Case Report Mycobacterial Central Venous Catheter Tunnel Infection: a Difﬁcult Problem

Bone Marrow Transplantation, (1999) 24, 325–329  1999 Stockton Press All rights reserved 0268–3369/99 $12.00 http://www.stockton-press.co.uk/bmt Case report Mycobacterial central venous catheter tunnel infection: a difﬁcult problem

MS Ward, KV Lam, PK Cannell and RP Herrmann

Haematology Department, Royal Perth Hospital, Perth, Australia

Summary: Case reports

We report our experience of non-tuberculous myco- Case 1 bacterial infection associated with the tunnel of Hick- man–Broviac central venous catheters in immunosup- A 28-year-old woman with AML achieved complete pressed patients with haematological malignancies remission with the first course of standard induction chemo- undergoing high-dose chemotherapy supported by therapy. Prolonged aplasia for several months complicated BMT. The problem is rare and difficult to treat. Our the first consolidation therapy. One year after diagnosis an cases are unique in developing tunnel site mycobac- abscess developed adjacent to the site of the previously terial infection well after the tunnelled catheters were removed Groshong catheter, with persistently negative removed. We diagnosed one case of Mycobacterium microbiological cultures (Figure 1). chelonae, which is a well-documented cause of such Fifteen months after initial diagnosis the AML relapsed. infections, and two cases of Mycobacterium haemophi- She went on to allogeneic bone marrow transplantation lum, which are the first reported cases in this setting. (1992) using busulphan and cyclophosphamide condition- Early wide surgical excision of the infected tunnel site ing with short-course methotrexate and cyclosporin graft- and prolonged antibiotic therapy is necessary. Despite versus-host disease (GVHD) prophylaxis. Grade II/III these measures recurrence occurred in two cases. Close cutaneous, hepatic and gastrointestinal GVHD was treated liaison with the microbiology laboratory is needed to with prednisolone 25 mg daily and Psoralen Ultra Violet A (PUVA) therapy. ensure the appropriate culture media and conditions Three months later the left chest wall lesion had deterio- are used for these fastidious organisms. Empiric anti- rated and numerous granulomata and acid-fast bacilli, later biotic regimens should be based on the likely organism. identified as Mycobacterium chelonae (abscessus), was Drugs active against M. chelonae and M. haemophilum isolated at biopsy. Wide surgical excision was undertaken. should be included. The lesion penetrated to within 0.5 cm of the pectoralis Keywords: mycobacteria; central venous catheter; Hick- major muscle and was extensive, necessitating subtotal man line; infection; M. chelonae; M. haemophilum mastectomy. Anti-mycobacterial therapy was i.v. amikacin 500 mg once daily and i.v. cefoxitin 2 g four times a day. A few months later M. chelonae recurred, just above the Tunnelled venous access devices are used commonly in scar, with 2–3 cm of induration. Local excision and imip- bone marrow transplantation to allow long-term venous enem were used initially and later amikacin and clarithro- access. These lines are associated with infectious and mycin (1 g twice a day). The area remained ulcerated for thrombotic complications. Mycobacterial infections are months, but was slowly healing. Intensive immunosuppres- rarely encountered. Tunnel-related infection is even less sion was continued because of progressive GVHD includ- frequently reported. We discuss the problem, outline our ing obliterative bronchiolitis. experience and review the literature. One year later a further lesion developed which was indurated, red and tender; the old lesion ulcerated. Amika- cin and cefoxitin were employed as the organism had become resistant to clarithromycin. Deafness was noted, probably due to cumulative toxicity of amikacin; although amikacin levels were never toxic. Death occurred 3 years after initial presentation from respiratory failure (obliterative bronchiolitis), disseminated mycobacterial Correspondence: Dr MS Ward, Haematology Department, Royal Perth sepsis (blood cultures positive for M. chelonae) and Hospital, Wellington Street, Perth WA 6000, Australia chronic GVHD. Received 15 September 1998; accepted 11 March 1999 Mycobacterial CVC tunnel infection MS Ward et al 326

Figure 1 Photograph of ulcerated area adjacent to previous Groshong catheter site (case 1).

Case 2 and no sign of recurrence. Immunosuppressive or immunomodulating therapy has been withheld indefinitely. A 29-year-old woman was diagnosed with intermediate- grade non-Hodgkin’s lymphoma (diffuse mixed small and large cell) at 26 weeks gestation. Skin, lung, liver and spleen were involved. The baby was delivered early, and Case 3 CEOP chemotherapy was commenced, without improve- ment. A salvage regimen containing dexamethasone, high- A 26-year-old woman with a past history of epilepsy was dose cytarabine and carboplatin was used, followed by dou- diagnosed with blastic transformation of CML and received ble autologous peripheral blood stem cell transplantation at two cycles of idarubicin, high-dose cytarabine and etopo- 3 and 5 months after diagnosis (1998). Interleukin-2 and side chemotherapy. Matched unrelated BMT was perfor- interferon-alpha were then employed. Three months after med 5 months after diagnosis (1998) using total body the second autograft an inflamed mass developed in the irradiation and cyclophosphamide conditioning followed by right supraclavicular region with some inflammation at the short-course methotrexate, Campath 1G, steroid and cyclo- CVC site. Cellulitis with lymphadenopathy were suspected sporin GVHD prophylaxis. and flucloxacillin commenced. The lesion became harder Three months after transplant the site of a previous Hick- and less inflamed but did not disappear. Fine needle aspir- man catheter (removed after only a few hours because of ation (FNA) cytology failed to demonstrate recurrent lym- accidental dislodgement) became ulcerated and infected, phoma or any microbial organism. Only neutrophils were associated with a purulent discharge. Flucloxacillin was seen. Since the lesions failed to respond, antibiotics were prescribed but swabs demonstrated leucocytes with no changed to ciprofloxacin and cephalexin. The lesion slowly microbial growth. The ulcer persisted and cultures from deteriorated, and new ulcerated lesions developed along the swabs grew abundant Proteus and Pseudomonas, which track of the previous catheter. Further FNA was perfomed were treated with ciprofloxacin. Numerous rapid-growing which demonstrated acid-fast bacilli. Wide surgical exci- mycobacteria were also detected after 5 days. Wide surgical sion of the catheter track and two adjacent lymph nodes excision was performed down to 2 cm from the muscle. confirmed granulomata with multi-nucleate giant cells. Oral clarithromycin 500 mg twice a day, i.v. amikacin One of the nodes contained a few acid-fast bacilli. Culture 500 mg daily and i.v. meropenem 500 mg three times a day confirmed Mycobacterium haemophilum after 7 (subsequently changed to i.v. cefoxitin 2 g twice a day) weeks incubation. Therapy consisted of amikacin, clari- were commenced. The immunosuppression was reduced. thromycin, ciprofloxacin and meropenem for 3 weeks, Identification of Mycobacterium haemophilum took 7 then amikacin and meropenem were ceased and ethambutol weeks. Figure 2 demonstrates numerous mycobacteria in started. She continues on this regimen with a healed wound the biopsy specimen. Mycobacterial CVC tunnel infection MS Ward et al 327 Table 1 Summary of published cases of catheter-related Mycobacterial infection

Study Cases Setting Species Site Therapy

Roy et al7 6/2241 BMT 3 M. chelonae 3 tunnel infection 1–6 months Ab: amik, clar, cipro, adult and paediatric 3 M. fortuitum 3 CRB cotrim. ± line removal. Surgery not mentioned. Raad et al8 15 Cancer 9 M. fortuitum 4 local line sepsis CRB: Catheter removal + Ab. 6 M. chelonae 11 CRB Tunnel infection: required surgical excision. Exit site infection: catheter removal only. Amik, cotrim, cefox. Flynn et al9 6 Haematological 4 M. chelonae Hickman 3 cases of tunnel infection recurred with malignancy (paediatric) 2 M. fortuitum catheter removal and Ab, which resolved with surgical excision. Holland et al10 1 BMT M. neoaurum Hickman catheter Ab (tic/clav, tob) then line removal. Esteban et al11 1 HD chemo, NHL M. aurum Tunnelled catheter line removal + Ab: amik, clari. Skeitz12 1 Cancer M. smegmatis CRB line removal + Ab. Groeger et al3 1/1431 Cancer M. chelonae tunnel infection No details. Ward et al (this 3/401 BMT 1 M. chelonae tunnel site infection surgical excision + Ab. paper) 2 M. haemophilum (catheter previously Amik, cefox, imip, cipro, clar. removed) Amik, clar, cipro, mero, etham. Amik, clar, rif, cipro.

CRB = catheter-related bacteraemia; Ab = antibiotic therapy; amik = amikacin; clar = clarithromycin; cotrim = co-trimoxazole (sulphamethoxazole/trimethoprim); cefox = cefoxitin; rif = rifampicin; tic/clav = ticarcillin/clavulanic acid; tob = tobramycin; imip = imipenem; mero = meropenem; etham = ethambutol; cipro = ciproﬂoxacin.

Two months later the ulcer reappeared and a few AFB relied upon catheter removal and antibiotics. They also were isolated. Repeat surgical excision (to the pectoralis noted, as we have, that tunnel infection required surgical muscle including a cuff of muscle) was undertaken. excision. Flynn et al9 also found that in three patients the The drugs were changed to oral rifampicin 600 mg daily, local infection relapsed after the catheter was removed and ciproﬂoxacin 750 mg twice a day and clarithromycin the patients had received parenteral therapy for Ͼ3 weeks. 500 mg twice a day. Cyclosporin and prednisolone Infection only resolved after surgical excision of the doses were reduced further. Even though amikacin levels infected tunnel.9 M. neoaurum infection of a Hickman cath- were never toxic, unfortunately deafness and vertigo eter was reported in an allogeneic BMT recipient with developed after 8 weeks. Currently the skin is intact and recovery following line removal and antimicrobial ther- therapy continues. apy.10 More recently a case of M. aurum related to a catheter resolved only after antibiotics and subsequent catheter removal.11 A case of M. smegmatis catheter-related bacteraemia in a cancer patient also responded to line removal and Discussion antibiotics.12 Only one episode of mycobacterial sepsis was noted in the large study of 1630 tunnelled venous access Broviac1–Hickman2 style cuffed tunnelled venous access catheters in 1431 cancer patients at the Memorial Sloan- catheters are frequently used in bone marrow transplan- Kettering Cancer Center3 due to M. chelonei, although no tation. Infection of tunnelled catheters are relatively com- details were given. Two cases of non-tuberculous Myco- mon, occurring in 13–43% of patients with cancer.3–5 The bacterial catheter-related sepsis13 and two cases of Myco- infections are mainly due to gram-negative bacilli and bacterium avium complex bacteraemia14 have been reported gram-positive cocci, with some caused by fungi.6 in children with cancer. Mycobacterial infections related to catheters have been Our cases illustrate the clinical problem of non-tubercu- rarely reported (Table 1). The largest study of mycobac- lous mycobacterial infections of the catheter tunnel site. terial infection following BMT, from the University of The problem is thankfully rare in the BMT setting: three Minnesota,7 reports only 11 cases of mycobacterial infec- cases since 1983 during which time we have transplanted tion of any site in 2241 BMT recipients over a 20-year 401 people (0.75% mycobacterial infection rate). The rate period. In total only six patients had Mycobacteria (all M. of all catheter-related mycobacterial infection at Minnesota chelonae or M. fortuitum) from blood culture or CVC tun- is 0.27%.7 There is little written in the literature regarding nel. Only three presented with CVC tunnel inﬂammation. tunnel-related infection. The major problem is that these All seemed to respond well to CVC removal and 1–6 mycobacteria are ubiquitous environmental organisms and months of antimicrobial therapy. Surgery was not men- infections occur in heavily immunosuppressed patients who tioned. Fifteen cases of catheter-related Mycobacterium for- are recovering from BMT, with reduced cell-mediated tuitum complex (fortuitum/chelonae) have been diagnosed immunity, and often with concomitant GVHD. There- at the MD Anderson Cancer Center.8 Successful treatment fore reduction of immunosuppressive therapy, although Mycobacterial CVC tunnel infection MS Ward et al 328

Figure 2 Photomicrograph of numerous Mycobacterium haemophilum from biopsy of case 3 (arrowed) (original magniﬁcation ×40; Ziel–Neilson stain).

desirable, is often not an option. Immunosuppression, M. fortuitum) should include surgical excision of all particularly loss of cell-mediated immunity and neutro- involved tissue. penia are the most important risk factors. Haematological Antibiotics recommended for M. fortuitum are amikacin, malignancy also seems to be more common than non- cefoxitin, ciprofloxacin and sulphonamides and for M. haematological malignancy. Foreign bodies and chelonae amikacin, doxycycline, erythromycin, imipenem, materials (such as catheters) also increase the risk for these and clarithromycin although M. chelonae tends to be more infections.8,9 resistant to drug therapy. M. haemophilum has a much more The therapy has toxicity, mainly attributable to the limited clinical database (with less than 100 reported cases). aminoglycosides. Two of our cases developed vestibulo- Responses to surgery, modification of immune suppression, cochlear damage. This problem seems to occur after some and antibiotics such as rifampicin, doxycycline and cotri- time (cumulative dose), and does not have to be associated moxazole have been reported. with toxic levels. Amikacin is thought to preferentially Our experience of tunnelled catheters in BMT patients affect auditory function rather than vestibular function,15 is that these infections, in contrast to other reported series, although one of our cases developed both, and never had occurred in relation to the tunnelled catheter site well after toxic levels with frequent assays. The onset can be abrupt. the line had been removed. Additionally wide surgical exci- Regular monitoring of function by audiometry would be sion was required, and our patients had ongoing infection prudent as the damage may be reversible if therapy is with documented recurrence despite antimicrobial therapy ceased before symptoms develop. and prompt wide surgical excision. We find these infections Mycobacteria are notoriously fastidious, and identifi- to be a difficult clinical problem causing significant mor- cation can take up to 2 months, with a longer delay for bidity. M. chelonae has been documented as causing these antibiotic sensitivity testing. The diagnosis needs to be sus- infections, but this is the first report of M. haemophilum in pected and cultures set-up on appropriate media. Empiric this setting. M. haemophilum is becoming an increasingly antibiotic cover, based on likely organisms, should be recognised cause of infection in the immunosuppressed. It initiated until the organism is later identified. PCR tech- predominantly causes cutaneous lesions, grows best in niques are now used to assist and speed the identification cooler environments, and is difficult to culture even on of standard Mycobacterium tuberculosis, but primers are appropriate media.18,19 It has also been implicated in bac- not available for the atypical mycobacteria. There is no teraemia, osteomyelitis, septic arthritis, pneumonitis in the ‘standard’ therapy for these organisms as they are rare, and immunosuppressed, and lymphadenitis in otherwise healthy display varying sensitivity to antibiotics. Current regimens children.20,21 The only other report of M. haemophilum in used include combinations of drugs most likely to be BMT recipients is from the Memorial Sloan-Kettering Can- active. According to Mandell, Douglas and Bennett’s cer Center. Three patients presented with the classical Principles and Practice of Infectious Diseases 4th edi- cutaneous lesions and survived, and two others developed tion,16,17 therapy for the rapid growers (M. chelonae and isolated pulmonary disease and died.22 M. haemophilum Mycobacterial CVC tunnel infection MS Ward et al ´ ´ 329 should be added to the list of possible causative organisms, 11 Esteban J, Fernandez-Roblas R, RomanAet al. Catheter- and the choice of empiric antibiotic therapy should take related bacteremia due to Mycobacterium aurum in an immun- this into account. compromised host. Clin Infect Dis 1998; 26: 496–497. 12 Skiest DJ, Levi ME. Catheter-related bacteremia due to Myco- bacterium smegmatis. South Med J 1998; 91: 36–37. 13 Rodgers GL, Mortensen JE, Blecker-Shelly D et al. Two case reports and review of vascular catheter-associated bacteremia References caused by nontuberculous Mycobacterium species. Ped Infect Dis J 1996; 15: 260–264. 1 Broviac JW, Cole JJ, Scribner BH. A silicone rubber atrial 14 Schelonka RL, Ascher DP, McMahon DP et al. Catheter- catheter for prolonged parenteral alimentation. Surg Gynecol related sepsis caused by Mycobacterium avium complex. Pedi- Obstet 1973; 136: 602–606. atr Infect Dis J 1994; 13: 236–237. 2 Hickman RO, Buckner CD, Clift RA et al. A modified right 15 Bendush CL. Ototoxicity: clinical considerations and com- atrial catheter for access to the venous system in marrow parative information. In: Whelton A, Neu HC (eds). The transplant recipients. Surg Gynecol Obstet 1979; 148: Aminoglycosides: Microbiology, Clinical Use and Toxicology. 871–875. Marcel Dekker: New York, 1982, pp 427–429. 3 Groeger JS, Lucas AB, Thaler HT et al. Infectious morbidity 16 Horowitz EA, Sanders Jnr WE. Other mycobacterium species. associated with long-term use of venous access devices in In: Mandell GL, Bennett JE, Dolin R (eds). Principles and patients with cancer. Ann Intern Med 1993; 119: 1168–1174. Practice of Infectious Diseases. Churchill Livingstone: New 4 Eastridge BJ, Lefor AT. Complications of indwelling venous York, 1995, pp 2269–2271. access devices in cancer patients. J Clin Oncol 1995; 13: 17 Alford RH, Wallace Jnr RJ. Antimycobacterial agents. In: 233–238. Mandell GL, Bennett JE, Dolin R (eds). Principles and Prac- 5 Howell PB, Walters PE, Donowitz GR, Farr BM. Risk fac- tice of Infectious Diseases. Churchill Livingstone: New York, tors for infection of adult patients with cancer who have 1995, pp 395–397. tunnelled central venous catheters. Cancer 1995; 75: 18 Woods GL, Washington JA. Mycobacteria other than Myco- 1367–1375. bacterium tuberculosis: review of microbiologic and clinical 6 Kiehn TE, Armstrong D. Changes in the spectrum of organ- aspects. Rev Infect Dis 1987; 9: 275–294. isms causing bacteraemia and fungaemia in immunocompro- 19 Kristjansson M, Bieluch VM, Byeff PD. Mycobacterium mised patients due to venous access devices. Eur J Clin haemophilum infection in immunocompromised patients: case Microbiol Infect Dis 1990; 9: 869–872. report and review of the literature. Rev Infect Dis 1991; 13: 7 Roy V, Weisdorf D. Mycobacterial infections following bone 906–910. marrow transplantantion: a 20 year retrospective review. Bone 20 Straus WL, Ostroff SM, Jernigan DB et al. Clinical and epide- Marrow Transplant 1997; 19: 467–470. miologic characteristics of Mycobacterium haemophilum,an 8 Raad II, Vartivarian S, Khan A, Bodey GP. Catheter-related emerging pathogen in immunocompromised patients. Ann infections caused by the Mycobacterium fortuitum complex: Intern Med 1994; 120: 118–125. 15 cases and review. Rev Infect Dis 1991; 13: 1120–1125. 21 Saubolle MA, Kiehn TE, White MH et al. Mycobacterium 9 Flynn PM, Hooser BV, Gigliotti F. Atypical mycobacterial haemophilum: microbiology and expanding clinical and geo- infections of Hickman catheter exit sites. Pediatr Infect Dis J graphic spectra of disease in humans. Clin Microbiol Rev 1988; 7: 510–513. 1996; 9: 435–447. 10 Holland DJ, Chen SC, Chew WW, Gilbert GL. Mycobacter- 22 White MH, Papadopoulos EB, Small TN et al. Mycobacterium ium neoaurum infection of a Hickman catheter in an immuno- haemophilum infections in bone marrow transplant recipients. suppressed patient. Clin Infect Dis 1994; 18: 1002–1003. Transplantation 1995; 60: 957–960.