Nutrition Programs in Older Heart Failure

Evidence

Two intervention studies examined structured nutrition programs for older adults with chronic heart failure, but they asked different questions. A July randomized trial tested a six-month program against death or heart-failure hospitalization. A May quasi-experimental study tested whether a two-month, dietary-diversity program could change food-group variety and nutritional assessment scores. Together, the studies separate changes in measured nutritional status from changes in major clinical events.

The BOCADOS-IC trial was a pragmatic, multicenter, prospective study at 14 Spanish hospitals. It enrolled outpatients older than 65 with chronic heart failure and a Mini Nutritional Assessment-Short Form score of 11 or less, indicating malnutrition or risk of malnutrition. Participants were randomly assigned to a structured strategy combining tailored dietary optimization and physical-exercise recommendations or to usual care. Endpoint assessment was blinded, although participants and care teams were not.

Recruitment ended early with 204 participants: 106 in the intervention group and 98 in usual care. At six months, the primary composite of all-cause death or heart-failure hospitalization occurred in 17.9% of the intervention group and 19.4% of controls. The hazard ratio was 0.90, with a 95% confidence interval from 0.48 to 1.70. That wide interval included no difference, and the result was not statistically significant.

Several secondary measures moved more clearly. Mini Nutritional Assessment-Short Form scores improved by 4.99 points with the structured program and 3.65 points with usual care. Mid-arm muscle circumference increased by 0.86 centimeters in the intervention group but fell by 1.38 centimeters in controls. Scores on the Minnesota Living with Heart Failure Questionnaire declined by 7.95 points versus 2.44 points; on that measure, a lower score represented better reported quality of life. These findings show differences in nutritional, body-size, and reported quality-of-life measures, not proof of fewer major events.

The second study took place in one tertiary hospital in Zunyi City, China. Researchers enrolled 88 older chronic-heart-failure patients who could eat orally. Forty-four patients admitted from July through August 2022 formed a nonconcurrent control group receiving conventional dietary guidance; 44 admitted from September through October received a program tailored to the nine-group Dietary Diversity Score and nutritional status, followed for two months after discharge. The design was quasi-experimental rather than randomized.

Eighty participants completed the full follow-up and were included in the per-protocol analysis, 40 per group. At two months, mean dietary-diversity scores were 7.95 in the intervention group and 6.40 in controls, a large standardized difference. Heart-failure-specific Mini Nutritional Assessment scores were 24.51 versus 23.00, a medium difference at that time point. However, the overall group-by-time interaction for that nutritional score was not significant, and BMI trajectories were also not significantly different. The study measured dietary and nutritional indicators; it was not designed to determine effects on hospitalization or survival.

Analysis — Measures improved before clinical outcomes

The cross-study pattern is a hierarchy of endpoints. Both programs changed measures closer to the intervention: dietary diversity in the Chinese study, and nutritional assessment, arm circumference, and reported quality of life in BOCADOS-IC. The randomized trial then tested a harder endpoint and did not detect fewer deaths or heart-failure hospitalizations over six months. Analysis: this convergence supports the narrower conclusion that structured programs can alter measurable nutritional status, while leaving uncertain whether those changes are large, durable, or causally connected to major clinical outcomes. The contrast also shows why a better score cannot be treated as a substitute for an event endpoint. An emerging, unproven direction is to test a prespecified chain from program adherence to dietary intake, body composition or muscle measures, functional capacity, and finally hospitalization or survival. That chain would need longer follow-up and enough participants to distinguish a small clinical effect from no effect. The studies do not establish one preferred nutrition program, and they do not support individualized treatment conclusions outside the populations studied.

Limitations

BOCADOS-IC provides the stronger design because it randomized participants across 14 hospitals and used blinded endpoint assessment. Even so, recruitment stopped early at 204 participants, reducing power for the primary composite outcome. Its 0.48-to-1.70 confidence interval is compatible with meaningfully lower or higher event risk, so the nonsignificant result is not evidence of equivalence. The ingested record is an abstract, which constrains assessment of adherence, missing data, component interventions, and subgroup methods.

The Chinese study was smaller, single-center, nonrandomized, and used sequential admission periods, leaving room for time-related or selection differences. Its analysis excluded people who did not complete follow-up, including deaths and losses, which can bias estimated effects. Food variety was recorded with a qualitative questionnaire that counted food groups but not quantities or nutrient intake. One baseline food-group measure differed between groups. Follow-up lasted only two months, BMI trajectories did not separate, and the overall nutritional-score interaction was not significant despite a difference at the final time point. The studies also used different nutritional scales, program components, countries, and follow-up periods, preventing a pooled effect estimate or broad generalization.