Desulfovibrio legallii Prosthetic Shoulder Joint Infection and Review of Antimicrobial Susceptibility and Clinical Characteristics of Desulfovibrio Infections

Abstract

We describe a case of shoulder hemiarthroplasty infection with Desulfovibrio legallii. Antimicrobial susceptibilities of 36 Desulfovibrio isolates are presented. Metronidazole and carbapenems exhibited reliable activity, although piperacillin-tazobactam did not. Eleven previous cases of Desulfovibrio infection are reviewed; most arose from a gastrointestinal tract-related source.

CASE REPORT

In July 2013, a 70-year-old woman with a 10-year history of palmoplantar psoriasis presented to our institution with a 6-month history of intermittent fever, chills, and left-shoulder pain. This had worsened over the preceding 2 weeks, and she had developed a diminished range of movement of her left shoulder. She had sustained a closed proximal humerus fracture which had been treated with a humeral endoprosthesis implanted 18 years prior to presentation. For about 8 years through August 2012, she had received treatment for psoriasis with twice-yearly intramuscular injections of 80 mg triamcinolone. In August 2012, she was placed on topical treatment and phototherapy. In attempt to improve her psoriasis, the patient reported soaking her hands and feet, which had many fissures and cracks, in Lake Superior and in the Pacific Ocean. Physical examination revealed a well-healed surgical scar over her left shoulder and a scant amount of swelling in the lower axilla. She had an elevated white cell count of 16 × 10⁹/liter, an erythrocyte sedimentation rate of 35 mm/h (normal, 0 to 29 mm/h), and a C-reactive protein level of 204.3 mg/liter (normal, ≤8 mg/liter). An indium¹¹¹ bone scan showed asymmetric uptake around her left shoulder suggestive of periprosthetic infection. The cell count in the aspirated shoulder joint fluid was 25,075 white blood cells/μl, with 99% neutrophils. Cultures of the aspirated joint fluid yielded moderate growth of small, pinpoint, clear colonies on CDC anaerobic sheep blood agar after 6 days of anaerobic incubation at 35°C (Fig. 1A); broth cultures were negative. The colonies measured about 1 mm in diameter and were gray-green and convex after 10 days of incubation (Fig. 1B). Gram stain showed spiral-shaped Gram-negative bacilli (Fig. 1C) which were motile, indole negative, and catalase positive with 15% H₂O₂. PrepMan Ultra (Life Technologies, Grand Island, NY) was used to prepare DNA for partial 16S rRNA gene PCR using primers 5′-TGGAGAGTTTGATCCTGGCTCAG-3′ and 5′-TACCGCGGCTGCTGGCAC-3′ with bidirectional amplification product sequencing performed using the same primers. A 496-bp sequence was generated, revealing 100% identity to Desulfovibrio legallii (strain H₁) (base pairs 4 to 499 of GenBank accession number FJ225426) (1, 2). Per our laboratory protocol, for species identification, a ≥99% agreement with a database strain with >0.8% separation from other species is required. The next-best matches were Desulfovibrio sp. canine oral taxon 070 clone 1P043 (GenBank accession number JQ295257) (96% match) and Desulfovibrio sp. canine oral taxon 070 clone 1J008 (GenBank accession number JQ294906) (95% match). Drawing from observations that growth of sulfate-reducing organisms is enhanced by the addition of magnesium sulfate to broth media (3), we also noted that the colony robustness on CDC anaerobic sheep blood agar was enhanced with supplementation with 250 μl of 10% magnesium sulfate. This isolate did not yield a genus- or species-level identification when subjected to matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS) using a Bruker MALDI Biotyper (Bruker Daltonics, Billerica, MA), with spectra analyzed using Bruker Biotyper 3.0 software and library version 3.3.1.0 (4,613 entries); the top match was Clostridium novyi (score, 1.457). The library used contained only a single Desulfovibrio entry, for Desulfovibrio fairfieldensis.

FIG 1.

Colony morphology, Gram stain appearance, and intraoperative findings in the reported case of Desulfovibrio legallii infection. (A) Desulfovibrio legallii colonies on anaerobic sheep blood agar plate at 6 days. (B) D. legallii colonies on anaerobic sheep blood agar plate at 10 days. (C) Gram stain of D. legallii with arrow showing spiral appearance. (D) Intraoperative findings showing yellow purulent fluid on exposure of shoulder joint.

The isolate was β-lactamase negative (cefinase disc; Becton, Dickinson, Franklin Lakes, NJ). Susceptibility testing by Etest methodology (AB Biodisk, Sweden) showed a MIC of 1 μg/ml (intermediate) for penicillin and a MIC of ≤0.5 μg/ml (susceptible) for both clindamycin and metronidazole (Table 1). The patient underwent resection shoulder arthroplasty. Purulent fluid was encountered intraoperatively (Fig. 1D). Histopathology of periprosthetic soft tissue revealed fibrovascular tissue with marked acute and chronic inflammation; operative cultures yielded D. legallii from three synovial-fluid and five tissue specimens as well as prosthetic-joint culture using vortex and sonication procedures (>100 CFU/10 ml sonicate fluid) (4, 5). The isolate obtained from sonication cultures had a susceptibility pattern identical to that of the preoperative isolate cultured from synovial fluid, except for the penicillin MIC, which was 2 μg/ml. Additional testing showed that it had the following MICs by Etest: ertapenem, 1 μg/ml; ceftriaxone, 2 μg/ml; amoxicillin-clavulanate, 0.25/0.125 μg/ml; piperacillin/tazobactam, >256/4 μg/ml; minocycline, 0.5 μg/ml; ciprofloxacin, >32 μg/ml. The patient underwent debridement 2 weeks after her initial surgery because of persistent drainage from the wound; no deep infection was noted. Cultures from this débridement yielded single colonies of Staphylococcus capitis and Candida glabrata from one of four tissues, results which were considered indicative of contamination. She was treated with 6 weeks of ertapenem followed by reimplantation with a reverse total shoulder arthroplasty, 7 weeks after the resection surgery. At 6 months postreimplantation, the patient remains well.

TABLE 1.

Desulfovibrio isolates evaluated in the Mayo Clinic Clinical Microbiology Laboratory, 1997 to 2013

Species [total no. of isolates from 1997 to 2013] and antimicrobial (no. tested)a	MIC (μg/ml)			No. (%) of susceptible isolates testedd
	Range	MIC50	MIC90
Desulfovibrio spp. [18]b
Clindamycin (8)	≤0.5 to 1	≤0.5	≤0.5	8 (100)
Penicillin (9)	≤0.5 to 32	8	8	2 (22)
Ampicillin-sulbactam (3)	≤0.5 to 0.5			3(100)
Piperacillin-tazobactam (1)	>256/4			0 (0)
Ertapenem (1)	≤0.5			1 (100)
Imipenem (3)	≤0.5 to 0.5			3 (100)
Metronidazole (9)	≤0.5	≤0.5	≤0.5	9 (100)
D. desulfuricans [40]c
Clindamycin (21)	≤0.5 to >256	≤0.5	1	20 (95)
Penicillin (22)	≤0.5 to 16	4	16	3 (14)
Piperacillin-tazobactam (15)	32/4 to >256/4	128/4	>256/4	2 (13)
Ertapenem (15)	≤0.5 to 1	≤0.5	0.5	15 (100)
Imipenem (2)	≤0.5 to 1			2 (100)
Metronidazole (23)	≤0.5	≤0.5	≤0.5	23 (100)
D. fairfieldensis [2]
D. intestinalis [2]
Clindamycin (1)	≤0.5			1 (100)
Penicillin (1)	2			0 (0)
Piperacillin-tazobactam (1)	32/4			1 (100)
Ertapenem (1)	≤0.5			1 (100)
Metronidazole (1)	≤0.5			1 (100)
D. legallii [2]
Clindamycin (2)	≤0.5			2 (100)
Penicillin (2)	1 to 2			0 (0)
Amoxicillin-clavunalate (1)	0.25/0.125			1 (100)
Piperacillin-tazobactam (1)	>256/4			0 (0)
Ceftriaxone (1)	2			1 (100)
Ertapenem (1)	1			1 (100)
Ciprofloxacin (1)	>32
Minocycline (1)	0.5
Metronidazole (2)	≤0.5			2 (100)
D. piger [3]
Clindamycin (1)	≤0.5			1 (100)
Penicillin (1)	8			0 (0)
Metronidazole (1)	≤0.5			1 (100)

^aNot all isolates were tested for antimicrobial susceptibility. The numbers in square brackets indicate the numbers of isolates isolated in our laboratory or submitted to our laboratory for identification; the numbers in round brackets indicate the numbers of isolates which underwent antimicrobial susceptibility testing. All isolates were from unique patients, with the exception of the two D. legallii isolates, which came from the same patient (as presented in this report). Susceptibility testing was performed via the Etest methodology (AB Biodisk). Sources of isolates: for the 18 Desulfovibrio isolates, blood (17), abdomen (1); for the 40 D. desulfuricans isolates, blood (35), abdomen (3), spine (1), urine (1); for the 2 D. fairfieldensis isolates, blood (2); for the 2 D. intestinalis isolates, blood (2); for the 2 D. legallii isolates (from the same patient, as reported in this paper), synovial fluid (1), prosthetic joint (1); for the 3 D. piger isolates, blood (2), abdomen (1).

^bThere was no species identification for these 18 isolates as they did not meet criteria for species identification by 16S rRNA gene sequencing.

^cOf 35 isolates tested for β-lactamase production (8 Desulfovibrio spp., 21 D. desulfuricans, 1 D. intestinalis, 2 D. legallii, 3 D. piger), only 1 D. desulfuricans isolate tested positive by the cefinase disk test.

^dInterpretive criteria per CLSI document M100-S24 (susceptible = S, intermediate = I, resistant = R; numbers refer to MIC in μg/ml): for amoxicillin/clavunalate, S, ≤4/2; I, 8/4; R, ≥16/8; for ceftriaxone, S, ≤16; I, 32; R, ≥64; for clindamycin, S, ≤2; I, 4; R, ≥8); for ertapenem, S, ≤4; I, 8; R, ≥16; for imipenem, S, ≤4; I, 8; R, ≥16; for metronidazole, S, ≤8; I, 16; R, ≥32; for penicillin, S, ≤0.5; I, 1; R, >2; for piperacillin-tazobactam, S, ≤32/4; I, 64/4; R, ≥128/4. There are no CLSI interpretive criteria for ciprofloxacin or minocycline.

Desulfovibrio species are non-spore-forming, nonfermentative, sulfate-reducing, anaerobic Gram-negative bacilli. Most are motile with polar flagella and are curved or spiral shaped. The type species, D. desulfuricans (formerly Spirillum desulfuricans) was described by the Dutch microbiologist Beijerinck in 1895 (6). Since then, 67 species (not including subspecies) have been described (7). They are found in anaerobic niches, including marine water, brackish water, freshwater, and wastewater and are important ecologically for both their harmful qualities (e.g., corrosion of steel) and their beneficial qualities (e.g., as bioremediators) resulting from their unique metabolic pathways (8). Desulfovibrio spp. are part of the oral and gastrointestinal flora of humans and animals (9, 10) and the vaginal flora of humans (11). A role for Desulfovibrio spp. in inflammatory bowel disease (12) and acute periodontitis (13) has been suggested; however, reported cases of human infection have been infrequent. This may be due to underrecognition of these organisms, which are slow growing; they are often involved in polymicrobial infections, and some clinical laboratories may lack the ability to identify them. Most infections have arisen from an abdominal source, although a case of polymicrobial brain abscess has been reported which was ascribed to an odontogenic or sinus source (14).

(This study was presented in part at the 114th General Meeting of the American Society for Microbiology, Boston, MA, 17 to 20 May 2014).

Since the last review of human Desulfovibrio infections from 1977 to 2003 (15), five additional cases of human Desulfovibrio infection have been reported, including three cases of D. fairfieldensis infection (16–18) and two of D. desulfuricans bacteremia (19, 20). All were associated with a gastrointestinal source (one case each of gastroenteritis [17], sigmoid diverticulitis [18], and cytomegalovirus colitis [20]); one case arose following endoscopic retrograde cholangiopancreatography for choledocholithiasis (16) and another from an infected sacral decubitus ulcer in a bedridden patient (19). Warren et al. also described the phenotypic characteristics of 17 human Desulfovibrio isolates (21). These included 15 isolates from intra-abdominal sources (2 D. piger, 9 D. fairfieldensis, 2 D. desulfuricans, and 2 D. vulgaris) and 2 from blood cultures (1 each of D. fairfieldensis and D. desulfuricans).

We report an unusual case of a prosthetic-joint infection with a novel Desulfovibrio species, D. legallii, and describe human Desulfovibrio isolates identified and subjected to antimicrobial susceptibility testing (AST) in our clinical microbiology laboratory. This study was approved by our Institutional Review Board (IRB 13-007249).

From 1997 to 2013, there were 13 Desulfovibrio isolates collected from Mayo Clinic patients, inclusive of two isolates from the case reported here (Table 2), and 54 isolates referred to us for identification and/or AST, which was performed using the Etest, on Brucella blood agar. Isolates were all identified by 16S rRNA gene sequencing. AST was performed on 10 of the Mayo Clinic isolates and 26 of the referred isolates (Table 1).

TABLE 2.

Twelve cases of Desulfovibrio infection from 1997 to 2013, Mayo Clinic, Rochester, Minnesota

Case	Yr	Genus/species	Source	Antimicrobial susceptibilitya		Coisolated organism(s)b	Age (yrs)/sexc	Diagnosis	Outcome
				Susceptible	Intermediate/Resistant
1	2013 (this report)	Desulfovibrio legallii	Synovial fluid, prosthetic joint, and periprosthetic tissue	CLI, MTZ, ETP, AMC, CRO	PCN, PTZ		70/F	Left-shoulder prosthetic-joint infection	Recovered
2	2012	D. desulfuricans	Urine from percutaneous nephrostomy	CLI, MTZ, PCN		Anaerobic Gram-positive bacilli, anaerobic Gram-positive cocci, Streptococcus agalactiae, Actinobaculum schaalii, Propionimicrobium spp.	66/F	Bilateral hydronephrosis with right emphysematous pyelitis, suspected coloureteral/vesical fistula, herpes encephalitis	Died, secondary to herpes encephalitis
3	2012	D. desulfuricans	Blood (positive at 76 h)	MTZ	PCN		76/M	Bacteremia presumed secondary to diverticulitis	Recovered
4	2011	D. desulfuricans	Spine tissue	CLI, MTZ, PCN		Mobiluncus curtisii, Candida albicans, Clostridium clostridioforme	60/M	Colonic rupture with colonic-dural fistula, Metastatic renal cell carcinoma; cervical, thoracic and lumbar spinal instrumentation, for bone metastases	Died
5	2008	D. piger	Blood (positive at 44 h)	CLI, MTZ	PCN	Ergerthella lenta, Bacteroides ovatus	73/F	Bacteremia secondary to pericolonic abscess, Sigmoid diverticulitis	Recovered
6	2005	D. desulfuricans	Blood (positive at 3 days)	CLI, MTZ PTZ, ETP	PCN		74/F	Bacteremia secondary to presumed gastrointestinal translocation; transient postoperative small-bowel obstruction complicating transverse colectomy for colon cancer	Recovered
7	2005	D. desulfuricans	Blood (positive at 4 days)	CLI, MTZ AMS, ETP	PCN	In peritoneal fluid: Escherichia coli, Klebsiella pneumoniae, anaerobic Gram-negative and -positive rods	57/M	Perforated acute appendicitis, HIV infection	Recovered
8	2004	Desulfovibrio sp.	Blood (positive at 3 days)	CLI, MTZ	PCN	E. lenta, anaerobic Gram-negative rod, Bacteroides fragilis	60/M	Perforated acute appendicitis	Recovered
9	2002	Desulfovibrio sp.	Blood (positive at 5 days)			Candida parapsilosis	74/M	Septic shock, complicated intra-abdominal infection post-aortic-aneurysm repair	Died
10	2000	Desulfovibrio sp.	Blood (positive at 5 days)			In subphrenic abscess: vancomycin-resistant enterococci	45/M	Subphrenic abscess/right empyema, complicated intra-abdominal infection following surgery for diverticulitis	Recovered
11	1999	Desulfovibrio sp.	Blood (positive at 6 days)	CLI, MTZ	PCN		93/F	Perforated sigmoid diverticulitis	Died
12	1997	D. piger	Peritoneal fluid			E. coli, Enterococcus sp., anaerobic Gram-negative rod	63/M	Perforated acute appendicitis	Recovered

^aCLI, clindamycin; MTZ, metronidazole; ETP, ertapenem; AMC, amoxicillin-clavulanate; CRO, ceftriaxone; PTZ, piperacillin-tazobactam; PCN, penicillin; AMS, ampicillin-sulbactam.

^bExcluding organisms deemed contaminants.

^cM, male; F, female.

D. desulfuricans was the most commonly isolated species (61%) in our study, in contrast to findings by Loubinoux et al., who reported D. piger to be the most common species in 100 consecutive specimens obtained from thoracoabdominal purulent material (22). In our series, Desulfovibrio spp. were most often isolated from blood cultures (87%) followed by abdominal sources (7.5%). Desulfovibrio infections occurred in patients with a median age of 74 years (range, 6 to 93 years), with males comprising 65% of cases.

In the Mayo Clinic series of 12 patients, two-thirds of the isolates were from blood. All of the cases but one involved patients who had a gastrointestinal portal of infection, the exception being the patient with D. legallii prosthetic-joint infection described here. We hypothesize that this patient, who had been previously immunosuppressed with high-dose corticosteroids and who did not have any symptoms or history referable to the gastrointestinal tract, may have acquired the organism exogenously through water exposure, with D. legallii entering through breaks in her psoriatic skin and seeding her prosthetic joint.

Our report expands on the species which may cause human infections, including D. legallii, which is a novel species first described in a Tunisian wastewater digester in 2011 (1), and also two cases of D. intestinalis bacteremia. Interestingly, D. intestinalis, which was first described in termite hindguts (23), was recently described as part of vaginal flora in 3% of women in a Japanese study (11). Unfortunately, we do not have clinical information on the two referred cases of D. intestinalis bacteremia. There were two isolates each of D. fairfieldensis and D. piger, with one of the D. piger isolates originating from our institution. D. fairfieldensis and D. piger have so far been described as only human-related bacteria (24).

Our results corroborate previous reports showing that Desulfovibrio spp. are predictably susceptible to metronidazole but that piperacillin-tazobactam, which is commonly used to treat anaerobic infections, is not reliably active. Regarding carbapenems, 23 isolates were tested in our study; ertapenem or imipenem or both were uniformly active. However, discrepant susceptibilities to carbapenems have been previously reported (17), so we suggest that AST be performed for the specific carbapenem used for treatment, particularly for isolates from sterile sources. Although AST was not performed on the two referred D. fairfieldensis isolates, this species has been associated with antimicrobial resistance, including resistance to ertapenem/meropenem (17, 21). D. fairfieldensis may be more virulent than other species, given reports of invasive infections (16, 25); however, putative virulence factors remain to be elucidated. Based our series, however, it appears that D. desulfuricans is the species most commonly associated with bacteremia and other sites of infection in humans.

Anaerobes are an uncommon cause of prosthetic-joint infection (26), and species of genus Desulfovibrio have been reported infrequently as causes of human infection. However, as illustrated by this case of D. legallii prosthetic-joint infection, such infections may occur in a susceptible host with risk factors. It is likely that with more widespread use of technologies such as sequencing and MALDI-TOF MS in the clinical laboratory (27), these types of organisms will be more readily identified, providing further insight into the role of Desulfovibrio spp. in human disease.