An Evidence Update
David S. Meyers, M.D.a; Heather Halvorson, M.D., M.P.H.a; and Sara Luckhaupt,
M.D., M.P.H.a
The authors of this article are responsible for its contents.
No statement in this article should be construed as an official position of
the Agency for Healthcare Research and Quality or the U.S. Department
of Health and Human Services.
Address correspondence to: Dr. David Meyers,
Center for Primary Care, Prevention, and Clinical Partnerships, Agency for
Healthcare Research and Quality, 540 Gaither Road, Rockville, MD 20850.
This article first appeared in Ann Intern Med 2007;147:134-41.
Select for copyright and source information.
Contents
Abstract
Introduction
Methods
Results
Contextual Questions
Discussion
References
Notes
Abstract
Background: Chlamydial infection is the most common sexually
transmitted bacterial infection in the United States, with an estimated 3 million
new cases annually. In 2001, the U.S. Preventive Services Task Force (USPSTF)
recommended that clinicians screen all sexually active women at increased risk
for infection for Chlamydia trachomatis.
Purpose: To summarize a systematic evidence review commissioned
by the USPSTF in preparation for an update of its 2001 recommendation.
Data Sources: English-language articles identified in PubMed
between July 2000 and July 2005. Additional articles were identified by bibliographic
reviews and discussions with experts. A total of 452 articles were identified.
Study Selection: Explicit inclusion and exclusion criteria
were used for each of 3 key questions. For studies of screening in nonpregnant
women at increased risk, review was limited to randomized, controlled trials.
For other groups, both randomized, controlled studies and nonrandomized,
prospective, controlled studies were included.
Data Extraction: By using standardized forms, staff of the
Agency for Healthcare Research and Quality abstracted data on study design,
setting, sample, randomization, blinding, results, and harms.
Data Synthesis: Only 1 new poor-quality study met inclusion
criteria. This study of the effectiveness of screening for chlamydial infection
among nonpregnant women at increased risk found that screening was associated
with a lower prevalence of chlamydial infection and fewer reported cases of
pelvic inflammatory disease at 1-year followup.
Limitations: No new evidence was found on screening in pregnant
women, nonpregnant women not at increased risk, or men.
Conclusions: A systematic review found a small amount of
new evidence to inform the USPSTF as it updates its recommendations regarding
screening for chlamydial infection. There are large gaps in the evidence regarding
the ability of screening in men to improve health outcomes in women.
Return to Contents
Introduction
Chlamydia trachomatis is the most common sexually transmitted bacterial
infection in the United States; more than 900 000 cases were reported in 2004 to
the Centers for Disease Control and Prevention (CDC), and an estimated 2 million
new cases go unreported each year.1,2
Although 75% of genital infections in women and 95% in men are asymptomatic,
up to 40% of untreated cases of C. trachomatis infection in women progress
to pelvic inflammatory disease (PID).3-5
It has been estimated that 20% of women with PID become infertile, 18% experience
chronic pelvic pain, and 9% may have a tubal pregnancy.6
Chlamydial infections are also related to adverse pregnancy outcomes, including
miscarriage, premature rupture of membranes, preterm labor, low birth weight,
infant mortality, neonatal chlamydial infection, and postpartum endometritis.3,7,8
In 2001, the U.S. Preventive Services Task Force (USPSTF) commissioned a systematic
review of the evidence regarding the benefits and harms of screening for chlamydial
infection (7 This
review considered screening in nonpregnant women, pregnant
women, and men and rated the overall body of evidence for all 3 groups as fair,
noting that many studies were performed under study conditions rather than
real-world conditions and that most studies did not use large screening samples
with low prevalence rates.7
The USPSTF concluded in 2001 that good evidence supports screening for
chlamydial infection among asymptomatic women at increased risk for infection,
including women at risk because of young age, and strongly recommended that this
group be screened nationwide. It found less evidence regarding screening of
pregnant women and, on the basis of estimates of benefits and harms, recommended
screening only for pregnant women at increased risk. At that time, the USPSTF found
a major gap in the evidence regarding the effectiveness of screening for men and made
no recommendation, concluding that the evidence was insufficient.
This article summarizes a full evidence update conducted for the USPSTF,
which reviewed information published since the USPSTF 2001 review.9 This
update focused on a search for direct evidence on the effect of screening asymptomatic
individuals on health outcomes. The USPSTF reviewed the evidence to update
its recommendations. (A description of how the Task Force develops and communicates
its recommendations also appears in the same issue of the Annals of Internal Medicine.10)
In preparing for the review, the USPSTF identified 3 critical key questions
for which they requested systematic evidence reviews. In addition, they identified
several subsidiary questions for which they requested nonsystematic reviews
to assist them with updating their recommendations and supporting materials.
This article focuses on the results of the systematic review and concludes
with a few highlights from the nonsystematic reviews.
Answers to the following critical questions were pursued by means of a
full-scale literature review.
- Does screening for chlamydial infection in nonpregnant women reduce
adverse health outcomes?
- Does screening for chlamydial infection in pregnant women reduce
adverse health outcomes?
- Does screening for chlamydial infection reduce adverse health outcomes
in men, reduce adverse health outcomes in women, or reduce the incidence
of infection in women?
In this context, health outcomes of interest were defined as follows: PID,
ectopic pregnancy, infertility, and chronic pelvic pain in nonpregnant women;
chorioamnionitis, premature rupture of membranes, preterm labor, preterm delivery,
spontaneous abortion, endometritis, and low birth weight in pregnant women;
and epididymitis, urethritis, prostatitis, chronic prostatitis, reactive arthritis,
and urethral strictures in men.
Return to Contents
Methods
Staff at the Agency for Healthcare Research and Quality conducted a systematic
evidence review for each of the critical key questions.
Data Sources
The search strategy included a review of English-language articles identified
from PubMed® between July 2000 and July 2005. Additional articles were found
through bibliographic reviews and discussion with experts. These searches identified
452 articles.
Study Selection
For key question 1, the review was limited to randomized, controlled trials
of nonpregnant women at increased risk for infection. For nonpregnant women
not at increased risk, the search was expanded to include both randomized,
controlled trials and nonrandomized, prospective, controlled studies. For key
questions 2 and 3 (screening in pregnant women and in men), the reviews were
limited to randomized, controlled trials and nonrandomized, prospective,
controlled studies.
Abstracts were reviewed by 2 staff members. All
abstracts that were clearly within the scope of this review and those with potential
or ambiguous relevance were retained. Eighteen articles were identified as potentially
meeting these broad inclusion and exclusion criteria
(Figure).
Key Summary Points
- There is direct evidence that screening women at increased risk for chlamydial infection improves health outcomes.
- There is no direct evidence that screening women not at increased risk for chlamydial infection improves health outcomes.
- There is fair evidence that screening pregnant women can identify chlamydial infection and that treatment of chlamydial infection improves birth outcomes.
- There is a significant evidence gap regarding the ability of screening in men for chlamydial infection to improve health outcomes in women.
|
Data Extraction and Quality Assessment
Two reviewers independently reviewed the full articles of all identified studies
to determine whether the articles met predetermined inclusion criteria. Additional
reviewers were consulted for consensus building around 2 articles that were
ultimately not included in this review. The 2 principal reviewers independently
abstracted data by using standardized forms from included articles to determine
study quality.
Role of the Funding Source
The work of the USPSTF is supported by the Agency for Healthcare Research
and Quality (AHRQ).
Return to Contents
Results
Key Question 1
Does screening for chlamydial infection in nonpregnant women reduce adverse
health outcomes?
Women at Increased Risk
In 2001, on the basis of a good-quality randomized, controlled trial by Scholes
and colleagues published in 1996,11 the USPSTF
gave an A rating to their recommendation that clinicians routinely screen all
sexually active women at increased risk for chlamydial infection. The study,
conducted in a large managed care organization in Seattle, Washington, concluded
that screening and treating young women at increased risk for chlamydial infection
significantly reduced the incidence of PID after 1 year of followup.11
Only 1 study12
identified in the current systematic review met the inclusion criteria
and addressed the effectiveness of screening for chlamydial infection among
nonpregnant women at increased risk. In a cluster randomized trial, Ostergaard and
colleagues12 found
that a 1-time, home-based screening intervention was associated with a lower
prevalence of chlamydial infection and fewer reported cases of PID at 1-year
followup. As part of a larger study, the team randomly assigned 17 high schools
in a Danish county to 1 of 2 groups.
Students in schools selected to be in the intervention group were offered a
single opportunity for home-based screening for chlamydial infection. Students in
the control schools were given the same educational information and were encouraged to
visit their physician for a free screening. Sexually active girls in both groups
were offered the opportunity to receive followup in 1 year.
Ostergaard and colleagues found that the intervention was associated with
a lower prevalence of chlamydial infection and fewer reported cases of PID
at 1-year followup. Of 443 girls in the intervention group who participated
in followup testing, 13 (2.9%) were found to have chlamydial infection and
9 (2.1%) reported receiving treatment for PID. Of 487 girls in the control
group, 32 (6.6%) were found to have chlamydial infection and 20 (4.2%) reported
receiving treatment for PID. Both of these differences were found to be statistically
significant.
Given the differences in initial screening rates between the groups
(93.4% of the intervention group vs. 7.6% of the control group), the USPSTF
chose to include this study as an example of a trial of screening versus not
screening. The effect of baseline screening in the control group would be expected
to decrease the ability of the intervention to demonstrate a difference between
screening and not screening. This factor gives additional weight to the study
findings. Although universal screening was offered, Ostergaard and colleagues
targeted high school-age female adolescents, who are defined by the USPSTF
as being at increased risk for chlamydial infection because of their age alone.
The overall initial prevalence rate of chlamydial infection among those screened
was 5.0%.
The USPSTF thus considered this to be a trial of screening women
at increased risk and not a trial of the effect of universal screening. The
study was deemed to be of poor quality because of an unaccounted loss of participants
in both groups for followup screening. Whereas 93.4% participants in the intervention
group and 100% of those in the control group agreed to followup screening,
only 51.1% in the intervention group and 58.5% in the control group actually
participated in followup screening. The researchers did not provide sufficient
information to assess the effects that this loss may have had on the results
of the study.
A study by Clark and colleagues.13 was closely
reviewed by the USPSTF. Although this study did not meet the criteria for the
systematic review because it was a nonrandomized trial of screening women at
increased risk, it is presented here as a good-quality study that contributes to
our understanding of screening for chlamydial infection. Clark and colleagues
conducted a nonrandomized cohort study examining hospitalization rates after
screening for chlamydial infection in female military recruits.13
A total of 7053 women were screened and treated
for chlamydial infection over 2 years on arrival at basic training. A group
of 21 021 women who were not screened upon their arrival were followed as a
comparison group. Eighty percent of the women studied were younger than age
25 years, and the overall prevalence rate of chlamydial infection among the
women screened was 9.1%. The average duration of followup for both cohorts
was more than 1.5 years. Results were adjusted for age, race, education, and
entrance aptitude score.
The investigators found a slight decrease in the adjusted
relative risk for hospitalization overall (0.94 [95% CI, 0.90 to 0.99]). The
investigators noted a lower adjusted relative risk for hospitalization for
PID, but this difference was not statistically significant (0.94 [CI, 0.69
to 1.29]). The relative risk for hospitalizations for chlamydia-related
sequelae (PID, infertility, or ectopic pregnancy) was nonstatistically higher
in the screened group than in the unscreened group (1.10 [CI, 0.85 to 1.43]).
The intervention and control groups differed significantly in terms of age,
education, and entrance aptitude scores: The screened group was slightly younger
and had lower educational level and aptitude scores than the control group.
Both groups were about 35% African American. The study examined military hospitalizations
only and did not capture civilian hospital use by either group. In addition,
the investigators could not include outpatient treatment for PID or other sequelae.
Although the research team adjusted for major known demographic confounders,
the trial is a nonrandomized study in groups with significant differences.
The time frame for this study may not have been adequate to detect the full
consequences of the long-term effects of chlamydial infection, but the findings
nonetheless remind us that dramatic benefits from a single screening test may
not be significant or may not be captured in a specific evaluation project.
Women Not at Increased Risk
In 2001, the USPSTF found no direct trials of screening women not at increased
risk for chlamydial infection that reported health outcomes. It found fair
evidence for each link in the analytic framework, noting that there was fair
evidence that screening women not at increased risk would find additional cases
of chlamydial infection. The USPSTF found little evidence of potential harms
of screening. However, given the low prevalence of chlamydial infection in
this population, the USPSTF made no recommendation for or against routinely
screening asymptomatic women not at increased risk for chlamydial infection.
The USPSTF concluded that the potential benefits of screening women not at
increased risk may be small and consequently may not justify the potential
harms of screening. The current systematic review found no new direct trials
of screening for chlamydial infection among women not at increased risk.
Key Question 2
Does screening for chlamydial infection in pregnant women reduce adverse health
outcomes?
In 2001, the USPSTF found fair evidence that screening asymptomatic pregnant
women can detect chlamydial infection and that treatment of chlamydial infection
during pregnancy improves health outcomes for both the mother and infant. The
USPSTF concluded that the potential benefits outweighed the potential harms
of screening pregnant women and recommended that clinicians routinely screen
all asymptomatic pregnant women age 25 years or younger, and other pregnant
women at increased risk, for chlamydial infection. The USPSTF considered the
potential net benefits of screening pregnant women who are not at increased
risk to be small, leading it to make no recommendation for or against screening
pregnant women not at increased risk for chlamydial infection. Evidence reviewed
for this report revealed no new randomized, controlled studies or nonrandomized,
prospective, controlled studies addressing this topic.
Key Question 3
Does screening for chlamydial infection reduce adverse health outcomes in
men, reduce adverse health outcomes in women, or reduce the incidence of infection
in women? In 2001, the USPSTF noted that the benefits to men of treating asymptomatic
infection are small, because longterm sequelae are rare and treatment of symptomatic
infection is effective. However, the USPSTF noted the potential for significant
benefit if screening in men can, in fact, decrease the incidence of infection
in women. Finding no evidence to support this hypothesis, the USPSTF concluded
that the evidence was insufficient to make a recommendation regarding screening
men for chlamydial infection. The current systematic review identified no randomized,
controlled studies or nonrandomized, prospective, controlled studies of screening
men for chlamydial infection and the ability of screening programs to reduce
the incidence of infection among women.
Return to Contents
Contextual Questions
The full evidence update report includes information on the following contextual
questions.
- Has the epidemiology of chlamydial infection in the United States changed
in significant ways since 2001, including in groups at increased risk?
- What are the harms of screening for chlamydial infection?
- Are screening tests for chlamydial infection accurate?
- What is the optimal screening frequency?
- Does chlamydial infection increase the risk for infection with HIV?
- What is the cost-effectiveness of screening for chlamydial infection?
Limited nonsystematic reviews were conducted for the subsidiary questions.
Literature reviews for subsidiary questions included review articles and topic-specific
searches of MEDLINE®. Articles reviewed during the critical key question reviews
were tagged if they addressed a subsidiary key question. Recommendations for
sentinel articles were also sought from content experts. The purpose of these
searches was to provide updated context for recommendations rather than to
serve as evidence for changes in the recommendations. The results of these
searches may be viewed in the full report at http://www.preventiveservices.ahrq.gov.
Selected highlights from the report are included below.
Epidemiology
The review found that the epidemiology of chlamydial infection in the United
States has not changed in significant ways in recent years. Age remains the
strongest predictor of risk in both men and women. Among women, the highest
rates of infection are reported among those 15 to 19 years of age, followed
by those 20 to 24 years of age. Other associated risk factors include both
behavioral and demographic factors, including having multiple sexual partners,
having a new sexual partner or an infected sexual partner, inconsistently using
barrier contraceptives, and having a history of previous or coexistent sexually
transmitted infections (STIs).7 The incidence
of chlamydial infection continues to be higher in many African-American and
Hispanic communities.
Harms of Screening
Several new qualitative studies have examined the adverse effects associated
with a diagnosis of chlamydial infection. A 2003 paper addressed the psychosocial
impact of the diagnosis of chlamydial infection through a qualitative study
using semistructured interviews with 17 Scottish women with recently diagnosed
chlamydial infection. The researchers found that these women perceived that
testing and diagnosis of chlamydial infection were associated with negative
stereotypes, such as contamination and delinquency, and they perceived a social
stigma attached to their diagnosis. The women expressed concern over the meaning
of their diagnosis to their future fertility and had significant anxiety about
the attitudes of their male partners. They also were concerned about notifying
both current and past partners.14
As part of a large trial of screening for chlamydial infection in England, Pimenta
and colleagues15
conducted in-depth interviews with more than 400
women who completed screening. Overall, participants were accepting of the
screening program, and most found screening beneficial. Participants with positive
test results commonly reported "feeling dirty, [feeling] ashamed at passing
on the infection, and [sensing others] suspicion about where the infection
originated." The investigators reported that for some women, this led
to "tension and suspicion within relationships," but no long-term
repercussions within relationships were identified.15
A qualitative study of 12 heterosexual men and 12 heterosexual women with
recently diagnosed chlamydial infection identified significant differences
in their responses to diagnosis. The women reported feeling anxious about their
future reproductive health, feared stigmatization, and blamed themselves for
contracting chlamydial infection. The men generally reported less concern and
were less willing to disclose their condition to sexual partners. Some of the
men, according to the investigators, blamed their partners for their infection
and avoided accepting responsibility themselves. The female participants experienced
blame and denial of the infection on the part of their male partners. The women
also reported concern about potential threats to their relationships. The investigators
concluded that a culture of the "blameless male and stigmatized female" continues
to persist around the issue of STIs. They note that avoidant attitudes and
behaviors among men should be accounted for in STI screening and treatment
programs.16
Accuracy of Screening Tests
The USPSTF 2001 systematic review on chlamydia screening found that nucleic
acid amplification tests (NAATs) had higher sensitivities and specificities
than older antigen detection tests and higher sensitivities than culture. However,
the USPSTF did not at that time provide clinical guidance as to which tests
should be used. In 2002, the CDC recommended that NAATs be used for screening both
women and men.17
It reached this conclusion on the basis of a systematic review that included, in
addition to many of the studies considered by the USPSTF, a large multicenter study
on screening technologies. In 2005, Cook and colleagues 18 presented
a systematic review of noninvasive testing for chlamydial infection in which
overall high sensitivities and specificities were found for polymerase chain
reaction and other NAATs from both genital and urine samples.
Chlamydial Infection and HIV
There is broad consensus in the literature that, as with other inflammatory
STIs, chlamydial infection facilitates the transmission of HIV infection in
both men and women. Sexually transmitted infections increase both the infectivity
of persons with HIV infection and the susceptibility of those with STIs to
HIV infection.19,20
The prevalence of chlamydial infection among men who have sex with men has
not been studied widely outside of HIV care settings. In addition, tests other
than culture, which are not widely available, have not been cleared by the
U.S. Food and Drug Administration for use with rectal or pharyngeal specimens.
Research Gaps and Emerging Issues
The systematic review uncovered no evidence on the effect of screening men
to reduce the prevalence of infection in women, but our understanding of the
ability of screening and treatment of a high-risk group to reduce community
prevalence of an STI may be informed by a study in a South African mining community.21 A research
team conducted 2 cross-sectional samples of male miners in
a mining town, where more than 90% of the miners live in single-sex hostels
near the mines. The intervention consisted of the establishment of a mobile
STI clinic for female sex workers and other local women with multiple sex partners.
Women enrolled at the clinic were encouraged to return for monthly visits and
were treated presumptively with azithromycin at each visit. For women who returned
to the clinic, rates of gonorrhea and chlamydial infection decreased with each
visit. At the end of the intervention, the rate of gonorrhea among miners decreased
from 10.9% to 6.2% (P < 0.001) and that of chlamydial infection decreased
from 6.6% to 3.5% (P = 0.005). Community records also showed that the miners
significantly decreased their number of visits to local medical facilities
for STI care.
Similar results were found in a study presented at a meeting of the International
Society for Sexually Transmitted Disease Research in Ottawa, Ontario, Canada,
in 2003.22
Researchers reported that chlamydial infection among women decreased
by 50% at a health center serving a population in which men who had been screened
and treated for chlamydial infection while incarcerated resided. If this type
of research continues, future recommendations regarding screening for chlamydial
infection in men may have an evidence base.
Another area lacking research is that of the potential harms associated with
screening for chlamydial infection. In 2005, the CDC began a study to examine
the psychosocial effect of a positive diagnosis of chlamydial infection (Walsh
C. Personal communication, 2005).
Return to Contents
Discussion
The evidence base supporting screening for chlamydial infection has not expanded
greatly since the USPSTF made its first recommendation regarding screening
in 2001. Table 1 summarizes
the combined evidence from the 2001 USPSTF systematic
review and the current update. The evidence is strongest regarding screening
in nonpregnant women at increased risk for infection. The effectiveness of
screening men for chlamydial infection to reduce the incidence of infection
and its sequelae in women remains a large gap in our current understanding
of screening for chlamydial infection.
There is an emerging literature about the potential harms of screening in
qualitative studies examining the implications of a positive chlamydia test
result. These studies find significant anxiety and concern for the future of
intimate relationships after a positive diagnosis of chlamydial infection.
This review identified no studies investigating screening intervals for chlamydial
infection. Therefore, there is no direct evidence to guide decisions about
repeated screening in those who have already been screened, regardless of the
outcome. Studies in this area will make a major contribution to improving the
evidence base for screening programs.
Table 2 shows hypothetical
outcomes of a screening program for chlamydial infection that are based on assumptions
from recent studies. It shows the results that might be expected from screening
programs among nonpregnant women with underlying prevalence rates of chlamydial
infection, including 0.1% (women not at increased risk), 1.0% (women age 25 to 29
years who are not otherwise at increased risk), 5.0% (sexually active female
adolescents), and 10.0% (women at significant increased risk, such as military
recruits). The positive predictive value increases from 8% to 91%, whereas the
number needed to screen to prevent 1 case of PID decreases from 3800 to 38 between
the groups with the lowest and highest prevalence rates.
Despite this evidence, many, if not most, women are not receiving this recommended
preventive health service. The National Committee for Quality Assurance noted
in their 2005 report that although there has been modest and steady improvement
in the rates of screening women for chlamydial infection, in 2004 the screening
rate was 32.6% among women age 16 to 20 years in commercial health plans and
45.9% among women enrolled in Medicaid managed care plans.23 For women age
21 to 25 years, the rates were 31.7% in commercial plans and 49.0% in Medicaid
plans.23 In
an editorial accompanying an article by Hu and colleagues.24
on the cost-effectiveness of screening for chlamydial infection, Stamm acknowledges
that "actual practice falls far short of recommended practice."25
The author suggests that before screening programs are expanded to men or to
women with lower prevalence rates of chlamydial infection, "sexually
active women 15 to 24 years of age... should have the highest priority for
screening."25
Return to Contents
Proceed to Next Section