Computed Tomography and Magnetic Resonance Imaging Parameters Associated with Poor Clinical Outcome in Spondylodiscitis

Sarah C. Foreman, Benedikt J. Schwaiger, Bernhard Meyer, Alexandra S. Gersing, Claus Zimmer, Jens Gempt, Jan S. Kirschke

Research output: Contribution to journalArticlepeer-review

9 Scopus citations

Abstract

Background: Prognostic factors for the disease course of patients with spondylodiscitis have not been well studied. Methods: The prognostic value of initial magnetic resonance imaging (MRI) and computed tomography imaging parameters was analyzed in 62 patients (47% women; mean age ± SD, 71.6 ± 9.6 years) with a confirmed diagnosis of spondylodiscitis. The disease course was separately evaluated during initial treatment response during hospitalization, relapse, and clinical short-term follow-up at 3 months. Results: Overall CT findings graded as definitely inflammatory (P = 0.006), reduced disc height on MRI (P = 0.044) and fluid-equivalent hyperintensity of discs on T2 short tau inversion recovery–weighted sequences (P = 0.047) were associated with poor initial treatment response. High initial C-reactive protein value (>10.1 mg/dL) was associated with a higher relapse rate (P = 0.038). Risk factors for poor outcome were infection with low-virulence bacteria (P = 0.040) and overall MRI findings atypical for infection (P = 0.027). Conclusions: Compared with MRI, CT imaging parameters have a higher prognostic value regarding the disease course. Patients infected with low-virulence bacteria and atypical MRI findings are at higher risk for poor clinical outcome and thus warrant closer monitoring.

Original languageEnglish
Pages (from-to)919-926.e2
JournalWorld Neurosurgery
Volume104
DOIs
StatePublished - Aug 2017

Keywords

  • CT
  • Infection
  • MRI
  • Outcome
  • Spine
  • Spondylodiscitis

Fingerprint

Dive into the research topics of 'Computed Tomography and Magnetic Resonance Imaging Parameters Associated with Poor Clinical Outcome in Spondylodiscitis'. Together they form a unique fingerprint.

Cite this