Journal of Clinical Virology 106 (2018) 44–48

Contents lists available at ScienceDirect

Journal of Clinical Virology

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

Diagnostic utility of hepatitis E virus antigen-specific ELISA versus PCR T

testing in a cohort of post liver transplant patients in a large university
hospital
⁎
G. Soothilla, , S. Hesseya, M. Erotocritoub, P. Griffithsa, S. Ijazc, D. Thorburna, M. Ankcornc,d,
D. Irisha
a
Royal Free Hospital, Pond St, London, NW3 2QG, UK
b
University College London, Gower St, Bloomsbury, London WC1E 6BT, UK
c
Blood Borne Virus Unit, Virus Reference Department, National Infection Service, Public Health England, London, UK
d
Transfusion Microbiology, National Health Service Blood and Transplant, London, UK

A R T I C LE I N FO A B S T R A C T

Keywords: Background: Hepatitis E virus (HEV) is an important infectious agent causing acute and chronic disease. Chronic
Hepatitis E virus hepatitis E affects immunocompromised people and serological testing is neither reliable nor sufficient to infer
Antigen-specific ELISA whether a patient has infection; therefore HEV RNA testing is the only reliable diagnostic test presently avail-
Liver transplant patients able. An HEV antigen-specific ELISA test is commercially available but is not yet in clinical use.
Screening
Objectives: 1) determine the prevalence of HEV infection in the Royal Free Hospital (RFH) liver transplant co-
Positive predictive value
hort; 2) compare the diagnostic utility of HEV antigen-detection against the current gold standard; 3) consider
Negative predictive value
screening strategies for HEV infection in immunocompromised groups.
Study design: The serum samples of 490 post liver transplant patients visiting the outpatient clinic at the RFH
over an eight-month period were tested for HEV with both an HEV antigen-specific ELISA and HEV RNA test.
Results: The prevalence of HEV infection was 0.20% (95% CI 0.0%–1.1%). The specificity of the ELISA was
98.2% with a positive predictive value of 10.0%. There was one true positive HEV case, which was picked up
correctly by the antigen-specific ELISA. These results were improved by incorporating a neutralisation step into
further ELISA tests.
Conclusions: The antigen-specific ELISA test gave no false negative results, supporting its utility as a screening
tool. There was one true antigen positive result. Further investigation including cost analysis is indicated to
determine the efficacy of HEV antigen-specific ELISA testing in a screening context and in the clinical in-
vestigation of HEV infection in immunocompromised patients.

1. Background 3, are being reported in Europe and North America [4]. Four genotypes
are well recognised: 1 and 2 (human viruses that infect humans via the
Hepatitis E virus (HEV) was first isolated by Mikhail Balayan in faeco-oral route) and 3 and 4 (animal viruses that infect humans zoo-
1983 and is a positive sense single-stranded non-enveloped RNA virus notically). Genotype 3 can affect animals such as pigs and deer and
of the Hepeviridae family [1]. It is recognised as a leading cause of acute transmission to humans is usually acquired from the diet, with pork
infectious hepatitis worldwide affecting around 20 million people a products carrying the greatest risk [4,5].
year and responsible for over 50,000 deaths [2]. It is currently the most The incubation period is 2–8 weeks and may be followed by acute
common cause of enteric acute hepatitis in the UK [3]. hepatitis; however, the majority of cases are asymptomatic [6]. Infec-
HEV was previously considered a disease of developing countries tion is usually self-limiting and the overall case-fatality for HEV-in-
but now an increasing number of human infections, mainly of genotype duced acute hepatitis is around 0.5–4% [6]. Fulminant hepatic failure

Abbreviations: HEV, Hepatitis E virus; ALT, alanine aminotransferase; AST, aspartate aminotransferase; HSCT, haematopoietic stem cell transplant; ELISA, enzyme-
linked immunosorbent assay; PCR, polymerase chain reaction; RFH, Royal Free Hospital; PPV, positive predictive value; NPV, negative predictive value; PHE, Public
Health England; VRD, Virus Reference Department; CI, confidence interval
⁎
Corresponding author.
E-mail address: germander.soothill@nhs.net (G. Soothill).

https://doi.org/10.1016/j.jcv.2018.07.007
Received 7 May 2018; Received in revised form 8 July 2018; Accepted 20 July 2018
1386-6532/ © 2018 Published by Elsevier B.V.
G. Soothill et al. Journal of Clinical Virology 106 (2018) 44–48

in pregnant women is a feature of genotype 1 infection and carries up to 2.3. Data collection
25% mortality [7]. A diagnosis of chronic HEV infection requires per-
sistence of the virus for over three months and is a feature of genotype 3 As part of routine care, post liver transplant patients attend the
virus infections [8], with some cases of genotype 4 persistence also hepatology clinic at RFH for regular follow-up. Time interval between
reported [9]. This manifests with symptoms of fatigue, diarrhoea and follow-up visits varies depending on time from transplant and/or any
arthralgia and persistent moderately elevated transaminases [10]. In hepatological complications. Serum samples are taken on these visits
solid organ transplant patients it may result in progressive liver injury for assessment of liver biochemistry and graft function. Once all re-
and cirrhosis in around 10% of cases [11]. Lack of awareness and a quested biochemistry tests have been performed by the laboratory any
largely asymptomatic infection means that chronic HEV infection can surplus samples would usually be discarded. These residual samples
remain undiagnosed [12]. The main risk factor for persistence of HEV is were taken to the Virology department and tested as part of this audit.
being immunocompromised and it is well recognised in transplant re- Blood samples from 490 patients who visited the outpatient clinic be-
cipients (renal, liver and haematopoietic stem cell (HSCT)), HIV-in- tween 31st January and 10th August 2017 were collected. Samples were
fected patients with low CD4 counts and patients with haematological aliquoted into two tubes, one for the HEV antigen ELISA and another
malignancies receiving chemotherapy [13]. for HEV RNA detection, which were stored separately at −80 °C, prior
The prevalence of HEV infection in Europe is higher than expected, to analysis.
with some studies showing around 25% of adults in the sixth and se- A liver transplant database, which included data on 458 of the pa-
venth decades of life to be seropositive [14]. Data from English blood tients included in the study (patients who had transplants at other
donors indicates 0.04% to be positive for active HEV infection [15]. The hospitals were not included in the database), was used to determine the
prevalence has been found to be higher amongst at-risk patient popu- demographics, baseline serology and the use of blood products peri-
lations such as transplant recipients [16]. Morbidity and mortality is operatively for the patients.
high in transplant populations and up to 60% of patients do not clear
the virus [17].
Current diagnostic testing for HEV includes the detection of HEV 2.4. Laboratory methods
antibodies (IgM and IgG) and HEV RNA. As antibody detection may be
unreliable in immunocompromised patients, HEV RNA testing must be HEV antigen was detected using the Fortress anti-HEV antigen
used for the screening and diagnosis of HEV infection in this patient ELISA assay (Fortress Diagnostics, Antrim, United Kingdom,
group [8]. BXE0903 A) which is a two-step, solid phase antibody sandwich ELISA.
An enzyme-linked immunosorbent assay (ELISA), which tests for The ELISA uses micro-well strips pre-coated with anti-HEV antibodies.
HEV antigen has been developed by Wantai and is now commercially In the first step, the test serum is placed onto the strips and incubated to
available. Testing carried out by the manufacturers determined the test allow for immune-complex formation and capture onto the solid phase.
to perform with a sensitivity of 66.7% and a specificity of 99.9% In the second step, anti-HEV antibodies conjugated to horseradish
(Wantai). This test has also been trialled in a transplant cohort with peroxidase are added, binding to the immobilised immune-complexes.
reported sensitivity of 65% for detecting acutely infected HEV RNA- After adding a peroxidase substrate the wells with antibody-antigen-
positive patients and sensitivity of 100% for detecting those chronically antibody ‘sandwich’ immune-complexes form coloured products, which
infected, with an overall specificity of 92% [18]. can be visualised and measured. The colour intensity or light absor-
The aims of this study were threefold: 1) determine the current bance of each sample is proportional to the amount of antigen present.
prevalence of HEV infection in the Royal Free Hospital (RFH) liver The absorbance value is read at wavelength 450/630 nm when a dual
transplant cohort; 2) compare the diagnostic utility of HEV antigen- filter instrument is used; the Biochrom ASYS Expert 96 plate reader was
detection against the current gold standard of HEV RNA detection; 3) used in this study.
use this data to consider screening strategies for HEV infection in im- The results from all the samples were expressed as a ratio of the
munocompromised groups. individual absorbance (S) to the cut-off (CO) (S/CO). Samples with a
ratio < 1, were classified as negative. Samples with a ratio of > 1 were
considered to be initially reactive. The manufacturer’s definition of a
2. Study design
positive result is at least two reactive tests. Hence, initially reactive
samples were subsequently retested and if they repeatedly (at least
2.1. Setting and population
twice) produced a ratio > 1, they were considered to be repeatedly re-
active and classified as positive. Initially equivocal results (S/
The study was conducted at the RFH, which is a large university
CO = 0.9–1.1) were also retested. Initially reactive or equivocal sam-
teaching hospital situated in North London serving a population of
ples, which were not reactive on repeat testing, were classified as ne-
about 1.6 million. The hospital has specialist tertiary referral hepa-
gative.
tology and liver transplant services. The study group included patients
All samples were tested individually for HEV RNA at the Virus
who had undergone a liver transplant (at RFH or elsewhere) and were
Reference Department (VRD), Public Health England (PHE), Colindale,
being followed up at the liver transplant outpatient clinic.
using a method as previously described [19,20]. Nucleic acid was ex-
tracted from 200 μl of each sample on the MagNA Pure 96 (Roche Di-
2.2. Ethics agnostics Ltd. Burgess Hill, UK; virus-specific cell-free protocol) before
HEV RNA detection and quantitation using an internally controlled
Ethical review was not required for this study since it was under- validated quantitative HEV PCR (limit of detection 22 IU/ml) [20].
taken as a clinical audit to establish the prevalence of HEV and validity To confirm the specificity of reactivity in the HEV antigen assay the
of an HEV ELISA test in liver transplant patients at the hospital. The reactive samples were referred for neutralisation, which was performed
liver transplant patients at RFH are given the opportunity to consent for at VRD. This technique, recently developed to differentiate between
their clinical samples to be tested to improve clinical services. The true reactivity and non-specific reactivity, was performed on samples
patient serum samples were assigned a number and all patient identi- that were HEV antigen reactive a third time at the reference laboratory
fying information was removed prior to antigen and PCR testing. In the and is described in detail elsewhere [21]. Samples were considered as
case of a positive PCR test, the protocol was to de-anonymise the truly positive in the HEV antigen assay when neutralisation exceeded
sample and inform the clinical team, in the best interest of the patient 50%.
and to review potential treatment options.

45
G. Soothill et al. Journal of Clinical Virology 106 (2018) 44–48

Table 1
Comparison of the baseline characteristics of patients with an initially reactive antigen test (HEV Ag test IR) to those without an initially reactive antigen test (HEV Ag
test not IR) and comparison of patients with repeatedly reactive antigen tests (HEV Ag test RR) to those without repeatedly reactive antigen tests (HEV Ag test not RR).
Gender, age, ethnicity and serology at the time of liver transplant are reported as total number (n) and percentage and compared using the Chi-squared test. Peri-
operative blood products used and serum parameters are reported as total number (n) and median with interquartile range and compared using the Mann Whitney U
test. Data on baseline characteristics were available from a liver transplant database for the majority of the 490 patients. Gender and age information was available
for 458 patients; information on peri-operative blood products was available for 456 patients and ethnicity and serology information was available for 455 patients.
(HIV Ab + ve = HIV antibody positive; HBsAg + ve = Hepatitis B surface antigen positive; Hep C Ab + ve = Hepatitis C antibody positive; HBsAg + Hep D
Ab + ve = Hepatitis B surface antigen and Hepatitis D antibody positive).
HEV Ag test not IR HEV Ag test IR P value HEV Ag test not RR HEV Ag test P value
(not IR v. IR) RR (not RR v. RR)

Gender (n = 443) (n = 15) 0.10 (n = 448) (n = 10) < 0.0001

(n, %) 284, 64.1% 13, 86.7% 288, 64.3% 9, 90.0%
Male 159, 35.9% 213.3% 160, 35.7% 1, 10.0%
Female
Age (n = 443) (n = 15) 0.87 (n = 448) (n = 10) 0.82
(n, %) 4, 0.9% 0, 0.0% 4, 0.9% 0, 0.0%
< 18 years 213, 48.1% 8, 53.3% 217, 48.4% 4, 40.0%
18-49 years 226, 51.0% 7, 46.7% 227, 50.7% 6, 60.0%
50 ≤ years
Ethnicity (n, %) (n = 440) (n = 15) 0.23 (n = 445) (n = 10) 0.25
Caucasian 310, 70.5% 14, 93.3% 314, 70.6% 10, 100.0%
Asian 88, 20.0% 0, 0.0% 88, 19.8% 0, 0.0%
Black 34, 7.7% 1, 6.7% 35, 7.9% 0, 0.0%
Other 8, 1.8% 0, 0.0% 8, 1.8% 0, 0.0%
Time since transplant (n = 444) (n = 15) 0.28 (n = 449) (n = 10) 0.81
(n, %) 94. 21.2% 1, 6.7% 94, 20.9% 1, 10.0%
< 1 year 75, 16.9% 2, 13.3% 75, 16.7% 2, 20.0%
1 ≤ years < 3 61, 13.7% 1, 6.7% 61, 13.6% 1, 10.0%
3 ≤ years < 5 26, 5.8% 0, 0.0% 26, 5,8% 0, 0.0%
5 ≤ years < 7 46, 10.4% 3, 20.0% 47, 10.5% 2, 20.0%
7 ≤ years < 9 142, 32.0% 8, 53.3% 146, 32.5% 4, 40.0%
9 ≤ years
Serology (n = 440) (n = 15) 0.85 (n = 445) (n = 10) 0.73
(n, %) 329, 74.8% 10, 66.7% 333, 74.8% 6, 60.0%
All negative 4, 0.9% 0, 0.0% 4, 0.9% 0, 0.0%
HIV Ab + ve 20, 4.5% 1, 6.7% 20, 4.5% 1, 1,0%
HBsAg + ve 78, 17.7% 4, 26.7% 79, 17.8% 3, 3.0%
Hep C Ab + ve 9, 2.0% 0, 0.0% 9, 2.0% 0, 0.0%
HBsAg + Hep D Ab + ve
Perioperative products (n = 441) (n = 15) 0.61 (n = 446) (n = 10) 0.42
(median, IQR) 13.0, 8.0-21.0 5.0, 4.5-8.0 0.21 3.0, 0.0-6.0 2.0, 0.0-4.8 0.99
Blood 11.0, 7.0-19.0 7.5, 5.3-9.8 0.89 3.0, 0.0-6.0 4.0, 0.5-5.5 0.89
Plasma 4.0, 3.0-5.0 1.0, 1.0-1.0 0.55 0.0, 0.0-2.0 0.0, 0.0-2.3 0.80
Platelets 6.0, 3.0-13.0 0.0, 0.0-0.0 0.0, 0.0-0.0 0.0, 0.0-0.0
Cryoprecipitate
Serum parameters (n = 475) (n = 15) 0.07 (n = 480) (n = 10) 0.12
(median, IQR) 4.0, 2.0-6.0 6, 3-7 0.93 4.0, 2.0-6.0 5.5, 3.5-7.8 0.87
Haemolysis index 16.0, 12.0-22.0 16, 12-22 0.81 16.0, 12.0- 22.0 17.5, 14.0 – 28.8 0.64
Icterus index 11.0, 8.0-17.0 11, 8-15.5 11.0, 8.0-17.0 12.5, 7.8-17.0
Lipidaemia index

2.5. Statistical methods product use or serological markers of other viruses (including hepatitis
B, C, D and HIV) between patients who had an initially reactive antigen
Statistical analysis was conducted in GraphPad Prism 7.0c. test, those who had a repeatedly reactive antigen test and those who did
Demographic data (gender, age, ethnicity and baseline serology) were not (Table 1).
expressed as percentages and quantitative variables as median and in- Samples from 15 patients were found to be initially reactive. On re-
terquartile ranges. The Mann Whitney U test was used to compare peat testing five of the samples had negative antigen test results and
nonparametric data and the Chi-squared test was used to compare ca- therefore did not fulfil the criteria of repeatedly reactive and were clas-
tegorical data. P value of < 0.05 was used to define a significant dif- sified as negative. The range of reactive value (S/CO) for the initially
ference. The Binomial "exact" calculation was used to calculate the 95% reactive patients was 0.92–10.88. Four of the five initially reactive sam-
confidence interval (CI) for prevalence of HEV. ples had a low-level reactive value in the initial test and one had a high-
level reactive value, which was not repeated on testing in-house or at
3. Results the reference laboratory. Of the ten patients whose samples were re-
peatedly reactive in-house, only one patient’s sample was HEV RNA
Data on baseline characteristics were available from a liver trans- positive at VRD, giving a prevalence of HEV RNA in our population of
plant database for 458 out of the 490 patients (Table 1). Time from 0.20% (95% CI 0.0%–1.1%). Neutralisation testing confirmed antigen
transplantation to inclusion in the study ranged between 12 days and reactivity in one of the nine reactive samples (which was also HEV RNA
28 years. The baseline demographics of patients who had an initially positive) suggesting non-specific binding in the other eight samples
reactive antigen test and repeatedly reactive antigen test were similar to (Table 2). There was a sample from one patient, which although posi-
those who did not. There was no significant difference between the tive on in-house testing, was negative on the HEV antigen test and on
serum haemolysis, icteric and lipaemic indices, peri-operative blood PCR testing at VRD.

46
G. Soothill et al. Journal of Clinical Virology 106 (2018) 44–48

Table 2
The anti-HEV antigen-specific ELISA test results of the initially reactive and repeatedly reactive patients are shown expressed as a ratio of the individual absorbance (S)
to the cut-off (CO) (S/CO). The PCR test results show that there was one true positive result. The neutralisation test results show that the true positive neutralised
whereas none of the other reactive results did suggesting non-specific binding (* sample diluted 1:300 for neutralisation; ** sample diluted 1:2 for neutralisation; ¥
neutralisation performed when HEV antigen reactive at reference laboratory).
Study no. HEV-Ag result 1 (S/CO1) HEV-Ag result 2 (S/CO2) Interpretation Reference lab PCR result Reference lab HEV-Ag result Neutralisation (%) ¥

0008 1.71 reactive 7.14 Repeatedly reactive Not detected 1.17 reactive 4.24
reactive
0059 2.73 reactive 1.93 reactive Repeatedly reactive Not detected 1.19 reactive −2.96
0241 2.32 reactive 2.86 reactive Repeatedly reactive Not detected 0.92 equivocal −27.74
0242 11.77 10.22 Repeatedly reactive Detected 17.29 103.35**
reactive reactive reactive
0292 9.05 reactive 9.79 Repeatedly reactive Not detected 15.86 reactive 5.01**
reactive
0294 8.37 reactive 9.19 Repeatedly reactive Not detected 4.56 reactive 13.73
reactive
0297 4.22 reactive 3.89 reactive Repeatedly reactive Not detected 1.19 reactive 13.56
0391 2.04 reactive 8.52 Repeatedly reactive Not detected 0.25 non-reactive –
reactive
0452 2.77 reactive 6.15 reactive Repeatedly reactive Not detected 1.21 reactive 38.54
0501 1.82 reactive 1.52 reactive Repeatedly reactive Not detected 1.42 reactive 11.65
0099 0.92 equivocal 0.60 non-reactive Initially reactive Not detected 0.26 non-reactive –
0123 1.43 reactive 0.69 non-reactive Initially reactive Not detected 0.52 non-reactive –
0234 3.34 reactive 0.68 non-reactive Initially reactive Not detected 0.00 non-reactive –
0239 1.65 reactive 0.96 equivocal Initially reactive Not detected 0.33 non-reactive –
0502 10.88 reactive 0.07 non-reactive Initially reactive Not detected 0.00 non-reactive –

Table 3
Comparison of the anti-HEV antigen-specific ELISA test versus the reference
real time PCR test for diagnosis of HEV (definition of a positive test being re-
peatedly reactive HEV antigen test).
HEV PCR HEV PCR Total
positive negative

HEV Ag test positive 1 (True 9 (False 10 (Total

(repeatedly reactive) positives) positives) positives)
HEV Ag test negative 0 (False 480 (True 480 (Total
(not repeatedly negatives) negatives) negatives)
reactive)
Total 1 489 490

Fig. 1. Graph showing that HEV antigen ELISA S/CO of the HEV positive pa-
Using the definition of a positive test as repeatedly reactive tests, the
tient fell with reduction in HEV RNA but remained detectable when HEV RNA
sensitivity was 100.0% and specificity was 98.2%. This gives a positive was not detectable. Each dot represents an individual serum sample from the
predictive value (PPV) of 10.0% and a negative predictive value (NPV) HEV infected patient in our dataset, which underwent HEV antigen ELISA and
of 100.0% in our population (Table 3). PCR testing.
The one patient who was found to have active HEV infection pre-
sented feeling generally unwell with abnormal liver biochemistry (bi-
Table 4
lirubin 198 μmol/L, aspartate transaminase (AST) 250 IU/L, alanine
The HEV antigen ELISA S/CO of the HEV positive patient fell with reduction in
transaminase (ALT) 211 IU/L). The patient was eight years post liver
HEV RNA. The liver biochemistry (bilirubin, AST and ALT) also improved over
transplant with a background of alcoholic liver disease. On admission this period.
the patient was investigated for cause of graft dysfunction and diag-
Date of HEV-Ag Reference lab PCR Bilirubin AST ALT
nosed with HEV infection. Subsequently, the patient was treated with a
sample result (S/ result (IU/ml) (μmol/L) (IU/L) (IU/L)
course of ribavirin as an outpatient. The sub-genotype of HEV was 3e. CO)
Serial HEV RNA and antigen testing were performed over the following
four months (Fig. 1). Of note, the HEV antigen remained detectable 26/04/17 10.8 5.05E+04 198 250 211
even after the PCR tests became negative. Reduction of the viral load in reactive
02/05/17 11.1 8.93E+02 122 217 186
response to treatment corresponded with a decrease in the antigen test reactive
reactive value and a normalisation of the patient’s liver biochemistry 11/05/17 10.8 Detected below 72 153 113
(Table 4). reactive limit of
The liver biochemistry tests of the 15 patients who had an initially quantitation
26/05/17 10.6 Not detected 47 77 58
reactive antigen test were examined to check for any sign of graft dys-
reactive
function, which might reflect a chronic HEV infection, even in the 14/07/17 6.5 Not detected 33 93 70
context of a negative HEV RNA test. Apart from the patient who was reactive
found to be HEV antigen as well as RNA positive, only one other initially 28/07/17 5.2 Not detected 23 91 65
reactive patient had derangement of liver biochemistry. This patient was reactive

diagnosed with a recurrence of primary sclerosing cholangitis and the

graft function improved with ursodeoxycholic acid and adjustment of
immunosuppression.

47
G. Soothill et al. Journal of Clinical Virology 106 (2018) 44–48

4. Discussion analysis under the supervision of DI. MA and SI performed the hepatitis
PCR and neutralisation testing at PHE. DT and PG advised on study
The HEV antigen-specific ELISA test performed well in this UK post design and preliminary results. GS wrote the initial manuscript. All
liver transplant population. There was a relatively low prevalence of authors edited and approved the final manuscript.
HEV infection in this cohort; however, as expected, the prevalence was
higher than in the general population. Conflicts of interests
The manufacturer’s definition of a positive antigen result (at least
two reactive tests) improved the specificity of the test. There was an There are no conflicts of interests.
acceptable level of false positive results, which were likely due to non-
specific reactivity, as demonstrated by the neutralisation testing. Funding
Interestingly, the antigen testing for the patient with RNA PCR con-
firmed HEV remained positive even after HEV RNA became un- Royal Free Charity.
detectable. Hence, the HEV antigen can persist following RNA clear-
ance, a phenomenon reported previously [21]. It has also been References
described that some HEV antigen screening tests can detect persistent
free forms of antigen in the serum, not associated with an infectious [1] M.S. Balayan, Evidence for a virus in non-A, non-B hepatitis transmitted via the
viral particle [22]. In cases of ELISA positive, PCR negative samples, fecal-oral route, Intervirology 20 (1) (1983) 23–31.
[2] R. Lozano, M. Naghavi, K. Foreman, et al., Global and regional mortality from 235
neutralisation testing can help determine whether this is due to recent/ causes of death for 20 age groups in 1990 and 2010: a systematic analysis for the
previous HEV infection or non-specific reactivity [21]. Global Burden of Disease Study 2010, Lancet 380 (2012) 2095–2128.
Due to the low prevalence of HEV in our population the PPV of the [3] J.H. Hoofnagle, Hepatitis E, NEJM 367 (13) (2012) 1237–1244.
[4] D. Lapa, M.R. Capobianchi, A.R. Garbuglia, Epidemiology of hepatitis E virus in
antigen test was low (10.0%), despite a specificity of 98.2%; hence only european countries, Int. J. Mol. Sci. 16 (10) (2015) 25711–25743.
one in ten of those defined as positive on antigen testing had PCR [5] B. Said, Hepatitis E virus in England and Wales: indigenous infection is associated
confirmed HEV. Therefore, if used as a screening tool all positive tests with the consumption of processed pork products, Epidemiol. Infect. 142 (7) (2014)
1467–1475.
would require confirmation with HEV RNA testing and/or the con-
[6] S. Mirazo, N. Ramos, V. Mainardi, et al., Transmission, diagnosis and management
firmatory neutralisation testing used in this study. The PPV would be of hepatitis E: an update, Hepat. Med. 6 (2014) 45–59.
less in a lower risk population. Hence, the validity of this test in other [7] S.A. Chaudhry, N. Verma, G. Koren, Hepatitis E infection during pregnancy, Can.
Fam. Physician 61 (7) (2015) 607–608.
populations of immunocompromised individuals, such as other organ
[8] N. Kamar, J. Izopet, H.R. Dalton, Chronic Hepatitis E Virus Infection and Treatment,
transplant recipients (renal and HSCT), would be dependent on pre- J. Clin. Exp. Hepatol. 3 (2) (2013) 134–140.
valence of HEV in the relevant populations. There were no false nega- [9] Y. Geng, H. Zhang, W. Huang, et al., Persistent hepatitis E virus genotype 4 infection
tive results giving a NPV of 100.0% in this cohort. This result supports in a child with acute lymphoblastic leukemia, Hepat. Mon. 14 (1) (2014) e15618.
[10] N. Kamar, H.R. Dalton, F. Abravanel, et al., Hepatitis E virus infection, Clin.
the utility of anti-HEV antigen ELISA as a screening tool; however, with Microbiol. Rev. 27 (1) (2014) 116–138.
only one true positive case in this cohort the evaluation of the sensi- [11] S. Lhomme, Hepatitis E virus quasispecies and the outcome of acute hepatitis E in
tivity of the antigen test in this study is not reliable. solid organ transplant patients, J. Virol. 86 (2012) 10006–10014.
[12] A.R. Murali, Chronic hepatitis e: a brief review, World J. Hepatol. 7 (19) (2015)
The sub-genotype of the HEV strain of the positive patient was 3e. 2194–2201.
There may be discrepancies in the sensitivity and specificity of the [13] L.J. Krain, K.E. Nelson, A.B. Labrique, Host immune status and response to hepatitis
ELISA test depending on the genotype or the sub-genotype of HEV. The e virus infection, Clin. Microbiol. Rev. 27 (1) (2014) 139–165.
[14] M.S. Faber, Hepatitis E virus seroprevalence among adults, Germany, Emerg. Infect.
parameters of the ELISA test according to HEV genotypes should be Dis. 18 (10) (2012) 1654–1657.
evaluated in future studies. [15] P.E. Hewitt, S. Ijaz, S.R. Brailsford, et al., Hepatitis E virus in blood components: a
We estimate that ELISA testing would cost 5–10 times less than PCR prevalence and transmission study in southeast England, Lancet 384 (9956) (2014)
1766–1773.
testing, although this requires a formal cost analysis. However, from a [16] M. Riveiro-Barciela, M. Buti, M. Homs, et al., Cirrhosis, liver transplantation and
practical viewpoint, larger laboratories set up with automated serology HIV infection are risk factors associated with hepatitis E virus infection, PLoS One 9
machines may not wish to switch to a plate-based serology assay. The (2014) e103028.
[17] N. Kamar, C. Garrouste, E.B. Haagsma, et al., Factors associated with chronic he-
addition of a neutralisation step improves the specificity but this
patitis in patients with hepatitis E virus infection who have received solid organ
"homemade" step would be difficult to standardise, and so difficult to transplants, Gastroenterology 140 (2011) 1481–1489.
extend to all the laboratories using the method, which is a limitation. [18] P. Behrendt, B. Bremer, D. Todt, et al., Hepatitis E virus (HEV) ORF2 antigen levels
Nevertheless, this assay offers an alternative to PCR where molecular differentiate between acute and chronic HEV infection, J. Infect. Dis. 214 (3) (2016)
361–368.
testing is not routinely available or prohibitively expensive. Further [19] Jothikumar, et al., A broadly reactive one-step real-time RT-PCR assay for rapid and
investigation, including cost analysis, seems justified to determine the sensitive detection of hepatitis E virus, J. Virol. Methods 131 (1) (2006) 65–71.
efficacy of anti-HEV antigen-specific ELISA testing in a screening con- [20] Garson, et al., Minor groove binder modification of widely used TaqMan probe for
hepatitis E virus reduces risk of false negative real-time PCR results, J. Virol.
text and clinical investigation of HEV infection. Methods 186 (1-2) (2012) 157–160.
[21] M.J. Ankcorn, S. Ijaz, B. Haywood, et al., Confirmation of specificity of reactivity in
Authors’ contributions a solid phase ELISA for the detection of hepatitis E viral antigen improves utility of
the assay, J. Virol. Methods 252 (2017) 42–48.
[22] C. Montepellier, C. Wychowski, I.M. Sayed, Hepatitis e virus lifecycle and
GS, SH and ME conducted the study including the collection of identificación of 3 forms of the ORF2 capsid protein, Gastroenterology 154 (1)
samples, performance of HEV antigen-specific ELISA testing and data (2018) 211–223.

48