Abstract P019: Identifying Laboratory Markers of Incident Myocardial Infarction through Electronic Health Records
Background: Patients on hemodialysis (HD) are at high risk of cardiac events including myocardial infarction (MI). Electronic health records (EHRs) present a unique opportunity to observe biomedical changes in patients over time.
Purpose: To assess whether there were any distinct changes in laboratory values prior to an incident MI.
Methods: Using EHR data from a large dialysis organization, we identified 6,468 patients with an incident MI and selected an equal number of patients receiving dialysis on the same date without an MI. We abstracted six months of 12 regularly collected laboratory measures for their potential utility as a marker for incident MI: albumin, calcium, co2, creatinine, ferritin, hemoglobin, iron saturation, phosphorus, platelet, potassium, single pool Kt/V and white blood cell count. We used natural cubic spline regression to determine whether the trajectory of the laboratory values differed amongst patients with and without an incident MI.
Results: Of the 12 laboratory measures, 7 showed a distinct change between patients with and without MI, 14 to 28 days before the incident event. Three measures were related to inflammation (white blood cell count, iron saturation, and ferritin; see Figure 1), 3 measures were related to malnutrition (albumin, creatinine, phosphorous), along with single pool kt/v.
Conclusion: Distinct changes in routinely collected laboratories are observed prior to an MI. Future studies based on our results will focus on whether these changes can be used in real-time to predict MI risk in HD patients.
Author Disclosures: B.A. Goldstein: None. T.I. Chang: None. T. Hastie: None. W.C. Winkelmayer: None.
This research has received full or partial funding support from the American Heart Association, National Center.
- © 2014 by American Heart Association, Inc.