Mycobacterium chelonae: A cause of nodular skin lesions with a proclivity for renal transplant recipients

Mycobacterium chelonae: A Cause of Nodular Skin Lesions with a Proclivity for Renal Transplant Recipients JOHN F. COOPER, SCHAFFNER,M.D.

M.D.,

MICHAELJ.

LICHTENSTEIN,

M.D.,

M.SC.,

BARNEY

S. GRAHAM,

M.D.,

WILLIAM

Nashvi//e,Tennessee

PURPOSE:Infections due to Mycobacterizun thelonae are uncommon. Several renal transplant recipients at our medical center have developed M. chelonae infections during the past several years, so we decided to review our recent experience with

M. chelonae infections. PATIENTSANDMETHODS: The clinicalmicrobiology laboratory records of four Vanderbilt University Affiliated Hospitals were reviewed. Ten patients with M. chelonae tissue or blood infections were identified between 1982 and July 1988. RESULTS:All infections involved the shin and subcutaneous tissue. Three infections developed at the sites of medical injections. The remaining seven infections occurred in renal transplant recipients and produced a clinically distinctive syndrome. All were indolent tender nodular lesions on the extremities, usually the lower legs. Systemic symptoms were absent, and white blood cell counts were within normal limits. Diagnosis required tissue biopsy and cultures that were incubated for a month. Therapy consisted of surgical excision combined with long-term antibiotics. Even so, some patients had a chronic, relapsing course. CONCLUSION: Although other diagnoses must be considered, the presumptive diagnosis of M. chelonae infection is suggested by the appearance of nodular erythematous lesions on the legs of a renal transplant recipient.

R

apidly growing mycobacteria, first described as agents of human disease in 1938 [l], are being recognized with increasing frequency as a cause of human infection [2-41. Although most of these organisms are saprophytes and do not have a proven association with human disease [2,3], two species, Mycobacterium chelonae and M. fortuitum, have recently been implicated as pathogens in a number of clinical settings [3,51. Over the past several years, it became apparent that several renal transplant recipients at this medical center had developed M. chelonae infections. This prompted us to review our entire recent experience with M. chelonae infections, with an emphasis on the distinctive clinical features occurring in renal transplant recipients.

PATIENTS AND METHODS The clinical microbiology laboratory records of the four Vanderbilt University Affiliated Hospitals (Vanderbilt University Hospital, Nashville Veterans Administration Hospital, St. Thomas Hospital, and Metropolitan Nashville General Hospital) were reviewed. All M. chelonae isolates recovered from tissue or blood specimens between 1982 and July 1988 were included. The medical records of patients with M. chelonae infections were reviewed, including both hospital and clinic charts. Of the 10 patients, seven had been examined by the authors during their illness. All surviving patients were contacted by telephone and questioned about the course of their infection.

RESULTS

From the Divisions of Infectious Diseases and Department of Medicine, Vanderbilt University ville, Tennessee. Requests for reprints should Lichtenstein, M.D., Division of General Internal Vanderbilt Clinic, Vanderbilt University nessee 37232. Manuscript submitted revised form November 28, 1988.

General Internal Medicine, School of Medicine, Nashbe addressed

Medicine,

to Michael J.

Room 2553, The

School of Medicine, Nashville, TenAugust 3. 1988, and accepted in

Ten patients with M. chelonae infection were identified (Table I). Their ages ranged from three to 77 years; six patients were female. All of the infections involved the skin and subcutaneous tissue; one patient was bacteremic. Seven of the ten patients had received renal transplants and were taking immunosuppressive medication. Among the transplant recipients, the average interval between transplantation and onset of infection was 4.5 years (range, two to 11 years). Strikingly, the cutaneous lesions in these seven renal transplant patients were confined entirely to their extremities (Table I). In six patients, the lesions occurred only on the legs, usually below the knee; one patient had lesions on both her arms and legs. Infection developed at the site of previous injections in the remaining three patients. The bloodstream infection was initiated by a cutaneous Broviac catheter exit site infection. An insulin-dependent diabetic patient had numerous lesions that corresponded to sites of prior injections in the posterolateral aspects of both hips and another patient developed lesions in the area February

1989

The American

Journal

of Medicine

Volume

86

173

MYCO/3MTH?/UM

CHELOMEINFECTIONS

TABLE I Clinical Characteristics

Patient

Sex

Age

/ COOPER

of M. chelonae

2

F”

i

26 38

5 6

infections

Renal Transplant

Disease

:

ET AL

University

Prednisone

Azathioprine 50 mg/day 125 mg/day 100 mg/day

MF

ESRD ESRD ESRD ESRD

;(:; Yes

lOmg/day 10 mg/day 20 mg/day 10 mg/day

50

F

ESRD

Yes

10 mg/day

50 mg/day

36

F

ESRD

Yes

20 mg/day

100 mg/day

ESRD Gout

Yes No

10 mg/day -

75 mg/day -

9

59

F

10

3

M

Yes

at the Vanderbilt

Trisomy 21

No

-

No

-

Hospitals (1982-1988)

Affiliated

Site of Infection

Years after Transplant

Right calf Right calf Right calf Left thigh and calf Left shin and ankle Left arm and hand, right leg Right leg Left foot and leg, near site of local steroid injections Insulin injection sites (hips, arms, abdomen) Broviac catheter, blood

Symptomatic interval before Diagnosis (months)

-

3

-

3

SRD = end-stage renal disease. Since most recent transplant.

TABLE ii Antimicrobial

Susceptibilities

Antibiotic

2

Amikacin Ansamycin Capreomycin Cefoxitin Ciprofloxacin Clofazimine Doxycycline Erythromycin lmipenem Kanamycin Minocycline Ofloxacin Sulfamethoxazole Tobramycin

of M. chelonea

3

8”

2 :;

G.25 0.25 2

320.25

4

5A* a

128 128 >32 >32

64 >128 >32 .l -

-4 >32 >512 16

isolates (MIC, pg/mi)

0.5 6:

>32 >%6

-

0.25 8 -

February

1989

The American

of Medicine

16 >128 16 a

-

>32 >32

32

512 1

>512 32

Susceptibilities

:

-

were not performed

881

9

8 -

>16 -

-

128 -

>%6 -

32 0.25

>32

a

Volume

8A*

-a >32

4

Journal

7

-

where she previously had received intra-articular injections of corticosteroids. The patients were remarkably tolerant of their infections. The symptomatic interval before diagnosis ranged from one to 12 months. The cutaneous lesions were characteristically nodular, erythematous, and often had an area of central necrosis. They were not associated with regional lymphadenopathy. No patients had fever or an elevated white blood cell count (only two subjects had blood cultures prior to administration of antibiotics). Three patients described their lesions as tender, but only one had substantial functional disability. Ascertaining the diagnosis of M. chelonae infection was difficult, despite the willingness of physicians to obtain biopsy specimens of the cutaneous lesions. Nine patients required a total of 14 biopsy specimens to establish the diagnosis; four patients required two or more specimens. Acute and chronic inflammation 174

128 >128 >32 0.25

>32 3:

38

by two laboratories.

8

a 1

i.25

MIC = minimal inhibitory concentrations. *Cases 5 and 8 had susceptibilities determined

Case 6

5B*

i16 2 1

10

>%6 4

-4.8 16 16 78 >128

>32

4 1

i:

16 -

,a >128 4

32 16 >128 >32 -

4

-2

on the isolate obtained from Patient 1.

were evident in each specimen, and three specimens contained microabscesses. Granulomas were not observed in any of, the specimens, and only four of the 14 biopsy specimens were acid-fast smear positive. In two patients, the initial Gram stains revealed beaded, gram-positive filamentous organisms that suggested Nocardia to experienced observers. Acid-fast stains demonstrated bacterial filaments that were actually mycobacteria, aligned end-to-end. Cultures subsequently grew mycobacteria. In each patient, the diagnosis of M. cheEonae infection was established by culturing the organism from a biopsy specimen. Nine of the 14 biopsy specimens grew M. chelonae, and in patient 10 M. chelonae was isolated from routine blood cultures (BACTEC radiometric system) after eight days of incubation. Subspeciation was performed on only three isolates (one was subspecies abcessus-Patient 3, two were subspecies chelonaePatients 5 and 9). 86

MYCOBACTERlUM CHELONAEINFECTIONS / COOPER ETAL

TABLEIII Clinical Course of A!. chelonae Infections at the Vanderbilt University Affiliated Hospitals (1982-1988) Patient 1

Monthsof Empiric

Treatment

2

Dicloxacillin, amoxicillin, clavulanate, oxacillin Cephalexin

3

Dicloxacillin

4

Lindane

5

TMP-SMX

6

-

;

-

9

Dicloxacillin, TMP-SMX

10

Cefaclor

Definitive

Treatment

Treatment

Surgical Debridement

Outcome

-

0

No

Died before treatment

Doxycycline, erythromycin Amikacin, rifampin, deoxycycline, erythromytin, cefoxitin, clofazimine Doxycycline

24

No

13

Twice

Fluctuating lesions; continuing antibiotics Cured (24-month followuPI

13

No

Cefoxitin, amikacin, tobramycin, erythromycin Minocycline, erythromycin TMP-SMX Cefoxitin, amikacin, erythromycin, deoxycycline, tobramycin, clofazimine Rifampin, erythromycin, tobramycin Amikacin: cefoxitin, erythromcyrn, IMP-SMX

2

Once

2

No

1 new lesion 8 months after stopping treatment Lesions progressing; still taking antibiotics Cured (17-month followup) Continuing antibiotics Fluctuating lesions; continuing antibiotics

1:

Once Twice

15

Numerous

1

Broviac catheter removed

Cured (3-month followuPI Cured (16-month followuPI

WP-SMX = trimethoprim-sulfamethoxazole

The results of antimicrobial susceptibility studies are summarized in Table II. A wide range of susceptibilities was reported, particularly with erythromycin, cefoxitin, and doxycycline. Two laboratories performed susceptibility determinations on the isolates from Patients 5 and 8, with the reported minimal inhibitory concentration susceptibilities varying by as much as seven dilutions. The patients’ clinical courses tended to be protracted, with both long-term administration of antimicrobials and surgical debridement necessary in patients in whom cures were ultimately achieved (Table III). Patient 1 died of other causes before treatment was initiated. The nine surviving patients received from one to six antibiotics for a duration of one to 24 months. Five patients required surgical debridement of their lesions, three of whom underwent debridement on two or more occasions. As this report was being prepared, cures had been achieved in four patients, and five continued to receive antibiotics.

CASEREPORTS The following two case summaries illustrate characteristic features of the diagnosis, course, and outcome of M. chelonae infections, one in a renal transplant recipient, the other in an elderly patient following injections. Patient 6 A 35-year-old woman was admitted to the hospital with chronic rejection of her fourth renal transplant. She was taking azathioprine 100 mg and prednisone 10 mg daily. Physical examination demonstrated multiple cutaneous nodules on her left arm, left hand, and right leg (Figure 1). The nodules were nontender, indurated, erythematous, and had developed approximately five months previously. Microscopic examination of a punch biopsy specimen of an arm nodule

demonstrated filamentous, gram-positive, acid-fast organisms in a bed of chronic inflammation (Figure 2). The patient was treated with minocycline for presumptive nocardiosis. The lesions substantially regressed and the minocycline was discontinued after six weeks of treatment. The biopsy culture subsequently grew M. chelonae. She was evaluated three months after discharge and a new lesion was discovered on her left forearm. The patient was treated with erythromycin for six additional weeks. During this period, her immunosuppressive agents were tapered and discontinued. The skin lesions resolved and she has remained well during 17 months of follow-up. The patient could recall no inciting event that might have produced the lesions on her extremities. Despite the presence of numerous lesions, they caused her little discomfort or systemic illness and she tolerated them for five months before they came to medical attention. The immunosuppressive regimen was a likely factor in predisposing this patient to M. chelonae infection. Withdrawal of immunosuppressive drugs, concurrent with prolonged antimicrobial therapy, resulted in resolution of the lesions. Patient 7 A 77-year-old woman was admitted with an eightweek history of left ankle pain and swelling. The patient had a history of gout and had received corticosteroid injections into her left Achilles tendon four months and 16 months prior to admission. An ulcerated lesion was present over her left Achilles tendon and multiple 5-mm nodules were on her left shin. The nodules were nontender, erythematous, and fluctuant, with an area of central necrosis. Deep venous thrombosis of the left leg was confirmed by venography. A punch biopsy specimen of one of the skin nodules February

1989

The American

Journal

of Medicine

Volume

86

175

MYCO&?/ICTERlUM

CHELOiVAElNFECTlONS

Figure 1. Nodular lesions of M. chelonae transplant recipient (Patient 6).

I COOPER

infection

ET AL

men was acid-fast smear negative, but culture of this specimen subsequently grew M. chelonae. She was treated in the hospital with amikacin, cefotaxime, and erythromycin for two months and was discharged taking erythromycin and doxycycline. The patient was admitted three more times during the ensuing four months, and despite the addition of trimethoprim-sulfamethoxazole, tobramycin, and clofazimine, her lesions persisted. After seven months of antibiotic treatment, a split-thickness skin graft was performed, but the lesions recurred at this site within two months. The area was debrided twice more, eight and 11 months after her initial diagnosis. The antibiotics were continued for a total of 15 months. She has remained well during the subsequent year. The M. chelonae infection likely was related to antecedent steroid injections, which were administered in the area where infection subsequently developed. Three biopsy specimens were required to make the diagnosis, since neither microscopic examination nor the cultures were consistently positive. The infection followed a protracted, fluctuating course. Three surgical procedures and 15 months of continuous antimicrobial therapy were necessary to eradicate the infection.

on left arm of renal

COMMENTS

Figure nodule

2. Touch preparation from punch biopsy (Patient 6). (Fite stain; original magnification

Figure 3. Ulcerated (Patient 7).

nodular

Although M. chelonae is an uncommon human pathogen, 10 patients at the Vanderbilt University Affiliated Hospitals recently have had infection with this organism. M. chelonae is an environmental mycobacterium and has been isolated from soil, domestic water supplies, and household dust [2,3,6]. Infection usually develops as a consequence of antecedent penetrating injury and often is associated with impaired cell-mediated immunity [2,3,6,7]. Despite its predilection for immunocompromised patients, disseminated infection with M. chelonae is rare. Two factors were identified in our patients that were associated with the subsequent development of infection. Each patient was either a renal transplant recipient or had developed infection at the site of an antecedent penetrating injury. Seven of the 10 patients were renal transplant recipients and each was taking immunosuppressive medication that results in the functional impairment of cell-mediated immunity [3,8]. Cell-mediated immunity is recognized as fundamental to host resistance against mycobacterial infection [9,10]. The absence of M. chelonae infections in recipients of other organ transplants, however, suggests that additional factors, as yet unknown, may predispose renal transplant recipients to infection with this pathogen. During the same time period, there have been nine isolates of M. fortuitum at Vanderbilt. Two of these were recovered from the peritoneal fluid of patients with chronic renal failure. There were no M. fortuitum isolates from skin biopsy specimens, and no M. fortuitum infections among renal transplant recipients. Each of the transplant recipients had multiple erythematous, subcutaneous nodules. These were located exclusively on the extremities, and most lesions were limited to the legs. This predilection for lower extremity involvement has been observed previously but remains unexplained [8]. None of the reported cases of M. chelonae infection involving renal transplant recipients had a recognized source of infection [3]. We reviewed the transplantation procedures in our insti-

specimen of arm X 1,000)

lesion of M. chelonaeinfection

on left shin

yielded negative stains for fungus, malignancy, and acid-fast bacilli; the cultures yielded no growth. The tissue specimen demonstrated microscopic evidence of acute inflammation. She was treated empirically with cephalexin for two weeks. One month later, she was readmitted for evaluation of a 3 X 2-cm necrotic, indurated, left pretibial nodule (Figure 3). A second biopsy specimen revealed numerous acid-fast organisms by microscopy, but the cultures again yielded no growth. A third biopsy speci176

February

1989

The American

Journal

of Medicine

Volume

86

MYCOtMCTERlUM

tution but were unable to identify any lower extremity trauma that might have served as the inciting event. Because postoperative wound infections due to rapidly growing mycobacteria typically develop within three months, the length of time between transplantation and onset of infection in the Vanderbilt patients speaks against a perioperative point source of infection [3,11]. Likewise, the patients lived at considerable distances from each other and had no shared experience that might have suggested either a common source or an entry site for infection. Thus, the clinical phenomenon remains enigmatic but is sufficiently distinctive in renal transplant patients to prompt rapid diagnostic efforts and empiric therapy. Two patients developed erythematous subcutaneous nodules at the sites of previous insulin or corticosteroid injections. The relationship between localized cutaneous infection and penetrating injury is well established. In another series, 31 of 34 cases of localized cutaneous infection with rapid growing mycobacteria developed following penetrating injuries [3]. A variety of types of injuries have been reported, including vaccinations, injections with home-sterilized needles, gunshot wounds, surgery, and farm machinery accidents [2,3,11-131. Each type of injury involved either direct injection of infected material or subsequent contamination of the wound with soil or water. Patient 9 developed infection as a consequence of reusing insulin syringes without adequate disinfection at home. One patient had mycobacteremia consequent to an indwelling central venous catheter. M. chelonae is not fastidious and grows well in standard aerobic BACTEC culture media, such that 80 percent of blood cultures of bacteremic patients are positive within five to eight days [3]. Our patient had purulent drainage from his catheter insertion site, which was erythematous and indurated [14]. He did not develop stigmata of endovascular infection or disseminated disease, although both of these complications of bacillemia have been reported in other series [3]. Cutaneous M. chelonae infections are characteristically indolent, and months often elapse before the diagnosis is confirmed [3,6]. Among our patients, the interval between onset of infection and diagnosis was as long as one year. None of the patients was either systemically ill or had an elevated white blood cell count. Establishing the diagnosis of M. chelonae infection was difficult. Multiple biopsy specimens were often necessary to confirm the diagnosis, as illustrated by four patients who required two or more specimens. M. chelonae will grow on routine laboratory media, such as trypticase soy agar, chocolate agar, or MuellerHinton broth. Good results are obtained when biopsy specimens are cultured on Lowenstein-Jensen agar, which is routinely incubated for a prolonged period. A plausible explanation for the difficulty in recovering M. chelonae is that most clinical microbacterial cultures were incubated at 35OC. At this temperature, M. chelonae may grow in four to six weeks or not grow at all 1151.If incubated at 28” to 32’C, M. chelonae will usually grow in seven days [15]. The optimal treatment of cutaneous mycobacterial infection has not been clearly defined. With the exception of the aminoglycosides, M. chelonae is resistant to antituberculous agents [16]. Susceptibility to other antimicrobials varies widely [16,17]. More disturbing, however, were the differences in results of susceptibili-

C/fF~ONAElNFECTlONS

/ COOPER

ET AL

ty testing of identical isolates by different reference laboratories. Susceptibility testing of rapidly growing mycobacteria is difficult and has not been standardized. Nevertheless, although results should be interpreted with caution, susceptibility data may still be used to guide selection of antimicrobials. A recent review of M. chelonae infections suggested that cefoxitin and amikacin should be administered as empiric therapy while susceptibility testing is being performed [3]. Another author has suggested that cefoxitin alone may be efficacious in treating cutaneous M. chelonae infection arising in a renal transplant recipient [18]. Our isolates were uniformly susceptible to amikacin, but six of nine isolates had minimal inhibitory concentrations for cefoxitin that exceeded 128 rglml. The data from the current series suggest that the combination of erythromycin and amikacin would be suitable as empiric therapy for cutaneous it4. chelonae infection. Although other mycobacteria, Nocardia species, and fungi must be considered, the clinical syndrome of nodular erythematous lesions on the legs of a renal transplant recipient is sufficiently distinctive that it should suggest the presumptive diagnosis of M. chelonae infection. Establishing the diagnosis requires culturing tissue biopsy specimens by a laboratory that has been alerted that rapidly growing mycobacterial infection is suspected so that the isolate may be incubated at an appropriate temperature. The patients are likely to experience a chronic, relapsing course; surgical debridement is often necessary and antimicrobial therapy must be prolonged. REFERENCES 1. Da Costa Cruz J: Mycobacferium forfuitum, urn nova bacilo acido-resistente pathogenic0 para o homen. Acta Med Rio de Janeiro 1938; 1: 297. 2. Hanson PJV, Thomas JM, Collins JV: Mycobacterium chelonae and abscess formation in soft tissues. Tubercle 1987; 68: 297-299. 3. Wallace RJ. Swenson JM. SilcoxVA, Good RC. Tschen JA. Stone MS: Soectrum of disease due to rapidly growing mycobacteria. Rev Infect Dis 1983; 5: 657-679. 4. Woods GL. Washington JA: Mycobacteria other than Mycobacterium tubercule sis:review of microbiologic and clinical aspects. Rev Infect Dis 1987: 9: 275-294. 5. Canilang B, Armstrong D: Subcutaneous abscesses due to Mycobacferium fortuifum. Am Rev Resoir Dis 1968: 97: 451-454. 6. Wallace RJ: Nontuberculous mycobacteria and water: a love affair with increasing clinical importance. Infect Dis Clin North Am 1987; 3: 677-686. 7. Lloveras J, Peterson PK, Simmons RL, Naiarian JS: Mycobacterial infections in renal transplant recipients: seven cases and a review of the literature. Arch Intern Med 1982; 142: 888-892. 8. Tice AD, Solomon RJ: Disseminated Mycobacferium cbelonae infection: response to sulfonamides. Am Rev Respir Dis 1979; 120: 197-201. 9. Goldblum SE. Reed WP: Host defenses and immunologic alterations associated with chronic hemodialvsis. Annlntern Med 1980: 93: 597-613. 10. Griffith DE: Environmental mycobacteria: an increasing problem. Hosp Pratt 1988: 15: 125-149. 11. Gremillion DH, Mursch SB. Lerner CJ: Injection site abscesses caused by Mycobacferium chelonae. Infect Control 1983; 4: 25-28. 12. Safranek TJ, Jarvis WR, Carson LA, et al: Mycobacferium chelonae wound infections after plastic surgery employing contaminated gentian violet marking solution. N Ennl J Med 1987: 317: 197-201. 13. Jackson PG, Keen H, Noble CJ, Simmons NA: Injection abscesses in a diabetic due to Mycobacterium chelonae var abscessus. Br Med J 1980; 281: 1105-l 106. 14. Flynn PM, Van Hooser B, Gigliotti F: Atypical mycobacterial infections of Hickman catheter exit sites. Pediatr infect Dis J 1988; 7: 510-513. 15. Silcox VA, Good RC, Floyd MM: Identification of clinically significant M. forfui turn complex isolates. J Clin Microbial 19&Q: 14: 686-691. 16. Swenson JM, Wallace RJ, SilcoxVA, Thornsberry C: Antimicrobial susceptibility of five subgroups of Mycobacferium fort&urn and Mycobacferium chelonae. Antimicrob Agents Chemother 1985: 28: 807-811. 17. Wallace RJ, Swenson JM, Silcox VA, Bullen MG: Treatment of nonpulmonary infections due to Mycobacterium forfuitum and Mycobacferium chelonae on the basis of in vitro SusceDtibitities. J Infect Dis 1985: 152: 500-514. 18. Mehta R. Oliver Ld, Melillo D, Milliorn K, Flye,W, Fish J: Disseminated Afyco bacterium cbelonae infection following cadaveric renal transplantation: favorable response to cefoxitin. Am J Kidney Dis 1983; 3: 124-128.

February

1989

The American

Journal

of Medicine

Volume

86

177