Journal of Clinical Virology 96 (2017) 60–63

Contents lists available at ScienceDirect

Journal of Clinical Virology

journal homepage: www.elsevier.com/locate/jcv

Full length article

Evaluation of clinical sensitivity and specificity of hepatitis B virus (HBV), T

hepatitis C virus, and human immunodeficiency Virus-1 by cobas MPX:
Detection of occult HBV infection in an HBV-endemic area
⁎
Jihye Ha, Younhee Park, Hyon-Suk Kim
Department of Laboratory Medicine, Yonsei University College of Medicine, Seoul, Republic of Korea

A R T I C L E I N F O A B S T R A C T

Keywords: Background: Transfusion-transmitted infectious diseases remain a major concern for blood safety, particularly
Cobas MPX with hepatitis B virus (HBV), hepatitis C virus (HCV), and human immunodeficiency virus (HIV). Nucleic acid
Blood donor screening testing (NAT) in donor screening shortens the serologically negative window period and reduces virus trans-
Occult HBV infection mission. The cobas MPX (Roche Molecular Systems, Inc., Branchburg, New Jersey) is a recently developed
Multiplex PCR
multiplex qualitative PCR system that enables the simultaneous detection of HBV, HCV, and HIV with improved
sensitivity and throughput using cobas 6800 and 8800 instruments.
Objectives: The aim of this study was to conduct an evaluation of the clinical sensitivity and specificity of cobas
MPX detection of HBV, HCV, and HIV in clinical specimens.
Study design: Among samples referred for HBV, HCV, and HIV-1 quantification at Severance Hospital, Yonsei
University College of Medicine, positive samples were selected to evaluate sensitivity. A total of 843 samples was
tested using both cobas MPX and COBAS AmpliPrep/COBAS TaqMan Tests for HBV, HCV, and HIV-1 using the
cobas 8800 system and a COBAS TaqMan 96 analyzer, respectively. Samples that showed discrepancies were
confirmed by nested PCR.
Conclusions: The cobas MPX achieved excellent sensitivity and specificity for the detection of HBV, HCV, and
HIV-1 in clinical samples. We found that the lower limit of detection (LOD) of blood screening by NAT actually
improves clinical sensitivity, and occult HBV infection prevalence among healthy employees of the hospital was
rather high.

1. Background multi-dye probe that eliminated additional discriminatory testing [5].

Transfusion-transmitted infectious diseases remain a major concern
for blood safety, particularly with hepatitis B virus (HBV), hepatitis C
virus (HCV), and human immunodeficiency virus (HIV). To reduce the
risk of transfusion-transmitted viral infection, serologic screening was
commonly performed in the past, although infectious blood units from
donors in the window period could still be transfused [1]. Nucleic acid
testing (NAT) in donor screening has shortened the serologically ne-
gative window period and reduced virus transmission [2,3]. Blood
screening laboratories handle large quantities of blood products that
require timely release, demanding automation and high-throughput.
Roche Molecular Systems, Inc. (RMS) (Branchburg, New Jersey) de-
veloped a fully automated multiplex NAT system for HBV, HCV, and
HIV detection, the cobas TaqScreen MPX test, which used a single-dye
probe for the target viruses necessitating additional discriminatory
testing [4]. The next version, the cobas TaqScreen MPX v2.0 test, used a
Recently, Roche released the new cobas MPX, a multiplex qualitative
PCR system detecting HBV, HCV, and HIV (HIV-1 Group M, HIV-1
Group O, and HIV-2), with improved throughput using cobas 6800 and
8800 systems and a low limit of detection (LOD) [6] (Table 1).
2. Objective
HBV, in particular, is widespread on the Asian continent, but until
now, the cobas MPX test has not been evaluated for clinical sensitivity
and specificity in HBV-endemic areas, which could have a great clinical
impact. We therefore aimed to evaluate the clinical sensitivity and
specificity of the cobas MPX test using clinical samples derived from
Korean patients and healthy volunteers.

⁎
Corresponding author.
E-mail addresses: KIMHS54@yuhs.ac, kimhs54@gmail.com (H.-S. Kim).

http://dx.doi.org/10.1016/j.jcv.2017.09.010
Received 6 March 2017; Received in revised form 13 August 2017; Accepted 18 September 2017
1386-6532/ © 2017 Elsevier B.V. All rights reserved.
J. Ha et al. Journal of Clinical Virology 96 (2017) 60–63

Table 1 (Plasmid #65459); this was a gift from Wang-Shick Ryu and had a
Claimed limit of detection by Probit analysis at a 95% hit rate for HIV, HCV, and HBV for known concentration [7]. The LOD of the assay was determined by
COBAS AmpliPrep/COBAS TaqMan Test, cobas MPX, and previous versions.
serially diluting the standard plasmid. DNA for the nested PCR was re-
Limit of detection, IU/mL extracted using a QIAamp Viral-RNA Mini Kit (Qiagen Inc., Valencia,
CA) according to the manufacturer's instructions.
HIV-1 HCV HBV

a 3.3. Serological assays for HBV infection

COBAS AmpliPrep/COBAS TaqMan v2.0 [15–17] 27.5 11
cobas Taqscreen MPX Test [4] 49 11
cobas Taqscreen MPX Test, v2.0 [5] 50.3 6.8
cobas MPX [6] 25.7 7 1.4
a
Numbers in parentheses indicate reference numbers.
3. Study design
3.1. Subjects and samples
Among samples referred for HBV, HCV, and HIV-1 quantification at
Severance Hospital, Yonsei University College of Medicine, positive
samples were selected to evaluate sensitivity. In total, 117, 111, and
105 positive samples by COBAS AmpliPrep/COBAS TaqMan tests for
HBV, HCV, and HIV-1 [14–16] using a COBAS TaqMan 96 analyzer
were collected to evaluate HBV, HCV, and HIV-1 sensitivity, respec-
tively. A further 510 samples collected during annual check-ups of
hospital employees with no clinical history of HBV, HCV, or HIV-1 in-
fection were used to evaluate specificity. The median age of 510 sub-
jects enrolled was 40 years (range 23–67). A total of 121 subjects were
male and 389 were female. This study was performed with the per-
mission of our hospital Institutional Review Board. The enrolled sub-
jects were anonymized for the study.
3.2. Molecular assays
In total, 843 samples were tested using both cobas MPX and COBAS
AmpliPrep/COBAS TaqMan tests for HBV, HCV, and HIV-1 [14–16]
using the cobas 8800 system and a COBAS TaqMan 96 analyzer, re-
spectively, following the manufacturer’s instructions. The limits of de-
tection of reagents used in this study are summarized in Table 1.
To calculate clinical sensitivity and specificity, concordant results
from both tests were assumed to be true results. Samples that showed
discrepant HBV results between cobas MPX and COBAS AmpliPrep/
COBAS TaqMan tests were confirmed by nested PCR (primers for first
round of PCR: BRB1, GATCGGATCCAGACTCGTGGACTTCTC; BRB2,
GATCTCTAGACCAAAAGACCCACAATT, primers for second round of
PCR: BRB1;BRB3, GATCTCTAGATGACATACTTTCCAATCAAT). Direct
sequencing of PCR products (primer: BRB1) of positive samples was
performed using a 3500xL Genetic Analyzer (Life Technologies,
Carlsbad, CA) to confirm the pathogen. To estimate the LOD of the
nested PCR used in this study, an HBV 1.3-mer WT replicon was used
9
3.8
2.3 Serum samples stored at −70 °C after performing the molecular
assays were used for serologic evaluation. HBsAg, anti-HBs, and anti-
HBc were detected using ARCHITECT HBsAg Qualitative II,
ARCHITECT anti-HBs, and ARCHITECT anti-HBc II assays (Abbott
Laboratories, IL, USA), respectively. Serological assays were performed
following the manufacturer’s recommendations without any modifica-
tions.
3.4. Statistical analyses
Analysis of the data was performed with analyze-it, version 3.80
(Analyze-it Software Ltd., Leeds, UK). A P value of less than 0.05 was
considered significant.
4. Results
4.1. Clinical sensitivity
All positive samples by COBAS AmpliPrep/COBAS TaqMan showed
positive results by cobas MPX. The sensitivity of cobas MPX was 100%
(333/333, 95% CI 98.90-100.00).
4.2. Clinical specificity
All HIV-1- and HCV-negative samples by COBAS AmpliPrep/COBAS
TaqMan were also shown to be negative by cobas MPX (510/510).
Therefore, specificities for HIV-1 and HCV were 100% (510/510, 95%
CI 99.28-100.00) (Table 2). However, for HBV, 17 samples from
healthcare workers that were HBV negative by COBAS AmpliPrep/
COBAS TaqMan showed positive results by cobas MPX. These dis-
crepant samples were tested by nested PCR and direct sequencing,
which showed 15 of the 17 samples as positive. Assuming that the two
samples that were positive by cobas MPX and negative by nested PCR
with direct sequencing were false positives, the specificity was 99.59%
(508/510, 95% CI 98.54-99.95) (Table 2). The LOD of nested PCR we
used for confirmation was within 100–101 copies/mL
4.3. Concordance
The overall concordance rate and kappa coefficient between COBAS
AmpliPrep/COBAS TaqMan and MPX were 98.0% (826/843, 95% CI

Table 2
Clinical sensitivity and specificity of cobas MPX.

Resultsa Total tested MPX results Clinical Sensitivity (%) 95% CI Clinical Specificity (%) 95% CI

HIV-1 Positive 117 Positive: 117 100 96.9–100.0 100 99.5–100.0

Negativeb 726 Negative: 726

HCV Positive 111 Positive: 111 100 96.7–100.0 100 99.5–100.0

Negativeb 732 Negative: 732

HBV Positivec 122 Positive: 122 100 97.2–100.0 99.7 99.0–99.9

Negativec 721 Positive: 2, Negative: 721

CI: confidence interval.

a
Concordant results by COBAS AmpliPrep/COBAS TaqMan Test v2.0, cobas MPX, and the nested PCR and sequencing results of discordant samples were assumed to be true.
b
Negative samples are consist of 510 of hospital employees with no history of HBV, HCV or HIV-1 and predefined positive samples for other targets.
c
15 samples of hospital employees showed positive in MPX and confirmed by nested PCR and sequencing was included in positive samples.

61
J. Ha et al. Journal of Clinical Virology 96 (2017) 60–63

Table 3
Serologic status of the subjects with suspected occult HBV infection.

Case No. Sex Age HBsAg Anti-HBc Anti-HBs Cobas MPX Nested PCR and Sequencing

1 M 47 Negative Positive Positive HBV HBV

2 M 52 Negative Positive Positive HBV HBV
3 F 38 Negative Positive Positive HBV HBV
4 F 35 Negative Positive Positive HBV HBV
5 F 51 Negative Positive Positive HBV HBV
6 M 63 Negative Positive Negative HBV HBV
7 F 45 Negative Negative Positive HBV Negative
8 F 32 Negative Negative Positive HBV Negative
9 F 41 Negative Negative Positive HBV HBV
10 F 46 Negative Negative Positive HBV HBV
11 F 33 Negative Negative Positive HBV HBV
12 F 36 Negative Negative Positive HBV HBV
13 M 67 Negative Negative Negative HBV HBV
14 F 30 Negative Negative Negative HBV HBV
15 F 34 Negative Negative Negative HBV HBV
16 F 43 Negative Negative Negative HBV HBV
17 F 51 Negative Negative Negative HBV HBV

96.7–98.8) and 0.96 (95% CI, 0.94–0.98, P < 0.0001), respectively.
HCV and HIV-1 testing showed a concordance rate of 100% (843/843,
95% CI 99.6–100.00) with kappa coefficient of 1.0. HBV testing showed
a concordance rate of 98.0% (826/843, 95% CI 96.8–98.8) with kappa
coefficient of 0.91 (95% CI, 0.87–0.95, P < 0.0001).
4.4. Serological assays for occult HBV infection samples
Among the 510 samples from healthy employees, anti-HBs were
detected in 307 (60.2%). Among 17 HBV positive samples by the cobas
MPX, HBsAg was negative for all samples. Anti-HBc and anti-HBs were
detected in 6 (35.3%) and 11 (64.7%) samples, respectively (Table 3).
5. Discussion
This study found that the lower LOD of blood screening by NAT
actually improves clinical sensitivity. Molecular study of healthy em-
ployee samples using the COBAS AmpliPrep/COBAS TaqMan test de-
tected no positive samples. On the contrary, using cobas MPX, 17 po-
sitive samples were detected. This implies that occult HBV infection
(OBI) could be better screened using cobas MPX by lowering the LOD in
endemic areas. Among 17 positive samples by cobas MPX, 15 were
confirmed to be HBV positive by nested PCR and subsequent sequen-
cing. The remaining two samples showed no bands in nested PCR
analysis. However, although we used nested PCR as a confirmatory test,
there is a possibility that the cobas MPX test has a lower LOD than
nested PCR. In our study, LOD of the nested PCR performed in the study
was between 100-101 copies/mL and the claimed LOD of cobas MPX
was 1.4 IU/mL (8.1 copies/mL under conversion factor of 5.82) [6].
Considering that the true results of the two discrepant samples showing
discrepancy remain unclear, the cobas MPX showed excellent clinical
sensitivity and specificity.
Whereas OBI infectivity depends on recipient immune status, blood
taken from donors with OBI was shown to be potentially infectious, and
the risk is increased when large volumes of plasma are transfused [8].
In HBV endemic areas, the residual risk of transfusion-transmitted HBV
infection originates from HBsAg-negative donors with OBI [1]. In this
study, the cobas MPX test detected HBV in 17 samples from healthy
workers, of which 15 samples were also confirmed by nested PCR and
subsequent sequencing. When these 15 positive samples were con-
sidered as OBIs, the prevalence of OBI among the tested subjects was
2.9% (15/510).
We reviewed literature that investigated OBI prevalence in the
absence of liver disease in Korea [9–12] (Table 4). Because anti-HBc has
been considered as a marker of previous HBV exposure, many studies
performed HBV DNA tests on anti-HBc positive subjects to investigate
the prevalence of OBI. However, Song et al. reported that anti-HBc was
positive in only 27% of OBI cases in Korean adult subjects [11]. In this
study, anti-HBc was detected in almost half of the samples (6/15,
40.0%). Therefore, we reviewed literature that investigated OBI pre-
valence regardless of anti-HBc. Overall, 17–67% of OBI cases from the
literature were anti-HBc positive, not counting the extremes. Interest-
ingly, the prevalence of OBI varied considerably from study to study.
Kim et al. detected 31 (16%) OBI cases from 195 outpatients of a gas-
trointestinal clinic with normal ALT levels. However, Song et al. re-
ported only 0.7% (7/1047) OBI prevalence for medical check-up sam-
ples. Various OBI prevalences were reported in Korea (Table 4). These
considerable differences in OBI prevalences could be explained mainly
by the characteristics of the subjects. The different OBI prevalences of
same population in a study were resulted from the analytical sensitivity
of HBV DNA detection methods [10,12]. Despite similar analytical
sensitivity, OBI prevalence in this study was four-times higher than that
was reported by Song et al. (2.9 vs. 0.7%) [11]. This difference is
consistent with other studies, which concluded that the risk of con-
tracting HBV by health care workers is four-times greater than that of
the general adult population [13,14].
In conclusion, the cobas MPX test achieved excellent sensitivity and
specificity for the detection of HBV, HCV, and HIV-1. We found that the
lower LOD of NAT blood screening actually improves clinical sensi-
tivity. Using the cobas MPX test, OBI was detected in 2.9% (15/510) of
healthy hospital employees. The excellent clinical sensitivity and spe-
cificity was accompanied by improved efficiency and reduction of the
costs and turn-around time of NAT for donor screening.
Competing interests
No potential conflicts of interest relevant to this article were re-
ported.
Funding
This research received no specific grant from any funding agency in
the public, commercial, or not-for-profit sectors.

62
J. Ha et al. Journal of Clinical Virology 96 (2017) 60–63

Ethical approval

55% (17/31)
70% (16/23)

65% (11/17)
100% (1/1)
75% (6/8)
86% (6/7)
67% (2/3)
Anti-HBs
This study was approved by Severance Hospital IRB (IRB No. 1-
2015-0047, Protocol No. RSS-MD-15-01).

References
52% (16/31)
17% (4/23)

53% (6/17)
100% (1/1)
29% (2/7)
67% (2/3)
0% (0/8)
Anti-HBc

[1] J.P. Allain, Occult hepatitis B virus infection: implications in transfusion, Vox Sang.
86 (2004) 83–91.
[2] P.F. Lindholm, K. Annen, G. Ramsey, Approaches to minimize infection risk in
blood banking and transfusion practice, Infect. Disord. Drug Targets 11 (2011)
45–56.
Analytical sensitivity (IU/mL)

[3] J.T. Wang, T.H. Wang, J.C. Sheu, L.N. Shih, J.T. Lin, D.S. Chen, Detection of he-
patitis B virus DNA by polymerase chain reaction in plasma of volunteer blood
donors negative for hepatitis B surface antigen, J. Infect. Dis. 163 (1991) 397–399.
[4] Roche, cobas® TaqScreen MPX Test: instruction manual, Roche Molecular Systems,
Inc., Branchburg, New Jersey, 2012.
[5] Roche, cobas® TaqScreen MPX Test version 2.0: instruction manual, Roche
Molecular Systems, Inc., Branchburg, New Jersey, 2014.
[6] Roche, cobas® MPX, Roche Molecular Systems, Inc., Branchburg New Jersey, 2015.
0.5−5.4

[7] H. Wang, S. Kim, W.S. Ryu, DDX3 DEAD-Box RNA helicase inhibits hepatitis B virus
73.2
3.8

1.4

reverse transcription by incorporation into nucleocapsids, J. Virol. 83 (11) (2009)

60

60
24

5815–5824.
[8] J.P. Allain, I. Mihaljevic, M.I. Gonzalez-Fraile, K. Gubbe, L. Holm-Harritshoj,
J.M. Garcia, et al., Infectivity of blood products from donors with occult hepatitis B
COBAS Amplicor HBV Monitor test

COBAS Amplicor HBV monitor test

virus infection, Transfusion (Paris) 53 (7) (2013) 1405–1415.

[9] S.M. Kim, K.S. Lee, C.J. Park, J.Y. Lee, K.H. Kim, J.Y. Park, J.H. Lee, H.Y. Kim,
J.Y. Yoo, M.K. Jang, Prevalence of occult HBV infection among subjects with
Cobas TaqScreen MPX test

normal serum ALT levels in Korea, J. Infect. 54 (2007) 185–191.

[10] C.I. Kwon, S.G. Hwang, S.J. Shin, S.W. Chang, S.Y. Kim, K.H. Ko, S.P. Hong,
P.W. Park, K.S. Rim, M.S. Kang, H.J. Chung, S.P. Hong, Occult hepatitis B virus
infection in pregnant woman and its clinical implication, Liver Int. 28 (2008)
667–674.
TaqMan PCR

TaqMan PCR
Nested PCR

Cobas MPX

[11] E.Y. Song, Y.M. Yun, M.H. Park, D.H. Seo, Prevalence of occult hepatitis B virus
infection in a general adult population in Korea, Intervirology 52 (2009) 57–62.
Method

[12] J.H. Yoo, S.G. Hwang, D.H. Yang, M.S. Son, C.-I. Kwon, K.H. Ko, S.P. Hong,
P.W. Park, K.S. Rim, Prevalence of occult hepatitis B virus infection in hemodialysis
patients, Korean J. Gastroenterol. 61 (2013) 209–214.
[13] L.A. Ciorlia, D.M. Zanetta, Hepatitis B in healthcare workers: prevalence, vaccina-
Age-years (range)

tion and relation to occupational factors, Brazilian J. Infect. Dis. 9 (2005) 384–389.
[14] E.B. Byrne, Viral hepatitis: an occupational hazard of medical personnel. Experience
of the Yalenew Haven Hospital, 1952–1965, JAMA 195 (1966) 362–364.
54 (15–94)

42 (14–72)
58 (25–86)
58 (25–86)
40 (23–67)

[15] Roche, COBAS® AmpliPrep/COBAS® TaqMan® HIV-1 Test, version 2.0: instruction
manual, Roche Molecular Systems, Inc., Branchburg, New Jersey, 2016.
[16] Roche, COBAS® AmpliPrep/COBAS® TaqMan® HCV Test, version 2.0: instruction
30.8
30.8

manual, Roche Molecular Systems, Inc., Branchburg, New Jersey, 2016.

[17] Roche, COBAS® AmpliPrep/COBAS® TaqMan® HBV Test, version 2.0: instruction
manual, Roche Molecular Systems, Inc., Branchburg, New Jersey, 2016.
Outpatient of a GI clinic with normal ALT

Healthy employees of our hospital

Previous reports of prevalence of occult HBV infection without overt liver disease in Korea.

Hemodialysis patients
Hemodialysis patients
Medical check-up
Pregnant women
Pregnant women
Subjects

11.4% (23/202)
OBI prevalence

0.7% (7/1047)

2.9% (15/510)
16% (31/195)

3.2% (3/94)
1.1% (1/94)
4% (8/202)

Yoo, Hwang et al, 2013 [12]

Yoo, Hwang et al, 2013 [12]
Reference, published year

Kwon et al, 2008 [10]

Kwon et al, 2008 [10]
Song et al, 2009 [11]
Kim et al, 2007 [9]

Present study
Table 4

63