Increasing incidence of syphilis among men living with HIV in Croatia during the COVID-19 years 2020 and 2021

We conducted a nationwide longitudinal observational study to estimate the incidence of syphilis in a cohort of male persons living with HIV (PLWH) in Croatia in the pre-COVID-19 and COVID-19 years. Data were reviewed and extracted from the clinical database. We analyzed 1187 PLWH (> 18 years) in care in Croatia from 2018 to 2021 and used Poisson regression to calculate rates. We observed a 91.4% increase in incidence between 2019 and 2020; the overall rate was 6.0/100 person-years, and the annual rate ranged from 3.3/100 person-years in 2018 to 9.3/100 person-years in 2021. We found higher rates in men who have sex with men, PLWH with a baseline history of syphilis, PLWH with a more recent HIV diagnosis, and a lower rate in those who had clinical AIDS. The rate of syphilis serological testing was 3.5% lower in 2020 compared to 2019. Recurrent syphilis was more likely asymptomatic compared to the �rst episodes. In conclusion, during the COVID-19 epidemic years there was a huge increase in syphilis. Results highlight the need for enhanced and novel prevention interventions.


Introduction
Soon after the rst reported case of the coronavirus disease (COVID-19) caused by a new strain of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in 2019 in China, the world faced a global health crisis.Considering all preventive measures to control the spread of SARS-CoV-2 infection through quarantine and social distancing, it was to be expected that there would be changes in sexual behaviors leading to a decrease in the incidence of sexually transmitted diseases (STDs).However, the data on STDs during the COVID-19 epidemic years is still evolving and the impact of the COVID-19 pandemic on STDs remains to be de ned and might be different in different countries.Some studies reported a decrease in the number of reported STDs particularly during the lockdown periods [1][2][3][4][5][6], whereas other studies found no signi cant change [7,8] or an increase in the number of reported STDs, including syphilis [9][10][11][12][13][14][15][16].
Syphilis is a highly contagious bacterial infection caused by gram-negative spirochaete Treponema pallidum and is typically spread through sexual contact.The World Health Organization estimated that seven million people got infected with syphilis worldwide In 2020 [17], the highest incidence was recorded among men who have sex with men (MSM) [18].
In the most recent European Centre for Disease Prevention and Control (ECDC) surveillance report 35 039 con rmed syphilis cases were reported in 29 EU/EEA Member States, with a crude noti cation rate of 7.4 cases per 100 000 population [19].The majority of cases (73%) were reported in MSM, and the trends in MSM showed a steep increase in reported cases from 2010 up to 2017, with a much slower increase between 2018 and 2019 [19].Syphilis is a noti able disease in Croatia, however, because of underreporting of cases the noti cation reports from the Croatian Nation Public Health Institute do not seem reliable.In the ECDC report the annual number of reported cases of syphilis ranged from 25 in 2015 to 35 in 2018 in Croatia.There have been two studies that estimated the prevalence of past syphilis among MSM measured by Treponema pallidum hemagglutination in Croatia.Both studies used a respondent-driven sampling method and reported a prevalence of syphilis of 10.6% in 2006 and 7.6% in 2011 [20].The aim of this study was to estimate the incidence of syphilis diagnoses in a cohort of male persons living with HIV (PLWH) in Croatia in the pre-COVID-19 and COVID-19 years and to identify syphilis risk factors in this population.

Setting
In Croatia, all PLWH are treated at the University Hospital for Infectious Diseases (UHID) in Zagreb.The HIV Outpatient Department at UHID has been established in 1997, it currently provides outpatient care for PLWH including treatment for common STDs such as gonorrhea, chlamydia, and syphilis.Antiretroviral therapy is also given from the hospital pharmacy.Because of the distance from the residence of PLWH to the treatment center, phone consultations were part of routine care since 1997 and a mHealth solution delivering laboratory results to the smartphones of PLWH has been introduced in 2016 [21].Antiretrovirals can also be delivered to the home address, this is usually done after a phone consultation.
Serologic testing for syphilis at UHID is usually routinely done annually or bi-annually, however, it may be done more or less frequently depending on the clinical ndings and sexual behavior history.Since 2004, the testing algorithm for syphilis screening includes rst a treponemal test (the Treponema pallidum hemagglutination assay [TPHA]), followed by a nontreponemal test (the Rapid Plasma Regain [RPR] test).

Population and de nitions
We studied all adult men (> 18 years) PLWH receiving care with at least 1 year of follow-up (n = 1187, Fig. 1).The observation period started on or after January 1st, 2018, and ended on December 31st, 2021, or earlier if PLWH died, moved, or were lost to follow-up.Because no women had syphilis, we restricted our analysis to men (Fig. 1).The person's data, including demographic characteristics, diagnoses, prescribed medications, number of syphilis tests, and laboratory values, were reviewed and extracted from the electronic database.

Statistical analysis
We used descriptive statistics to characterize our study population.Baseline socio-demographic, laboratory, and clinical characteristics were compared between PLWH with and without incident syphilis with chi-square tests or Fisher's exact tests for categorical variables and Wilcoxon rank-sum tests for continuous variables.The chi-square trend test was used to examine the proportion of asymptomatic syphilis cases from 2018 to 2021.We also examined whether the rst syphilis episode after inclusion into care was more likely asymptomatic in PLWH with a baseline history of past or current syphilis.
Using Poisson regression, we calculated crude incidence rates for incident syphilis ( rst and repeated episodes) by sociodemographic and clinical characteristics.To examine factors related to syphilis diagnoses, incident rate ratios with 95% con dence intervals (CI) were calculated.The calendar year was annualized, and we used a generalized estimating equations (GEE) model with an exchangeable correlation structure to account for recurrent events.Age, viral load, prior episode of syphilis, clinical AIDS and hepatitis C infection on follow-up were time-updated variables whereas other variables were xed.We modeled the total number of events using an offset for the person-year follow-up time.Crude analyses included the outcome and one predictor variable whereas multivariable analyses included more than one predictor.Statistical analyses were done using SAS version 9.4 (SAS Institute, Inc., Cary, North Carolina).
P-values were two-sided and statistical signi cance was set at the conventional P-value of < 0.05.

Descriptive analysis
Of 1187 participants three were transgender women and all, except one man, were Caucasians.The HIV transmission risk was: sex between men (n = 976, 82.2%), heterosexual (n = 142, 12.0%), injection drug use (n = 28, 2.4%), recipient of blood or blood products (n = 5, 0.4%), mother to child transmission (n = 4, 0.3%) and unknown (n = 32, 2.7%).At baseline, the median age was lower among PLWH who had incident syphilis compared to those who did not (40.4 years versus 42.0 years) (Table 1).PLWH who had syphilis on follow-up were more likely MSM, living in Zagreb, diagnosed with HIV in the period 2016 to 2020, and did not have clinical AIDS (Table 1).The median CD4 lymphocyte count at baseline was high (599.5 per mm 3 ), 75.2% had < 50 copies of HIV1-RNA per ml, and 91.8% were ART-experienced (

Longitudinal analysis
Of 1187 PLWH, 197 (16.6%) experienced 254 rst-time or reinfection syphilis events during 4224.2person-years of follow-up (PYFU).The median follow-up was 4.0 years, ( rst quartile [Q1] to third quartile [Q3], 3.3-4.0years), and the overall incidence rate was 6.0/100 PYFU (95% con dence interval [CI] 5.3-6.8).Of 254 syphilis events, 116 (45.7%) were the rst-ever diagnoses.The incidence rate for rst-ever syphilis events was 5.0 (95% CI, 4.2-5.9)per 100 PYFU, and for the recurrent event, it was 8.3 (95%, 6.8-10.0)per 100 PYFU.There was an increase in the number of PLWH with syphilis and an increase in the incidence rate in the period 2018 to 2021 (Table 2, Fig. 2A).This increase was particularly high between 2019 and 2020 (91.4%,Table 3).Evidence of syphilis at baseline or on follow-up.IRR are based on a generalized estimating equations model with an exchangeable correlation structure.The overall incidence rate per 100 person-years was 6.0 (95% CI, 5.3-6.8).The 95% con dence intervals for the rate per 100 personyears of observation was calculated using the Rothman/Greenland estimation.The number of syphilis events, follow-up time, events per 100 PYFU, and incidence rate ratios are shown in Table 2 and Supplementary Fig. S1.On crude analysis we observed evidence of a higher rate of a new episode of syphilis among MSM, those aged less than 40, those diagnosed with HIV and syphilis more recently, living in Zagreb and in PLWH who had past evidence of syphilis (Table 2, Supplementary Fig. S1).Syphilis diagnoses per 100 person-years were similar in the age time-updated groups 18 to 29 years (N = 26, rate = 6.9 [95% CI, 4.7-10.1])compared to the 30 to 49 years (N = 169, rate, 7.0 [95% CI, 6.0-8.1]) and higher compared to those older than 50 (N = 59, rate, 4.2 [95% CI, 3.2-5.4]).On multivariable analysis, many of the relationships found on crude analysis were less pronounced, statistical signi cance was found for MSM versus others, more recent calendar year of HIV diagnosis (2016 to 2020 versus ≤ 2010), more recent calendar years of syphilis diagnosis, and having a previous episode of syphilis (Table 2, Supplementary Fig. S1).A previous clinical AIDS diagnosis was associated with a decreased risk of having syphilis on follow-up (Table 2, Supplementary Fig. S1).The annual rate of asymptomatic and symptomatic syphilis diagnoses and the percent increase compared to the previous years are shown in Table 3.
We performed several sensitivity multivariable analyses which included only observations with at least one syphilis test done in a calendar year or included only MSM (Supplementary Figures S2-S4).In those analyses, the rates of syphilis events were higher than in our main analysis.The overall rate per 100 PYFU was as follows: 7.2 (95% CI, 6.  2B).The proportion of PLWH who had at least one syphilis test per calendar year ranged from 80 to 90% and was 5.8% lower in 2020 compared to 2019 (Fig. 2B).

Clinical presentation
The major clinical ndings are described in Table 4. Overall, of 254 syphilis episodes 154 (60.6%) were symptomatic, secondary syphilis was diagnosed in 90 (35.4%), and primary in 64 (25.2%) cases (Table 4).The proportion of asymptomatic cases ranged from 35.5% in 2018 to 42.7% in 2021, however, this trend was not statistically signi cant (P = 0.483, chi-square trend test).The use of intramuscular benzathine penicillin G 2.4 million units in a single dose for the treatment of primary, secondary, and early latent syphilis increased from 2018 to 2021 (Table 4).Of 116 rst syphilis episodes 81 (69.8%) were symptomatic whereas of 138 recurrent episodes 73 (52.9%) were symptomatic (P = 0.006) (Fig. 3A and 3B).PLWH with syphilis at baseline or a history of previous syphilis were more likely to have an asymptomatic rst syphilis event on follow-up (OR 2.49, 95% CI, 1.38-4.49).The median number of serological tests before the rst syphilis event on follow-up was the same in PLWH with a baseline history of syphilis and those who did not have past syphilis (3, Q1-Q3, 2-4).a Treatment of primary, secondary and early latent syphilis with one shot of benzathine benzylpenicilline 2.4 million I.U.im.
Values are frequencies and percentages in parentheses.

Discussion
In this nationwide study of men living with HIV, the incidence of syphilis increased by 91.4% in 2020 compared to 2019, and this increase was mainly driven by an increase in symptomatic cases (106.1%)(Table 3).We also found a 30.7% increase in the incidence of syphilis in 2021 compared to the previous year, but this increase was larger in asymptomatic syphilis cases.In our models, the incident rate of syphilis was at least > 2.5 times higher in 2021 compared to 2018.The increase in syphilis between 2020 and 2019 occurred in the background of a somewhat higher syphilis testing rate in 2019 (1.14 tests per one person-years) compared to 2020 (1.10 tests per one person-years).
The overall incidence rate of syphilis in our men living with HIV was 6.0/100 PYFU and the annual incidence rate ranged from 3.3/100 PYFU in 2018 to 9.3/100 PYFU in 2021.Similarly, to our ndings in 2021, a retrospective cohort study from Australia reported an incidence of early syphilis of 9.3 per 100 person-years in MSM living with HIV in the period 2013-2019 [22].Another study from Thailand found an incidence of syphilis of 10.2 per 100 person-years in a predominantly MSM cohort diagnosed with acute HIV infection in the period April 2009 through December 2018 [23].A very high incidence of syphilis in PLWH was reported from Mexico (16.0/100PYFU, years covered: 2011-2015) [24] and in MSM living with HIV from Argentina (14.9/100PYFU) in the period 1 March 2015 to 29 February 2016 [25].Some studies reported somewhat lower incidence rates in PLWH than ours.A study from the HIV Swiss cohort reported data from October 2017 to November 2019 and found an incidence of syphilis of 4.4 cases per 100 person-years in male PLWH and of 6.3 cases per 100 person-years in MSM [26].A recent analysis from the French Dat'AIDS HIV cohort including men and women, found an incidence of syphilis of 3.8 per 100 person-years for infection and 6.5 per 100 person-years for re-infection in the period 2010 to 2019 [27].Data from four US clinical sites reported an incidence rate of 5.6/100 PYFU in men and 6.9/100 PYFU in MSM in the period 2014 to 2018 [28].There are reports on an increasing number of cases of syphilis in some countries during the COVID-19 years 2020 and 2021 [9][10][11], however, data on the incidence density rates are lacking.
This huge increase in syphilis in PLWH occurred in Croatia one to two decades after similar outbreaks were reported in North America and Western Europe.Outbreaks of syphilis among MSM have been reported from the US since 2000 [29] and ECDC reported a sharp increase (> 50%) of noti able cases in several European countries (Germany, Denmark, Greece, Luxembourg, Malta, Norway, Portugal, and Sweden) between 2008 and 2013 [30].This increase in syphilis cases in Europe continued for at least up to 2019 [19].ECDC reported that 68% of syphilis cases were in MSM in 2019 and that among the countries that reported the HIV status 43% of cases were in PLWH [19].It has been suggested that the high incidence of syphilis among MSM and PLWH could be due to risk perception, risky sexual behaviors such as unprotected anal sex, occasional sexual partners, concurrent sexual partners, HIV serosorting, chemex, transactional sex, methamphetamine use, group sex or contacting partners over the internet, and anonymous couples [31][32][33][34][35] HIV cluster of which the largest included 53 persons [39].We could expect that similar clusters of syphilis have occurred during the current outbreak.
Risk factors associated with syphilis such as being MSM and having a history of previous syphilis have also been reported previously [22,23,27,[40][41][42][43].Younger age was also associated with syphilis; however, this relationship was not as strong as in other studies [40,43].Interestingly, in the French Dat'AIDS cohort, age > 35 years was associated with a rst or re-infection syphilis episode [28] whereas, in the recent study from the US, syphilis incidence was highest in the 30 to 39 age group [28].Of note, having a past history of clinical AIDS was related to a reduced risk for syphilis, this was also found in the univariable but not multivariable analysis in the French Dat'AIDS cohort [27].There was no association of detectable HIV-1 viral load and syphilis which implies that the risky sexual behavior leading to syphilis will not increase in HIV cases.Some recent studies found an association of HCV infection with syphilis [23,28]; this was not the case in our study possibly because of a small number of HCV infections.Living in Zagreb the capital of Croatia was associated with syphilis mainly in our univariable analyses, syphilis outbreaks are especially common in large cities with signi cant populations of MSM [23,24,28,37,42].The incidence of syphilis was higher in persons diagnosed with HIV in more recent years which might suggest that newer generations of PLWH have a different perception of risk and risky behavior.Similarly in a recent study from four US sites, a more recent entry into the cohort was associated with a higher incidence of syphilis [28].
There was overall an increasing trend in the syphilis serologic testing rate from 2018 to 2021, however, the rate was lower in 2020 by 3.5% compared to 2019.Similarly, the proportion of men who had at least one annual syphilis test was lower in 2020 compared to 2019.It seems that measures to curb the COVID-19 epidemic had a negative impact on syphilis testing, similar suggestions on sexual health screening and patients' referrals to STD clinics have also been made [44,45].We have con rmed ndings from others [46,47] that repeated episodes of syphilis are more frequently asymptomatic.However, it is still debated whether this observation is a result of an immune response to a previous infection or an increased frequency of serological testing [47].We found an increased likelihood of early latent syphilis in PLWH with a repeated syphilis episode in the background of a similar testing rate.
Our study has limitations.This was a retrospective longitudinal observational study, so causal inferences cannot reliably be made.We did not collect data on speci c sexual practices, condom use, or other risky behaviors such as chemsex, methamphetamine use, group sex, transactional sex, or use of dating sites to nd new partners.Hence, we cannot report on the risky behaviors that contributed to the rise in syphilis cases.

Conclusions
We report a 91% rise in the incidence of syphilis in men living with HIV in Croatia between 2019 and 2020.MSM were mainly affected.Traditional syphilis control strategies such as timely diagnosis and treatment, partner noti cation and treatment, and education/awareness campaigns might not be su cient to curb the epidemic in PLWH.We should strongly consider introducing doxycycline preexposure prophylaxis which has been shown to signi cantly reduce syphilis in 3 randomized trials [48][49][50].Since many cases of syphilis are asymptomatic, serological screening should be increased particularly in those who had a previous syphilis episode.We showed that repeated episodes of syphilis are more frequently asymptomatic and not related to the frequency of testing.Longitudinal studies on the host humoral and cellular immune response in syphilis during index and repeated syphilis episodes could inform us about the underlying mechanisms of different clinical presentations.

Figure 1 Flow
Figure 1

Figure 2
Figure 2 Incident syphilis was de ned as a new syphilis diagnosis, it included both rst-time and reinfection events.We de ned incident cases of syphilis based on clinical diagnosis, prescribed treatment, and serological testing.Syphilis was considered present if any of the following were observed: both a documented syphilis clinical diagnosis and treatment for syphilis, a positive syphilis TPHA screen when a previous test was negative, and a 4-fold increase in RPR titer.Episodes of syphilis at inclusion into the study (n = 18) were not considered incident cases.Past history of syphilis at inclusion into care was de ned if treatment was recalled by the patient or the TPHA test was positive.
The Ethics Committee of the University Hospital for Infectious Diseases (UHID), Zagreb, Croatia approved the study.The study was performed following the Declaration of Helsinki and the Code of Medical Ethics

Table 1
Main baseline characteristics of 1187 men living with HIV with and without incident syphilis on follow-up, Croatia, 2018-2021.Clinical AIDS in the baseline calendar year b 61 had chronic, 3 had acute hepatitis B. c Based on history and/or positive Treponema pallidum hemagglutination assay.
).Of 284 PLWH who had clinical AIDS, 6 had clinical AIDS in a follow-up calendar year.A total of 67 PLWH had hepatitis C infection (HCV), 62 at baseline, and 5 acquired HCV infection during follow-up.There were no new hepatitis B infections during the follow-up.Values are frequencies with percentages and median with rst and third quartile in parenthesis.MSM, men who have sex with men.ART, antiretroviral therapy.aClinical AIDS in the baseline calendar year b 61 had chronic, 3 had acute hepatitis B. c Based on history and/or positive Treponema pallidum hemagglutination assay.Values are frequencies with percentages and median with rst and third quartile in parenthesis.MSM, men who have sex with men.ART, antiretroviral therapy.aClinical AIDS in the baseline calendar year b 61 had chronic, 3 had acute hepatitis B. c Based on history and/or positive Treponema pallidum hemagglutination assay.a

Table 2
Rates of Incident Syphilis in men living with HIV by Sociodemographic and Clinical Characteristics and Univariable and Multivariable Syphilis Diagnosis, Croatia, 2018 to 2021.
MSM, men who have sex with men; IRR, incidence rate ratio; CI, con dence intervals; a Time updated; b Not included in the multivariable regression model c At baseline, d Evidence of syphilis at baseline or on follow-up.IRR are based on a generalized estimating equations model with an exchangeable correlation structure.The overall incidence rate per 100 person-years was 6.0 (95% CI, 5.3-6.8).The 95% con dence intervals for the rate per 100 personyears of observation was calculated using the Rothman/Greenland estimation.MSM, men who have sex with men; IRR, incidence rate ratio; CI, con dence intervals; a Time updated; b Not included in the multivariable regression model c At baseline, d Evidence of syphilis at baseline or on follow-up.IRR are based on a generalized estimating equations model with an exchangeable correlation structure.The overall incidence rate per 100 person-years was 6.0 (95% CI, 5.3-6.8).The 95% con dence intervals for the rate per 100 personyears of observation was calculated using the Rothman/Greenland estimation.MSM, men who have sex with men; IRR, incidence rate ratio; CI, con dence intervals; a Time updated; b Not included in the multivariable regression model c At baseline, d

Table 3
Rates of new symptomatic and asymptomatic syphilis among men living with HIV, Croatia, 2018 to 2021.

Table 4
Major clinical ndings and treatment of 254 episodes of syphilis in men, Croatia, 2018 to 2021.
. Cl37]d sexual networks are important contributors to syphilis outbreaks[36,37].As we have not collected data on sexual risks, we do not know whether those practices changed during the COVID-19 pandemic.However, lockdown measures and measures of social distancing with other precautions might have given a false sense of security.In Croatia, the sexual network dynamic may have changed during 2019 and 2020 leading to a local syphilis epidemic.There is little data on risk behaviors in PLWH from Croatia, however, a study conducted in 2005/2006 reported that 20% of MSM