1186. Recovery of Propionibacterium spp. in Cultures from the Central Nervous System – Clinical and Laboratory Criteria to Distinguish Infection from Contamination
Session: Poster Abstract Session: Clinical Infectious Diseases: CNS Infection
Friday, October 28, 2016
Room: Poster Hall
  • IDweekPoster.10.17.16.pt.pdf (401.0 kB)
  • Background:

    Propionibacterium spp. (P. spp) are commensal bacteria, but are implicated in post-operative central nervous system infections (CNS) and may be underappreciated pathogens in patients without prior neurosurgery. Laboratory and clinical characteristics that predict infection rather than contamination when P. spp. are isolated from the CNS are poorly understood.


    We identified P. spp. isolates from specimens in the Barnes-Jewish Hospital microbiology lab from 2011-2014. Chart review was performed to distinguish CNS infection from contamination. We defined P. spp as infectious if treating physicians administered ≥ 7 days of directed antibiotic therapy against P. spp.


    Of 416 all-site P. spp. isolates, 100 were from CNS specimens. Twenty-nine (29%) CNS isolates were considered infections; 20 (69%) were P. acnes and 9 (31%) were P. spp. Of these, cultures were from CSF (n=13), wounds involving CNS (n=7), CNS abscess (n=4), CNS aspirates (n=2), CNS tissue, hardware, or cyst (n=1 each). Seventy-one CNS isolates were considered contaminants, 44 (62%) were P. acnes and 27 (38%) were P. spp. Species did not affect likelihood of treatment (p=0.51). Patients with 4+ WBCs on Gram stain (n=3) or organisms on Gram stain (n=6) were considered infections. Compared to patients with contaminated cultures, P. spp. infections were more often polymicrobial [9/29 (31%) vs. 2/71 (2.8%) non-infections; p<0.001] or in patients with prior neurosurgical procedure, hardware, or tumor [25 (86%) vs. 16 (22.5%); p<0.001]. Patients with infections had higher CSF nucleated cell counts (median 36 cells/mcl vs. 4 cells/mcl; p=0.002) and protein (median 99.5 mg/dl vs. 41 mg/dl; p=0.010), and higher quantification of organisms in culture (median 2+ vs. 1+; p<0.001), Gram stain (median 0 vs. 0; p<0.001), and WBCs on Gram stain (median 1+ vs. 0; p=0.008). CSF glucose was not different between groups (median 66 mg/dl vs. 69 mg/dl; p=0.49). Only 5.7% (n=2) of patients with 0 cells/mcl (n=35) in the CSF were treated; both had CNS hardware. Of the 31 patients with 0 cells/mcl from CSF, 4 had hardware.


    Patients with no CNS instrumentation or pathology, negative Gram stain, and normal CSF parameters are less likely to have a true infection if P. spp. are isolated.

    Jason Burnham, MD, Washington University in Saint Louis, Saint Louis, MO, Carey-Ann D. Burnham, PhD, Pathology & Immunology, Washington University, St. Louis, MO and David K. Warren, MD, MPH, FIDSA, FSHEA, Division of Infectious Diseases, Washington University School of Medicine, St. Louis, MO


    J. Burnham, None

    C. A. D. Burnham, None

    D. K. Warren, None

    Findings in the abstracts are embargoed until 12:01 a.m. CDT, Wednesday Oct. 26th with the exception of research findings presented at the IDWeek press conferences.