Dose-response association of dietary sodium intake with all-cause and cardiovascular mortality: a systematic review and meta-analysis of prospective studies.

Abstract

High Na intake has been associated with different health problems. However, serious controversies exist over studies investigating associations of Na intake with mortality from all-causes and CVD. The present systematic review and meta-analysis was done to investigate, for the first time, the dose-response association of dietary Na intake with all-cause and CVD mortality among prospective studies. Relevant papers published up to August 2017 were searched in MEDLINE, EMBASE and Google Scholar databases. Prospective cohort studies on the association of dietary Na intake with all-cause or/and CVD mortality were included. Linear and non-linear dose-response associations between Na intake and CVD and all-cause mortality were examined. Overall, twenty publications met inclusion criteria. A significant non-linear association (P<0·001) was found between Na intake and CVD mortality risk among studies assessing urinary Na excretion, with a relatively steep slope at Na intakes above 2400mg/d. However, the association was not significant in studies using dietary Na intake (P=0·61). Additionally, the non-linear association of Na intake with all-cause mortality was also non-significant. No linear association (effect size; 95 % CI; I 2) was seen between 100mg/d increment in Na intake and CVD mortality (1·01; 0·97, 1·05; 98·4 %) or all-cause mortality (1·01; 1·00, 1·02; 89·2 %). Following subgroup analyses, the association between Na intake and CVD mortality was observed only among studies conducted in the USA (0·99; 0·99, 1·00; 20·0 %). The study showed a direct association between urinary Na excretion and CVD mortality which was more considerable at intakes above 2400mg/d. In contrast, no significant association was found between Na intake and all-cause mortality. Further long-term prospective studies on different populations are required to confirm these findings.