Show-Hong Duh

Performance evaluation of blood glucose monitoring devices.

Many new technologies are being applied to measure blood glucose concentrations, but there is a lack of a standardized approach to evaluate performance of these devices. We sought to identify the key elements in evaluating the performance of devices for measuring blood glucose. We examined these elements in a multicenter study of four brands of glucose meters that are commonly used by diabetic patients. We tested control materials, spiked whole blood specimens, and 461 heparinized whole blood specimens in triplicate by each of the four brand glucose meters, and analyzed the plasma glucose concentrations of these specimens by a hexokinase (HK) method that incorporated reference materials developed by National Institute of Standards and Technology. Testing with glucose meters was performed at three sites, with multiple operators, meters, and representative lots of reagents. We evaluated the systematic bias, random error, and clinical significance of glucose meters. Meters were precise with a coefficient of variation of <4% across a wide range of glucose concentrations. Slopes significantly different from 1.0 were observed for two meters with 11-13% and -11% to -13% at the 95% confidence interval level by the linear regression of meter results versus the HK method from 33 to 481 mg/dL (correlation coefficient >0.98 and standard error of estimation S(y/x) <13 mg/dL for both meters). Analysis of the clinical significance of bias by Clarke Error Grid showed that results of the four meters were outside the accurate zone (26.5%, 2.4%, 1.5%, and 5.6%). Only a small number of the results showed clinically significant bias, mostly in the hypoglycemic range. Meters performed consistently throughout the study and, generally, were precise, although precision varied at extremely high or low glucose concentrations. Two of the glucose meters had substantial systematic bias when compared with an HK method, indicating a need for improving calibration and standardization. Analytical performance varied over the physiological range of glucose values so that separate accuracy and precision goals should be defined for hypoglycemic, normoglycemic, and hyperglycemic ranges. This study describes the current state of performance of blood glucose monitoring devices and points out those factors that should be assessed during evaluation of new devices.

Interpreting Cardiac Troponin Results from High-Sensitivity Assays in Chronic Kidney Disease without Acute Coronary Syndrome

BACKGROUND Quantification and comparison of high-sensitivity (hs) cardiac troponin I (cTnI) and cTnT concentrations in chronic kidney disease (CKD) have not been reported. We examined the associations between hs cTnI and cTnT, cardiovascular disease, and renal function in outpatients with stable CKD. METHODS Outpatients (n = 148; 16.9% with prior myocardial infarction or coronary revascularization) with an estimated glomerular filtration rate (eGFR) of <60 mL · min⁻¹ · (1.73 m²)⁻¹ had serum cTnI (99th percentile of a healthy population = 9.0 ng/L), and cTnT (99th percentile = 14 ng/L) measured with hs assays. Left ventricular ejection fraction (LVEF) and mass were assessed by echocardiography, and coronary artery calcification (CAC) was determined by computed tomography. Renal function was estimated by eGFR and urine albumin/creatinine ratio (UACR). RESULTS The median (interquartile range) concentrations of cTnI and cTnT were 6.3 (3.4-14.4) ng/L and 17.0 (11.2-31.4) ng/L, respectively; 38% and 68% of patients had a cTnI and cTnT above the 99th percentile, respectively. The median CAC score was 80.8 (0.7-308.6), LV mass index was 85 (73-99) g/m², and LVEF was 58% (57%-61%). The prevalences of prior coronary disease events, CAC score, and LV mass index were higher with increasing concentrations from both hs cardiac troponin assays (P < 0.05 for all). After adjustment for demographics and risk factors, neither cardiac troponin assay was associated with CAC, but both remained associated with LV mass index as well as eGFR and UACR. CONCLUSIONS Increased hs cTnI and cTnT concentrations are common in outpatients with stable CKD and are influenced by both underlying cardiac and renal disease.

Relation of temporal creatine kinase-MB release and outcome after thrombolytic therapy for acute myocardial infarction

Measuring biochemical marker release after acute myocardial infarction helps in estimating infarct size and prognosis. We sought to relate in-hospital outcomes and curve-fitted creatine kinase (CK)-MB variables after thrombolysis. We measured CK-MB mass initially and at 30 and 90 minutes, and at 3, 8, and 20 hours after thrombolysis in 130 patients also undergoing cardiac catheterization at 90 minutes and at 5 to 7 days. Data were fitted, and maximums and curve areas calculated. CK-MB maximums related to infarct location (p = 0.014) and time to therapy (p = 0.002); curve area did not. Neither maximums nor curve area related to Thrombolysis in Myocardial Infarction trial flow grade at 90 minutes. Maximums related to ejection fraction at 90 minutes (p = 0.0004) and at 5 to 7 days (p = 0.0014), as did curve area (p = 0.0076 and 0.030, respectively). Maximums related to infarct zone function at 90 minutes (p = 0.024) and at 5 to 7 days (p = 0.042); curve area related only at 90 minutes (p = 0.027). Both maximums and curve area predicted congestive heart failure (p = 0.008 and p = 0.042, respectively) and a composite of congestive heart failure or death (p = 0.004 and p = 0.047, respectively); however, after adjusting for maximums, curve area no longer predicted congestive heart failure (p = 0.92). Maximums predicted the composite outcome after adjustment for curve area, and showed a trend toward predicting congestive heart failure (p = 0.089). We conclude that CK-MB maximums relate to infarct zone function, left ventricular function, and in-hospital outcomes after thrombolysis for acute myocardial infarction.

Assessment of Coronary Reperfusion After Thrombolysis With a Model Combining Myoglobin, Creatine Kinase–MB, and Clinical Variables

BACKGROUND Several biochemical markers have been investigated for the noninvasive assessment of reperfusion after myocardial infarction. Because myoglobin is released very soon after myocardial injury and clears rapidly after reperfusion, it may prove to be an excellent marker of occlusion and reperfusion. METHODS AND RESULTS We examined the relation between various myoglobin measures and Thrombolysis In Myocardial Infarction (TIMI) flow grade in 96 patients enrolled in a study of front-loaded thrombolysis who underwent 90-minute angiography. We also combined myoglobin measures with models that include clinical and creatine kinase-MB variables. The myoglobin level measured within 10 minutes of acute angiography showed the best overall performance and was used for later analyses. Of the clinical variables examined, only time from symptom onset to thrombolysis and chest pain grade at angiography discriminated among TIMI flow grades. Combining the 90-minute myoglobin level and these clinical variables showed a significant difference (P<.0001) between both TIMI 3 versus TIMI 0 through 2 and TIMI 2 or 3 versus TIMI 0 or 1 flow. When the 90-minute myoglobin level was added to an established predictive model containing clinical variables and creatine kinase-MB measures, its contribution remained significant (P=.044). The area under the receiver operator characteristic curve for this combined model was .88. CONCLUSIONS A single myoglobin measurement obtained 90 minutes after the start of thrombolysis, combined with select clinical variables and creatine kinase-MB levels, enhances the noninvasive prediction of reperfusion after myocardial infarction.

Fibroblast Growth Factor 23, High-Sensitivity Cardiac Troponin, and Left Ventricular Hypertrophy in CKD

BACKGROUND Detectable levels of cardiac troponins are common in individuals with chronic kidney disease (CKD), even in the absence of symptomatic cardiovascular disease. Abnormal cardiac troponin values are associated with coronary artery disease and left ventricular hypertrophy (LVH) and predict poor clinical outcomes. Elevated levels of fibroblast growth factor 23 (FGF-23) contribute to LVH in CKD. We investigated the association of FGF-23 and hs-cTnI (high-sensitivity cardiac troponin I) and hs-cTnT (high-sensitivity cardiac troponin T) levels in CKD and examined the role of LVH in this association. STUDY DESIGN Cross-sectional observational study. SETTING & PARTICIPANTS 153 stable outpatients with non-dialysis-dependent CKD. PREDICTOR The primary predictor was FGF-23 level. OUTCOMES hs-cTnI, hs-cTnT. MEASUREMENTS FGF-23, hs-cTnI, hs-cTnT; left ventricular mass index (LVMI) assessed by echocardiography; coronary artery calcification (CAC) measured by computed tomography. LVMI and CAC were evaluated as potential mediators of the effect of FGF-23 on hs-cTnI/T. RESULTS Mean age was 64 ± 12 (SD) years, mean estimated glomerular filtration rate was 34 ± 11 mL/min/1.73 m(2), median FGF-23 level was 120 (25th-75th percentile, 79-223) reference unit (RU)/mL, median hs-cTnI level was 6.5 (25th-75th percentile, 3.5-14.5) pg/mL, and median hs-cTnT level was 16.8 (25th-75th percentile, 11.1-33.9) pg/mL. cTnI and cTnT concentrations were higher than the 99th percentile of a healthy population in 42% and 61% of patients, respectively. In unadjusted and multivariable-adjusted analyses, hs-cTnI/T levels were associated significantly with FGF-23 levels. Adjusting for LVMI, but not CAC, weakened the association of FGF-23 and hs-cTnI/T levels. LIMITATIONS Vitamin D levels were not measured. The prevalence of coronary artery disease may have been underestimated because it was ascertained by self-report. CONCLUSIONS Minimally elevated cTnI and cTnT levels, detectable by high-sensitivity assays, are associated with elevated FGF-23 levels in stable outpatients with CKD. FGF-23-associated LVH may contribute to detectable hs-cTnI/T levels observed in non-dialysis-dependent patients with CKD.

Characteristics of myoglobin, carbonic anhydrase III and the myoglobin/carbonic anhydrase III ratio in trauma, exercise, and myocardial infarction patients

Carbonic anhydrase III (CA III) is an enzyme present in skeletal muscle which is released into circulation following injury. Myoglobin (Mb) is a heme protein located in skeletal, smooth, and cardiac muscle which is also released after injury. Because CA III is not present in myocardium, combining serum CA III and Mb measurements may improve the specificity of Mb as an early diagnostic marker for myocardial infarction (MI) provided that a fixed ratio of Mb and CA III is released from skeletal muscle following cell injury. We examined release of Mb and CA III for exercise subjects (n=12), trauma patients (n=18), and MI patients (n=10) following emergency department admission. A fixed ratio of Mb/CA III had medians of 3.505 (range: 1.05-6.76) and 2.890 (range: 0.97-3.97) for exercise and trauma subjects, respectively, in samples collected within 5 h of the event. The Mb/CA III ratio was significantly higher (P<0.001) in MI patients (median: 35.395; range: 8.65-170.45) during this same time. This study confirmed that Mb and CA III are released in a fixed ratio following exercise, showed no significant difference in the ratio for trauma patients, and demonstrated significant ratio elevation for MI patients. These data suggest the ratio to be a useful diagnostic indicator of MI.

Evidence based approach to practice guides and decision thresholds for cardiac markers

There is substantial evidence for use of CK-MB, myoglobin, cardiac troponin T (cTnT) and cardiac troponin I (cTnI) for diagnosis of myocardial infarction (MI) and risk stratification of the acute coronary syndromes. Timing of specimen collection, patient population, and decision limits are important considerations. Serial CK-MB measurement must be considered a benchmark of cardiac markers. The National Heart Attack Alert Program (NHAAP) determined that high quality clinical trials show that serial CK-MB measurements are very accurate for diagnosis of MI and have large clinical impact. Meta-analysis of CK-MB mass for retrospective MI diagnosis yielded a clinical sensitivity of 96.8% and a specificity of 89.6% for sampling at 12-48 hours. Numerous studies of myoglobin have demonstrated a high negative predictive value and high sensitivity 2-6 hours after presentation. The NHAAP group, however, rated performance of myoglobin as modestly accurate with small clinical impact. cTnI has a clinical sensitivity and specificity for diagnosis of MI in the range of 90% and 97%, respectively. However decision limits for cTnI may differ by up to 10-fold for different assays, indicating need for standardization. Meta-analysis showed that cTnT had a sensitivity of 98.2% (CI: 97-99%) for diagnosis of MI with lower specificity due to detection of minor myocardial injury. Both cTnT and cTnI are more useful for stratifying risk in acute coronary syndrome patients than CK-MB mass. The predictive ability of both cTnT and cTnI is optimized by sampling > 6-12 hours after symptoms onset. But in unstable angina, the odds ratios for predicting outcomes of MI/short-term mortality with 12-24 hour sampling for cTnT was 2.7 the odds ratio for cTnI did not differ significantly at 4.2.

Impact of Increased Body Mass Index on Accuracy of B-Type Natriuretic Peptide (BNP) and N-Terminal proBNP for Diagnosis of Decompensated Heart Failure and Prediction of All-Cause Mortality

BACKGROUND BNP and N-terminal proBNP (NT-proBNP) concentrations may be depressed in patients with increased body mass index (BMI). Whether increased BMI affects accuracy of these biomarkers for diagnosing decompensated heart failure (HF) and predicting outcomes is unknown. METHODS We measured BNP and NT-proBNP in 685 patients with possible decompensated HF in a free-living community population subdivided by BMI as obese, overweight, and normal weight. HF diagnosis was adjudicated by a cardiologist blinded to BNP and NT-proBNP results. We tabulated all-cause mortality over a median follow-up of 401 days and assessed marker accuracy for HF diagnosis and mortality by ROC analysis. RESULTS Of the 685 patients, 40.9% were obese (n = 280), 28.2% were overweight (n = 193), and 30.9% had normal BMI (n = 212). Obese patients had lower BNP and NT-proBNP compared with overweight or normal-weight individuals (P < 0.001) and decreased mortality compared with normal-weight individuals (P < 0.001). Both biomarkers added significantly to a multivariate logistic regression model for diagnosis of decompensated HF across BMI categories. NT-proBNP outperformed BNP for predicting all-cause mortality in normal-weight individuals (chi(2) for BNP = 6.4, P = 0.09; chi(2) for NT-proBNP = 16.5, P < 0.001). Multivariate regression showed that both biomarkers remained significant predictors of decompensated HF diagnosis in each BMI subgroup. CONCLUSIONS In this study population, obese patients had significantly lower BNP and NT-proBNP that reflected lower mortality. BNP and NT-proBNP can be used in all BMI groups for decompensated HF diagnosis, although BMI-specific cutpoints may be necessary to optimize sensitivity.

Clinical performance characteristics of a new photometric lithium assay: a multicenter study

BACKGROUND Therapeutic monitoring of lithium is important because of its narrow therapeutic range and therapeutic index, low protein binding and single route of elimination. We characterized a new photometric method that avoids the specialized requirements of ion-specific electrode (ISE), atomic absorption and flame emission methods. METHODS Minimum detectable concentration (MDC), linearity and calibration drift over 65 days were determined. Within-run, between-run and total imprecision were assessed over 20 days in accordance with NCCLS EP5. Interference studies were conducted for 46 endogenous and exogenous compounds. Two production lots of the new photometric method (LI) were compared on the Dimension(R) RxL system and two ISE methods [Ciba Corning (n=124) and DuPont (n=131)], an established photometric method (Vitros) 950 system; n=63) and atomic absorption (Thermo-Jerell Ash; n=63). RESULTS The MDC was 0.04 mmol/l. Linearity was demonstrated from 0.12 to 5.8 mmol/l by the regression equation: observed=(1.01 x expected)-0.0005 mmol/l, S(y/x)=0.03 mmol/l, r=0.999. Drift for the lithium calibrators over the 65-day study period was <5%, except for the lowest calibrator, which showed 0.04 mmol/l drift. None of the 46 potential interfering substances showed greater than a 6.5% difference between control and test solutions. ISE method comparisons showed the following: LI=(1.08 x Ciba Corning ISE)-0.15 mmol/l, S(y/x)=0.05, r=0.999, and LI=(1.03 x DuPont ISE)+0.00 mmol/l, S(y/x)=0.06 mmol/l, r=0.999. Comparison of the LI method with the atomic absorption and Vitros system showed proportionality error <10%. Bias between the LI method and atomic absorption was 7%, substantially less than that documented in proficiency surveys for the Vitros and other systems. No lot-to-lot or site-to-site differences were observed. CONCLUSION This new photometric method is an attractive alternative for Li measurement and is adaptable to instruments having spectrophotometric capability.

Cardiovascular disease testing on the Dimension Vista® system: Biomarkers of acute coronary syndromes

OBJECTIVES Performance characteristics of the LOCI cTnI, CK-MB, MYO, NTproBNP and hsCRP methods on the Dimension Vista System were evaluated. DESIGN AND METHODS Imprecision (following CLSI EP05-A2 guidelines), limit of quantitation (cTnI), limit of blank, linearity on dilution, serum versus plasma matrix studies (cTnI), and method comparison studies were conducted. RESULTS Method imprecision of 1.8 to 9.7% (cTnI), 1.8 to 5.7% (CK-MB), 2.1 to 2.2% (MYO), 1.6 to 3.3% (NTproBNP), and 3.5 to 4.2% (hsCRP) were demonstrated. The manufacturers claimed imprecision, detection limits and upper measurement limits were met. Limit of Quantitation was 0.040 ng/mL for the cTnI assay. Agreement of serum and plasma values for cTnI (r=0.99) was shown. Method comparison study results were acceptable. CONCLUSIONS The Dimension Vista cTnI, CK-MB, MYO, NTproBNP, and hsCRP methods demonstrate acceptable performance characteristics for use as an aid in the diagnosis and risk assessment of patients presenting with suspected acute coronary syndromes.