Diagnostic Work-Up for Detection of Paroxysmal Atrial Fibrillation After Acute Ischemic Stroke
Cross-Sectional Survey on German Stroke Units
Background and Purpose—Multiple methods to detect paroxysmal atrial fibrillation (pAF) in patients with acute stroke are available. However, it is unknown which approaches are currently used in clinical routine and guidelines remain vague to the extent of cardiac monitoring. We characterize diagnostic efforts for pAF detection on German stroke units (SU).
Methods—A standardized anonymous questionnaire was sent to all clinical leads of certified SUs in Germany. The questionnaire focused on basic characteristics of SUs, procedures to detect AF, and estimates on AF detection.
Results—One hundred seventy-nine SU leads participated (response rate 71.6%). All patients undergo continuous bedside ECG monitoring. A percentage of 77.6 SUs initiate additional 24-hour Holter ECG in >50% of patients without known AF. Patients with transient ischemic attack are monitored significantly shorter than patients with ischemic stroke. Independent of SU type or size, 67.6% of leads assumed to fail detecting pAF in 5% to 20% of patients. In cryptogenic stroke, additional ECG monitoring is recommended by 90.2% but only 13.8% of SUs perform routine ECG follow-up visits. The use of implanted event recorders is low (1–10 patients/y by 60.7% of SUs; 28.1%: no use). A percentage of 83.9 do not use external event recorders.
Conclusions—Our survey demonstrates substantial heterogeneity among German SUs on diagnostic work-up for pAF. Future prospective multicenter studies should systematically evaluate the impact of different methods to uncover pAF.
An important goal of stroke unit (SU) treatment is to clarify stroke cause. Detection of paroxysmal atrial fibrillation (pAF) can be challenging. Longer ECG monitoring increases pAF detection,1 but optimal duration of ECG monitoring is controversial and current stroke guidelines are vague.2,3 We characterized current diagnostic efforts to detect AF on certified SUs in Germany.
We conducted an anonymous cross-sectional survey using a standardized questionnaire (online-only Data Supplement) that was sent to clinical leads of all 250 SU, which had been certified by the German Stroke Society. No individual patient-level data were collected. Details of statistical methods are available in the online-only Data Supplement.
One hundred seventy-nine SU leads participated (response rate, 71.6%; SU characteristics: Table I in the online-only Data Supplement). Diagnostic procedures are presented in the Table. A 12-channel ECG is performed in >95% of patients on the majority of SUs (83.6%). Repeated daily 12-channel ECGs to detect AF are performed in 18% of SUs. The majority of SUs initiate 24-hour Holter ECG in >50% of patients without known AF. No association was present between the use of 24-hour Holter ECG and SU type (odds ratio [OR], 1.02; 95% confidence interval [CI], 0.55–1.90) or number of SU beds (OR, 1.12; 95% CI, 0.61–2.05), respectively.
Continuous ECG Monitoring
All patients undergo CEM, but duration of CEM differed between patients with ischemic stroke and transient ischemic attack (TIA; 66.8% versus 22.2% undergo CEM for >48 hours; OR, 7.09; 95% CI, 4.41–11.37).
The CEM system has an integrated AF alarm in 44.4% of the SUs; 15.2% of SUs use a telemetric wireless system and 31.1% a dedicated analysis software to detect AF.
The majority of SUs (62.3%) perform daily ECG trace visits of stored ECG data. These visits are more often performed on SUs with an AF alarm in their monitoring system (73.1% versus 53.6%; OR, 2.35; 95% CI, 1.24–4.46) and in those using dedicated AF software compared with those without (81.5% versus 53.3%; OR, 3.85; 95% CI, 1.77–8.35). The assumed proportion of failed detection of pAF was not associated with the performance of daily monitor visits (OR, 1.74; 95% CI, 0.90–3.35), the presence of a telemetric system (OR, 2.11; 95% CI, 0.92–4.83), a software-based analysis of monitor traces (OR, 1.10; 95% CI, 0.57–2.11), or a dedicated alarm for AF (OR, 1.12; 95% CI, 0.61–2.07).
Independent of SU type, 67.6% of leads assumed that they failed to detect AF in 5% to 20% of patients although the presence of AF was suspected (Table II in the online-only Data Supplement).
The proportion of cryptogenic strokes was estimated to be up to 30% by almost all unit leads (Table III in the online-only Data Supplement). Subsequent additional ECG monitoring is recommended by 90.2% in patients with cryptogenic stroke. However, only 13.8% of SUs have an established pathway for routine outpatient follow-up visits for these patients.
Event Recorders and Perceived Value of Methods to Diagnose AF
Implanted event recorders are used in 1 to 10 patients per year by ≈60% of SUs. A percentage of 11.2 reported to implant even >10 patients per year. However, almost one third of SUs (28.1%) does not use implanted event recorders at all, which was independent of SU type or size (OR, 1.06; 95% CI, 0.54–2.07 and OR, 1.12; 95% CI, 0.58–2.15, respectively). Most participating SUs (83.9%) do not use external event recorders.
Half of the SU leads regard event recorders to be the best option for detecting pAF, whereas Holter-ECG and CEM are regarded second best methods to detect pAF by most SU leads (Table IV in the online-only Data Supplement). In contrast, single 12-channel ECGs are not considered helpful by 53.6% of clinicians.
Efforts to unveil pAF differ among SUs, and CEM is acknowledged a valuable tool to detect pAF.4 In our study, the SUs that already use methods to improve pAF detection by CEM, including AF alarms or dedicated AF detection software, more often perform daily ECG trace visits of stored CEM data. This indicates that efforts to make the most of the monitoring data vary substantially across participating SUs.
The estimated proportion of undetected pAF in patients with stroke was up to 20%. The majority of leads recommend ECG monitoring in cryptogenic stroke after hospital discharge. However, only a minority of SUs have an established pathway for routine outpatient follow-up visits. Preferences of SU leads for further diagnostic procedures varied. Presumably, this uncertainty reflects the heterogeneity of approaches tested in various clinical studies.5–7 Furthermore, clinical guidelines remain vague on the choice and extent of ECG monitoring techniques in patients with acute ischemic stroke or TIA.2,3
SU leads appraised event recorders as one of the most sensitive methods for pAF detection. This perception is consistent with recent findings from randomized controlled trials in cryptogenic stroke.6,7 However, implementation of such techniques was low, and selection criteria of patients remain to be established. An important finding of our survey is that the work-up for pAF after TIA is less extensive than in ischemic stroke. Studies addressing cardiac evaluation of patients with TIA are limited.8 Furthermore, there is less common practice of ECG monitoring for patients with TIA than for those with stroke in different healthcare systems,9 and hospitalization rates in patients with TIA vary among hospitals and regions.8,9 Nevertheless, both groups of patients should be evaluated in the same manner for the presence of AF. Additional studies should compare the impact of different cardiac diagnostic procedures in cryptogenic stroke and particularly TIA to improve standardization of procedures.
The high response rate in our survey supports a representative picture of the current practice of AF detection across German SUs. National reimbursement practices for inpatient stroke care in the German healthcare system may have influenced our findings and translation to other healthcare settings may be limited. Furthermore, our results are based on estimations by SU leads rather than on individual patient data of the participating centers. Thus, we cannot exclude that the variations seen by center characteristics are caused by chance alone.
In conclusion, our study reveals substantial heterogeneity among German SUs on diagnostic work-up for pAF detection. Additional studies should compare different methods to detect pAF during inpatient management and thereafter.
We thank all participating stroke unit leads for their cooperation.
Sources of Funding
Dr Endres receives funding from the Deutsche Forschungsgemeinschaft, Bundesministerium für Bildung und Forschung, European Union, Volkswagen-Foundation, and Corona-Foundation.
Dr Rizos received consulting and speakers honoraria from BMS-Pfizer and Boehringer-Ingelheim. Dr Haeusler received consulting honoraria, research support, and speakers’ honoraria from Bayer-Healthcare, Sanofi, Pfizer, and Bristol-Myers Squibb. Dr Endres received consulting honoraria, research support, travel grants, and speakers’ honoraria from AstraZeneca, Bayer, Boston Scientific, Bristol-Myers Squibb, Boehringer-Ingelheim, Ever, Glaxo-Smith-Kline, MSD, Novartis, Pfizer, Roche, and Sanofi. Dr Veltkamp received consulting honoraria, research support, travel grants, speakers’ honoraria from Bayer, Boehringer-Ingelheim, BMS-Pfizer, Daiichi Sankyo, Roche-Diagnostics, St. Jude Medical, ApoplexMedical, and Sanofi. The other authors report no conflicts.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.115.009374/-/DC1.
- Received March 9, 2015.
- Revision received March 9, 2015.
- Accepted March 23, 2015.
- © 2015 American Heart Association, Inc.
- Seet RC,
- Friedman PA,
- Rabinstein AA
- Jauch EC,
- Saver JL,
- Adams HP Jr,
- Bruno A,
- Connors JJ,
- Demaerschalk BM,
- et al
- Rizos T,
- Güntner J,
- Jenetzky E,
- Marquardt L,
- Reichardt C,
- Becker R,
- et al
- Grond M,
- Jauss M,
- Hamann G,
- Stark E,
- Veltkamp R,
- Nabavi D,
- et al
- Easton JD,
- Saver JL,
- Albers GW,
- Alberts MJ,
- Chaturvedi S,
- Feldmann E,
- et al
- Giles MF,
- Rothwell PM