Plant protein tracked lower death rates, animal did not

The advice to swap some meat for beans is old, and the evidence behind it comes mostly from people already old enough for the outcome to arrive quickly.

That leaves a gap. Diets that matter for how long you live are presumably set decades earlier, in the years when nobody is dying and so nothing can be measured.

A cohort at the University of Navarra has been working on the gap since 1999. Its latest analysis, in the European Journal of Nutrition, followed 17,989 Spanish graduates whose mean age at baseline was 38 years for a median of 12 years, and asked whether the source of their protein predicted anything.

Plant protein did. Animal protein did not.

What the protein comparison actually measured

Protein is protein, chemically, and the body takes it apart into amino acids regardless of origin. What differs is everything that arrives alongside it: fiber and polyphenols with the lentils, saturated fat and heme iron with the steak.

So a study of plant against animal protein is never only about protein. It is a study of two eating patterns, indexed by the nutrient that separates them most cleanly.

The team measured intake with a semi-quantitative food frequency questionnaire (FFQ) previously validated in Spain, sorted people into quarters by how much of each type they ate, and compared the top quarter against the bottom.

What repeated protein measurement adds

The methodological contribution here is small and matters. A limited number of studies have performed repeated measurements of protein intake.

Almost every cohort of this kind records diet once, at recruitment, and then attributes twenty years of outcomes to that single snapshot. Anyone who changed how they ate is filed under what they used to eat.

This analysis evaluated the relationship between plant versus animal protein intake at baseline and at the ten-year follow-up. Two readings, a decade apart, in a population young enough for the second to still precede most of the deaths.

What happened to death rates over twelve years

During 251,363 person-years of follow-up, 460 deaths were identified.

The top quarter for plant protein had a 35% lower risk of all-cause mortality compared to those in the lowest, after adjusting for potential confounders. The trend across all four quarters ran the same way, which is more persuasive than a single top-versus-bottom contrast.

Animal protein produced nothing. No association was found between animal protein intake and mortality, with the estimate sitting close to no difference and the range around it comfortably spanning both harm and benefit.

The gap between those two results is the story. It is also, on this evidence, a gap between one thing that showed an association and one thing that showed none, which is weaker than a demonstration that one is better than the other.

The plant protein result that does not fit

A benefit in total deaths should show up somewhere in the causes. Here it did not. There was no association between plant protein intake and cardiovascular or cancer mortality, and those two categories account for most deaths in a European population.

Small numbers are the likely explanation. Split 460 deaths across causes and the subgroups become too thin to detect a moderate effect, so the absence is uninformative rather than contradictory.

It is still a loose thread, and worth naming rather than leaving out. A protective association that appears only in the total, and in none of its parts, is the pattern you would also expect from a chance finding.

Who ate the plant protein in this cohort

The cohort, known as SUN, after the Seguimiento Universidad de Navarra study, follows 23,321 Spanish university graduates; a total of 17,989 subjects were included in the analysis. Around half of the participants (55.5%) were health professionals, and the global retention rate is 90.6%.

Retention that high is rare and valuable, because people lost from a cohort are usually the ones doing worst. The education level is the trade-off. The cohort’s own investigators state it plainly: the sample is not representative of the general Spanish population, which could limit the external validity of the results, noting the relatively low BMI and high educational level of the people in it.

There is a further wrinkle specific to this question. In a Mediterranean population, the lowest quarter of plant protein intake is still eating more legumes, nuts and olive oil than a low-plant-protein group would in Britain or the United States. That compresses the contrast, and if anything it argues the effect could be larger elsewhere.

What a food questionnaire cannot show about protein

This is an observational cohort, so it can report that a way of eating traveled with living longer. It cannot show that the protein did it.

People who take much of their protein from beans, nuts and whole grains differ from people who take it from meat in ways no statistical adjustment recovers: they smoke less, move more, and are generally more attentive to health. The authors adjusted for what they could measure, which is never the whole of it.

The interval around the 35% figure is also wide, running from a large benefit down to a slight one. Direction is better established than magnitude.

What to do about your own protein

The practical read is undramatic, and it agrees with almost every other line of evidence: shifting part of your protein toward legumes, nuts, seeds and whole grains is a reasonable move, and this study is one more reason to make it.

What it does not support is cutting protein. Total intake matters for muscle, particularly past 60, and nothing here suggests eating less of it.

The useful part of this paper is who was in it. A benefit visible in people whose average age was 38 says the choice is worth making long before the outcome is on the horizon.