Accuracy of ambulatory blood pressure monitors: a systematic review of validation studies.
Hodgkinson JA., Sheppard JP., Heneghan C., Martin U., Mant J., Roberts N., McManus RJ.
BACKGROUND: Recent research and guidelines recommend the routine use of ambulatory blood pressure monitoring for the diagnosis of hypertension, so accuracy of such monitors is more important than ever. AIM: : To systematically review the literature regarding the accuracy of ambulatory monitors currently in use. METHODS: Medline, Embase, Cinahl, the Cochrane database, Medion and the dabl Educational Trust website were searched until February 2011. No language or publication date limits were applied. Data were extracted separately by two independent reviewers. Methodological quality was assessed by whether a validation protocol had been used and followed correctly. RESULTS: From 5420 journal articles identified, 108 met the inclusion criteria. Excluding studies assessing monitors no longer in use, 40 relevant studies were found using 21 different monitors. Thirty-eight (95%) studies used a validation protocol of which 28 studies assessed a monitor in the general population. Of these, protocols were passed in 24 of 28 studies, but 12 of 24 (50%) found a difference of at least 5 mmHg systolic between the test device and the reference standard for 30% or more of the readings. Of the 10 studies conducted in special population groups (e.g. pregnancy, elderly people), only four devices passed the protocols. Only six (16%) studies correctly adhered to the protocols. CONCLUSION: Published validation studies assessed most ambulatory monitors as accurate, but many failed to adhere to the underlying protocols, undermining this conclusion and peer review standards. Furthermore, most monitors which 'passed' validation showed significant variation in blood pressure from the reference standard, highlighting inadequacies in older validation protocols. Future validation studies should use protocols with simpler methodologies but more rigorous accuracy criteria.