A Molecular Host Response Assay to Discriminate Between Sepsis and Infection-Negative Systemic Inflammation in Critically Ill Patients: Discovery and Validation in Independent Cohorts

Abstract

Background: Systemic inflammation is a whole body reaction having an infection-positive (i.e., sepsis) or infection-negative origin. It is important to distinguish between these two etiologies early and accurately because this has significant therapeutic implications for critically ill patients. We hypothesized that a molecular classifier based on peripheral blood RNAs could be discovered that would (1) determine which patients with systemic inflammation had sepsis, (2) be robust across independent patient cohorts, (3) be insensitive to disease severity, and (4) provide diagnostic utility. The goal of this study was to identify and validate such a molecular classifier.

Methods and findings: We conducted an observational, non-interventional study of adult patients recruited from tertiary intensive care units (ICUs). Biomarker discovery utilized an Australian cohort (n = 105) consisting of 74 cases (sepsis patients) and 31 controls (post-surgical patients with infection-negative systemic inflammation) recruited at five tertiary care settings in Brisbane, Australia, from June 3, 2008, to December 22, 2011. A four-gene classifier combining CEACAM4, LAMP1, PLA2G7, and PLAC8 RNA biomarkers was identified. This classifier, designated SeptiCyte Lab, was validated using reverse transcription quantitative PCR and receiver operating characteristic (ROC) curve analysis in five cohorts (n = 345) from the Netherlands. Patients for validation were selected from the Molecular Diagnosis and Risk Stratification of Sepsis study (ClinicalTrials.gov, NCT01905033), which recruited ICU patients from the Academic Medical Center in Amsterdam and the University Medical Center Utrecht. Patients recruited from November 30, 2012, to August 5, 2013, were eligible for inclusion in the present study. Validation cohort 1 (n = 59) consisted entirely of unambiguous cases and controls; SeptiCyte Lab gave an area under curve (AUC) of 0.95 (95% CI 0.91-1.00) in this cohort. ROC curve analysis of an independent, more heterogeneous group of patients (validation cohorts 2-5; 249 patients after excluding 37 patients with an infection likelihood of "possible") gave an AUC of 0.89 (95% CI 0.85-0.93). Disease severity, as measured by Sequential Organ Failure Assessment (SOFA) score or Acute Physiology and Chronic Health Evaluation (APACHE) IV score, was not a significant confounding variable. The diagnostic utility of SeptiCyte Lab was evaluated by comparison to various clinical and laboratory parameters available to a clinician within 24 h of ICU admission. SeptiCyte Lab was significantly better at differentiating cases from controls than all tested parameters, both singly and in various logistic combinations, and more than halved the diagnostic error rate compared to procalcitonin in all tested cohorts and cohort combinations. Limitations of this study relate to (1) cohort compositions that do not perfectly reflect the composition of the intended use population, (2) potential biases that could be introduced as a result of the current lack of a gold standard for diagnosing sepsis, and (3) lack of a complete, unbiased comparison to C-reactive protein.

Conclusions: SeptiCyte Lab is a rapid molecular assay that may be clinically useful in managing ICU patients with systemic inflammation. Further study in population-based cohorts is needed to validate this assay for clinical use.

Fig 1. Flow diagram for selection of patients composing the validation cohorts.

From the MARS study, only patients admitted to the ICU between the dates of November 30, 2012, and August 5, 2013, were eligible for possible inclusion in the present study. A detailed description of the inclusion and exclusion criteria, and the classification algorithm, is given in S2 Text. In the last step of the cohort selection process as described in this figure, 14 patients were reassigned as controls because the attending physicians had retrospectively adjudicated the patients to have an infection likelihood of none for their sepsis events. S.I., systemic inflammation.

Fig 2. Analysis of behavior of PLAC8, LAMP1, PLA2G7, and CEACAM4 in the discovery cohort.

(A) Heat map representation of the discovery cohort (74 cases, 31 controls). Normalized expression levels of the individual genes comprising the SeptiCyte Lab classifier (color) are plotted versus disease status (dendrogram position) using unsupervised clustering with equally weighted Euclidean distance. The normalization scale (expression level Z-score) for up-regulation (red) or down-regulation (green) is shown in the insert at the left of the heat map. (The Z-score is the number of standard deviations a value lies away from the mean. Higher absolute Z-scores correspond to lower p-values. A Z-score of ±1.96 equates to a p-value of 0.05 in a two-tailed test.) In the cases (sepsis), two genes are predominately up-regulated (PLAC8 and LAMP1), whilst two are predominantly down regulated (PLA2G7 and CEACAM4). (B) Scatterplot representation of microarray expression levels for individual genes in the SeptiCyte Lab classifier, for the discovery cohort. The expression level on log₂ scale (y-axis) is presented for PLAC8, CEACAM4, LAMP1, and PLA2G7 in individual patients (red for cases, black for controls). Each gene contributes to the ability of the SeptiCyte Lab classifier to separate the cases and controls. S.I., systemic inflammation.

Fig 3. Performance of SeptiCyte Lab in validation cohort 1.

Left: Scatterplot of SeptiScores for 24 cases (sepsis; red) versus 35 controls (infection-negative systemic inflammation [S.I.]; black). The blue dashed lines denote SeptiScore values of 3.1, 4.0, 6.0, and 9.0, which are used for subsequent calculations. Right: ROC curve. The grey shading denotes the 95% confidence area.

Fig 4. Performance of SeptiCyte Lab in validation cohorts 3, 4, and 5.

In each panel, the left side presents a scatterplot of SeptiCyte Lab scores for cases (sepsis; red), controls (infection-negative systemic inflammation [S.I.]; black), and individuals with an infection likelihood of possible (green), and the right side presents the corresponding ROC curve. In the scatterplots, the blue dashed lines denote SeptiScore values of 3.1, 4.0, 6.0, and 9.0, which are used for subsequent calculations. On the right, the grey shading denotes the 95% confidence area for the ROC curve. Patients with an infection likelihood of possible have been excluded from the calculation of ROC curves. (A) Validation cohort 3 (29 cases, 77 controls). (B) Validation cohort 4 (20 cases, 47 controls), consisting of patients sequentially admitted to the ICU. (C) Validation cohort 5 (21 cases, 25 controls), consisting of black and Asian patients sequentially admitted to the ICU.

Fig 5. Test for disease severity as a potential confounding variable.

Validation cohorts 1–5 (excluding patients with an infection likelihood of possible) were combined, and then stratified on APACHE IV score or SOFA score. (A) ROC curve analysis was performed on the combined dataset. (B) Separate ROC curve analyses were performed over sub-ranges of the APACHE IV score (left) or SOFA score (right). The error bars indicate the 95% CI for the AUC as estimated by resampling. The results for the various strata of APACHE IV and SOFA score indicate that disease severity has a relatively weak influence (AUC = 0.52–0.65) on the discrimination of cases from controls, while SeptiCyte Lab has a much stronger influence (AUC = 0.87–0.91). Additional details of the calculations are given in S5 Data.

Fig 6. Performance of PCT, clinical parameters, and SeptiCyte Lab for the discrimination of sepsis versus infection-negative systemic inflammation.

This comparative analysis used the largest set of patients (n = 157) for which values were available for all comparisons. Patients having an infection likelihood of possible were excluded. Performance is reported as AUC. The error bars indicate the 95% CI for the AUC, obtained from a 50 × 2 cross-validation with training and test samples selected in a 1:1 ratio. The significance levels (p-value by t-test) for pairwise comparisons to SeptiCyte Lab are indicated in the figure. Details of the analysis are given in S7 Data. NS, not significant; Top 3, top three clinical parameters; Top 5, top five clinical parameters.