Sample to sample carryover: a source of analytical laboratory error and its relevance to integrated clinical chemistry/immunoassay systems

BACKGROUND

Integrated systems that combine clinical chemistry and immunoassay analyzers are used routinely. Sample to sample carryover is an inherent risk and can cause erroneously high patient test results for immunoassays. IVD manufacturers and laboratories must be aware of this phenomenon and guard against it.

METHODS

We used a sample carryover protocol that directs the clinical chemistry module to process samples with very high immunoassay analyte concentrations followed by samples with very low concentrations for the same analyte. Low concentration samples were then tested by the immunoassay module to determine if the clinical chemistry module caused primary sample tube to primary sample tube carryover of the immunoassay analyte.

RESULTS

Sample carryover was assessed on the Abbott ci8200 for HBsAg, AFP, beta-hCG, and PSA. Observed HBsAg carryover met the design specification of <0.1 ppm. Carryover for the other analytes was <0.1 ppm or below the assay limit of detection.

CONCLUSIONS

IVD manufacturers must design integrated systems to minimize primary specimen tube carryover and avoid analytical laboratory error that can impact patient safety. Carryover testing is difficult for clinical laboratories to perform in order to verify system performance. Laboratories must consider the potential for specimen carryover and its impact on results whether moving primary sample tubes between separate analyzers or using an integrated system.

Nanodiagnostics: a new frontier for clinical laboratory medicine

BACKGROUND

The use of nanotechnologies for diagnostic applications shows great promise to meet the rigorous demands of the clinical laboratory for sensitivity and cost-effectiveness. New nanodiagnostic tools include quantum dots (QDs), gold nanoparticles, and cantilevers. QDs, which are the most promising nanostructures for diagnostic applications, are semiconductor nanocrystals characterized by high photostability, single-wavelength excitation, and size-tunable emission. QDs and magnetic nanoparticles can be used for barcoding of specific analytes. Gold and magnetic nanoparticles are key components of the bio-barcode assay, which has been proposed as a future alternative to the PCR.

METHODS

We examined articles published over the past 10 years investigating the use of QDs, gold nanoparticles, cantilevers, and other nanotechnologies in promising diagnostic applications.

RESULTS

Several nanodiagnostic assays have been developed, including a QD-based assay capable of detecting biotinylated prostate-specific antigen (PSA) at 0.38 ng/L, a bio-barcode assay capable of detecting 30 amol/L PSA in a 10-microL sample, and another able to detect 50 molecules of the Alzheimer marker amyloid beta-derived diffusible ligand in 10 microL of cerebrospinal fluid.

CONCLUSIONS

Nanodiagnostics promise increased sensitivity, multiplexing capabilities, and reduced cost for many diagnostic applications as well as intracellular imaging. Further work is needed to fully optimize these diagnostic nanotechnologies for clinical laboratory setting and to address the potential health and environmental risks related to QDs.

Conflict between guideline methodologic quality and recommendation validity: a potential problem for practitioners

BACKGROUND

It is not clear if good methodologic quality in current practice guidelines necessarily leads to more valid recommendations, i.e., those that are supported with consistent research evidence or, when evidence is conflicting or lacking, with sufficient consensus among the guideline development team. To help clarify this issue, we assessed whether there is a link between methodologic quality and recommendation validity in practice guidelines for the use of laboratory tests in the management of patients with non-small cell lung cancer (NSCLC).

METHODS

We conducted a systematic review of data on laboratory tests in NSCLC published in English or in French within the last 10 years and retrieved 11 practice guidelines for the use of these tests. The guidelines were critically appraised and scored for methodologic quality and recommendation validity based on the Appraisal of Guidelines Research and Evaluation (AGREE) criteria and on the systematic review.

RESULTS

Overall, these 11 guidelines had considerable shortcomings in methodologic quality and, to a lesser extent, in recommendation validity. Practice guidelines with the best methodologic quality were not necessarily the most valid in their recommendations, and conversely.

CONCLUSIONS

Poor methodologic quality and lack of recommendation validity in laboratory medicine call for methodologic standards of guideline development and for international collaboration of guideline development agencies. We advise readers of guidelines to critically evaluate the methods used as well as the content of the recommendations before adopting them for use in practice.

Evaluation of the VITROS chemistry products dHDL slides for direct measurement of high density lipoprotein cholesterol

BACKGROUND

We evaluated the new VITROS direct high density lipoprotein cholesterol (dHDL) slide assay, based on the precipitation of apolipoprotein B-containing lipoproteins by phosphotungstic acid/magnesium chloride (PTA/MgCl2).

METHODS

We determined linearity, within-run and between-run imprecision for the VITROS dHDL slide assay, and compared it to the RANDOX two-step PTA/MgCl2 method for 300 fresh sera. Moreover, triglyceride and cholesterol interference were tested.

RESULTS

The VITROS dHDL slide assay was linear from 0 to 3.13 mmol/l. Within-run and between-run imprecision were 1.6%, 1.8%, 3.3% and 3.2% for target values at 1.00 and 1.40 mmol/l, respectively. For 300 fresh serum samples, the linear regression equation between the VITROS dHDL slide assay (y) and the RANDOX PTA/MgCl2 2 method (x) was: y (mmol/l)=0.934 x+0.043 (CI=0.916-0.952 for slope and 0.017-0.069 for intercept) . The bias of the VITROS assay compared to the RANDOX PTA/MgCl2 method was less than 5% for HDL cholesterol concentrations used for medical decision (1.04 and 1.50 mmol/l). Total error remained below 13% for triglyceride and LDL cholesterol concentrations below 6.6 mmol/l.

CONCLUSION

The VITROS dHDL slide assay is rapid, linear over a broad range of concentrations, and satisfies the National Cholesterol Education Program goals for precision and accuracy.

Accuracy of a user-friendly centrifuge for measuring creamatocrits on mothers' milk in the clinical setting

The creamatocrit (CRCT), a simple, accurate, and inexpensive technique for the estimation of lipid and caloric content in mothers' milk, has been used extensively in lactation research, but has not been integrated into the routine management of clinical lactation problems such as slow weight gain in mothers' milk-fed preterm and term infants. The Creamatocrit Plus is a lightweight, noiseless centrifuge with an embedded reader that automatically calculates lipid and calories from the CRCT value, making it ideal for use in the clinical setting. This study compared intra-user and inter-user reliability, the equivalence of the CRCT values obtained with the Creamatocrit Plus to the two standard techniques for performing CRCTs: the standard laboratory centrifuge with a hematocrit reader and the standard laboratory centrifuge with digital calipers, and the predictive accuracy of the Creamatocrit Plus for estimating the lipid and caloric content in mothers' milk. CRCTs were performed using the three techniques on 36 milk specimens from 12 women. Laboratory analyses of lipid and calories were performed by investigators blinded to CRCT values. The mean absolute intra-user and inter-user differences were all <1% CRCT, and the mean CRCT measures were nearly identical for the three measurement techniques. Linear correlations between CRCT and laboratory measures for lipid (r = 0.95) and calories (r = 0.94) were very high. The authors conclude that the Creamatocrit Plus can replace cumbersome laboratory equipment for measuring CRCTs in the clinical setting.

[Determination of nicotine and cotinine in human biological materials and their significance in toxicological studies]

The aim of this study was the preparation of reliable procedure of the determination of nicotine and cotinine both in classic (serum, urine) and alternative biological materials (hair, saliva) and evaluation of their significance for clinical and forensic toxicology. Biological material samples (blood, urine, saliva) were taken from patients after Percutaneous Trans-luminal Coronary Angioplasty (PTCA). The determination of cotinine and nicotine concentration in the biological material should be optimized depending on the aim of analysis. Liquid-liquid extraction procedure and high performance liquid chromatography HPLC/UV-DAD are reliable, specific and relatively cheap. Serum and saliva are valuable biological materials which allow to determine temporary nicotine and cotinine content on the similar level of concentrations. In the near future it will be able to replace blood with saliva sample because of an easy and non-invasive way of sampling. Evaluation of cotinine concentration in urine allows to distinguish the passive from the active tobacco smokers. Hair analysis allows to control a nicotine abstinence as well as a long-term evaluation of the history of smoking. However usage of hair is limited because of difficulty with sampling. Interpretation of results in analysis of alternative materials (hair, saliva) pose a problem because of lack of sampling standardization and lack of standardization of final analysis method.

Status of iodine nutrition in pregnant and lactating women in national capital district, Papua New Guinea

Urinary Iodine excretion is a useful and important indicator of the iodine status of a population. This study attempts to determine the urinary iodine concentration of non-pregnant, pregnant and lactating women, resident in the National Capital District of Papua New Guinea, so as to evaluate their status of iodine nutrition. The study population was made up of 56 non-pregnant, 40 lactating and 212 pregnant women. Of the 212 pregnant women, 14 were in the first, 64 in the second, and 134 in the third Trimester of pregnancy. Casual urine samples were collected and analysed for urinary iodine by Sandell-Kolthoff reaction. The median urinary iodine concentration for the non-pregnant, lactating and pregnant women was 163.0 micro g/L, 134.0 micro g/L and 180.0 micro g/L, respectively. Median urinary iodine for the first, second and third trimesters were 165.0 micro g/L, 221.5 micro g/L and 178.0 micro g/L, respectively. The 20th percentile urinary iodine values were higher than 50 micro g/L for all the groups. This indicates adequate intake of dietary iodine and optimal status of iodine nutrition amongst women in the various groups. Mild to severe status of iodine nutrition was found in 30.4% of non-pregnant, 35.0% of lactating, 22.2% of pregnant women, 28.5% of women in the first, 18.8% in the second, and 23.1% in the third trimester of pregnancy. To achieve optimal iodine nutrition in pregnant and lactating women, an increase in their intake of dietary iodine is recommended.

Gene expression analysis and clinical diagnosis

BACKGROUND

A new basis for diagnostic tests is being provided by the vast amount of data on gene expression that are now becoming available through large-scale measurement of mRNA abundance. The insights gained from these resources are most likely going to provide both a better basic understanding of disease mechanisms, and to identify molecular markers for more precise diagnoses and for prediction of prognosis and treatment response.

METHODS

Some quantitative RT-PCR assays are utilized today for diagnosis of both malignant and non-malignant disease, but the use of gene expression measurements in clinical medicine can be expected to increase dramatically.

CONCLUSIONS

There are important technical issues that must be adequately solved in order to obtain robust assays, such as standardized protocols with appropriate quality controls that ensure reliable data for the specific samples being analysed and good inter-laboratory reproducibility.

Clinical analysis by microchip capillary electrophoresis

Clinical analysis often requires rapid, automated, and high-throughput analytical systems. Microchip capillary electrophoresis (CE) has the potential to achieve very rapid analysis (typically seconds), easy integration of multiple analytical steps, and parallel operation. Although it is currently still in an early stage of development, there are already many reports in the literature describing the applications of microchip CE in clinical analysis. At the same time, more fully automated and higher throughput commercial instruments for microchip CE are becoming available and are expected to further enhance the development of applications of microchip CE in routine clinical testing. To put into perspective its potential, we briefly compare microchip CE with conventional CE and review developments in this technique that may be useful in diagnosis of major diseases.

Analytical evaluation of HgbA1c, microalbumin, CRP, and RF on Architect ci8200 integrated system and workflow performance evaluation using computer simulation

BACKGROUND

Recently, hemoglobin A1c (HgbA1c), microalbumin (MA), C-reactive protein (CRP) and rheumatoid factor (RF) have been introduced on high throughput general chemistry system. We evaluated analytical performance of these assays on an integrated clinical chemistry and immunoassay analyzer and studied the impact of testing these assays on these systems on the overall efficiency of the analyzer, via computer simulation.

METHODS

The analytical performance was measured by determining precision, linearity and correlation of patient sample results with in-house testing methodology. MedModel simulation software is used to develop simulation model and process efficiency is determined by measuring turnaround times and resource utilization.

RESULTS

Between-days CVs ranged from 8.59% for MA to 3.22% for HgbA1c level 1 controls. Less than 2% carryover for all 4 methods was observed on the integrated analyzer. For HgbA1c on HPLC analyzer, the minimum and maximum TAT for a batch of 50 samples was 3.78 and 160 min, respectively, while for the integrated system it was 28.2 and 35.1 min, respectively. Labor utilization for the 2 processes ranged from 3.21% to 3.75%.

CONCLUSION

Chemistry module on an integrated system can be used to determine the HgbA1c and other serum proteins.

[Application of benchmark dose on occupational epidemiology research in lead exposure]

OBJECTIVE

To use the data of occupational epidemiology to estimate the benchmark dose (BMD) of renal dysfunction induced by lead.

METHODS

Blood lead was considered as an exposure biomarker, while urinary total protein (TP), urinary beta(2)-microglobulin (beta(2)-MG) and urinary N-Acetyl-beta-D-glucosaminidase (NAG) were considered as effect biomarkers reflecting the damage of renal function. The dichotomized (binary) data was used as effect endpoints. The BMD and BMD lower limit (BMDL) of blood lead were estimated at the 10% benchmark response using BMDS version 1.3.1.

RESULTS

There was an increased prevalence of hyper-TP-uria, hyper-beta(2)-MG-uria and hyper-NAG-uria with an increasing blood lead concentration. There was obviously dose-response relationship between blood lead and TP, beta(2)-MG and NAG, respectively. The BMD and BMDL of blood lead affecting renal function were estimated to be 323.6 - 754.3 microg/L and 274.2 - 541.5 microg/L. The BMDL of blood lead was ranged from low to high as NAG, TP and beta(2)-MG. The urinary NAG activity might be served as a sensitive biomarker in detecting early renal dysfunction.

CONCLUSION

It should be feasible to use the BMD approach to set up the reference dose (RfD) and reference concentration (RfC). BMD approach might provide a new and better way for setting up the RfD/RfC.

Detection of outliers in reference distributions: performance of Horn's algorithm

BACKGROUND

Medical laboratory reference data may be contaminated with outliers that should be eliminated before estimation of the reference interval. A statistical test for outliers has been proposed by Paul S. Horn and coworkers (Clin Chem 2001;47:2137-45). The algorithm operates in 2 steps: (a) mathematically transform the original data to approximate a gaussian distribution; and (b) establish detection limits (Tukey fences) based on the central part of the transformed distribution.

METHODS

We studied the specificity of Horn's test algorithm (probability of false detection of outliers), using Monte Carlo computer simulations performed on 13 types of probability distributions covering a wide range of positive and negative skewness. Distributions with 3% of the original observations replaced by random outliers were used to also examine the sensitivity of the test (probability of detection of true outliers). Three data transformations were used: the Box and Cox function (used in the original Horn's test), the Manly exponential function, and the John and Draper modulus function.

RESULTS

For many of the probability distributions, the specificity of Horn's algorithm was rather poor compared with the theoretical expectation. The cause for such poor performance was at least partially related to remaining nongaussian kurtosis (peakedness). The sensitivity showed great variation, dependent on both the type of underlying distribution and the location of the outliers (upper and/or lower tail).

CONCLUSION

Although Horn's algorithm undoubtedly is an improvement compared with older methods for outlier detection, reliable statistical identification of outliers in reference data remains a challenge.

Preanalytical variability in laboratory testing: influence of the blood drawing technique

The predominant technique used to draw blood for laboratory testing is a conventional straight needle attached to an evacuated tube system. However, alternative tools might be advantageous in exceptional circumstances. The use of butterfly devices has been traditionally discouraged for reasons of costs and due to the high risk of obtaining unsuitable samples, but there is no convincing evidence to support the latter indication. The purpose of this study was to compare results of hematological and clinical chemistry testing, after drawing blood into evacuated tubes, employing either a traditional 21-gauge straight needle or a 21-gauge butterfly device with 300-mm-grade polyvinyl chloride tubing. Blood samples and complete sets of data were successfully obtained for 30 consecutive outpatients. Of the 43 hematological and clinical chemistry parameters measured, means for paired samples collected by the two alternative drawing techniques did not differ significantly, except for serum sodium, white blood cells and platelets counts. Bland-Altman plots and limits-of-agreement analysis showed mean bias of between -7.2% and 1.7% and relative coefficients of variation ranging from 0.2% to 21.2%. The 95% agreement interval in the set of differences was acceptable and was mostly within the current analytical quality specifications for desirable bias. The rate of hemolysis in plasma was not statistically different between the two collection techniques. Taken together, the results of the present investigation suggest that, when a proper technique is used and within certain limitations, the butterfly device may be a reliable alternative to the conventional straight needle to draw blood for laboratory testing.

A negative D-dimer assay does not rule out cerebral venous thrombosis: a series of seventy-three patients

BACKGROUND AND PURPOSE

Cerebral venous thrombosis (CVT) is an infrequent variety of cerebrovascular disease with a wide spectrum of clinical presentations and a notoriously difficult diagnosis. Previous reports have emphasized the potential clinical utility of D-dimer assay in CVT diagnosis.

METHODS

A rapid sensitive D-dimer assay was performed at entry in 73 patients with CVT <30 days duration, examined in our institution between 1999 and 2003.

RESULTS

The mean value of D-dimer levels was 1521 ng/mL; 7 patients (10% of all patients and 26% of those presenting with isolated headache) had values <500 ng/mL. In a multivariate analysis, isolated headache was the only variable associated with a negative D-dimer assay.

CONCLUSIONS

A negative D-Dimer assay does not confidently rule out CVT, particularly in the setting of recent isolated headache.

Evaluation of analytical methods and workflow performance of the Architect ci8200 integrated serum/plasma analyzer system

BACKGROUND

The Architect ci8200 is an integrated serum analyzer for photometric, electrochemical and immunological assays. Several assays of each category and the workflow performance of the system were compared with established laboratory procedures in two laboratories.

METHODS

Measurements were compared with the ELECSYS 2010 (Roche Diagnostics) for CEA, PSA, FPSA, AFP, folate, vitamin B12, with the CENTAUR (Bayer) for TSH, T4, FT4, FSH and Estradiol, with the LIAISON (DiaSorin) for TSH, FT4 and FT3, with the Behring Nephelometer BN II (Dade-Behring) for ferritin, and with the INTEGRA 800 (Roche Diagnostics), and the AU640 (Olympus) for clinical chemistry assays. Workflow studies were performed to compare times of analysis required for defined analytical workloads.

RESULTS

The coefficients of variation (CVs) for within-run imprecision were between 3% and 6% for CEA, PSA, FPSA, AFP and ferritin, and between 3% and 11% for TSH, FT4, FT3, folate and vitamin B12. The CVs for day-to-day imprecision for immunoassays were between 3% and 10%, except for vitamin B12 (CVs 11-13%) and FT4 (CV 10% -13%). For clinical chemistry tests corresponding CVs for within-run imprecision were < 1%, except for HDL, triglyceride, creatinine, ALT, LD and lipase (CVs<2%) and bicarbonate (CV 3%-6%) and magnesium (CV < 3%). The CVs for day-to-day imprecision for clinical chemistry tests were < 1%, except for sodium, CO(2), magnesium, phosphorus, glucose, uric acid, HDL, triglyceride, ALT, AST CK, lipase with CVs < 6% and for CO(2)<11%. Dilutional linearity testing of seven immunoassays and five clinical chemistry analytes resulted in recovery rates of 90-110%. Correlation studies with 15 immunoassays and 25 clinical chemistry tests showed acceptable agreements with established methods. Work flow analyses demonstrated a net gain in time of analysis up to 109 min depending on the size of the sample batch analyzed with the Architect ci8200 as the main analyzer as compared to the currently installed routine laboratory equipment. Median turn-around times were 7 and 30 min for chemistry assays and immunoassays, respectively, when ordered as STAT analyses, and 18 min when chemistry assays were ordered as routine determinations.

CONCLUSIONS

Assays on the Architect ci8200 performed well, fulfilling quality control requirements as defined for instance by German quality control guidelines (RiliBAK). Method comparisons showed acceptable agreements with established assays. Workflow studies using the Architect ci8200 documented shorter times of analyses as compared with the conventionally established laboratory routine demonstrating the potential of integrated chemistry/immunoassay analyzers to provide faster and more efficient performance.

[Non-specific reactions in four methods measuring (1-->3)-beta-D-glucan levels in plasma]

We detected non-specific reactions in the measurement of (1-->3)-beta-D-glucan levels (beta-glucan) in plasma, and the influences of the non-specific reactions on sensitivity and specificity of measurement methods were examined. In this study, 460 plasma samples from 174 patients at Kawasaki Medical School Hospital were used, and the plasma beta-glucan levels were measured by different four methods. The methods included the dilution-heating-endpoint (DHE), dilution-heating-turbidimetric (DHT), alkaline-kinetic (AK), and alkaline-endpoint method (AE) with and without 4-amidinophenyl benzoate hydrochloride (APB) of a protease inhibitor blocking the Limulus reaction. Non-specific reactions were detected from the calculated value under conditions with APB. Therefore, both of the actual values and the values equivalent to non-specific reactions were calculated. The incidence of non-specific reactions was 2.4% in DHE method, 0% in DHT, 53.3% in AK, and 99.3% in AE. The sensitivity and specificity in the methods were 35.7% and 96.0%, 28.6% and 96.0%, 78.6% and 80.1%, and 57.1% and 84.1%, respectively. When subtracted the non-specific reaction values from the actual values in AK and AE method, the specificity was increased by 91.4% and 94.0%, respectively. In these two methods, the non-specific reaction was considered to be a major cause of the low specificity. Finally, to measure plasma beta-glucan levels accurately, non-specific reactions should be excluded as possible by further improvement of measurement methods.

Biomimetic autoseparation of leukocytes from whole blood in a microfluidic device

Leukocytes comprise less than 1% of all blood cells. Enrichment of their number, starting from a sample of whole blood, is the required first step of many clinical and basic research assays. We created a microfluidic device that takes advantage of the intrinsic features of blood flow in the microcirculation, such as plasma skimming and leukocyte margination, to separate leukocytes directly from whole blood. It consists of a simple network of rectangular microchannels designed to enhance lateral migration of leukocytes and their subsequent extraction from the erythrocyte-depleted region near the sidewalls. A single pass through the device produces a 34-fold enrichment of the leukocyte-to-erythrocyte ratio. It operates on microliter samples of whole blood, provides positive, continuous flow selection of leukocytes, and requires neither preliminary labeling of cells nor input of energy (except for a small pressure gradient to support the flow of blood). This effortless, efficient, and inexpensive technology can be used as a lab-on-a-chip component for initial whole blood sample preparation. Its integration into microanalytical devices that require leukocyte enrichment will enable accelerated transition of these devices into the field for point-of-care clinical testing.

Requirements for improving quality in the measurement of bone markers

BACKGROUND

A number of biochemical assays readily detect molecules released from the bone matrix and collagen degradation in both serum and urine specimens. The variability of bone biochemical markers is still an issue of practical concern limiting their wider clinical use, and compromising the management of the individual patient. The aim of the present study was to define quality specifications of bone markers measurement in the total testing process.

METHODS

The different sources of biological variability of biochemical bone marker were assessed according to the current knowledge and categorized as pre-analytical, intra-analytical and post-analytical.

RESULTS

The identification of factors influencing bone marker measurements allowed recommendations to be made for decreasing pre- and intra-analytical variation in order to improve laboratory performance. The assessment of post-analytical sources of variation, moreover, involves the practical use of bone markers in identifying individual women at risk of fractures and the identification of non-responders, providing adequate limit value, least significant change value or optimal threshold of bone marker change.

CONCLUSIONS

Laboratory services may improve the potential clinical applications of biochemical markers of bone remodeling by improving method standardization, defining and reducing pre-analytical and analytical variables and providing reports that specify analytical and biological variation or cut-off values, thus facilitating data interpretation or reports.