Globicatella sanguinis Is an Etiological Agent of Ventriculoperitoneal Shunt-Associated Meningitis

Abstract

Globicatella sanguinis is a very rare isolate in clinical samples. We present a case of meningitis in a 69-year-old female patient after implantation of an external left ventricular drainage due to a hydrocephalus. She recovered after antibiotic treatment with ceftriaxone.

CASE REPORT

In October 2005 a 69-year-old female patient presented with a 1-year history of worsening weakness of her left leg and left arm and increasing problems with walking. She complained of being confused sometimes. Additionally, she had a longstanding history of depressive disorders for which she had been in psychiatric care for several years.

A computed tomography brain scan showed a dilated ventricular system, and the preliminary diagnosis of normal-pressure hydrocephalus was made. To confirm the diagnosis and to assess the suitability of the implantation of a ventriculoperitoneal shunt, an external left ventricular drainage was implanted on 26 October 2005. The diagnosis could be confirmed, and the drainage was removed on 28 October 2005. In the morning of 30 October 2005 the patient was somnolent, threw up several times, and presented clinical signs of meningitis. A control computed tomography scan showed an unchanged ventricular configuration. The differential diagnosis of decompensated hydrocephalus or infectious meningitis was made, and a new external ventricular drainage was implanted. In a sample of cerebrospinal fluid, lactate levels were elevated to 6.35 mmol/liter, and subsequently a gram-positive coccus was grown. The patient was put on intravenous ceftriaxone (2 g once daily for 10 days) and improved rapidly. After negative cerebrospinal fluid cultures, a definitive ventriculoperitoneal shunt was implanted on 9 November 2005. On the 22 November the patient was transferred to a specialized rehabilitation facility.

The culture of the first cerebrospinal fluid sample (Columbia agar supplemented with 5% sheep blood and pyridoxal, 37°C, 5%CO₂) resulted in faint growth of rough, alpha-hemolytic colonies after 24 h, which was more pronounced after 48 h (Fig. 1). Gram staining showed a gram-positive coccus (short chains). A catalase test was negative. An optochin test (5-μg disk; Oxoid) was negative (no inhibition zone). The rapid ID32strep test resulted in the finding of a Globicatella sp. The Phoenix system (PMIC/ID-56) confirmed the API result and identified the bacterium as Globicatella sanguinis (certainty of 99%). Surprisingly, a 16S rRNA sequence analysis showed 99% identity with the 16S rRNA sequence of G. sulfidifaciens (GenBank accession no. AJ297627) and 95% identity with the16S rRNA sequence of G. sanguinis (GenBank accession no. S50214).

FIG. 1.

Globicatella sanguinis (48 h of growth, CO₂, 37°C, Columbia agar supplemented with 5% sheep blood and pyridoxal).

Looking into the literature, we found that the biochemical profile of G. sanguinis is based on a total of 29 isolates (1a, 3). The biochemical profile of G. sulfidifaciens is based on a total of eight isolates (7). Of 33 reactions for our isolate, only one was not in line with G. sanguinis, but six reactions were not in line with G. sulfidifaciens, one of which is the production of sulfide (Table 1). In the original publication (7), the authors stressed the importance of this parameter for the differentiation between G. sanguinis (negative) and G. sulfidifaciens (positive). Our isolate is negative. On the other hand, there is only one 16S rRNA sequence for each Globicatella sp. in GenBank. Additionally, these sequences are very similar, and there is evidence that currently the database for the identification of Globicatella spp. based on 16S rRNA is too small to be helpful (2). Therefore, we named our isolate G. sanguinis in accordance with the biochemical profile.

TABLE 1.

Differences in biochemical profiles of G. sulfidifaciens, G. sanguinis, and our strain^a

Test	Resulta for:
	G. sulfidifaciens	G. sanguinis	Our strain
Mannitol	Neg	Pos	Weakly pos
β-Galactosidase	Neg	Pos	Pos
β-Glucuronidase	Pos	Neg	Neg
Hippurat	Neg	Pos	Pos
H2S (Kligler)	Pos	Neg	Neg
Arginine dihydrolase	ND	ND	Neg
β-Glucosidase	ND	ND	Neg
Sorbitol	Neg	Pos	Weakly pos
Lactose	Neg	Pos	Neg
α-Galactosidase	Pos	ND	Pos
Alkaline phosphatase	Neg	ND	Neg
Ribose	ND	Pos	Pos
Trehalose	Pos	Pos	Pos
Raffinose	Pos	Pos	Pos
Saccharose	Pos	ND	Pos
l-Arabinose	Var	ND	Pos
d-Arabitol	Neg	ND	Neg
Cyclodextrin	ND	ND	Neg
Voges-Proskauer	Neg	ND	Neg
Alanine-phenylalanine-proline arylamidase	ND	ND	Pos
β-Galactosidase	Var	ND	Pos
Pyroglutamic acid arylamidase	ND	ND	Pos
N-Acetyl-β-glucosamidase	ND	ND	Pos
Glycyl-tryptophane arylamidase	ND	ND	Neg
Glycogen	Pos	ND	Pos
Pullulan	ND	ND	Pos
Maltose	Pos	Pos	Pos
Melibiose	Pos	Pos	Pos
Melezitose	Neg	ND	Neg
Methyl-β-d-glucopyranoside	ND	ND	Neg
Tagatose	Neg	ND	Neg
β-Mannosidase	ND	ND	Neg
Urease	Neg	ND	Neg

^aND, not done; Pos, positive; Neg; negative; Var, variable. Results for our strain were determined by API 32strep and Phoenix testing. Data for G. sulfidifaciens and G. sanguinis are from references 7 and 1a, respectively. Boldface indicates differences between results for our strain and for G. sulfidifaciens.

Antimicrobial susceptibility testing was done according to the 2005 CLSI guidelines (1), using the established breakpoints for Streptococcus spp. other than S. pneumoniae. Broth microdilution was performed for penicillin G, cefotaxime, clarithromycin, clindamycin, tetracycline, telithromycin, gatifloxacin, and levofloxacin. MICs of vancomycin, linezolid, imipenem, erythromycin, and ciprofloxacin were determined using the Etest. The results are shown in Table 2.

TABLE 2.

MICs

Antimicrobial agent	MIC (μg/ml)	Interpretationa
Penicillin G	0.06	Sensitive
Cefotaxime	0.5	Sensitive
Imipenem	0.25	No breakpoint
Erythromycin	2	Resistant
Clarithromycin	1	Resistant
Telithromycin	≤0.03	No breakpoint
Clindamycin	≤0.12	Sensitive
Ciprofloxacin	0.38	No breakpoint
Levofloxacin	<0.25	Sensitive
Gatifloxacin	≤0.06	No breakpoint
Vancomycin	0.125	Sensitive
Tetracycline	1.0	Sensitive
Linezolid	0.5	Sensitive

^aAccording to reference 1.

There are two well-known mechanisms for macrolide resistance. The presence of the mef gene is associated with macrolide-only resistance, whereas the ermB gene causes combined macrolide/lincosamide resistance. There are no published reports on the origin of macrolide resistance in Globicatella isolates. To investigate the resistance mechanism in our isolate, a mef PCR and an ermB PCR were performed (4, 5). The mef PCR was positive, whereas the ermB PCR was negative. These results confirm the macrolide resistance and the lincosamide sensitivity of our isolate.

G. sanguinis was first described in 1992 by Collins and coworkers, who named it G. sanguis (1a). It was renamed G. sanguinis in 1997 by Trüper and de'Clari (6). A PubMed search for Globicatella revealed that G. sanguinis was described as a cause of meningoencephalitis in lambs by Vela and coworkers in 2000 (8). Whether our patient had contact with sheep could not be elucidated. In 2001, Shewmaker and coworkers (3) published the first susceptibility testing results for G. sanguinis, which are in good accordance with our own results. Interestingly, 48% of their 27 strains showed resistance to cefotaxime, with a MIC₅₀ of 1.0 mg/liter and a MIC₉₀ of 4.0 mg/liter. Fortunately for our patient, our isolate was susceptible to cefotaxime.