Residual diabetic foot osteomyelitis after surgery leads to poor clinical outcomes: A systematic review and meta-analysis

Abstract

The aim of this meta-analysis is to compare the clinical outcomes in patients with and without residual osteomyelitis (ROM) after surgical bone resection for diabetic foot osteomyelitis (DFO). We completed a systematic literature search using PubMed, Scopus, and Embase using keywords DFO, Residual OM (ROM), and positive bone margins. The study outcomes included wound healing, antibiotic duration, amputation, and re-infection. Five hundred and thirty patients were included in the analysis; 319 had no residual osteomyelitis (NROM), and 211 had ROM. There was not a significant difference in the proportion of wounds that healed 0.6 (p = 0.1, 95% confidence intervals [95% CI] 0.3-1.3). The risk of infection was 2.0 times higher (OR = 2.0, p = 0.02, 95% CI 1.1-3.4), and the risk of amputation was 4.3 times higher (OR = 4.3, p = 0.0001, 95% CI 2.4-7.6) in patients with ROM. Patients with ROM received antibiotics significantly longer. The mean difference was 16.3 days (p = 0.02, 95% CI 11.1-21.1).

Keywords: amputation; diabetes; foot ulcer; infection; neuropathy; osteomyelitis.

FIGURE 1

Funnel Plot measuring publication bias Egger's test indicates that there is no strong evidence of publication bias in the meta‐analysis. The high p‐value (0.74) suggests that the observed results are likely not significantly skewed by a tendency to publish studies with positive or statistically significant findings.

FIGURE 2

Preferred Reporting Items for Systematic Reviews and Meta‐Analysis (PRSIMA) flow chart of search process. The figure presents the records extracted from three different scientific databases and registrars: PubMed, Scopus, and Embase. All articles were screened based on the inclusion/exclusion criteria and 11 studies were evaluated in full review. Three studies were excluded from this list, yielding eight papers for meta‐analysis.

FIGURE 3

Forest plot of antibiotic duration in patients with and without residual osteomyelitis. Forest plot depicting the mean difference in antibiotic duration in patients with and without residual osteomyelitis following surgical resection. Patients with residual osteomyelitis had longer antibiotic treatment with an average mean difference across all studies of 16.3 days (p = 0.02, 95% CI 11.1–21.1). There was no evidence of heterogeneity, Cochrane Q = 4.5 (p = 0.2) and I ² = 33.2 (95% CI 0–76.3).

FIGURE 4

Forest plot of re‐amputation in patients with and without residual osteomyelitis. Forest plot depicting re‐amputation in patients with and without residual osteomyelitis following surgical resection. The risk of amputation was 4.3 times higher when there was residual osteomyelitis, (OR = 4.3, p = 0.0001, 95% CI 2.4–7.6). There was no evidence of heterogeneity, Cochrane Q = 1.4 (p = 0.96) and I ² = 0.0 (95% CI 0.0–0.0).

FIGURE 5

Forest plot of wound healing rates in patients with and without residual osteomyelitis. Forest plot depicting the association of healing in patients with and without residual osteomyelitis following surgical resection. The average OR across studies was calculated as 0.6 (p = 0.1, 95% CI 0.3–1.3), indicating that residual osteomyelitis does not have a deleterious effect on wound healing. There was no evidence of heterogeneity, Cochrane Q = 1.5 (p = 0.7) and I ² = 0.0 (95% CI 0.0–70.4).

FIGURE 6

Forest plot of re‐infection in patients with and without residual osteomyelitis. Forest plot depicting re‐infection in patients with residual osteomyelitis following surgical resection. The risk of re‐infection was 2.0 times higher in patients with residual osteomyelitis, (OR = 2.0, p = 0.02, 95% CI 1.1–3.4). There was no evidence of heterogeneity, Cochrane Q = 4.4 (p = 0.35) and I ² = 9.11 (95% CI 0.0–81.1).