Salt and Cardiovascular Disease: Insufficient Evidence to Recommend Low Sodium intake

Martin O'Donnell; Andrew Mente; Michael H. Alderman; Adrian J.B. Brady; Rafael Diaz; Rajeev Gupta; Patricio López-Jaramillo; Friedrich C. Luft; Thomas F. Lüscher; Giuseppe Mancia; Johannes F.E. Mann; David McCarron; Martin McKee; Franz H. Messerli; Lynn L. Moore; Jagat Narula; Suzanne Oparil; Milton Packer; Dorairaj Prabhakaran; Alta Schutte; Karen Sliwa; Jan A. Staessen; Clyde Yancy; Salim Yusuf


Eur Heart J. 2020;41(35):3363-3373. 

In This Article

Salt and Health


The importance of sodium to human physiology suggests that its association with health is likely to be J-shaped.[46] Sodium is an essential nutrient, required for numerous physiologic processes, and is tightly regulated by numerous processes (renal, endocrine, immune, and neural, in the regulation of serum osmolality) to maintain serum sodium within normal range.[8] Sodium exchange with potassium is a key part of the action potential of human cells. Salt appetite, first described by Richter in 1936, refers to the physiological pursuit of salt in diet, and falls under regulatory controls; this concept is supported by ecological evidence of stability in populations.[47] In most people, normal kidneys are capable of excreting large amounts of sodium when intake is high, or can excrete very little sodium when intake is low. This ability to excrete sodium is impaired in the presence of low potassium diets.[48]

Our understanding of the physiology of sodium intake, storage, and excretion continues to evolve. For example, recent evidence suggests that sodium excretion demonstrates circa-septan variability and that the majority of sodium is stored in skin, subcutaneous lymphatic networks and in muscle and bone, and sodium storage is partly regulated by the immune system (suggesting a potential evolutionary antiseptic role), for which sodium appears to have a modulating effect.[49,50] The level of sodium intake affects the renin-angiotensin-aldosterone system (RAAS); low sodium intake activates the RAAS (Figure 2), as shown in short-term studies,[53,54] and increased plasma renin levels are associated with increased cardiovascular risk.[55,56] Activation of the sympathetic nervous system with impairment of reflex homeostatic control has been reported with low sodium intake.[57,58] Large reductions in sodium intake also affect cardiovascular dynamics, including baroreflex-mediated regulation. In animal models, low sodium intake is associated with attenuated pressor responses during stress (e.g. septic shock),[59,60] and with impairment in baroreflex function and adverse neuro-adrenergic effects. Low sodium intake also appears to affect membrane sodium transport, with reductions in sodium efflux in animal studies.[61] Niche hunter-gatherer populations with very low sodium intake have been reported to have a low prevalence of hypertension and excessive activation of the RAAS,[62] but the long-term effects of low sodium intake on health cannot be evaluated in these populations, as their mean life expectancy was short (~40 years). In animal studies, high sodium intake is associated with aortic hypertrophy and development of extracellular matrix, leading to increased arterial stiffness and alternations in vascular smooth muscle cells, which are reversed with reduced sodium intake. The association of high sodium intake and arterial stiffness appears to be independent of blood pressure effects.[63] In human studies, high sodium intake is associated with vascular structure and function, with high sodium intake is also associated with increased carotid intimal wall thickness, arterial stiffness index, but the pattern of association suggests a J-shaped relationship.[64]

Figure 2.

(A and B) Association of sodium intake with (A) plasma renin activity and systolic blood pressure; and (B) mortality and cardiovascular events (adapted from and O'Donnell et al. 51 and Brunner and Gavras52).

Historical Perspective

Salt has been credited with facilitating the transition from hunter-gatherer to settled communities, as it permitted preservation of food during winter months. The concept that salt is essential to life resulted in its prominence in many rituals and religious ceremonies in diverse cultures. Intuitively, manual workers understood the need for replacement of salt and water, as both were obvious constituents of sweat, and the innate behavioural drive for both water and salt is now known to be driven by robust neurohormonal control.[7] The word 'salary' (from the Latin 'salarium') originated as payment to Roman soldiers to purchase salt, reflecting its importance to daily life. In selecting the salt tax to protest British rule, Gandhi stated that 'next to water and air, salt is perhaps the most vital to health', appreciating its visceral appeal to the general population. The British colonizers of India understood the critical importance of salt to life and so imposed heavy taxes on salt to increase their revenues, resulting in large surges in mortality in parts of India under British rule during the hot months.[65] Ecologically, the countries with the highest mean life expectancy are many of those with the highest mean sodium intake, although a curvilinear association is suggested by an analysis of 182 countries.[66] Sodium reduction has fallen in and out of favour in the management of heart failure since and is now finally being evaluated in large clinical trials.[67]

An important dietary transition has been the change in sources of sodium intake, which may be derived from discretionary (e.g. table salt added to meals) or non-discretionary sources (part of pre-prepared foods, e.g. bread). There has been a gradual increase in dietary sodium from non-discretionary sources, most marked in high-income countries. Non-discretionary sources of salt intake account for 75–80% of salt intake in Europe and North America, but lower in other regions (e.g. China and Africa).[68] The food industry adds salt to foods to extend shelf-life and/or to improve taste. With the increase in non-discretionary sources of sodium, there is less individual-level control of intake and reduced ability to estimate sodium intake in the diet (as it is hidden within foods). However, sodium intake appears to be tightly regulated through a central neural mechanism, which appears to have translated into stable mean intake of sodium in population, despite transitions in sources of sodium intake.[8,46]


Blood Pressure. Overall, there is convincing epidemiologic evidence of a monotonic, curvilinear association of sodium intake and blood pressure at the population level. INTERSALT was the first large international study to examine the association of sodium intake and blood pressure cross-sectionally, included randomly sampled participants aged 20–50 years from 52 centres in 39 countries (n = 10 079) who provided a 24-h urine collection for sodium, and reported in 1988. The INTERSALT study[12,69] reported a positive association between sodium intake and blood pressure in 33 of 53 centres (statistically significant in 8), and the Scottish Heart study,[70] published at the same time, reported no significant association. The largest international study to report on the association of sodium intake and blood pressure was the Prospective Urban Rural Epidemiology (PURE) study,[71] which included 102 216 adults from 18 countries. PURE reported a positive curvilinear association of sodium intake with blood pressure (2.11/0.78 mmHg per 1 g/day increase in sodium), which was only significant for sodium intakes above 3 g/day and strongest in those with consumption above 5 g/day (2.58 mmHg per 1 g/day increase in sodium). The association between sodium intake and blood pressure was stronger in older populations, those with hypertension and those consuming lower potassium diets. The largest cohort study to report on the association of sodium intake and blood pressure was the recent UK-Biobank study,[32] which also found a positive association of sodium intake with blood pressure, with an effect size generally consistent with results of prior observational studies. The EPOGH/FLEMINGHO study[72] examined the association of baseline sodium intake (measured with 24-h urine collections) with change in blood pressure after 6 years follow-up (n = 1499) and reported an increase in systolic blood pressure (1.71 mmHg per 2.3 g/day increase in sodium intake), after adjusting for covariates.

Some individuals are more sensitive to the pressor effects of increased sodium intake, a condition termed 'salt sensitivity'.[73] Identification of genetic polymorphisms that are associated with both hypertension and 'salt sensitivity' is an intensive area of research.[74] For example, studies report a gene–diet interaction between salt intake and the angiotensin I-converting enzyme insertion-deletion polymorphism (CAE I/D) for development of hypertension.[75] The current clinical relevance of salt sensitivity is uncertain, as the definition and identification of salt-sensitive populations lack uniformity, although people who develop hypertension appear to have a greater increase in blood pressure for a given increase in sodium intake, compared to those who do not develop hypertension. It is suggested, however, that identification of key polymorphisms may facilitate a future personalized approach to sodium intake recommendations.

Cardiovascular Events. Numerous prospective cohort studies have evaluated the association of sodium intake (using different measurement methodologies) with cardiovascular events and mortality.[76] Most meta-analyses of these studies compared the lowest quantile of sodium intake with highest quantile, rather than evaluating the totality of data, thereby assuming a linear association.[77,78] In contrast, Graudal et al.[28] included all categories of intake and found a J-shaped association of sodium intake with mortality and cardiovascular events, with an increased risk above 4.6 g/day and below 2.7 g/day, compared with moderate intake. Findings from this meta-analysis were consistent across methods used to estimate sodium intake. Since that meta-analysis, two large studies have been published,[29,32] one international study (PURE, n = 101 945, 7.2 years of follow-up),[29] and the UK-Biobank (n = 322 624, 7.0 years of follow-up),[32] both employing formula-derived estimates of 24-h urinary sodium excretion (proxy for intake). The PURE study reported a J-shaped association between sodium excretion and cardiovascular event incidence (CVD) incidence and mortality, with lowest risk between 3 and 5 g of sodium per day, consistent with the previous meta-analysis.[28] The PURE study reported that the increased risk associated with high sodium intake was largely confined to those with hypertension,[29] a finding also supported by the PREVEND study,[79] and the PURE study also found that the risk associated with high sodium intake was mitigated in those with high potassium intake and higher-quality diets.[80] The PURE study also found a community-level (n = 369 communities) association of mean sodium intake with cardiovascular risk, observing that only communities with high sodium intake (>5 g/day) had increased cardiovascular risk associated with sodium intake.[81] The UK-Biobank study reported no significant association of sodium intake with cardiovascular events, although there was a suggested J-shaped pattern of association with mortality.[32,82]

The association of low sodium intake with increased risk of cardiovascular events and mortality has been questioned on the basis that it has only been reported in studies using formula-derived estimates of sodium intake derived from spot urines.[5,40,43,44] However, it is important to note that the increased risk associated with low sodium intake (compared to moderate intake) has been reported in a number of studies that employed 24-h urine collections, namely the EPOGH/FLEMINGHO study (n = 3681),[72] PREVEND study[31] (n = 7330), the CRIC study (n = 3757 for myocardial infarction),[83] and a study by Alderman et al.[84] (n = 2937).

In 2019, Dietary Reference Intakes for Sodium and Potassium[5] released by the US National Academy of Medicine (NAM) did not take into account evidence from the above cohort studies and meta-analyses that supported an increased risk of cardiovascular events or mortality with low sodium intake (J-shaped relationship). Instead, the report focused on observational follow-up of the control group of the TOHP trial (n = 2275; 193 cardiovascular events or deaths).[85] The report concluded that there was a linear association of sodium intake with cardiovascular events, although there was lack of a statistically significant difference in risk comparing those with low sodium intake (<2.3 g/d) to moderate sodium intake (2.3–4.6 g/d), i.e. conclusions were not supported by evidence of a statistical difference between these intake groups.[85] The decision to include this analysis from TOHP as the main representative of observational research appeared to be based on its use of repeated 24-h urinary estimates of sodium intake.

The discordant associations of sodium intake with blood pressure (positive) and cardiovascular events (inverse) when sodium intake transitions from moderate intake to low intake levels likely relates to competing 'off-target' adverse effects of low sodium intake on RAAS, other neurohormonal activation, baroreflex integrity, potentially dysfunction in membrane sodium transport, and perhaps effects of other related dietary factors (e.g. potassium intake) (Figure 2). Inherent limitations of observational research studies, including the potential for residual confounding and reverse causation, preclude definitive conclusions on a causal association of sodium intake and clinical outcomes, which require clinical trials.[86]

Clinical Trials

Blood Pressure. Numerous clinical trials have evaluated the effects of reducing sodium intake on blood pressure. Most (>95%) were short-term trials (<6 months) with relatively small samples sizes.[53,54]

Feeding Trials: The largest 'feeding' clinical trial was the DASH-Sodium trial, which included 412 participants with pre-hypertension.[18] This was a 3 × 2 factorial trial that evaluated three different levels of sodium intake (3.3, 2.5, and 1.5 g/day) consumed for 30 days and compared a DASH diet pattern with a control diet pattern. It reported a reduction in blood pressure with lower sodium intake (change in systolic blood pressure of −2.1 mmHg on reduction from 3.3 g/day to 2.5 g sodium, −4.6 mmHg on further reduction to 1.5 g sodium when consuming the control diet, and −1.3 and −1.7 mmHg when consuming the DASH diet, respectively). The blood pressure reduction was most marked in those consuming the control diet, in which the background potassium intake (1.56 g/day) was lower than in the typical US diet (2.6 g/day),[87] which may have enhanced the effects of sodium reduction on blood pressure in the trial.[88] Findings from the DASH-Sodium trial were very influential on the subsequent specific recommendation of a sodium intake of <1.5 g/day as the optimal target for the entire adult population.

Intensive Dietary Counselling: Of clinical trials evaluating a counselling/educational intervention, the TOHP-II trial is the largest (n = 2382) to examine the effect of longer-term sodium reduction on blood pressure (mean duration of follow-up was 36 months).[17] TOHP-II employed a 2 × 2 factorial design that also evaluated an educational intervention for weight loss. Mean sodium intake at 18 months was 3.1 g/day (3.2 g/day at 36 months) despite targeting a sodium intake of <1.8 g/day in the intervention group, indicating that a target of <2.3 g/day was not achievable even in the controlled setting of a clinical trial making substantial efforts to lower sodium. The mean sodium intake in the control group was 3.9 g/day (4.0 g/day at 36 months). The difference in systolic blood pressure between sodium intake groups was 2.9 mmHg at 6 months, 2.0 mmHg at 18 months, and 1.2 mmHg at 36 months. The reduction in frequency of hypertension in the intervention group also diminished over time [risk ratio of 0.61 (P = 0.04) at 6 months, 0.88 (P = 0.28) at 18 months, and 0.82 (P = 0.05) at study end].[17] The primary outcome measure in TOHP-II, mean change in diastolic blood pressure, was not statistically significantly different between sodium reduction and control groups at 36 months (−0.6 mmHg; P = 0.1).

The TONE trial[16] evaluated a similar counselling/educational intervention in older adults (60–80 years) with controlled hypertension on a single antihypertensive agent (n = 975). Among obese participants (n = 585), the trial design was a 2 × 2 factorial, evaluating both sodium reduction diet and weight loss programme compared with control. In the non-obese group, the trial compared sodium reduction diet to control. At 90 days after the introduction of the intervention, blood pressure was reduced by 3.4 mmHg systolic in the intervention group (P<0.01). Withdrawal of antihypertensive therapy began after 90 days of the intervention. Thirty months later, a larger proportion of patients in the sodium intervention group were taking no antihypertensive therapy or had a blood pressure <150 mmHg systolic and <90 mmHg diastolic (and no evidence of cardiovascular disease) compared to control (38% vs. 24%; P < 0.001).

Meta-analyses of clinical trials have reported mean reductions in blood pressure with reductions in sodium intake generally consistent with findings from prospective cohort studies.[78,89] In one meta-analysis of 36 blood pressure clinical trials (n = 6736), a reduction in sodium intake was associated with a reduction in blood pressure (−3.39/−1.54 mmHg), greater in participants with hypertension (−4.06/2.26 mmHg) than those without hypertension (−1.38/0.58 mmHg).[78] Subgroup analyses performed by differing duration of clinical trial follow-up[78] reported between-group differences in blood pressure of −4.07/1.67 mmHg for trials at <3 months (31 trials; n = 3351), −1.91/1.33 mmHg for trials at 3–6 months (5 trials; n = 2817), and −0.88/0.45 mmHg at more than 6 months (3 trials; n = 2862), raising the possibility that long-term compensatory mechanisms may counteract the initial reduction in blood pressure from lowering salt intake.

These data demonstrate the challenges of achieving, and sustaining, a low sodium intake (<2.3 g/day) in free-living populations, despite relatively intense dietary counselling, and indicate that reducing sodium intake does reduce blood pressure, but the antihypertensive effects of sodium reduction may diminish over time. In both the TONE and TOHP-II trials,[16,17] the control group did not receive a dietary counselling intervention (e.g. advice on healthy diet without a focus on salt intake), thus the investigators were unable to determine the independent effects of sodium reduction, unrelated to overall improvements in diet quality.

Household/Community Level: An educational intervention targeting school children to reduce household salt intake in a cluster randomized controlled trial in China (28 schools, 832 members of families) reported a reduction in sodium intake (mean effect on sodium intake for intervention versus control was −0.76 g/day in children and −1.16g/day in adults (P<0.001). While systolic blood pressure increased in both intervention and control groups during follow-up, the increase was significantly less in the sodium reduction intervention group, with a mean effect −0.8mmHg in children (P=0.51) and −2.3mmHg in adults (P<0.05).[90] An earlier trial (n = 2211) that compared an educational programme (employing mass media techniques) in one town with a control town in Belgium reported a reduction in sodium intake (0.58 g/day) but no difference in mean change in blood pressure between intervention and control town after 5 years of follow-up.[91]

Salt Substitution: Clinical trials have evaluated the blood pressure effects of salt substitutes (low sodium salt), usually achieved by reducing the proportion of NaCl by about 45–65% and replacing it with KCl or MgSO2, an intervention most applicable in populations with high discretionary salt use (e.g. China). A meta-analysis of these clinical trials (five clinical trials, n = 1974) reported a reduction in blood pressure (−4.9/−1.5 mmHg) with use of salt substitutes, over a period of 6 months to 2 years.[92]

A cluster randomized controlled trial in China (China Rural Health Initiative-Sodium Reduction Study) randomized 120 villages to the intervention (community-based health education programme and access to salt substitution-reduced sodium and added potassium) or control. Thirty of the intervention villages had price subsidies, as salt substitute is more expensive than regular salt.[93] Among 1295 participants in 60 intervention villages (1063 with urine assessment), there was a 0.32 g lower sodium intake (P = 0.03) on follow-up compared to control (1272 participants, 1001 with urine assessments). There was no statistically significant difference in blood pressure (−1.1/0.7 mmHg, P = 0.39 and 0.34, respectively) between groups.

Cardiovascular Events/Heart Failure

No large individual-level randomized controlled trial designed to determine the effect of low sodium intake on clinical outcomes, including cardiovascular events and mortality, has been completed or published.

Primary Prevention of CVD. A cluster randomized controlled trial (n = 1981, 30 months of follow-up) conducted in the kitchens of five veteran's retirement homes in Taiwan, which increased potassium consumption and reduced sodium consumption through use of potassium-enriched salt, found a reduction in cardiovascular mortality (HR 0.59; 95%CI 0.37–0.95) in those assigned to the higher potassium-lower sodium group.[94] The trial was analysed using individuals as the unit, rather than using cluster-level data. Moreover, in that trial, sodium intake was estimated to be reduced from about 5.2 to 3.8 g/day (based on a subsample); with an increase in potassium intake in the intervention group.

Meta-analyses of clinical trials of dietary salt reduction that reported cardiovascular events, either during the intervention period or during extended observational follow-up, have come to differing conclusions. A Cochrane review of clinical trials reporting cardiovascular events, including the observational follow-up of the TOHP trials[97] and the trial by Chang et al.[94] concluded 'there is insufficient power to confirm clinically important effects', but did report a 19% reduction in CVD in one analysis (RR 0.81; 0.66–0.98; six clinical trials, n = 5762) with no significant effect on mortality.[95] Another meta-analysis of these clinical trials (without one small trial included by the Cochrane review) by the recent NAM panel also reported a clinically important risk reduction in cardiovascular events (RR 0.72; 0.59–0.89; five clinical trials) and suggested a reduction in mortality (RR 0.89; 0.78–1.01; four clinical trials).[7] The evidence was judged to be of moderate strength and was referenced as a key piece of evidence to support recommending low sodium intake. In a more recent umbrella review ('meta-analysis of meta-analyses'), the authors reported moderate-certainty evidence for reduced salt intake to decrease risk of all-cause mortality in normotensive individuals (RR 0.90; 0.85–0.95) and cardiovascular mortality in hypertensive individuals (RR 0.67; 0.46–0.99).[96]

While findings from these meta-analyses suggest benefit from sodium reduction strategies (in conjunction with other dietary changes, e.g. increased potassium intake), they do not address the effect of low sodium intake (<2.3 g/day) per se, since neither the observational follow-up of the TOHP-II trial nor the trial by Chang et al., the two largest studies in the meta-analysis), achieved the currently recommended low sodium intake target in the intervention groups. Moreover, there was a 23% loss to follow-up in the TOHP trial for cardiovascular outcomes, but complete follow-up for mortality.[97]

Findings from these meta-analyses provide an impetus to carry-out large definitive long-term randomized controlled trials to determine the effect of lowering sodium intake on cardiovascular events in order to provide high-quality evidence to inform future guideline recommendations.

The Salt Substitute and Stroke Study (SSaSS), a large cluster randomized controlled trial of 600 villages with over 21 000 participants followed over 5 years, is being undertaken in China to evaluate the effects of salt substitution with high potassium on stroke risk in individuals with either prior stroke or at high risk of stroke.[98] Findings from the SSaSS trial will have important implications for population-level approaches to prevention of stroke in high-risk populations with high discretionary sodium intake. However, as this trial is evaluating salt substitution in a region with high sodium intake, it is unlikely to provide an answer to the effects of low sodium intake, or distinguish the effect of lowering sodium intake alone, vs. that due to increased potassium on stroke risk.

Secondary Prevention of Cardiovascular Events. No clinical trial has reported the independent effect of lowering sodium intake on cardiovascular events in patients with prior cardiovascular events.

Heart Failure. Ongoing research is examining the role of reducing sodium intake in heart failure, both to prevent events (recurrent hospitalization and mortality) and to achieve symptomatic improvement. To date, evidence to support, or dispute, low sodium intake for patients with heart failure is inconsistent. An observational analysis of the Heart Failure Adherence and Retention trial (HART)[99] reported a higher risk of death and hospitalization for heart failure among those randomized to restricted sodium intake, compared with those randomized to unrestricted sodium intake, delivered within a multi-component intervention. In contrast, Colin-Ramirez et al.[100] reported lower mortality and re-hospitalization rates in the group allocated to restricted sodium intake (2–2.4 g/day) compared with an unrestricted recommendation. A meta-analysis of nine studies (n = 479) evaluating reduced sodium intake in patients with heart failure found no evidence to support or refute benefit or harm.[101] Two ongoing randomized controlled trials are evaluating intensive dietary counselling to reduce sodium intake to a target of 1.5 g/day in the Sodium-HF[102] (n = 1000) and <2.0 g/day in the SALT trial[103] (n = 250), compared with usual care.

Guideline recommendations for sodium intake in patients with heart failure have evolved with the new evidence, and the Heart Failure Association of the European Society of Cardiology (ESC) removed the formal guideline recommendation to reduce salt intake in 2012, due to lack of evidence.[104] In the 2016 ESC heart failure guideline,[105] the panel report 'little evidence that specific lifestyle advice improves quality of life or prognosis'. While a sodium intake target was not included as a specific 'level of evidence' recommendation, the panel listed reducing sodium intake (in patients with a sodium intake of >2.4 g/day) as a key topic to include in patient education sessions.


  1. There is a positive curvilinear association of sodium intake and blood pressure, with the largest magnitude of effect observed in those consuming high sodium, older adults, hypertensives and those consuming low potassium diets.

  2. The increased risk of cardiovascular events associated with higher sodium intake is observed when sodium intake exceeds 5 g/day.

  3. Moderate sodium intake (2.3–4.6 g/day) has been consistently associated with lower cardiovascular risk, compared to both high and low sodium intake, in prospective cohort studies.

  4. The cardiovascular risk associated with high sodium intake is modified by hypertension status.

  5. The risk of high sodium intake is modified by potassium intake, in that the risk of high sodium intake is diminished in those consuming more potassium (higher intakes of fruits, vegetables, and nuts).

  6. The optimal intake of sodium in patients with heart failure is unknown.