Impact of Pathogen Status on Sepsis-Associated Acute Respiratory Distress Syndrome Outcomes

Introduction

Acute respiratory distress syndrome (ARDS) affects approximately 10% of patients in intensive care units (ICUs). With sepsis being the leading cause in over 40% of cases, the mortality rate of sepsis-associated ARDS remains alarmingly high, ranging from 30% to 50%, despite advances in critical care management [1–3]. Clinically, patients with sepsis-associated ARDS present with severe hypoxemia, bilateral pulmonary infiltrates, and respiratory distress, necessitating prompt diagnosis and intervention. Diagnostic approaches typically rely on clinical criteria, including the Berlin Definition, and supportive imaging and laboratory findings; treatment strategies focus on lung-protective ventilation, fluid management, and addressing the underlying infection, although outcomes remain suboptimal [4–6]. Therefore, it is particularly important to explore the risk factors for sepsis to cause different subtypes of ARDS and provide effective prevention and treatment.

The occurrence of ARDS due to sepsis, as defined by Sepsis 3.0, includes identification of the causative organism and suspicion of infection and the absence of the causative organism captured. Despite extensive research on sepsis and ARDS, which is traditionally associated with detectable pathogenic microorganisms, emerging evidence suggests that a substantial proportion of cases present without identifiable pathogens, posing unique therapeutic challenges. Significant gaps remain in understanding the role of pathogenic microorganisms in shaping the clinical course and prognosis of sepsis-associated ARDS. While some studies have highlighted the correlation between microbiological findings and patient outcomes, there is a paucity of data on the distinct characteristics and prognostic differences between microbiologically positive and negative sepsis-associated ARDS cases. Furthermore, the identification of risk factors specific to microbiologically confirmed sepsis-associated ARDS is crucial for improving risk stratification and therapeutic strategies [7,8].

The purpose of this study was to investigate the clinical characteristics and prognostic differences between patients with microbiologically positive and negative sepsis-associated ARDS. Additionally, we aimed to identify prognostic risk factors specific to microbiologically confirmed cases, providing insights that can guide more targeted and effective management approaches.

Material and Methods

STUDY DESIGN AND SETTING:

This was a retrospective, multicenter cohort study conducted between January 2008 and December 2019. The study aimed to investigate the clinical characteristics and prognostic differences between patients with microbiologically positive and negative sepsis-associated ARDS.

STUDY POPULATION:

The study population included all adult patients (≥18 years) diagnosed with sepsis-associated ARDS according to the Berlin Definition and Sepsis 3.0 criteria [1,9]. Patients administered mechanical ventilation were included in this study from the Medical Information Mart for Intensive Care IV (MIMIC IV v2.2) and electronic Intensive Care Unit (eICU 2.0) databases. The patients who had a hospital stay of less than 24 h, were younger than 18 years, and were diagnosed with acute heart failure, cardiogenic pulmonary edema, massive pleural effusion, and other diseases that affected the diagnosis of ARDS were excluded (Figure 1). The MIMIC IV database is a single-center database with data from 2008 to 2019 [10]. The eICU database contains multicenter data from 2014 to 2015 [11]. The MIMIC IV v2.2, eICU-CRD v2.0, was approved by the Institutional Review Board of the Beth Israel Deaconess Medical Center (2001-P001699/14) and Massachusetts Institute of Technology (No. 0403000206). All patient information were unidentified, so patient informed consent was not required.

DATA COLLECTION:

In this study, patients were searched for baseline data, including age, sex, site of infection, pathogenic microorganisms, and co-morbid conditions; clinical data: temperature, heart rate, systolic blood pressure, diastolic blood pressure, median levels of respiratory-related indicators during mechanical ventilation, worst level of function of coagulation function, kidney function, indicators of infection during mechanical ventilation, and prognostic indicators: mortality, length of ICU stay, Simplified Acute Physiology Score (SAPS) II, Sequential Organ Failure Assessment (SOFA) score, use of inotropic/vasopressor support, and renal replacement therapy.

STATISTICAL ANALYSIS:

Continuous variables were assessed for normality using the Shapiro-Wilk test and were found to be non-normally distributed. These variables are presented as median and interquartile range (IQR). Categorical variables are expressed as frequencies and percentages. To compare baseline characteristics and outcomes between the 2 groups (patients with microbiologically positive and negative sepsis-associated ARDS), the following statistical tests were applied: Wilcoxon rank-sum test (Mann-Whitney U test) was used for continuous variables, due to their non-normal distribution, and the Fisher exact test was used for categorical variables, particularly when expected cell counts were small (Tables 1–3). Multivariate Cox proportional hazards regression analysis was performed to identify independent risk factors associated with prognosis in patients with microbiologically positive sepsis-associated ARDS. The results of the Cox analysis are presented as hazard ratios (HRs) with 95% confidence intervals (CIs) and visualized using a forest plot (Figure 1). Independent risk factors for prognosis were further analyzed using logistic regression, and the results are summarized in Table 4 and illustrated in Figure 2. Missing data were handled using multiple imputation to account for potential bias introduced by incomplete records. The sample size for this study was determined by the availability of data within the multicenter ICU database during the study period. Data were extracted using structured query language (SQL) to include all eligible patients meeting the inclusion criteria, ensuring a comprehensive and representative sample. All statistical analyses were performed using R Studio software (version 4.3.1, R Foundation for Statistical Computing, Vienna, Austria). To address the imbalance in group sizes (negative for microbial infection, n=1273; positive for microbial infection, n=655), stratified analysis and multivariable regression models were used to adjust for potential confounding factors. Sensitivity analyses were conducted to confirm the robustness of the findings. A 2-sided P value <0.05 was considered statistically significant.

Results

BASELINE CHARACTERISTICS AND OUTCOME OF SEPSIS-ASSOCIATED ARDS:

A total of 10 495 patients met the diagnosis of Sepsis 3.0 and Berlin Definition of ARDS. A total of 8567 patients with congestive heart failure (n=4259), cardiogenic pulmonary edema (n=943), large pleural effusion (n=11), hospital stay of less than 24 h (n=1680), missing blood oxygen-related indexes (n=1210), and other diseases affecting the diagnosis of ARDS (n=464) were excluded, and 1928 patients were included in the analysis (Figure 1). Although the proportions of the 2 groups (negative for microbial infection, n=1273; positive for microbial infection, n=655) were not equal, this reflects the natural distribution of microbial infection status in the study population. To ensure the credibility of our results, we used multivariable regression models to adjust for potential confounding factors. Sensitivity analyses were also conducted to confirm the robustness of the findings.

Table 1 compares the baseline data on pathogenic microorganisms-positive versus pathogenic microorganisms-negative sepsis-associated ARDS. The results in Table 1 show that the incidence of pulmonary infection with pathogenic microorganisms-positive sepsis-associated ARDS was higher (P=0.01), and the level of partial pressure of oxygen in the arterial blood (PaO2, P=0.01), PaO2/FiO2 (fraction of inspired O2, P<0.001), SpO2 (oxygen saturation)/FiO2 (P=0.01), and respiratory rate oxygenation (P=0.01) were lower than that in patients with pathogenic microorganisms-negative sepsis-associated ARDS. Patients with pathogenic microorganism-positive ARDS had worse renal function than did patients with pathogenic microbial-negative disease (P=0.06). There were no significant differences in age, sex, co-morbid conditions, platelet, hemoglobin, prothrombin time (PT), international normalized ratio (INR), white blood cell count (WBC), glucose, lactate, and other indexes between the 2 groups (P>0.05). Despite the difference in group sizes, the baseline characteristics of the 2 groups were balanced, as shown in Table 1.

The results of Table 2 showed that patients with pathogenic microorganisms-positive sepsis-associated ARDS had longer ICU stays (P<0.001), higher SOFA scores (P=0.001), and higher 28-day (P<0.001) and 90-day (P=0.001) mortality than did patients with pathogenic microorganism-negative sepsis-associated ARDS. Additionally, more patients in the pathogenic microorganism-positive group were treated with vasoactive drugs (P=0.003). These results suggest that patients with sepsis-associated ARDS positive for pathogenic microorganisms had a worse prognosis.

PROGNOSTIC ANALYSIS OF PATIENTS WITH PATHOGENIC MICROORGANISM-POSITIVE SEPSIS-ASSOCIATED ARDS:

The results in Table 3 show that compared with patients in the survival group, patients in the non-survival group with pathogenic microorganism-positive sepsis-associated ARDS were older; more patients were infected with Acinetobacter baumannii; more patients had chronic liver disease, kidney disease, chronic obstructive pulmonary disease, and immunosuppressive disease; more patients had a higher level of temperature, heart rate, PT, INR, APTT, creatinine, BUN, WBC, glucose, lactate, and SOFA score; more patients had lower levels of SpO2, PaO2, and SpO2/FiO2; and more patients were treated with vasoactive drugs and continuous renal replacement therapy (CRRT).

MULTIVARIATE COX REGRESSION ANALYSIS IN PATIENTS WITH MICROBIOLOGICAL-POSITIVE SEPSIS-ASSOCIATED ARDS:

Multivariate COX regression analysis revealed that age (HR=1.015, 95% CI: 1.005–1.025, P=0.004), INR (HR=1.188, 95% CI: 1.052–1.342, P=0.005), lactate level (HR=1.076, 95% CI: 1.027–1.128, P=0.002), A. baumannii infection (HR=2.024, 95% CI: 1.495–2.742, P<0.001), CRRT treatment (HR=2.256, 95% CI: 1.644–3.098, P<0.001), and SOFA score (HR=1.069, 95% CI: 1.022–1.119, P=0.004) were independent risk factors for mortality in patients with pathogenic microorganism-positive sepsis-associated ARDS.

ROC CURVE OF PATHOGENIC MICROORGANISM-POSITIVE PROGNOSIS IN PATIENTS WITH SEPSIS-ASSOCIATED ARDS:

As shown in Figure 2, age, A. baumannii, INR, lactate, CRRT, and SOFA scores were fitted to the model. Figure 3 and Table 4 show that the area under the curve (AUC) of the ROC curve of the model was significantly higher than that of other indicators (AUC=0.83, P<0.001), and the model had a strong discrimination ability for patients with sepsis-associated ARDS who were positive for pathogenic microorganisms.

:

Figure 4 shows the comparison of SOFA, SAPS II, PaO2, and PaO2/FiO2 levels between patients with survival and non-survival among cases of pathogenic microorganism-positive sepsis-associated ARDS. The figure shows that patients with sepsis-associated ARDS with non-survival had higher SOFA scores and lower PaO2 levels than did patients with survival (P<0.001).

Discussion

Patients with ARDS, especially those associated with sepsis, have a high mortality rate [6]. In this study, the 28-day mortality rate of patients with sepsis-associated ARDS was 25.3%, and the 90-day mortality rate was 26.1%, suggesting that patients with sepsis-associated ARDS still had a high mortality rate, which is consistent with the results of previous studies [12–14]. Further analysis in the present study found that the mortality rate of patients with positive pathogenic microbial-positive sepsis-associated ARDS was significantly higher than that of patients negative for pathogenic microorganism-related ARDS, and the 28-day mortality and 90-day mortality rates were 30.2% and 31%, respectively. This difference can be attributed to several factors. First, patients with pathogenic microorganism-positive sepsis-associated ARDS exhibited worse oxygenation levels (eg, lower PaO2/FiO2 ratios) and higher SOFA scores, indicating more severe lung injury and multi-organ dysfunction. These factors are well-established predictors of mortality in critically ill patients. Second, the presence of pathogenic microorganisms can exacerbate systemic inflammation and immune dysregulation, leading to a more severe clinical course and higher mortality. Finally, pathogenic microorganism-positive patients often require more aggressive interventions, such as broad-spectrum antibiotics and advanced respiratory support, which may not always be effective in reversing the underlying pathophysiology. These findings underscore the importance of early identification and targeted management of pathogenic microorganism-positive sepsis-associated ARDS. Future research should focus on exploring risk factors for microbial infection in sepsis-associated ARDS, as well as on developing personalized treatment strategies to improve outcomes in this high-risk population.

This study further identified that patients with pathogenic microorganism-positive sepsis-associated ARDS who did not survive had worse vital signs, blood oxygen indices, coagulation function, renal function, and lactate levels, compared with survivors. Additionally, a higher proportion of non-survivors required vasoactive drugs and CRRT, indicating more severe multisystem dysfunction. Notably, A. baumannii infection was more prevalent among non-survivors, and further analysis revealed that A. baumannii infection is an independent risk factor for mortality in patients with pathogenic microorganism-positive sepsis-associated ARDS. These findings align with previous studies demonstrating that A. baumannii infection is associated with high mortality rates due to its multidrug resistance and ability to cause severe systemic inflammation [15–18]. The high mortality rate observed in patients with A. baumannii infection underscores the need for more aggressive and targeted therapeutic strategies. For example, early identification of A. baumannii infection, combined with higher doses of sensitive antibiotics and personalized treatment regimens, can improve outcomes in this high-risk population. Furthermore, our results highlight the importance of implementing stringent infection control measures to prevent the spread of A. baumannii in ICU settings. These findings not only confirm the poor prognosis associated with A. baumannii infection but also emphasize the need for further research to explore effective interventions for managing sepsis-associated ARDS in the context of multidrug-resistant infections.

Medical models are popular for predicting the occurrence and prognosis of diseases, and many models have improved disease diagnosis and prediction [19,20]. Further findings from this study suggest that age, INR, lactate, A. baumannii infection, CRRT treatment, and SOFA were independent risk factors for mortality in patients with pathogenic microorganism-positive sepsis-associated ARDS. On the basis of this study, the above risk factors were further fitted into a model. This model had a strong predictive power for non-survival patients with pathogenic microorganism-positive sepsis-associated ARDS, which can be replicated in the clinic and further externally validated.

This study has several limitations. First, it is a retrospective study, so the results of the study do not illustrate causality, only phenomena. Second, the model has not been validated internally or externally, and further validation is required in the future before it can be applied to clinical practice. In addition, the duration of mechanical ventilation is an important indicator for patients with ARDS, but it was not included in this study. Finally, despite these limitations, the results of this study were obtained from a multicenter cohort and provide valuable reference data.

Conclusions

This study provides critical insights into the clinical characteristics and prognostic outcomes of patients with pathogenic microorganism-positive sepsis-associated ARDS. As a phenomenon study, it cannot confirm cause-and-effect relationships, but it offers important evidence of differences in the prognosis of ARDS events in patients. Our findings reveal that these patients exhibit significantly worse oxygenation indices and higher mortality rates than do those with microbiologically negative sepsis-associated ARDS. Notably, infections caused by A. baumannii were associated with particularly poor prognoses, highlighting the need for aggressive and targeted therapeutic interventions. These findings not only contribute to the growing understanding of the prognostic implications of microbiological status in sepsis-associated ARDS but also provide a foundation for future research aimed at developing more effective, personalized treatment protocols. By integrating these insights into clinical practice, healthcare providers can enhance survival rates and reduce the burden of this life-threatening condition.