Does Atrial Fibrillation Affect the Automated Oscillometric Blood Pressure Measurement?
To the Editor:
We read with interest the article by Pagonas et al1 on the accuracy of oscillometric blood pressure (BP) monitors in atrial fibrillation (AF). An important point raised by the authors is that AF impedes both oscillometric and auscultatory BP measurement1 and disagreement between them does not necessarily mean that the latter is the correct one. Thus, intra-arterial measurements were used as reference.1
The main issue with this analysis is that established validation criteria have not been applied. We re-evaluated these results using the Association for Advancement of Medical Instrumentation (AAMI) validation criteria of accuracy (mean BP difference ≤5 mm Hg) and precision (SD ≤8 mm Hg).2
Using the oscillometric arm device, the BP bias (versus intra- arterial measurement) was −7.4±11.3/8.3±8.1 mm Hg (systolic/diastolic) in AF and −3.9±10.4/8.6±9.8 in sinus rhythm (SR).1 Thus, the AAMI criterion of systolic BP measurement accuracy was fulfilled in SR but not in AF. The difference in systolic BP bias between the 2 groups (almost double) was not significant because of high SD values. The precision criterion was not fulfilled in both SR and AF. Regarding the wrist device, this failed the accuracy and precision criteria in both groups.
These findings suggest that the accuracy of the arm device in measuring systolic BP is negatively affected by AF. Although not statistically significant (underpowered comparison), this difference is clinically important and leads to a different conclusion when considering the AAMI criterion (pass in SR, fail in AF). Moreover, the similar BP bias with the wrist device in SR and AF is too uncertain evidence to imply that the measurement accuracy is not affected by AF. The wrist device failed the AAMI criteria, which cannot be ignored. Old studies of oscillometric versus intra-arterial measurements showed inconsistent results, suggesting important differences in the accuracy of different devices.3,4
Another issue is that because of the different direction of bias for systolic versus diastolic BP, the bias in pulse pressure is considerable (12.5–16.0 mm Hg),1 which is unacceptable. Moreover, to ameliorate the impact of BP variability, triplicate measurements were averaged (n=52). Although an interesting approach, additional analysis of individual readings would be more powerful (n=156) and might offer further insight to the methods’ disagreement, which otherwise might be hidden.
In conclusion, the exact answer on how AF affects the oscillometric BP measurement remains uncertain. The available evidence is imperfect and suggests that the accuracy of the oscillometric devices is negatively influenced by AF and differs among devices. Thus, results for a single device cannot be extrapolated to the oscillometric technique in general. Given that AF mainly affects the elderly and presents a particularly difficult condition for accurate auscultatory measurement, we agree with the recommendation by Pagonas et al1 that oscillometric devices should be used for self-home, but not for office measurement.5
George S. Stergiou
Third University Department of Medicine
- © 2013 American Heart Association, Inc.
- Pagonas N,
- Schmidt S,
- Eysel J,
- Compton F,
- Hoffmann C,
- Seibert F,
- Hilpert J,
- Tschöpe C,
- Zidek W,
- Westhoff TH
- 2.↵American National Standards Institute. Noninvasive Sphygmoma nometers. Part 2: Clinical Validation of Automated Measurement Type. ANSI/AAMI/ISO 81060–2. 2009. http://webstore.ansi.org/RecordDetail.aspx?sku=ANSI%2FAAMI%2FISO+81060-2%3A2009. Accessed 7 September 2013.
- White WB,
- Lund-Johansen P,
- McCabe EJ,
- Omvik P