blood pressure and low sodium diet

a new analysis of several older studies looked at the association of sodium urinary excretion, a marker of sodium intake, finding a complex relationship between that and cardiovascular events in patients both with and without hypertension (see htn sodium cv events lancet2016 in dropbox, or doi.org/10.1016/S0140-6736(16)30467-6​). details:

--133,118 individuals (63,559 with hypertension; 69,559 without) from 4 large prospective studies: PURE (prospective urban rural epidemiological study, 156424 people; EPIDREAM with 17453 people, ONTARGET with 25620 people, TRANSCEND with 5926 people)

--median age 55 (58.6 in those with hypertension, 50.5 in those without), from 49 countries, mean sodium excretion 4850 mg/d, 50% women, 45% from Asia, 40% <high school education, BMI 26, 60% nonsmokers/20% current smokers (though less in hypertensives: 16%), 40% medium/40% high physical activity, 33% current drinkers/63% never, 14% diabetic (more in hypertensives), mean BP 148/76 in hypertensives, 121/76 in non, 20% with CVD, 16% on statins (22% in hypertensives). median follow-up of 4.2 years. 

--primary composite outcome: death and major cardiovascular events (MI, stroke, heart failure)

--findings:

    --in those with hypertension (6835 clinical events, in 11%) vs normotensive:

        --increased sodium intake is associated with 2.08 mmHg change in systolic BP per g sodium increase

        --as compared to sodium excretion of 4-5 g/day (25% of the hypertensives):

            --sodium excretion of ≥​7 g/d (7060 people, 11% of the hypertensives) there was a 23% increase in CV events [HR 1.23 (1.11-1.37), p<0.001]

        ​    ​--sodium excretion of <3  g/d (7006 people, 11% of the hypertensives) there was a 34% increase in CV events [HR 1.34 (1.23-1.47), p<0.001]

    --in those normotensive (3021 events, in 4%), comparing those with 4-5 g/d sodium (18508 people, 27% of the population):

        --those excreting ≥​7 g/d (6271 people, 9% of the population) had no increased risk of cardiovasc events

        --those excreting <3 g/d (7547 people, 11% of the population), had a 26% increased risk of events [HR 1.26 (1.10-145), p-0.009]

--therefore, low sodium intake is associated with more deaths and major cardiovasc events in all comers, but high sodium intake only in hypertensives

so, this brings up several issues:

--several prior epidemiological studies have found that there is a "U"-shaped relationship between sodium intake and cardiovascular events or mortality, including different populations: diabetics, those with vascular disease, and in the general population. other studies have found that the relationship of sodium intake with cardiovasc events is much more so in those who are hypertensive

--several caveats:

    --they used a single morning fasting urine sample to estimate 24-h urinary sodium excretion. though there is a reasonably good correlation with a 24-hour specimen, older studies found that the best assessment was with three 24-hour specimens, given the daily variability of sodium intake

    ​--there are always concerns about epidemiological observational studies (and therefore they cannot prove causation). Is the relationship between low sodium and increased clinical events because of the low sodium ingestion? or is the sodium ingestion low because they have underlying diseases that lead to decreased food or decreased salt consumption? although there do seem to be some "salt-sensitive" people who get more increased blood pressure with more salt ingestion (and, conversely, "non-salt-sensitive "people who don't), does this one-time measurement of the blood pressure at baseline truly reflect what would happen over time in those on a very high salt diet? the bottom line is that it would be a tad hard to devise a real nutrition intervention study, with people randomized to long-term differences in salt consumption (or pretty much any other dietary intervention)

    --another question is about the relationship between salt and cardiovasc events themselves: are the cardiovasc events from the increase in blood pressure in those who become hypertensive? or, is it because of high renin, aldosterone, catecholamine levels in those on low salt diets (which by themselves may be associated with cardiac events)? or some combination? or something else which may be associated with the hypertension? the RAS system causation would be supported by the fact that low sodium intake in the above meta-analysis is associated with cardiac events in both hypertensives and normotensives.

    --also, a surprisingly low number of patients (pretty high risk, it seems) were on statins. Would using more statins decrease the cardiovasc events? Perhaps moreso in the hypertensive group since they have a higher absolute risk of cardiovasc disease.

 --and, we do need to be careful in extrapolating from our (perhaps primitive) models of disease leading to major public policy shifts. high sodium can cause hypertension (at least short-term in some individuals). should that be extrapolated to major public policy (perhaps activation of renin-angiotensin-aldosterone and the sympathetic nervous system is more harmful!!). or, the fact that some lipids cause heart disease. should that be translated into low lipid/high carbohydrate diets as happened in the 1970's? (the short answer here: NO)

--perhaps the good news is that even when we recommend very low salt intake, on a population basis, that doesn't seem to actually happen..... (see prior blog below). But there are pretty consistent data that a very low salt and perhaps even a lowish salt diet are not necessary. Not sure exactly what the cut off should be. But berating patients to eliminate any trace of salt in the diet is probably not appropriate.

see http://gmodestmedblogs.blogspot.com/2013/11/dietary-sodium-and-disease.html  for an Institute of Medicine report from 2013 which questioned lowering sodium intake too much and suggested a higher target (eg 2300 mg/d) in both hypertensive and normotensive people. includes reference to MMWR which showed that overall sodium intake is increasing in the US and there there are in fact minimal differences in intake in those who are supposed to be taking 2300 mg or 1500 mg/d