Background: Differential diagnosis of diabetes insipidus is challenging. The most reliable approach is hypertonic saline-stimulated copeptin measurements. However, this test is based on the induction of hypernatraemia and requires close monitoring of plasma sodium concentrations. Arginine-stimulated copeptin measurements might provide an alternative, simple, and safe test.
Methods: In this prospective diagnostic study, we recruited a development cohort from University Hospital Basel, Basel, Switzerland, and a validation cohort from five centres in Basel, Aarau, Luzern, Bern, and St Gallen, Switzerland, and the University Hospital Würzburg, Würzburg, Germany. For both cohorts, patients were eligible for inclusion if they were aged 18 years or older, were newly referred with polyuria (>50 mL/kg bodyweight per day) or had a known diagnosis of central diabetes insipidus or primary polydipsia. We also recruited a comparator cohort of healthy controls in parallel to each cohort, comprising adults (aged 18 years and older, with normal drinking habits, and no history of polyuria) and children who underwent arginine stimulation to diagnose growth hormone deficiency (children were only included in the comparator cohort to the development cohort as proof of concept). Patients and healthy controls underwent arginine stimulation with measurement of plasma copeptin at baseline and 30, 45, 60, 90, and 120 min. The primary objective in the development cohort was to determine the diagnostic accuracy of plasma copeptin concentrations to discriminate between diabetes insipidus and primary polydipsia, and in the validation cohort was to confirm those results. Adverse effects of the test were monitored in all participants, with tolerability of the test rated using a visual analogue scale (VAS) that ranged from no (0) to maximum (10) discomfort. This trial is registered with ClinicalTrials.gov, number NCT00757276.
Findings: Between May 24, 2013, and Jan 11, 2017, 52 patients were enrolled in the development cohort (12 [23%] with complete diabetes insipidus, nine [17%] with partial diabetes insipidus, and 31 [60%] with primary polydipsia) alongside 20 healthy adults and 42 child controls. Between Oct 24, 2017, and June 27, 2018, 46 patients were enrolled in the validation cohort (12 [26%] with complete diabetes insipidus, seven [15%] with partial diabetes insipidus, and 27 [59%] with primary polydipsia) alongside 30 healthy adult controls (two patients in this cohort were excluded from the main analysis because of early vomiting during the test). In the pooled patient and control datasets, median arginine-stimulated copeptin concentrations increased in healthy adult controls (from 5·2 pM [IQR 3·3-10·9] to a maximum of 9·8 pM [6·4-19·6]) and in participants with primary polydipsia (from 3·6 pM [IQR 2·4-5·7] to a maximum of 7·9 pM [5·1-11·8]), but only minimally in those with diabetes insipidus (2·1 pM [IQR 1·9-2·7] to a maximum of 2·5 pM [1·9-3·1]). In the development cohort, a cutoff of 3·5 pM at 60 min provided the highest diagnostic accuracy of 94% (95% CI 84-98). The accuracy of this cutoff in the validation cohort was 86% (95% CI 73-94). By pooling the data from both cohorts, an optimal accuracy of 93% (95% CI 86-97) was reached at a cutoff of 3·8 pM copeptin at 60 min (sensitivity 93%, 95% CI 86-98; specificity 92%, 95% CI 84-100). The test was safe and well tolerated, with median VAS scores of 3·5 (IQR 2-4) in patients with diabetes insipidus, 3 (2-4) in those with primary polydipsia, 1 (1-3) in healthy adults, and 1 (0-5) in healthy children in the pooled participant dataset.
Interpretation: Arginine-stimulated copeptin measurements are an innovative test for diabetes insipidus with high diagnostic accuracy, and could be a simplified, novel, and safe diagnostic approach to diabetes insipidus in clinical practice.
Funding: Swiss National Science Foundation and University Hospital Basel.
Copyright © 2019 Elsevier Ltd. All rights reserved.