Chlamydia, gonorrhoea, trichomoniasis and syphilis: global prevalence and incidence estimates, 2016

Abstract Objective To generate estimates of the global prevalence and incidence of urogenital infection with chlamydia, gonorrhoea, trichomoniasis and syphilis in women and men, aged 15–49 years, in 2016. Methods For chlamydia, gonorrhoea and trichomoniasis, we systematically searched for studies conducted between 2009 and 2016 reporting prevalence. We also consulted regional experts. To generate estimates, we used Bayesian meta-analysis. For syphilis, we aggregated the national estimates generated by using Spectrum-STI. Findings For chlamydia, gonorrhoea and/or trichomoniasis, 130 studies were eligible. For syphilis, the Spectrum-STI database contained 978 data points for the same period. The 2016 global prevalence estimates in women were: chlamydia 3.8% (95% uncertainty interval, UI: 3.3–4.5); gonorrhoea 0.9% (95% UI: 0.7–1.1); trichomoniasis 5.3% (95% UI:4.0–7.2); and syphilis 0.5% (95% UI: 0.4–0.6). In men prevalence estimates were: chlamydia 2.7% (95% UI: 1.9–3.7); gonorrhoea 0.7% (95% UI: 0.5–1.1); trichomoniasis 0.6% (95% UI: 0.4–0.9); and syphilis 0.5% (95% UI: 0.4–0.6). Total estimated incident cases were 376.4 million: 127.2 million (95% UI: 95.1–165.9 million) chlamydia cases; 86.9 million (95% UI: 58.6–123.4 million) gonorrhoea cases; 156.0 million (95% UI: 103.4–231.2 million) trichomoniasis cases; and 6.3 million (95% UI: 5.5–7.1 million) syphilis cases. Conclusion Global estimates of prevalence and incidence of these four curable sexually transmitted infections remain high. The study highlights the need to expand data collection efforts at country level and provides an initial baseline for monitoring progress of the World Health Organization global health sector strategy on sexually transmitted infections 2016–2021.


Introduction
Sexually transmitted infections are among the most common communicable conditions and affect the health and lives of people worldwide. The World Health Organization (WHO) periodically generates estimates to gauge the global burden of four of the most common curable sexually transmitted infections: chlamydia (etiological agent: Chlamydia trachomatis), gonorrhoea (Neisseria gonorrhoeae), trichomoniasis (Trichomonas vaginalis) and syphilis (Treponema pallidum). [1][2][3][4][5][6] The estimates provide evidence for programme improvement, monitoring and evaluation.
These sexually transmitted infections cause acute urogenital conditions such as cervicitis, urethritis, vaginitis and genital ulceration, and some of the etiological agents also infect the rectum and pharynx. Chlamydia and gonorrhoea can cause serious short-and long-term complications, including pelvic inflammatory disease, ectopic pregnancy, infertility, chronic pelvic pain and arthritis, and they can be transmitted during pregnancy or delivery. Syphilis can cause neurological, cardiovascular and dermatological disease in adults, and stillbirth, neonatal death, premature delivery or severe disability in infants. All four infections are implicated in increasing the risk of human immunodeficiency virus (HIV) acquisition and transmission. 7 Moreover, people with sexually transmitted infections often experience stigma, stereotyping, vulnerability, shame and gender-based violence. 8 In May 2016, the World Health Assembly adopted the Global health sector strategy on sexually transmitted infections, 2016-2021. 9 This strategy includes rapid scale-up of evidencebased interventions and services to end sexually transmitted infections as public health concerns by 2030. The strategy sets targets for reductions in gonorrhoea and syphilis incidence in adults and recommends the establishment of global baseline incidences of sexually transmitted infections by 2018. The primary objectives of this study were to estimate the 2016 global and regional prevalence and incidence of chlamydia, gonorrhoea, trichomoniasis and syphilis in adult women and men.

Chlamydia, gonorrhoea and trichomoniasis
We generated estimates for these three infections through systematic reviews using the same methods as for the 2012 estimates. 6 Estimates of four sexually transmitted infections, 2016 Jane Rowley et al.
We searched for articles published between 1 January 2009 and 29 July 2018 in PubMed® without language restrictions. We used PubMed Medical subject heading (MeSH) terms for individual country names combined with: "chlamydia"[MeSH Te r m s ] O R "c h l a m y d i a" [ A l l Fields], "gonorrhoea"[All Fields] OR "gonorrhea"[MeSH Terms] OR "gonorrhea"[All Fields], "trichomonas infections"[MeSH Terms] OR ("trichomonas"[All Fields] AND "infections"[All Fields]) OR "trichomonas infections"[All Fields] OR "trichomoniasis"[All Fields]). We also asked WHO regional sexually transmitted infection advisors and other leading experts in the field for additional published and unpublished data.
To be eligible, studies had to collect most specimens between 2009 and 2016 or be published in 2010 or later if specimen collection dates were not available. Other study inclusion criteria were: sample size of at least 100 individuals; general population (e.g. pregnant women, women at delivery, women attending family planning clinics, men and women selected for participation in demographic and health surveys); and use of an internationally recognized diagnostic test with demonstrated precision using urine, urethral, cervical or vaginal specimens.
To reduce bias in the estimation of general population prevalence, we excluded studies conducted among the following groups: patients seeking care for sexually transmitted infection or urogenital symptoms, women presenting at gynaecology or sexual health clinics with sexually transmitted infection related issues, studies restricted to women with abnormal Papanicolaou test results, remote or indigenous populations, recent immigrant or migrant populations, men who have sex with men and commercial sex workers.
Two investigators independently reviewed all identified studies to verify eligibility. When more than one publication reported on the same population, we retained the publication with the most detailed information. For each included study, we calculated prevalence as the number of individuals with a positive test result divided by the total number tested. We then standardized these values by applying adjustment factors for the accuracy of the laboratory diagnostic test, study location (rural versus urban) and the age of the study population. If the adjustments resulted in a negative value, we replaced the value with 0.1% when doing the meta-analysis. The methods and adjustment factors were identical to those used to generate the 2012 estimates. 6 We obtained estimates for 10 geographical areas (referred to as estimation regions). 6 Estimates for high-income North America (Canada and United States of America), were based on the latest published United States estimates that used data from multiple sources. 10,11 For the other nine estimation regions, we calculated a summary prevalence estimate by meta-analysis if there were three or more data points. 12 There were sufficient data to generate an estimate for chlamydia in women in all regions, but not for gonorrhoea or trichomoniasis. For regions with insufficient data for gonorrhoea and trichomoniasis, we assumed that prevalence was a multiple of the prevalence of chlamydia. The infection specific multiples were based on those studies that met the 2016 inclusion criteria (available from the data repository). 13 For men, when there were insufficient data for meta-analysis, the prevalence of an infection was assumed to be proportional to the prevalence in women. The male-to-female ratios were infection-specific and were set at the same values as in 2012 estimates. 6 To reflect the contribution of populations at higher risk of infection (e.g. men who have sex with men and commercial sex workers), who are likely to be under-represented in general population samples, we increased prevalence estimates by 10%, as in the 2012 estimates, 6 for each estimation region, apart from high-income North America.
We performed the meta-analyses using a Bayesian approach with a Markov Chain Monte Carlo algorithm implemented with the software BRrugs in R package (R foundation, Vienna, Austria). 14 For each infection, the software generated 10 000 samples from the posterior distribution for the expected mean prevalence in each estimation region based on the β-binomial model, and used these to calculate the 2.5 and 97.5 uncertainty percentiles. 15 We calculated global and regional prevalence estimates for each infection by weighting each of the 10 000 samples from estimation regions according to population size, using United Nations population data for women and men aged 15-49 years. 16 We present results by WHO region, 2016 World Bank income classification 17 and 2017 sustainable development goal (SDG) region. 18 All analyses were carried out using R statistical software (R foundation).

Syphilis
We based syphilis estimates on the WHO's published 2016 maternal prevalence estimates. 19 These estimates were generated by using Spectrum-STI, a statistical trend-fitting model in the publicly available Spectrum suite of health policy planning tools 20 and country specific data from the global Spectrum-STI syphilis database (available from the corresponding author). As in the 2012 estimation, 6 we assumed that the prevalence of syphilis in all women 15-49 years of age in each country was the same as in pregnant women. We then increased the estimate by 10% to reflect the contribution of populations at higher risk. The men to women prevalence ratio of syphilis was set at 1.0 and assumed to have a uniform distribution ± 33% around this value, in agreement with data from a recent global meta-analysis of syphilis. 21 We generated regional and global estimates by weighting the contribution of each country by the number of women and men aged 15-49 years. Regional and global 95% uncertainty intervals (UIs) were generated using the delta method; 22 uncertainties were assumed to be independent across countries.

Incidence estimation
We calculated incidence estimates for each infection by dividing prevalence by the average duration of infection for all estimation regions except highincome North America where published estimates were used. 10,11 Estimates of the average duration of infection were those used in the 2012 estimation 6 and assumed to have a uniform distribution of ± 33.3% around the average duration. We calculated uncertainty in incidence for a given region, sex and infection at the national level using the delta method; 22 uncertainty in the prevalence estimate was multiplied by uncertainty in the estimated duration of infection. Regional and global uncertainty intervals were generated assuming uncertainties were independent across countries.

Chlamydia, gonorrhoea and trichomoniasis
Of the 7244 articles screened, 112 studies met the inclusion criteria for one or more of the three infections ( Fig. 1 (Table 1 and Table 2). Only 34 studies in women and four studies in men provided information on all three infections. The included studies contained 100 data points in women for chlamydia, 64 for gonorrhoea and 69 for trichomoniasis. In men, there were 16 data points for chlamydia, 11 for gonorrhoea and seven for trichomoniasis (Table 3).
For women, a total of 43 (21.0%) of 205 countries, territories and areas had one or more data points for chlamydia, 32 (15.6%) for gonorrhoea and 29 (14.1%) for trichomoniasis. For men, only 15 (7.3%) countries, territories and areas had one or more data points for chlamydia, 10 (4.9%) for gonorrhoea and 7 (3.4%) for trichomoniasis. For women there were sufficient data to generate summary estimates for chlamydia for the nine estimation regions, but not for gonorrhoea or trichomoniasis (Table 4).  Table 5 shows prevalence estimates for the WHO regions for 2016. Based on prevalence data from 2009 to 2016, the estimated pooled global prevalence of chlamydia in 15-49-year-old women was 3.8% (95% UI: 3.3-4.5) and in men 2.7% (95% UI: 1.9-3.7), with regional values ranging from 1.5 to 7.0% in women and 1.2 to 4.0% in men. For gonorrhoea, the global estimate was 0.9% (95% UI: 0.7-1.1) in women and 0.7% (95% UI: 0.5-1.1) in men, with regional values in women ranging from 0.3 to 1.9% and from 0.3 to 1.6% in men. The estimates for trichomoniasis were 5.3% (95% UI: 4.0-7.2) in women and 0.6% (95% UI: 0.4-0.9) in men, with regional values ranging from 1.6 to 11.7% in women and from 0.2 to 1.3% in men. For syphilis, the global estimate in both men and women was 0.5% (95% UI: 0.4-0.6) with regional values ranging from 0.1 to 1.6%. The WHO African Region had the highest prevalence for chlamydia in men, gonorrhoea in women and men, trichomoniasis in women and syphilis in men and women. The WHO Region of the Americas had the highest prevalence of chlamydia in women and of trichomoniasis in men.

Prevalence and incidence estimates
These prevalence estimates correspond to the totals of 124.3 million cases of chlamydia, 30.6 million cases of gonorrhoea, 110.4 million cases of trichomoniasis and 19.9 million cases of syphilis (available from the data repository). 13 Using the World Bank classification, high-income countries, territories and areas had the lowest estimated prevalence, and low-income countries, territories and areas had the highest prevalence of gonorrhoea, trichomoniasis and syphilis. For chlamydia, estimated prevalence was highest in upper-middle income countries, territories and areas (Fig. 2). The SDG grouping showed the highest prevalence of all four sexually transmitted infections in Oceania region, that is, Pacific island nations excluding Australia and New Zealand (available from the data repository). 13 We estimated the global incidence rate for chlamydia in 2016 to be 34 cases per 1000 women (95% UI: 25-45) and 33 per 1000 men (95% UI: 21-48); for gonorrhoea 20 per 1000 women (95% Fig. 1 Fig. 3). The WHO Region of the Americas had the highest incidence rate for chlamydia and syphilis in both women and men, while the WHO African Region had the highest incidence rates for gonorrhoea and trichomoniasis in women and men. Incidence rates by income category and SDG regions are available from the data repository. 13 These incidence rates translate globally into 127

Comparison of estimates
Comparing the 2012 estimates with the estimates presented here shows that more data points were available in women for the 2016 estimates. The number increased from 69 to 100 for chlamydia, 50 to 64 for gonorrhoea and 44 to 69 for trichomoniasis. For men, the number of data points fell from 21 to 16 for chlamydia and from 12 to 11 for gonorrhoea, but increased from one to seven for trichomoniasis. The period of eligibility for both estimates was eight years with an overlap of four years (2009 to 2012); in women 27 data points were included in both estimates for chlamydia, 18 for gonorrhoea and 20 for trichomoniasis. In men, these overlaps were six, five and one, respectively.  Global male-tofemale ratio a In consultation with advisors on sexual transmitted infections for the World Health Organization (WHO) Region of the Americas, we decided to use the midpoint between the 2016 estimate generated by applying the global male-to-female ratio (7.5%) and the 2012 estimate for the region (2.1%). We deemed the former to be too high and the latter too low. b Following discussions with the United States Centers for Disease Control and Prevention, we based our estimates on the latest published United States national estimates 21,22 and assumed they remained constant over time and that estimates for 15-39-year-old people could be extrapolated to the 15-49-year age range. We did not apply the adjustments used for other Regions in the WHO estimates process. The figures for the United States were also applied to Canada. c The estimate based on the three available data points was over 4%, considerably higher than the 2012 estimate. Following discussions with regional experts we decided not to use this estimate, but instead to use the trichomoniasis to chlamydia ratio for low and lower middle-income countries, territories and areas. d This estimation region is made up of countries from East Asia and South East Asia. We used the higher and upper-middle income gonorrhoea to chlamydia ratio for East Asia and the low and lower-middle income ratio for South East Asia. " Estimates of four sexually transmitted infections, 2016 Jane Rowley et al.
Spectrum data set. 19 For all infections in both women and men, the 2016 global prevalence estimate was within the 95% UI for 2012. At the regional level, the 95% UIs for prevalence overlapped for all four infections in both men and women, apart from gonorrhoea in men in the WHO African Region which was higher in 2016 than in 2012.

Discussion
We estimated a global total of 376.4 million new curable urogenital infections with chlamydia, gonorrhoea, trichomoniasis and syphilis in 15-49-year-old women and men in 2016. This estimate corresponds to an average of just over 1 million new infections each day. The number of individuals infected, however, is smaller as repeat infections and co-infections are common. 152 The estimates of prevalence and incidence in 2016 were similar to those in 2012, both globally and by region, showing that sexually transmitted infections are persistently endemic worldwide. Grouping countries, territories and areas according to SDG regions revealed that the prevalence and incidence of all four sexually transmitted infections, in both women and men, were highest in the Oceania Region. The small island states in this SDG region are part of the WHO Western Pacific Region, which is dominated by China (owing to its population size). Therefore, the levels of sexually transmitted infections and need for infection control in these island states are masked when viewing the estimates only by WHO Region. When using the World Bank classification of countries, the prevalence of gonorrhoea, trichomoniasis and syphilis were highest in low-income countries, territories and areas. The prevalence of chlamydia was highest in the upper middle-income countries, territories and areas, partly due to high estimates in some Latin American countries. Further research is needed to determine whether these estimates reflect methodological factors or differences in C. trachomatis transmission.
The 2016 estimates for chlamydia, gonorrhoea and trichomoniasis were based on a systematic review of the literature complemented by outreach to experts using the same methods as in 2012. The aim was to reduce bias and insure comprehensiveness in the Table 5. South-East Asia Region The study has limitations. First, limited prevalence data were available, despite an eight-year time window for data inclusion. Estimates for a given infection and region are therefore extrapolated from a small number of data points and ratios were used to generate estimates for some regions. For men, the lack of data was particularly striking. For syphilis, most data were from pregnant women, which might not reflect all women aged 15-49 years, or men. Second, the source studies include people in different age groups and used a range of diagnostic tests, so adjustment factors were applied to standardize measures across studies. Third, owing to the absence of empirical studies, incidence estimates were derived from the relationship between prevalence and duration of infection, and data on the average duration of infection for each of the four infections are also limited. Finally, because only studies among the general population were used, the prevalence and incidence in areas where key populations contribute disproportionately to sexually transmitted infection epidemics may have been underestimated despite the applied correction factor. These limitations have been discussed previously in detail. 6 This study has implications for sexually transmitted infection programming and research. The quantity and quality of prevalence and incidence studies for sexually transmitted infections in representative samples of the general population, for both women and men, need improvement. Identifying opportunities to integrate data collection with clinical care platforms, such as HIV, adolescent, maternal, family planning and immunization is crucial. The recently developed WHO protocol for assessing the prevalence of sexually transmitted infections in antenatal care settings 153 provides a framework and consistent methods that can be adapted for women and men. Comparing data across studies requires better understanding of the performance characteristics of diagnostic tests, and implications for estimates of the average duration of infection for each infection. The processes for producing future prevalence estimates could be made timelier and more efficient through continually updated systematic reviews, 154 as well as technological solutions that automate searching of databases and facilitate high quality updates of reviews.
The global estimates of prevalence and incidence of four curable sexually transmitted infections are important in the broader global context, highlighting a continuing public health challenge. Prevalence and incidence data play an important role in the design and evaluation of programmes and interventions for sexually transmitted infections and in interpreting changes in HIV epidemiology. The global threat of antimicrobial resistance, particularly the emergence of N. gonorrhoeae resistance to the few remaining antimicrobials recommended for treatment, further highlights the importance of investing in monitoring prevalence and incidence. 155 Estimates of prevalence and incidence are essential for calculations of the burden of disease due to sexually transmitted infections, which are needed to advocate for funding to support sexually transmitted infection programmes. These burden estimates can also be used to promote innovation for point-of-care diagnostics, new therapeutics, vaccines and microbicides. The WHO Global Health Sector  Estimates of four sexually transmitted infections, 2016 Jane Rowley et al.