Reproducibility and validity of a quantitative FFQ designed for patients with type 2 diabetes mellitus from southern Brazil

To evaluate the reproducibility and validity of a previously constructed FFQ to assess the usual diet of patients with type 2 diabetes mellitus (T2DM). Cross-sectional survey using two quantitative FFQ (1-month interval) supported by a food photograph portfolio, a 3 d weighed diet record (WDR) and urinary N output measurement (as a biomarker of protein intake). Group of Nutrition in Endocrinology, southern Brazil. Out-patients with T2DM. From a total of 104 eligible T2DM patients, eighty-eight were included in the evaluation of FFQ reproducibility and seventy-two provided data for the validity study. The intakes estimated from the two FFQ did not differ (P > 0·05) and the correlation coefficients were significant (P < 0·01) for energy and nutrients, ranging from 0·451 (soluble fibre) to 0·936 (PUFA). Regarding the validity evaluation, data from the FFQ were higher than those from the WDR for total (28·3%), soluble (27·4%) and insoluble fibres (29·1%), and SFA (13·5%), MUFA (11·1 %) and total lipids (9·2%; all P < 0·05). There were significant correlation coefficients between the FFQ and WDR for most nutrients, when adjusted for energy intake and de-attenuated. Also, the Bland-Altman plots between the FFQ and WDR for energy and macronutrient intakes showed that the FFQ may be used as alternative method to the WDR. The validity coefficient (using the method of triads) for the FFQ protein intake was 0·522 (95% CI 0·414, 0·597). This quantitative FFQ was valid and precise to assess the usual diet of patients with T2DM, according to its validity and reproducibility.