Syphilis in the state of São Paulo , Brazil , 2011 – 2017 Sífilis no estado de São Paulo , Brasil , 2011 – 2017

REV BRAS EPIDEMIOL 2020; 23: E200103 ABSTRACT: Objective: To analyze how syphilis detection rates evolved from 2011 to 2017 according to sex, age and place of residence in the state of São Paulo. Methods: A historical series was organized with data from the Notification Disease Information System. The acquired syphilis detection rates (ASDR) per 100,000 inhabitants and the acquired syphilis detection rates including pregnant women with syphilis (PASDR) per 100,000 inhabitants were described. For a trend analysis of the rates in the studied period, the Poisson Jointpoint (inflection point) model was performed, and the annual percentage change (APC) per segment and the average annual percentage change (AAPC) were estimated, with respective 95% confidence intervals (95%CI). Results: A total of 205,424 cases of acquired syphilis and syphilis in pregnant women in the period were reported. The ASDR per 100,000 inhabitants ranged from 26.0 to 84.6 between 2011 and 2017 and the PASDR per 100,000 inhabitants ranged from 33.7 to 108.9; the trend was increasing in both, and an inflection point was identified dividing the ASDR and PASDR curve into two periods: 2011 to 2013 and 2013 to 2017: the AAPC found for ASDR was 21.0% (95%CI 15.5 – 26.4) and the PASDR was 21.2% (95%CI 16.4 – 26.1), in the age groups up to 24 years old, there was a significant growth in both sexes. A heterogeneity in the evolution of rates by region of the state was observed between 2011 and 2017. Conclusions: The increasing trend in acquired syphilis detection rates can be attributed to better adherence to notification and disproportionate involvement of young people.


INTRODUCTION
Syphilis is a sexually transmitted infection (STI) that continues to be a serious public health problem in Brazil and around the world, despite effective measures for diagnosis, treatment, prevention and control that have been around since the 1950s 1-3 . During pregnancy, congenital syphilis can occur, causing serious consequences for the fetus, such as: fetal death, stillbirth, abortion, malformations, prematurity and low birth weight [3][4][5][6] . Syphilis is also associated with the human immunodeficiency virus (HIV), as it increases the risk of getting it, especially if there lesions appear in the primary and secondary stages of the disease 3, 6 .
In 2012, approximately six million cases of syphilis were estimated to have occurred worldwide 7 . The Sustainable Development Goals (SDGs) set for 2030 aim at establishing strategies that prevent and control STIs, by reducing the worldwide incidence of infection of Treponema pallidum, in relation to the estimated incidence in 2018, and the elimination of congenital syphilis. In order to obtain accurate information about the scope of the disease, an STI surveillance system based on case notification is necessary. Such a surveillance system is capable of analyzing the population with reported cases according to age, sex and location 8 .
In Brazil, the National Compulsory Notification List (Lista Nacional de Notificação Compulsória -LNNC) 9 contains the following diseases related to syphilis: congenital syphilis, syphilis in pregnant women, and acquired syphilis.

REV BRAS EPIDEMIOL 2020; 23: E200103
In 2017, 119,800 cases of acquired syphilis were reported and the detection rate was 58.1 per 100 thousand inhabitants. In that year, the state of São Paulo (SSP) was responsible for 30% of notified cases of acquired syphilis and 21.5% of syphilis cases in pregnant women in the country 10 .
Despite being the same disease, syphilis in pregnant women and acquired syphilis are treated as different diseases in the Notifiable Diseases Information System (Sistema de Informação de Agravos de Notificação -SINAN) 11 . Consequently, to estimate the magnitude of syphilis and analyze the trend of it in the general population, excluding congenital syphilis, one must take into account the set of cases of acquired syphilis and syphilis in pregnant women. The aim of this study was to analyze how it evolved from 2011 to 2017, in addition to the detection rates of syphilis reported by sex, age group and region of residence in the SSP.

METHODS
An ecological study of a historical series was conducted with secondary notification data including all cases of acquired syphilis and syphilis in pregnant women in the SSP from 2011 to 2017. The main source of data for this study was the SINAN. Data on population estimates were obtained from the São Paulo State Data Analysis Foundation (Fundação Sistema Estadual de Análise de Dados -SEADE) in São Paulo.
SINAN is a nationwide information system, used in Health Surveillance and it has been managed by the Ministry of Health (MoH) since 1998 12 . The municipality is responsible for putting data entry into the system and the data are extracted from notification and investigation forms that are specific to each disease, and which are included in the LNNC of diseases, grievances and public health events 9 . The information is subsequently consolidated and analyzed by local, municipal, state and national levels. The data in this study are related to notifications of acquired syphilis and syphilis in pregnant women. SINAN presents data related to identifying cases of syphilis and syphilis in pregnant women. However, more information about the case only appears for cases of syphilis in pregnant women All reported cases of acquired syphilis (International Classification of Diseases [ICD10] -A53.9) and syphilis in pregnant women (ICD10 -O98.1) were included.
For reporting purposes, the definition of acquired syphilis used was "Every asymptomatic individual or every individual with clinical evidence of primary or secondary syphilis (presence of a hard canker or lesions compatible with secondary syphilis), with a non-treponemal test reagent and any titration and test treponemic reagent" 11 . The definition of syphilis in pregnant women was: "Suspected case -pregnant woman who during the prenatal period presented clinical evidence of syphilis, or a non-treponemal reagent test with any titration. Confirmed case -pregnant woman who had a non-treponemal reagent test and any titration and reagent treponemal test, regardless of any clinical evidence of syphilis, performed during prenatal care; and pregnant woman who had a reactive treponemal test and a non-reactive or unreactive non-treponemal test, with no previous treatment record" 6 . REV BRAS EPIDEMIOL 2020; 23: E200103 The variables used in the study were: detection rates of acquired syphilis (ASDR) per 100 thousand inhabitants, detection rates of acquired syphilis including pregnant women with syphilis (PASDR) per 100 thousand inhabitants, sex, age group, year of diagnosis (2007 to 2017) and place of residence, considering the Regional Health Department (Departamento Regional de Saúde -DRS).
General and sex-specific detection rates were calculated. To calculate the PASDR per 100 thousand inhabitants, the quotient between the number of all cases of acquired syphilis was performed. All cases of syphilis in pregnant women in a given population and year (numerator) were added up, with the respective population and the corresponding year as the denominator. This result was then multiplied by a constant 100 thousand 13 .
The data were disaggregated for analysis according to the following subpopulations: • All cases reported with acquired syphilis; • All cases of acquired syphilis added to all cases of syphilis in pregnant women.
When disaggregating by sex, the following results were obtained: • All female cases of acquired syphilis; • All female cases of acquired syphilis plus cases of syphilis in pregnant women; • All male cases of acquired syphilis; The municipalities of residence were grouped according to the areas covered by the 17 Regional Health Departments (DRS) that make up the administrative structure of the São Paulo State Department of Health: Araçatuba, Araraquara, Baixada Santista, Barretos, Bauru, Campinas, Franca, Greater São Paulo, Marília, Piracicaba, Presidente Prudente, Registro, Ribeirão Preto, São João da Boa Vista, São José do Rio Preto, Sorocaba and Taubaté.
The data used for syphilis came from SINAN -State Program STI/ AIDS-SP 13 . Microsoft ® Excel was used to organize the data.
The first stage of the analysis included the creation of dispersion diagrams generated by STATA software, version 14.0. A residual analysis was performed, fulfilling the assumptions of independence and constant error variance (assumption of true homoscedasticity). The trend analysis of the detection rate curves was conducted using Poisson's Jointpoint (inflection point) model, using the Jointpoint Regression Program, version 4.7.0.0 (http:// surveillance.cancer.gov/joinpoint /) 14 . To assess trends in detection rates, the Joinpoint regression analysis (with up to three inflection points) was used to identify the years (independent variable) in which statistically significant changes in the detection rate trend occurred (variable dependent). This method identified the number of inflection points in a historical series. Each inflection point denoted a statistically significant change in the trend. The degree of this change is presented by means of the annual percentage change (APC). The degree of the general change, over the entire period, is presented by means of the average of the annual percentage change (AAPC). The significance test used was the Monte Carlo permutation method with a Bonferroni adjustment for multiple comparisons and p<0.050 as statistically significant 15 .

REV BRAS EPIDEMIOL 2020; 23: E200103
Using the number of inflections and the respective segments of a curve, the APC per segment and the AAPC of a segmented linear regression of all segments were estimated, including the respective 95% confidence intervals (95%CI).
All data contained in this publication is open and public. Only aggregated data will be presented and no subjects will be identified. It will be freely available for consultation. Due to these characteristics, it was not necessary to undergo an evaluation from the Research Ethics Committee (REC) system under the terms of Law No. 12,527, of November 18, 2011.

RESULTS
In 2017, 516 (80.0%) of the 645 municipalities in São Paulo presented at least one notified case of syphilis (Table 1). Between 2011 and 2017, there was an increase of 86.3% in the number of municipalities with notifications of acquired syphilis and an 78.8% increase in the number pregnant women with syphilis. There was an increase of 119.5% with regard to reporting units.
In 2017, the ASDR was 84.6 per 100 thousand inhabitants, an increase of 225.4% compared to 2011. When adding the numbers of syphilis cases in pregnant women and those of acquired syphilis, a PASDR of 108.9 per 100 thousand inhabitants was calculated in 2017, representing an increase of 223.1% in relation to 2011. The evolution of detection rates according to sex and age group showed a positive percentage increase in all age groups in both sexes between 2011 and 2017. The greatest increase occurred in the age group of up to 24 years old (male -387.1% -and female -570.9%). For PASDR, the greatest increase also occurred in the age group of up to 24 years old (417.0%) ( Table 2).
By analyzing the trend curves of detection rates from 2011 to 2017, there was an inflection point that divided the ASDR and PASDR curve into two periods: from 2011 to 2013 and from 2013 to 2017. In the first period, the APC was 31.8% (95%CI 3.1 -68.6) for    Figure 1). Table 3  Throughout the period, the highest AAPC found was found in females up to 24 years old (36.5%; 95%CI 28.5 -45.1). Table 4 shows the distribution of ASDR and PASDR in 2011 and 2017, according to place of residence -classified by DRS. It was observed that the DRS of São João da Boa Vista and Table 3. Average of the annual percentage and annual percentage variation of the acquired syphilis detection rate and acquired syphilis detection rate including pregnant women with syphilis per 100 thousand inhabitants, according to sex and age group. State of São Paulo, 2011 to 2017*.

DISCUSSION
This study showed an increasing trend and similar behavior of the curves with regards to the rates of detection of acquired syphilis, regardless of the inclusion of syphilis cases in pregnant women, between 2011 and 2017 in the SSP. Additionally, it was highlighted that the period that had the highest APC was from 2011 to 2013. The trend of Detection Rate (DR) in females increased throughout the period from 2011 to 2017, with no inflection point in the curve. The trend curves of DR in young people presented the highest APC averages over the entire period analyzed, in both sexes.
There was an increasing trend in DR, between 2011 and 2017, in the SSP that accompanied the increase in rates throughout Brazil and in the federation units (FU) 10 . The DR of syphilis acquired in the SSP was 1.7 times higher than that reported in all of Brazil and corresponded to the fifth highest rate in the country in 2017 10 . The number of municipalities with resident cases and the number of reporting services also increased positively, pointing to a reduction in underreporting as an explanation for this trend. Another possible explanation is related to the expansion of testing availability in the SSP 16,17 .
The syphilis DR in the SSP was much higher than that found in other countries: the median incidence rate per 100 thousand inhabitants, in 2014, was 25.1 (0.1-1.664) in 55 countries around the globe, and 34.1 (4.3-227.7) specifically in 18 countries in the Americas 18 . In 2016, in Europe, the incidence rates of syphilis in adults ranged from 0.6, in Portugal, to 9.9 per 100 thousand inhabitants, in the United Kingdom 19 . In the United States, the incidence rates of primary and secondary syphilis in 2016 were 27.3 and 9.5 cases per 100 thousand inhabitants, respectively 20 .
It must be taken into account that comparing syphilis notification data between countries is very difficult, especially due to the different definitions of a case and the differences in relation to adherence to reporting. In Brazil, the definition of a case of syphilis includes active cases of the disease, and excludes cases of serological scarring and false positives; however, it was not possible to check for the consistency of compliance with this definition, as there is no data entry for the investigation carried out regarding the reported cases of acquired syphilis 6 . Another issue that must be taken into account in order to make comparisons is that, in Brazil, syphilis notification is subdivided into two conditions: acquired syphilis and syphilis in pregnant women. Therefore, in order to assess the real magnitude of the condition and make comparisons with estimates from other countries, it is important to add the reported cases into these different databases.
The trend curve of the syphilis DR showed the highest APC in the period from 2011 to 2013, which is attributed, especially, to the reduction in underreporting. However, despite the reduction in the second period, the APC increase annually. A possible explanation for this change in the trend in the second period may be due to the expansion of rapid testing coverage for syphilis in the SSP, carried out starting in 2012. In 2017, the state testing campaign was focused initially on the diagnosis of HIV. However, it was expanded to offer syphilis tests. Of the 380,000 rapid tests performed, 31.3% were for syphilis, and 383 municipalities reported at least one reagent syphilis test 16,17 . The expansion of access to the rapid test for syphilis in campaigns and in the routines of basic health units may have increased the detection of this condition and, consequently, the reporting of cases of syphilis in pregnant women and the reporting of acquired syphilis. The hope is that by increasing testing for syphilis, it will be possible to stabilize the rising curve of syphilis.
When disaggregating the analysis by sex and age, it was observed that the behavior of the trend curve was different for females, who did not present an inflection point during the entire period investigated.
The disproportionate involvement of young people up to 24 years of age in both sexes was worth noting. This included pregnant women who reported having syphilis, and who had a higher APC of the detection rate in the observed period, which is very similar to the pattern observed in the entire country 10 . The United Nations (UN) considers young people to be those aged between 10 and 24 years old 21 .
In a study conducted in Brazil, between 2016 and 2017, with 8,562 participants aged up to 25 years old in the 26 Brazilian state capitals and in the Federal District, the prevalence of reporting having syphilis at least once was 12.8%: 10.7% in women and 14.2% in men. The factors associated with STI reporting were: men who have sex with men (MSM) and use of drugs and alcohol 22 . In Brazil, in 2016, a study carried out with 38,247 conscripts, aged 17 to 22 years old, found a prevalence of 1.09% of confirmed syphilis. The following factors were independently associated with syphilis: early sexual activity, report of a previous STI and MSM 23 . In this population, there was an increase in the prevalence of positive syphilis test results from 0.53%, in 2007, to 1.63%, in 2016 23,24 . Another population-based study conducted in the city of São Paulo found 6.3% of participants with a history of STIs (4.3% among women and 8.2% among men) 25 .
One of the main prevention strategies for syphilis transmission is the use of condoms. In research on behavior, attitudes and practices covering the entire country, it was found that there was little variation in the frequency of condom use in young people in all sexual relations in the last 12 months: a reduction from 39.0%, in 2004, to 36.9%, in 2013 26,27 .
In the SSP, in 2017, 80% of the municipalities reported at least one case of acquired syphilis, but when disaggregating by place of residence -DRS of residence -there was heterogeneity in the reporting of the disease. The highest rates were observed in the metropolitan regions of the SSP, and especially in Baixada Santista. With regard to the percentage increase between the detection rates of 2011 and 2017, there was important variability. In two regions of the SSP, percentage increases and rates below the state average were observed. The increase in DR did not occur homogeneously, with possible continuation of REV BRAS EPIDEMIOL 2020; 23: E200103 underreporting in some regions of the State, or regional differences in the expected prevalence. A representative sample study conducted in Brazil from 2011 to 2012, with 23,894 hospital-based parturients, found a prevalence of cases of 1.03% in the Southeast Region 28 , however there are no regional studies on the prevalence of syphilis with a sample that is representative of the general population.
The heterogeneity in the detection rates observed in the SSP regions can also be attributed to the differentiated adherence to other prevention actions developed in the State, such as the expansion of testing for syphilis, in addition to the commitment and adherence of municipal surveillance to notification 16,17 .
One of the main limitations of this study was related to the source of information and the type of study.
The SINAN only includes case identification data: date of diagnosis, date of reporting, sex, date of birth, race and education, and it is not possible to verify compliance with the definition of a "case". When reporting acquired syphilis, it is important to distinguish the primary, secondary and recent latent phases from the others, as these are phases that correspond to the acute disease, approximating the estimated incidence of the disease in the population -as occurs in the surveillance of syphilis in United States 20 .
Additionally, another existing limitation is related to the difference between the case definitions of syphilis in pregnant women and acquired syphilis. In this study, the populations who reported acquired syphilis and syphilis in pregnant women were added together. In the data published on syphilis in pregnant women in 2016 in the SSP, only 641 pregnant women with syphilis did not meet the same case definition criteria for acquired syphilis 13 , which corresponded to only 3% of the women who reported, and did not influence the analysis of the disease trend.
Despite the limitations mentioned, the results obtained are consistent with the literature and support public policies and control strategies, such as expanding the coverage of rapid testing for syphilis, especially in the young population.
This growing trend in syphilis detection rates observed in the period studied can be attributed to the increased availability and implementation of rapid testing in the SSP and better adherence to reporting. The disproportionate increase in the detection of syphilis in young men and women, over the entire period, may represent a real increase in the incidence of the infection in young people. It is urgent to formulate more effective public policies aimed at preventing STIs in young people and actions aimed at reducing the underreporting of syphilis in some regions of the state.