walkitout (walkitout) wrote,

JAMA study, sodium in urine


Anyone got access to the whole thing?

I'm trying to find out whether they checked to see if people were taking diuretics.

[ETA: It turns out they were not. No one being treated for hypertension was allowed in. Anyone who developed and received treatment for hypertension was excluded. See below.]

"Of 3681 participants without CVD, 2096 were normotensive at baseline and 1499 had BP and sodium excretion measured at baseline and last follow-up (2005-2008)."

I think that means that the 1499 weren't normotensive [ETA: totally wrong], but I don't see any indication as to whether the 2096 were normotensive because they were taking antihypertensive drugs like diuretics or unmedicated. I think it might matter a lot, because another who is normotensive because they are taking diuretics is probably excreting a whole lot more sodium than someone who is not hypertensive and not taking diuretics -- and that latter person would be expected to have a higher mortality risk during the duration of the study. But the two people in question might have _exactly_ the same diet, so you wouldn't really be measuring what's in the diet but rather the effects of medication.

I would hope they would have allowed for this possibility (because it seems like a really obvious, big problem), but I don't see anything in the abstract about it. From the Results in the Abstract, this:

"Incident hypertension was 187 (27.0%; HR, 1.00; 95% CI, 0.87-1.16) in the low, 190 (26.6%; HR, 1.02; 95% CI, 0.89-1.16) in the medium, and 175 (25.4%; HR, 0.98; 95% CI, 0.86-1.12) in the high sodium excretion group."

would be _utterly_ compatible with my hypothesis (viz. taking diuretics works to lower hypertension, reduce CVD risk, and it does it all by making you dump sodium).

[ETA: So that theory was completely wrong. However, the reality turned out to be far, far worse.]

ETA: I _cannot_ believe this. A google news search on "jama sodium" (not in quotes in the search) turns up unbelievable amounts of coverage. "jama sodium diuretic" turns up nothing. Come on, people: just tell me they _accounted_ for this and I'm totally wrong. That's all I want to know.



I _luuuurrrrvve_ the interwebs.

This is the symposium version of the article, listed as the work of Katarzyna Stolarz-Skrzypek, MD, PhD here:


It is a powerpoint, so I uploaded it into google docs which sort of mangles it but it is mostly readable.

The people recruited into the original study and available at _baseline_ were 3836 white Europeans from five countries who had never been treated for hypertension. At baseline.

Here's what happened to cut that down to only 1499 people who had 24h urine at baseline and followup at 6.1 years and did the questionnaire and so forth.

(1) 222 died
(2) 20 severely ill
(3) 269 moved out
(4) 386 refused
(5) 590 24 h urine missing

separate from those, they excluded "346 24h urine" and undercollected "504 AHT drugs"

If you add up all those numbers, you do indeed get the difference between 3836 and 1499.

Let's think about that.

You excluded the people who died, were sick, or were taking AHT drugs (<--- antihypertension, as in diuretics). (Moving out, refusing and missing samples we will assume happens randomly, altho it does _not_ since some of that moving out might also be being dead). Those three categories are _half_ the size of your resulting 1499 that supposedly demonstrated that salt has no discernable effect on mortality or blood pressure.

I am completely flabbergasted. This study in no way says what people think it says. You cannot reasonably produce the kind of abstract/press release that went out and got so much coverage and _never_ mention the 504 people who got removed from the data because they were taking antihypertensives and your study is only looking at people never treated for antihypertensives. Hell, that's 13% of your _baseline_. I honestly cannot _imagine_ a signal from sodium that could survive this study structure. Well, maybe the kind of signal that tobacco constituted, but no one thinks salt is that bad.

ETA Still more: The five countries are : B I CZ RUS PL which I think means Belgium, Italy, Czech Republic, Russia, Poland. Over half of the participants seem to be female (in CZ and RU, 60.9% female and 63.6% female). BMIs at baseline are listed by country, but I don't know if it's median or average -- either way, they are all under 25. Baseline blood pressures are _really_ low: 121.6/ 74.3 119.9/77.2 116.7/74.3 115.8/73.4 120.2/75.2. This is a _heavily_ sorted population by this point. Followup bps: 124.4/ 77.7 126.0/83.5 116.7/77.7 119.3/77.1 130.0/90.0. That suggests to me that within this sorted population, the blood pressure increase typical of increasing age in adults dominates any other effect. They did the usual adjustments but also for "use of female sex hormones and NSAIDs".

Amazingly (I think this may be the most shocking part), here are the symposium slide conclusions:

"Salt intake, as exemplified by 24-h UVNa, was consistently and positively associated with SBP in cross-sectional analyses both at baseline and follow-up."

"Changes in SBP over time follow changes in 24-h UVNa."

More salt = higher systolic blood pressure.
Changes in systolic blood pressure track urinary sodium.

How did this get turned into "dietary salt has no impact on blood pressure"?

I can't speak to the assertions about CVD mortality, but given who got ripped out before the analysis occurred, I also cannot imagine it matters.
  • Post a new comment


    default userpic

    Your reply will be screened

    Your IP address will be recorded 

    When you submit the form an invisible reCAPTCHA check will be performed.
    You must follow the Privacy Policy and Google Terms of use.