Corticosteroids in Life-threatening Varicella Pneumonia

Corticosteroids in Life-threatening Varicella Pneumonia* Mervyn Mer, MBBCh; and Guy A Richards, MBBCh, FCCP

Background: Varicella pneumonia that results in respiratory failure or progresses to the institution of mechanical ventilation carries a significant morbidity and mortality despite intensive respiratory support and antiviral therapy. There has been no reported study of the role of corticosteroids in life-threatening varicella pneumonia. Design and methods: This was an uncontrolled retrospective and prospective study of all adult patients with a diagnosis of varicella pneumonia who were admitted to the ICUs of the Johannesburg group of academic hospitals in South Africa between 1980 and 1996. Patient demographics, clinical and laboratory features, necessity for mechanical ventilation, and complications were reviewed. The outcome and therapy of varicella pneumonia was evaluated with particular reference to the use of corticosteroids. Patients with comorbid disease and those already taking immunosuppressive agents were excluded. Key endpoints included length of ICU and hospital stay and mortality. Measurements and results: Fifteen adult patients were evaluated, six of whom received corticosteroids in addition to antiviral and supportive therapy. These six patients demonstrated a clinically significant therapeutic response. They had significantly shorter hospital (median difference, 10 days; p<0.006) and ICU (median difference, 8 days; p=0.008) stays and there was no mortality, despite the fact that they were admitted to the ICU with significantly lower median ratios between Pa0 2 and fraction of inspired oxygen than those patients (n=9) who did not receive corticosteroid therapy (86.5 vs 129.5; p=0.045). Conclusion: When used in addition to appropriate supportive care and early institution of antiviral therapy, corticosteroids appear to be of value in the treatment of previously well patients with life-threatening varicella pneumonia. The observations presented in this study are important and should form the basis for a randomized controlled trial, as no other relevant studies or guidelines are available. (CHEST 1998; 114:426-431) Key words: acyclovir; corticoste roids; interstitial pneumonitis; respiratory failure; varicella pneumonia Abbreviations: CI =confidence interval; Flo 2 = fraction of inspired oxygen

varicella pneumonia, initially described in 1942,l represents a severe complication of varicella and usually occurs in adults. Estimates of its incidence vary, with the highest recorded incidence being 50% of all cases of adult chickenpox.2 -6 Varicella pneumoFor editorial comment see page 355 nia has variously been reported to carry an overall mortality of between lO and 30%:5·7 However, where respiratory failure occurs and mechanical ventilation *From the Department of Medicine, Division of Pulmonology, University of th e Witwatersrand Medical School, Johannesburg, South Africa. Manuscript received Octobe r 7 , 1997; revision accepted February 3, 1998 Correspondence to: Mervyn Mer, MBBCh, Department of Medicine, University of the Witwatersrand Medical School, 7 York Rd, Parktown 2193, Johannesburg, South Africa

426

is necessary, mortality is as high as 50%8 ·9 despite appropriate supportive and antiviral therapy. Risk factors for the development of pneumonia in adults with primary varicella include cigarette smoking, pregnancy (in which maternal and fetal mortality is high),5 •10•11 immunocompromise,12 and male sex. 2 Corticosteroids have the potential to modulate the intrapulmonary inflammatory response and have been used successfully in some diseases to limit the degree of respiratory failure. 13 - 17 As far as we are aware, there has been no study of the use of corticosteroids as an adjunct to current therapy in severe fulminant varicella pneumonia. This study was initiated after corticosteroids were used in one critically ill patient who was deteriorating despite antiviral and supportive therapy. The response was so clinically significant that their use in subsequent patients who had failed to respond to conventional therapy appeared to be justified. In Clinical Investigations

addition to documenting demographic and clinical data, we recorded the therapy, outcome, and potential role of corticosteroids in patients with lifethreatening varicella pneumonia.

METHODS

The study consisted of an uncontrolled r etrospective and prospective review of all adult patients with the diagnosis of varicella pneumonia who were admitted to the ICUs of the Johann esburg group of academic hospitals between 1980 and 1996. The data collected consisted of patient demographics, clinical and laboratory features, necessity for mechanical ventilation, therapy (ie, use of acyclovir and corticosteroids ), complications, length of hospital and ICU stay, and mortality. Patients with comorbid disease or those already receiving immunosuppressive agents were excluded because it was felt that the natural history of their disease might be different. Varicella infection was diagnosed on clinical and serologic parameters (ie, IgM response on enzyme-linked immunosorbent assay testing or seroconve rsion). Life-threatening varicella pneumonia was defin ed as the development of respiratory symptoms and signs within 1 week followin g the onset of clinically evident varicella infection, th e presence of infiltrates on the c hest r adiograph, and hypoxia, defin ed as a ratio between Pa0 2 and fraction of inspired oxygen (Flo 2 ) of less than 200. This study was approved unconditionally bythe Committee for Research on Humans of the University of the Witvvatersrand. Statistical Analysis

Key endpoints for comparison between the corticosteroid- and noncorticosteroid-treated groups included length of ICU stay, length of hospital stay, and mortality. Ages, Pa02f'Flo 2 ratios and

platelet counts on admission to ICU, as well as length of ICU stays and total hospital stays, were compared using the MannWhitney U test. The Fisher's exact test (two-tailed ) was used to compare categorical variables between the corticosteroid- and noncOJticosteroid-treated groups, including sex, race, smoking status, treatment with acyclovir, presence/absence of thrombocytopenia, and mortality. The estim ated m edian differences and 95% confidence intervals (Cis) for the median differences in hospital and ICU stays behveen the two groups was calculated with non parametric statistical methods. 18 A p value of < 0.05 was considered significant.

RESULTS

Patient Demographics

The study population (see Table 1) consisted of 15 adult patients. Eleven were white and four were black; nine were men and six were women. The median age was 30 years (range, 23 to 63 years). Another three patients who died were included only in the mortality analysis because detailed clinical records were not available (full demographic, clinical, and laboratory data). These three patients were not included in the demographic and clinical analysis. Eleven patients (73%) were current smokers, accounting for all six of the women and five of the men (55%). Of the remaining men, two were ex-smokers, one had never smoked, and one's smoking status was unknown. Seven patients (46%) had documented contacts, mostly with children with chickenpox, but

Table ! - Demographic, Clinical, and Therapeutic Data of 15 Patients With Life-Threatening Varicella Pneumonia*

Age, yr Sex, M:F No. of cu rrent smokers Race, black:white No. with bilateral infiltrates seen on chest r adiograph No. intubated and ventilated!CPAP No. given acyclovir No. g vi en antibiotics PaO!Flo 2 ratio on ICU admission No. with thrombocytopenia on ICU admission Platelet count on ICU admission, X 109/L Other complicationsf Neurologic Acute renal failure Myocarditis Hepatitis Pancreatitis Orchitis Myos itis DIC

Total (n=15 )

Noncorticosteroid Group (n= 9)

Corticosteroid Group (n=6)

30 (23-63) 9:6

30 (23-55) 5:4

ll 5:10

3:6

29.5 (23-63) 4:2 4 2:4 6 4 6 6 86.5 (51.5-147.4) 4 108 (18-164)

7

15 10 13 15 122.85 (51.5-197.1 )

129.5 (78.0-197.1 )

9

5

113 (18-699)

179 (47-699)

3 4 3 4 2

3 3 2 2 2 1 1 3

1 3

9 6 7

9

p Value 1.00 1.00 1.00 0.580 0.580 0.486 0.045 1.00 0.316

0 1 2 0 0 0 0

*Values expressed as median (range). All other values expressed as absolute numbers. CPAP=continuous positive airway pressure; DIC =disserninated intravascular coagulation. fIn view of extremely small sample sizes, no s tatistical analyses were performed and these are regarded as trends. CHEST /114 /2 I AUGUST, 1998

427

one patient's wife had herpes zoster. One patient was pregnant and one was HIV-positive without any AIDS-defining illness. Clinical Features

Six patients (40%) had oral mucosal lesions in addition to skin lesions. Nine (60%) had thrombocytopenia (platelet count less than 150,000X 109 /L), with the lowest value being 18X 109/L. All patients had bilateral pulmonary infiltrates seen on a chest radiograph. Other complications included encephalitis/cerebellitis in three patients (20%), two of whom were left with permanent neurologic sequelae. Disseminated intravascular coagulation occurred in three, acute renal failure in four, myocarditis in three, hepatitis in four, pancreatitis in two, orchitis in one, and myositis in one. Therapy Ventilation: All the patients had PaOJFio2 ratios of less than 200; ie, they were defined as having acute respiratory distress syndrome. 19 Eight patients (53%) required mechanical ventilation and two (13%) were managed with continuous positive airway pressure masks alone with the amount of airway pressure applied ranging from 5 to 10 em H 2 0 . The patients in the corticosteroid-treated group were admitted to the ICU with significantly lower median PaOJFio2 ratios than the noncorticosteroid-treated group (86.5 VS 129.5; p=0.045). Acyclovir: Thirteen of the patients (86%) received parenteral acyclovir in recommended dosages for a minimum period of 7 days. One patient who did not had been admitted in 1980 prior to the drug's availability in Johannesburg; in the early 1980s, in another patient who was pregnant, the agent was omitted because of concern over possible teratogenic side effects. All the patients received broad-spectrum antibiotics designed to cover usual communityacquired pathogens. Corticosteroids: Six patients (40%) received corticosteroids. One patient received only a single dose before transfer from another hospital. The rest received 200 mg of hydrocortisone every 6 h for a period of 48 h. Corticosteroids were administered early (within 24 h of admission) in the course of the ICU stay.

Hospital Stay

The median hospital stay (see Table 2) for the 15 patients was 12 days (range, 8 to 60 days). The cmticosteroid-treated group had a median stay of 10 days (range, 8 to 13 days) and the noncorticosteroid-treated group was hospitalized a median of 20 days (range, 12 to 60 days). Comparison using the Mann-Whitney U test revealed strong evidence of a between-group difference in mediar1 stay (p<0.006). The estimated median difference between the two groups was 10 days (95% CI, 2 to 21 days). ICU Stay

The median ICU stay (see Table 2) overall was 8 days (range, 2 to 60 days). For the corticosteroidtreated group it was 5.5 days (range, 2to 7 days) and for the noncorticosteroid-treated group it was 12 days (range, 4 to 60 days). Statistical analysis again provided strong evidence of a difference between the two groups (p=0.008). The estimated median difference between the two groups was 8 days (95% CI, 3 to 2.5 days) . Mortality

The three patients for whom detailed records could not be obtained were included in the mortality analysis. Although the limitations of including these three patients were recognized, it was felt that they should be included as all of them received conventional therapy and none received cmticosteroid therapy. With the inclusion of these three patients, the mortality rate for the group as a whole was 22% (four of 18). Of note is the fact that no patient who received corticosteroids died. Mortality in the noncorticosteroid-treated group was 33% (four of 12). The difference in mortality benveen the corticosteroid- and noncorticosteroidtreated groups did not reach statistical significance (p=0.245), but lack of study power may be a factor, and a larger study is needed to confirm these observations. Rate of Response to Corticosteroids

Those patients who received corticosteroids in addition to antiviral therapy and broad-spectrum antibiotics made striking recoveries, particularly in

Table 2-Hospital and ICU Stays: Corticosteroid vs Noncorticosteroid Group*

Hospital stay, d ICU stay, d

Corticosteroid Group (n=6)

N oncorticosteroid Group (n = 9)

Median Difference Between th e Two Groups

95% CI for Median Difference

p Value

10 (8-13) 5.5 (2-7)

20 (12-60) 12 (4-60)

lO 8

2-21 3-25

< 0.006 0.008

*Values expressed as median (range).

428

Clinical Investigations

terms of gas exchange. There was also rapid radiologic improvement (Figs 1-3). The initial patient was on pressure-controlled ventilation with inverse ratios, an Flo2 of 1.0, and a pulmonary capillary wedge pressure maintained in the region of 10 to 12 mm

FIGURE 2. Case 2: Top: chest radiograph in a patient with life-threatenin g vari cella pneumonia showing a significant bilateral pulmonaty infiltrate on admission to th e ICU. Bottom: chest radiograph in th e same patient showing a marked improve ment l day after initiation of corti costeroid th erapy.

FIGURE l. Case l: Top: chest radiograph revealing a significant bilateral pulmonmy infiltrate in a patient with life-threatening vmicella pneumonia on admission to the ICU. Bottom: chest radiograph in the same pati ent showing improve ment 4 clays after initiation of corticos teroid therapy.

Hg. Despite this, the Pa0 2 was 46.4 mm Hg initially and deteriorated. Within 6 h of initiation of corticosteroid therapy, the Flo 2 had decreased to 0.9 and the Pa0 2 had increased to 112 mm Hg. Within 12 h, the Flo 2 had been further reduced to 0.5. Patient 2 was managed initially with a rebreathing mask at 15 Umin of oxygen; his Pa0 2 was 46.8 mm CHEST I 114 I 2 I AUGUST, 1998

429

FIGURE 3. Case 3: Top: chest radiograph in a pati ent with life-threatenin g varicella pneum onia showing bilateral pulmonary infiltrates on admission to the ICU . Bottom: ches t r adiograph in th e same patient showing s ignifi cant improvement 2 days after initiation of corticos teroid th erapy.

Hg and oxygen saturation was 83%. Corticosteroids were begun, and 4 h later the Pa0 2 had increased to 71 .6 mm H g with a saturation of 94%. This rate of recovery continued until discharge. The other patients treated with corticosteroids all responded in a similar fashion , with extremely rapid improvement.

DISCU SSION

This study examined the effects of corticosteroids in patients with severe varicella pneumonia who were also receiving acyclovir and broad-spectrum antibiotics. The 430

rapidity of the clinical response in each case was remarkable. All important vmiables in the corticosteroid and noncmiicosteroid groups were addressed and only the median PaOJFio2 ratio was noted to be statistically different, demonstrating that the two groups were well matched except for the significantly lower PaOi Fio2 ratio in the corticosteroid group. This indicated that initially there was greater impairment in pulmonary gas exchange in those who received corticosteroids than in those who did not. In addition, tl1e corticosteroid group had no mortality and significantly shorter ICU and hospital stays, and tl1ere was a trend towards the development of fewer complications. Therapy to date for life-tl1reatening varicella pneumonia has been r elatively unsuccessful, consisting primmily of cardiorespiratory support and antiviral therapy. The benefit of the latter, however, is controversial. 7 ·9 ·20·2 1 In a r etrospective study of 99 adults with vaticella pneumonia prior to and after the availability of acyclovir, no conclusive benefit was demonstrated.10 This was confirmed by fairly constant mmtality statistics for England and Wales from 1967 to 1985, the period in which antivirals were introduced.12 It is possible, however, that acyclovir may hasten improvement in patients who are less ill and do not require ventilation.1·9 Despite the realization that acyclovir has limited efficacy, 10·21 it is still widely recommended as early primary therapy, particularly in tl1ose patients who are extremely ill. Corticosteroids have been used in the management of several pulmonary disorders, both infective and noninfective in etiology, with some success . Examples of these include overwhelming and miliary TB 15,16 Pneurrwcystis carinii pneumonia 13·14 (in patients with AIDS and other immunocompromised states), vasculitides,22 gastric acid aspiration,23·24 and the fibroproliferative phase of the acute respiratory distress syndrome.25,26 Varicella pneumonia causes an interstitial pneumonitis with subsequent impairment of pulmonary gas exchange, specifically oxygenation capacity. Pathologic changes manifest as a florid immune reaction characterized by an interstitial pneumonitis with mononuclear cell infiltrates, capillary endothelial cell destruction, intra-alveolar exudates and hemorrhage, septal wall invasion by mononuclear cells, and inflammatory changes in the bronchioles.5 This pneumonitis is deemed t o bedue to the host response rather than to specific virally mediated tissue injury. Corticosteroids may potentially modifY the inflammatory response 27 when administered early. At a molecular level, corticosteroids inhibit transcription of genes coding for cytokines, resulting in a d ecrease in the release of macrophage-derived proinflammatory cytokines such as interleukin-1-~ and tumor necrosis factor-a. 28 •29 T-cell function is Clinical Investigations

directly inhibited, 28 as is neutrophil adherence to epithelial cells and complement-induced granulocyte aggregation. 30 Production of membrane-derived products such as leukotrienes and prostaglandins by inflammatory cells is also inhibited, 31 with a consequent decrease in edema and vascular permeability. 32-35 It is also possible that corticosteroids may favorably alter pulmonary hemodynamics. 24 Chickenpox is predominantly a disease of childhood, and in most developed countries less than 5% of reported cases occur in adults. However, more than one third of all deaths from varicella take place in this group, of which varicella pneumonia is the most frequent complication. Recent evidence indicates that there is a marked upward shift in the age at which chickenpox is contracted. 36 Although varicella pneumonia can occur in patients of any age, 75% of cases occur between the third and fifth decades and 90% of cases occur after the age of 19 years. 5 If there is an increasing trend for this disease to occur later in life, it is possible that ICU s and hospitals may be seeing more critically ill patients with varicella pneumonia. Based upon the clinical observations of this study and the fact that the total dose of corticosteroids used would have minimal side effects, we believe that this therapy should be considered in addition to antiviral therapy and appropriate supportive care in any previously well patient with life-threatening varicella pneumonia. This study should form the basis for a randomized controlled trial, as no other relevant studies or guidelines are available. REFERENCES 1 Waring JJ, Neubuerger K, Geever EF. Severe forms of chickenpox in adults. Arch Intern Med 1942; 69:384-404 2 Weber DM, Pellecchia JA. Varicella pneumonia: study of prevalence in adult men. JAMA 1965; 192:572-73 3 Mermelstein RH, Freireich AW. Varicella pneumonia. Arch Intern Med 1961; 55:456-63 4 Hockberger RS, Rothstein RJ. Varicella pneumonia in adults: spectrum of disease. Ann Emerg Med 1986; 15:931-34 5 Triebwasser JH, Harris RE, Bryant RE, e t al. Varicella pneumonia in adults: report of seven cases and a review of literature. Medicine 1967; 46:409-23 6 Weinstein L, Meade RH. Respiratory manifestations of chickenpox. Arch Intern Med 1956; 98:91-99 7 Schlossberg D, Littman M. Varicella pneumonia. Arch Intern Med 1988; 148:1613-32 8 Feldman S. Varicella-zoster virus pneumonitis. Chest 1994; 106(suppl):22S-27S 9 Haake DA, Zakowski PC, Haake DC, et al. Early treatment with acyclovir for varicella pneumonia in othef\vise healthy adults: retrospective controlled study and review. Rev Infect Dis 1990; 12:788-98 10 Esmonde TF, Herdman G, Anderson G. Chickenpox pneumonia: an association with pregnancy. Thorax 1989; 44:812-15 11 Harris RE, Rhoades ER. Varicella pneumonia complicating pregnancy: report of a case and review of literature. Obstet Gynecol 1965; 25:734-40 12 Joseph CA, Noah ND. Epidemiology of chickenpox in

England and Wales, 1967-85. Br Med J 1988; 296:673-76 13 Gagnon S, Ahmad M, Boota ND, et al. Corticosteroids as adjunctive therapy for severe Pneunwcystis carinii pneumonia in the acquired immunodeficiency syndrome. N Eng! J Med 1990; 323:1444-50 14 Bozette SA, Sattler FR, Chiu J, et al. A controlled trial of early adjunctive treatment with corticosteroids for Pneumocystis carinii pneumonia in the acquired immunodeficiency syndrom e. N Eng! J Med 1990; 323:1451-57 15 Johnson JR, Taylor BG, Morrisey JF. Corticosteroids in pulmonary tuberculosis. Am Rev Respir Dis 1965; 92:376-91 16 Senderovitz T, Viskun K. Corticosteroids in tuberculosis. Respir Med 1994; 88:561-65 17 Rupp ME, Schwartz ML, Bechard DE. Measles pneumonia: treatment of a near-fatal case with corticosteroids and vitamin A. Chest 1993; 103:1625-26 18 Hollander M, Wolfe DA. Nonparametric statistical methods. New York: John Wiley, 1973; 78-82 19 Bernard GR, Artigas A, Brigham KL, e t a!. The AmericanEuropean Consensus Conference on ARDS: definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med 1994; 149:818-24 20 Clark GPM, Dobson PM, Thickett A, et al. Chickenpox pneumonia, its complications and management. Anaesthesia 1991; 46:376-80 21 Davidson RN, Lynn W, Savage P, et al. Chickenpox pneumonia: experience with antiviral treatment. Thorax 1988; 43:627-30 22 Leavitt RH, Fauci AS. Pulmonary vasculitis. Am Rev Respir Dis 1986; 134:149-66 23 Bannister WK, Sattilano AJ, Otis RD. Therapeutic aspects of aspiration pneumonitis in experimental animals. Anesthesiology 1961; 23:440-45 24 Toung TJK, Bordos D , Benson DW, et al. Aspiration pneumonia: experimental evaluation of albumin and steroid therapy. Ann Surg 1976; 183:179-84 25 Meduri GU, Belenchia JM, Estes RJ, et al. Fibroproliferative phase of ARDS: clinical findings and effects of corticosteroids. Chest 1991; 100:943-52 26 Meduri GU, Headley S, Carson S, et al. Methylprednisolone treatment of late ARDS (abstract]. Am J Respir Crit Care Med 1997; 155:A391 27 Meduri CU. The role of the host defence response in the progression and outcome of ARDS: pathophysiological correlations and response to glucocorticoid treatment. Eur Respir J 1996; 9:2650-70 28 Barnes PJ, Adcock I. Antiinflammatory actions of steroids: molecular mechanisms. Trends Pharmacol Sci 1993; 14:436-41 29 Huang ZB, Eden E. Effect of corticosteroids on ILli3 and TN Fa release by alveolar macrophages from patients with AIDS and Pneunwcystis cminii pneumonia. Chest 1993; 104:751-55 30 Hammerschmidt D, White J, Craddock P, et al. Corticosteroids inhibit complement-induced granulocyte aggregation. J Clin Invest 1979; 63:798-803 31 HongS, Levine L. Inhibition of arachidonic acid release fi·om cells as the biochemical actions of anti-inflammatory steroids. Proc Nat! Acad Sci USA 1976; 73:1730-33 32 Farrukh IS, Michael JR, Peters SP, et al. The role of cyclooxygenase and lipoxygenase mediators in oxidant-induced lung injury. Am Rev Hespir Dis 1988; 137:1343-49 33 Farrukh IS, Spannhake EvV, Sciuto AM, et a!. Leukotriene D4 increases pulmonary vascular permeability and pressure by different mechanisms in the rabbit. Am Rev Respir Dis 1986; 134:229-32 34 Ryrfeldt A, Bannenberg G, Moldeus P. Free radicals and lung disease. Br Med Bull 1993; 49:558-603 35 Boscheto P, Rogers DF, Fabbri LM, et al. Corticosteroid inhibition of airway microvascular leakage. Am Rev Respir Dis 1991; 143:605-09 36 Schoub BD. Chickenpox in adults. Virus SA 1992; 1:1-3 CHEST 1114 I 2 I AUGUST, 1998

431