[Risk factors of death in patients with bloodstream infections in a medical intensive care unit]

To analyze the incidence, etiological characteristics, and risk factors for death in patients with bloodstream infection in a medical intensive care unit (ICU).

A retrospective case-control study was conducted. Adult patients with bloodstream infection admitted to the medical ICU of The Third Xiangya Hospital of Central South University from January 2015 to December 2023 were enrolled. Data were collected from the hospital's electronic medical record system, including demographic characteristics [gender, age, body mass index (BMI)], clinical features (comorbidities, pathogen types, source of bloodstream infection, antibiotic use), laboratory parameters [routine blood test, liver and kidney function, results of bacterial culture and drug susceptibility test, C-reactive protein (CRP), procalcitonin (PCT), etc.], clinical management (red blood cell transfusion and usage of immunoglobulin, thymosin α1, glucocorticoid, empirical antibiotic therapy, mechanical ventilation, renal replacement therapy, etc.), and prognostic indicators (length of medical ICU stay and length of hospital stay after bloodstream infection). All patients were followed up for at least 3 months after the onset of bloodstream infection. Based on follow-up outcomes, patients were divided into survivor and non-survivor groups. Univariate analysis was used to identify variables associated with 3-month all-cause death in patients with bloodstream infection. Variables with P<0.05 in the univariate analysis were entered into multivariate Logistic regression analysis to identify independent risk factors for 3-month all-cause death. The Hosmer-Lemeshow test was used to assess the goodness-of-fit of the regression model.

A total of 106 of the 4 820 medical ICU patients (2.2%) ultimately included in the study developed bloodstream infection during the 9 years. Among the 106 patients with bloodstream infection, the majority were male (71 cases, 67.0%), with an advanced age [(64.2±17.1) years]. The BMI was 23.0 (21.0, 23.0) kg/m². The most common concomitant diagnosis was pulmonary infection (99 cases, 93.4%). Common comorbidities included type 2 diabetes mellitus (39 cases, 36.8%), gastrointestinal bleeding (26 cases, 24.5%), malignancy (23 cases, 21.7%), and chronic obstructive pulmonary disease (21 cases, 19.8%). The most common source of bloodstream infection was the lung (53 cases, 50.0%). At the time of bloodstream infection onset, patients presented with a clinical state characterized by immunosuppression, systemic inflammatory response, malnutrition and potential kidney injury. Septic shock was present at the time of bloodstream infection onset in 71 patients (67.0%). Forty patients (37.7%) received appropriate empirical antibiotic therapy after bloodstream infection onset. Mechanical ventilation was required in 83 patients (78.3%), and renal replacement therapy was required in 26 patients (24.5%). The median length of medical ICU stay and the median total hospital stay after bloodstream infection were 10.0 (4.0, 20.3) days and 16.0 (7.0, 28.3) days, respectively. A total of 113 bacterial or fungal strains were isolated from the blood of the 106 patients with bloodstream infection, among which Gram-negative bacteria were the predominant pathogens (73 strains, 64.6%). The most common pathogens were Enterobacteriaceae, predominantly Klebsiella pneumoniae (29 strains, 25.7%). Among the 106 patients, 68 patients (64.2%) died within 3 months after bloodstream infection onset, while 38 patients (35.8%) survived. Univariate analysis showed that, compared with the survivor group, the non-survivor group had significantly higher proportions of patients aged≥75 years, those with malignancy, those with carbapenem-resistant Gram-negative bacterial bloodstream infection, those with LYM≤0.5×10⁹/L at bloodstream infection onset, those with septic shock at bloodstream infection onset, those requiring mechanical ventilation and renal replacement therapy. In contrast, the proportions of patients with BMI>23 kg/m² and those receiving appropriate empirical antibiotic therapy after bloodstream infection onset were significantly lower in the non-survivor group (all P<0.05). Multivariate Logistic regression analysis showed that LYM≤0.5×10⁹/L at bloodstream infection onset [odds ratio (OR)=3.268, 95% confidence interval (95%CI) was 1.196-8.927, P=0.021], septic shock at bloodstream infection onset (OR=3.637, 95%CI was 1.293-10.235, P=0.014), and the need for mechanical ventilation (OR=4.111, 95%CI was 1.242-13.610, P=0.021) were independent risk factors for 3-month all-cause death in patients with bloodstream infection, while receipt of appropriate empirical antibiotic therapy after bloodstream infection onset was a protective factor (OR=0.316, 95%CI was 0.118-0.844, P=0.022). The Hosmer-Lemeshow test yielded a χ ² value of 5.082 with 6 degrees of freedom (P=0.533), indicating no evidence of overfitting.

Bloodstream infection is a common complication among patients in the medical ICU, with Enterobacteriaceae being the most frequently isolated pathogens. The mortality of bloodstream infection is high. LYM≤0.5×10⁹/L at bloodstream infection onset, septic shock at bloodstream infection onset, and the need for mechanical ventilation are independent risk factors for 3-month all-cause death in patients with bloodstream infection, while receipt of appropriate empirical antibiotic therapy after bloodstream infection onset is a protective factor.