Evaluating the confounding effects of medical therapies on potassium intake assessment in patients with heart failure.

BACKGROUND AND AIMS: Potassium-wasting (loop diuretics [LD]) and potassium-sparing (spironolactone) medications used for heart failure (HF) may alter renal potassium handling and confound the use of twenty-four-hour (24-h) urine collections as a surrogate marker for potassium intake, an effect that has been observed with dietary sodium assessment. The objective was to determine the strength of association between 24-h urine collections and weighed food records in assessing potassium intake in HF patients stratified by LD usage and spironolactone usage. METHODS AND RESULTS: Stable outpatients with HF simultaneously completed two 24-h urine collections and two weighed food records on consecutive days. Analyses compared patients stratified by LD and/or spironolactone use. Pearson's correlation and the Bland-Altman method of agreement assessed the relationship between the techniques. Overall, 109 patients (61 ± 11 yrs, 74% male) were included. The mean difference in dietary potassium estimated between 24-h urine collections and food records was -353 ± 1043 mg (p < 0.01) for all patients, with no differences between measures among subgroups. The association between the two methods was r = 0.551 (95% CI, 0.373 to 0.852, p < 0.001) for LD users; r = 0.287 (95% CI, 0.01 to 0.570, p = 0.050) for LD non-users; r = 0.321 (95% CI, 0.13 to 0.798, p = 0.043) for spironolactone users, and; r = 0.534 (95% CI, 0.331 to 0.747, p < 0.001) for spironolactone non-users. There were no significant mean biases identified as part of the Bland-Altman analysis. CONCLUSION: Among HF patients, potassium-wasting and potassium-sparing medications do not influence the agreement between the two methods in the assessment of potassium intake.