Secondary Logo

Journal Logo

Automated blood pressure measurement in atrial fibrillation: a systematic review and meta-analysis

Cheng, Hao-Mina,b,c; Tufanaru, Catalina; Pearson, Alana; Chen, Chen-Huana,b,c,d

doi: 10.1097/HJH.0b013e32835ae9a3

aThe Joanna Briggs Institute, Faculty of Health Sciences, The University of Adelaide, Adelaide, Australia

bDepartment of Medical Research and Education, Taipei Veterans General Hospital

cDepartment of Medicine, National Yang-Ming University

dInstitute of Public Health and Community Medicine Research Center, National Yang-Ming University, Taipei, Taiwan

Correspondence to Hao-Min Cheng, MD, The Joanna Briggs Institute, Faculty of Health Sciences, The University of Adelaide, Adelaide 5005, Australia. Tel: +61 8 8410 4809; fax: +61 8 8410 4809; e-mail:

We read with interest the article by Stergiou et al.[1] on the ‘Automated blood pressure measurement in atrial fibrillation: a systematic review and meta-analysis’. Although appreciative of the contribution this article makes to the field, we wish to comment on the method used in the meta-analysis related to the measurement comparison of a tested technique with an established one [2]. Bland and Altman, the methodological leaders in this field, proposed methods to investigate agreement that have become the validation standards for the measurement accuracy of blood pressure monitors [2–4]. They have pointed out more than once that the use of a correlation coefficient is inappropriate in the analysis of measurement method comparison data [3,5]. To describe agreement, it is necessary to emphasize not only the mean difference between two measurements but also the 95% limits of agreement, which can be calculated by the equation: (mean difference) ± 1.96 × (standard deviation of differences) [3]. It is insufficient to present only the mean difference because if the two methods agree on average, the mean difference is small regardless of the agreement for individuals [5]. In this regard, there should be two outcome measures that the meta-analysis of method comparison studies should synthesize: the pooled estimates of systematic bias (mean differences) and random error (standard deviation of differences) [6]. Williamson et al.[6] have developed a concise methodology for the meta-analysis of method comparison studies, which had been utilized in previous systematic reviews with meta-analysis [7–9]. The outcome measures for meta-analysis in the present article [1] are correlation coefficients and averaged blood pressure difference, and the random error was only evaluated with the variability of individual studies as presented in its Table 2 [1]. The 95% confidence interval shown in the article is actually the pooled estimate of the mean differences and has nothing to do with the limits of agreement (random error). We think that the limits of agreement is an important measure for this study to describe the extent of agreement between blood pressure measurements, which should also be meta-analysed on the basis of the previously established methodology [6–9]. Through the presentation of pooled estimates of random errors (the 95% limits of agreement), readers can have a better understanding and assessment on the accuracy of automated blood pressure monitors in atrial fibrillation.

On the basis of the data provided in Figures 5 and 6 [1] and the random effect model of meta-analysis for method comparison studies established by Williamson et al.[6], the pooled estimates of agreement for SBP and DBP (Table 1) were 0.04 ± 5.5 mmHg (95% limits of agreement −10.7∼10.8) and 2.5 ± 8.5 mmHg (95% limits of agreement −14.1∼19.2), respectively. However, after excluding the study by Farsky et al.[10], which used a dual control device with both Korotkov and oscillometric methods, the pooled estimates of agreement for SBP and DBP are 1.0 ± 11.3 mmHg (95% limits of agreement −21.1∼23.1) and 3.0 ± 9.7 mmHg (95% limits of agreement −16.0∼22.0), respectively. Both exceed the recommended range for agreement on the basis of the international standard of 5.0 ± 8.0 mmHg [4]. Thus, the accuracy of the automatic blood pressure monitors in measuring SBP needs to be improved.



Back to Top | Article Outline


Conflicts of interest

There are no conflicts of interest.

Back to Top | Article Outline


1. Stergiou GS, Kollias A, Destounis A, Tzamouranis D. Reply. J Hypertens 2013; 31:215–216.
2. Altman DG, Bland JM. Measurement in medicine: the analysis of method comparison studies. Statistician 1983; 32:307–317.
3. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; 1:307–310.
4. Stergiou GS, Karpettas N, Atkins N, O’Brien E. European Society of Hypertension International Protocol for the validation of blood pressure monitors: a critical review of its application and rationale for revision. Blood Press Monit 2010; 15:39–48.
5. Bland JM, Altman DG. Comparing two methods of clinical measurement: a personal history. Int J Epidemiol 1995; 24 (Suppl 1):S7–S14.
6. Williamson PR, Lancaster GA, Craig JV, Smyth RL. Meta-analysis of method comparison studies. Stat Med 2002; 21:2013–2025.
7. Craig JV, Lancaster GA, Williamson PR, Smyth RL. Temperature measured at the axilla compared with rectum in children and young people: systematic review. BMJ 2000; 320:1174–1178.
8. Craig JV, Lancaster GA, Taylor S, Williamson PR, Smyth RL. Infrared ear thermometry compared with rectal thermometry in children: a systematic review. Lancet 2002; 360:603–609.
9. Cheng HM, Lang D, Tufanaru C, Pearson A. Measurement accuracy of noninvasively obtained central blood pressure by applanation tonometry: a systematic review and meta-analysis. Int J Cardiol 2012 [Epub ahead of print].
10. Farsky S, Benova K, Krausova D, Sirotiakova J, Vysocanova P. Clinical blood pressure measurement verification when comparing a Tensoval duo control device with a mercury sphygmomanometer in patients suffering from atrial fibrillation. Blood Press Monit 2011; 16:252–257.
© 2013 Wolters Kluwer Health | Lippincott Williams & Wilkins