Diagnostic value of potassium level in a spot urine sample as an index of 24-hour urinary potassium excretion in unselected patients hospitalized in a hypertension unit

BACKGROUND: Primary hyperaldosteronism may be associated with elevated 24-hour urinary potassium excretion. We evaluated the diagnostic value of spot urine (SU) potassium as an index of 24-hour urinary potassium excretion. METHODS: We measured SU and 24-hour urinary collection potassium and creatinine in 382 patients. Correlations between SU and 24-hour collections were assessed for potassium levels and potassium/creatinine ratios. We used the PAHO formula to estimate 24-hour urinary potassium excretion based on SU potassium level. The agreement between estimated and measured 24-hour urinary potassium excretion was evaluated using the Bland-Altman method. To evaluate diagnostic performance of SU potassium, we calculated areas under the curve (AUC) for SU potassium/creatinine ratio and 24-hour urinary potassium excretion estimated using the PAHO formula. RESULTS: Strongest correlation between SU and 24-hour collection was found for potassium/creatinine ratio (r = 0.69, P<0.001). The PAHO formula underestimated 24-hour urinary potassium excretion by mean 8.3±18 mmol/d (95% limits of agreement -28 to +44 mmol/d). Diagnostic performance of SU potassium/creatinine ratio was borderline good only if 24-hour urinary potassium excretion was largely elevated (AUC 0.802 for 120 mmol K(+)/24 h) but poor with lower values (AUC 0.696 for 100 mmol K(+)/24 h, 0.636 for 80 mmol K(+)/24 h, 0.675 for 40 mmol K(+)/24 h). Diagnostic performance of 24-hour urinary potassium excretion estimated by the PAHO formula was excellent with values above 120 mmol/d and good with lower values (AUC 0.941 for 120 mmol K(+)/24 h, 0.819 for 100 mmol K(+)/24 h, 0.823 for 80 mmol K(+)/24 h, 0.836 for 40 mmol K(+)/24 h). CONCLUSIONS: Spot urine potassium/creatinine ratio might be a marker of increased 24-hour urinary potassium excretion and a potentially useful screening test when reliable 24-hour urine collection is not available. The PAHO formula allowed estimation of the 24-hour urinary potassium excretion based on SU measurements with reasonable clinical accuracy.