sodium and hypertension/mortality

2 international studies were published in new engl journal of medicine last week, from the PURE investigators (Prospective Urban Rural Epidemiology).

1. 102K people from 18 countries (low, middle and high income countries, on 5 continents) estimating 24-h urinary sodium and potassium excretion (which reflects intake) from a single fasting urine sample and blood pressure (see htn sodium potassium excretion nejm 2014 in dropbox, or DOI: 10.1056/NEJMoa1311989)​. findings:

--regression analysis showed 2.11 mmHg increase in systolic and 0.78 in diastolic for each 1-g increment in estimated sodium excretion. (average sodium excretion 4.93 g/d)

--the association was nonlinear (and similar results were found for diastolic readings): 

        --in those excreting >5gm/d, there was an increment of 2.58 mmHg systolic

        --in those excreting 3-5 gm/d, there was an increment of 1.74 mmHg systolic

        ​--in those excreting <3 gm/d, there was an increment of 0.74 mmHg systolic

--in those with hypertension, the association was stronger (2.49 mmHg/gm) than in those without (1.30 mmHg/gm)

--in those >55 years old, the association was also stronger (2.97 mmHg/gm), than in those 45-55 (2.43 mmHg/gm), and in those <45yo (1.96 mmHg/gm)

--potassium excretion was inversely associated with hypertension, with greater association in those with hypertension and older people (average potassium excretion 2.12 gm/d).

        --for each increment of 1g of potassium/d, there was a decrement of 1.08 mmHg in systolic and 0.09 mmHg in diastolic

--there was a strong linear association between the sodium-to-potassium ratio and both systolic and diastolic pressures: a 1-SD increment in sodium-to-potassium ratio (3.26) was associated with a 2.30/0.78 increment in blood pressure
--these associations were observed in all geographic regions, though the slope was less steep in the middle east (iran, turkey, UAE, palestine).

so, this data basically supports the institute of medicine report (blog on that report attached below), also suggesting a non-linear relationship and purporting potential risks of lowering sodium intake too much. there have been a slew of epidemiologic studies over the past several decades supporting the beneficial association of a high potassium diet. it certainly makes sense from the epidemiologic observational studies that in hypertensive patients we should decrease sodium intake in those with very high intake and encourage high potassium intake (i have encouraged people to use a potassium instead of sodium supplement, with some success), but bearing in mind that there are no intervention studies (that i know of) which have proven that strategy.

2. same study, though in 17 countries, which looked at mortality and cardiovascular events (see cardiovasc mortality sodium potassium nejm 2014 in dropbox, or DOI: 10.1056/NEJMoa1311889​), with a followup of 3.7 years. findings, for composite outcome of death and major cardiovascular events:

--composite outcome in 3317 people (3.3%)

--as compared to excretion of 4-5.99 gm/d, those with sodium excretion >7gm/d had increased composite outcome with OR 1.15 (1.02-1.30), with elevations of both cardiovasc events and mortality individually, but this risk was attenuated and not significant in nonhypertensives

--in those with hypertension, with increased risk at excretion of 6 gm/d --

--but, sodium excretion of <3 gm/d was also assoc with increased risk (vs 4-5.99 gm/d), with  OR 1.27 (1.12-1.44)

--for potassium excretion, higher than 1.5 gm/d assoc with reduced risk of composite endpoints

unclear reason why there was a J-shaped curve here. ?reverse causation (people with cardiovascular disease reducing their sodium intake). in this study those with diabetes and a history ofcardiovasc disease did in fact have lower sodium intake, but the vast majority of  people did not have reported cardiovasc history. no difference in outcome if exclude these patients, or wait 2 years before analysis (which should exclude some of those with more severe disease). both the association with sodium and potassium were significantly attenuated by controlling for hypertension, suggesting that hypertension may be the mediator of the association with mortality and cardiovascular disease. There is a potentially important methodologic issue here in that prior studies from the 1970s, as I remember, found that accurate sodium intake required assessing three 24-hour urine samples and not just one 24-hour sample (let alone one fasting sample only)

geoff

link: https://www.dropbox.com/sh/0bmvtita8mzms11/XDTwHySFFg

go to "clinic", then to either "clinical stuff" for articles, or "powerpt presentations" for the powerpoint presentations

feel free to share this email with others

blog from 6/4/13 -- IOM report:

somewhat striking report by the Institute of Medicine on the health effects of limiting sodium consumption  (See sodium and disease IOM report 2013 for the summary, or the 170 page full report as sodium and disease IOM full report 2013 in dropbox).  Current guidelines suggest overall lowering sodium intake to 2300mg/d for those over 14yo,  with target of 1500 mg/d in higher risk people (African-Americans, those with diabetes/hypertension/chronic kidney dz, and those over 51 yo – ie, a majority of the population). The concern is that lowering sodium intake may not be entirely positive.  For example, low sodium intake is assoc with high plasma renin activity (PRA), which in some studies is itself assoc with increased CVD risk.  In one study of high risk pts with atherosclerosis or diabetes, PRA was independently assoc with hypertension, left ventricular hypertrophy, abnormal lipids, and insulin resistance.  This detailed review suggested:

-- there are significant methodologic issues with the studies overall (quantitation of sodium intake, most studies are observational, many done in countries with very different diets, hard to disentangle sodium consumption from consumption of other nutrients)

-- no consistent evidence of relationship between sodium intake and noncardiovasc outcomes. Some data from outside US where very high sodium intake that there may be more gastric cancer

-- In terms of blood pressure, there seems to be quite large variability, with some patients more salt sensitive than others. In the aggregate, there is a positive relationship between sodium intake and blood pressure, esp down to 2300 mg/d. not enough evidence to conclude lowering sodium below 2300 mg/d is beneficial. 

--for prehypertensive people, some data that decreasing sodium to 2300 mg/d is beneficial, but no benefit and maybe harm in the 1500-2300mg/d range, esp in those with diabetes, chronic kidney disease or pre-existing CVD. No relevant data for the other “high-risk” groups of African-Americans or those >51yo.

--for CHF, some data it might be harmful to lower sodium too much: one study achieving 1840 mg/d assoc in pts with mod to severe heart failure with low ejection fraction on aggressive med regimen found worse outcomes (though using different regimen than we use in US, so needs more studies)

--for CKD progression, data inconsistent that lowering Na helps

--for DM or metabolic syndrome, studies inadequate to draw conclusions

Turns out that we are not in serious danger of too little sodium – since the guidelines were published to decrease sodium intake with the 2300mg and 1500 mg targets, the MMWR found that sodium intake was actually increasing, with minimal difference between those who should be taking 2300 vs 1500 mg/d….   so, good that there are no consistent data supporting decreasing the level to below2300mg for anyone, even high risk.