Mycobacterium Marinum Treatment Literature Review Abstract Background Mycobacterium marinum is an atypical mycobacterium that can be found in water environment. It is the agent of a characteristic skin disease known also as fish tank granuloma. In some occasions it can spread as a nodular lymphangitis, extend to deep structures as well as in exceptional cases disseminate systemicaly . The infection is adquired after contact with fish or contaminated water mainly from aquaria or swimming pools. Although the real incidence is understimate, it is a uncommon infection that needs high clinical suspicion to be diagnosed. As a consequence, delay in the diagnosis is common. Mycobacterium marinum is intrinsically a multidrug resistant mycobacterium.There is no clear consesus in the management of this infection. Usually it is managed either with monotherapy or with combination of antibiotics plus surgery in selected cases. Methods Retrospective study of cases of fish tank granuloma collected from 2000 to 2009 in the dermatology surgery of The Hospital of Tropical Diseases (HTD) of the University College London Hospitals. Systematic review of the literature with the terms Mycobacterium marinum and fish tank granuloma from 1999 to 2009. Results From the HTD dermatology surgery were collected 7 cases. Identification of M. marinum was possible only in 5. The response to treatment was good at least in 5 of them. In the literature review there were reported a total of 516 cases. From those 133 could be analyzed as individual cases. Identification was possible in 89.5% of the cases. Cured was reported in 82% of the cases. Conclusion The key of the diagnosis is to collect accurately the history of exposure. Histology, although no pathognomonic, will rise the suspicion if granuloma formation are found and will facilitate the differential diagnosis. Identification is done after culture with classical biochemical tests. Molecular biology techniques have the advantage of accelerate the procedure. There is not enough evidence to propose any specific treatment. Currently, recommendations are based in experts opinions. A prospective, randomized controlled clinical trial would be valuable to propose a base evident treatment. Contents Pages Introduction M. marinum is an environmental atypical mycobacterium ubiquitous in fresh, salt, and brackish water. It is known that infects temperate and tropical species of fish of at least 150 species, including ornamental fish. But also affects frogs, eels, oysters, aquatic mammals, toads and snakes [2, 11]. It cause tuberculosis-like disease in fish, its natural host [129]. The infection in fish has an average incubation period of 3 months. It affects viscera and produce anorexia with emaciation, skin defects, distension of the abdomen, being cause of sudden death [11, 12, 66]. In 1904, Alexander described for first time lesions in a cod fish that were associated with acid fast bacilli. But it was Aronson in 1926 that isolated the bacteria from tubercles of fish that died in an aquarium of Philadelphia. He described the bacilli as acid fast, chromogenic, pleomorphic and growing best at 18Ãâà º-20Ãâà ºC. Aronson suggested the name of Mycobacterium marinum [130]. M. marinum is as well the causative agent of the human disease called fish tank granuloma, also known as swimming pool granuloma or fish fancierÃâà ´s finger syndrome. In 1951 Norden and Linell reported for the first time the human disease in a swimming pool outbreak in ÃÆ'ââ¬ârebro, Sweden. They described the lesions as chronic papulous ulcerations, usually located in the elbows. The pathogen was isolated from the walls of that swimming pool as well as from the lesions of the patients. They called the pathogen isolated Mycobacterium balnei [131]. It was not until 1959 when Bojalil demonstrated that M. marinum and M. balnei were in fact, the same mycobacterium [132]. M. marinum is the most frequent cause of skin infection among the environmental mycobacterium that affects humans [86, 133]. Nowadays the frequency of human infections is mainly sporadic. But in the past, outbreaks related with swimming pools were not uncommon [134]. As an example, one of the biggest epidemics was in Glenwood springs pool, Colorado in 1956 with 262 cases reported [135]. That was before chlorination became a common practice. Chlorination makes water safer. As was seen recently in Bologna, were water from swimming pools were free of M. marinum. But still could be isolated in 4.5% of the samples from the shower floor of the same [136]. The real global incidence of the disease is not know because the number of cases are underreported, due probably to the difficulties in the diagnosis [94]. It is world wide distributed but with a tendency to aggregate geographically [137]. Like in Chesapeake bay, Maryland, where there is an incidence of 4 cases per 100000 population per year [138]. Meanwhile in California was estimated in 0.27 cases per 100000 adults [139]. Or in Satowan, Micronesia, with an estimated prevalence of 10% of the population [115]. Apparently the global annual incidence remains small and stable [140]. Even though an increase number of reported cases has been noticed in The United States, going from an average of 40 cases per year in the 80s to an average of 198 cases per year in the 90s [141]. What seems clearer is that differing from other atypical mycobacteria the prevalence of M. marinum has not increased with the HIV epidemic [140]. Opposite to humans, the incidence in fish is increasing in hatchery fish, probably due to the high population density of fish. Transmission is possible fish to fish and between fish and amphibians. In addition it has been proposed transmission through eggs and through practice of feeding fish with fish carcasses [140]. There are 2 groups or clusters of M. marinum with different pathogenicity. Cluster I is characterized by producing acute disease and death in fish and also for affecting humans. On the contrary, cluster II only affects fish producing the classical chronic disease with granuloma formation [9]. This is also supported by a study done in Israel were it was seen that only certain strains of M. marinum affected humans. They also demonstrate that in Israel strains affecting humans came from ornamental fish and not from local fish for consumption [5]. The mode of transmission to humans is mainly waterborne and fish borne. Person to person transmission has not been documented [137]. However, It has been described indirect transmission via fomites in at least 3 cases. Two very small children and one infant who acquired the infection after bathing in containers that were previously used to clean the family fish tanks of tropical fish [27, 40, 142]. As other environmental mycobacterium, M. marinum has commonly low pathogenicity. For this reason in normal conditions only affects disrupted skin [8]. The main risk factor to contract the infection consists in having lesions or abrasions in the skin with exposure to non chlorinated water or marine animals infected [140]. The most frequently nowadays is the exposure to private aquaria. But some times the source of exposure is unknown. As a consequence, after the description of cases following injuries with plants, it has been suggested the possibility that could be other reservoirs different from water and fish. Although at the present moment this possibility has not been demonstrated [30, 43]. The incubation period is usually 3 to 4 weeks [135]. Following, the most common manifestation is a cutaneous lesion at the site of inoculation. It initiates as a solitary nodule or pustule that eventually evolutes to an ulcer , abscess or verrucous plaque [143]. It affects more frequently the extremities, probably because the pathogen grows better at low temperatures [144]. The severity of the disease depends, among other factors, on the number of microorganisms inoculated [134] In 20% of the cases the cutaneous lesions spread along ascending lymphatic vessels. This is called sporotrichoid spread or nodular lymphangitis [143]. As a result of direct extension invasion of deep structures as tendons, articulations and bones occurs in 29% of the cases [144]. Systemic dissemination is unusual but has been described in immunocompromised patients [140]. Spontaneous resolution , usually with scaring , has been documented from months after the infection up to 2 years [133, 135]. The diagnosis is based on the history of exposure and risk factors along with the characteristic clinical features. It is supported with histopathology, culture and bacteriological identification that in some cases require molecular biology techniques [94, 120]. The fact is that the diagnosis is not easy and in most of the cases is either delayed or remain being presumptive based in the history and response to treatment. The objective of the treatment is to increase the speed of resolution and prevent progression of the disease [1]. With this purpose different combinations of antibiotics plus the support, in selected cases, of surgery are the common practices in the treatment of this infection. Although the election of the drugs still depends of the preference of individual authors and is not based on controlled evidence [143]. Aims and Objectives The overall aim of the project is to determine the current state of evidence for the diagnosis and treatment of M. marinum infection. Being the specific objective to review the literature and the series of HTD cases with the purpose of suggest appropriate diagnosis and case management of Mycobacterium marinum infection. Material and Methods Cases with diagnosis of fish tank granuloma were review. The cases were diagnosed and treated from 2000 to 2009 in the dermatology surgery of The Hospital of Tropical Diseases of the University College London Hospitals. There were included patients with either culture positive for M. marinum or clinical diagnosis plus response to appropriate treatment. The files of those patients were reviewed and data were collected in a questioner that included: anthropological data, past medical and drug history, risk factors and exposure, description and location of the lesions, spread or deep extension of the infection, incubation period , delay in diagnosis, diagnosis, treatment and evolution (see questioner in annex). Additional information considered relevant was as well collected. A literature review in Medline and Cochrane databases was done. The review included the combination of the following terms: Mycobacterium marinum or fish tank granuloma. It was limited to papers from 1999 to July 2009, English and Spanish literature and humans. The papers obtained in the search were divided in 3 categories. First, case reports in which it was possible to collect data from individual cases. Information of those papers was collected in the questionnaire previously mentioned. Second, case series of 14 or more cases in which data from individual cases was not reported. And finally, a miscellaneous category of papers that were considered relevant for the diagnosis and treatment of this infection. The information gathered in the questioners was computerized in Microsoft Excel 2007. No statistical analysis was done in view of the cases were publish not with this purpose. As a consequence probably important reporting bias would invalidate any statistical outcome. Simple description of the results was done. Results Cases of The Hospital of Tropical Diseases From 2000 to July 2009 there were collected 7 cases with diagnosis of Mycobacterium marinum infection. The 7 cases included 6 males and 1 female and their ages ranged form 31 to 65 years. All the cases admitted to have had contact with an aquarium. In two cases the exposure was occupational. One of them worked in the London Zoo being responsible of the management of the aquaria. The second was a cook in a restaurant that store crustacean in an aquarium. All the cases presented nodular lesions located in one of the upper limbs, six of them with sporotrichoid spread and one with a solitary nodule (see photos is annex). One patient that was taking oral prednisolone due to severe atopic eczema presented with tenosynovitis of the left hand, sporotrichoid spread and palpable regional lymphonodes. The rest of the patients were not immunosuppressed. Biopsy of the lesions was performed in all the cases. However AFB were found just in one case. Whereas cultures were positive in 5 cases for M. marinum. In the other two cases the diagnosis was done based on history of exposure, clinical characteristics and positive response to treatment. Different combinations of antibiotics were used with no clear advantage of any regimen in particular. In 4 patients it was need to change the regiment. The reasons were drug intolerance in one case and lack of improvement in the rest. There was also one case that initially was improving with a regimen of rifampicin plus minocycline. But after simplification of the regiment to clarithromycin monotherapy presented worsening of the lesions. This case was eventually controlled switching to minocycline plus ethambutol. There was one patient who presented a relapse after one year of the previous infection. Finally it was cured with 6 months of rifampicin plus clarithromycin. Susceptibility test was done in 3 cases. It was detected resistance to trimethoprim-sulfamethoxazole and rifampicin in one case and susceptibility to clarithromycin, ethambutol and doxycycline in 3 cases. The average time of duration of the treatment was 6 months with a range of 2 to 9 months. Surgery was not required in any patient. The final outcome was good in 5 patients being the other 2 lost of follow up (see table 1). Case No Age(y) /Sex Medical history Drug history Exposure Location Clinical characteristics Histology AFB results Culture Treatment Duration (months) Evolution 1 9/M No Aquarium Upper limb Multiple nodules sporotrichoid spread Granulomatous inflammation AFB + 1-RIF+DOX 2-RIF+EMB+DOX 3-RIF+EMB+CLR 7 Cured 2 61/F Psoriasis Chronic paronychia Aquarium Upper limb 3 nodules Sporotrichoid spread AFB + 1-MIN 2 Improving Lost of follow up 3 64/M No Aquarium Upper limb Multiple nodules Sporotrichoid spread Noncaseating granuloma 1-EMB+TET 2-RIF+EMB+INH 9 Relapse after 1year 3(*) 65/M Fish tank granuloma Aquarium Upper limb Multiple nodules Sporotrichoid spread 1-RIF+CLR 6 Cured 4 59/M Severe atopic eczema Systemic steroids Aquarium Upper limb Multiple nodules Sporotrichoid spread Tenosynovitis AFB + 1-DOX 2-RIF+EMB 3-RIF+EMB+CLR 4-ERI+MIN ? Lost of follow up 5 44/M No Aquarium Upper limb 5 nodules Purulent discharge Sporotrichoid spread Necrotizing granulomatous inflammation AFB- + 1-RIF+MIN 2-CLR 3-MIN+EMB 3.5 Cured 6 31/M No Aquaria (London Zoo) Finger Solitary nodule AFB + 1-CLR+EMB 4 Cured 7 49/M No Aquarium (Restaurant) Upper limb 5 nodules Sporotrichoid spread + 1-RIF+EMB 5 Cured (*) Notice that case number 3 is repeated. It belongs to the same patient that the one above but one year later. The patient presented a relapse after one year of being cured. Results from the literature review From 1999 to July 20009 there were found 233 results in Medline database and zero In Cochran Library. From those, 127 were considered relevant and consequently analysed in this review. From the 127 papers reviewed, 108 contained case reports and the 19 remaining were a miscellaneous of reviews or original articles covering issues related with diagnosis and treatment. No clinical trials or randomized control trials were found. In those years the literature reported a total of 516 cases of M. marinum infection. From those cases reported only 133 could be analyzed as individual cases. The rest of cases were reported as series of cases (See figure 1). The number of cases reported per year since 1999 up to July 2009 have been variable with a minimum of 6 cases per year in 2004 to a maximum of 88 cases in 2000. No clear tendency to increase neither decrease has been notice (see Figure 2). The majority of the cases have been reported in Europe, North America and South East Asia. No cases have been reported in Africa and only one case in South America (see Figure 3). Results from papers that could be analyzed as individual cases The reports included 82 men and 51 women. The average age was 46.7 years with a minimum of 18 months and a maximum of 87 years (see Figure 4). 70% of the patients had no relevant past medical history. 9% of the patients were immunosuppressed: 5 patients had HIV infection, 4 were recipients of solid organ transplant, 1 patient had a myelodisplastic syndrome, 1 had Non-Hodking Lymphoma and 1 had Chronic Lymphocytic Leukaemia. Among other relevant pathologies were reported 12 diabetic patients, 11 with rheumatic diseases and finally 5 that were suffering from other problems as: asthma, bullous pemphigoid, myasthenia gravis, CronhÃâà ´s disease and sarcoidosis. The majority of the patients were not taking any relevant drug. However, 20 of them were on systemic steroids, 12 had received steroids as local injection, 10 took methotrexate and 9 TNF-blockers. The most frequent exposure referred was the contact with an aquarium usually of tropical fish, it was reported in 51% of the cases. It was followed in 26 cases (20%) by other kind of contact with fish mainly referred as contact while cooking or cleaning fish. Finally 16 cases (14%) had other kind of contact with water environment. That included mostly fishermen or recreational sailors. Opposite to papers from the 60s, only in 2 patients referred contact with swimming pools. In 15 cases (11%) the source of exposure was not recall or reflected in the papers. Injury related with plants was reported in 3 cases. In one case the exposure was a bucket. The bucket was used to bath a child of 18 months after being used to put fish from an aquarium. As a result the child got infected [40]. History of trauma with skin barrier impairment was referred in 46 cases (71%) of the patients. Among those, 18 cases (39%) recall direct injury with either fin fish, fish hook or crab bite. Occupational exposure was reported in 20 cases (15%). The most frequent occupational risk was to be cook and have injuries while cleaning fish. Another common job of high risk was to work in a pet shop, with the duty of cleaning aquaria. Finally fishermen are evidently in direct contact with water and fish and prone to suffer injures with fish hooks. The incubation period was documented only in 30 patients. It went from 1 day up to 4 months with and average of 48.9 days. The cases that presented an incubation period less than two weeks had in common to have suffered penetrating injuries with fish (figure 5). The upper limbs were affected in 120 case (90.2%), being the fingers the most frequent location. The lower limbs were affected in 11 cases (8.3%). The face in 7 cases (5.3%) and that included delicate locations as nostrils, eyelid and cornea. Finally, cutaneous dissemination was reported in 7 cases (5.3%). Sporotrichoid spread was found in 53 cases (39.8%). Lymphonodes affectation was recorded in only in 7 cases (5.3%) of the cases. The lesions were described as nodules in 63 cases (7%), plaques in 23 (17.35%), papules in 13 (9.8%) and ulcers in 22 (16.5%). There was purulent discharge in 32 cases (24.1%), as well as swelling and tenderness in 37 (27.8%) and 42 (72.4%) cases respectively. Involvement of deep structures was referred in 45 cases (33.8%) of the cases being the most frequent tenosynovitis with 34 cases (75.6%), followed by arthritis with 12 (26.7%) and osteomyelitis with 6 (13.3%). Systemic dissemination with documented bacteraemia was reported in 3 cases. The 3 of them were males from 66 to 87 years. In 2 cases the patients were on systemic steroids, due to myasthenia gravis[103] in one case and polymyalgia rheumatica[62] in the other. Sadly the third case that initially was not taking drugs, after being misdiagnosed as rheumatoid arthritis was put on systemic steroids, infliximab (TFN blocker) and methotrexate [59]. From the patients with invasive disease, including involvement of deep structures or systemic dissemination, 21 (46.6%9 were taken some kind of immunosuppressive drug. Meanwhile only 13 (15.5%) of the rest of the patients were taken them (see figure 6). Other interesting characteristic of the patients with invasive disease was that 15 (33.3%) of them referred direct fish injury. Only 3 (3.4%) of the rest had this exposure (see figure 7). Patients that for any reason were taken immunosuppressive drugs presented different characteristics from the rest of the patients. More than half of them presented invasive disease compared with only 20% of the rest (see table 2). Patients taking immunosuppressive drugs (*) Patients no taking any drugs Number of patients 36 91 Average age (years) 53.2 44.9 Female % 14 / 38.9% 36 / 39.6 % Sporotrichoid spread 13 / 36.1% 39 / 42.8% Involvement deep structures 21 / 58.3% 19 / 20.8% Systemic dissemination 3 / 8.3% 0 AFB positive 19 / 52.7% 33 / 36.3% Culture 33 / 91.6% 75 / 82.4% Bad evolution 2 / 5.5% 3 / 3.3% (*)Immunosuppressive drugs including: TNF-Blockers, systemic steroids, local injectable steroids, methotrexate and azathioprine. The time of evolution until the patients presented for consultation was as short as 4 days and as long as 18 years with an average of 8.6 months. The time until the diagnosis was finally done was only reflected in 17 patients. However the delay went from 21 days to 2 years with an average of 6.3 months. Tuberculosis skin test was only done in 19 patients. From those 86.4% were positive. Biopsy of the lesions was done in 120 cases (90.2%). Aspirate was reported only in 19 cases (1 4.3%). Histology characteristics suggestive of mycobacterial infection with granuloma formation were found in 45.5% of the biopsies. However, only 21% of those were described as caseating granulomas. Other frequent finding reported was mix infiltrates with chronic and acute inflammatory cells. In some cases a wrong diagnosis was done due to confusion with rheumatoid nodules, Sweetà ¢Ã¢â ¬Ã¢â ¢s syndrome, foreign body granuloma or interstitial granuloma annulare (See table 3). Table 3: Histology Findings Histology description Number of patients Non caseating granuloma 19 Caseating granuloma 12 Granuloma( type not specified) 26 Infiltrates of chronic and acute inflammatory cells (lymphocytes , neutrophils, histiocytes) 18 Granulation tissue 6 Abscess formation 9 Necrotizing folliculitis 1 Focal dermal necrosis 3 Fibrinoid degeneration 1 Necrotizing paniculitis 2 Cystic degeneration 1 Pseudoepitheliomatous hyperplasia 4 Acute suppurative paniculitis 1 Fibrinous exudates 3 Lichenoid inflammation 1 Acanthosis in epidermis 2 Necrotic Corneal Stroma 1 Dermal fibrosis 2 à ¢Ã¢â ¬Ã ¢ Pseudocarcinomatous hyperplasia of follicles à ¢Ã¢â ¬Ã ¢ Nodular and perifollicular infiltrate of neutrophils and histiocytes à ¢Ã¢â ¬Ã ¢ Dermal fibrosis à ¢Ã¢â ¬Ã ¢ Collections of neutrophils within follicles 1 patient with lesions of 18 years evolution Confusion with other pathologies Rheumatoid Arthritis ( rheumatoid nodule) 2 SweetÃâà ´s syndrome 2 Foreign body granuloma 2 Interstitial gra nuloma annulare 2 In the samples collected AFB was found in 41.7%, was negative in 34.6% and was not reported in 23.6%. In total identification of M. marinum was possible in 119 cases (89.5%). Culture was positive in 114 cases (85.7%). The time until the cultures grew went from 8 to 56 days, with an average of 23.3 days. Identification with PCR was done in 19 cases (14.2%). The imaging techniques were useful to diagnose extension of the infection. Radiographies were used in 25 cases (18.7%) to rule out bone involvement. Magnetic resonance imaging was used in 16 cases (12%) resulting in the diagnosis of tenosynovitis, abscess, join effusions or osteomyelitis. About the treatment, 126 patients were treated with antibiotics; in the rest of the cases the management is not mentioned. Surgery was need in 38 patients (84.4%) with affectation of deep structures and in 21 patients (25%)with cutaneous lesions. Monotherapy was used in 54 cases (42.8%), bitherapy in 38 cases (29.4%), triple therapy in 20 cases (15%) and combination of 4 or more drugs in 5 cases (4%). Finally combinations of drugs that included classical tuberculosis treatment were used in 10 cases (8.7%). In 41 patients the regimen of drugs needed to be change, either for non effectiveness or non tolerance. The regiment was change one time in 29 cases (21.8%), two times in 10 cases (7.5%), and up to 3 times in 2 patients (1.5%). The drug more frequently used as monotherapy was clarithromycin, followed by minocycline, doxycycline, ciprofloxacin and trimethoprim-sulfamethoxazole. The combinations of drugs more frequently used were rifampicin + ethambutol followed by clarithromycin + rifampicin and clarithromycin + ethambutol. (Effectiveness of the different regimens depending of the extension of the disease can be seen in table 4) Susceptibility test were reported in 34 patients. Rifampicin was susceptible in 86.4% of the test, ethambutol in 91.3%, clarithromycin in 95% and minocycline in 62.5%. Isoniazid was resistant in 100% of the tests done and streptomycin in 66.6% (see table 5). The average time of duration of antibiotic treatment was 5.4 months, with a range of 12 days to 15 months. After completion of the treatment the final evolution of 109 patients (81.9%) was reported as good outcome or cured. Only in 8 patients (6%) the evolution was reported as bad outcome. No mention about the evolution was done in the rest of cases. Among 12 patients in which long follow up was reported, only one patient presented recurrence of the infection after 3 months course of doxycycline. (Characteristics of the cases with bad outcome are resumed in table 6. Table 4:: Antibiotic Combinations Used Depending On The Extension of The Disease Patients with only cutaneous lesions Patients with Invasive disease Number of patients Effectiveness Number of patients Effectiveness Monotherapy 38 52.5% 16 75% CLR 8 75% 5 100% MIN 10 70% 1 100% DOX 8 50% 1 100% CIP 4 25% 2 50% CTX 2 100% 3 33.3% AZI 3 0% 0 AMK 1 0% 0 ERI 0 1 0% LEV 0 1 0% MOX 1 0% 0 OFL 1 0% 2 50% Combinations of 2 drugs 26 92% 11 83.3% CLR + EMB 6 100% 2 100% CLR + CIP 1 100% 0 CLR + MIN 1 100% 0 CLR + RIF 5 80% 0 CLR +CTX 1 100% 0 RIF + EMB 9 100% 5 60% RIF + CTX 0 1 100% RIF + INH 0 1 100% CIP +DOX 1 100% 0 CIP + EIR 1 0% 0 DOX +CTX 0 2 100% CIP + EMB 1 100% 0 Combinations of 3 drugs 13 72.7% 7 66.6% CLR + EMB + CIP 0 1 0% CLR + EMB + RIF 2 100% 5 100% CLR + EMB +RFB 0 1 0% CLR +CTX + CIP 1 0% 0 EMB + AZI + MIN 1 0% 0 RIF + CLR + AMK 1 100% 0 RIF + INH + CLR 1 0% 0 RIF + EMB + CTX 5 100% 0 RIF + EMB + DOX 1 0% 0 RIF + EMB + MOX 1 0% 0 Combinations of > than 3 drugs 2 100% 3 33.3% CIP+RIF + EMB + CLR + RFB + AMK 0 1 100% CLR + DOX + RIF + EMB 0 1 0% RIF + EMB + CLR + AMK + IMP 0 1 0% RIF + EMB + AZ I+ CTX 1 100% 0 RIF + EMB + CLR + CIP 1 100% 0 Combinations with TB treatment 1 100% 9 55.5% INH + RIF + EMB + CLR 1 100% 1 0% INH + RIF + EMB 0 3 100% INH + RIF + EMB + PZA 0 4 50% INH + RIF + EMB + PZA + CLR 0 1 0% TABLE 5: PATTERN OF SUSCEPTIBILITY Drug Number of patients Susceptible Resistant Indeterminate Isoniazid 9 9 Rifampicin 22 19 3 Ethambutol 23 21 2 Pyrazinamide 1 1 Streptomycin 6 1 4 1 Rifabutin 2 2 Azithromycin 4 1 3 Clarithromycin 20 19 1 Minocycline 8 5 1 2 Doxycycline 6 6 Tetracycline 2 1 1 Trimethoprim-sulfamethoxazole 6 5 1 Ciprofloxacin 11 6 1 Levofloxacin 1 1 Moxifloxacin 4 4 Gatifloxacin 1 1 Amikacin 9 9 Linezolid 1 1 Imipenem 2 2 Erythromycin 2 1 1 Cefotaxime 2 1 1 Kanamycin 3 3 Ethionamide 3 3 Ansamycin 1 1 The table reflects in how many patients each drug was tested and in how many it resulted as susceptible, resistant or indeterminate. TABLE 6: CHARACTERISTICS OF CASES WITH BAD OR POOR OUTCOME References Age Sex Past medical history Type of infection Treatment Duration treatment (months) Evolution [25] 67/M DM Cutaneous disseminated RIF+EMB+CLR+CIP 3.7 Deceased Secondary bacterial infection [26] 50/M No Tenosynovitis RIF+EMB Synovectomy 3 Dysfunctional index finger [46] 62/F No Tenosynovitis CLR 4 debridements 6 Amputation index finger [46] 26/M No Tenosynovitis DOX+CTX 4 debridements 3.5 Persistent infection Need of grafting [73] 56/M No Tenosynovitis Local gentamicin Synovectomy 3 drainages ? Deformity [83] 47/M HIV infection Osteomyelitis 1-INH+RIF+EMB+PZA 2-Avobe + Fluoroquinolone 3-RFB+EMB+CIP 4-RFB+CIP ? Amputation Knee [88] 60/F NHL Systemic steroids Cutaneous disseminated 1-Ofloxacine 2-RIF+EMB+LEV+CLR 3-Above+STR+IG 4-CTX+MOX ? Continue with signs of active infection [103] 81/M Myasthenia gravis Systemic steroids Cutaneous and systemic dissemination 1-CIP 2-Hyperthermia 3-DOX intravenous 5 Deceased Bone marrow infected by M.marinum Results from papers that contain series of 14 or more cases There were 14 papers that contained series of cases with no available information of individual cases. As a consequence, individual cases could not be analyzed separately. There were 2 papers that reported the same series of cases in different years, the cases were counted just once. The principal characteristics of the papers are described in the table 7. There were 363 cases reported, 68% were males with an average of 44.3 years. In most of the cases the past medical history was not relevant; only 4 cases of HIV infection were reported. The most frequent exposure was to own an aquarium in 134 cases (37%). There was other kind of contact with fish in 37 cases (10%), swimming pool contact in 7 cases (2%) and other kind of contact with water in 58 cases (16%). In the rest of cases exposure was not mention. Finally occupational exposure was referred in 68 cases (19%). The incubation period was not reported. The upper limbs were affected in 245 cases (67%), the lower limbs in 78 cases (21%), the face in 3 cases (0.8%) and there was cutaneous dissemination in 10 cases (3%). Sporotrichoid spread was reported in 46 cases (12%). The lesions were described as plaques in 92 cases (25%) and nodules in 54 cases (15%). Deep structures were affected in 45 cases (12%): 41 tenosynovitis (91%), 8 arthritis (18%) and 3 osteomyelitis (6%). There were no cases of systemic dissemination described in any of the series. The average in the delay until the diagnosis was done was 3.8 months. The drugs more commonly used as monotherapy were doxycycline, minocycline, trimethoprim-sulfamethoxazole and clarithromycin. The combinations more prescribed were rifampicin plus ethambutol, rifampicin plus clarithromycin and clarithromycin plus ethambutol. The effectiveness of the treatments were not frequently reported. Consequently is not possible to point the advantage of any concrete regimen (see table 8). The average duration of the treatment was 5 months. Surgery was reported in 73 cases (20%). The evolution was good in 225 cases (62%), poor in 15 cases (4%) and not mention in the rest. There were 5 papers that contained series of cases of atypical mycobacterial infecti