Accurate diagnosis and treatment of sexually transmitted infections (STIs) prevents health complications such as pelvic inflammatory disease, ectopic pregnancy, and infertility.1 HIV-infected populations are at higher risk for acquiring STIs.2 The HIV epidemic has disproportionally affected resource-limited settings, and new infections continue to outpace treatment access,3 making effective STI screening and treatment essential to comprehensive HIV care and prevention programs. However, in resource-limited settings, STI management has proven particularly challenging given limited funds, personnel, laboratory capacity, and space. Many HIV care clinics are not able to offer routine screening for STIs.4
The World Health Organization (WHO) algorithm for syndromic management of genital infections provides a strategy to help clinics in resource-limited settings address patient complaints of abnormal genital symptoms without relying on laboratory diagnostics and has been incorporated into many national guidelines.5–7 However, syndromic management is limited because it depends on a patient presenting with a complaint. Effectiveness depends on the patient’s comfort and willingness to address potentially sensitive subjects with their health care provider, which can be especially difficult in cultures where such discussion, especially between sexes, is considered taboo. Furthermore, women may not focus on reproductive health complaints when asked general or HIV-related screening questions and may only discuss reproductive health complaints if a clinician asks a more directed question. These limitations are complicated by the fact that many STIs are asymptomatic; there are few clinical predictors to identify who is at risk8 and relaying solely on symptoms such as vaginal discharge still underdiagnoses the actual number infected.9 Despite this limitation, we believe that eliciting a more extensive history from patients could still help diagnose more disease. We hypothesized that the inclusion of focused questions, along with a basic physical examination, would increase the sensitivity and specificity of syndromic management for nonulcerative STIs among HIV-infected women in an HIV clinic in Kenya.
This cross-sectional study was carried out among a cohort of HIV-infected women enrolled at Family AIDS Care and Education Services clinic10 in Kisumu, Kenya, between October 2010 and April 2011. Approval was received from the Committee for Human Research at the University of California, San Francisco, and the Kenya Medical Research Institute Ethical Review Committee, and women provided written informed consent before participation in the study. Women were recruited from the cervical cancer screening program that included nonpregnant women older than 23 years. Clinical and demographic characteristics were collected for each participant (age, latest CD4+ count, WHO stage, antiretroviral use, contraception use, and number of sexual partners).
During their routine clinic visit, participants’ HIV primary care provider asked them to report any health complaints with the following question: do you have any complaints today? At the end of their regular clinic visit, all participants underwent screening for abnormal vaginal discharge in 4 steps per our study protocol. First, participants were asked if they had any general vaginal symptoms (brief prompt). Next, all participants were asked specifically about complaints of genital pruritis and vaginal odor and discharge (detailed prompt). Because women who spontaneously reported vaginal symptoms had essentially already answered the brief and detailed prompt, we categorized them as positive. Participants then underwent a physical examination with a speculum examination, which included specimen collection for STI testing and assessment of cervical friability, mucopurulent discharge, erythema, or abnormal-appearing vaginal discharge (Fig. 1).
In this proof-of-concept study, we chose to focus only on nonulcerative cervical STIs, testing for Neisseria gonorrhoeae, Chlamydia trachomatis, and Trichomonas vaginalis (TV). Women were excluded if they had ulcerative or condylomatous genital lesions or evidence of candidiasis on speculum examination. Cervical swabs were evaluated for N. gonorrhoeae and C. trachomatis via nucleic acid amplification test (Roche Amplicor TM, Indianapolis, IN). Participants were tested for TV by wet mount (57.7%) and InPouch (Biomed Diagnostics, White City, OR) (42.3%) from swabs collected from the posterior vaginal fornix.
Using laboratory confirmation as the gold standard, sensitivity, specificity, and positive predictive value were calculated for each method of ascertainment: unprompted, brief prompt, detailed prompt, and gynecologic examination. Statistical analysis was performed using STATA 11 (StataCorp LP, College Station, TX).
Demographic and clinical information for the 334 women enrolled in the study are reported in Table 1. The overall prevalence of a nonulcerative STI was 12.3% (41/334). The prevalence of N. gonorrhoeae was 1.8% (6/334), that of C. trachomatis was 0.9% (3/334), and that of TV was 9.9% (33/334). Only 13 women (3.9%) reported symptoms unprompted, after a general health assessment. When using the clinician-initiated diagnostic strategies, 61 (18.3%) women reported symptoms after a brief prompt and 63 (18.8%) after a detailed prompt. Twenty-three (6.9%) women had a pelvic examination suggestive of cervicovaginal infection (Table 2). Sensitivity, specificity, and positive predictive values for any measured STI were 0%, 95.6%, and 0% based only on self-report. With the use of clinician-initiated diagnostic strategies, sensitivity, specificity, and positive predictive values were 19.5%, 81.9%, and 13.1% for brief prompt; 19.5%, 81.2%, and 12.7% for detailed prompt; and 31.7%, 96.6% and 56.5%, respectively, for the pelvic examination (Fig. 2)
The proportion of STIs detected by brief prompt was significantly different from self-report alone (difference = 0.1951, P < 0.01). There was no significant difference between proportions of STIs detected on brief and detailed prompting, and in fact, the proportions were nearly identical. Thus, further analysis was done with only using results from the brief prompt. The proportion of STIs detected by pelvic examination was also significantly different from results found by unprompted report (difference = 0.244, P < 0.01) and brief prompt (difference = 0.122, P < 0.05).
Of the 13 women self-reporting STI-related symptoms unprompted, none had laboratory-confirmed disease. The false-positive rate was 86.9% (53/61) for brief prompt and 43.4% (10/23) for gynecologic examination. Seventy-eight percent (32/41) of women with positive STI laboratory testing result reported no symptoms at all, and 31.3% (10/32) of these women would have been diagnosed as having an STI on clinical examination. The false-negative rate was 12.7% (41/321) for the unprompted self-report, 12.1% (33/273) for the brief prompt, and 9.0% (38/311) for the gynecologic examination.
Our study showed that reliance on unprompted self-report in response to a general health question resulted in a very poor sensitivity for detecting nonulcerative STIs in HIV-infected women. In our sample, no women who self-reported symptoms had one of the infections we were testing for, and no women with the infections we were looking for reported symptoms without directed questioning. Thus, basing management of STIs in resource-limited settings solely on unprompted patient-initiated report via the syndrome management protocol appears inadequate.
The addition of clinician-initiated STI-specific screening strategies improved the detection of treatable STIs, with no difference observed between the use of brief questions and the use of detailed questions. The addition of a gynecologic examination led to the greatest improvement in sensitivity and specificity. Our experience supports the introduction of a single, simple question to improve the accuracy of screening for STIs in the absence of laboratory testing capability. Where feasible, our findings also support a basic gynecologic examination, including speculum, especially for detection of TV.
False-positive rates were high for all diagnostic methods, which is likely attributable to the overall low prevalence of disease. It is concerning, however, that management based on unprompted patient report of symptoms or the current standard of care in Kenya and much of sub-Saharan Africa had a 100% false-positive rate.11 Many antibiotics used to treat these STIs have low barriers to resistance, and syndromic management is implemented in settings where access to second-line antibiotics is limited, making the judicious use of the available therapies essential.12
One limitation to our study was that we did not have the resources to test for bacterial vaginosis (BV), one of the most common causes of abnormal vaginal discharge worldwide including sub-Saharan Africa.13 Although excluding BV would have no impact on the high proportion of false negatives, it is likely that the sensitivities for all groups would have been increased if we had included BV in our study.
Almost three quarters of the women found to have an abnormal vaginal discharge on examination had a laboratory-diagnosed STI, indicating that physical examination was an effective way of identifying women likely to have an STI. The inclusion of a genitourinary examination in STI screening in a resource-limited setting is not always possible because there is a need for equipment, sterilization capacity, and a private examination area. However, in settings where an examination is feasible, performing one could greatly increase the accuracy of STI screening for asymptomatic women.
Although alternative solutions such as point-of-care diagnostics that do not require special equipment or laboratory capacity are still needed, our study demonstrates that improved clinical ascertainment of STI signs and symptoms in women can be undertaken in the meantime to improve the diagnosis and treatment of STIs in women infected with HIV. In particular, the simple addition of even just 1 clinician-initiated, focused question and/or a speculum examination greatly increased the ability to accurately diagnose STIs in this vulnerable population of HIV-infected women.
1. Soper DE. Pelvic inflammatory disease. Obstet Gynecol 2010; 116: 419–428.
2. Rottingen JA, William Cameron D, Garnett GP. A systematic review of the epidemiologic interactions between classic sexually transmitted diseases and HIV. Sex Transm Dis 2001; 28: 579–597.
4. Spaar A, Graber C, Dabis F, et al.. Prioritising prevention strategies for patients in antiretroviral treatment programmes in resource-limited settings. AIDS Care 2010; 22: 775–783.
7. Vishwanath S, Talwar V, Prasad R, et al.. Syndromic management of vaginal discharge among women in a reproductive health clinic in India. Sex Transm Infect 2000; 76: 303–306.
8. Lewis DA, Chirwa TF, Msimang VM, et al.. Urethritis/Cervicitis pathogen prevalence and associated risk factors among asymptomatic HIV-infected patients in South Africa. Sex Transm Dis 2012; 39: 531–536.
9. Milsana K, Maicke N, Werner L, et al.. Symptomatic vaginal discharge is a poor predictor of sexually transmitted infections and genital tract inflammation in high-risk women in South Africa. J Infect Dis 2012; 206: 6–14.
10. Kulzer JL, Penner JA, Marima R, et al.. Family model of HIV care and treatment: A retrospective study in Kenya. J Int AIDS Soc 2012; 15: 8.
11. Liambila W, Awuor C, Natecho A, et al. Revitalizing the national STI/RTI control activities in Kenya. Report of high level Consultative Meeting. NASCOP; 2009. Available at: www.nacc.or.ke/attachments/article/102/STI.pdf
. Accessed September 1, 2011.
12. Chisholm SA, Neal TJ, Alawattegama AB, et al.. Emergence of high-level azithromycin resistance in Neisseria gonorrhoeae
in England and Wales. J Antimicrob Chemother 2009; 64: 353–358.
13. Sewankambo N, Gray RH, Wawer MJ, et al.. HIV-1 infection associated with abnormal vaginal flora morphology and bacterial vaginosis. Lancet 1997; 350: 546–550.