This looks like it’s going to be a research sort of day, so since I’m going to be hanging out on PubMed anyways…
I couldn’t check the paper cited on Wiki* for the “only 33% respond” figure, since the article was in German… But I have found a number of meta-reviews. This one dates back a while, and it generally states that reduced sodium intake results in:
Another single study, looking at the effects of one particular low-sodium diet, found that:
A more recent meta-review is consistent with the other two citations, though it adds some more details and qualifications:
But in all of those studies, I haven’t come across anything that would support the idea that a low sodium diet has no effect on a large majority of hypertensive patients. That may be an artifact of how you define “hypertensive” – it’s a somewhat arbitrary line. The AMA defines hypertension as having a systolic blood pressure above 140. So, it might depend on how you slice the data and define your outcomes. For example, you could have a study that observes a consistent 5 mm hg drop in blood pressure. But if a majority of subjects start with a blood pressure above 145, you could also legitimately state that that those patients dropped from “hypertensive” to “still hypertensive”. That could be easily misinterpreted by some random wiki editor as being equivalent to “low salt diet doesn’t help most people”. Which is bogus, since all of the trials I’ve seen so far indicate that reducing sodium results in a statistically and clinically significant reduction in blood pressure.
After doing a bit more research, there’s a review that looks at the variation in normotensive patients. That is, if you take healthy population and feed them increased or reduced sodium, some fraction will be insensitive to the change (and won’t see any changes in blood pressure), while another fraction will be hypersensitive to dietary sodium. That’s not the same as what you quote from wiki, however.
*I’m starting to realize that, while wiki citations are usually reliable, they’re usually a very small and selective subset of actual research on the topic. I think that means that one wiki contributor that isn’t familiar with the field will read a single paper, and stick that citation in, so you don’t really see the whole context. In my subfield, for example, wiki has a not-horrible article on the topic. But the citations represent a really tiny and non-representative sample of the research – just a single paper, which is presented by the wiki editor as definitive. But it’s just one small part of the story, and the conclusions from the entire field are really spread over dozens of papers.