3.1 Demographic data
During the eight-year study period, 435 patients with MP were finally recruited from the initial 1137 possible MP patients, and 82 (18.9%) of them developed secondary MRSA-BSI (MP-BSI) (Fig.1). The demographic data were listed in Table1. The median age was 62 (IQR,51,72) years, and 74.5% (324/435) of them were male. The most common comorbidity was cerebrovascular accident or traumatic brain injury (39.1%), followed by diabetes (15.4%) and chronic heart failure (8.5%). 58.6% (255/435) of patients were exposed to antibiotics before the onset of pneumonia, and 51.7% (225/435) of them received surgery before MRSA-BSI. Nosocomial MP and community-acquired MP (CA-MP) accounted for 81.4% and 18.6%, respectively.
Compared with MP, patients with MP-BSI showed a high proportion of immunosuppression (13.4% vs. 1.7%, p<0.001) and CA-MP (35.4% vs. 14.7%, p<0.001), a higher SOFA score (median, 5 vs. 4, p=0.018), faster heart rate (HR) (median, 102 vs. 87, p<0.001), accelerated respiratory rate (RR) (median, 20 vs. 18, p<0.001), higher mean arterial pressure (median, 89.7 vs. 84.9, p=0.015), and a lower oxygenation index (median, 248.4 vs. 280.0, p=0.007), but there were no differences in other parameters (all p>0.05) (Table1).
3.2 Biological Parameters
A comparison of biological parameters between groups of MP and MP-BSI was illustrated in Table 2. Patients with MP-BSI were positively correlated with alanine aminotransferase (ALT) (median, 38 vs. 31, p=0.011), gamma glutamyl transpeptidase (γ-GT) (median, 54.5 vs. 41.0, p=0.018), lactic dehydrogenase (LDH) (median, 311.5 vs. 278.0, p=0.023), direct bilirubin (DBil) (median, 5.4 vs. 4.2, p=0.017), but negatively correlated with albumin (mean, 30.01 vs. 32.00, p=0.006). Blood lactic acid was higher in MP-BSI than MP alone (median, 1.6 vs. 1.3, p=0.001).
3.3 Anti-infection Strategy
The anti-infection strategy was summarized in Table 3. All participants received initial antibiotic therapy and only a small proportion (10.3%) of them received empirical anti-MRSA therapy. The most administered targeted antibiotics in MP-BSI were glycopeptides (69.5%), followed by linezolid (22.0%) and other anti-MRSA antibiotics (8.5%), whereas they were linezolid (47.0%), glycopeptides (45.0%) and others (7.9%) in MP group (p<0.001). Compared with MP group, the MP-BSI group used a significantly lower proportion of linezolid than glycopeptides on targeted antibiotics (p<0.001). Additionally, patients with MP-BSI were delay in initiation of targeted antibiotics (median days, 4 vs. 3, p<0.001).
3.4 Independent Risk Factors For Secondary MRSA-BSI
High SOFA score (adjusted odds ratio[aOR], 1.192; 95% confidence interval [CI], 1.065–1.333; p=0.002), immunosuppression (aOR, 13.599; 95% CI, 4.063–45.521; p<0.001), CA-MP (aOR, 2.827; 95% CI, 1.496–5.343; p=0.001), time from initial to targeted antibiotics (aOR, 1.304; 95% CI, 1.136–1.497; p<0.001), accelerated RR (aOR, 1.135; 95% CI, 1.066–1.209; p<0.001), elevated γ-GT (aOR, 1.004; 95% CI, 1.001–1.008; p=0.016) were independent risk factors for secondary MRSA-BSI, while linezolid as targeted antibiotic was a protective factor (aOR, 0.224; 95% CI, 0.115–0.438; p<0.001) (Table 4).
3.5 Outcomes
Compared with MP, patients with MP-BSI required a longer hospital LOS (median days, 35.5 vs. 25, p=0.015), but there were no significant differences in LOS of intensive care unit (ICU) (median days, 9.5 vs. 12, p=0.893) and days of mechanical ventilation (median, 8 vs. 7, p=0.252) (Table 5). The all-cause hospital mortality in the entire cohort was 17% (74/435), which was substantially higher in patients with MP-BSI than those with MP alone (26.8% vs. 14.7%, p=0.009). Similarly, the 28-day all-cause mortality was also substantially higher in MP-BSI than in MP alone (24.4% vs. 11.0%, p=0.001), which was consistent with the results shown by survival curve (Fig.2).