Article Text

Download PDFPDF

Original article
A cross-sectional analysis of Trichomonas vaginalis infection among heterosexual HIV-1 serodiscordant African couples
  1. Aaron F Bochner1,2,
  2. Jared M Baeten1,2,3,
  3. Alison S Rustagi4,
  4. Edith Nakku-Joloba5,
  5. Jairam R Lingappa2,3,6,
  6. Nelly R Mugo2,7,
  7. Elizabeth A Bukusi2,8,
  8. Saidi Kapiga9,
  9. Sinead Delany-Moretlwe10,
  10. Connie Celum1,2,3,
  11. Ruanne V Barnabas1,2,3
  12. for the Partners in Prevention HSV/HIV Transmission Study and Partners PrEP Study Teams
    1. 1 Department of Epidemiology, University of Washington, Seattle, Washington, USA
    2. 2 Department of Global Health, University of Washington, Seattle, Washington, USA
    3. 3 School of Medicine, University of Washington, Seattle, Washington, USA
    4. 4 Department of Medicine, University of California, San Francisco, California, USA
    5. 5 School of Public Health, Makerere University, Kampala, Uganda
    6. 6 Department of Pediatrics, University of Washington, Seattle, Washington, USA
    7. 7 Center for Clinical Research, Kenya Medical Research Institute, Nairobi, Kenya
    8. 8 Center for Microbiology Research, Kenya Medical Research Institute, Nairobi, Kenya
    9. 9 Department of Infectious Disease Epidemiology, London School of Hygiene and Tropical Medicine, London, UK
    10. 10 Wits Reproductive Health and HIV Institute, University of Witwatersrand, Johannesburg, South Africa
    1. Correspondence to Dr Aaron Bochner, University of Washington, 325 9th Ave, Box 359932, Seattle, WA 98104, USA; bochner{at}uw.edu

    Abstract

    Objectives Trichomonas vaginalis is the most prevalent curable STI worldwide and has been associated with adverse health outcomes and increased HIV-1 transmission risk. We conducted a cross-sectional analysis among couples to assess how characteristics of both individuals in sexual partnerships are associated with the prevalence of male and female T. vaginalis infection.

    Methods African HIV-1 serodiscordant heterosexual couples were concurrently tested for trichomoniasis at enrolment into two clinical trials. T. vaginalis testing was by nucleic acid amplification or culture methods. Using Poisson regression with robust standard errors, we identified characteristics associated with trichomoniasis.

    Results Among 7531 couples tested for trichomoniasis, 981 (13%) couples contained at least one infected partner. The prevalence was 11% (n=857) among women and 4% (n=319) among men, and most infected individuals did not experience signs or symptoms of T. vaginalis. Exploring concordance of T. vaginalis status within sexual partnerships, we observed that 61% (195/319) of T. vaginalis-positive men and 23% (195/857) of T. vaginalis-positive women had a concurrently infected partner. In multivariable analysis, having a T. vaginalis-positive partner was the strongest predictor of infection for women (relative risk (RR) 4.70, 95% CI 4.10 to 5.38) and men (RR 10.09, 95% CI 7.92 to 12.85). For women, having outside sex partners, gonorrhoea, and intermediate or high Nugent scores for bacterial vaginosis were associated with increased risk of trichomoniasis, whereas age 45 years and above, being married, having children and injectable contraceptive use were associated with reduced trichomoniasis risk. Additionally, women whose male partners were circumcised, had more education or earned income had lower risk of trichomoniasis.

    Conclusions We found that within African HIV-1 serodiscordant heterosexual couples, the prevalence of trichomoniasis was high among partners of T. vaginalis-infected individuals, suggesting that partner services could play an important role identifying additional cases and preventing reinfection. Our results also suggest that male circumcision may reduce the risk of male-to-female T. vaginalis transmission.

    • TRICHOMONAS
    • AFRICA
    • HORMONAL CONTRACEPTION
    • BACTERIAL VAGINOSIS
    • CIRCUMCISION

    Statistics from Altmetric.com

    Introduction

    The protozoan parasite Trichomonas vaginalis is the most prevalent curable sexually transmitted infection (STI) worldwide, with an estimated 187 million infected individuals aged 15–49 years old.1 Trichomoniasis has been associated with adverse outcomes in women, including pelvic inflammatory disease, low birth weight and preterm delivery.2 ,3 In addition, observational data suggest that T. vaginalis may increase women's risk of HIV-1 acquisition by 1.5 to 3-fold.4 For both men and women with HIV-1, T. vaginalis co-infection has been associated with increased genital shedding of HIV-1,4 suggesting that T. vaginalis increases HIV-1 transmission risk.5

    Trichomoniasis symptoms for women can include vaginal discharge, dysuria, itching, vulvar irritation and abdominal pain while men may experience urethritis.6 Though trichomoniasis can be easily treated with a 1 to 7-day course of metronidazole,7 identifying T. vaginalis cases is challenging because the infection is usually asymptomatic.8 Previously, T. vaginalis screening has relied on wet mount or culture diagnostic assays, which are time-consuming and lack sensitivity. Recently developed nucleic acid amplification tests (NAATs) have improved diagnostic sensitivity, especially among men.9 However, the cost and infrastructure requirements for NAAT testing currently precludes widespread clinical use, and affordable point-of-care T. vaginalis diagnostics are needed to support widespread testing, diagnosis and treatment.

    The prevalence of trichomoniasis worldwide is higher among women than men.1 A cause of this disparity is that the average duration of untreated infection is longer for women than men. Though data are limited, the average duration of untreated infection was estimated to be 18 months for women compared with 1.5 months for men.10 The high prevalence and relatively short duration of infection suggests that for T. vaginalis to be sustained in populations, individuals (especially men) are likely to be repeatedly infected by sexual partners. Although ongoing sexual partnerships are a central component of T. vaginalis epidemiology, few recent studies have explored the prevalence and correlates of trichomoniasis within couples.11–13

    Since widespread screening for trichomoniasis is not currently practical in most settings, identifying and targeting interventions towards modifiable risk factors for trichomoniasis may present the best approach to reduce the burden of T. vaginalis and T. vaginalis-associated morbidity. To identify correlates of T. vaginalis infection, we conducted a large cross-sectional analysis among African HIV-1 serodiscordant heterosexual couples. We assessed how characteristics of both members of sexual partnerships were associated with the prevalence of male and female T. vaginalis infection.

    Methods

    Study population

    Data from two prospective cohorts of African heterosexual HIV-1 serodiscordant couples were included in this analysis. The Partners in Prevention HSV/HIV Transmission Study was a randomised, placebo-controlled trial of daily acyclovir as herpes simplex virus type 2 (HSV-2) suppressive therapy to reduce HIV-1 transmission from HSV-2/HIV-1 co-infected individuals to their HIV-1 uninfected partners. Between November 2004 and April 2007 the study enrolled 3408 couples from 14 sites in seven African countries.14 The Partners PrEP Study was a randomised, placebo-controlled trial of pre-exposure prophylaxis (PrEP) to prevent HIV-1 acquisition. Serodiscordant couples were randomised to daily tenofovir disoproxil fumarate, combination tenofovir disoproxil fumarate and emtricitabine, or placebo. Between July 2008 and November 2010, 4747 heterosexual HIV-1 serodiscordant couples were enrolled at nine sites in Kenya and Uganda.15

    Enrolment and exclusion criteria were similar in both parent studies: participants were aged 18 years and above, and eligible couples reported regular sexual intercourse with their study partner in the 3 months prior to enrolment and intention to remain together for the duration of the study. Exclusion criteria at enrolment included that the HIV-1 infected partner could not be on antiretroviral therapy or be eligible for initiation of antiretroviral therapy according to national guidelines, and women were excluded if they were pregnant. For the Partners in Prevention HSV/HIV Transmission Study, all HIV-1 infected partners were HSV-2 seropositive. There were no exclusions for individuals with syndromic or aetiological STI diagnoses.

    At the enrolment visit, interviewer-administered standardised questionnaires were used to collect information on demographics and sexual behaviours from both study partners. Genital exams were performed to identify STI symptoms. Clinician-collected endocervical swabs and urine samples were gathered from women and men, respectively, to test for trichomoniasis, chlamydia and gonorrhoea while blood samples were collected for serological assays.

    Both study protocols, including planned analyses for STI transmission risk factors, were approved by the University of Washington Human Subjects Review Committee as well as ethics review committees at each study site. All study participants provided written informed consent.

    Laboratory methods

    All samples used for diagnostic testing were collected at the study enrolment visit, which both members of the couple attended together. Testing for T. vaginalis was done using NAAT by APTIMA TV transcription-mediated amplification assay (seven sites in the Partners PrEP Study; Gen-Probe, San Diego, California, USA), a research version of the APTIMA assay (all sites in the Partners in Prevention HSV/HIV Transmission Study), or by culture using InPouch TV (two sites in the Partners PrEP Study; Biomed Diagnostics, White City, Oregon, USA).

    HIV-1 serostatus was determined by rapid testing, with positive results confirmed by western blot or ELISA. Neisseria gonorrhoeae and Chlamydia trachomatis testing was performed using APTIMA Combo 2 (Gen-Probe, San Diego, California, USA) or COBAS Amplicor (Roche Diagnostics, Indianapolis, Indiana, USA). Serological testing for syphilis was by rapid plasma reagin test, confirmed with the use of a treponema-specific assay. Gram-stained smears were prepared from vaginal swabs to evaluate vaginal microbiota. Normal microbiota, intermediate microbiota and bacterial vaginosis were defined by Nugent scores of 0–3, 4–6 and 7–10, respectively.

    Statistical methods

    All couples in which both partners provided a sample for T. vaginalis testing at study enrolment were included in this analysis. The Wilcoxon matched-pairs signed-rank test was used to assess if the prevalence of T. vaginalis differed in men and women. Poisson regression models with robust standard errors were used to identify factors associated with the risk of a prevalent T. vaginalis infection. The association between genital symptoms and T. vaginalis infection was assessed in bivariate models and in models adjusting for the presence of other STIs. Correlates of a prevalent T. vaginalis infection for men and women were also analysed using Poisson regression. Factors associated with T. vaginalis infection in bivariate models (p<0.05) were included in multivariable models. For factors with more than two categories, likelihood ratio tests were used to determine inclusion in multivariable models. Analyses were performed using Stata V.13.1 (Stata Corporation, College Station, Texas, USA).

    Results

    Of the 8155 HIV-1 serodiscordant couples enrolled in the Partners in Prevention HSV/HIV Transmission and Partners PrEP Studies, 92% (7531/8155) of couples were tested for trichomoniasis at enrolment. This included 4872 couples (65%) with a HIV-1 seropositive female partner and 2659 couples (35%) with a HIV-1 seropositive male partner (table 1). The median age was 30 years for women and 36 years for men. Most couples were married, living together and had been together for more than 5 years. Unprotected sex in the prior month was reported by 33% of couples, while 0.9% of women and 10% of men reported sex with individuals besides their primary partner in the prior month.

    Table 1

    Characteristics of study participants

    T. vaginalis infection was detected in 1176 (7.8%) study participants, with the prevalence of infection 2.7 times higher among women than men (4.2% vs 11.4%, p<0.001). At least one partner tested positive for trichomoniasis in 13% (981/7531) of couples: 8.8% (662/7531) of couples had only the female partner positive, 1.6% (124/7531) of couples had only the male partner positive and 2.6% (195/7531) of couples had both partners positive. Exploring concordance of T. vaginalis status within sexual partnerships, we observed that 61% (195/319) of men with trichomoniasis had an infected female partner and 23% (195/857) of women with trichomoniasis had an infected male partner.

    Women with abnormal vaginal discharge (RR 1.58, 95% CI 1.35 to 1.84), cervical bleeding (RR 1.65, 95% CI 1.41 to 1.92) or adnexal, cervical or uterine tenderness (RR 1.42, 95% CI 1.04 to 1.95) at their clinical exam were more likely to have trichomoniasis than women without those respective signs and symptoms (table 2). These associations remained statistically significant after adjusting for other STIs. However, the majority (65%, 555/852) of women with trichomoniasis had none of these three signs or symptoms present, with 20% (173/852) of infected women experiencing abnormal vaginal discharge, 20% (169/852) of infected women experiencing cervical bleeding and 4% (34/852) of infected women experiencing adnexal, cervical or uterine tenderness. Among men, urethral discharge was uncommon.

    Table 2

    Signs and symptoms of Trichomonas vaginalis infection at examination

    Several factors were associated with the risk of a woman having a prevalent T. vaginalis infection (table 3). Multivariable analysis found that women with outside sex partners (RR 1.78, 95% CI 1.15 to 2.76), HIV-1 (RR 1.19, 95% CI 1.02 to 1.38), N. gonorrhoeae (RR 1.65, 95% CI 1.23 to 2.23) and intermediate (RR 2.04, 95% CI 1.70 to 2.44) or high (RR 1.96, 95% CI 1.67 to 2.30) Nugent scores for bacterial vaginosis had increased risk of T. vaginalis infection. Being married (RR 0.74, 95% CI 0.59 to 0.93), having children (RR 0.80, 95% CI 0.69 to 0.92) and using injectable contraception (RR 0.77, 95% CI 0.64 to 0.94) were associated with reduced T. vaginalis risk. In addition, women aged 45 years and above were found to have a lower risk of T. vaginalis (RR 0.63, 95% CI 0.44 to 0.89) compared with women aged 18–24 years. Women whose male partner was circumcised (RR 0.82, 95% CI 0.72 to 0.94), had nine or more years of education (RR 0.72, 95% CI 0.63 to 0.82) or who earned income (RR 0.82, 95% CI 0.71 to 0.94) were less likely to have trichomoniasis.

    Table 3

    Correlates of female Trichomonas vaginalis infection

    For men, concurrent HIV-1 infection (RR 0.69, 95% CI 0.53 to 0.90) was associated with decreased risk of T. vaginalis infection (table 4). Men whose samples were tested by NAAT had increased risk of being identified as having trichomoniasis (RR 2.43, 95% CI 1.27 to 4.65) compared with men tested by culture.

    Table 4

    Correlates of male Trichomonas vaginalis infection

    For both women and men, having a partner with trichomoniasis was the factor most strongly associated with T. vaginalis infection risk. In unadjusted analyses, women with a T. vaginalis-positive male partner had 6.66 times the prevalence (95% CI 5.94 to 7.46) of T. vaginalis infection, while men with a T. vaginalis-positive female partner had 12.25 times the prevalence (95% CI 9.89 to 15.16) of T. vaginalis infection compared with individuals whose primary partner was T. vaginalis-negative. The corresponding absolute change in the prevalence of T. vaginalis associated with having a T. vaginalis-positive partner was also large, with the prevalence increasing from 9% among women whose partner was T. vaginalis-negative to 61% among women whose partner was T. vaginalis-positive (risk difference = 52%). For men, the prevalence of T. vaginalis increased from 2% among men whose partner was T. vaginalis-negative to 23% among men whose partner was T. vaginalis-positive (risk difference = 21%). In multivariable analyses, women with a T. vaginalis-positive male partner had 4.7 times the prevalence (95% CI 4.10 to 5.38) of T. vaginalis infection, while men with a T. vaginalis-positive female partner had 10.09 times the prevalence (95% CI 7.92 to 12.85) of T. vaginalis infection compared with individuals whose primary partner was T. vaginalis-negative.

    Discussion

    Using data from two large trials that enrolled HIV-1 serodiscordant couples living in sub-Saharan Africa, this exploratory analysis identified characteristics of both individuals in heterosexual couples that were associated with trichomoniasis. For both men and women, having a T. vaginalis-infected partner was the factor most strongly associated with the likelihood of a prevalent T. vaginalis infection. Our finding that 23% of women and 61% of men with T. vaginalis had a partner with a concurrent T. vaginalis infection highlights the importance of treating sexual partners of individuals diagnosed with T. vaginalis. This will be increasingly important as the use of NAAT testing for T. vaginalis expands and more male T. vaginalis cases are identified, since the prevalence of trichomoniasis was approximately threefold higher among partners of T. vaginalis-positive men compared with partners of T. vaginalis-positive women. We also found that the prevalence of trichomoniasis among women was 2.7 times higher than among men, which is strikingly different from the existing WHO African region estimates that women are 10 times more likely than men to have trichomoniasis.1 WHO had few data points available to estimate the prevalence of T. vaginalis among men and used a complex methodology that attempted to adjust for the sensitivity and specificity of specific diagnostic tests to generate their estimates,10 which our findings suggest may not accurately reflect the distribution of trichomoniasis among women and men.

    This study found that women with gonorrhoea, HIV-1 and intermediate or high Nugent scores for bacterial vaginosis were more likely to have trichomoniasis. The association with bacterial vaginosis has the potential to be especially important from a public health perspective because a large proportion of women in some populations have intermediate or high Nugent scores (48% of this study population). Cohort studies have found that women with intermediate or high Nugent scores were more likely to acquire T. vaginalis,9 ,16–19 and our analysis confirms this association while adjusting for characteristics of both members of sexual partnerships. In addition, the unadjusted analysis found that men whose female partners had intermediate or high Nugent scores had over a twofold increased risk of T. vaginalis infection. Though this association was no longer significant in the multivariable model which adjusted for female partner T. vaginalis status, the expected mediator of the association, this finding highlights the potential impact of bacterial vaginosis on the prevalence of T. vaginalis among men.

    Additionally, this analysis found that injectable contraceptive use was associated with a decreased risk of trichomoniasis. Several recent well-powered studies have also found injectable contraceptive use associated with decreased risk of T. vaginalis infection.16 ,17 ,20–22 It has been hypothesised that injectable contraceptive use reduces T. vaginalis risk though inhibition of exogenous oestrogen and androgen receptors or by reducing iron availability through decreased menstrual flow.22 ,23 Additional research is needed to better understand the mechanism underlying this association.

    This study confirms the results of a randomised controlled trial which observed that women with circumcised male partners were less likely to have trichomoniasis.24 Two longitudinal studies that assessed this association reported inconstant results, though these studies relied on female partners to report the circumcision status of their male partner and used lower-sensitivity wet mount assays to detect T. vaginalis infection.25 ,26 We did not find an association between male circumcision and the risk of trichomoniasis among men, though data from other studies suggest a modest protective effect.27–30 The mechanism through which male circumcision reduces the risk of male-to-female T. vaginalis transmission is currently unknown, though it has been hypothesised that moisture in the subpreputial space in uncircumcised men may facilitate T. vaginalis survival, increasing female exposure to the pathogen.24

    We identified few factors associated with trichomoniasis risk among men. Our data suggest that a man's risk of T. vaginalis was largely driven by having a T. vaginalis-positive female partner, consistent with evidence that the average duration of infection among men is short and thus recent exposure to T. vaginalis is a prerequisite for infection. It is worth noting that though we found a protective effect of HIV-1 on men's risk of T. vaginalis in our multivariable model, the study population consisted of HIV-1 serodiscordant couples so all HIV-1-negative men had HIV-1-positive female partners.

    Limitations of this analysis should be considered when interpreting these results. Since the analysis was cross-sectional, we are unable to assess temporality and cannot determine if factors are associated with incident T. vaginalis infections. Though NAATs were used for the majority of T. vaginalis testing, 16% of couples were tested using culture. Culture is less sensitive at detecting T. vaginalis infections, especially among men, so we may have underestimated the prevalence of T. vaginalis.8 ,31 Including only those individuals tested by NAAT, the prevalence of T. vaginalis among women rose from 11% to 12% (770/6339) and among men rose from 4% to 5% (303/6339). Lastly, these cohorts of HIV-1 serodiscordant couples in stable partnerships are unique, and some results such as the prevalence estimates may not be generalisable to other populations.

    Strengths of this analysis include the fact that we explored the epidemiology of T. vaginalis within sexual partnerships. Adjusting for characteristics of the male partner, the female partner and the couple allowed us to thoroughly explore factors associated with T. vaginalis risk. Interviewing and performing genital exams for both members of sexual partnerships, rather than relying on one individual to report their partner's characteristics, likely reduced misclassification.32 Our ability to analyse data from 7531 couples provided us with the statistical power to identify associations between T. vaginalis and factors that have weaker associations or are less common. In addition, much of the existing data on T. vaginalis concordance within couples come from STD clinics, where patients are likely to be experiencing symptoms. Most T. vaginalis cases are asymptomatic,8 so our concordance data may better reflect the general population of T. vaginalis cases.

    In summary, trichomoniasis was more prevalent among women than men and was common in partners of T. vaginalis-infected individuals, while most infected individuals did not have signs or symptoms of infection. These findings illustrate the importance of notifying the sexual partners of individuals diagnosed with trichomoniasis to identify additional cases and prevent reinfection. Our findings that female partners of circumcised men had an 18% reduced risk of prevalent trichomoniasis compared with women with uncircumcised partners suggests that male circumcision may be a useful public health intervention to prevent T. vaginalis infections.

    Key messages

    • For both men and women, the strongest predictor of a prevalent Trichomonas vaginalis infection is having a T. vaginalis-infected sexual partner.

    • Concurrent treatment of sexual partners is critical to prevent reinfection.

    • Male circumcision programmes and expanded treatment for bacterial vaginosis have the potential to reduce the prevalence of trichomoniasis.

    Acknowledgments

    The authors thank all the co-investigators and staff who worked on the Partners in Prevention HSV/HIV Transmission and Partners PrEP Studies, both at study sites and at the University of Washington. The authors also thank the HIV-1 serodiscordant couples who enrolled in these studies.

    References

    Footnotes

    • Handling editor Jackie A Cassell

    • Collaborators Partners PrEP Study Team: University of Washington Coordinating Center and Central Laboratories, Seattle: Connie Celum (principal investigator, protocol cochair), Jared M Baeten (medical director, protocol co-chair), Deborah Donnell (protocol statistician), Robert W Coombs, Lisa Frenkel, Craig W Hendrix, Jairam Lingappa, and M Juliana McElrath. Study sites and site principal investigators: Eldoret, Kenya (Moi University; Indiana University): Kenneth Fife, Edwin Were; Kabwohe, Uganda (Kabwohe Clinical Research Center): Elioda Tumwesigye; Jinja, Uganda (Makerere University; University of Washington): Patrick Ndase, Elly Katabira; Kampala, Uganda (Makerere University): Elly Katabira, Allan Ronald; Kisumu, Kenya (Kenya Medical Research Institute, University of California San Francisco): Elizabeth Bukusi, Craig Cohen; Mbale, Uganda (The AIDS Support Organization; Centers for Disease Control and Prevention, Uganda): Jonathan Wangisi, James Campbell, Jordan Tappero; Nairobi, Kenya (University of Nairobi; University of Washington): James Kiarie, Carey Farquhar, Grace John-Stewart; Thika, Kenya (University of Nairobi; University of Washington): Nelly Rwamba Mugo; Tororo, Uganda (Centers for Disease Control and Prevention, Uganda; The AIDS Support Organization): James Campbell, Jordan Tappero, Jonathan Wangisi. Data management for the HIV-serodiscordant couples studies was provided by DF/Net Research, and site laboratory oversight was provided by Contract Laboratory Services (University of the Witwatersrand, Johannesburg, South Africa). Study medication was donated by Gilead Sciences. The Partners in Prevention HSV/HIV Transmission Study Team: University of Washington Coordinating Center and Central Laboratories, Seattle, USA: Connie Celum (principal investigator), Anna Wald (protocol co-chair), Jairam R Lingappa (medical director), Jared M Baeten, Mary S Campbell, Lawrence Corey, Robert W Coombs, James P Hughes, Amalia Magaret, M Juliana McElrath, Rhoda Morrow, James I Mullins. Study site principal investigators and study coordinators at sites contributing data and samples to this study: Cape Town, South Africa (University of Cape Town): David Coetzee; Eldoret, Kenya (Moi University, Indiana University): Kenneth Fife, Edwin Were; Gaborone, Botswana (Botswana Harvard Partnership): Max Essex, Joseph Makhema; Kampala, Uganda (Infectious Disease Institute, Makerere University): Elly Katabira, Allan Ronald; Kisumu, Kenya (Kenya Medical Research Institute, University of California San Francisco): Elizabeth Bukusi, Craig Cohen; Moshi, Tanzania (Kilimanjaro Christian Medical College, Harvard University): Saidi Kapiga, Rachel Manongi; Nairobi, Kenya (University of Nairobi, University of Washington): Carey Farquhar, Grace John-Stewart, James Kiarie; Kitwe, Zambia (Rwanda Zambia HIV Research Group, and Emory University): Susan Allen, William Kanweka; Ndola, Zambia (Rwanda Zambia HIV Research Group, and Emory University): Susan Allen, Mubiana Inambao; Orange Farm, South Africa (Reproductive Health Research Unit, University of the Witwatersrand): Sinead Delany-Moretlwe, Helen Rees; Soweto, South Africa (Perinatal HIV Research Unit, University of the Witwatersrand): Guy de Bruyn, Glenda Gray, James McIntyre; Thika, Kenya (University of Nairobi, University of Washington): Nelly Rwamba Mugo.

    • Contributors AFB, JMB, ASR and RVB designed the study. JMB, EN-J, JRL, NRM, EAB, SK, SD-M and CC contributed to data collection. AFB developed the statistical analysis plan, performed the statistical analyses and led manuscript development. All authors contributed to the writing of this manuscript. The authors designed and executed the study, had full access to the raw data, performed all analyses, wrote the manuscript, and had final responsibility for the decision to submit for publication.

    • Funding The Bill and Melinda Gates Foundation (grant ID #26469 and #47674).

    • Competing interests None declared.

    • Patient consent Obtained.

    • Ethics approval University of Washington Human Subjects Review Committee (institutional review board ID: STUDY00000172 and STUDY00000867) as well as ethics review committees at each study site.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Data sharing statement Detailed information on study data is available upon request from RVB at rbarnaba@uw.edu.

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.