Prevalence and predictors of asymptomatic Chlamydia trachomatis and Neisseria gonorrhoeae in a Ugandan population most at risk of HIV transmission.

Abstract

The aim of this study was to establish the prevalence of asymptomatic Chlamydia trachomatis (CT) and Neisseria gonorrhoeae (NG) in key populations at increased risk of transmission of HIV. Additionally, we aimed to identify the associated risk factors for asymptomatic sexually transmitted infections (STIs) and evaluate the acceptability of self-collected samples. Asymptomatic people living with HIV (PLHIV) in the following categories were offered testing: discordant couples; young adults; pregnant patients and those attending the ‘most-at-risk-population’ clinic. Patients provided first-pass urine, self-collected vaginal swabs or both to test for NG and CT by polymerase chain reaction using BD ProbeTec™. Patients also completed an acceptability questionnaire, including the negative partner of an HIV positive participant. Three hundred and sixty-three PLHIV had an STI screen. Asymptomatic STIs were only diagnosed in women (prevalence 5.7%), overall prevalence 3.9% (n=14). Factors independently associated with an STI in women were: being under 25 years (OR 9.63 95% CI 1.56 – 59.5) and having more than one sexual partner (OR 8.06 95% CI 1.07 −60.6). Four hundred and seven completed the acceptability questionnaire. More than 95% of patients found self-sampling easy and comfortable and 83.8% would believe the results. Women significantly preferred the option of self-sampling, 56.9% versus 29.3% of men (p< 0.001). Acceptability of self- sampling was high. Young women with or at risk of HIV are an important target for STI testing regardless of symptoms. There is need for diagnostic tests that are inexpensive, rapid and accurate especially in resource limited settings.

INTRODUCTION

Chlamydia trachomatis (CT) and Neisseria gonorrhoeae (NG) are common treatable sexually transmitted infections (STIs). The World Health Organisation (WHO) estimates that each year there are approximately 376 million new treatable STIs in adults aged 15 to 49 years: 127 million cases of CT, 87 million cases of NG, 156 million cases of trichomoniasis (TV) and 6 million cases of syphilis [1]. The majority of patients with these infections remain asymptomatic; 70–80% of women and up to 50% of men with CT are asymptomatic and up to 50% of women and 10% of men with NG are asymptomatic [2, 3].

Current management of STIs in Uganda is based on World Health Organisation (WHO) guidance,[4] using the syndromic approach. This employs identification of a consistent group of symptoms and easily recognized signs with treatment given for the most likely infections, without the use of laboratory tests. The approach is most effective for the management of urethral discharge in men (87–99%) and genital ulcer disease in men and women (68–98%)[5]. In contrast, the syndromic approach in women with vaginal discharge has poor sensitivity (cervical infection 27.4 – 90.1%; vaginal infections 55%−90%). However there is superior specificity for vaginal infections 41.4 – 99.8% than with cervical infection 35.3 – 84.9%. The wide variation is dependent on the algorithms for syndromic treatment utilised, as well as type of infection [6]. The major limitation of syndromic management is that asymptomatic patients will remain undiagnosed and untreated which could lead to long-term sequelae like pelvic inflammatory disease, and infertility [7]

There is limited data on the prevalence of asymptomatic STIs in high-risk HIV-positive Ugandan adults, as accurate molecular diagnostic tests are not routinely available. A study in rural Uganda (Rakai, 1998) reported a prevalence of non-ulcerative STIs: men NG 0.9%, CT 2.1%; women NG 1.5%, CT 2.4%, TV 23.8%, using ligase chain reaction (LCR) on urine, and demonstrated that if management was based only on symptoms, 80% of men and 72% of women with NG and CT would have been missed [8]. Whilst effective ART with virological suppression prevents HIV transmission regardless of STIs, we have not yet reached the target of 90% virological suppression in Africa [9], and only around 50–70% of young people are achieving viral suppression [10,11,12]. Therefore, STIs are important as they still increase chance of HIV transmission [13].

There are specific groups who have been identified as key populations and vulnerable groups by WHO as they are more likely to be exposed or transmit HIV so are key in controlling the epidemic [14].

The primary aim of this study was first to establish the prevalence of asymptomatic CT and NG in the following populations: HIV-discordant couples, young people with HIV, pregnant HIV-positive women; and the most-at-risk populations (MARPs); Secondly, we aimed to determine the risk factors for asymptomatic STIs and finally to assess if self-collected samples were acceptable in these populations.

METHODS

Study population

This was a prospective cross sectional study conducted at the Adult Infectious Disease Clinic (AIDC) based at the Infectious Disease Institute (IDI), which cares for approximately 8,000 people living with HIV (PLHIV), in Kampala, Uganda. Anyone belonging to a Ministry of Health key population or most at risk population attending the clinic was included in the study population. At the AIDC key populations include those attending separate clinic sessions to support: HIV discordant couples; young people (aged 15–24 years) transitioning from paediatric to adult HIV care (transition clinic); a sexual and reproductive health (SRH) clinic offering Prevention of Mother to Child Transmission (PMTCT) and other reproductive health services to HIV positive women. MARPs are seen by a dedicated physician, and defined as per the Ministry of Health (MOH) [15] definitions: fisher folk, men who have sex with men, transactional sex workers and their partners, mobile men with money (e.g. boda-boda drivers), bar workers and long distance lorry drivers. AIDC primarily supports PLHIV, however some HIV negative individuals who were partners of the PLHIV in a discordant couple are provided with pre-exposure prophylaxis (PrEP) and sexual health services.

Inclusion criteria

Participants were included if they signed an informed consent form, were aged 14 years and older, could undergo requested testing, and had no STI associated symptoms. Male and female attendees determined to be discordant by the clinic, and with evidence of discordant HIV test results in the last 6 months. During the days the study team were at AIDC, all consecutive willing patients from the clinics were offered to participate during the specified time period.

Study procedures

Study staff obtained written informed consent from all participants prior to any study procedures. Sexual history was obtained using a standardized form for males and females. Patients were not examined. Non-pregnant women provided a self-collected vaginal swab and first-pass urine. Self-collected vaginal swabs were chosen as there is no statistical difference in sensitivity of these compared to physician collected vaginal swabs in asymptomatic women (87% vs 84.8% for CT 95.7% vs 95.7% for GC) [16]. Men and pregnant women provided approximately 30mls of first pass urine for molecular testing. All patients were given a questionnaire (supplementary material)which collected demographic data (sex and age), symptom status, previous STI, ease and comfort of self-sampling, whether they believed the results and preferable method in the future (self-collected, clinician taken, doesn’t matter or don’t know).

Laboratory procedures

Nucleic acid amplification testing

Testing was done at the Makerere University – Johns Hopkins University core laboratory which is a College of American Pathology (CAP) certified laboratory and Division of Acquired Immunodeficiency Syndrome (DAID’s) National Institute of Health (NIH) quality certified testing facility. Urine and vaginal samples were tested for NG and CT by PCR assay (BD Probetec™ ET CT/NG test; BD Diagnostics, Sparks, MD, USA). All tests were carried out according to the manufacturer’s instructions, which included one positive and one negative control in each assay run. NG and CT were confirmed if nucleic acid amplification test was positive for NG and CT respectively. Treatment for CT was azithromycin 1gram as a single dose and for NG, cefixime 400mg as a single dose. The same treatment was offered to all partners that attended. All treatment was offered free of charge.

Statistical analysis

Our analysis was designed to estimate the STI prevalence, to determine predictors for an STI and assess acceptability of self-collected samples. We compared sociodemographic (age, education level, employment status, marital status), sexual behaviour (having more than one sexual partner, condom use with the main partner), and HIV-related differences by gender using chi-square test for categorical variables and Wilcoxon rank-sum test for not normally distributed variables. Similarly, we used logistic regression to assess the adjusted correlates between STI test results and other variables, adjusting for underlying differences in demographic and HIV treatment history variables for both the crude and adjusted odds ratios. We reported the corresponding Odds Ratios (ORs) and 95% Confidence Interval (CI). Age and CD4 count were modelled as continuous variables while all the rest were categorical. Only variables with a probability value less than 0.2 at the unadjusted level and those variables deemed theoretically important to adjust for potential confounders were considered for the adjusted level. All analyses were conducted using STATA 14.2 (StataCorp, College Station, Texas).

Ethical statement

Written approval was obtained to conduct the study from Makerere University School of Public Health Research and Ethics Committee (IRB no HREC 275), Uganda National Council for Science and Technology (HS 1743), and Johns Hopkins University School of Medicine Institutional Review Board (IRB00061162). Uganda National Council of Science and Technology defines “mature minors” as (individuals 14–17 years) who may have a sexually transmitted infection may independently provide informed consent to participate in research [17].

RESULTS

During the study period, 18th March 2015 to the 1^st August 2015, there were 1,271 patients who attended the key and MARPs clinic at AIDC. All patients attending on days that a study nurse or doctor was available during this period were approached for screening. A total of 407 patients were screened for CT and NG., 363 were HIV positive and included in the analysis.

The majority (67.8%; 246/363) were women. Men were older, median age 39 for males and 26 for females (p <0.001). The majority 86.7% were employed; 60% reported having an education above secondary level. Males were more likely to report having more than one sexual partner compared to females 17.9% vs 4.5% (p=<0.001). Females were more likely to report no condom use with their main partner 39.4% compared to 13.7% (p=<0.001). The median CD4 cell count was 505 cells/μL (IQR 350 – 670) with 92% on ART. Of those patients with a viral load result (n=284), 90% of patients had achieved viral suppression (HIV RNA <75 copies/mL).

Overall prevalence of an asymptomatic STIs was 3.9% (n=14), and individual prevalence for NG 2.2% (n=9) and CT 1.7% (n=7). In addition two HIV negative partners of PLHIV in the clinic had an STI both female one had CT and the other NG. The overall prevalence of asymptomatic STIs in HIV positive patients in the sample was 5.7% in women, excluding men. We therefore restricted the analysis on predictors for STIs in women. The following factors were independently associated with asymptomatic STIs and were included in the final logistic regression model; age under 25 year (OR 9.63 95% CI 1.56 – 59.5) and having more than one sexual partner (OR 8.06 95% CI 1.07 – 60.6), Table 2. Nineteen percent (n=3) women with asymptomatic chlamydia were newly diagnosed with HIV, all under the age of 25 years. All women were notified of their results by phone and all attended for treatment, in addition 10 (58.8%) partners were reported to have received treatment.

Table 2:

Factors associated with having an asymptomatic STI, restricted to HIV positive women

	Unadjusted model			Adjusted modely
Characteristics	ORƗ	95 CI%Ɨ	P-valueƗ	ORƗƗƗ	95 CI%ƗƗƗ	P-valueƗƗƗ
Age
25 years and above	1.00	·	·	1.00	·	·
Less than 25 years	4.28	1.34–13.65	0.014	9.63	1.56–59.5	0.015
Education level
None	1.00	·	·	1.00	·	·
Primary	1.09	0.12–9.53	0.939	0.44	0.04–5.1	0.509
Secondary	0.87	0.09–7.58	0.898	0.18	0.01–2.3	0.184
Tertiary	0.37	0.02–6.29	0.491	0.13	0.01–2.9	0.195
Employment status
Employed	1.00	·	·	1.00	·	·
Not employed	2.59	0.85–7.89	0.093	1.27	0.35–4.6	0.713
Marital status
Married a	1.00	·	·	1.00	·	·
Not married b	2.13	0.77–5.90	0.146	1.26	0.36–4.3	0.716
Have more than one sexual partner
No	1.00	·	·	1.00	·	·
Yes	3.91	0.77–19.8	0.100	8.06	1.07–60.6	0.043
Condom use with the main partner
Yes	1.00	·	·	1.00	·	·
No	0.98	0.34–2.77	0.966	0.81	0.24–2.7	0.740
CD4 count, Median(IQR)	0.98	0.99–1.01	0.346	0.99	0.99–1.0	0.116
Viral load suppression, n(%)
Not suppressed	1.00	·	·	1.00	·	·
Suppressed	0.56	0.11–2.79	0.477	1.40	0.22–9.1	0.722
Missing c	0.85	0.14–4.96	0.853	0.63	0.05–8.5	0.725
ART status
On ART	1.00	·	·	1.00	·	·
ART Naïve	3.06	0.78–11.9	0.107	5.43	0.44–66.6	0.186

Note: Only females considered; Sex and group were excluded variables from the bivariate and multivariable analysis; Ɨ Unadjusted model; ƗƗ; y Adjusted model with HIV related factors;

^aMarried participants included cohabiting couples;

^bNot married participants included single, separated, divorced and widowed;

^c56 participants did not have viral load results at the time of the study

Four hundred and seven patients completed the questionnaire on acceptability of self-collected vaginal swabs and first-catch urine. 393 (96.9%) reported that it was easy to perform and 389 (95.6%) found it comfortable; 341 (83.8%) believed that the result they would receive would be accurate. When asked about a preferred method if future testing was required 193 (47.2%) preferred self-collected samples, 95 (23.3%) preferred a doctor or nurse to take the samples and 119 (29.2%) had no preference. Females significantly preferred self-collected sampling 56.9% versus 29.3% of men (p< 0.001). A higher proportion 55.8% of young adults aged between 17–24 years reported a preference for self-collected samples compared to their older counterparts aged 25–39 years (47.2%) and those aged 40 years and above (35%), the difference in proportions was statistically significant (p=0.002). There was no difference in ease and comfort in performing the test or believing the results by sex or age.

DISCUSSION

This study enrolled attendees at an urban clinic serving PLHIV and their partners. The study population was taken from those in key and MARPs populations; young people were over sampled as this is a group known to be at high risk of STIs. Prevalence of an asymptomatic STI in women was 5.7% (prevalence of NG was 3.2% and CT was 2.5%); there were no men with asymptomatic infections. The study found that under the age of 25 with more than one sexual partner were more likely to have an asymptomatic STI.

The overall prevalence of NG was higher than CT in our cohort. This is similar to other studies done in key and MARPs HIV-negative populations in Uganda, in patients with and without symptoms: adolescents in an urban setting [18] and urban women involved in high-risk sexual behaviour [19]. Kakaire et al. [20] reviewed asymptomatic STIs in HIV-positive females opting for intrauterine contraception in Uganda. The overall prevalence was 11% with individual prevalence’s of TV, NG and CT 5.9%, 5.4% and 0.9%, respectively. In South Africa Lewis et al [21] found a prevalence of TV 7.6%, Mycoplasma genitalium (MG) 6.1%; NG 5.4% and CT 2.1% amongst asymptomatic men and women attending for routine HIV care.

Our data confirm that of other studies in sub-Saharan Africa; the highest rates of undiagnosed STI’s are among women. Singa et al. [22] screened 1,661 adults in Kenya and found HIV-positive males had significantly lower prevalence of STIs in both symptomatic and asymptomatic participants. Among HIV-infected patients in South Africa, Lewis at al. [21] found that women had a significantly higher risk of an asymptomatic STI.

We found that females under 25 years were 10 times more likely than women over 25 years to have asymptomatic CT or NG. Worldwide the highest rates of STIs are in 20–24 year olds (young adults) followed by 15–19 year olds (adolescents) [23]. The immaturity of the adolescent reproductive sexual tract has been cited as a reason for increased susceptibility to STIs. Interestingly there were no infections in males under 25 years of age, though this group was under-represented as the majority of young men attending the transition clinic were not sexually active. Of concern, 19% of women had a concurrent new HIV diagnosis. In sub-Saharan Africa, women comprise 56% of new HIV infections among adults (15 and older); and the proportion is higher among young women aged 15–24, who make up 67% of new infections among young people [24]. In Uganda, HIV prevalence is almost four times higher among females than males aged 15 to 19 and 20 to 24 [25] and 570 young women aged 15 – 24 years of age get infected with HIV every week [26]. Untreated STIs and earlier sexual debut are likely to be contributing to this disproportionately high rate of new HIV acquisition among young women.

STIs can increase the risk of HIV acquisition three-fold [13]. Research suggests that STIs increase the susceptibility of the HIV-negative person to acquiring HIV, and increases HIV transmission from HIV-positive people. Biological reasons for this: direct mucosal disruption; recruitment of HIV target cells to the genital tract and increased HIV viral load in plasma and genital secretions [27]. Treatment as prevention has a role in reducing HIV transmission in discordant couples but will not reduce all HIV acquisition as in some settings 20–30% of HIV incidence is from an external partner [28].

There are very few sub-Saharan studies assessing the acceptability of self-collected vaginal swabs and first-catch urine for the diagnosis of chlamydia and gonorrhoea [29]. The results were encouraging with over 95% of patients finding it easy and comfortable and 83% of patients would believe the result. However, despite this, less than 50% of women and those under 40 preferred self-collected sampling. This could be due to a high level of faith in medical professionals, or that the concept of an asymptomatic infection is understood in Uganda in the context of HIV, but not other infections. From our study there is a suggestion men prefer an examination with a urethral swab.

The cornerstone of management for STIs is syndromic in resource-constrained settings [4]. This management strategy has the advantage of being cheap, rapid and allows for immediate treatment. The major drawback is that asymptomatic patients will not be treated. Concerns around patients returning for treatment was not demonstrated in our study. All patients provided a mobile contact number and were able to be contacted and all were offered and attended for treatment. Asymptomatic patients risk being inadequately treated with antibiotics prescribed for other reasons. This has the potential to create a reservoir of resistance fuelling the global crisis of antimicrobial resistance particularly for NG. In addition, the syndromic approach will over treat a large number of women as this approach has poor sensitivity and positive predictive value for detecting cervical infections for women presenting with vaginal discharge [5]. There is a need for cheap point-of-care tests utilising nucleic acid amplification technology that can be used in a variety of anatomical sites and provide rapid results.

The strength of our study is that we sampled multiple high-risk groups with HIV and at high risk of HIV acquisition. However, this led to a limitation, which is that the population was heterogeneous. Our study had a number of other limitations. Firstly, we did not test for trichomonas which studies have revealed is the most common asymptomatic STI in women [19, 21, 22]. Secondly we relied on self-reported history of lack of symptoms, number of partners and condom use which would potentially bring in social desirability bias. Thirdly, males under the age of 25 and females in MARPs were under-represented, though this was a reflection of the population in the clinic.

In conclusion, our study demonstrated a significant prevalence of asymptomatic CT and NG in young women. These women would have not been identified or offered treatment if the syndromic approach had been utilised. In countries where resources are limited, there is a need for targeted testing for STIs of at risk asymptomatic patients particularly young adult women. Diagnostic tests that are inexpensive and provide quick and accurate results will be needed to move beyond syndromic patient management to algorithms based on diagnostic certainty.

Table 1:

Background characteristics of the PLHIV enrolled into the study stratified by sex including only HIV positive patients

Characteristics N=363	Males (n=117)	Females (n=246)	p-value
Group, n (%)
Discordant	43(36.7)	36(14.6)	<0.001
Young adults	27(23.1)	94(38.2)
MARPS	47(40.2)	23(9.34)
Pregnant	·	93(37.8)
Age, Median(IQR)	39(28–47)	26(23–32)	<0.001
Education level, n (%)
None	10(8.6)	12(4.9)	0.578
Primary	39(33.3)	84(34.1)
Secondary	49(41.9)	111(45.1)
Tertiary	19(16.2)	39(15.9)
Employment status, n (%)
Employed	111(94.9)	204(82.9)	0.002
Not employed	6(94.9)	42(17.1)
Marital status, n (%)
Married a	82(70.1)	143(58.1)	0.028
Not married b	35(29.9)	103(41.9)
Have more than one sexual partner, n (%)
Yes	21(17.9)	11(4.5)	<0.001
No	96(82.1)	235(95.5)
Condom use with the main partner, n (%)
Yes	101(86.3)	149(60.6)	<0.001
No	16(13.7)	97(39.4)
CD4 count, Median(IQR)	430(317–582)	555(378–711)	<0.001
Viral load suppression, n(%)c
Not suppressed	4(3.4)	24(9.8)	0.064
Suppressed	90(76.9)	166(67.5)
Missing	23(19.7)	56(22.7)
ART status
On ART	111(94.9)	224(91.1)	0.203
ART Naïve	6(5.1)	22(8.9)
STI test result
Negative	117(100.0)	232(94.3)	0.004
Positive	0(0.0)	14(5.7)

^aMarried participants included cohabiting couples

^bNot married participants included single, separated, divorced and widowed

^c79 participants did not have viral load results at the time of the study

Abbreviations: ART: antiretroviral therapy; IQR: interquartile range; MARPs: most at risk populations; STI: sexually transmitted infections