Style A 36 by 48 tall

1
Development of an oral fluid assay capable of
differentiating recent from established HIV
infection Niel Constantine 1, Ligia Peralta 1,
Anne Sill 2, Kristen Kreisel 1, Bethany Griffin
Deeds 1, and the ATN for HIV/AIDS Intervention. 1
University of Maryland, Baltimore, School of
Medicine 2University of Maryland, Baltimore,
Institute of Human Virology
Abstract
Methods

• Samples
• 15 participating Adolescent Trials Network
  (ATN) sites provided serum/oral fluid pairs
• Serum was collected by standard venipuncture,
  whereas the oral fluid was collected using the
  OraSure device that targets
• mucosal transudate containing IgG Ab
• Sample Size (selected on the obtainment of at
  least 65 samples from recently infected persons)
  
• Total sample size 246
• Phase I 40 (20 Recents and 20 Establsihed as
  classified by the serum DV)
• Phase II 206 (62 Recents and 144 Established as
  classified by the seurm DV)
• Phase I and II 246 (82 Recents and 164
  Established as classified by the seurm DV)
• Tests
• HIV status
• Serum EIA methods and Western blot as
  per each site
• Oral Fluid Oral Fluid Vironostika HIV-1
  Microelisa System (sensitive mode)
• Testing for Recent or Established Infection
  (S/LS)
• Serum Vironostika HIV-1 Microelisa
  System (less-sensitive mode)
• Oral Fluid Oral Fluid Vironostika HIV-1
  Microelisa System, modified to a less-sensitive
  mode
• Oral Fluid S/LS Calibration
• Oral fluid dilutions during Phase I
  calibration 1110, 125, 150, 160, 180,
  1100

Background/Objective Identifying persons who
have been recently infected with HIV is important
for epidemiologic investigations, including the
estimation of incidence and for the enrollment of
acutely infected individuals into appropriate
intervention programs. Currently, modified
serologic methods (sensitive/less-sensitive
assays, S/LS) are available to classify recently
infected persons (3-6 months), but they require
the collection of blood that often limits their
application, including among adolescents.
Simpler methods of specimen collection are needed
to increase compliance and offer an alternative
to invasive procedures that increase the risk of
occupational exposure to HIV. We sought to
develop, calibrate, and validate a S/LS assay
that uses easy-to-collect oral fluid. Methods
Serum/oral fluid pairs were collected from 246
HIV-infected adolescents newly enrolled in HIV
care programs of the Adolescent Trials Network
from 15 sites in the United States and Puerto
Rico. The serum was collected by venipuncture,
whereas the oral fluid was collected by the
OraSure device that targets mucosal transudate
containing IgG antibody. Reference testing was
performed on all serum specimens by the
Vironostika S/LS test using an optical density
(OD) cutoff of 1.0 to distinguish recent from
established infections. We chose the first 20
recent and 20 established samples, as classified
by the serum Vironostika S/LS, and tested the
complementary oral fluid specimens at various
dilutions to calibrate the modified Vironostika
Oral Fluid assay an optimal dilution of 150 and
an OD cutoff of 0.238 were selected to yield a
concordance of 100 for recent samples and 85
for established samples (OK). Using these
parameters, the concordance of the serum S/LS and
the oral fluid S/LS assay was then determined
using the remaining 206 serum/oral fluid pairs.
Of the 206 specimens, 61 were classified as being
from recently infected individuals and 145 from
those with established HIV infection according to
the serum S/LS assay. Results There was a
correlative trend of the OD readings of both
assays when testing the 206 serum/oral fluid
pairs (r.658, plt.001). The concordance rate
between the serum and oral fluid based S/LS
assays using the parameters selected during the
calibration phase was lower than anticipated (57
for recents and 92 for established). However,
when reassessing the results from the entire
specimen pool (n246 serum/oral fluid pairs), the
best agreement between the oral fluid and the
serum S/LS results remained at an oral fluid
dilution of 150, but with a new cutoff of 0.280.
This adjustment to the OD cutoff of the oral
fluid assay gave the highest kappa value
(?.703), and a concordance rate of 85 for the
recent samples and 87 for the established
samples as compared with the serum S/LS test.
Conclusions The HIV oral fluid S/LS test has
exhibited good concordance with the routinely
used serum S/LS test. This S/LS test strategy
using oral fluids offers a simple and safe
collection method that is suitable for use in a
variety of epidemiologic investigations.
Oral Mucosal Transudate from tooth-gum margin
Results

• The oral fluid (OF) dilution of 150 and OF S/LS
  OD cutoff of 0.280 gave the
• highest rate of agreement with the reference
  serum S/LS assay when testing
• all 246 samples
• The OD readings of the oral fluid S/LS assay and
  the serum S/LS assay when
• testing 246 oral fluid/serum pairs, showed a
  correlative trend (r.658, plt.001).
• The concordance rate of the OF S/LS assay and
  the serum S/LS assay was 85
• for samples from recently infected persons
  and 87 from those with
• established infection.

Serum S/LS Designations
Recent Established
Recent 70 (85.4) 22 (13.4)
Established 12 (14.6) 142 (86.6)
Background
Oral fluid S/LS Designations

• Oral fluid assays for HIV are widely available
  and have been validated as accurate.
• EIA, Western blot, and rapid tests for HIV using
  oral fluid for testing are licensed in the U.S.
• The test medium for oral fluid assays is oral
  mucosal transudate (rich in IgG).
• The applications for an oral fluid sensitive/less
  sensitive assay (OF S/LS) include
• Oral fluid collection is preferred by clients
  (Peralta, Constantine, et al. Arch Peds Adol
  Med, 2001, 155 838-843)
• Collection of samples is easier in populations
  from which blood is difficult to obtain (IVDU,
  Obese clients, children)
• gt Oral fluid collection is safer than blood
  collection.

Iterative concordance rates of the oral fluid
S/LS and serum S/LS at varying oral fluid cutoffs
Considerations

• It should be noted that the Vironostika serum
  S/LS assay itself is believed to be only about
  85 accurate when tested on
• seroconversion panel specimens. (Constantine
  et al. 2003. JAIDS 3294-103)
• It is possible that the true sensitivity for
  correctly identifying those with recent and
  established HIV infection using the oral
• fluid S/LS method could be better, equivalent,
  or worse than the serum-based S/LS assay.
• We are in the process of obtaining serum and
  oral fluid pairs from seroconverting individuals
  such samples will yield
• definitive information on the utility of both
  tests.
• The HIV oral fluid S/LS test has exhibited good
  concordance with the routinely used serum S/LS
  test.
• This S/LS test strategy using oral fluids offers
  a less invasive and safe collection method that
  is suitable for use in a variety of
• epidemiologic investigations.
• The collection of oral fluid will likely
  increase compliance levels among those being
  tested.

Objectives

• To develop, calibrate, and validate an oral fluid
  S/LS EIA method to perform equivalently to the
  reference S/LS test.
• Phase I
• Calibration of the oral fluid S/LS assay with
  the Vironostika serum S/LS assay
• Proof of Principle that the oral fluid S/LS
  acts equivalently to the serum based S/LS assay
• Phase II
• Validation of findings of the Phase I
  calibration
• Modifications of the oral fluid S/LS assay
  parameters to enhance concordance
• Define parameters that yield optimal
  concordance between the OF S/LS and the serum
  S/LS tests.

Conclusions