A Nomogram for Early Prediction of Inflammation, Catabolism, and Immunosuppression Syndrome in Critically Ill Patients

TL;DR

This study creates a tool to predict a dangerous syndrome in ICU patients using nine admission factors, which could help identify low-risk patients.

Contribution

The first nomogram for predicting inflammation-immunosuppression-catabolism syndrome in ICU patients using admission variables.

Findings

A nomogram with nine predictors (age, BMI, SOFA score, etc.) was developed and validated for ICS prediction.

The nomogram showed good discrimination with C-indices of 0.763 in training and 0.735 in validation sets.

The tool reliably identifies low-risk ICU patients with a negative predictive value of 0.87.

Abstract

Background: Chronic critical illness (CCI) affects ~7.6% of ICU patients worldwide and is associated with poor outcomes, including 25% in-hospital and 50% one-year mortality. A proposed key mechanism is the inflammation-immunosuppression-catabolism (ICS) triad, which contributes to multiple organ failure and independently increases mortality. Although early identification of ICS could improve risk stratification, no clinically applicable predictive tool currently exists. This study aimed to develop and validate a prognostic nomogram to predict ICS development in ICU (Intensive Care Unit) patients. Methods: This real-world analysis used electronic health records from the Russian Intensive Care Dataset (RICD). ICS was defined as C-reactive protein > 20 mg/L, albumin < 30 g/L, and lymphocyte count < 0.8 × 109/L. Variables with >30% missing data were excluded, and remaining missing values were handled by multiple imputation. A Cox proportional hazards model was used to construct the nomogram. Internal validation was performed using an 8:2 training–validation split. Results: Among 1963 eligible patients, 540 (27.5%) developed ICS. LASSO (Least Absolute Shrinkage and Selection Operator) regression identified nine significant predictors: age, body mass index, SOFA (Sequential Organ Failure Assessment) and FOUR (Full Outline of UnResponsiveness) scores at admission, pneumonia and anemia at admission, platelet count, total protein, and creatinine. The nomogram showed good discrimination, with C-indices of 0.763 (95% CI: 0.741–0.783) in the training set and 0.735 (95% CI: 0.689–0.784) in the validation set. At the optimal cutoff, sensitivity was 0.75, specificity was 0.63, positive predictive value was 0.43, and negative predictive value was 0.87. Conclusions: This study presents the first nomogram for predicting ICS in ICU patients, using nine admission variables to reliably identify low-risk individuals. Further external validation is required.

Figures