Prolonged activated partial thromboplastin time (APTT) is a frequent laboratory finding that may reflect a broad spectrum of underlying conditions, ranging from benign laboratory abnormalities to clinically relevant hemostatic disorders. Clot waveform analysis (CWA), automatically generated during routine APTT testing by optical coagulation analyzers, provides additional quantitative and qualitative information on clot formation dynamics. The APTTO model is a previously developed two-step predictive algorithm based on CWA features designed to estimate the probability of a pathological cause of prolonged APTT and to differentiate lupus anticoagulant from intrinsic pathway factor deficiency or von Willebrand disease. Internal validation has demonstrated good discrimination and calibration. This multicenter observational study aims to perform an external validation of the APTTO model in independent patient cohorts, assessing its discrimination, calibration, and decision-analytic performance without model updating.
This multicenter observational cohort study is designed to externally validate the APTTO predictive model in patients with prolonged APTT and normal prothrombin time evaluated in routine clinical practice across multiple hospitals. All laboratory data, including CWA parameters and waveform morphology, are generated as part of standard diagnostic workflows. No additional blood sampling, laboratory testing, or modifications to clinical management are introduced for research purposes. The study focuses on evaluating model performance in independent cohorts by applying the original APTTO model coefficients and predefined cut-offs without recalibration or re-estimation. Model discrimination, calibration, and decision-analytic measures will be assessed. Secondary analyses will explore model performance across predefined subgroups and analytical robustness. This study adheres to the TRIPOD statement for validation of prediction models.
Study Type
OBSERVATIONAL
Enrollment
1,500
Application of the APTTO predictive models (APTTO1 and APTTO2) to clot waveform analysis parameters generated during routine activated partial thromboplastin time testing, for research purposes only. The model output does not influence clinical management or surgical decision-making during the study.
Hospital Universitario Severo Ochoa
Leganés, Madrid, Spain
NOT_YET_RECRUITINGHospital de la Santa Creu i Sant Pau
Barcelona, Spain
NOT_YET_RECRUITINGHospital Universitario Arnau De Vilanova
Lleida, Spain
NOT_YET_RECRUITINGHospital Universitario 12 de Octubre
Madrid, Spain
NOT_YET_RECRUITINGHospital Universitario Clínico San Carlos
Madrid, Spain
NOT_YET_RECRUITINGHospital Universitario Fundación Jiménez Díaz
Madrid, Spain
RECRUITINGHospital Universitario Gregorio Marañón
Madrid, Spain
NOT_YET_RECRUITINGHospital Universitario Puerta de Hierro
Madrid, Spain
NOT_YET_RECRUITINGHospital Universitario Ramón y Cajal
Madrid, Spain
NOT_YET_RECRUITINGHospital Clínico Universitario Virgen de la Arrixaca
Murcia, Spain
NOT_YET_RECRUITING...and 6 more locations
Discriminatory performance of the APTTO model
Discrimination of the APTTO1 and APTTO2 models for identifying the cause of prolonged activated partial thromboplastin time (APTT), assessed by the area under the receiver operating characteristic curve (AUC) in an independent multicenter cohort.
Time frame: Baseline (at the time of prolonged APTT laboratory assessment)
Calibration-in-the-large of the APTTO model
Calibration-in-the-large of the APTTO1 and APTTO2 models assessing agreement between predicted and observed probabilities in an independent multicenter cohort.
Time frame: Baseline (at the time of prolonged APTT laboratory assessment)
Calibration slope of the APTTO model
Calibration slope of the APTTO1 and APTTO2 models, evaluating the relationship between predicted and observed risk across the probability spectrum.
Time frame: Baseline (at the time of prolonged APTT laboratory assessment)
Overall prediction error of the APTTO model
Overall prediction error of the APTTO1 and APTTO2 models, assessed using the Brier score.
Time frame: Baseline (at the time of prolonged APTT laboratory assessment)
Clinical utility of the APTTO model assessed by decision curve analysis
Net clinical benefit of the APTTO models compared with investigate-all and investigate-none strategies, assessed by decision curve analysis across clinically plausible threshold probabilities.
Time frame: Baseline (at the time of prolonged APTT laboratory assessment)
Diagnostic accuracy of predefined APTTO cut-offs
Sensitivity, specificity, positive predictive value, and negative predictive value of predefined APTTO cut-offs for identifying pathological causes of prolonged APTT.
Time frame: Baseline (at the time of prolonged APTT laboratory assessment)
Interobserver agreement in clot waveform morphology classification
Agreement between local investigators and central reader (blinded to clinical data) for qualitative clot waveform morphology classification, assessed using Cohen´s kappa or Fleiss´kappa statistics, as appropriate.
Time frame: Baseline (at the time of prolonged APTT laboratory assessment)
Estimated impact of the APTTO algorithm on preoperative workflow
Estimated potential reduction in time to surgical cleareance and avoidance of additional etiologic testing based on application of the APTTO algorithm.
Time frame: Baseline (model-based estimation using timing data from index APTT laboratory assessment through surgical clearance)
This platform is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional.