DIAGN MICROBIOLINFECT DIS 1991;14:233-237
233
Correlation of the Percent of Positive Chlamydia trachomatis Direct Fluorescent Antibody Detection Tests with the Adequacy of Specimen Collection Cynthia Howard, Debbie L. Friedman, Jean K. Leete, and Mary L. Christensen
Chlamydia trachomatis is an obligate, intracellular parasite infecting the columnar and transitional cells lining the endocervix, uterus, fallopian tubes, rectum, urethra, and epididymis. We determined if the percent of specimens positive for C. trachomatis in the Microtrak Direct Specimen Test depended on the quality of specimens obtained. Female genital slides (649) were evaluated by the direct fluorescent antibody (DFA) test for the presence and numbers of (a) C. trachoma~s elementary bodies and (b) columnar, transitional and squamous epithelial cells, and polymorphonuclear neutrophils (PMNs). Only 138 (21.3%) of the 649 slides were considered to be adequately taken, that is, containing columnar/transitional cells either alone or in conjunction with squamous cells
and/or PMNs. Of the 138 adequate slides, 10 (7.2%) were C. trachomatis positive. However, 511 (78.7%) of the 649 slides were judged inadequate; 395 contained only squamous cells and/or PMNs, 19 were too thick to determine cell types, 46 contained only cell debris, and 51 contained neither cells nor debris. Only four (0.78%) of 511 were C. trachomatis positive. Thus adequate specimens containing columnar~transitional cells for C. trachornatis detection had a tenfold increase in the percent of positive results compared to inadequately collected specimens. By using the DFA test, one has the advantage of determining the adequacy of the specimens obtained as well as the presence of chlamydiae.
INTRODUCTION
urethritis (Holmes et al., 1975), a n d proctitis (Goldmeier and Darougar, 1977). If not diagnosed and treated, C. trachomatis infection can lead to endometritis (Schachter and Grossman, 1981), salpingitis (Mardh et al., 1977), ectopic p r e g n a n c y (Westron, 1985), and epididymitis (Berger et al., 1978), and can result in sterility in w o m e n a n d men. It is an obligate intracellular parasite (Schachter a n d Grossman, 1981) that infects columnar a n d transitional (cuboidal) cells lining the endocervix, uterus, fallopian tubes, rectum, urethra, a n d epididymis (Holmes, 1981). This organism is f o u n d in columnar a n d transitional cells in two forms: smaller infectious, nonreplicating elementary bodies, a n d larger, noninfectious, metabolically active replicating reticulate (initial) bodies (Holmes, 1981; Schachter and Grossman, 1981). To diagnose accurately C. trachomatis infection in the laboratory by the direct fluorescent antibody (DFA)
Chlamydia trachomatis infection is one of the most prevalent sexually transmitted diseases today, with estimates of 500,000 to 3 million n e w cases occurring annually (Holmes, 1981). Chlamydia trachomatis is a significant cause of cervicitis (Oriel et al., 1978),
From the Department of Pathology, The Children's Memorial Hospital (C.H., D.L.F., M.L.C.); the Departments of Pathology and Pediatrics, Northwestern University Medical School (M.L.C.); and the Central Pathology Regional Laboratories (J.K.L.), Chicago, Illinois, USA. Address reprint requests to Dr. M.L. Christensen, The Children's Memorial Hospital, 2300 Children's Plaza, Chicago, IL 60614, USA. Present address of Dr. Leete: Diagnostics Division, Abbott Laboratories, North Chicago, IL 60064. Received and accepted 1 August 1990. © 1991 Elsevier Science Publishing Co., Inc. 655 Avenue of the Americas 0732-8893/91/$3.50
234
test, enzyme immunoassay (EIA), or culture, one must obtain adequate specimens containing columnar and/or transitional cells (Holmes, 1981). The DFA test, being a detection test using microscopic examination, is an excellent method by which to evaluate the quality of specimens for the presence of appropriate epithelial cells as well as for C. trachomatis organisms. In this study, 649 female genital swab specimens smeared on slides were examined using a fluorescein-conjugated monoclonal antibody to C. trachomatis. By this method, we could determine the percent of specimens adequately obtained, that is, containing columnar/transitional cells, and we correlated this with the percent of specimens positive for C. trachomatis.
C. Howard et al.
TABLE 1 Distribution of Numbers of Specimens from Females in Various Age Groups Age (Years)
No. of Specimens
13-14 15-19 20-24 25-29 30-34 35-39 40--44 4549 50-55 70-75 88 Age unspecified
2 41 61 79 57 26 11 4 3 2 1 362
Total
649
MATERIALS AND METHODS Specimens A total of 649 female genital specimens were obtained from (a) private physicians, (b) health maintenance organizations (HMO), and (c) hospital clinics in the Chicago metropolitan area that were submitted by a reference laboratory, Central Pathology Regional Laboratories, to the Virology Laboratory at the Children's Memorial Hospital (CMH) in Chicago, Illinois, for staining and evaluation. DFA has the advantage of detecting chlamydiae more frequently than does culture when transit of specimens from other geographic sites causes a delay in culturing and a loss of viability (Mahony and Chernesky, 1985; Williams et al., 1985). The ages of the patients from w h o m specimens were obtained (Table 1) ranged from 13 years to 88 years. Data on the clinical status of the patients were not available.
Specimen Collection and Slide Preparation Specimens were collected using sterile swabs and slides from the Syva MicroTrak Specimen Collection Kits, which were provided to physicians, HMOs, and hospital clinics. Directions for specimen collection are provided in the kit and stress that collection of an adequate specimen is critical. For cervical specimens, directions include wiping excess mucous from the exocervix and inserting and rotating the collection swab in the endocervical canal so that most of the tip is not visible. This is important as the squamocolumnar junction recedes more deeply into the endocervix with increased age. Directions for slide preparation are included in the kit and include (a) having visible material on the well and (b) air drying the specimen completely before fixing (as opposed to Papanicolaou smears, which are fixed wet), using fixative provided in the collection kit. After the slides
were received at the CMH Virology Laboratory they were stained with MicroTrak fluorescein-conjugated monoclonal antibody to C. trachomatis according to kit directions.
Microscopic Evaluation of Slides Immediately after mounting, the slides were read by a trained microscopist as follows. The slides were scanned at a magnification of x 400 (low power) in a regular, overlapping manner from edge to edge and any observed C. trachomatis elementary bodies were confirmed at a magnification of x 1000 (high power). The presence of 10 or more elementary bodies on a slide was considered positive for C. trachomatis infection. The presence of 1-9 elementary bodies on a slide was considered suspicious for C. trachomatis infection, and no observed elementary bodies were considered negative. The slides were also evaluated for the types and numbers of host cells, that is, columnar/transitional epithelial cells, squamous epithelial cells, PMNs, or cell debris, all counterstained with Evan's blue. Numbers of specific host cells were graded as none, rare to occasional (150 per well or >3 per low-power field). The positive detection rates of the two groups of specimens, the adequately, and inadequately collected groups were compared using chi-square statistics.
RESULTS Results of the microscopic evaluation of 649 female genital specimens are shown in Table 2. Only 138
Adequate Chlamydia trachomatis Test Specimens
TABLE 2
235
Correlation of the Positive Rate of Chlamydia trachomatis Detection with Host-Cell Type Obtained in Specimens from Females
Specimen Group Based on Cell Type
No. (%) Tested
Adequate slides (total) Colum/trans. cells only Colurn/trans. + squamous cells Colum/trans. + PMNs ColutWtrans. + squamous + PMNs
138 (21.3%) 78 (12.0%) 26 (4.0%) 22 (3.4%) 12 (1.9%)
10 6 1 0 3
(7.2%)* (4.3%) (0.7%) (0%) (2.2%)
Inadequate slides (total) Squamous cells only PMNs only Squamous + PMNs Too thick to read cell type Cell debris and/or mucus No cells on slide
511 (78.7%) 236 (36.4%) 111 (17.1%) 48 (7.4%) 19 (2.9%) 46 (7.1%) 51 (7.8%)
4 0 1 2 1 0
(0.78%) (0%) (0.195%) (0.39%) (0.195%) (0%) (0%)
Total slides
649
No. (%) Positive
o
(100%)
14
(2.2%)
Colum/trans., columnar/transitional. *p < 0.05.
(21.3%) of the 649 slides were considered to be adequately collected and contained columnar and/or transitional cells, either (a) alone or (b) mixed with squamous epithelial cells and/or PMNs. In this group, 78 (12.0%) contained columnar and/or transitional cells alone, 26 (4.0%) contained these cells plus squamous epithelial cells, 22 (3.4%) contained these cells plus PMNs, and 12 (1.9%) contained these cells plus PMNs and squamous epithelial cells. Of the 138 adequate slides, 10 (7.2%) were positive for 10 or more C. trachomatis elementary bodies. However, 511 (78.7%) of the 649 female genital slides were considered inadequate, containing no visible columnar or transitional cells. Of the 511, 236 (36.4%) contained only squamous epithelial cells, 111 (17.1%) contained only PMNs, and 48 (7.4%) contained a mixture of squamous cells plus PMNs. On 19 (2.9%) of the slides, smears were too thick to be able to determine the cell type, whereas 46 (7.1%) of the slides contained only cell debris and/or mucous, and 51 (7.8%) of the slides had nothing, that is, no cells or debris on them. Of these 511 slides considered to be inadequately prepared, only four (0.78%) were positive for C. trachomatis elementary bodies. Thus there was an approximate tenfold reduction in the percent of positive slides that were not correctly obtained, versus those that were appropriately obtained (p < 0.05). The correlation of numbers of positive specimens to the numbers of columnar/transitional cells present in specimens from females are seen in Table 3. Of the specimen slides, 36 contained only rare to occasional columnar/transitional cells, of which one
TABLE 3
Correlation of the Positive Rate of Chlamydia trachomatis Detection with Numbers of Columnar/Transitional Cells Present on Specimen Slides from Females
Nos. of Columnar/Transitional Cells on Slide No. of Isolates No. (%) Positive Rare to occasional (1-24) Moderate to many (i>25)
36
1 (2.8%)
102
9 (8.8%)
Total
138
10 (7.2%)
(2.8%) was positive; 102 specimen slides contained moderate to many columnar/transitional cells, of which nine (8.8%) were positive. Thus there appears to be a correlation between the numbers of columnar/transitional cells observed and the number of positive chlamydial specimens. In the same study, 171 adolescent and adult male urethral slides were examined. Sampling directions specify inserting and rotating the swab 2-4 cm into the urethra to obtain columnar cells, as the terminal section of the male urethra is lined with squamous epithelium. Patients should not have urinated within 1 hr of sampling. Of these 171 slides, 73 (42.7%) were adequate, containing columnar cells either alone or in combination with other cell types. Of these 73, five (6.8%) were positive. However, 98 (57.3%) of
236
171 were considered inadequately collected, containing no columnar cells and only three (3.1%) of 98 were positive. Of these 98, 17 (9.9%) contained only squamous ceils, 28 (16.4%) contained only PMNs, 12 (7.0%) contained squamous cells and PMNs, 1 (0.6%) was too thick to read the cell type, 12 (7.0%) contained only cell debris and/or mucus, and 28 (16.4%) had no cells on the slide. However, no significant difference between the adequately and inadequately collected groups were observed, due to the fewer numbers of positive specimens occurring in both groups.
DISCUSSION When 649 female genital smear slides were examined by a direct fluorescent antibody technique for adequateness of specimen collection as well as for the presence of C. trachomatis, only 21% were appropriately taken, containing columnar/transitional epithelial cells, either alone or in conjunction with squamous epithelial cells and/or PMNs. Obtaining specimens from anatomic sites in which columnar/ transitional cells are found is important as these are the host cells that the obligate intracellular parasite C. trachomatis infects. In this study, - 3 . 7 times as many specimens were inadequately obtained compared to those adequately obtained (511 vs 138). Only 0.78% of inadequately obtained specimens were positive, versus 7.2% of adequately obtained specimens. Thus, theoretically, 6.4% of specimens or 32 of 511 were being misdiagnosed as negatives that may have been true positives. Because the MicroTrak reagent used contains the counterstain Evan's blue, which stains mammalian cells and debris, the DFA technique has the advantage of determining the quality of the specimens obtained, and is an inherent part of evaluating slides for C. trachomatis. Adequately obtained specimens containing columnar/transitional cells should also be obtained for other C. trachomatis detection methods, that is, EIA and culture. However, provisions for evaluating the quality of the specimens are not an inherent part of these two other techniques. Better specimen sampling may be obtained with cervical cytobrush (Linder et al., 1987), although this technique cannot be used in pregnant women, and may cause transient bleeding in patients with chlamydial cervicitis due to the friability of these cervices (Moscicki et al., 1987). In addition, PMN-containing mucus needs to be wiped away prior to endocervical sampling, even though PMNs are associated with chlamydial infection of the cervix (Kiviat et al., 1985; Moscicki et al., 1987). Inadequately obtained specimens for chlamydia detection appears to be a major problem, at least in
C. Howard et al.
this study. In one other study (Phillips et al., 1987), 45% of the specimens had no columnar cells and 38% of the specimens had some mucous partially obscuring the slide; in the study, the sensitivity of the DFA test compared to culture was 70%, but increased to 92% if five or more columnar cells were observed on a slide and decreased to 40% if fewer than five columnar cells were observed. In two other reports, 12% of the specimens obtained for DFA were reported as inadequately taken, but chlamydia resuits were not reported to the physicians (Lipkin et al., 1986; Schafer et al., 1986). The relatively low percentage of inadequate specimens in these two reports may have been due to the specimens being submitted from a very limited number of sites (1-3 clinics) in which clinicians may have had more thorough training in specimen collection. Slide variation may be greater as the number of collection locations increases, leading to variability in clinician training and collection techniques. The MicroTrak directions suggest reporting "indeterminate" results when inadequate specimens lacking columnar/transitional cells are obtained. However, we do not reject specimens that are inadequate as (a) we found that close to 1% of these specimens were chlamydiae positive and (b) we prefer to report a description of cell types observed along with an explanation that chlamydiae replicate in columnar/transitional epithelial cells. Our reporting on the type and amount of cells, as well as the presence of cell debris, mucus, and/or lack of cells gives more information to the physicians and may alert them as to problems encountered in obtaining specimens for chlamydial detection. Problems with false-negative results due to sampling errors have also been reported for cervical smears taken for the detection of cervical intraepithelial neoplasia. Gay et al. (1985) found a falsenegative rate of 63% due to sampling failure; Duguid (1986) reported that >50% of the smears taken by general practitioners contained
233
Correlation of the Percent of Positive Chlamydia trachomatis Direct Fluorescent Antibody Detection Tests with the Adequacy of Specimen Collection Cynthia Howard, Debbie L. Friedman, Jean K. Leete, and Mary L. Christensen
Chlamydia trachomatis is an obligate, intracellular parasite infecting the columnar and transitional cells lining the endocervix, uterus, fallopian tubes, rectum, urethra, and epididymis. We determined if the percent of specimens positive for C. trachomatis in the Microtrak Direct Specimen Test depended on the quality of specimens obtained. Female genital slides (649) were evaluated by the direct fluorescent antibody (DFA) test for the presence and numbers of (a) C. trachoma~s elementary bodies and (b) columnar, transitional and squamous epithelial cells, and polymorphonuclear neutrophils (PMNs). Only 138 (21.3%) of the 649 slides were considered to be adequately taken, that is, containing columnar/transitional cells either alone or in conjunction with squamous cells
and/or PMNs. Of the 138 adequate slides, 10 (7.2%) were C. trachomatis positive. However, 511 (78.7%) of the 649 slides were judged inadequate; 395 contained only squamous cells and/or PMNs, 19 were too thick to determine cell types, 46 contained only cell debris, and 51 contained neither cells nor debris. Only four (0.78%) of 511 were C. trachomatis positive. Thus adequate specimens containing columnar~transitional cells for C. trachornatis detection had a tenfold increase in the percent of positive results compared to inadequately collected specimens. By using the DFA test, one has the advantage of determining the adequacy of the specimens obtained as well as the presence of chlamydiae.
INTRODUCTION
urethritis (Holmes et al., 1975), a n d proctitis (Goldmeier and Darougar, 1977). If not diagnosed and treated, C. trachomatis infection can lead to endometritis (Schachter and Grossman, 1981), salpingitis (Mardh et al., 1977), ectopic p r e g n a n c y (Westron, 1985), and epididymitis (Berger et al., 1978), and can result in sterility in w o m e n a n d men. It is an obligate intracellular parasite (Schachter a n d Grossman, 1981) that infects columnar a n d transitional (cuboidal) cells lining the endocervix, uterus, fallopian tubes, rectum, urethra, a n d epididymis (Holmes, 1981). This organism is f o u n d in columnar a n d transitional cells in two forms: smaller infectious, nonreplicating elementary bodies, a n d larger, noninfectious, metabolically active replicating reticulate (initial) bodies (Holmes, 1981; Schachter and Grossman, 1981). To diagnose accurately C. trachomatis infection in the laboratory by the direct fluorescent antibody (DFA)
Chlamydia trachomatis infection is one of the most prevalent sexually transmitted diseases today, with estimates of 500,000 to 3 million n e w cases occurring annually (Holmes, 1981). Chlamydia trachomatis is a significant cause of cervicitis (Oriel et al., 1978),
From the Department of Pathology, The Children's Memorial Hospital (C.H., D.L.F., M.L.C.); the Departments of Pathology and Pediatrics, Northwestern University Medical School (M.L.C.); and the Central Pathology Regional Laboratories (J.K.L.), Chicago, Illinois, USA. Address reprint requests to Dr. M.L. Christensen, The Children's Memorial Hospital, 2300 Children's Plaza, Chicago, IL 60614, USA. Present address of Dr. Leete: Diagnostics Division, Abbott Laboratories, North Chicago, IL 60064. Received and accepted 1 August 1990. © 1991 Elsevier Science Publishing Co., Inc. 655 Avenue of the Americas 0732-8893/91/$3.50
234
test, enzyme immunoassay (EIA), or culture, one must obtain adequate specimens containing columnar and/or transitional cells (Holmes, 1981). The DFA test, being a detection test using microscopic examination, is an excellent method by which to evaluate the quality of specimens for the presence of appropriate epithelial cells as well as for C. trachomatis organisms. In this study, 649 female genital swab specimens smeared on slides were examined using a fluorescein-conjugated monoclonal antibody to C. trachomatis. By this method, we could determine the percent of specimens adequately obtained, that is, containing columnar/transitional cells, and we correlated this with the percent of specimens positive for C. trachomatis.
C. Howard et al.
TABLE 1 Distribution of Numbers of Specimens from Females in Various Age Groups Age (Years)
No. of Specimens
13-14 15-19 20-24 25-29 30-34 35-39 40--44 4549 50-55 70-75 88 Age unspecified
2 41 61 79 57 26 11 4 3 2 1 362
Total
649
MATERIALS AND METHODS Specimens A total of 649 female genital specimens were obtained from (a) private physicians, (b) health maintenance organizations (HMO), and (c) hospital clinics in the Chicago metropolitan area that were submitted by a reference laboratory, Central Pathology Regional Laboratories, to the Virology Laboratory at the Children's Memorial Hospital (CMH) in Chicago, Illinois, for staining and evaluation. DFA has the advantage of detecting chlamydiae more frequently than does culture when transit of specimens from other geographic sites causes a delay in culturing and a loss of viability (Mahony and Chernesky, 1985; Williams et al., 1985). The ages of the patients from w h o m specimens were obtained (Table 1) ranged from 13 years to 88 years. Data on the clinical status of the patients were not available.
Specimen Collection and Slide Preparation Specimens were collected using sterile swabs and slides from the Syva MicroTrak Specimen Collection Kits, which were provided to physicians, HMOs, and hospital clinics. Directions for specimen collection are provided in the kit and stress that collection of an adequate specimen is critical. For cervical specimens, directions include wiping excess mucous from the exocervix and inserting and rotating the collection swab in the endocervical canal so that most of the tip is not visible. This is important as the squamocolumnar junction recedes more deeply into the endocervix with increased age. Directions for slide preparation are included in the kit and include (a) having visible material on the well and (b) air drying the specimen completely before fixing (as opposed to Papanicolaou smears, which are fixed wet), using fixative provided in the collection kit. After the slides
were received at the CMH Virology Laboratory they were stained with MicroTrak fluorescein-conjugated monoclonal antibody to C. trachomatis according to kit directions.
Microscopic Evaluation of Slides Immediately after mounting, the slides were read by a trained microscopist as follows. The slides were scanned at a magnification of x 400 (low power) in a regular, overlapping manner from edge to edge and any observed C. trachomatis elementary bodies were confirmed at a magnification of x 1000 (high power). The presence of 10 or more elementary bodies on a slide was considered positive for C. trachomatis infection. The presence of 1-9 elementary bodies on a slide was considered suspicious for C. trachomatis infection, and no observed elementary bodies were considered negative. The slides were also evaluated for the types and numbers of host cells, that is, columnar/transitional epithelial cells, squamous epithelial cells, PMNs, or cell debris, all counterstained with Evan's blue. Numbers of specific host cells were graded as none, rare to occasional (150 per well or >3 per low-power field). The positive detection rates of the two groups of specimens, the adequately, and inadequately collected groups were compared using chi-square statistics.
RESULTS Results of the microscopic evaluation of 649 female genital specimens are shown in Table 2. Only 138
Adequate Chlamydia trachomatis Test Specimens
TABLE 2
235
Correlation of the Positive Rate of Chlamydia trachomatis Detection with Host-Cell Type Obtained in Specimens from Females
Specimen Group Based on Cell Type
No. (%) Tested
Adequate slides (total) Colum/trans. cells only Colurn/trans. + squamous cells Colum/trans. + PMNs ColutWtrans. + squamous + PMNs
138 (21.3%) 78 (12.0%) 26 (4.0%) 22 (3.4%) 12 (1.9%)
10 6 1 0 3
(7.2%)* (4.3%) (0.7%) (0%) (2.2%)
Inadequate slides (total) Squamous cells only PMNs only Squamous + PMNs Too thick to read cell type Cell debris and/or mucus No cells on slide
511 (78.7%) 236 (36.4%) 111 (17.1%) 48 (7.4%) 19 (2.9%) 46 (7.1%) 51 (7.8%)
4 0 1 2 1 0
(0.78%) (0%) (0.195%) (0.39%) (0.195%) (0%) (0%)
Total slides
649
No. (%) Positive
o
(100%)
14
(2.2%)
Colum/trans., columnar/transitional. *p < 0.05.
(21.3%) of the 649 slides were considered to be adequately collected and contained columnar and/or transitional cells, either (a) alone or (b) mixed with squamous epithelial cells and/or PMNs. In this group, 78 (12.0%) contained columnar and/or transitional cells alone, 26 (4.0%) contained these cells plus squamous epithelial cells, 22 (3.4%) contained these cells plus PMNs, and 12 (1.9%) contained these cells plus PMNs and squamous epithelial cells. Of the 138 adequate slides, 10 (7.2%) were positive for 10 or more C. trachomatis elementary bodies. However, 511 (78.7%) of the 649 female genital slides were considered inadequate, containing no visible columnar or transitional cells. Of the 511, 236 (36.4%) contained only squamous epithelial cells, 111 (17.1%) contained only PMNs, and 48 (7.4%) contained a mixture of squamous cells plus PMNs. On 19 (2.9%) of the slides, smears were too thick to be able to determine the cell type, whereas 46 (7.1%) of the slides contained only cell debris and/or mucous, and 51 (7.8%) of the slides had nothing, that is, no cells or debris on them. Of these 511 slides considered to be inadequately prepared, only four (0.78%) were positive for C. trachomatis elementary bodies. Thus there was an approximate tenfold reduction in the percent of positive slides that were not correctly obtained, versus those that were appropriately obtained (p < 0.05). The correlation of numbers of positive specimens to the numbers of columnar/transitional cells present in specimens from females are seen in Table 3. Of the specimen slides, 36 contained only rare to occasional columnar/transitional cells, of which one
TABLE 3
Correlation of the Positive Rate of Chlamydia trachomatis Detection with Numbers of Columnar/Transitional Cells Present on Specimen Slides from Females
Nos. of Columnar/Transitional Cells on Slide No. of Isolates No. (%) Positive Rare to occasional (1-24) Moderate to many (i>25)
36
1 (2.8%)
102
9 (8.8%)
Total
138
10 (7.2%)
(2.8%) was positive; 102 specimen slides contained moderate to many columnar/transitional cells, of which nine (8.8%) were positive. Thus there appears to be a correlation between the numbers of columnar/transitional cells observed and the number of positive chlamydial specimens. In the same study, 171 adolescent and adult male urethral slides were examined. Sampling directions specify inserting and rotating the swab 2-4 cm into the urethra to obtain columnar cells, as the terminal section of the male urethra is lined with squamous epithelium. Patients should not have urinated within 1 hr of sampling. Of these 171 slides, 73 (42.7%) were adequate, containing columnar cells either alone or in combination with other cell types. Of these 73, five (6.8%) were positive. However, 98 (57.3%) of
236
171 were considered inadequately collected, containing no columnar cells and only three (3.1%) of 98 were positive. Of these 98, 17 (9.9%) contained only squamous ceils, 28 (16.4%) contained only PMNs, 12 (7.0%) contained squamous cells and PMNs, 1 (0.6%) was too thick to read the cell type, 12 (7.0%) contained only cell debris and/or mucus, and 28 (16.4%) had no cells on the slide. However, no significant difference between the adequately and inadequately collected groups were observed, due to the fewer numbers of positive specimens occurring in both groups.
DISCUSSION When 649 female genital smear slides were examined by a direct fluorescent antibody technique for adequateness of specimen collection as well as for the presence of C. trachomatis, only 21% were appropriately taken, containing columnar/transitional epithelial cells, either alone or in conjunction with squamous epithelial cells and/or PMNs. Obtaining specimens from anatomic sites in which columnar/ transitional cells are found is important as these are the host cells that the obligate intracellular parasite C. trachomatis infects. In this study, - 3 . 7 times as many specimens were inadequately obtained compared to those adequately obtained (511 vs 138). Only 0.78% of inadequately obtained specimens were positive, versus 7.2% of adequately obtained specimens. Thus, theoretically, 6.4% of specimens or 32 of 511 were being misdiagnosed as negatives that may have been true positives. Because the MicroTrak reagent used contains the counterstain Evan's blue, which stains mammalian cells and debris, the DFA technique has the advantage of determining the quality of the specimens obtained, and is an inherent part of evaluating slides for C. trachomatis. Adequately obtained specimens containing columnar/transitional cells should also be obtained for other C. trachomatis detection methods, that is, EIA and culture. However, provisions for evaluating the quality of the specimens are not an inherent part of these two other techniques. Better specimen sampling may be obtained with cervical cytobrush (Linder et al., 1987), although this technique cannot be used in pregnant women, and may cause transient bleeding in patients with chlamydial cervicitis due to the friability of these cervices (Moscicki et al., 1987). In addition, PMN-containing mucus needs to be wiped away prior to endocervical sampling, even though PMNs are associated with chlamydial infection of the cervix (Kiviat et al., 1985; Moscicki et al., 1987). Inadequately obtained specimens for chlamydia detection appears to be a major problem, at least in
C. Howard et al.
this study. In one other study (Phillips et al., 1987), 45% of the specimens had no columnar cells and 38% of the specimens had some mucous partially obscuring the slide; in the study, the sensitivity of the DFA test compared to culture was 70%, but increased to 92% if five or more columnar cells were observed on a slide and decreased to 40% if fewer than five columnar cells were observed. In two other reports, 12% of the specimens obtained for DFA were reported as inadequately taken, but chlamydia resuits were not reported to the physicians (Lipkin et al., 1986; Schafer et al., 1986). The relatively low percentage of inadequate specimens in these two reports may have been due to the specimens being submitted from a very limited number of sites (1-3 clinics) in which clinicians may have had more thorough training in specimen collection. Slide variation may be greater as the number of collection locations increases, leading to variability in clinician training and collection techniques. The MicroTrak directions suggest reporting "indeterminate" results when inadequate specimens lacking columnar/transitional cells are obtained. However, we do not reject specimens that are inadequate as (a) we found that close to 1% of these specimens were chlamydiae positive and (b) we prefer to report a description of cell types observed along with an explanation that chlamydiae replicate in columnar/transitional epithelial cells. Our reporting on the type and amount of cells, as well as the presence of cell debris, mucus, and/or lack of cells gives more information to the physicians and may alert them as to problems encountered in obtaining specimens for chlamydial detection. Problems with false-negative results due to sampling errors have also been reported for cervical smears taken for the detection of cervical intraepithelial neoplasia. Gay et al. (1985) found a falsenegative rate of 63% due to sampling failure; Duguid (1986) reported that >50% of the smears taken by general practitioners contained
