Urine potassium per hour as a marker for renal potassium losses

Hypokalemia, serum potassium (K) < 3.5 mEq/L, is a serious and common clinical problem. The traditional diagnosis of renal potassium losses is 24-hr urine potassium (24U(K)) > or = 20 mEq/day during hypokalemia. Immediate replacement of potassium is often required to prevent complication but may normalize serum K during 24-hr urine collection and render the test inconclusive.

The authors examined the ability of urinary indices including 24U(K), transtubular potassium gradient (TTKG), fractional excretion of potassium (FE(K)), urine potassium-creatinine ratio (U(K/Cr)) and spot U(K) and introduced urine potassium per hour during the first 8 hours (U(K)/hr) as a novel index for evaluation of hypokalemia during treatment. Any serum K level > or = 4 mEq/L during urine collection was defined as normalized serum K. In the present study, the final classification of renal K losses in non-normalized 24-hr serum K group was made when 24U(K) > or = 20 mEq/day. In normalized group, the final classification of renal or non-renal K losses was based on the majority of the results of four urine indices including TTKG, FE(K) U(K/Cr) and spot U(K).

Of 61 patients (renal:non-renal = 50:11), 51% and 18% met the criteria of normalized 24-hr and 8-hr serum K. Over all, the U(K)/hr > or = 0.9 mEq/hr can indicate renal K losses with a sensitivity of 96% and specificity of 72.7% compared with the 24U(K) > or = 20 mEq/day of 100% and 54.5%, respectively. In a subgroup of normalized 24-hr serum K, the sensitivity and specificity of U(K)/hr = 95.5% and 77.8% whereas 24U(K) = 100% and 44.4%, respectively

U(K)/hr is a new practical, simple, and reliable marker that can be applied to evaluate hypokalemic patients during treatment with comparable sensitivity and specificity with 24U(K).