Nuts to you (... and you, and you)

A generation ago the prudent diet for preventing coronary disease was dominated by negative advice. Poverty and food rationing in the 1930-50s had led to the promised land of the 1960s, but it was a monotonous greasy landscape of cakes, pies, chips, sausages, and fry ups, dominated by dairy and processed foods. The “you never had it so good” life was bad for the heart. Prophets of doom emerged from communion with molecules, denouncing cholesterol and issuing dietary commandments, almost all phrased “Thou shalt not eat x.” They were less explicit on what should be eaten. It seemed to be what remained after eliminating the favourites or cutting off the fatty bits. Foods were judged in one dimension—what they did to blood cholesterol.¹ Meanwhile a vegetarian subculture, regarded as cranky and unscientific (as many of its adherents were), was promoting fresh and natural foods of vegetable origin.

The cholesterol monolith was evidence based, but confrontation with sceptics, and powerful unscrupulous elements of the food industry, led to a siege mentality among some adherents, reluctant to concede that the diet-heart story might be multidimensional. A key element in that concession, still not categorically established, was the antioxidant-vitamin theory. Fruit and vegetables had a role separate from providing low fat calories. Antioxidant-vitamins protected low density lipoprotein (LDL) cholesterol from oxidation to its atherogenic product.² Further dietary dimensions have been provided through haemostatic factors,³ non-vitamin antioxidants such as flavonoids,⁴ folic acid in green vegetables reducing blood levels of homocysteine,⁵ plant sterols,⁶ and alcohol.⁷ Popular health mythology has embraced red wine, garlic, and green tea. And now we have nuts.

Like food grains, nuts are concentrated sources of food energy and micronutrients, built to carry life forward into a new generation, but their dietary image has been mixed. Commonly eaten salted, as self indulgent snacks, and containing significant quantities of fat, they figure in the “cut out...” diet lists for obesity and hypertension. Yet the nut roast is the archetypal vegetarian meal. In the naive stereotyping of foods, which belies the sophisticated doctrine that it is the variety and balance that determine a good diet, nuts could be classified as both healthy and unhealthy. This ambiguity could continue had we not the experience of 84 409 American nurses followed for 14 years in this week’s BMJ (p 1341).⁸

Earlier work on Seventh Day Adventists suggested that nut eating was associated with diminished coronary risk.⁹ Feeding experiments tested the effect of dietary substitution of almonds and walnuts on serum lipid concentrations and blood pressure—still an essential initiation for foods wishing to be considered beneficial. What the nurses’ health study has done is to relate nut consumption to subsequent coronary risk in a very large study.⁸

One nurse in 20 ate five or more helpings of nuts a week whereas a third hardly ever ate them. Fourteen year coronary risk in the first group was half of that in the latter after adjusting for age and remained significantly reduced by one third even after correcting for all available risk factors. Intermediate nut consumption groups had intermediate risk. Reduced risk from nut eating was maintained across subgroups stratified to test for major confounding factors. The effect therefore appears large and significant, as great or greater than that in the statin trials,¹⁰ and unexplained by anything else.

Is the effect causal, and if so, to what constituent of the nuts can it be attributed? Unfortunately an observational study lacks the power of a randomised controlled trial to determine causality, as there is always the worry of residual confounding. Frequent nut eaters differed in many respects from the others. Of the four demarcated consumption groups, the high consumption group came out most favourably for 16 of 23 factors, most of which would be independent rather than a consequence of nut eating. These benefits have been corrected for, but possibly not enough. The nurses’ health study is a huge (and extremely productive) postal study whose participants have never been subjected to the clinical routines of other American cohort studies. Self reported hypertension was marginally least common in the high consumption group, whereas hypercholesterolaemia was marginally most common. These could not explain a halving of risk. The authors list eight constituents of nuts that might contribute to nutritional benefit but cannot be more specific, a common problem after identifying potentially beneficial dietary habits.¹¹

Further corroboration and investigation are obviously needed, but nuts may meanwhile be rehabilitated as a likely constituent of a prudent diet. Red wine drinkers can claim, with apparent authority, that they are taking their daily dose of flavonoids. Those eating nuts (perhaps in combination with red wine) will currently have greater difficulty in identifying specifically what they are doing, other than eating nuts. They should thank the American nurses none the less.

Papers p 1341