Dietary counselling alone improved nutrition and cognition, and the exercise arm added nothing to it

Design

Three-arm randomised controlled trial with intention-to-treat linear regression analysis (PROMED-EX, NCT05166564)

Population

105 community-dwelling older adults at risk of undernutrition and cognitive decline; mean age 67.7, 69% women, baseline Mini Nutritional Assessment 22.5

Primary outcome

Change in nutritional status at 6 months on the Mini Nutritional Assessment (0-30)

Effect

Counselling alone +2.7 (95% CI 1.3-4.2) and counselling plus exercise +2.9 (1.5-4.3) vs control, both p<0.001; neurocognitive z-score +0.3 (0.1-0.5) and +0.2 (0.0-0.4)

Undernutrition in older adults is associated with cognitive decline, and the interventions offered for it are usually supplement drinks. PROMED-EX tested something different: personalised counselling towards a protein-enriched Mediterranean diet, with or without home-based exercise, against a healthy eating leaflet, in 105 community-dwelling adults at risk of both undernutrition and cognitive decline. Mean age 67.7, 69% women, baseline Mini Nutritional Assessment 22.5.

At six months both intervention arms beat the leaflet on the primary outcome. Mini Nutritional Assessment scores improved by 2.7 points (95% CI 1.3-4.2) with counselling alone and 2.9 (1.5-4.3) with counselling plus exercise, both p<0.001. Diet quality scores rose by about 4 points in both arms. Neurocognitive test battery z-scores improved by 0.3 (0.1-0.5) with counselling and 0.2 (0.0-0.4) with counselling plus exercise. Adherence to the exercise component was low, and despite that there were additional benefits for physical performance and quality of life.

The finding that should change something is the comparison between the two intervention arms. Adding a home exercise programme did not improve the nutritional or cognitive outcome over counselling alone - and the exercise was poorly adhered to, so this is as much a statement about deliverability as about biology.

So where time and money are limited, spend them on personalised dietary counselling rather than on adding an unsupervised home exercise component to it. A leaflet is not counselling; the control arm got the leaflet and did not improve. The Mediterranean pattern needs local translation - pulses, curd, groundnut, seasonal vegetables and adequate protein at each meal will do the same work in an Indian diet, and the trial's principle is personalisation, not the food list. Note this is 105 people over six months, with cognition as a secondary outcome, so treat the cognitive signal as supportive rather than proven.

• Refer for personalised dietary counselling rather than handing out a leaflet or a supplement drink
• Do not assume adding unsupervised home exercise improves the nutritional result - it did not here
• Target protein at every meal, not just total daily intake
• Translate the Mediterranean pattern into what the patient actually eats; the principle is personalisation
• Reassess with the same nutritional score at six months to see whether it worked

The statistics, in plain English

A 2.7-point gain on a 30-point scale from a baseline of 22.5 moves a typical participant meaningfully within the 'at risk' band and, for some, out of it - the confidence intervals (1.3 to 4.2) exclude zero comfortably. The cognitive z-score differences of 0.3 and 0.2 are small, and the 0.2 interval reaches 0.0, so the cognitive claim is the weaker of the two. Cognition was a secondary outcome, which means the trial was not sized to detect it and the result should prompt a larger trial rather than settle the question. With 105 participants in three arms, the comparison between the two intervention arms rests on about 35 people each and cannot exclude a modest exercise benefit.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.