Outcomes, Microbiology and Antimicrobial Usage in Pressure Ulcer-Related Pelvic Osteomyelitis: Messages for Clinical Practice

Abstract

Introduction: Pressure ulcer-related pelvic osteomyelitis is a relatively under-studied entity in the field of bone infection. We sought to add to the limited evidence base for managing this challenging syndrome. Methods: Cases were identified retrospectively from a surgical database and hospital discharge codes at a U.K. tertiary centre (2009-2018). Risk factors associated with outcomes were analysed by logistic regression. Results: We identified 35 patients (mean age 57.4 years), 69% managed with a combined medical and surgical approach, with mean follow-up of 3.7 years from index admission. Treatment failure (requiring further surgery or intravenous antimicrobials) occurred in 71% and eventual ulcer healing in 36%. One-year mortality was 23%. Lack of formal care support on discharge, post-traumatic (asensate) neurological deficit and index CRP (>184mg/L) were associated with treatment failure (p=0.001). Age (>59.5 years), lack of attempted soft tissue coverage, haemoglobin (<111g/L) and albumin (<25g/L) were associated with non-healing ulcers (p=0.003). Superficial wound swabs had low sensitivity and specificity compared to deep bone microbiology. Infection (based on deep bone microbiology from 46 infection episodes) was usually polymicrobial (87%), commonly involving S. aureus, Enterococci, GNB and anaerobes. Antimicrobial duration ranged from 0-103 days (mean 54) and was not associated with subsequent treatment failure. Conclusions: Attempted soft tissue coverage after surgical debridement, ensuring appropriate support for personal care after discharge and nutritional optimisation could improve outcomes. Superficial wound swabs are uninformative and deep bone sampling should be pursued. Long antimicrobial courses do not improve outcomes. Clinicians should engage patients in anticipatory care planning.

Keywords: osteomyelitis; pelvis; pressure ulcer; spinal cord injuries.

Figure 1

Antimicrobial usage. a) The total antimicrobial duration at index presentation was not associated with subsequent treatment failure (groups compared by unpaired t test). b) Antimicrobial agents used during management of index episode. c) Duration of each antimicrobial agent during index episodes (number is mean). Antimicrobial data available for 33 patients. Graphs show mean and standard deviation.

Figure 2

Clinical outcomes. a) Kaplan-Meier analysis of patient survival. b) Log rank analysis of patient survival between those with healed and non-healed pressure ulcers at time of final follow-up. c) Treatment failure. d) Admission C-reactive protein (mg L^-1) at index presentation in the prediction of treatment failure.