Noncompletion of Nighttime Ambulatory Blood Pressure Monitoring: Potential for Selection Bias in Analyses of Nondipping

    loading  Checking for direct PDF access through Ovid


Nondipping of blood pressure refers to the absence of at least a 10% reduction in average nighttime (asleep) ambulatory blood pressure (ABP) when compared with average daytime (awake) ABP (1–4). Nondipping may predict increased cardiovascular risk (1), and identifying correlates of nondipping has been the topic of a number of studies (2,5–7). Indeed, in the recent article by Rodriguez et al. (2), nondipping was more common in black non-Hispanic participants who reported low perceived racism in a sample of 180 community-dwelling Hispanic participants. This paper used an ABP monitoring (ABPM) device for measuring nondipping, and we were intrigued by the statement that “Only one participant had to repeat the ABP assessment due to initial poor-quality recordings. Otherwise, there were no protocol violations, technical failures, or insufficient number of ABP readings (p. 842)” (2).
To assess nondipping, the ABPM device must be worn during the night, with readings typically taken every 30 minutes. In most studies, there are some participants who remove the device before going to sleep or sometime in the middle of the night because of discomfort or anticipated/actual sleep disruption. Indeed, an analysis of pooled data from several studies stated that 19% of participants (1909/9828) were excluded because of nighttime ABPM noncompliance (3). Furthermore, ABPM device-related sleep disruption has been previously reported (4).
In light of these findings, we hypothesized that this disruption of sleep could be a reason for ABPM noncompliance in studies of nondipping. Furthermore, we were interested in examining whether certain participant characteristics and important correlates of nondipping differ between those who do and do not comply with nighttime ABPM. We examined these questions using data from the Masked Hypertension Study (8).
Our analytic sample comprised 878 adults (M [SD] age = 44.3 [10.4] years, 59.0% female, 12.1% Hispanic, 7.3% black) who completed 24-hour ABPM and a self-reported sleep quality questionnaire as part of the Masked Hypertension Study, a prospective cohort study of masked hypertension and target organ damage (8). Participants with less than five nighttime readings—usually due to removing the device during sleep—were classified as nighttime ABPM noncompleters. Sleep quality was reported on a 0 (much worse than usual) to 4 (much better than usual) scale the next day; poor sleep was defined as a rating of less than 2. Using nonparametric testing methods, we assessed whether completion status was related to race, ethnicity, age, sex, body mass index, depression, sleep quality the night of ABPM measurement, Pittsburgh Sleep Quality Index score, discrimination, and daytime average systolic blood pressure BP and diastolic BP.
Approximately 13.2% (n = 116) of participants were nighttime ABPM noncompleters based on the accepted definition of completion; of these noncompleters, 71.6% (n = 83) had no nighttime readings. Of the 33 nighttime ABPM noncompleters with partial data (i.e., 1–4 nighttime readings), 24.2% (n = 8) were nondippers, with a mean SBP decrease of 13.3 mm Hg and a mean DBP decrease of 14.7 mm Hg. These participants completed, on average, three nightly readings. In contrast, the 762 nighttime ABPM completers provided 14 nightly readings on average. Twenty-five percent (n = 190) of nighttime ABPM completers exhibited nondipping, with a mean SBP decrease of 16.7 mm Hg and a mean DBP decrease of 15.4 mm Hg. Approximately 39.0% (n = 343) of the analytic sample reported poor sleep. Nighttime ABPM noncompleters were less likely to report poor sleep on the night of the assessment than completers (p = .002, Fisher exact test). The common correlates of nondipping (sex, age, race, ethnicity, depression, body mass index, usual sleep quality, and reported discrimination) were not significantly related to noncompletion (all p > .09).
    loading  Loading Related Articles