JMM
False positive influenza A and B detections in clinical samples due
to contamination with live attenuated influenza vaccine
Curran et al. (2012) described the detection
of inactivated influenza A vaccine (IAV)
contamination in clinical samples being
sent for respiratory virus PCR testing. The
authors concluded that the administration
of IAV and clinical sampling in the same
room resulted in vaccine strains
contaminating surveillance swabs collected
from patients with influenza-like illness
(ILI)/respiratory illness. Herein we describe
our recent similar experience in Scotland
where we detected the recently licensed live
attenuated influenza vaccine (LAIV) in
clinical samples taken from patients with ILI.
Between October and November 2013, the
West of Scotland Specialist Virology
Centre in Glasgow detected 14 throat swab
samples which, using real-time PCR
methods (Gunson & Carman, 2011),
contained low level [cycle threshold (Ct)
.35] influenza positive signals (see Table
1). Some of the samples were positive for a
single influenza A or B virus whereas
others were positive for .1 influenza virus.
Eight of the 14 samples were also positive
for non-influenza respiratory viruses. The
14 samples were all collected from
individual patients that had been sent to us
from 10 different GP practices. All GP
practices were participating in the Scottish
GP spotter influenza surveillance system
(HPS, 2013a) and were situated in six
different Scottish Health Boards (including
Greater Glasgow and Clyde, Fife, Tayside,
Lothian and Ayrshire and Arran). On
repeat testing all sample were positive and
no negative controls were detected as
influenza positive during this time.
At the time of testing, influenza A and B
activity was very low and detections were
not being reported through the Scottish
GP spotter surveillance or in other
European countries (PHE, 2013). This and
the fact that all the positive detections were
very weak and that some contained .1
target suggested to us that these low
positive results may not represent ‘true’
influenza infection and may in fact be the
466 DOI 10.1099/jmm.0.000035
C orrespondence
result of contamination from recent
vaccinations. In Scotland the influenza
vaccinations began in October and
continued throughout the month.
Furthermore, this year differed from
previous years in that it was the first time that
all 2- and 3-year-old children in Scotland
had been offered the LAIV (HPS, 2013b).
We investigated whether these detections
were actually due to LAIV by testing the 14
samples using a recently published LAIV
specific real-time PCR assay (Shcherbik
et al., 2014) (see Table 1). This assay
targets the matrix region of the influenza A
component of the LAIV. In our hands the
assay was highly sensitive with the same
end point detection limit as our generic
influenza A assay. The assay was also found
to be specific for H2N2 when challenged
with various influenza A subtypes, water
controls (n548) and other non-influenza
respiratory pathogens. Of the 14 samples
suspected of containing LAIV seven were
positive by the H2N2 specific assay. This
included one sample that on the initial
influenza screen was weakly positive for
influenza B only. Rather than consider this
a false positive H2N2 result we believe that
this result actually represents a false
negative influenza A matrix test result. The
failure of the matrix assay to detect this
weak sample as positive is not surprising as
tests with similar detection limits can often
give discrepant results when testing such
weakly positive samples. The remaining
seven samples were negative. This may
have been because the samples contained
LAIV contamination at levels beyond the
cut off of the LAIV assay. However, we
cannot rule out the possibility that these
samples contained wild-type virus or
perhaps IAV derived virus. To clarify
testing these samples were subject to
Sanger or next-generation sequencing.
However, both methods were unable to
detect and therefore sequence the virus.
Vaccine history was available for 12 of the
14 patients. Of the 12 patients with a
G 2015 belongs to the NHS Greater Glasgow and Clyde
vaccine history, three were given IAV and
the remaining nine had not received any
influenza vaccination. The two patients
without a vaccine history were 29 and
75 year olds and therefore would not have
been offered LAIV. Since none of the
individuals were given LAIV we can rule
out the positives being a result of vaccine
replication/shedding post vaccination. The
false-positive results may be due to
transmission of LAIV from recently
vaccinated individuals as previous studies
have also shown that live vaccine can be,
albeit rarely, transmitted between those
given vaccine and close contacts. However,
none of the 12 patients had known contact
with a child who had been recently
vaccinated with LAIV.
We feel the most likely explanation for
these results is that the intranasal
administration of LAIV allows large
amounts of influenza A and B virus to be
aerosolised, which in turn can contaminate
the local environment. Subsequently,
samples taken from patients attending with
respiratory infection can then become
contaminated with LAIV. To give an
indication of the level of influenza virus
present in LAIV the control used in this
study was an RNA extract of LAIV. This
extract was tested and found to have Ct
,10. We contacted each of the GP
practices and found that all had been
carrying out LAIV vaccinations in children
on site, and all stated that it was likely that
vaccinations and sampling had taken place
in the same room. In the paper by Curran
et al. (2012), they describe the detection of
IAV in environmental samples taken from
areas where vaccination clinics had taken
place, and they showed that vaccine could
be detectable on surfaces for up to 66 days.
Environmental sampling was not carried
out in this instance but would have been
informative.
To conclude, our results highlight those
laboratories using highly sensitive real-
time PCR methods should be aware of the
Printed in Great Britain
 Correspondence

risk of LAIV contamination occurring

within clinical samples taken at the time of

interpretation

Unknown
Unknown

Unknown

Unknown
Unknown

Unknown
Unknown
vaccination protocols/roll out. Public

Vaccine
Vaccine

Vaccine

Vaccine
Vaccine

Vaccine
Vaccine
Vaccine
Final
health bodies should also be aware of this
issue when interpreting laboratory
surveillance data. Laboratories should
consider using a LAIV-specific assay to

Parainfluenza 2
Rhinovirus confirm all weak influenza A or B positive

Rhinovirus
Rhinovirus
Rhinovirus
Rhinovirus

Rhinovirus
Rhinovirus
detection
Other

results, especially those detected

–
N

N
N

N
N
N
concurrently with vaccination
programmes. Using such a test will ensure
that this type of contamination will not
affect public health surveillance data or
H2N2 (vaccine)
specific assay

patient management and will prevent

36.77
39.34

38.74

38.06
35.42

36.19
28.02
9.18
unnecessary laboratory expenses should
N
N

N
N

N
N
such results be mistaken for PCR
contamination. The results also highlight
that services offering influenza
vaccinations should take special care to
decontaminate the environment prior to
(generic test)

sampling patients with respiratory illness.

Flu B

36
34

35
27
–
–
–
–
–
–
–
–
–
–

Susan Bennett,1 Alasdair R. MacLean,1

Arlene Reynolds,2
Beatrix von Wissmann2
typing assay

and Rory N. Gunson1

H1N1

38

38
37

39
33
36

35
37
37
38
29
–

1
West of Scotland Specialist Virology
Centre, New Lister Building, Glasgow
Royal Infirmary, Level 5, 10–16,
(generic test)

Alexander Parade, Glasgow G31 2ER,

UK
Flu A

39
39

38

34

38
29
–

–
–
–
–
–

2
Health Protection Scotland, NHS
National Services Scotland, Meridian
Court, 5 Cadogan Street, Glasgow
G2 6QE, UK
Known contact with

Correspondence: Susan Bennett

LAIV recipient
Table 1. Results of the 14 weakly positive influenza samples

Unknown

Unknown

(susan.bennett@ggc.scot.nhs.uk)
No
No
No
No
No
No
No
No
No

No

No
No
–

Abbreviations: Ct, cycle threshold; IAV,

inactivated influenza A vaccine; ILI,
influenza-like illness; LAIV, licensed live
attenuated influenza vaccine.
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
GP

Curran, T., McCaughey, C., Ellis, J., Mitchell,

S. J., Feeney, S. A., Watt, A. P., Mitchell, F.,
Vaccinated

Unknown

Unknown

Fairley, D., Crawford, L. & other authors (2012).

IAV

IAV
No

No
No
No
No
No
No
No

No
No
–

False-positive PCR results linked to

administration of seasonal influenza vaccine.
J Med Microbiol 61, 332–338.
Gunson, R. N. & Carman, W. F. (2011). During
(yrs)
Age

15
63
63
43

37
28
26
68
29

75
65
1

6
–

the summer 2009 outbreak of ‘‘swine flu’’ in

Scotland what respiratory pathogens were
LAIV vaccine

diagnosed as H1N1/2009? BMC Infect Dis 11, 192.

Sample ID

Negative.

HPS (2013a). Monthly national influenza

report. Summary of surveillance of influenza
and other seasonal respiratory illnesses.
10
11
12
13
14
1
2
3
4
5
6
7
8
9

N,

Month ending 27 October 2013-week 43.

http://jmm.sgmjournals.org 467
Correspondence
http://www.documents.hps.scot.nhs.uk/
respiratory/seasonal-influenza/flu-update/
2013-10-31.pdf.
HPS (2013b). Scottish vaccine update. Issue 24.
August 2013. http://www.documents.hps.scot.
nhs.uk/immunisation/scottish-vaccine-update/
issue24-2013-08.pdf.
468
PHE (2013). Weekly national influenza
report. Summary of UK surveillance of influenza
and other seasonal respiratory illnesses. 31
October 2013-week 44 report (up to week 43
data). https://www.gov.uk/government/uploads/
system/uploads/attachment_data/file/336644/
National_flu_report_31_October_2013.pdf
Shcherbik, S., Sergent, S. B., Davis, W. G., Shu,
B., Barnes, J., Kiseleva, I., Larionova, N.,
Klimov, A. & Bousse, T. (2014). Application of
real time RT-PCR for the genetic homogeneity
and stability tests of the seed candidates for live
attenuated influenza vaccine production. J Virol
Methods 195, 18–25.
Journal of Medical Microbiology 64