Screening for Gonorrhea

Update of the Evidence

---

Nancy Glass, Ph.D., M.P.H., R.N.; Heidi D. Nelson, M.D., M.P.H.; Kim Villemyer, B.A.

Address correspondence to Nancy Glass, Oregon Evidence-based Practice Center, 3181 SW Sam Jackson Park Road, Portland, OR 97239.

Select for copyright, source, and reprint information. The USPSTF recommendation based on this update is available online.

---

This update of the evidence report examines evidence of the effectiveness of screening for gonorrhea in asymptomatic sexually active men and women including adolescents and pregnant women. It reviews studies of screening strategies, individual and population-level risk factors, characteristics and accuracy of tests used for screening, harms of preventive medication treatment for newborns, and cost effectiveness of universal and targeted screening strategies. This review is an update and includes only studies published since the last recommendations of the U.S. Preventive Services Task Force (USPSTF) were released in 1996.¹

---

Contents

Structured Abstract
Introduction
Methods
Results
   Asymptomatic Men and Women Including Adolescents
   Pregnant Women
   Newborn Chemoprophylaxis
Discussion
   Summary of Evidence
   Outcomes Table
   Limitations of the Evidence
   Future Research
Acknowledgments
References
Notes

---

Structured Abstract

Context: Gonorrhea, caused by Neisseria gonorrhoeae, is second only to chlamydia in the number of sexually transmitted disease (STD) cases reported to the U.S. Centers for Disease Control and Prevention (CDC) annually. The prevalence of gonococcal infection varies depending on the population and setting, and the highest rates are reported among young males and females, African Americans, and men who have sex with men.

Objective: To examine the effectiveness of screening for gonorrhea in asymptomatic sexually active men and women including adolescents and pregnant women. This systematic review focuses on studies of screening strategies and their outcomes, individual and population-level risk factors, characteristics and accuracy of tests used for screening, adverse effects of preventive medication treatment for newborns, and cost effectiveness of universal and targeted screening strategies.

Data Sources: MEDLINE® database (January 1966-July 2004), reviews, Web sites, and experts.

Study Selection: English-language abstracts were dual-reviewed for eligibility and only studies published in 1996 or later were included. Papers were selected for full review if they addressed key questions in the target populations.

Data Extraction: Relevant data were extracted from each study and summarized in evidence tables. Predefined criteria from the USPSTF were used to assess the internal validity of included studies.

Results: No new evidence was identified that evaluated the effectiveness of population screening to reduce transmission and improve health outcomes. Individual-level risk factors include young age (<25 years), African American race, multiple sex partners or an infected sex partner, inconsistent use of barrier contraceptives, previous or coexistent STDs, douching, use of drugs, and history of incarceration. Contextual risk factors include sexual networks, sexual mixing within a community or neighborhood with high prevalence of STDs, and residence in a community with limited social capital or markers of physical deterioration. New testing technologies, such as nucleic acid amplification tests and nucleic acid hybridization tests, demonstrate high sensitivity and specificity, although studies are methodologically limited. Sensitivity is lower using urine specimens for some tests, and may vary by symptom status.

Conclusions: Recent evidence only addresses key questions about risk factors and new tests. These studies are limited by descriptive, cross-sectional designs focusing on highly prevalent communities and settings, such as inner city STD clinics, that may not generalize to primary care.

Return to Contents

Introduction

Burden of Condition and Epidemiology

Gonorrhea, caused by Neisseria gonorrhoeae, was second only to urogenital Chlamydia trachomatis infection in the number of sexually transmitted disease (STD) cases reported to the U.S. Centers for Disease Control and Prevention (CDC) in 2003.² The most recent data indicate that up to 335,104 new gonococcal infections occurred in 2003.² The overall prevalence of gonococcal infection in a national representative sample of over 14,000 young adults aged 18 to 26 years was 0.43 percent (95% confidence interval [CI], 0.29%-0.63%),³ although prevalence varies from negligible to 41 percent depending on the population and setting.^3-15 Prevalence rates in the U.S. are currently highest among populations of African Americans and men who have sex with men (MSM).^2,3

There are several similarities between gonococcal and chlamydial infections. The age distribution of individuals with gonorrhea is similar to that for chlamydia with the highest rates reported among female adolescents 15 to 19 years old, and adult males and females 20 to 24 years old.^3,16,17 Expert reviewers for this report noted that the peak age for gonorrhea is about five years older than for chlamydial inventions.

In women and men, uncomplicated gonorrhea is usually confined to the mucosa of the cervix, urethra, rectum, and throat and is often asymptomatic.² Pharyngeal and rectal infections are typically asymptomatic. In women, gonorrhea is a major cause of cervicitis, and its complications include pelvic inflammatory disease (PID), ectopic pregnancy, infertility, and chronic pelvic pain.¹⁶ Among young women, infections often do not produce recognizable symptoms until complications such as PID have occurred. Infections are also related to adverse pregnancy outcomes such as chorioamnionitis, premature rupture of membranes, and preterm labor.¹⁶ Perinatal transmission to infants can cause severe conjunctivitis resulting in blindness if untreated, and, rarely, sepsis with associated meningitis, endocarditis, or arthritis.¹⁶

In men, gonorrhea can result in symptomatic urethritis, epididymitis, and prostatitis.16 Most reports indicate that the majority of genital infections among men are symptomatic and lead to treatment early enough to prevent serious complications, but not transmission to others.² However, one study reported that only 40 percent of men who screened positive for gonococcal infection were symptomatic and another study concluded that less than 5 percent of men with urethral gonorrhea reported dysuria and none reported penile discharge during the 24 hours prior to screening.^3,18

Rarely, local gonococcal infection disseminates to cause an acute dermatitis tenosynovitis syndrome, which can be complicated by arthritis, meningitis, or endocarditis.¹⁶ Gonococcal infection may increase susceptibility to and transmission of human immunodeficiency virus (HIV) in both men and women.¹⁹

Health Care Interventions

Several tests for infection with N. gonorrhoeae are currently available (Table 1).^2,17,20 Culture analysis using swab specimens was traditionally considered the diagnostic gold standard. The advantages of culture include low cost, use with different specimen types (endocervix, urethra, pharynx, or rectum), and the ability to retain the isolate for additional testing, such as for antibiotic resistance. Culture technology poses methodological shortcomings, however, including variation in sensitivity and specificity, need for careful handling to maintain viable organisms, a wait of 2 to 3 days for presumptive results, and the need for invasive sampling.^17,21

Non-culture tests using swab specimens were developed to improve upon some of the limitations of culture. These tests initially included antigen detection tests such as enzyme immunoassay (EIA). The performance and cost associated with EIA have not made them competitive with culture analysis for detecting N. gonorrhoeae.¹⁷ Newer technologies are based on amplified DNA or RNA assays (nucleic acid amplification tests [NAATs]). The amplification methods and the target nucleic acid sequences differ by manufacturer and include polymerase chain reaction (PCR), strand displacement amplification (SDA), and transcription-mediated amplification (TMA).

Non-culture tests offer advantages of improved sensitivity, wider availability, less stringent handling requirements that are not dependent upon living organisms, and timely results.¹⁷ NAATs can be used with either urine or swab specimens for men and women providing a more acceptable method of noninvasive specimen collection, and allowing for screening in non-traditional settings (e.g., school-based clinics, job training programs, substance abuse treatment programs). These tests can detect C. trachomatis or N. gonorrhoeae in a single specimen.

A disadvantage of some NAATs, specifically PCR platforms, is that specimens can contain amplification inhibitors that result in false negative results. NAATs require a high level of technical laboratory expertise to perform the test. Also, non-culture techniques cannot be used for antibiotic resistance testing. The sensitivity of NAATs can vary by specimen type with particularly low sensitivity for PCR using urine samples from women.¹⁷ Some experts recommend confirming non culture test results because of potential false positive results, particularly among asymptomatic individuals.^22-24

Nucleic acid hybridization (nucleic acid probe) tests are also available commercially, and both of the U.S. Food and Drug Administration (FDA) approved tests, PACE® 2 and the Hybrid Capture II®, can detect C. trachomatis or N. gonorrhoeae in a single specimen.¹⁷ The advantage of the nucleic acid hybridization test is the ability to store and transport specimens for up to seven days without refrigeration before testing by the laboratory. These tests also require a high level of technical laboratory expertise to perform the test, and have lower sensitivity than NAATs.¹⁷

All currently available tests require sending the specimen to a laboratory. There are few options for point of care testing for N. gonorrhoeae. Gram stain is most reliable for the presumptive identification in urethral exudates for men, but is not recommended for women because of low sensitivity and variable specificity.¹⁷ However, its application to screening is limited because it requires use of intraurethral swab specimens if no discharge is present.¹⁷

Although there are a variety of screening tests, the performance of routine STD screening in medical practice is low. A random sample of 7,300 physicians in five specialties (obstetrics/gynecology, internal medicine, general practice or family medicine, emergency medicine, and pediatrics) indicated that fewer than one-third routinely screened for STDs.²⁵ Surveyed physicians reported screening 13 percent of men, 30 percent of non-pregnant women, and 31 percent of pregnant women for gonorrhea in their practices.²⁵

Most screening programs target young women in STD or family planning clinics because of the relatively high prevalence rates among patients in these settings, and to take advantage of the opportunity to obtain diagnostic tests in the context of other services.²⁶ Young men have been much more difficult to screen and study. Screening programs have been implemented in emergency departments, school-based clinics, juvenile detention centers and jails, and job training programs.^18,27-38 Community-based gonorrhea screening using self-collected mailed specimens has been studied for feasibility and acceptability.^39,40

The most recent treatment recommendations by the CDC were published in 2002.⁴¹ Approved treatments for uncomplicated gonococcal infections of the cervix, urethra, and rectum include one of the following antibiotic regimens:

• Cefixime 400 mg orally in a single dose.
• Ceftriaxone 125 mg IM in a single dose.
• Ciprofloxacin 500 mg orally in a single dose.
• Ofloxacin 400 mg orally in a single dose.
• Levofloxacin 250 mg orally in a single dose.

Also, if co-existing genital chlamydia infection is not ruled out, the CDC recommends presumptive treatment with azithromycin 1 g orally in a single dose or doxycycline 100 mg orally twice daily for 7 days. Gonococcal infections of the pharynx are more difficult to eradicate than infections at urogenital and anorectal sites. Few antimicrobial regimens can reliably cure >90 percent of infections. The CDC recommended antibiotics for pharyngeal gonococcal infections include ceftriaxone 125 mg IM in a single dose or ciprofloxacin 500 mg orally in a single dose. Also, if co-existing genital chlamydia infection is not ruled out, the CDC recommends presumptive treatment with azithromycin 1 g orally in a single dose or doxycycline 100 mg orally twice daily for 7 days. To prevent gonococcal ophthalmia neonatorum, a prophylactic agent should be instilled into the eyes of all newborn infants. This procedure is required by law in most states. All of the recommended prophylactic regimens prevent gonococcal ophthalmia and include a single application of silver nitrate (1 percent) aqueous solution, erythromycin (0.5 percent) ophthalmic ointment, or tetracycline ophthalmic ointment (1 percent).⁴¹

Experts caution that fluoroquinones are not recommended in young adolescents, and fluoroquinone resistant strains of gonorrhea are emerging. The CDC recommends that fluoroquinones not be used in MSM and in patients who acquired their infections in California, Hawaii, Asia, or other areas with increased resistance to fluoroquinones.^42,43

Prior Recommendations of the USPSTF

In 1996, the U.S. Preventive Services Task Force (USPSTF) recommended that clinicians should routinely screen for gonorrhea in asymptomatic women at high risk for infection (e.g. commercial sex workers, those with a history of repeated episodes of gonorrhea, and young women under the age 25 with two or more sexual partners in the past year) ("B" recommendation).¹

There was insufficient evidence to recommend for or against screening high-risk men (e.g., young, sexually active) for gonorrhea, and routine screening of men or women was not recommended in the general population of low-risk adults ("D" recommendation).

For pregnant women, the USPSTF recommended screening at the first prenatal visit for those who fall into one of the high-risk categories and an additional test in the third trimester for those at continued risk for acquiring gonorrhea ("B" recommendation). There was insufficient evidence to recommend for or against universal screening of pregnant women ("C" recommendation).

In the case of preventive medication against transmission in newborns, the USPSTF recommended that erythromycin 0.5 percent ophthalmic ointment, tetracycline 1 percent ophthalmic ointment, or 1 percent silver nitrate solution should be applied topically to the eyes of all newborns as soon as possible after birth and no later than 1 hour after birth ("A" recommendation).

Analytic Frameworks and Key Questions

The analytic frameworks in Figures 1-3 indicate the strategy used to guide the literature search for evidence of the effectiveness of screening for gonorrhea in asymptomatic sexually active men and women including adolescents and pregnant women. The accompanying key questions correspond to selected numbered arrows in the three analytic frameworks. Key questions were identified by the USPSTF as areas with unresolved issues relevant to clinical practice with potentially new studies since the last USPSTF recommendations were published in 1996. These include:

Key Questions: Asymptomatic men and women including adolescents

1A. Does screening women reduce complications and transmission of disease?
1B. Does screening men reduce complications and transmission of disease?
2A. What individual-level risk factors identify groups at higher risk for gonococcal infection?
2B. What population-level characteristics identify groups at higher risk for gonococcal infection?
2C. What individual-level risk factors identify groups at higher risk for gonococcal infection when used in conjunction with population-level or provider-level characteristics?
2D. What are the screening tests and their performance characteristics?
2E. What is the yield of screening in different risk populations?
2F. Does performance of screening tests vary by specimen type?
2G. What is the role of screening for gonococcal infection among men who have sex with men (MSM)?
3A. What is the evidence on cost effectiveness for universal vs. targeted strategies?
3B. Are dual chlamydia-gonorrhea screening tests cost-effective?

Key Questions: Pregnant women

1A. Does screening reduce adverse maternal/pregnancy outcomes (septic abortion, stillbirth, preterm delivery/low birth weight)?
1B. Does screening reduce adverse neonatal outcomes (gonococcal conjunctivitis, blindness)?
2A. Does screening reduce maternal complications (chorioamnionitis, premature rupture of membranes, preterm labor)?
2B. Does screening reduce transmission to the newborn?
3. What is the evidence on cost effectiveness for universal vs. targeted strategies?

Key Questions: Newborn preventive medication

1. What are the adverse effects of treatment?

Return to Contents

Methods

Literature Search Strategy

The topic of gonorrhea was searched in the MEDLINE® database (January 1966 through July 2004) by a research librarian. A total of nine searches were performed on prevalence, screening programs, risk factors, screening tests and test performance, and cost. Searches specifically related to pregnancy included maternal and neonatal complications and outcomes. A specific search on neonatal preventive medication was also performed. Detailed electronic search strategies are presented in Appendix 1. Periodic hand searching of relevant medical journals and reference lists, and suggestions from experts supplemented the electronic searches. Relevant systematic reviews, policy statements, and other papers with contextual value were also obtained.

Inclusion and Exclusion Criteria

English-language abstracts were dual-reviewed for eligibility. Only studies published in 1996 or later were included in this update. Papers were selected for full review if the abstracts were about:

• Screening strategies in the target populations.
• Individual and population-level risk factors.
• Characteristics and accuracy of tests used for screening.
• Adverse effects of preventive medication treatment for newborns.
• Evidence on cost effectiveness for universal and targeted screening strategies.

Studies were included if they were conducted in the U.S., Australia, Canada, and Western Europe because of similar epidemiology and management of gonorrhea in these countries. Studies of non-human subjects and those without original data were excluded. Foreign language papers were considered if they were randomized controlled trials related to a key question and the abstract was in English.

Studies of screening strategies and programs were included if they met additional criteria. Screening is defined as testing in asymptomatic persons, and "case finding" in those found to have another sexually transmitted infection. Universal screening means testing everyone regardless of symptoms or risk factors; targeted screening indicates that only those who meet specific criteria are tested. Studies about screening programs were included if they described the study population (number screened, sex, age range, setting, presence of symptoms, and other available socio-demographic factors), features of the screening program (duration, type of testing, followup), and outcome measures.

Studies of risk factors for gonococcal infection were included if they reported the number screened, sex, age, setting, reason for visit, screening criteria (universal vs. targeted), type of gonococcal test, other forms of data collection (e.g., questionnaire), and prevalence rates of the tested populations. Results included odds ratios for gonococcal infection from univariate or multivariate regression analysis and significance levels for comparisons between infected and non-infected women and/or men. Risk factors that were not significantly related to gonococcal infection were noted when reported.

This review focused on the new nucleic acid amplification tests obtained by both swab and urine specimens published since 1996. Studies of test performance were included in the summary table only if they met quality criteria at the fair or good-quality level including:

• The test was appropriately performed in a standardized manner.
• The gold standard was appropriately used.
• The study population was adequately described.
• Data were sufficient to determine the sensitivity and specificity of tests.

Outcome measures included sensitivity, specificity, positive predictive value, and negative predictive value of tests evaluated.

Data Extraction and Synthesis

Relevant data were extracted from each study and summarized in evidence tables. In general, these include descriptions of the study population and setting, characteristics of the screening program or test, and outcomes. Studies of risk factors reported associations between infections and risk factors. Predefined criteria from the USPSTF were used to assess the internal validity of included systematic reviews, randomized controlled trials, and observational studies (Appendix 2).⁴⁴ Studies were also considered for applicability to the population that would be identified by screening.

Size of Literature Reviewed

Investigators reviewed 1576 abstracts identified by the searches (Appendix 3). From the searches, 310 full-text articles were reviewed. An additional 12 non-duplicate articles identified from reference lists and experts were also reviewed. The draft report was reviewed by task force members, and content and methodology experts (Appendix 4).

Return to Contents

Results

Asymptomatic Men and Women Including Adolescents

Key Question 1A. Does screening women reduce complications and transmission of disease?

No studies meeting inclusion criteria addressed this question.

Key Question 1B. Does screening men reduce complications and transmission of disease?

No studies meeting inclusion criteria addressed this question.

Key Question 2A. What individual-level risk factors identify groups at higher risk for gonococcal infection?

Fifteen studies published since 1996 describe individual-level risk factors for gonococcal infection among men tested in the military, and heterosexual women and men tested in community, primary care, family planning, and STD clinics (Evidence Table 1).^{5,11,14,15,18,28-31,35-37,45-47} Studies specifically about men who have sex with men are presented under Key Question 2G. Studies of risk factors are descriptive and focus on urban high prevalence populations in community and clinical settings. Studies do not specifically define risk assessment criteria appropriate for screening most heterosexual men and women presenting to primary care settings. No studies prospectively test risk criteria in an asymptomatic screening population to determine its accuracy. The feasibility of using behavioral risk factors for screening is compromised by the difficulty of obtaining this information and its questionable reliability compared with demographic data.

Table 2 summarizes significant findings from the included studies. Ten of 11 studies, representing a wide range of settings and prevalence rates (0.1% to 100%), reported young age as an important predictor.^{5,14,29-31,35-37,47} Age was usually expressed as under 21 or under 25 years old. Five of 6 studies for which race was an analyzed factor reported African American and other non-white race as significantly associated with gonorrhea.^{5,14,30,36,46}

Frequently cited behavioral risk factors have included multiple partners,^{5,11,14,27,30,31,48} partner with symptoms of STD,³¹ inconsistent or no use of barrier contraception,^14,28,,31 drug use,^5,14,28,36 and incarceration.¹⁴ Personal history of PID,^5,27 STD,^{5,18,30,45,46} douching,³¹ and oral contraceptive use²⁹ were also noted as risk factors in some studies. The variation in methods, definitions, and individual-level risk factors assessed across studies do not allow further synthesis of the results. Physical findings on examination such as discharge and co-existent chlamydial infection were also predictive of gonorrhea infection.^{5,18,30,45,46} These factors, however, would necessitate gonorrhea testing for reasons other than screening and would not be helpful in forming a targeted screening strategy.

Specific studies of individual risk factors. A cross-sectional survey examined characteristics of urethral gonococcal infection in male army recruits (n=2,245).¹⁸ The mean age of subjects was 20.6 years (range 17-35 years), 89 percent of participants were under 25 years old, and 60 percent were white. A majority (87 percent) reported ever having vaginal sex, 33 percent reported having sex with more than one partner in past 90 days, and 34 percent reported having sex with a new partner in past 90 days. Twenty percent reported using condoms every time they had sex, and 2.4 percent reported a previous diagnosis of gonorrhea. The overall prevalence of gonorrhea was 0.6 percent, and 7.5 percent of participants with gonococcal infection were co-infected with chlamydia. Of those testing positive for gonorrhea, 40 percent reported having symptoms of any kind. Of these, 60 percent were co-infected with chlamydia. Young age (<25 years) was not a significant predictor of gonorrhea, however, this finding may be the result of the limited age range of the male participants.

A case-control study examined risk factors for male acquisition of gonorrhea in 214 men age 15 to 29 years old seen in STD clinics in Newark, New Jersey.¹⁴ Men with culture confirmed gonorrhea (cases) were compared with controls that had no STDs. A previous diagnosis of gonorrhea was reported by 41 percent of cases and 29 percent of controls, and a history of another STD was reported by 17 percent of cases and 25 percent of controls. Cases were more likely than controls to be African American (odds ratio [OR]=4.2; 95% CI, 1.5-11.5), younger (OR=2.6; 95% CI, 1.2-5.4), or to ever have spent a night in jail (OR=2.3; 95% CI, 1.4-3.9). There were no differences in the percentages of those who finished 12 years of school or were employed. Compared with controls, cases reported a least one casual sex partner within the preceding month (OR=3.2; 95% CI, 1.8-5.7), sex after using marijuana during the preceding month (OR=2.4; 95% CI, 1.1-11.2), a history of incarceration (OR=2.1; 95% CI, 1.2-3.7), and age 15 to 19 years (OR=2.1; 95% CI, 1.0-4.2). Inconsistent condom use was highly prevalent for both case (63 percent) and control (50 percent) groups.

A prospective study of adolescent and adult women (n=477) identified behaviors associated with gonorrhea and chlamydia infections following a behavioral risk reduction intervention for Mexican American and African American women.³¹ The majority (70 percent) of the urban sample was under 25 years of age (age range 14-45 years), and all had low income and limited education. The intervention, Project SAFE (Sexual Awareness for Everyone), focused on five modifiable sexual risk behaviors (e.g., sex with untreated partner, not mutually monogamous, unsafe sex, rapid partner turnover, douches after sex). Infection rates were 18 percent for the intervention group vs. 26 percent for the control group at the end of the 12-month study, and the regression model demonstrated that behaviors correctly predicted infection rates in 75 percent of participants. Unprotected sex with an untreated/incompletely treated partner had the strongest association with infection (cumulative adjusted OR=5.6; 95% CI, 3.0-10.5). Unsafe sex (e.g., no or inconsistent condom use) (OR=1.9; 95% CI, 1.1-3.3), and rapid partner turnover were significantly associated with gonococcal infection (OR=2.7; 95% CI, 1.6-4.8).

Two prospective studies and a case-control study were conducted in urban clinic settings to examine the risk for recurrent gonococcal infection in sexually active females and males.^45,46,49 A current diagnosis or recent history of another STD were significant predictors in all three studies, and unsafe sex behaviors had little impact on subsequent risk for gonococcal infection in two of the studies.^45,46

Key Question 2B. What population-level characteristics identify groups at higher risk for gonococcal infection?

Four studies of population-level characteristics and their associations with gonococcal infection met inclusion criteria (Evidence Table 2).^50-53

In a study using data from all U.S. states, associations between social capital, poverty, income inequality, and four infectious diseases, including gonorrhea, were examined.⁵¹ Predictor variables included:

1. Social capital, defined by Putnam's public use dataset including 14 variables that span the domains of community organizational life, involvement in public affairs, volunteerism, informal sociability and social trust.
2. Poverty, defined by the percentage of the state population living below the poverty line.
3. Income inequality, measured as the ratio of mean income for the top earning one-fifth of families to the bottom one-fifth.

The outcome variables were defined by the 1999 federal surveillance of STDs. Low social capital was significantly correlated to all STDs studied, including gonorrhea (p<0.01).51 High poverty was significantly correlated with chlamydia, and income inequality was significantly correlated with chlamydia and AIDS case rates but not gonorrhea.

A retrospective geographic/regional analysis in San Francisco included patients 14 to 35 years old with initial infections with gonorrhea (n=12,506) and chlamydia (n=9,461), and investigated whether core groups of transmitters existed.⁵⁰ Over 5 years, 8,613 cases of recurrent gonorrhea occurred among males (17 percent) and 3,893 among females (19 percent). Recurrences were more likely in geographically defined populations, independent of race and ethnicity, suggesting core groups of transmitters. These cores have been furthered studied using a geographic information system (GIS) linked to disease surveillance, providing additional information on the geographic epidemiology of gonorrhea and other STDs.⁵³

In a cross-sectional study in New Orleans, LA, researchers performed a regression analysis indicating that traditional variables associated with gonorrhea risk such as poverty, race and unemployment are not as predictive of gonorrhea rates as markers of neighborhood deterioration (e.g., broken windows) (p=0.005).⁵²

Key Question 2C. What individual-level risk factors identify groups at higher risk for gonococcal infection when used in conjunction with population-level or provider-level characteristics?

No studies meeting inclusion criteria addressed this question.

Key Question 2D. What are the screening tests and their performance characteristics?

A total of 25 studies published since 1996 addressing one or more of the key questions about screening tests and their performance characteristics were identified by the literature search (Evidence Table 3).^54-78

Of these, three were rated good quality,^56,76,78 13 fair,^{55,57,60,61,63-65,67,68,70,72,75,77} and nine poor.^{54,58,59,62,66,69,71,73,74} Most studies rated poor quality were limited by inappropriate use, or lack, of a confirmatory or discrepant test.

Most studies were conducted in populations that differed from the target population for this review, specifically prisoners, patients in STD and family planning clinics, contacts of known cases, and other high-risk individuals. Most studies included both symptomatic and asymptomatic individuals and few studies reported results by symptom status. No study specifically focused on adolescents, and the one study that included women age 15 to 44 years did not report results by age.⁶⁹ Sensitivity and specificity of tests are likely to be lower than reported in studies when generalized to clinical practice because of less rigorous specimen collection, increased transport time with deterioration of samples, and variability in the quality of the laboratories performing the tests. Most studies had too few positive samples to be considered statistically valid, and none of the performance parameters were provided with confidence intervals to express the level of uncertainty with the estimates. Confirmatory or discrepant testing was inconsistent across studies. Considering these limitations, studies are too heterogeneous for statistical meta-analysis and are presented descriptively in this report.

Five fair or good quality studies reported sensitivity and specificity of cervical or urethral culture specimens.^{57,65,70,75,77} Specificity was considered 100 percent because most studies defined culture as the gold standard, and sensitivity varied from 65.2 percent to 92.6 percent.^{57,65,70,75,77} Results for cervical, urethral, and urine specimens using NAATs included: PCR sensitivity 42.3 percent to 100 percent, specificity 95.9 percent to 100 percent, five studies.^{57,60,70,72,76} SDA sensitivity 83.7 percent to 100 percent, specificity 94.7 percent to 100 percent, three studies^56,65,72 and TMA sensitivity, 87.5 percent to 100 percent, specificity 98.1 percent to 99.6 percent, two studies.^77,78 Two studies of nucleic acid hybridization tests (nucleic acid probe) reported sensitivities ranging from 92.2 percent to 92.6 percent, and specificities from 98.5 percent to 99.8 percent.^64,75

Key Questions 2E and 2F. What is the yield of screening in different risk populations? Does performance of screening tests vary by specimen type?

Studies of tests examined new technologies in women, men, and combined groups using endocervical, vaginal, urethral, urine, rectal, and pharyngeal specimens. Further analysis based on age or risk was not provided. Results of studies rated good or fair quality are shown in Figures 4 to 9 and described below.

Studies in women using endocervical and vaginal specimens. Ten studies reported test results for women using endocervical and vaginal specimens;^{56-58,60,64,65,70,76-78} some included more than one type of test. Five reported culture results,^{57,58,65,70,77} four PCR,^57,60,70,76 two SDA,^56,65 two TMA,^77,78 and two DNA probe^58,64 (Figure 4). Most studies reported sensitivity of 90 percent or above, and specificity of 97 percent or above. Outliers included sensitivity below 90 percent in four of five studies using culture.^{58,65,65,70,77}

Studies in women using urine specimens. Six studies evaluated tests for women using urine specimens, some included more than one type of test, including four studies of PCR,^60,70,72,76 two of SDA,^56,72 and one of TMA⁷⁸ (Figure 5). Sensitivity was below 90 percent in the one study of TMA,⁷⁸ two of four studies of PCR,^60,76 and one study of SDA;⁵⁶ specificity was below 97 percent in one study of PCR.⁷⁰ In five studies evaluating both cervical and urine specimens, sensitivity was lower for urine specimens when using PCR,^60,70,76 TMA,⁷⁸ and SDA,⁵⁶ while specificity was generally comparable between specimen types.

Studies in men using urethral specimens. Four studies evaluated test technologies using urethral specimens in men, some included more than one type of test (Figure 6).^56,60,70,76 Culture was evaluated in one study,70 PCR in three,^60,70,76 and SDA in one.56 Sensitivity was below 90 percent in studies of PCR⁷⁶ and culture;⁷⁰ specificity was below 97 percent in one study of SDA.⁵⁶

Studies in men using urine specimens. Five studies evaluated tests for men using urine specimens, some included more than one type of test, including four studies of PCR^60,70,72,76 and two of SDA^56,72 (Figure 7). Sensitivity was below 90 percent in one study of SDA⁵⁶ and one study of PCR;⁷⁶ specificity was below 97 percent in one study of SDA.⁵⁶ In four studies evaluating both urethral and urine specimens in men, sensitivity was slightly lower for urine specimens when using PCR^60,70,76 and similar with SDA,⁵⁶ while specificity was generally comparable between specimen types.

Studies in men who have sex with men using pharyngeal and rectal specimens. A study of testing in a sample of 161 men who have sex with men (MSM) used the PACE® 2 assay (nucleic acid hybridization test) and culture to test pharyngeal and rectal swab specimens.⁶⁷ For PACE® 2, sensitivity was 94.1 percent and specificity 100 percent with rectal specimens, and sensitivity was 86.4 percent and specificity 100 percent with pharyngeal specimens (Figures 8 & 9). For culture, sensitivity was 88.2 percent for rectal specimens and 59 percent for pharyngeal specimens.

Studies comparing test results of symptomatic vs. asymptomatic individuals. A study of TMA in women indicated lower sensitivity among asymptomatic vs. symptomatic women (96.9 percent vs. 100 percent for cervical specimens, 87.5 percent vs. 92.6 percent for urine), but no differences in specificity.⁷⁸ For SDA, sensitivity was slightly higher in asymptomatic vs. symptomatic women (97.4 percent vs. 96.1 percent for cervical specimens, 86.5 percent vs. 83.7 percent for urine), with no differences in specificity.⁵⁶

In men, the sensitivity of PCR for both urethral and urine specimens was lower among asymptomatic vs. symptomatic men (73.1 percent vs. 98.1 percent for urethral specimens, 42.3 percent vs. 94.1 percent for urine), with no differences in specificity.⁷⁶ Results for SDA indicated slightly higher sensitivity and specificity for asymptomatic vs. symptomatic men (100 percent sensitivity and 99.5 percent specificity vs. 98.4 percent and 94.8 percent for urethral specimens, 100 percent sensitivity and 100 percent specificity vs. 97.9 percent and 94.4 percent for urine).⁵⁶

Studies of new sampling techniques. Three studies evaluated the feasibility and acceptability of self-collected specimens for gonorrhea testing.^13,39,40 The use of self-collected vaginal swabs for gonorrhea and chlamydia testing was examined in adolescent females seen for non-gynecological symptoms in a school-based clinic.¹³ The median age of subjects was 16 years, 46 percent were African American and 47 percent were white. Two percent of the participants tested positive for gonorrhea. Use of self-collected vaginal swabs was reported as easy to perform (99 percent) and preferable to a gynecological examination (84 percent). Nearly all (97 percent) stated that they would undergo testing at frequent intervals if self-testing was available.

The feasibility and acceptability of urine-based gonorrhea retesting using mailed specimens was evaluated in heterosexual patients 14 years or older who had recent positive tests for gonorrhea and/or chlamydia.⁴⁰ One hundred and twenty-two patients were randomized to two groups:

• Clinic retesting only.
• Option for clinic retesting or mailing urine specimens to the clinic.

Of those randomized to mail/clinic option,70 percent chose clinic retesting and 30 percent chose mail retesting. Age, race, gender and STD diagnosis did not differ between groups. The majority of those who chose to retest by mail were successfully retested within one month of initial diagnosis.

A descriptive study of home-based testing used free urine test kits available from local businesses that were then mailed by the user to the health department.³⁹ The study was conducted among a population of predominantly MSM with higher than average rates of STDs (Castro neighborhood of San Francisco, CA). A total of 209 kits were picked up, and 80 (38 percent) were returned to the health department. Results indicated one chlamydia infection and 3 gonorrhea infections. Respondents' biggest concern about this type of testing was confidentiality.

Key Question 2G. What is the role of screening for gonococcal infection among men who have sex with men (MSM)?

No studies meeting inclusion criteria addressed this question.

Studies relevant to gonorrhea infection in men who have sex with men. Two studies examined trends in rates of gonorrhea infection, decline in safe sex practices, and associations with the availability of highly active antiretroviral therapy (HAART).^79,80 A retrospective analysis of MSM seeking care in clinics (n=8,000) examined rates of gonorrhea from 1982 to 2001, and compared rates during pre-HAART (1990 to 1995) and post-HAART (1996 to 2001) periods.⁸⁰ Although gonorrhea rates for all individuals declined between 1982 and 1998 and stabilized at low rates, rates for MSM increased after 1995. Rates for MSM were significantly higher during the post-HAART (12.9 percent) than pre-HAART periods (8.1 percent p<0.0001). In MSM known to be HIV positive, gonorrhea rates increased from 11.6 percent in the pre-HAART to 24 percent in the post-HAART period (p<0.00001).

A retrospective analysis of behavioral factors and changes in incidence of male rectal gonorrhea was conduced in 21,587 MSM.⁸¹ Rectal gonorrhea declined from 1990 to 1993, and increased from 1994 to 1997 from 21 to 38 per 100,000 adult men (p<0.01). This increase was observed in all racial/ethnic and age groups but was highest among men aged 25 to 34 years.

Three studies examined individual-level risk factors for gonococcal infection in MSM.^82-84 A cross-sectional survey identified behavioral risk factors associated with rectal gonorrhea in MSM (n=564) by HIV serostatus.⁸⁴ The median age of participants was 33 years (range 18 to 74),⁶⁵ percent were white, and 78 percent reported some college education. MSM were included in the study if they reported receptive anal sex in the past six months. Twenty percent of the sample was HIV positive, 21 percent reported unknown HIV status, 90 percent reported no rectal symptoms, and 7 percent had rectal gonorrhea. HIV positive men were significantly more likely to have rectal gonorrhea than men with unknown or negative HIV status (OR=3.5, 95% CI, 1.9-5.8). HIV positive men who engaged in anonymous sex were at highest risk for rectal gonococcal infection. Men with unknown or negative HIV status were at highest risk if using drugs during anal sex.

A medical record review of MSM (n=1,253) seen in an urban STD clinic assessed behavioral and demographic determinants of STD acquisition.⁸² Oral insertive intercourse was independently associated with urethral gonorrhea (OR 4.4, 95% CI, 1.4-13.4). A cross-sectional study examined the prevalence of urethral infections in MSM (n=566) and found no cases of gonorrhea and few cases of chlamydia, even among those with multiple sexual partners.⁸²

Key Question 3A. What is the evidence on cost effectiveness for universal vs. targeted strategies?

A decision analysis compared standard emergency department (ED) screening practice to four enhanced screening strategies in a theoretical cohort of 10,000 female and male patients aged 18 to 31 years.⁸⁵ The five screening strategies included:

• Standard practice in which emergency clinicians rely on history and physical examination to decide whether to screen and treat.
• Universal screening.
• Selective screening for patients with risk factors combined with standard ED practice.
• Screening all patients aged 18 to 31 years combined with standard ED practice.
• Mass treatment of all patients aged 18 to 31 years with antibiotics (e.g., single dose of 1 gm azithromycin and 500 mg ciprofloxacin).

The outcomes were untreated gonorrhea or chlamydia cases and their sequellae, transmission to a partner, congenital outcomes, and cost to prevent a case of gonorrhea or chlamydia.

For women, each enhanced screening strategy was associated with less costs for clinical sequellae because of greater numbers of detected and treated infections than standard practice. Including programmatic costs and overhead, mass treatment of all women aged 18 to 31 years was the most cost-saving strategy and involved treatment of the most cases. Even with the side effects of medication accounted for, treating all women aged 18 to 31 years saved $436.54 per case treated compared with standard practice, and resulted in treatment of 1,005 additional cases of gonorrhea and chlamydia. In this modeling exercise, screening all women aged 18 to 31 years for both chlamydia and gonorrhea was found to be more cost effective than selective screening when the combined prevalence of gonorrhea and chlamydia was 7 percent to 17.5 percent.

For men, standard ED practice for detection and treatment of gonorrhea and chlamydia was more cost-saving than enhanced screening. This is most likely related to the lower costs of treatment and management of infections in men missed by screening, and the higher rates of symptomatic infections.

Although mass treatment without testing for gonorrhea and chlamydia was found to cost less for women in this analysis, the generalizability of this finding is limited because the study focused on an urban ED serving a high prevalence population. In considering the study's relevance to gonorrhea screening, it should be remembered that the reported savings are likely to have been driven by chlamydia with its higher prevalence rates. While this study did not consider the potential costs of antibiotic resistance associated with mass treatment, it also did not consider the acceptability of mass treatment to both patients and health care providers.

Key Question 3B. Are dual chlamydia-gonorrhea screening tests cost-effective?

No studies meeting inclusion criteria addressed this question.

Studies relevant to testing and treatment of dual infections. Patients infected with gonorrhea are frequently co-infected with chlamydia and routine dual treatment of patients with gonococcal infection is recommended and frequently practiced. A decision analysis examined the cost-effectiveness of routine dual treatment of women with gonococcal infection, with or without separate testing for chlamydia.⁸⁶ Three options were compared:

• Co-treat: test for gonorrhea, do not test for chlamydia, presumptively treat women with positive gonorrhea tests for both infections (baseline).
• Test: test for both infections, treat women with positive tests for their specific infections.
• Test/co-treat: test for both infections, treat women who test positive for any infection for both infections.

Three tests for gonorrhea and chlamydia were considered including culture, nucleic acid amplification assay (ligase chain reaction [LCR]), and the combination nucleic acid hybridization test (PACE® 2). Program costs and new costs of the testing and treatment algorithms were calculated. The outcome for each model was the number of cases of PID prevented. Regardless of the screening tests considered, including the combination test, the test/co-treat algorithm averted the greatest number of cases of PID, followed by the test algorithm. The co-treat algorithm averted the fewest cases of PID.⁸⁷

The decision analysis indicates that determination of the optimal algorithm should be based upon the prevalence rate of chlamydial infection, not the co-infection rate. The prevalence of gonorrhea will be lower than the prevalence of chlamydia in most settings. A relatively high prevalence of gonorrhea infection is found among STD clinic and hospital emergency department patients, jail and prison inmates, and other populations. However, in other settings such as family planning and prenatal clinics, the prevalence of gonorrhea is typically low and lower than the prevalence of chlamydia, although the co-infection rate is often in the range of 20 percent to 40 percent suggested in the 1998 CDC guidelines. Even if the co-infection rate is high, the majority of women infected with chlamydia will not be treated if treatment is determined by the outcome of a gonorrhea test as outlined in the co-treat algorithm.

Return to Contents

Pregnant Women

Key Question 1A. Does screening reduce adverse maternal/pregnancy outcomes (septic abortion, stillbirth, preterm delivery/low birth weight)?

No studies meeting inclusion criteria addressed this question.

Studies relevant to third trimester screening. Several professional groups, including the American College of Obstetricians and Gynecologists (ACOG) and the CDC, recommend repeat screening for gonorrhea during the third trimester for at-risk patients.^16,88 Two studies provide new information on third trimester screening since the last USPSTF recommendation. A retrospective chart review of clinic records over a 29-month period was conducted to determine the value of late pregnancy (34 weeks' gestation) testing for gonorrhea after a negative initial test at the beginning of prenatal care.⁸⁹ Of 751 participants, 38 women (5.1 percent) had positive gonorrhea tests at the first screening, and 19 (2.5 percent) women had positive tests only at the second screening. A prospective study evaluated screening for gonorrhea using risk factors and routine third-trimester testing in a clinic setting.⁹⁰ Five hundred and forty-two women entering prenatal care participated in the study. The risk factors examined included age less than 20 years, marital status, history of STD or hepatitis, drug use, and gestational age at entry into prenatal care. In this study, 4 percent of the third-trimester tests for gonorrhea and chlamydia were positive after an initial negative test. The presence of risk factors, such as a history of STD, drug use, and age less than 20 years, increased the risk for a positive third-trimester test 7-fold in the study sample.

Key Question 1B. Does screening reduce adverse neonatal outcomes (gonococcal conjunctivitis, blindness)?

No studies meeting inclusion criteria addressed this question.

Key Question 2A. Does screening reduce maternal complications (chorioamnionitis, premature rupture of membranes, preterm labor)?

No studies meeting inclusion criteria addressed this question.

Key Question 2B. Does screening reduce transmission to the newborn?

No studies meeting inclusion criteria addressed this question.

Key Question 3. What is the evidence on cost effectiveness for universal vs. targeted strategies?

No studies meeting inclusion criteria addressed this question.

Return to Contents

Newborn Chemoprophylaxis

Key Question 1. What are the adverse effects of treatment?

No studies meeting inclusion criteria addressed this question.

Return to Contents
Proceed to Next Section