1. INTRODUCTION

Bloodstream infections (BSIs) are often caused by a single specific pathogen; however, the proportion of polymicrobial BSIs increases with certain patient characteristics, such as old age, immunocompromised status, malignancy, and neutropenia.1 Polymicrobial BSIs have been reported to account for as much as 6% to 34% of BSIs2–4; the mortality rate is approximately twice that of BSIs with a single specific pathogen, ranging from 21% to 63%.3 Although gram-negative BSIs have gradually declined over the past few decades, Pseudomonas aeruginosa, a glucose-nonfermenting gram-negative bacterium with minimal survival requirements, is an important pathogen of hospital-acquired BSIs.5 Notably, a recent study has shown that P. aeruginosa also accounts for up to 5% of community-onset BSIs.6 The gradual spread of P. aeruginosa morbidity to the community is associated with a high incidence of mortality, with previous reports showing that hospital mortality rates associated with P. aeruginosa bloodstream infections (Pa-BSIs) range from approximately 18% to 61%.7–11 On the premise of this high mortality, polymicrobial Pa-BSIs are also of concern to clinicians. Previous studies have found that patients with polymicrobial Pa-BSI were more severely infected, with a higher incidence of septic shock, than those with monomicrobial Pa-BSI,12,13 but the risk factors for polymicrobial Pa-BSI were not clearly described. As previous studies are dated and there is a lack of relevant studies in mainland China, it is necessary to reanalyze the clinical characteristics and risk factors for polymicrobial Pa-BSI to help clinicians identify associated risk factors in China and high-risk patients and implement preventive measures. Such efforts may also improve the prognosis of these patients and reduce the burden of polymicrobial BSIs.

2. METHODS 2.1. Patients and study design

This double-center retrospective cohort study was conducted from January 2013 to June 2022 at Taizhou Municipal Hospital and the Second Affiliated Hospital, Zhejiang University School of Medicine. The Ethics Committee of the Taizhou Municipal Hospital approved the retrospective analysis of patient data without obtaining patient consent. In addition, a statement of permission from patients for submission of the present study was not needed, as the study did not include any personal information.

The inclusion criteria included patients who had at least one positive blood culture for P. aeruginosa accompanied by signs and symptoms of infection. For patients with more than two positive blood cultures, the clinical data at only the first positive test were included. When pathogens other than P. aeruginosa were isolated from the same blood culture, the case was considered a polymicrobial Pa-BSI, and the clinical data were recorded at the same time. The exclusion criteria were as follows: (1) age <18 years; (2) incomplete or missing case data; and (3) pregnancy. In the event of isolation is of a potentially colonizing skin organism, such as coagulase-negative Staphylococcus, it was considered a contaminant unless found again in consecutive blood cultures on the same day.

2.2. Data collection

Clinical data were collected from electronic medical records. Demographic data were recorded, including age and sex; clinical data, including underlying diseases; scores reflecting the severity of the disease, including the Sequential Organ Failure Assessment (SOFA) score; Pitt bacteremia score; Charlson Comorbidity Index (CCI); Acute Physiology, Age and Chronic Health Evaluation (APACHE) II score in the first 24 hours following onset of BSI; hospitalization wards; nosocomial infection or not; previous exposures (before hospital stay, previous treatment such as surgical procedures, parenteral nutrition, mechanical ventilation, renal replacement therapy, blood transfusion); and outcomes (intensive care unit [ICU] and hospital length of stay, days of mechanical ventilation after BSI onset, occurrence of septic shock and 28-day mortality). Microbiological data were also collected, including the likely source of BSI and sensitivity to antibiotics.

2.3. Species identification and antibiotic sensitivity testing

An automated BacT/ALERT 3D system (Becton-Dickinson, Sparks, MD) was used to culture blood. Species identification was performed using Bruker Daltonics Data Analysis. As recommended by Clinical & Laboratory Standards Institute, antibiotic susceptibility testing was performed using the VITEK 2 (Card number: AST-GN16; AST-GP67) system or the Kirby-Bauer disk diffusion method (Oxoid, Hampshire, UK).

2.4. Definitions

Onset of BSI was defined as the time when the first positive blood culture was obtained. Diagnosis of Pa-BSI was based on the Centers for Disease Control (CDC) definition of BSI.14 Polymicrobial Pa-BSI was defined as simultaneous isolation of P. aeruginosa and one or more other organisms from blood cultures. Nosocomial BSI was defined as a BSI developing ≥48 hours after hospitalization.15 Appropriate empirical antimicrobial therapy was defined as treatment performed within 24 hours of obtaining blood culture samples, including use of any antimicrobial agent to which P. aeruginosa and other copathogens were susceptible. Neutropenia was defined as absolute neutrophil counts of 1000/mm3 or below when bacteremia occurred. Sepsis and septic shock were defined using the new Sepsis-3 definition.16 Primary BSIs were BSI events without another primary source of infection (as defined by National Healthcare Safety Network [NHSN] criteria). Secondary BSI was defined as a BSI thought to be seeded from a site-specific infection at another body site.14

2.5. Statistical analysis

SPSS 26.0 (IBM Corp, Armonk, NY) software was used to perform statistical analysis. In the case of continuous variables with a normal distribution, the mean and standard deviation are presented, whereas the median and interquartile range are presented in the case of nonnormal distributions. Continuous variables were compared by Student’s t test or the Mann-Whitney U test, and enumeration variables were compared by Pearson χ2 or Fisher exact test, where appropriate. Variables that had significance at a p < 0.05 level in univariate analysis were considered candidates for the building of stepwise logistic regression multivariate models. A two-tailed p < 0.05 was considered statistically significant.

3. RESULTS 3.1. Demographic and clinical characteristics

A total of 661 blood culture specimens containing P. aeruginosa were first enrolled, and 278 cases were eventually recruited, including 77 cases of polymicrobial Pa-BSI and 201 cases of monomicrobial Pa-BSI (Fig. 1). Table 1 summarizes the clinical and demographic characteristics of all the patients. There were 204 (73.4%) males, and the mean age was 58.19 years. The most common comorbidity was trauma (32.4%), followed by solid tumor (17.6%). There was no significant difference between the two groups in age or sex. In patients with polymicrobial Pa-BSIs, the proportion of those with a history of a cerebrovascular accident (CVA) was significantly higher than that in patients with monomicrobial Pa-BSI (p < 0.05). There was no significant difference between the two groups in other comorbidities. Compared with patients with monomicrobial Pa-BSI, patients with polymicrobial Pa-BSI presented higher SOFA scores (7.0 [4-11] vs 5 [3-9]), higher APACHE II scores (16 [12-25] vs 16 [11-20]), higher Pitt bacteremia scores (5 [3-6.5] vs 3 [1-6]), higher rates of admission to the ICU (74% vs 59.7%), more use of mechanical ventilation (75.3% vs 52.5%), and more nosocomial infections (97.4% vs 88.1%) (all p < 0.05).

Table 1 - Demographic and clinical characteristics of the patients with polymicrobial Pa-BSIs compared with monomicrobial Pa-BSIs Characteristics Total (n = 278) Monomicrobial Pa-BSI (n = 201) Polymicrobial Pa-BSI (n = 77) p Age, y (mean ± SD) 58.19 ± 16.41 59.31 ± 15.99 55.27 ± 17.22 0.07 Male sex 204 (73.4%) 145 (72.1%) 59 (76.6%) 0.45 Co-morbidities  Diabetes mellitus 33 (11.9%) 23 (11.4%) 10 (13.0%) 0.72  Chronic kidney disease 21 (7.6%) 15 (7.5%) 6 (7.8%) 0.93  Chronic liver disease 4 (1.4%) 3 (1.5%) 1 (1.3%) 1  COPD or severe asthma 11 (4%) 8 (4%) 3 (3.9%) 1  Chronic cardiac insufficiency 22 (7.9%) 18 (9.0%) 4 (5.2%) 0.46  Solid tumor 49 (17.6%) 34 (16.9%) 15 (19.5%) 0.62  Trauma 90 (32.4%) 65 (32.3%) 25 (32.5%) 0.98  Burn injury 44 (15.8%) 30 (14.9%) 14 (18.2%) 0.51  Cerebrovascular accident 31 (11.2%) 16 (8%) 15 (19.5%) <0.01  Long-term corticoid treatment 7 (2.5%) 6 (3.0%) 1 (1.3%) 0.42  HIV or immunosuppressed status 6 (2.2%) 5 (2.5%) 1 (1.3%) 0.54  Hematological diseases 13 (4.7%) 11 (5.5%) 2 (2.6%) 0.31 CCI, median (IQR) 3 (1-5) 3 (1-5) 3 (1-4) 0.50 APACHE II score, median (IQR) 17 (11-21) 16 (11-20) 18 (12-25) 0.02 SOFA score, median (IQR) 6 (3-9.25) 5 (3-9) 7 (4-11) 0.03 Pitt bacteremia score, median (IQR) 4 (1-6) 3 (1-6) 5 (3-6.5) <0.01 Hospitalization ward  ICU 177 (63.7%) 120 (59.7%) 57 (74%) 0.03 Previous treatment  Hyperalimentation 138 (49.6%) 100 (49.8%) 38 (49.4%) 0.95  Mechanical ventilation 163 (58.6%) 105 (52.2%) 58 (75.3%) <0.01  Antibiotic exposure 267 (99.3%) 191 (99%) 76 (100%) 0.37   Use of carbapenems 115 (41.4%) 78 (38.8%) 37 (48.1%) 0.16  Surgery 133 (47.8%) 90 (44.8%) 43 (55.8%) 0.10  Chemotherapy/radiation 6 (2.2%) 5 (2.5%) 1 (1.3%) 0.54  Renal replacement therapy 51 (18.3%) 34 (16.9%) 17 (22.1%) 0.32  Blood transfusion 85 (30.6%) 58 (28.9%) 27 (35.1%) 0.32 Invasive devices  Central line 194 (69.8%) 140 (69.7%) 54 (70.1%) 0.94  Indwelling urinary catheter 209 (75.2%) 145 (72.1%) 64 (83.1%) 0.06  Intraperitoneal drainage 36 (12.9%) 25 (12.9%) 10 (13%) 0.99 Prior hospital stay, median days (IQR) 13 (8-26) 13 (6.5-25) 16 (10-27) 0.17 Nosocomial infection 252 (90.6%) 117 (88.1%) 75 (97.4%) 0.02 Neutropenia 20 (7.2%) 17 (8.5%) 3 (3.9%) 0.30

Bold indicates p < 0.05.

APACHE = Acute Physiology, Age and Chronic Health Evaluation; CCI = Charlson Comorbidity Index; COPD = chronic obstructive pulmonary disorder; ICU = intensive care unit; IQR = interquartile range; Pa-BSI = Pseudomonas aeruginosa bloodstream infections; SOFA = sequential organ failure assessment.

Fig. 1:

Flowchart of study participant enrollment. Pa-BSI = Pseudomonas aeruginosa bloodstream infection.

3.2. Biological indicators

Table 2 compares biological indicators between polymicrobial and monomicrobial Pa-BSIs. In comparison to patients with monomicrobial Pa-BSI, patients with polymicrobial Pa-BSI had worse liver function, as evidenced by significant increases in glutamic-oxaloacetic transaminase (median, 45 vs 35, p < 0.01) and lactic dehydrogenase (median, 299 vs 244, p ≤ 0.01). However, there was no significant difference in procalcitonin between the two groups or for any of the other biochemical parameters assessed.

Table 2 - Comparison of biological indicators between polymicrobial Pa-BSI and monomicrobial Pa-BSI Biological indicators Total (n = 278) Monomicrobial Pa-BSI (n = 201) Polymicrobial Pa-BSI (n = 77) p Blood routine test  WBC (×109/L), median (IQR) 10.5 (6.5-15.68) 10.7 (6.25-16.15) 9.3 (6.85-13.40) 0.40  Hematocrit (%), median (IQR) 26.5 (21.48-31.6) 26.9 (22.15-32) 23.7 (21.15-29) 0.06  Platelet (×109/L), median (IQR) 151.5 (72.75-239.5) 153 (87-245.5) 136 (49.5-233) 0.18  ANC, median (IQR) 9.05 (5.2-13.93) 9.5 (5.2-15) 7.8 (5.2-11.85) 0.30 Liver and kidney function  Albumin, g/L (mean ± SD) 31.04 ± 5.52 31.17 ± 5.71 30.70 ± 5.03 0.53  GPT (U/L), median (IQR) 33 (20-59) 32 (19-56) 37 (21.5-62) 0.28  GOT (U/L), median (IQR) 38 (22-62) 35 (21-57) 45 (29.5-83) <0.01  ALP (U/L), median (IQR) 121 (98-157) 121 (101.5-162.5) 121 (86.5-131.5) 0.04  γ-GT (U/L), median (IQR) 42.5 (23-87.25) 48 (23-96) 31 (22-67.5) 0.14  LDH (U/L), median (IQR) 251.5 (203-391.5) 244 (195-363) 299 (221.5-545) <0.01  TBil, µmol/L, median (IQR) 18.55 (11.05-35.05) 17 (10.75-35.1) 21.3 (11.6-34.8) 0.45  SCr, µmol/L, median (IQR) 62.5 (44-96.25) 59 (44-93) 68 (44-123.5) 0.18  PCT, ng/mL, median (IQR) 1.3 (0.41-6.61) 1.2 (0.35-5.83) 1.71 (0.55-7.49) 0.21

Bold indicates p < 0.05.

γ-GT = gamma glutamyl transpeptidase; ALP = alkaline phosphatase; ANC = absolute neutrophil count; GOT = glutamic-oxaloacetic transaminase; GPT = glutamic-pyruvic transaminase; IQR = interquartile range; LDH = lactic dehydrogenase; Pa-BSI = Pseudomonas aeruginosa bloodstream infections; PCT = procalcitonin; SCr = serum creatinine; TBil = total bilirubin; WBC = white blood count.

3.3. Bacteriology, sources of polymicrobial Pa-BSI and antibiotic resistance

Blood culture results for 77 polymicrobial Pa-BSIs showed that gram-negative bacteria accounted for 77.4% of the cases, with Acinetobacter baumannii as the main copathogen, followed by Klebsiella pneumoniae; Enterococcus was the main gram-positive copathogen (Table 3). Among patients with polymicrobial BSI, three kinds of microorganisms, including P. aeruginosa, were isolated from blood cultures of three patients, and four were isolated from one patient.

Table 3 - Characteristics of 82 copathogens isolated in 77 cases of polymicrobial BSI Microorganisms Polymicrobial BSI cases (n = 77) Number of agents (associated with Pa-BSI)  1 73 (94.8%)  2 3 (3.9%)  3 1 (1.3%) Gram-positive bacteria 18 (22%)  Coagulase-negative Staphylococci 3 (3.7%)   Staphylococcus aureus 5 (6.1%)   Enterococcus sp. 9 (11%)   S. agalactiae 1 (1.2%) Gram-negative bacteria 61 (74.4%)   Acinetobacter baumannii 25 (30.5%)   Klebsiella pneumoniae 21 (25.6%)   Enterobacter sp. 8 (9.8%)   Serratia sp. 7 (8.5%) Fungus 2 (2.4%)   Anaerobic bacteria 1 (1.2%)

BSI = bloodstream infections; Pa-BSI = Pseudomonas aeruginosa bloodstream infections; S. agalactiae = Streptococcus agalactiae.

Regarding the source of infection, pneumonia (36.3%, 101/278) was the main source of Pa-BSI, followed by skin/soft tissue infection (16.5%, 46/278) and primary BSI (16.5%, 46/278). Compared with monomicrobial Pa-BSI, the main source of polymicrobial Pa-BSI was pneumonia (49.4% vs 31.3%, p < 0.01), whereas the main source of monomicrobial Pa-BSI was primary BSI (21.9% vs 2.6%, p = 0.04) (Table 4). Details of antimicrobial resistance and antimicrobial therapy are shown in Table 4. Pseudomonas aeruginosa showed up to 62% resistance to imipenem. We did not find significant differences in drug resistance between the two groups. In addition, only 3.6% (10/349) of patients did not receive appropriate therapy within 24 hours after the release of antibiotic susceptibility results, with no difference between the two groups (3.0% vs 5.2%, p = 0.38) (Table 4).

Table 4 - Comparison of the microbiological characteristics with monomicrobial Pa-BSI and polymicrobial Pa-BSI Total (n = 278) Monomicrobial Pa-BSI (n = 201) Polymicrobial Pa-BSI (n = 77) p Source of BSIs  Intra-abdominal 26 (9.4%) 19 (9.5%) 7 (9.1%) 0.93  Primary BSI 46 (16.5%) 44 (21.9%) 2 (2.6%) <0.01  Pneumonia 101 (36.3%) 63 (31.3%) 38 (49.4%) <0.01  Skin and soft tissue infection 46 (16.5%) 30 (14.9%) 16 (20.8%) 0.24  Central venous catheter 27 (9.7%) 18 (9%) 9 (11.7%) 0.49  Urinary tract infection 15 (5.4%) 13 (6.5%) 2 (2.6%) 0.25  Intracranial 1 (0.4%) 1 (0.5%) 0 (0%) 1  Biliary 13 (4.3%) 11 (5.5%) 2 (2.6%) 0.53  Joint 3 (1.1%) 2 (1.0%) 1 (1.3%) 1 Antibiotic resistancea  Amikacin (200 vs 76)b 27 (9.8%) 19 (9.5%) 8 (10.5%) 0.80  Aztreonam (138 vs 55)b 76 (39.4%) 52 (37.7%) 24 (43.6%) 0.45  Ciprofloxacin (201 vs 77)b 119 (42.8%) 87 (43.3%) 32 (41.6%) 0.80  Piperacillin/tazobactam (201 vs 77)b 109 (39.2%) 78 (38.8%) 31 (40.3%) 0.82  Gentamicin (201 vs 77)b 73 (26.3%) 53 (26.4%) 20 (26%) 0.97  Cefepime (201 vs 77)b 96 (34.5%) 69 (34.3%) 27 (35.1%) 0.91  Tobramycin (195 vs 72)b 79 (29.6%) 60 (30.8%) 19 (26.4%) 0.49  Levofloxacin (199 vs 76)b 103 (37.5%) 78 (39.2%) 25 (32.9%) 0.33  Imipenem (197 vs 75)b 171 (62.9%) 125 (63.5%) 46 (63.1%) 0.75 Treatment after the onset of BSIs  Delayed antibiotic therapy 10 (3.6%) 6 (3.0%) 4 (5.2%) 0.38  Proportion of appropriate empirical therapy 268 (96.4%) 195 (97.0%) 73 (94.8%) 0.38

Bold indicates p < 0.05.

BSI = bloodstream infections; Pa-BSI = Pseudomonas aeruginosa bloodstream infections.

aNot all agents listed tested in all isolates.

bThe numbers in parentheses represent the total numbers of Pseudomonas aeruginosa isolates performed susceptibility test.

3.4. Independent risk factors for polymicrobial Pa-BSI

Univariate analysis showed that comorbidities such as CVA, ICU admission, nosocomial infection, use of mechanical ventilation, and high SOFA, APACHE II, and Pitt bacteremia scores were associated with polymicrobial Pa-BSI. As shown in multivariate analysis (Table 5), a history of CVA (adjusted odds ratio [OR], 3.62; 95% CI, 1.47-8.92) was the only independent risk factor associated with polymicrobial Pa-BSI. Interestingly, primary BSI might be more associated with monomicrobial Pa-BSI (aOR, 0.08; 95% CI, 0.02-0.38).

Table 5 - Multivariable logistic regression of factors associated with polymicrobial Pseudomonas aeruginosa bloodstream infections Variable Unadjusted OR (95% CI) p Adjusted OR (95% CI) p Cerebrovascular accident 0.36 (0.17-0.77) <0.01 3.62 (1.47-8.92) <0.01 APACHE II score 1.04 (1.00-1.08) 0.03 SOFA score 1.08 (1.02-1.14) 0.01 Pitt bacteremia score 1.19 (1.08-1.30) <0.01 ICU stay 0.52 (0.29-0.93) 0.03 Pneumonia 2.13 (1.25-3.65) <0.01 Primary BSI 0.10 (0.02-0.40) <0.01 0.08 (0.02-0.38) <0.01 Nosocomial infection 0.20 (0.05-0.85) 0.03 Mechanical ventilation