EURO-SCORE study

EURO-SCORE study

The EURO-SCORE study produced a widely cited risk model for predicting operative mortality after adult cardiac surgery, originating from a multicenter European collaboration involving centers such as Royal Brompton Hospital, Papworth Hospital, Hôpital Européen Georges-Pompidou, Karolinska University Hospital and institutions connected to European Association for Cardio-Thoracic Surgery and Society of Thoracic Surgeons. Developed in the late 1990s during an era marked by comparative initiatives like the Framingham Heart Study and linked to registries similar to STS National Database and National Adult Cardiac Surgery Audit, the study influenced practice across NHS, INSERM, Azienda Ospedaliera, and other European healthcare centers.

Background and development

The project emerged from collaborative networks involving clinicians from United Kingdom, France, Sweden, Italy, Germany, and Spain with input from centers including Royal Brompton Hospital, Papworth Hospital, Hôpital Européen Georges-Pompidou, Karolinska University Hospital, and Universitätsklinikum Heidelberg. Influenced by risk modeling precedents such as the Framingham Heart Study and initiatives in the United States like the Society of Thoracic Surgeons database, investigators aimed to create a parsimonious tool suitable for perioperative audit across diverse systems including the NHS and continental European healthcare institutions. The development team performed retrospective data collection from multiple centers, engaging clinical leads and registries akin to EuroSCORE Community partners and drawing methodological inspiration from statistical work at institutions like Imperial College London and University of Cambridge.

Methodology and scoring system

Derivation used logistic regression on pooled data from adult patients undergoing cardiac surgery across participating centers, following conventions seen in predictive work from Harvard Medical School, Johns Hopkins Hospital, Mayo Clinic, and Cleveland Clinic. Candidate predictors included demographic and clinical covariates recorded in surgical records from centers such as Royal Brompton Hospital, Papworth Hospital, Karolinska University Hospital, Hôpital Européen Georges-Pompidou and variables analogous to those used in databases like the STS National Database. The original additive and logistic versions produced point-based scores, allowing bedside calculation similar in intent to tools like the CHADS2 score and prognostic indices developed at Massachusetts General Hospital. The scoring system stratified risk across factors including age, comorbidity profiles recorded in charts from centers linked to University College London Hospitals NHS Foundation Trust and anatomical and procedural variables reflecting practices taught at institutions such as Guy's Hospital and St Thomas' Hospital.

Validation and performance

External validation exercises compared predicted versus observed mortality in datasets assembled by groups at Royal Brompton Hospital, Papworth Hospital, Universitätsklinikum Heidelberg, and national registries like those coordinated by National Institute for Health and Care Excellence and agencies analogous to Agence technique de l'information sur l'hospitalisation. Performance assessment used calibration and discrimination metrics, producing results similar to validation studies for models from Framingham Heart Study, PREDICT registry, and STS risk models. Subsequent recalibration efforts—akin to model updates performed at institutions such as University of Toronto and Duke University Hospital—yielded revised euroSCORE II–style algorithms to address shifts in practice documented in audits from NHS and registries in Italy and Germany.

Clinical applications and impact

Clinicians at centers including Royal Brompton Hospital, Papworth Hospital, Karolinska University Hospital, and Hôpital Européen Georges-Pompidou adopted the score for preoperative counseling, audit, and benchmarking, paralleling uses of the STS risk calculator and prognostic tools from Mayo Clinic. Health technology assessment bodies and guideline committees in organizations like National Institute for Health and Care Excellence and professional societies such as European Association for Cardio-Thoracic Surgery referenced the model when framing recommendations and comparing institutional outcomes, similar to how evidence from the Framingham Heart Study informs policy. The score influenced resource allocation decisions in settings including NHS Trusts, university hospitals tied to University of Oxford, University of Cambridge, and national cardiac networks across France, Sweden, and Italy.

Limitations and controversies

Critiques mirrored concerns raised in debates surrounding predictive models from Framingham Heart Study and registries like the STS National Database: limited generalizability to populations outside the derivation cohort such as patients in United States tertiary centers like Cleveland Clinic and varying performance in subgroups including those treated at specialized centers like Mayo Clinic. Methodological controversies involved case-mix adjustment, coding variability across hospitals such as Royal Brompton Hospital and Papworth Hospital, and temporal drift requiring recalibration similar to updates implemented by teams at University of Toronto and Duke University Hospital. Ethical and policy discussions engaged stakeholders from European Association for Cardio-Thoracic Surgery, national audit agencies, and guideline bodies like National Institute for Health and Care Excellence regarding appropriate use in high-stakes individual decision-making versus population-level audit.

Category:Cardiac surgery studies