Kong et al. 2025 — Dietary salt intake and cardiovascular outcomes: an umbrella review
Citation
Kong F, Liu Q, Zhou Q, Xiao P, Bai Y, Wu T, Xia L. “Dietary salt intake and cardiovascular outcomes: an umbrella review of meta-analyses and dose-response evidence.” Annals of Medicine 2025;57(1). doi:10.1080/07853890.2025.2582065. PMID 41243115. Published online November 16, 2025.
Study design
Umbrella review — a meta-analysis of meta-analyses. Synthesized 21 meta-analyses covering 91 distinct cardiovascular and mortality outcomes, drawing on both randomized controlled trials and observational cohort studies. Dose-response analysis performed where data permitted.
Key findings
Low sodium intake → protective outcomes
| Outcome | Effect measure | Finding |
|---|---|---|
| All-cause mortality | RR | 0.88 (−12%) |
| Cardiovascular mortality | RR | 0.83 (−17%) |
| Stroke mortality | RR | 0.74 (−26%) |
| Systolic BP | change (mmHg) | −3.39 |
| Diastolic BP | change (mmHg) | −1.54 |
High sodium intake → adverse outcomes
| Outcome | Effect measure | Finding |
|---|---|---|
| Cardiovascular disease | RR | 1.13 (+13%) |
| Hypertension | OR | 1.33 |
| Stroke mortality | OR | 1.40 (+40%) |
| Ischemic stroke | OR | 1.60 (+60%) |
Dose-response
Per +1 g/day sodium:
- Stroke risk: +6% (significant)
- Cardiovascular risk: ~+4%
- Systolic BP: +0.60 mmHg
The dose-response was linear, with no J-shaped or U-shaped pattern detected. This directly refutes the PURE J-curve hypothesis (O’Donnell/Mente 2014 NEJM) for mortality outcomes.
RAAS counter-regulatory response at low sodium
At low sodium intake, RAAS activation was mild:
- Aldosterone: +73.20 pmol/L
- Plasma renin activity: +2.09 ng/mL/h
- Catecholamines (epinephrine, norepinephrine): no significant change
Authors interpret this as evidence that the body adapts under sustained moderate restriction without adverse neurohormonal compensation sufficient to reverse the BP benefit.
Regional heterogeneity
The BP-lowering effect of low sodium was significant in Western Pacific, European, and Southeast Asian regions but not significant in the Americas (systolic −2.80 mmHg, p>0.05). Stroke risk from high sodium was strong in Asia (RR 1.13 per +1 g/day) but non-significant in Europe (1.01) and Americas (1.08).
Authors’ attribution: Asian vulnerability likely reflects higher baseline sodium intake (soy sauce, pickled foods, preserved vegetables) combined with lower potassium counter-intake.
Salt sensitivity subgroup: Black hypertensive individuals showed systolic BP reduction of −7.83 mmHg versus −2.11 mmHg in White normotensives — a ~3.7-fold difference consistent with higher prevalence of low-renin salt-sensitive hypertension in Black populations.
Authors’ conclusions
High salt intake is associated with adverse cardiovascular outcomes; low salt intake is associated with protective outcomes. A linear dose-response favors lower intake without evidence for a J-curve optimum. Recommend <5 g salt/day (<2 g sodium/day). Special attention warranted for high-baseline-intake regions (Asia).
Limitations
- Ecological heterogeneity: the umbrella review aggregates meta-analyses that themselves aggregate primary studies with heterogeneous designs, populations, and sodium measurement methods (24-hour urine vs. spot urine vs. dietary recall)
- Reverse causality in observational studies: individuals with illness (diabetes, CKD, heart failure) may self-restrict sodium, creating an apparent association between low sodium and poor outcomes that is not causal
- Sodium measurement imprecision: spot urine underestimates sodium intake in low-intake populations (circadian variation); 24-hour collections are the gold standard but logistically difficult
- Publication bias in included meta-analyses cannot be excluded
Relationships
- Directly refutes: odonnell-2014-pure-sodium (J-curve hypothesis)
- Supported by: neal-2021-ssass-salt-substitute (hard endpoints)
- Contextualizes: sacks-2001-dash-sodium (mechanistic RCT)
- See also: sodium-restriction (synthesis page)