A novel diagnostic score for diagnosing arginine vasopressin deficiency (central diabetes insipidus) or primary polydipsia with basal laboratory parameters and a novel diagnostic score: results from two international multicentre prospective diagnostic studies

Abstract

Background

Distinguishing arginine vasopressin deficiency (central diabetes insipidus) from primary polydipsia is challenging. There is no validated initial laboratory assessment or diagnostic score to rule-in or rule-out arginine vasopressin deficiency during the first consultation. Therefore, this study aimed to evaluate the diagnostic potential of basal laboratory parameters and to develop a practical diagnostic score.

Methods

Data from two international multicentre studies of patients with arginine vasopressin deficiency and primary polydipsia undergoing the hypertonic saline test were used to evaluate the diagnostic potential of basal laboratory tests and to develop a score incorporating laboratory results, symptoms, and medical history. CODDI was a non-randomised, controlled, diagnostic, international, multicentre non-inferiority study in 11 tertiary medical centres in Switzerland, Germany, and Brazil. CARGOx was a randomised, controlled, cross-over, diagnostic, international, multicentre non-inferiority study across seven tertiary medical centres in Switzerland, Germany, the Netherlands, Italy, the UK, and Brazil. Participants were adult patients with polydipsia (>3 L per day) and hypotonic polyuria (>50 mL/kg bodyweight in 24 h and urine osmolality <800 mOsm/kg) and adult patients with a previous diagnosis of arginine vasopressin deficiency. Data were derived from the initial consultation and a basal laboratory test. For each laboratory parameter, the cutoffs resulting in the highest specificity at 100% sensitivity and the highest sensitivity at 100% specificity were identified. For the diagnostic score, the overall best cutoff, high-sensitivity cutoff (≥95% sensitivity), and high-specificity cutoff (≥95% specificity) were identified. Each cutoff was derived from the first study (development), and their performance was determined in the second study (validation). The final score included the sum of: basal plasma sodium multiplied by plasma osmolality, divided by 100; –50 points for plasma copeptin more than 4·9 pmol/L; +30 points for nycturia (≥3 times per night) or +20 points for nycturia (2 times per night); +20 points for sudden polyuria or polydipsia onset; +30 points for drinking more than 1 L per night; +50 points for anterior pituitary dysfunction and +50 points for pituitary surgery history. The diagnostic performance in predicting arginine vasopressin deficiency was examined by the receiver operating characteristic (ROC) area under the curve (AUC) and by sensitivity and specificity. The studies were registered with ClinicalTrials.gov (NCT01940614 and NCT03572166).

Findings

299 patients who underwent the hypertonic saline test from July 1, 2013, to Sept 30, 2022 were included in this analysis. 141 patients were in the development cohort (59 [42%] had arginine vasopressin deficiency; 82 [58%] had primary polydipsia) and 158 patients were in the validation cohort (69 [44%] had arginine vasopressin deficiency; 89 [56%] had primary polydipsia). In the development cohort, the median age of patients with arginine vasopressin deficiency was 45 years (IQR 33–53), with 38 (64%) of 59 being female and 21 (36%) male, compared with a median age of 32 years (IQR 24–44) and 55 (67%) of 82 being female and 27 (33%) male in the group of patients with primary polydipsia. In the validation cohort, patients with arginine vasopressin deficiency had a median age of 42 years (IQR 32–54), with 38 (55%) of 69 being female and 31 (45%) male, compared with a median age of 37 years (IQR 28–50) and 68 (76%) of 89 being female and 21 (24%) male for patients with primary polydipsia. In the validation cohort, basal plasma sodium of more than 145 mmol/L identified arginine vasopressin deficiency with 100% specificity (95% CI 61–100), whereas primary polydipsia was identified by sodium less than 135 mmol/L with 100% specificity (34–100) and by copeptin more than 5·6 pmol/L with 100% specificity (74–100). In the validation cohort, the clinical score had an AUC of 91% (87–96), a cutoff of more than 441 points provided an overall accuracy of 86% (80–91) for diagnosing arginine vasopressin deficiency. In the validation cohort, the high-specificity cutoff of less than 415 points had 93% specificity (87–99) for diagnosing primary polydipsia, and the high-specificity cutoff of more than 461 points had 93% specificity (88–98) for diagnosing arginine vasopressin deficiency. This stepwise approach enabled diagnosis in 223 (75%) of 299 patients.

Interpretation

We introduce a stepwise diagnostic approach, starting with basal laboratory tests and rule-in and rule-out criteria for immediate treatment. For intermediate cases, the novel score aids in identifying arginine vasopressin deficiency or primary polydipsia with high accuracy. This approach could lead to shortening the diagnostic timeline and reducing dependence on stimulation or dynamic tests.

Funding

Swiss National Science Foundation.

Translation

For the German translation of the abstract see Supplementary Materials section.