Dietary Na +, K +, and Na +/K + ratioĭ.J.A.J. No associations were seen between adherence to either diet and blood pressure amongst participants on the dietary portfolio diet ( Supplemental Table 4). Among participants on the DASH-type control diet, no significant correlations were seen between portfolio adherence and blood pressure outcomes however, DASH adherence showed significant negative correlations with systolic (ρ = −0.35, P = 0.001), diastolic (ρ = −0.24, P = 0.033) and mean arterial pressure (ρ = −0.30, P = 0.005). Change in all three key dietary portfolio components were also significantly related to reductions in mean arterial pressure: nuts (ρ = −0.15, p = 0.016), soy (ρ = −0.16, p = 0.015), and viscous fiber (ρ = −0.17, p = 0.009), similar associations were seen for systolic and diastolic blood pressures ( Table 2). No significant associations were seen with DASH adherence. Using spearman correlations, and pooling test and control treatments, change in dietary portfolio adherence was related significantly to change in diastolic (ρ = −0.18, p = 0.006) and mean arterial pressure (ρ = −0.16, p = 0.011). Table 2 Change in dietary variables versus change in blood pressure, CHD, and CVD risk. A sensitivity analysis using generalized estimating equations for the reduction in systolic blood pressure values were significant at 2.1 mm Hg (CI, 4.1 to 0.1 mm Hg) (p = 0.040). 1) with no significant changes on the control diet, resulting in relative reductions in blood pressure on the dietary portfolio: systolic 2.1 mm Hg (CI, 4.2 to −0.1 mm Hg) (p = 0.056) diastolic, 1.8 mm Hg (CI, 3.2 to 0.4 mm Hg) (p = 0.013) and mean arterial pressure, 1.9 mm Hg (CI, 3.4 to 0.4 mm Hg) (p = 0.015) ( Table 1). Using completer data from 12 to 24 weeks in the unadjusted model for the primary analysis systolic, diastolic, and mean arterial pressure were reduced on the portfolio diet by 2.5 mm Hg (CI, 3.7 to 1.2 mm Hg) (p < 0.001), 2.0 mm Hg (CI, 2.8 to 1.2 mm Hg) (p < 0.001), and 2.1 mm Hg (CI, 3.0 to 1.3 mm Hg) (p < 0.001), respectively, ( Fig. No changes were observed in use of blood pressure medications throughout the study.
In the portfolio and control diet groups, 28 (18%) and 20 (24%) participants, respectively, were taking blood pressure medications. There was no significant difference in the number of people taking blood pressure medications by treatment (Fisher's exact test, p = 0.235). Table 1 Anthropometric and serum measures of portfolio and control diets, and the between-treatment differences. The mixed models were also run excluding participants on blood pressure medications, as well as using all participants providing at least one post-randomization measure (intention-to-treat). A sensitivity analysis was conducted using generalized estimating equations (PROC GENMOD) with an exchangeable correlation structure to validate the findings obtained from the mixed-effects model. C-reactive protein (CRP) values were log transformed to satisfy distributional assumptions. The results are presented both with and without adjustment for changes in covariates including waist, BMI, age, sex, blood pressure medication and baseline values. A repeated measures mixed-effects model, using compound symmetry covariance (PROC MIXED, SAS version 9.4), was used to estimate least squares mean within and between-treatment changes for all variables. This measure combines systolic and diastolic blood pressure in the ratio 1 SBP: 2 DBP. The primary outcome was change in mean arterial pressure from baseline to week 24 in the completers. Therefore the two dietary portfolios were considered here as a single treatment with 241 participants completing the study.