Introduction: Undiagnosed atrial fibrillation (UAF) is a major burden in ischemic stroke. However, randomised trials have partly shown astonishingly low AF detection rates (e. g. in the CRYSTAL-AF study). This may be due to differences in baseline and stroke characteristics between studies.
Hypothesis: We hypothesized that stroke patients in a randomised controlled trial have less severe strokes than patients in an observational trial with similar inclusion and exclusion criteria.
Methods: We used data from the Find-AF observational (NCT 01855035) and the Find-AF randomised controlled trial (ISRCTN 46104198). We included only patients at study site Goettingen of the Find-AF randomised controlled trial (n=153) and only included patients from the Find-AF observational trial that fulfilled the inclusion/exclusion criteria of Find-AF randomised (n=90). We compared baseline characteristics of screened versus included patients in Find-AF randomised and baseline characteristics and stroke severity parameters between both studies. Data are shown as mean (Standard Deviation) or Median (25%; 75% percentile) and were compared by chi-square, t-test or Mann-Whitney U test.
Results: Table 1 shows as comparison between baseline characteristics of both studies Comparing patients fulfilling the inclusion/exclusion criteria of Find-AF randomised, but unwilling to give informed consent to randomised patients showed a significant difference in age (77 ± 11 vs. 74 ± 8; p< 0.001), but no difference in gender (p=0.581).
Conclusions: Both studies included patients similar in age, gender and many comorbidities. Major differences occurred in NIHSS and MRS. As NIHSS is a predictor of atrial fibrillation, randomized controlled trials investigating AF detection should include a minimum NIHSS to avoid a selection bias towards less severe strokes.