The Brief Case: Strictly Anaerobic and Staining Acid Fast

CASE

A 38-year-old male with a past medical history of poorly controlled diabetes mellitus type 2 and hidradenitis suppurativa, a chronic inflammatory condition affecting apocrine glands of the axilla and groin, presented with 1 week of posterior neck swelling and pain. He noted having chronic, stable draining lesions in the axilla and groin but developed a new area of swelling and tenderness in the left posterior neck over the week prior to presentation. On presentation he was afebrile with a heart rate of 130 beats per min (bpm). Other vital signs were within normal limits. His exam was notable for a left-sided occiput large indurated area with mild erythema and tenderness to palpation without active drainage. Laboratory studies were notable for white blood cells, 22.2 × 10³/mm³ (reference range, 4.0 × 10³ to 11.0 × 10³/mm³), and hemoglobin, 10.8 g/dL (reference range, 14.0 to 17.0 g/dL). Computed tomography of the soft tissues of the neck with contrast demonstrated a rim-enhancing hypodense collection in the subcutaneous tissues of the posterior neck measuring 3.1 by 6.1 by 3 cm with surrounding fat infiltration and phlegmonous change (Fig. 1). He was taken to the operating room by the Otolaryngology Department for incision and drainage. In the operating room he was found to have a left occipital abscess with copious purulence expressed with multiple pockets of pus and murky fluid. He was initially given one dose of vancomycin and piperacillin-tazobactam, which were stopped, and ampicillin-sulbactam was started postoperatively.

FIG 1.

Computed tomography of soft tissues of neck with contrast demonstrating rim-enhancing hypodense collection in the subcutaneous tissues of the posterior neck measuring 3.1 by 6.1 by 3 cm with surrounding fat infiltration and phlegmonous change.

Purulent fluid and a tissue specimen were collected during the incision and drainage procedure for the posterior neck abscess and were submitted for routine bacterial aerobic and anaerobic, fungal, and acid-fast bacillus (AFB) cultures. The direct Gram stain of the fluid showed Gram-variable rods with poor retention of crystal violet (Fig. 2A), but no organisms were observed on the Gram stain from tissue. Auramine-rhodamine (AR) stain of the fluid was read as positive (4+, >30 organisms seen per 400× field) for acid-fast bacilli (AFB), and the tissue was also AR positive but with considerably fewer organisms (1+, 2 to 20 organisms seen in a total of 50 400× fields). The aerobic (sheep’s blood, chocolate, and MacConkey agar, 3 days of incubation) and anaerobic (CDC anaerobic agar, 5 days of incubation in anaerobic environment) cultures showed no growth. Similarly, fungal cultures of the neck fluid aspirate showed no growth after 28 days, and AFB cultures on 7H11 and LJ-Gruff agar media and BD MGIT broth medium had no growth after 12 weeks (extended culture time because of positive smear). To confirm the AR result when the organism did not grow in culture, both Kinyoun (weakly positive) and modified Kinyoun (strongly positive) (Fig. 2B) stains were performed on the fluid aspirate. Because no organism grew in culture, the fluid specimen was sent to a reference lab (University of Washington) to be tested for tuberculous and nontuberculous mycobacteria by rpoB and hsp65 PCR and sequencing. Mycobacterium-specific targets were negative, but Lawsonella clevelandensis was identified via 16S rRNA primer sets included with the nontuberculous mycobacterium evaluation and subsequent 16S amplicon sequencing.

FIG 2.

Gram stain (A) demonstrating poor crystal violet retention and modified Kinyoun stain (B) of neck fluid aspirate demonstrating aggregating bacteria. Magnification, ×1,000.

He was treated with 7 days of intravenous ampicillin-sulbactam, 3 g every 6 h, while admitted as an inpatient. He was discharged with an additional 10 days of oral amoxicillin-clavulanate, 875 to 125 mg every 12 h. At outpatient follow-up his posterior neck swelling was completely resolved but he was noted to still have some purulent drainage from the left cheek, and a culture was sent. Gram stain showed very few Gram-variable rods, and the modified Kinyoun stain showed 1+ modified acid-fast-positive bacilli. He was given an additional 4-week course of amoxicillin-clavulanate, 875 to 125 mg every 12 h, with improvement in his symptoms.

DISCUSSION

Lawsonella clevelandensis is a recently described Gram-positive, anaerobic, and partially acid-fast bacterium first reported in the literature by Harrington et al. (1), describing four case reports where L. clevelandensis was positively identified by 16S rRNA sequencing. Phylogenetic analysis of the 16S rRNA gene sequences generated using ClustalW and the neighbor joining algorithm from MEGA5 along with cellular fatty acid analysis determined that L. clevelandensis is within the suborder of Corynebacterineae with its nearest phylogenetic neighbor being Dietzia timorensis (1). The combination of being strictly anaerobic and containing cell wall mycolic acids sufficient to stain acid fast or partially acid fast is unique to L. clevelandensis among organisms that have been described to cause human disease. Previous reports have described it as either acid fast by Kinyoun method from direct culture or partially acid fast by modified Kinyoun method from subculture. In our case it was weakly positive by Kinyoun stain and strongly positive by modified Kinyoun stain from the direct specimen. Harrington et al. (1) speculate that the difference in staining from culture versus subculture could be due to the effects of a purulent sample on the staining of the organism. There is also consideration given to the fact that L. clevelandensis tends to be isolated from fatty tissue, such as fatty liver or soft tissue in obese patients. Further characterization of the makeup of this organism’s cell wall and staining properties would be academically interesting and novel material. The organism is further described as non-spore forming, catalase positive, and pleomorphic, with shapes that vary from cocci to bacilli based on sizes that range from 1.0 μm to 2 μm in length. Reports generally describe L. clevelandensis as being isolated from soft tissue abscesses and difficult to isolate in conventional cultures due to its fastidious nature and the prolonged incubation time required for culture (2). As clinical reports of L. clevelandensis are rare, and mycobacterial cultures are slow, AFB smear-positive but culture-negative cases may still not justify taking specific measures to routinely attempt recovery of L. clevelandensis, but it may be justifiable to extend anaerobic cultures in some cases and to retain direct samples for possible 16S rRNA identification. An extended anaerobic culture was attempted in our case after identification via 16S rRNA sequencing; however, recovery was unsuccessful due to nonviability of the organism from a direct specimen.

The growth of the organism on solid medium was not observed in our case in standard 5-day anaerobic culture on CDC anaerobic agar. Previous reports indicate that L. clevelandensis can be recovered on CDC anaerobic blood agar (supplemented with hemin and vitamin K) after 5 to 7 days of incubation at 35°C using an anaerobic gas pack in an anaerobe jar, producing pinpoint and waxy colonies, about 2 to 3mm in diameter (2). There is no growth in fastidious anaerobic broth (2). Although we did not grow the organism in our lab, it is worth noting that our primary identification system is matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS) and that L. clevelandensis is not currently listed in the most up-to-date Bruker or bioMérieux MALDI_TOF MS databases, and it is also not included in the Bruker research-use-only database as of the time of writing.

An ecological niche has not been identified for Lawsonella. It is suggested that it may be a part of the oral microbiota (1) or human nasal microbiota (3), but it has mainly been isolated from abscesses, some described as within the breast, liver, spine, and peritoneum in the cases published from the United States and Canada (2). Of note, all 4 case reports by Harrington et al. (1) describe monomicrobial abscesses, as was seen in our case report. Patients are described as having various degrees of compromised immune function such as rheumatologic diseases and neoplastic conditions. They are also described as having extensive past medical histories, particularly including being diagnosed with diabetes mellitus. The patient in this case report also had a significant history of uncontrolled diabetes mellitus, as well as essential hypertension and hidradenitis suppurativa.

In this case study, the organism was ultimately identified via 16S rRNA sequencing, as was common in many of the other published case studies. While previous reports have demonstrated branching Gram-positive bacilli from direct specimens (but not from subculture) (1), an aggregating structure reminiscent of the mycobacterial cording phenomenon was observed in our patient’s specimens (Fig. 2B). Susceptibility testing for this organism has not been described previously, but the published reports describe an improving clinical course with treatment with amoxicillin-clavulanic acid along with incision and drainage of the abscess. Treatment duration is listed as anywhere between 4 weeks and 6 months twice a day (1, 4). A different study reports using a 6-week course of intravenous vancomycin and oral doxycycline, followed by chronic suppression with doxycycline and cefadroxil with resolved symptoms and an uncomplicated postoperative recovery (5). Our patient was treated with intravenous ampicillin-sulbactam, 3 g every 6 h during his admission, and was discharged on a 10-day course of amoxicillin-clavulanate, 875 to 125 mg every 12 h, in addition to doxycycline at 100 mg every 12 h for his hidradenitis. He was briefly lost to follow-up, but when he returned to clinic 6 weeks after discharge from the hospital, he was successfully treated with a 4-week course of amoxicillin-clavulanate, 875 to 125 mg every 12 h, for the L. clevelandensis infection.

In summary, L. clevelandensis is a recently described anaerobic, acid-fast bacterium that has been isolated from abscesses in patients with a certain level of immunocompromise, particularly with a diagnosis of diabetes mellitus. The bacterium is not readily identified by conventional microbiological detection methods, but extended anaerobic culture and methods such as 16S rRNA gene sequencing may help with identifying this clinically significant bacterium. Ultimately, the retention of the direct specimen is key in preserving the ability to send out for 16S rRNA identification in laboratories that do not routinely do this specialized testing. Extended aerobic and anaerobic cultures are probably not justified, but as always, a discussion with the patient’s clinical team about the impact on patient care can help to further guide the lab in deciding the priority of next steps for identifying fastidious organisms such as L. clevelandensis.

SELF-ASSESSMENT QUESTIONS

• Which of the following descriptions is consistent with L. clevelandensis?

  • a.Gram-positive branching or aggregating bacilli, growth under aerobic or anaerobic conditions, modified Kinyoun stain negative, usually identified by MALDI-TOF MS
  • b.Gram-positive branching or aggregating bacilli, growth under anaerobic conditions only, modified Kinyoun stain positive, usually identified by 16S rRNA sequencing
  • c.Gram-positive bacilli, modified Kinyoun stain positive, growth under anaerobic conditions or on routine AFB medium cultured aerobically, usually identified by 16S rRNA sequencing
  • d.Gram-positive bacilli that grow only in anaerobic broth culture, modified Kinyoun stain positive, usually identified by MALDI-TOF MS
• 2.Among clinically relevant bacteria, what is the closest relative of L. clevelandensis?
  • a.Mycobacterium spp.
  • b.Corynebacterium spp.
  • c.Dietzia spp.
  • d.Lactobacillus spp.
• 3.Which of the following statements is most accurate in relation to the specimen type/source that L. clevelandensis has been isolated from according to the limited available literature?
  • a.L. clevelandensis has been isolated in blood cultures.
  • b.L. clevelandensis has been isolated from ear swabs.
  • c.L. clevelandensis has been isolated from abscesses of various locations.
  • d.L. clevelandensis is found isolated only from abscesses of the head and neck region.

ANSWERS TO SELF-ASSESSMENT QUESTIONS

• 1.Which of the following descriptions is consistent with L. clevelandensis?
  • a.Gram-positive branching or aggregating bacilli, growth under aerobic or anaerobic conditions, modified Kinyoun negative, usually identified by MALDI-TOF MS
  • b.Gram-positive branching or aggregating bacilli, growth under anaerobic conditions only, modified Kinyoun positive, usually identified by 16S rRNA sequencing
  • c.Gram-positive bacilli, modified Kinyoun positive, growth under anaerobic conditions or on routine AFB media cultured aerobically, usually identified by 16S rRNA sequencing
  • d.Gram-positive bacilli that grow only anaerobic broth culture, modified Kinyoun positive, usually identified by MALDI-TOF MS

Answer: b. Introductory work by Harrington et al. describe a branching Gram-positive rod that supported the description of a new genus and species of the suborder Corynebacterineae (3). A subsequent article by Bell et al. described the naming of L. clevelandensis and reiterated its presence as a novel member of the suborder Corynebacterineae (2). They listed its phenotypic characteristics as: Gram stain positive, partially acid fast, non-spore forming, anaerobic, catalase positive and pleomorphic branching bacteria. They also describe detecting mycolic acids via HPLC and one dimensional TLC giving the bacteria its acid fast qualities. This combination of being anaerobic and acid fast is a unique feature of the bacteria.

• 2.Among clinically relevant bacteria, what is the closest relative of L. clevelandensis?
  • a.Mycobacterium spp.
  • b.Corynebacterium spp.
  • c.Dietzia spp.
  • d.Lactobacillus spp.

Answer: c. Bell et al. describe a number of methods that were utilized to identify and classify L. clevelandensis. They include analysis of 16S rRNA sequencing, utilizing a phylogenetic tree of identified bacteria sequences using the neighbor joining method, with topology assessed by bootstrap analysis of 2000 replicates using molecular evolutionary genetics analysis software (MEGA5). Whole-cell hydroxylates of isolated strains were also detected and characterized using HPLC and one dimensional TLC and were found to produce an alpha-mycolic acid similar in chain length and Rf value to those of Dietzia. Ultimately, similarity values of <95% to the nearest neighbor categorized L. clevelandensis as belonging within the suborder of Corynebacterineae but classified as a novel bacteria with Dietzia as its closest phylogenetic neighbor.

• 3.Which of the following statements is most accurate in relation to the specimen type/source that L. clevelandensis has been isolated from according to the limited available literature?
  • a.L. clevelandensis has been isolated in blood cultures
  • b.L. clevelandensis has been isolated from ear swabs
  • c.L. clevelandensis has been isolated from abscesses or various anatomic locations
  • d.L. clevelandensis is only found isolated from abscesses of the head and neck region

Answer: c. Very little is known about the ecological niche of this organism, but reports demonstrate identification of isolates only from human sources, all in different geographic locations in the United States and Canada. Literature describes isolation of this bacteria primarily from abscesses of different sources such as breast, liver, spine and peritoneum. In our case, L. clevelandensis was isolated from an abscess found on the posterior neck.