Derivation and validation of generalized sepsis-induced acute respiratory failure phenotypes among critically ill patients: a retrospective study

Abstract

Background: Septic patients who develop acute respiratory failure (ARF) requiring mechanical ventilation represent a heterogenous subgroup of critically ill patients with widely variable clinical characteristics. Identifying distinct phenotypes of these patients may reveal insights about the broader heterogeneity in the clinical course of sepsis, considering multi-organ dynamics. We aimed to derive novel phenotypes of sepsis-induced ARF using observational clinical data and investigate the generalizability of the derived phenotypes.

Methods: We performed a multi-center retrospective study of ICU patients with sepsis who required mechanical ventilation for ≥ 24 h. Data from two different high-volume academic hospital centers were used, where all phenotypes were derived in MICU of Hospital-I (N = 3225). The derived phenotypes were validated in MICU of Hospital-II (N = 848), SICU of Hospital-I (N = 1112), and SICU of Hospital-II (N = 465). Clinical data from 24 h preceding intubation was used to derive distinct phenotypes using an explainable machine learning-based clustering model interpreted by clinical experts.

Results: Four distinct ARF phenotypes were identified: A (severe multi-organ dysfunction (MOD) with a high likelihood of kidney injury and heart failure), B (severe hypoxemic respiratory failure [median P/F = 123]), C (mild hypoxia [median P/F = 240]), and D (severe MOD with a high likelihood of hepatic injury, coagulopathy, and lactic acidosis). Patients in each phenotype showed differences in clinical course and mortality rates despite similarities in demographics and admission co-morbidities. The phenotypes were reproduced in external validation utilizing the MICU of Hospital-II and SICUs from Hospital-I and -II. Kaplan-Meier analysis showed significant difference in 28-day mortality across the phenotypes (p < 0.01) and consistent across MICU and SICU of both Hospital-I and -II. The phenotypes demonstrated differences in treatment effects associated with high positive end-expiratory pressure (PEEP) strategy.

Conclusion: The phenotypes demonstrated unique patterns of organ injury and differences in clinical outcomes, which may help inform future research and clinical trial design for tailored management strategies.

Keywords: Acute respiratory failure; Critical care; Phenotype; Sepsis-induced ARF; Unsupervised machine learning.

Fig. 1

Overall study pipeline of our ARF phenotyping approach showing data extraction, preprocessing, feature reduction, model development for clustering, phenotype analysis and potential usage

Fig. 2

UMAP projections showing distribution of the derivation data for ARF phenotyping and feature variations. a UMAP representing all ARF phenotypes, and b–f UMAP representations showing variabilities in P/F ratio (mmHg), bilirubin total (mg/dL), creatinine (mg/dL), BNP (pg/mL), and platelets (× 10³/µL), respectively

Fig. 3

Visualization of feature variations across different ARF phenotypes of the derivation data. a–f boxplot representations of different phenotypes showing variabilities in creatinine (mg/dL), bilirubin total (mg/dL), P/F ratio (mmHg), FiO₂, BNP (pg/mL), and platelets (× 10³/µL), respectively to show fitness of individual organs. Presented phenotyping results were significant (p < 0.001, Kruskal–Wallis test) for each of the variables

Fig. 4

Kaplan–Meier curve showing 28-day survival rates for ARF patients stratified by phenotypes developed for a derivation set: Emory MICU data, and b validation set: Grady MICU data, c validation set: Emory SICU data, and d validation set: Grady SICU data. Survival was analyzed from the time of intubation. Survival probabilities (solid line) with their confidence intervals (faded region) are represented for different phenotypes via different color-codes

Fig. 5

Kaplan–Meier survival curves showing treatment effects of high PEEP (PEEP ≥ 10) and low PEEP (< 10) regimes on propensity matched ARF phenotypes from the derivation dataset. Survival probabilities (solid line) with their 95% confidence intervals (faded region) are represented for treated and untreated cohorts within each phenotype