CLB-08711; No. of pages: 3; 4C: Clinical Biochemistry xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Clinical Biochemistry journal homepage: www.elsevier.com/locate/clinbiochem
1
Short Communication
4Q1
Urs Wilgen a,b,⁎, Carel J. Pretorius a,b, Rehna S. Gous a, Cameron Martin a, Vincent J. Hale a, Jacobus P.J. Ungerer a
O
F
3
Hook effect in Abbott i-STAT β-human chorionic gonadotropin (β-hCG) point of care assay☆
5 6
a
7
a r t i c l e
8 9 10 11 12
Article history: Received 7 April 2014 Received in revised form 30 April 2014 Accepted 5 May 2014 Available online xxxx
13 14 15 16 17
Keywords: Human chorionic gonadotropin (β-hCG) Point-of-care testing Hook effect Gestational trophoblastic disease
i n f o
R O
Pathology Queensland, Department of Chemical Pathology, Royal Brisbane and Women's Hospital, Herston Road, Herston 4029, Queensland, Australia University of Queensland, School of Medicine, Brisbane, St. Lucia, 4072 Queensland, Australia
a b s t r a c t
D
P
Point-of-care testing for β-hCG has been widely advocated to allow rapid diagnosis/exclusion of pregnancy in the emergency department. A quantitative blood β-hCG assay has the additional benefit of being able to monitor the viability of pregnancy, using serial measurements, to determine the appropriate expected increase in β-hCG levels over time (e.g. ectopic pregnancy), and aiding in determining if an intrauterine gestational sac should be visible on sonographic imaging. Objectives: Evaluation of the newly released Abbott i-STAT β-hCG point-of-care assay with the Beckman Coulter β-hCG laboratory assay in use. Design and methods: Whole blood, plasma and serum samples with a wide range of β-hCG concentrations were analysed by both methods. Results: The Abbott I-STAT β-hCG compares favourably, can be performed on heparinised whole blood, plasma and serum, and shows acceptable accuracy and precision. However a hook effect at elevated β-hCG was shown in gestational trophoblastic disease as well as normal pregnancies. Conclusions: The i-STAT β-hCG performs acceptably in its intended use in the early detection of pregnancy, but results should always be interpreted within the clinical context, as a hook effect may occur. © 2014 The Canadian Society of Clinical Chemists. Published by Elsevier Inc.
E
C
T
b
E
2
42 43 44 45 46 47 48 49 50
R
N C O
40 41
Human chorionic gonadotropin (β-hCG) is a glycoprotein hormone produced predominantly in syncytiotrophoblasts, and is commonly used to confirm/exclude the diagnosis of pregnancy. HCG can be measured in urine or serum by immunoassay giving either a qualitative result (using a cut-off of N25 IU/L), or quantitative result. Concentrations of β-hCG in non-pregnant women are b5 IU/L, and rise rapidly during the first weeks of pregnancy, approximately doubling every two days, and reaching a peak of 100,000–500,000 IU/L at 8–10 weeks of gestation. Samples with β-hCG results b5 IU/L are interpreted as “negative”, 5–25 IU/L as “indeterminant” with a suggestion to repeat the test 2 days later, and N 25 IU/L as “positive” [1]. Testing for β-hCG at the point-of-care (POC) has been advocated to facilitate rapid diagnosis and decision making in the emergency department [2], but many of
U
38 39
R
Background
☆ Declarations: Abbott Diagnostics provided the β-hCG test cartridges for the evaluation free of charge. ⁎ Corresponding author at: Pathology Queensland, Department of Chemical Pathology, Royal Brisbane and Women's Hospital, Herston Road, Herston 4029, Queensland, Australia. Fax: +61 7 3646 3417. E-mail address: [email protected] (U. Wilgen).
18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 36 34 33 35
the qualitative hCG devices used at the POC are susceptible to the hook effect giving false negative results [3]. Apart from excluding/confirming the diagnosis of pregnancy with a qualitative β-hCG test, a quantitative β-hCG assay can prove useful for monitoring the viability of pregnancy, using serial measurements to determine the appropriate expected increase in β-hCG levels over time (e.g. ectopic pregnancy), and aiding in determining if an intrauterine gestational sac should be visible on sonographic imaging [4]. Abbott diagnostics recently released a β-hCG test cartridge to allow quantitative β-hCG testing at the POC on the i-STAT analyser. The test cartridge uses a two-site enzyme linked immunosorbent assay (ELISA) method for use with heparinised whole blood or plasma samples. The i-STAT assay reports quantitative β-hCG values from 5 to 2000 IU/L. Samples with concentrations less than 5 IU/L are reported as “b5 IU/L”, and exceeding 2000 IU/L are reported as “N 2000 IU/L” according to the manufacturer's package insert. Our statewide pathology service network utilises 270 Abbott i-STAT analysers, with 25 in small laboratory settings operated by laboratory personnel, and the remainder in non-laboratory POC settings (e.g. emergency departments) operated by medical and nursing staff. Serum is the preferred sample type in our laboratories for immunoassays. The i-STAT hCG assay is approved for use in Europe, New Zealand and Australia, and pending FDA approval for the USA and Canada. We evaluated the method for possible use in our pathology service.
http://dx.doi.org/10.1016/j.clinbiochem.2014.05.005 0009-9120/© 2014 The Canadian Society of Clinical Chemists. Published by Elsevier Inc.
Please cite this article as: Wilgen U, et al, Hook effect in Abbott i-STAT β-human chorionic gonadotropin (β-hCG) point of care assay, Clin Biochem (2014), http://dx.doi.org/10.1016/j.clinbiochem.2014.05.005
37 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74
U. Wilgen et al. / Clinical Biochemistry xxx (2014) xxx–xxx
2000
1000
500
1000
1500
2000
Beckman Coulter DxI800 hCG (IU/L) Fig. 1. Passing–Bablok regression of samples analysed for β-hCG on the Beckman Coulter DxI800 on serum and Abbott i-STAT on heparinised whole blood (open circles) (95% CI for slope and intercept: 0.89 to 0.97 and −1.05 to 4.31 respectively; n = 40) and serum (closed circles) (95% CI for slope and intercept: 1.04 to 1.12 and −0.87 to 4.50 respectively; n = 40).
114
R O
0
O
DxI800 serum vs iSTAT serum DxI800 serum vs iSTAT whole blood Passing Bablok: iSTAT=1.07 DxI800+1.66 Passing Bablok: iSTAT=0.92 DxI800+1.94
0
Results
F
500
93 94
P
iSTAT hCG (IU/L)
1500
β-hCG results up to ~2000 IU/L (measured by the DxI800) were selected at this stage due to the manufacturer's stated reportable range. Thereafter 16 additional serum samples with elevated β-hCG received for routine analysis were retrieved and analysed on the i-STAT. These samples were stored at 4 °C until analysis, which was usually the same day, but no longer than 24 h later. Three levels of the manufacturer's quality control materials were run to ensure the batch of Abbott i-STAT β-hCG cartridges performed within specifications. Accuracy and precision evaluation was performed analysing 3 levels of quality control material (24.7 IU/L; 929.5 IU/L; 1419.8 IU/L) ten times each. A series of dilutions using a sample with a β-hCG of 520,000 IU/L was performed on both methods using Beckman Coulter wash buffer (cat no. 8547197) as diluent. This is as per manufacturer's instruction for the DxI800. Diluting with β-hCG negative male serum demonstrated insignificant differences to wash buffer. The DxI800 will routinely dilute samples with β-hCG concentrations N1000 IU/L using a 1 in 200 dilution using wash buffer. Samples with concentrations exceeding 200,000 IU/L require a manual dilution. Abbott gives no instruction for dilutions, as the assay is intended to be as simple as possible for use at the point-of-care.
Analysis of 40 paired whole blood and serum samples on the i-STAT and DxI800 respectively showed excellent correlation using Passing Bablok regression (i-STAT = 0.92 ∗ DxI800 + 1.94, r2 = 0.994). Correlation between the DxI800 and i-STAT β-hCG when 40 serum samples were analysed by both methods was similar (i-STAT = 1.07 ∗ DxI800 + 1.66, r2 = 0.992) (see Fig. 1). Concordant interpretation would be achieved in all but 2 samples (DxI800 = 4.5 IU/L, i-STAT = 5.5 IU/L and DxI800 = 22.0 IU/L, i-STAT = 26.5 IU/L). Accuracy and precision are acceptable with %CVs of less than 7% across all 3 levels of quality control materials (Level 1 (24.7 IU/L): %CV 4.7%; Level 2 (929.5 IU/L): %CV 2.0%; Level 3 (1419.8 IU/L): %CV 6.6%). Five serum samples with elevated concentrations of β-hCG (measured on the DxI800) showed erroneously low values on the i-STAT, where numerical values were displayed below 2000 IU/L when the result should have been reported as “N 2000 IU/L”. This occurred when the β-hCG levels measured on the DxI800 exceeded 260,000 IU/L, demonstrating a hook effect on the i-STAT β-hCG, (see Table 1). The hook effect was present in samples from normal pregnancies as well as gestational trophoblastic disease. Dilutions of a serum sample with a measured β-hCG (on the DxI800) of 520,000 IU/L gave erroneously low results reported on the i-STAT up to a diluted concentration of 218,000 IU/L (see Table 2).
D
2
Design and methods
76 77
92
Human chorionic gonadotropin levels were measured on the Beckman Coulter UniCel DxI800 (Beckman Coulter Diagnostics, Brae CA USA) and an Abbott i-STAT POC Analyser (Abbott Diagnostics, IL USA). Forty samples (heparinised whole blood and serum) were collected in addition to routine bloods at visits to the antenatal clinic of the Royal Brisbane Women's Hospital. These samples were from pregnant women ≤6 week gestation. Whole blood and serum samples were collected in Becton Dickinson heparinised and serum separator tubes respectively. Whole blood samples were analysed on the i-STAT immediately upon receipt in the laboratory, and serum samples were allowed to clot for approximately 30 min, centrifuged and then analysed on both the DxI800 and the i-STAT. Our institutional ethics committee gave approval for the study, and written consent was obtained. A further 40 serum samples received in the laboratory for routine analysis and measured on the DxI800 were selected and retrieved from sample storage and analysed on the i-STAT. Only samples with
t1:1 t1:2 t1:3
Table 1 Patient serum samples with markedly elevated β-hCG performed on the Beckman Coulter DxI800 and Abbott i-STAT. The i-STAT should display numerical values exceeding 2000 IU/L as “N2000 IU/L”. Values exceeding 1000 IU/L will automatically be run in dilution on the Beckman Coulter DxI800.
90 91
T
C
E
R
88 89
R
86 87
O
84 85
C
82 83
t1:4
Patient sample
t1:5 t1:6 t1:7 t1:8 t1:9 t1:10 t1:11 t1:12 t1:13 t1:14 t1:15 t1:16 t1:17 t1:18 t1:19 t1:20
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
N
80 81 Q2
Clinical diagnosis
U
78 79
E
75
Ectopic pregnancy Ectopic pregnancy Ectopic pregnancy Miscarriage Miscarriage 1st Trimester bleed Hyperemesis gravidarum Abdominal pain in pregnancy Hyperemesis gravidarum Hyperemesis gravidarum 1st Trimester bleed 1st Trimester bleed Gestational trophoblastic disease Hyperemesis gravidarum Gestational trophoblastic disease Gestational trophoblastic disease
Gestational age (weeks)
Beckman coulter DxI800 (IU/L)
Abbott i-STAT (IU/L)
Interpretation of reported i-STAT result
7 6 6 6 8 9 10 10 12 14 10 8 – 13 – –
2846 18,000 21,194 22,826 140,229 210,000 240,000 250,000 260,000 290,000 310,000 320,000 360,000 390,000 520,000 1,200,000
N2000 N2000 N2000 N2000 N2000 N2000 N2000 N2000 N2000 1816 N2000 N2000 1421 1805 1195 717
Correct Correct Correct Correct Correct Correct Correct Correct Correct Hook effect Correct Correct Hook effect Hook effect Hook effect Hook effect
Please cite this article as: Wilgen U, et al, Hook effect in Abbott i-STAT β-human chorionic gonadotropin (β-hCG) point of care assay, Clin Biochem (2014), http://dx.doi.org/10.1016/j.clinbiochem.2014.05.005
95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113
115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136
U. Wilgen et al. / Clinical Biochemistry xxx (2014) xxx–xxx
Neat 1:1 2:3 1:4 1:150 1:200 1:250
520,000 277,265 218,712 117,000 4514 3344 2705
1195 1187 1702 N2000 N2000 N2000 N2000
Hook effect Hook effect Hook effect Correct Correct Correct Correct
Discussion
138 139
Human chorionic gonadotropin (hCG) is a heterodimeric glycoprotein hormone consisting of two α subunits and two β subunits with eight carbohydrate side chains, which results in considerable heterogeneity of the structure of hCG. The hCG molecule is produced by trophoblasts in the placenta, gestational trophoblastic disease and some other malignancies [5]. Immunoassays measure the quantity of β-hCG by means of a sandwich assay which includes a capture antibody and a signal antibody bound to different antigenic sites on the captured antigen molecule. Different β-hCG assays use antibodies to different sites on the β-hCG molecule, and therefore may measure very different combinations of β-hCG related molecules. Most immunoassays are designed to detect primarily the regular form of β-hCG produced by the placenta [6], but standardisation of immunoassays for heterogeneous antigens such as β-hCG is notoriously problematic [7]. Not only does hCG exist in numerous forms (intact, free α, free β, nicked, β-core fragment, and varying degrees of glycosylation) [8], but also the composition changes as pregnancy progresses, and different isoforms may be produced by neoplastic tissues [9]. According to the manufacturer “the assay is intended for use in the early detection of pregnancy only”. The principal reason for testing for β-hCG is to confirm/exclude pregnancy. Other indications include determination of pregnancy viability (e.g. ectopic pregnancy, miscarriage, foetal demise), where serial quantitative values are required, and to assist in the diagnosis and management of gestational trophoblastic disease. A quantitative POC β-hCG test is helpful in guiding decision making in the emergency department. In a woman of child bearing age with lower abdominal pain and bleeding with a positive pregnancy test the diagnosis of an ectopic pregnancy should always be borne in mind. Visualisation of an intrauterine pregnancy by sonographic examination is dependant on gestational age, and dating according to last menstrual period is notoriously unreliable. Studies have shown that an intrauterine gestational sac should be visible on endovaginal ultrasound when serum β-hCG levels exceed 1500 IU/L. Conversely, the absence of an intrauterine gestational sac would make an ectopic pregnancy more likely [4]. The i-STAT β-hCG assay will perform satisfactorily within its intended use in the early detection of pregnancy, but there may be instances in clinical practice where an erroneously low value is reported due to a hook effect. Numerous case reports of false negative urine, serum and both urine and serum β-hCG results in patients with gestational trophoblastic disease have been published [10–13]. The hook effect is an uncommon phenomenon seen in sandwich immunoassays when the antigen (in this case β-hCG) is present in sufficiently high concentrations to cause saturation of both the capture and signal antibodies, preventing
154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182
C
152 153
E
150 151
R
148 149
R
146 147
N C O
144 145
U
142 143
Conclusions
The Abbott i-STAT β-hCG assay used at the point of care shows good correlation with the laboratory method used. We have shown that in addition to heparinised whole blood and plasma, serum is an acceptable sample type. We have also demonstrated that erroneously low β-hCG results may be obtained on the i-STAT when concentrations approach 218,000 IU/L, and results within the reportable range (b 2000 IU/L) may be reported in samples with much higher concentrations due to a hook effect. In instances where the i-STAT values do not correlate with the clinical picture, the sample should be sent to the laboratory for analysis, and results should be interpreted with caution if gestational trophoblastic disease is suspected.
T
137
F
t2:5 t2:6 t2:7 t2:8 t2:9 t2:10 t2:11
O
Interpretation of reported i-STAT result
R O
Abbott i-STAT (IU/L)
P
Beckman Coulter DxI800 (IU/L)
D
t2:4
Dilution (serum:diluent)
140 141
the formation of a “sandwich”. As a result the signal generated is less than what would be expected at that concentration, and the analyser reports an erroneously low value [14]. The Abbott i-STAT β-hCG product insert states that “no significant hook effect was detected in samples up to 300,000 IU/L”. We have shown that samples exceeding 218,000 IU/L demonstrated a hook effect, and that values as low as 717 IU/L may be reported, which is well below the stated upper limit of reporting of 2000 IU/L, with actual concentrations exceeding 1,200,000 IU/L. Grossly haemolysed specimens are known to decrease β-hCG detection. None of the samples demonstrating a hook effect were haemolysed. In instances where the i-STAT values do not correlate with the clinical picture the sample should be sent to the laboratory, where the analysis of the sample in dilution will usually yield the correct result. Furthermore, reported β-hCG concentrations should be interpreted with caution, especially if a clinical suspicion of gestational trophoblastic disease exists.
Table 2 Dilutions of a serum sample with a markedly elevated β-hCG analysed on the DxI800 and i-STAT.
E
t2:1 t2:2 t2:3
3
183 184 Q3 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209
References
210
[1] Ashwood ER, Knight GJ. Clinical chemistry of pregnancy. In: Burtis CA, Ashwood ER, Bruns DE, editors. Tietz textbook of clinical chemistry and molecular diagnostics. 4th ed. St. Louis: Saunders; 2006. p. 2153–206. [2] Mishalani SH, Seliktak J, Braunstain GD. Four rapid serum-urine combination assays of chorionic gonadotropin (hCG) compared and assessed for their utility in quantitative determinations of hCG. Clin Chem 1994;40:1944–9. [3] Nerenz RD, Haowei Song, Gronowski AM. Screening to evaluate point-of-care human chorionic gonadotropin (hCG) devices for susceptibility to the hook effect by hCG β core fragment: evaluation of 11 devices. Clin Chem 2014;60:667–74. [4] Kohn MA, Kerr K, Malkevich D, O'Niel N, Kerr MJ, Kaplan BC. Beta-human chorionic gonadotropin levels and the likelihood of ectopic pregnancy in emergency department patients with abdominal pain or vaginal bleeding. Acad Emerg Med 2003; 10:119–26. [5] Seki K, Matsui H, Sekiya S. Advances in the clinical laboratory detection of gestational trophoblastic disease. Clin Chim Acta 2004;349:1–13. [6] Cole LA, Khanlian SA, Sutton JM, Davies S, Stevens ND. Hyperglycosylated hCG (invasive trophoblastic antigen) a key antigen for early pregnancy detection. Clin Biochem 2003;36:647–55. [7] Stenman U. Immunoassay standardisation: Is it possible, who is responsible, who is capable? Clin Chem 2001;47:815–20. [8] Montagnana M, Trenti T, Aloe R, Cerevellin G, Lippi G. Human chorionic gonadotropin in pregnancy diagnostics. Clin Chim Acta 2011;412:1515–20. [9] O'Connor JF, Ellish N, Kakuma T, Schlatterer J, Kovalevskaya G. Differential urinary gonadotropin profiles in early pregnancy and early pregnancy loss. Prenat Diagn 1998;18:1232–40. [10] O'Reilly SM, Rustin GJ. Mismanagement of choriocarcinoma due to false low hCG measurement. Int J Gynaecol Cancer 1993;3:186–8. [11] Tabas JA, Strehlow M, Isaacs E. A false negative pregnancy test in a patient with hydatidiform molar pregnancy. N Eng J Med 2003;349:2172–3. [12] Davidson CM, Kaplan RM, Wenig LN, Burmeister D. Qualitative beta-hCG urine assays may be misleading in the presence of molar pregnancy: a case report. J Emerg Med 2004;27:43–7. [13] Er TK, Yong YJ, Tsai EM, Huang CL, Chow HW, Zheng BH, et al. False negative pregnancy test in hydatidiform mole. Clin Chem 2006;52:1616–7. [14] Rodbard D, Feldman Y, Jaffe L, Miles LE. Kinetics of the two-site immunoradiometric (sandwich) assays. Studies on the nature of the “high dose hook effect”. Immunochemistry 1978;15:77–82.
211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247
248
Please cite this article as: Wilgen U, et al, Hook effect in Abbott i-STAT β-human chorionic gonadotropin (β-hCG) point of care assay, Clin Biochem (2014), http://dx.doi.org/10.1016/j.clinbiochem.2014.05.005
Contents lists available at ScienceDirect
Clinical Biochemistry journal homepage: www.elsevier.com/locate/clinbiochem
1
Short Communication
4Q1
Urs Wilgen a,b,⁎, Carel J. Pretorius a,b, Rehna S. Gous a, Cameron Martin a, Vincent J. Hale a, Jacobus P.J. Ungerer a
O
F
3
Hook effect in Abbott i-STAT β-human chorionic gonadotropin (β-hCG) point of care assay☆
5 6
a
7
a r t i c l e
8 9 10 11 12
Article history: Received 7 April 2014 Received in revised form 30 April 2014 Accepted 5 May 2014 Available online xxxx
13 14 15 16 17
Keywords: Human chorionic gonadotropin (β-hCG) Point-of-care testing Hook effect Gestational trophoblastic disease
i n f o
R O
Pathology Queensland, Department of Chemical Pathology, Royal Brisbane and Women's Hospital, Herston Road, Herston 4029, Queensland, Australia University of Queensland, School of Medicine, Brisbane, St. Lucia, 4072 Queensland, Australia
a b s t r a c t
D
P
Point-of-care testing for β-hCG has been widely advocated to allow rapid diagnosis/exclusion of pregnancy in the emergency department. A quantitative blood β-hCG assay has the additional benefit of being able to monitor the viability of pregnancy, using serial measurements, to determine the appropriate expected increase in β-hCG levels over time (e.g. ectopic pregnancy), and aiding in determining if an intrauterine gestational sac should be visible on sonographic imaging. Objectives: Evaluation of the newly released Abbott i-STAT β-hCG point-of-care assay with the Beckman Coulter β-hCG laboratory assay in use. Design and methods: Whole blood, plasma and serum samples with a wide range of β-hCG concentrations were analysed by both methods. Results: The Abbott I-STAT β-hCG compares favourably, can be performed on heparinised whole blood, plasma and serum, and shows acceptable accuracy and precision. However a hook effect at elevated β-hCG was shown in gestational trophoblastic disease as well as normal pregnancies. Conclusions: The i-STAT β-hCG performs acceptably in its intended use in the early detection of pregnancy, but results should always be interpreted within the clinical context, as a hook effect may occur. © 2014 The Canadian Society of Clinical Chemists. Published by Elsevier Inc.
E
C
T
b
E
2
42 43 44 45 46 47 48 49 50
R
N C O
40 41
Human chorionic gonadotropin (β-hCG) is a glycoprotein hormone produced predominantly in syncytiotrophoblasts, and is commonly used to confirm/exclude the diagnosis of pregnancy. HCG can be measured in urine or serum by immunoassay giving either a qualitative result (using a cut-off of N25 IU/L), or quantitative result. Concentrations of β-hCG in non-pregnant women are b5 IU/L, and rise rapidly during the first weeks of pregnancy, approximately doubling every two days, and reaching a peak of 100,000–500,000 IU/L at 8–10 weeks of gestation. Samples with β-hCG results b5 IU/L are interpreted as “negative”, 5–25 IU/L as “indeterminant” with a suggestion to repeat the test 2 days later, and N 25 IU/L as “positive” [1]. Testing for β-hCG at the point-of-care (POC) has been advocated to facilitate rapid diagnosis and decision making in the emergency department [2], but many of
U
38 39
R
Background
☆ Declarations: Abbott Diagnostics provided the β-hCG test cartridges for the evaluation free of charge. ⁎ Corresponding author at: Pathology Queensland, Department of Chemical Pathology, Royal Brisbane and Women's Hospital, Herston Road, Herston 4029, Queensland, Australia. Fax: +61 7 3646 3417. E-mail address: [email protected] (U. Wilgen).
18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 36 34 33 35
the qualitative hCG devices used at the POC are susceptible to the hook effect giving false negative results [3]. Apart from excluding/confirming the diagnosis of pregnancy with a qualitative β-hCG test, a quantitative β-hCG assay can prove useful for monitoring the viability of pregnancy, using serial measurements to determine the appropriate expected increase in β-hCG levels over time (e.g. ectopic pregnancy), and aiding in determining if an intrauterine gestational sac should be visible on sonographic imaging [4]. Abbott diagnostics recently released a β-hCG test cartridge to allow quantitative β-hCG testing at the POC on the i-STAT analyser. The test cartridge uses a two-site enzyme linked immunosorbent assay (ELISA) method for use with heparinised whole blood or plasma samples. The i-STAT assay reports quantitative β-hCG values from 5 to 2000 IU/L. Samples with concentrations less than 5 IU/L are reported as “b5 IU/L”, and exceeding 2000 IU/L are reported as “N 2000 IU/L” according to the manufacturer's package insert. Our statewide pathology service network utilises 270 Abbott i-STAT analysers, with 25 in small laboratory settings operated by laboratory personnel, and the remainder in non-laboratory POC settings (e.g. emergency departments) operated by medical and nursing staff. Serum is the preferred sample type in our laboratories for immunoassays. The i-STAT hCG assay is approved for use in Europe, New Zealand and Australia, and pending FDA approval for the USA and Canada. We evaluated the method for possible use in our pathology service.
http://dx.doi.org/10.1016/j.clinbiochem.2014.05.005 0009-9120/© 2014 The Canadian Society of Clinical Chemists. Published by Elsevier Inc.
Please cite this article as: Wilgen U, et al, Hook effect in Abbott i-STAT β-human chorionic gonadotropin (β-hCG) point of care assay, Clin Biochem (2014), http://dx.doi.org/10.1016/j.clinbiochem.2014.05.005
37 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74
U. Wilgen et al. / Clinical Biochemistry xxx (2014) xxx–xxx
2000
1000
500
1000
1500
2000
Beckman Coulter DxI800 hCG (IU/L) Fig. 1. Passing–Bablok regression of samples analysed for β-hCG on the Beckman Coulter DxI800 on serum and Abbott i-STAT on heparinised whole blood (open circles) (95% CI for slope and intercept: 0.89 to 0.97 and −1.05 to 4.31 respectively; n = 40) and serum (closed circles) (95% CI for slope and intercept: 1.04 to 1.12 and −0.87 to 4.50 respectively; n = 40).
114
R O
0
O
DxI800 serum vs iSTAT serum DxI800 serum vs iSTAT whole blood Passing Bablok: iSTAT=1.07 DxI800+1.66 Passing Bablok: iSTAT=0.92 DxI800+1.94
0
Results
F
500
93 94
P
iSTAT hCG (IU/L)
1500
β-hCG results up to ~2000 IU/L (measured by the DxI800) were selected at this stage due to the manufacturer's stated reportable range. Thereafter 16 additional serum samples with elevated β-hCG received for routine analysis were retrieved and analysed on the i-STAT. These samples were stored at 4 °C until analysis, which was usually the same day, but no longer than 24 h later. Three levels of the manufacturer's quality control materials were run to ensure the batch of Abbott i-STAT β-hCG cartridges performed within specifications. Accuracy and precision evaluation was performed analysing 3 levels of quality control material (24.7 IU/L; 929.5 IU/L; 1419.8 IU/L) ten times each. A series of dilutions using a sample with a β-hCG of 520,000 IU/L was performed on both methods using Beckman Coulter wash buffer (cat no. 8547197) as diluent. This is as per manufacturer's instruction for the DxI800. Diluting with β-hCG negative male serum demonstrated insignificant differences to wash buffer. The DxI800 will routinely dilute samples with β-hCG concentrations N1000 IU/L using a 1 in 200 dilution using wash buffer. Samples with concentrations exceeding 200,000 IU/L require a manual dilution. Abbott gives no instruction for dilutions, as the assay is intended to be as simple as possible for use at the point-of-care.
Analysis of 40 paired whole blood and serum samples on the i-STAT and DxI800 respectively showed excellent correlation using Passing Bablok regression (i-STAT = 0.92 ∗ DxI800 + 1.94, r2 = 0.994). Correlation between the DxI800 and i-STAT β-hCG when 40 serum samples were analysed by both methods was similar (i-STAT = 1.07 ∗ DxI800 + 1.66, r2 = 0.992) (see Fig. 1). Concordant interpretation would be achieved in all but 2 samples (DxI800 = 4.5 IU/L, i-STAT = 5.5 IU/L and DxI800 = 22.0 IU/L, i-STAT = 26.5 IU/L). Accuracy and precision are acceptable with %CVs of less than 7% across all 3 levels of quality control materials (Level 1 (24.7 IU/L): %CV 4.7%; Level 2 (929.5 IU/L): %CV 2.0%; Level 3 (1419.8 IU/L): %CV 6.6%). Five serum samples with elevated concentrations of β-hCG (measured on the DxI800) showed erroneously low values on the i-STAT, where numerical values were displayed below 2000 IU/L when the result should have been reported as “N 2000 IU/L”. This occurred when the β-hCG levels measured on the DxI800 exceeded 260,000 IU/L, demonstrating a hook effect on the i-STAT β-hCG, (see Table 1). The hook effect was present in samples from normal pregnancies as well as gestational trophoblastic disease. Dilutions of a serum sample with a measured β-hCG (on the DxI800) of 520,000 IU/L gave erroneously low results reported on the i-STAT up to a diluted concentration of 218,000 IU/L (see Table 2).
D
2
Design and methods
76 77
92
Human chorionic gonadotropin levels were measured on the Beckman Coulter UniCel DxI800 (Beckman Coulter Diagnostics, Brae CA USA) and an Abbott i-STAT POC Analyser (Abbott Diagnostics, IL USA). Forty samples (heparinised whole blood and serum) were collected in addition to routine bloods at visits to the antenatal clinic of the Royal Brisbane Women's Hospital. These samples were from pregnant women ≤6 week gestation. Whole blood and serum samples were collected in Becton Dickinson heparinised and serum separator tubes respectively. Whole blood samples were analysed on the i-STAT immediately upon receipt in the laboratory, and serum samples were allowed to clot for approximately 30 min, centrifuged and then analysed on both the DxI800 and the i-STAT. Our institutional ethics committee gave approval for the study, and written consent was obtained. A further 40 serum samples received in the laboratory for routine analysis and measured on the DxI800 were selected and retrieved from sample storage and analysed on the i-STAT. Only samples with
t1:1 t1:2 t1:3
Table 1 Patient serum samples with markedly elevated β-hCG performed on the Beckman Coulter DxI800 and Abbott i-STAT. The i-STAT should display numerical values exceeding 2000 IU/L as “N2000 IU/L”. Values exceeding 1000 IU/L will automatically be run in dilution on the Beckman Coulter DxI800.
90 91
T
C
E
R
88 89
R
86 87
O
84 85
C
82 83
t1:4
Patient sample
t1:5 t1:6 t1:7 t1:8 t1:9 t1:10 t1:11 t1:12 t1:13 t1:14 t1:15 t1:16 t1:17 t1:18 t1:19 t1:20
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
N
80 81 Q2
Clinical diagnosis
U
78 79
E
75
Ectopic pregnancy Ectopic pregnancy Ectopic pregnancy Miscarriage Miscarriage 1st Trimester bleed Hyperemesis gravidarum Abdominal pain in pregnancy Hyperemesis gravidarum Hyperemesis gravidarum 1st Trimester bleed 1st Trimester bleed Gestational trophoblastic disease Hyperemesis gravidarum Gestational trophoblastic disease Gestational trophoblastic disease
Gestational age (weeks)
Beckman coulter DxI800 (IU/L)
Abbott i-STAT (IU/L)
Interpretation of reported i-STAT result
7 6 6 6 8 9 10 10 12 14 10 8 – 13 – –
2846 18,000 21,194 22,826 140,229 210,000 240,000 250,000 260,000 290,000 310,000 320,000 360,000 390,000 520,000 1,200,000
N2000 N2000 N2000 N2000 N2000 N2000 N2000 N2000 N2000 1816 N2000 N2000 1421 1805 1195 717
Correct Correct Correct Correct Correct Correct Correct Correct Correct Hook effect Correct Correct Hook effect Hook effect Hook effect Hook effect
Please cite this article as: Wilgen U, et al, Hook effect in Abbott i-STAT β-human chorionic gonadotropin (β-hCG) point of care assay, Clin Biochem (2014), http://dx.doi.org/10.1016/j.clinbiochem.2014.05.005
95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113
115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136
U. Wilgen et al. / Clinical Biochemistry xxx (2014) xxx–xxx
Neat 1:1 2:3 1:4 1:150 1:200 1:250
520,000 277,265 218,712 117,000 4514 3344 2705
1195 1187 1702 N2000 N2000 N2000 N2000
Hook effect Hook effect Hook effect Correct Correct Correct Correct
Discussion
138 139
Human chorionic gonadotropin (hCG) is a heterodimeric glycoprotein hormone consisting of two α subunits and two β subunits with eight carbohydrate side chains, which results in considerable heterogeneity of the structure of hCG. The hCG molecule is produced by trophoblasts in the placenta, gestational trophoblastic disease and some other malignancies [5]. Immunoassays measure the quantity of β-hCG by means of a sandwich assay which includes a capture antibody and a signal antibody bound to different antigenic sites on the captured antigen molecule. Different β-hCG assays use antibodies to different sites on the β-hCG molecule, and therefore may measure very different combinations of β-hCG related molecules. Most immunoassays are designed to detect primarily the regular form of β-hCG produced by the placenta [6], but standardisation of immunoassays for heterogeneous antigens such as β-hCG is notoriously problematic [7]. Not only does hCG exist in numerous forms (intact, free α, free β, nicked, β-core fragment, and varying degrees of glycosylation) [8], but also the composition changes as pregnancy progresses, and different isoforms may be produced by neoplastic tissues [9]. According to the manufacturer “the assay is intended for use in the early detection of pregnancy only”. The principal reason for testing for β-hCG is to confirm/exclude pregnancy. Other indications include determination of pregnancy viability (e.g. ectopic pregnancy, miscarriage, foetal demise), where serial quantitative values are required, and to assist in the diagnosis and management of gestational trophoblastic disease. A quantitative POC β-hCG test is helpful in guiding decision making in the emergency department. In a woman of child bearing age with lower abdominal pain and bleeding with a positive pregnancy test the diagnosis of an ectopic pregnancy should always be borne in mind. Visualisation of an intrauterine pregnancy by sonographic examination is dependant on gestational age, and dating according to last menstrual period is notoriously unreliable. Studies have shown that an intrauterine gestational sac should be visible on endovaginal ultrasound when serum β-hCG levels exceed 1500 IU/L. Conversely, the absence of an intrauterine gestational sac would make an ectopic pregnancy more likely [4]. The i-STAT β-hCG assay will perform satisfactorily within its intended use in the early detection of pregnancy, but there may be instances in clinical practice where an erroneously low value is reported due to a hook effect. Numerous case reports of false negative urine, serum and both urine and serum β-hCG results in patients with gestational trophoblastic disease have been published [10–13]. The hook effect is an uncommon phenomenon seen in sandwich immunoassays when the antigen (in this case β-hCG) is present in sufficiently high concentrations to cause saturation of both the capture and signal antibodies, preventing
154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182
C
152 153
E
150 151
R
148 149
R
146 147
N C O
144 145
U
142 143
Conclusions
The Abbott i-STAT β-hCG assay used at the point of care shows good correlation with the laboratory method used. We have shown that in addition to heparinised whole blood and plasma, serum is an acceptable sample type. We have also demonstrated that erroneously low β-hCG results may be obtained on the i-STAT when concentrations approach 218,000 IU/L, and results within the reportable range (b 2000 IU/L) may be reported in samples with much higher concentrations due to a hook effect. In instances where the i-STAT values do not correlate with the clinical picture, the sample should be sent to the laboratory for analysis, and results should be interpreted with caution if gestational trophoblastic disease is suspected.
T
137
F
t2:5 t2:6 t2:7 t2:8 t2:9 t2:10 t2:11
O
Interpretation of reported i-STAT result
R O
Abbott i-STAT (IU/L)
P
Beckman Coulter DxI800 (IU/L)
D
t2:4
Dilution (serum:diluent)
140 141
the formation of a “sandwich”. As a result the signal generated is less than what would be expected at that concentration, and the analyser reports an erroneously low value [14]. The Abbott i-STAT β-hCG product insert states that “no significant hook effect was detected in samples up to 300,000 IU/L”. We have shown that samples exceeding 218,000 IU/L demonstrated a hook effect, and that values as low as 717 IU/L may be reported, which is well below the stated upper limit of reporting of 2000 IU/L, with actual concentrations exceeding 1,200,000 IU/L. Grossly haemolysed specimens are known to decrease β-hCG detection. None of the samples demonstrating a hook effect were haemolysed. In instances where the i-STAT values do not correlate with the clinical picture the sample should be sent to the laboratory, where the analysis of the sample in dilution will usually yield the correct result. Furthermore, reported β-hCG concentrations should be interpreted with caution, especially if a clinical suspicion of gestational trophoblastic disease exists.
Table 2 Dilutions of a serum sample with a markedly elevated β-hCG analysed on the DxI800 and i-STAT.
E
t2:1 t2:2 t2:3
3
183 184 Q3 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209
References
210
[1] Ashwood ER, Knight GJ. Clinical chemistry of pregnancy. In: Burtis CA, Ashwood ER, Bruns DE, editors. Tietz textbook of clinical chemistry and molecular diagnostics. 4th ed. St. Louis: Saunders; 2006. p. 2153–206. [2] Mishalani SH, Seliktak J, Braunstain GD. Four rapid serum-urine combination assays of chorionic gonadotropin (hCG) compared and assessed for their utility in quantitative determinations of hCG. Clin Chem 1994;40:1944–9. [3] Nerenz RD, Haowei Song, Gronowski AM. Screening to evaluate point-of-care human chorionic gonadotropin (hCG) devices for susceptibility to the hook effect by hCG β core fragment: evaluation of 11 devices. Clin Chem 2014;60:667–74. [4] Kohn MA, Kerr K, Malkevich D, O'Niel N, Kerr MJ, Kaplan BC. Beta-human chorionic gonadotropin levels and the likelihood of ectopic pregnancy in emergency department patients with abdominal pain or vaginal bleeding. Acad Emerg Med 2003; 10:119–26. [5] Seki K, Matsui H, Sekiya S. Advances in the clinical laboratory detection of gestational trophoblastic disease. Clin Chim Acta 2004;349:1–13. [6] Cole LA, Khanlian SA, Sutton JM, Davies S, Stevens ND. Hyperglycosylated hCG (invasive trophoblastic antigen) a key antigen for early pregnancy detection. Clin Biochem 2003;36:647–55. [7] Stenman U. Immunoassay standardisation: Is it possible, who is responsible, who is capable? Clin Chem 2001;47:815–20. [8] Montagnana M, Trenti T, Aloe R, Cerevellin G, Lippi G. Human chorionic gonadotropin in pregnancy diagnostics. Clin Chim Acta 2011;412:1515–20. [9] O'Connor JF, Ellish N, Kakuma T, Schlatterer J, Kovalevskaya G. Differential urinary gonadotropin profiles in early pregnancy and early pregnancy loss. Prenat Diagn 1998;18:1232–40. [10] O'Reilly SM, Rustin GJ. Mismanagement of choriocarcinoma due to false low hCG measurement. Int J Gynaecol Cancer 1993;3:186–8. [11] Tabas JA, Strehlow M, Isaacs E. A false negative pregnancy test in a patient with hydatidiform molar pregnancy. N Eng J Med 2003;349:2172–3. [12] Davidson CM, Kaplan RM, Wenig LN, Burmeister D. Qualitative beta-hCG urine assays may be misleading in the presence of molar pregnancy: a case report. J Emerg Med 2004;27:43–7. [13] Er TK, Yong YJ, Tsai EM, Huang CL, Chow HW, Zheng BH, et al. False negative pregnancy test in hydatidiform mole. Clin Chem 2006;52:1616–7. [14] Rodbard D, Feldman Y, Jaffe L, Miles LE. Kinetics of the two-site immunoradiometric (sandwich) assays. Studies on the nature of the “high dose hook effect”. Immunochemistry 1978;15:77–82.
211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247
248
Please cite this article as: Wilgen U, et al, Hook effect in Abbott i-STAT β-human chorionic gonadotropin (β-hCG) point of care assay, Clin Biochem (2014), http://dx.doi.org/10.1016/j.clinbiochem.2014.05.005
