Evaluation of 2 methods for sodium intake assessment in cardiac patients with and without heart failure: the confounding effect of loop diuretics.

Background: Twenty-four-hour urine collections are considered the optimal method for sodium intake assessment. Whether a diagnosis of heart failure (HF) or the use of loop diuretic (LD) therapy for HF compromises the validity of 24-h urine collections as a surrogate marker for sodium intake is unknown.
Objective: The objective was to determine the strength of association between 24-h urine collections and food records for sodium intake assessment in non-HF cardiac patients and in HF patients stratified by LD usage.
Design: Food records and 24-h urine collections were simultaneously completed for 2 consecutive days. Correlation coefficients and the Bland-Altman method of agreement described the relation between the techniques.
Results: Non-HF cardiac patients (n = 96; mean ± SD age: 65 ± 11 y), HF patients who were not taking an LD (n = 47; 62 ± 11 y), and HF patients who were taking an LD (n = 62; age: 60 ± 12 y) were included. Correlation coefficients for sodium intake between food records and urine collections were r = 0.624 (P < 0.001) for non-HF cardiac patients and r = 0.678 (P < 0.001) for HF patients who were not taking an LD. However, no significant association (r = 0.132, P = 0.312) was observed for HF patients who were taking LDs. The 95% limits of agreement between the non-HF cardiac patients and the HF patients who were not taking LDs were similar but were ≈50% wider for HF patients who were taking LDs.
Conclusions: For the assessment of sodium intake, food records agree well with 24-h urine collections in non-HF patients with cardiovascular disease and in HF patients who are not receiving LD but not for HF patients who are taking LDs. Therefore, food records may provide a better estimate of sodium intake in HF patients who are receiving LD therapy.