500 DIABETES CARE, VOLUME 23, NUMBER 4, APRIL 2000 D iabetic ketoacidosis (DKA) is a serious complication of diabetes that is associ- ated with considerable mortality and m o rb id ity. Although the mortality rate diminished from 44% in the 1930s (1) to 16% in the 1970s (2) and to 3–5% in the 1980s (3,4), little improvement in these rates has been evident in recent years, and the incidence of DKA remains 20% in patients 65 years of age (5). Delays in the diagnosis and treatment of DKA are associ- ated with an increase in morbidity and mor- tality (6). The diagnosis of DKA is not always ea sy. Between 1 and 8% of patients are ketoacidotic without significant hypergly- cemia (6,7). Ketostix (Bayer Diagnostics, Stoke Poges, Slough, U.K.) measure urinary acetoacetate but not - hyd ro xyb ut yra t e ( -HBA), which is the predominant ketone body in DKA, and Ketostix may give false- negative results even if blood is tested. Lab- o ra to ry measurement of -HBA is not routinely available and takes too long to be of practical use in the emergency diagnosis and management of DKA. A combined glu- cose and ketone sensor that produces an electrical current pro p o rtional to blood -HBA concentration has been developed (MediSense/Abbott Laboratories, Abingdon, U.K.) (Fig. 1). By using a 10-μl capillary blood sample, this hand-held sensor pro- duces -HBA results within 30 s. The aim of this study was to compare the accuracy and precision of the ketone sensor against an established laboratory enzymatic ref erence method (Pro c ed ure No. 310-UV 1994; Sigma Diagnostics, St. Louis, MO). RESEARCH DESIGN AND M E TH O D S Two groups of patients who were suspected of having abnorma li- ties of ketone metabolism were studied: patients with suspected DKA (group D) and patients treated with a very-lo w-ca lo r ie diet who attend a weight management clinic (group W). All subjects gave their info rmed consent, and the local ethics committee approved the study. For group D, 500-μl capillary blood or 2-ml venous whole blood samples were col- lected during routine management. For group W, a single 500-μl capillary blood sample was obtained during the clinic visit. Sensor measurements were made in dupli- cate within 20 min for capillary samples and at the earliest opportunity (always within 12 h) for venous samples. Veno us samples were refrigerated at 4°C before test- ing. One res ea rcher (K.L.T.) perfo rmed all mea s urements. An electrochemical strip was ins erted into the sensor to which 10 μl of whole blood were applied (Fig. 1). The -HBA, in the presence of hydro xybut yr a t e dehydrogenase, was oxidized to acetoacetate with the concomitant reduction of NAD t o NADH. The NADH was reoxidized to NAD by a redox mediator, and the current generated was directly pro po rtional to the -HBA concentration. After 30 s, the -HB A concentrations (mmol/l) were displayed. Hemoglobin was measured using a HemoCue AB analyzer (Angelholm, Swe- den); room temperature and humidity were rec o rded using a Vaisala meter (Va isa la , S uffolk, U.K.). After initial testing with the ketone sensor, the samples were cen- trifuged, and the plasma was stored at 20°C pending analysis with an enzymatic la b o ra to ry method (Cobas Fara; Roche Diagnostics, Welwyn Garden City, U.K.). To evaluate the precision of the ketone s ens o r, three venous whole blood samples with -HBA concentrations of 1, 1.1–3.0, and 3.1–6.0 mmol/l as determined by the la bo r a t o ry method were tested 20 times each within a 30-min period. Accuracy was assessed by comparing the sensor res ult with the reference method. F rom the Departments of Diabetes and Endocrinology (H.A.B., T.L.D., J.P.N.), and Clinical Biochemistry ( K .L .T.), Hope Hospital, Salford; and the Medical Statistics Unit (S.H.), Department of Mathematics and Sta- tistics, Lancaster University, Lancaster, U.K. Ad d ress correspondence and reprint requests to J.P . New, Department of Diabetes and Endocrinology, Hope Hospital, Stott Ln., Salford M6 8HD, U.K. Received for publication 28 June 1999 and accepted in revised form 7 December 1999. H.A.B. and K.L.T. are employed through an educational grant from MediSense; H.A.B., K.L.T., T.L.D., and J .P .N. have received resea rch grant funding from MediSense; and T.L.D. and J.P .N. have received consulting fees from MediSense, which manufactures the sensor discussed in this article. Abbreviatio ns : -HBA, - hyd roxybutyrate; CV, coefficient of variation; DKA, diabetic ketoacidosis. A table elsewhere in this issue shows conventional and Système International (SI) units and conversion factors for many substances. Evaluation of an Electrochemical Sensor for M easuring Blood Ketones O R I G I N A L A R T I C L E O BJE C TI V E To evaluate the perf ormance of a hand-held ketone sensor that is able to measu re blood -h ydroxybutyrate ( -HBA) concentrations within 30 s in patients with diabetic ketoacidosis (DKA) and patients who attend a weight management clinic. RESEARCH DESIGN AND M ETHODS Two groups of patients were studied: 19 patients admitted with DKA and 156 patients attending a weight management clinic. Paired capillary and venous whole blood samples were measured using the ketone sensor and also using an enzymatic laboratory ref erence method. RE SU LTS The ketone sensor accurately measured -HBA concentrations in patients with DKA (limits of agreement 0.9 to 1.0 mmol/l) or starvation-induced ketonemia (limits of agreement 0.5 to 0.5 mmol/l). C O N C L U SI O N S This ketone sensor accurately measures whole blood -HBA concen- trations within 30 s. Diabetes Care 23:500–503, 2000 HILARY A. BYRNE, MRCP KENNETH L. TIESZEN, PHD SALLY HOLLIS, MSC TIM L. DORNAN, FRCP J OHN P. N EW , MRCP Emerging Treatm ents a nd T e c h n o l o g i e s