Download Mycobacterium terrae: a case report

Journal of Orthopaedic Surgery 2009;17(1):103-8
Infectious arthritis of the knee caused by
Mycobacterium terrae: a case report
BW Milne, MH Arnold, B Hudson, MRJ Coolican
Royal North Shore Hospital, University of Sydney, Sydney, Australia
INTRODUCTION
ABSTRACT
Mycobacterium terrae is ubiquitous in our environment.
M terrae infections most commonly involve tendon
sheaths, bones, bursae, and joints. We report a case
of infectious arthritis of the knee caused by M terrae
in a 21-year-old man who had non-specific chronic
synovitis. No organism was seen on microscopy or
isolated from cultures until months later. Initially
the M terrae culture was considered a contaminant
and specific anti-mycobacterial treatment was not
advised. The patient was commenced on suppressive
therapy for persistent effusion and discomfort.
Eventually, the M terrae infection was confirmed and
he was commenced on clarithromycin, ciprofloxacin,
and ethambutol. The triple antibiotic regimen was
continued for 2 years. The knee improved but never
completely settled. The patient chose to cease all
antibiotic medication. The knee remained swollen
and irritable, with little chance of eradicating the
organism.
Key words: arthritis, infectious; mycobacteria, atypical
Mycobacterium terrae complex (M terrae, M
nonchromogenicum, and M triviale) has been classified
as a non-pathogenic saprophyte, but is increasingly
considered a pathogen. It is one of approximately 50
species of atypical mycobacteria or non-tuberculous
mycobacteria capable of causing disease. These
bacteria are ubiquitous in our environment and
produce 6 major clinical syndromes1: chronic
bronchopulmonary disease, lymphadenitis, skin and
soft tissue disease, skeletal (bone, joint, and tendon)
disease, and disseminated and catheter-related
infections. The most common presentation of M
terrae infection is infection of tendon sheaths, bones,
bursae, and joints, usually after direct inoculation.
We present a case of infectious arthritis of the knee
caused by M terrae in a 21-year-old man. The modes
of transmission, difficulties with diagnosis, and
treatment options are reviewed.
CASE REPORT
In April 1997, a 21-year-old man presented with a
10-month history of a painful, swollen left knee.
Address correspondence and reprint requests to: Dr BW Milne, Department of Orthopaedics and Trauma Surgery, Royal North
Shore Hospital, St Leonards, NSW, 2065, Australia. E-mail: [email protected]
Journal of Orthopaedic Surgery
104 BW Milne et al.
He had undergone 2 arthroscopies: the first in
1994 for a partial medial meniscectomy after a
sporting injury, and the second in February 1995
for a hamstring autograft anterior cruciate ligament
(ACL) reconstruction after another sporting injury.
The knee was stable but persistently swollen and
uncomfortable. An arthroscopy confirmed an active
chronic synovitis with a possible granulomatous
component. The inflammatory fluid contained no
organisms or crystals. No acid-fast bacilli were seen.
He gave no history of a penetrating injury to the
knee while working as a landscape gardener, and
was provisionally diagnosed with an unexplained
inflammatory mono-arthritis. Differential diagnoses
included sarcoidosis, tuberculosis, and a foreign
body synovitis. Further radiographic and serological
investigations were unremarkable and he was referred
for physiotherapy.
Six weeks later, the knee remained swollen and
inflamed. An aspiration confirmed an inflammatory
effusion with no organisms seen on microscopy. Intraarticular triamcinolone was injected and suppressive
therapy with salazopyrin was commenced. 10 weeks
later, M terrae was isolated from the culture and the
salazopyrin was ceased.
An urgent microbiological review and aspiration
again yielded no mycobacteria. It was unknown
whether the M terrae was a commensal, pathogen
or contaminant. The microbiologist was reluctant to
prescribe prolonged anti-tuberculous therapy without
further bacteriological and histological evidence. The
patient therefore underwent a diagnostic arthroscopy
and biopsy by an orthopaedic surgeon.
Repeat radiographs demonstrated a small
cyst in the lateral femoral condyle, minor medial
compartment wear, and features of the previous ACL
reconstruction (Fig. 1). Bone scans confirmed a mild
(a)
Figure 1
(b)
(c)
to moderate diffuse synovitis of the left knee (Fig. 2).
In December 1997, the patient underwent a
further arthroscopy, synovial biopsy, and arthroscopic
synovectomy for moderate synovitis (Fig. 3).
Inflammatory fluid and mycobacterial cultures
remained negative 12 weeks later. A histopathological
examination showed an active chronic synovitis
with numerous epithelioid granulomata, which was
consistent with mycobacterial infection, foreign body
reaction (despite no foreign material on polarisation)
or sarcoidosis.
Specific anti-mycobacterial treatment was not
advised as the original M terrae culture was considered
a contaminant. The patient was commenced on
suppressive therapy with methotrexate at 6 weeks
for the persistent effusion and discomfort, but this
was ceased after a 6-month trial as the symptoms
persisted. The patient moved out of the area and was
referred to a rheumatologist for follow-up.
In July 2001, the patient was referred back
with a 3-month history of a swollen left knee. The
rheumatologist had isolated M terrae from aspirates.
Radiographs demonstrated the cystic changes in
the lateral femoral condyle and marked swelling
consistent with a joint effusion (Fig. 1). His blood
tests remained unremarkable. The inflammatory
fluid was negative on microscopy but positive
on cultures 7 weeks later. He was commenced on
clarithromycin and ciprofloxacin before sensitivity
results became available. The organism was sensitive
to clarithromycin and ethambutol but resistant to
ciprofloxacin and rifampicin. Ethambutol was then
added. Magnetic resonance imaging at 3 months
showed an extensive erosive arthropathy with
marked synovial proliferation and erosions in the
femoral condyles and posteriorly along both tibial
plateaux (Fig. 4).
(d)
Radiographs of the left knee in (a) April 1997, (b) October 1997, (c) May 2001, and (d) May 2001 (magnified view).
Vol. 17 No. 1, April 2009
Infectious arthritis of the knee caused by Mycobacterium terrae 105
Figure 3 Arthroscopy of the left knee in December 1997
showing moderate synovitis.
Figure 2 Bone scans in November 1997 showing mild to
moderate diffuse synovitis of the left knee.
The triple antibiotic regimen was continued for
2 years until November 2003. The knee improved
but never completely settled. From December 2002
onwards, improvement was minimal, with a persistent
granulomatous synovitis on serial magnetic resonance
imaging. In February 2005, the patient underwent
synovectomy owing to ongoing pain and swelling in
his knee. Once again, the straw-coloured inflammatory
fluid collected did not yield any organisms on both
microscopy and initial culturing. A histopathological
examination revealed a chronic non-caseating
granulomatous synovitis. In the early postoperative
period, the patient developed a pulmonary embolus
and was commenced on warfarin for a 6-month
period. 12 weeks later, the cultures became positive
for M terrae, which was sensitive to clarithromycin,
amikacin, ciprofloxacin, and rifampicin. The patient
was re-commenced on clarithromycin, ciprofloxacin,
and ethambutol indefinitely. As the effusion remained
6 months later, linezolid (an oxazolidinone) was
trialled in combination with clarithromycin and
ethambutol, once the warfarin had been ceased. The
linezolid was stopped because the patient developed
peripheral neuropathy in his legs and clofazimine was
then added. This was stopped after 3 months because
Figure 4 Magnetic resonance imaging of the left knee in September 2001 showing an extensive erosive arthropathy with
marked synovial proliferation and erosions in the femoral condyles and posteriorly along both tibial plateaux.
106 BW Milne et al.
of possible adverse effects. In November 2005, the
patient chose to cease all antibiotic medication.
His knee remained swollen and irritable, and there
appeared little chance of eradicating the organism.
DISCUSSION
M terrae was first isolated in 1950 from radish washings
and was thus known as the radish bacillus.2 Based on
its laboratory behaviour and culture characteristics, it
was classified as non-pathogenic. In 1966, the radish
bacillus was isolated from human sputum and gastric
lavage samples. This soil organism occasionally
inhabits human hosts’ secretions as a non-pathogenic
coloniser. It was designated M terrae because of
its ubiquitous presence in soil. In 1967, the Centers
for Disease Control in the United States reported 5
clinically significant cases of M terrae infection. Over
50 cases of M terrae infection have been reported,
involving bone and other synovial structures,3–18
lungs,19–22 skin,19,23 gut,24 urinary tract, lymph nodes or
disseminated disease.19,23,25 The most common site is
the tenosynovium of the hand and wrist. Two cases of
M terrae infection in large joints have been reported.
One involved the hip in a 6-month-old infant3 and
another involved the knee in a 57-year-old man with
rheumatoid arthritis.5 Mycobacterial infection of
larger joints is usually due to M tuberculosis or other
non-tuberculous mycobacteria such as M intracellulareavium complex, M kansasii or M marinum.5,7,13 Most
patients are otherwise healthy with no concurrent
medical illnesses or immunosuppression.
Inoculation of M terrae usually occurs through
direct trauma or a penetrating injury. Direct
inoculation is reported in up to 75% of cases.15 In
our patient, inoculation was either through the ACL
surgery or a minor penetrating injury while working
as a landscape gardener. Haematogenous spread from
a distant site is possible, but has never been reported
in musculoskeletal M terrae infection.
Nosocomial infections of atypical mycobacteria
have been reported.26,27 They are environmentally
ubiquitous organisms and therefore present in most
municipal water supplies. A breach of sterilisation
protocols is a possible explanation, for example
during arthroscopy or the ACL reconstruction.
Contamination of medical equipment by tap water
during the cleaning process has caused numerous
outbreaks and pseudo-outbreaks of non-tuberculous
mycobacteria. Inadequate cleaning of instruments,
inappropriate disinfectant concentrations, outdated
disinfectants, insufficient cleaning time, and the use
of tap water can cause a breach in sterilisation. These
Journal of Orthopaedic Surgery
bacteria are more resistant to elevated temperatures,
disinfectants, and ultra-violet light because of their
high lipid content and multi-layered cell walls. A
rise in the prevalence of such infections has been
reported when hospitals lower the temperature in the
hot water systems to prevent scalding. Sterilisation
equipment should be cleaned and serviced regularly,
as mycobacteria can survive in biofilms inside the
plastic and rubber tubing. Some atypical mycobacteria
are resistant to chlorine, 2% formaldehyde and
alkaline glutaraldehyde solutions, organomercurials,
and other commonly used disinfectants. Our
case was an isolated event with no other atypical
mycobacterial infections reported during this period.
The sterilisation system used during arthroscopy and
ACL reconstruction was the STERRAD Sterilization
Systems (Johnson & Johnson Gateway), which has
not been linked to atypical mycobacterial outbreaks
or pseudo-outbreaks.
M terrae infection of synovial structures is
difficult to diagnose because it is uncommon
and non-specific and is easily mistaken for other
systemic or regional rheumatological conditions.4,7,28
Diagnosis can be delayed by use of erroneous or
ineffective treatments involving systemic or local
corticosteroids, non-steroidal anti-inflammatory
drugs, immunosuppressive agents, splintage,
physiotherapy, or even amputation.7,8,11 The mean
time for diagnosing M terrae infection from symptom
onset has been reported as 12.6 months, with more
than half taking over 6 months.16 M terrae may not
be consistently isolated from tissue or synovial fluid
samples; even so it is unknown whether the M terrae
is a commensal, pathogen or contaminant. Acid-fast
bacilli are often not seen on microscopy and grown in
cultures until weeks to months later. The histological
appearance of synovial fluid samples is non-specific
and can only provide supportive evidence.28
Non-tuberculous mycobacteria should always be
considered in the differential diagnosis of a chronic
mono-arthritis. Chronic infections of any type should
be excluded before injection of, or treatment with,
corticosteroids or other immunosuppressives. The
criteria for diagnosis of nontuberculous mycobacterial
infection are22: (1) repeated isolation of the particular
organism at high concentrations over several weeks,
(2) manifestation of a disease that could be caused by
such an organism, and (3) lack of an associated cause
for the disease.
It was unknown whether the initial M terrae
culture was a contaminent or a pathogen. An outbreak
of 163 false positive M terrae infections secondary to
contaminated hospital water supply29 and a pseudooutbreak of M terrae secondary to contamination of
Vol. 17 No. 1, April 2009
Infectious arthritis of the knee caused by Mycobacterium terrae 107
specimens in the laboratory30 have been reported. A
single isolation of M terrae is easily dismissed as a
contaminant or non-pathogenic coloniser, because it
is ubiquitous in temperate climates.
Management of M terrae infection is controversial,
with a variety of options reported: surgery alone,
antimycobacterial therapy alone, monotherapy or
multi-drug therapy, or a combination of surgical
and antimycobacterial treatment. A combination of
debridement and multiple chemotherapeutic agents
for a prolonged period provides the best results.16,17
Initially, our patient responded well to drug therapy
alone, thus arthroscopic synovectomy was delayed.
The organism was not eliminated, however, despite
prolonged antibiotic treatment and multiple
arthroscopic synovectomies.
M terrae can be multi-drug resistant when cultured.
Sensitivity results may vary between laboratories
because there is no standardised system for defining
organism sensitivity and resistance.1 Sensitivities
differ between in vivo and in vitro conditions.16,23 No
empirical treatment regimen exists because of the
paucity of in vitro testing of M terrae isolates.16 No
single agent or drug combination is significantly
better than others, but a combination of ethambutol
and rifampicin is tending toward significance.16 The in
vitro susceptibility testing has found that ethambutol
is most effective on agar plates, and in minimum
inhibitory concentration studies, semisynthetic
macrolides such as clarithromycin are particularly
effective, as are ethambutol and aminoglycosides.16
Use of an empirical regimen of clarithromycin,
rifampicin, and ethambutol is recommended; it can
be altered based on individual sensitivity testing and
should be continued for at least 12 months after the first
signs of clinical benefit.16 This regimen is also effective
for other non-tuberculous mycobacterial infections.
In our case, ciprofloxacin was continued despite its in
vitro resistance, because in vitro sensitivity testing is
not always predictable. Linezolid and clofazimine are
useful for treating other mycobacterial infections,31,32
but were not successful in our patient.
REFERENCES
1. Brown BA, Wallace RJ Jr. Infections due to nontuberculous mycobacteria. In: Mandell GL, Bennett JE, Dolin R, editors.
Mandell, Douglas, and Bennett’s principles and practices of infectious diseases. Vol 2. 5th ed. Philadelphia: Churchill
Livingstone; 2000:2630–6.
2. Richmond L, Cummings MM. An evaluation of methods of testing the virulence of acid-fast bacilli. Am Rev Tuberc
1950;62:632–7.
3. Dechairo DC, Kittredge D, Meyers A, Corrales J. Septic arthritis due to Mycobacterium triviale. Am Rev Respir Dis
1973;108:1224–6.
4. Deenstra W. Synovial hand infection from Mycobacterium terrae. J Hand Surg Br 1988;13:335–6.
5. Dijkmans BA, Mouton RP, Macfarlane JD, et al. Bacterial arthritis caused by Mycobacterium terrae. Infection 1981;9:204–
7.
6. Edwards MS, Huber TW, Baker CJ. Mycobacterium terrae synovitis and osteomyelitis. Am Rev Respir Dis 1978;117:161–
3.
7. Fodero J, Chung KC, Ognenovski VM. Flexor tenosynovitis in the hand caused by Mycobacterium terrae. Ann Plast Surg
1999;42:330–2.
8. Halla JT, Gould JS, Hardin JG. Chronic tenosynovial hand infection from Mycobacterium terrae. Arthritis Rheum
1979;22:1386–90.
9. Huskisson EC, Doyle DV, Fowler EF, Shaw EJ. Sausage digit due to radish bacillus. Ann Rheum Dis 1981;40:90–1.
10. Karthigasu KT, Spagnolo DV, Gow BL. Mycobacterium terrae tenosynovitis. Pathology 1990;22:106–7.
11. Kremer LB, Rhame FS, House JH. Mycobacterium terrae tenosynovitis. Arthritis Rheum 1988;31:932–4.
12. Love GL, Melchior E. Mycobacterium terrae tenosynovitis. J Hand Surg Am 1985;10:730–2.
13. May DC, Kutz JE, Howell RS, Raff MJ, Melo JC. Mycobacterium terrae tenosynovitis: chronic infection in a previously
healthy individual. South Med J 1983;76:1445–7.
14. Mehta JB, Hovis WM. Tenosynovitis of the forearm due to Mycobacterium terrae (radish bacillus). South Med J 1983;76:1433–
5.
15. Rougraff BT, Reeck CC Jr, Slama TG. Mycobacterium terrae osteomyelitis and septic arthritis in a normal host. A case report.
Clin Orthop Relat Res 1989;238:308–10.
16. Smith DS, Lindholm-Levy P, Huitt GA, Heifets LB, Cook JL. Mycobacterium terrae: case reports, literature review, and in
vitro antibiotic susceptibility testing. Clin Infect Dis 2000;30:444–53.
17. Zenone T, Boibieux A, Tigaud S, Fredenucci JF, Vincent V, Chidiac C, et al. Non-tuberculous mycobacterial tenosynovitis: a
review. Scand J Infect Dis 1999;31:221–8.
18. Eskesen AN, Skramm I, Steinbakk M. Infectious tenosynovitis and osteomyelitis caused by Mycobacterium
nonchromogenicum. Scan J Infect Dis 2007;39:179–80.
19. Carbonara S, Tortoli E, Costa D, Monno L, Fiorentino G, Grimaldi A, et al. Disseminated Mycobacterium terrae infection in
a patient with advanced human immunodeficiency virus disease. Clin Infect Dis 2000;30:831–5.
108 BW Milne et al.
Journal of Orthopaedic Surgery
20. Kuze F, Mitsuoka A, Chiba W, Shimizu Y, Ito M, Teramatsu T, et al. Chronic pulmonary infection caused by Mycobacterium
terrae complex: a resected case. Am Rev Respir Dis 1983;128:561–5.
21. Spence TH, Ferris VM. Spontaneous resolution of a lung mass due to infection with Mycobacterium terrae. South Med J
1996;89:414–6.
22. Tonner JA, Hammond MD. Pulmonary disease caused by Mycobacterium terrae complex. South Med J 1989;82:1279–
82.
23. Peters EJ, Morice R. Miliary pulmonary infection caused by Mycobacterium terrae in an autologous bone marrow transplant
patient. Chest 1991;100:1449–50.
24. Fujisawa K, Watanabe H, Yamamoto K, Nasu T, Kitahara Y, Nakano M. Primary atypical mycobacteriosis of the intestine: a
report of three cases. Gut 1989;30:541–5.
25. Cianciulli FD. The radish bacillus (Mycobacterium terrae): saprophyte or pathogen? Am Rev Respir Dis 1974;109:138–
41.
26. Phillips MS, von Reyn CF. Nosocomial infections due to nontuberculous mycobacteria. Clin Infect Dis 2001;33:1363–74.
27. Wallace RJ Jr, Brown BA, Griffith DE. Nosocomial outbreaks/pseudo-outbreaks caused by nontuberculous mycobacteria.
Annu Rev Microbiol 1998;52:453–90.
28. Marchevsky AM, Damsker B, Green S, Tepper S. The clinicopathological spectrum of non-tuberculous mycobacterial
osteoarticular infections. J Bone Joint Surg Am 1985;67:925–9.
29. Lockwood WW, Friedman C, Bus N, Pierson C, Gaynes R. An outbreak of Mycobacterium terrae in clinical specimens
associated with a hospital potable water supply. Am Rev Respir Dis 1989;140:1614–7.
30. Bettiker RL, Axelrod PI, Fekete T, St John K, Truant A, Toney S, et al. Delayed recognition of a pseudo-outbreak of
Mycobacterium terrae. Am J Infect Control 2006;34:343–7.
31. Fung HB, Kirschenbaum HL, Ojofeitimi BO. Linezolid: an oxazlidinone antimicrobial agent. Clin Ther 2001;23:356–91.
32. Reddy VM, O’Sullivan JF, Gangadharam PR. Antimycobacterial activities of riminophenazines. J Antimicrob Chemother
1999;43:615–23.