Detection of Aspergillus galactomannan in BAL in the

1
Detection of Aspergillus galactomannan in BAL in
the Platelia Aspergillus EIA, as performed at
MiraVista Diagnostics LJ Wheat1, AM LeMonte1, MM
Durkin1, SL Swartzentruber1, KK Knox2, CA Hage2,
C Bentsen3, S Husain5, N Singh5, CJ Clancy4, MH
Nguyen4, HL Leather4, JR Wingard4 1MiraVista
Diagnostics, Indianapolis, IN,2Indiana University
School of Medicine, Indianapolis, IN,3Bio-Rad
Laboratories, Redmond, WA, 4University of Florida
College of Medicine, Gainesville, FL.5University
of Pittsburgh, Pittsburgh, PA
MiraVista Diagnostics L. Joseph Wheat,
M.D. President and Director jwheat_at_miravistalabs.c
om Tele [phone] www.miravistalabs.com
Poster 67
Reproducibility of Positive Results in Cases and
Controls
Table 2. Comparison of precision in serum and BAL
METHODS
ABSTRACT



Serum is the only specimen type that is
FDA-cleared for testing in the Bio-Rad Platelia
Aspergillus EIA kit. The sensitivity for
diagnosis of invasive pulmonary aspergillosis
(IPA) was shown to be 2-fold higher in BAL than
serum in hematologic patients (MJ Becker Br J
Hematol 2003). We have validated BAL in the
Platelia Aspergillus EIA. The detection limit
was 0.5 ng/ml of galactomannan (GM) in BAL and
serum. Precision and reproducibility were similar
in BAL and serum. In non-immunosuppressed (NIS)
patients, lung transplant (LuTP), and solid organ
transplant (SOT) patients (n17) with IPA, and
controls (N239), sensitivity was 88 and
specificity 84 at a GMI cutoff of 0.5. At a 1.0
cutoff, sensitivity was 88 and specificity 91.
The false-positive results were attributed to
colonization or infection with Aspergillus spp.
or cross reactive molds, but in 4 LuTP patients
IPA was subsequently diagnosed. We have
compared results in patients for whom serum and
BAL were tested within one week of each other and
at least one was positive. BAL was positive in
all, and serum in 9/33 (27). The GMI in BAL was
gt1.0 in 29/33 (88). Our results indicate that
the Platelia Aspergillus EIA, validated using
the MVista Process, exhibits comparable
performance characteristics in BAL and serum, and
that testing BAL improves the sensitivity for
diagnosis of IPA over antigenemia, cytology or
culture in LuTP and all SOT patients, but not in
NIS patients. As with serum, results between 0.5
and 1.0 GMI are less specific, but should not be
regarded as negative. Additional studies are
needed to validate these findings in larger
numbers of patients with IPA.

• Clinical specimens. Cases with IPA and controls
  have been described (S Husain Transplantation
  2007, CJ Clancy J Clin Microbiol 2007, MH Nguyen
  J Clin Microbiol 2007). Specimens from 56
  healthy subjects with asymptomatic HIV infection
  but without pulmonary disease or suspected IPA
  also were tested. Specimens had been stored at
  -70 C for up to three years before use in this
  study.
• Platelia Aspergillus EIA. The test was
  performed on BAL according to the manufacturers
  recommendations for serum. BAL specimens were
  extracted with EDTA at 104 C for 4 minutes,
  followed by centrifugation. Precision,
  reproducibility, and calibrations in BAL and
  serum were determined in specimens to which the
  Aspergillus GM positive kit control had been
  added. Results are expressed as units, by
  dividing specimen OD by the mean cutoff OD, and
  results greater than 0.5 units are considered
  positive.

Specimens spiked with different concentrations of
GM were aliquoted and tested on the same day
(intra-assay comparison) or on separate days
(intra-assay comparison). The EDTA/104 C
extraction procedure was performed separately on
each aliquot. Data are expressed as the mean,
plus or minus one standard deviation (coefficient
of variation).
SUMMARY

Sensitivity and Specificity

• Analytical characteristics are similar in serum
  and BAL, except that results in serum are
  significantly lower than in BAL.
• Clinical sensitivity in IPA is higher in BAL
  (88) than serum (44).
• Specificity was 100 in asymptomatic HIV
  infected controls without pulmonary disease.
• Positive results were more common in specimens
  from patients deemed colonized with yeast or mold
  (18.2) than in those without colonization
  (3.9).
• Positive results were seen in BALs from 3
  symptomatic patients receiving piperacillin-tazoba
  ctam who did not have IPA, but 2 of whom were
  colonized with mold.

Specimens yielding positive results are retested
for confirmation before they are reported as
positive. The EDTA/104 C extraction procedure
was performed separately for the initial and
repeat test. Results of test 1 and test 2 agreed
closely by linear regression analysis.
Table 1. Positivity rate in IPA and controls

INTRODUCTION

• Detection of Aspergillus galactomannan in BAL
  can be a sensitive method for diagnosis of
  invasive pulmonary aspergillosis (IPA).
• The sensitivity for detection of GM in
  hematology patients was at least two-fold higher
  in BAL than in serum using the Platelia
  Aspergillus EIA in published studies (MJ Becker
  Br J Hematol 2003, W Meersseman Am J Respir Crit
  Care Med 2008).
• However, the Platelia Aspergillus EIA is not
  FDA cleared for evaluation of BAL specimens.
• The purpose of this study was to analyze the
  performance characteristics of the Platelia
  Aspergillus EIA on BAL for diagnosis of IPA.

CONCLUSIONS

• The Platelia Aspergillus EIA is an accurate
  method for detection of GM in BAL.
• BAL is more sensitive than serum for diagnosis
  of IPA.
• Causes for positive results without IPA require
  further investigation, and may include
  undiagnosed invasive mold infection or treatment
  with piperacillin-tazobactam .

Sensitivity was 88 in 17 patients with proven
IPA. Specificity was 100 in HIV-infected
controls without pulmonary disease and 97.3 in
lung transplant patients undergoing
post-transplant BAL for surveillance. Serum was
positive for GM in 4 of 9 (44) IPA cases in whom
it was tested.
1Two positive episodes in one patient