A clinical risk score to predict the time to first appropriate device therapy in recipients of implantable cardioverter defibrillators

Pacing Clin Electrophysiol. 2007 Mar;30(3):385-9. doi: 10.1111/j.1540-8159.2007.00679.x.

Abstract

Background: To develop a risk score to predict the occurrence of appropriate defibrillator [implantable cardioverter-defibrillator (ICD)] therapies. A simple clinical score predicting the risk of appropriate ICD therapy is lacking.

Methods: A Cox regression model was developed from a database of ICD patients at a single tertiary center to predict the time to appropriate ICD therapy defined as shock or antitachycardia pacing. A risk score was derived from this model using half of the database and was validated using the other half.

Results: A total of 399 patients were entered into the database between July 2001 and February 2004. There were no statistically significant differences between the derivation (n = 200) and validation (n = 199) groups in any of the demographic or clinical variables recorded. The risk score included three independent variables: indication for ICD implantation (P = 0.03), serum creatinine level (P = 0.015), and QRS width (P = 0.028). The observed risk scores were highly predictive of time to ICD therapy in the validation group (P = 0.02).

Conclusion: We describe a new clinical risk score that predicts the time to appropriate device therapy in ICD recipients of a single tertiary center hospital. The performance of this risk score needs to be investigated prospectively in a larger patient population.

Publication types

  • Controlled Clinical Trial
  • Validation Study

MeSH terms

  • Arrhythmias, Cardiac / blood
  • Arrhythmias, Cardiac / epidemiology*
  • Arrhythmias, Cardiac / prevention & control*
  • Creatine / blood
  • Defibrillators, Implantable / statistics & numerical data*
  • Disease-Free Survival
  • Electric Countershock / statistics & numerical data*
  • Female
  • Humans
  • Incidence
  • Male
  • Middle Aged
  • Outcome Assessment, Health Care / methods*
  • Pennsylvania / epidemiology
  • Prognosis
  • Proportional Hazards Models*
  • Reproducibility of Results
  • Risk Assessment / methods*
  • Risk Factors
  • Sensitivity and Specificity

Substances

  • Creatine