The Epidemiology of Infection in Trachoma - Investigative ...

Investigative Ophthalmology & Visual Science, Vol. 30, No. 8, August 1989
Copyright © Association for Research in Vision and Ophthalmology
The Epidemiology of Infection in Trachoma
HughPV Taylor,* Peter A. Rapozaf Sheila West,* Shirley Johnson,* Bearriz Munoz,*
Sidney Katala4 and B. B. O. Mmbaga^:
Specimens for chlamydial isolation culture and direct fluorescent antibody cytology (DFA) were
collected from 1671 women and children from a trachoma-endemic area in Central Tanzania. Trachoma
was graded using the new World Health Organization grading scheme, and 54% of the children
and 9% of the women had inflammatory trachoma (TF or TI). DFA, using the presence of five
elementary bodies as the criterion for a positive test, had a sensitivity of 88.0% and a specificity of
87.5% compared to culture and a sensitivity of 54.7% and specificity of 92.8% compared to clinical
diagnosis. Altogether, 52.9% of those with trachoma grade TF were positive on either or both culture
and DFA versus 77.0% of those with TI. Twenty-nine isolates were serotyped; 18 were serovar A, ten
were serovar B, and one was serovar Ba. Positive cultures or DFA were obtained in 6.9% of those
graded clinically as not having TF or TI and in a smaller number of those without any perceptible
evidence of disease. Conversely, organisms could not be demonstrated in a number of people with
severe inflammation (TI) even though some became positive after multiple repeated culture. These two
findings of infection without disease and disease without evidence of infection suggest the importance
of the immunologic response to infection in determining the clinical status. DFA was found to be an
appropriate test for future field studies of trachoma. Further studies of those with disease but without
agent and of those with agent but without disease will help understand the dynamics of infection and
transmission and the role of the immune response in this important blinding disease. Invest Ophthalmol
Vis Sci 30:1823-1833,1989
Repeated episodes of infection with Chlamydia
trachomatis are believed to be important in maintaining
the sustained inflammatory disease recognized
clinically as active or inflammatory trachoma.
1 ' 2 Much of this reinfection probably occurs
within the family setting, especially between children
and women involved in child care. 3 The clinical disease
in chlamydial eye infection has been characterized
as a delayed-type hypersensitivity reaction and
can be induced by the topical application of a purified
chlamydial antigen. 4 - 5 This immunologic response is
seen to lead to conjunctival fibrosis and ultimately to
trichiasis and blindness. However, several observations
concerning the biology of infection require further
elucidation. For example, even in those with the
From *the International Center for Epidemiologic and Preventive
Ophthalmology, The Dana Center of The Wilmer Institute and
The School of Public Health The Johns Hopkins University, Baltimore,
Maryland, fHelen Keller International, Inc., New York,
New York, and the ^Kongwa Primary Eye Health Care Project,
Kongwa, Tanzania.
Supported by grants from The Edna McConnell Clark Foundation,
the National Eye Institute (EY-03324), and The Syva Company.
Submitted for publication: November 15, 1988; accepted February
28, 1989.
Reprint requests: Dr. Hugh R. Taylor, The Johns Hopkins Hospital,
600 North Wolfe Street, Baltimore, MD 21205.
most severe clinical disease, chlamydia frequently
cannot be identified in laboratory tests. Only 60% to
70% of those with severe disease and 10% to 20% of
those with mild inflammatory disease will be positive
on laboratory testing. 6 " 13 Similar observations have
been made in a monkey model where, despite the
requirement for weekly inoculation with viable organisms
to sustain continuing disease, chlamydial
cultures are persistently negative after the first few
months. 2 Even though direct fluorescent antibody cytology
(DFA) using monoclonal antibody continued
to detect the presence of organisms in repeatedly inoculated
monkeys for four weeks longer than culture,
it too eventually became negative. 14 Similarly, organisms
have occasionally been isolated from people
who do not have obvious clinical disease. 9 " 15
The following study was undertaken to assess the
distribution of chlamydial culture and DFA results in
families in an area of hyperendemic trachoma to further
clarify the association between the presence of
identifiable chlamydia and clinical disease. In addition,
a newer test for detecting chlamydia, DFA, was
compared to chlamydial culture. DFA offers a number
of logistic advantages over chlamydial culture in
that it does not require the maintenance of a cold
chain between collection and testing and is a cheaper,
more rapid and less complex test. Lastly, a comparison
was made of the frequency with which chlamydia
1823

1824 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / August 1989 Vol. 30
could be identified within the different grades of trachoma
defined by the new simplified grading scheme
developed by the World Health Organization
(WHO). 16
Study Population
Materials and Methods
Specimens were obtained from 1671 women and
children who were examined as part of an epidemiologic
survey of risk factors for trachoma in Central
Tanzania. 17 Briefly, a stratified random sample of 20
villages was drawn. Within each village, a cluster
sample of preschool children (aged 1 to 7 years inclusive)
and their mothers (or female caretakers) were
examined. In the first nine villages, members of every
second household had specimens collected from their
right eye and two photographs taken of the everted
upper lid.
Verbal informed consent was obtained from each
person or their guardian before entry into this study.
The method of obtaining consent and the study procedures
had been reviewed and approved by the Joint
Committee on Clinical Investigation of the Johns
Hopkins University School of Medicine.
Trachoma Grading
One trained examiner used the new simplified
WHO trachoma grading scheme throughout the survey.
1618 Each person was examined with a X2.5
loupe. The presence of fivesigns were graded for each
eye: TF—trachomatous inflammation-follicular; TI
—trachomatous inflammation-intense; TS—trachomatous
scarring; TT—trachomatous trichiasis; and
CO—corneal opacity. Photographs of the tarsal conjunctiva
were graded independently in a masked fashion
with a X4 loupe using the same grading
scheme.
A sample of 135 photographs was reexamined
using a finer grading scheme for the signs TF and TI.
The sample included 47 subjects with positive laboratory
tests but without the clinical signs of TF or TI
and a random sample of 88 other subjects who had
either positive laboratory tests or TF or TI. The regarding
was done without knowledge of either the
laboratory results or the previous clinical grade in
order to reduce grader bias. For this regrading, each
sign was graded as: 0—definitely absent; 1—equivocally
present; 2—mild disease definitely present but
less than defined by the WHO grade; and 3—equivalent
to the WHO grade. This gave a finegrading of TF
(FTF) and a fine grading of TI (FTI).
Specimen Collection
Conjunctival scrapings were obtained from the
right superior palpebral conjunctiva of each subject
using sterile, dry dacron swabs on plastic shafts. Topical
anesthetic was not used. An initial swab on a
metal shaft was gently rubbed across the conjunctiva
to remove mucus and debris from the tissue surface.
A second dacron swab was vigorously stroked across
the conjunctiva five times. It was promptly rolled
across each half of the central 8 mm well of a fluorescence
microscopy slide (MicroTrak Collection Kit,
Syva Co., Palo Alto, CA) and then placed into a plastic
cryogenic vial containing 1 ml chlamydial transport
media composed of 25% fetal calf serum, 10%
Eagle's MEM, and 5% DMSO with vancomycin 10
Mg/ml, gentamicin 10 ng/m\, and mycostatin 10
buffered to pH 7.2.
Chlamydial Culture
Cryogenic vials containing inoculated transport
medium were kept at ambient temperature for less
than one hour prior to refrigeration at 4°C. One to 8
hr later, vials were snap frozen to -196°C in liquid
nitrogen refrigerators. Refrigerators were maintained
at this temperature and shipped to Baltimore. Two
weeks to 8 months later, specimens were thawed and
100 Ail inoculated onto cycloheximide-treated McCoy
cell monolayers in each of four wells of a microtiter
plate. 14 One set of duplicate wells was stained at 2
days (first passage) and another set passed at 2 days
and stained at 4 days (second passage) with fluorescein-conjugated
monoclonal antibodies to C. trachomatis
(MicroTrak Tissue Confirmation Reagents,
Syva Co.). Specimens were examined at X500 magnification
and scored positive if one or more typical
inclusion bodies were present. The degree of positivity
was graded as follows: 1 +—one to nine inclusions
in the well; 2+—10 to 20 inclusions in the well;
3+—1 to 10 inclusions per X500 magnification field;
4+—greater than 10 inclusions per X500 magnification
field. The examiner was masked from knowledge
of the clinical grading and DFA results. Chlamydial
cultures were judged inadequate if the cell monolayers
from both the first and second passages were
lysed.
There were 69 people who had TI but who were
negative on routine tissue culture. Residual collection
material was available in 40 of these, and in these
instances, the remaining half of the specimen (400 n\)
was diluted in culture medium and cultured in three
1-dram vials containing DEAE-Dextran-treated
McCoy cells grown on coverslips that were incubated
at 37°C in 5% CO 2 . A one-on-one second passage was

No. 8 EPIDEMIOLOGY OF TRACHOMA / Toylor er a I 1825
made after 72 hr, and one coverslip was stained with
fluorescein-conjugated monoclonal antibody. If it
was negative, the remaining two vials were combined
and split to inoculate three fresh vials of cells. A total
of six serial passages was finallymade before the specimen
was considered negative, at which time all three
coverslips were examined.
Previously, we had established the comparability of
the two culture methods in our laboratory. In an unpublished
substudy, replicate swabs were collected
from seven monkeys at days 0, 7, 14, 21, 28, 42 and
56 after ocular inoculation with 5000 inclusion
forming units (IFUs) of a serovar B organism (TW-5).
Microtiter plate and dram vial cultures were performed
as described above, except only two passages
were used in each culture. The two culture methods
gave 86% agreement (42 out of 49 paired cultures). In
one case, the vial culture became positive on the second
passage while the plate culture remained negative.
In six cases, the vial cultures were negative while
the plate cultures were positive (once in the first passage
and five times in the second passage). In only
one of these seven discrepancies did DFA show ten or
more elementary bodies (EB). These data suggest
that, in our hands, chlamydial culture in a microtiter
plate is at least as sensitive if not more sensitive than
culture in dram vials. The few discrepancies that occurred
did so at lower levels of infection.
DFA Cytology
After collection, slides were air-dried and then
flooded repeatedly with acetone for 5 min. Fixed
slides were kept at ambient temperature for 1 to 8 hr
while collections proceeded. Slides were stored at 4°C
for up to 6 weeks prior to shipment to Baltimore.
Slides were then stored at -20°C until processmg 2
weeks to 8 months later. Prior to staining, slides were
allowed to thaw for 30 min, immersed in acetone for
5 min, and air-dried. Thirty microliters of fluorescein-conjugated
monoclonal antibody to the major
outer membrane protein of C. trachomatis (Micro-
Trak Direct Reagent, Syva Co.) were placed over the
specimen which was then incubated in a moist
chamber for 30 min at room temperature, rinsed
twice with distilled water, and airdried. Slides were
read at X500 magnification under oil immersion by
fluorescence microscopy. Each field of each slide was
examined. Fluorescent particles were examined at
X500 and confirmed at XI250. Typical apple-green
fluorescing EB were enumerated. When less than ten
EB were seen in the entire slide, the number of EB
was recorded. If ten or more EB were present, the
result was graded as follows: 1+—10 to 49 EB per
well; 2+—50 to 100 EB per well; 3+—1 to 10 EB per
X500 magnification; or 4+—greater than 10 EB per
X500 magnification. 19 The examiner was masked
from knowledge of the clinical grading and culture
results. Greater than 200 epithelial cells per slide were
required as a criterion of an adequate DFA specimen.
Serotyping
Isolates for serotyping were cultured in 1-dram
vials. When 80% infective titers were reached, EB
were harvested, washed and resuspended in 0.2 ml of
formalin-TWEEN-80-PBS solution (0.02 formaldehyde,
0.01% TWEEN-80, [Sigma Chemical, St.
Louis, MO] in sterile PBS pH 7.0). Serotyping was
performed using the MicroIF system of Wang and
coworkers. 20 ' 21 Equal parts (0.1 ml) of the EB suspension
and a 3% formalinized yolk sac were combined
and pinpoint-sized dots placed on a clean glass slide.
These were overlayed with monoclonal antibodies
specific for C. trachomatis serovars (Monoclonal Antibody
Typing Kit, Washington Research Foundation,
Seattle, WA). 21
Statistics
Chi-square and Fisher's exact tests were used to
assess differences in proportions. Comparison of
DFA and culture tests were done using standard
measures of sensitivity, specificity, and predictive
values. Unless explicitly stated otherwise, all comparisons
and analyses use the routine microtiter plate
culture results. Spearman's Rank correlation coefficient
was used to assess the correlation between the
laboratory tests and between the laboratory results
and the clinical grading. For this analysis, specimens
positive on first passage were scored according to
their inclusion grading. Specimens that were only
positive on second passage were arbitrarily scored as
0.5 as they were obviously positive but presumably of
an infectious titer insufficient to be positive on first
passage. For DFA results, specimens with 1 to 4 EB
were scored as 0.25, 5 to 9 EB as 0.75, and then
according to their EB grading.
Results
In all, specimens from 1671 subjects were available
for laboratory evaluation for the presence of C. trachomatis
(Table 1). Of these specimens, 1090 came
from children aged 1 to 7 years (51% were girls; and
54% of all children had inflammatory trachoma), and
581 of the specimens came from adult women who
were the children's mothers or caretakers (9% of
women had inflammatory trachoma). Altogether,

1826 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / Augusr 1989 Vol. 30
Table 1. Age and sex distribution and prevalence
of inflammatory trachoma in the
1671 study subjects
Age
1
2
3
4
5
6
7
Total
8-24
25-34
35-44
45+
Total
Number
72
80
79
94
86
103
46
560
133
256
142
50
581
Female
Percent
with TF
and/or TI
47%
61%
62%
56%
62%
58%
58%
57%
9%
10%
6%
6%
9%
Number
73
73
84
94
77
83
46
530
—
Male
Percent
with TF
and/or TI
42%
56%
57%
54%
51%
44%
34%
50%
265 specimens were positive on routine chlamydial
culture, 237 (89.4%) on the first passage and 28
(10.7%) on the second passage. There were 386 positive
specimens on DFA testing, 32 (8.3%) with five to
nine EB, and 354 (91.7%) with ten EB or more. In 51
DFA smears, intracytoplasmic inclusions were identified.
The number of inclusions ranged from one
to 25.
Adequacy of Specimens
Cultures were considered to be inadequate if the
cell monolayer had been completely destroyed in
both the firstand second passage (23 cases) or if it had
been completely destroyed in one and had been partially
destroyed in the other (12 cases). In most cases,
there was frank evidence for bacterial infection.
Overall, 35 (2.1%) cultures were inadequate. Inadequate
cultures occurred in all age groups without a
clear age-specific or sex trend. They were equally
common in eyes with TF or TI but were more common
in eyes with either TF or TI (combined rates
Table 2. Performance of DFA cytology with different
criteria for a positive test result (threshold number
of elementary bodies)—tissue culture is
used as the reference standard
Criterion
(number ofEB) Sensitivity
10
85.3%
5
88.0%
3 88.0%
1 88.8%
Specificity
89.6%
87.5%
84.9%
82.9%
—
—
Predictive value
Positive
64.0%
60.4%
55.6%
52.8%
Negative
96.6%
97.1%
97.0%
97.2%
3.0%) than in eyes without inflammation (1.5%) (X 2
= 4.45, P = 0.04). Inadequate cultures were equally
common in DFA-positive and DFA-negative (2.3%)
cases but less common in DFA-inadequate specimens
(0.5%) (Fisher's exact test, P = 0.017).
DFA specimens were considered inadequate if less
than 200 cells were seen in the smear. In all, 188
smears (11.3%) were inadequate. Two of these smears
were still positive, having more than ten EB, and
these have been included in the subsequent analyses.
Inadequate smears were more common from the
youngest children; 20% of specimens were inadequate
in 1- to 2-year-olds compared to 13% in 3- to
7-year-olds and 4% in those over age 7 years (X 2
= 52.3, P < 0.001). Inadequate smears were less
common in eyes with either TF or TI (combined rates
8.9%) than those without any inflammation (12.7%)
(X 2 = 5.94, P = 0.02). Five times as many inadequate
smears corresponded to a negative culture (13.1%) as
to a positive culture (2.6%) (X 2 = 24.0, P < 0.001).
This suggests that an inadequate specimen may have
been collected and the culture was artifactitiously
negative. However, seven of the 188 inadequate
smears (3.8%) were coupled with a positive culture,
showing that this did not occur invariably. Only one
specimen was judged as being inadequate for both
culture and DFA.
Comparison of DFA and Culture
Altogether, both culture and cytology specimens
were adequate for 1451 people. The performance of
DFA was highly comparable to that of culture (Table
2). As would be expected, the sensitivity, specificity
and predictive values changed with the criterion for
DFA positivity. The criterion for DFA positivity of
five or more EB seemed to give the optimal performance
as it maximized the balance between sensitivity
and specificity. Thus, this criterion has been used
in subsequent analyses. There was a strong correlation
between the number of inclusions seen on culture
and the number of EB seen on DFA (Spearman's
Rank correlation coefficient 0.665, P < 0.0001)
(Table 3).
Correlation of Clinical Disease with DFA and Culture
The frequency of a positive culture or DFA in
those with inflammatory trachoma showed no difference
by age or sex, although there was a marked increase
in the number of positive laboratory results
with increasing severity of inflammatory trachoma
(Fig. 1). Overall, DFA was positive more frequently
than culture (Table 4). If the results of both laboratory
tests were combined, 52.9% of those with TF and
77.0% of those with TI were positive, whereas only

No. 8 EPIDEMIOLOGY OF TRACHOMA / Toylor er ol 1827
Table 3. Correlation between degree of positivity of culture and DFA for 1451 paired specimens
First passage
DFA grading
Negative
Negative*
(2nd passage positive)
7+
2+
3+
4+
Total
Negative
1-4 EB
5-9 EB
1 +
2+
3+
4+
988
56
25
52
19
44
9
1193
6
0
1
2
3
10
6
28
19
2
4
28
12
45
14
124
1
0
0
4
2
19
5
31
3
0
2
2
6
30
21
64
0
0
0
0
0
3
8
11
1017
58
32
88
42
151
63
1451
* Spearman's Rank correlation coefficient 0.665, P < 0.0001
23.0% of those with TS and 22.2% of those with TT in
the absence of inflammatory trachoma were positive.
The performance of the two laboratory tests were
compared with the clinical diagnosis of inflammatory
trachoma (either TF or TI) (Table 5). DFA, using five
or more EB as the criterion for a positive result, was
more sensitive and only slightly less specific than culture.
If the criterion for a positive DFA was changed
to ten or more EB, the sensitivity declined to 52.5%
while the specificity increased to 94.9%. For both
culture and DFA, there was a strong correlation between
infectious load (as measured by inclusion
count and the number of EB) and increasing severity
of disease (ranked normal, TF, and TI) (Spearman's
Rank correlation coefficient 0.421, P < 0.001 for
culture; 0.531, P < 0.0001 for DFA).
Examination of Those not Graded as TF or TI but
Laboratory Test-Positive
The photographs of 47 subjects who were graded
clinically as not having inflammatory trachoma but
who were positive by either chlamydial culture or
DFA were regarded using the fine grading scheme as
described in Methods. Twenty-seven people were
positive by culture; 17 of these were also DFA-positive.
The remaining 20 people were positive on DFA
alone. Nineteen of the 27 positive cultures (70%) were
positive in the first passage. The DFA results were not
predictive as to whether the culture would be positive
on the first or second passage, although those who
were DFA-negative had lower infectious titers on
culture. The 17 people who were positive on both
culture and DFA had more EB seen on DFA than the
20 people who were positive on DFA but negative on
culture. Thus, in each case, the discrepancies between
culture and DFA tended to occur with lower levels of
organism. There was no overall difference in the age
or sex distribution of people who were positive in
culture or in DFA nor was there a difference in the
occurrence or severity of inflammatory trachoma in
other household members. Therefore, for the next
analysis, these subgroups were combined and placed
in a matrix on the basis of their fine grading score for
FTF and FTI (Table 6). Altogether, six people had
completely normal appearing conjunctiva but were
positive on either or both culture or DFA cytology
(Tables 6, 7). A total of 15 had an equivocal status
(one or both fine signs graded as 1), and 26 had definite
signs of disease that were less than the criteria for
being classified as TF or TI (one or both fine signs
graded as 2).
These 47 people with positive laboratory tests represent
4.6% of the 1029 people who had neither TF
nor TI. Nearly two-thirds of these people were
women (Table 7). The rate of positive laboratory tests
for those without TF or TI was lowest for girls (2.1 %),
was intermediate for boys (4.5%), and was highest for
women (5.7%) (odds ratio for boys 2.1 [0.77-6.4] and
for women 2.7 [1.02-7.0] compared to girls). Those
Table 4. Frequency of positive chlamydial culture or DFA cytology by clinical trachoma status
None
Inflammatory
TF only
trachoma
77
Cicatricial trachoma
TS only
TT*
Number
positive
Number
total
Culture
DFA
Culture or
DFA
34/927 (3.7%)
48/813(5:9%)
55/800 (6.9%)
150/481 (31.2%)
221/447(49.4%)
229/433 (52.9%)
73/142(51.4%)
100/140(71.4%)
104/135(77.0%)
7/77(9.1%)
16/76(21.1%)
17/74(23.0%)
1/9(11.1%)
1/9(11.1%)
2/9 (22.2%)
265
386
407
1636
1485
1451
• Without TF or TI.

1828 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / August 1989 Vol. 30
Table 5. Performance
[TF and/or TI] is the
of chlamydial culture and DFA cytology (clinical grading of inflammatory trachoma
reference standard)
Predictive Value
Sensitivity
Specificity
Positive
Negative
(Number)
Culture
DFA
Culture or DFA
35.8
54.7
58.6
95.9
92.8
91.6
84.2
83.2
81.2
70.8
75.8
77.5
(1636)
(1485)
(1451)
with either no inflammation or with equivocal disease
had less scarring (TS) than the general population,
but the proportion coming from families in
which at least one other person had TF was not significantly
different from the group as a whole.
Examination of Those with Severe Clinical Disease
but Laboratory Test-Negative
There were 69 people who were clinically graded as
having TI with or without TF for whom routine cultures
were negative. In some wells of negative cultures,
debris with an appearance similar to free EB
was seen, suggesting that chlamydia may have been
present but had not been detected by culture. To see
if further blind passage of material might lead to the
establishment of inclusions and a positive culture, the
remaining collection material was serially cultured in
dram vials for the 40 people with TI and negative
culture for whom residual material was available. In
all, 16 became culture-positive; of 17 DFA positive
specimens, seven became culture-positive (41%) and
of 23 DFA negative specimens, nine became culturepositive
(39%). Although 13 of the 16 became positive
by the third passage, three were only positive on
the final sixth passage. There was no clear indication
of titration of infectivity as nine had six or more inclusions
per high-power field on their first positive
passage. For children, there was no difference in the
age or sex distribution of those who became positive
Table 6. Matrix showing distribution of fine grading
status (FTF and FTI) of 47 people who were
positive on laboratory testing but who
did not have TF or TI
FTF
0
12
Total
0
6
91
16
FTI
1
3
3
5
11
2
5
5
10
20
Total
14
17
16
47
compared to those who remained negative. Four
adult women were retested; two became positive
(aged 25 and 40 years) and two did not (aged 22 and
70 years). If those who became positive after reculturing
in vials were added to those who were positive
on routine culture, the overall rate of positive culture
in people with TI would become 89 out of 142 (63%).
Similarly, the number of positives on first passage
would become 84.3%, with 10.0% being detected on
second passage and 5.7% on third to sixth passage.
Distribution of Serotypes
Specimens from 69 patients with TF that had been
positive on tissue culture were selected for serotyping.
They formed two blocks of consecutive specimens
collected at the start and the end of the specimen
collection. It was possible to raise sufficient material
to serotype 29 of these; 18 were serovar A, ten were
serovar B, and one was serovar Ba. There was no
tendency for the serovar to be grouped by sex or by
disease severity. Serovar A tended to be more common
in younger children; 15 of the 18 isolates from
the 1- to 5-year-olds were serovar A, whereas only
three of the 11 isolates from those 6 years or older
were serovar A (Fisher's exact test, P = 0.003). The
serovar Ba came from a 1-year-old girl. Three isolates
came from the same family: a 26-year-old woman
and her two sons aged 7 and 4 years. Each was serovar
B.
Discussion
This study investigated the biology of infection
with C. trachomatis in an area of hyperendemic trachoma.
It studied a population-based sample of children
and their mothers and compared the clinical
status with chlamydial culture and DFA. Because it
was population-based, the findings relating to sensitivity
and predictive value have relevance for other
populations in which trachoma is endemic.
Overall, DFA was found to be a very satisfactory
test for use in field studies of trachoma. The results
obtained with DFA showed a good correlation with

No. 8 EPIDEMIOLOGY OF TRACHOMA / Taylor er al 1829
Table 7. Summary of characteristics of 47 people who were positive on laboratory testing
but who did not have TF or TI
Status Number Boys Girls Mothers TS
Cultured,
DFA+
Culture*,
DFA-
Culture-,
DFA+
Family
members
with TF (%)
Normal
(both 0)
Equivocal
(at least one
fine grade 1)
Mild disease
(at least one
fine grade 2)
15
26
11
14
0
1 (7%)
6 (27%)*
2
3
12
2
4
4
2
8
10
67%
80%
81%
* One woman aged 28 had both TS and TT.
both culture results and with clinical status. DFA offered
one advantage over culture as it was obvious
when an inadequate DFA specimen had been collected.
However, inadequate DFA specimens were
recognized in 11% of cases. Inadequate DFA specimens
were more common in young children and in
those without ocular inflammation. DFA specimens
can be masked by mucus; therefore it is important to
clear any gross discharge from the eye before obtaining
the specimen. Inadequate DFA specimens contained
insufficient cells, and this occurs either when
the specimen is not collected with enough vigor or
when it is not transferred firmly enough from the
swab to the slide. That the latter occurs is demonstrated
by the finding of more than five EB or the
occurrence of positive culture tests in specimens
where less than 200 cells were seen in the DFA smear.
Careful attention to specimen collection can reduce
the rate of inadequate specimens.
With chlamydial culture, it is not possible to tell if
an inadequate specimen has been collected. Cultures
recognized as being inadequate usually result from
bacterial contamination, most commonly from concurrent
bacterial infection, and antibiotics are included
in the collection and culture media to reduce
this. The rate of 2.1% inadequate cultures is acceptable
and compares very favorably with other reports,
albeit using genital specimens. 22 Our culture specimens
were frozen in liquid nitrogen, and this probably
caused some reduction in infectious titer of chlamydia
23 ; but, once frozen, the variation in the duration
of storage should not have led to a change in
titer. 8 Cultures were stained with a DFA reagent,
which is reported to be more sensitive than iodine
stain. 14 - 2425
Recently, some investigations have reported improved
results with the DFA test if methanol is used
as a fixative instead of acetone, 26 - 27 and this change
has subsequently been recommended by the manufacturer.
It is thought that methanol removes some
surface components that may expose more epitopes
for the monoclonal antibody reagent to recognize.
This effect is not thought to occur with a brief acetone
fixative but may occur with prolonged acetone fixation.
We used a prolonged acetone fixation in the
field, which was also repeated in the laboratory prior
to staining, and our results compare more than favorably
to those using methanol refixation. 27
There was good agreement between the two laboratory
tests. As would be expected, the performance of
DFA compared to culture varied with the criterion
for a positive test. Initially, a finding of ten EB was
recommended for the criterion for a positive DFA
test. 28 Many have used this cutoff, 101113 ' 26 - 29 although
others have suggested five, 30 " 33 three, 1434 two 35 or
even one EB I1>36 as being sufficient for the diagnosis
80
70
60
LU 50
w 40
o
0_
55 30
20
10
Culture
DFA
Culture or DFA
none TFonly TI TS only TT *
CLINICAL TRACHOMA GRADING
• without TF or TI
Fig. 1. Proportion of positive laboratory tests by clinical trachoma
grading.

1830 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / August 1989 Vol. 30
of a positive smear. Our data would suggest that five
EB is the optimal cut-off for ocular specimens, although
it should be recognized that bacterial contamination
is much more common with genital and rectal
specimens and a higher cutoff may still be appropriate
for these specimens.
Both laboratory tests correlated well with the clinical
grading. This is the first report to compare the
microbiologic status with the new simplified WHO
trachoma grading scheme, 16 and our findings provide
a laboratory validation of this system. An increasing
laboratory identification rate with increasing clinical
severity of inflammation has been reported by
others. 6 " 91213 ' 37 Others have also identified a small
number of people who were seemingly normal but
from whom organisms can be identified—"asymptomatic
carriers." Rates of positive tests in such people
have varied from about 2% 7 to 4% to 5%. 912 It is
possible that in some cases this could represent laboratory
error or contamination, which Schachter has
reported can occur at a rate of 0.5%. 38 However, we
found 47 people who did not have TF or TI and who
were positive on laboratory testing, some by culture
alone and some by DFA alone, but 17 (36%) by both
tests. Over half of these people had mild clinical disease
that was not of sufficient severity to meet the
criteria to be graded as either TF or TI and another
third had equivocal clinical findings. However, of the
47 people, there were six (13%) who had entirely normal-appearing
eyes five were women, none of whom
had trachomatous scarring. It is interesting to speculate
whether these six people were incubating their
initial infection, as suggested by Nichols, 7 or whether
they may be true asymptomatic carriers whose immunologic
status is somehow altered so they can tolerate
the presence of chlamydia without producing an
immunopathologic respose. Clearly, such a group of
patients is worthy of further investigation.
Of equal interest was the finding that 33% of patients
with severe clinical disease (TI) were initially
negative by laboratory testing. This finding is ia line
with previous reports, although the actual percentage
varies somewhat as different clinical grading schemes
and laboratory tests were used. 6 " 12 Each laboratory
test used to detect chlamydia clearly does not have
100% sensitivity, 1429 although culture with serial passage
does offer the possibility of amplification of infectious
particles that would lead to an increased sensitivity
that is not possible with other tests. Most specimens
found to be positive on culture were positive
during the first passage. The first passage has a reported
sensitivity which can vary between 65% to
97% 40 ' 41 compared to the overall culture results. A
second blind passage is often used routinely for chlamydial
culture. 14 ' 22 ' 38 We were interested to see if additional
blind passages of specimens from cases with
severe inflammatory disease (TI) would lead to a further
increase in positive cultures as has been reported
by others for genital cultures. 2240 Although others
have shown the general comparability of chlamydial
culture in microtiter plates and vials, 38 ' 42 we believed
it important to first validate the methods we use in
our laboratory as outlined above.
With the material from the patients in Tanzania,
we found an enhanced sensitivity of culture in vials
when serial passage was used. Of the 40 specimens
from patients with TI that were initially negative, 19
became positive when passed up to six times in vials.
The reason for increased sensitivity with multiple
blind passage is unclear. It is possible that the infectious
material could be clumped in the collection medium
and was not included in the original inoculum.
This tendency should be minimized by vortexing
prior to specimen transfer. Preformed tear antibodies
or other cytokines could prevent organism replication,
but given the marked dilutional effect and the
fact it only occurs in some specimens, this seems unlikely.
Length of storage was not correlated with a
delayed culture, although unrecorded minor variations
in initial specimen preparation and handling
cannot be excluded. It is possible that some biologic
variation in the ease with which the organism adapts
to cell culture could account for the appearance of
delayed positives. Multiple blind passage did increase
the number of culture positive specimens; however,
we would not recommend this labor-intensive and
extensive effort for the routine culture of chlamydia
and believe that in most circumstances a second blind
passage is adequate.
It is of interest that the DFA results did not predict
which specimens became positive on reculture in
vials, although the finding that some people with a
positive DFA test had negative cultures suggests the
presence of nonviable organisms in these smears.
However, despite the intensive efforts to identify organisms,
there were still 14 people with severe inflammatory
disease who did not have demonstrable organism
by either culture or DFA and ten who had
organism seen on DFA but who were culture-negative.
These people form a most interesting subgroup,
further study of whom may provide insights into the
pathogenesis of the disease.
Parallels for the presence of clinical disease in the
absence of demonstrable organisms have been found
in the animal models of trachoma. 2 Despite the need
for repeated weekly inoculation of viable organisms
to sustain chronic disease, chlamydia cannot be isolated
or identified after the first few months. 14 ' 43 A

No. 8 EPIDEMIOLOGY OF TRACHOMA / Taylor er al 1831
triton-extractable chlamydial antigen can elicit a delayed-type
hypersensitivity reaction in the conjunctiva
that is characteristic of trachoma. 4 ' 5 This suggests
that organisms that are not actively replicating may
be able to elaborate or release sufficient antigenic material
to sustain the conjunctival inflammatory response.
It is known that penicillin, for example, inhibits
chlamydial replication but does not stop production
of the triton-extractable antigen; rather,
penicillin treatment leads to an excessive production
(R. Morrison and H. Caldwell, personal communication,
1987). Antichlamydial antibodies rapidly appear
in the tears of those with trachoma and are capable
of neutralizing chlamydia. 44 However, despite the
presence of high titers of tear antibodies, chlamydia
can routinely be isolated from the conjunctiva, and it
seems unlikely that antibodies are responsible for the
observed phenomenon. Recently, gamma interferon
has been shown to inhibit the growth of chlamydia.
45 " 48 It is possible that this or other cytokines
could lead to the arrest of chlamydial replication with
subsequent antigen release. This suggests that the immune
response may be enough to block chlamydial
replication but not enough to completely eliminate
infection; and clearly, the immune response is not
capable of preventing further exposure to new inoculations
of infectious organisms even if it were capable
of preventing the establishment of new episodes of
infection. Hence, antigenic products could be released
by either persistent intracellular organisms or
freshly inoculated but rapidly neutralized organisms.
These antigens could further enhance the immune
response. However, the immunopathogenetic mechanisms
still require elucidation, and those patients
with severe disease in the absence of demonstrable
chlamydia form an important group for further
study.
Finally, it was of interest to examine the serotype of
chlamydia associated with trachoma in this part of
Africa. The finding of serovars A and B is consistent
with reports from other areas. 17 ' 49 " 51 Although the
members of only one family were typed, all shared
the same serovar; and similar findings have been reported
from Saudi Arabia 7 and Taiwan. 1 ' 52 The finding
of a single serotype within a family supports the
contention that intrafamily transmission is of prime
importance. There was no correlation between serovar
and disease severity. The increased frequency of
serovar B in older people is unexplained but was also
reported in Saudi Arabia. 7
This study demonstrated the advantages of the recently
developed DFA method over chlamydial culture
for use in trachoma field studies. The simplicity
of DFA, particularly the lack of the requirement for a
cold chain, greatly facilitates field work. We have
confirmed the presence of chlamydial infection with
the serovars A and B in this population in Central
Tanzania. The identification of groups of people with
proven infection in the absence of clinical disease and
those with clinical disease in the absence of demonstrable
agent provides important targets for further
investigations on the dynamics of infection within the
family transmission unit.
Key words: trachoma, laboratory diagnosis, culture, direct
fluorescent antibody cytology, clinical grading
Acknowledgments
The authors wish to thank all those who assisted with this
project, in particular, Hon. Dr. A. Chiduo, MP, Minister of
Health; Mr. Mapunda, former Regional Development Director,
Dodoma; Dr. J. M. Temba, former Regional Medical
Officer, Dodoma, and now Assistant Chief Medical Officer
and Director of Preventive Services, Ministry of Health;
Mr. Mluwande, Mpwapwa District CCM Chairman; Dr.
Kibauri, former District Medical Officer, and Mr. Mwasakieni,
Administrative Officer-in-Charge, Kongwa Subdistrict,
both of Mpwapwa; all the Village Chairmen and Secretaries
of the villages that were surveyed; Dr. Joseph Taylor,
Moshi, Africa Representative of Christoffel
Blindenmission; the staff of Helen Keller International,
Inc., New York; and Mesdames Vivian Velez, Lori DeJong,
Ruth Barfield, Celeste Wilson and Alice Flumbaum and
Mr. Dan Fisher of The Johns Hopkins University. This
study was conducted under the auspices of the National
Prevention of Blindness Committee of Tanzania and was
supported by a grant from The Edna McConnell Clark
Foundation.
References
1. Grayston JT and Wang S: New knowledge of chlamydiae and
the diseases they cause. J Infect Dis 132:87, 1975.
2. Taylor HR, Johnson SL, Prendergast RA, Schachter J, Dawson
CR, and Silverstein AM: An animal model of trachoma: II.
The importance of repeated reinfection. Invest Ophthalmol
Vis Sci 23:507, 1982.
3. Taylor CE, Gulati PV, and Harinarain J: Eye infections in a
Punjab village. Am J Trop Med Hyg 7:42, 1958.
4. Watkins NG, Hadlow WJ, Moos AB, and Caldwell HD: Ocular
delayed hypersensitivity: A pathogenetic mechanism of
chlamydial conjunctivitis in guinea pigs. Proc Natl Acad Sci
USA 83:7480, 1986.
5. Taylor HR, Johnson SL, Schachter J, Caldwell HD, and Prendergast
RA: Pathogenesis of trachoma: The stimulus for inflammation.
J Immunol 38:3023, 1987.
6. Sowa S, Sowa J, Collier LH, and Blyth W: Trachoma and allied
infections in a Gambian village. Medical Research Council,
Special Report Series 308. London, Her Majesty's Stationery
Office, 1965, pp. 1-88.
7. Nichols RL, Von Fritzinger K, and McComb DE: Epidemiological
data derived from immunotyping of 338 trachoma
strains isolated from children in Saudi Arabia. In Trachoma
and Related Disorders, Nichols RL, editor. Amsterdam, Excerpta
Medica, 1971, pp. 337-358.

1832 INVESTIGATIVE OPHTHALMOLOGY b VISUAL SCIENCE / Augusr 1989 Vol. 30
8. Nichols RL, Murray ES, Scott PP, and McComb DE: Trachoma
isolation studies in Saudi Arabia from 1957 through
1969. In Trachoma and Related Disorders, Nichols RL, editor.
Amsterdam, Excerpta Medica, 1971, pp. 517-528.
9. Dawson CR, Daghfous T, Messadi M, Hoshiwara I, and
Schachter J: Severe endemic trachoma in Tunisia. Br J Ophthalmol
60:245, 1976.
10. Wilson MC, Millan-Velasco F, Tielsch JM, and Taylor HR:
Direct-smear fluorescent antibody cytology as a fielddiagnostic
tool for trachoma. Arch Ophthalmol 104:688, 1986.
11. Mabey DCW and Booth-Mason S: The detection of Chlamydia
trachomatis by direct immunofluorescence in conjunctival
smears from patients with trachoma and patients with
ophthalmia neonatorum using a conjugated monoclonal antibody.
J Hyg Camb 96:83, 1986.
12. Mabey DCW, Robertson JN, and Ward ME: Detection of
Chlamydia trachomatis by enzyme immunoassay in patients
with trachoma. Lancet 2:1491, 1987.
13. Barsoum IS, Mostafa MSE, Shihab AA, El Alamy M, Habib
MA, and Coley DG: Prevalence of trachoma in school children
and ophthalmological outpatients in rural Egypt. Am J Trop
Med Hyg 36:97, 1987.
14. Taylor HR, Agarwala N, and Johnson S: The detection of
Chlamydia trachomatis eye infection in conjunctival smears
and in tissue culture with a fluoresceinconjugated monoclonal
antibody. J Clin Microbiol 20:391, 1984.
15. Nabli B: The infectivity of trachoma: Stage IV. Rev Int Trach
50:19, 1973.
16. Thylefors B, Dawson CR, Jones BR, West SK, and Taylor HR:
A simple system for the assessment of trachoma and its complications.
Bull WHO 65:477, 1987.
17. West S, Lynch M, Turner V, Munoz B, Rapoza P, Mmbaga
BBO, and Taylor HR: Water availability and trachoma. Bull
WHO 67:71, 1989.
18. Taylor HR, West SK, Katala S, and Foster A: Trachoma:
Evaluation of a new grading scheme in the United Republic of
Tanzania. Bull WHO 65:485, 1987.
19. Taylor HR and Velez VL: Clearance of chlamydial elementary
bodies from the conjunctival sac. Invest Ophthalmol Vis Sci
28:1199, 1987.
20. Wang S-P and Grayston JT: Immunologic relationship between
genital trie, lymphogranuloma venereum, and related
organisms in a new microtiter indirect immunofluorescence
test. Am J Ophthalmol 70:367, 1970.
21. Wang S-P, Kuo CC, Barnes RC, Stephens RS, and Grayston
JT: Immunotyping of Chlamydia trachomatis with monoclonal
antibodies. J Infect Dis 152:791, 1985.
22. Schachter J and Martin DH: Failure of multiple passages to
increase chlamydial recovery. J Clin Microbiol 25:1851, 1987.
23. Sabet SF, Lambert F, and Dalton HP: Infectivity index of
Chlamydia trachomatis serotypes and relevance to sensitivities
of available diagnostic methods. Abstracts of the Annual Meeting
for the American Society of Microbiology, 1987, p. 345.
24. Nowinski RC, Tarn MR, Goldstein LC, Stong L, Kuo C-C,
Corey L, Stamm WE, Handsfield HH, Knapp JS, and Holmes
KK: Monoclonal antibodies for diagnosis of infectious diseases
in humans. Science 219:637, 1983.
25. Stamm WE, Tarn M, Koester M, and Cles L: Detection of
Chlamydia trachomatis inclusions in the McCoy cell cultures
with fluorescein-conjugated monoclonal antibodies. J Clin
Microbiol 17:666, 1983.
26. Dawson C, Sheppard J, Moncada J, Courtright P, Hafez S,
Schachter J, and Lorrinez A: Laboratory diagnosis of chlamydial
infection in hyperendemic trachoma. ARVO Abstracts. Invest
Ophthalmol Vis Sci 29(Suppl):227, 1988.
27. Dean D, Pant C, and O'Hanley P: Use of a Chlamydia trachomatis
DNA probe in comparison with fluorescent antibody
and culture as a field diagnostic technique for trachoma. Abstracts
of the Annual Meeting of the American Society for
Microbiology, 1988, p. C-235.
28. Tarn MR, Stamm WE, Handsfield HH, Stephens R, Kuo CC,
Holmes KK, Ditzenberger K, Krieger M, and Nowinski RC:
Culture-independent diagnosis of Chlamydia trachomatis
using monoclonal antibodies. N Engl J Med 310:1146, 1984.
29. Lipkin ES, Moncada JV, Shafer M-A, Wilson TE, and
Schachter J: Comparison of monoclonal antibody staining and
culture in diagnosing cervical chlamydial infection. J Clin Microbiol
23:114, 1986.
30. Bell TA, Kuo C-C, Stamm WE, Tarn MR, Stephens RS,
Holmes KK, and Grayston JT: Direct fluorescentmonoclonal
antibody stain for rapid detection of infant Chlamydia trachomatis
infections. Pediatrics 74:224, 1984.
31. Rapoza PA, Quinn TC, Kiessling LA, Green WR, and Taylor
HR: Assessment of neonatal conjunctivitis with a direct immunofluorescent
monoclonal antibody stain for Chlamydia.
JAMA 225:3369, 1986.
32. Potts MJ, Paul ID, Roome APCH, and Caul EO: Rapid diagnosis
of Chlamydia trachomatis infection in patients attending
an ophthalmic casualty department. Br J Ophthalmol 70:677,
1986.
33. Phillips RS, Hanff PA, Kauffman RS, and Aronson MD: Use
of a direct fluorescent antibody test for detecting Chlamydia
trachomatis cervical infection in women seeking routine gynecologic
care. J Infect Dis 156:575, 1987.
34. Toomey KE, Rafferty MP, and Stamm WE: Unrecognized
high prevalence of Chlamydia trachomatis cervical infection in
an isolated Alaskan Eskimo population. JAMA 258:53, 1987.
35. Stamm WE, Harrison HR, Alexander ER, Cles LD, Spence
MR, and Quinn TC: Diagnosis of Chlamydia trachomatis infections
by direct immunofluorescence staining of genital secretions.
Ann Intern Med 101:638, 1984.
36. Thomas BJ, Evans RT, Hawkins DA and Taylor-Robinson D:
Sensitivity of detecting Chlamydia trachomatis elementary
bodies in smears by use of a fluoroscein labeled monoclonal
antibody: Comparison with conventional chlamydial isolation.
J Clin Pathol 37:812, 1984.
37. Daghfous MT, Dawson CR, Kamoun M, Romdhane K, and
Djerad A: Action de la doxycycline par voie orale et de la
tetracycline pommade 1% dans le traitment du trachome severe.
Rev Int Trach Pathol Ocul Trop Subtrop Sante Publique
1-2:47, 1985.
38. Schachter J: Immunodiagnosis of sexually transmitted disease.
Yale J Biol Med 58:443, 1985.
39. Wingerson L: Two new tests for chlamydia get quick results
without culture. JAMA 250:2257, 1983.
40. Jones RB, Katz BP, van der Pol B, Caine VA, Batteiger BE,
and Newhall WJ: Effect of blind passage and multiple sampling
on recovery of Chlamydia trachomatis from urogenital
specimens. J Clin Microbiol 24:1029, 1986.
41. Bidwell CE, Brumback BG, Morris MV, and Bailey JL: Comparison
of single inoculation with blind passage for the isolation
of Chlamydia trachomatis in a microculture system. Abstracts
of the Annual Meeting for the American Society of
Microbiology, 1987, p. 345.
42. Yoder BL, Stamm WE, Koester CM, and Alexander ER: Mi-

No. 8 EPIDEMIOLOGY OF TRACHOMA / Taylor er al 1833
crotest procedure for isolation of Chlamydia trachomatis. J
Clin Microbiol 14:325, 1981.
43. Taylor HR, Prendergast RA, Dawson CR, Schachter J, and
Silverstein AM: Animal model of trachoma: III. The necessity
of repeated exposure to live chlamydia. In Chlamydial Infections,
Mardh P-A, Holmes KK, Oriel JD, Piot P, and
Schachter J, editors. Amsterdam, Elsevier Biomedical Press,
1982, pp. 387-390.
44. Barenfanger J and MacDonald AB: The role of immunoglobulin
in the neutralization of trachoma infectivity. J Immunol
113:1607, 1974.
45. Rothermel CD, Rubin BY, and Murray HW: Gamma-interferon
is the factor in lymphokine that activates human macrophages
to inhibit intracellular Chlamydia psittaci replication. J
Immunol 131:2542, 1983.
46. Byrne GI and Krueger DA: In vitro expression of factor-mediated
cytotoxic activity generated during the immune response
to Chlamydia in the mouse. J Immunol 134:4189,
1985.
47. Rothermel CD, Rubin BY, Jaffe EA, and Murray HW: Oxygen-independent
inhibition of intracellular Chlamydia psittaci
growth by human monocytes and interferon-7 activated macrophages.
J Immunol 137:689, 1986.
48. Williams DM, Grubbs B, Marshall TJ, Schachter J, and Byrne
GI: Gamma interferon-mediated cytotoxicity related murine
Chlamydia trachomatis infection. Abstracts of the Annual
Meeting for the American Society of Microbiology, 1988,
p. D-14.
49. Graham DM: A report on microbiological findings. In The
National Trachoma and Eye Health Program of The Royal
Australian College of Ophthalmologists. Sydney, The Royal
Australian College of Ophthalmologists, 1980, pp. 29-36.
50. Treharne JD: The community epidemiology of trachoma. Rev
Infect Dis 7:760, 1985.
51. Mabey DCW, Forsey T, and Treharne JD: Serotypes of Chlamydia
trachomatis in the Gambia. Lancet 2:452, 1987.
52. Grayston JT, Wang S-P, Yeh L-J, and Kuo C-C: Importance of
reinfection in the pathogenesis of trachoma. Rev Infect Dis
7:717, 1985.