This paper, which presents a systematic evaluation of popular approaches to the causal diagnosis of hyponatremia, is complex and has many clinical lessons. It should be read in the original if you can access it. Some key points in the paper (and key impressions on my part after reading it):
Inexperienced clinicians guided by an algorithm performed better than seasoned clinicians (“senior attendings”). (There are two ways you might interpret that finding. Some might be tempted to think cookbook medicine trumps clinical judgment. I think it means a systematic approach beats flying by the seat of your pants!).
Clinical volume assessment tends to be unreliable.
The diagnostic approach is confounded in patients with comorbidities and various external factors, particularly the use of diuretics.
Assessment of the fractional excretion of urate may add to the diagnostic power of the algorithm.
SIADH was said to be over diagnosed in this study. It must be acknowledged, however, that while many hospitalized patients had an identifiable non-SIADH etiology for their hyponatremia many hospitalized patients, and most if not all of those with complex illnesses, have elevated vasopressin levels. Those patients may not be labeled as SIADH in the usual sense but the elevated vasopressin level contributes to hyponatremia. SIADH is often said to be a diagnosis of exclusion, nevertheless many patients with non-SIADH etiologies for hyponatremia likely have elevated vasopressin levels as a contributing factor. This is why the construct of SIADH is confusing and needs better definition.
Patients with primary polydipsia may not maximally dilute their urine, and the urine osmolality threshold for excluding primary polydipsia should be raised to greater then 200. This makes sense, because compulsive water drinkers who have normal diluting capacity shouldn't get hyponatremic unless their water intake is enormous. Normal kidneys can dump a lot of free water.
Don't overlook adrenal insufficiency.
(By the way, I think there may be an error in the algorithm in the figure. A urine osmolality of greater than 100 indicates impaired renal dilution, not concentration.
More commentary from DB here.