Journal of Clinical Virology - Volume 82 - Supplement 1

S48

Abstracts / Journal of Clinical Virology 82S (2016) S1–S142

provided a positive microorganism identification in 21% (23/106)

of cases where SOC test results were negative.

Conclusion:

Testing of febrile infants with the FilmArray FI

Panel resulted in a pathogen detection in 33% (80/245) of tests,

most of which would not require antibiotic intervention. Addition-

ally, the FilmArray FI Panel detected 17 bacteria directly fromblood

with no upfront enrichment step. These results suggest that the Fil-

mArray FI Panel couldbe a useful systemto rapidly aid in identifying

pathogens causing fever in infants.

This abstract contains information regarding assays that have

not been reviewed by regulatory agencies for

in vitro

diagnostic

use.

http://dx.doi.org/10.1016/j.jcv.2016.08.093

Abstract no: 55

Presentation at ESCV 2016: Poster 54

HIV incidence assays: Evaluation of three HIV

Avidity enzyme immunoassays

J. Hassan

1 ,

∗

, G. Murphy

, F. Reid

, H. Tuite

D. Igoe

, C. De Gascun

National Virus Reference Laboratory, University

College Dublin, Dublin 4, Ireland

Public Health England, London on behalf of the

Consortium for Performance and Evaluation of HIV

Incidence Assays (CEPHIA), United Kingdom

Galway University Hospital, Galway, Ireland

Health Protection Surveillance Centre, Gardiner

Street, Dublin 1, Ireland

Background:

The development of assays for detection of recent

HIV infections is crucial for analysing trends in infection in different

populations for surveillance and prevention measures. Several HIV

Avidity assays have been developed to distinguish between recent

HIV infections and long term established infections.

Objectives:

To identify and validate a suitablemethod for deter-

mining recent HIV infections in Ireland and to incorporate this assay

as a routine diagnostic tool in the NVRL.

Study design

: We compared three currently available manual

HIV avidity enzyme immunoassays: Sedia HIV Limiting Antigen

Avidity assay (LAg), Sedia BED HIV-1 Incidence immunoassay

(BED) and a modified Bio-Rad Genetic Systems HIV-1/HIV-2 plus

O enzyme immunoassay. A total of 50 samples from the Consor-

tium for the Evaluation and Performance of HIV Incidence Assays

(CEPHIA) which included 15 recent and 35 long term HIV-1 infec-

tions were tested. All assays were performed in both screening

(tested in singlet) and confirmatory (tested in triplicate) modes.

The normalised cut-off in the screening assay was OD = 2 and in the

confirmatory assay was OD = 1.5.

Results:

All patients were HIV subtype B. The age range was 23-

64 years and included 47 males. HIV viral loads ranged from 709

to 9.44

copies/ml. None of the patients were on antiretroviral

therapy at the time of sampling. In the screening mode, the pos-

itive predictive value for the LAg, BED and Bio-Rad assays were

98%, 90% and 93.3% respectively, sensitivity was 100% for all assays

however, specificity was 100%, 100% and 97.2% respectively. The

modified Bio-Rad assay incorrectly identified 1 sample as recent

infection. HIV viral loads were significantly higher in recent infec-

tions (

< 0.001) and seroconversion intervals were longer in long

term infected individuals (

< 0.001). Chi-square analysis revealed

that more long term infected patients had Fiebig stage of V+

compared to recently infected individuals (

< 0.001). Significant

correlationwas observedwhen 24 samples from Irish patientswere

tested at the NVRL and Public Health England laboratory,

= 0.96,

p < 0.001.

Conclusions:

The method identified for use in the NVRL is the

Sedia HIV-1 Limiting Antigen Avidity enzyme immunoassay. This

method is recommended by the WHO and is used by Public Health

England to distinguish recent HIV infection from long-term infec-

tion. This assay will be used in the NVRL to test all new HIV

diagnoses from January 2016 onwards.

http://dx.doi.org/10.1016/j.jcv.2016.08.094

Abstract no: 57

Presentation at ESCV 2016: Poster 55

Evaluation of the H-DiaCMVQ kit

for detecting

and quantifying CMV-DNA in plasma and in

whole blood samples

Catherine Mengelle

1 ,

∗

, Jean-Michel Mansuy

Laeitita Houles

, Karine Sandres-Saune

Jacques Izopet

Department of Virology, Toulouse University

Hospital, Toulouse, France

Department of Virology, Toulouse University

Hospital, Department of Physiopathology, Unité

Inserm U563, Toulouse, France

Aims:

To validate the analytical characteristics of the H-

DiaCMVQ kit

for the detection and the quantification of CMV-DNA

in plasma and in whole blood. To assess the clinical performances

and to analyze the influence of blood compartment by comparing

the H-DiaCMVQ kit

and the in-house method on 150 samples.

Material and methods:

Whole blood samples were tested with

the in-housemethod. 50 positive (mean viral load: 2.79 log

IU/ml,

range: 6–1.85 log

IU/ml) and 100 negative samples were selected

for analysis and plasma collection.

Nucleic acids were extracted with the MagNA Pure 96 DNA and

Viral NA Small Volume kit

on the MagNA Pure 96

instrument

according to the Pathogen Universal 200 protocol (plasma) and

DNA Blood SV protocol (whole blood) (Roche Molecular Diagnos-

tics, Meylan, France).

For comparison, sampleswere testedwith the referencemethod

targeting the UL83 (limit of detection 74 IU/ml (1.87 log

IU/ml)

and with the H-DiaCMVQ kit

(Diagenode, Seraing, Belgium), both

on the Light Cycler 480

. Results were calculated in log

IU/ml.

Results:

The analytical performances of the H-DiaCMVQ kit

were very satisfactory on either plasma or whole blood: specificity

was 100% and a very high range of linearity was obtained.

Intra-assay reproducibility was 0.20 and 0.03 in two plasma

samples (3.04 and 6.10 log

IU/ml) it was 0.33 and 0.07 in two

whole blood samples (3.17 and 6.87 log

IU/ml). Inter-assay repro-

ducibility was 0.30 and 0.13 in two plasma samples (3.58 and

5.38 log

IU/ml) it was 0.47, 0.19 in two whole blood samples (3.62

and 5.51 log

IU/ml).

CMV-DNA correlated well in plasma between both methods

(

= 24):

= 0.82,

< 0.0001, (slope of Deming regression 0.912 [CI

95%: 0.63–1.19] and y-intercept 1.12 [CI 95%: 0.30–1.94]. Simi-

larly CMV-DNA correlated well in whole blood (

= 36):

= 0.86,

< 0.0001, (slope of Deming regression 1.071 [CI 95%: 0.86–1.28]

and y-intercept 0.84 [CI 95%: 0.18–1.5].

We analyzed the influence of the blood compartment by com-

paring the results obtained on whole blood with the reference

method with those obtained with the H-DiaCMVQ kit

on corre-

sponding plasma samples (

= 150). 142 samples gave concordant

results (36 positive and 106 negative). Eight samples gave dis-

crepant results: 2 samples were positive in whole blood and