Fish Tank Exposure and Cutaneous Infections Due to Mycobacterium marinum: Tuberculin Skin Testing, Treatment, and Prevention

Transcription

1 MAJOR ARTICLE Fish Tank Exposure and Cutaneous Infections Due to Mycobacterium marinum: Tuberculin Skin Testing, Treatment, and Prevention Felicia M. T. Lewis, Bryan J. Marsh, and C. Fordham von Reyn Infectious Disease Section, Department of Medicine, Dartmouth-Hitchcock Medical Center, Lebanon, New Hampshire In the present study, 8 patients with soft tissue infection due to Mycobacterium marinum are described, and contemporary data on treatment are reviewed. Six patients had positive cultures, all patients had cutaneous exposure to fish tanks, 7 had sporotrichoid lesions, and 2 had deep infection. All 7 tested patients had tuberculin skin test reactions 10 mm. Six patients with disease limited to the skin were successfully treated with 2- drug combination therapy, including clarithromycin, ethambutol, and rifampin. Optimal treatment should include 2 drugs for 1 2 months after resolution of lesions, typically 3 4 months in total. Deeper infections may require more prolonged treatment and surgical debridement. Positive tuberculin reactions may be due to infection with M. marinum. Persons with open skin lesions or immunosuppression should avoid cutaneous contact with fish tanks. Mycobacterium marinum is a slow-growing environmental mycobacterium that was first isolated from dead fish in a Philadelphia aquarium in 1926 [1] and was identified as a human pathogen in 1951 after isolation from granulomatous skin lesions in patients from Sweden. M. marinum is distributed widely in aquatic environments [2], especially in relatively still or stagnant water, such as in fish tanks and swimming pools, and in naturally occurring bodies of water [3, 4]. Infection is acquired by direct inoculation with the bacterium through broken skin in an aquatic environment. Adequate chlorination has drastically reduced the number of cases acquired from swimming pools, and recent reports emphasize acquisition from fish tanks but do not offer recommendations for prevention [3, 5 10]. Although more potent antimicrobials are now available for the treatment of M. marinum infection, only Received 20 December 2002; accepted 14 April 2003; electronically published 18 July Reprints or correspondence: Dr. C. F. von Reyn, Infectious Disease Section, Dartmouth-Hitchcock Medical Center, Lebanon, NH (fvr@hitchcock.org) Clinical Infectious Diseases 2003; 37: by the Infectious Diseases Society of America. All rights reserved /2003/ $15.00 limited numbers of cases are available on which to analyze the effects of newer antimicrobials on the outcome of treatment. Large series typically include reviews of cases treated by numerous providers, rather than cases treated by the authors. In addition, although recent studies have emphasized the influence of asymptomatic infection with environmental nontuberculous mycobacteria (NTM), such as Mycobacterium avium complex (MAC), on the tuberculin skin test [11], detailed data have not been presented on the effect of symptomatic infections with NTM, such as M. marinum, on the tuberculin skin test. We present our personal experience with 8 cases of soft-tissue infection due to M. marinum and emphasize the mode of acquisition and implications for prevention, the effect on tuberculin skin testing, and responses to combination antibiotic therapy, including macrolides. We also review other recent studies on treatment to provide recommendations for current management of infections due to M. marinum. METHODS Medical records at the Dartmouth-Hitchcock Medical Center and records of skin test studies from the Infec- 390 CID 2003:37 (1 August) Lewis et al.

2 tious Disease Section were reviewed from May 1990 to 2000 to identify patients with a diagnosis of skin or soft tissue infection due to M. marinum. Data then were abstracted from patient records and reports of laboratory tests. All case patients were seen by one of the authors. Case patients were accepted for inclusion if culture of a cutaneous lesion was positive for M. marinum or if a patient without a culture performed had a characteristic chronic cutaneous lesion with sporotrichoid spread and a plausible aquatic exposure history. The literature on treatment of M. marinum was reviewed by use of MEDLINE for the period using the following keywords: M. marinum, fish tank granuloma, swimming pool granuloma, and aquarium granuloma. Intradermal skin tests were approved by the Dartmouth Committee for the Protection of Human Subjects and were performed with 0.1 ml M. avium sensitin (MAS; 10/2; State Serum Institute) and 0.1 ml (5 TU) Mycobacterium tuberculosis PPD (Connaught Laboratories). Skin tests were performed at the time that patients were seen by one of the authors. Tests were administered intradermally by use of the Mantoux technique and were read by trained personnel h after placement, as described elsewhere [11]. Cultures for M. marinum were performed by use of both solid medium (Middlebrook 7H11; Lowenstein-Jensen) and broth medium (Middlebrook 7H9; MB/BacT) at both 30 C and 35 C when a sufficient sample was available. RESULTS Clinical features. Clinical features of the 8 patients are summarized in table 1. Four cases occurred in men and 4 in women (age range, years; mean age, 45 years). All patients reported cutaneous contact with fish tanks at home or work before the onset of infection. In 1 instructive case (patient 2), a nurse working on a medical ward cleaned a fish tank on several occasions while she had an open lesion on her finger from a recent wart removal. Cultures were positive for M. marinum for all 6 patients who had cultures performed. One patient (patient 7) was diagnosed on the basis of a positive acid-fast bacillus (AFB) stain and typical clinical features. One patient (patient 8) was diagnosed on the basis of a typical cutaneous lesion with sporotrichoid spread and fish tank exposure. Overall, AFB stains were positive in 3 patients, and biopsy specimens showed granuloma formation in 6 patients. Infections began on the finger or hand in all 8 patients, and, in 6 patients, remained confined to the skin. Deep tissue infection developed in 2 patients. Patient 5 had sporotrichoid spread (figure 1) and extensive deep infection, including osteomyelitis. Patient 7 developed tenosynovitis of the right index finger. Six patients (75%) demonstrated sporotrichoid spread with secondary nodules along proximal lymphatic channels. Three patients had significant underlying medical conditions, including both patients with deep infection extending to a site beyond the skin. Patient 5 who had extensive deep disease had psoriasis and was receiving treatment with prednisone. Patient 6 who had cutaneous disease had rheumatoid arthritis and was receiving treatment with plaquenil and nonsteroidal anti-inflammatory agents. Patient 7 who had tenosynovitis had type II diabetes mellitus. Chest radiograph findings were negative for all 4 patients who had x-rays performed in this study (patients 1 and 5 7). Human immunodeficiency virus (HIV) risk factors were not noted in any patient, and HIV testing was not performed. Skin test results. Seven patients had skin testing with PPD and MAS antigens. Of these 7 patients, all 7 (100%) had PPD reactions 10 mm in diameter, and 2 (29%) had reactions 15 mm in diameter. Three (50%) of 6 tested patients had MAS reactions 10 mm in diameter. Three patients had PPD and MAS reactions differing by!5 mm (nondominant), 2 had larger MAS reactions by 5 mm in size, and 1 had a PPD reaction that was greater by 5 mm. None of the patients were foreign born, and there was no known history of bacille Calmette- Guérin (BCG) immunization. Four patients had a history of a previous negative tuberculin skin test result, and 1 patient had a history of a previous positive tuberculin skin test result (patient 6). Treatment results. Treatment regimens and outcomes are summarized in table 1. All 6 patients with cutaneous disease alone responded completely to antibiotic therapy with or without excisional biopsy. Four of these patients received therapy with clarithromycin and ethambutol for 3 6 months. Two received therapy with rifampin and ethambutol, one for 2 months and another for 14 months before resolution. Of the 2 patients who had extension of disease, patient 7 responded to 2 months treatment with of clarithromycin and ethambutol, and patient 5 with metastatic melanoma had not resolved after 2 years of multiple excisions and a drug regimen that included clarithromycin, ethambutol, and rifabutin. For most patients, duration of therapy was typically 1 2 months after resolution of lesions. DISCUSSION The present study demonstrates that cutaneous exposure to fish tanks is the principal contemporary source of M. marinum infections in our inland area of the United States and is consistent with other recently published reports [6, 12]. Fish and shellfish injuries are responsible for a small proportion of cases in other studies, but acquisition from swimming pools is now unusual [6, 12, 13]. As demonstrated by patients 2 and 5 (who had psoriasis), fish tank related infections may be related to immersion of an open cutaneous lesion. Despite the known Fish Tanks, Tuberculin Tests, and M. marinum CID 2003:37 (1 August) 391

3 Table 1. Characteristics of 8 patients with Mycobacterium marinum infection. Patient Age in years/sex Type of exposure Underlying disease or condition Skin test reaction diameter, mm PPD MAS AFB stain/ culture Infection type Pathologic findings Treatment regimen (duration) Outcome 1 25/M Fish tank None ND ND +/+ Cutaneous, sporotrichoid Mild acute inflammation Rifampin and ethambutol (2 months); excisional biopsy 2 38/F Fish tank None /+ Cutaneous, sporotrichoid Granuloma Clarithromycin and ethambutol (4 months) 3 40/M Fish tank None /+ Cutaneous, sporotrichoid Granuloma Clarithromycin and ethambutol (3 months) 4 57/M Fish tank None 20 0 /+ Cutaneous, sporotrichoid Early histiocytic infiltrate 5 55/M Fish tank Psoriasis, melanoma, prednisone therapy 6 36/F Fish tank Rheumatoid arthritis, hydroxychloroquine therapy 11 9 /+ Osteomyelitis, sporotrichoid Clarithromycin and ethambutol (6 months) Granuloma Ciprofloxacin, rifampin, ethambutol, clarithromycin, rifabutin, and amikacin; excision of nodules /+ Cutaneous Granuloma Ethambutol and rifampin (14 months) 7 50/F Fish tank Diabetes mellitus /ND Tendonitis, sporotrichoid Granuloma Clarithromycin and ethambutol (2 months); tenosynovectomy 8 59/F Fish tank None 13 ND ND Cutaneous, sporotrichoid ND Clarithromycin and ethambutol (6 months) Resolving (lost to follow-up) NOTE. AFB, acid-fast bacillus; MAS, Mycobacterium avium sensitin; ND, not done; +, positive;, negative.

4 Figure 1. Mycobacterium marinum nodules and inflammation of proximal interphalangeal joints (top) and wrist and forearm (bottom) in a patient with underlying psoriasis (patient 5). association of fish tanks with M. marinum infections our informal review of national aquarium supply stores indicates that instructions for home fish tanks do not typically recommend avoidance of water contact or the use of waterproof gloves for persons with open skin lesions. M. marinum should be suspected when a characteristic skin lesion occurs 8 30 days (median, 21 days; 30 days in 35% of cases) after fish tank exposure [6]. Primary lesions typically present as a bluish-red papule, nodule, or plaque with a verrucous surface, most often at the site of a small abrasion or cut on the dominant hand [14, 15]. As in the present study, most patients have a sportotrichoid presentation on the upper extremity [15]. Lesions may have purulent drainage, and satellite lesions may appear and coalesce. Lesions are painful in fewer than one-half of cases [16], lymphadenopathy is rare and typically mild [3, 17], and systemic symptoms are distinctly unusual [18]. Infection is usually confined to the skin, because of preferential growth of M. marinum at temperatures!37 C Fish Tanks, Tuberculin Tests, and M. marinum CID 2003:37 (1 August) 393

5 [19]. However, as demonstrated here, deeper extension of infection may lead to tenosynovitis, arthritis, bursitis, and osteomyelitis [20 24]. Disseminated infection is rare and almost always occurs in immunosuppressed individuals [18, 25, 26]. Our experience demonstrates that infection with these organisms also is associated with sufficient tuberculin skin test reactivity to trigger consideration of treatment for latent tuberculosis. This issue has not been emphasized in recent studies. Jolly and Seabury [27] reported tuberculin reactions 10 mm in diameter in 14 (67%) M. marinum infected patients tested with PPD-S, and Mollohan and Romer [7] reported positive reactions to the relatively nonspecific Vollmer patch test in 55 (77%) of 71 patients. In our study, all tested patients had tuberculin reactions 10 mm in diameter (some would be candidates for treatment of latent tuberculosis if they were in defined risk groups), and 2 had reactions 15 mm in diameter (both would be candidates for treatment of latent tuberculosis) [28]. Although one-half our patients did not have preinfection skin tests to prove unequivocally that their reactions were due to M. marinum, most healthy adults living in the United States have negative reactions to tuberculin [11], and infection with various NTMs often results in tuberculin reactivity [29]. Evaluation for tuberculosis always should be performed in those with positive tuberculin skin test results, and treatment should be provided on the basis of current guidelines [28]. However, positive tuberculin reactions in patients with M. marinum infection can be attributed to this infection and should not always be considered an indication for treatment of latent tuberculosis. The cross -reactions that we observed to tuberculin in patients with M. marinum infection were larger than tuberculin crossreactions that we have observed in patients with MAC disease. This is consistent with genetic studies showing that M. marinum is closely related to M. tuberculosis [30]. Reactions 10 mm to the MAS skin test were observed in 3 (60%) of 5 patients. This is higher than the 30% 40% rate of MAS reactions expected in a general US population [11], but lower than the rate of positive PPD reactions in patients with M. marinum infection. This is consistent with the greater genetic difference between MAC and M. marinum. In the study by Jolly and Seabury [27], which used a different M. avium antigen, 14 (67%) of 21 patients had reactions 10 mm in diameter. In their study and the present study, most dual skin test reactions with these 2 antigens were nondominant (i.e., neither the PPD nor the MAS antigen produced a consistently larger reaction than the other). Distinction between infection with M. tuberculosis and M. marinum with a single tuberculin skin test, dual skin tests with tuberculin and an NTM antigen, or newer in vitro assays of IFN-g production may not be possible [31]. There have been no controlled trials that compare treatment regimens for M. marinum infection, only case reports and case series. Recent treatment reports that include 110 patients are summarized in table 2. Combination therapy, typically with 2 drugs, as reported in several small series, appears to have a low failure rate in superficial infection and is generally recommended [2, 6, 32 37]. A review of 44 cases concluded that combination therapy with ethambutol and rifampin might be highly effective, on the basis of successful outcomes for 8 of 9 patients treated with this regimen [38]. Combination therapy, most commonly with clarithromycin plus rifampin, was used for 40 of 63 patients from France who were treated for an average of 3.5 months during [12]. Surgical debridement or excision was required for 30 (48%) patients. Deep infection was present in 18 (29%) patients and was found to be the principal determinant of outcome. Infection resolved in 42 (93%) of 45 patients with localized infection versus 13 (72%) of 18 with deep structure infection. Monotherapy failed more often in deep than in cutaneous infections. Failure of clarithromycin-containing regimens was reported in only 2 (10%) of 20 patients with infection limited to the skin [12]. Success with azithromycin in combination with ethambutol and minocycline has been reported for the treatment of skin and soft tissue infection in 1 lung transplant recipient [39]. We believe that clarithromycin is the optimal first agent in combination treatment of M. marinum. Experience in the treatment of other NTMs suggests that azithromycin might be a reasonable alternative. Published series and our experience support ethambutol as a reasonable second agent, whereas in vitro studies and published clinical series suggest that rifampin also would be appropriate. There is not enough clinical experience from the treatment of infection with either M. marinum or other NTMs with any of the other agents that have demonstrated activity in vitro to judge their potential efficacy as second agents. Even if the organism is deemed susceptible to an antibiotic, we believe that monotherapy should be avoided because of the unreliability of susceptibility testing for this organism and the possibility of emerging resistance. Published studies and our experience suggest that optimal combination therapy would therefore be clarithromycin and either rifampin or ethambutol. In a patient who cannot tolerate clarithromycin, the combination of rifampin and ethambutol may be considered, although our experience with patient 6 demonstrates that response to this regimen may sometimes be delayed. The optimal duration of therapy is not known, and results of published studies have been highly variable. The drugs recommended for treatment do not have prominent doserelated or time-dependent side effects. Thus, a reasonable approach is to treat with 2 agents for 1 2 months after the resolution of all lesions (typically 3 4 months in total) [35, 36]. Susceptibility testing is not routinely recommended and should be reserved for cases of treatment failure. As evidenced by patient 5, treatment of deep or extensive 394 CID 2003:37 (1 August) Lewis et al.

6 Table 2. Published studies on treatment of Mycobacterium marinum soft tissue infection. Reference No. of patients (no. with deep infection) n Duration Monotherapy Combination therapy Specific therapy, no. of patients (% cured) n Duration Specific therapy, no. of patients (% cured) Surgery performed, no. of patients (% cured) Comments [12] 66 (18) 23 Median, 3.5 months Minocycline, 19 (NA); clarithromycin, 4 (NA) 40 Median, 3.5 months Clarithromycin + rifampin, 20 (NA); cycline + clarithromycin, 11 (NA) 30 (NA) Resolution in 55 (87%) of 66 overall but only 13 (72%) of 18 with deep infection [13] 39 (NA) 20 NA Minocycline 12 (NA); rifampin, 4 (NA) 7 NA Clarithromycin + ethambutol, 7 (NA) NA Resolution within 2 4 months of treatment in most patients [37] 38 (NA) 22 Mean, 15 weeks TMP-SMZ, 19 (93); minocycline, 3 (100) [35] 31 (0) 19 Mean, 4 months Minocycline, 14 (71); doxycycline, 3 (67); tetracycline, 1 (100); TMP- SMZ, 1 (100) [41] 12 (6) 6 7 Mean, 8 months Doxycyline, 4 5 (100); TMP-SMZ, 2 (100); rifampin, 1 (100) 12 Mean, 15 weeks TMP-SMZ + minocycline, 5 (100) 1 (100) Of patients with adequate follow up, none worsened after treatment; only 2 failed to improve 8 Mean, 5 months Ethambutol + rifampin, 5 (100); other, 3 (100) 5 6 Mean, 8 months Rifampin + ethambutol, 3 4(100); rifampin, ethambutol, and ciprofloxacin, 1 (100) 1 (NA) Relapse in 2 patients after 4 months of minocycline; recommend treatment for 2 months after resolution for minimum of 6 months 12 (100) Surgical series: resolution in all patients with long-term antibiotic therapy and debridement NOTE. Includes studies with minimum of 10 patients published since 1966, not including case series from literature surveys. NA, not available; TMP-SMZ, trimethoprim-sulfamethoxazole.

7 disease, especially in an immunocompromised patient, can be challenging. Rifampin may be a preferred second or third agent in bone or joint infection, where this drug has been shown to improve outcome with infections due to other susceptible organisms [40]. It is possible that patients with disseminated infection or extensive deep infection, and the potential for a large burden of organisms, may benefit from treatment with 3 active agents, at least until clinical and microbiological response has been documented. Treatment of deep infection, especially in the immunocompromised host, typically requires surgical debridement. Multicenter controlled trials are needed to define the optimal regimen and duration of treatment for both superficial and deep infection. In summary, fish-tank exposure is the source of most cases of cutaneous M. marinum infections and may be preventable through the use of waterproof gloves for persons with acute or chronic open skin lesions. Tuberculin skin test reactivity is common after infection with M. marinum and should not generally be attributed to tuberculosis. Treatment with a 2-drug regimen that includes a macrolide should be successful in most patients with disease limited to the skin. Acknowledgments We wish to thank Kaare Haslov (State Serum Institute, Copenhagen), for donation of M. avium sensitin skin tests, and Richard Waddell and Sue Tvaroha, for assistance with the study. References 1. Arsonson JD. Spontaneous tuberculosis in saltwater fish. J Infect Dis 1926; 39: Huminer D, Pitlik SD, Block C, Kaufman L, Amit S, Rosenfeld JB. Aquarium-borne Mycobacterium marinum skin infection: report of a case and a review of the literature. Arch Derm 1986; 122: Hautmann G, Lotti T. Atypical mycobacterial infections of the skin. Dermatol Clin 1994; 12: Falkinham JO, Norton CD, LeChevallier MW. Factors influencing numbers of Mycobacterium avium, Mycobacterium intracellulare, and other mycobacteria in drinking water distribution systems. Appl Environ Microbiol 2001; 67: Kullavanijaya P, Sirimachan S, Bhuddhavudhikrai P. Mycobacterium marinum cutaneous infections acquired from occupations and hobbies. Int J Derm 1993; 32: Jernigan JA, Farr BM. Incubation period and sources of exposure for cutaneous Mycobacterium marinum infection: case report and review of the literature. Clin Infect Dis 2000; 31: Mollohan CS, Romer MS. Public health significance of swimming-pool granuloma. Am J Public Health 1961; 51: Dailloux M, Hartemann P, Beurey J. Study on the relationship between isolation of mycobacteria and classical microbiological and chemical indicators of water quality in swimming pools. Zentralbl Bakteriol Mikrobiol Hyg [B] 1980; 171: Emde KM, Chomyc SA, Finch GR. Initial investigation on the occurrence of Mycobacterium species in swimming pools. J Environ Health 1992; 54: Iivanainen E, Northrup J, Arbeit RD, Ristola M, Katila ML, von Reyn CF. Isolation of mycobacteria from indoor swimming pools in Finland. APMIS 1999; 107: von Reyn CF, Horsburgh CR, Olivier KN, et al. Skin test reactions to Mycobacterium tuberculosis purified protein derivative and Mycobacterium avium sensitin among health care workers and medical students in the United States. Int J Tuberc Lung Dis 2001; 5: Aubry A, Chosidow O, Caumes E, Robert J, Cambau E. Sixty-three cases of Mycobacterium marinum infection. Arch Intern Med 2002; 162: Casal M, del Mar Casal M. Multicenter study of incidence of Mycobacterium marinum in humans in Spain. Spanish Group of Mycobacteriology. Int J Tuberc Lung Dis 2001; 5: Califano L, Cannavo SP, Malara G. Verrucous nodule of the finger. Arch Derm 1998; 134:365 6, Gluckman SJ. M marinum. Clin Dermatol 1995; 13: Heller HM, Swartz MN. Nodular lymphangitis: clinical features, differential diagnosis and management. Curr Clin Top Infect Dis 1994: Phillips SA, Marya SKS, Dryden MS, Samuel AW. Mycobacterium marinum infection of the finger. J Hand Surg [Br] 1995; 20: Tchornobay AM, Claudy AL, Perrot JL, Levigne V, Denis M. Fatal disseminated Mycobacterium marinum infection. Int J Derm 1992; 31: Silcox VA, David HL. Differential identification of Mycobacterium kansasii and Mycobacterium marinum. Appl Microbiol 1971; 21: Patel S, Duke O, Harland C. Septic arthritis due to Mycobacterium marinum. J Rheumatol 1995; 22: Harth M, Ralph ED, Faraawi R. Septic arthritis due to Mycobacterium marinum. J Rheumatol 1994; 21: Shih J, Hsueh PR, Chang YL, Chen MT, Yang PC, Luh KT. Osteomyeliltis and tenosynovitis due to Mycobacterium marinum in a fish dealer. J Formos Med Assoc 1997; 96: Barton A, Bernstein RM, Struthers JK, O Neill TW. Mycobacterium marinum causing septic arthritis and osteomyelitis. Br J Rheumatol 1997; 36: Gatt R, Cushieri P, Sciberras C. An unusual case of flexor sheath tenosynovitis. J Hand Surg [Br] 1998; 23: Parent LJ, Salam MM, Appelbaum PC, Dossett JH. Disseminated Mycobacterium marinum infection and bacteremia in a child with severe combined immunodeficiency. Clin Infect Dis 1995; 21: Holmes GF, Harrington SM, Romagnoli MJ, Merz WG. Recurrent, disseminated Mycobacterium marinum infection caused by the same genotypically defined strain in an immunocompromised patient. J Clin Microbiol 1999; 37: Jolly HW, Seabury JH. Infections with Mycobacterium marinum. Arch Derm 1972; 106: American Thoracic Society. Targeted tuberculin testing and treatment of latent tuberculosis infection. Am J Respir Crit Care Med 2000; 161: S von Reyn CF, Williams D, Horsburgh CR, et al. Dual skin testing with Mycobacterium avium sensitin and purified protein derivative to discriminate pulmonary disease due to M. avium complex from pulmonary disease due to Mycobacterium tuberculosis. J Infect Dis 1998; 177: Tonjum T, Welty DB, Jantzen E, Small PL. Differentiation of M. ulcerans, M. marinum, and M. hemophilum: mapping of their relationships to M. tuberculosis by fatty acid profile analysis, DNA-DNA hybridization, and 16S rrna gene sequence analysis. J Clin Microbiol 1998; 36: Arend SM, Meijgarden KE, de Boer K, et al. Tuberculin skin testing and in vitro T cell responses to ESAT-6 and culture filtrate protein 10 after infection with Mycobacterium marinum or M. kansasii. J Infect Dis 2002; 186: Van Dyke JJ, Lake KB. Chemotherapy for aquarium granuloma. JAMA 1975; 233: Engbaek HC, Thormann J, Vergmann B. Aquarium-bourne Mycobacterium marinum granulomas. Scand J Infect Dis 1980; 12: CID 2003:37 (1 August) Lewis et al.

8 34. Donta SM, Smith PW, Levitz RE, Quintiliani R. Therapy of Mycobacterium marinum infections. Arch Intern Med 1986; 146: Edelstein H. Mycobacterium marinum skin infections: report of 31 cases and review of the literature. Arch Intern Med 1994; 154: Alloway JA, Evangelisti SM, Sartin JS. Mycobacterium marinum arthritis. Semin Arthritis Rheum 1995; 24: Ang P, Rattana-Apiromyakij N, Goh CL. Retrospective study of Mycobacterium marinum skin infections. Int J Dermatol 2000; 39: Wolinsky E, Gomez F and Zimpfer F. Sporotrichoid Mycobacterium marinum infection treated with rifampin-ethambutol. Am Rev Respir Dis 1972; 105: Torres F, Hodges T, Zamora MR. Mycobacterium marinum infection in a lung transplant recipient. J Heart Lung Transplant 2001; 20: Norden CW, Fierer J, Bryant RE. Chronic staphylococcal osteomyelitis: treatment with regimens containing rifampin. Rev Infect Dis 1983; 5(Suppl 3):S Kozin SH, Bishop AT. Atypical Mycobacterium infections of the upper extremity. J Hand Surg [Am] 1994; 19: Fish Tanks, Tuberculin Tests, and M. marinum CID 2003:37 (1 August) 397