Amy M. Wockenfus

Precision and Reliability of 5 Platelet Function Tests in Healthy Volunteers and Donors on Daily Antiplatelet Agent Therapy

BACKGROUND Anticoagulation protocols used during mechanical circulatory support call for titration of antiplatelet agents. We compared the precision and reliability of 5 platelet function tests in healthy volunteers and donors on daily antiplatelet therapy to distinguish their efficacy for titrating antiplatelet therapy. METHODS We assessed arachidonic acid-induced platelet function by light transmission aggregometry (LTA), Multiplate impedance aggregometry, VerifyNow, and platelet mapping by thromboelastography (TEG PM). We assessed ADP-induced platelet function by the same methods and flow cytometry. Forty healthy volunteers and 10-13 volunteers on daily aspirin and/or clopidogrel therapy were evaluated. We compared tests for intraassay precision, interassay precision (samples from 2 separate blood draws), and reliability coefficient. RESULTS For arachidonic acid-induced platelet aggregation in healthy volunteers, intra- and interassay CVs were ≤ 10% for all methods. Intra- and interassay precision among donors on daily aspirin was ≤ 30% for all methods except LTA (38% interassay CV) and TEG PM (95% intraassay and 104% interassay CV). For ADP-induced platelet function, intra- and interassay precision was ≤ 10% and ≤ 30% for all methods. Only Multiplate demonstrated moderate or greater (R > 0.40) reliability coefficients for arachidonic acid-induced platelet function among all subjects. All methods of ADP-induced platelet function, except TEG PM, demonstrated substantial or greater (R > 0.60) reliability among all subjects. CONCLUSIONS TEG PM is least suited to monitor effects of antiplatelet agents. Multiplate impedance aggregometry was the only method to demonstrate an acceptable reliability coefficient among healthy volunteers and donors on both aspirin and clopidogrel therapy.

BD rapid serum tubes reduce false positive plasma troponin T results on the Roche Cobas e411 analyzer

OBJECTIVES In an attempt to reduce false positive results and improve turnaround time, we investigated the BD Rapid Serum Tube as an alternate sample type to lithium heparin plasma for Roche Troponin T analysis on the Roche Cobas e411 analyzer. DESIGN AND METHODS BD Plasma Separator Tubes (PST) and Rapid Serum Tubes (RST) were collected in tandem from Emergency Department patients who had clinical orders for Troponin T over a 1 month period. RESULTS RST and PST samples yielded analytically and clinically concordant Troponin T results on the Roche Cobas e411. Rare false positive results in lithium heparin samples were not observed with rapid clot serum tubes. CONCLUSIONS RST samples are appropriate for stat Troponin T analysis, and appear to reduce the incidence of rare false positive Troponin T results obtained with lithium heparin samples.

Accuracy of whole blood glucose measurement when venous catheter blood samples are used on glucose meters.

BACKGROUND Previous studies have found positive bias and frequent outliers when central venous catheter (CVC) whole blood is used to dose glucose meters. We designed a study to determine whether positive bias and outliers with CVC whole blood glucose samples are due to exogenous glucose contamination of CVC samples, inherent bias and imprecision of glucose meters, or properties of CVC whole blood that interfere with the function of some glucose meters. METHODS We studied the relationship between venous whole blood and venous plasma glucose drawn by venipuncture to CVC whole blood and CVC plasma glucose in 50 hospitalized patients. In 27 patients whole blood glucose was measured on both the Accu-Chek Inform (Roche Diagnostics, Indianapolis, IN) and StatStrip (Nova Biomedical, Waltham, MA). RESULTS By comparing CVC plasma to venous plasma glucose, we determined that contamination of CVC samples with exogenous glucose was uncommon. On the Inform meter outliers were approximately twice as common with CVC whole blood compared to venous whole blood. In 27 patients who had CVC whole blood analyzed by both Inform and StatStrip, outliers occurred approximately twice as often on the Inform compared to the StatStrip. Accounting for CVC samples contaminated with exogenous glucose, outliers on the StatStrip did not occur significantly more often using CVC whole blood compared to venous whole blood. CONCLUSIONS Properties unique to CVC whole blood differentially affect glucose meter bias and imprecision. Device selection is critical in practices that wish to use CVC whole blood to monitor glucose concentration in hospitalized patients.

Analytical performance of three whole blood point-of-care lactate devices compared to plasma lactate comparison methods and a flow-injection mass spectrometry method

OBJECTIVES Point of care (POC) whole blood lactate testing may facilitate rapid detection of sepsis. We evaluated three POC methods against both plasma lactate comparison methods and a flow-injection mass spectrometric (MS) method. DESIGN AND METHODS Nova StatStrip, Abbott i-STAT CG4+ and Radiometer ABL90 POC lactate methods were evaluated against the mean of Cobas Integra 400 and Vitros 350 plasma lactate. POC methods were also compared to a flow-injection mass spectrometric assay measuring lactate in ZnSO4-precipitated whole blood extracts. Intra- and inter-assay precision was determined using quality control material. Method comparison included specimens from normal donors at rest, after exertion, and after spiking with lactic acid. RESULTS Intra- and inter-assay coefficient of variation was <5% for i-STAT and ABL90; but ranged from 3.1-8.2% on two StatStrip meters. Mean (±SD) bias between POC and plasma lactate ranged from -0.2±0.9 (i-STAT and ABL90) to -0.4±1.2 (StatStrip) mmol/L. At concentrations >6mmol/L, all POC methods showed proportional negative bias compared to plasma methods; but this bias was not observed when compared to the MS method. Despite proportional negative bias, all POC methods demonstrated acceptable concordance (94-100%) with plasma lactate within the reference interval (<2.3mmol/L) and >4mmol/L, commonly used clinical cut-offs for detection of sepsis. CONCLUSIONS POC lactate methods demonstrate acceptable concordance with plasma lactate across commonly used clinical cut-offs for detection of sepsis. Due to systematic negative bias at higher lactate concentrations, POC and plasma lactate should not be used interchangeably to monitor patients with elevated lactate concentrations.

Evaluation of lactate, white blood cell count, neutrophil count, procalcitonin and immature granulocyte count as biomarkers for sepsis in emergency department patients

BACKGROUND Lactate, white blood cell (WBC) and neutrophil count, procalcitonin and immature granulocyte (IG) count were compared for the prediction of sepsis, and severe sepsis or septic shock, in patients presenting to the emergency department (ED). METHODS We prospectively enrolled 501 ED patients with a sepsis panel ordered for suspicion of sepsis. WBC, neutrophil, and IG counts were measured on a Sysmex XT-2000i analyzer. Lactate was measured by i-STAT, and procalcitonin by Brahms Kryptor. We classified patients as having sepsis using a simplification of the 1992 consensus conference sepsis definitions. Patients with sepsis were further classified as having severe sepsis or septic shock using established criteria. Univariate receiver operating characteristic (ROC) analysis was performed to determine odds ratio (OR), area under the ROC curve (AUC), and sensitivity/specificity at optimal cut-off for prediction of sepsis (vs. no sepsis), and prediction of severe sepsis or septic shock (vs. no sepsis). RESULTS There were 267 patients without sepsis; and 234 with sepsis, including 35 patients with severe sepsis or septic shock. Lactate had the highest OR (1.44, 95th% CI 1.20-1.73) for the prediction of sepsis; while WBC, neutrophil count and percent (neutrophil/WBC) had OR>1.00 (p<0.05). All biomarkers had AUC<0.70 and sensitivity and specificity <70% at the optimal cut-off. Initial lactate was the best biomarker for predicting severe sepsis or septic shock, with an odds ratio (95th% CI) of 2.70 (2.02-3.61) and AUC 0.89 (0.82-0.96). CONCLUSION Traditional biomarkers (lactate, WBC, neutrophil count, procalcitonin, IG) have limited utility in the prediction of sepsis.

Discordance between urine pH measured by dipstick and pH meter: Implications for methotrexate administration protocols

OBJECTIVES To minimize toxicity of high-dose methotrexate (MTX) therapy, urinary alkalinization with frequent monitoring of urine pH is required. Urine pH is usually assessed by fast and convenient dipstick methods. When urine color interferes with dipstick measurement, as occurs in patients receiving MTX, alternative methods such as pH meters are used. Nursing staff caring for patients on high-dose MTX reported that urine pH results from dipstick and pH analyzers were often clinically discordant. As a result urine pH by dipstick and pH meter were compared in patients on high-dose MTX therapy and patients with normal-colored urine samples. DESIGN AND METHODS We measured urine pH by dipstick and pH meter in 116 urine samples from 4 patients receiving high-dose MTX therapy, and in 50 normal-colored urine samples from 50 patients not on MTX therapy. RESULTS In patients on MTX therapy the mean (±standard deviation) bias between dipstick and pH meter urine pH was 0.7±0.4, compared to 0.4±0.3 in patients not on MTX. For patients on MTX clinical concordance between dipstick and pH meter urine results was poor around a clinical cut-off of pH 8.0. Of the 92 samples with a meter urine pH≤8.0, 72 had a discordant value by dipstick (pH>8). CONCLUSIONS Urine pH readings by dipstick and pH meter are not equivalent, and the bias between them is exacerbated in patients on MTX. Institutions with high-dose MTX therapy protocols should not alternate between dipstick and pH meter urine pH monitoring.

Agreement between whole blood and plasma sodium measurements in profound hyponatremia

OBJECTIVES We compared two different methods of whole blood sodium measurement to plasma sodium measurement using samples in the profoundly hyponatremic range (Na < 120 mmol/L). DESIGN AND METHODS Whole blood pools with a range of low sodium values were generated using combinations and dilutions of pooled electrolyte-balanced lithium heparin samples submitted for arterial blood gas analysis. Each pool was analyzed five times on a Radiometer 827 blood gas analyzer and iSTAT analyzer. Pools were centrifuged to produce plasma, which was analyzed five times on a Roche Cobas c501 chemistry analyzer. An additional 40 fresh (analyzed on day of collection) excess lithium heparin arterial blood gas samples from 36 patients were analyzed on the Radiometer 827, iSTAT, and Cobas c501 as described above. The setting was a tertiary referral center. Blood samples were collected from a combination of patients in the intensive care unit, operating theaters and emergency room. RESULTS All methods demonstrated excellent precision, even in the profoundly hyponatremic measurement range (Na < 120 mmol/L using a plasma reference method). However, agreement between the methods varied with the degree of hyponatremia. In the profoundly hyponatremic range, Radiometer whole blood sodium values were nearly identical to plasma reference sodium, while iSTAT whole blood sodium showed a consistent positive bias relative to plasma sodium in this range. CONCLUSION If whole blood direct sodium measurements are compared with plasma sodium in profoundly hyponatremic patients consideration should be given to the use of Radiometer blood gas analyzers over iSTAT since the latter shows a positive bias relative to a plasma comparative method.

Comparison of two point of care devices for capillary lipid screening in fasting and postprandial adults

OBJECTIVES The aim of this study was to assess the performance of two point of care (POC) devices for capillary lipid screening in fasting and post-prandial adults. DESIGN AND METHODS Fasting and post-prandial capillary whole blood samples collected from 57 adult donors were analyzed simultaneously on Cholestech LDX Lipid Profile (Alere San Diego, Inc., San Diego, CA) cassettes and CardioChek Lipid Panel (Polymer Technology Systems, Indianapolis, IN) strips. Paired serum samples were collected from the same donors and analyzed with CDC-certified methods for total cholesterol, high density lipoprotein cholesterol (HDL-C) and non-blanked triglycerides. Non-HDL-C (total cholesterol minus HDL-C) and low density lipoprotein cholesterol (LDL-C) were calculated. Mean bias between capillary whole blood and serum laboratory lipids was calculated. RESULTS HDL-C measurements were not affected by triglyceride content on either device. However, both devices exhibited significant variability in triglyceride measurement relative to the reference method. Compared to reference methods, Cholestech was more accurate than CardioChek for non-HDL-C while CardioChek was more accurate for HDL-C. Among the calculated cardiovascular risk parameters (LDL-C and non-HDL-C), Cholestech-calculated non-HDL-C exhibited the least average bias in both fasting and postprandial samples. CONCLUSIONS The optimal approach to capillary lipid screening may be to use Cholestech non-HDL cholesterol; as it exhibited little bias relative to CDC reference methods in both fasting and postprandial samples, facilitating lipid screening in non-fasting adults.

Comparing analytical outliers and the percent of emergency department patients with results above the 99th percentile upper reference limit for 2 conventional and one high sensitivity troponin assay

OBJECTIVES We compared rates of analytical outliers, and percent of emergency department (ED) patients with cardiac troponin (cTn) values above the 99th percentile upper reference limit (URL), for two conventional and one high sensitivity cTn assay. METHODS We measured 3008 samples from 1931 ED patients by Roche e411 4th generation Troponin T (cTnT); and Abbott STAT Troponin I (cTnI) and high sensitivity troponin I (hscTnI) on an Architect i2000. Within 24h of initial measurement, samples were aliquoted, re-centrifuged, and repeated in duplicate by all methods. Outliers were defined as one or both replicates exceeding initial value by a critical difference (CD): where CD=z×2×SDanalytical (z=3.29 at a probability of 0.0005), and at least one replicate on a different side of 99th percentile URL compared to initial value. We also assessed percent of ED patients with values >99th percentile by all methods (excluding outliers), using both sex-neutral and sex-specific hscTnI URL. RESULTS The outlier rate for cTnI (3.66%) was significantly higher than the outlier rate for either cTnT (0.33%) or hscTnI (0.47%) (p<0.0001). More ED patients (33%) had elevated cTnT values compared to either cTnI (25%) or hscTnI (29%). Application of sex-specific URL did not change the percent of ED patients with >99th percentile hscTnI values. CONCLUSION Abbott STAT cTnI had more analytic outliers than Roche cTnT or Abbott hscTnI. Compared to cTnT, use of hscTnI will significantly decrease the percent of ED patients with elevated cTn values without increasing analytical outliers.

Blood gas sample spiking with total parenteral nutrition, lipid emulsion, and concentrated dextrose solutions as a model for predicting sample contamination based on glucose result

OBJECTIVE Evaluate the effects of blood gas sample contamination with total parenteral nutrition (TPN)/lipid emulsion and dextrose 50% (D50) solutions on blood gas and electrolyte measurement; and determine whether glucose concentration can predict blood gas sample contamination with TPN/lipid emulsion or D50. DESIGN AND METHODS Residual lithium heparin arterial blood gas samples were spiked with TPN/lipid emulsion (0 to 15%) and D50 solutions (0 to 2.5%). Blood gas (pH, pCO2, pO2), electrolytes (Na+, K+ ionized calcium) and hemoglobin were measured with a Radiometer ABL90. Glucose concentration was measured in separated plasma by Roche Cobas c501. Chart review of neonatal blood gas results with glucose >300 mg/dL (>16.65 mmol/L) over a seven month period was performed to determine whether repeat (within 4 h) blood gas results suggested pre-analytical errors in blood gas results. Results were used to determine whether a glucose threshold could predict contamination resulting in blood gas and electrolyte results with greater than laboratory-defined allowable error. RESULTS Samples spiked with 5% or more TPN/lipid emulsion solution or 1% D50 showed glucose concentration >500 mg/dL (>27.75 mmol/L) and produced blood gas (pH, pO2, pCO2) results with greater than laboratory-defined allowable error. TPN/lipid emulsion, but not D50, produced greater than allowable error in electrolyte (Na+,K+,Ca++,Hb) results at these concentrations. Based on chart review of 144 neonatal blood gas results with glucose >250 mg/dL received over seven months, four of ten neonatal intensive care unit (NICU) patients with glucose results >500 mg/dL and repeat blood gas results within 4 h had results highly suggestive of pre-analytical error. Only 3 of 36 NICU patients with glucose results 300-500 mg/dL and repeat blood gas results within 4 h had clear pre-analytical errors in blood gas results. CONCLUSION Glucose concentration can be used as an indicator of significant blood sample contamination with either TPN/lipid emulsion or D50 solution. NICU blood gas samples with glucose ≥300 mg/dL should be considered potentially contaminated, and samples with glucose >500 mg/dL have a risk for contamination.