1. Varicella-Zoster Virus

Varicella-zoster virus (VZV) is a human α-herpesvirus in a group with herpes simplex virus type 1 (herpes labialis) and herpes simplex virus type 2 (herpes genitalis). All three have an affinity for nerve tissue and cause primary infection followed by latency and delayed reactivation resulting in secondary disease. VZV is markedly efficient at damaging cells that it has infected.

VZV causes varicella (chicken pox) as a primary infection. In temperate climate regions, this most often affects children and, in the USA, approximately 95% of individuals are seropositive by the age of 20 years.[1,2] In tropical areas, varicella more often affects adolescents and young adults and, typically, <50% are seropositive by adulthood, especially in rural areas.[3,4]

Following varicella, virus from skin lesions migrates within primary afferent nerve tissue to spinal ganglia and cranial sensory nerves where it persists in a clinically latent form. During latency, it is believed that subclinical viral activity may stimulate enhancement of VZV-specific cell mediated immunity (CMI) — endogenous boosting — which delays clinical reactivation of the virus.[5] Inability of VZV-specific CMI to contain latency results in viral replication and can lead to clinical herpes zoster (HZ).

2. Herpes Zoster (HZ)

2.1 Development of HZ

In HZ, clinical reactivation and spread of VZV results in inflammatory, viral and immune-response damage to the ganglion, the primary afferent nerve and skin. Prodromal pain or itching often occurs over several days and is followed by a unilateral dermatomal vesicular rash, usually accompanied by pain, which affects the thoracic dermatomes in about half of cases. The second most commonly affected location is the trigeminal nerve, particularly the ophthalmic division. The rash usually heals in 2–4 weeks but often scarring and pigmentation changes persist. HZ lesions are infectious for several days (fewer when antiviral drugs are used) and can result in varicella in VZV-naive (seronegative) contacts. In two studies, clinical HZ diagnosis was shown to be incorrect in up to 20% of cases, with herpes simplex being the most common condition mistaken for HZ.[6,7] During the eruptive phase, acute local neuritic pain occurs in up to 90% of immunocompetent individuals and is more common with increasing age. The acute pain of HZ, which may be constant or intermittent, is variously described as sharp, stabbing, shooting, burning, throbbing, tender, boring, itching and/or hot and may be accompanied by allodynia (pain experienced in response to a normally innocuous stimulus such as light contact with clothing or a cool breeze). The disordered primary afferent (pain transmitting) nerve pathway and the associated segment of the spinal cord may undergo complex pathophysiological changes resulting in chronic neuropathic pain known as postherpetic neuralgia (PHN) [see section 3].[8]

2.2 Who Gets HZ?

Two-thirds of individuals developing HZ are ≥60 years of age whereas the mean age of the population in the developed world is approximately 35 years.[9–11] The vast majority of cases of PHN also occur within this older group of patients. HZ also affects individuals from infancy and throughout life, but with a lower incidence and usually lesser severity. People who are immunocompromised as a result of malignancies such as lymphoma, chemotherapy or radiotherapy, HIV and use of therapeutic immunosuppressant drugs (following organ transplant or for autoimmune disease), including corticosteroids, are at significantly increased risk relative to healthy immunocompetent people of similar age. Diabetes mellitus may also facilitate increased incidence. A recent study of 22 294 cases of HZ in Israel and 88 895 control subjects showed that diabetes was associated with an increased risk of HZ (odds ratio 1.53; 95% CI 1.44, 1.62).[12] It is believed that the increased incidence of HZ in the elderly results from progressive decline in VZV-specific CMI with advancing age, i.e. immunosenescence. VZV-specific CMI appears to be greatly enhanced following HZ as second attacks are rare (probably <5%) and often occur several years after the first attack.[9,13–16]

2.3 Epidemiology of HZ

Any person who has had varicella (seropositive for VZV) is at risk of developing HZ when VZV-specific CMI declines below a threshold level (figure 1).[5] In immunocompetent individuals, increasing age is the main risk factor for developing HZ. In the USA, Canada and Europe, the overall annual incidence of HZ is estimated to be 3–4 per 1000 population.[10,11,14,17–20] The annual risk of HZ begins to increase markedly at around 50 years of age, rising to >10/1000 person-years among persons >75 years of age. According to Hope-Simpson’s seminal papers on HZ epidemiology in Cirencester, UK, the overall annual incidence of HZ was 3.4/1000 population per year, increasing from 2.4/1000 person-years at 40–49 years to 5.6/1000 at 50–59 years, 6.8/1000 at 60–69 years, 7.2/1000 at 70–79 years and 11.0/1000 in individuals >80 years.[9,17] Similar trends, although with slightly higher age-specific incidence rates, were found in a recent population-based study conducted in Olmsted County, Minnesota, USA.[10]

Fig. 1
figure 1

Lifetime changes in cell-mediated immunity to varicella-zoster virus (VZV). Varicella is the primary infection caused by VZV, and its resolution is associated with the induction of memory T cells specific for VZV (black line). Memory immunity to VZV may be boosted periodically by exposure to varicella or silent reactivation from latency (grey peaks). Memory T cells specific for VZV decline with age. Decline below a threshold (dashed black line) correlates with an increased risk of herpes zoster (HZ). Occurrence of HZ, in turn, is associated with an increase in T cells specific for VZV. Administering zoster vaccine to older people may prevent T cells specific for VZV from dropping below the threshold for HZ occurrence (dashed grey line) [reproduced from Arvin,[5] with permission. Copyright© 2005 Massachusetts Medical Society. All rights reserved].

Full size image

Many studies have shown lower incidence of HZ in males than in females.[18,21] However, this is not universal: in the Shingles Prevention Study, the incidence of HZ in the placebo group was similar in men and women (10.7 and 11.8/1000 person-years, respectively).[6] It may be that differing data-collection methods between studies and a lower tendency for males to seek health advice explain most of these discrepancies. Based on current incidence rate estimates, approximately 1 million cases of HZ occur every year in the USA (300 million population), 1.7 million in the European Union (460 million population), 100 000 in Canada (32 million population) and 80 000 in Australia and New Zealand (24 million population). The number of HZ cases is expected to increase with the aging of the population and the increased prevalence of immunocompromised individuals. It has also been suggested that widespread varicella vaccination, by decreasing the circulation of VZV in the population, may reduce the opportunity for exogenous boosting of VZV-immunity by exposure to natural varicella, which may result in further HZ incidence increase in older adults over the next few decades.[22,23] However, a longer term decline in HZ incidence is probable as the pool of individuals infected by wild-type (natural) VZV is reduced by varicella vaccination. The current lifetime risk of developing HZ is estimated to be 20–30%.[10,24]

HZ usually occurs only once in an individual’s lifetime but immunocompetent subjects may experience a second episode or more of HZ. The rate of recurrence has not been well studied and is quoted in the literature to vary from 1.5% to 12.5% of HZ patients, typically close to 4–5%.[9,13–16] However, recurrence of HZ is frequent in immunocompromised individuals.

2.4 Complications of HZ

The most common complication of HZ is PHN, the result of VZV-induced pathophysiological changes in the primary afferent (pain transmitting) nerve pathway and the associated segment of the spinal cord leading to an ongoing neuropathic pain syndrome (see section 3). Less common but serious complications of HZ include secondary bacterial infection, ophthalmic damage, motor paresis, cerebral angiitis, Guillain-Barré syndrome and visceral dissemination. Serious complications of HZ are more common in immunocompromised patients. The reductions in quality of life and activities of daily living resulting from HZ and PHN are substantial (see section 4).

3. Postherpetic Neuralgia (PHN)

PHN is a neuropathic pain syndrome resulting from a combination of inflammatory and viral damage to primary afferent fibres of sensory nerves and the corresponding levels of the spinal cord, as well as peripheral and central sensitization.

While there is no agreed definition of what constitutes PHN, current research tends to define the condition as pain persisting 3–4 months after zoster rash onset. In addition, some studies require that pain should be ‘significant’; for example, scoring ≥3 on a 0–10 scale, where 0 = no pain and 10 = worst imaginable pain.[6] The majority of patients with PHN describe characteristic patterns of pain, often consisting of at least two of the following: (i) spontaneous, constant, deep burning, throbbing, aching pain; (ii) intermittent sharp, stabbing, shooting, lancinating pain, which may also be spontaneous; and (iii) evoked allodynia that usually lasts well beyond the duration of the stimulus (hyperpathia). Allodynia, which is present in at least 70% of PHN patients, is often described as the most distressing and debilitating component of PHN. Identified risk factors for PHN include advancing age, greater acute pain, severe rash, prodromal pain, ophthalmic location and possibly female sex.[25]

The prevalence of PHN is estimated to range from 500 000 to 1 million cases in the USA and from 100 000 to 200 000 in the UK.[9,10] The proportion of HZ patients reported to develop PHN varies across studies depending on the definition used and the age distribution of the population studied. The Shingles Prevention Study used a definition based on a pre-specified level of pain (≥3 on a 0–10 scale in which 0 = no pain and 10 = worst possible pain) measured using the Zoster Brief Pain Inventory (ZBPI), a validated pain instrument.[6] The proportion of HZ patients with pain score ≥3 in the placebo group was 31% (196/642) at 30 days, 18% (113/642) at 60 days and 12% (80/642) at 90 days. Most patients with HZ in this study had received antiviral medications during the acute phase of HZ; thus, the incidence of PHN may have been higher in an untreated population. The frequency of PHN increases with age.[9]

4. Impact of HZ and PHN on Patients and Healthcare Costs

Life expectancy has risen considerably over the decades and is projected to rise further. Although medicine and social change have added ‘years to life’ they have so far been less successful in adding ‘life to years’ — very elderly people are often not gaining quality years of additional life. Also, in the elderly, a short period of disease-induced debility, e.g. shingles, often leads to prolonged loss of ability for activities of daily living and independence.[26]

HZ results in significant suffering and cost, both to patients and healthcare providers. A recent study of 27 225 patients aged ≥50 years in the UK reported a mean direct cost of £103 per HZ patient and £397 per PHN episode (3-month definition) [2006 values].[21] Both HZ and PHN costs increased markedly with pain severity. A pilot study in East London, UK, showed that mean costs (2003 values) for HZ management during the first 6 months were similar in younger (age <65 years, £526.30) and older (age ≥65 years, £519.90) patients, but with medical costs being highest in older subjects and societal costs higher in younger patients.[27] Hospitalization was a major contributing factor in the elderly.

Recent data indicate that the annual rate of hospitalization with a primary diagnosis of HZ ranges from 4 to 10/100 000 population.[11,28–31] Morbidity associated with HZ leads to approximately 2–3% of HZ cases being hospitalized, including up to 10% of HZ patients aged >65 years. Approximately 70–80% of HZ-associated hospitalizations occur in individuals aged ≥50 years. The need for hospitalization is not limited to immunocompromised patients as 70–80% of patients hospitalized with HZ are immunocompetent. With advanced age, HZ may be life-threatening. The annual mortality due to HZ is estimated to be 0.6–1 per million population, with almost all deaths occurring in patients aged >65 years.[11,28–31]

5. Management of HZ and PHN

5.1 Treatment of HZ

The aims of treatment of HZ are to relieve acute symptoms and prevent complications such as secondary bacterial infection, PHN and, in the case of zoster ophthalmicus, damage to the eye.

There is abundant evidence that antiviral drugs (aciclovir, valaciclovir, famciclovir) administered within 72 hours of rash appearance are significantly more effective than placebo in controlling acute pain and hastening rash healing.[32] Brivudin is also effective in this respect but greatly enhances the toxicity of fluorouracil with potentially fatal results[33] and should not be used in immunocompromised patients.[34] Although evidence for benefit of commencing administration of antiviral drugs after 72 hours is lacking and despite licensed indications, expert consensus opinion advises that it is appropriate to commence treatment beyond this point in cases of ophthalmic HZ and severe HZ where there is clinical evidence of ongoing viral activity.[35] The more bioavailable drugs valaciclovir, famciclovir and brivudin have a more favourable dosage and administration schedule than aciclovir (table I) and have been shown to be more effective.[36] However, in severely immunocompromised patients, intravenous aciclovir is the treatment of choice and has been shown to accelerate disease resolution and reduce the risk of VZV dissemination.[35] Oral antiviral drugs may be considered in less severely immunocompromised patients. Topical antiviral drugs have no place in the management of HZ except in addition to systemic administration in some cases of zoster ophthalmicus. Addition of analgesics is usually necessary and paracetamol (acetaminophen) with or without weak opioid is commonly used. Application of cold packs and avoidance of irritant clothing may provide additional relief.[35] If pain relief remains inadequate, strong opioids, tricyclic antidepressants (TCAs) or α2δ-ligand antiepileptics (gabapentin, pregabalin) should be prescribed.[35]

Table I
figure Tab1

Guidance for use of oral antiviral drugs in herpes zoster

Full size image

5.2 Prevention of PHN

Prevention of PHN has been the subject of many studies. Contending treatments are antiviral drugs with or without concomitant oral corticosteroids, early use of neuropathic pain active drugs and nerve blocks of various types.

Antiviral drugs significantly reduce the duration of pain from the time of rash onset at all time intervals up to 6 months. Beyond 90 days after rash onset, such pain would normally be termed PHN.[32,37,38] It is unknown whether antivirals given within 72 hours of rash onset reduce the risk of PHN persisting for ≥1 year. Addition of oral corticosteroids enhances the short-term benefit of antiviral drugs but does not add any protection against PHN.[39,40] It is important to note that the diagnosis of HZ is often established several days after viral replication (and, therefore, pathological damage) has commenced. Indeed, during the prodromal period, it is often difficult to diagnose HZ and investigation for other conditions such as myocardial infarction and cholecystitis may be initiated. Once the rash appears, it may take another day or two before medical consultation is achieved. Thus, by the time treatment is instituted, the virus has been active within nerve tissue for several days. This probably explains the relative lack of efficacy of antiviral drugs and corticosteroids in preventing the development of PHN. Importantly, corticosteroids may also increase the risk of adverse events.[39]

A small placebo-controlled study indicated that amitriptyline administered during the acute stage of HZ could significantly reduce the risk of PHN.[41] However, no large-scale study of use of drugs such as gabapentin or pregabalin, strong opioids or TCAs for prevention of PHN has been conducted. A single epidural injection of local anaesthetic and corticosteroid reduces acute pain but has no protective effect against PHN.[42] Studies of prolonged nerve blockade (autonomic or systemic, with or without corticosteroid) suggest that this approach may reduce the duration of pain.[43,44] However, in most situations, this type of prolonged treatment carries additional risk and is not practical within the healthcare system.[45]

5.3 Treatment of PHN

Systematic review of data from well designed trials of therapy for PHN gives a clear indication of those treatments that are appropriate to offer patients (table II).[46–49] However, such trials have been undertaken over many years, are of differing design, involve intense observation of drug effects and patient behaviour by the researchers and are typically of limited duration. For some treatment modalities, very few patients have been studied. Furthermore, many PHN patients have co-morbidities and often take medication for them. Most of the drug groups investigated may lead to significant adverse effects, particularly in the elderly.

Table II
figure Tab2

Guidance for use of drugs in postherpetic neuralgia (reproduced from Johnson,[50] with permission)

Full size image

Management of PHN remains difficult despite the increasing number of treatment options, probably because single treatments do not cover all the multiple pain mechanisms involved in PHN. Furthermore, few studies make direct comparisons of one drug with another to inform decisions about the most effective drugs for pain relief with the most tolerable adverse effect profile. Determination of numbers needed to treat (NNT) has been, despite shortcomings, a useful measure that allows indirect comparisons between therapies. NNT describes the number of patients that need to be treated with a given drug to observe one patient with a defined degree of pain relief relative to placebo. Usually an NNT for ≥50% pain relief is used because this is considered to indicate a useful clinical effect. NNTs for ≥50% pain relief, and the number of patient episodes for each treatment investigated, are shown in table III and figure 2.

Table III
figure Tab3

Number of patients needed to treat (NNT) for a defined benefita and number of patients needed to cause harm (NNH)b estimated for effective drug treatments of postherpetic neuralgia[46] [reproduced from Johnson,[50] with permission].

Full size image
Fig. 2
figure 2

Numbers needed to treat (NNT) for drugs against postherpetic neuralgia (PHN), defined as the number of patients needed to treat to observe one patient with ≥50% pain relief. The NNT for each drug is <5, providing significant evidence for analgesic efficacy. The widths of the bars correlate with the number of patient episodes that are accounted for each drug, i.e. the thicker the bars, the more patient episodes that were included, which increases the significance of the findings. *NNTs for topical lidocaine (lignocaine) ranged from 2 to 4.4 in differently designed studies that included neuropathic pain patients with and without PHN, and therefore these data need to be interpreted with care (reproduced from Johnson et al.,[26] with permission from Future Medicine Ltd).

Full size image

5.3.1 Antidepressants

TCAs inhibit reuptake of monoaminergic transmitters, thus potentiating their effects in CNS pain-modulating pathways. They also block voltage dependent sodium channels. PHN has been shown to respond to TCAs such as amitriptyline, desipramine and nortriptyline. Indeed, pooled data from four placebo-controlled trials reporting on 248 patient episodes showed significant benefit associated with TCAs, with an average NNT of 2.6.[46,51–53]

Three studies comparing the efficacy of different antidepressants in PHN showed that (a) amitriptyline led to considerable pain reduction in 47% of patients and maprotiline in 38% of patients, without significant differences between the two treatments;[54] (b) amitriptyline and nortriptyline had a similar effect in about 50% of patients;[55] and (c) the efficacy of amitriptyline was not influenced by combining it with fluphenazine, which itself was no better than placebo.[56] In a comparison of amitriptyline, desipramine and fluoxetine in antidepressant-naive PHN patients, all three drugs resulted in pain reduction, with desipramine and amitriptyline showing superior efficacy and tolerability compared with fluoxetine.[57]

TCAs have significant adverse effects which limit their use, especially in the elderly. They can produce dry mouth, sedation, memory loss, orthostatic hypotension, urinary retention, exacerbation of glaucoma, constipation and cardiac conduction abnormalities. High doses of TCAs are associated with a risk of sudden cardiac death.[58] Doses required for neuropathic pain management are usually low or moderate but caution is necessary. Relatively selective noradrenaline reuptake inhibitors, such as desipramine and nortriptyline, have fewer anticholinergic adverse effects than the nonselective serotonin-noradrenaline reuptake inhibitor amitriptyline, and may be better tolerated.[55] Slow dose titration to a minimum effective dose is recommended to reduce the impact of antidepressant adverse effects.[26,46]

5.3.2 Antiepileptics

Gabapentin and pregabalin act on the α2δ-subunit of calcium channels located on the spinal presynaptic terminals of primary afferent nociceptive neurons.

Two placebo-controlled trials of gabapentin accounting for 559 patient episodes showed that gabapentin is effective for PHN with an average NNT of 4.4.[46,59,60] Daily drug doses of gabapentin ranged from 1200–3600 mg, titrated over 1–2 weeks. Two controlled studies (411 patient episodes) of pregabalin showed efficacy with a pooled NNT of 4.9.[61,62] The efficacy of pregabalin was confirmed in a controlled study of 338 patients with neuropathic pain, including 178 PHN patients, who were treated with either flexible or fixed doses of pregabalin or placebo (NNT for the entire group was 3.8).[63] Daily doses of pregabalin in the three studies ranged from 150 mg to 600 mg. A significant advantage of pregabalin over gabapentin is its superior bioavailability, making it easier to establish without the need for long titration periods. Both drugs reduced sleep disturbance and improved overall mood and other measures of quality of life in patients with neuropathic pain.[60,61]

Dizziness, somnolence, ataxia and peripheral oedema are the most commonly reported adverse events with gabapentin and pregabalin, especially during upward titration to target doses.[59–62] Both drugs have low potential for drug-drug interactions and no negative impact on cardiac function, properties that favour their use, especially in the elderly, despite the fact that they may add more to the cost of treatment than some TCAs. It is unwise to discontinue these drugs abruptly as a case study has reported encephalopathic changes following sudden withdrawal of pregabalin.[64]

Sodium channel-blocking antiepileptics such as carbamazepine, while effective for some other neuropathic pain conditions, have no role in the treatment of PHN.[65]

5.3.3 Tramadol and Strong Opioid Analgesics

One of the major metabolites of tramadol, a norepinephrine and serotonin reuptake inhibitor, is an opioid μ receptor agonist. Efficacy for oral tramadol has been demonstrated in a controlled study of patients with PHN (108 patient episodes).[66] The NNT was 4.8, with an average daily titration dose of 275 mg. The study showed a very high placebo responder rate.

Use of opioid analgesics for patients with chronic neuropathic pain is controversial but positive results of trials of oral strong opioid analgesics in various neuropathic pain conditions have been reported.[67] Double-blind, placebo-controlled studies have demonstrated that acute infusions of morphine or fentanyl provide significant pain relief in patients with PHN.[68] Oral oxycodone was effective in a controlled crossover study of 38 PHN patients with an NNT of 2.5; the daily drug dose ranged from 20 mg to 60 mg.[69] A three-period crossover study compared treatment with oral morphine (or methadone, if morphine was not tolerated), nortriptyline (or desipramine) and placebo in 50 patients with PHN.[51] Mean daily maintenance doses were morphine 91 mg or methadone 15 mg, and nortriptyline 89 mg or desipramine 63 mg. Controlled-release morphine provided statistically significant benefits relative to placebo for pain and sleep, but not for physical function and mood. The efficacy of opioids and TCAs for pain relief was similar (NNT opioids 2.8, NNT TCAs 3.7). There was no correlation in response rate between drugs, indicating that different mechanisms are active in these PHN patients.

Patients with neuropathic pain can be successfully and safely treated with opioids on a long-term basis using stable doses but their use requires caution in patients with a history of chemical dependence or pulmonary disease. Following dose titration with short-acting drugs, conversion to long-acting opioid analgesics (e.g. sustained release morphine or oxycodone preparation) is desirable. Many physicians prefer to reserve opioid treatment for patients with unsatisfactory responses to appropriate trials of TCAs and antiepileptics. Prophylactic treatment of common adverse effects (e.g. nausea, dry mouth and constipation) is necessary and improves patient compliance. Other common adverse effects include dizziness, sedation and pruritus. Psychotic symptoms may also occur.

Withdrawal of opioids should occur by gradual reduction in dosage. Drugs for an individual patient should be prescribed by one physician only. Guidelines for the use of opioids in chronic non-malignant pain provide important information on their safe and effective use.[70,71]

5.3.4 Topically Acting Agents

Topical Capsaicin

Capsaicin is an agonist at transient receptor potential vanilloid 1 (TRPV1), which is present on the terminals of primary nociceptive afferents.

Pooled data for 175 patient episodes of PHN from two placebo-controlled studies demonstrated the efficacy of 0.075% capsaicin cream applied three to four times daily within the painful area with an NNT of 3.3.[72,73] On initial application, capsaicin has an excitatory action and produces burning pain and hyperalgesia, making study-blinding and patient compliance difficult. However, with repeated or prolonged application, capsaicin inactivates the receptive terminals of nociceptors. Thus, in the experience of many practitioners, many patients reject this treatment because of the transient burning sensation and the inconvenience of three or four times daily application but gain benefit if encouraged to persevere. This therapy is reasonable in patients whose pain is maintained by anatomically intact sensitized nociceptors. More recently, pilot studies of higher concentration capsaicin (8%) following application of topical local anaesthetic suggest that a single application may produce prolonged (3 months) reduction in the intensity of PHN in some patients.[74] Use of topical capsaicin is not associated with any systemic adverse effects or drug interactions.

Topical Lidocaine (Lignocaine)

Local anaesthetics block voltage-dependent sodium channels. Although the site of action of membrane-stabilizing drugs for relief of pain has not been proven in patients, in vitro studies have shown that ectopic impulses generated by damaged primary afferent nociceptors are abolished by concentrations of local anaesthetics much lower than that required to block normal axonal conduction.

Studies report pain relief with topically applied special formulations of local anaesthetic (5% lidocaine [lignocaine] patches) in PHN and other neuropathies.[75,76] Randomized controlled trials of varying design indicate an NNT of between 2 and 4.4. Use of up to four patches (5% lidocaine, 700 mg/patch, patch size 10 × 14 cm) for 12 hours/day is recommended.[75–78] Blood levels of lidocaine after topical application are considerably below those required for an antiarrhythmic effect and, therefore, systemic effects of the drug are avoided. Minor local adverse effects, such as rash and pruritus, are occasionally associated with application of the patch itself. Lidocaine patch therapy is a safe and well tolerated supplemental treatment for PHN.

5.3.5 Miscellaneous Treatments

Intrathecal administration of lidocaine and methylprednisolone combined appears to be associated with remarkable benefit in PHN patients.[79] However, intrathecal methylprednisolone has potentially dangerous short- and long-term adverse effects and the trial that reported this result has not yet been replicated. Further high-quality controlled trials of this therapy are necessary before its use can be recommended.

Based on knowledge of pathophysiological pain mechanisms, NMDA receptor antagonists might be pain-relieving candidates. However, data from three controlled studies (131 patient episodes) of dextromethorphan did not demonstrate a superior efficacy for this agent over placebo.[80]

Antiepileptics such as lamotrigine, the selective serotonin-noradrenaline reuptake inhibitors venlafaxine and duloxetine, and cannabinoids might also be considered, as they have been effective in neuropathic pain conditions other than PHN.[48]

Transcutaneous electrical nerve stimulation, which has minimal adverse effects, may be effective in some cases. In severely affected PHN patients who have gained little benefit from other treatment, spinal cord stimulation might also be considered as an option. However, evidence for benefit from this invasive treatment option is limited.

Supportive therapy, including consideration of psychosocial factors, is essential for successful management of chronic pain disorders, including PHN.[81,82]

5.3.6 Treatment Guidelines

Drug management of PHN consists of three main classes of oral medication (serotonin/norepinephrine modulating antidepressants, calcium-modulator antiepileptics, tramadol and opioids) and two categories of topical medications (capsaicin and lidocaine). In table III, the NNT and number needed to harm (NNH) values of drugs for treatment of PHN are shown in relation to the number of patients who have been treated with each drug to indicate the validity of the different trials. The NNH indicates the number of patients who need to be treated for one patient to drop out of a study because of adverse effects. It has been suggested that an NNT <5 provides significant evidence of useful clinical analgesic efficacy.[46] The precise NNH at which use of a drug becomes unsatisfactory will depend upon physician experience, the requirement for treatment as well as the individual patient. Choice of medication for the individual patient requires clinical judgement.

Since more than one mechanism of PHN is involved in most patients, a combination of two or more analgesic agents to cover such mechanisms will often produce greater pain relief than monotherapy, and will also produce fewer adverse effects, since lower doses of each agent may be utilized. Therefore, early use of combinations of two or three compounds from different classes may be more appropriate, rather than a stepwise strategy with successive monotherapies. A recent controlled four-period crossover trial in patients with PHN and painful diabetic neuropathy showed that gabapentin and morphine combined achieved better analgesia at lower doses of each drug than either given as a single agent.[83] However, the advantages of using drug combinations must be weighed against the possibility of severe complications resulting from interactions between medications. A potentially life-threatening interaction can occur between drugs such as TCAs in combination with tramadol, leading to an additive increase in central serotonin concentration. Tramadol, in addition to its analgesic effects, acts also as a monoamine reuptake inhibitor. Induction of a serotonin syndrome, which is clinically characterized by visual hallucinations, muscle rigidity, myoclonus or hypotension, needs to be considered when tramadol or other phenylpiperidine series opioids, such as pethidine, methadone and dextromethorphan, are combined with selective serotonin reuptake inhibitors or monoamine oxidase inhibitors, which the patient may be receiving as therapy for depression.[84]

Medication-related adverse effects are important factors not only with respect to drug safety but also in terms of their effects on patient compliance. Recent meta-analyses have considered this by calculating the NNH. As noted above, the NNH indicates the number of patients that need to be treated for one patient to drop out of a study because of adverse effects. NNHs were all considered to be acceptable in that they indicated that the drugs may safely be given to most patients but that individual patients may experience adverse effects for the drugs listed in table III. The significance of NNH values for clinical assessment is limited, partly because drug trial protocols often exclude patients with defined contraindications from clinical studies. In the review by Hempenstall et al.,[46] the NNH for minor events (i.e. reported adverse effects that did not lead to withdrawal from the trial) for TCAs (combining trials) was 5.67 (3.34, 18.58) and the NNH for major events (i.e. reported adverse effects that led to withdrawal of the study drug) was 16.9 (8.85, 178). The wide confidence intervals reflect the relatively small numbers of incidents reported. Most major events were sedation and other anticholinergic effects but one patient receiving desipramine developed left bundle branch block and another developed erythema multiforme. For gabapentin (combining trials), the NNH for minor harm was 4.07 (3.15, 5.74) and for major harm was 12.25 (7.69, 30.2). For pregabalin, the NNH for minor harm was 4.27 (2.78, 9.18). The NNH for major harm could not be calculated because of heterogeneity of data. For opioids, the NNH (combining trials) was 3.57 (2.16, 10.23) for minor harm and 6.29 (4.16, 12.8) for major harm.

An algorithm for the prevention (see also section 6) and treatment of PHN is provided in figure 3.

Fig. 3
figure 3

Algorithm for prevention and treatment of postherpetic neuralgia. Scientific evidence for vaccination in children and early neuropathic pain treatment is scanty. The combination of gabapentin and morphine (designated small square connecting both classes of medication) has been shown to be more effective than either drug alone.[83] The arrows indicate the potential advantage of early drug combination, with a preference for antiepileptic and opioid combined based on recent study results (small square connecting both classes of medication) [reproduced from Johnson et al.,[26] with permission from Future Medicine Ltd].

Full size image

6. Prevention of Varicella and HZ

In 1974, Takahashi successfully attenuated VZV obtained from a child with varicella whose family name was Oka. Subsequently, the Oka strain has been used in development of both vaccines to prevent varicella and, more recently, for attenuation and prevention of HZ and PHN.

Vaccination of young children is effective in preventing varicella and has become routine in the US since 1995. Disease incidence, varicella-related hospitalizations and deaths have all been drastically reduced as a result.[85] Surveillance of vaccinated subjects for subsequent HZ (caused by either wild- or vaccine-type virus) has to date not identified any increase in incidence of HZ in vaccinated individuals.[86] Based on observations of immunocompromised children, it is even likely that varicella vaccination will decrease the incidence of HZ in vaccinated people.[87,88] Ultimately, the proportion of seropositive individuals in the community with latent wild-type VZV will decline and HZ is likely to become less common. There is a theoretical possibility that reduced incidence of varicella amongst children may deprive seropositive adults of protective CMI enhancement (exogenous boosting) resulting from contact with infected children. This could result in an increase in HZ incidence among older adults. To date, there is no evidence that this is the case.

Stimulation of CMI by vaccines declines with increasing age. A vaccine using the same VZV Oka strain of the virus, but several-fold stronger, has been developed and studied for the prevention of HZ and PHN. The Shingles Prevention Study investigated 38 546 immunocompetent subjects aged ≥60 years who were randomly assigned to receive active vaccine or placebo vaccine.[6] Prolonged surveillance was undertaken to detect cases of HZ, those who developed PHN and the pain burden of disease resulting from HZ. In addition, minor and serious adverse events were studied and laboratory studies of immune responses undertaken. In all cases, diagnosis of suspected cases of HZ was rigorously confirmed. Patient retention in the study was >95%. HZ vaccine reduced HZ pain burden of illness (the area under the curve of daily worst pain over 6 months) by 61.1% (95% CI 51.1, 69.1; p < 0.001) and reduced the incidence of PHN by 66.5% (95% CI 47.5, 79.2; p < 0.0001) [table IV]. The incidence of HZ was also reduced by 51.3% (95% CI 44.2, 57.6; p < 0.001). Although the vaccine was less effective in reducing the incidence of HZ in subjects aged ≥70 years, protective effects against PHN and pain burden of illness were maintained, suggesting attenuation of HZ severity even when HZ was not prevented. The vaccine neither caused nor induced HZ. The vaccine was well tolerated and injection site reactions were generally mild.

Table IV
figure Tab4

Vaccine efficacy for herpes zoster: effects on pain burden of illness (BOI), incidence of postherpetic neuralgia (PHN) and incidence of herpes zoster (HZ)[6]

Full size image

In May 2008, the US Centers for Disease Control and Prevention recommended that (immunocompetent) adults aged ≥60 years should be vaccinated against HZ with one dose for all adults, including those who have previously had shingles. Of note, zoster vaccine is indicated for the prevention of HZ and PHN, not as a therapeutic intervention for existing cases.

7. Public Knowledge of HZ and PHN

There may be a need to increase disease awareness as, if knowledge of HZ is poor, those developing early symptoms of the disease may not feel a need to seek urgent medical attention, thereby reducing the opportunity for antiviral drug use during the very early days after rash appearance. In addition, if older adults are to request or accept vaccination against HZ, it is also important that they are aware that they are at risk of developing the condition and that the disease and its complications are of sufficient severity to make prevention desirable.

A survey undertaken on behalf of the International Federation on Ageing questioned 1808 adults aged ≥55 years in six countries about their knowledge of shingles.[89] Most had heard of the condition but few believed that they were at risk of developing HZ or that HZ was associated with ongoing pain. A national survey of primary care physicians in the US showed that most perceived a high level of burden from HZ and PHN and generally favoured HZ vaccination while recognizing barriers to its acceptance.[90] The greatest barrier appeared to be related to cost issues although health economic studies indicate that use of the vaccine is cost effective by currently accepted standards.

8. Future Directions

Changes in longevity, the size of the immunocompromised community and the effects of childhood varicella vaccination and adult zoster vaccination programmes will determine the extent to which HZ and PHN will affect individuals and healthcare systems. It is as yet unclear whether childhood varicella vaccination will result in a medium-term increase in HZ incidence and whether adult zoster vaccination will produce long-lasting immunity for its recipients. Ongoing epidemiological surveillance of HZ in populations with and without childhood varicella vaccination and/or zoster vaccination programmes should assist estimation of the ongoing burden of HZ and PHN. Improved patient and physician awareness of VZV-related disease will facilitate optimal use of available prophylactic and therapeutic options, such as vaccines, antiviral drugs and analgesics. In addition, scientific advances may provide more effective antiviral drugs, possibly helicase primase inhibitors, and make truly mechanism-based pharmacological management of PHN a reality. Anticipated research may confirm that early use of drugs, such as opioids and α2δ-ligands, can reduce the incidence of PHN.

9. Conclusion

HZ and PHN are most common in the elderly. Although antiviral drugs are highly beneficial in controlling the acute symptoms of HZ and reducing the duration of pain, they do not prevent PHN occurrence in all patients. PHN remains difficult to manage despite improved understanding of its pathology and drugs in neuropathic pain. Many patients experience poor pain relief and significant adverse effects from PHN medication. It is possible that newer antiviral drugs, such as helicase primase inhibitors, will achieve greater protection against PHN and that newer therapies for neuropathic pain will be developed. In the long term, HZ is likely to become less common subsequent to widespread introduction of varicella vaccination. In the meantime, vaccination of older adults against HZ and PHN seems an effective, cost-effective and safe strategy. Long-term surveillance of both varicella and HZ is essential, and strategies to reduce the burden of HZ are extremely important in immunocompetent and immunocompromised individuals.