(1) Background: Longevity Blue Zones (LBZs) are populations characterized by exceptional longevity. The purpose of this cross-sectional study was to compare the food habits of two representative samples of the oldest old subjects from the population residing in the LBZs of Nicoya peninsula (Costa Rica) and in the mountainous part of Ogliastra (Sardinia, Italy). (2) Methods: Data were collected using validated tools, including a food frequency questionnaire, Basic Activities of Daily Living (BADL) and Instrumental Activities of Daily Living (IADL) scales for functional autonomy, body mass index, and waist and limbs circumferences. (3) Results: A total of 210 subjects, 60 (31 male) from Nicoya (age range 80–109 years), and 150 (61 male) from Ogliastra (age 90–101 years) were included in the study. In both populations, the highest frequencies of consumption were recorded for plant-derived foods (cereals 60–80% daily, legumes ≥ 80% daily in Nicoya, ≥ 60% 2–5 servings/week in Ogliastra), followed by those of animal origin (dairy products, meat) ≥60% and 80% daily, in Nicoya and Ogliastra, respectively. The frequency of milk consumption showed a positive correlation with BADL (ρ = 0.268 for Nicoya and ρ = 0.214 for Ogliastra) and IADL scores (ρ = 0.466 for Nicoya and ρ = 0.471 for Ogliastra), whereas legumes consumption correlated negatively with self-rated health (ρ = −0.264) and IADL (ρ = −0.332). (4) Conclusions: Our results indicate that the dominant dietary model among the elderly of Nicoya and Ogliastra is a plant-based diet complemented by a non-negligible consumption of animal products, mostly dairy products. Further prospective studies are needed to ascertain a possible cause–effect relationship between food habits and increased likelihood of reaching advanced age.

Background: Dietary habits in long-lived populations are often interpreted as determinants of exceptional longevity, although most evidence is based on ecological or cross-sectional studies. In this study we examined whether diet measured after extreme old age was reached was associated with subsequent survival among Sardinian Blue Zone residents aged 90 years or older. Methods: The cohort included 200 participants aged 90–106 years who completed a qualitative food-frequency questionnaire covering 17 foods or food groups and were followed for up to seven years. A posteriori dietary patterns were derived using principal component analysis with Varimax rotation. Associations with all-cause mortality were assessed through Kaplan–Meier analysis and Cox models adjusted for age, sex, and comorbidity. Results: Four patterns were identified: dairy–traditional fat, red-meat, chicken–carbohydrate, and plant–starchy food. Greater conformity to the dairy–traditional-fat pattern was associated with lower mortality (HR per 1-SD increase, 0.839; 95% CI, 0.727–0.969, p = 0.017), although this pattern-level association did not meet the 5% false-discovery-rate (FDR) threshold (q = 0.066). The highest versus lowest tertile showed a similar association (HR, 0.682; 95% CI, 0.480–0.969). No other pattern was significantly associated with mortality. In secondary food-specific analyses, higher frequencies of milk, cheese, and chicken consumption were inversely associated with mortality, whereas wine consumption was positively associated with mortality; these four associations remained significant after FDR correction across the 17 foods examined. Conclusions: In adults aged 90 years or older, greater conformity to a dairy-based late-life dietary pattern was associated with longer subsequent survival, although the pattern-level association did not remain significant after FDR correction. These findings concern diet in extreme old age and should not be extrapolated to lifelong dietary factors influencing exceptional longevity.