Key points are not available for this paper at this time.
OBJECTIVES: To establish the reproducibility of the nocturnal systolic blood pressure (SBP) change in elderly subjects and to examine the use of cumulative sums (cusums) analysis in the assessment of circadian SBP variation. SUBJECTS: Forty-two untreated elderly subjects (35 hypertensive, 7 normotensive) of mean age 75.5 years from the hypertension clinic at a large teaching hospital participated in a reproducibility study. METHODS: Twenty-four-hour ambulatory blood pressure monitoring was performed and repeated at a median interval of 2 months (range 2 weeks to 9 months). OUTCOME MEASURES: Reproducibility of circadian SBP variation from fixed time analysis of day-night SBP difference and from cusums-based parameters. RESULTS: Twenty-four-hour SBP values were highly reproducible with a coefficient of variation of 5.8%. However, the day-night SBP difference for fixed time periods was poorly reproducible, with a coefficient of variation > 130%. A substantial proportion of subjects (36-43%) altered their 'dipping status' between visits. The use of cusums analysis improved the reproducibility of measures of circadian SBP change (cusums plot height and maximum circadian variation) with coefficients of variation falling to 40 and 38%, respectively. CONCLUSIONS: The use of fixed time definitions results in poor reproducibility of the circadian SBP change in the elderly, which will lead to regression dilution bias when studying the relationship of circadian SBP variation to outcome measures in hypertension. The notion of dipping and non-dipping circadian blood pressure patterns should be abandoned in favour of more reproducible cusums-based measures of circadian blood pressure variation.
James et al. (Sun,) studied this question.