Background/aims: Twenty-four hour urinary sodium excretion (24-h UNa) of greater than 78 mmol/day is important in the management of cirrhotic ascites. Although the random urine sodium-to-potassium ratio (UNa/K) is strongly correlated with 24-h UNa, and approximately 95% of patients with a random UNa/K greater than 1 have 24-h UNa greater than 78 mmol, few data have been published on the correlation between 24-h UNa and random UNa/K. We evaluated diagnostic value of morning and afternoon random UNa/K (AM UNa/K and PM UNa/K, respectively) with 24-h UNa.
Methods: A total of 42 male patients were enrolled from October 2007 to March 2008. Each patient collected 5 mL of urine twice at random times during 24-h urine collection (at 10-12 a.m. and 3-5 p.m.). ROC curve analysis was performed to evaluate the feasibility of AM and PM UNa/K for differentiating 24-h UNa greater than 78 mmol/day.
Results: Forty patients with a 24-h urinary creatinine of greater than 15 mg/kg were analyzed. The 24-h UNa, AM UNa/K, and PM UNa/K were 107.9+/-91.2 mmol (mean+/-SD), 3.44+/-3.64, and 3.97+/-4.60, respectively. When compared with 24-h UNa greater than 78 mmol, AUROC values for AM and PM UNa/K were 0.861 (95% CI, 0.715-0.950) and 0.929 (95% CI, 0.802-0.986), respectively (P=0.0001). No difference was found between the AUROC for AM and PM UNa/K (95% CI, -0.161-0.153, P=0.113). UNa/K greater than 1.25 was sensitive and specific for prediction of 24-h UNa greater than 78 mmol.
Conclusions: The results suggest that anytime random UNa/K greater than 1.25 is an accurate, cost-effective, and convenient method for replacing 24-h UNa. Large multicentered cohort studies are needed to confirm our results.