Rare case of Mycobacterium marinum in a patient presenting with tenosynovitis

Abstract

A 72-year-old man presented with tenosynovitis of the left hand’s extensor tendons that had been present for several months. He was initially treated with corticosteroids, first by local injection then systemically, but with no effect. When re-evaluated, the patient had developed a rash, and the symptoms had spread locally to surrounding structures. At this point, the patient added to the medical history that he had been stung by a sculpin a month before the debut of symptoms. Based on this, the patient’s involved area was biopsied, and subsequent microbiology findings proved consistent with Mycobacterium marinum infection. By the time of diagnosis, the patient had soft tissue involvement, arthritis and osteomyelitis with an overlying rash. This case emphasises the need for reassessment when treatment is not effective and for further investigations of the medical history to establish the correct diagnosis and treatment.

Background

Mycobacterium marinum is a rare cause of infection in humans with a reported incidence of 0.04 per 100 000 patient years.¹ M. marinum is normally a skin infection, but the infection can spread to tendons, joints and bones and be difficult to treat.^1–4 Diagnosis is often delayed since M. marinum is rarely considered as a differential diagnosis when a patient presents with tenosynovitis or arthritis.² M. marinum is a frequent infection in fishes and amphibious animals and is found in contaminated water. Human infections normally occur after contact with infected fish or contaminated water, especially when the skin barrier is breached.^1–3

Case presentation

A 72-year-old immunocompetent man with hypertension and diabetes was seen at the department of rheumatology with a swelling of the left hand for 3 months. He had also noted a mild rash on the dorsum of the left hand that had not responded to topical corticosteroid therapy. The rash was red, confluent and macular. There were no nodules. The patient was initially diagnosed with tenosynovitis of unknown origin of the left hand’s extensor tendons and was treated with a local ultrasound-guided corticosteroid injection (20 mg triamcinolone hexacetonide mixed with lidocaine) and afterwards systemic corticosteroids (25 mg prednisolone daily for 5 days and then 12.5 mg daily for 10 days). Both treatments had a short-term effect on the symptoms. A month later, the mild rash had evolved and was accompanied by ulcerations (figure 1A), and the swelling had spread to nearby joints. The patient was shortly treated with dicloxacillin for erysipelas. Since the patient had not responded adequately to the corticosteroid or antibiotic treatments, the patient was questioned again to explore further causes. The patient revealed that he often went salt water fishing and that he had been stung by a sculpin a month before the debut of symptoms. After this new important information, the patient was referred to the department of orthopaedic surgery and biopsies were taken from the infected soft tissue and abscess of the left dorsal hand. Microbiological analyses of the biopsy material were positive for M. marinum, approximately 6 months after the first symptoms had developed. The infection had now spread to surrounding structures and affected the soft tissue, joints and bones of the patient’s left hand.

Figure 1.

The left hand (A) prior to biopsies and drainage, (B) after second surgical drainage and (C) 7 months after the end of the treatment.

Investigations

Blood samples were drawn from the patient at his first visit to the department of rheumatology. No inflammation was present: total leucocyte count was 7.55 (3.5–8.8)×10⁹/L and serum C reactive protein level was 1.1 (<6.0) mg/L, serum urate level was 0.39 (0.16–0.40) mmol/L and negative findings for anticyclic citrullinated peptides, IgM rheumatoid factor, myeloma protein, hepatitis and parvovirus serology. Immunoglobulins were slightly decreased, serum immunoglobulin G level was 4.6 (6.1–15.7) g/L and IgM level was 0.21 (0.40–2.30) g/L. X-ray of the left hand showed normal bone structure.

At the second visit, C reactive protein level was elevated at 11 mg/L and later topped at 33 mg/L combined with a slight elevation of total leucocyte count which was 10.2 (3.5–8.8)×10⁹/L and absolute neutrophil count was 7.65 (1.5–7.5)×10⁹/L. After initiation of treatment, the values quickly normalised. Pathology findings indicated granulomatous inflammation, with necrosis and formation of a giant cell.

MRI showed involvement of soft tissue (figure 2, left).

Figure 2.

Left: MRI of the hand at the time of diagnosis showing soft tissue involvement. Middle: X-ray before osteomyelitis in wrist (hamate bone). Right: X-ray 7 months after the end of the treatment showing sequelae from osteomyelitis in wrist (hamate bone).

Later X-rays showed development of osteomyelitis in the wrist, with no signs of osteomyelitis first and sequelae from osteomyelitis in the hamate bone at the end of the treatment (figure 2, middle and right).

The patient underwent surgery twice and biopsies were taken on both occasions. At the first surgery, two biopsies and material from an abscess were obtained. At the second surgery, four biopsies were taken. At the time of the two surgeries, the patient was treated with dicloxacillin. The material was sent to a central national laboratory, where all analyses for mycobacteria are conducted. Samples were stored and sent in sterile tubes and on arrival, all samples were cultured directly in both Mycobacteria Growth Indicator Tubes and Lövenstein medium at 33°C and 38°C. The following procedure was applied to all biopsy materials: (1) initial staining with auramine–rhodamine, if positive then the material was tested with (2) PCR for M. tuberculosis. If the M. tuberculosis test was negative, then the material was analysed for non-tuberculous mycobacteria with (3) PCR ‘Hain CM V.2.0 kit’ and if positive then (4) PCR test with ‘Hain Genotype AS V.1.0 kit’ to detect M. marinum. After 1 week, a single biopsy from the first surgery showed 5–10 acid-fast rods per high-power field and after 2 weeks the sample was PCR positive for M. marinum. All four biopsies from the second surgery had acid-fast rods and were PCR positive for M. marinum. Furthermore, all samples were cultured for 56 days, no bacterial growth was found.

Differential diagnosis

At first visit, the patient presented with a tenosynovitis of the left hand’s extensor tendons. Commonly considered differential diagnoses for tenosynovitis seen in the department of rheumatology are tendon overuse and rheumatic diseases such as rheumatoid arthritis or gout.⁵ No blood samples, objective signs or elements of the patient’s medical history gave suspicion of an underlying disease, including rheumatic diseases. The tenosynovitis was interpreted as result of overuse and was treated with corticosteroids.

After local and systemic corticosteroid treatment, the patient’s symptoms worsened, and he developed arthritis and a prominent local rash, but there were no other symptoms except in the left upper limb. Sudden and pronounced exacerbation of symptoms is not typical for a tenosynovitis due to overuse. There are three characteristics of the case that should lead to extra caution and consideration of other diseases: (1) it is uncommon for rheumatic diseases to be locally aggressive, without any systemic signs; (2) the symptoms did not improve with corticosteroid treatment, which is uncommon for inflammatory diseases; and (3) the rash covered most of the dorsal hand and was confluent and ulcerated (figure 1A). It was thus dissimilar to other rashes associated with rheumatic diseases such as erythema nodosum or vasculitis.

When the initial diagnoses were rejected, the patient was further questioned, and he revealed that he was an angler and had been stung by a sculpin 1–2 months prior to the debut of symptoms. Infection was then considered the primary diagnosis. A number of waterborne bacteria can cause infection, such as Aeromonas hydrofila and some Vibrio species, but other common infections in wounds after aquatic injuries include Streptococcus and Staphylococcus.^{6 7} Most of these infections have a more rapid and aggressive course. The symptoms and the indolent course of the current case fitted well with a Mycobacterium infection, and the link to fishing made M. marinum the primary suspected Mycobacterium.

Treatment

The patient was treated with rifampicin 600 mg once daily, ethambutol 1200 mg once daily and clarithromycin 500 mg two times per day for 6 months. Shortly after the PCR proved positive for M. marinum, the patient was administered antibiotics. He underwent surgical revision three times with drainage of abscesses but was not treated surgically for the bone infections. The surgical wounds were difficult to stitch due to the vulnerable skin, so the wounds were closed with negative pressure wound therapy. Alongside the medical and surgical treatment, the patient received physiotherapy and occupational therapy.

Outcome and follow-up

After 6 months of antibiotic treatment and surgical interventions, the patient no longer had ulcers, abscesses, tenosynovitis, arthritis or rash. C reactive protein had normalised, and the patient was considered cured of the M. marinum infection. The patient had a mild pain in his left hand and slightly impaired flexion of the left hand’s four ulnar fingers (especially the second finger), scar tissue where he was incised and diarrhoea. The patient was followed weekly and hereafter monthly at either the department of infectious diseases or the department of orthopaedic surgery. Regular blood tests including C reactive protein were taken in the first 6 months. X-rays of the hand were performed at the beginning of treatment and after 1, 3 and 6 months. The faeces were investigated for pathogens including Clostridium difficile, but all cultures were negative. Except for the pain in his left hand, the patient had no sequelae in terms of function (figure 1C).

Discussion

M. marinum is a rare cause of infection in humans although some case series have been published. In 2002, Aubry published a case series of 63 patients from across France in a 2-year period. In most cases, M. marinum was limited to the skin, but in 29% of the cases M. marinum had spread to deeper structures with a significant correlation to treatment failure. Of the cases with spread to deeper structures, 83% had tenosynovitis, 39% osteomyelitis and 17% arthritis. Only 3 (5%) cases were located on the lower limb, and 48 (76%) had involvement of a hand. Time from first symptoms to diagnosis was not reported.¹ A case series from a US tertiary referral centre reported 28 cases in an 18-year period. The patients were prone to infections in deeper structures (68%). Most patients had been exposed to fish or contaminated water (87%), and there was a statistically significant correlation between infection through fishes or boating and invasive disease. Invasive M. marinum resulted in tenosynovitis (95%) and was located on the upper limb especially the hands or fingers (93%). The median time to diagnosis was 3.5 months, ranging from 1 to 24 months.³ A Taiwanese case series from a university hospital with 2200 beds found 32 cases of invasive Mycobacterium infections during a 9-year period. Of these cases, 53% had tenosynovitis, and M. marinum was the most common cause of this (41%). Most cases were located on the upper limb (88%). The mean time from fish-related injury to diagnosis was 3.3 months (ranging from 10 days to 4 months).⁴ These case reports show that M. marinum is an infrequent cause of tenosynovitis but it must be considered when a patient presents with indolent tenosynovitis on the upper limb, especially distal. Compared with the case reports, the patient had a severe M. marinum infection which is normally correlated to treatment failure. The relatively long timespan from trauma to diagnosis was also a feature of the current case. This underscores that M. marinum is difficult to diagnose, probably because doctors do not suspect it in their everyday clinical work.

M. marinum is a member of the non-tuberculous mycobacteria family and is an intracellular pathogen. It is a non-motile approximately 2×0.4 mm rod that can be coloured by typical acid-fast staining. The natural habitat for M. marinum is fishes and amphibious animals. Humans get infected after contact with infected fishes or contaminated water, especially with a breached skin barrier.³ Human infections normally result in papulonodular skin infection, but can spread into a sporotrichoid form, or deeper infections such as tendinitis, arthritis, bursitis or osteomyelitis.⁸ It requires specific techniques to culture M. marinum, hence it is important to order specific investigations as previously described.⁸ The incubation period of M. marinum is considered to be 2–4 weeks, however it may range between 5 and 270 days.⁹ The incubation period in the present case is difficult to estimate, because the patient had the infection for some time, before seeking medical attention, but the patient reported approximately 1 month from trauma to debut of symptoms. The biopsy specimen in the current case was examined with culture and PCR analysis as recommended by the American guidelines.¹⁰ Recommended histological analysis was also performed, showing acid-fast bacilli, which is compatible with a Mycobacterium. The culture was performed at 33°C which is warmer than the recommended 30°C.⁸ The PCR analyses were positive for M. marinum after 2 weeks. This was considered sufficient to start treatment. Note that only one PCR analysis was positive when treatment was started. This shows the value of taking several biopsies and inoculations during surgery. All seven cultures were negative after 56 days, which emphasises that M. marinum can be difficult to culture and underscores the value of performing several analyses. It is noteworthy that all cultures were taken while the patient was treated with dicloxacillin, which could influence the result of the culture. In a review of 166 patients treated for M. marinum infection, only 40% had a positive culture, confirming that negative cultures are common in M. marinum infections.⁷ Even though M. marinum infection is strongly suspected, it is relevant to rule out infections caused by other mycobacteria because of their different treatment regimens.¹⁰

Most commercial tests for M. tuberculosis are based on an interferon γ test against the two antigens, early secreted antigenic target-6 and culture filtrate proteins 10. These antigens are shared by some non-tuberculous mycobacteria, including M. marinum. A Japanese study showed that out of 12 patients with M. marinum infection, 8 were positive, 1 was intermediate positive and 3 were negative when tested with QuantiFeron-TB GOLD test.¹¹ Another Japanese study group has published a case study, in which M. marinum infection was detected by a QuantiFeron-TB 2G test screening before the start of the treatment with adalimumab.¹² These studies suggest that it is relevant to perform a QuantiFeron test if M. marinum is suspected, even though the test is not 100% sensitive. A tuberculin skin test can also be used as it was 100% sensitive in a case series of eight patients.¹³ In another case series, the test was less sensitive,¹⁴ and it can be falsely positive in populations vaccinated against tuberculosis with BCG.¹⁵

The American Thoracic Society and the Infectious Disease Society of America have developed guidelines for treatment of M. marinum but do not recommend a specific regime for antibiotics or surgery.¹⁰ Treatment with cyclines and rifampin is suggested,⁸ or a combination of clarithromycin and ethambutol or rifampicin and ethambutol.¹⁶ The patient in the present case was treated with rifampicin, ethambutol and clarithromycin. All agents are recommended, but normally as a part of a two-drug therapy. The patient had severe infection and tolerated the medication well, except for developing diarrhoea. This regime thus seemed to be appropriate for this patient.

Surgical drainage of soft tissue is recommended in severe cases of M. marinum,^{8 10 16} but there are no recommendations concerning osteomyelitis. The patient’s abscesses were drained three times, but the osteomyelitis was not treated surgically. The patient did not have any symptoms from the area where osteomyelitis was located, and surgical treatment was considered risky in regards of sequelae. Therefore, it was decided to treat the osteomyelitis solely with antibiotics and to follow the patient closely with clinical visits. Other surgical intervention showed that the infection was limited to the abscess, other structures were left untouched. The patient recovered, and it appears to have been the correct strategy.

The patient initially had limited symptoms, that evolved into a severe infection with tenosynovitis, arthritis and osteomyelitis. This course may be the natural progression of this particular infection; however, the infection may have been facilitated by the use of immunosuppressants (corticosteroids), which are known to worsen the course of M. marinum infection.² In general, immunosuppressed individuals are more likely to get infections including M. marinum. The patient was 72 years old and had both diabetes and hypertension. The patient received corticosteroid therapy, which is a risk factor, furthermore, the patient was exposed to a fish-related injury, which is an established risk factor.

When diagnosis was confirmed after 2 months, appropriate medical and surgical treatment was initiated, and the infection stopped progressing. After 6 months of antibiotic treatment and three surgical revisions, the patient had recovered and had a minimum of sequelae despite the severe M. marinum infection. The disease would probably have been less widespread and of shorter duration, if M. marinum had been considered earlier, but none of the involved doctors suspected a rare cause of a common clinical presentation, until regular treatment failed.

Patient’s perspective.

The worst part about the disease was the uncertainty, from the start of symptoms to diagnosis. I saw several doctors, and no one could tell me what was wrong.

While I was in treatment, it was difficult not to use my left hand. You find out that most daily activities require two hands, like unbuttoning a shirt or visiting the restroom. In my local rural area, it is also difficult to go anywhere without a car.

The damage is now limited, and the pain and impaired movement does not affect my everyday living. The diarrhoea caused by the antibiotics has an impact on my life, but luckily it is getting better.

Learning points.

• Other differential diagnoses must be considered when treatment response is lacking. It is crucial to re-evaluate the medical history, objective examinations and laboratory tests. In this case, the conclusive clue was hidden in the medical history.

• Mycobacterium marinum is a rare cause of tenosynovitis but must be suspected when tenosynovitis is accompanied by an indolently progressing rash, located distally in the upper limb, is unresponsive to commonly used treatment and has a compatible background story.

• When suspecting a diagnosis, it is important to secure relevant specimens for further examinations to establish the correct diagnosis and most appropriate treatment. In the current case, the patient had two biopsies and one inoculation at the first surgical visit, and only one biopsy was positive.

• When suspecting M. marinum, attention to the correct procedure for culture and PCR analysis is of utmost importance.