CLIN. CHEM. 41/2, 306-311 (1995) #{149} General Clinical Chemistry 306 CLINICAL CHEMISTRY, Vol. 41, No. 2, 1995 Analytical Artifacts in Hematocrit Measurements by Whole-Blood Chemistry Analyzers Richard A. W. Stott,���4 Glen L. Hortin,���2 Timothy R. Wilhite,2 Steven B. Miller,3 Carl H. Smith,���2 and Michael Landt���2���6 Compact analyzers suited to near-patient testing esti- mate hematocrit by measuring the conductivity of undi- luted blood. We evaluated the accuracy of hematocrit determination of one such analyzer (Instrumentation Lab- oratory BGE Analyzer) against an automated cell counter (EPC) and packed cell volume (PCV) microhematocrit. When specimens (n = 34) from outpatient and ward patients were analyzed with all three methods, the BGE analyzer correlated well with both EPC and PCV hemat- ocrit determinations (BGE = 1.00 PCV + 0.3%, S = 1.6% BGE = 1.04 EPC + 0.4%, S, = 2.0%), suggest- ing that all three methods are similar in performance for most patients. However, a patient with increased plasma osmolality showed significant decreases in BGE and PCV hematocrits relative to the EPC method. The differences in hematocrit measurements could be reproduced by adding solutes to blood in vitro or by modifying the plasma osmolality of rats in vivo. Samples from patients undergoing cardiac surgery, whose blood had large changes in protein concentration, showed discrepancies between hematocrits by conductivity and other methods similar effects could be produced by changes in protein concentration or in vitro addition of polyethylene glycol. We conclude that conductivity measurements provide accurate hematocrit results for physiologically normal subjects but not for some intensive-care and surgical patients. Indexing Terms: conductivity/near-patient testing/osmolaiity/ intermethod comparison /analyticai error Several recently developed analyzers- e.g. BGE1ec- trolytes (BGE Instrumentation Laboratory, Lexing- ton, MA),7 Stat Profile series instruments (Nova Bio- medical Corp., Waltham, MA), i-STAT (i-STAT, Princeton, NJ), Gemstat (Mallinckrodt Sensor Sys- tems, Ann Arbor, MI)-provide hematocrit measure- ment in combination with such commonly requested chemistry analyses as Na, K, chloride, ionized calcium, ���Department of Pathology, 2The Edward Mallinckrodt De- partment of Pediatrics, and3 Department of Medicine, Washing- ton University, School of Medicine at St. Louis Children���s Hospi- tal, One Children���s Place, St. Louis, MO 63110. 4Present address: Department of Clinical Chemistry, Don- caster Royal Infirmary, Doncaster, South Yorkshire, UK Present address: Department of Pathology, University of Alabama at Birmingham, 618 S. 18th St., Birmingham, AL 35233. 6Author for correspondence. Fax 314-367-3765. 7Nonstandard abbreviations: BGE, blood gas electrolyte ana- lyzer EPC, electronic particle counter PCV, packed cell volume and PEG, polyethylene glycol. Received August 11, 1994 accepted October 21, 1994. glucose, blood gases, and calculated variables. These instruments require only small volumes of whole blood and are frequently used in near-patient testing situa- tions such as intensive care, operating areas, and emergency rooms (1-4). Placing instruments in these locations ensures immediate availability of results in critical situations, avoids specimen transport prob- lems, and reduces the burden of ���stat��� requests on the main hospital laboratory. The availability of hemato- crit measurement on these instruments avoids the need to send a separate specimen to the hematology laboratory. However, the analytical method used by these instruments differs from those used in conven- tional manual or automated hematocrit measurements (see below) and has been reported to give relatively poor correlation with hematocrit results obtained from major laboratory analyzers and other conductivity- based instruments (1-4). The electrical measurement of hematocrit derives from the effect of particles on the measured resistance of a liquid (strictly, the measured quantity is the impedance, given that ac voltage is used to avoid electrolytic effects at the electrodes). Addition of parti- cles that have higher resistivity than the suspending medium causes an increase in resistance measured between fixed electrodes (decreased conductivity). Maxwell (5) originally studied suspensions of spherical particles and derived an equation relating bulk resis- tivity (resistance per unit length) to the resistivity of the individual components and the volume fraction occupied by the particles. In a suspension of multiple particle types, all particles contribute to the bulk resistivity according to their individual resistivity and volume fraction. The original equation was modified by Fricke (6) to account for nonspherical particles by inclusion of a farm factor. Under these conditions the equation simplifies to: 1 + (f - 1)(v/100) P = Pm 1 - (v/100) where p = bulk resistivity, pm = resistivity of the medium, f = form factor, and v = volume fraction (%). Owing to their nonconductive lipid membrane, blood cells have very high resistivity and therefore provide the majority of the particulate contribution to the resistivity of whole blood. Erythrocytes ac- count for most of the volume fraction, and the con- tribution of other cell types can be ignored in healthy individuals. This principle is applied in electronic particle

CLINICAL CHEMISTRY, Vol. 41, No. 2, 1995 307 counters (EPC) (e.g., the Coulter Counter Coulter Diagnostics, Hialeah, FL). Whole blood is diluted (1: 6250) in isotonic saline and forced through a narrow orifice that reduces the flow of cells to single file past a pair of electrodes. Continuous monitoring of the resis- tance between these electrodes allows detection and measurement of the effective volume of individual cells. Because of the high dilution, the value of Pm is governed entirely by the saline diluent, with no signif- icant contribution from plasma components such as nonionized molecules, electrolytes, and proteins. The direct calculation of hematocrit from measure- ments of the resistivity of undiluted whole blood (basis of analysis in the BGE and similar instruments) is also based on the simplified form of the Maxwell-Fricke equation (7,8). In contrast to EPC, these whole-blood analyzers do not dilute the sample and thus cannot directly control matrix resistivity by dilution in saline instead, these analyzers derive an expected conductiv- ity value by measuring the plasma cations. The mea- sured conductivity is compared with the expected con- ductivity calculated from the measurement of cation concentrations, and the difference is calibrated to spec- imen hematocrit. This calculation assumes that the bulk resistivity contributions from proteins, nonionized molecules, unmeasured cations, and their paired an- ions are either constant between individuals or linearly related to the cation activity measured in each speci- men. The BGE instrument bases the calculation on sodium activity alone, whereas both sodium and potas- sium are used in Nova and i-STAT instruments. Because direct resistivity methods rely on assump- tions concerning the composition of whole blood, we anticipated that specimens from several important categories of patients likely to be served by such analyzers might yield erroneous hematocrit determina- tions. We therefore sought to determine the reliability of hematocrit measurements on one such instrument, the BGE analyzer, in specimens with abnormal amounts of electrolytes or plasma proteins by compar- ison with determinations from electronic cell counters and with packed cell volume (PCV) measurement de- termined by centrifugation. Materials and Methods Study subjects. Patient correlation studies were per- formed with hepariized whole-blood specimens re- maining after completion of routine or emergency blood gas analyses initiated by the clinical staff of the hospi- tal. Some studies utilized fresh blood drawn by stan- dard venipuncture techniques from healthy volunteers. Blood specimens were also drawn from male Sprague- Dawley rats (Harlan Co., Indianapolis, IN) fed ad libitum. All studies were conducted in accordance with protocols approved by the Human Studies Committee (or the Animal Care Committee for the rat studies) of Washington University. Specimen preparation. The effects of plasma solutes on hematocrit were systematically investigated with hepariized blood samples obtained from healthy vol- unteers. Concentrated stock solutions of the solute of interest (NaC1 1.0 moIIL, sorbitol 2.0 moIIL, glucose 1.0 molfL) were prepared in either water or isotonic saline (as appropriate to the experiment). Pooled whole blood was divided into aliquots and small volumes of stock solution were added, with simultaneous addition of diluent to maintain a constant dilution of the blood of 1:20 or less. After the samples were mixed gently, hematocrit was measured by the methods described below, in randomized order. The concentration (Na, glucose) and osmatic effects of the additive were deter- mined with the remaining aliquot of sample. Whole- blood K was also measured as a check for hemolysis. Animal studies. Nonfasted animals were weighed and then surgically prepared under light ether anes- thesia for study in the awake state by a modification of the technique described by Cotlove (9). One catheter (PE 50) was placed in the right femoral artery for the collection of blood and another in the left femoral vein for infusion. Pairs of rats were placed in Plexiglas holders and allowed to recover from the effects of anesthesia. A baseline blood sample (0.6 mL) was obtained in a heparinized arterial blood sampling sy- ringe, and infusion was begun at a constant rate of 1.3 mL/h per 100 g of body weight (syringe infusion pump 22 Harvard Apparatus, South Natick, MA). Experi- mental rats received sorbitol (3.0 mol/L) in isotonic saline control rats received saline alone. Blood sam- ples were collected hourly and analyzed without delay for hematocrit and plasma osmolality (centrifugation at ilOOg for 5 mm yielded suitable plasma). Each animal was weighed again at the end of the experi- ment. Analytical methods. Hematocrit was estimated by three methods: the BGE analyzer the S-Plus Jr ana- lyzer (Coulter Diagnostics), an EPC that dilutes speci- mens 6250-fold with an isotonic solution of 68.4 mmolJL Na2SO4, 68 mmolIL NaCl, 11 mmolIL dimethy- lurea, and 0.4 mmoIIL procaine HCI and by centrifu- gation for 3-15 mm at 13 700g in an Autocrit Ultra 3 (Becton Dickinson, Parsippany, NJ) with use of Crito- caps (Sherwood Medical Co., St. Louis, MO) to deter- mine PCV. Osmolality was measured by freezing-point depression in a Micro-Osmometer 3MO (Advanced In- struments, Needham Heights, MA). Sodium and potas- sium were determined in conjunction with hematocrit analyses on the BGE analyzer. Protein in plasma was measured with an Ektachem 700 analyzer (Kodak, Rochester, NY). Results A preliminary correlation of BGE hematocrit results with the EPC and PCV methods found close agreement between the three methods for specimens from outpa- tient and inpatient (non-intensive-care) sources. BGE and PCV values were highly concordant (BGE = 1.00 PCV + 0.3% S = 1.6%, n = 34) EPC values averaged