Isolated insular strokes and plasma MR-proANP levels are associated with newly diagnosed atrial fibrillation: a pilot study

Abstract

Introduction: In this study, we assessed the relationship of insular strokes and plasma MR-proANP levels with newly diagnosed atrial fibrillation (NDAF).

Methods: This study is based on a prospective acute stroke cohort (http://www.clinicaltrials.gov, NCT00390962). Patient eligibility was dependent on the diagnosis of acute ischemic stroke, absence of previous stroke based on past medical history and MRI, no history of AF and congestive heart failure (cohort A) and, additionally, no stroke lesion size ≥ 20 mL (sub-cohort A*). AF, the primary endpoint, was detected on 24-hour electrocardiography and/or echocardiography. Involvement of the insula was assessed by two experienced readers on MRI blinded to clinical data. MR-proANP levels were obtained through a novel sandwich immunoassay. Logistic-regression-models were fitted to estimate odds ratios for the association of insular strokes and MR-proANP with NDAF. The discriminatory accuracy of insular strokes and MR-proANP was assessed by a model-wise comparison of the area under the receiver-operating-characteristics-curve (AUC) with known predictors of AF.

Results: 104 (cohort A) and 83 (cohort A*) patients fulfilled above-mentioned criteria. Patients with isolated insular strokes had a 10.7-fold higher odds of NDAF than patients with a small ischemic stroke at any other location. The AUC of multivariate logistic regression models for the prediction of NDAF improved significantly when adding stroke location and MR-proANP levels. Moreover, MR-proANP levels remained significantly elevated throughout the acute hospitalization period in patients with NDAF compared to those without.

Conclusions: Isolated insular strokes and plasma MR-proANP levels on admission are independent predictors of NDAF and significantly improve the prediction accuracy of identifying patients with NDAF compared to known predictors including age, the NIHSS and lesion size. To accelerate accurate diagnosis and enhance secondary prevention in acute stroke, higher levels of MR-proANP and insular strokes may represent easily accessible indicators of AF if confirmed in an independent validation cohort.

Conflict of interest statement

Competing Interests: The authors have read the journal's policy and have the following conflicts: B. Müller received financial research support as well as payments for lectures from B.R.A.H.M.S. part of Thermo Fisher Scientific. A. Gass received payments for lectures including service on speaker bureaus from BiogenIdec, Novartis, Bayer and Merck. M. Katan received within the last 3 years (not related to this manuscript) an unconditional research grant form B.R.A.H.M.S. part of Thermo Fisher Scientific. K. Frontzek, F. Fluri,J. Siemerkus and M. Christ-Crain have nothing to disclose. The authors have read the journals’s policy on competing interests and the conflicts mentioned do not alter the authors' adherence to all the PLOS ONE policies on sharing data and materials.

Figures

Figure 1. Patient flow chart.

The upper part of the figure (as divided by 2 horizontal bars) denotes the study cohort of the original COSMOS cohort as previously described. The lower part of the scheme indicates the eligibility criteria for the cohorts analyzed in this study.

Figure 2. Timecourse of plasma MR-proANP levels.

Plasma lnMR-proANP levels are plotted on the ordinate, lnMR-proANP levels were significantly different between patient groups throughout the monitored period (day 0 denotes the day of admission; ** p