Clinica Chimica Acta 446 (2015) 165–170

Contents lists available at ScienceDirect

Clinica Chimica Acta journal homepage: www.elsevier.com/locate/clinchim

Analytical performance evaluation of the i-STAT Total β-human chorionic gonadotropin immunoassay Aleksandra M. Sowder a,1, Melanie L. Yarbrough b,1, Robert D. Nerenz b, John V. Mitsios b, Rachel Mortensen c, Ann M. Gronowski b,d, David G. Grenache a,e,⁎ a

Department of Pathology, University of Utah School of Medicine, Salt Lake City, UT, USA Department of Pathology and Immunology, Washington University School of Medicine, St. Louis, MO, USA ARUP Laboratories, Salt Lake City, UT, USA d Department of Obstetrics and Gynecology, Washington University School of Medicine, St. Louis, MO, USA e ARUP Institute for Clinical and Experimental Pathology, Salt Lake City, UT, USA b c

a r t i c l e

i n f o

Article history: Received 26 October 2014 Received in revised form 8 April 2015 Accepted 10 April 2015 Available online 25 April 2015 Keywords: Human chorionic gonadotropin Quantitative Whole blood Point-of-care Pregnancy test

a b s t r a c t Background: The ability to perform quantitative hCG testing in whole blood at the point-of-care is desirable. The purpose of this study was to perform an analytical validation of the Abbott i-STAT Total β-hCG test. Methods: Whole blood, plasma, and serum samples were prepared by the addition of hCG and were used to evaluate precision, linearity, analytical sensitivity, accuracy, the high-dose hook effect, and dilution recovery. Results: Imprecision was highest with whole blood (CV = 16.0% and 6.7% at 10 and 1184 IU/l, respectively) and lowest in serum (CV = 8.1% and 4.3% at 11 and 1305 IU/l, respectively). The limits-of-quantitation were 8 and b 5 IU/l for whole blood and both plasma and serum, respectively. The assay was linear between 5 and 2000 IU/l in all sample types (R2 ≥ 0.998). i-STAT results agreed most closely with the Architect Total β-hCG assay and with greater differences observed with Beckman DxI Total βhCG and Roche Cobas e601 hCG+β assays (mean differences across all sample types were 9.3% and 12.3%, respectively). A high-dose hook effect was observed at concentrations N400,000 IU/l. Accuracy was achieved in samples diluted with serum but not saline. Conclusions: The i-STAT Total β-hCG test demonstrates acceptable performance for quantifying hCG in whole blood, plasma and serum. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Tests for human chorionic gonadotropin (hCG) are commonly performed in healthcare settings to identify pregnant patients in order to prevent interventions that could potentially harm a fetus. In certain clinical situations, a rapid assessment of pregnancy status is important to choose the proper clinical course of action. hCG is a glycoprotein hormone produced by placental trophoblast cells following implantation of a fertilized egg. As a biomarker for the diagnosis of early pregnancy, serum hCG is more sensitive than urine hCG [1]. While quantitative tests for hCG performed on serum are able to detect very low concentrations and provide the earliest possible detection of pregnancy, these tests are most often performed in a centralized laboratory. As such, the time required for specimen collection, transport, processing, and analysis contribute to what may be an unacceptably long turnaround time [2]. Alternatively, qualitative urine hCG tests can

⁎ Corresponding author at: ARUP Laboratories, Dept. 115, 500 Chipeta Way, Salt Lake City, UT 84108, USA. Tel.: +1 801 583 2787. E-mail address: [email protected] (D.G. Grenache). 1 Aleksandra M. Sowder and Melanie L. Yarbrough contributed equally to the work, and both should be considered as first authors.

http://dx.doi.org/10.1016/j.cca.2015.04.025 0009-8981/© 2015 Elsevier B.V. All rights reserved.

be performed at the point-of-care, provide rapid results, and are granted waived status by the Clinical Laboratory Improvement Amendments. However, they are less analytically sensitive than quantitative tests and are prone to producing false-negative results [3–5]. The ability to perform rapid, quantitative hCG testing using whole blood would provide the benefits of quantitative serum testing while avoiding the risks of false-negative results associated with qualitative urine tests. The i-STAT Total β-hCG cartridge (Abbott Diagnostics) is a quantitative hCG test to be used with whole blood or plasma. It is intended to be used for the early detection of pregnancy. The objective of this study was to perform an analytical validation of this test using whole blood, plasma, and serum. 2. Materials and methods 2.1. Assay According to the manufacturer, the i-STAT hCG cartridge is an immunometric test that utilizes 2 antibodies to quantify hCG. A solidphase capture antibody is bound to an electrochemical sensor and a signal antibody is conjugated to alkaline phosphatase. hCG in the sample combines with both antibodies during a seven-minute incubation

166

A.M. Sowder et al. / Clinica Chimica Acta 446 (2015) 165–170

followed by a wash step to remove unbound signal antibody. The wash fluid contains a substrate that is catalyzed by alkaline phosphatase to release an electrochemical product that is amperometrically measured by the sensor. Signal intensity is proportional to hCG concentration. Calibration verification materials are traceable to the WHO 5th International Standard (IS) for hCG. The cartridge contains reagents designed to minimize interferences caused by interfering antibodies. The assay utilizes the same capture and signal antibodies as the Architect Total β-hCG (Abbott Diagnostics) reagent platform. As such, it is expected that the i-STAT test will recognize the same hCG variants as the Architect assay (intact, nicked, free β, and nicked free β). A total of four different i-STAT instruments were used for this study. Two were used by investigators at the University of Utah and 2 were used by investigators at Washington University in St. Louis. 2.2. Study samples

result from the regressions with the best fit [6]. Based on the professional opinion of the authors, allowable nonlinearity was defined as 2 IU/l between 5 and 20 IU/l and 10% at concentrations N20 IU/l. 2.6. Validation of accuracy Accuracy was evaluated by adding hCG from clinical serum samples to 40 aliquots each of hCG-free whole blood and serum to create a set of samples to span the measuring range of the i-STAT hCG cartridge (5–2000 IU/l). The hCG concentration range in the samples tested was targeted to be 31–1885 IU/l (mean, 901 IU/l) using a random number generator. The samples were stored at 4–8 °C for up to 2 days before being analyzed with the i-STAT. Immediately after testing the whole blood, samples were centrifuged to obtain plasma and the plasma analyzed with the i-STAT. Residual aliquots of serum and plasma were stored at −80 °C for up to 31 days before being tested on each of 3 comparative methods (Architect Total β-hCG; DxI Total βhCG, Beckman Coulter, Inc.; and Cobas e601 hCG+β, Roche Diagnostics), all calibrated with material traceable to the 4th IS for hCG, in duplicate, on a single day.

Samples were prepared by the addition of hCG to hCG-free whole blood, plasma, and/or serum specimens collected from non-pregnant volunteers and stored at 4–8 °C (whole blood) or −80 °C (plasma and serum). Sources of hCG included de-identified, residual clinical serum and/or plasma specimens obtained from females that were sent to the laboratory for physician-ordered hCG testing. The hCG concentration in these samples was determined using the Architect Total β-hCG that was calibrated with material traceable to the 4th IS for hCG. Another hCG source was a commercial hCG preparation (Scripps Laboratories) re-suspended in 100 mmol/l Tris, 1 g/l bovine serum albumin, pH 8.0 to achieve a concentration of 2 × 107 IU/l and stored at −80 °C. This material is derived from human pregnancy urine and is ≥99% pure by SDSPAGE. Institutional review board approval was received for this study.

The high-dose hook effect was evaluated by adding commercially purified hCG to eight hCG-free serum samples to achieve hCG concentrations between 8000 and 2,000,000 IU/l. Each sample was tested in duplicate on the i-STAT on a single day. Five of these samples were also tested in duplicate with the Architect Total β-hCG assay. The upper limits of the assays' measuring ranges are 2000 and 15,000 IU/l for i-STAT and Architect, respectively.

2.3. Validation of precision

2.8. Evaluation of sample dilution

Precision was determined from 2 samples of whole blood, plasma, and serum to which hCG from clinical samples was added to obtain target concentrations of 10 and 1500 IU/l. Each sample was analyzed in duplicate, twice per day for 20 days using 2 different lots of hCG cartridges and 2 different i-STAT instruments. Samples were stored at 4–8 °C for the duration of the precision study (24–29 days). Based on the professional opinion of the authors, allowable total imprecision was defined as 2 IU/l between 5 and 20 IU/l and 10% at concentrations N 20 IU/l.

The effects of sample dilution were evaluated by adding commercially purified hCG to hCG-free serum or Architect Multi-assay Manual Diluent (phosphate buffered saline (PBS)) (Abbott Laboratories) to create a set of 20 samples with an hCG concentration range of 2252 to 929,759 IU/l and tested in duplicate on the Architect (11 samples with an hCG concentration expected to exceed the measuring range of the Architect (N15,000 IU/l) were manually diluted with PBS by a factor of 1:20 prior to analysis). All samples were then diluted by either a factor of 2 (N = 10), 10 (N = 15), 100 (N = 15), or 500 (N = 5) with either serum or PBS and tested in duplicate on the i-STAT and the Architect. Percent recovery for each sample was calculated by dividing the observed mean concentration upon dilution by the mean concentration of its corresponding undiluted sample as measured on the Architect. Based on the professional opinion of the authors, allowable error for recovery was defined as 90–110%.

2.4. Validation of analytical sensitivity Analytical sensitivity was determined from a set of 5 whole blood, plasma, and serum samples prepared by adding hCG from clinical samples to contain 0, 5, 10, 15, and 20 IU/l of hCG and each analyzed in 10 replicates on a single day. The limit-of-blank (LOB) was defined as the mean + 3 SD of the 0 IU/l sample. The limit-of-detection (LOD) was defined as LOB + 3 SD of the 5 IU/l sample. The limit-of-quantitation (LOQ) was determined by plotting the average hCG concentration versus its corresponding CV and performing nonlinear (second-order polynomial) regression to fit the data. The hCG concentration that yielded a CV of 20% was defined as the LOQ.

2.7. Evaluation of the high-dose hook effect

2.9. Statistical analysis Precision metrics were determined using StatisPro (Clinical and Laboratory Standards Institute). All other metrics were determined using Prism 5 (GraphPad Software, Inc.).

2.5. Validation of linearity

3. Results

Linearity was evaluated by adding hCG from clinical samples to hCGfree whole blood, plasma, and serum, to achieve a target concentration close to the i-STAT hCG test's upper measuring limit of 2000 IU/l. Samples were serially diluted to produce a set of six samples, that spanned the claimed analytical measuring range of 5–2000 IU/l and tested in duplicate on a single day. Linear and nonlinear (third-order polynomial) regression were used to define the relationship between expected and measured concentrations. Nonlinearity was calculated as the difference between the mean measured result and the predicted

3.1. Precision Precision data for whole blood, plasma, and serum at 2 concentrations are shown in Table 1. There was a significant degradation of hCG in the whole blood samples after 14 days of storage compared to the first 14 days (r = −1.0, p b 0.0001 vs. −0.58, p = 0.11, respectively). For this reason, precision in whole blood was determined from samples analyzed in duplicate, twice per day on 9 separate days over a 14-day period. Imprecision was greatest with whole blood (within-lab CV of

A.M. Sowder et al. / Clinica Chimica Acta 446 (2015) 165–170 Table 1 Precision at 2 concentrations of hCG in whole blood, plasma, and serum. Samples were analyzed in duplicate, twice per day for 9 (whole blood) or 20 (plasma and serum) days.

Mean hCG (IU/l) Allowable total SD (IU/l) Repeatabilitya SD (IU/l) Repeatability CV (%) Within-labb SD (IU/l) Within-lab CV (%)

Whole Blood

Plasma

10 2.0 1.5 14.8 1.6 16.0

12 2.0 0.7 6.1 1.0 8.3

1184 118.4 79.7 6.7 79.7 6.7

Serum 1274 127.4 53.7 4.2 61.0 4.8

11 2.0 0.8 7.2 0.9 8.1

1305 130.5 50.0 3.8 56.7 4.3

a The agreement between results of successive measurements of the same measurand and carried out under the same conditions of measurement [15]. b An intermediate measure of precision where measurements are obtained in the same laboratory but under some different operating conditions (e.g. time) [15].

16.0% and 6.7% at 10 and 1184 IU/l, respectively) and lowest with serum (within-lab CV of 8.1% and 4.3% at 11 and 1305 IU/l, respectively). Imprecision was less than the allowable total imprecision in all sample types at the concentrations tested within and across both reagent lots. 3.2. Analytical sensitivity Analytical sensitivity data was determined for whole blood, plasma, and serum. The LOB for all sample types was 0 IU/l (the output of the iSTAT was “0” for all LOB samples tested). The LODs were 3, 2, and 3 IU/l, respectively. The LOQs were 8, b1, and 4 IU/l, respectively. When extrapolated from the nonlinear regression curve produced by the plot of hCG concentration versus CV, the CVs at a concentration of 5 IU/l were 22.6%, 9.9%, and 17.0%, respectively. 3.3. Linearity Assay linearity in whole blood, plasma, and serum was evaluated by linear regression analysis which generated the following: y = 0.99(x) + 5.2, R2 = 0.99 (whole blood); y = 1.01(x) + 1.0, R2 = 0.99 (plasma); and y = 1.01(x) + 11.5, R2 = 0.99 (serum) (Fig. 1). Allowable nonlinearity was not exceeded for any sample type. 3.4. Accuracy The accuracy of the i-STAT hCG cartridge was evaluated in whole blood, plasma, and serum by comparison to hCG concentrations in plasma and serum determined by 3 quantitative methods. Difference plots are shown in Fig. 2. Across all methods, hCG measured by the i-STAT tended to produce results that were biased high for all sample types. The range of the mean differences was 2.1 to 13.1%. i-STAT hCG measurements were very similar to corresponding concentrations determined by the Architect Total β-hCG reagent platform (mean difference 3.1%, range 2.1–4.1%). Greater positive differences were observed with the DxI Total βhCG (mean difference 9.3%, range 7.4–10.6%) and the Cobas e601 hCG+β (mean difference 12.3%, range 11.7–13.1%) reagent platforms. The standard deviations of the relative differences were greatest with whole blood (~ 14%) followed by plasma (~ 9%) and serum (~5%). Correlation plots are shown in Supplemental Fig. 1.

167

Deming regression produced slopes that were close to 1.0 (range 0.902–1.138) and Spearman correlations were all 0.99 with all comparative methods but constant systematic errors were high as indicated by y-intercepts that were N12 IU/l or b16 IU/l with the exception of the comparison to the Architect method (y-intercept −2.9 IU/l). As expected, the y-intercepts decreased considerably (range − 5.4–3.5 IU/l) when regression analysis was restricted to samples with an hCG concentration b170 IU/l (N = 5). 3.5. High-dose hook effect In order to evaluate the susceptibility of the i-STAT hCG assay to exhibit a high dose hook effect, serum samples with hCG concentrations between 8000 and 2,000,000 IU/l were analyzed in duplicate. Samples with hCG concentrations ≤400,000 IU/l produced the expected result of N2000 IU/l on the i-STAT. Samples with a concentration ≥600,000 IU/l produced results between 617 and 1704 IU/l, indicating the occurrence of a high-dose hook effect at a concentration between 400,000 and 600,000 IU/l (Fig. 3). None of the 4 samples with an hCG concentration ≥600,000 produced a result of b5 IU/l therefore all samples were correctly interpreted as being qualitatively positive. The high-dose hook effect did not occur when samples with an hCG concentration range of 8000– 2,000,000 IU/l were measured using the Architect (Fig. 3). 3.6. Sample dilution In order to quantify samples with hCG concentrations N 2000 IU/l on the i-STAT, serum samples with hCG concentrations ranging from 2252 to 929,759 IU/l were diluted into hCG-negative serum or PBS and measured in duplicate on the i-STAT and the Architect. When compared to the hCG concentrations of the undiluted samples measured on the Architect, the mean recovery of hCG from samples diluted into serum was 93% (range 79–106%) on the i-STAT. This was not significantly different (p = 0.22) from the mean recovery of 96% on the Architect (range 76–117%) (Fig. 4) and was within allowable error for recovery. Conversely, samples added to PBS produced a mean recovery of 121% (range 87–151%) on the i-STAT which exceeded allowable error for recovery and was significantly greater (p b 0.0001) than the 93% mean recovery observed on the Architect (range 61–131%) and the 93% mean recovery of serum-diluted samples tested by the i-STAT (p b 0.0001). 4. Discussion The ability to use a blood sample for rapid, point-of-care, quantitative hCG testing is desirable, as such a method could provide the benefits of testing a whole blood or serum matrix without the risks of falsenegative results associated with testing urine [4,5,7,8]. False-negative urine hCG test results may occur if the hCG concentration is below the detection thresholds (~10–25 IU/l) of qualitative devices or due to the high-dose or variant hook effects [5,7,8]. Some authors have described the use of whole blood instead of urine or serum for the qualitative detection of hCG [9,10], an inappropriate practice that is fraught with difficulties due to analytical and regulatory matters [11]. The i-STAT Total

Fig. 1. Linearity of the i-STAT β-hCG assay in whole blood, plasma, and serum.

168

A.M. Sowder et al. / Clinica Chimica Acta 446 (2015) 165–170

Fig. 2. Difference plots of i-STAT β-hCG test results in whole blood, plasma, and serum relative to plasma and serum measurements in 3 comparative methods. The dashed line represents the mean difference.

β-hCG test is designed to rapidly quantify hCG in whole blood or plasma and is intended to be used for the early detection of pregnancy. This study describes the analytical performance of the test in whole blood, plasma, and serum. Overall, the i-STAT hCG cartridge demonstrated acceptable performance characteristics for quantifying hCG in whole blood, plasma, and

serum. Imprecision was greatest in whole blood compared to plasma or serum but was consistently less than the allowable total imprecision in all sample types at the concentrations tested. Likewise, the analytical sensitivity of a whole blood matrix was observed to be greater than that of plasma or serum. While the LODs for all 3 sample types were below the lower limit of the measuring range of the assay (5 IU/l), the LOQ

Fig. 3. The high-dose hook effect was observed with the i-STAT hCG assay at a concentration between 400,000 and 600,000 IU/l but not observed with the Architect Total β-hCG reagent platform. hCG concentrations in samples tested by i-STAT and Architect were created to be 8000, 40,000, 200,000, 400,000, 600,000, 800,000, 1,000,000, and 2,000,000 IU/l and 8000, 40,000, 200,000, 1,000,000, and 2,000,000, respectively.

A.M. Sowder et al. / Clinica Chimica Acta 446 (2015) 165–170

169

Fig. 4. Mean percent recovery of hCG in serum samples diluted with either hCG-negative serum or Architect diluent (phosphate buffered saline). The error bars represent the standard deviation. Mean recovery was significantly greater with saline-diluted samples tested on the i-STAT compared to either serum-diluted samples tested by the i-STAT and to saline- and serum-diluted samples tested by the Architect (p b 0.0001). N equals the number of samples included. Note that due to the i-STAT's smaller measuring range, percent recovery could not be calculated for specimens that produced a result of “N2000 IU/l.”.

for whole blood was 8.0 IU/l which was higher than either plasma or serum at b1 and 4 IU/l, respectively. The expected serum hCG concentration in a non-pregnant female is generally accepted to be ≤5 IU/l. Although the LOQ was 8.0 IU/l for whole blood, the CV at a concentration of 5 IU/l was 22.6% and the LOD was 3 IU/l. Taken together, we believe that the i-STAT hCG test is capable of clinically acceptable performance at the frequently used cutoff of 5 IU/l. As compared to the 3 laboratory hCG assays, results from the i-STAT hCG cartridge demonstrated the most agreement with the Architect Total β-hCG reagent platform. This was expected because the i-STAT and Architect assays utilize the same capture and signal antibodies and therefore recognize identical epitopes on hCG molecules. As such, it is expected that the i-STAT assay would detect the same hCG variants as the Architect Total β-hCG reagent platform (all clinically relevant variants except the β core fragment) [12]. It should be noted that the i-STAT and Architect assays use calibrator materials that are traceable to different IS (WHO 4th and 5th IS for Architect and i-STAT, respectively). One study has demonstrated a ~ 20% positive bias in a quantitative serum hCG assay that is traceable to the 5th IS compared to assays traceable to the 4th IS [13]. We speculate that any possible bias between i-STAT and Architect hCG results is minimized by the i-STAT's implementation of an adjustment factor that improves agreement with the Architect assay. Also expected were the greater differences between i-STAT and both the DxI Total βhCG and Cobas e601 hCG+β results, this is likely due to the differences in antibody pairs that each assay utilizes and/or differences in the traceability of each assay to different IS for hCG (after this work was completed, the calibrator material used in the DxI assay was modified to one that is traceable to the 5th IS and, as such, the bias reported here likely will not represent the current performance). The slopes from Deming regression analysis were within 14% of 1.0 for all sample types but the y-intercepts were unexpectedly high with the exception of the comparison of serum results between the i-STAT and Architect assays and is likely due to the lack of assay harmonization. By comparing the difference plots shown in Fig. 2 to the correlation plots shown in Supplemental Fig. 1, it can be appreciated that the yintercepts are furthest from zero with comparisons that show the greatest proportional systematic error. However, as expected, these decreased considerably when analysis was restricted to samples that had an hCG concentration b 170 IU/l. The product insert for the i-STAT hCG assay indicates that the highdose hook effect does not occur in samples with an hCG concentration

up to 300,000 IU/l. Our data support that claim. No hook effect was observed at an hCG concentration ≤400,000 IU/l but was observed in samples with concentrations ≥ 600,000 IU/l. In contrast, a study that investigated the i-STAT hCG for the hook effect found that it occurred in a sample with an hCG concentration as low as 218,712 IU/l [14]. Although the i-STAT hCG assay is susceptible to the hook effect, it is important to note that the qualitative interpretations of the quantitative results (up to 2,000,000 IU/l, the highest concentration evaluated in this study) were unaffected and were still identified as “positive” by the instrument. This suggests that the i-STAT can still be used as a qualitative test well outside of the assay's measuring range. The upper limit of the i-STAT hCG analytical measuring range is 2000 IU/l. While this is similar to the measuring ranges of many laboratory hCG reagent platforms, laboratory platforms employ automatic instrument-performed dilution protocols to provide an absolute quantitative result, something that the i-STAT is unable to do. A result of N2000 IU/l from the i-STAT may be insufficiently informative when an absolute hCG concentration is clinically necessary. Notably, the concentration of hCG in blood is likely to exceed 2000 IU/l by 4 to 6 weeks of gestation as measured from the date of the last menstrual period. Recognizing that some healthcare delivery environments may lack immediate access to a centralized laboratory to perform quantitative hCG testing, we sought to evaluate the accuracy of i-STAT hCG measurements with diluted serum samples. The data support the use of hCGnegative serum rather than PBS as a diluent. hCG recovery in serum was not significantly different with the i-STAT and Architect hCG assays. Conversely, when PBS was used as a diluent there was significant overrecovery of hCG by the i-STAT compared to the Architect. These data support the use of hCG-negative serum for diluting serum samples with an hCG concentration N 2000 IU/l prior to being analyzed by the i-STAT. Although plasma samples were not evaluated for dilution recovery, it is unlikely that they would perform differently than serum. We would not advocate for the dilution of whole blood samples prior to testing on the i-STAT. In conclusion, the Abbott i-STAT Total β-hCG assay can be used to accurately quantify hCG in whole blood, plasma, and serum. Falsely decreased results due to the high-dose hook effect are possible when the hCG concentration is N400,000 IU/l. If needed, serum or plasma samples with an hCG concentration that exceed 2000 IU/l can be diluted with hCG-free serum and reanalyzed on the i-STAT. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.cca.2015.04.025.

170

A.M. Sowder et al. / Clinica Chimica Acta 446 (2015) 165–170

Acknowledgments This research was supported by a grant from Abbott Point of Care, Inc. References [1] Lohstroh PN, Overstreet JW, Stewart DR, et al. Secretion and excretion of human chorionic gonadotropin during early pregnancy. Fertil Steril 2005;83:1000–11. [2] Furtado LV, Lehman CM, Thompson C, Grenache DG. Should the qualitative serum pregnancy test be considered obsolete? Am J Clin Pathol 2012;137:194–202. [3] Cervinski MA, Lockwood CM, Ferguson AM, et al. Qualitative point-of-care and overthe-counter urine hCG devices differentially detect the hCG variants of early pregnancy. Clin Chim Acta 2009;406:81–5. [4] Greene DN, Schmidt RL, Kamer SM, Grenache DG, Hoke C, Lorey TS. Limitations in qualitative point of care hCG tests for detecting early pregnancy. Clin Chim Acta 2013;415:317–21. [5] Gronowski AM, Cervinski MA, Stenman U-H, Woodworth A, Ashby L, Scott MG. False-negative results in point-of-care qualitative human chorionic gonadotropin (hCG) devices due to excess hCGbeta core fragment. Clin Chem 2009;55:1389–94. [6] CLSI. Evaluation of the linearity of quantitative measurement procedures: a statistical approach; approved guideline. Clinical and Laboratory Standards Institute; 2003.

[7] Hunter CL, Ladde J. Molar pregnancy with false negative β-hCG urine in the emergency department. West J Emerg Med 2011;12:213–5. [8] Griffey RT, Trent CJ, Bavolek RA, Keeperman JB, Sampson C, Poirier RF. “Hook-like effect” causes false-negative point-of-care urine pregnancy testing in emergency patients. J Emerg Med 2013;44:155–60. [9] Habboushe JP, Walker G. Novel use of a urine pregnancy test using whole blood. Am J Emerg Med 2011;29 [840.e3–4]. [10] Fromm C, Likourezos A, Haines L, Khan ANGA, Williams J, Berezow J. Substituting whole blood for urine in a bedside pregnancy test. J Emerg Med 2012;43:478–82. [11] Grenache DG, Gronowski AM, Fantz CR. Inappropriate use of qualitative, point-ofcare urine human chorionic gonadotropin test. Am J Emerg Med 2013;31:992–3. [12] Whittington J, Fantz CR, Gronowski AM, et al. The analytical specificity of human chorionic gonadotropin assays determined using WHO International Reference Reagents. Clin Chim Acta 2010;411:81–5. [13] Greene DN, Petrie MS, Pyle AL, Kamer SM, Grenache DG. Performance characteristics of the Beckman Coulter total βhCG (5th IS) assay. Clin Chim Acta 2015;439:61–7. [14] Wilgen U, Pretorius CJ, Gous RS, Martin C, Hale VJ, Ungerer JPJ. Hook effect in Abbott i-STAT β-human chorionic gonadotropin (β-hCG) point of care assay. Clin Biochem 2014;47:1320–2. [15] CLSI. Evaluation of precision performance of quantitative measurement methods; approved guideline. Second ed. Clinical and Laboratory Standards Institute; 2004.

Analytical performance evaluation of the i-STAT Total β-human chorionic gonadotropin immunoassay.

The ability to perform quantitative hCG testing in whole blood at the point-of-care is desirable. The purpose of this study was to perform an analytic...
1MB Sizes 1 Downloads 10 Views