A Predictive Model to Identify High Risk of Cardiac Arrest or In-Hospital Mortality among Patients with Suspected Acute Coronary Syndromes: An IMMEDIATE Trial Sub-study.
Ray, Madhab.
2014
-
Abstract: Background: Intravenous glucose insulin potassium (GIK) solution has been shown to have a protective role in preventing myocardial damage from ischemic insult in laboratory animals. Its effectiveness in human subjects has long been debated and clinical studies have produced divergent results. The recently completed IMMEDIATE Trial showed benefit of GIK in the composite outcome of cardiac... read morearrest or in-hospital mortality in patients with suspected acute coronary syndrome (ACS). To aid the identification of patients most likely to experience this benefit, this ancillary analysis of the IMMEDIATE Trial was performed to develop a predictive model that could facilitate recognition of these high-risk patients for prioritization for GIK therapy. Methods: Multivariable logistic regression was used to develop a predictive model for the composite endpoint cardiac arrest or in-hospital mortality using patients of the control arm of the IMMEDIATE Trial, i.e., patients not treated with GIK. Tertiles of predicted risk categories then were created for all patients in the trial (treated and not treated) using the developed predictive model and risk categories were tested for interactions with GIK on the outcome. Results: Four variables were significantly associated with the study endpoint: advanced age, low systolic blood pressure, ST elevation in the presenting electrocardiogram (ECG), and duration of symptoms from the onset. The model developed using these variables predicted cardiac arrest or death with good performance, corresponding to a C statistic of 0.75. Calibration of the predictive instrument for observed and predicted rates of cardiac arrest or in-hospital mortality was very good. No significant interaction on the odds ratio scale with GIK was identified for any candidate variables or the predicted risk categories in the study population. However, based on constant odds ratio (OR) across the risk categories there was an absolute risk reduction of 8.6% with GIK in high-risk tertile with the corresponding number needed to treat (NNT) of 12 to prevent one cardiac arrest or in-hospital mortality. The corresponding figures for the low-risk tertile were 0.8% and 125 respectively. Thus there appears to be a pronounced benefit of GIK in the high-risk group compared to the low-risk group when absolute risk reduction was used to measure the effect of treatment with GIK compared to placebo. Conclusions: A multivariable predictive model was developed that identifies patients at high risk of cardiac arrest or death. When classified by the predictive model, the greatest benefit of GIK is demonstrated in the high-risk tertile.
Thesis (M.S.)--Tufts University, 2014.
Submitted to the Dept. of Clinical & Translational Science.
Advisor: Harry Selker.
Committee: Robin Ruthazer, Joni Beshansky, and David Kent.
Keyword: Health sciences.read less - ID:
- 0v838b682
- Component ID:
- tufts:20518
- To Cite:
- DCA Citation Guide EndNote
