REF-1125 · Baer, Gebauer & Novotny, 2016 · The Journal of Nutrition, · Paper · PubMed

Walnuts Consumed by Healthy Adults Provide Less Available Energy than Predicted by the Atwater Factors

B+moderate evidenceMeasured on non-athletes and recreational people

Baer et al. measured how much energy walnuts actually deliver when eaten as part of a typical American diet. Eighteen healthy adults ate the same controlled base diet for two 3-week periods in random order, once unsupplemented and once with 42 g/d of walnuts (the base foods were reduced to keep intake isocaloric).

After 9 days of adaptation all urine and faeces were collected for about a week, and diets, walnuts and excreta went through bomb calorimetry. One 28-g serving of walnuts provided 146 kcal (5.22 kcal/g), 39 kcal less than the 185 kcal (6.61 kcal/g) calculated from the Atwater factors: the metabolizable energy was 21% less than predicted (P < 0.0001).

The same group had found the same direction for almonds (REF-1124).

Protocol

Randomised crossover, fully controlled feeding; two treatment periods of 3 weeks each (from the abstract).
Participants: healthy adults, n = 18, mean age 53.1 y, BMI 28.8.
Treatments: the same base diet unsupplemented, or supplemented with 42 g walnuts/d; base foods reduced in equal proportions in the walnut period to keep food intake isocaloric.
After a 9-day acclimation, all urine and faeces were collected for about 1 week (Brilliant Blue faecal marker). Diets, walnuts, faeces and urine analysed by bomb calorimetry; walnut metabolizable energy (ME) calculated from these data.

Key Results

OutcomeResult (abstract)
Measured ME of walnuts146 kcal per 28-g serving (5.22 kcal/g)
Atwater prediction185 kcal per serving (6.61 kcal/g)
Difference39 kcal per serving less than calculated; ME 21% less than predicted by the Atwater factors (P < 0.0001)

Mechanisms

Measured: more of the energy eaten as walnuts appeared in faeces and urine than the Atwater factors assume; the abstract does not split the loss into fat, protein or fibre.
Inference, not tested here: the same mechanism as for almonds (REF-1124, where faecal fat rose with dose), with fat held inside intact nut cell walls.
Authors' interpretation: consistent with other tree nuts, and could help explain why nut eaters do not gain excess weight; weight was not an outcome of this study.

Practical Application

Walnuts give the body about a fifth less energy than their label says: 146 vs 185 kcal per 28-g serving, 21% less than the Atwater prediction, in 18 adults on a controlled diet (B+, from the abstract). With almonds (REF-1124, 32% Atwater overestimate), whole tree nuts are the clearest case where a label calorie and an absorbed calorie part company, always with the label higher.
In practice the gap is tens of kilocalories a day: 39 kcal per serving, and the study dose was 42 g/d (1.5 servings). This matters for the honesty of a food log, not for the training plan.
Do not take the 21% beyond whole walnuts: walnut oil, butter or finely ground forms were not tested.
Use it as a reason to read the body-mass trend, not to correct labels by hand (SYNTH-87).

Limitations

n = 18 healthy, mostly overweight, middle-aged adults; no athletes.
One food, one dose (42 g/d), whole walnuts only.
Entry written from the abstract: the macronutrient breakdown and the spread between individuals are not in it; funding is not in the abstract, but PubMed lists non-government support alongside the USDA. Treat it with the same caution as the Almond Board-funded REF-1124.
Weight and body composition were not outcomes.
How this entry was checked and changed
Why this level

controlled crossover, n = 18 (rubric: crossover ≥15 = B+); from the abstract (the open publisher copy could not be retrieved)