Nonlinear mixed effects modeling of the diurnal blood pressure profile in a multiracial population

Am J Hypertens. 2013 Sep;26(9):1103-13. doi: 10.1093/ajh/hpt088.

Abstract

Background: Cardiac and cerebrovascular events in hypertensive patients are related to specific features of the 24-hour diurnal blood pressure (BP) profile (i.e., daytime and nighttime BP, nocturnal dip (ND), and morning surge (MS)). This investigation aimed to characterize 24-hour diurnal systolic BP (SBP) with parameters that correlate directly with daytime and nighttime SBP, ND, and MS using nonlinear mixed effects modeling.

Methods: Ambulatory 24-hour SBP measurements (ABPM) of 196 nontreated subjects from three ethnic groups were available. A population model was parameterized in NONMEM to estimate and evaluate the parameters baseline SBP (BSL), nadir (minimum SBP during the night), and change (SBP difference between day and night). Associations were tested between these parameters and patient-related factors to explain interindividual variability.

Results: The diurnal SBP profile was adequately described as the sum of 2 cosine functions. The following typical values (interindividual variability) were found: BSL = 139 mm Hg (11%); nadir = 122 mm Hg (14%); change = 25 mm Hg (52%), and residual error = 12 mm Hg. The model parameters correlate well with daytime and nighttime SBP, ND, and MS (R (2) = 0.50-0.92). During covariate analysis, ethnicity was found to be associated with change; change was 40% higher in white Dutch subjects and 26.8% higher in South Asians than in blacks.

Conclusions: The developed population model allows simultaneous estimation of BSL, nadir, and change for all individuals in the investigated population, regardless of individual number of SBP measurements. Ethnicity was associated with change. The model provides a tool to evaluate and optimize the sampling frequency for 24-hour ABPM.

Keywords: blood pressure; blood pressure profile; cardiovascular risk; ethnicity; hypertension; nonlinear mixed effects modeling (NONMEM); variability..

Publication types

  • Randomized Controlled Trial

MeSH terms

  • Adult
  • Asian People / statistics & numerical data
  • Bayes Theorem
  • Black People / statistics & numerical data
  • Blood Pressure / physiology*
  • Circadian Rhythm / physiology*
  • Humans
  • Hypertension / epidemiology
  • Hypertension / ethnology*
  • Hypertension / therapy
  • Middle Aged
  • Models, Statistical
  • Netherlands / epidemiology
  • Suriname / ethnology
  • Systole
  • White People / statistics & numerical data