain severity and rash severity appear to be correlated with incidence of PHN, accurate predictors for PHN have not been defined (Johnson 2003). About 20% of people with herpes zoster develop PHN. Its incidence is between 9% and 14% one month after the herpes zoster eruption. The most established risk factor is age. As age increases, the risk and duration of PHN also rise (Griffin 1998; Rosler 1996). The incidence of PHN after an outbreak of shingles is 10% in people over 40 years, and 20% to 50% in people over 60 years. PHN is rarely seen in people under 30 years. Other possible risk factors for the development of PHN are ophthalmic zoster, prodromal pain before the appearance of skin lesions and an immunocompromised state (Stankus 2000).

There is a tendency for PHN to improve with time and as few as 3% of people are left with severe PHN after one year. However, some series report that as many as 40% of people with PHN will continue to have long-term problems because of incomplete or no pain relief from treatments. There is no way of predicting who will recover (de Moragas 1957).

Varicella-zoster virus is a highly contagious DNA virus. It is thought that the varicella virus passes to the dorsal root ganglion via the skin during the initial infection (chicken pox) and lies dormant. The latent virus becomes reactivated when immune mechanisms are impaired and it is manifested by the rash and the pain. The pathophysiology of PHN remains unclear. However, pathologic studies have demonstrated damage to the sensory nerves, sensory dorsal root ganglia and dorsal horns of the spinal cord. The presence of PHN may reflect the persistence of more than the low amounts of varicella-zoster virus found during latency with continued inflammation. If this is the case, there may be a rationale for the aggressive treatment with aciclovir, and perhaps corticosteroids, of people who have zoster (Mahalingam 1993; Smith 1978).

The treatment of herpes zoster has three major objectives: (1) treatment of the acute viral infection, (2) treatment of the acute pain associated with herpes zoster, and (3) prevention of PHN. Antiviral agents, oral corticosteroids and adjunctive individualised pain-management modalities are used to achieve these objectives.

Treatment of PHN is difficult and a variety of treatments are offered without consensus about their effectiveness. The complexity of the underlying changes might account for the lack of efficacy of a single therapeutic approach (Alper 2002; Dworkin 2000; Johnson 2003). The effectiveness of antiviral agents in preventing PHN has been evaluated in a separate Cochrane review; it concluded that oral aciclovir was ineffective in reducing the incidence of PHN while insufficient evidence was found to recommend other antiviral treatments to prevent PHN (Li 2009).

Some older studies designed to evaluate the effectiveness of corticosteroids such as prednisolone or triamcinolone prednisone therapy in preventing PHN have suggested decreased pain at three and 12 months (Eaglstein 1970; Keczkes 1980). Other studies have demonstrated no significant benefit (Lancaster 1995; Volmink 1996). Another two large randomised, placebo-controlled trials evaluated the combination of corticosteroids and the antiviral agent aciclovir. One claimed that the addition of prednisone reduced the incidence and severity of acute pain but provided no additional benefit for long-term pain over aciclovir alone (Wood 1994b). The other suggested that aciclovir and a corticosteroid did not significantly alter the course of long-term zoster-associated pain but might improve quality of life (Whitley 1996). Despite the lack of clear evidence, corticosteroids are commonly used in the treatment of herpes zoster. A systematic review of corticosteroids for preventing PHN is needed to try and clarify the evidence and include in any meta-analysis the results of past and future trials.

The first version of this Cochrane review, published in 2008, indicated that there was insufficient evidence to draw any conclusion about the efficacy of corticosteroids in preventing PHN. In this update, which includes more up to date data analysis methodology, we conclude that there is moderate quality evidence that there is no benefit of corticosteroids in preventing postherpetic neuralgia, but that this evidence is limited.

Objectives

To examine the efficacy of corticosteroids in preventing PHN.

Methods

Criteria for considering studies for this review

Types of studies

We searched for all randomised controlled trials (RCTs) for corticosteroids for preventing PHN after an acute herpes zoster infection, irrespective of any language restrictions. We did not include quasi-RCTs (trials in which a systematic method of randomisation such as alternation or hospital number was used).

Types of participants

We included people of all ages with herpes zoster of all degrees of severity, within seven days after the onset.

Types of interventions

We included all kinds of corticosteroids including hydrocortisone, prednisone, prednisolone, triamcinolone and dexamethasone given by an oral, intramuscular or intravenous route during the acute stage (starting within one week of the onset of the rash). We included trials which compared corticosteroids with no treatment or placebo but not with other treatments. It is intended that another review will include comparisons of corticosteroids with antiviral agents. We also included trials which compared corticosteroids plus routine treatment with placebo plus routine treatment. Other forms of administration of corticosteroids such as epidural injection or topical administration were not included.

Types of outcome measures

Primary outcomes

The primary outcome measure was the presence of PHN six months after the onset of the acute herpetic rash. PHN was defined according to clinical diagnostic criteria as persisting or recurring pain at the site of shingles at least one month after the onset of the acute rash (MacDonald 2000).

Secondary outcomes

Secondary outcome measures were as follows.

  1. Pain severity measured by a validated visual analogue scale or numerical descriptive scale after three, six and 12 months.

  2. Quality of life measured with the Short Form-36 Health Survey questionnaire (SF-36) (Ware 1998) after six months.

  3. Adverse events during or within two weeks after stopping treatment. Adverse events were categorised as serious or not serious. Serious adverse events were those which were life-threatening, required or prolonged hospitalisation, or caused death.

Outcomes for inclusion in a 'Summary of findings' table

We included a 'Summary of findings' table to illustrate the findings of the review and our assessment of the quality of the evidence for the key outcomes:

  • presence of PHN six months after the onset of the acute herpetic rash;

  • serious adverse events;

  • non-serious adverse events.

We graded the quality of the evidence from the included RCTs as high, moderate, low or very low based on the GRADE criteria. These start from a grading of high for RCTs. Reasons for downgrading are: study limitations, consistency of effect, imprecision, indirectness and publication bias. Evidence from downgraded RCTs can be upgraded for a large effect size in cases when all plausible confounding would tend to underestimate the size of the effect or when there is a dose-response gradient (GRADE working group 2004; Schünemann 2011; Schünemann 2011b).

Search methods for identification of studies

We searched for all RCTs of corticosteroids for preventing PHN after an acute herpes zoster infection irrespective of the language of publication.

We searched the Cochrane Neuromuscular Disease Group Specialized Register for RCTs (16 April 2012). The following search terms were used singly and in appropriate combinations: 'herpes zoster', 'shingles', 'postherpetic' or 'post-herpetic', 'neuralgia' or 'neuropathy' or 'pain', 'glucocorticoids', 'adrenal cortex hormones', 'corticosteroid', 'steroid', 'prednisolone', 'triamcinolone', 'dexamethasone', 'hydrocortisone' and 'prednisone'. We adapted this strategy to search CENTRAL (2012, Issue 3), MEDLINE (January 1966 to April 2012), EMBASE (January 1980 to April 2012), LILACS (January 1982 to April 2012) and the Chinese Biomedical Retrieval System (January 1978 to February 2012). We also reviewed the bibliographies of the trials identified, contacted the authors and known experts in the field, and approached pharmaceutical companies to identify additional published or unpublished data.

The detailed search strategies are in the appendices: Appendix 1 (MEDLINE), Appendix 2 (EMBASE), Appendix 3 (CENTRAL), Appendix 4 (LILACS) and Appendix 5 (Chinese Medical Retrieval System).

Data collection and analysis

Selection of studies

Two review authors scrutinised titles and abstracts identified from the register. The review authors obtained the full text of all potentially relevant studies for independent assessment. Three review authors scrutinised all possible published and unpublished trials for inclusion. We resolved any disagreement by discussion.

Data extraction and management

Two review authors extracted data on participants, methods, interventions, outcomes and results independently and then entered the data into the Review Manager (RevMan 5) Cochrane authoring and statistical software. We obtained missing data from the study authors whenever possible. We extracted data on the number of participants with each outcome event, by allocated treatment group, irrespective of compliance with the protocol, and whether or not the participant was subsequently deemed ineligible or otherwise excluded from treatment or follow-up, so that the data could be analysed on an intention-to-treat basis. We resolved disagreement by discussion.

Assessment of risk of bias in included studies

Two review authors (YH, JZ) assessed the risk of bias in each trial. The assessment of risk of bias took into account security of randomisation, allocation concealment, blinding, completeness of outcome data, selective outcome reporting, and any other potential sources of bias. These items were assessed by two authors independently according to the Cochrane Collaboration standard scheme (Higgins 2011). All included trials were judged for each item as 'High risk of bias', 'Low risk of bias' or 'Unclear risk of bias'. We used 'Unclear risk of bias' if: there was insufficient detail to assess risk of bias, information was available but the risk of bias was unclear, or when the entry was not relevant to the study. We resolved any disagreement by discussion, with reference to a third author if necessary.

Data synthesis

We used RevMan 5 for the statistical analysis and reported data according to the Cochrane Collaboration criteria. Where meta-analysis was possible, we pooled results of clinically and statistically homogeneous trials to provide estimates of efficacy. We planned to analyse all the primary and secondary outcomes under consideration. For dichotomous outcomes we expressed the results as risk ratios (RRs) while for continuous outcomes we compared means and calculated mean differences (MDs), all with 95% confidence intervals (CIs).

To avoid unit-of-analysis error resulting from combining results of more than one time point for each study in a standard meta-analysis, we evaluated outcomes based on the periods of follow-up (six months after disease onset). For studies that compared more than two intervention groups, we selected the relevant pair of intervention groups to include in the analyses.

Subgroup analysis and investigation of heterogeneity

We performed the following subgroup analyses.

  1. Treatment started sooner or later after onset of herpes zoster (24 hours or less after onset; more than 24 hours up to 72 hours after onset; and more than 72 hours after onset).

  2. Younger and older (adults 49 years of age or less; adults aged 50 years or more).

Sensitivity analysis

We assessed heterogeneity among trials using the Chi2 test with a 10% level of statistical significance (P < 0.1) and I2 > 50% (Higgins 2002; Higgins 2003). When significant heterogeneity was present, we planned to undertake sensitivity analyses by repeating the calculation after omitting the trials which had a high risk of bias. We used a fixed-effect model for meta-analysis unless unexplained heterogeneity was identified when we planned to use a random-effects model analysis. For trials that were clinically heterogeneous or provided insufficient information for pooling, we performed a descriptive analysis.

Results

Description of studies

See tables: Characteristics of included studies and Characteristics of excluded studies.

Results of the search

For the 2012 update, the electronic searches retrieved a number of references: 20 from the Cochrane Neuromuscular Disease Group Specialized Register, 73 from CENTRAL, 760 from MEDLINE, 164 from EMBASE, 6 from LILACS, and 338 from the Chinese Biomedical Retrieval System. After scrutinising these titles and abstracts, we selected 43 possible RCTs, 35 of which were the same as those in the first version of this review. We found three additional potentially relevant trials (Brusco 1993; Huang 2010; Zeng 2011) during the search for the 2012 update, but these were found not to be RCTs on further checking. No other new trials were found by searching other sources. We excluded four trials that were awaiting assessment in the previous update as the participants' course of disease from onset to start of treatment exceeded seven days (Lin 2002; Wang 2004; Xu 1999; Zhang 2005); one study is still awaiting assessment (Hu 2001).

We excluded 37 trials (see Characteristics of excluded studies). Of these, two trials (Benoldi 1991; Keczkes 1980) compared corticosteroids with other treatment. Nineteen studies were found not to be RCTs by contacting the authors (Brusco 1993; Cui 2002; Hao 2002; Huang 2004; Huang 2010; Jiang 2008; Li 2000; Li 2002; Lin 2005; Ma 2000; Ma 2002; Shi 2008; Song 2009; Tang 2004; Yang 2010; Zeng 2011; Zhang 2003; Zhou 2000; Zhou 2008). In seven trials, participants received therapy that was mismatched in dosage, course of treatment or basal medication between corticosteroids and control groups (Guo 2001; Jiang 2005; Yang 2000; Yin 2004; Yin 2005; Zhang 2004; Zheng 2004). Four trials defined PHN as pain persisting at the site of shingles two weeks after the onset of the acute rash (Chang 2004; Liu 2003; Liu 2005; Yang 2002) and another trial defined PHN as pain persisting one week after total decrustation (Liao 2005); the follow-up of each of these studies was less than one month. As described above, four trials exceeded our seven days in the treatment criterion (Lin 2002; Wang 2004; Xu 1999; Zhang 2005).

The study currently awaiting classification did not clarify the exact course of disease from onset of herpes zoster to receipt of treatment (Hu 2001). We contacted the study authors by mail but obtained no reply. Another article (Levinson 1985), which was classified as a study awaiting assessment in the first version of this Cochrane review, was a review of the previous studies. Although the authors said they were investigating the feasibility of a multicentre trial in this field, we did not find any subsequent trial report so we removed the reference from studies awaiting assessment.

Five trials (Clemmensen 1984; Eaglstein 1970; Esmann 1987; Whitley 1996; Wood 1994a) fulfilled the selection criteria (see Characteristics of included studies).

For the 2012 update we did not find any new trials for inclusion. Although we found two more trials published in the last two years in the Chinese databases (Huang 2010; Zeng 2011) and one trial in LILACS (Brusco 1993), we excluded all three because their participants were not truly randomly assigned to groups.

Trial design

The included trials were all randomised, double-blind, placebo-controlled parallel studies. Two of these were performed in a single centre (Clemmensen 1984 conducted in Denmark; Eaglstein 1970 conducted in a Miami dermatology inpatient service) and the others were performed in multiple centres (Esmann 1987 conducted in Aarhus and Copenhagen, Denmark; Whitley 1996 conducted in 15 university hospitals or affiliated clinics in USA; Wood 1994a conducted in four centres in the United Kingdom).

Participants

A total of 787 participants were enrolled in the five included studies. Four trials (Clemmensen 1984; Eaglstein 1970; Esmann 1987; Whitley 1996) reported the range of ages as 16 to 91 years old. Four trials (Clemmensen 1984; Esmann 1987; Whitley 1996; Wood 1994a) stated the gender distribution (male 307, female 427) and mean age of participants. All of the five trials defined explicit inclusion criteria. Among them, one trial (Eaglstein 1970) included participants with early, severely painful zoster; four trials (Clemmensen 1984; Esmann 1987; Whitley 1996; Wood 1994a) included participants with herpes or pain of different grades of severity. One trial (Esmann 1987) included participants aged at least 60 years and with onset of herpes zoster less than 96 hours before admission. One trial (Whitley 1996) included immunocompetent adults older than 50 years of age who fell ill less than 72 hours before study enrolment. One trial (Wood 1994a) included adults who fell ill less than 72 hours before study enrolment. All five included trials also defined explicit exclusion criteria. They excluded participants with peptic ulcer, psychosis, malignant disease, hypertension, diabetes, cardiac insufficiency, adrenocortical disease, tuberculosis, lymphomas, leukaemias, bacterial infections, pregnancy, or those who were on corticosteroid treatment. The time of onset to start of treatment was zero to seven days for four included trials (Clemmensen 1984; Esmann 1987; Whitley 1996; Wood 1994a). One trial (Eaglstein 1970) only stated that the mean time was five days, and more details could not be obtained.

Interventions

The treatment regimens varied between studies (see Characteristics of included studies). Two trials (Clemmensen 1984; Eaglstein 1970) compared corticosteroids with placebo. One trial used oral triamcinolone 16 mg three times daily for seven days, 8 mg three times daily for seven days, and 8 mg twice daily for seven days (Eaglstein 1970). One trial administered corticosteroid orally or adrenocorticotropic hormone (ACTH) intramuscularly (Clemmensen 1984). Prednisone was given in doses of 45 mg daily during the first week, 30 mg daily during the second week, and 15 mg daily tapered to zero during the third week. ACTH (Synacthen depot, 1 mg) was given intramuscularly three times a week (Monday, Wednesday, Friday) amounting to a total of seven injections. Placebo tablets or injections that were indistinguishable from the active medication were used (Clemmensen 1984). Three trials used aciclovir in combination with corticosteroids versus aciclovir in combination with placebo (Esmann 1987; Whitley 1996; Wood 1994a). One trial used 800 mg aciclovir orally five times daily for seven days and coded tablets containing either prednisolone or calcium lactate for 21 days. The dose of prednisolone was 40 mg daily for seven days, 30 mg for four days, 20 mg for three days, 10 mg for four days, and finally 5 mg for three days (Esmann 1987). One trial used prednisone or a matched placebo orally, 60 mg/day for days one to seven, 30 mg/day for days eight to 14, and 15 mg/day for days 15 to 21 (Whitley 1996). Aciclovir or a matched placebo was administered orally at 800 mg five times daily for 21 days. Matched medications were identical in taste and appearance. The four treatment regimens given were aciclovir plus prednisone, aciclovir plus prednisone placebo, prednisone plus aciclovir placebo, and placebos for both aciclovir and prednisone (Whitley 1996). Another trial administered aciclovir 800 mg orally five times daily, beginning on day zero (Wood 1994a). The participants in the groups assigned to seven days of aciclovir therapy (with or without corticosteroid) received matching placebo beginning on day seven. Prednisolone was administered according to the following schedule: on days zero through six, 40 mg per day; days seven through 10, 30 mg per day; days 11 through 14, 20 mg per day; days 15 through 18, 10 mg per day; and days 19 through 21, 5 mg per day (total dose 535 mg). Prednisolone was given as 5 mg tablets. The participants in the groups not receiving corticosteroid received matching placebo tablets. The four treatment regimens given were: aciclovir for seven days with corticosteroids, aciclovir for seven days without corticosteroids, aciclovir for 21 days with corticosteroids, and aciclovir for 21 days without corticosteroids (Wood 1994a). All trials followed the participants for at least six months or until the PHN ended or the participant no longer returned. One trial monitored all participants for three years (Eaglstein 1970).

Outcome measures

The outcome measures used differed between trials. Four trials reported the duration of PHN or presence of PHN at six months after the onset of the acute herpetic rash (Eaglstein 1970; Esmann 1987; Whitley 1996; Wood 1994a). Two trials did not provide separate information on the number of participants with PHN at six months and were therefore not included in the meta-analysis (Whitley 1996; Wood 1994a). All five included trials did not report, as a separate outcome, pain severity measured by a validated visual analogue scale or numerical descriptive scale after three, six and 12 months or the quality of life measured with the SF-36 Health Survey after six months. All five included trials reported adverse events during treatment or within two weeks after stopping treatment. Adverse events were categorised as serious or not serious.

Risk of bias in included studies

See: the 'Risk of bias' table of each included study in Characteristics of included studies.

All included trials were randomised, double-blind, placebo-controlled parallel studies. The method of randomisation was reported in three included trials (Eaglstein 1970; Whitley 1996; Wood 1994a). The Eaglstein study (Eaglstein 1970) used randomisation with a centralised code generated at a pharmacy. A supply of medication for each participant was assigned a different code number and distributed to the participants; the code of each participant was opened after a three-year follow-up. The Whitley and Wood studies both used a computer-generated randomisation code to randomly assign participants; in Wood 1994a the randomisation code was stratified by study centre to assign participants in blocks of eight to either group so that allocation concealment might be performed; the method of allocation concealment in Whitley 1996 was unclear. Clemmensen 1984 and Esmann 1987 did not describe the method of randomisation, and it was not clear from the reports if there was adequate allocation concealment in either study. The five included trials were all double-blind, using placebo in the control group. One of these trials used lactose as placebo (Eaglstein 1970), another used calcium lactate (Esmann 1987), and the other three studies only stated placebo tablets indistinguishable from the active medication but did not describe their composition (Clemmensen 1984; Whitley 1996; Wood 1994a). Three of the five trials (Esmann 1987; Whitley 1996; Wood 1994a) included aciclovir as routine treatment. All trials considered baseline clinical features. In four trials the baseline clinical features were similar between groups (Clemmensen 1984; Esmann 1987; Whitley 1996; Wood 1994a). In one trial baseline differences were not described (Eaglstein 1970) but the authors reported no serious imbalances in baseline prognostic factors between groups.

All five included studies reported the time of follow-up; three used six months (Esmann 1987; Whitley 1996; Wood 1994a), one 10 months (Clemmensen 1984), and one three years (Eaglstein 1970). In one study (Esmann 1987) six participants were withdrawn but it was unclear whether the lack of compliance was due to inefficacy or side effects. The group that those participants were first assigned to was not specified so bias from incomplete outcome data was possibly an issue in this study. The other four studies all reported information about follow-up and dropouts and described the reasons for dropping out clearly. Only one study claimed that an intention-to-treat analysis was used (Whitley 1996); the other four included studies did not state whether, or not, the analysis was by intention to treat, but there was sufficient information in three trial reports (Clemmensen 1984; Eaglstein 1970; Wood 1994a) to restore them to the correct group and perform an intention-to-treat analysis in our review.

For each included study, outcomes listed in the methods section were all reported. Publication bias should be taken into account since most of the included and excluded studies were published in English or Chinese, although we have attempted to do our best to search all probable literature without any language restrictions and have contacted investigators to get more information. The included studies used different cut-off times to definite PHN so, although we clearly stated that PHN was pain persisting, or recurring, at the site of shingles at least one month after the onset of the acute rash, we have not restricted inclusion to studies using the same definition in order not to introduce more missing data.

For this update, according to the new 'Risk of bias' assessments (Higgins 2011), we updated 'blinding' and performed the analyses in two parts: blinding of participants and personnel, and blinding of outcome assessment. Two (Eaglstein 1970; Wood 1994a) of the trials were rated as at low risk of bias, two (Clemmensen 1984; Whitley 1996) as unclear risk of bias, and one (Esmann 1987) as at high risk of bias (Figure 1).

Figure 1.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Effects of interventions

See: Summary of findings for the main comparison Corticosteroids for acute herpes zoster to prevent postherpetic neuralgia

Primary outcome measure

Presence of PHN six months after the onset of the acute herpetic rash

One trial comparing triamcinolone with placebo (Eaglstein 1970) provided data on the presence of PHN six months after the onset of the acute herpetic rash. There was no significant difference in the number of participants with PHN six months after the onset of the acute herpetic rash between those in the corticosteroids group (2/15,13.3%) and those in the placebo group (2/20, 10.0%). However, the wide CIs meant we could not rule out significant benefit or harm (RR 1.33, 95% CI 0.21 to 8.41). This study also reported the presence of PHN at other time points during follow-up, including one and four months after the onset, which are also commonly used to evaluate incidence of PHN. The presence of PHN was not statistically different between groups (9/15 versus 14/20 at one month, P = 0.55; 2/15 versus 4/20 at four months, P = 0.61). Another trial comparing prednisolone plus routine treatment with placebo plus routine treatment (Esmann 1987) provided data for our primary outcome. The trial compared aciclovir plus corticosteroids with aciclovir plus placebo. The presence of PHN six months after the onset of the acute herpetic rash following corticosteroids plus antiviral agents (9/42, 21.4%) was not significantly different from its presence following placebo plus antiviral agents (9/37, 24.3%) (RR 0.88, 95% CI 0.39 to 1.98).

We conducted a meta-analysis combining relevant data from the above two trials, with a total of 114 participants, and the results showed that oral corticosteroids did not play a part in preventing PHN six months after the onset of herpes (RR 0.95, 95% CI 0.45 to 1.99, P = 0.89) (Analysis 1.1; Figure 2). The Clemmensen study (Clemmensen 1984) used a cut-off time of six weeks for defining PHN, concluding that prednisone did not decrease the incidence of PHN. Since the numbers of participants with PHN six months after the rash onset could not be obtained, these data were not included in any meta-analysis for effects of corticosteroids (Summary of findings for the main comparison).

Figure 2.

Forest plot of comparison: 1 Corticosteroids versus placebo or no treatment, outcome: 1.1 The presence of PHN six months after the onset of the acute herpetic rash.

Two other trials provided relevant data for this outcome although not in a format which permitted inclusion in our meta-analysis. In a trial with 201 participants (Whitley 1996), a Cox regression model analysis of the main effect of prednisone compared with no prednisone showed no significant difference in the time to cessation of zoster-associated pain (RR 1.26, 95% CI 0.91 to 1.75). From their Cox regression model, we used the generic inverse variance approach to calculate the main effect of prednisone compared with no prednisone and confirmed that there was no significant difference in the time to cessation of zoster-associated pain (RR 1.11, 95% CI 0.96 to 1.27) (Analysis 1.2; Figure 3). In the trial with 400 participants, of whom 349 completed the study (Wood 1994a), the investigators did not detect significant differences between any of the treatment groups in the time to complete cessation of pain. The median time to cessation of pain was 147 and 120 days in the groups receiving 7-day and 21-day aciclovir without a corticosteroid and 146 and 120 days in the two groups who received aciclovir with corticosteroids. Thus, the results of both these large trials agreed with the conclusion from the meta-analysis of the two smaller trials that corticosteroids did not significantly affect the presence of PHN after six months.

Figure 3.

Forest plot of comparison: 1 Corticosteroids versus placebo or no treatment, outcome: 1.2 The main effect of prednisone compared with no prednisone on six months evaluation of pain (generic inverse variance).

Secondary outcome measures

(1) Pain severity measured by a validated visual analogue scale or numerical descriptive scale after three, six and 12 months

Four included trials (Clemmensen 1984; Eaglstein 1970; Esmann 1987; Wood 1994a) evaluated pain intensity after using corticosteroids to treat acute herpes zoster, but three of them (Clemmensen 1984; Eaglstein 1970; Wood 1994a) recorded the data only during the first month and they used different pain evaluation methods; so we were not able to include the data in a meta-analysis even when attempting to convert outcomes to dichotomous data. The Clemmensen study graded pain from zero (no pain) to three (insufferable pain). There was no significant difference in mean pain score between prednisone and placebo during the 21-day treatment period, and the score was significantly lower in the ACTH group during the first four days of the trial (P = 0.02 to 0.03) but not after (Clemmensen 1984). One trial reported that in participants more than 60 years old, pain resolved spontaneously but more rapidly with corticosteroids (Eaglstein 1970). In the Wood trial (Wood 1994a) the reduction in pain score was significantly larger in the corticosteroids groups than the no-corticosteroids groups on days 7 and 14 (P < 0.01). Only one trial reported pain at six months: 18 participants had pain, of whom 15 had light pain; two had moderate pain; and one had severe pain. Among them nine had received prednisone and nine placebo but the severity of pain was not reported by group (Esmann 1987).

(2) Quality of life measured with the SF-36 Health Survey after six months

None of the trials reported separate data on quality of life measured with the SF-36 Health Survey after six months.

(3) Adverse events during or within two weeks after stopping treatment

Adverse events were categorised as serious or not serious. Serious adverse events were those which were life-threatening, required or prolonged hospitalisation, or caused death. Details for individual studies have been given in the Characteristics of included studies table.

(a) Serious adverse events

Two of the included trials explicitly recorded the absence of serious adverse effects attributable to the experimental treatment (Clemmensen 1984; Eaglstein 1970). The other three trials all reported several serious adverse events during or within two weeks after stopping treatment, including acute cardiac insufficiency (Esmann 1987), myocardial infarction (Whitley 1996), pneumonia or bronchopneumonia (Whitley 1996; Wood 1994a), chest infection (Wood 1994a), haematemesis (Wood 1994a) and death from other unspecified causes (Wood 1994a). In the meta-analysis, the incidence of serious adverse events for corticosteroids (6/376, 1.6%) was not significantly different from that for placebo (3/379, 0.8%) (RR 1.65, 95% CI 0.51 to 5.29, P = 0.40) (Analysis 1.3; Figure 4).

Figure 4.

Forest plot of comparison: 1 Corticosteroids versus placebo, outcome: 1.3 Serious adverse events.

(b) Non-serious adverse events

All five trials reported details of non-serious adverse events, including clinical manifestation or laboratory results, the number of participants experiencing each adverse event, and their distribution between groups. The most frequently reported non-serious adverse events were gastrointestinal symptoms (such as dyspepsia, nausea, vomiting and diarrhoea), dizziness, headache, sweats, rash, oedema, hyperglycaemia, and increase of serum aspartate glutamyltransferase. In our meta-analysis of data from the five trials (Analysis 1.4; Figure 5), the overall incidence of non-serious adverse events for corticosteroids (55/376, 14.6%) was not statistically significant compared to placebo (43/379, 11.3%) (RR 1.30, 95% CI 0.90 to 1.87, P = 0.16).

Figure 5.

Forest plot of comparison: 1 Corticosteroids versus placebo, outcome: 1.4 Non-serious adverse events.

Subgroup analyses

Time from onset of herpes zoster to start of treatment (24 hours or less after onset, more than 24 hours up to 72 hours after onset, and more than 72 hours after onset)

This information was not available from the published reports.

Younger and older (adults 49 years of age or less, adults aged 50 years or more)
(a) Adults aged 50 years or more

The Eaglstein study (Eaglstein 1970) reported duration of pain for each participant in bar charts which clearly showed the age of each participant. Two other trials (Esmann 1987; Whitley 1996) only enrolled participants aged more than 60 or 50 years old, respectively. So a subgroup analysis involving only adults aged 50 years or more was potentially possible for these three trials. Unfortunately the Whitley study was not included in the meta-analysis since detailed numbers of events could not be extracted from this article. Thus in two trials (Eaglstein 1970; Esmann 1987) including 107 participants aged 50 years or more, the presence of PHN six months after the onset of the acute herpetic rash was similar in the corticosteroids group (11/53, 20.8%) to that in the placebo group (11/54, 20.4%) (RR 0.97, 95% CI 0.47 to 2.04) (Analysis 2.1; Figure 6).

Figure 6.

Forest plot of comparison: 2 Subgroup analysis, outcome: 2.1 The presence of postherpetic neuralgia six months after the onset of the acute herpetic rash.

The Clemmensen study (Clemmensen 1984) reported 33 participants in the group of participants aged over 55 years, of whom nine developed PHN (four in the prednisone group, one in the placebo group, and four in the other treatment group) without significant differences between groups. However, the study defined PHN using a six-week cut-off time and only reported the mean duration of PHN as 4.2 months (range 1.5 to 10.0 months), so we were not able to obtain the separate data on participants with PHN after six months follow-up.

(b) Adults 49 years of age or less

In one trial with six participants (Eaglstein 1970), none developed PHN after six months follow-up. In the Clemmensen study (Clemmensen 1984) none of the 22 participants aged 55 years or less developed PHN (the mean duration of PHN was 4.2 months). Relevant data could not be obtained from the other trials.

One of the participants in the Eaglstein study was not included in either the younger or older subgroup as she was withdrawn because of possible side effects. Her age was not reported in the article (Eaglstein 1970).

Sensitivity analyses

Heterogeneity amongst trials was assessed for each comparison, but no significant heterogeneity ( I2 > 50%) was present. Furthermore, although a trial of poor quality was included in the meta-analyses the results of each included trial were all similar with no significant differences in the outcomes relevant to this review. Therefore, we did not undertake any sensitivity analysis.

Discussion

Clinical therapeutic effect

Our aim was to review the evidence from RCTs on the effectiveness and safety of corticosteroids in preventing PHN. Only five studies examining the prevention effects of corticosteroids in a total of 787 participants were suitable for this review. This is a relatively small number in relation to the known variability in the outcome of PHN.

The meta-analysis of two trials which provided data for our primary outcome measure showed no significant difference between the corticosteroid group and the placebo group in the number of participants with PHN six months after the onset of the acute herpetic rash. There was no significant difference in time to cessation of pain in two larger trials, one with 201 and one with 359 participants (Whitley 1996; Wood 1994a).

Established PHN may be intractable and lead to considerable disability in social and domestic activities. Pain evaluation is the key step to controlling neuropathic pain. Doctors must make a detailed and full-scale pain evaluation during the period of treatment and follow-up, including pain character, intensity, position and scope. The most used evaluation methods are the validated visual analogue scale (VAS) or numerical descriptive scales (NRS). Four of the trials included in this review evaluated pain intensity changes but used different pain evaluation methods, so we were unable to combine the data in our meta-analysis. Two trials (Clemmensen 1984; Esmann 1987) reported that corticosteroid treatment did not give additional pain relief during the three or six months of follow-up. One trial (Eaglstein 1970) reported that pain tended to resolve spontaneously without therapy but more rapidly with corticosteroid therapy, but this trial only included 35 participants.

Despite the lack of benefit in terms of reduction of PHN, which was the focus of this review, there is some evidence of a beneficial effect on short-term outcomes. A trial with 359 participants (Wood 1994a) showed that pain intensity after two and three weeks had reduced from baseline significantly more in those who received corticosteroids than in those who did not. Another trial with 201 participants (Whitley 1996) showed significantly faster recovery with corticosteroids than without. On evaluation after one month, with the Cox regression model, the main effect of prednisone compared with no prednisone was significant for all four outcomes reported:

  1. time to cessation of acute neuritis (RR 2.28, 95% CI 1.35 to 3.86);

  2. time to return to uninterrupted sleep (RR 1.65, 95% CI 1.14 to 2.41);

  3. time to return to 100% usual daily activity (RR 1.74, 95% CI 1.21 to 2.51);

  4. time to total cessation of analgesic therapy (RR 2.25, 95% CI 1.42 to 3.54).

The results of these two papers suggest that corticosteroids may have a significant effect in accelerating healing and reducing acute zoster pain. This conclusion differs from the lack of effect of corticosteroids on the persistence of PHN and suggests that the relationship between acute inflammation, pain and PHN is not simple.

Adverse events

All five included trials reported adverse events, but these were not significantly more common in corticosteroid than in placebo participants.

Subgroup analyses

Although the most established risk factor for PHN is age, accurate predictors for PHN have not been defined (Johnson 2003). We conducted subgroup analyses according to age of participants but did not find a significant benefit either in those older than 50 years or those younger. Small numbers make this conclusion very uncertain. There was a lack of sufficient detail to permit extraction of all required data concerning most subgroups of interest. In the absence of a significant effect in the primary outcome measures, and the fact that individual trials were too small to detect moderate effects, more extensive subgroup analysis would have been unreliable. We hope that publication of this review will encourage authors of future PHN prevention trials to collect and publish data which will allow the analysis of subgroups in subsequent systematic reviews.

Outcome measures

In the trials reviewed, the outcome measures involved crude clinical endpoints or a simple pain scale which may be insufficiently responsive to detect meaningful clinical effects. None of the trials reported separate data on pain severity measured by validated visual analogue scales or numerical descriptive scales after three, six or 12 months. However, even the results using these scales are affected by difficulties in standardisation, which make it difficult to draw useful conclusions.

None of the trials reported separate data on quality of life measured with the SF-36 questionnaire after six months.The SF-36 questionnaire is used to assess physical functioning, bodily pain, vitality, social functioning, emotional role and mental health, but can also be used for evaluating quality of life among pain patients (Ware 1998). None of the four trials used this questionnaire to assess the quality of life among participants with pain, and only one trial evaluated aspects of quality of life (Whitley 1996). Future randomised controlled trials (RCTs) should monitor long-term quality of life regularly, which could allow better evaluation of the efficacy of corticosteroids. Regular monitoring of participants' quality of life would improve understanding of the natural progression of PHN with respect to its impact on the physical, emotional and social well-being of those affected.

No trial has incorporated cost-effectiveness calculations.

Future trials

The meta-analysis of two small studies (Eaglstein 1970; Esmann 1987) and the analysis of another two large studies (Wood 1994a; Whitley 1996) showed that corticosteroids did not reduce pain at six months after the onset of the acute rash more than the control group. However, the results of the Wood and Whitley studies suggested that corticosteroids may reduce the acute pain in herpes zoster and may improve quality of life.

The relationship between acute inflammation, pain and PHN is complicated. The effect of corticosteroids is not clear at different stages in the transition from acute pain to PHN. Dworkin et al have recommended the use of systemic corticosteroids as soon as possible after diagnosis of herpes zoster for people with at least moderately severe pain and no contraindications, and that these should be initiated only in combination with antiviral therapy (Dworkin 2007). However, this recommendation is based on expert opinion and the evidence for the effect of corticosteroids on preventing PHN at six months would not support this recommendation. Trials addressing short-term pain relief with corticosteroids are lacking. If they were performed they should also include a long-term follow-up to assess the transition from short-term to long-term pain. Further high quality RCTs using validated and generally accepted outcome measures should be considered for both short and long-term pain prevention and treatment in PHN. In a condition where the natural course is spontaneous recovery in the majority of cases, the number of participants has to be large enough to give the study the power to detect clinically relevant improvements.

Authors' conclusions

Implications for practice

Short courses of corticosteroids do not result in significantly more adverse events in participants with acute herpes zoster but they are ineffective in preventing PHN persisting at six months.

Implications for research

Moderate quality evidence does not support the use of corticosteroids in acute herpes zoster infection for preventing PHN. However, large trials with sufficient power to detect a meaningful difference and which include validated and approved pain outcomes have not been performed. Further high quality studies to assess the effect of corticosteroids on both short-term pain and longer-term PHN are required. This may provide information about the mechanisms of transition from acute pain to long-term PHN.

Acknowledgements

We would like to acknowledge the Chinese Evidence-Based Medicine Center and the Cochrane Neuromuscular Disease Group for their technical support. We thank Professor Richard Hughes, Kate Jewitt, Janice Fernandes, Ruth Brassington, Rachel Barton and Dr Tony Swan for their advice and encouragement in the preparation of this review.

The editorial base of the Cochrane Neuromuscular Disease Group is supported by the MRC Centre for Neuromuscular Diseases.

Data and analyses

Download statistical data

Comparison 1. Corticosteroids versus placebo
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 The presence of PHN six months after the onset of the acute herpetic rash2114Risk Ratio (M-H, Fixed, 95% CI)0.95 [0.45, 1.99]
2 The main effect of prednisone compared with no prednisone on six months evaluation of pain (generic inverse variance)1 RR Ratios (Fixed, 95% CI)Totals not selected
3 Serious adverse events5755Risk Ratio (M-H, Fixed, 95% CI)1.65 [0.51, 5.29]
4 Non-serious adverse events5755Risk Ratio (M-H, Fixed, 95% CI)1.30 [0.90, 1.87]
Analysis 1.1.

Comparison 1 Corticosteroids versus placebo, Outcome 1 The presence of PHN six months after the onset of the acute herpetic rash.

Analysis 1.2.

Comparison 1 Corticosteroids versus placebo, Outcome 2 The main effect of prednisone compared with no prednisone on six months evaluation of pain (generic inverse variance).

Analysis 1.3.

Comparison 1 Corticosteroids versus placebo, Outcome 3 Serious adverse events.

Analysis 1.4.

Comparison 1 Corticosteroids versus placebo, Outcome 4 Non-serious adverse events.

Comparison 2. Subgroup analysis
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 The presence of PHN six months after the onset of the acute herpetic rash2 Risk Ratio (M-H, Fixed, 95% CI)Subtotals only
1.1 Adults aged 50 years or more2107Risk Ratio (M-H, Fixed, 95% CI)0.97 [0.47, 2.04]
1.2 Adults 49 years of age or less16Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
Analysis 2.1.

Comparison 2 Subgroup analysis, Outcome 1 The presence of PHN six months after the onset of the acute herpetic rash.

Appendices

Appendix 1. MEDLINE (OvidSP) search strategy

Database: Ovid MEDLINE(R) <1946 to April Week 1 2012>
Search Strategy:
--------------------------------------------------------------------------------
1 randomized controlled trial.pt. (324061)
2 controlled clinical trial.pt. (83872)
3 randomized.ab. (228422)
4 placebo.ab. (130165)
5 drug therapy.fs. (1520746)
6 randomly.ab. (165133)
7 trial.ab. (236319)
8 groups.ab. (1087415)
9 or/1-8 (2821643)
10 exp animals/ not humans.sh. (3698513)
11 9 not 10 (2395603)
12 exp Herpes Zoster/ (8963)
13 herpes zoster.mp. (10273)
14 shingle$.mp. (710)
15 neuralgia/ (6430)
16 (postherpetic or post herpetic).mp. (1817)
17 15 and 16 (724)
18 PHN.tw. (944)
19 postherpetic neuralgia.mp. (1109)
20 post herpetic neuralgia.mp. (465)
21 post-herpetic neuralgia.mp. (465)
22 or/12-14,17-21 (11470)
23 Glucocorticoids/ (43624)
24 glucocorticoid$.mp. (75141)
25 adrenal cortex hormone/ (49254)
26 adrenal cortex hormone$.mp. (49446)
27 corticosteroid$.mp. (65614)
28 exp Steroids/ (658271)
29 steroid$.mp. (236281)
30 Prednisolone/ (27327)
31 prednisolone$.mp. (35082)
32 TRIAMCINOLONE/ (3003)
33 triamcinalone$.mp. (25)
34 DEXAMETHASONE/ (41472)
35 dexamethasone$.mp. (52796)
36 triamcinolone$.mp. (8540)
37 HYDROCORTISONE/ (58559)
38 hydrocortisone$.mp. (62689)
39 PREDNISONE/ (32495)
40 prednisone$.mp. (41193)
41 or/23-40 (883394)
42 11 and 22 and 41 (760)

Appendix 2. EMBASE (OvidSP) search strategy

Database: Embase <1980 to 2012 Week 15>
Search Strategy:
--------------------------------------------------------------------------------
1 crossover-procedure.sh. (33524)
2 double-blind procedure.sh. (108172)
3 single-blind procedure.sh. (15702)
4 randomized controlled trial.sh. (319715)
5 (random$ or crossover$ or cross over$ or placebo$ or (doubl$ adj blind$) or allocat$).tw,ot. (856935)
6 trial.ti. (128545)
7 or/1-6 (982129)
8 (animal/ or nonhuman/ or animal experiment/) and human/ (1169403)
9 animal/ or nonanimal/ or animal experiment/ (3254561)
10 9 not 8 (2698485)
11 7 not 10 (899899)
12 limit 11 to embase (696033)
13 exp *Herpes Zoster/ or exp herpes zoster/pc (9000)
14 herpes zoster.tw. (7301)
15 shingle$.tw. (971)
16 neuralgia/ (5689)
17 (postherpetic or post herpetic).tw. (2648)
18 16 and 17 (371)
19 PHN.tw. (1278)
20 postherpetic neuralgia.tw. (1698)
21 post herpetic neuralgia.tw. (756)
22 post-herpetic neuralgia.tw. (756)
23 or/13-15,18-22 (13391)
24 Glucocorticoid/ (56327)
25 glucocorticoid$.tw. (53251)
26 adrenal cortex hormone/ (153885)
27 adrenal cortex hormone$.tw. (475)
28 corticosteroid$.tw. (84499)
29 exp Steroid/ (987285)
30 steroid$.mp. (270140)
31 Prednisolone/ (81643)
32 prednisolone$.tw. (21934)
33 TRIAMCINOLONE/ (9114)
34 triamcinalone$.tw. (27)
35 DEXAMETHASONE/ (92367)
36 dexamethasone$.tw. (45944)
37 triamcinolone$.tw. (6002)
38 HYDROCORTISONE/ (85948)
39 hydrocortisone$.tw. (14239)
40 PREDNISONE/ (112202)
41 prednisone$.tw. (23775)
42 or/24-41 (1087022)
43 12 and 23 and 42 (166)
44 remove duplicates from 43 (164)

Appendix 3. The Cochrane Library CENTRAL search strategy

#1 MeSH descriptor Herpes Zoster explode all trees
#2 "herpes zoster"
#3 shingle*
#4 MeSH descriptor Neuralgia, Postherpetic, this term only
#5 PHN
#6 (postherpetic or "post herpetic" or "post-herpetic") and neuralgia
#7 (#1 OR #2 OR #3 OR #4 OR #5 OR #6)
#8 MeSH descriptor Glucocorticoids explode all trees
#9 glucocorticoid*
#10 MeSH descriptor Adrenal Cortex Hormones explode all trees
#11 "adrenal cortex hormone"
#12 corticosteroid*
#13 MeSH descriptor Steroids explode all trees
#14 steroid*
#15 MeSH descriptor Prednisolone explode all trees
#16 prednisolone
#17 MeSH descriptor Triamcinolone explode all trees
#18 triamcinalone
#19 MeSH descriptor Dexamethasone explode all trees
#20 dexamethasone
#21 triamcinolone
#22 MeSH descriptor Hydrocortisone explode all trees
#23 hydrocortisone
#24 MeSH descriptor Prednisone explode all trees
#25 prednisone
#26 (#8 OR #9 OR #10 OR #11 OR #12 OR #13 OR #14 OR #15 OR #16 OR #17 OR #18 OR #19 OR #20 OR #21 OR #22 OR #23 OR #24 OR #25)
#27 (#7 AND #26)

Appendix 4. LILACS search strategy

(Mh Herpes Zoster OR Tw herpes zoster OR Tw shingle$ OR (Mh neuralgia AND (postherpetic OR post herpetic)) OR Tw PHN OR Tw postherpetic neuralgia OR Tw post herpetic neuralgia OR post-herpetic neuralgia) [Words] and 

(Mh Glucocorticoids OR Tw glucocorticoid$ OR Mh adrenal cortex hormone OR Tw adrenal cortex hormone$ OR Tw corticosteroid$ OR Mh Steroids OR Tw steroid$ OR Mh Prednisolone OR Tw prednisolone$ OR  Mh TRIAMCINOLONE OR Tw  triamcinalone$ OR Mh DEXAMETHASONE Or Tw dexamethasone$ OR Tw triamcinolone$ OR Mh HYDROCORTISONE OR Tw hydrocortisone$ OR Mh PREDNISONE OR Tw prednisone$) [Words] and 

((Pt randomised controlled trial OR Pt controlled clinical trial OR Mh randomised controlled trials OR Mh random allocation OR Mh double-blind method OR Mh single-blind method) AND NOT (Ct animal AND NOT (Ct human and Ct animal)) OR (Pt clinical trial OR Ex E05.318.760.535$ OR (Tw clin$ AND (Tw trial$ OR Tw ensa$ OR Tw estud$ OR Tw experim$ OR Tw investiga$)) OR ((Tw singl$ OR Tw simple$ OR Tw doubl$ OR Tw doble$ OR Tw duplo$ OR Tw trebl$ OR Tw trip$) AND (Tw blind$ OR Tw cego$ OR Tw ciego$ OR Tw mask$ OR Tw mascar$)) OR Mh placebos OR Tw placebo$ OR (Tw random$ OR Tw randon$ OR Tw casual$ OR Tw acaso$ OR Tw azar OR Tw aleator$) OR Mh research design) AND NOT (Ct animal AND NOT (Ct human and Ct animal)) OR (Ct comparative study OR Ex E05.337$ OR Mh follow-up studies OR Mh prospective studies OR Tw control$ OR Tw prospectiv$ OR Tw volunt$ OR Tw volunteer$) AND NOT (Ct animal AND NOT (Ct human and Ct animal))) [Words]

Appendix 5. Chinese Biomedical Retrieval System Database search strategy

(NB. all of the search terms were translated to Chinese terms when we conducted the searches)
1. herpes zoster
2. postherpetic neuralgia
3. PHN
4. shingle
5. 1-4/or
6. herpes
7. neuralgia
8. 6 and 7
9. 5 or 8
10.corticorsteroid
11.hormone
12.steroid
13.prednisone
14.prednisolone
15.triamcinolone
16.dexamethasone
17.hydrocortisone
18.10-17/or
19.random
20.control
21.clinical trial
22.blind procedure
23.placebo
24.19-23/or
25.9 and 18 and 24

What's new

DateEventDescription
23 October 2012New citation required but conclusions have not changedUpdated searches integrated but no new trials identified. Change in first author.
3 September 2012New search has been performedSearches updated to April 2012. We also updated the methods of assessment of risk of bias. Some edits throughout; abstract and plain language summary revised.

History

Protocol first published: Issue 1, 2006
Review first published: Issue 1, 2008

DateEventDescription
8 November 2010New citation required but conclusions have not changedNew authors have joined the review update team
17 July 2010New search has been performedFor the 2010 update we updated the searches, but found no new trials. We assessed risk of bias using the new methods, added a 'Summary of findings' table and revised the review.
14 November 2007New citation required and conclusions have changedSubstantive amendment

Contributions of authors

Ying Han and Jingjing Zhang identified the studies, reassessed their risk of bias and made a revision for the new version. Ning Chen and Mi Yang performed the previous update of the review in 2010. Li He is a consultant and expert in this field. She is the contact author, originally suggested the review and contributed to writing the final version of the protocol and review. She also contributed to updating the review and offering expert advice. Dongping Zhang developed and wrote the protocol and first review, and entered the text into Review Manager. Muke Zhou contributed to developing and writing the final version of the protocol and review. Cairong Zhu contributed to writing the section of the manuscript that refers to the planned statistical analysis.

Declarations of interest

None known

Sources of support

Internal sources

  • None, Not specified.

External sources

  • None, Not specified.

Differences between protocol and review

In the 2010 update we assessed risk of bias using new methods and added a 'Summary of findings' table. We introduced a change from the protocol (He 2006) to exclude quasi-RCTs, because RCTs are thought to be the only way to prevent systematic differences in baseline characteristics between intervention groups, according to the latest Cochrane Handbook for Systematic Reviews of Interventions. However, no quasi-RCTs were found.

In the 2012 update we split 'blinding' into assessments of blinding of participants and personnel (performance bias) and blinding of outcome assessors (assessment bias) and revised the wording of assessments to high, low or unclear (Higgins 2011). We added information about the 'Summary of findings' table to the methods section.

Changes to the team of authors have taken place. See Other published versions of this review.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Clemmensen 1984

MethodsSingle centre RCT, double-blind, placebo-controlled, parallel-group study (methods not described).
Participants

60 patients (33 males and 22 females) within 7 days of onset of HZ.

Age range: 16 to 86 years, 33 were 55 years or older. Among them 20 participants received intramuscular ACTH (Synacthen depot). 5 participants dropped out.

Exclusion criteria: (1) duration of symptoms (pain and/or cutaneous signs) beyond 7 days; (2) age under 16 years; (3) generalised HZ (more than 50 vesicles outside the affected dermatome); (4) history of, or current, malignant disease; (5) treatment with cytostatics and corticosteroids; (6) history or findings of peptic ulcer, psychosis, cardiac decompensation, hypertension, diabetes mellitus, adrenocortical disease or with symptoms of osteoporosis; (7) pregnancy.

InterventionsACTH: ACTH depot 1 mg intramuscularly 3 times a week (Monday, Wednesday, Friday) amounting to a total of 7 injections.
Prednisone: orally, 45 mg daily during the 1st week, 30 mg daily during the 2nd week, and 15 mg daily tapered to zero during the 3rd week.
Comparison treatment: placebo.
OutcomesPrimary outcome (mean duration of PHN: 4.2 months, which ranged from 1.5 to 10 months) and secondary outcomes (serious adverse events and non-serious adverse events) were available.
NotesConducted in Denmark
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskMethod of randomisation was not described.
Allocation concealment (selection bias)Unclear riskMethod of allocation concealment was not described.
Blinding of participants and personnel (performance bias)
All outcomes
Low riskIt was stated that a "double-dummy" administration technique was used: matched oral and/or parenteral placebo was given to each participant. Placebo tablets or injections indistinguishable from the active medication.
Blinding of outcome assessment (detection bias)
All outcomes
Low risk"Double-dummy" administration technique.
Incomplete outcome data (attrition bias)
All outcomes
Low risk5 participants dropped out, 1 in the prednisone group (because of increasing blood sugar), 1 in the placebo group (discontinued by the participant without specific reason), 3 in the ACTH group (2 participants developed uncomfortable dizziness and 1 participant moderate periorbital oedema).
Selective reporting (reporting bias)Low riskOutcomes listed in the methods section were all reported.
Other biasLow riskNo other potential bias was found.

Eaglstein 1970

MethodsSingle centre, randomised, double-blind, placebo-controlled parallel design.
Participants

35 patients with early, severely painful HZ admitted to the dermatology inpatient service.

Age: from 21 to 91 years, 24 of them older than 59 years of age. One participant dropped out.
Exclusion criteria: hypertension, tuberculosis, lymphoma, leukaemia, bleeding peptic ulcers, diabetes, cardiac disease, or bacterial infections.

InterventionsUnmarked red capsules containing either 8 mg of triamcinolone or lactose, 2 capsules 3 times daily (48 mg/day) for 7 days, and 1 capsule 3 times daily (24 mg/day) for 7 days, and 1 capsule twice daily (16 mg/day) for 7 days.
OutcomesPrimary outcome (the presence of PHN six months after the onset of the acute herpetic rash) and secondary outcomes (serious adverse events and non-serious adverse events) were available.
NotesConducted in Miami, USA dermatology inpatient service
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskRandom numbers were used; each patient was assigned a different code number and distributed to a group in a random fashion.
Allocation concealment (selection bias)Low riskThe code for each participant was opened after a 3-year follow-up and evaluation.
Blinding of participants and personnel (performance bias)
All outcomes
Low riskParticipants were treated with unmarked red capsules prepared by the hospital pharmacy.
Blinding of outcome assessment (detection bias)
All outcomes
Low riskUnmarked red capsules were prepared by the hospital pharmacy. Comment: a double-blind method was probably done.
Incomplete outcome data (attrition bias)
All outcomes
Low riskOnly one participant in the control group dropped out after five days because of a sudden increase in her BP.
Selective reporting (reporting bias)Low riskOutcomes listed in the methods section were all reported.
Other biasLow riskNo other potential bias was found.

Esmann 1987

MethodsMulticentre randomised, double-blind, placebo-controlled parallel design. Number of losses to follow up: all participants were evaluated at week 26 except for one from the prednisolone group, who was last seen at week 10. She had not had pain since day 5.
Participants

84 patients (25 males and 53 females) within 4 days of onset of HZ.

Age at least 60 years, mean age: intervention group 72.8 years (SD 7.5); control group 71.4 (SD 8.1).
Exclusion criteria: immunocompromise, pituitary or adrenal dysfunction, diastolic BP > 105 mm Hg on entry day, signs of cardiac insufficiency, insulin dependent diabetes, bacterial infections, bleeding peptic ulcers, severe mental confusion, serum creatinine 150 mmol, receiving corticosteroid treatment.

Interventions800 mg aciclovir orally 5 times daily for 7 days and coded tablets containing either prednisolone or calcium lactate for 21 days. The dose of prednisolone was 40 mg daily for 7 days, 30 mg for 4 days, 20 mg for 3 days, 10 mg for 4 days, and finally 5 mg for 3 days.
OutcomesThe primary outcome (the presence of PHN six months after the onset of the acute herpetic rash) and secondary outcomes (serious adverse events and non-serious adverse events) were available.
NotesConducted in Aarhus and Copenhagen, Denmark
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskMethod of randomisation was not described.
Allocation concealment (selection bias)Unclear riskMethod of allocation concealment was not described.
Blinding of participants and personnel (performance bias)
All outcomes
Low riskThe trial report stated that double-blind method was used, and all participants were given coded tablets containing either prednisolone or calcium lactate.
Blinding of outcome assessment (detection bias)
All outcomes
Low riskDouble-blind method was used.
Incomplete outcome data (attrition bias)
All outcomes
High riskSix participants were withdrawn, either because the inclusion criteria could not be upheld upon subsequent scrutiny or because of lack of compliance during the first 1-2 weeks. One of the six participants dropped out because of a possible side effect of prednisolone, and she could be included in the intention-to-treat analysis. But whether the lack of compliance of the other participants was due to inefficacy or side effects, and which group those five participants were firstly assigned to, were not specified.
Selective reporting (reporting bias)Low riskOutcomes listed in the methods section were all reported.
Other biasLow riskNo other potential bias was found.

Whitley 1996

MethodsMulticentre randomised, double-blind, placebo-controlled parallel study with a 2 x 2 factorial design.
Participants

208 immunocompetent patients older than 50 years of age who had localised herpes zoster that developed less than 72 hours before study enrolment. Five randomly assigned participants were not included in this analysis because they never received study medication; no case record forms were submitted. Two other participants were proven to have herpes simplex virus infection and were not included in the analysis.

Of the 201 participants included in the analysis, 51 received aciclovir plus prednisone (24 males and 27 females, mean age 63), 48 received aciclovir plus prednisone placebo (21 males and 27 females, mean age 62), 50 received prednisone plus aciclovir placebo (26 males and 24 females, mean age 60), and 52 received aciclovir and prednisone placebo (25 males and 27 females, mean age 61). 32 participants were lost to follow-up.
Exclusion criteria: immunosuppressive therapy; cancer; women capable of conceiving and bearing a child; history of hypertension (diastolic BP >100 mm Hg) or receiving antihypertensive therapy; osteoporosis or insulin-dependent diabetes mellitus; receipt of other antiviral drugs or immunoglobulin products in the 4 weeks before the study began; history of glycosuria or hyperglycaemia.

Interventions

Prednisone or a matched placebo orally 60 mg/d for days 1 to 7, 30 mg/d for days 8 to 14, and 15 mg/d for days 15 to 21.

Aciclovir or a matched placebo orally as 800 mg 5 x daily, for 21 days. The four treatments regimens were aciclovir plus prednisone, aciclovir plus prednisone placebo, prednisone plus aciclovir placebo, and placebos for both aciclovir and prednisone.

OutcomesThe primary outcome (six-month evaluation of pain (time to cessation of zoster-associated pain)) and secondary outcomes (one-month evaluation of quality of life; serious adverse events and non-serious adverse events) were available.
NotesConducted in 15 university hospitals or affiliated clinics in USA.
Participants discontinued therapy because of influenza, conjunctivitis or iritis, nausea and vomiting, complete resolution of disease, cutaneous dissemination, hyperglycaemia, and bacterial pneumonia.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskA computer-generated randomisation code randomised participants to one of the four treatment groups.
Allocation concealment (selection bias)Unclear riskMethod of allocation concealment was not described.
Blinding of participants and personnel (performance bias)
All outcomes
Low riskAll research personnel remained blinded to drug assignment until the study was completed and the database was locked. All matched medications were identical in taste and appearance.
Blinding of outcome assessment (detection bias)
All outcomes
Low riskAll research personnel remained blinded to drug assignment until the study was completed and the database was locked.
Incomplete outcome data (attrition bias)
All outcomes
Low risk32 participants dropped out: 7/51 in the aciclovir plus prednisone group, 6/48 in the aciclovir plus prednisone placebo group, 9/50 in the prednisone plus aciclovir placebo group, and 10/52 in the wholly placebo-treated group. Missing data were equal among the treatment groups, and an intention-to-treat analysis was performed.
Selective reporting (reporting bias)Low riskOutcomes listed in the methods section were all reported.
Other biasLow riskNo other potential bias was found.

Wood 1994a

  1. a

    ACTH: adrenocorticotrophic hormone

    BP: blood pressure

    HZ: herpes zoster

    PHN: postherpetic neuralgia

    RCT: randomised controlled trial

    SD: standard deviation

MethodsMulticentre randomised, double-blind, placebo-controlled parallel study. Losses to follow up: 2 participants in 7-day aciclovir with corticosteroids; 3 participants in 7-day aciclovir without corticosteroids; 2 participants in 21-day aciclovir with corticosteroids; 3 participants in 21-day aciclovir without corticosteroids.
Participants

Adults over 18 years of age without immune dysfunction due to cancer or immunosuppressive therapy, who presented with a clinical diagnosis of HZ as confirmed by one of the investigators and had a rash for 72 hours or less and at least moderate pain.

400 participants enrolled and assigned to 4 groups:

aciclovir for 7 days with corticosteroids (99 participants, 37 males and 62 females, mean age 59); aciclovir for 7 days without corticosteroids (101 participants, 39 males and 62 females, mean age 58); aciclovir for 21 days with corticosteroid (99 participants, 39 males and 60 females, mean age 60); aciclovir for 21 days without corticosteroid (101 participants, 38 males and 63 females, mean age 59).

51 participants were withdrawn.
Exclusion criteria: pregnant women and women of childbearing potential who were not adequately protected by contraception; renal insufficiency (serum creatinine concentration, more than 1.8 mg per dL), hypertension (diastolic BP >110 mmHg), insulin-dependent diabetes, or random blood glucose > 216 mg/dL (12 mmol/L); history of peptic ulceration, severe psoriasis, or hypersensitivity to aciclovir; and patients receiving barbiturates, anticonvulsant drugs, systemic corticosteroids, rifampicin, or specific antiviral therapy for the present infection.

Interventions

Aciclovir (800 mg orally) five times daily, beginning on day 0. The participants in the groups assigned to 7 days of aciclovir therapy (with or without corticosteroid) received matching placebo beginning on day 7.

Prednisolone (5 mg tablets) according to the following schedule: on days 0 to 6, 40 mg per day; days 7 to 10, 30 mg per day; days 11 to 14, 20 mg per day; days 15 to 18,10 mg per day; and days 19 to 21, 5 mg per day (total dose 535 mg).

OutcomesThe primary outcome (up to month six to assess PHN, and time to complete cessation of pain) and secondary outcomes (serious adverse events and non-serious adverse events) were available.
NotesConducted in 4 clinical centres in the United Kingdom.
51 participants were withdrawn:
14 in aciclovir for 7 days with corticosteroid group, 10 in aciclovir for 7 days without corticosteroid group, 13 in aciclovir for 21 days with corticosteroid group, 14 in aciclovir for 21 days without corticosteroid group.
Reasons were deviation from protocol, adverse events, loss to follow-up, no reason given, and death.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskA computer-generated randomisation code was used.
Allocation concealment (selection bias)Low riskThe randomisation code was stratified by study centre to assigned participants in blocks of eight to either group; it indicated that allocation concealment might be performed.
Blinding of participants and personnel (performance bias)
All outcomes
Low riskIt was stated that it was a double-blind study, and the participants in the groups not receiving corticosteroid received matching placebo tablets.
Blinding of outcome assessment (detection bias)
All outcomes
Low riskDouble-blind method was used.
Incomplete outcome data (attrition bias)
All outcomes
Low riskWithdrawals of participants and the reasons were balanced equally across all groups.
Selective reporting (reporting bias)Low riskOutcomes listed in the methods section were all reported.
Other biasLow riskNo other potential bias was found.

Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion
  1. a

    PHN: postherpetic neuralgia

    RCT: randomised controlled trial

Benoldi 1991RCT, but the control group was treated with carbamazepine.
Brusco 1993Confirmed not a true RCT.
Chang 2004PHN was defined as pain persisting at the site of shingles two weeks after the onset of acute rash.
Cui 2002Confirmed not truly randomised by contacting the study author.
Guo 2001The routine treatments were mismatched between two groups.
Hao 2002Confirmed not truly randomised by contact with study author.
Huang 2004Confirmed not truly randomised by contacting study author.
Huang 2010Confirmed not a true RCT.
Jiang 2005The routine treatments were mismatched between two groups.
Jiang 2008Confirmed not a true RCT, and the duration of herpes zoster was not specified.
Keczkes 1980The control group was treated with carbamazepine.
Li 2000Confirmed not a true RCT.
Li 2002Confirmed not truly randomised by contact with study author.
Liao 2005PHN was defined as pain persisting one week after total decrustation.
Lin 2002This trial exceeded our seven days to treatment criterion.
Lin 2005Confirmed not truly randomised by contact with study author.
Liu 2003PHN was defined as pain persisting at the site of shingles two weeks after the onset of acute rash.
Liu 2005PHN was defined as pain persisting at the site of shingles two weeks after the onset of acute rash.
Ma 2000Confirmed not a true RCT.
Ma 2002Confirmed not truly randomised by contacting study author.
Shi 2008Confirmed not a true RCT.
Song 2009Confirmed not a true RCT.
Tang 2004Confirmed not truly randomised by contacting study author.
Wang 2004This trial exceeded our seven days to treatment criterion.
Xu 1999This trial exceeded our seven days to treatment criterion.
Yang 2000The routine treatments were mismatched between two groups.
Yang 2002PHN was defined as pain persisting at the site of shingles two weeks after the onset of acute rash.
Yang 2010Confirmed not a true RCT, and PHN was not clearly defined. The duration of herpes of some participants was longer than seven days after the rash onset.
Yin 2004The routine treatments were mismatched between two groups.
Yin 2005The routine treatments were mismatched between two groups.
Zeng 2011Confirmed not a true RCT.
Zhang 2003Confirmed not truly randomised by contacting study author.
Zhang 2004The routine treatments were mismatched between two groups.
Zhang 2005This trial exceeded our seven days to treatment criterion.
Zheng 2004The routine treatments were mismatched between two groups.
Zhou 2000Confirmed not truly randomised by contacting study author.
Zhou 2008Not a true RCT.

Characteristics of studies awaiting assessment [ordered by study ID]

Hu 2001

MethodsSingle centre, randomised controlled parallel trial.
Participants45 patients with confirmed herpes zoster and without contraindications to glucocorticosteroids were enrolled. They were randomly assigned to the aciclovir plus dexamethasone group (20 participants) or the aciclovir alone group (25 participants).
InterventionsIn the aciclovir plus dexamethasone group, participants were given 0.25 g aciclovir intravenously every 8 hours for a total of 10 days, and 5 mg dexamethasone intravenously daily. In the control group, aciclovir alone was administered in the same way as with the other group.
OutcomesTime to cessation of new herpes eruption, time to crusting, and the presence of PHN at one month after the rash onset.
NotesThe exact course of disease from onset of herpes zoster to receipt of treatment was not specified. We tried to contact the author, but have not yet received a reply.

Ancillary