Varicella is a common contagious infection in childhood with increasing incidence in adults. Pneumonia, although rare, is the most serious complication that commonly affects adults. Over the last two decades there have been major advances in the understanding of Varicella infections, management and prevention. This review discusses the epidemiology, pathogenesis, pulmonary manifestation, morbidity, long-term clinical consequences and current state of management of Varicella pneumonia in adults. Prevention and other disease-modifying therapy are also discussed.
Varicella (chickenpox) is a common infection of childhood typically affecting children aged 2–8 yrs and usually follows benign outcome. However, data from Europe and North America have shown that the incidence of chickenpox in adults has doubled in recent years and this has been paralleled with an increase in hospital admissions 1–5 and mortality 6. The reason for this age shift is not known, though it may be the result of less exposure to the Varicella-Zoster virus (VZV), smaller family size or increased virus virulence. Varicella pneumonia is the most common and serious complication of chickenpox infection in adults 7, 8, with a reported incidence in healthy adults that is 25-fold greater than in children 9. In the USA chickenpox is the leading cause of death among vaccine-preventable diseases 10. Currently, the proportion of adults in industrial nations susceptible to chickenpox is ∼7%, evident from negative Varicella-Zoster antibodies 11, 12. Varicella pneumonia is so uncommon that large-scale studies are difficult to conduct and most published studies represent either collections of small case series or retrospective analyses over many years.
This article will review aspects relating to Varicella pneumonia and its natural history in adults.
All relevant published articles relating to the epidemiology, natural history and treatment of Varicella pneumonia since 1964 were identified. The Medline database was first searched using the terms “Varicella pneumonia” and “chickenpox pneumonia”. For the treatment section, the terms “acyclovir” and “immunoglobulin” were searched both as single words and in combination. The effect of acyclovir treatment was assessed on the basis of single outcome, with mortality versus survival as an end-point. Studies investigating risk factors for the development of pneumonia were also combined according to each factor and the odds ratio was calculated. All studies were included, even case reports.
Although most data suggested increasing incidence of Varicella in adults, this was largely in the 1980s and early 1990s and more recent data suggests a decrease in incidence of Varicella in adults 13. The frequency of Varicella pneumonia in chickenpox has been difficult to determine in most populations. However, it is estimated to occur in one in 400 cases of chickenpox infection 14. Recent data from the USA has shown an incidence of pneumonia of 2.3 in 400 cases 5. The incidence of pneumonia in adults has been reported in two studies and varied between 0.32–1.36 cases per 100,000 person years 5, 15. Adults who develop chickenpox are at much greater risk of developing changes on their chest radiograph, with an incidence of 5–50% 3, 15–19. A report of 110 chest radiographs from 114 USA army recruits who developed Varicella during basic training showed that 18 (16.3%) had interstitial and/or nodular infiltrates indicative of pneumonia 19. Pregnant females were also reported to develop this implication in 9% of cases 20. Patients with impaired immune status and patients with chronic lung disease who develop primary Varicella-Zoster infection have an increased risk of developing pneumonia 21–23.
Varicella pneumonia usually presents 1–6 days after the onset of the rash and is associated with tachypnoea, chest tightness, cough, dyspnoea, fever and occasionally with pleuretic chest pain and haemoptysis. However, chest symptoms may start before the appearance of the skin rash 16, 24, 25. Physical findings are often minimal and chest radiographs typically reveal nodular or interstitial pneumonitis 16, 19, as shown in figure 1⇓. The presence of new chest symptoms has been shown to be strongly associated with the documentation of radiological pneumonia 16, 26, 27. With the exception of hypoxia, physical signs are a poor guide of severity 18. The risk of developing respiratory failure requiring artificial ventilation is difficult to predict early in the disease. Radiographic abnormalities were detected in nearly 16% of enlisted military personnel who developed Varicella in an outpatient setting, yet only one-quarter of these had a cough and only 10% of those with radiographic abnormalities developed evidence of tachypnoea 19. Therefore, the data indicate that asymptomatic pneumonitis may exist more commonly than might be supposed.
Table 1⇓ summarises factors shown to be associated with the development of pneumonia, from prospective studies only 16, 28–31. There is a strong correlation between pneumonia and the development of new respiratory symptoms. Previous or current smokers were also at increased risk of developing pneumonia. It is possible that smokers have an enhanced primary viraemia, secondary to the effects of smoking on the nasal mucosa, and this predisposes pneumonia. Furthermore, a previous report has shown that smoking renders human alveolar macrophages more susceptible to infection by herpes viruses 30, which could be relevant pathogenetically, although this requires further study. Increased number of skin spots (>100 spots), i.e. severity of rash, was a factor that increased the risk of developing pneumonia, which may be a reflection of enhanced viraemia. A history of contact with an index case was another factor predisposing to the development of pneumonia. This could be a consequence of these patients having closer contact with the index case and therefore receiving a larger “infecting dose” with an enhanced primary viraemia. Previously, it has been demonstrated that children who get chickenpox from siblings usually have a worse disease with more spots 31. The third trimester of pregnancy was shown in univariate, but not multivariable, analysis to be associated with an increased incidence of pneumonia 28. This is likely to be the result of both immune tolerance during pregnancy and the increased demand and pressure of the baby on the respiratory system. Duration of fever was shown to be independently associated with the incidence of pneumonia (odds ratio (OR) 5.6, 95% confidence interval (CI) 2.4–13.0) in a retrospective setting 32. Immunocompromised hosts and patients with chronic lung disease are also at increased risk of developing pneumonia 21–23. It has been reported that patients with chronic obstructive pulmonary disease develop more severe pulmonary complications in comparison with healthy individuals 33.
The course of Varicella appears to be more aggressive in immunocompromised hosts. In a recent series on human immunodeficiency virus-infected patients from South Africa, seven of 12 patients who were hospitalised with chickenpox developed clinically severe pneumonia and, despite receiving antiviral within 12 h of admission, three (43%) died 21. A review of 38 cases of adult renal allograft recipients with disseminated Varicella infection, reported in 15 different studies, found that of the 29 patients with primary Varicella, 29% developed Varicella pneumonia and the overall mortality was 34%. However, the mortality from 1981–1990 was 53% and from 1990–2000 was 22% 23. The improvement in mortality rate was most likely to have been influenced by the availability of specific antiviral therapy, since antiviral therapy was used in 33% of cases before 1990 and 74% of cases after 1990.
The data available on the pathogenesis of Varicella is limited. Infection with VZV usually occurs by an airborne route, with high infectivity resulting in a secondary attack rate of >90% in households with susceptible individuals 31. Airborne spread in hospitals is also well documented 34. Uncertainty exists as to whether the portal of entry is the conjunctiva, pharynx or lungs. Postinfection primary viraemia starts at ∼96 h, probably following replication in the regional lymph nodes; the use of Varicella-Zoster immune globulin (VZIG) is usually successful in preventing progression to disease if administered before that time 9. The second stage of viral replication takes place in the lymph nodes, lungs, bone marrow, liver, pancreas and adrenal glands, and this mainly takes place in the macrophages. Two studies showed that among the white blood cells, only the monocyte/macrophage will support VZV replication 35, 36. The pulmonary lesions caused by acute Varicella consist of endothelial damage in small blood vessels, with focal haemorrhagic necrosis, mononuclear infiltration of alveolar walls and fibrinous exudates with macrophages in the alveoli, which contain eosinophilic intranuclear inclusions 37. Lung involvement in Varicella infection seems to occur through the bloodstream rather than local extension through the respiratory tree, as shown by the use of monoclonal antibodies 38.
Acyclovir has been associated with successful treatment of Varicella pneumonia; the first case was reported in 1980 39. It has become standard therapy for patients with or at risk of developing complications of Varicella infection, despite the fact that there have been no randomised controlled trials for its use in the treatment of Varicella pneumonia. However, trials have shown a clear benefit of acyclovir in reducing severity of the skin rash in immunocompetent adults when administered at <24 h of rash onset 40. Currently, the consensus is to use acyclovir daily for 7–10 days and this use should be tailored to each patient's clinical assessment. Table 2⇓ summarises published cases with chickenpox pneumonia and antiviral therapy use 5, 15–19, 25, 28, 29, 32, 39–75. This comprised 46 reports, which included 272 patients with Varicella pneumonia. The data is selective; 11 of 179 treated with acyclovir died compared with 17 of 89 patients who received no treatment and none of the other six patients who had other antiviral treatment. There was a significant difference between the groups, with a mortality rate 3.6-fold higher in the group who did not receive acyclovir (OR 3.6, 95% CI 1.63–7.95; p=0.001). There are several limitations to this analysis: 1) only published reports are included; 2) patient selection procedure; 3) lack of randomised controlled trials; 4) no adjustment for the severity of the pneumonia; and 5) no adjustment for ventilatory support. In the most recent series there were no fatalities 16, 28, which may be explained on the basis of having better experience using acyclovir and availability of improved intensive care unit (ICU) support for patients with severe disease. Data from >1,400 pregnant women did not show increased foetal abnormalities as a result of acyclovir treatment 76. Acyclovir is not currently licensed for use in pregnancy, however, the risks from withholding treatment, particularly in the second half of pregnancy when severe complicated chicken pox are more common, probably outweigh the risks of adverse drug effects on the foetus or mother.
The antiviral agents licensed for treatment of VZV include aciclovir, valaciclovir and famciclovir. Valaciclovir is the valine ester of aciclovir and improves oral bioavailability from ∼15% to ∼75%. It is broken down after absorption to valine and aciclovir. Famciclovir is similarly well absorbed. VZV is ∼10-times less sensitive to aciclovir compared with the Herpes Simplex virus. With serious illness, such as pneumonia, it is important to appreciate the necessity to use i.v. therapy (by slow i.v. infusion). However, some clinical response in uncomplicated chickenpox and/or shingles has been demonstrated with oral aciclovir, valaciclovir and famciclovir, the better absorbed preparations probably being preferable. The earlier in the illness the agent is given, the greater the likely benefit.
Corticosteroid use adjunctive to current therapy was assessed in one study that included 15 patients, who were all admitted to ICUs. Six received corticosteroids and showed significantly shorter hospital and ICU stays, and no mortality was recorded 77. The study was small, mainly retrospective, but may form the basis for a prospective, randomised and controlled trial.
Varicella pneumonia can progress rapidly to fulminant respiratory failure despite maximum conventional support; this type of respiratory failure is potentially refractory. This argues for the use of extracorporeal membrane oxygenation/life support (ECMO/ECLS), which have been shown to be beneficial. Of 20 patients treated using ECMO/ECLS, 16 patients survived (60%) 78, 79. Bugge and Tanbo 80 reported a case that was successfully treated with nitric oxide as an alternative to ECMO/ECLS. These interventions have not been subjected to prospective trials to show they improve the outcome. However, they should be considered as an option in those patients who develop fulminant respiratory failure as a result of Varicella pneumonia.
The authors suggest that maximum ventilatory support should be offered to patients who progress to respiratory failure. Acyclovir reduces mortality and should be used early in the course of illness in patients with suspected or proven chickenpox pneumonia.
Morbidity and mortality
Very few reports have addressed the long-term morbidity of Varicella pneumonia 16, 29, 81. Many studies have reported diffused milliary calcifications, which are thought to be clinically insignificant 72. Three published studies investigated the effect of Varicella pneumonia on respiratory function. They included only 33 patients with chickenpox pneumonia and 41 patients without pneumonia who were assessed prospectively 16, 29, 81. Two of these studies did not indicate whether the patients studied had any history of previous pneumonia or immunosuppression and the follow-up varied between 2–13 months 29, 81. However, the remaining study followed 38 immunocompetent patients, with no history of previous pneumonia or lung disease, until recovery or ≥12 months from hospital discharge 16. The main abnormality observed in all studies was a defect in the single breath transfer factor of the lung for carbon monoxide (TL,CO) in patients with and without pneumonia, although the reduction was significantly greater in the pneumonia group. Mohsen et al. 16 reported that patients with pneumonia showed no evidence of further improvement in TL,CO >5 months after hospital discharge. Bocles et al. 81 studied four patients with possible chickenpox pneumonia retrospectively and demonstrated a defect of diffusion ≤8 yrs after the acute illness, which raises the possibility that these changes might be permanent.
There were temporary reductions in forced vital capacity (FVC) and forced expiratory volume in one second (FEV1), although FEV1:FVC remained normal in 92% of patients in their initial lung function tests in two studies 16, 81. However, this was not found in a study by Ellis et al. 29. These changes are probably the result of diffuse inflammation of lung tissue, affecting patients with and without radiological evidence of pneumonia. The abnormalities indicate that chickenpox may be associated with a restrictive lung disease pattern in the acute and recovery phase. There was no evidence to suggest that chickenpox causes obstructive small airway disease in the recovery period.
Mortality rates from chickenpox pneumonia seem to be improving, with an average of 19% in the 1960s and 1970s compared with 6% in more recent data (table 2⇓). The improvement in mortality is likely to be the result of several factors including better experience and availability of chemotherapy, better respiratory support on ICUs and early diagnosis and institution of aciclovir therapy.
There has not yet been a study with sufficient power to address whether antiviral therapy will prevent complications, such as pneumonia. However, the use of acyclovir as prophylaxis/treatment during the incubation period of chickenpox has been shown to prevent or modify illness 82. Although it is not currently licensed, there is indication that such intervention may be used more in the future.
Three immunoglobulin preparations were studied for the prevention and modification of Varicella infection. Currently, the only formula available is VZIG, which is derived from outdated plasma in blood banks. VZIG has proved to be effective 83, and its use is recommended for susceptible immunosuppressed patients and pregnant women who have been in contact with a proven source of Varicella infection 69. It should be emphasised that VZIG should be administered <96 h after exposure to be maximally effective.
Recent evidence suggests that Varicella vaccine is effective in preventing or modifying the severity of Varicella infection if used <3 days and possibly even <5 days after exposure and in outbreak control 84–86. The United States Advisory Committee on Immunization Practices provides a recommendation about administration of Varicella vaccine after exposure, however, it is contraindicated in pregnancy 87.
Immunocompromised hosts are at increased risk of developing severe disease and immunisation of susceptible patients, such as those requiring chemotherapy, prior to profound immunosuppression may be useful (the immunisation available currently uses a live attenuated strain of Varicella). Furthermore, those who have been in contact with a Varicella infection source, and therefore at risk of developing Varicella, should be offered Varicella-Zoster immunoglobulin and monitored closely. The use of acyclovir as prophylaxis post-exposure may be of value but has not yet been formally assessed in this group.
↵Previous articles in this series: No. 1: Tärnvik A, Berglund L. Tularaemia. Eur Respir J 2003; 21: 361–373. No. 2: Mabeza GF, Macfarlane J. Pulmonary actinomycosis. Eur Respir J 2003; 21: 545–551. No. 3: Marrie, T.J. Coxiella burnetii pneumonia. Eur Respir J 2003; 21: 713–719.
- Received November 11, 2002.
- Accepted December 18, 2002.
- © ERS Journals Ltd