The study had a median follow-up of 5.5 years. Baseline characteristics of the study population are shown in Table . Randomly selected controls (n = 1,448) were expectedly healthier at baseline than cases, as indicated by a lower prevalence of hypertension, diabetes, and CVD.
Baseline characteristics of the study population
Baseline urinary excretions and dietary intakes are presented in Table . In the random sample, 24-h urinary sodium excretion estimated from overnight urine collection was 117 mmol (i.e., 2.7 g/day, which corresponds to a NaCl intake of 6.8 g/day). Urinary potassium excretion was 45 mmol/24 h (1.8 g/day), which was half the amount estimated by food frequency questionnaire (3.6 g/day). The correlation between urinary and dietary potassium was 0.21 (P < 0.001).
Baseline urinary excretions and dietary intakes of Dutch men and women aged 55 years and over: The Rotterdam Study
RR for cardiovascular events and all-cause mortality per 1-SD increase in 24-h urinary sodium are presented in Table . Urinary sodium was not significantly associated with incident myocardial infarction, incident stroke, or overall mortality. For CVD mortality, however, a borderline significant inverse association was observed (RR = 0.77 (0.60–1.01) per 1-SD, model 3) but the relationship was attenuated after excluding subjects with a history of CVD or hypertension (RR = 0.83 (0.47–1.44) per 1-SD, model 3). In subjects initially free of CVD, the risk of all-cause mortality was also examined across quartiles of 24-h urinary sodium (median values: 45, 87, 125 and 190 mmol, respectively). RR in consecutive quartiles, using the lower quartile as the reference, were 0.80 (0.43–1.49), 0.66 (0.34–1.27) and 0.98 (0.54–1.78), respectively (model 3). In a subgroup analysis of CVD free subjects with a body mass index ≥25 kg/m2, the association of urinary sodium with CVD mortality or all-cause mortality was neither statistically significant (RR = 0.91 (0.44–1.89) and RR = 1.19 (0.86–1.66) per 1-SD, respectively; model 3).
Relative risk of urinary sodium with cardiovascular events and all-cause mortality in Dutch men and women aged 55 years and over
Findings for potassium are presented in Table . Urinary potassium tended to be positively associated with incident CVD events or mortality, especially in subjects who were initially free of CVD and hypertension. After full adjustment for confounders (model 3), however, none of these associations were statistically significant. Urinary potassium did neither predict all-cause mortality. For dietary potassium, similar results were obtained except for risk of all-cause mortality that was significantly reduced both in the entire cohort (RR = 0.78 (0.65–0.94 per 1-SD) and in subjects initially free of CVD and hypertension (RR = 0.71 (0.51–1.00), model 3).
Relationship of urinary and dietary potassium with cardiovascular events and all-cause mortality in Dutch men and women aged 55 years and over
Data for urinary sodium/potassium ratio (Table ) showed no relationship with CVD events and mortality. When restricting this analysis to CVD free subjects with a body mass index ≥25 kg/m2, urinary sodium/potassium ratio was significantly associated with all-cause mortality (RR = 1.19 (1.02–1.39) per unit, model 3), but not with CVD mortality (RR = 0.86 (0.60–1.25)).
Relationship of urinary sodium/potassium ratio with cardiovascular events and all-cause mortality in Dutch men and women aged 55 years and over