Reply to Baughman et al
Joel I. Ward
0
1
2
3
James D. Cherry
0
1
2
3
Martin Lee
0
1
2
3
Swei-Ju Chang
0
1
2
3
0
University of California at Los Angeles (UCLA) Center for Vaccine Research, Los Angeles Biomedical Research Institute, Harbor-UCLA Medical Center, David Geffen School of Medicine UCLA
, Torrance,
California
1
Henry F. Chambers Department of Medicine, University of California-San Francisco
2
Peter Horby Oxford University Clinical Research Unit, National Institute of Infectious and Tropical Diseases, Bach Mai Hospital
, Hanoi,
Vietnam
3
for Vaccine Research, Los Angeles Biomedical Research In- stitute, Harbor-UCLA Medical Center, David Geffen School of Medicine UCLA
, 1124 W. Carson St., Torrance,
CA
(joelward @ucla.edu). Clinical Infectious Diseases 2007
;
44:150-1 2006 by the Infectious Diseases Society of America. All rights reserved. 1058-4838/2007/4401-0030$15.00
D o w n l o a d e d f r o m h t t p : / / c .i d o x f o r d j o u r n a .l s o r g / b y g u e s t o n N o v e m b e r 3 , 2 0 1 4
-
To the EditorAs reviewed in our
articles reporting on our acellular pertussis
vaccine trial [15], it is difficult to assess
the incidence of pertussis in older persons.
Beyond variations in incidence
attributable to outbreaks, age, and regional and
seasonal factors, there is difficulty in
recognizing or diagnosing pertussis. Cultures
and PCR are rarely performed early in the
course of illness, when they are most likely
to have positive results, and serological
tests for pertussis have limitations.
Consequently, one can easily underestimate or
overestimate the infection burden.
Approximately one-half of the pertussis cases
in our trial were diagnosed by culture and
PCR. Additional cases and asymptomatic
infections were identified by systematic
prospective serologic evaluations. To
maximize sensitivity (i.e., to detect true cases)
and specificity (i.e., avoid false-positive
results), we employed exacting serologic
methods. First, we prospectively collected
serial blood specimens from each subject,
and we assayed 9 pertussis antibodies as
titer changes in sequential serum samples
from the same individual (i.e., the assays
were self-controlled).
A multitude of assay variables exist that
result in chance inaccurate quantitations.
Our assays were developed to exacting US
Food and Drug Administration
specifications and were based on standards they
provided. The validation of the assays
included standardization of reagents, plates,
dilution accuracy, and use of quantitative
positive controls. All titer increases were
determined with paired serum samples
from the same subjects (i.e., they were
selfcontrolled). All paired serum samples
from the same individual were run on the
same day and on the same microtiter
plates to minimize assay variation.
In an efficacy trial, one wants to
maximize case detection but also avoid
falsepositive results. To accomplish this,
besides using precise assays, we evaluated
remaining assay variability to determine
the cut-off points for positivity (i.e., 2- or
4-fold increase). With same-plate
evaluations of each set of paired serum samples,
the coefficient of variation for the various
pertussis antibody assays (pertussis toxin,
filamentous hemagglutinin, pertactin, and
fimbriae) ranged from only 0% to 18%
for IgG assays and from 0% to 22% for
IgA assays. Therefore, the 2-fold and
4fold cut-off points selected far exceeded
the variation inherent in these assays. With
these low coefficients of variation and our
serologic criteria, one would expect !1
false-positive result per 10,000 assays.
Our estimate of a 1.1%2.7% infection
rate per year derives from examining the
36 permutations of results shown in table
1 of our article [1]. Independent of the
criteria employed, we found the observed
incidence values to be remarkably
consistent. Undoubtedly, there are variations in
different populations at different times
and over longer intervals, but we believe
few studies have been as rigorous as ours.
Lastly, among control subjects, there
were no titer differences between the
samples obtained before and at 1 month after
hepatitis A immunization. Therefore, our
incidence estimate does not differ whether
we use the preimmunization (12-month
interval) or the 1-month
postimmunization sample (11-month interval). We used
the 11-month interval so that our
comparisons between samples from subjects
who had received acellular pertussis
vaccine and control samples from
unvaccinated subjects were identical.
In conclusion, we find little meaningful
difference between an estimate of 1.1%
2.7% and an estimate of 2.9%. An
advantage of our study is that we could compare
the infection rate to the active,
prospectively ascertained disease incidence. We
found a 5:1 ratio of asymptomatic to
symptomatic cases, and this is, to our
knowledge, the first estimate of this type
to be assessed prospectively in a national
cohort.
Acknowledgments
Potential conflicts of interest. All authors: no
conflicts.
1. Ward JI, Cherry JD, Chang S- (...truncated)