Fatal varicella in steroid-dependent asthma

Allergen

VOLUME 81 NUMBER 1

26. Russell G, Jones SP. Selection of skin tests in childhood asthma. Br J Dis Chest 197670: 104. 27. Marsh DG, BiasWB. BasalserumIgE levelsand HLA antigen frequencies in allergic subjects. II. Studies in people sensitive to rye grass group 1 and ragweed antigen E and of postulated immuneresponse(Ir) lociin the HLA region.Immunogenetics 1977;5:235.

Appendix When the frequency distribution of threshold doses is expressedas percentage of responders against log dose, it elicits a sigmoid curve (Fig. 5). If the threshold doses are normally distributed, percentage of responderscan be expressed in probits (see Table I in reference 15), and the curve is transformed to a straight line (Fig. 5). The fitting of a probit regression line is computed by maximum like-

Boston.

curve

lihood estimation.” A chi-squaretest for goodnessof fit is used to test if the line is an adequate representationof the data. The dose that will produce a response in half the population is defined as ED,,. The comparison of the potency of allergenic extractscan be performed when the seriesof data elicit parallel regressions against the log dose. Thus, the difference between dosesthat produce the same percentageof responseis constant and is called R. With an allergen extract as a standard (for instance, grass in Table I), R representsthe multiple (or the fraction) of the standard extract that has the same effect of the others. R is > =
Fatal varicella in steroid-dependent Howard Antonio

dose-response

asthma

J. Silk, MD, Lisa Guay-Woodford, MD, R. Perez-Atayde, MD, Raif S. Geha, MD, and Martin

0. Broff, MD

Mass.

Disseminated varicella infection is a potentially life-threatening complication of chronic high-dose corticosteroid (CS) or immunosuppressive therapy. A review of the literature indicates that, with one possible exception, this complication has not occurred in a G-dependent subject with asthma. We present in this article the clinical features and autopsy jindings of a steroid-dependent subject with asthma who died of acute, disseminated varicella. A 16-year-old poorly compliant, steroid-dependent subject with asthma received two courses of high-dose intravenous methylprednisolone during a 3-week period, followed by a tapering schedule of oral prednisone. During this time, she was exposed to chickenpox. She subsequently developed a classic varicella rash, severe back pain, rapidly progressive hepatic failure, pneumonitis, and encephalopathy. Death ensued 3 days after the onset of the rash. Evidence of disseminated varicella infection was confirmed at autopsy. This case illustrates that a small number of subjects with severe asthma receiving high-dose CS need to be considered as a separate, high-risk group for developing disseminated varicella. We recommend that the immune status of these patients to varicella-zoster virus be assessed by a serum titer. If these patients are nonimmune, they would be candidates for varicella-zoster immune globulin on exposure, and for acyclovir therapy should varicella dissemination occur. (J ALLERGY CWN IMMUNOL 1988;81:47-51.)

From the Divisions of Allergy and Pathology, Children’s Hospital, and the Department of Pediatrics, Harvard Medical School, Boston, Mass. Supported by National Institues of Health Training Grants HD T32, HD-07321-01, and National Institute of Allergy and Infectious Diseases Grant ROl AI-2037. Received for publication Feb. 2, 1987. Accepted for publication July 11, 1987. Reprint requests: Howard J. Silk, MD, Atlanta Allergy Clinic, 6667 Vemonwoods Dr., Atlanta, GA 30328.

The complications of CS use in severe asthma are well-known. ’ Fortunately, life-threatening complications are rare. Disseminated varicella infection is a potentially life-threatening complication of longterm, high-dose CS or immunosuppressive therapy, which is extremely rare in asthma.’ Our review of the English literature reveals only one possible case of fatal varicella in a CS-treated subject with asthma.3 We present the first clinical and pathologic findings 47

48

J. ALLERGY CLIN. IMMUNOL. JANUARY 1988

Silk et al.

Abbreviations used CS: Corticosteroid VZlG: Varicella-zoster VZV: Varicella-zoster

immune globulin virus

of fatal varicella in a long-term steroid-dependent subject with asthma, as well as a discussion of the literature. CASE REPORT K. T., a 16-year-old black female adolescent, with longterm steroid-dependent asthma, was initially diagnosed with asthma at 4 years of age. She was first observed at the Children’s Hospital allergy clinic at age 6 years. She was skin test positive to cockroach, dust, and mite. CSs were first used for control of her asthma at 7% years of age. From the ages of 7% to 16 years, she was hospitalized 36 times for status asthmaticus, as well as numerous emergency room visits for treatment of acute attacks. In addition to bronchodilator therapy, she was treated with CSs during all of these admissions and many of the outpatient visits. For long periods of time, she was receiving alternate-day prednisone. A major problem was poor compliance with medications. Despite a 60-day stay in the psychosomatic unit, there was no difference in her frequency of hospitalization or her compliance with medications. Review of the medical record reveals a cumulative steroid dose of >20,500 mg before her final hospitalization. There was no history of recurrent or opportunistic infections. During the thirty-fourth hospital admission on Dec. 18, 1984, she received intravenous methylprednisolone, 1 mgikglday, for 3 days, and she was discharged on a tapering schedule of oral prednisone, starting at 40 mg daily. On December 23, her 5-year-old brother developed varicella. The patient was readmitted for the thirty-fifth time on Jan. 8, 1985, when intravenous methylprednisolone, 2 mg/kg/day, was administered for 4 days, followed by another tapering schedule of prednisone. Two days previously, her sister developed varicella. On Jan. 18, 1985, she presented to the emergency room with severe back pain of l-day duration. The pain was spasmodic, paraspinal, and radiating to the legs. She was unable to walk. Blood pressure was 116/80; pulse, 80; and respiration, 18. Her skin demonstrated a few scattered nonerythematous papules and vesicles. Result of the examination of the mucous membranes was notable only for a single vesicle on the gingiva. Her chest examination revealed mild wheezes, the heart was normal, and abdomen was soft and flat with mild right upper quadrant tenderness and normal bowel sounds. There was tenderness in the paraspinal area. Neurologic examination results were grossly normal. Laboratory studies demonstrated a white blood count of 27,600 with 74 neutrophils, 10 bands, 7 lympho-

cytes, 4 monocytes, 3 metamyelocytes, and 2 atypical lymphocytes. Her serum glutamic oxalacetic transaminase and serum glutamic pyruvic transaminase were elevated at 1 IS mu/ml and 81 mu/ml, respectively (normal values, 2 to 21 mU/ml and 2 to 22 mu/ml), lactic dehydrogenase of 506 units (normal, 2 to 88), and creatine phosphokinase of 62 mu/ml (normal, 5 to 30 mUim1). Three doses of meperidine were required for control of pain, and she was admitted. Her hospital course was notable for progressively increasing liver enzyme levels, fever, and a white blood count of 66,400, with 39 neutrophils, 28 bands, 9 lymphocytes, 14 monocytes, 6 metamyelocytes, and 3 eosinophils. A Tczanck preparation of a vesicular lesion was positive for intranuclear inclusion bodies. Disseminated intravascular coagulopathy developed, and she was transferred to the mobile intensive care unit where acyclovir was started, intravenously, as well as ticarcillin and tobramycin, fresh-frozen plasma, and hydrocortisone at physiologic replacement doses. An interstitial pneumonitis developed with a progressively increasing oxygen requirement and increasing alveolar-arterial oxygen gradient. Hepatic function continued to deteriorate; the coagulopathy was refractory to treatment. Myocardial function deteriorated, and hypotension developed that was unresponsive to colloid boluses or dopamine hydrochloride infusion. After the acute onset of confusion and encephalopathic signs, her respiratory effort became irregular. Brachycardia developed, and then asystole. Despite a prolonged resuscitative effort, her blood pressure and heart rate could not be maintained, and resuscitative efforts were suspended. Permission for autopsy was obtained.

RESULTS Pathology At autopsy, evidence of a disseminated herpes simplex viral infection was observed. The skin had the characteristic varicella rash with lesions at different stages of progression. Histologically, the maculopapular and vesicular lesions had intranuclear inclusions in acantholytic keratinocytes and in characteristic, multinucleated giant cells. The liver had extensive geographic and discrete areas of necrosis. Necrotic and degenerating hepatocytes demonstrated characteristic

eosinophilic

intranuclear

inclusions.

Electron

microscopic studies of the liver revealed numerous intranuclear and intracytoplasmic viruses with cuboidal shape and ultrastructural features of herpes. The nuclei of infected cells had margination and dissolution of the chromatin with numerous viral nucleoids and nucleocapsids (Fig. 1). The lymphoid system had generalized, marked lymphoid depletion and lympholysis. The lungs had evidence of chronic bronchitis, with features of asthma, and diffuse intra-alveolar edema.

Fatal

VOLUME 81 NUMBER 1

varicella

49

Fig. 1. Electron

micrograph of an infected hepatocyte demonstrating effacement of chromatin and marked degeneration and transformation of cytoplasmic organelles. Numerous viral particles are present within the nucleus (arrows) and cytoplasm, and extracellularly (uranyl acetate and lead citrate; original magnification x7700). The inset demonstrates higher magnification of a hepatocyte nucleus containing numerous viral nucleoids (small arrow) and nucleocapsides (large arrow) (uranyl acetate and lead citrate; original magnification x18.900).

DISCUSSION The risk of disseminated or fatal varicella infection in immunocompromised patients is well recognized, with mortality ranging from 7% to 50%.3-6 Haggerty and Eley3 first reported two cases of fatal varicella in patients receiving cortisone therapy. These patients developed shock and died several days after the appearance of vesicles. Anecdotal reports of eleven additional fatalities were discussed, one of which was a 4-year-old boy with asthma who was taking 50 mg of cortisone daily (equivalent to 10 mg prednisone). He died 3 weeks after onset of vesicles, secondary to acute encephalitis. Scabs were present at the time of death. This is the only reported case of fatal varicella in a (X-treated subject with asthma. However, the late onset of encephalopathy is more consistent with Reye’s syndrome, as varicella encephalitis occurs between 2 to 4 days after initial signs of dissemination7 Neither the case history nor the pathology was discussed in detail to confirm the diagnosis of fatal varicella. It has become evident that the mortality risk in CStreated.patients who develop varicella is directly related to the underlying disease process. Patients at particular risk are patients with acute lymphoblastic leukemia or other malignancies or with nephrosis, and transplant recipients .4, 8-‘2 The two major reviews of

allergic patients’, I2 have not reported any additional deaths in CS-treated atopic patients or patients with asthma. Falliers and EllisI reported 59 cases of varicella in hospitalized patients with asthma. Twenty-one patients were receiving CS at the time of their infection. There were no cases of dissemination, and all patients had unremarkable disease courses. Girsh et al.* reviewed the records of 474 hospitalized patients with varicella, 16 of whom were receiving CS therapy. The two atopic patients in this group recovered uneventfully. Varicella infection occurs after seeding of the virus in the nasopharynx, followed by a primary viremia with infection of the reticuloendothelial system. This is followed by a secondary viremia with dissemination to the skin, and possibly other organs as we11.6 Viremia has been described in both normall and immunocompromised patients,14 detectable up to 5 days before onset of rash and clinical disease. Host responses to VZV infection include both humoral and cell-mediated immune responses. Specific VZV antibodies are detected 2 to 5 days after onset of rash, with maximal elevation 2 to 3 weeks later, although antibodies can be detected years after the primary infectionI The significance of humoral immunity is questionable, however, since children with

50

Silk

et al.

agammaglobulinemia do not develop disseminated disease,6 and passive immunization with either maternal antibody or VZIG modifies, but does not prevent, infection. 16,” Our patient’s initial presentation is consistent with the clinical course of disseminated varicella in immunocompromised patients.’ In 15 reported cases of disseminated varicella, all had fever, rash, hepatitis, and pneumonitis. Eleven of 15 patients, or 73%, had severe abdominal pain requiring narcotic analgesia, and five of 15 patients had severe back pain or myalgias. Fever, rash, and abdominal or back pain were always the presenting symptoms and heralded severe visceral involvement. Our patient was a poorly compliant, steroiddependent patient with asthma who had received multiple courses of high-dose steroids since the age of 4 years. There was no previous history to suggest a primary immune deficiency or a noniatrogenic cause for suppressed cell-mediated immunity. Her siblings developed clinical varicella infection 21 and 5 days before the onset of our patient’s rash. She had been hospitalized twice for status asthmaticus and received a cumulative dose of 680 mg of intravenous methylprednisolone, in association with both the primary and secondary viremias. The immunosuppressive effects of the high-dose CS on her ability to contain the viremia, presumably resulted in the subsequent visceral dissemination and the fatal outcome of this patient’s disease. CONCLUSION Varicella infection is virtually ubiquitous in humans. Most cases, including patients with asthma, recover uneventfully. The patient with asthma receiving moderate doses of alternate-day CS, or those on short pulses of CS, do not need their steroid doses altered if they contract chickenpox, as they are at very slight risk for disseminated disease. However, as this case illustrates, a small group of CS-dependent subjects with severe asthma need to be considered as a separate, high-risk group, similar to other immunosuppressed children. Although it is difficult to predict which patient is at greater risk for dissemination, patients who have recently received, or are receiving, daily high-dose CS do exhibit impaired delayed-type skin reactivity, as compared to patients on alternate-day CS regimens. Is These patients may therefore be considered at greatest risk for dissemination. Since VZV titers are measurable for a prolonged period of time after infection,15 we recommend that all CS-dependent patients with severe asthma should be assessed for the status of their immunity by a serum titer for VZV. If the titer does not indicate immunity, the patient should

J. ALLERGY CLIN. IMMUNOL. JANUARY 1988

avoid contact with subjects having clinical varicella and immediately report any exposure. If a nonimmune patient is aware of a varicella exposure, passive immunization with VZIG should be administered within 72 hours. VZIG (available through local American Red Cross offices) prophylactically reduces disease severity and complication rate in immunocompromised patients, whereas it is not effective therapeutically.‘6, I7 If varicella develops, the patient should be closely monitored for early symptoms of dissemination, especially severe abdominal or back pain. Acyclovir has been proved to be the drug of choice for disseminated varicella in immunocompromised patients,” and therapy should be initiated promptly if signs of dissemination develop. Finally, nonimmune CS-dependent subjects with asthma should be considered candidates for varicella vaccination when it becomes generally available. We thank Mrs. Adrienne Siscoand Miss Ann MacDonald for secretarialassistancein preparing this manuscript.

REFERENCES 1.

ShapiroGG. Corticosteroidsin the treatmentof allergicdisease: principlesand practice.PediatrClin North Am 1983;

30:955. 2. Girsh LS, Yu M, Jones J, Schulaner FA. A study of the risk of mortality of varicella in patients with bronchial asthma. Ann Allergy 1966;24:690. RJ, Eley RC. Varicella and cortisone. Pediatrics 3. Haggerty 1956;18:160. 4. Feldman S, Hughes WT, Daniel CB. Varicella in children with cancer: seventy-seven cases. Pediatrics 1975;58:388. 5. Dolin R, Reichman RC, Mazur MH, Whitley RJ. Herpes zoster-varicella infections in immunocompromised patients. Ann Intern Med 1978;89:378. 6. Weller TH. Varicella and herpes zoster. N Engl J Med 1983;309:1362. I. Morgan ER, Smalley LA. Varicella in immunocompromised children: incidence of abdominal pain and organ involvement. Am J Dis Child 1983;137:883. 8. Finkel KC. Mortality form varicella in children receiving adenocorticosteroids. Pediatrics 1961;28:436. 9. Gershon A, Brunell PA, Doyle EF. Steroid therapy and varicella. J Pediatr 1972;81:1034. 10. Close GC, Houston IB. Fatal hemorrhagic chicken pox in a child on long-term steroids. Lancet 1981;2:480. 11. Feldhoff CM, Balfour HH Jr, Simmons RL, Najarian JS, Mauer SM. Varicella in children with renal transplant. J Pediatr 1981;98:25. 12. Falliers EJ, Ellis EF. Corticosteroids and varicella. Arch Dis Child 1965;40:593. 13. Asano Y, Italcura N, Hiroishi Y, et al. Viremia is present in incubation period in nonimmunocompromised children with varicella. J Pediatr 1985;106:69. 14. Feldman S, Epp E. Isolation of VZ virus from blood. J Pediatr 1976;88:265. 15. Brunell P, Gershon A, Uduman S, Steinberg S. Varicella-zoster immunoglobulins during varicella, latency, and zoster. J Intect Dis 1975;132:49.

Fatal

VOLUME 81 NUMBER 1

16. Orenstein W, Heymann D, Ellis RJ, Rosenberg RL, et al. Prophylaxis of varicella in high-risk children. J Pediatr 1981; 99:368. 17. Committee on infectious diseases. Peter G. Varicella zoster infections. In: Report of the Committee on Infectious Diseases. 20th ed. Evanston, Ill.: American Academy of Pediatrics, 1986:399-400.

varicella

18. MacGregor R, Sheagren J, Lipsett M, Wolff S. Alternate-day prednisone therapy. N Engl J Med 1969;280: 1427. 19. Shepp DH, Dandliker PS, Meyers JD. Treatment of varicellazoster virus infection in severe immunocompromised patients. N Engl J Med 1986;314:208.

Circadian and circannual rhythms of allergic rhinitis: An epidemiologic study involving chronobiologic methods Alain Reinberg, MD, Pierre Gervais, MD, Francis Levi, MD, Michael Smolensky, PhD,* Luis Del Cerro, PhD, and Catherine Ugolini, PhD Paris, France, and Houston, Texas Seven hundred sixty-five patients, living in France and suffering from allergic rhinitis (eg, with positive skin tests to various antigens), agreed to self-rate (visual analog scales}, four times daily, symptoms such as sneezing, stuffy or blocked nose, runny nose, itchy nose, itchy eyes, wheeze, or cough. Despite acute symptoms, patients did not take medications of any kind by any route during 36 hours. Several statistical methods (eg, Student’s t test, analysis of variance, cosinor, chi-square, etc.) were used to validate both circadian and circannual rhythms of these symptoms in the group as a whole, as well as in subgroups related to age, sex, etc. Large-amplitude circadian rhythms with early morning peak times (eg, -6 AM) were validated for sneezing, stuffr nose, and runny nose (with p < 0.0001) but notfor wheeze or cough. Such time-dependent changes were related neither to age (from 10 to 80 years) nor to sex. However, small differences were observed in subgroups sorted with regard to duration of disease (old versus new cases), smoking habits, and geographic location (north versus south France). Reanalysis of data taking into account interindividual differences revealed that the respective peak times of the three major symptoms occurred in the early morning in about 60% to 70% of the patients. Annual changes were validated as well with the annual peak time being January to April. The proposed interpretation of both circadian and circannual rhythms suggests taking into account endogenous component rhythms (eg, involving metabolic, immunologic, and endocrine systems), since they contribute to time-dependent changes in the human susceptibility to antigens. In addition the elevated severity of symptoms in the morning experienced by 60% to 70% of patients should serve as a guide to individually optimize dosing time(s) of medications, such as antihistamines. (J ALLERGY CLIN IMMUNOL 1988;81:51-62.)

Periodic changes in the occurrence and intensity of the symptoms of rhinitis as a function of the time of From E. R. Chronobiologie Chronopharmacologie, National Center for Scientific Research and Medecine Inteme Allergologie, Fondation A. de Rothschild, and Policlinique Medicale, Hopital Femand Widal, Parts, France, and *University of Texas, School of Public Health and Graduate School of Biomedical Sciences, Health Science Center at Houston, Houston, Texas. Supported by National Center for Scientific Research and Laboratories Rhone-Poulenc Pharmuka-SM, Rhone, France. Received for publication Dec. 1, 1986. Accepted for publication July 11, 1987. Reprint requests: Michael Smolensky, University of Texas, School of Public Health and Graduate School of Biomedical Sciences, Health Science Center at Houston, PO Box 20186, Houston, TX 77225.

Abbreviations

used

AU: Arbitrary unit ANOVA: Analysis of variance day during the 24 hours were initially reported by Trousseau in 1865. ’ The morning preponderance of symptoms as well as their rhythmicity led Trousseau to consider rhinitis as a form of asthma affecting the nose (“L’asthme du nez”). However, the temporal pattern observed by Trousseau was not objectively quantified. More than 100 years later, Nicholson and Bogie’ reported the findings of their study on 246 “hay fever” sufferers, noting that the commencement of 51