A prediction model for 5-year cardiac mortality in patients with chronic heart failure using 123I-metaiodobenzylguanidine imaging

Kenichi Nakajima, Tomoaki Nakata, Takahisa Yamada, Shohei Yamashina, Mitsuru Momose, Shu Kasama, Toshiki Matsui, Shinro Matsuo, Mark I. Travin, Arnold F. Jacobson

Research output: Contribution to journalArticle

29 Scopus citations

Abstract

Purpose: Prediction of mortality risk is important in the management of chronic heart failure (CHF). The aim of this study was to create a prediction model for 5-year cardiac death including assessment of cardiac sympathetic innervation using data from a multicenter cohort study in Japan. Methods: The original pooled database consisted of cohort studies from six sites in Japan. A total of 933 CHF patients who underwent 123I-metaiodobenzylguanidine (MIBG) imaging and whose 5-year outcomes were known were selected from this database. The late MIBG heart-to-mediastinum ratio (HMR) was used for quantification of cardiac uptake. Cox proportional hazard and logistic regression analyses were used to select appropriate variables for predicting 5-year cardiac mortality. The formula for predicting 5-year mortality was created using a logistic regression model. Results: During the 5-year follow-up, 205 patients (22 %) died of a cardiac event including heart failure death, sudden cardiac death and fatal acute myocardial infarction (64 %, 30 % and 6 %, respectively). Multivariate logistic analysis selected four parameters, including New York Heart Association (NYHA) functional class, age, gender and left ventricular ejection fraction, without HMR (model 1) and five parameters with the addition of HMR (model 2). The net reclassification improvement analysis for all subjects was 13.8 % (p<0.0001) by including HMR and its inclusion was most effective in the downward reclassification of low-risk patients. Nomograms for predicting 5-year cardiac mortality were created from the five-parameter regression model. Conclusion: Cardiac MIBG imaging had a significant additive value for predicting cardiac mortality. The prediction formula and nomograms can be used for risk stratifying in patients with CHF.

Original languageEnglish (US)
Pages (from-to)1673-1682
Number of pages10
JournalEuropean Journal of Nuclear Medicine and Molecular Imaging
Volume41
Issue number9
DOIs
StatePublished - Sep 2014

    Fingerprint

Keywords

  • Cardiac mortality
  • Chronic heart failure
  • Prediction model
  • Risk stratification

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging

Cite this