Original Articles. Pathologic Features of Mycobacterium kansasii Infection in Patients With Acquired Immunodeficiency Syndrome

Transcription

1 Original Articles Pathologic Features of Mycobacterium kansasii Infection in Patients With Acquired Immunodeficiency Syndrome Michael B. Smith, MD; Claudia P. Molina, MD; Vicki J. Schnadig, MD; Michael C. Boyars, MD; Judith F. Aronson, MD Context. Mycobacterium kansasii is a slow-growing photochromogenic mycobacterium that may infect patients with human immunodeficiency virus (HIV) late in the course of acquired immunodeficiency syndrome (AIDS). The clinical features of pulmonary and extrapulmonary infections have been described in the literature; however, the pathology of infection has not been adequately addressed. Objective. This report describes the pathologic features of 12 cases of M kansasii infection in patients with AIDS. Design. The medical records, autopsy protocols, cytologic material, and histologic material from patients with AIDS and concomitant M kansasii infection at a tertiarycare medical center during were reviewed. Results. Twelve cases were identified, 6 by autopsy, 5 of which were diagnosed postmortem. Four of the 12 cases had cytologic material and 4 cases had histologic biopsies available for review. Pulmonary infection was most common (9/12), and all patients in whom thoracic lymph nodes were assessed showed involvement (7/7). Abdominal infection was less frequent, with only 1 of 6, 2 of 6, and 2 of Mycobacterium kansasii is a photochromogenic, slowly growing mycobacterium species that can cause pulmonary infection in patients with predisposing lung diseases, such as chronic obstructive pulmonary disease, bronchiectasis, pneumonoconiosis, previous tuberculosis, or bronchogenic carcinoma. 1 Disseminated infections, primarily in immunodepressed patients, also occur. 1,2 Mycobacterium kansasii infections show an increased incidence in urban areas, and in the United States, the highest incidence occurs in the central states and across the southern states from coast to coast. 3 Infections in patients with the acquired immunodeficiency syndrome (AIDS) also occur, and M kansasii has been reported to infect AIDS patients with an incidence of 0.14% in the United States, with a higher (0.44%) incidence in the central and southern sections of the country. 4 Of nontuberculous mycobacteria, it is second only to Mycobacterium avium complex (MAC) as a cause of dissemi- 6, demonstrating liver, spleen, and abdominal lymph node infection, respectively. Isolated infections without documented pulmonary infection included brain abscess (n 1), ulnar osteomyelitis (n 1), and paratracheal mass (n 1). Cytologic and histologic material showed a wide range of inflammatory reactions, including granulomas with and without necrosis, neutrophilic abscesses, spindlecell proliferations, and foci of granular eosinophilic necrosis. The M kansasii bacillus was characteristically long, coarsely beaded, and frequently showed folded, bent, or curved ends. Intracellular bacilli were randomly or haphazardly distributed within histiocytes. Conclusion. Mycobacterium kansasii infection produces predominately pulmonary infection in late-stage AIDS with a high incidence of thoracic lymph node involvement and a much lower incidence of dissemination to other sites. Infection is manifest as a wide variety of inflammatory reactions on cytology and histology; however, the characteristic appearance of the bacillus on acid-fast bacilli stain and its intracellular arrangement in histiocytes can allow a presumptive identification. (Arch Pathol Lab Med. 2003;127: ) nated mycobacterial disease in this setting. 4 In addition to pulmonary and disseminated infection, it has also been reported to cause diverse focal infections, such as osteomyelitis, gastrointestinal infection, gingivomandibular disease, appendicitis, and brain abscess. 3,5 8 Although M kansasii represents an important cause of infection in AIDS patients, the literature to this point has dealt with the clinical aspects, and the pathologic manifestations of the infection have been mentioned only briefly or not at all The morphology of M kansasii has been briefly described in the cytology literature; however, to the best of our knowledge, a systematic review of pathologic findings, including patterns of organ involvement and histologic or cytologic findings in the setting of AIDS, does not exist in the literature. 17 To address this discrepancy, we reviewed all the cases of AIDS-associated, culture-proven M kansasii infection that had available autopsy, cytology, and surgical pathology material at our institution, a tertiary-care medical center serving a large population of AIDS patients. Accepted for publication November 25, From the Departments of Pathology (Drs Smith, Molina, Schnadig, and Aronson) and Internal Medicine (Dr Boyars), University of Texas Medical Branch, Galveston. Reprints: Michael B. Smith, MD, Department of Pathology, University of Texas Medical Branch, 301 University Blvd, Galveston, TX ( mismith@utmb.edu). MATERIALS AND METHODS Culture records in the mycobacteriology laboratory at the University of Texas Medical Branch, Galveston, were reviewed for the years , and cultures positive for M kansasii were identified. These cultures were correlated with archive files for autopsies, surgical pathology cases, and cytology cases for the 554 Arch Pathol Lab Med Vol 127, May 2003 Mycobacterium kansasii in AIDS Patients Smith et al

2 Patient No. Table 1. No. of CD4, Cells/ L Patients With Acquired Immunodeficiency Syndrome Infected With Mycobacterium Kansasii* Comorbid Pulmonary Condition Pulmonary Infection Other Organs Infected 1 Unknown None Yes Thoracic and abdominal lymph nodes, skin pack/year 3 14 duration unknown 4 9 Smoking history duration unknown No Thoracic and abdominal lymph nodes Coinfections/Comorbid Conditions Pseudomonas aeruginosa bronchopneumonia, PCP Pseudomonas aeruginosa GI ulcers and sepsis, MAC Outcome Yes None HCV Resolved on Yes Thoracic and abdominal lymph nodes, liver, spleen, bone marrow Disseminated Histoplasma capsulatum, HCV 5 21 None Yes Thoracic lymph nodes B-cell lymphoma of the brain, HCV 6 3 Smoking history Yes Thoracic lymph Staphylococcus aureus duration unknowsis, nodes bronchopneumonia/sep- MAC, CMV, HCV, hepatitis B 7 28 None Unknown Right ulna None Resolved on ; died 1.5 years later of Enterococcal sepsis 8 7 COPD and 75 pack/year pack/year Yes Thoracic lymph nodes and spleen 10 9 None Yes Cervical lymph nodes pack/year Asthma and duration unknown CMV, aspiration pneumonia Yes Tracheal mass HCV Resolved on Thrush, HSV-2 Resolved on No Brain HCV Resolved on Role of M kansasii in Outcome Incidental finding Primary cause of death None Yes None None Continued ; no evidence of disease 1 year after diagnosis * COPD indicates chronic obstructive pulmonary disease; PCP, Pneumocystis carinii pneumonia; GI, gastrointestinal; MAC, Mycobacterium avium complex; HCV, hepatitis C virus; CMV cytomegalovirus; and, not available. same period, and cases from patients with concomitant AIDS and M kansasii infection were identified. Available hematoxylin-eosin, Ziehl-Neelsen acid-fast, Fite modified acid-fast, Gomori methenamine silver (GMS), and Brown-Brenn Gram-stained histology sections were reviewed. If slides were not available, paraffin blocks were recut and stained as appropriate and examined. For cytology specimens, Papanicolaou, modified rapid Romanowsky stain, Kinyoun acid-fast stain, and GMS-stained slides were reviewed. Patients charts were reviewed for demographic data, past medical history, and clinical history at the time of presentation of M kansasii infection. RESULTS Twelve cases were identified (Table 1). Ten of the 12 patients were men and 2 were women. The median age was 36 years (range, years). Ten patients were African American, 1 was white, and 1 was Hispanic. Risk factors for HIV infection were intravenous drug abuse in 6 cases, intravenous drug abuse and homosexuality in 1, homosexuality in 1, heterosexual sex with an infected partner in 2, and in 2 the risk factors were unknown. Eight of the 12 admitted to tobacco use, although quantitation was recorded for only 4, who had 20, 20, 25, and 75 packyear histories of smoking. Two patients were listed as having a chronic lung disease, with chronic obstructive pulmonary disease in 1 and asthma requiring bronchodilator in 1 (Table 1). Occupational exposure to or a profession associated with chronic dust exposure was not recorded in any of the patient s charts. The median CD4 count for the patients in whom the information was available (n 11) was 14/ L ( /L) with a range of 3 83/ L ( /L), and the mean CD4 count was 24/ L ( /L). Of the 12 patients, 7 are known to have died, 6 of whom had autopsies. Of these 7, M kansasii infection was the primary cause of death in 1 patient, a contributing factor in 4, and an incidental finding in 1. One patient died of unrelated causes after treatment and apparent resolution of her infection. 18 Arch Pathol Lab Med Vol 127, May 2003 Mycobacterium kansasii in AIDS Patients Smith et al 555

3 Cytologic Findings Five cytologic preparations from 4 patients, including 1 bronchoalveolar lavage, 1 bronchial wash, 1 sputum, 1 fine-needle aspiration (FNA) of an ulnar lesion, and 1 FNA of a cervical lymph node, were available for review. All 4 patients had pulmonary infiltrates by chest radiography on admission to the hospital, but only 3 gave a history of pulmonary symptoms. In these 3 patients, M kansasii bronchopneumonia was confirmed by bronchoscopy in 2 and by a sputum in 1. For the fourth patient, a case of ulnar osteomyelitis (patient 7), a bronchopneumonic process seen on a radiograph was not investigated. All of the cases showed a mixed inflammatory infiltrate composed of varying combinations of neutrophils, necrosis, macrophages, and lymphocytes. Two cases, the ulnar FNA and the cervical lymph node FNA, showed granulomas without giant cells in a background of necrosis, neutrophils, and nuclear debris. In the ulnar FNA granulomas were plentiful, and in the lymph node FNA granulomas were infrequent (Figure 1). Kinyoun acid-fast stain demonstrated long slender bacilli, which were often curved or folded and showed irregular coarse beading. Bacilli were numerous in the FNAs of the ulna, lymph node, and sputum, but much less so in the bronchoalveolar lavage and bronchial wash specimens. In those specimens with a large number of histiocytes, bacilli were predominately intracellular, while in the specimens with extensive necrosis, the bacilli were found mostly in an extracellular location (Figure 2). The bacilli were seen with GMS stain in 3 cases and showed an identical morphology, but beading was less prominent than with the Kinyoun stain (Figure 3, a). Conventional Gram stain was performed in only 1 case (FNA of the lymph node), and the bacilli did not stain. Negative images of the bacilli were seen in 2 cases, in the FNA of the ulna on Papanicolaou stain and in the FNA of the lymph node on Romanowsky stain (Figures 3, b, and 4). Autopsy and Histologic Findings A complete postmortem examination was performed in 6 cases. Chest radiography showed infiltrates consistent with pneumonia in 3 of 4 patients for whom radiologic records were available. The diagnosis of M kansasii infection was made postmortem in all patients except one (patient 1), who had been diagnosed as having M kansasii bronchopneumonia with dissemination to the skin 5 months prior to death; the patient had been treated with isoniazid, rifampin, and pyrazinamide during this period. Pulmonary involvement was identified in 5 of 6 patients. In 4 of the 5 cases, variably sized foci characterized by a large central area of noncaseous necrosis with extensive nuclear debris and neutrophils were seen. The rim of the foci consisted either of epithelioid histiocytes (with a spindled appearance in some cases) admixed with large histiocytes with foamy cytoplasm or of fibrosis (Figure 5). Multinucleated giant cells were not present. Acid-fast bacilli (AFB) were numerous, both in an extracellular location in the necrosis and in an intracellular location in histiocytes. The staining pattern, morphology, and arrangement in cells of the bacilli was relatively uniform from case to case. Bacilli were well-stained with the Ziehl-Neelsen stain, but stained more intensely and in greater numbers with the Fite stain. Organisms were seen with the Brown-Brenn modification of the Gram stain (Figure 6, a), periodic acid Schiff stain with diastase (Figure 6, b), and the GMS stain (Figure 7, a), although the beading characteristic of this species was poorly demonstrated with these stains. Bacilli frequently showed crooked, bent, or folded ends (Figure 7, b) and were arranged in a haphazard manner when seen intracellularly. In the remaining case (patient 4), small granulomas without necrosis were seen scattered throughout the lungs with most granulomas showing numerous intracellular yeast (Histoplasma capsulatum) and only a few intracellular AFB. Lymph nodes were involved in all autopsy cases, including the case that showed no lung involvement at autopsy (patient 2), with perihilar, peribronchial, or mediastinal nodes infected. Patient 2 and 1 other patient (patient 4) also showed involvement of abdominal lymph nodes. Three primary patterns of inflammatory reaction to the infection were seen in lymph nodes, although a single pattern predominated in a given patient. Patterns included granulomatous foci with necrosis and fibrosis, as described in the lung; foci of eosinophilic granular necrosis with a few small areas of viable epithelioid histiocytes (Figure 8); and histiocytes with a spindle-cell pseudotumor pattern with scattered clusters of neutrophils (Figure 9). When present in histiocytes with a spindle cell morphology, the bacilli were arranged in parallel, but the bent or crooked ends were still demonstrable. Four of the autopsy cases showed small, well-formed granulomas in the liver, with a predominant location in portal areas, although only 1 patient showed the scattered intracellular bacilli characteristic of M kansasii (patient 4) in the granulomas. Similarly, bacilli with this appearance were seen in granulomas in the spleens of patients 4 and 8. No necrosis was associated with the granulomas in patient 8, and while patient 4 did show necrosis, this patient also had extensive involvement of the spleen with Hcapsulatum. The bone marrow was examined in 3 of the autopsied patients, and AFB were not seen in any of the cases, although granulomas with H capsulatum were noted in patient 4 and M kansasii was cultured from the bone marrow of this patient at autopsy. The surgical pathology specimens were a transbronchial biopsy, brain biopsy, skin biopsy, and tracheal biopsy. These specimens showed granulomatous inflammation with a mononuclear infiltrate in the case of the first two and a neutrophilic infiltrate in the case of the tracheal biopsy. Acid-fast bacilli were rare in the transbronchial biopsy and frequent in the brain and tracheal biopsies. All AFB were intracellular and were present singly or in small clusters. Granulomatous inflammation with areas of eosinophilic necrosis containing cellular debris were present in the dermis and subcutaneous fat of the skin biopsy, along with areas showing a spindle cell histology. Acid-fast bacilli were numerous, both intracellularly in histiocytes and extracellularly in necrotic areas. In 1 case (patient 5), periaortic lymph nodes demonstrated involvement with a mycobacterium distinct from M kansasii; this organism was short, straight, and beaded, and was present in foamy histiocytes diffusely effacing normal lymph node architecture, characteristic of MAC. The mycobacteria in the granulomas in the liver of patients 1 and 5 and spleen of patient 5 also demonstrated AFB with this morphology. 556 Arch Pathol Lab Med Vol 127, May 2003 Mycobacterium kansasii in AIDS Patients Smith et al

4 Figure 1. Mixed inflammatory infiltrate with necrosis in a fine-needle aspiration of a cervical lymph node (patient 10) (Papanicolaou, original magnification 400). Figure 2. Numerous long, beaded acid-fast bacilli present in histiocytes and extracellularly in a fine-needle aspiration of a cervical lymph node (patient 10) (Kinyoun, original magnification 400). Figure 3. Fine-needle aspiration of ulna (patient 7). a, Beaded appearance and curved ends of bacillus seen on Gomori methenamine silver stain (original magnification 1000). b, Negative images of bacilli can sometimes be seen even in fixed, Papanicolaou-stained material (original magnification 1000). Figure 4. Fine-needle aspiration of cervical lymph node (patient 10). a, Negative images seen on air-dried rapid Romanowsky-stained smear (original magnification 720). b, Large histiocytes with numerous intracellular negative images (original magnification 720). Figure 5. Abscess from the lung of patient 8 (hematoxylin-eosin, original magnification 125). Figure 6. Lung (patient 8). a, Mycobacterium kansasii stains gram-positive (Brown-Brenn, original magnification 1000). b, Bacilli are also stained by periodic acid Schiff with diastase (original magnification 1000). Arch Pathol Lab Med Vol 127, May 2003 Mycobacterium kansasii in AIDS Patients Smith et al 557

5 Figure 7. a, Staining of bacilli by Gomori methenamine silver was less prominent and did not demonstrate prominent beading like that seen on cytologic preparations (lung, patient 8) (original magnification 720). b, The characteristic appearance of the Mycobacterium kansasii bacillus seen in a brain biopsy from patient 11 (Ziehl-Neelsen, original magnification 1000). Figure 8. Eosinophilic necrosis with nuclear debris seen in lymph node of patient 2 (hematoxylin-eosin, original magnification 125). Figure 9. Spindle cell proliferation in a thoracic lymph node from patient 8 containing Mycobacterium kansasii with characteristic morphology (inset) (hematoxylin-eosin, original magnification 125; inset, Ziehl-Neelsen, original magnification 500). COMMENT The AIDS patients in this series demonstrated a clinical profile consistent with that described in previously published clinical series of AIDS patients infected with M kansasii, that is, late-stage AIDS (mean CD4 count of 24/ L [ /L]) in our series) with a majority of patients ( 90% of cases in most series) suffering from pulmonary infection (9/12 of our patients). Cavitation of the lung was described as occurring with high frequency in the first case series published by Levine et al 5 ; however, this finding has not been confirmed in subsequent reports, and none of our patients showed cavitary pneumonia. 3,10,11 Involvement of organs other than the lung has been reported to occur in 5% to 40% of patients in most series and was frequent in our group of patients (Table 1) All the patients autopsied showed hilar or mediastinal lymph node involvement, even the patient in whom lung infection by M kansasii could not be identified. This finding contrasts with reports that suggest that thoracic lymph node involvement is infrequent with this infection in AIDS patients. 16 Abdominal lymph node involvement, however, proved to be less frequent and was identified in only 2 of 6 autopsied patients in our series. The reduced frequency of abdominal node involvement relative to thoracic node involvement and the infrequency of confirmed liver involvement (only 1 patient in this series) would suggest that acquiring the organism via the gastrointestinal route from water, with subsequent spread either through the abdominal lymphatics or portal venous system, is not a major route of infection and dissemination, as has been suggested. 9,10 A gastrointestinal route of infection has been associated with MAC, and gastrointestinal infection is more frequent than pulmonary disease in AIDS patients with MAC. 19 With the exception of a single case series of 3 patients who were reported to have duodenal infection and a case of appendicitis, reports of M kansasii causing gastrointestinal disease are unusual relative to the great preponderance of patients with pulmonary infection. 7,20 None of the 6 patients autopsied in our series showed gastrointestinal infection. The suggestion that M kansasii isolated from stool may serve as a predictor of disease, similar to the strong correlation seen with MAC, has been proposed based on isolation in stool from up to one third of infected patients. 10 The differences in the frequency of involved organs with M kansasii and MAC when infection is present, however, suggest that such a correlation may not have a pathophysiologic basis. Presumptive differentiation of MAC and the other mycobacterium commonly infecting AIDS patients, Mycobacterium tuberculosis complex (MTBC), from M kansasii on cytologic or histologic specimens can have important therapeutic and public health implications. While culture is required for definitive speciation, all of these species are slow growing, and it may require several weeks for a laboratory to grow and identify the species. Additionally, coinfections with several species are not rare, and in one series, 13 of 49 patients with M kansasii infection also were infected with MAC. 9 Inourseries,4ofthe6autopsied patients showed evidence of MAC coinfection, 2 who were culture positive for MAC and 2 patients who showed histologic evidence of MAC infection. Since for MAC usually involves one of the newer macrolides and for M kansasii involves rifampin and ethambutol, sometimes with high-dose isoniazid, recognition that multiple 558 Arch Pathol Lab Med Vol 127, May 2003 Mycobacterium kansasii in AIDS Patients Smith et al

6 Table 2. Summary of Cytologic and Histologic Features of M. kansasii infection in Acquired Immunodeficiency Syndrome Patient No. Specimen Reviewed* Predominant Histologic Reaction 1 Premortem skin biopsy, autopsy Skin: mixture of abscesses, granulomas, and spindle cell proliferations Autopsy: abscesses and granulomas 2 Autopsy Foci of granular, eosinophilic necrosis with nuclear debris and clusters of epithelioid histiocytes 3 Bronchoalveolar lavage Mixed inflammatory infiltrate, neutrophil predominant 4 Autopsy Granulomas with a neutrophilic infiltrate 5 Autopsy Abscesses and spindle cell proliferations 6 Autopsy Granulomas with necrosis, neutrophils, and nuclear debris 7 FNA of ulna Granulomas with necrosis, neutrophils, and nuclear debris 8 Autopsy Abscesses, spindle cell proliferation, and granulomas 9 Biopsy of tracheal mass Granuloma with neutrophilic infiltrate 10 Sputum, FNA of cervical lymph nodes Sputum: mixed inflammatory infiltrate, neutrophil predominant FNA: mixed inflammatory infiltrate with neutrophils and necrosis predominant; rare granulomas 11 Biopsy of brain abscess Granuloma with a lymphocytic infiltrate 12 Bronchial wash, transbronchial biopsy Wash: mixed inflammatory infiltrate, lymphocytic and macrophage predominant Biopsy: granulomas with a lymphocytic infiltrate * FNA indicates fine-needle aspiration. species are present prior to culture identification can also have important therapeutic and prognostic implications. Differentiation of M kansasii from MAC in cytologic preparations is assisted by the longer length of M kansasii, more prominent beading on acid-fast stain, and lack of necrosis with MAC in most cases. 17 In our experience, the tendency of the bacillus of M kansasii to fold or bend back on itself, a feature than can also be appreciated when bacilli are identified as negative images, is extremely helpful. The longer length and folding are also helpful in differentiating M kansasii from MTBC, although a similar type of necrosis, that is, granular with neutrophils and nuclear debris, can be seen with both organisms in the setting of AIDS. The histologic morphology of the M kansasii bacillus is similar and useful in presumptively identifying it in histologic specimens. The histologic inflammatory reactions associated with infection are protean (Table 2) and overlap some of the reactions associated with both MAC and MTBC in the setting of AIDS. Five types of reactions were seen in this study: abscesses; granulomas without giant cells but with large areas of central eosinophilic necrosis with numerous neutrophils and nuclear debris; well-organized granulomas without giant cells or necrosis, but with a mononuclear cell infiltrate; areas of eosinophilic granular necrosis with scattered clusters of epithelioid histiocytes; and spindle-cell proliferations with scattered clusters of neutrophils. A sixth pattern that was not seen in our study, that of sheets of foamy histiocytes containing AFB, was described by Marinelli et al 21 in their description of a single case. As granulomas with noncaseating necrosis and neutrophils can be seen with MTBC, granulomas without necrosis with MTBC or MAC, sheets of foamy macrophages with MAC, and spindle-cell pseudotumors with either MAC or MTBC, 22,23 presumptive identification Table 3. Organs Involved in Mycobacterium kansasii Infection Organ system involved Pulmonary Thoracic lymph nodes Abdominal lymph nodes Liver Spleen Bone marrow Skin Brain Bone Blood No. of Patients Infected/No. of Patients Examined* 9/12 7/7 2/6 1/6 2/6 1/3 1/6 1/7 1/1 3/12 * Includes histologic, cytologic, or radiologic study with culture or culture alone. Five patients showed bilateral pneumonia, and 3 patients showed unilateral pneumonia. Patient s bone marrow was culture positive for M kansasii. of M kansasii is best accomplished by observing both the morphology of the bacilli and their arrangement in histiocytes when intracellular. The morphology of relatively long, beaded AFB, often with bent or folded ends, is demonstrable even when bacilli are present in spindle cells. Additionally, the bacilli are often arranged in a haphazard, intersecting orientation when seen in large nonepithelioid histiocytes, rather than the parallel stacking associated with MAC. The utilization of special stains, such as GMS, tissue Gram stain, and periodic acid Schiff, does not appear to be helpful, except to further delineate morphology of the bacilli if necessary, as all mycobacteria can potentially stain with GMS and Gram, and both MAC and Arch Pathol Lab Med Vol 127, May 2003 Mycobacterium kansasii in AIDS Patients Smith et al 559

7 M kansasii stain with periodic acid Schiff, although staining appears more intense with MAC. In summary, M kansasii is an important mycobacterial pathogen infecting patients late in the course of AIDS. Although most frequently involving the pulmonary system and hilar/mediastinal lymph nodes, infection can occur in a variety of locations, sometimes without discernible pulmonary involvement. Widespread dissemination to many organs in a given case was not seen in our series (Table 3). Although the cytologic and histologic manifestations of infection are diverse, knowledge of the possible pathologic appearances of the inflammatory reaction and the unique morphology of the bacillus can allow the pathologist to render a presumptive diagnosis long before the slow-growing species is identified by culture, facilitating appropriate. References 1. Lillo M, Orengo S, Cernorch P, Harris R. Pulmonary and disseminated infection due to Mycobacterium kansasii. Rev Infect Dis. 1990;12: Jacobson K, Teira R, Libshitz H, et al. Mycobacterium kansasii infections in patients with cancer. Clin Infect Dis. 2000;30: Bloch K, Zwerling L, Pletcher M, et al. Incidence and clinical implications of isolation of Mycobacterium kansasii: results of a 5-year, population-based study. Ann Intern Med. 1998;129: Horsbaugh C, Selik R. The epidemiology of disseminated nontuberculous mycobacterial infection in the acquired immune deficiency syndrome (AIDS). Am Rev Respir Dis. 1989;139: Levine B, Chaisson R. Mycobacterium kansasii: a cause of treatable pulmonary disease associated with advanced human immunodeficiency virus (HIV) infection. Ann Intern Med. 1991;114: Courrier B, Mauprivez C, Coignard S, et al. Gingivomandibular infection due to Mycobacterium kansasii in a patient with AIDS. Clin Infect Dis. 1998;26: Enani M, Frayha H, Halim M. An appendiceal abscess due to Mycobacterium kansasii in a child with AIDS. Clin Infect Dis. 1998;27: Gordon S, Blumber H. Mycobacterium kansasii brain abscess in a patient with AIDS. Clin Infect Dis. 1992;14: Witzig R, Fazal B, Mera R, et al. Clinical manifestations and implications of coinfection with Mycobacterium kansasii and human immunodeficiency virus type 1. Clin Infect Dis. 1995;21: Klein J, Corbett E, Slade P, Miller R, Coker R. Mycobacterium kansasii and human immunodeficiency virus co-infection in London. J Infect. 1998;37: Valainis G, Cardona L, Greer D. The spectrum of Mycobacterium kansasii disease associated with HIV-1 infected patients. J AIDS. 1991;4: Bamberger D, Driks M, Gupta M, et al. Mycobacterium kansasii among patients infected with human immunodeficiency virus in Kansas City. Clin Infect Dis. 1994;18: Campo R, Campo C. Mycobacterium kansasii disease in patients infected with human immunodeficiency virus. Clin Infect Dis. 1997;24: Pintado V, Gomez-Mompaso E, Martin-Davila P, et al. Mycobacterium kansasii infection in patients infected with the human immunodeficiency virus. Eur J Clin Microbiol Infect Dis. 1999;18: Lortholary O, Deniel F, Boudon P, et al. Mycobacterium kansasi infection in a Paris suburb: comparison of disease presentation and outcome according to human immunodeficiency virus status. Int J Tuberc Lung Dis. 1999;3: Fishman J, Schwartz D, Sais G. Mycobacterium kansasii pulmonary infection in patients with AIDS: spectrum of chest radiographic findings. Radiology. 1997;204: Jannota FS, Sidaway MK. The recognition of mycobacterial infections by intraoperative cytology in patients with the acquired immune deficiency syndrome. Arch Pathol Lab Med. 1989;113: Schnadig VJ, Fiaz S, Boyvat F, Borucki M. Mycobacterium kansasii osteomyelitis presenting as a solitary lytic lesion of the ulna: fine-needle aspiration findings and morphologic comparison with other mycobacteria. Diagn Cytopathol. 1998;19: Wallace JM, Hannah JB. Mycobacterium avium complex infection in patients with the acquired immunodeficiency syndrome. Chest. 1988;93: Sherer R, Sable R, Sonnenberg M, et al. Disseminated infection with Mycobacterium kansasii in the acquired immunodeficiency syndrome. Ann Intern Med. 1986;105: Marinelli DL, Albelda SM, Williams TM, Kern JA, Iozzo R, Miller WT. Nontuberculous mycobacterial infection in AIDS: clinical, pathologic, and radiographic features. Radiology. 1986;160: Sekosan M, Cleto M, Senseng C, Farokan M, Sekosan J. Spindle cell pseudotumors in the lungs due to Mycobacterium tuberculosis in a transplant patient. Am J Surg Pathol. 1994;18: Wood C, Mickloff BJ, Todes-Taylor NR. Pseudotumor resulting from atypical mycobacterial infection: a histoid variety of Mycobacterium avium-intracellulare complex infection. Am J Clin Pathol. 1985;83: Arch Pathol Lab Med Vol 127, May 2003 Mycobacterium kansasii in AIDS Patients Smith et al