Judith A. Aberg, MD; Linda M. Mundy, MD; and William G. Powderly, MD

Pulmonary Cryptococcosis in Patients Without HIV Infection* Judith A. Aberg, MD; Linda M. Mundy, MD; and William G. Powderly, MD Purpose: To further elucidate the diagnostic and therapeutic approaches to patients with pulmonary cryptococcosis who are not HIV-infected. Subjects: All of the patients without HIV infection who received care at two Midwest hospitals between January 1986 and February 1996 and had a respiratory isolate of Cryptococcus neoformans. Methods: The medical records of the study patients were reviewed for demographic data, host immune status, respiratory symptoms, diagnostic studies, treatment, and follow-up. Results: Forty-two patient presentations comprised the overall study group. Thirty-six patients (85.7%) had no evidence of dissemination, and six patients (14.3%) had disseminated disease. Seven of the 36 patient presentations were definitive pulmonary cryptococcosis, 15 were presumptive disease, and 14 were colonization with C neoformans. Neither the baseline demographic parameters nor the immune status appeared to discriminate the patients with disease from the patients with colonization. A serum cryptococcal antigen (scrag) was positive for 7 of 18 patients, 3 of whom were proven by culture to have a disseminated infection. A negative scrag was observed in 11 patients, one of whom had proven dissemination. Fifteen patients underwent a lumbar puncture as part of their evaluation, and cryptococcal meningitis was diagnosed in three of these patients, all of whom had positive blood cultures for C neoformans. The majority of the patients did not receive antifungal therapy. Conclusion: In the majority of the patients, the lung appeared to be the sole organ involved, and a workup for systemic infection was rarely helpful. A positive scrag was not specific for dissemination. Antifungal therapy should be reserved for symptomatic patients, for patients with a positive scrag, and for patients with underlying immunosuppression. (CHEST 1999; 115:734 740) Key words: antifungal therapy; cryptococcal antigen; HIV negative; pulmonary cryptococcosis Abbreviations: CSF cerebrospinal fluid; scrag serum cryptococcal antigen Pulmonary cryptococcal infection is relatively rare. Although several types of yeast can be isolated from the respiratory tract of immunocompetent hosts, Cryptococcus neoformans is not considered to For editorial comment see page 610 be a normal respiratory flora in humans or animals. As a consequence, its isolation from respiratory specimens warrants attention from physicians. In some cases, the C neoformans isolate can represent a *From the Division of Infectious Diseases, Department of Medicine (Drs. Mundy and Powderly), Washington University School of Medicine, St. Louis, MO; and the AIDS Clinical Trials Unit, San Francisco General Hospital AIDS Program (Dr. Aberg), University of California, San Francisco, CA. Manuscript received May 20, 1998; revision accepted October 14, 1998. Correspondence to: Judith A. Aberg, MD, San Francisco General Hospital AIDS Program, 995 Potrero Avenue, Building 80, Ward 84, San Francisco, CA 94110; e-mail: jaberg@sfaids.ucsf.edu pulmonary infection, either alone or in association with a disseminated disease; in other cases, it can represent an asymptomatic carriage. 1 The distinction between disseminated cryptococcal disease, pulmonary cryptococcosis, and airway colonization with C neoformans has important diagnostic and therapeutic implications. However, this distinction may be difficult to ascertain. 2,3 Most of the previous reviews of pulmonary cryptococcosis were conducted before the current epidemic of HIV infection, a situation that clearly increased the prevalence of cryptococcal infection. Pulmonary cryptococcal infection in the HIV-infected population should generally be handled in a manner similar to that of disseminated disease. The previous literature on pulmonary cryptococcosis suggested that isolated pulmonary cryptococcosis in immunocompetent patients does not require therapy. 2 However, over the last 10 years, several orally active, well-tolerated antifungal agents have emerged 734 Clinical Investigations

that could have a role in the management of pulmonary cryptococcosis. After reviewing our experience with pulmonary cryptococcal infection in patients who were not HIV-infected, we attempt to clarify the diagnostic and therapeutic approaches to this patient population. Patients Materials and Methods All of the patients who received care at Cleveland Clinic Hospital, Cleveland, OH, or at Barnes Hospital, St. Louis, MO, between January 1986 and February 1996 were eligible for study enrollment. The inclusion criteria accepted patients with a growth of C neoformans from a respiratory specimen noted in the clinical mycology laboratory log book of their respective institution. The respiratory specimens were defined as expectorated sputum, induced sputum, BAL, lung biopsy, or lung histopathology specimen. The patients were excluded if they had a known HIV infection or unavailable medical records. Chart Abstraction The medical records of the study patients were reviewed for demographic data, host immune status, confirmation of C neoformans and the isolate site, respiratory symptoms (asymptomatic, fever, tachypnea, shortness of breath, cough, thick pulmonary secretions, and hemoptysis), additional diagnostic studies such as serum cryptococcal antigen (scrag), evaluation of the cerebrospinal fluid (CSF), chest radiographs, CT scans, treatment, and follow-up (range, 1 to 7 years). Study Definitions The patients were categorized as having either pulmonary cryptococcosis or disseminated C neoformans. For pulmonary cryptococcosis, the case definitions used in this report are as follows: (1) definite, the isolation of C neoformans in a respiratory specimen with radiographic evidence of disease and clinical symptoms without any other proven etiologies; (2) presumptive, the isolation of C neoformans in a respiratory specimen with radiographic evidence of disease, clinical symptoms, and concomitant respiratory pathogens or other noninfectious respiratory disease (bronchitis, aspiration pneumonia, bronchiectasis, interstitial pulmonary fibrosis, congestive heart disease, or connective tissue disease), so that one cannot solely attribute the patient s presentation to cryptococcosis; and (3) colonization, the isolation of C neoformans in a respiratory specimen with either (a) a normal chest radiograph, (b) an asymptomatic state with abnormal chest radiograph, or (c) a postmortem confirmation of another etiology without evidence of cryptococcal disease. A disseminated disease was defined as an isolation of C neoformans in the blood, in sterile body fluid, or at an extrapulmonary site. The growth of C neoformans from pleural fluid without evidence of systemic disease was classified as a localized disease. Results Sixty of 174 patients (34.5%) with isolates of the Cryptococcus spp in their clinical mycology laboratories had respiratory specimens and were eligible for study enrollment. Nineteen patients were excluded, including 11 with a known HIV infection, 5 with a final isolate identified as Cryptococcus albidus, and 3 with unavailable medical records. Forty-two patient presentations comprised the overall study group. Thirty-six patients (85.7%) had no evidence of dissemination, and 6 patients (14.3%) had disseminated disease. One patient with pulmonary cryptococcosis was categorized twice. Initially, this patient had presumptive disease; however, it evolved to definitive disease after steroid therapy was initiated for membranous proliferative glomerulonephritis. Of the 41 patients, the above-mentioned patient was the only one having a recurrent isolate of C neoformans. Seven of the 36 patient presentations were definitive pulmonary cryptococcosis, 14 were presumptive disease, and 15 were colonization with C neoformans. There were 27 men (64.2%), 38 whites (90.5%), and the median age was 63 years (range, 22 to 82 years). Eighteen patients (42.9%) were immunocompetent, and the rest were immunosuppressed. Among the six patients with disseminated cryptococcosis, all were immunocompromised (Table 1). Oral steroids (dosage range, 30 to 60 mg po qd) were being administered to three patients: one had seronegative rheumatoid arthritis; the second had Crohn s disease; and the third had crescentic glomerular nephritis. The median age was 61.5 years (range, 22 to 75 years). Five patients (83.3%) had pulmonary symptoms that included shortness of breath, and two patients (33.3%) had documented fever. The blood cultures of all of the patients grew C neoformans. Three of four patients tested positive for scrag, and three of five patients had positive CSF cryptococcal antigen titers and cultures. The one negative scrag had been sent to a reference laboratory 2 weeks before the collection of blood cultures. The other scrag specimens were collected concurrently with the blood cultures. All six patients had abnormal chest radiographs. The single CT scan provided no additional information. Five of the six patients received treatment. Three patients died (including the untreated patient), and disseminated cryptococcosis was confirmed postmortem in all. Of the seven patients with definite pulmonary cryptococcosis, four were immunosuppressed (Table 2). All seven patients had respiratory symptoms (since this was required in the case definition), and hemoptysis was reported in five patients (71.4%). Two of the five patients (40%) tested positive for scrag, and each of these patients had a normal CSF analysis. A chest radiograph demonstrated interstitial and/or pleural effusion in six patients (85.7%), and additional information was CHEST / 115 / 3/ MARCH, 1999 735

Table 1 Disseminated Cryptococcosis* Patient No. Age, yr Race Gender Host Status Respiratory Symptoms CXR CT scrag, titer LP, titer RX F/U, yr 1 75 W F Steroids, RA Chest pain, Nodule ND ND ND None D 1 SOB 2 22 W F ALL Asymptomatic Nodule, interstitial ND 1:1,024 cx-pos, A Fl 7 1:1,024 3 54 W M Renal tx, Fever, SOB Interstitial ND 1:16,000 cx-pos, A 5FC D 2 lymphoma 1:16,000 then Fl 4 63 W F Cirrhosis SOB Pleural effusion Pleural ND NA A 1 effusion 5 56 W M Steroids, Cough, fever, Interstitial ND 1:64 NA A 5FC D 3 Crohn s SOB, chills 6 64 W M Steroids, crescentic GN Cough, SOB Interstitial ND Neg cx-pos, 1:512 A Fl 1 *All patients had C neoformans isolated from blood. CXR chest radiograph; LP lumbar puncture; RX treatment; F/U follow-up; W white; F female; M male; RA seronegative rheumatoid arthritis; ALL acute lymphocytic leukemia; tx transplant; GN glomerular nephritis; SOB shortness of breath; ND not done; Neg negative; NA no abnormalities, culture negative and CSF cryptococcal antigen negative; Cx-pos culture positive; A Fl amphotericin B followed by fluconazole; A amphotericin B; Fl fluconazole; 5FC flucytosine; D 1 autopsy revealed disseminated cryptococcosis including lung tissue; D 2 autopsy revealed disseminated cryptococcosis including lung tissue with immediate death secondary to Gram-negative sepsis; D 3 autopsy revealed disseminated cryptococcosis including lung tissue. Greater than 3 months on glucocorticosteroids at doses above physiological requirements. obtained for all five patients who underwent chest CT scans, including one who had a normal chest radiograph. Five patients (71.4%) received systemic antifungal therapy. Two of the three immunocompetent patients did not receive therapy and had no subsequent clinical sequelae. Of the 15 patients with presumptive pulmonary cryptococcosis, 6 were immunocompromised (Table 3). All but one patient had pulmonary symptoms, two of whom had hemoptysis. One renal transplant patient was asymptomatic; however, this patient did have a cavitary lung lesion without any other etiology and a positive scrag that suggested cryptococcosis. One of five other patients tested for scrag was positive at a titer of 1:4. All of the patients had abnormal chest radiographs, with interstitial and pleural effusion being the most common findings. The CT scans performed on six patients provided no additional information. Three patients were screened with lumbar punctures; all had normal CSF parameters and negative cultures. None of these patients died of cryptococcal infection. Four of the 15 patients (26.6%) were treated, 1 with a surgical pleurectomy and 3 with antifungal agents. Ten patients (71.4%) survived for more than 1 year. The four deaths that occurred in close proximity to the time of the C neoformans isolate were attributed to other etiologies. Table 2 Definitive Pulmonary Cryptococcosis* Patient No. Age, yr Race Gender Host Status Respiratory Symptoms CXR CT scrag, titer LP RX F/U, yr 1 51 W F Normal Cough, SOB Interstitial, Pleural-based nodular ND ND A 3 pleural effusion densities, 2 45 W M Lung tx Cough, SOB, Pleural effusion, ND Neg ND Fl 3 1 2 hemoptysis nodular infiltrate 3 68 W M Steroids Cough, SOB, Patchy consolidation Nodule 1:128 NA Fl 3 hemoptysis 4 49 AA F Liver tx Cough, hemoptysis Lobar consolidation Mass 1:64 NA A Fl 1 5 22 W M Normal Cough, fever, Interstitial Adenopathy Neg ND None 1 hemoptysis 6 56 W M Lung tx Cough, SOB Interstitial infiltrate, ND Neg ND A Fl D(a) pneumothorax 7 69 W M Normal Hemoptysis Normal Lobulated mass, atelectasis ND ND None 4 *AA African American; D(a) autopsy without evidence of cryptococcosis; immediate cause of death secondary to Staphylococcus aureus sepsis; died 7 months after treatment for pulmonary cryptococcosis. For definitions of other abbreviations used, see Table 1. 736 Clinical Investigations

Table 3 Presumptive Pulmonary Cryptococcosis* Patient No. Age, yr Race Gender Host Status Respiratory Symptoms CXR CT scrag, titer LP RX Other Diagnosis F/U, yr 1 82 W M Normal SOB Pleural effusion ND ND ND None COPD 1.5 2 65 W M Normal Hemoptysis Mass Mass Neg NA None Adeno CA Rec 2.5 3 54 W M Steroids SOB Interstitial Adenopathy,, pleural thickening 4 67 H F Ovarian cancer SOB Nodules, pleural effusion Neg NA K CTD 4 ND ND ND None Metastases D 5 76 W M Normal Respiratory arrest Interstitial ND ND ND None ARDS D D 6 74 W M Normal Cough, SOB, fever, sputum 7 38 W M Steroids Cough Interstitial Nodules, cavitary lesions Interstitial Interstitial Neg ND None CHF, pneumonia, ARDS 1:4 ND Fl Alveolar proteinosis 8 36 W F Steroids SOB Pleural effusion ND ND ND None MCTD 7 9 42 W M T-cell Cough, hemoptysis Nodule ND Neg NA None MAI 6 10 70 W M Normal Cough, SOB Pleural effusion, atelectasis 11 80 W F Normal SOB Interstitial, pleural effusions Loculated pleural effusion 12 75 W M Normal Cough Nodule Infiltrates, lobar consolidation 13 72 W F Normal SOB Interstitial, pleural effusions ND ND Pleur Mesothelioma 4 ND ND ND None Sepsis, pneumonia ND ND None Adeno CA 2 ND ND ND None CHF, MI D 14 54 W M Lung tx SOB Interstitial ND ND ND None Bacterial pneumonia 15 63 W M Renal tx Asymptomatic Cavitary lesion Cavitary lesion, 1 2 2.5 1:4 NA A Fl None 1 *H Hispanic; K ketoconazole; CTD connective tissue disease; MCTD mixed connective tissue disease; MAI Mycobacterium avium-intracellulare; CHF congestive heart failure; MI myocardial infarction; adeno CA adenocarcinoma of lung; D died in 1 year; 1 less than 1 year since diagnosis; Rec time to recurrence; pleur pleurectomy; T-cell immunodeficiency without HIV infection. For definitions of other abbreviations used, see Tables 1 and 2. CHEST / 115 / 3/ MARCH, 1999 737

Of the 14 patients classified as having a C neoformans colonization, 7 were immunosuppressed (Table 4). Five patients (36%) were recipients of solid organ transplants. The median age was 63 years (range, 22 to 77 years). Only four patients (26.6%) had respiratory symptoms. There were no other positive scrag results among the patients categorized as having a C neoformans colonization. The CSF was normal in the five patients tested. Two patients were treated with systemic antifungal therapy. All four of the transplant patients did not receive therapy. Four patients (26.6%) in this category died; none of the deaths were attributed to cryptococcosis (two of which were confirmed at autopsy). A scrag study was performed in 18 of the 42 patients reviewed and was positive in 7 patients. Among these 7 patients, a disseminated infection were proven by culture in three patients. A negative test was observed in 11 patients, one of whom had proven dissemination. Fifteen patients underwent a lumbar puncture for routine analysis that included a CSF cryptococcal antigen and a fungal culture as part of their evaluation; cryptococcal meningitis was diagnosed in three patients, all of whom had positive blood cultures for C neoformans. Using univariate analysis, neither the baseline demographic parameters nor the immune status appeared to discriminate the patients with disease from the patients with colonization. Discussion Cryptococcosis is usually a systemic invasive fungal infection that typically presents as meningitis. In such circumstances, the diagnosis and treatment are relatively straightforward. It is assumed that C neoformans generally gains access to the host by inhalation and that the respiratory system is the usual portal of entry. The evaluation of the significance of respiratory cryptococcal isolates becomes an important part of the decision-making process. Critical decisions in the clinical management of such patients include the following: (1) an assessment of whether cryptococcal disease is present in the lung or elsewhere, (2) a determination of the degree of workup that is appropriate for asymptomatic hosts, and (3) the establishment of the necessity for treatment. These determinations have been difficult to make for a number of reasons. First, because the condition is relatively unusual, there is little or no prospective data. Second, the decision-making process has often been influenced by the presence or absence of concomitant immunosuppression. Finally, most of the reports have come from the era before the availability of azole antifungals, and therapeutic guidelines may have been influenced by the relative toxicity of amphotericin B. Our retrospective study suggests that C neoformans plays a pathogenic role in most of the patients in whom it is isolated from a respiratory source. In the majority of these patients, the lung appeared to be the sole organ involved, and a workup for systemic infection was rarely helpful. This conclusion may be biased because the patients were not evaluated or treated in a systematic manner; therefore, not all of the patients received an appropriate or complete workup for systemic infection. However, a long-term follow-up (range, 1 to 7 years) is available for the majority of the patients, and it does not appear that we have missed any cases of subsequent disseminated disease. The immune status did predict dissemination: disseminated cryptococcosis occurred in 6 of the 24 immunocompromised HIV-negative patients (25%) with pulmonary isolates, as opposed to none of the 18 immunocompetent patients. Interestingly, the presence or absence of pulmonary disease was not correlated with immune function. How much evaluation should be performed in patients with pulmonary cryptococcosis? The current general practice is to proceed with a workup that includes serologic and CSF analyses for cryptococcal meningitis. Certainly there are data 4 6 that a screening scrag in febrile patients with advanced HIV disease is useful and that high titers are suggestive of invasive disease. This correlation of scrag and systemic cryptococcal disease is less certain in patients without AIDS. Although the numbers screened in this fashion in our study are small, a positive test does not appear to be specific for dissemination. The recommendation that all patients with a respiratory cryptococcal isolate undergo a lumbar puncture 7,8 is more controversial. Our limited data would suggest that this is not necessary in all patients, and we would limit a more extensive diagnostic evaluation (with blood cultures, serum antigen testing, and lumbar puncture) to patients whose immune status puts them at a higher risk for dissemination. Even if that workup is negative, it is important to recognize that in immunosuppressed patients the progression to a disseminated infection can still occur. The patients in our cohort were not treated in any systematic fashion; in fact, the majority of them were not treated. Thus, it is difficult to make definitive conclusions about therapy. Based on a review of 41 cases seen before 1982, Kerkering and colleagues 2 concluded that immunocompetent hosts with isolated pulmonary cryptococcosis do not require antifungal therapy. We concur that this approach is very 738 Clinical Investigations

Table 4 Pulmonary Colonization of Cryptococcosis* Patient No. Age, yr Race Gender Host Status Respiratory Symptoms CXR CT scrag LP RX Other Diagnosis F/U, yr 1 22 W F Steroids Asymptomatic Nodule Nodule Neg NA Fl SLE 2 2 77 W M Normal Tachypnea, fever Interstitial infiltrate ND ND NA None Aspiration D(a) pneumonia 3 77 W M Normal Thick tracheal secretions 4 76 W F NHL Tachypnea, fever, SOB Normal ND ND ND None Epidermoid cancer Interstitial, 5 30 AA F Normal Asymptomatic Interstitial 6 63 W M Lung tx Asymptomatic Interstitial Interstitial, pleural effusion, atelectasis, ND ND None NHL D(a) Adenopathy ND NA None Sarcoid (CNS) 4 ND ND ND None PCP 4 7 66 W F Normal Asymptomatic Nodule Cavitary lesion ND ND Lobectomy Adeno cancer 1.5 8 32 W M Lung tx Asymptomatic Interstitial infiltrate ND Neg ND None Lung tx 2 9 41 W M Cirrhosis Asymptomatic Nodule Nodule, interstitial Neg NA Fl None 1 infiltrate, 10 73 W M Normal Asymptomatic Nodule Nodule ND ND None Metastatic adeno CA 11 56 W M Liver tx Asymptomatic Interstitial, pleural effusion ND ND ND None Liver failure D 12 66 W F Normal Cough Normal Pleural thickening ND ND None None 1.5 13 56 W F Lung tx Asymptomatic Interstitial ND ND ND None Lung tx 1 14 43 ME M Normal Asymptomatic Patchy density Nodule, ND ND Wedge resection 1 D None 1 *SLE systemic lupus erythematosus; NHL non-hodgkin s lymphoma; PCP Pneumocystis carinii pneumonia; ME middle-eastern. For definitions of other abbreviations used, see Tables 1, 2, and 3. CHEST / 115 / 3/ MARCH, 1999 739

reasonable for asymptomatic patients. Patients who are symptomatic, patients who have a positive scrag, and patients who have underlying immunological disorders should be treated. The optimal therapy depends on the severity of illness, and azole therapy is probably sufficient in most of the cases. However, as this study demonstrates, there is a need for prospective studies to address with greater clarity which patients are at risk for developing disease, and what therapeutic regimens are the most effective. References 1 Rozenbaum R, Goncalves AJR. Clinical epidemiological study of 171 cases of cryptococcosis. Clin Infect Dis 1994; 18:369 380 2 Kerkering TM, Duma RJ, Shadomy S. The evolution of pulmonary cryptococcosis: clinical implications from a study of 41 patients with and without compromising host factors. Ann Intern Med 1981; 94:611 616 3 Miller GPG. The immunology of cryptococcal disease. Semin Respir Infect 1986; 1:45 52 4 Kovacs JA, Kovacs AA, Polis M, et al. Cryptococcosis in the acquired immunodeficiency syndrome. Ann Intern Med 1985; 103:533 538 5 Eng RH, Bishburg E, Smith SM, et al. Cryptococcal infections in patients with acquired immunodeficiency syndrome. Am J Med 1986; 81:19 23 6 Chuck SL, Sande MA. Infections with Cryptococcus neoformans in the acquired immunodeficiency syndrome. N Engl J Med 1989; 321:794 799 7 Mitchell TG, Perfect JR. Cryptococcosis in the era of AIDS: 100 years after the discovery of Cryptococcus neoformans [review]. Clin Microbiol Rev 1995; 8:515 548 8 Kauffman CA. Opportunistic fungal infections: cryptococcosis. J Respir Dis 1995; 16:1074 1079 740 Clinical Investigations