Varicella-Zoster Virus Detection by Polymerase Chain Reaction Using Bronchoalveolar Lavage Specimens

CLAYTON T. COWL, UDAYA B. S. PRAKASH, P. SHAWN MITCHELL, and MICHAEL R. MIGDEN

Division of Pulmonary and Critical Care Medicine, Division of Clinical Microbiology, and Department of Dermatology, Mayo Medical Center and Mayo Medical School, Rochester, Minnesota

We report the use of polymerase chain reaction (PCR) techniques from a bronchoalveolar lavage (BAL) specimen for the successful early diagnosis in a case of atypical but severe varicella-zoster virus (VZV) pneumonitis. Atypical presentations of disseminated VZV frequently prolong the time required for accurate diagnosis, resulting in increased morbidity, or mortality. Although further investigation is necessary to determine the sensitivity and positive predictive value of this test, PCR analysis of bronchoscopic specimens may expedite the diagnosis of disseminated VZV.

Varicella-zoster virus (VZV) causes a mild, self-limiting childhood disease that may reactivate years later as shingles. In immunocompromised patients with altered cell-mediated immunity, and rarely in healthy individuals, VZV results in a life-threatening infection (1, 2). Disseminated VZV infections frequently involve the central nervous system (CNS) and the pulmonary system in addition to the typical dermatological findings. Rapid and adequate diagnosis of VZV has previously been hampered by the shortcomings of standard virological methods (3 to 10 d for cytopathic effect in cell culture and 2 to 5 d for shell vial techniques). Recently, polymerase chain reaction (PCR) techniques have been successfully implemented in the rapid (24 h) diagnosis of VZV in patients with CNS involvement (35), in asymptomatic patients with persistent pulmonary nodules during their recovery phase (6), in individuals with varicella arthritis (7), and in pregnant women with transplacental tranmission of the virus (8). In this report, PCR was utilized to rapidly confirm the diagnosis of an atypical varicella-zoster pneumonitis in an immunocompromised patient.

CASE REPORT
A 70-year-old female with history of chronic lymphocytic leukemia presented with an erythematous lesion of the right deltoid region. Initially, the patient believed she had been bitten by a spider while camping the previous week. She was evaluated locally and treated conservatively with a presumed diagnosis of a brown recluse spider bite. Over the next several days, the lesion expanded, became associated with vesicular lesions at its periphery, and a necrotic center rapidly developed. She was referred to our institution after she developed small (1 to 4 mm), tender, erythematous maculopapular lesions around the initial lesion and across her anterior thorax. Her initial workup included multiple biopsies of the right shoulder lesion that were examined under both direct and indirect immunofluorescence and cultured, but were initially

nondiagnostic. In the subsequent 24 h, the lesions became more widespread over the patients neck, face, and genitalia followed by the development of thick, whitish, gingival lesions. Her respiratory status decompensated rapidly; she became hypoxemic (PO2 of 52 mm Hg on 5 L/min supplemental oxygen) and developed bilateral diffuse nodular infiltrates and pleural effusions (Figure 1). The patient was transferred to the intensive care unit where she was intubated and underwent fiberoptic bronchoscopy with bronchoalveolar lavage. Multiple cultures were performed from the bronchial specimens, including those for community-acquired bacteria, viruses, Pneumocytsis carinii, atypical bacteria, and acid-fast bacilli (none of which was eventually reported positive). Computed tomography of the thorax revealed dense consolidation involving the lower lobes. Because of the nature of the skin lesions, varicella was the leading pathogen on the differential diagnosis at the time of the patients decompensation. However, because other infectious sources had not been ruled out the patient received both empiric antibiotic and antiviral therapy. Because of the patients tenuous clinical status, PCR analysis of the patients bronchoalveolar lavage fluid was performed in an effort to obtain a rapid definitive diagnosis. We utilized a 200-l bronchoalveolar lavage sample from our patient, in which DNA was extracted using a chaotropic extraction protocol with 20 g of glycogen added to each sample as a carrier during isopropanol precipitation (Isoquick; OCRA Research, Inc., Bothell, WA). RNase-free water known to be negative for VZV DNA was processed and amplified to demonstrate that the extraction and amplification reagents are free of contaminants. A VZV-positive control was extracted with the patient sample. This will indicate that the extraction technically was adequate to yield DNA for PCR. Oligonucleotide primers directed to a 287-base pair region in the gene 29 sequence of VZV were used to detect target DNA in the sample. These primers do not cross-react with other members of the Herpesviridae family and are specific for VZV as determined by GenBank analysis. The test is able to detect as few as 50 copies per PCR reaction. VZV-DNA present in the specimen was amplified by PCR and identified by gel electrophoresis and subsequent Southern Blot hybridization. The BAL was positive for VZV DNA within 24 h (Figure 2). Three days later, viral cultures from the skin lesions were confirmed positive for VZV. With continued antiviral therapy (acyclovir 800 mg intravenously every 4 h), the patients respiratory status improved over the ensuing 5 d and she was successfully extubated, leaving the hospital after a total of 21 d. Follow-up 6 mo after discharge revealed a complete recovery with no evidence of recurring skin eruptions.

(Received in original form December 9, 1999 and in revised form February 11, 2000) Correspondence and requests for reprints should be addressed to Clayton T. Cowl, M.D., Division of Pulmonary and Critical Care Medicine, Mayo Medical Center, 200 First St. SW, Rochester, MN 55905. E-mail: cowl.clayton@mayo.edu Am J Respir Crit Care Med Vol 161. pp 753754, 2000 Internet address: www.atsjournals.org

DISCUSSION
Pneumonitis may occur in up to half of the VZV infections in adults (9). The presentation is often quite variable and mis-

754

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

VOL 162

2000

dictive value. One such test involves the use of restriction fragment length polymorphism (RFLP) analysis of whole viral DNA, which likely meets the requirements of accuracy. However, the process is time consuming and expensive. Modified techniques using this technology have been employed to successfully identify the pathogen in histologic samples, CNS fluid, and the vesicular fluid of skin lesions (1012). Promising new techniques utilizing PCR may make the diagnosis of VZV even more rapid and less expensive (13, 14). We report the first use of PCR analysis to diagnose disseminated VZV infection using bronchoalveolar lavage samples in the English literature. In Japan, Saito and colleagues recently reported a case of VZV pneumonitis in a young male somewhat refractory to antiviral therapy who was diagnosed by PCR analysis of bronchoalveolar lavage fluid late in the patients clinical course (15). In our patient, the need to obtain an expedited diagnosis due to her rapidly deteriorating clinical condition prompted us to attempt this method. A definitive diagnosis allowed us to focus our antiviral and antimicrobial therapies and avoid unnecessary diagnostic studies that could potentially further jeopardize the patients condition. Although further investigation is necessary to determine the specificity and positive predictive value of this test, PCR analysis of bronchoscopic specimens may expedite the diagnosis of disseminated VZV.
References
Figure 1. Bilateral diffuse nodular infiltrates and small pleural effusions developed rapidly with associated hypoxemia.
1. Feldman, S. 1994. Varicella-zoster virus pneumonitis. Chest 106(Suppl. 1):22S27S. 2. Ruben, F. L., and M. L. Nguyen. 1991. Viral pneumonitis. Clin. Chest Med. 12(2):223235. 3. Iten, A., P. Chatelard, P. Vuadens, J. Miklossy, R. Meuli, R. Sahli, and P. R. Meylan. 1999. Impact of cerebrospinal fluid PCR on the management of HIV-infected patients with varicella-zoster virus infection of the central nervous system. J. Neurovirol. 5(2):172180. 4. Echevarria, J. M., I. Casas, A. Tenorio, F. de Ory, and P. Martinez-Martin. 1994. Detection of varicella-zoster virus-specific DNA sequences in cerebrospinal fluid from patients with acute aseptic meningitis and no cutaneous lesions. J. Med. Virol. 43(4):331335. 5. Bergstrom, T. 1996. Polymerase chain reaction for diagnosis of varicella zoster virus central nervous system infections without skin manifestations. Scand. J. Infect. Dis. Suppl. 100:4145. 6. Miyashita, M. Y., H. Koike, A. Misawa, H. Shimizu, K. Yoshida, T. Yasutomi, N. Kanai, and T. Kagami. 1997. Multiple pulmonary nodules after recovery from chickenpox in a patient with chronic lymphocytic leukemia. Respirology 2(2):135138. 7. Chen, M. K., C. C. Wang, J. J. Lu, C. L. Perng, and M. L. Chu. 1999. Varicella arthritis diagnosed by polymerase chain reaction. J. Formos. Med. Assoc. 98(7):519521. 8. Mouly, F., V. Mirlesse, J. F. Meritet, F. Rozenberg, M. H. Poissonier, P. Lebon, and F. Daffos. 1997. Prenatal diagnosis of fetal varicella-zoster infection with polymerase chain reaction of amniotic fluid in 107 cases. Am. J. Obstet. Gynecol. 177(4):894898. 9. Conti, J. A., and M. S. Karetzky. 1989. Varicella pneumonia in the adult. N. Engl. J. Med. 86(6):475478. 10. Mahalingam, R., S. Kido, M. Wellish, R. Cohrs, and D. H. Gilden. 1995. In situ polymerase chain reaction detection of varicella zoster virus in infected cells in culture. J. Virol. Methods 52:2133. 11. Takayama, M., N. Takayama, N. Inoue, and Y. Kameoka. 1996. Application of long PCR method of identification of variations in nucleotide sequences among varicella-zoster virus isolates. J. Clin. Microbiol. 34: 28692874. 12. Hawrami, K., D. Harper, and J. Breuer. 1996. Typing of varicella zoster virus by amplification of DNA polymorphisms. J. Virol. Methods 57(2):169174. 13. Yoshida, M., and T. Tamura. 1999. An analytical method for R5 repeated structure in varicella-zoster virus DNA by polymerase chain reaction. J. Virol. Methods 80(2):213215. 14. Hawrami, K., and J. Breuer. 1999. Development of a fluorogenic polymerase chain reaction assay (TaqMan) for the detection and quantitation of varicella zoster virus. J. Virol. Methods 79(1):3340. 15. Saito, S., T. Yano, H. Koga, K. Arikawa, N. Koyanagi, and K. Oizumi. 1999. Case of varicella-zoster pneumonia with bronchoalveolar lavage confirmed by the detection of VZV DNA in the bronchial washing by the polymerase chain reaction. Kansenshogaku Zasshi 73(4):346350.

leading, and the infection may cause considerable morbidity and mortality in predisposed groups, especially hosts with deficiencies of cell-mediated immunity, individuals with malignancy, neonates, and pregnant women. Although judicious use of antiviral medications such as acyclovir has been advocated to reduce mortality in this patient population, the key to its efficacy has been early diagnosis and institution of therapy (9). In light of its variable presentation with often rapid clinical decompensation, rapid molecular screening tests for disseminated VZV have been sought with adequate sensitivity and positive pre-

Figure 2. Autoradiograph (left) and agarose gel (2%, right) demonstrate the presence of the 287-base pair PCR product of VZV. Samples were (from top to bottom): PCR master-mix only (blank, lane 1), VZV negative control (normal, lane 2), DNA ladder (lane 3), VZV positive BAL (lanes 4 and 5), VZV negative specimens (lanes 611), VZV positive control (lane 15).