The Original Water Deprivation Test from 1963

In the 1940s, in order to distinguish between primary polydipsia and diabetes insipidus a number iof dehydration tests were proposed  However, it was only in 1963 that Dashe et al published in JAMA a standardized water deprivation test, solving the problem of differential diagnosis of polyuric syndromes.

How to Perform a Water Deprivation Test according to the original Dashe's protocol: 

1.      On the morning of the test day, the patient is allowed to drink fluids ad lib until 8:30 am, and to eat breakfast prior to that time, but coffee, tea, and smoking are interdicted.

2.      For 7 hours after 8:30 am, the subjects are allowed no food or fluids.

3.      Serum and urine specimens are obtained for determination of osmolality

4.       Body weight is measured at the beginning and end of the experiment

5.      The subject voids urine at 8:30 am, and urine is subsequently collected by voiding at 9:30 and 11:30 am, and 12:30, 2:30 and 3:30 pm.

6.      Serum osmolality and serum sodium levels are to be measured at 9:00 am, 12 noon, and 3:00 pm

Dashe et al. compared changes in in the ratio Uosm-to-Sosm in several groups:

1)      Group A: Normal subjects. There weren’t significant changes in serum osmolality (285->286 mOsm/Kg). All normal subjects were able to excrete a highly concentrated urine (range of final-hour urine concentration, 756 to 1,496 mOsm/kg). The ratio of urine to serum osmolality ( urine-to-serum ) for the last hour of the test was 3.8 ± 0.9, with a range from 1.9 to 5.2

2)      Group B: patients with known diabetes insipidus. Six of the 13 patients had initial serum osmolalities within or below the normal range; however, all subjects developed a serum osmolality of 300 mOsm/kg or greater by the end of the test. The final osmolar concentration of the serum exceeded the initial level by an average of 12.0 mOsm/kg; the greatest change was +47 and the least was + 1.  Not surprisingly, patients with diabetes insipidus excreted urine which was less concentrated. The mean urine-to-serum ratio was  0.93 ± 0.45.

3)      Group C: Primary polydipsia. There was no change in serum osmolality during the test. The urine-to-serum ratio was comparable to group 1 ( > 2).

Interestingly, whereas the test is currently stopped when body weight decreases by 3% from baseline, in the original study by Dashe et al. an average of 2.6% of body weight was lost by the patients with severe diabetes insipidus. Indeed, the authors used a fixed duration of 6.5 h

These are Dashe’s conclusions, reporting clinical results for each diagnosis

1)      Mildest degree of diabetes insipidus currently identifiable:  Stability of serum osmolality with subnormal urine-to-serum osmolal ratio

2)      Moderate diabetes insipidus: supernormal serum concentration or supernormal increment in serum osmolar concentration. The serum concentration, however, is not greater than 300 mOsm/kg at any time.

3)      Severe diabetes insipidus: a serum osmolality which is greater than 300 at the end of the experiment

4)      A urine-to-serum ratio above 1.0 excludes the diagnosis of diabetes insipidus. Although some patients with rather severe diabetes insipidus may achieve a urine-to-serum ratio greater than 1.0, they do so only at a time when serum osmolalities were greater than 300 mOsm/kg.

Values before and after a Water Deprivation Test according to Dashe et al.

Diagnosis	Initial Sosm	Post-test Uosm-to-Sosm ratio
Normal subjects	Normal	> 1
Diabetes Insipidus	High (or High Normal)	< 1
Primary polydipsia	Low (or Low Normal)	> 1

Reference: Dashe et al: A Water Deprivation Test for the Differential Diagnosis of Polyuria. JAMA Vol. 185, No 9. Aug 31, 1963.