Table 5.
Nested case control analysis: multivariate analysis to identify predictors of persistent bacteremia in patients with repeat cultures
| Variable | Odds ratio | 95 % CI | P value |
|---|---|---|---|
| Sex | |||
| Male | 2.59 | 1.28–5.25 | 0.008 |
| Femalea | 1.00 | 1.00 | 1.00 |
| Admitting service | |||
| Medical in-patient | 2.80 | 1.34–5.84 | 0.006 |
| Heme-onc in-patient | 1.21 | 0.44–3.33 | 0.71 |
| Surgical in-patienta | 1.00 | 1.00 | 1.00 |
| Organism | |||
| S. aureus | 4.49 | 1.88–10.73 | <0.001 |
| Other Gram-positive bacteria | 0.80 | 0.35–1.82 | 0.59 |
| Anaerobes | 0.96 | 0.15–6.38 | 0.97 |
| All Gram-negative bacteriaa | 1.00 | 1.00 | 1.00 |
| Source of bacteremia | |||
| Endovascular | 7.66 | 2.30–25.48 | <0.001 |
| Epidural abscess/discitis | 26.99 | 1.91–391.08 | 0.015 |
| Other extravascular source | 3.02 | 0.97–9.41 | 0.057 |
| Unknown sourcea | 1.00 | 1.00 | 1.00 |
NOTE: Multivariable logistic regression was performed including all variables associated with persistent bacteremia in bivariate analyses in the nested case control analysis, with stepwise backward selection for variable reduction until all retained variables were statistically significant at p-value threshold <0.05 (statistically significant values are bolded). Referent categories are indicated with ‘a’. Heme-onc hematology-oncology in-patient ward