The associations of self-reported salt-intake and spot urine sodium with home blood pressure

Abstract

Purpose: A limited number of studies have suggested a nonlinear association between spot urine (SU) sodium concentration and office blood pressure (BP). We examined how SU sodium concentration and dietary salt obtained from a food frequency questionnaire are associated with more accurately measured home BP in a large, nationwide population sample. Materials and methods: We included 1398 participants in cross-sectional and 851 participants in 11-year longitudinal analyses. We investigated associations between baseline salt/sodium variables and (i) baseline and follow-up home BP; and (ii) prevalent and incident hypertension with linear and logistic regression models. Results: We observed positive associations (β ± standard error) between salt/sodium variables and BP in unadjusted models. SU sodium concentration associated with baseline systolic (0.04 ± 0.01, p < 0.001) and diastolic (0.02 ± 0.01, p < 0.001) BP and follow-up systolic (0.03 ± 0.01, p = 0.003) and diastolic (0.02 ± 0.01, p < 0.001) BP. Dietary salt intake was associated with baseline (0.52 ± 0.19, p = 0.008) and follow-up (0.57 ± 0.20, p = 0.006) systolic BP. Compared to the lowest quintile of SU sodium concentration, the highest quintile had greater odds of prevalent hypertension (odds ratio [OR] 1.57, 95% confidence interval [CI] 1.12–2.19) and the second highest quintile with incident hypertension (OR 1.86, 95% CI 1.05–3.34). Unadjusted odds of incident hypertension were higher in the highest as compared to the lowest quintile of dietary salt intake (OR 1.83, 95% CI 1.01–3.35). After adjustments for sex, age, plasma creatinine concentration and alcohol intake, none of the aforementioned associations remained statistically significant. We found no evidence of a J-shaped association between the salt/sodium variables and BP or hypertension. Conclusion: SU sodium concentration and dietary salt intake are associated with home BP and hypertension only in some of the unadjusted models. Our results underscore that feasible estimation of sodium intake remains challenging in epidemiology.