Diagnosis of STI

Yeast Infection No More

Natural Solution for Candida Albicans

Get Instant Access

Clinical and syndromic diagnosis

Sexually transmitted pathogens cause several common syndromes. Infection with Neisseria gonorrhoea or Chlamydia trachomatis frequently results in urethritis, cervicitis, or the constellation of symptoms and signs that suggest the presence of pelvic inflammatory disease. HSV, Treponema pallidum, and Haemophilus ducreyi are common agents of ulcerative genital disease, while vaginal discharge is commonly caused by infection with Trichomonas vaginalis or Candida spp. or by bacterial vaginosis.

observational studies generated a summary estimate of the relative risk of HIV acquisition in the context of another sexually transmitted infection to be 3-7 (95% CI 2-7-5-0%) (Figure 10.1).12

However, STIs other than HIV infection may also result in chronic medical illness or long-term complications. Genital chlamydia infection is associated with tubal infertility,13 ectopic pregnancies,14,15 and chronic pelvic pain.16,17 HPVs are strongly associated with cervical and anal cancers.18,19 Infection of pregnant individuals with sexually transmitted pathogens may increase the risk of premature delivery, and may cause severe illness in the newborn.20-25

The ability of clinicians to accurately diagnose infections caused by specific pathogens without the use of diagnostic tests appears poor. For example, a study involving a cohort of 446 men presenting to a New Orleans clinic found the clinical diagnosis of the causative agents of genital ulcer disease to be highly sensitive (94-98%), but non-specific (31-35%) when compared with culture, microscopy, and serologic diagnosis (Table 10.1).26 Studies comparing clinical diagnosis of genital ulcer disease with the use of multiplex PCR have found similar limitations in clinician diagnostic accuracy.27-29

The accuracy of bedside diagnosis of vaginitis based on clinical features and simple bedside

Table 10.1 Sensitivity and specificity of ulcer appearance in identifying specific aeiologic agents of genital ulcer disease (modified from ref 26 with permission of the publisher)

Pathogen Ulcer feature Sensitivity (%) Specificity (%)

Table 10.1 Sensitivity and specificity of ulcer appearance in identifying specific aeiologic agents of genital ulcer disease (modified from ref 26 with permission of the publisher)

Pathogen Ulcer feature Sensitivity (%) Specificity (%)

Herpes simplex virus

3 or more lesions

63

64

Shallow ulcer

6O

88

Moderate tenderness on examination

60

50

All of the above features present

35

94

Haemophilus ducreyi

Undermined lesion border

85

68

Moderate or severe tenderness on examination

57

52

Purulent ulcer

64

75

All of the above features present

34

94

Treponema pallidum

Indurated ulcer

47

95

Non-purulent ulcer

82

53

Ulcer painless or minimally painful

67

58

All of the above features present

31

98

tests (i.e. pH testing, whiff test, microscopic evaluation of 'wet preps') also appears limited when compared with more comprehensive laboratory-based evaluations.30 In a study performed in 153 women presenting to a clinic in Israel with vaginal discharge, only the finding of vaginal pH < 4-5 was associated with infection by a particular pathogen (yeast); the positive predictive value of low vaginal pH for vaginal candidiasis was 68%.

None the less, the limited availability of laboratory diagnostics in areas where STI are prevalent combined with concern that patients will not return for treatment has resulted in the development of the "syndromic" approach to diagnosis and treatment. In this approach, the presence of a given clinical history or constellation of physical exam findings results in the provision of broad-spectrum therapy targeting multiple treatable organisms.31,32

Relatively simple diagnostic algorithms exist for such syndromes as genital ulceration, lower abdominal discomfort, and genital discharge.

The term "sensitivity" as applied to these algorithms indicates the proportion of individuals with infections diagnosed by laboratory methods who receive appropriate therapy as a result of algorithm use.

A recent review evaluated studies of syndromic diagnosis and management of STI; this review included no controlled trials comparing diagnostic approaches.33 Rather, attempts were made to validate algorithms using more comprehensive laboratory testing as a gold standard. Algorithms used alone have been associated with high sensitivity for urethral discharge (91-97%) and in genital ulcer diseases from syphilis or chancroid (68-100%), and vaginal discharge syndromes. However, diagnostic sensitivity is achieved at a cost of low specificity (as low as 7% in diagnosis of urethral discharge) and low positive predictive values. Thus the decision to use algorithms in settings where diagnostic tests are unavailable needs to be based on the prevalence and health impact of a given infection in the local population, and balanced against the potential

Table 10.2 Use of urine leukocyte esterase for the diagnosis of gonorrhea or chlamydia In men

Population or specimen source

Prevalence

Study gold standard

(%)

Reference

55 male STD clinic patients,

Gonorrhea

40%

Gonorrhea detected by

96

38

334

Mwanza Region, Tanzania

Chlamydia

7%

culture, chlamydia by EIA

1095 ambulatory emergency

Gonorrhea

25%

Gonorrhea and chlamydia

41

90

335

room patients, Atlanta, Georgia

Chlamydia

39%

detected by culture

479 male college students,

Gonorrhea

02%

Gonorrhea and chlamydia

26

11

40

Songkla Province, Thailand

Chlamydia

40%

detected by PCR

consequences and costs of unnecessary antibiotic treatment.

Basic laboratory testing for urethritis and cervicitis

Non-specific laboratory tests for the presence of gonorrheal and chlamydial cervicitis and urethritis include assessment of cervical, urethral and vaginal white blood cell counts, urine leukocyte esterase testing, and the use of Gram stains. Most of these modalities have proven disappointing. For example, a study evaluating the use of cervical or vaginal white blood cell counts for the identification of gonorrheal or chlamydial cervicitis found no white blood cell cut-off to be both sensitive and specific. The area under receiver operating curves created using a range of white blood cell cut-offs was < 0-6 for the presence of either type of infection, suggesting that such tests provide little additional information (i.e. a random guess would have an area of 0-5).34 Although specificity can be enhanced by the use of white blood cell cut-offs in concert with clinical findings of cervical erythema and mucopus, sensitivity of such testing remains poor, especially for chlamydia (sensitivity 41-52% for > 10 polymorphonuclear cells/high powered field).35'36

In men, urine leukocyte esterase testing has had variable sensitivity and specificity in the diagnosis of urethritis (Table 10.2), while the evaluation of urethral Gram stain findings for leukocytes has low sensitivity (-67%) for the presence of chlamydia.37

In experienced hands, the use of urethral Gram stain for the identification of Gram-negative diplococci appears to be an extremely sensitive and specific tool for the identification of gonorrhea in men. An extremely high degree of correlation between Gram stain results and nucleic acid amplification-based testing was reported in more than 7000 specimens submitted to a sexually transmitted disease program in Houston (k= 0-99).33 The ability to perform Gram stain evaluations on clinical specimens may markedly enhance the diagnostic usefulness of clinical algorithms, as described above. For example, in a study evaluating the diagnostic performance of an algorithm for urethritis, the addition of the Gram stain on urethral discharge markedly improved the specificity of algorithm diagnosis of gonorrhea (from 15% to 99%).39

The so-called "two glass test" (passage of -50 mL of urine into the first glass, with the remainder passed into the second) has traditionally been used to distinguish infection in the anterior urethra from more proximal infection (anterior urethritis is thought to be present when only the first glass specimen has a cloudy appearance). The sensitivity and specificity of this test for the

Table 10.3 Estimated sensitivity of culture for Neisseria gonorrhoea relative to newer nucleic-acid based tests

Population

Gonorrhea Culture source prevalence (%)

Reference

Female commercial sex-trade workers Endocervical in Benin, South Africa, and Thailand Male STD clinic attendees, Urethral

Baltimore, Maryland, USA

Female STD clinic attendees, Endocervical

Baltimore, Maryland, USA

Female hospital emergency department Endocervical attendees, Omaha, Nebraska, USA

Females using Duke University Endocervical health system, North Carolina, USA

IG (51-81) ll(66-86) 65 (46-8G) 89 (11-98) 93 (l6-99)

4l l

diagnosis of either gonococcal or chlamydial infection were 57% and 83% respectively in a cohort of Thai men.40

Identification of individual pathogens

Recent years have seen an explosion in the use of molecular diagnostic tests, particularly nucleic acid based methods such as the polymerase chain reaction (PCR), in the diagnosis of STI caused by fastidious pathogens. These newer testing methods may not only improve test sensitivity, but also permit the use of techniques that overcome traditional barriers to testing for sexually transmitted pathogens. For example, newer tests may yield satisfactory results when specimens are obtained via self-sampling, which may increase test acceptability.41,42

However, because newer tests may be more sensitive than the traditional "gold standards" (culture or microscopic visualization of an individual pathogen), calculation of sensitivity and specificity relative to a gold standard has become problematic. Furthermore, the use of additional tests to resolve discrepancies between negative culture tests and positive non-culture tests may introduce a form of verification bias, resulting in overestimation of sensitivity and specificity.43 Such difficulties need to be taken into account in the interpretation of the data provided below.

Neisseria gonorhoea

Culture of Neisseria gonorrhoea from the genital tract is regarded as pathognomonic for infection. The sensitivity of N. gonorrhoea culture for genital tract specimens is relatively low when compared with newer nucleic acid amplification tests (Table 10.3).44-47 Lack of sensitivity may be due in part to loss of viability associated with delays in transport; for example, a decline in sensitivity of culture testing from 89% to 78% was seen when on-site and off-site cultures were compared.46 Specimen source also contributes to the sensitivity of culture, which is as low as 30-47% when specimens are obtained from the pharynx.48,49 In a study in which 16 individuals had rectal gonorrhea identified by nucleic acid amplification, the organism was not identified by culture in any.49

More sensitive, non-culture methods for the diagnosis of gonococcal infection include enzyme immunoassay (EIA), nucleic acid hybridization ("probe") tests, and nucleic acid amplification tests. These tests have been the subject of a recent systematic review.50 Nucleic acid amplification tests identified in this review were highly sensitive and specific in the diagnosis of gonococcal infections of the cervix (sensitivity 91-100%, specificity 97-100%), male urethra (sensitivity 98-100%, specificity 98-100%), and in urine testing (94-100%, specificity 98-100%). Studies of nucleic acid amplification tests not included in this review have reported similar test characteristics.44'45,51-54

Nucleic acid hybridization or "probe" tests were also highly sensitive and specific in the diagnosis of gonococcal infections of the cervix (sensitivity 91-100%, specificity 97-100%), male urethra (sensitivity 98-100%, specificity 98-100%), and in urine testing (94-100%, specificity 98-100%).50 Although not approved for use on samples from non-genital sources, such as pharynx and rectum, probe tests and nucleic acid based tests appear to be sensitive and specific relative to culture testing, with sensitivity ranging from 86-94% and specificity 98-100% at these sites.55,56

A wide range of sensitivities (50-100%) have been reported for EIA for gonococcus, when performed on genital specimens, although these tests do appear specific (95-99%) for gonococcal infection when compared with culture.57-62 Factors that may depress the sensitivity of gonococcal EIA include disturbances of normal vaginal flora associated with concommittant vaginitis in women,62 and the degree of dilution of urinary sediment when these tests are used on urine.58

Chlamydia trachomatis

Sensitivity of culture for the recovery of Chlamydia trachomatis is more limited than is seen with gonococcus, as the former organism must be grown in cell culture, and recovery is influenced by the expertise of the testing laboratory, composition of the collection swab, and timely transport to the microbiology laboratory. The limited sensitivity of chlamydia culture for organism detection has resulted in substantial efforts being devoted to the development of non-culture methods for the diagnosis of chlamydial infection. Such methods include antigen detection methods such as direct fluorescent antigen assays (DFA), EIA, nucleic acid hybridization ("probe") tests, and nucleic acid amplification tests (including ligase chain reaction (LCR) and PCR).

A systematic review of the characteristics of non-culture tests for the identification of C. trachomatis in asymptomatic men and women aged 14-40 has been performed63 The gold standard for diagnosis in this review was considered to be either culture of chlamydia or the identification of chlamydia by two different methods. All screening tests except the use of leukocyte esterase testing of urine (see above) were highly specific for the diagnosis of chlamydia (specificity 96-100%).

In this review, the sensitivity of non-culture tests for chlamydia was compared with the sensitivity of gold standard tests by calculation of odds ratios, expressed as the odds of a false-negative test with the non-culture test divided by the odds of a false-negative test with the gold standard assay. By these criteria, a test that is more sensitive than the gold standard will have an odds ratio < 1, while a test that is inferior to the gold standard will have an odds ratio > 1. The summary odds ratios for non-culture methods are presented in Figure 10.2. While this review did not include testing in individuals with symptomatic chlamydial infection, nucleic acid amplification tests appear to be equally sensitive and retain their high specificity in symptomatic individuals.51,64,65

Available evidence suggests that the sensitivity and specificity of nucleic acid amplification tests and probe testing for chlamydia on self-collected vaginal specimens or submitted tampons is similar to that seen with clinician-collected cervical specimens, and self-collection may have the advantage of greater acceptability or convenience in some circumstances.66-69 The use

Proportion false negative Test Gold Standard

Odds Ratio (95% CI)

Urine LCR (5 studies)

48/413

112/413

0 33 (0-13-0-80)

Urine PCR

(7 studies)

41/336

45/354

0 84 (0^37-1^89)

Cervix PCR (12 studies)

30/375

96/367 ^^^^

0 26 (0^12-0^56)

Urine probe (1 study)

7/39

13/39

044 (0^15-1^26)

Cervix probe

(4 studies)

20/128

20/128

116 (0^25-547)

EIA urine (4 studies)

40/98

36/98

186 (0•39-8•75)

EIA cervix (9 studies)

20/128

20/128

4^10 (1^15-14^59)

DFA cervix

5/40

(2 studies)

6/40 4

► 105 (0^09-12^93)

^ Favors test Favors gold standard ►

Figure 10.2 Pooled odds ratios for false negative results in screening tests for Chlamydia trachomatis infection. The authors of a meta-analysis of testing methods for asymptomatic chlamydial infection found most testing methods to be highly specific. Pooled sensitivity estimates were assessed by calculating the ratio of the odds of a false-negative result with each testing method (listed on the Y axis) and the odds of a false-negative result with gold standard tests. Odds ratios and 95% confidence limits are represented by black diamonds and listed on the right side of the figure; arrows indicate confidence limits that extend beyond the scale presented on the X axis. The sensitivity of PCR on cervical specimens and urine LCR were significantly better than those of gold standard assays, while the sensitivity of EIA on cervical specimens was significantly worse. Modified from J Med Microbiol 2000;51(12):1027, with permission from the Society for General Microbiology.

of self-collected specimens may open the way to approaches such as mail-in sampling for population-based screening. In a study performed in general practices in Denmark, testing of pooled self-collected mail-in specimens had a sensitivity and specificity comparable to that seen with testing of pooled physician-collected cervical and urethral swabs (sensitivity 96-100%, specificity of 93-100% with self-collected specimens; sensitivity 91%, specificity 100% with clinician-collected specimens).70

Pelvicinflammatory disease

Clinical assessment remains the mainstay of diagnosis of pelvic inflammatory disease (PID), a spectrum of pathological conditions including endometritis, salpingitis, tubo-ovarian abscess, and pelvic peritonitis. "Gold standard" tests (for example, endometrial biopsy, laparoscopy) are invasive and not readily available in many clinical settings. The triad of lower abdominal discomfort, cervical motion tenderness, and adnexal tenderness has been suggested to represent minimal diagnostic criteria for PID.71

Table 10.4 Sensitivity and specificity of ultrasonographic detection of fluid-tilled fallopian tubes in the diagnosis of pelvic inflammatory diseases

Type of

Study gold

Sensitivity (%)

Specificity (%)

Population

sonography

standard

(95% CI)

(95% CI)

Reference

51 non-pregnant

Transvaginal

Plasma cell

85 (55-98)

100 (91-100)

336

outpatients in

endometritis on biopsy

Helsinki, Finland

30 consecutive individuals

Transvaginal

Presence of PID

81 (58-95)

78 (40-97)

337

hospitalized for suspected

at laparoscopy

PID in Helsinki, Finland

55 women with suspected

Transvaginal

Presence of PID at

32 (13-57)

97 (85-100)

338

PID in Providence,

laparoscopy or histological

Rhode Island, USA

endometritis on biopsy or culture of N. gonorrhoea or C trachomatis from upper genital tract specimen

lower genital tract had a sensitivity and specificity of 77% for the presence of PID.74

Two recent studies have used the presence of plasma cell endometritis, rather than laparoscopic evidence of PID, as the gold standard for the diagnosis of PID.75'76 One study found the CDC "minimal diagnostic criteria" to be only 33% sensitive for the presence of plasma cell endometritis, but 88% specific,75 while a second study found the CDC criteria to be more sensitive (83%) but less specific (22%).76

When available, ultrasonography may aid in the diagnosis of PID. The finding of fluid-filled fallopian tubes on ultrasound appears to be specific for the presence of PID, although the sensitivity of this finding has varied between studies (Table 10.4).

Trichomonas vaginalis

Trichomonas vaginalis is a unicellular flagellated organism that causes vaginitis in women and urethritis in men. The importance of diagnosis and treatment relates to the association between infection with this organism and adverse

A systematic review evaluated the sensitivity and specificity of historical, clinical, and laboratory findings for PID, when compared with laparoscopic diagnosis.72 This review found no evidence that historical information (for example, history of irregular menses or history of intrauterine device use) can reliably identify the presence of PID in cohorts of women with abdominal pain and other signs of genital tract infection. The presence of individual clinical signs, such as purulent vaginal discharge or a palpable adnexal mass on examination in an individual with a complaint of abdominal tenderness, was both insensitive and nonspecific.72 In a study performed in Sweden in the 1960s, the presence of at least four clinical signs (such as pelvic tenderness, pelvic mass, fever, and abnormal vaginal discharge) was found to be specific (91%) for laparoscopically diagnosed PID but had a sensitivity of only 39%.73

The detection of gonorrhea or chlamydia may be helpful in the diagnosis of PID in individuals with compatible signs and symptoms. In a study performed in a cohort of women with abdominal pain and tenderness on bimanual examination, the isolation of one of these organisms from the outcomes in pregnancy, as well as enhanced HIV transmission.20,77 A meta-analysis of test characteristics associated with simple bedside tests, such as the use of 'wet mounts', and the use of Papanicolaou smear testing, found the sensitivity of these methods to be low (wet mount sensitivity 68% [95% CI 62-74]; Papanicolaou smear sensitivity 58%, [95% CI 43-73%]).78

Superior sensitivity is seen with other testing modalities, including culture using special media and PCR-based testing. A systematic review and meta-analysis found that culture using special media has a sensitivity of 90% (95% CI 77-93%), while PCR has a sensitivity of 95% (95% CI 91-99%) and specificity of 98% (95% CI 96-100%) relative to culture. Other non-culture tests, including DFA testing (sensitivity 85%, specificity 99%) and ELISA (sensitivity 82%, specificity 73%) also have good test performance, and are less expensive than culture methods.79

It has been noted that the high sensitivity of culture remains intact even with the use of standard transport media,80 and despite delays in inoculation into special media.81 This has led to the suggestion that a two-step process might be more efficient, with inexpensive and highly specific 'wet mount' testing used initially, and more expensive and sensitive tests reserved for specimens that test negative by wet mount.82 However, it should be noted that such an approach would provide few advantages in settings where the prevalence of T. vaginalis i nfection is low.

Chancroid

Chancroid is an ulcerative genital disease caused by Haemophilus ducreyi. This organism has a distinct microscopic appearance, and direct Gram staining of purulent material from the ulcer base may reveal chains of short, Gramnegative bacilli. Such a finding had a sensitivity of 60% compared with culture in a cohort of individuals with genital ulcer disease attending a sexually transmitted diseases clinic in Nairobi.83

Of 37 individuals who did not have H. ducreyi isolated by culture, 18 had Gram stain findings suggestive of H. ducreyi, suggesting either lack of sensitivity of culture or lack of specificity of Gram stain. When compared with the use of concurrent PCR assays for H. ducreyi-specific sequences, culture for H. ducreyi had a sensitivity ranging from 63% to 87%.84-86

Initial studies evaluating the use of PCR for the identification of H. ducreyi in clinical settings estimated sensitivity to be as low as 62% relative to culture.87 However, subsequent technical improvements in specimen preparation have increased the sensitivity of PCR,87 and more current estimates of the sensitivity of PCR for detection of H. ducreyi range from 79-98%, with specificity of 92-100% relative to culture.85,86,88

The use of PCR for the identification of H. ducreyi has provided important insights into the epidemiology of chancroid; for example, it has been observed that H. ducreyi may be present in ulcers co-infected with herpes viruses or T. pallidum.27-29,86 Further, the phenomenon of asymptomatic carriage of H. ducreyi has been observed in 2% of commercial sex workers in The Gambia without signs or symptoms of chancroid.89

Other diagnostic modalities, including an indirect immunofluorescent assay, and an enzyme immunoassay, may also have value in the diagnosis of chancroid.84

Herpes simplex viruses

HSV are the most common agents of ulcerative genital disease in the developed world, and are increasingly recognized in the developing world as well.90 The gold standard test for diagnosis of genital herpes has traditionally been culture of virus from genital lesions. If viral culture is not available, infection may be diagnosed by evaluating ulcer scrapings for the presence of multinucleated giant cells ("Tzank smear"). The sensitivity of Tzank smear relative to culture is

52-80% in anogenital lesions, with higher sensitivity in men than in women; the corresponding specificity is reported as 93%.91 When used for orolabial herpes, the Tzanck smear has a reported sensitivity of 54% and a specificity of 100% relative to culture.92

Enzyme immunoassays provide a rapid and sensitive alternative to culture for identification of HSV. The sensitivity of these tests has been estimated to be 80-96%, while their specificity has been reported as 93-100%.93-96 Direct immunofluorescent assays may also be useful for the diagnosis of HSV in the genital tract, and provide a more timely diagnosis than would be obtained with culture. Reported sensitivity is 74-80%, and specificity is 85-98% relative to culture.97-98

As has been noted, the quantification of the sensitivity and specificity of newer assays (for example, nucleic acid amplification-based assays) is difficult, since these assays are more sensitive than culture, the traditional gold standard. For example, in studies using PCR as the gold standard, viral culture has a sensitivity of 72-88%,85'93'99'100 while EIA has a sensitivity of 65%.93

Older serologic assays for anti-herpes simplex antibody were unable to reliably distinguish between infection with HSV-1 and HSV-2.101 More recent serologic assays, such as glycoprotein G-based Western blot, can differentiate the response to infection with these two viruses, and are more than 90% sensitive if performed 21 days or more after primary infection.102 Based on individuals prospectively followed in the setting of randomized controlled trials, it can be estimated that approximately 40% of those who acquire HSV-2 infection (as evidenced by seroconversion) actually develop genital herpes.103

Newer ELISA-based assays for type-specific antibodies against HSV-1 and HSV-2 are sensitive and specific104,105 when compared with Western blot assays, and are less expensive to perform. The role of antibody testing in the diagnosis of genital herpes remains poorly defined, but such tests might be used in counselling couples,106 and in pregnancy-related screening.107,108

Syphilis

Primary and secondary syphilis may be diagnosed by visualization of spirochetes from ulcers, condylomata lata, and mucous patches using dark-field microscopy. Such diagnostic methods require both technical competence and experience; in the hands of an experienced microscopist, the sensitivity of dark-field microscopy has been estimated to range from 74 to 81% when compared with various reference standards.85,109-111 The finding of motile spirochetes by dark-field microscopy in a sample from a genital lesion might be expected to be pathognomonic for syphilis, but other non-pathogenic genital tract spirochetes may lead to false positive test results.112 Antibody-based assays and PCR may also be used to detect the presence of T. pallidum in lesions of primary or secondary syphilis, and may offer improved sensitivity in detection of treponemes (Table 10.5).

Serological testing is the mainstay of syphilis diagnosis in adults with non-primary disease; the characteristics of these tests have been reviewed in detail elsewhere.112,113 Such tests can be classified as non-treponemal tests, which identify antibodies not directed against treponemes, and treponemal tests, which identify antibodies directed at treponemal components. Non-treponemal tests may be positive in the presence of a primary chancre, but are less than 90% sensitive in primary syphilis. Sensitivity is higher in secondary and early latent syphilis. By contrast, the fluorescent treponemal antibody absorbed assay (FTA-Abs), a treponemal test, is usually positive within a week of the development of a primary chancre (Figure 10.3).

Table 10.5 Diagnostic characteristics of commonly used tests for the detection of Treponema pallidum in early syphilis

Population or specimens source

Prevalence (%)

Comparator or study gold standard

Sensitivity (%)

Specificity (%)

Reference

Darkfield microscopy

128 individuals with anogenital lesions attending an STD clinic in Edmonton, Alberta, Canada

52

Positive darkfield evaluation or positive serologic test for syphilis

79

100

111

350 specimens taken from individuals with lesions suggestive of syphilis (>1 specimen per individual)

34

"Subsequent diagnosis of syphilis"

74

97

110

302 individuals with genital ulcer disease in Pune, India

14

Multiplex PCR

39

82

28

188 individuals with genital lesions attending STD clinics in Brooklyn, New York, and Seattle, Washington, USA

34

Direct fluorescent monoclonal antibody testing

85

96

122

295 men presenting to a New Orleans STD clinic with genital ulcer

25

Multiplex PCR

81

100

85

241 individuals assessed at county clinics in San Francisco and Los Angeles with lesions suggestive of primary syphilis

22

Direct fluorescent antibody testing

85

97

339

Direct fluorescent antibody test

241 individuals assessed at county clinics in San Francisco and Los Angeles with lesions suggestive of primary syphilis

18

Darkfield microscopy

86

93

339

156 individuals with genital ulcer disease from Malawi

17

PCR with dot-blot hybridization

85

97

340

128 individuals with anogenital lesions attending an STD clinic in Edmonton, Alberta, Canada

52

Positive darkfield evaluation or positive serologic test for syphilis

79

100

111

350 specimens taken from individuals with lesions suggestive of syphilis (>1 specimen per individual)

34

"Subsequent diagnosis of syphilis"

86

100

110

(Continued)

Table 10.5 Continued

Population or specimens source

Prevalence (%)

Comparator or study gold standard

Sensitivity (%)

Specificity (%)

Reference

188 individuals with genital lesions attending STD clinics in Brooklyn, New York, and Seattle, Washington, USA

34

Dark-field microscopy

91

93

109

PCR

295 men presenting to a New Orleans STD clinic with genital ulcer

22

Dark-field microscopy

100

99

85

due to "prozone" phenomena, which occur when extremely high titers of antibody disrupt the assay. This results in a false-negative test result, which becomes positive upon dilution.114 Non-treponemal tests revert to negative over time in approximately 30% of untreated individuals;115 treponemal tests may uncommonly revert to negative, a phenomenon that appears to be more common in individuals with HIV-associated immune dysfunction.116

The specificity of non-treponemal tests is problematic, and reports of falsely positive non-treponemal tests in the presence of other infectious diseases, rheumatologic diseases, and pregnancy are common.117 The relative risk of a false-positive non-treponemal test in individuals with underlying HIV infection was 8-4 (95% CI 4-2-13-6%) in a Spanish cohort.118 None the less, non-treponemal tests remain useful as screening tests because of their low cost, and because a reduction in titer following treatment is a useful indicator of microbiological cure.119 Treponemal tests are more specific than non-treponemal tests, although false-positive test results are reported.117 The characteristics of tests for the serologic diagnosis of syphilis are presented in further detail in Table 10.6.

The diagnosis of neurosyphilis is challenging. While VDRL testing of cerebrospinal fluid (CSF

1GG-I

Primary Secondary

Late

2G 3G Years

Primary Secondary

Late

3 e Weeks

9 12

2G 3G Years

Figure 10.3 Timing of serological test positivity in syphilis. Comparison of timing of test positivity for a non-treponemal test (rapid plasma reagin or RPR), and two treponemal tests (fluorescent treponemal antibody absorbed (FTA-ABS) and microhemagglutination assay for T. pallidum (MHA-TP)). Both the RPR and FTA-ABS are positive in most individuals with a primary chancre, but FTA-ABS is more sensitive in primary syphilis. The two treponemal tests remain positive over time, while RPR will revert to negative in approximately one-third of untreated individuals. Reproduced from Larsen SA, Steiner BM, Rudolph AH. Laboratory diagnosis and interpretation of tests for syphilis. Clin Microbiol Rev 1995;8(1):1-21 with permission from the American Society for Microbiology.

A small proportion of individuals with syphilis have a negative non-treponemal test for syphilis

Table 10.6 Ranges of sensitivity and specificity reported for serological tests for syphilis by stage

Sensitivity

Table 10.6 Ranges of sensitivity and specificity reported for serological tests for syphilis by stage

Sensitivity

Test

Primary

Secondary

Latent

Late

Specificity (%)

Reference

Non-treponemal tests

VDRL

74-87

100

88-100

34-94

96-99

115,341,342

RPR

77-100

100

95-100

73

93-99

115,342

TRUST

77-86

100

95-100

98-99

115

USR

72-88

100

88-100

99

115

Treponemal tests

FTA-Abs

70-100

100

99-100

96

84-100

115,341-343

MHA-TP

69-90

99-100

97-100

94

98-100

115,341-343

Non-standard tests

ELISA

82-100

91-100

86-100

100

89-100

344-349

Western blot

78-100

98-100

83-100

100

97-100

350-352

Abbreviations: VDRL, Venereal Disease Research Laboratory; RPR, Rapid Plasma Reagin; TRUST, toluidine red unheated serum test; USR, unheated serum reagin; FTA-Abs, fluorescent treponemal antibody absorbed; MHA-TP, microhemagglutination assay for T. pallidum; ELISA, enzyme-linked immunosorbent assay. Modified with permission from Larsen SA, Pope V, Johnson RE, Kennedy EJ, eds. A Manual of Tests for Syphilis, 9th ed. 13-18. Copyright © 1998 by American Public Health Association.

Abbreviations: VDRL, Venereal Disease Research Laboratory; RPR, Rapid Plasma Reagin; TRUST, toluidine red unheated serum test; USR, unheated serum reagin; FTA-Abs, fluorescent treponemal antibody absorbed; MHA-TP, microhemagglutination assay for T. pallidum; ELISA, enzyme-linked immunosorbent assay. Modified with permission from Larsen SA, Pope V, Johnson RE, Kennedy EJ, eds. A Manual of Tests for Syphilis, 9th ed. 13-18. Copyright © 1998 by American Public Health Association.

VDRL) is often advocated, the sensitivity of this test is poor. A retrospective study was performed in 38 individuals with positive cerebrospinal fluid FTA-Abs (a test thought to be sensitive but non-specific for the diagnosis of neurosyphilis). Fifteen of 38 had likely neurosyphilis on the basis of a compatible clinical history and other CSF abnormalities (for example, leukocytosis or elevated protein), but only four of these 15 individuals had a positive CSF-VDRL (sensitivity 27%).120

The use of the "TPHA index" has been suggested as a more sensitive means of diagnosing neurosyphilis. This index is based on an antibody test (MHA-TP) that is more sensitive than CSF VDRL. False-positive test results are reduced by adjusting for CSF protein concentration, which in turn helps to control for blood contamination of the CSF sample.121 However, a study in individuals co-infected with HIV and syphilis found high index values in only five of 40 individuals with possible neurosyphilis, and three of five individuals with positive CSF VDRL tests, suggesting that the TPHA index may also be relatively insensitive for active central nervous system infection.122

Existing evidence does not support the routine use of PCR for the diagnosis of neurosyphilis in adults, and published studies have yielded inconsistent results.123,124 In a study conducted in infants born to mothers with untreated syphilis in Dallas, Texas, CSF PCR had a sensitivity of 65% when compared with a gold standard of rabbit infectivity testing; in this study PCR of blood or serum was more sensitive than CSF PCR for the presence of central nervous system disease (94%).125

Diagnosis of genital warts and human papillomavirus infection

The diagnosis of genital warts is usually made clinically, but rigorous studies of the sensitivity and specificity of clinical diagnosis are lacking. Although the intuition of experienced clinicians was more sensitive and specific than the use of a standardized diagnostic instrument in a small study of extragenital warts, the gold standard used in this study was the clinical judgment of one of the study investigators.126

Acetic acid (3-5%) has been used as an adjunct to the clinical diagnosis of genital warts, and whitening with acid application is said to signify the presence of underlying HPV infection. The application of acetic acid has also been advocated for the identification of subclinical warty lesions. However, whitening appears to be non-specific for the presence of HPV infection; in a cohort of Swedish army conscripts, HPV DNA was detected by PCR in only 17 of 39 biopsy specimens taken from aceto-white areas, and there was no difference in the detection of HPV DNA in urethral brushings from men with and without aceto-white lesions.127 In another study HPV DNA was detected in only 55 of 91 aceto-white lesions detected by penoscopy, with other aceto-white biopsy specimens having histology suggestive of eczema.128

Furthermore, aceto-white lesions appear to be insensitive for the presence of HPV infection: in a cohort of Swedish women undergoing colposcopy, the finding of an aceto-white vulvar lesion had a sensitivity of 44% for the detection of HPV DNA by PCR.129 Finally, many clinically typical genital warts do not turn white with the application of acetic acid. In a study of 202 men in Chandigarh, India, all hyperplastic warts turned white with the application of acetic acid, but only one of 12 typical verruca vulgaris-type lesions, and 15 of 59 flat warts did so.130 Thus, the poor sensitivity and specificity of acetic acid testing for small or subclinical genital warts, combined with the lack of evidence to suggest that treatment of such lesions changes long-term outcome, makes it difficult to advocate the routine use of acetic acid testing for external genital warts.

Similarly, no evidence exists currently to support the use of HPV DNA testing in the clinical diagnosis of external genital warts. However, such testing may contribute substantially to cervical cancer screening programs. The presence of "high-risk" HPV DNA in genital tract specimens of women with atypical squamous cells of undetermined significance (ASCUS) on Papanicolaou smear is highly sensitive for the presence of underlying cervical neoplasia.131-133 Mathematical models based on available screening data suggest that the incorporation of HPV DNA testing into screening practices would likely be cost-effective relative to current practices.134-136 A more complete review of the relationship between human papillomavirus and cervical neoplasia is available elsewhere.137

Was this article helpful?

0 0
How To Get Rid Of Yeast Infections Once And For All

How To Get Rid Of Yeast Infections Once And For All

No more itching, odor or pain or your money is refunded! Safe and DRUG FREE Natural Yeast Infection Solutions Are you looking for a safe, fast and permanent cure for your chronic yeast infection? Get Rid of that Yeast Infection Right Now and For Good!

Get My Free Ebook


Post a comment