|
.gif)
|
|
Cervical findings
suggesting chlamydial or gonococcal infection have higher positive
predictive values in women less than 25 years old.
Predicting chlamydial and gonococcal
cervical infection: implications for management of cervicitis.
Marrazzo JM, Handsfield HH, Whittington WLH.
Obstet Gynecol. 2002;100:579-84
Summary:
Question
How well do age, clinical signs of cervical infection, and inflammation on
endocervical Gram stain predict infection with C. trachomatis and N.
gonorrhoeae among women diagnosed by culture?
Design
This paper describes a retrospective study of women attending an STD
clinic. All women had endocervical Gram stains and cultures for C.
trachomatis and N. gonorrhoeae performed. Predictive values of
age, cervical signs, and inflammation on Gram stain for the presence of C.
trachomatis or N. gonorrhoeae by culture were calculated.
Participants
Records for 6230 new problem visits to Public Health, Seattle and King
County STD Clinic by women who underwent pelvic examination, did not have
signs of pelvic inflammatory disease, and were tested for C.
trachomatis and N. gonorrhoeae were included in the analysis.
Of these 6230 women, 430 (6.9%) had C. trachomatis and 133 (2.1%)
had N. gonorrhoeae detected by culture.
Description of Tests and Diagnostic
Standard
Clinicians performed Gram stain smears of endocervical secretions and
quantified polymorphonuclear cells per 1000X field (PMN/HPF) by averaging
the PMN number in each of at least 10 fields. Inflammation was defined as
>30 PMN/HPF on Gram stain of endocervical secretions. Mucopurulent
cervicitis was defined as the presence of either clinical sign (mucopus or
easily induced bleeding at the cervix) and of cervical inflammation on
Gram stain. C. trachomatis and N gonorrhoeae were identified
by endocervical cultures according to standard methods.
Main Outcome Measures
The positive predictive values (PPV) and confidence intervals (CI) by age
of the presence of either clinical sign, of cervical inflammation, and of
mucopurulent cervicitis for cervical chlamydial or gonococcal infection as
determined by culture were calculated.
Main Results
The PPVs of features associated with cervical chlamydial or gonococcal
infection by age group are presented in the table. The PPV of all cervical
findings and of Gram stain smear of endocervical secretions for cervical
infection were significantly higher in women younger than 25 years than in
women 25 years or older.
Inflammation on endocervical Gram stain
was relatively insensitive in detecting infection in women without either
cervical sign (sensitivity 26.5%, seen in 79 of 308 infections), and,
among all 520 infected women, provided the only indication of chlamydial
or gonococcal infection in 79 (15.2%).
| Positive predictive
values of features associated with cervical chlamydial or gonococcal
infection by age group |
| Feature |
Num≤ber
infected/Number with characteristic, PPV,
and (95% CI) |
| ≤19
y |
20-24 y |
25-29 y |
30-44 y |
≥45
y |
Total |
| Bleeding
or mucopus |
51/158
39%
(31, 47) |
80/294
27%
(22, 33) |
26/232
11%
(7.6, 16) |
28/293
9.6%
(6.6, 14) |
2/51
3.9%
(0.7, 15) |
227/1028
20%
(17, 22) |
| ≥30
PMN/HPF alone |
63/175
36%
(29, 44) |
80/307
26%
(21, 31) |
36/210
17%
(12, 23) |
28/227
12%
(8.5, 18) |
2/40
5.0%
(0.9, 18) |
209/959
22%
(19, 25) |
Muco-purulent
cervicitis |
33/81
41%
(30, 52) |
57/151
38%
(30, 46) |
20/100
20%
(13, 29) |
15/102
15%
(8.7, 23) |
2/21
9.5%
(1.7, 32) |
127/455
28%
(24, 32) |
Authors' Conclusions
The data support empiric treatment along
with diagnostic testing for chlamydial or gonococcal cervical infections
in women younger than 25 years in whom induced bleeding or cervical
mucopus are detected, and only diagnostic testing for women 25 years and
older. The performance of Gram stain smear of endocervical secretions in
predicting chlamydial or gonococcal cervical infection does not justify
its routine use.
Source of funding: In part by National
Institutes of Health grant R29-AI41153-04
For correspondence: Jeanne M. Marrazzo,
Harborview Medical Center, Box #359931, 325 Ninth Avenue, Seattle, WA
98104. E-mail address: jmm2@u.washington.edu
|
