Human Immunology xxx (2016) xxxxxx

Contents lists available at ScienceDirect

www.ashi-hla.org

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

Validation of a novel real-time PCR assay for detection of HLA-B*15:02

allele for prevention of carbamazepine Induced Stevens-Johnson
syndrome/Toxic Epidermal Necrolysis in individuals of Asian ancestry
Dinh Van Nguyen a,b,c,, Christopher Vidal a,, Hieu Chi Chu d, Nga Thi Quynh Do e, Tu Thi Linh Tran f,
Huong Thi Minh Le g, Richard B. Fulton a, Jamma Li a,h, Suran L. Fernando a,b,h
a
ImmunoRheumatology Laboratory, Pathology North-Sydney, St Leonards, Australia
b
Sydney Medical School Northern, University of Sydney, Sydney, Australia
c
Department of Allergy and Clinical Immunology, Hanoi Medical University, Hanoi, Viet Nam
d
Center of Allergology and Clinical Immunology, Bach Mai Hospital, Hanoi, Viet Nam
e
Department of Immunology and Molecular Biology, National Institute of Hygiene and Epidemiology, Hanoi, Viet Nam
f
Hanoi Heart Hospital, Hanoi, Viet Nam
g
Department of Allergy, Immunology and Rheumatology, National Hospital of Pediatrics, Hanoi, Viet Nam
h
Department of Clinical Immunology and Allergy, Royal North Shore Hospital, Sydney, Australia

a r t i c l e i n f o a b s t r a c t

Article history: Objectives: Screening for the HLA-B*15:02 allele has been recommended to prevent carbamazepine (CBZ)
Received 9 June 2016 induced Stevens-Johnson syndrome (SJS) and Toxic Epidermal Necrolysis (TEN) in individuals with
Revised 11 August 2016 Asian ancestry. We aimed, therefore, to develop and validate a robust and inexpensive method for detec-
Accepted 16 August 2016
tion of the HLA-B*15:02 allele.
Available online xxxx
Methods: Real-time PCR using TaqMan probes followed by SYBR Green was used to detect the HLA-
B*15:02 allele prior to treatment with CBZ therapy.
Keywords:
Results: A total of 121 samples were tested. The assay has a sensitivity of 100% (95% CI: 76.84100.0%), a
Carbamazepine hypersensitivity
HLA-B antigens
specificity of 100% (95% CI: 96.61100%), a positive predictive value of 100% (95% CI: 76.84100%) and a
Real-time polymerase chain reaction negative predictive value of 100.0% (95% CI: 96.61100.0%), respectively. There was 100% agreement
Stevens Johnson syndrome between our results and genotyping using Luminex SSO/SBT/SSP. The lowest limit of detection of the
Toxic Epidermal Necrolysis TaqMan probe is 0.05 ng/ll and the SYBR Green is 0.5 ng/ll of DNA. The unit cost of using the
Severe cutaneous adverse reactions TaqMan probe followed by SYBR Green is only $4.7 USD.
SCARs Conclusion: We developed a novel assay for the detection of the HLA-B*15:02 allele, which is robust,
inexpensive and suitable for screening individuals of Asian ancestry in the prevention of CBZ-induced
SJS/TEN.
2016 American Society for Histocompatibility and Immunogenetics. Published by Elsevier Inc. All rights
reserved.

1. Introduction can cause a variety of hypersensitivity reactions in 510% of indi-

Carbamazepine (CBZ) is classified as an aromatic amino-
anticonvulsant. It has been in use since 1965 and is commonly pre-
scribed for multiple indications such as epilepsy, mental health
disorders and neuropathic pain [1,2]. This medication, however,
Corresponding author at: Sydney Medical School Northern, University of
Sydney, Sydney, Australia (D. Van Nguyen). ImmunoRheumatology Laboratory,
Pathology North-Northern Sydney, RNSH, St Leonards, NSW 2065, Australia
(C. Vidal).
E-mail addresses: vngu3162@uni.sydney.edu.au (D.V. Nguyen), christophervid@
gmail.com (C. Vidal).
viduals treated. The most serious of these are severe cutaneous
adverse reactions (SCARs), such as Stevens-Johnson syndrome
(SJS) and Toxic Epidermal Necrolysis (TEN) which have been
observed to be most prevalent in Asian populations. SJS/TEN cause
significant mortality (1050%) and morbidity (60%) with severe
ocular complications, alopecia, mucosal alteration of the gastroin-
testinal and respiratory tracts, and psychological sequelae [3].
The HLA-B*15:02 allele appears to be the major genetic suscep-
tibility factor for CBZ-induced SJS/TEN. A strong genetic association
was detected in a cohort of Han Chinese in Taiwan where HLA-
B*15:02 allele was found in 100% of 44 patients with CBZ-SJS/
TEN patients (OR: 2504 [95% CI: 126-49552]) [4]. These findings

http://dx.doi.org/10.1016/j.humimm.2016.08.010
0198-8859/ 2016 American Society for Histocompatibility and Immunogenetics. Published by Elsevier Inc. All rights reserved.

Please cite this article in press as: D.V. Nguyen et al., Validation of a novel real-time PCR assay for detection of HLA-B*15:02 allele for prevention of car-
bamazepine Induced Stevens-Johnson syndrome/Toxic Epidermal Necrolysis in individuals of Asian ancestry, Hum. Immunol. (2016), http://dx.doi.org/
10.1016/j.humimm.2016.08.010
2 D.V. Nguyen et al. / Human Immunology xxx (2016) xxxxxx
have been replicated in an extended cohort of subjects of Chinese
descent from various geographic areas of China [58], Taiwan [9],
Hong Kong [10], Thailand [11,12], India [13], Malaysia [14], Singa-
pore [15] and Vietnam [16].
A weak association between CBZ-induced SJS/TEN and HLA-
A*31:01 has been reported in a meta-analysis by Genin et al.
[17], with a pooled OR of 3.94 (95% CI: 1.4-11.5), however, this
association has been mainly observed in Caucasians [1719]and
Japanese [20].
In order to reduce the mortality and morbidity due to SJS/TEN
caused by CBZ, in 2007, the Food and Drug Administration (FDA)
[21] recommended that physicians screen patients of Asian
backgrounds for HLA-B*15:02 prior to prescribing CBZ. In 2011,
Chen et al. [22] screened 4877 patients for HLA-B*15:02 prior to
commencing CBZ. Based upon the historical prevalence of CBZ-
induced SJS/TEN in Taiwan, their results indicated that approxi-
mately 10 cases of SJS/TEN were prevented by screening.
In order to be applied in routine clinical practice, a screening
assay has to be simple, rapid and inexpensive [23]. In this consid-
eration, there has been a number of proposed methods that are
potentially used for HLA screening prior to treatment with CBZ.
However, these approaches still have some limitations incompati-
ble with a screening test. In 2009, Cheng et al. [24] introduced a
new technique for detection of the HLA-B*1502 allele using
loop-mediated isothermal amplification (LAMP). This approach
has some advantages such as low cost and visibility to the naked
eye. However, as being very sensitive, cross-contamination from
sample to sample could occur in LAMP [25], increasing the risk
for false positive results. In 2012, Virakul et al. [26] reported a
method using nested PCR to detect HLA-B*15:02 that can be imple-
mented in clinical practice. However, the major drawback of this
method is the high risk of cross-contamination between samples
in the second PCR due to high concentration of PCR products.
Currently, a commercially available kit (PG1502 Detection kit,
Pharmigene Inc, Taiwan) using real-time PCR with SYBR Green
has been widely employed as a screening test. This kit is, however,
quite expensive ($15 USD/test) and developing countries such as
Vietnam could not afford to implement the screening test nation-
wide to prevent CBZ-induced SJS/TEN. Another minor disadvantage
of the kit is that the PCRs for target gene and control are performed
separately. This approach is not ideal for avoiding potential false
negative results. Finally, the FastGel kit requires more than 25 ng
DNA/reaction, making it difficult to perform PCR on genomic
DNA extracted from low DNA yielding samples such as buccal
swabs. More recently, in May, 2016, Jaruthamsophon et al. [25]
proposed a new in house based, 2 step technique using allele
specific PCR (AS-PCR) followed by direct dot blot hybridization
(DDB). Although this approach is inexpensive, time consumption
is a major limitation of this method in clinical screening for
HLA-B*15:02.
In view of all this, we developed a robust and inexpensive test
using real- time PCR to detect individuals carrying HLA-B*15:02
suitable for use in the prevention of SJS/TEN caused by CBZ in those
of Asian ancestry.
2. Materials and methods
2.1. DNA Samples
A group of samples for optimization and validation (n = 121,
including 14 HLA-B*15:02 positive samples) was randomly col-
lected from those submitted for testing to ImmunoRheumatology
Department, Pathology North, Royal North Shore Hospital, Sydney,
Australia (101 samples) and from Vietnamese individuals in
Sydney and Vietnam (20 samples). Power analysis shows that to
Please cite this article in press as: D.V. Nguyen et al., Validation of a novel real-time PCR assay for detection of HLA-B*15:02 allele for prevention of car-
bamazepine Induced Stevens-Johnson syndrome/Toxic Epidermal Necrolysis in individuals of Asian ancestry, Hum. Immunol. (2016), http://dx.doi.org/
10.1016/j.humimm.2016.08.010
assess the correlation between two methods, with a power of
80% (<alpha> = 0.05; <beta> = 0.20, at least 9 positive individuals
within the group of participants are needed, if the anticipated
correlation coefficient is 0.8. The sample size needed to calculate
the positive and negative predictive values as described by Hanley
et al. [27], using the receiver operating characteristic (ROC) curve,
consists of 9 positive cases and 81 negatives, for a power of 80%
(<alpha> = 0.05; <beta> = 0.20, Area under ROC curve is 0.8, null
hypothesis value is 0.5, ratio of negative/positive is 9). The
estimated prevalence of HLA-B*15:02 positive samples is approxi-
mately 10%. Thus the total number of DNA samples needed for the
validation of the new methodology is at least 90, including 9
positive samples. Peripheral blood (3 ml) was collected into EDTA
anti-coagulated tubes. Participants completed an informed consent
which was approved by both the Northern Sydney Local Health
District HREC, St Leonards, NSW, Australia (HREC/15/HAWKE/86)
and Bach Mai Hospital, Hanoi, Vietnam.
2.2. DNA extraction
Genomic DNA was extracted from peripheral blood leucocytes
using AccPrep Genomic DNA extraction kit (Bioneer Corp, Korea).
DNA concentration and purity were measured using NanoPho-
tometerTM, and the average DNA concentration was approximately
46 ng/ll and A260/280 ratio was over 1.7.
High resolution HLA-B typing using LuminexSSO/SBT/SSP was
performed blindly on each sample at the New South Wales
Transplantation and Immunogenetics Service (Australian Red Cross
Blood Service).
2.3. Real-time PCR with Taqman probe
Using the Immunogenetics project HLA database (IMGT) [28],
we compared the HLA-B*15:02 genomic sequence (IMGT/HLA
Acc no: HLA00165) with other HLA-B allele sequences to identify
a hyperpolymophic region which can be targeted to specifically
amplify the HLA-B*15:02 allele while eliminating the rest of the
HLA-B alleles (Fig. 1). A set of oligonucleotide primers and probe
were designed using IDTs PrimerQuest incorporates Primer3 soft-
ware [29,30] (version 2.2.3) (Integrated DNA Technologies, Inc),
yielding a 76 bp amplicon (Table 1). The sequence of the forward
primer (B15:02-F-Primer) is 50 -GGGCCGGAGTATTGGGA-30 , the
reverse primer (B15:02-R-Primer) is 50 -GGTTCCGCAGGCTCTCT-30
and the probe labelled with FAM (6-corboxylfluorescein) is 50 -CG
GAACACACAGATCTCCAA-30 . A set of primers and probe amplifying
the housekeeping gene actin (ACTB) was also designed using the
above mentioned software, yielding a 107 bp amplicon. The
sequences of primers and probe are ACTB-F 50 -CTG TGC TGT GGA
AGC TAA GT-30 , ACTB-R 50 -GAT GTC CAC GTC ACA CTT CA-30 and
HEX (6-carboxy-20 -4-40 -50 -7-70 -hexachlorofluorescein) labelled
probe 50 -ATG CCT GAG AGG GAA ATG AGG GC-30 , respectively.
All primers and probes were synthesized by Geneworks
(Geneworks Pty Ltd, South Australia).
The optimal multiplex PCR was performed in a total volume of
20 ll, containing input DNA and PCR mixture (10 ll of SsoFastTM
Probes Supermix, 50 pmol of ACTB primer; 5 pmol of ACTB probe;
100 pmol of HLA-B*15:02 primer, 10 pmol of ll HLA-B*15:02
probe and PCR water). The PCRs were performed in Rotor-
Gene-Q Cycler (Qiagen, Hilden, Germany) under the optimal profil-
ing conditions: initial hot-start 95 C for 3 min followed by 40
cycles of 96 C for 15 s, 65 C for 60 s and 72 C for 30 s. A series
of four ten-fold dilutions ranging from 50 to 0.05 ng/ll was pre-
pared to evaluate the efficiency and the lowest limit of detection
of the multiplex real-time PCR with the TaqMan probe.
 D.V. Nguyen et al. / Human Immunology xxx (2016) xxxxxx 3

Fig. 1. Binding sites of the set of primers and comparison of the different allele sequences of the target region. Note: Numbered nucleotide positions as per IMGT/HLA
accession number HLA00165 [28]. The HLA-B*15:02:01 sequence is aligned against the selected alleles that can generate false positive results with both PCRs and against
other frequent alleles in the Asian population. The latter includes HLA-B*46:01, which is highly frequent (11.5%) in the Vietnamese population [35]. This described method
should ensure the elimination of common false positive results method. The primers and probe targeted regions within exon 2 and 3 harbouring hypervariable
polymorphisms. More importantly, the probe binds at the polymorphic region allowing the first real-time PCR with TaqMan Probe amplifies only HLA-B*15:02 and HLA-
B*18, whilst eliminating the rest of the HLA-B alleles. The second PCR with a specific set of primers does not amplify HLA-B*18 by 2 mismatches at 30 prime end of the forward
primer (C and A) and 3 mismatches at 30 prime end of the forward primer (C, A and T).

Table 1
Specific primers and TaqMan probe for detection of HLA-B*15:02.

Name Target gene Sequences (50 -30 ) Amplicon (bps) Tm* (C) Potential amplification
B*15:02 F-Primer HLA-B*15:02 GGGCCGGAGTATTGGGA 76 79.8 HLA-B*18
B*15:02 FAM-Probe CGGAACACACAGATCTCCAA
B*15:02 R-Primer GGTTCCGCAGGCTCTCT
B*15:02 F-SBYR HLA-B*15:02 CGCGAGTCCGAGGATGGC 423 91.6 HLA-B*13:01; 15:13; 15:25; 15:36
B*15:02 R-SBYR GCCATACATCCTCTG GAT
ACTB F-Primer ACTB GATGTCCACGTCACACTT CA 107 81.4 None
ACTB HEX-Probe ATG CCT GAG AGG GAA ATG AGG GC
ACTB R-Primer ATGCCTGAG AGGGAAATG AGGGC
*
Using oligo calculator to estimate the theoretical Tm of PCR products.

2.4. Real-time PCR with SYBR Green
A set of specific primers to amplify the HLA-B*15:02 allele used
for the second real-time PCR with SYBR Green was previously
reported by Yu et al. [31]. The forward primer (B1502-F-SYBR) tar-
gets exon 2 with the sequence of 50 -CGC GAG TCC GAG GAT GGC-30 ,
and the reverse primer (B1502 R-SYBR) targets exon 3 with the
sequence of 50 -GCC-ATA-CAT-CCT-CTG-GAT-30 . The amplicon is
423 bps and melting temperature (Tm) at 91.6 C. The same set
of primers used in the TaqMan probe master mix was also used
for the SYBR PCR. The Tm of amplicon is 81.4 C. To calculate reac-
tion efficiencies, the singleplex PCRs for HLA-B*15:02 and ACTB
were performed in a series of five ten-fold dilutions ranging from
50 to 0.005 ng/ll. The optimal annealing temperature for both sets
of primers is 63 C. The PCR products were confirmed by elec-
trophoresis on 1.2% agarose gel (FlashGel Cassette-Lonza, USA)
and by direct DNA sequencing using BigDye terminators v3.1
(Applied Biosystems).
A multiplex PCR was performed in a total volume of 25 ll, con-
sisting of genomic DNA and PCR master mix (12.5 ll of SensiMixTM-
SYBR No-ROX Kit, 35 pmol of HLA-B*15:02-specific primers,
Please cite this article in press as: D.V. Nguyen et al., Validation of a novel real-time PCR assay for detection of HLA-B*15:02 allele for prevention of car-
bamazepine Induced Stevens-Johnson syndrome/Toxic Epidermal Necrolysis in individuals of Asian ancestry, Hum. Immunol. (2016), http://dx.doi.org/
10.1016/j.humimm.2016.08.010
20 pmol of ACTB primers and PCR water). PCRs were performed
in a Rotor-Gene-Q cycler (Qiagen, Hilden, Germany) under the fol-
lowing optimal conditions: initial hot-start 95 C for 3 min fol-
lowed by 40 cycles of 96 C for 15 s, 63 C for 15 s and 72 C for
60 s. The lowest limit of detection of multiplex PCR was deter-
mined by running in different concentrations of DNA (Supplemen-
tary data 1).
2.5. Data analysis
Raw PCR data were analysed using Rotor-Gene v6 software
(Qiagen, Hilden, Germany). Analyses for sensitivity, specificity,
positive predictive value (PPV) and negative predictive value
(NPV) were performed using MedCalc v12.7.0.0. A Cohens Kappa
for measuring agreement between our assay and the Luminex
SSO/SBT/SSB genotyping was analysed using SPSS software version
22 (Chicago, USA).
Direct sequencing data obtained from Applied Biosystems 3130
Genetic Analyser were analysed using Chromas software 2.5.0
(Applied Biosystems).
4 D.V. Nguyen et al. / Human Immunology xxx (2016) xxxxxx

3. Results
3.1. Optimization of real-time PCR with TaqMan probe
For the optimization stage, we performed testing on positive
controls with the HLA-B*15:02 allele and a negative control to
identify an annealing temperature for each set of primers and a
probe for the target and reference genes, respectively. The optimal
annealing temperature for HLA-B*15:02 primers and probe was
65 C. The best ratio of final concentrations of primers to probe
in the reaction was 10:1. In order to optimize a multiplex PCR reac-
tion, different ratios of primers and probe concentrations were
tested. The best results were observed when the concentration of
ACTB primers and probe were half those of HLA-B*15:02 (for
HLA-B*15:02: efficiency: 92%; R2: 0.993; slope: 3.53; lowest Ct:
23.6; highest Ct: 34.5, threshold: 0.02. For ACTB: efficiency: 91%,
slope: 3.53; lowest Ct: 21.28; highest Ct: 32.04; threshold:
0.02) (Fig. 3). The lowest limit of detection is 0.05 ng/lL. Total time
for this PCR is 110 min.
3.2. Optimization of real-time PCR with SYBR Green
For the optimization stage, the best annealing temperature was
63 C for both HLA-B*15:02 primers (efficiency: 99%; R2: 0.99;
slope: 3.34; threshold: 0.02) and ACTB primers (efficiency: 90%,
R2: 0.99; slope: 3.58; threshold: 0.02) (Fig. 2). To optimize the
multiplex PCR using SYBR green, different ratios of HLA-B*15:02
and ACTB primers concentrations were also tested. The best ratio
observed was 0.35:0.20 by comparing dissociation curves. The dis-
sociation curve was used to determine HLA-B*15:02 and ACTB
with Tm of 91.4 C and 81.4 C, respectively (Table 1). The thresh-
old for these curves was set at 0.45 to reduce the chance of a false
negative result (Fig. 3). The assay was able to detect the target allele
in a DNA sample with a concentration as low as 0.5 ng/ll of DNA.
Total time for this PCR is 128 min.
3.3. Results of validation
The total number of 121 samples was tested using the method
of two PCRs (Fig. 3). After the first step, we detected 27 positive
samples by using the first real-time PCR with the Taqman Probe
Ct = 24.08 0.41 (Mean Standard Error of Mean). The average Ct
of all samples for ACTB was 22.82 0.21. In the second step, the
positive group was tested by real- time with SYBR Green and only
14 samples carrying HLA-B*15:02 were detected with two peaks of
Tm at 91.31 0.02 C (Mean Standard Error of Mean) (HLA-
B*15:02 allele) and 80.85 0.04 (ACTB). The 13 negative samples
showed only one peak for ACTB (81.31 0.04 C) (Fig. 3). No false
positives or false negatives were observed. When compared with
the results of using the Luminex SSO/SBT/SSP, the first PCR gave
13 false positive samples with genotype of HLA-B*18. No false pos-
itive was observed after the second PCR.
The final result has perfect agreement (k = 1, p < 0.0001) with
Luminex SSO/SBT/SSP. The combined assay has a sensitivity of

Fig. 2. Standard curves of real-time PCR with TaqMan probe and real-time PCR with SYBR Green.
Please cite this article in press as: D.V. Nguyen et al., Validation of a novel real-time PCR assay for detection of HLA-B*15:02 allele for prevention of car-
bamazepine Induced Stevens-Johnson syndrome/Toxic Epidermal Necrolysis in individuals of Asian ancestry, Hum. Immunol. (2016), http://dx.doi.org/
10.1016/j.humimm.2016.08.010
 D.V. Nguyen et al. / Human Immunology xxx (2016) xxxxxx 5

Fig. 4. The diagram and results of validation progress.

method can be used for the development of a screening test with

Fig. 3. Detection HLA-B*15:02. Note: (A) Detection HLA-B*15:02 using TaqMan
Probe (Ct at threshold = 0.02). (B) Detection HLA-B*15:02 using dissociation curves
(melting curves).
100% (95% CI: 76.84100.0%), specificity of 100 % (95% CI: 96.61
100%), PPV of 100% (95% CI: 76.84100%) and NPV of 100.0%
(95% CI: 96.61100.0%), respectively.
In the panel of 107 samples negative for HLA-B*15:02 after two
PCRs, there are 30 different HLA allele groups identified by Lumi-
nex SSO/SBT/SSP. No false negative results were detected in this
group (Supplementary data 2).
4. Discussion
This paper demonstrates the development of a robust and inex-
pensive method for the detection of HLA-B*15:02 allele, combining
realtime PCR with TaqMan probe followed by SYBR Green. This
the aim of preventing CBZ-induced SJS/TEN in susceptible individ-
uals of Asian ancestry.
When compared with Luminex SSO/SBT/SSP, which is currently
used for tissue typing, our PCR using the TaqMan probe success-
fully detected a number of HLA-B*15:02 positive individuals with
an assay sensitivity of 100% (95% CI: 76.84100.0%). Importantly,
we did not observe any false negative results with negative predic-
tive value of 100.0% (95% CI: 96.15100.0%). The HLA-B*18 allele,
however, was also amplified. Using the second PCR (SYBR), only
HLA-B*15:02 samples were detected and none of the HLA-B*18
alleles were amplified. Ideally, the second PCR should be per-
formed to eliminate any false positive samples for HLA-B*18. In a
population with a very low prevalence of HLA-B*18 (Table 2,
Fig. 4) [32], however, the second real-time PCR might not be nec-
essary because the possibility of very rarely occurring false posi-
tives would be acceptable for screening purposes and its
omission would reduce the cost and the time taken to perform
the test.
Our method, using multiplex real-time PCR, has some technical
advantages when compared with the commercially available test-
ing kit (PG1502 Detection kit, Pharmigene Inc, Taiwan) and other
methodologies (nested PCR, AS-PCR with DDB). Multiplexing is
also an optimal way of reducing the incidence of a false negative
result due to low DNA quality or the presence of potential PCR inhi-
bitors. Furthermore, it is more likely to be accurate and robust in

Table 2
Frequency of HLA-B*15:02 and HLA-B*18 in Asian populations.

No. Population HLA-B*15:02 allele HLA-B*18 allele References

1 Vietnam 0.13500.1385 0.0100 Allele frequency net 2015 update
2 China Up to 0.3580 <0.0900* (http://www.allelefrequency.net) [32]
3 Hong Kong 0.1020 0.0010**
4 Taiwan Up to 0.1800 <0.0090
5 Singapore Up to 0.1160 0.0100***
6 Thailand 0.08200.0850 0.03900.0800
7 Malaysia Up to 0.1710 0.02000.1600****
8 India 0.0010-0.0600 <0.0630
*
In China, subpopulations with the frequency of HLA-B*18:01 higher than 0.001 are China Southwest Dai (0.0160; sample size is 124), China Shanxi (0.0230; sample size is
22) and China Uyghur (0.0260, sample size is 19). The rest has less than 0.0090 of HLA-B*18.
**
The frequency refers to selected alleles.
***
The frequency of HLA-B*18 allele is less than 0.010 in Chinese Singaporeans and 0.099 only in Malaysian Singaporeans, respectively.
****
The frequency of HLA-B*18 allele is 0.280 in Malaysia Perak Grik Jehai (sample size is 25).

Please cite this article in press as: D.V. Nguyen et al., Validation of a novel real-time PCR assay for detection of HLA-B*15:02 allele for prevention of car-
bamazepine Induced Stevens-Johnson syndrome/Toxic Epidermal Necrolysis in individuals of Asian ancestry, Hum. Immunol. (2016), http://dx.doi.org/
10.1016/j.humimm.2016.08.010
6 D.V. Nguyen et al. / Human Immunology xxx (2016) xxxxxx

Table 3
Comparing the present method using TaqMan probe with other current screening methods in Asians.

Properties of method Nested PCR LAMP method Commercial kit* AS-PCR and DDB TaqMan probe followed by SYBR
[26] [24] [25] (Our method)
Time consumption Estimated assay 240 min 45 min Approximately 300 min 110 min and 128 min
time 80 min
Time for result 35 min (loading 1 min Time for DCt 30 min (loading 0 min
on gel) calculation or on gel)
loading on gel
Total time 275 min 46 min N/A 330 min 238 min
Cost Estimated cost of $6.5 USD $3.8 USD $15 USD $7 USD $4.7 USD**
reaction unit
Robustness Sensitivity 100% 100% N/A 100% 100%
Lowest limit of N/A N/A 25 ng/reaction N/A 0.05 ng/ll (TaqMan Probe) 0.5 ng/ll
detection (DNA (for SYBR)
concentration)
False positive alleles Rare alleles in Heterozygote of Some (less than HLA-B*15:31 and Heterozygote of HLA-B*18 with any
detected in method HLA-B*15 HLA-B*15:25 and 1% frequent) rare alleles in alleles such as HLA-B*13:01; 15:13;
group B*15:15 alleles HLA-B*15 group 15:25; 15:36

SBT: Sequence based-typing; SSO: Sequence-specific oligonucleotide; LAMP: Loop-mediated isothermal amplification; AS-PCR: Allele specific PCR; DDB: Direct dot blot
hybridization; NPV: Negative predictive value; N/A: not applicable.
*
PG1502 DNA Detection Kit Users Manual.
**
The cost of reaction unit was calculated by cost of DNA extracted kit, master mixes, oligos primers and fluorescent dyes.

the detection of the target PCR product using the melting curves
rather than by calculating the delta Ct or by loading an agarose
gel. Finally, our assay is robust and very sensitive in the detection
of the target allele in samples with very low DNA concentrations
(0.5 ng DNA). By using 2 ll input DNA for each reaction, both PCRs
are able to detect successfully the target allele at only 1 ng. This
allows for DNA to be extracted from specimens obtained by non-
invasive methods such as buccal swabs, which can have a lower
yield of DNA. Buccal swabs have the advantages of being more por-
table, less invasive and less likely to become infected, making them
suitable for DNA sample collection for HLA screening aimed at pre-
venting drug hypersensitivity reactions [33].
In comparison with current screening methods in terms of time
and cost, our method seems to be appropriate for clinical applica-
tion. We calculated the unit cost of our method, including DNA
extraction reagents, master mixes, oligos primers and probe
labelled with fluorescent dyes. It costs only $4.7 USD/sample, being
cheaper than nested PCR, AS-PCR with DDB but slightly more
expensive when compared to LAMP (Table 3). In consideration of
time, even though our method requires two PCRs (TaqMan probe
followed by SYBR), the total time taken of 238 min still makes it
suitable for use in screening, as it will not influence significantly
the overall turnaround time. Our report is only addressing the time
taken for the test to be completed from the time it is received and
accepted at the laboratory to the time the result is issued. This does
not take into account, however, other factors that determine the
overall turnaround time such as the time taken for the sample to
reach the laboratory or the time taken for the result to reach the
clinician after the lab issues the report. As shown in Table 3,
although our method is only longer second to LAMP both method-
ologies are suitable to be used as screening tests since results from
both methods can be obtained within a day after sample reception
and there is no need to wait for a couple of days or weeks to run
the tests in batches.
In their original paper [31], the authors previously reported that
this set of specific primers used for SYBR can also amplify other
alleles such as HLA-B*15:13, B*15:20-21, B*15:25, B*15:36 and
B*13:01. When testing 94 HLA-B*15:02 negative samples using
SYBR Green, only the presence of HLA-B*15:25 and HLA-
B*13:01 alleles resulted in a false positive result. Our assay then
can only give a false positive result in heterozygous individuals,
who harbour the HLA-B*18 allele in combination with any of the
above mentioned HLA alleles.
Please cite this article in press as: D.V. Nguyen et al., Validation of a novel real-time PCR assay for detection of HLA-B*15:02 allele for prevention of car-
bamazepine Induced Stevens-Johnson syndrome/Toxic Epidermal Necrolysis in individuals of Asian ancestry, Hum. Immunol. (2016), http://dx.doi.org/
10.1016/j.humimm.2016.08.010
One limitation of our assays is that both were validated using
one commercially available master mix (SensiMixTM SYBR No-ROX
Kit) and SSOFastTM Probes Supermix (Bio-RAD). Re-validation will
be needed if these kits are discontinued by suppliers. Re-
validation of the assay will also be required prior to implementa-
tion on other real-time PCR cyclers. Finally, another limitation
common to any other existing molecular test for HLA is related
to the continuous need for sequence updates owing to the highly
polymorphic nature of the HLA-B gene. New polymorphisms that
could be present in the primer binding sites of the assay could
result in allele drop-out. These problems were also reported by
Dello Russo et al. [34]. We also suggest that this assay should be
re-validated before it is introduced for screening in other
populations.
In conclusion, we developed and validated a rapid, robust and
inexpensive method to use for screening for the HLA-B*15:02 allele
in the prevention of CBZ-induced SJS/TEN in susceptible Asian
populations.
Funding sources
Pathology North Immunology Trust Fund for funding this pro-
ject and Australia Awards Scholarship for funding Dr. Dinh Van
Nguyen.
Conflicts of interest
None.
Acknowledgements
We would like to acknowledge all the people that contributed
or helped in any way to make this study possible. We would like
to thank Ms Anne Proos for the use of facilities at the Department
of Molecular Genetics, Pathology North Sydney, RNSH, St Leonards.
We would like to thank Doctor Nguyet Minh Nguyen, Ha Thi Thu
Nguyen, Hanh Hong Chu, Huyen Thanh Nguyen for collecting
DNA samples. We are also grateful to the Pathology North
Immunology Trust Fund for funding this project and the Australia
Awards Scholarship program for funding Dr Dinh Van Nguyen.
 D.V. Nguyen et al. / Human Immunology xxx (2016) xxxxxx
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.humimm.2016.
08.010.
References
[1] L.P. Hsieh, C.Y. Huang, Trends in the use of antiepileptic drugs in Taiwan from
2003 to 2007: a population-based national health insurance study, Epilepsy
Res. 96 (12) (2011) 8188.
[2] C.J. Landmark et al., Prescription patterns of antiepileptic drugs in patients
with epilepsy in a nation-wide population, Epilepsy Res. 95 (12) (2011) 51
59.
[3] J.C. Roujeau, R.S. Stern, Severe adverse cutaneous reactions to drugs, N. Engl. J.
Med. 331 (19) (1994) 12721285.
[4] W.H. Chung et al., Medical genetics: a marker for Stevens-Johnson syndrome,
Nature 428 (6982) (2004) 486.
[5] X.T. Wu et al., Association between carbamazepine-induced cutaneous adverse
drug reactions and the HLA-B*1502 allele among patients in central China,
Epilepsy Behav. 19 (3) (2010) 405408.
[6] Q. Wang et al., Association between HLA-B*1502 allele and carbamazepine-
induced severe cutaneous adverse reactions in Han people of southern China
mainland, Seizure 20 (6) (2011) 446448.
[7] Y. Zhang et al., Strong association between HLA-B*1502 and carbamazepine-
induced Stevens-Johnson syndrome and toxic epidermal necrolysis in
mainland Han Chinese patients, Eur. J. Clin. Pharmacol. 67 (9) (2011) 885887.
[8] D. Sun et al., Association of HLA-B*1502 and *1511 allele with antiepileptic
drug-induced Stevens-Johnson syndrome in central China, J. Huazhong Univ.
Sci. Technol. Med. Sci. 34 (1) (2014) 146150.
[9] S.I. Hung et al., Genetic susceptibility to carbamazepine-induced cutaneous
adverse drug reactions, Pharmacogenet. Genomics 16 (4) (2006) 297306.
[10] P.K. Kwan, M.H. Ng, S.V. Lo, Association between HLA-B*15:02 allele and
antiepileptic drug-induced severe cutaneous reactions in Hong Kong Chinese:
a population-based study, Hong Kong Med. J. 20 (Suppl. 7) (2014) 1618.
[11] C. Locharernkul et al., Carbamazepine and phenytoin induced Stevens-Johnson
syndrome is associated with HLA-B*1502 allele in Thai population, Epilepsia
49 (12) (2008) 20872091.
[12] W. Tassaneeyakul et al., Association between HLA-B*1502 and
carbamazepine-induced severe cutaneous adverse drug reactions in a Thai
population, Epilepsia 51 (5) (2010) 926930.
[13] T.Y. Mehta et al., Association of HLA-B*1502 allele and carbamazepine-induced
Stevens-Johnson syndrome among Indians, Indian J. Dermatol. Venereol.
Leprol. 75 (6) (2009) 579582.
[14] C.C. Chang et al., Association of HLA-B*1502 allele with carbamazepine-
induced toxic epidermal necrolysis and Stevens-Johnson syndrome in the
multi-ethnic Malaysian population, Int. J. Dermatol. 50 (2) (2011) 221224.
[15] K.W. Chong et al., Association of carbamazepine-induced severe cutaneous
drug reactions and HLA-B*1502 allele status, and dose and treatment duration
Please cite this article in press as: D.V. Nguyen et al., Validation of a novel real-time PCR assay for detection of HLA-B*15:02 allele for prevention of car-
bamazepine Induced Stevens-Johnson syndrome/Toxic Epidermal Necrolysis in individuals of Asian ancestry, Hum. Immunol. (2016), http://dx.doi.org/
10.1016/j.humimm.2016.08.010
7
in paediatric neurology patients in Singapore, Arch. Dis. Child. 99 (6) (2014)
581584.
[16] D.V. Nguyen et al., HLA-B*1502 and carbamazepine-induced severe cutaneous
adverse drug reactions in Vietnamese, Asia Pac. Allergy 5 (2) (2015) 6877.
[17] E. Genin et al., HLA-A*31:01 and different types of carbamazepine-induced
severe cutaneous adverse reactions: an international study and meta-analysis,
Pharmacogenomics J. 14 (3) (2014) 281288.
[18] M. McCormack et al., HLA-A*3101 and carbamazepine-induced
hypersensitivity reactions in Europeans, N. Engl. J. Med. 364 (12) (2011)
11341143.
[19] U. Amstutz et al., HLA-A 31:01 and HLA-B 15:02 as genetic markers for
carbamazepine hypersensitivity in children, Clin. Pharmacol. Ther. 94 (1)
(2013) 142149.
[20] T. Ozeki et al., Genome-wide association study identifies HLA-A*3101 allele as
a genetic risk factor for carbamazepine-induced cutaneous adverse drug
reactions in Japanese population, Hum. Mol. Genet. 20 (5) (2011) 10341041.
[21] P.B. Ferrell Jr., H.L. McLeod, Carbamazepine, HLA-B*1502 and risk of Stevens-
Johnson syndrome and toxic epidermal necrolysis: US FDA recommendations,
Pharmacogenomics 9 (10) (2008) 15431546.
[22] P. Chen et al., Carbamazepine-induced toxic effects and HLA-B*1502 screening
in Taiwan, N. Engl. J. Med. 364 (12) (2011) 11261133.
[23] R. Pavlos, S. Mallal, E. Phillips, HLA and pharmacogenetics of drug
hypersensitivity, Pharmacogenomics 13 (11) (2012) 12851306.
[24] S.H. Cheng et al., New testing approach in HLA genotyping helps overcome
barriers in effective clinical practice, Clin. Chem. 55 (8) (2009) 15681572.
[25] K. Jaruthamsophon et al., Comparison of a new in-house and three published
HLA-B*15:02 screening methods for prevention of carbamazepine-induced
severe drug reactions, PLoS One 11 (5) (2016) e0155907.
[26] S. Virakul et al., A nested sequence-specific primer-polymerase chain reaction
for the detection of HLA-B*15:02, Tissue Antigens 79 (4) (2012) 295301.
[27] J.A. Hanley, B.J. McNeil, The meaning and use of the area under a receiver
operating characteristic (ROC) curve, Radiology 143 (1) (1982) 2936.
[28] J. Robinson et al., The IMGT/HLA database, Nucleic Acids Res. 41 (2013)
D1222D1227 (Database issue).
[29] T. Koressaar, M. Remm, Enhancements and modifications of primer design
program Primer3, Bioinformatics 23 (10) (2007) 12891291.
[30] A. Untergasser et al., Primer3-new capabilities and interfaces, Nucleic Acids
Res. 40 (15) (2012) e115.
[31] N. Yu et al., Typing of HLA-B*15 alleles using sequence-specific primers, Tissue
Antigens 52 (3) (1998) 260269.
[32] F.F. Gonzalez-Galarza et al., Allele frequency net, Update: new features for HLA
epitopes, KIR and disease and HLA adverse drug reaction associations, Nucleic
Acids Res. 43 (2015) D784D788 (Database issue).
[33] C. Badulli et al., A new approach to safely type for HLA the HIV infected people
eligible to abacavir therapy: saliva or buccal swab as reliable DNA sources,
Clin. Chim. Acta 412 (2122) (2011) 19951998.
[34] C. Dello Russo et al., Detection of HLA-B*57:01 by real-time PCR:
implementation into routine clinical practice and additional validation data,
Pharmacogenomics 15 (3) (2014) 319327.
[35] B.K. Hoa et al., HLA-A, -B, -C, -DRB1 and -DQB1 alleles and haplotypes in the
Kinh population in Vietnam, Tissue Antigens 71 (2) (2008) 127134.