Original Contribution

diagnostic utility of the genital Gram stain in ED patients^?

Peter Stefanski BS ^a, John W. Hafner MD, MPH ^b,c,?, Shanda L. Riley MD ^d,

Kharmene L.Y. Sunga MD ^e, Timothy J. Schaefer MD ^b,c

^aUniversity of Illinois College of Medicine at Peoria, Peoria, IL 61637, USA

^bEmergency Medicine, University of Illinois College of Medicine at Peoria, Peoria, IL 61637, USA ^cDepartment of Emergency Medicine, OSF Saint Francis Medical Center, Peoria, IL 61637, USA ^dDepartment of Emergency Medicine, Wesley Medical Center, Wichita, KS 62714, USA

^eMayo School of Graduate Medical Education, Mayo Clinic College of Medicine, Rochester, MN 55902, USA

Received 12 December 2007; revised 25 August 2008; accepted 5 September 2008

Abstract

Objective: The study aimed to determine the diagnostic usefulness of the genital Gram stain in an emergency department (ED) population.

Methods: A linked-query of an urban, tertiary-care, university- affiliated hospital laboratory database was conducted for all completed Chlamydia trachomatis and Neisseria gonorrhoeae DNA probes, Trichomonas vaginalis wet preps, and genital Gram stains performed on ED patient visits between January and December 2004. Positive criteria for a Gram stain included greater than 10 white blood cells per high-power field, gram-negative intracellular/extracellular diplococci (suggesting N gonorrhoeae), clue cells (suggesting T vaginalis), or direct visualization of T vaginalis organisms. DNA probes were used as the gold standard definition for N gonorrhoeae and C trachomatis infection. Results: Of 1511 initially eligible ED visits, 941 were analyzed (genital Gram stain and DNA probe results both present), with a prevalence of either C trachomatis or N gonorrhoeae of 11.4%. A positive genital Gram stain was 75.7% sensitive and 43.3% specific in diagnosing either C trachomatis and/or N gonorrhoeae infection, and 80.4% sensitive and 32.2% specific when the positive cutoff was lowered to more than 5 white blood cells/high-power field. No Gram stains were positive for T vaginalis (with 47 positive wet mounts), and clue cells were noted on 117 Gram stains (11.6%).

Conclusion: Gram stains in isolation lack sufficient Diagnostic ability to detect either C trachomatis or

N gonorrhoeae infection in the ED.

(C) 2010

Introduction

^? Presented as a poster presentation at the American College of Emergency Physicians 2005 Research Forum, September 2005.

* Corresponding author. Emergency Medicine, University of Illinois

College of Medicine at Peoria, Peoria, IL 61637, USA.

E-mail address: [email protected] (J.W. Hafner).

In the United States alone, more than 19 million Sexually transmitted diseases are diagnosed every year, and it is estimated that more than half of all Americans will acquire an STD at some point in their lifetime [1-3]. Data from the 1992 to 1994 National Hospital Ambulatory Medical Care Survey estimates that women of reproductive age make 1.4

0735-6757/$ - see front matter (C) 2010 doi:10.1016/j.ajem.2008.09.016

million visits to the emergency department (ED) per year for gynecologic reasons, with almost half relating to infectious diseases, including STDs, lower genital tract infection, and pelvic inflammatory disease [4]. Pelvic inflammatory disease is the most regarded complication of STDs and can lead to Ectopic pregnancy or infertility [5,6]. In the United States, more than 8 billion dollars annually is spent diagnosing and treating STDs and their complications [7].

Often, testing and treatment of STDs occurs in the ED, which serves as a medical safety net for underserved patient populations. Many patients at highest risk for STDs will present to the ED for care. When facing lengthy differentials for complaints like pelvic or abdominal pain, emergency medicine (EM) physicians require rapid, inexpensive, and effective laboratory testing for accurate and Timely diagnosis. The Gram stain is a rapid and economical test that has historically been used to diagnose the etiology of suspected undifferentiated genital infections. Although the genital Gram stain has not been extensively studied within the setting of an ED, some research has been done to ascertain its utility in other settings. Several studies have shown conflicting results for diagnostic effectiveness [11-34]. Our study gauges the usefulness of the genital Gram stain in an ED population vs the current gold standard, the DNA probe.

Materials and methods

This study was conducted at OSF Saint Francis Medical Center in Peoria, IL, an urban tertiary-care, university- affiliated teaching hospital associated with the University of Illinois College of Medicine at Peoria. The ED receives more than 65 000 visits annually and is associated with a long- standing EM residency program. All genital samples are analyzed on site using the OSF Saint Francis Medical Center Regional Laboratory.

A query of the Medical Center’s Regional Laboratory database was performed for all Neisseria gonorrhoeae and Chlamydia trachomatis DNA probes, Trichomonas vaginalis wet preps, and genital Gram stains performed on ED patients between January and December of 2004. Either a board- certified attending emergency physician or a resident physician had collected all specimens using standard technique. Collection swabs obtained during routine care remained consistent throughout the course of the study, and treating physicians were unaware of the study. Laboratory personnel performed all Gram stains, using a standardized 4- step staining protocol with Hucker’s modification and Becton Dickinson (Becton Dickinson and Company, Frank- lin Lakes, NJ) Gram stain kits. The Becton Dickinson Gram stain kits contain 4 single-use standard reagents: crystal violet primary stain, L-polyvinylpyrollidone-iodide mordant, acetone/isopropanol decolorizer, and a safranin counterstain. Standard inorganic iodine is rapidly oxidized and loses its effectiveness as a mordant, and therefore the more stable

organic iodine L-polyvinylpyrollidone-iodide is the preferred modern reagent for performing Gram stains [8]. The overall accuracy of the Gram stain in general has been previously noted to be dependent on the training and skill of the laboratory personnel performing the stain and the micro- biologist interpreting the slide [9]. Laboratory personal perform Clinical Laboratory Improvement Amendments- regulated quality control measures for gram-staining reagents and technique using slides of known gram-positive (Staphylococcus aureus) and gram-negative (Escherichia coli) organisms with each new box of kits used. Emergency department visits were eligible for inclusion in the study if both a genital Gram stain and an associated gold standard test was available.

A priori positive criteria for a genital Gram stain was defined as a report of greater than 10 white blood cells (WBCs) per high-power field (HPF), a report of gram- negative intracellular/extracellular diplococci (suggesting N gonorrhoeae), a report of clue cells (suggesting Gardnerella vaginalis), and/or a direct visualization of T vaginalis. DNA probes results were used as the gold standard for N gonorrhoeae and C trachomatis genital infections. Becton Dickinson ProbTec ET DNA amplification assays (Becton Dickinson and Company) were used for laboratory DNA probe diagnosis of urethral and endocervical swabs. The reported manufacturer sensitivity/specificity is 92.8/98.1% for C trachomatis and 96.6/99.5% for N gonorrhoeae [10]. Direct microscopic visualization of organisms on a separate wet mount prep was considered the gold standard for T vaginalis infection. Descriptive analysis of the data and sensitivity, specificity, negative and positive predictive values, and positive and negative likelihood ratios were calculated using SPSS 12.0 (SPSS, Chicago, Ill) and correlated with 95% confidence intervals (CIs).

Results

During the study period, 1511 separate ED visits were eligible for analysis, and 941 (62.3%) of them fulfilled inclusion criteria (Fig. 1). Of the visits analyzed, 90.7% were by female patients. Of the 140 males present in the initial study population, only 6 ED visits reported both a Gram stain and a DNA probe (sensitivity 100%, specificity 100%).

The prevalence of C trachomatis and/or N gonorrhoeae infection among analyzed visits was 11.4%. The combined overall sensitivity of the Gram stain for diagnosing either C trachomatis or N gonorrhoeae was 75.7% (95% CI, 67.3%-

82.6%) and 43.3% (95% CI, 42.2%-44.2%) specific

(Table 1). The Gram stain was 71.1% sensitive (95% CI, 55.6%-82.9%) and 41% specific (95% CI, 41.0%-42.1%) in

diagnosing strictly N gonorrhoeae (Table 2). Similarly, the Gram stain was 75.6% sensitive (95% CI, 66.0%-83.3%) and 42.8% specific (95% CI, 41.8%-43.6%) in diagnosing C trachomatis (Table 3). The genital Gram stain positive and

DNA probe +		DNA probe -	Total
Gram stain +	27	527	554
Gram stain -	11	376	387
Total	38	903	941
Sensitivity	71.1%	95% CI, 55.6%-82.9%
Specificity	41.6%	95% CI, 41.0%-42.1%
Positive predictive	4.9%	95% CI, 3.8%-5.7%
value
Negative predictive	97.2%	95% CI, 95.6%-98.3%
value
Positive likelihood	1.22	(95% CI, 0.94-1.43)
ratio
Negative likelihood	0.70	(95% CI, 0.41-1.08)
ratio

Fig. 1 Case assignment.

negative likelihood ratios indicated a limited clinical value in diagnosing N gonorrhoeae or C trachomatis (Tables 1-3).

When the cutoff for a positive genital Gram stain test was lowered to more than 5 WBCs/HPF, the sensitivity increased to 80.4% (95% CI, 72.4%-86.6%) and the specificity

decreased to 30.2% (95% CI, 29.2%-31.0%). This lower

cutoff produced a positive likelihood ratio of 1.15 (95% CI, 1.02%-1.26) and a negative likelihood ratio of 0.65 (95% CI, 0.43-0.95). No Gram stains were positive for T vaginalis, with 47 positive wet mounts. Clue cells, indicative of G vaginalis, were seen on 117 Gram stains (11.6% of total Gram stains).

Discussion

Genital Gram stains have been studied both as part of clinical algorithms to detect cervical infection, and for their independent diagnostic contribution. This study demon- strates that genital Gram stains in isolation lack sufficient diagnostic ability to detect C trachomatis and N gonorrhoeae infection in the ED. As expected, with a lower threshold for a positive genital Gram stain (from 10 to 5 WBCs/HPF), the combined overall sensitivity rose (75.7% vs 80.4%), and the

Table 1 Overall effectiveness of Gram stain for diagnosing either N gonorrhoeae or C trachomatis

Table 3 Effectiveness of Gram stain for diagnosing C trachomatis

Table 2 Effectiveness of Gram stain for diagnosing N gonorrhoeae

DNA probe +		DNA probe -	Total
Gram stain +	81	473	554
Gram stain -	26	361	387
Total	107	834	941
Sensitivity	75.7%	95% CI, 67.3%-82.6%
Specificity	43.3%	95% CI, 42.2%-44.2%
Positive predictive	14.6%	95% CI, 13.0%-16.0%
value
Negative predictive	93.3%	95% CI, 91.0%-95.2%
value
Positive likelihood	1.34	95% CI, 1.16-1.48
ratio
Negative likelihood	0.56	95% CI, 0.39-0.78
ratio

DNA probe +		DNA probe -	Total
Gram stain +	65	489	554
Gram stain -	21	366	387
Total	86	855	941
Sensitivity	75.6%	95% CI, 66.0%-83.3%
Specificity	42.8%	95% CI, 41.8%-43.6%
Positive predictive	11.7%	95% CI, 10.2%-12.9%
value
Negative predictive	94.6%	95% CI, 92.4%-96.3%
value
Positive likelihood	1.32	95% CI, 1.14-1.48
ratio
Negative likelihood	0.57	95% CI, 0.38-0.81
ratio

specificity fell (43.3% vs 30.2%). Past studies have both supported and refuted use of the endocervical Gram stain (EGS). They have examined different populations in various environments, using an assortment of tests as standards for comparison.

Some early studies support the use of EGS to diagnose chlamydial and gonococcal infections. Studies by Brunham et al [11] and Katz et al [12] demonstrated that 10 or more polymorphonuclear neutrophil (PMN)/HPF on Gram stain correlated with cervical chlamydia infection, and Moscicki et al found EGS to be useful in identifying populations at risk for C trachomatis infection (sensitivity, 91%; specificity, 65%; PPV, 31%) [13]. Other studies showed that EGS was highly specific for diagnosis of gonorrhea infection [14-17]. The Gram stain has been extensively studied in female sex workers (FSW). Deceuninck et al [18] analyzed the use of Gram stain among FSW in Ghana and discovered it to be 42.4% sensitive and 98.5% specific for diagnosis of gonorrhea, as compared to PCR; they also noted nearly equivalent performance of vaginal and cervical Gram stains. However, among 374 sex workers in Senegal, Ndoye et al noticed a significant association between STDs and 10 or more leukocytes on EGS (odds ratio, 3.45; 95% CI, 2.04- 5.85). Although this was encouraging, the EGS observed only 66% sensitivity, 64% specificity, and 1.83 LR [19]. Most of the studies performed in FSW show a significant association between number of PMNs on EGS and chlamydial/gonococcal infection [20-22]. This is likely due

to high disease prevalence in these populations.

The genital Gram stain has also been widely studied in STD clinic and family planning clinic populations, and its utility in diagnosing chlamydial or gonococcal infection has been challenged. In 495 STD clinic patients, Knud-Hansen et al found that EGS diagnosed gonorrhea infection with a sensitivity and specificity of 76% and 44% and a combined gonorrhea or chlamydia sensitivity and specificity of 47% and 75%, as compared to culture [23]. In another study of clinic patients with mucopurulent endocervical discharge and more than 10 PMNs/HPF on EGS, Sellors et al [24] found a specificity of 97% (sensitivity, 18.9%; positive predictive value, 29.2%) for chlamydial cervicitis as compared to culture and EIA. Myziuk et al compared EGS to culture and direct fluorescent antibody in females visiting an STD clinic and found equal numbers of C trachomatis isolates in females with both b10 PMNs/HPF (sensitivity and specificity 52% and 71%) and N10 PMNs/HPF (sensitivity and specificity 48% and 29%). However, when combined with clinical information, EGS was statistically significant for the presence of C trachomatis (96%-98% specificity) [25]. Our study did not assess the contribution of clinical information but agrees that the Gram stain is inadequate when used in isolation (sensitivity, 75.6%; specificity, 42.8%). Other studies have investigated EGS in the context of clinical algorithms and have yielded conflicting data regarding the correlation between chlamydial or gonorrheal infection and PMNs on Gram stain [26-32].

Our results tend to support the skepticism of these more recent studies and oppose the studies that suggest use of the Gram stain to aid in diagnosis of cervical infection. Our study pits the genital Gram stain against the clinical gold standard of our day-the DNA probe. Few studies have compared the genital Gram stain to currently accepted and readily available standards for diagnosis. Moore et al used a ligase chain reaction assay as a standard for comparison in diagnosis of C trachomatis cervicitis but used culture as the standard for gonorrhea infection. They reported that the EGS was only useful as a single presumptive diagnostic test for gonorrhea or chlamydia, in settings with prevalence of infection of 9% or more [33]. Although prevalence of C trachomatis and/or N gonorrhoeae infection in our study was relatively high (11.4%), the sensitivity and specificity of the Gram stain were not high enough to justify its use.

More recent studies argue against use of the genital Gram stain as a screening tool. In a retrospective analysis of women visiting Seattle STD clinics, Marrazzo et al revealed that EGS, in the absence of mucopurulent cervicitis, was only 26% sensitive (too low to suggest routine use) with a PPV of 21% (too low to guide therapy) [34]. Our evidence against using the Gram stain in isolation may apply to screening efforts; conversely, EM physicians can invoke results of screening studies to guide management in the ED. Although screening for STDs does not occur in the ED per se, the diagnostic tests used in an ED share with Screening tests a common goal of being reliable and inexpensive.

The genital Gram stain has not been studied extensively in an ED setting, where it is often used to guide therapy of suspected STD. A more recent study includes the vaginal Gram stain as part of a clinical model proposed to identify cervical infection and to direct Empiric treatment [35]. Our data indicate that the Gram stain does not exhibit sufficient sensitivity or specificity to warrant its use in the ED and may be obsolete in the era of the DNA probe. If use of Gram stain to diagnose common STDs can be eliminated, costs would be reduced. Moreover, EM physicians would save time, a known commodity in the ED.

As diagnostic methods improve, noninvasive nucleic acid amplification tests of Urine samples may supplant DNA probes as the preferred mode of detecting STDs. In a systematic review of 3 types of urine-based nucleic acid amplification tests (polymerase chain reaction, transcription- mediated amplification, and strand displacement amplifica- tion assays), Cook et al [36] determined that sensitivities and specificities of urine-based tests for C trachomatis were nearly identical to those based on cervical and urethral samples, and found that, in testing for N gonorrhoeae, only the sensitivity of the polymerase chain reaction assay in women was too low to recommend its routine use. noninvasive testing may be of particular benefit males reluctant to undergo urethral swabs and may have allowed for a higher inclusion rate among males in our study. Because of the apparent trend toward noninvasive testing, further study is warranted.

This study had several limitations. One weakness is the low inclusion of males (9% of analyzed visits). Our conclusions may not be relevant in populations with different disease prevalence or demographic characteristics, such as female sex workers. Our study also did not assess the utility of the Gram stain in combination with clinical factors, so our conclusions may not apply to clinical algorithms that include genital Gram stains. Finally, the use of Becton Dickinson ProbTec ET DNA amplification assays as the gold standard was chosen as our gold standard. Although these tests have been shown to be highly sensitivity and specificity for C trachomatis and N gonorrhoeae, and are regularly available in the standard clinical setting, other laboratory modalities may have incurred different findings.

Conclusions

In conclusion, the Gram stain, in isolation, does not sufficiently detect Chlamydia Trachomatis and Neisseria gonorrhea infection in the ED. In our study, genital Gram stains did not improve the diagnosis of T vaginalis, and Gram stains frequently demonstrated clue cells, indicative of G vaginalis.

References

1. American Social Health Association. Sexually transmitted diseases in America: how many cases and at what cost? Research Triangle Park (NC): American Social Health Association; 1998.
2. Weinstock H, Berman S, Cates W. Sexually transmitted diseases among American youth: incidence and prevalence estimates, 2000. Perspect Sex Reprod Health 2004;36(1):6-10.
3. Koutsky L. Epidemiology of genital human papillomavirus infection. Am J Med 1997;102(5A):3-8.
4. Curtis KM, Hillis SD, Kieke Jr BA, Brett KM, et al. Visits to emergency departments for gynecologic disorders in the United States, 1992-1994. Obstet Gynecol 1998;91(6):1007-12.
5. Yeh JM, Hook 3rd EW, Goldie SJ. A refined estimate of the average lifetime cost of pelvic inflammatory disease. Sex Transm Dis 2003;30 (5):369-78.
6. Ness RB, Randall H, Richter HE, et al. Condom use and the risk of recurrent pelvic inflammatory disease, chronic pelvic pain, or infertility following an episode of pelvic inflammatory disease. Am J Public Health 2004;94:1327-9.
7. Institute of Medicine. In: Eng TR, Butler WT, editors. The hidden epidemic-confronting sexually transmitted disease. Washington (DC): National Academy Press; 1997. p. 330-56.
8. Magee CM, Rodenheaver G, Edgerton MT, Edlich RF. A more reliable Gram staining technique for diagnosis of surgical infections. Am J Surg 1975;130:341-6.
9. Kruczak-Filipov P, Shively RG. Gram stain procedures. In: Isenberg HD, editor. Clinical microbiology procedures handbook. Washington (DC): American Society for Microbiology; 1994. p. 1.5.1-1.5.18.
10. Van Der Pol B, Ferrero DV, Buck-Barrington L, et al. Multicenter evaluation of the BDProbeTec ET System for detection of Chlamydia trachomatis and Neisseria gonorrhoeae in urine specimens, female endocervical swabs, and male urethral swabs. J Clin Microbiol 2001; 39:1008-16.
11. Brunham RC, Paavonen J, Stevens CE, et al. Mucopurulent cervicitis

-the ignored counterpart in women of urethritis in men. N Engl J Med 1984;311(1):1-6.

1. Katz BP, Caine VA, Jones RB. Diagnosis of mucopurulent cervicitis among women at risk for Chlamydia trachomatis infection. Sex Transm Dis 1989;16(2):103-6.
2. Moscicki B, Shafer MA, Millstein SG, et al. The use and limitations of endocervical Gram stains and mucopurulent cervicitis as predictors for Chlamydia trachomatis in female adolescents. Am J Obstet Gynecol 1987;157(1):65-71.
3. Wald ER. Gonnorrhea: diagnosis by gram stain in the female adolescent. Am J Dis Child 1977;131(10):1094-6.
4. Lossick JG, Smeltzer MP, Curran JW. The value of the cervical gram stain in the diagnosis and treatment of gonorrhea in women in a venereal disease clinic. Sex Transm Dis 1982;9(3): 124-7.
5. Manis Jr RD, Harris B, Geiseler PJ. Evaluation of Gonozyme, an enzyme immunoassay for the rapid diagnosis of gonorrhea. J Clin Microbiol 1984;20(4):742-6.
6. Ison CA, McLean K, Gedney J, Easmon CS, et al. Evaluation of a direct immunofluorescence test for diagnosing gonorrhea. J Clin Pathol 1985;38(10):1142-5.
7. Deceuninck G, Asamoah-Adu C, Khonde N, et al. Improvement of clinical algorithms for the diagnosis of Neisseria gonorrhoeae and Chlamydia trachomatis by the use of Gram-stained smears among female sex workers in Accra, Ghana. Sex Transm Dis 2000;27(7): 401-10.
8. Ndoye I, Mboup S, De Schryver A, et al. Diagnosis of sexually transmitted infections in female prostitutes in Dakar, Senegal. Sex Transm Infect 1998;74:S112-7.
9. Wi T, Mesola V, Manalastas R, et al. Syndromic approach to detection of gonococcal and chlamydial infections among female sex workers in two Philippine cities. Sex Transm Infect 1998;74: S118-22.
10. Diallo MO, Ghys PD, Vuylsteke B, et al. Evaluation of simple diagnostic algorithms for Neisseria gonorrhoeae and Chlamydia trachomatis cervical infections in female sex workers in Abidjan, Cote d’Ivoire. Sex Transm Infect 1998;74:S106-11.
11. Davies SC, Otto B, Partohudoyo S, et al. sexually transmitted infections among female sex workers in Kupang, Indonesia: searching for a screening algorithm to detect cervical gonococcal and chlamydial infections. Sex Transm Dis 2003;30(9):671-9.
12. Knud-Hansen CR, Dallabetta GA, Reichart C, et al. Surrogate methods to diagnose gonococcal and chlamydial cervicitis: comparison of Leukocyte esterase dipstick, endocervical gram stain, and culture. Sex Transm Dis 1991;18(4):211-6.
13. Sellors J, Howard M, Pickard L, Jang D, et al. Chlamydial cervicitis: testing the practice guidelines for presumptive diagnosis. Can Med Assoc J 1998;158(1):41-6.
14. Myziuk L, Romanowski B, Brown M. Endocervical Gram stain smears and their usefulness in the diagnosis of Chlamydia trachomatis. Sex Transm Infect 2001;77(2):103-6.
15. Vuylsteke B, Laga M, Alary M, et al. Clinical algorithms for the screening of women for gonococcal and chlamydial infection: evaluation of pregnant women and prostitutes in Zaire. Clin Infect Dis 1993;17:82-8.
16. Alary M, Baganizi E, Guedeme A, et al. Evaluation of clinical algorithms for the diagnosis of gonococcal and chlamydial infections among men with urethral discharge or dysuria and women with vaginal discharge in Benin. Sex Transm Infect 1998;74:S44-9.
17. Daly CC, Wangel AM, Hoffman IF, et al. Validation of the WHO diagnostic algorithm and development of an alternative scoring system for the management of women presenting with vaginal discharge in Malawi. Sex Transm Infect 1998;74:S50-8.
18. Mayaud P, Uledi E, Cornelissen J, et al. Risk scores to detect cervical infections in urban antenatal clinic attenders in Mwanza, Tanzania. Sex Transm Infect 1998;74:S139-46.
19. Ryan CA, Courtois BN, Hawes SE, et al. Risk assessment, symptoms, and signs as predictors of vulvovaginal and cervical infections in an urban US STD clinic: implications for use of STD algorithms. Sex Transm Infect 1998;74:S59-76.
20. Sanchez SE, Koutsky LA, Sanchez J, et al. Rapid and inexpensive approaches to managing abnormal vaginal discharge or lower abdominal pain: an evaluation in women attending gynaecology and family planning clinics in Peru. Sex Transm Infect 1998;74:S85-94.
21. Tyndall MW, Kidula N, Sande J, et al. Predicting Neisseria gonorrhoeae and Chlamydia trachomatis infection using risk scores, physical examination, microscopy, and leukocyte esterase Urine dipsticks among asymptomatic women attending a family planning clinic in Kenya. Sex Transm Dis 1999;26:476-82.
22. Moore SG, Miller WC, Hoffman IF, et al. Clinical utility of measuring white blood cells on vaginal wet mount and endocervical gram stain for the prediction of chlamydial and gonococcal infections. Sex Transm Dis 2000;27:530-8.
23. Marrazzo JM, Handsfield HH, Whittington WL. Predicting chlamydial and gonococcal cervical infection: implications for management of cervicitis. Obstet Gynecol 2002;100(3):579-84.
24. Reed JL, Mahabee-Gittens EM, Huppert JS. A decision rule to identify adolescent females with cervical infections. J Womens Health (Larchmt) 2007;16(2):272-80.
25. Cook RL, Hutchison SL, Ostergaard L, et al. Systematic review: noninvasive testing for Chlamydia trachomatis and Neisseria gonor- rhoeae. Ann Intern Med 2005;142(11):914-25.