BACKGROUND: Many older hospitalized patients are at nutritional risk or malnourished and the nutritional condition is often further impaired during hospitalization. When discharged to own home, a "Nutrition Gap" often occurs, causing inadequate dietary intake, and potentially impeded recovery. Previously, cross-sectorial studies of single component nutritional intervention have shown a limited effect on clinically relevant outcomes. We hypothesized that a multimodal nutritional intervention is necessary to elicit a beneficial effect on clinically relevant outcomes. METHODS: A randomized controlled trial was performed for a period of 16 weeks. At discharge, the intervention group (IG) received dietetic counselling including a recommendation of daily training, an individual nutrition plan and a package containing foods and drinks covering dietary requirements for the next 24 h. Further, a goodie-bag containing samples of protein-rich milk-based drinks were provided. Information regarding recommendations of nutritional therapy after discharge was systematically and electronically communicated to the municipality. The dietician performed telephone follow-ups on day 4 and 30 and a home visit at 16 weeks. The control group (CG) received standard treatment. The primary outcome was readmissions within 6 month, secondary outcomes were Length of Stay (LOS), Health Related Quality of Life (EQ-5D-3L), nutritional status, physical function (30s-CST) and mortality. This trial was registered under ClinicalTrials.gov Identifier no. NCT03488329. RESULTS: We included 191 patients (IG: n = 93). No significant difference was seen in readmissions within 6 month (IG: 45% vs. CG: 45%, Risk Ratio (RR): 0.96 0.71-1.31, p = 0.885). At the 16-weeks follow-up more patients in the IG reached at least 75% of energy and protein requirements (82% vs. CG: 61%, p = 0,007). The energy (kcal) and protein intake (g) per kg was significantly higher in the IG (26.4 kcal/kg (±7.4) vs. 22.6 (±7.4), p = 0.0248) (1.1 g/kg (±0.3) vs. 0.9 g/kg (±0.3). Furthermore, significant lower weight loss was seen in IG (0.7 (±4.3) vs. -1.4 (±3.6), p = 0.002). A significant and clinically relevant difference was found in the EQ-5D-3L VAS-score (IG: mean 61.6 ± 16.2 vs. CG: 53.3 ± 19.3, p = 0.011) (Δ14.3 (±15.5) vs. Δ5.6 (±17.2), p = 0.002). A significant difference in mean 30s-CST in IG was also found (7.2 (±4.3) vs. 5.3 (±4.1), p = 0.010). The improvements in physical function were of clinical relevance in both groups, but significantly higher in the IG (Δ4.2 (±4.4) vs. Δ2.2 (±2.5), p = 0.008). In fact, 86% in IG experienced improvements in the 30s-CST compared with 68% in the CG (p = 0.022). LOS was found to be lower at all time points, however not significant (30 days: -3 (-8.5 to 2.5), p = 0.276, 16 weeks: -4 (-10.2 to 2.2, p = 0.204), 6 months: -3 (-9.3 to 3.3, p = 0346)). All-cause mortality was not different between groups, however RR showed a non-significantly 47% reduction at day 30 (0.53 (0.14-2.05, p = 0.499)) and a 17% reduction at 16 weeks (0.83 (0.40-1.73, p = 1.000)) in IG. Per protocol (PP) analysis revealed a non-significant decrease of 32% in readmission at 6 months (RR: 0.68 (0.42-1.08), p = 0.105). CONCLUSION: The present study, using a multimodal nutritional approach, revealed no significant effect on readmissions however a significant positive effect on nutritional status, quality of life and physical function was found. The improvements in quality of life and physical function were of clinical relevance. No significant effect was found on LOS and mortality.