Original ContributionsDiagnostic Accuracy of Tests for Polyuria in Lithium-Treated PatientsKinahan, James Conor MD, MRCPI, MRCPsych*; NiChorcorain, Aoife MB, MRCPsych†; Cunningham, Sean MSc, PhD‡; Freyne, Aideen MD, MRCPsych§; Cooney, Colm MD, FRCPI, MRCPsych§; Barry, Siobhan MD, MRCPsych∥; Kelly, Brendan D. MD, MA, MSc, PhD, DGov, PhD, FRCPI, FRCPsych*Author Information From the *Department of Adult Psychiatry, UCD School of Medicine and Medical Science, University College Dublin, Mater Misericordiae University Hospital, Dublin; †Department of Psychiatry, Cork University Hospital, Cork; Departments of ‡Pathology and Laboratory Medicine, and §Old Age Psychiatry, St Vincent's University Hospital; and ║Cluain Mhuire Community Mental Health Service, Dublin, Ireland. Received August 9, 2014; accepted after revision April 27, 2015. Reprints: James Conor Kinahan, MD, MRCPI, MRCPsych, Department of Adult Psychiatry, UCD School of Medicine and Medical Science, University College Dublin, Mater Misericordiae University Hospital, 63 Eccles St, Dublin 7, Ireland (e-mail: [email protected]). Journal of Clinical Psychopharmacology: August 2015 - Volume 35 - Issue 4 - p 434-441 doi: 10.1097/JCP.0000000000000350 Buy Metrics Abstract In lithium-treated patients, polyuria increases the risk of dehydration and lithium toxicity. If detected early, it is reversible. Despite its prevalence and associated morbidity in clinical practice, it remains underrecognized and therefore undertreated. The 24-hour urine collection is limited by its convenience and practicality. This study explores the diagnostic accuracy of alternative tests such as questionnaires on subjective polyuria, polydipsia, nocturia (dichotomous and ordinal responses), early morning urine sample osmolality (EMUO), and fluid intake record (FIR). This is a cross-sectional study of 179 lithium-treated patients attending a general adult and an old age psychiatry service. Participants completed the tests after completing an accurate 24-hour urine collection. The diagnostic accuracy of the individual tests was explored using the appropriate statistical techniques. Seventy-nine participants completed all of the tests. Polydipsia severity, EMUO, and FIR significantly differentiated the participants with polyuria (area under the receiver operating characteristic curve of 0.646, 0.760, and 0.846, respectively). Of the tests investigated, the FIR made the largest significant change in the probability that a patient experiences polyuria (<2000 mL/24 hours; interval likelihood ratio, 0.18 and >3500 mL/24 hours; interval likelihood ratio, 14). Symptomatic questioning, EMUO, and an FIR could be used in clinical practice to inform the prescriber of the probability that a lithium-treated patient is experiencing polyuria. Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.