Validation of the diabetes Chronic Kidney Disease (CKD) and Heart Failure (HF) model in the HARMONY Outcomes trial

Validation of the diabetes Chronic Kidney Disease (CKD) and Heart Failure (HF) model in the HARMONY Outcomes trial

Congestive Heart Failure (CHF) and Chronic Kidney Disease (CKD) are highly prevalent in patients with Type 2 Diabetes (T2DM) and are associated with heavy disease burden and high mortality. CHF and CKD progression, especially ESRD, carry poor prognosis for diabetes since they occur over several years and early clinical identification is difficult. Models predicted CHF mainly focused on patients with CHF history, while diabetes patients free of CHF should be paid more attention in terms of disease complications and financial burdens. Therefore, this study is to develop CKD progression and CHF incidence prediction models using the Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial data and validate the models use the HARMONY data. Physicians and patients could take advantage of this risk prediction tool for efficient alert and good pre-evaluation, especially pay attention to the physical condition and healthcare utilization in the current year, in turn to better prevent severe complications and decrease the financial burdens. Dynamic prediction modelling has the potential to enhance and maintain model accuracy by updating existing prediction models continually as new data are accrued. The dynamic prediction models will also use time-updated data over time to perform external validation. Most of the diabetes CKD and Heart Failure (HF) models only used baseline characteristics to conduct prediction and were developed without external validation. These model development process decreased the model accuracy and could not demonstrate the generalizability. Although, there are a few validation studies evaluated the model performance of the previous developed models in recently external data. The prediction accuracy did not demonstrate good results, possibly because of the updating therapies, or the external data that do not fit the model's target population.The CKD progression and HF incidence risk prediction models were patient-level dynamic models. Our model was developed using data from ACCORD trial, which has more recent data. Time-varying dynamic prediction CKD and HF models will be developed. Using the Harmony Outcomes longitudinal patient level data, the researchers will examine the external validation of the CKD progression and HF incidence prediction models. The Harmony Outcomes trial is an international, double-blind, randomized, placebo-controlled trial, testing the cardiovascular (CV) safety of albiglutide, a GLP-1 agonist drug for T2DM treatment, among patients with diabetes, who were at high risk of CV events. In addition, CKD is prevalent among diabetes population. The Harmony Outcomes trial as the target population will fit the model external validation very well. A series of methods will be utilized to evaluate the performance of the models. Brier score, which measures the accuracy of probabilistic predictions, will demonstrate overall performance. The False Positive, False Negative, True Positive, True Negative and related metrics will be presented as classification metrics. Concordance statistics (C statistics) is a "global" index for validating the predictive ability of a survival model, which equals to Area under Curve (AUC). C statistics will be computed as measures of discrimination. For calibration, the Hosmer-Lemeshow test for goodness of fit and calibration plots will be applied to compare the observed vs predicted risk. These calibration measurements assess whether or not the observed event rates match expected event rates in the model.

[{ "PostingID": 19799, "Title": "GSK-GLP116174", "Description": "A long term, randomised, double blind, placebo-controlled study to determine the effect of albiglutide, when added to standard blood glucose lowering therapies, on major cardiovascular events in patients with Type 2 diabetes mellitus" }]

Lin, Y, Shao, H, Shi, L, Anderson, AH, Fonseca, V.
Predicting incident heart failure among patients with type 2 diabetes mellitus: The DM-CURE risk score. Diabetes Obes Metab. 2022; 24( 11): 2203- 2211
doi:10.1111/dom.14806