dIPC-RFI provides a simple 24-hour postoperative framework for identifying incomplete recovery across inflammation, perfusion, and coagulation after ATAAD surgery, achieving an AUC of 0.873 in temporal validation and outperforming simplified and extended clinical models.
Key Findings
Results
Severe organ dysfunction after 24 hours occurred in 40.6% of ATAAD surgical patients overall, with similar rates in derivation and validation cohorts.
330 total patients were included from The University of Hong Kong-Shenzhen Hospital between August 2021 and December 2025.
134 of 330 patients (40.6%) developed severe organ dysfunction after 24 hours.
Event rate was 39.0% in the derivation cohort (first 70%) and 44.4% in the temporal validation cohort (last 30%).
The difference in event rates between cohorts was not statistically significant (p = 0.392).
Only patients alive, free from the primary outcome, and with complete variables at 24 hours were included.
Results
The dIPC-RFI achieved good discrimination for severe organ dysfunction in the temporal validation cohort with an AUC of 0.873.
AUC in the validation cohort was 0.873 (95% CI, 0.794–0.935).
Brier score in the validation cohort was 0.1387.
dIPC-RFI outperformed the simplified clinical model (AUC = 0.762) and the extended clinical model (AUC = 0.735).
Model performance was assessed using AUC, Brier score, calibration curve, calibration intercept and slope, and decision curve analysis.
Results
A combined model integrating dIPC-RFI with the simplified clinical model further improved discrimination and calibration.
The combined model achieved an AUC of 0.908 in the validation cohort.
The Brier score for the combined model was reduced to 0.1243.
This represented improvement over dIPC-RFI alone (AUC 0.873, Brier score 0.1387) and the simplified clinical model alone (AUC 0.762).
Methods
The dIPC-RFI is composed of three domains—inflammation, perfusion, and coagulation—each including one static 24-hour variable and one dynamic change variable.
Variables were drawn from routinely available preoperative, immediate postoperative, and 24-hour postoperative laboratory and blood gas measurements.
Cutoffs for each component were derived using receiver operating characteristic curves and the Youden index in the derivation cohort, then fixed for the validation cohort.
The index was designed to capture early postoperative pathophysiological recovery that preoperative risk scores alone may not reflect.
The index integrates inflammatory, perfusion, and coagulation variables across three time points: before surgery, immediately after surgery, and 24 hours after surgery.
Results
Component-specific analyses showed acceptable discrimination for major outcome components including CRRT, prolonged mechanical ventilation, neurological complications, and MODS.
Components analyzed included continuous renal replacement therapy (CRRT), mechanical ventilation for ≥72 hours, neurological complications, and multiple organ dysfunction syndrome (MODS).
Acceptable discrimination was noted for components with sufficient validation events.
The primary outcome was defined as any of: CRRT, mechanical ventilation ≥72 hours, neurological complications, MODS, in-hospital death after 24 hours or 30-day death, intra-aortic balloon pump support, ECMO, or tracheostomy.
Results
Sensitivity analyses supported the robustness of the main dIPC-RFI findings.
A no-CRRT sensitivity analysis was performed, which supported the robustness of the main findings.
A death-or-MODS sensitivity analysis was also conducted and supported the main findings.
These analyses were conducted to address potential concerns about specific outcome components driving overall model performance.
Conclusions
The dIPC-RFI is intended as a risk-restratification aid at 24 hours postoperatively, providing low-, intermediate-, and high-risk strata, not as a stand-alone treatment trigger.
The score provides clear low-, intermediate-, and high-risk strata in the temporal validation cohort.
It is intended to help ICU teams identify patients requiring intensified monitoring, repeated lactate and coagulation assessment, early multidisciplinary review, and preparation for organ support.
The authors explicitly state it should not be used for preoperative decision-making.
Prospective multicenter validation is required before routine clinical implementation.
What This Means
This research suggests that after emergency surgery for acute type A aortic dissection—a life-threatening tear in the main artery from the heart—identifying which patients will deteriorate remains difficult using only pre-surgery risk assessments. Researchers developed and tested a new scoring tool called the dIPC-RFI (dynamic inflammation-perfusion-coagulation recovery failure index) that uses routine blood test results collected before surgery, immediately after surgery, and 24 hours after surgery to predict which patients are at risk for serious complications such as kidney failure requiring dialysis, prolonged dependence on a breathing machine, brain complications, or death. In a group of 330 patients treated at a single hospital over about four years, roughly 40% experienced severe organ complications after the first 24 hours. The new scoring tool correctly identified high-risk patients far better than existing clinical models, achieving strong discriminatory performance (AUC of 0.873) in the validation group, and performed even better when combined with a simple clinical model (AUC of 0.908).
This research suggests that tracking how a patient's inflammation, blood flow (perfusion), and clotting (coagulation) markers change in the first 24 hours after surgery—rather than relying only on pre-surgery information—can meaningfully improve the ability to predict who is heading toward serious organ failure. The tool is built entirely from tests already routinely performed in intensive care units, meaning it could potentially be applied without requiring new equipment or procedures. The authors emphasize that this score is designed to help ICU teams decide who needs closer monitoring or earlier preparation for organ support, not to replace clinical judgment or trigger automatic treatment decisions.
Importantly, the study was conducted at a single center in China, and the authors caution that prospective testing across multiple hospitals is needed before the tool could be adopted broadly. The study design was retrospective, meaning it looked back at existing patient records rather than prospectively following patients in a planned trial. The findings are promising for improving post-surgical care of a very high-risk patient population, but further validation will be essential to confirm whether the tool performs as well in other clinical settings and patient populations.
Yuan S, Li L, Wei M, Wang Z, Zhang R, Ma Y, et al.. (2026). A dynamic recovery failure index after acute type A aortic dissection surgery.. Annals of medicine. https://doi.org/10.1080/07853890.2026.2724602