Diagnostic accuracy of fingerstick β-hydroxybutyrate for ketonuria in pregnant women with nausea and vomiting

Abstract

Objectives The aim of this study was to determine whether a rapid, fingerstick β-hydroxybutyrate (β-OHB) test predicts ketonuria in pregnant women with nausea and vomiting. Methods A convenience sample of 82 pregnant women who presented to the Barnes-Jewish Hospital emergency department (ED) with complaints of nausea or vomiting were enrolled into an institutional review board-approved, prospective, observational study. Exclusion criteria were temperature of >38.3°C, altered mental status, prisoners, and >1 L of any intravenous (IV) fluid before screening. Subjects had fingerstick β-OHB tests, with results reported in increments of 0.1 mmol/L. Urine ketone results were made available once reported as part of standard care and varied from 0 (trace) to +4 values. Ketonuria was defined as a urine ketone value of 3+ or 4+ and was analyzed as a dichotomous, categorical variable. A contingency table and receiver operator characteristic (ROC) curve were constructed for comparing the β-OHB values to those urine ketone values for each patient. Results Mean (±SD) β-OHB was 0.43 (±0.52) mmol/L, and median urine was 1+ for the cohort as a whole. The area under the ROC curve equaled 0.95. The sensitivity and specificity for a fingerstick test of >0.45 mmol/L were 89 and 94.5%, respectively. The positive likelihood ratio (LR) was 16, and the negative LR was 0.12. Conclusions Fingerstick β-OHB is a rapid and reliable diagnostic tool that correlates well with ketonuria and identifies ketonemia in pregnant women with nausea and vomiting. Fingerstick β-OHB testing has the potential to increase triage efficiency, shorten length-of-stay times, and positively affect patient outcomes in an ED setting. © 2013 by the Society for Academic Emergency Medicine.