Accounting for culinary practices to improve FFQ-based vitamin C estimates in epidemiological studies: a cross-sectional analysis

PURPOSE: Conventional vitamin C intake estimations via food frequency questionnaires (FFQ) typically ignore cooking techniques, which can significantly alter nutrient content. This study evaluates the effect of incorporating a home cooking frequency questionnaire (HCFQ) alongside a FFQ to calculate vitamin C intake and its impact on adherence to dietary recommendations. METHODS: We conducted a cross-sectional analysis in the PREDIMAR study, a randomized trial that aims to evaluate the effect of a Mediterranean diet supplemented with extra virgin olive oil on atrial fibrillation recurrence after ablation. Vitamin C intake was estimated in two ways: derived from (1) raw food using a FFQ and the Spanish food composition tables, and (2) from raw and cooked food using a FFQ+HCFQ. Adherence to estimated average requirements and a recommended intake of 200 mg/day were used to assess nutritional adequacy. Paired t-Test compared vitamin C mean intakes derived from both methods, and paired McNemar tests assessed differences in adequacy rates. RESULTS: Among 447 patients (25.1% female; mean age = 59.7 years (10.1)), mean vitamin C intake (SD) was significantly higher when derived solely from the FFQ (173.7 mg/day (71.2)) vs. FFQ+HCFQ (160.1 mg/day (68.8); p < 0.001). The largest decreases of vitamin C, when cooking techniques were considered, were observed in legumes (- 88.0%), tubers (- 62.7%), and vegetables (- 12.5%). The main source of between-person variability in vitamin C intake from vegetables and tubers was raw vegetables and boiled potatoes, respectively. Adequacy of vitamin C intake significantly dropped using FFQ+HCFQ (21.5%) vs. FFQ alone (30.4%, p < 0.001). CONCLUSIONS: Ignoring cooking methods may lead to overestimation by approximately 8% of both vitamin C intake and adequacy prevalence in epidemiologic research. Incorporating a cooking-frequency questionnaire yields more conservative and potentially accurate estimates, improving nutritional epidemiology precision.