Development Of Clinical Prediction Models Research Articles

Harnessing the power of proteomics in precision diabetes medicine.

Precision diabetes medicine (PDM) aims to reduce errors in prevention programmes, diagnosis thresholds, prognosis prediction and treatment strategies. However, its advancement and implementation are difficult due to the heterogeneity of complex molecular processes and environmental exposures that influence an individual's disease trajectory. To address this challenge, it is imperative to develop robust screening methods for all areas of PDM. Innovative proteomic technologies, alongside genomics, have proven effective in precision cancer medicine and are showing promise in diabetes research for potential translation. This narrative review highlights how proteomics is well-positioned to help improve PDM. Specifically, a critical assessment of widely adopted affinity-based proteomic technologies in large-scale clinical studies and evidence of the benefits and feasibility of using MS-based plasma proteomics is presented. We also present a case for the use of proteomics to identify predictive protein panels for type 2 diabetes subtyping and the development of clinical prediction models for prevention, diagnosis, prognosis and treatment strategies. Lastly, we discuss the importance of plasma and tissue proteomics and its integration with genomics (proteogenomics) for identifying unique type 2 diabetes intra- and inter-subtype aetiology. We conclude with a call for action formed on advancing proteomics technologies, benchmarking their performance and standardisation across sites, with an emphasis on data sharing and the inclusion of diverse ancestries in large cohort studies. These efforts should foster collaboration with key stakeholders and align with ongoing academic programmes such as the Precision Medicine in Diabetes Initiative consortium.

Predicting lung cancer survival prognosis based on the conditional survival bayesian network.

Lung cancer is a leading cause of cancer deaths and imposes an enormous economic burden on patients. It is important to develop an accurate risk assessment model to determine the appropriate treatment for patients after an initial lung cancer diagnosis. The Cox proportional hazards model is mainly employed in survival analysis. However, real-world medical data are usually incomplete, posing a great challenge to the application of this model. Commonly used imputation methods cannot achieve sufficient accuracy when data are missing, so we investigated novel methods for the development of clinical prediction models. In this article, we present a novel model for survival prediction in missing scenarios. We collected data from 5,240 patients diagnosed with lung cancer at the Weihai Municipal Hospital, China. Then, we applied a joint model that combined a BN and a Cox model to predict mortality risk in individual patients with lung cancer. The established prognostic model achieved good predictive performance in discrimination and calibration. We showed that combining the BN with the Cox proportional hazards model is highly beneficial and provides a more efficient tool for risk prediction.

Development of clinical prediction models for renal and cardiovascular outcomes and mortality in patients with type 2 diabetes and chronic kidney disease using time-varying predictors

AimsTo develop a set of prediction models for end-stage kidney disease (ESKD), cardiovascular outcomes, and mortality in patients with type 2 diabetes (T2D) and chronic kidney disease (CKD) using commonly measured clinical variables. MethodsWe studied 1432 participants with T2D and CKD enrolled in the Chronic Renal Insufficiency Cohort, followed for a median period of 7 years. We used Cox proportional-hazards models to model the six outcomes (ESKD, stroke, myocardial infarction (MI), congestive heart failure (CHF), death before ESKD, and all-cause mortality). We internally evaluated these models using concordance and calibration measures. ResultsThe newly developed six prediction models included 15 predictors: age at diabetes diagnosis, sex, blood pressure, body mass index, hemoglobin A1c, high density lipoprotein cholesterol, urine protein-to-creatinine ratio, estimated glomerular filtration rate, smoking status, and history of stroke, MI, CHF, ESKD, and amputation. The resulting models demonstrated good/strong discrimination (cross-validation C-index range: 0.70 to 0.90) and calibration. ConclusionsThis study provided an internally validated and useful tool for predicting individual adverse outcomes and mortality in patients with T2D and CKD. These models may inform optimal use of targeted health interventions.

Updating Clinical Prediction Models: An Illustrative Case Study.

The performance of clinical prediction models tends to deteriorate over time. Researchers often develop a new prediction if an existing model performs poorly at external validation. Model updating is an efficient technique and promising alternative to the de novo development of clinical prediction models. Model updating has been recommended by the TRIPOD guidelines. To illustrate several model updating techniques, a case study is provided for the development and updating of a clinical prediction model assessing postoperative anxiety in data coming from two double-blinded placebo-controlled randomized controlled trials with a very similar methodological framework. Note that the developed model and updated model are for didactic purposes only. This paper discusses some common considerations and caveats for researchers to be aware of when planning or applying updating of a prediction model.

Adaptive sample size determination for the development of clinical prediction models

BackgroundWe suggest an adaptive sample size calculation method for developing clinical prediction models, in which model performance is monitored sequentially as new data comes in.MethodsWe illustrate the approach using data for the diagnosis of ovarian cancer (n = 5914, 33% event fraction) and obstructive coronary artery disease (CAD; n = 4888, 44% event fraction). We used logistic regression to develop a prediction model consisting only of a priori selected predictors and assumed linear relations for continuous predictors. We mimicked prospective patient recruitment by developing the model on 100 randomly selected patients, and we used bootstrapping to internally validate the model. We sequentially added 50 random new patients until we reached a sample size of 3000 and re-estimated model performance at each step. We examined the required sample size for satisfying the following stopping rule: obtaining a calibration slope ≥ 0.9 and optimism in the c-statistic (or AUC) < = 0.02 at two consecutive sample sizes. This procedure was repeated 500 times. We also investigated the impact of alternative modeling strategies: modeling nonlinear relations for continuous predictors and correcting for bias on the model estimates (Firth’s correction).ResultsBetter discrimination was achieved in the ovarian cancer data (c-statistic 0.9 with 7 predictors) than in the CAD data (c-statistic 0.7 with 11 predictors). Adequate calibration and limited optimism in discrimination was achieved after a median of 450 patients (interquartile range 450–500) for the ovarian cancer data (22 events per parameter (EPP), 20–24) and 850 patients (750–900) for the CAD data (33 EPP, 30–35). A stricter criterion, requiring AUC optimism < = 0.01, was met with a median of 500 (23 EPP) and 1500 (59 EPP) patients, respectively. These sample sizes were much higher than the well-known 10 EPP rule of thumb and slightly higher than a recently published fixed sample size calculation method by Riley et al. Higher sample sizes were required when nonlinear relationships were modeled, and lower sample sizes when Firth’s correction was used.ConclusionsAdaptive sample size determination can be a useful supplement to fixed a priori sample size calculations, because it allows to tailor the sample size to the specific prediction modeling context in a dynamic fashion.

Risk Factors Associated With All-Cause 30-Day Mortality in Nursing Home Residents With COVID-19

The coronavirus disease 2019 (COVID-19) pandemic has severely affected nursing homes. Vulnerable nursing home residents are at high risk for adverse outcomes, but improved understanding is needed to identify risk factors for mortality among nursing home residents. To identify risk factors for 30-day all-cause mortality among US nursing home residents with COVID-19. This cohort study was conducted at 351 US nursing homes among 5256 nursing home residents with COVID-19-related symptoms who had severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection confirmed by polymerase chain reaction testing between March 16 and September 15, 2020. Resident-level characteristics, including age, sex, race/ethnicity, symptoms, chronic conditions, and physical and cognitive function. Death due to any cause within 30 days of the first positive SARS-CoV-2 test result. The study included 5256 nursing home residents (3185 women [61%]; median age, 79 years [interquartile range, 69-88 years]; and 3741 White residents [71%], 909 Black residents [17%], and 586 individuals of other races/ethnicities [11%]) with COVID-19. Compared with residents aged 75 to 79 years, the odds of death were 1.46 (95% CI, 1.14-1.86) times higher for residents aged 80 to 84 years, 1.59 (95% CI, 1.25-2.03) times higher for residents aged 85 to 89 years, and 2.14 (95% CI, 1.70-2.69) times higher for residents aged 90 years or older. Women had lower risk for 30-day mortality than men (odds ratio [OR], 0.69 [95% CI, 0.60-0.80]). Two comorbidities were associated with mortality: diabetes (OR, 1.21 [95% CI, 1.05-1.40]) and chronic kidney disease (OR, 1.33 [95%, 1.11-1.61]). Fever (OR, 1.66 [95% CI, 1.41-1.96]), shortness of breath (OR, 2.52 [95% CI, 2.00-3.16]), tachycardia (OR, 1.31 [95% CI, 1.04-1.64]), and hypoxia (OR, 2.05 [95% CI, 1.68-2.50]) were also associated with increased risk of 30-day mortality. Compared with cognitively intact residents, the odds of death among residents with moderate cognitive impairment were 2.09 (95% CI, 1.68-2.59) times higher, and the odds of death among residents with severe cognitive impairment were 2.79 (95% CI, 2.14-3.66) times higher. Compared with residents with no or limited impairment in physical function, the odds of death among residents with moderate impairment were 1.49 (95% CI, 1.18-1.88) times higher, and the odds of death among residents with severe impairment were 1.64 (95% CI, 1.30-2.08) times higher. In this cohort study of US nursing home residents with COVID-19, increased age, male sex, and impaired cognitive and physical function were independently associated with mortality. Understanding these risk factors can aid in the development of clinical prediction models of mortality in this population.

Fostering reproducibility and generalizability in machine learning for clinical prediction modeling in spine surgery

As the use of machine learning algorithms in the development of clinical prediction models has increased, researchers are becoming more aware of the deleterious effects that stem from the lack of reporting standards. One of the most obvious consequences is the insufficient reproducibility found in current prediction models. In an attempt to characterize methods to improve reproducibility and to allow for better clinical performance, we utilize a previously proposed taxonomy that separates reproducibility into 3 components: technical, statistical, and conceptual reproducibility. By following this framework, we discuss common errors that lead to poor reproducibility, highlight the importance of generalizability when evaluating a ML model's performance, and provide suggestions to optimize generalizability to ensure adequate performance. These efforts are a necessity before such models are applied to patient care.

Early detection of type 2 diabetes mellitus using machine learning-based prediction models

Most screening tests for T2DM in use today were developed using multivariate regression methods that are often further simplified to allow transformation into a scoring formula. The increasing volume of electronically collected data opened the opportunity to develop more complex, accurate prediction models that can be continuously updated using machine learning approaches. This study compares machine learning-based prediction models (i.e. Glmnet, RF, XGBoost, LightGBM) to commonly used regression models for prediction of undiagnosed T2DM. The performance in prediction of fasting plasma glucose level was measured using 100 bootstrap iterations in different subsets of data simulating new incoming data in 6-month batches. With 6 months of data available, simple regression model performed with the lowest average RMSE of 0.838, followed by RF (0.842), LightGBM (0.846), Glmnet (0.859) and XGBoost (0.881). When more data were added, Glmnet improved with the highest rate (+ 3.4%). The highest level of variable selection stability over time was observed with LightGBM models. Our results show no clinically relevant improvement when more sophisticated prediction models were used. Since higher stability of selected variables over time contributes to simpler interpretation of the models, interpretability and model calibration should also be considered in development of clinical prediction models.

Predicting the Risk of Recurrent Venous Thromboembolism: Current Challenges and Future Opportunities.

Acute venous thromboembolism (VTE) is a commonly diagnosed condition and requires treatment with anticoagulation to reduce the risk of embolisation as well as recurrent venous thrombotic events. In many cases, cessation of anticoagulation is associated with an unacceptably high risk of recurrent VTE, precipitating the use of indefinite anticoagulation. In contrast, however, continuing anticoagulation is associated with increased major bleeding events. As a consequence, it is essential to accurately predict the subgroup of patients who have the highest probability of experiencing recurrent VTE, so that treatment can be appropriately tailored to each individual. To this end, the development of clinical prediction models has aided in calculating the risk of recurrent thrombotic events; however, there are several limitations with regards to routine use for all patients with acute VTE. More recently, focus has shifted towards the utility of novel biomarkers in the understanding of disease pathogenesis as well as their application in predicting recurrent VTE. Below, we review the current strategies used to predict the development of recurrent VTE, with emphasis on the application of several promising novel biomarkers in this field.

Dynamic clinical prediction models for discrete time-to-event data with competing risks-A case study on the OUTCOMEREA database.

The development of clinical prediction models requires the selection of suitable predictor variables. Techniques to perform objective Bayesian variable selection in the linear model are well developed and have been extended to the generalized linear model setting as well as to the Cox proportional hazards model. Here, we consider discrete time-to-event data with competing risks and propose methodology to develop a clinical prediction model for the daily risk of acquiring a ventilator-associated pneumonia (VAP) attributed to P. aeruginosa (PA) in intensive care units. The competing events for a PA VAP are extubation, death, and VAP due to other bacteria. Baseline variables are potentially important to predict the outcome at the start of ventilation, but may lose some of their predictive power after a certain time. Therefore, we use a landmark approach for dynamic Bayesian variable selection where the set of relevant predictors depends on the time already spent at risk. We finally determine the direct impact of a variable on each competing event through cause-specific variable selection.

A genetic programming approach to development of clinical prediction models: A case study in symptomatic cardiovascular disease.

BackgroundGenetic programming (GP) is an evolutionary computing methodology capable of identifying complex, non-linear patterns in large data sets. Despite the potential advantages of GP over more typical, frequentist statistical approach methods, its applications to survival analyses are rare, at best. The aim of this study was to determine the utility of GP for the automatic development of clinical prediction models.MethodsWe compared GP against the commonly used Cox regression technique in terms of the development and performance of a cardiovascular risk score using data from the SMART study, a prospective cohort study of patients with symptomatic cardiovascular disease. The composite endpoint was cardiovascular death, non-fatal stroke, and myocardial infarction. A total of 3,873 patients aged 19–82 years were enrolled in the study 1996–2006. The cohort was split 70:30 into derivation and validation sets. The derivation set was used for development of both GP and Cox regression models. These models were then used to predict the discrete hazards at t = 1, 3, and 5 years. The predictive ability of both models was evaluated in terms of their risk discrimination and calibration using the validation set.ResultsThe discrimination of both models was comparable. At time points t = 1, 3, and 5 years the C-index was 0.59, 0.69, 0.64 and 0.66, 0.70, 0.70 for the GP and Cox regression models respectively. At the same time points, the calibration of both models, which was assessed using calibration plots and the generalization of the Hosmer-Lemeshow test statistic, was also comparable, but with the Cox model being better calibrated to the validation data.ConclusionUsing empirical data, we demonstrated that a prediction model developed automatically by GP has predictive ability comparable to that of manually tuned Cox regression. The GP model was more complex, but it was developed in a fully automated way and comprised fewer covariates. Furthermore, it did not require the expertise normally needed for its derivation, thereby alleviating the knowledge elicitation bottleneck. Overall, GP demonstrated considerable potential as a method for the automated development of clinical prediction models for diagnostic and prognostic purposes.

Development of Clinical Prediction Models for Surgery and Complications in Crohn's Disease.

Crohn's disease-related complications account for a substantial proportion of inflammatory bowel disease-associated health care expenditure. Identifying patients at risk for complications may allow for targeted use of early therapeutic interventions to offset this natural course. We aimed to develop risk prediction models for Crohn's disease-related surgery and complications. Using data from the Randomised Evaluation of an Algorithm for Crohn's Disease cluster-randomised clinical Trial [REACT], which involved 1898 patients from 40 community practices, separate prediction models were derived and internally validated for predicting Crohn's disease-related surgery and disease-related complications [defined as the first disease-related surgery, hospitalisation, or complication within 24 months]. Model performance was assessed in terms of discrimination and calibration, decision curves, and net benefit analyses. There were 130 [6.8%] disease-related surgeries and 504 [26.6%] complications during the 24-month follow-up period. Selected baseline predictors of surgery included age, gender, disease location, Harvey-Bradshaw Index [HBI] score, stool frequency, antimetabolite or 5-aminosalicylate use, and the presence of a fistula, abscess, or abdominal mass. Selected predictors of complications included those same factors for surgery, plus corticosteroid or anti-tumour necrosis factor use, but excluded 5-aminosalicylate use. Discrimination ability, as measured by validated c-statistics, was 0.70 and 0.62 for the surgery and complication models, respectively. Score charts and nomograms were developed to facilitate future risk score calculation. Separate risk prediction models for Crohn's disease-related surgery and complications were developed using clinical trial data involving community gastroenterology practices. These models could be used to guide Crohn's disease management. External validation is warranted.

Reliability of Clinical Assessments in Older Adults With Syncope or Near Syncope.

Clinical prediction models for risk stratification of older adults with syncope or near syncope may improve resource utilization and management. Predictors considered for inclusion into such models must be reliable. Our primary objective was to evaluate the inter-rater agreement of historical, physical examination, and electrocardiogram (ECG) findings in older adults undergoing emergency department (ED) evaluation for syncope or near syncope. Our secondary objective was to assess the level of agreement between clinicians on the patient's overall risk for death or serious cardiac outcomes. We conducted a cross-sectional study at 11 EDs in adults 60 years of age or older who presented with unexplained syncope or near syncope. We excluded patients with a presumptive cause of syncope (e.g., seizure) or if they were unable or unwilling to follow-up. Evaluations of the patient's past medical history and current medication use were completed by treating provider and trained research associate pairs. Evaluations of the patient's physical examination and ECG interpretation were completed by attending/resident, attending/advanced practice provider, or attending/attending pairs. All evaluations were blinded to the responses from the other rater. We calculated the percent agreement and kappa statistic for binary variables. Inter-rater agreement was considered acceptable if the kappa statistic was 0.6 or higher. We obtained paired observations from 255 patients; mean (±SD) age was 73 (±9) years, 137 (54%) were male, and 204 (80%) were admitted to the hospital. Acceptable agreement was achieved in 18 of the 21 (86%) past medical history and current medication findings, none of the 10 physical examination variables, and three of the 13 (23%) ECG interpretation variables. There was moderate agreement (Spearman correlation coefficient, r = 0.40) between clinicians on the patient's probability of 30-day death or serious cardiac outcome, although as the probability increased, there was less agreement. Acceptable agreement between raters was more commonly achieved with historical rather than physical examination or ECG interpretation variables. Clinicians had moderate agreement in assessing the patient's overall risk for a serious outcome at 30 days. Future development of clinical prediction models in older adults with syncope should account for variability of assessments between raters and consider the use of objective clinical variables.

Synthesis of clinical prediction models under different sets of covariates with one individual patient data.

BackgroundRecently, increased development of clinical prediction models has been reported in the medical literature. However, evidence synthesis methodologies for these prediction models have not been sufficiently studied, especially for practical situations such as a meta-analyses where only aggregated summaries of important predictors are available. Also, in general, the covariate sets involved in the prediction models are not common across studies. As in ordinary model misspecification problems, dropping relevant covariates would raise potentially serious biases to the prediction models, and consequently to the synthesized results.MethodsWe developed synthesizing methods for logistic clinical prediction models with possibly different sets of covariates. In order to aggregate the regression coefficient estimates from different prediction models, we adopted a generalized least squares approach with non-linear terms (a sort of generalization of multivariate meta-analysis). Firstly, we evaluated omitted variable biases in this approach. Then, under an assumption of homogeneity of studies, we developed bias-corrected estimating procedures for regression coefficients of the synthesized prediction models.ResultsNumerical evaluations with simulations showed that our approach resulted in smaller biases and more precise estimates compared with conventional methods, which use only studies with common covariates or which utilize a mean imputation method for omitted coefficients. These methods were also applied to a series of Japanese epidemiologic studies on the incidence of a stroke.ConclusionsOur proposed methods adequately correct the biases due to different sets of covariates between studies, and would provide precise estimates compared with the conventional approach. If the assumption of homogeneity within studies is plausible, this methodology would be useful for incorporating prior published information into the construction of new prediction models.

Development of clinical prediction models for good or bad response to classic systemic drugs, anti-TNFs, and ustekinumab in psoriasis, based on the BIOBADADERM cohort

Background: Identifying patients likely to have very good or bad results from systemic psoriasis therapy could improve efficiency of therapy. Objective: To develop prognostic models for good or bad response to classic systemic drugs, anti-TNFs, and ustekinumab in psoriasis. Methods: Multivariable logistic regression of a prospective multicenter cohort of psoriatic patients in clinical practice (6449 person-years of follow-up). We used as possible predictors demographic characteristics, comorbidities, characteristics of the psoriasis (type, PASI, arthritis), history of past therapy at entry in the cohort, and history of response to previous cycles while in the cohort. We defined good response to a treatment cycle as either cycle end due to disease remission or a cycle longer than 2 years that does not end later due to inefficacy in the follow-up period. Bad response to a treatment cycle was defined as a cycle that is finished due to inefficacy, based on the physician judgment, after more than 3 months of treatment. Results: Patients with fewer previous therapies, lower body mass index, older at start of therapy, and with previous history of good responses to therapy are more likely to have positive results of therapy. However, the predictive characteristics of models are poor. Conclusion: Predictive models of clinical response to systemic drugs in psoriasis with the studied variables do not seem to outperform drug selection by a dermatologist.

PRM115 - Automatic development of clinical prediction models with genetic programming: A case study in cardiovascular disease

Objectives Genetic programming is an Evolutionary Computing technique, inspired by biological evolution, capable of discovering complex non-linear patterns in large datasets. Despite the potential advantages of genetic programming over standard statistical methods, its applications to survival analysis are at best rare, primarily because of the difficulty in handling censored data. The aim of this study was to demonstrate the utility of genetic programming for the automatic development of clinical prediction models using cardiovascular disease as a case study. Methods We compared genetic programming and the commonly used Cox regression technique in the development of a cardiovascular risk score using data from the SMART study, a prospective cohort study designed to identify predictors of future cardiovascular events in patients with symptomatic cardiovascular disease. The primary outcome was any cardiovascular event, comprising cardiovascular death and non-fatal stroke and myocardial infarction. The predictive ability of the model was assessed in terms of discrimination and calibration. Results 3,873 patients were enrolled in the study 1996–2006, aged 19–82 years and with 460 cardiovascular events. The discrimination of both models was comparable; the C-index of the genetic programming model being smaller (0.65; 95% CI: 0.63–0.66) but not significantly different from that of the Cox regression model (0.71; 0.67–0.75). The calibration of both models was also comparable, indicating similar disagreement between observed and predicted risks. Conclusions Using empirical data, we demonstrated that a prediction model developed by the novel technique of genetic programming has a comparable predictive ability to that of Cox regression. The genetic programming model was more complicated but was developed in an automated fashion and did not require the expertise needed for survival analysis. Genetic programming seems a promising technique for the automated development of clinical prediction models for diagnostic and prognostic purposes.

Formal and informal prediction of recurrent stroke and myocardial infarction after stroke: a systematic review and evaluation of clinical prediction models in a new cohort

BackgroundThe objective of this study was to: (1) systematically review the reporting and methods used in the development of clinical prediction models for recurrent stroke or myocardial infarction (MI) after ischemic stroke; (2) to meta-analyze their external performance; and (3) to compare clinical prediction models to informal clinicians’ prediction in the Edinburgh Stroke Study (ESS).MethodsWe searched Medline, EMBASE, reference lists and forward citations of relevant articles from 1980 to 19 April 2013. We included articles which developed multivariable clinical prediction models for the prediction of recurrent stroke and/or MI following ischemic stroke. We extracted information to assess aspects of model development as well as metrics of performance to determine predictive ability. Model quality was assessed against a pre-defined set of criteria. We used random-effects meta-analysis to pool performance metrics.ResultsWe identified twelve model development studies and eleven evaluation studies. Investigators often did not report effective sample size, regression coefficients, handling of missing data; typically categorized continuous predictors; and used data dependent methods to build models. A meta-analysis of the area under the receiver operating characteristic curve (AUROCC) was possible for the Essen Stroke Risk Score (ESRS) and for the Stroke Prognosis Instrument II (SPI-II); the pooled AUROCCs were 0.60 (95% CI 0.59 to 0.62) and 0.62 (95% CI 0.60 to 0.64), respectively. An evaluation among minor stroke patients in the ESS demonstrated that clinicians discriminated poorly between those with and those without recurrent events and that this was similar to clinical prediction models.ConclusionsThe available models for recurrent stroke discriminate poorly between patients with and without a recurrent stroke or MI after stroke. Models had a similar discrimination to informal clinicians' predictions. Formal prediction may be improved by addressing commonly encountered methodological problems.

Function of restored jejunum lengthened by a degradable spring

CONCLUSIONS: Intra-abdominal injury is associated with extraabdominal injuries that are identifiable on plain x-ray. These injuries serve as clinical predictors of intra-abdominal injury and should be included in future prospective development of clinical prediction models. Specific attention to mechanism is essential as injury patterns may differ with these factors and may increase the specificity of these prediction models, particularly for femur fracture as a predictor of intra-abdominal injury.

Circulating Apoptotic Endothelial Cells and Apoptotic Endothelial Microparticles Independently Predict the Presence of Cardiac Allograft Vasculopathy

Maintenance of endothelial homeostasis may prevent the development of cardiac allograft vasculopathy (CAV). This study investigated whether biomarkers related to endothelial injury and endothelial repair discriminate between CAV-negative and CAV-positive heart transplant recipients. CAV is the most important determinant of cardiac allograft survival and a major cause of death after heart transplantation. Fifty-two patients undergoing coronary angiography between 5 and 15 years after heart transplantation were recruited in this study. Flow cytometry was applied to quantify endothelial progenitor cells (EPCs), circulating endothelial cells (CECs), and endothelial microparticles. Cell culture was used for quantification of circulating EPC number and hematopoietic progenitor cell number and for analysis of EPC function. The EPC number and function did not differ between CAV-negative and CAV-positive patients. In univariable models, age, creatinine, steroid dose, granulocyte colony-forming units, apoptotic CECs, and apoptotic endothelial microparticles discriminated between CAV-positive and CAV-negative patients. The logistic regression model containing apoptotic CECs and apoptotic endothelial microparticles as independent predictors provided high discrimination between CAV-positive and CAV-negative patients (C-statistic 0.812; 95% confidence interval: 0.692 to 0.932). In a logistic regression model with age and creatinine as covariates, apoptotic CECs (p = 0.0112) and apoptotic endothelial microparticles (p = 0.0141) were independent predictors (C-statistic 0.855; 95% confidence interval: 0.756 to 0.953). These 2 biomarkers remained independent predictors when steroid dose was introduced in the model. The high discriminative ability of apoptotic CECs and apoptotic endothelial microparticles is a solid foundation for the development of clinical prediction models of CAV.