Abstract
Background
Appropriate antibiotic therapy is a key determinant of outcomes in Pseudomonas aeruginosa bloodstream infection (BSI), but the impact of adequate source control on prognosis has not been well defined in large cohorts.
Methods
We conducted a retrospective multicenter study including hospitalized patients with P. aeruginosa BSI across 14 Italian hospitals between January 2021 and December 2022.
Results
Overall, 639 BSI episodes were analyzed. The most frequent sources were an unknown source (31.8%), central venous catheter BSI (22.5%), hospital-acquired/ventilator-associated pneumonia (20.3%), urinary tract infections (11.4%), and intra-abdominal infections (6.9%). Source control was required in 34.0% of cases and was achieved early and adequately in 49.3% of these. The 30-day all-cause mortality rate was 23.9%. Mortality was significantly lower in patients who received early adequate source control (6.5%) compared with those without early adequate source control (23.6%) or in whom it was deemed unnecessary (28.4%) (P < .0001). In multivariable analysis, age, corticosteroid therapy, and septic shock were independently associated with increased mortality, whereas prior P. aeruginosa colonization, early adequate source control, and appropriate antibiotic therapy were protective. A propensity score–matched analysis confirmed the association between early adequate source control and reduced 30-day mortality (adjusted hazard ratio 0.15; P = .0024).
Conclusions
Early adequate source control is a critical determinant of survival in P. aeruginosa BSI. Outcomes are optimized by a combined approach integrating timely source control with appropriate antibiotic therapy.
Keywords: bloodstream infections, mortality, Pseudomonas aeruginosa, source control
Adequate source control, along with appropriate antimicrobial therapy, is a key determinant of clinical outcomes in severe infections [1], including intravascular catheter-related [2], intra-abdominal [3], and urinary tract infections [4]. Multiple studies have underscored its prognostic value in bloodstream infections (BSIs) caused by a broad range of pathogens, such as Gram-positive (eg, Staphylococcus aureus, Streptococcus spp.) [5, 6] and Gram-negative bacteria (eg, Enterobacterales) [7], as well as fungi (eg, Candida spp.) [8].
Pseudomonas aeruginosa is a major nosocomial pathogen [9] and ranks as the third most common cause of hospital-acquired Gram-negative BSI [10], with mortality rates reaching up to 60% [11–16]. While the importance of timely and appropriate antimicrobial therapy has been well established [17, 18], the potential impact of early source control in P. aeruginosa BSI remains largely underexplored. To date, only a few small, single-center studies have specifically addressed this issue [19, 20], and none have systematically evaluated the interplay between appropriate antimicrobial therapy, adequate source control, and 30-day all-cause mortality.
This study aims to address this knowledge gap by identifying predictors of clinical outcomes in patients with P. aeruginosa BSI, with a particular focus on the impact of adequate source control on 30-day all-cause mortality.
METHODS
This multicenter retrospective cohort study was conducted across 14 public hospitals in 8 Italian regions (Liguria, Piedmont, Lombardy, Veneto, Emilia-Romagna, Lazio, Apulia, and Sicily) over a 2-year period (2021–2022). The study was approved by the Ethics Committee of the City of Health and Science of Turin (Protocol No. 202/2023, PROT. No. 0066633).
The study population was identified using computerized microbiology laboratory databases from all participating centers. Inclusion criteria consisted of adult (≥ 18 years old) patients with at least one peripheral blood culture positive for P. aeruginosa. Exclusion criteria included polymicrobial BSI and recurrent P. aeruginosa BSI, with only the first episode per patient considered in the analysis.
Data Collection
Baseline characteristics and process-of-care variables were retrieved from the automated hospital medical records, microbiology databases, and pharmacy databases of each participating hospital. Baseline characteristics included: age, gender, hospital ward at the time of P. aeruginosa BSI onset and underlying diseases, summarized using the Charlson Comorbidity Index [21]. Additional variables considered were immunosuppression (chemotherapy, corticosteroid therapy or other immunosuppressive therapy prior to BSI onset), neutropenia (absolute neutrophil count of <500/µL at the time of BSI onset), and invasive procedures performed within 30 days before P. aeruginosa BSI. These procedures included central venous catheter (CVC) placement, urinary catheterization, any type of surgery, and percutaneous endoscopic gastrostomy. Infection-related variables included previous P. aeruginosa colonization within 90 days, infection source, antimicrobial susceptibility profile of the P. aeruginosa isolate, severity of the disease at clinical presentation, antibiotic therapy, adequacy of source control, intensive care unit (ICU) admission after P. aeruginosa BSI onset, need for continuous renal replacement therapy (CRRT) after P. aeruginosa BSI onset, and 30-day all-cause mortality.
Endpoint
The primary endpoint was 30-day all-cause mortality following the first positive blood culture for P. aeruginosa.
Exposure Variable
The primary exposure variable was early adequate source control.
Definition of Infection Source
The source of P. aeruginosa BSI was classified by the investigators at each participating center during data collection, before the statistical analysis, according to predefined study criteria and based on the clinical, microbiological, and imaging information available in the medical records. CVC-related BSI was defined by at least one of the following: (1) presence of exudate at the CVC exit site with isolation of P. aeruginosa (same phenotypic strain) from both the exudate and bloodstream; (2) a semi-quantitative catheter tip culture yielding >15 colony-forming units of the same P. aeruginosa strain isolated from blood, or (3) a differential time to positivity of ≥2 hours between blood cultures drawn from the CVC and a peripheral vein. The abdomen was considered to be the origin of the P. aeruginosa BSI when a patient had clinical and radiological evidence of abdominal infection and (1) a positive culture from an intra-abdominal site was obtained during surgery or needle aspiration or (2) no other apparent sources of P. aeruginosa BSI were detected. Urinary source was identified by P. aeruginosa isolation from urine culture or affected tissue, along with urologic conditions such as urinary tract manipulation or obstruction. Hospital-acquired pneumonia (HAP) or ventilator-associated pneumonia (VAP)–associated BSI was diagnosed when respiratory cultures yielded the same phenotypic strain, in the presence of clinical and radiological findings consistent with pneumonia [22]. BSI of unknown source was defined as a BSI for which no infectious source was identified by the participating investigators. This classification did not presume a specific pathogenesis and could therefore include neutropenic patients with possible gastrointestinal translocation, as well as patients with other unidentified sources.
Early Adequate Source Control
The need for source control was classified retrospectively at each participating center according to the predefined criteria detailed below. All decisions regarding whether to perform source control had been made by the treating physicians during routine clinical care, independently of the study, based on the patient's clinical condition, procedural feasibility, and local practice. Source control was deemed necessary in clinical scenarios including the following: (1) removal of infected hardware (eg, endovascular prostheses, CVC) in BSIs associated with intravascular devices; (2) resolution of urinary tract obstruction in patients with urinary tract-related BSI; (3) surgical or image-guided drainage of infected collections and/or debridement of necrotic material in patients with intra-abdominal infections (eg, abscess, peritonitis), complicated pneumonia (eg, pneumonia with abscess or pleural empyema), and/or wound infections; (4) joint fluid drainage via arthrotomy or arthroscopy in cases of septic arthritis and/or osteomyelitis; and (5) cardiac surgery for endocarditis when indicated by current guidelines [23]. Source control was considered unnecessary in cases of BSI of unknown source or uncomplicated pneumonia. In all other cases, source control was deemed necessary. The timing of source control was defined as the interval between BSI onset and the performance of a definitive infection-resolving procedure. Early adequate source control was defined as intervention within 72 hours of P. aeruginosa BSI onset.
Other Definitions
Bloodstream infection onset was defined as the date on which the first positive blood culture for P.aeruginosa was collected. Appropriate antibiotic therapy was defined as the administration of at least one antimicrobial agent at an adequate dosage with confirmed in vitro activity against the infecting isolate. The timing of appropriate antibiotic therapy was defined as the interval between P. aeruginosa BSI onset and the first intravenous dose of an active antimicrobial agent. Severity of illness was assessed by the local investigators at each participating center through review of the medical records, based on the presence of severe sepsis or septic shock at P. aeruginosa BSI onset [24]. The susceptibility profile of P. aeruginosa was classified as multidrug resistant (MDR), extensively drug resistant (XDR), or pandrug resistant (PDR), according to the standardized international definitions proposed by Magiorakos et al [25].
Microbiological Studies
Blood cultures, organism identification, and susceptibility testing were performed at each participating center following standard operating procedures. Antimicrobial susceptibility was interpreted by local laboratories according to the European Committee on Antimicrobial Susceptibility Testing (EUCAST) recommendations [26].
Statistical Analysis
Baseline demographic and clinical characteristics, as well as subsequent outcomes, were compared between patients who died within 30 days of P. aeruginosa BSI onset and those who survived. Categorical variables were expressed as counts and percentages, while continuous variables were reported as medians with interquartile ranges (IQRs). Comparisons were performed using the χ2 test or Fisher's exact test for categorical variables and the Wilcoxon rank sum test for continuous variables.
The impact of variables on 30-day all-cause mortality was evaluated through Cox proportional hazards regression, estimating both unadjusted hazard ratios (HRs) and multivariate-adjusted HRs (aHRs). Variables with P < .1 in unadjusted models were included in the multivariate analysis, along with characteristics associated with source control status (source control deemed not necessary, early adequate source control, inadequate source control), as identified through 3-group comparison. To identify independent predictors of 30-day all-cause mortality, a backward stepwise selection was then performed (model A). The same set of variables initially included in the backward stepwise selection for model A was re-entered into a new regression model, replacing fixed source control status with time-dependent source control to better address immortal time bias; a new backward selection was then performed to construct model B. In contrast, the set of variables retained in model A was used to build 3 additional models: model C, which applied a landmark analysis restricting inclusion to patients who were alive at day 4 after BSI onset (chosen because early adequate source control is defined within 72 hours), also to mitigate immortal time bias; model D, which included the study center as a shared frailty term to account for center-level heterogeneity; and model E, which evaluated source control (early adequate source control vs not), regardless of whether it was deemed necessary, as an additional robustness analysis. In addition, day-4 landmark Kaplan–Meier survival curves, including only patients who were alive at day 4 after P.aeruginosa onset, were generated to visually compare unadjusted 30-day survival between patients, stratified by source control status. Differences between the 3 groups were assessed using the log-rank test.
A sensitivity analysis restricted to patients with source control deemed necessary was conducted using the same methodology, to specifically address the subgroup for whom source control was indicated. Additionally, as part of the sensitivity analysis, the effect of early adequate source control on 30-day all-cause mortality was assessed through a propensity score (PS)–matching analysis. The PS was defined as the probability of receiving early adequate source control; it was estimated using logistic regression, including all variables with a P value of <.15 in the unadjusted comparison between patients with and without early adequate source control, as well as septic shock at BSI onset. Patients were matched 1:1 using nearest neighbor matching without replacement and a caliper of 0.1. Covariate balance was assessed using standardized differences. In an exploratory analysis, mortality was also compared between patients who received appropriate antibiotic therapy and those who did not, stratified by early source control adequacy, in the subgroup with source control deemed necessary. To reduce immortal time bias, only patients who were alive at day 4 after BSI onset were included.
Time to adequate source control, ICU admission, need for CRRT after P. aeruginosa BSI onset, and time to appropriate antibiotic therapy were treated as time-dependent covariates in the Cox regression models. The proportional hazards assumption was verified for all fixed-effect Cox regression models, using cumulative sums of martingale residuals and Kolmogorov-type supremum tests. Missing data among baseline variables were limited to 1 missing age value, which was imputed using the cohort median. A P value of ≤.05 was considered statistically significant, and all tests were 2-sided. Statistical analyses were conducted using SPSS version 28.0.1.0 (SPSS Software, Chicago, IL) and SAS software version 9.4 (SAS Institute Inc., Cary, NC).
RESULTS
During the study period, 639 episodes of P. aeruginosa BSI were identified. Bloodstream infection of unknown source was the most frequent category, accounting for 203 episodes (31.8%), followed by CVC-related BSI (144 cases, 22.5%), HAP/VAP-related BSI (130 cases, 20.3%), urinary tract infection (73 cases, 11.4%), and intra-abdominal infection (44 cases, 6.9%). Among patients with BSI of unknown source, 36 (17.7%) were neutropenic. Their clinical characteristics are detailed in Supplementary Table 1.
Overall, source control was deemed necessary in 217 patients (34.0%) and was performed adequately and early in 107 of these cases (49.3%). The median time from P. aeruginosa BSI onset to adequate source control was 3 days (IQR 1–5). The baseline characteristics of the 3 patient groups are reported in Supplementary Table 2.
Table 1 provides an overview of the entire cohort, including a comparison between patients who died and those who survived. Among the 639 patients, 153 (23.9%) died within 30 days. The unadjusted 30-day all-cause mortality rate was significantly lower in patients who received early adequate source control (6.5%), while no significant difference was observed between those who did not undergo early source control (23.6%) and those for whom source control was deemed not necessary (28.4%) (Supplementary Table 2). Univariate and multivariate analyses of factors associated with 30-day all-cause mortality are presented in Table 2. The backward stepwise multivariable analysis (model A) identified age (aHR 1.03; CI 1.02–1.04, P < .0001), corticosteroid therapy (aHR 1.99; CI 1.41–2.81, P = .0001) and septic shock at P. aeruginosa BSI onset (aHR 3.85; CI 2.77–5.35, P < .0001) as factors independently associated with higher mortality. Conversely, previous history of P. aeruginosa colonization (aHR 0.60; CI .40–.90, P = .0142), early adequate source control (aHR 0.26; CI .11–.60, P = .0016), and appropriate antibiotic therapy (aHR 0.27; CI .18–.40, P < .0001) were factors found to be protective against mortality. To account for potential immortal time bias affecting source control, models B and C were also performed. In particular, landmark analysis (model C) confirmed age (aHR 1.03; CI 1.01–1.05, P = .0005), septic shock (aHR 2.44; CI 1.51–3.96, P = .0003), prior P. aeruginosa colonization (aHR 0.52; CI .29–.91, P = .0227), early adequate source control (aHR 0.35; CI .14–.90, P = .0293), and appropriate antibiotic therapy (aHR 0.41; CI .21–.78, P = .0067) as significant predictors of 30-day all-cause mortality (Table 3; Figure 1). Including center as shared frailty did not alter the results of the model. Similarly, model E evaluating early adequate source control regardless of whether it was deemed necessary, confirmed the robustness of the main findings, as reported in Table 3.
Table 1.
Demographic and Clinical Characteristics of the Entire Population With Pseudomonas aeruginosa BSI (N = 639), with a Comparison Between Alive and Dead Patients for All-Cause Mortality at 30 day
| Variables | Total n = 639 |
Alive n = 486 |
Dead n = 153 |
P Value |
|---|---|---|---|---|
| Age (y), median (IQR) | 68 (57–78) | 66 (55–76) | 72 (65–79) | <.0001 |
| Male sex, n (%) | 431 (67.5) | 334 (68.7) | 97 (63.4) | .2202 |
| Hospital ward at the time of P. aeruginosa BSI onset, n (%) | .0086 | |||
| Internal medicine | 279 (43.7) | 219 (45.1) | 60 (39.2) | |
| Surgery | 100 (15.7) | 82 (16.9) | 18 (11.8) | |
| Oncology-hematology | 71 (11.1) | 56 (11.5) | 15 (9.8) | |
| ICU | 183 (28.6) | 127 (26.1) | 56 (36.6) | |
| Others | 6 (0.9) | 2 (0.4) | 4 (2.6) | |
| Underlying diseases, n (%) | ||||
| Cardiovascular disease | 207 (32.4) | 147 (30.3) | 60 (39.2) | .0387 |
| Neurological disease | 179 (28.0) | 134 (27.6) | 45 (29.4) | .6585 |
| Diabetes mellitus | 125 (19.6) | 93 (19.1) | 32 (20.9) | .6285 |
| Solid malignancy | 133 (20.8) | 95 (20.0) | 38 (24.8) | .1599 |
| Gastrointestinal disease | 110 (17.2) | 80 (16.5) | 30 (19.6) | .3685 |
| Hematological malignancy | 104 (16.3) | 82 (16.9) | 22 (14.4) | .4662 |
| Chronic renal failure | 99 (15.5) | 75 (15.4) | 24 (15.7) | .9396 |
| Chronic obstructive pulmonary disease | 74 (11.6) | 53 (10.9) | 21 (13.7) | .3418 |
| Hematopoietic stem cell transplant | 35 (5.5) | 29 (6.0) | 6 (3.9) | .3322 |
| Solid organ transplant | 28 (4.4) | 23 (4.7) | 5 (3.3) | .4402 |
| Liver cirrhosis | 21 (3.3) | 17 (3.5) | 4 (2.6) | .5929 |
| Splenectomy | 10 (1.6) | 8 (1.7) | 2 (1.3) | 1.0000g |
| HIV | 8 (1.3) | 7 (1.4) | 1 (0.7) | .6873g |
| Charlson comorbidity index, median (IQR) | 4 (2–6) | 3 (2–5) | 4 (2–6) | .0082 |
| Corticosteroid therapy, n (%) | 142 (22.2) | 93 (19.1) | 49 (32.0) | .0008 |
| Chemotherapy, n (%) | 82 (12.8) | 60 (12.4) | 22 (14.4) | .5119 |
| Other immunosuppressive therapy, n (%) | 72 (11.3) | 59 (12.1) | 13 (8.5) | .2139 |
| Neutropenia, n (%) | 70 (11.0) | 51 (10.5) | 19 (12.4) | .5062 |
| Intermittent hemodialysis, n (%) | 36 (5.6) | 30 (6.2) | 6 (3.9) | .2922 |
| Invasive proceduresd, n (%) | ||||
| Central venous catheter | 411 (64.3) | 319 (65.6) | 92 (60.1) | .2149 |
| Urinary catheter | 419 (65.6) | 311 (64.0) | 108 (70.6) | .1342 |
| Any surgery | 172 (26.9) | 133 (27.4) | 39 (25.5) | .6482 |
| Percutaneous endoscopic gastrostomy | 40 (6.3) | 32 (6.6) | 8 (5.2) | .546 |
| Previous P. aeruginosa colonizatione, n (%) | 171 (26.8) | 141 (29.0) | 30 (19.6) | .0219 |
| Source of bloodstream infectionb, n (%) | <.0001g | |||
| Unknown source | 203 (31.8) | 138 (28.4) | 65 (42.5) | |
| CVC-related BSI | 144 (22.5) | 124 (25.5) | 20 (13.1) | |
| HAP/VAP | 130 (20.3) | 89 (18.3) | 41 (26.8) | |
| Urinary tract infections | 73 (11.4) | 67 (13.8) | 6 (3.9) | |
| Intra-abdominal infections | 44 (6.9) | 31 (6.4) | 13 (8.5) | |
| Skin and soft tissue infection | 26 (4.1) | 23 (4.7) | 3 (2.0) | |
| Osteomyelitis | 3 (0.5) | 3 (0.6) | 0 (0.0) | |
| Central nervous system infection | 2 (0.3) | 1 (0.2) | 1 (0.7) | |
| Othersf | 14 (2.2) | 10 (2.1) | 4 (2.6) | |
| Susceptibility profilec, n (%) | .9074 | |||
| MDR | 57 (9.1) | 42 (8.7) | 15 (10.2) | |
| XDR | 40 (6.4) | 30 (6.2) | 10 (6.8) | |
| PDR | 7 (1.1) | 6 (1.2) | 1 (0.7) | |
| Clinical presentation, n (%) | <.0001 | |||
| Sepsis | 312 (48.8) | 245 (50.4) | 67 (43.8) | |
| Septic shock | 129 (20.2) | 65 (13.4) | 64 (41.8) | |
| No sepsis | 198 (31.0) | 176 (36.2) | 22 (14.4) | |
| Appropriate antibiotic therapy, n (%) | 578 (90.5) | 459 (94.4) | 119 (77.8) | <.0001 |
| Source control status, n (%) | <.0001 | |||
| Deemed unnecessary | 422 (66.0) | 302 (62.1) | 120 (78.4) | |
| Deemed necessary but not performed early | 110 (17.2) | 84 (17.3) | 26 (17.0) | |
| Deemed necessary and adequately performed early | 107 (16.7) | 100 (20.6) | 7 (4.6) | |
| Time to source control (d)a, median (IQR) | 3 (1–5) | 3 (1–5) | 2.5 (1–4) | .4906 |
| Outcomes, n (%) | ||||
| Need for ICU admission | 23 (3.6) | 14 (2.9) | 9 (5.9) | .0822 |
| Need for CRRT after P. aeruginosa BSI | 20 (3.1) | 9 (1.9) | 11 (7.2) | .0023 |
The reported P values are derived from the Wilcoxon rank sum test for continuous variables and χ2 or Fisher's exact test for categorical variables. Bold values indicate statistical significance at the predefined level (α = .05). One patient alive at 30 d had missing age data, which was imputed using the median for subsequent analyses. The median, first and third quartiles, and P value reported in the table were identical before and after imputation. No other missing values were present in this table.
Abbreviations: BSI, bloodstream infection; CRRT, continuous renal replacement therapy; CVC, central venous catheter; ICU, intensive care unit; IQR, interquartile range; HAP, hospital-acquired pneumonia; HIV, human immunodeficiency virus; MDR, multidrug resistant; PDR, pandrug-resistant; VAP, ventilator-associated pneumonia; XDR, extensively drug resistant.
aAnalysis conducted on a total of 172 patients who underwent source control during the study period.
bTest calculated using Monte Carlo simulation.
cAnalysis conducted on a total of 629 patients with a known susceptibility profile.
dWithin 30 d before P. aeruginosa BSI onset.
eWithin 90 d before P. aeruginosa BSI onset.
fIncludes mediastinitis, endocarditis, infection of spinal fixation hardware, community-acquired pneumonia, infection of prosthetic femur, lung abscess, pleural empyema, left ventricular assist device (LVAD) infection.
gFisher's exact test.
Table 2.
Univariable, Multivariable, and Backward Selection Multivariable Analyses of Factors Associated With 30-d All-Cause Mortality in the Entire Study Population (n = 639)
| Univariable | Multivariable | Multivariable Backward (Model A) |
||||
|---|---|---|---|---|---|---|
| Variables | HR (95% CI) | P Value | aHR (95% CI) | P Value | aHR (95% CI) | P Value |
| Age (y) | 1.03 (1.02–1.05) | <.0001 | 1.03 (1.01–1.04) | <.0001 | 1.03 (1.02–1.04) | <.0001 |
| Male sex | 0.80 (.57–1.11) | .1792 | 0.84 (.60–1.18) | .3192 | … | … |
| Cardiovascular disease | 1.40 (1.01–1.94) | .0424 | 1.20 (.84–1.70) | .3155 | … | … |
| Neurological disease | 1.07 (.76–1.52) | .6899 | … | … | … | … |
| Diabetes mellitus | 1.07 (.73–1.58) | .7296 | … | … | … | … |
| Solid malignancy | 1.36 (.94–1.96) | .1021 | … | … | … | … |
| Gastrointestinal disease | 1.26 (.85–1.89) | .2494 | … | … | … | … |
| Hematological malignancy | 0.91 (.58–1.42) | .6705 | … | … | … | … |
| Chronic renal failure | 1.10 (.71–1.71) | .6585 | … | … | … | … |
| Chronic obstructive pulmonary disease | 1.31 (.83–2.08) | .2519 | … | … | … | … |
| Hematopoietic stem cell transplant | 0.67 (.29–1.50) | .3266 | … | … | … | … |
| Solid organ transplant | 0.72 (.30–1.75) | .4682 | … | … | … | … |
| Liver cirrhosis | 0.87 (.32–2.34) | .7749 | … | … | … | … |
| Splenectomy | 0.80 (.20–3.24) | .7585 | … | … | … | … |
| HIVa | ||||||
| Charlson comorbidity index | 1.08 (1.02–1.13) | .0044 | 1.00 (.95–1.06) | .9862 | … | … |
| Corticosteroid therapy | 1.83 (1.30–2.57) | .0005 | 2.00 (1.41–2.86) | .0001 | 1.99 (1.41–2.81) | .0001 |
| Chemotherapy | 1.23 (.78–1.93) | .3710 | … | … | … | … |
| Other immunosuppressive therapy | 0.74 (.42–1.31) | .3031 | … | … | … | … |
| Neutropenia | 1.22 (.76–1.97) | .4165 | … | … | … | … |
| Intermittent dialysis treatment | 0.64 (.28–1.44) | .2797 | 0.51 (.22–1.17) | .1128 | … | … |
| Central venous cathetere | 0.73 (.53–1.02) | .0622 | 0.90 (.63–1.29) | .5686 | … | … |
| Urinary cathetere | 1.19 (.84–1.69) | .3253 | … | … | … | … |
| Any surgerye | 0.87 (.61–1.26) | .4674 | … | … | … | … |
| Percutaneous endoscopic gastrostomye | 0.70 (.34–1.43) | .3284 | … | … | … | … |
| Previous P. aeruginosa colonizationf | 0.57 (.38–0.85) | .0059 | 0.63 (.41–.96) | .0320 | 0.60 (.40–.90) | .0142 |
| Septic shock | 3.53 (2.56–4.87) | <.0001 | 4.18 (2.99–5.84) | <.0001 | 3.85 (2.77–5.35) | <.0001 |
| Appropriate antibiotic therapyc | 0.26 (.18–.38) | <.0001 | 0.29 (.20–.43) | <.0001 | 0.27 (.18–.40) | <.0001 |
| Time to appropriate antibiotic therapy (d)d | 0.89 (.76–1.03) | .1143 | … | … | … | … |
| Source control status (vs deemed necessary but not performed early) | … | .0001 | … | <.0001 | … | <.0001 |
| Deemed necessary and adequately performed early | 0.26 (.11–.61) | .0017 | 0.24 (.10–.55) | .0009 | 0.26 (.11–.60) | .0016 |
| Deemed unnecessary | 1.33 (.87–2.03) | .1864 | 1.46 (.95–2.25) | .0876 | 1.49 (.97–2.29) | .0656 |
| Time to source control (d)b | 0.90 (.73–1.12) | .3487 | … | … | … | … |
| Need for ICU admission | 2.13 (.79–5.76) | .1380 | … | … | … | … |
| Need for CRRT after P. aeruginosa BSIa | … | … | … | … | … | |
The reported P values are derived from Cox regression analysis. Bold values indicate statistical significance at the predefined level (α = .05). The variables hospital ward, source of bloodstream infection, and susceptibility profile were excluded from the analysis. The clinical presentation variable was replaced with the binary variable septic shock (Yes vs No). The variables time to source control, need for ICU admission, need for CRRT after P. aeruginosa BSI, and time to appropriate antibiotic therapy were treated as time-dependent. For 3 patients with missing ICU discharge dates, the hospital discharge date was used to define the end of ICU exposure.
Abbreviations: aHR, adjusted hazard ratio; BSI, bloodstream infection; CRRT, continuous renal replacement therapy; HIV, human immunodeficiency virus; HR, hazard ratio; ICU, intensive care unit; IQR, interquartile range.
aAnalysis not valid due to an insufficient number of events.
bAnalysis conducted on a total of 172 patients who underwent source control during the study period.
cVariable treated as fixed due to its clinical relevance, a median initiation time of 0 d (IQR: 0–1), the lack of a significant effect of time to appropriate antibiotic therapy (P = .1143), and the absence of interaction with time (P = .1525) in the Cox model, ensuring methodological consistency.
dAnalysis conducted on a total of 578 patients who received appropriate antibiotic therapy.
eWithin 30 d before P. aeruginosa BSI onset.
fWithin 90 d before P. aeruginosa BSI onset.
Table 3.
Multivariable Analyses of Factors Associated With 30-d All-Cause Mortality: Time-Dependent Source Control (Model B), 4-d Landmark Analysis (Model C, n = 562), Shared Frailty Model (Model D), and Model With Source Control Regardless of Whether It Was Deemed Necessary (Model E)
| Multivariable Backward (Model B) |
Landmark Analysis (Model C, n = 562) |
Shared Frailty Model (Model D) |
Model E | |||||
|---|---|---|---|---|---|---|---|---|
| Variable | aHR (95% CI) | P Value | aHR (95% CI) | P Value | aHR (95% CI) | P Value | aHR (95% CI) | P Value |
| Source control status (vs Deemed necessary but not performed early) | .0142 | <.0001 | ||||||
| Deemed necessary and adequately performed early | 0.35 (.14–.90) | .0293 | 0.25 (.11–.58) | .0013 | ||||
| Deemed unnecessary | 1.23 (.72–2.13) | .4519 | 1.56 (1.01–2.39) | .0439 | ||||
| Source control (vs Deemed necessary but not adequately performed)a | .0001 | |||||||
| Deemed necessary and adequately performed | 0.33 (.15–.73) | .0066 | ||||||
| Deemed unnecessary | 1.42 (.88–2.29) | .1513 | ||||||
| Deemed necessary and adequately performed source control (vs Not) | 0.19 (.09–.40) | <.0001 | ||||||
| Age (y) | 1.03 (1.02–1.05) | <.0001 | 1.03 (1.01–1.05) | .0005 | 1.03 (1.02–1.04) | <.0001 | 1.03 (1.02–1.04) | <.0001 |
| Corticosteroid therapy | 1.93 (1.32–2.83) | .0008 | 1.46 (.87–2.43) | .1484 | 2.03 (1.43–2.88) | <.0001 | 2.00 (1.41–2.82) | <.0001 |
| Previous P. aeruginosa colonization | — | — | 0.52 (.29–.91) | .0227 | 0.59 (.39–.89) | .0116 | 0.61 (.41–.92) | .0170 |
| Septic Shock | 3.32 (2.32–4.76) | <.0001 | 2.44 (1.51–3.96) | .0003 | 4.01 (2.87–5.59) | <.0001 | 3.83 (2.76–5.31) | <.0001 |
| Appropriate antibiotic therapyb | 0.30 (.19–.47) | <.0001 | 0.41 (.21–.78) | .0067 | 0.26 (.18–.39) | <.0001 | 0.28 (.19–.41) | <.0001 |
The reported P values are derived from Cox regression analysis. Bold values indicate statistical significance at the predefined level (α = .05). An empty cell indicates variables not included in the model, while a dash (—) indicates variables not selected by the model.
Abbreviations: aHR, adjusted hazard ratio; BSI, bloodstream infection; CRRT, continuous renal replacement therapy; HR, hazard ratio; ICU, intensive care unit.
aTreated as time-dependent, see Methods section.
bVariable treated as fixed due to its clinical relevance, a median initiation time of 0 d (IQR: 0–1), the lack of a significant effect of time to appropriate antibiotic therapy (P = .1143), and the absence of interaction with time (P = .1525) in the Cox model, ensuring methodological consistency.
Figure 1.
Landmark Kaplan–Meier curves including only patients who were alive at day 4 after bloodstream infection (BSI) onset, stratified by source control status (deemed unnecessary, early adequate, not early adequate).
Clinical Characteristics of Patients in Whom Source Control Was Deemed Necessary and Differences Between Those Who Underwent the Procedure and Those Who Did Not
Table 4 presents the clinical characteristics of the 217 patients for whom source control was deemed necessary. No statistically significant differences were observed between those who received early adequate source control (n = 107) and those who did not (n = 110), except for a higher prevalence of intermittent dialysis in patients who did not receive early adequate source control (10.9% vs 3.7%, P = .0433). In the Cox multivariable regression model (Table 5, model A2) 30-day all-cause mortality was associated with septic shock (aHR 6.16; CI 3.03–12.51, P < .0001). Conversely, early adequate source control (aHR 0.22; CI .09–.50, P = .0004) and male sex (aHR 0.45, CI .22–.90, P = .0240) were the only factors associated with improved outcomes. Similar results were observed in models B2, C2, and D2 (Supplementary Table 3).
Table 4.
Demographic and Clinical Characteristics of the Study Population With Source Control Deemed Necessary (n = 217); Comparison Between Patients Who Received Early Adequate Source Control and Those Who Did Not
| Variables | Total n = 217 |
Early Adequate Source Control n = 107 |
No Early Adequate Source Control n = 110 |
P Value |
|---|---|---|---|---|
| Age (y), median (IQR) | 67 (56–76) | 66 (55–74) | 69.5 (57–77) | .2021 |
| Male sex, n (%) | 132 (60.8) | 72 (67.3) | 60 (54.6) | .0545 |
| Hospital ward at the time of P. aeruginosa BSI onset, n (%) | … | … | … | .4920 |
| Internal medicine | 103 (47.5) | 47 (43.9) | 56 (50.9) | |
| Surgery | 27 (12.4) | 12 (11.2) | 15 (13.6) | |
| Oncology-Hematology | 23 (10.6) | 11 (10.3) | 12 (10.9) | |
| ICU | 63 (29.0) | 36 (33.6) | 27 (24.6) | |
| Others | 1 (0.5) | 1 (0.9) | 0 (0.0) | |
| Underlying diseases, n (%) | ||||
| Cardiovascular disease | 71 (32.7) | 34 (31.8) | 37 (33.6) | .7702 |
| Neurological disease | 55 (25.4) | 31 (29.0) | 24 (21.8) | .2258 |
| Diabetes mellitus | 38 (17.5) | 15 (14.0) | 23 (20.9) | .1818 |
| Solid malignancy | 48 (22.1) | 23 (21.5) | 25 (22.7) | .8270 |
| Gastrointestinal disease | 39 (18.0) | 21 (19.6) | 18 (16.4) | .5315 |
| Hematological malignancy | 35 (16.1) | 14 (13.1) | 21 (19.1) | .2291 |
| Chronic renal failure | 40 (18.4) | 16 (15.0) | 24 (21.8) | .1923 |
| Chronic obstructive pulmonary disease | 24 (11.1) | 9 (8.4) | 15 (13.6) | .2198 |
| Hematopoietic stem cell transplant | 11 (5.1) | 5 (4.7) | 6 (5.5) | .7930 |
| Solid organ transplant | 12 (5.5) | 4 (3.7) | 8 (7.3) | .2548 |
| Liver cirrhosis | 4 (1.8) | 2 (1.9) | 2 (1.8) | 1.0000f |
| Splenectomy | 5 (2.3) | 2 (1.9) | 3 (2.7) | 1.0000f |
| HIV | 3 (1.4) | 2 (1.9) | 1 (0.9) | .6181f |
| Charlson comorbidity index, median (IQR) | 4 (2–6) | 4 (2–6) | 4 (2–6) | .3156 |
| Corticosteroid therapy | 48 (22.1) | 28 (26.2) | 20 (18.2) | .1564 |
| Chemotherapy | 28 (12.9) | 13 (12.2) | 15 (13.6) | .7439 |
| Other immunosuppressive therapy | 29 (13.4) | 14 (13.1) | 15 (13.6) | .9049 |
| Neutropenia | 18 (8.3) | 8 (7.5) | 10 (9.1) | .6664 |
| Intermittent dialysis treatment | 16 (7.4) | 4 (3.7) | 12 (10.9) | .0433 |
| Invasive proceduresc, n (%) | ||||
| Central venous catheter | 157 (72.4) | 82 (76.6) | 75 (68.2) | .1639 |
| Urinary catheter | 134 (61.8) | 67 (62.6) | 67 (60.9) | .7958 |
| Any surgery | 49 (22.6) | 23 (21.5) | 26 (23.6) | .7061 |
| Percutaneous endoscopic gastrostomy | 13 (6.0) | 6 (5.6) | 7 (6.4) | .8145 |
| Previous P. aeruginosa colonizationd, n (%) | 66 (30.4) | 34 (31.8) | 32 (29.1) | .6673 |
| Source of bloodstream infectiona, n (%) | … | … | … | .0969f |
| CVC-related BSI | 144 (66.4) | 80 (74.8) | 64 (58.2) | |
| HAP/VAP | 1 (0.5) | 0 (0.0) | 1 (0.9) | |
| Urinary tract infections | 28 (12.9) | 14 (13.1) | 14 (12.7) | |
| Intra-abdominal infections | 23 (10.6) | 8 (7.5) | 15 (13.6) | |
| Skin and soft tissue infection | 11 (5.1) | 3 (2.8) | 8 (7.3) | |
| Osteomyelitis | 3 (1.4) | 1 (0.9) | 2 (1.8) | |
| Central nervous system infection | 1 (0.5) | 0 (0.0) | 1 (0.9) | |
| Otherse | 6 (2.7) | 1 (0.9) | 5 (4.6) | |
| Susceptibility profileb, n (%) | … | … | … | .4435f |
| MDR | 22 (10.3) | 10 (9.4) | 12 (11.2) | |
| XDR | 12 (5.6) | 4 (3.8) | 8 (7.5) | |
| PDR | 1 (0.5) | 0 (0.0) | 1 (0.9) | |
| Clinical presentation, n (%) | … | … | … | .8642 |
| Sepsis | 105 (48.4) | 52 (48.6) | 53 (48.2) | |
| Septic Shock | 42 (19.4) | 22 (20.6) | 20 (18.2) | |
| No sepsis | 70 (32.3) | 33 (30.8) | 37 (33.6) | |
| Appropriate antibiotic therapy, n (%) | 198 (91.2) | 100 (93.5) | 98 (89.1) | .2552 |
| Outcomes, n (%) | ||||
| Need for ICU admission | 6 (2.8) | 3 (2.8) | 3 (2.7) | 1.0000f |
| Need for CRRT after P. aeruginosa BSI | 4 (1.8) | 1 (0.9) | 3 (2.7) | .6216f |
| 30-d all-cause mortality | 33 (15.2) | 7 (6.5) | 26 (23.6) | .0005 |
| Recurrence of BSI | 6 (2.8) | 2 (1.9) | 4 (3.6) | .6833f |
The reported P values are derived from the Wilcoxon rank sum test for continuous variables and χ2 or Fisher's exact test for categorical variables; bold values indicate statistical significance at the predefined level (α = .05). One patient in the early adequate source control group had missing age data, which was imputed using the median for subsequent analyses. The median and interquartile range reported in the table were identical before and after imputation, while the P value was virtually unchanged (reported: available cases; after imputation: P = .2029). No other missing values were present in this table.
Abbreviations: BSI, bloodstream infection; CRRT, continuous renal replacement therapy; CVC, central venous catheter; ICU, intensive care unit; IQR, interquartile range; HAP, hospital-acquired pneumonia; HIV, human immunodeficiency virus; MDR, multidrug resistant; PDR, pandrug-resistant; VAP, ventilator-associated pneumonia; XDR, extensively drug resistant.
aTest calculated using Monte Carlo simulation.
bAnalysis conducted on a total of 213 patients with a known susceptibility profile.
cWithin 30 d before P. aeruginosa BSI onset.
dWithin 90 d before P. aeruginosa BSI onset.
eIncludes mediastinitis, endocarditis, infection of spinal fixation hardware, community-acquired pneumonia, infection of prosthetic femur, lung abscess, pleural empyema, LVAD infection.
fFisher's exact test.
Table 5.
Univariable, Multivariable, and Backward Selection Multivariable Analyses of Factors Associated With 30-d All-Cause Mortality in the Study Population With Source Control Deemed Necessary (n = 217)
| Univariable | Multivariable | Multivariable Backward (Model A2) |
||||
|---|---|---|---|---|---|---|
| Variable | HR (95% CI) | P Value | aHR (95% CI) | P Value | aHR (95% CI) | P Value |
| Age (y) | 1.02 (1.00–1.05) | .0704 | 1.02 (.99–1.04) | .1632 | … | … |
| Male sex | 0.45 (.22–.89) | .0224 | 0.45 (.22–.90) | .0245 | 0.45 (.22–.90) | .0240 |
| Cardiovascular disease | 1.72 (.87–3.41) | .1209 | … | … | … | … |
| Neurological disease | 1.45 (.70–2.99) | .3145 | … | … | … | … |
| Diabetes mellitus | 0.74 (.29–1.93) | .5424 | … | … | … | … |
| Solid malignancy | 0.92 (.40–2.13) | .8517 | … | … | … | … |
| Gastrointestinal disease | 0.65 (.23–1.84) | .4112 | … | … | … | … |
| Hematological malignancy | 0.36 (.09–1.49) | .1564 | … | … | … | … |
| Chronic renal failure | 0.99 (.41–2.39) | .9729 | … | … | … | … |
| Chronic obstructive pulmonary disease | 0.87 (.26–2.84) | .8121 | … | … | … | … |
| Hematopoietic stem cell transplanta | … | … | … | … | … | … |
| Solid organ transplanta | … | … | … | … | … | … |
| Liver cirrhosisa | … | … | … | … | … | … |
| Splenectomya | … | … | … | … | … | … |
| HIVa | … | … | … | … | … | … |
| Charlson comorbidity index | 1.06 (.95–1.18) | .3231 | … | … | … | … |
| Corticosteroid therapy | 0.72 (.30–1.75) | .4693 | … | … | … | … |
| Chemotherapy | 0.70 (.21–2.29) | .5540 | … | … | … | … |
| Other immunosuppressive therapya | … | … | … | … | … | … |
| Neutropenia | 0.81 (.19–3.38) | .7696 | … | … | … | … |
| Intermittent dialysisa | … | … | … | … | … | … |
| Central venous cathetere | 0.59 (.29–1.18) | .1371 | … | … | … | … |
| Urinary cathetere | 1.85 (.83–4.10) | .1309 | … | … | … | … |
| Any surgerye | 1.43 (.68–3.00) | .3509 | … | … | … | … |
| Percutaneous endoscopic gastrostomye | 1.53 (.47–5.03) | .4823 | … | … | … | … |
| Previous P. aeruginosa colonizationf | 0.63 (.28–1.40) | .2582 | … | … | … | … |
| Septic shock | 4.54 (2.29–9.00) | <.0001 | 6.16 (3.02–12.57) | <.0001 | 6.16 (3.03–12.51) | <.0001 |
| Deemed necessary and adequately performed early source control | 0.26 (.11–.60) | .0017 | 0.21 (.09–.50) | .0004 | 0.22 (.09–.50) | .0004 |
| Time to source control (d)b | 0.90 (.73–1.12) | .3487 | … | … | … | … |
| Appropriate antibiotic therapyc | 0.51 (.20–1.32) | .1663 | … | … | … | … |
| Time to appropriate antibiotic therapy (d)d | 0.79 (.56–1.11) | .1727 | … | … | … | … |
| Need for ICU admissiona | … | … | … | … | … | … |
| Need for CRRT after P. aeruginosa BSIa | … | … | … | … | … | … |
The reported P values are from the Cox regression analysis. Bold values are significant at the selected level of significance (α = .05). Variables source of bloodstream infection, hospital ward at the time of P. aeruginosa BSI onset, susceptibility profile, and recurrence of BSI were excluded from the analysis. Variables time to source control and time to appropriate antibiotic therapy were treated as time-dependent.
Abbreviations: aHR, adjusted hazard ratio; BSI, bloodstream infection; CRRT, continuous renal replacement therapy; HR, hazard ratio; ICU, intensive care unit.
aAnalysis not done due to insufficient events.
bAnalysis conducted on a total of 172 patients who underwent source control during the study period.
c Appropriate antibiotic therapy was treated as fixed covariate due to its clinical relevance, the median initiation time of 0 (IQR: 0–2) day, the lack of a significant effect of time to appropriate antibiotic therapy (P = .1727), and the absence of interaction with time (P = .3189) in the Cox model.
dAnalysis conducted on a total of 198 patients who received appropriate antibiotic therapy.
eWithin 30 d before P. aeruginosa BSI onset.
fWithin 90 d before P. aeruginosa BSI onset.
We also performed a PS-based analysis, in which 70 pairs of patients who received early adequate source control were matched to those who did not. The standardized differences between the 2 groups were ≤0.2 for all variables except hematological malignancy, indicating that the groups were well balanced. The Cox proportional hazards model in the PS-matched cohort, adjusted for hematological malignancy, confirmed that early adequate source control was associated with a lower 30-day all-cause mortality rate (aHR 0.15, 95% CI .04–.51, P = .0024) compared to patients who did not receive it. The results in the 4-day landmark analysis were largely consistent with the PS findings.
Impact of Source Control and Antibiotic Therapy on Mortality
Among patients with source control deemed necessary who survived at least 4 days from BSI onset, 30-day mortality rates were low in those undergoing early adequate source control, irrespective of antibiotic appropriateness (0% vs 6.2%, P = 1.0000). In contrast, among patients who did not receive early adequate source control, mortality was 16.5% among those who received appropriate antibiotic therapy and 22.2% in those who did not; this difference was not statistically significant (P = .6475) (Supplementary Figure 1).
DISCUSSION
The key findings of this multicenter cohort study can be summarized as follows: (1) approximately one in 3 patients with P. aeruginosa BSI required source control; (2) early adequate source control was associated with a significantly lower risk of 30-day all-cause mortality; and (3) in the exploratory analysis restricted to patients requiring source control, mortality was numerically lower with appropriate antibiotic therapy among patients without early adequate source control, whereas mortality remained low overall among those who received early adequate source control.
P. aeruginosa infections are associated with high mortality, as highlighted in a large observational study on sepsis conducted in 198 European ICUs, where it was the only pathogen independently linked to mortality [27]. Several studies confirm the poor prognosis of P. aeruginosa BSI, with attributable mortality rates reaching up to 60% [13–16]. In recent years, efforts to improve outcomes have primarily focused on early diagnosis [15], optimized antibiotic dosing [28], and combination therapy in both empirical [16] and targeted regimens [29]. In contrast, adequate source control-a cornerstone of infection management, encapsulated in the ancient principle ubi pus, ibi evacua-has received little attention [1]. Few studies have evaluated its impact [19, 20], and current guidelines make only a brief mention of its role in P. aeruginosa BSI [30, 31].
Yet, adequate source control remains essential for a substantial proportion of patients with P. aeruginosa BSI [19, 20]. A recent study by Papadimitriou-Olivgeris et al [19] reported that 50% of P. aeruginosa BSI cases required source control. However, their analysis considered catheter removal in all patients with a CVC, regardless of whether it was identified as the actual infection source [19]. To better reflect clinical practice and guideline-based infection management, we considered CVC removal necessary only when the catheter was confirmed as the likely infection source [2]. Using this more stringent definition, our findings indicate that approximately one-third of P. aeruginosa BSI cases required source control, underscoring the need for a more structured and evidence-based approach to its management.
Although the absence of early source control has been identified as an independent predictor of mortality in BSI caused by various pathogens [5–8], this critical aspect of P. aeruginosa BSI management has often been overlooked in mortality analyses [32–34]. To our knowledge, only one retrospective study, conducted at a single center in Switzerland and including 278 patients with P. aeruginosa BSI, has thoroughly addressed this issue [19], demonstrating a significant survival benefit associated with early adequate source control. In that study, 14-day mortality was markedly lower in patients who underwent source control compared to those who did not (7% vs 21%) [19]. Our study reinforces and expands these findings by providing a national multicenter perspective and including a higher proportion of patients with septic shock due to P. aeruginosa (20% vs 0%) [19]. Additionally, our cohort had a higher prevalence of MDR, XDR or PDR strains (overall 16.6% vs 7.0%) [19], further emphasizing the broader clinical relevance of early source control in improving patient outcomes.
Of importance, among patients who achieved early adequate source control, 30-day mortality rates were very low, with no statistically significant difference by antibiotic appropriateness (P = 1.0000; Supplementary Figure 1). Conversely, among patients without early adequate source control, mortality was numerically lower in those receiving appropriate antibiotic therapy (16.5% vs 22.2%), although the difference was not statistically significant (P = .6475). These exploratory findings highlight that, among patients without early adequate source control, mortality remained substantial even when appropriate antibiotic therapy was administered. From a clinical perspective, these findings emphasize the importance of promptly identifying and managing the infectious focus, as recommended for other severe infections (eg, S. aureus and Candida BSI) [35–37]. In this context, infectious disease specialists play a key role, as previous studies have shown their contribution to improving infection source identification and, consequently, patient outcomes [20, 38].
Our findings regarding other mortality risk factors align with recent studies identifying older age [32, 34, 39], corticosteroid therapy [32, 34], and septic shock at presentation [32, 34, 39] as key predictors of 30-day all-cause mortality in P. aeruginosa BSI. Notably, in contrast with prior studies [40, 41], we observed a protective effect of prior P. aeruginosa colonization. This may be explained by clinicians' heightened suspicion of P. aeruginosa infection in colonized patients, leading to earlier interventions, including source control measures and prompt initiation of appropriate antibiotic therapy based on prior microbiological data. However, susceptibility data from previous colonizing isolates and their concordance with empirical therapy were not available; therefore, this hypothesis could not be directly assessed.
Our study has other limitations that should be acknowledged. First, multiple confounding factors may have influenced the observed association between early interventions and improved survival, including decisions to limit or withdraw care and the availability and willingness of surgeons or interventional radiologists to perform adequate source control procedures. However, our findings were confirmed in the landmark analysis, which included only patients who were alive at day 4 after P. aeruginosa BSI onset. Second, infection source attribution and the need for source control were assessed by the local investigators at each participating center based on the available clinical information. As a result, some patients classified as having BSI of unknown source may have had an occult catheter-related, urinary, pulmonary, gastrointestinal, or another potential focus that was not identified. The high mortality rate observed in this group comparable to that among patients for whom source control was indicated but not performed supports this hypothesis. This potential misclassification should be considered when interpreting the outcomes of patients in whom source control is deemed unnecessary and further reinforces the importance of a systematic approach to P. aeruginosa BSI and the need for actively identifying and addressing the infection source. Third, although PS-matching does not fully adjust for time-dependent bias, it was included just as a complementary, exploratory analysis to support the robustness of the main findings. Fourth, the use of a 4-day landmark analysis, applied to assess the interplay between source control adequacy and appropriate antibiotic therapy, helped mitigate immortal time bias by including only patients who were alive at day 4 after BSI onset. However, this approach inherently excludes early deaths and may therefore underestimate the potential benefit of early interventions in the most critically ill patients. Fifth, we collected information only on the first source control intervention. Data on additional source control procedures were not recorded; therefore, their potential impact on patient outcomes could not be assessed. Lastly, we did not assess the impact of adequate source control on the duration of appropriate antibiotic therapy [42] or its potential role in reducing the risk of in vivo antimicrobial resistance development [43]-both of which may represent additional benefits of effective source control. However, our study was not designed to address these aspects, and further research is needed to explore them.
By addressing the knowledge gap regarding the role of appropriate source control in P. aeruginosa BSI outcomes, our study provides valuable insights into the multifaceted approach required for its effective management. Our findings underscore the critical importance of early identification and appropriate implementation of source control in patients with P. aeruginosa BSI, highlighting its potential to improve patient care and reduce mortality. Further research is warranted to refine source control strategies and assess their impact on antimicrobial resistance and treatment duration.
Supplementary Material
Notes
Acknowledgments. Members of the OPERA SITA GIOVANI study group: Accurso Giuseppe, Bavaro Davide Fiore, Chiappetta Stefania, Faliero Domenico, Fumarola Benedetta, Geremia Nicholas, Giovannenze Francesca, Maccaro Angelo, Marino Andrea, Merli Marco, Pagani Gabriele, Sardanelli Alessia, Scaglione Vincenzo, Tontodonati Monica, and Vassia Veronica.
Author contributions. A. V., M. B., and M. Bav.: Formal analysis, writing—original draft, writing—review and editing. M. M.: Methodology, writing—review and editing. S. C. and P. V.: Methodology. All authors: Data collection and data curation. A. V., M. Bav., S. C., P. V., and M. B.: Conceptualization, formal analysis, project administration, resources, study design, supervision, writing—review and editing.
Data availability. The data underlying this article will be shared on reasonable request to the corresponding author.
Financial support. This study was supported by an independent medical grant from Pfizer Inc., Shionogi & Co., Ltd., and Tillotts Pharma AG. No grant numbers were assigned. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Contributor Information
Antonio Vena, Department of Health Sciences (DISSAL), University of Genoa, Genoa, Italy; Clinica Malattie Infettive, AOM - IRCCS Ospedale Policlinico San Martino, Genoa, Italy.
Martina Bavastro, Department of Health Sciences (DISSAL), University of Genoa, Genoa, Italy; Clinica Malattie Infettive, AOM - IRCCS Ospedale Policlinico San Martino, Genoa, Italy.
Marco Muccio, Clinica Malattie Infettive, AOM - IRCCS Ospedale Policlinico San Martino, Genoa, Italy.
Michela Schenone, Clinica Malattie Infettive, AOM - IRCCS Ospedale Policlinico San Martino, Genoa, Italy.
Silvia Corcione, Department of Medical Sciences, Infectious Diseases, University of Turin, Turin, Italy; Department of Infectious Diseases, Tufts University School of Medicine, Boston, Massachusetts, USA.
Renato Pascale, Infectious Diseases Unit, IRCCS-Sant’ Orsola Polyclinic, Bologna, Italy; Department of Medical and Surgical Sciences, University of Bologna, Bologna, Italy.
Daniele Roberto Giacobbe, Department of Health Sciences (DISSAL), University of Genoa, Genoa, Italy; Clinica Malattie Infettive, AOM - IRCCS Ospedale Policlinico San Martino, Genoa, Italy.
Simone Mornese Pinna, Department of Medical Sciences, Infectious Diseases, University of Turin, Turin, Italy.
Bianca Pari, Department of Medical Sciences, Infectious Diseases, University of Turin, Turin, Italy.
Francesca Giovannenze, Department of Laboratory and Infectious Sciences, IRCCS A. Gemelli University Polyclinic Foundation, Rome, Italy.
Nicholas Geremia, Unit of Infectious Diseases, Department of Clinical Medicine, Dell’Angelo Hospital, Venice, Italy.
Malgorzata Mikulska, Department of Health Sciences (DISSAL), University of Genoa, Genoa, Italy; Clinica Malattie Infettive, AOM - IRCCS Ospedale Policlinico San Martino, Genoa, Italy.
Eleonora Taddei, Department of Laboratory and Infectious Sciences, IRCCS A. Gemelli University Polyclinic Foundation, Rome, Italy.
Flavio Sangiorgi, Department of Security and Bioethics-Infectious Diseases Section, Catholic University of the Sacred Heart, Rome, Italy.
Davide Fiore Bavaro, Department of Biomedical Sciences, Humanitas University, Milan, Italy; Infectious Disease Unit, IRCCS Humanitas Research Hospital, Rozzano, Milan, Italy.
Vincenzo Scaglione, Infectious and Tropical Diseases Unit, Padua University Hospital, Padua, Italy.
Silvia Sisto, Department of Health Sciences (DISSAL), University of Genoa, Genoa, Italy; Clinica Malattie Infettive, AOM - IRCCS Ospedale Policlinico San Martino, Genoa, Italy.
Lucia Taramasso, Clinica Malattie Infettive, AOM - IRCCS Ospedale Policlinico San Martino, Genoa, Italy.
Veronica Vassia, Infectious and Tropical Disease Unit, Mauriziano Umberto I Hospital, Turin, Italy; Infectious and Tropical Disease Unit, Civile Hospital, Ivrea, Italy.
Michele Bartoletti, Department of Biomedical Sciences, Humanitas University, Milan, Italy; Infectious Disease Unit, IRCCS Humanitas Research Hospital, Rozzano, Milan, Italy.
Francesco Giuseppe De Rosa, Department of Medical Sciences, Infectious Diseases, University of Turin, Turin, Italy.
Maddalena Giannella, Infectious Diseases Unit, IRCCS-Sant’ Orsola Polyclinic, Bologna, Italy; Department of Medical and Surgical Sciences, University of Bologna, Bologna, Italy.
Pierluigi Viale, Infectious Diseases Unit, IRCCS-Sant’ Orsola Polyclinic, Bologna, Italy; Department of Medical and Surgical Sciences, University of Bologna, Bologna, Italy.
Matteo Bassetti, Department of Health Sciences (DISSAL), University of Genoa, Genoa, Italy; Clinica Malattie Infettive, AOM - IRCCS Ospedale Policlinico San Martino, Genoa, Italy.
SITA GIOVANI (Young Investigators Group of the Società Italiana Terapia Antinfettiva:
Accurso Giuseppe, Bavaro Davide Fiore, Chiappetta Stefania, Faliero Domenico, Fumarola Benedetta, Geremia Nicholas, Giovannenze Francesca, Maccaro Angelo, Marino Andrea, Merli Marco, Pagani Gabriele, Sardanelli Alessia, Scaglione Vincenzo, Tontodonati Monica, and Vassia Veronica
Supplementary Data
Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
References
- 1. Evans L, Rhodes A, Alhazzani W, et al. Surviving sepsis campaign: international guidelines for management of sepsis and septic shock 2021. Crit Care Med 2021; 49:e1063–143. [DOI] [PubMed] [Google Scholar]
- 2. Mermel LA, Allon M, Bouza E, et al. Clinical practice guidelines for the diagnosis and management of intravascular catheter-related infection: 2009 update by the Infectious Diseases Society of America. Clin Infect Dis 2009; 49:1–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Reitz KM, Kennedy J, Li SR, et al. Association between time to source control in sepsis and 90-day mortality. JAMA Surg 2022; 157:817–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Nicolle LE. A practical guide to the management of complicated urinary tract infection. Drugs 1997; 53:583–92. [DOI] [PubMed] [Google Scholar]
- 5. Papadimitriou-Olivgeris M, Caruana G, Senn L, Guery B. Predictors of mortality of Staphylococcus aureus bacteremia among patients hospitalized in a Swiss University Hospital and the role of early source control; a retrospective cohort study. Eur J Clin Microbiol Infect Dis 2023; 42:347–57. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Fourre N, Zimmermann V, Senn L, Aruanno M, Guery B, Papadimitriou-Olivgeris M. Predictors of mortality of streptococcal bacteremia and the role of infectious diseases consultation: a retrospective cohort study. Clin Infect Dis 2024; 78:1544–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Falcone M, Russo A, Iacovelli A, et al. Predictors of outcome in ICU patients with septic shock caused by Klebsiella pneumoniae carbapenemase-producing K. pneumoniae. Clin Microbiol Infect 2016; 22:444–50. [DOI] [PubMed] [Google Scholar]
- 8. Bassetti M, Vena A, Meroi M, et al. Factors associated with the development of septic shock in patients with candidemia: a post hoc analysis from two prospective cohorts. Crit Care 2020; 24:117. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Antimicrobial Resistance Collaborators . Global burden of bacterial antimicrobial resistance in 2019: a systematic analysis. Lancet 2022; 399:629–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Wisplinghoff H, Bischoff T, Tallent SM, Seifert H, Wenzel RP, Edmond MB. Nosocomial bloodstream infections in US hospitals: analysis of 24,179 cases from a prospective nationwide surveillance study. Clin Infect Dis 2004; 39:309–17. [DOI] [PubMed] [Google Scholar]
- 11. Strich JR, Mishuk A, Diao G, et al. Assessing clinician utilization of next-generation antibiotics against resistant Gram-negative infections in U.S. hospitals: a retrospective cohort study. Ann Intern Med 2024; 177:559–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Bassetti M, Vena A, Russo A, Croxatto A, Calandra T, Guery B. Rational approach in the management of Pseudomonas aeruginosa infections. Curr Opin Infect Dis 2018; 31:578–86. [DOI] [PubMed] [Google Scholar]
- 13. Zhang Y, Li Y, Zeng J, et al. Risk factors for mortality of inpatients with Pseudomonas aeruginosa bacteremia in China: impact of resistance profile in the mortality. Infect Drug Resist 2020; 13:4115–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Ammerlaan HS, Harbarth S, Buiting AG, et al. Secular trends in nosocomial bloodstream infections: antibiotic-resistant bacteria increase the total burden of infection. Clin Infect Dis 2013; 56:798–805. [DOI] [PubMed] [Google Scholar]
- 15. Osih RB, McGregor JC, Rich SE, et al. Impact of empiric antibiotic therapy on outcomes in patients with Pseudomonas aeruginosa bacteremia. Antimicrob Agents Chemother 2007; 51:839–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Vena A, Schenone M, Corcione S, et al. Impact of adequate empirical combination therapy on mortality in septic shock due to Pseudomonas aeruginosa bloodstream infections: a multicentre retrospective cohort study. J Antimicrob Chemother 2024; 79:2846–53. [DOI] [PubMed] [Google Scholar]
- 17. Kang CI, Kim SH, Kim HB, et al. Pseudomonas aeruginosa bacteremia: risk factors for mortality and influence of delayed receipt of effective antimicrobial therapy on clinical outcome. Clin Infect Dis 2003; 37:745–51. [DOI] [PubMed] [Google Scholar]
- 18. Schechner V, Gottesman T, Schwartz O, et al. Pseudomonas aeruginosa bacteremia upon hospital admission: risk factors for mortality and influence of inadequate empirical antimicrobial therapy. Diagn Microbiol Infect Dis 2011; 71:38–45. [DOI] [PubMed] [Google Scholar]
- 19. Papadimitriou-Olivgeris M, Senn L, Jacot D, Guery B. Predictors of mortality of Pseudomonas aeruginosa bacteraemia and the role of infectious diseases consultation and source control; a retrospective cohort study. Infection 2025; 53:117–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Chiong F, Wasef MS, Liew KC, et al. The impact of infectious diseases consultation on the management and outcomes of Pseudomonas aeruginosa bacteraemia in adults: a retrospective cohort study. BMC Infect Dis 2021; 21:671. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987; 40:373–83. [DOI] [PubMed] [Google Scholar]
- 22. Kalil AC, Metersky ML, Klompas M, et al. Executive summary: management of adults with hospital-acquired and ventilator-associated pneumonia: 2016 clinical practice guidelines by the Infectious Diseases Society of America and the American Thoracic Society. Clin Infect Dis 2016; 63:575–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Otto CM, Nishimura RA, Bonow RO, et al. 2020 ACC/AHA guideline for the management of patients with valvular heart disease: a report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation 2021; 143:e72–227. [DOI] [PubMed] [Google Scholar]
- 24. Singer M, Deutschman CS, Seymour CW, et al. The third international consensus definitions for sepsis and septic shock (sepsis-3). JAMA 2016; 315:801–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Magiorakos AP, Srinivasan A, Carey RB, et al. Multidrug-resistant, extensively drug-resistant and pandrug-resistant bacteria: an international expert proposal for interim standard definitions for acquired resistance. Clin Microbiol Infect 2012; 18:268–81. [DOI] [PubMed] [Google Scholar]
- 26.The European Committee on Antimicrobial Susceptibility Testing. Breakpoint tables for Interpretation of MICs and Zone Diameters. Version 14.0. 24. https://www.eucast.org/clinical_breakpoints.
- 27. Vincent JL, Sakr Y, Sprung CL, et al. Sepsis in European intensive care units: results of the SOAP study. Crit Care Med 2006; 34:344–53. [DOI] [PubMed] [Google Scholar]
- 28. Yadav R, Rogers KE, Bergen PJ, et al. Optimization and evaluation of piperacillin-tobramycin combination dosage regimens against Pseudomonas aeruginosa for patients with altered pharmacokinetics via the hollow-fiber infection model and mechanism-based modeling. Antimicrob Agents Chemother 2018; 62:e00078-18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Onorato L, Macera M, Calo F, Cirillo P, Di Caprio G, Coppola N. Beta-lactam monotherapy or combination therapy for bloodstream infections or pneumonia due to Pseudomonas aeruginosa: a meta-analysis. Int J Antimicrob Agents 2022; 59:106512. [DOI] [PubMed] [Google Scholar]
- 30. Paul M, Carrara E, Retamar P, et al. European Society of Clinical Microbiology and Infectious Diseases (ESCMID) guidelines for the treatment of infections caused by multidrug-resistant Gram-negative bacilli (endorsed by European Society of Intensive Care Medicine). Clin Microbiol Infect 2022; 28:521–47. [DOI] [PubMed] [Google Scholar]
- 31. Tamma PD, Heil EL, Justo JA, Mathers AJ, Satlin MJ, Bonomo RA. Infectious Diseases Society of America 2024 guidance on the treatment of antimicrobial-resistant Gram-negative infections. Clin Infect Dis 2024; [DOI] [PubMed] [Google Scholar]
- 32. Babich T, Naucler P, Valik JK, et al. Risk factors for mortality among patients with Pseudomonas aeruginosa bacteraemia: a retrospective multicentre study. Int J Antimicrob Agents 2020; 55:105847. [DOI] [PubMed] [Google Scholar]
- 33. Frem JA, Doumat G, Kazma J, et al. Clinical predictors of mortality in patients with pseudomonas aeruginosa infection. PLoS One 2023; 18:e0282276. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Yuan F, Li M, Wang X, Fu Y. Risk factors and mortality of carbapenem-resistant Pseudomonas aeruginosa bloodstream infection in haematology department: a 10-year retrospective study. J Glob Antimicrob Resist 2024; 37:150–6. [DOI] [PubMed] [Google Scholar]
- 35. Cardozo C, Cuervo G, Salavert M, et al. An evidence-based bundle improves the quality of care and outcomes of patients with candidaemia. J Antimicrob Chemother 2020; 75:730–7. [DOI] [PubMed] [Google Scholar]
- 36. Vena A, Bouza E, Corisco R, et al. Efficacy of a “checklist” intervention bundle on the clinical outcome of patients with Candida bloodstream infections: a quasi-experimental pre-post study. Infect Dis Ther 2020; 9:119–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Lopez-Cortes LE, Del Toro MD, Galvez-Acebal J, et al. Impact of an evidence-based bundle intervention in the quality-of-care management and outcome of Staphylococcus aureus bacteremia. Clin Infect Dis 2013; 57:1225–33. [DOI] [PubMed] [Google Scholar]
- 38. Ramanathan S, Albarillo FS, Fitzpatrick MA, et al. Infectious disease consults of Pseudomonas aeruginosa bloodstream infection and impact on health outcomes. Open Forum Infect Dis 2022; 9:ofac456. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Morata L, Cobos-Trigueros N, Martinez JA, et al. Influence of multidrug resistance and appropriate empirical therapy on the 30-day mortality rate of Pseudomonas aeruginosa bacteremia. Antimicrob Agents Chemother 2012; 56:4833–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Meschiari M, Orlando G, Kaleci S, et al. Combined resistance to ceftolozane-tazobactam and ceftazidime-avibactam in extensively drug-resistant (XDR) and multidrug-resistant (MDR) Pseudomonas aeruginosa: resistance predictors and impact on clinical outcomes besides implications for antimicrobial stewardship programs. Antibiotics (Basel) 2021; 10:1224. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Mahar P, Padiglione AA, Cleland H, Paul E, Hinrichs M, Wasiak J. Pseudomonas aeruginosa bacteraemia in burns patients: risk factors and outcomes. Burns 2010; 36:1228–33. [DOI] [PubMed] [Google Scholar]
- 42. Ranganath N, Hassett LC, Saleh OMA, Yetmar ZA. Short versus prolonged duration of therapy for Pseudomonas aeruginosa bacteraemia: a systematic review and meta-analysis. J Hosp Infect 2024; 148:155–66. [DOI] [PubMed] [Google Scholar]
- 43. Tamma PD, Beisken S, Bergman Y, et al. Modifiable risk factors for the emergence of ceftolozane-tazobactam resistance. Clin Infect Dis 2021; 73:e4599–606. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.