Urine tests are very helpful for the diagnosis of renal and electrolyte disorders. However, ordering and interpreting urine tests properly have been an Achilles heel for many primary care physicians and residents. It is particularly difficult for chronic hypokalemia because the pathophysiological mechanisms behind it can be rather complicated. Traditionally, we use transtubular K+ gradient (TTKG) or urinary K+/creatinine ratio (UK/Cr) to identify the source of K+ loss,1, 2 thinking that if it is an extrarenal loss, TTKG or UK/Cr would be low, because the K+ secretion from the collecting duct would be compensatorily diminished. On the other hand, if K+ is lost via the kidney, TTKG or UK/Cr would be high, as more K+ is delivered to the collecting duct. In this issue, Wu et al3 reported that TTKG and UK/Cr are not reliable for predicting source of K+ loss, based on a cohort of 99 patients with chronic hypokalemia without hypertension, collected over 5 years at a tertiary hospital in Taipei. They came up with a rather clever approach for identifying sources of K+ loss.
In their cohort, 33 patients had gastrointestinal loss of potassium due to chronic vomiting (anorexia/bulimia) or diarrhea (laxative abuse) and 66 patients with renal potassium loss associated with diuretic abuse, hereditary or acquired Batter syndrome, Gitelman syndrome, or distal renal tubular acidosis.3 It should be mentioned that most hereditary renal tubular disorders were verified with genetic testing in this study, and that it is the largest series of chronic hypokalemia in the era of molecular diagnosis of renal tubular disorders.
In this study, authors confirmed that renal loss is associated with elevated TTKG (>3) and UK/Cr (>2 mmol/mmol),2 except that 33% of patients with surreptitious diuretic use showed low TTKG and UK/Cr. The absence of elevated K+ excretion is associated with remote diuretic use, therefore, diuretic effects were already washed off. On the other hand, for patients with K+ loss due to anorexia/bulimia nervosa, a very low percentage had low TTKG (5%) and UK/Cr (29%). For those with laxative abuse, UK/Cr appears better than TTKG for predicting the gastrointestinal loss, with 91% showing low UK/Cr, and only 36% showing low TTKG.3 Gladziwa et al4 reported similar results with urinary K+ concentration >20 mmol/L in patients with surreptitious vomiting and laxative abuse. The elevated urinary K+ excretion in patients with chronic vomiting is probably due to bicarbonaturia, while in patients with chronic diarrhea it may be due to hypomagnesemia and hyperaldosteronism.
The authors proposed to use urine-sodium-to-chloride ratio (UNa/Cl) to predict the source of K+ loss. The reason is that renal Na+ excretion is always coupled with Cl− excretion in renal tubular disorders as listed above, thus, the UNa/Cl is close to one. It is therefore extremely useful for patients with remote diuretic abuse, because their TTKG and UK/Cr may be low, but their urinary Na+ excretion is still coupled to Cl−. For patients with K+ loss due to vomiting, UNa/Cl is high because of Cl− deficiency and bicarbonaturia; while in K+ loss due to chronic diarrhea, UNa/Cl would be low because of hyperchloremia. Using UNa/Cl >1.6 for anorexia/bulimia nervosa, the sensitivity was 95.2% and specificity 98.7%. Using UNa/Cl <0.7 for laxative abuse, the sensitivity is 86.5% and specificity 100%.3 The UNa/Cl seems to be a reasonable tool for distinguishing extrarenal K loss from renal loss.
However, we should notice that the study cohort does not include patients with proximal renal tubular acidosis. Those patients have a high UNa/Cl because of bicarbonaturia, particularly during the correction of metabolic acidosis.5 The presence of metabolic acidosis can separate those from patients with anorexia/bulimia nervosa who always have metabolic alkalosis. Another cause for uncoupling urinary Na+ to Cl− excretion associated with renal K+ loss is primary hyperaldosteronism, which was excluded from the study because of hypertension. The UNa/Cl ratio is low in patients with hyperaldosteronism because of excessive renal absorption of Na+ and secretion of K+ with relatively unchanged Cl−.6 The low UNa/Cl is similar to patients with laxative abuse who tend to have elevated renin and aldosterone levels due to volume depletion, while patients with primary hyperaldosteronism have hypertension and suppressed plasma renin activity.
In conclusion, UNa/Cl, but not TTKG or UK/Cr, is quite useful for predicting the source of K+ loss in patients with chronic hypokalemia: coupling Na+ to Cl− excretion (UNa/Cl is close to 1) indicates a renal loss, while uncoupling indicates an extrarenal loss with 2 exceptions: proximal renal tubular acidosis and primary hyperaldosteronism. For difficult cases of chronic hypokalemia, ordering urinary Na+ and Cl− levels can tell you how the K+ is lost.
To read this article in its entirety please visit our website.
-Yeong-Hau H. Lien, MD, PhD
This article originally appeared in the July 2017 issue of The American Journal of Medicine.
To read “Identification of the Causes for Chronic Hypokalemia,” another article related to this topic, follow this link.
