Soft Tissue Infection Caused by a Novel Pigmented, Rapidly Growing Mycobacterium Species

Abstract

Molecular techniques are playing an important role in the diagnosis of nontuberculous mycobacterial infections. This case report describes a chronic soft tissue infection in an immunocompetent patient caused by a previously undescribed pigmented, rapidly growing Mycobacterium species, emphasizing the importance of clinical suspicion and effective laboratory techniques in the diagnosis and treatment of infection.

CASE REPORT

A previously healthy 39-year-old male presented in May 2001 to emergency with an isolated soft tissue injury to his left ankle. He had fallen off his motorcycle, and the foot pedal penetrated the skin postero-inferiorly to his left medial malleolus, leaving a wound contaminated with dirt, grass, and gravel. Stress testing of the ankle as well as X-rays excluded any ligamentous, bony, or articular abnormalities. The wound was irrigated, and he received 1 g of ceftizoxime intravenously, followed by 1 week of oral cephalexin four times a day.

Ten days later, he presented to the emergency room with increasing erythema and sanguineous discharge from the wound, but without systemic complaints. A wound culture grew a member of the Enterobacter agglomerans group that was sensitive to cefoxitin, cefazolin, cotrimoxazole, ampicillin, and gentamicin. Over the next month, he received a 7-day course of cloxacillin four times a day and a 4-day course of 1 g of ceftizoxime every 12 h parenterally, followed by a 2-week course of cephalexin four times a day. Despite this antimicrobial therapy, local swelling, clear drainage, and pain persisted. A repeat ankle X-ray and a bone scan performed 2 weeks later revealed soft tissue swelling, normal joint spaces, and no evidence of osteomyelitis. Laboratory investigations included a white blood cell count of 3.8 × 10⁹/liter and an erythrocyte sedimentation rate of 4 mm/h.

The patient was subsequently referred for assessment by an infectious diseases consultant with complaints of persistent serosanguineous drainage, minimal improvement in swelling, and some night sweats. There was pain in the ankle with walking. A 2-cm by 2-cm crusted ulcer with a rim of erythema and copious clear discharge was present posterior to his medial malleolus. The foot was neurovascularly intact, with restricted dorsiflexion because of pain and medial joint line synovial thickening. Swabs for aerobes, anaerobes, fungi, and acid-fast bacilli were obtained from the wound ulcer, and he was maintained on cephalexin four times a day. Specimen processing for mycobacteria included monitoring for growth in liquid medium with the MB/BacT Alert 3D continuous monitoring system (bioMérieux, Inc., St, Louis, Mo.) at 37°C, plus inoculation of Lowenstein-Jensen medium for incubation at 31°C and 37°C. The initial Auramine O fluorescent stain for acid-fast bacilli was negative.

After 1 week, the wound was smaller and nonerythematous but still draining clear, non-foul-smelling discharge, and the ankle remained painful. He remained on cephalexin and was instructed to return to the clinic in 2 weeks. Bone scan results revealed moderately increased uptake in the superior half of the left calcaneus, which was worrisome for osteomyelitis. In addition, there was increased uptake at the talonavicular joint, the first metatarsal base, and the second to fifth metatarsophalangeal joints. Initial cultures grew coagulase-negative staphylococci at 1+ (on a 0 to 4+ scale), and there was no evidence of fungal growth.

One month later, the laboratory noted the presence of a pigmented Mycobacterium sp. on Lowenstein-Jensen medium at 31°C only. AccuProbe assays (Gen-Probe, Inc., San Diego, Calif.) for the most common pigmented mycobacterial species, i.e., M. gordonae and M. kansasii, were performed and were negative. The organism, designated 01-154, was submitted for 16S rRNA gene sequence analysis for rapid identification. Concurrently, the susceptibility profile of the organism was determined with Etest (AB Biodisk, Solna, Sweden) and showed in vitro susceptibility to cefoxitin (MIC of <0.016 μg/ml), amikacin (0.023 μg/ml), ciprofloxacin (<0.002 μg/ml), clarithromycin (<0.016 μg/ml), sulfamethoxazole (0.032 μg/ml), tobramycin (4 μg/ml), and doxycycline (<0.016 μg/ml) and resistance to imipenem (>32 μg/ml).

Cephalexin therapy was then discontinued, and the patient was prescribed clarithromycin twice a day and ethambutol three times a day, with follow-up in 1 month. There was no clinical improvement, and when he was reassessed 6 weeks later, ethambutol was discontinued and doxycycline twice a day was initiated. Over the next 2 weeks, he improved markedly, with resolution of pain and swelling and no further drainage. Doxycycline was discontinued after 4 weeks of therapy because of gastrointestinal intolerance. Clarithromycin was discontinued at the same time. He has remained well since discontinuing antibiotic therapy.

Sequencing of the nearly complete 16S rRNA gene of strain 01-154 was performed as previously described (12) and revealed a unique organism distantly related to the Mycobacterium smegmatis group. The sequence was compared against those of reference strains determined in-house and against two public sequence repositories, NCBI (GenBank) and Ribosomal Differentiation of Medical Microorganisms (RIDOM) (7). Sequence distances determined with ClustalW from Escherichia coli bp 54 to 510, which encompasses the hypervariable regions of the gene, indicated the highest identity with M. madagascariense ATCC 49865^T (97.4% identity; 13 bp variations), followed by M. smegmatis ATCC 19420^T, M. goodii ATCC 700504^T, and M. confluentis ATCC 49920^T, all with an identity value of 97.0% (Fig. 1).

FIG. 1.

Phylogenetic relationships of “M. manitobense” strain 01-154^T with its closest relatives based on the partial (bp 54 to 510 of E. coli) 16S rRNA gene sequence. Multiple sequence alignments were determined with ClustalW in the Megalign component of Lasergene version 5.03 (DNAStar, Inc., Madison, Wis.). The tree was rooted with the slow-growing species Mycobacterium tuberculosis as the out-group sequence. Sequences were determined in our laboratory unless indicated by a GenBank accession number.

PCR-restriction enzyme analysis was performed as previously described (10), with amplification of a 441-bp fragment of the hsp65 gene followed by BstEII and HaeIII restriction enzyme digestion. The resulting fragments were 205 and 235 bp in size with BstEII and 140, 85, and 60 bp in size with HaeIII, identical to the pattern in M. obuense and M. chubuense (2, 5). However, these two species are very distant by 16S ribosomal DNA analysis in comparison to “M. manitobense,” as shown by low identity values: 93.5% identity or 27 bp variations for M. chubuense and 93.7% identity or 26 bp variations for M. obuense, comparing only the bp 54 to 510 region of the 16S rRNA gene.

Mycolic acid analysis by high-pressure liquid chromatography (HPLC) was performed according to the standard method (3). The HPLC profile consisted of a widely separated double-peak cluster of prominent early peaks followed by late peaks (Fig. 2), resembling those of M. neoaurum, M. mucogenicum, M. rhodesiae, M. sphagni, and patterns designated as M. terrae/M. nonchromogenicum group (http://hplcusersgroup.tripod.com/) (4).

FIG. 2.

Mycolic acid pattern of “M. manitobense” strain 01-154^T, obtained by HPLC analysis. The relative retention time is indicated for each peak. HMW, high-molecular-weight internal standard.

Sequencing of intergenic spacer region ITS1 directly from the PCR product was performed as described previously (11) and was successful, unlike the results for most rapidly growing nontuberculous mycobacterial species, which generally have two rRNA operons. Blast analysis did not result in successful pairwise alignments of the complete ITS1 sequence with GenBank submissions due to the high sequence variability of ITS1 between them. Analysis with RIDOM indicated the highest identity was with M. murale, M. tokaiense, M. parafortuitum, and M. peregrinum, albeit with very low identity scores (<70%). The sequence of the 16S rRNA gene and the ITS1 region was submitted to GenBank and given accession number AY082001.

The organism grew best at 25°C and 31°C, followed by 37°C, and not at all at 42°C and showed a bright orange pigment in both light and dark conditions. Biochemical test results showed a negative 3-day arylsulfatase test and no growth on MacConkey agar without crystal violet, differentiating this rapid grower from the M. fortuitum complex. Test results were negative for nitrate reductase, semiquantitative catalase, beta-glucosidase, tellurite reduction, tolerance to 5% NaCl, and niacin. The organism did not utilize sodium citrate as a sole carbon source. Results were positive for urease, iron uptake, acid phosphatase, Tween 80 hydrolysis, and pyrazinamidase. Acid was produced from fructose, mannitol, inositol, and sorbitol. The biochemical profile did not correspond to any well-established species.

Environmentally derived mycobacteria, encompassing most species in the genus Mycobacterium (8), may be considered opportunistic pathogens and are often associated with disease in both immunocompetent and immunocompromised patients. Infections with rapidly growing mycobacteria are almost exclusively caused by members of the M. fortuitum complex (M. fortuitum and its biovariants, M. peregrinum, M. abscessus, and M. chelonae) (6, 15), while other rapidly growing species are normally not considered clinically significant (15). However, with the help of sequence-based identification or mycolic acid analyses, M. neoaurum, a pigmented, rapidly growing mycobacterium first isolated from soil, has been identified as a causative agent of catheter-related bacteremia and peritonitis (9, 16).

Cutaneous infections account for a majority of cases caused by rapidly growing mycobacteria, often as a result of surgical procedures or accidental penetrating trauma (15). The latter usually occurs several weeks earlier, with possible water or soil contamination, revealing cellulitis, abscess formation, and drainage of small amounts of watery fluid (15). Our patient showed evidence of infection consistent with this, with a history of dirty, traumatic, percutaneous chronic injury that included localized pain, swelling, and drainage. This possibility is often overlooked by physicians during initial examination, delaying proper diagnosis. The initial course of antimicrobials did not resolve the infection, and consequently, mycobacteriology cultures were only included upon further evaluation.

The American Thoracic Society guidelines advocate susceptibility testing for all clinically significant isolates due to the differences in susceptibilities among and within a species (13). Unfortunately, only select species have a well-established susceptibility profile, and antituberculosis first-line drugs, i.e., isoniazid, rifampin, pyrazinamide, and ethambutol, have no effect on rapid growers with the exception of ethambutol, to which all M. smegmatis isolates are susceptible (13, 14). The initial ethambutol and clarithromycin combination therapy did not aid in resolving the infection despite in vitro susceptibility to clarithromycin. The paucity of clinical studies to determine treatment outcome of nontuberculous mycobacteria has made it difficult to correlate in vitro antimicrobial sensitivity results with in vivo efficacy, even with well-known species (1, 17). Often, susceptibility testing can only serve as a guide to predict treatment success. In our case, the clarithromycin and doxycycline combination was clinically effective.

The introduction of more advanced molecular diagnostic methods has dramatically improved the ability to identify less common species that may occasionally cause disease. We have reported a case of chronic soft tissue infection caused by a previously undescribed pigmented, rapidly growing Mycobacterium species as a result of the appreciation of the clinical picture. Any mycobacterial species may in fact be clinically significant in some cases, with less common species often being misidentified due to a lack of adequate methods in some laboratories. Correct identification of clinically relevant mycobacteria is important for proper antimicrobial treatment and the establishment of a detailed taxonomic system. We propose the name “M. manitobense” for this particular species, in reference to the Canadian province in which it was isolated. Strain 01-154 has been designated the type strain for this species (= ATCC BAA-545^T = DSM 44615^T).