Temporal trend in the incidence of tuberculosis-HIV coinfection in Brazil, by macro-region, Federative Unit, sex and age group, 2010-2021

ABSTRACT Objective To analyze the temporal trend in the incidence of tuberculosis-HIV coinfection in Brazil, by macro-region, Federative Unit, sex and age group, from 2010 to 2021. Methods: This was a time series study using surveillance data to estimate average annual percentage changes (AAPC), and 95% confidence intervals (95%CI) via joinpoint regression. Results: 122,211 cases of tuberculosis-HIV coinfection were analyzed; a falling trend was identified for Brazil as a whole (AAPC = -4.3; 95%CI -5.1;-3.7), and in the country’s Southern (AAPC = -6.2; 95%CI -6.9;-5.5) and Southeast (AAPC = -4.6; 95%CI -5.6;-3.8) regions, even more so during the COVID-19 pandemic (2020-2021); the greatest falling trend was seen in Santa Catarina (AAPC = -9.3; 95%CI -10.1;-8.5), while the greatest rising trend was found in Tocantins (AAPC = 4.1; 95%CI 0.1;8.6); there was a rising trend among males, especially in Sergipe (AAPC = 3.9; 95%CI 0.4;7.9), and those aged 18 to 34 years, especially in Amapá (AAPC = 7.9; 95%CI 5.1;11.5). Conclusion The burden and trends of tuberculosis-HIV coinfection were geographically and demographically disparate.


INTRODUCTION
T u b e r c u l o s i s ( T B ) a n d h u m a n immunodef iciency virus (HIV) infection overburden health systems, especially in countries with less availability of economic, human and structural resources. 1International agreements, expressed through the United Nations Sustainable Development Goals (SDGs), have been established to end the HIV and TB transmission -and therefore, TB-HIV coinfection as public health problem by 2030. 1,2 is estimated that a quarter of the world's population is infected with TB.These are cases of infection that, eventually, can progress to the disease itself. 3Furthermore, TB persists as one of the main infectious causes of mortality in the global population, especially among those living with HIV. 3 These people, compared to those not infected with HIV, have a high risk, up to 20 times greater, of progression from TB-infection to TB-disease, in addition to being more susceptible to unfavorable TB outcomes, such as death. 1,4e World Health Organization (WHO) compiled a list of 30 countries with the highest burden of TB-HIV coinfection.Developing countries and those with the largest populations, including Brazil, stood out on that list. 3Worldwide, in 2021, of the 6.4 million TB cases registered, 6.7% were people living with HIV and death was the outcome for 187,000 of them. 3In Brazil, the proportion of TB-HIV coinfection was 10.3% in 2019, with variations between national macro-regions and states. 57][8][9] Furthermore, the absence and/or non-adoption of resources for prevention, diagnosis and treatment, both at the individual and programmatic level, may be related to higher incidence of dual infection. 6,7 this sense, the layout and regionalization of the health care network and its sociospatial, economic and political inequalities must be considered.[12] It is necessary to consider the way in which different contexts influence the incidence of TB-HIV coinfection, especially regarding infections

METHODS
This was an ecological study of time series of TB-HIV coinfection incidence in Brazil, by macro-region (North, Northeast, Midwest, South and Southeast) and FU, based on data from the Mortality Information System (Sistema de Informações sobre Mortalidade -SIM), the Notif iable Health Conditions Information System (Sistema de Informação de   13 The study population consisted of new cases reported on the SINAN-TB system, regardless of clinical form, with the "HIV" variable coded as "positive" or the "AIDS" (acquired immune def iciency syndrome) variable coded as as "yes"; or TB cases notified on the TB databases without one of these variables having been filled out, but for whom diagnosis had been recorded on the HIV databases, or who had a laboratory result on the SISCEL, or who had antiretroviral medication dispensation recorded on the Siclom. 5ses from 2010 to 2021 were included in the study, considering DATHI availability of data on people aged 18 to 59, given that this age group corresponds to the majority of cases of TB-HIV coinfection (± 91.9%); children, adolescents and elderly people were not included, as they have particularities that would make it impossible

ORIGINAL ARTICLE
Tuberculosis-HIV coinfection trend in Brazil (2010-2021)   to understand these specif icities.Twelve records with the "sex" variable not filled out were excluded, given that this was one of the variables analyzed by this study.Initially, we obtained crude incidence coefficients, year by year, by dividing the total number of new cases of TB-HIV coinfection by the resident population, in the same period and location; and multiplying the result by 100,000 inhabitants.After exploratory analysis of the data, we decided to calculate the incidence coefficients by sex (male; female) and age group (in years: 18-34; 35-59), taking the denominator as the population with the same demographic characteristics.
Trend analysis was performed using joinpoint regression, which allows checking whether straight segments would better explain the series than a single straight line.Given that twelve points were analyzed (one point for each year), we def ined a maximum of two joinpoints, as established in the literature. 14The overall and stratified trend (sex and age group) was estimated for each macro-region and FU, assuming the influence of individual and programmatic aspects on the epidemiology of infections.
The annual incidence coeff icients of TB-HIV coinfection, transformed by a natural logarithmic function (ln) due to better interpretation and comparison of results, were taken as the dependent variable (y); while the calendar years of the period were taken as the independent variable (x).The log-linear models [ln(y) = x'beta + error] were adjusted by the standard errors of the incidence coefficients and by correcting first-order autocorrelation, verified based on the data. 14e final models, estimated via grid search, were chosen by the lowest value of the weighted Bayesian information criterion.For each final model, we used the quantile-empirical method to calculate, (i) annual percentage change (APC), referring to the change in the values of the incidence coefficients at each joinpoint, (ii) average annual percentage change (AAPC), relative to the geometric averages of the APCs, and (iii) the 95% confidence intervals (95%CI) of the APCs/AAPCs. 14en interpreting the calculated values, positive APCs/AAPCs indicated a rising trend in TB-HIV coinfection incidence coefficients, while negative APCs/AAPCs indicated a falling trend.APCs/AAPCs values with 95%CIs that did not include the null value (zero) were considered to be significant.Non-significant changes were interpreted as having a stationary trend.The analyses were performed using version 5.0.2. of the Joinpoint Regression Program®. 14

RESULTS
We analyzed 122,211 new cases of TB-HIV coinfection notified between 2010 and 2021 in the population aged 18 to 59 years in Brazil.The annual incidence coefficients of TB-HIV coinfection in the period, for each national macro-region and for the country as a whole, are shown in Figure 1.Table 1 presents the crude incidence coefficients of dual infection, year by year, by FU.
The analysis by FU showed a rising trend in TB-HIV coinfection incidence in Tocantins, Sergipe and Espírito Santo, throughout the time series.The states of Acre, Piauí, Pernambuco, Bahia, Minas Gerais, Rio de Janeiro, São Paulo, Paraná, Santa Catarina, Rio Grande do Sul, Mato Grosso and Goiás had fallings trends, from 2010 to 2021.The majority of FU (14; 51.8%) recorded a drop in TB-HIV coinfection incidence with effect from 2018 or 2019 (Table 2).
A falling trend was identif ied in the coefficients among the female population, in Brazil as a whole and its macro-regions, from 2010 to 2020.A greater increase in trends was seen for Maranhão (AAPC = 3.3; 95%CI 1.3;5 .6)and greater decrease for Acre (AAPC = -15.8;95%CI -27.6;-9.0).Some FU (seven) registered positive APCs in incidence among the female population, such as Acre, Amazonas, Bahia and Espírito Santo (Table 3).
As for the male population, there was a falling trend in incidence coefficients in Brazil as a whole and in most macro-regions; the exception was the Northern region, where the trend proved to be stable.The most pronounced rising and falling trends were, respectively, in Amapá (AAPC = 5.6; 95%CI 3.3;8.7)and Santa Catarina (AAPC = -8.7;95%CI -10, 7;-7.4).Nine FU -including Pará, Rondônia, Amazonas, Piauí and Espírito Santo -registered positive APCs (Table 3).TB-HIV coinfection incidence in the 18-34 age group showed a falling trend for Brazil as a whole, and also in the South, Southeast and Northeast macro-regions.Still with regard to this age group, a greater increase and a greater decline in TB-HIV incidence were seen, respectively, in Amapá (AAPC = 7.9; 95%CI 5.1;11.5)and in Santa Catarina (AAPC = -9.7;95%CI -12.0;-7.7).Furthermore, for the same age group, ten FU had positive APCs in segments of the series, such as Amazonas, Mato Grosso, Ceará, Alagoas and Rio Grande do Norte (Table 4).
In the population aged between 35 and 59 years, there was a falling trend in the dual infection incidence coefficients in Brazil as a whole and in all its macro-regions.None of the states presented positive AAPC, with Santa Catarina (AAPC = -9.4:95%CI -10.3;-8.7)being the state that showed the greatest falling trend in TB-HIV incidence in this age group; ten FU recorded periods with positive APCs, such as Pará, Maranhão and Sergipe (Table 4).

DISCUSSION
This TB-HIV coinfection time series study revealed that the states of Rio Grande do Sul, Amazonas, Pernambuco and Santa Catarina had the highest levels of incidence in Brazil between 2010 and 2021.Falling trends were seen mainly in the FU in the Southern and Southeastern regions.Increases in incidence were recorded, especially in the male population and in those aged 18 to 34 years.There was a downward trend in most FU during the COVID-19 pandemic.
The Brazilian response to the HIV and TB epidemics, historically, is related to political and budgetary issues.It is known that since 2013 Brazil has faced unprecedented socioeconomic adversities, the rise of inequalities and the impact of these obstacles on the Brazilian National Health System (Sistema Único de Saúde -SUS), 15 associated with aspects of regional disparity in the distribution of and access to services. 11,12This is a reality to be considered when interpreting the different trends found in this study.
It is essential to recognize the development of collaborative strategies between TB and HIV programs in Brazil, such as: 5,16,17 recommending antiretroviral therapy (ART) for people with HIV, regardless of their lymphocyte count, with effect from 2011; incorporation of rapid molecular testing for TB into the health care network in 2014; and strengthening the detection and treatment of TB infection, especially in people living with HIV, following the publication of the surveillance protocol in 2018.
Internationally, in the United Kingdom, a falling trend in TB-HIV coinfection incidence was also seen between 2000 and 2014. 18hat fall was linked to the increase in the lymphocyte count threshold for starting ART in 2008, contributing to the improvement of quality of life of people with HIV and reducing susceptibility to TB. 18 In Brazil, it is assumed that the indication of ART for all people with HIV may have influenced the reduction in cases of coinfection with TB.
The decrease in the levels of TB-HIV coinfection incidence may also be related to the expansion of TB infection diagnosis and treatment actions, both among the general population and among people living with HIV.This is one of the fundamental strategies for eliminating TB as a public health problem, in Brazil and around the world, 19 as it can result in a reduction in the incidence of TB disease cases and contribute to interrupting the Koch bacillus transmission chain. 20 is also important to highlight the crucial role that pre-exposure prophylaxis (PrEP) can play in addressing the HIV epidemic.It is a fact that the expansion of prevention options in Brazil, since 2017, such as PrEP and methods from the perspective of combined prevention, can culminate in effective control of infection. 21herefore, it is inferred that the reduction in new HIV cases could reduce the incidence of TB coinfection, due to the smaller number of potentially susceptible people.
The growing trend in TB-HIV coinfection, however, is linked to (i) the greater circulation of etiological agents, which favors the transmission of infections, and/or (ii) the greater supply of tests to detect HIV and TB, which increases the

ORIGINAL ARTICLE
Tuberculosis-HIV coinfection trend in Brazil (2010-2021) number of people diagnosed.In Brazil, testing for HIV in people with TB and investigating TB in people with HIV (dual testing) is a strategy of the health care network, especially in Primary Health Care and Specialized Care Services. 16 this sense, it is considered that the expansion of TB testing among people with HIV can lead to substantial increases in the number of cases of coinfection, culminating in rising trends in incidence.As an example, a study carried out in a region of Ghana showed that the strengthening of collaborative actions between TB and HIV control programs, such as the provision of tests, resulted in significant periods of increased coinfection in the time series from 2008 to 2018. 22e dual testing policy is particularly relevant in Brazil, given the high number of people who discover their HIV infection only as a result of TB.In 2020, data resulting from linkage between the SINAN, SIM, SISCEL and SICLOM databases showed that 47.9% of registered coinfection cases were diagnosed with HIV due to TB. 19 This is a warning sign for possible late diagnosis of HIV, which affects just over a quarter of the cases of infection registered in Brazil. 23 should be highlighted that identification of a stationary trend in TB-HIV coinfection raises another alert, regarding possible weaknesses in TB care, such as ineffective contact assessment and active tracing.This situation results in an epidemiological plateau, given that, even with the diagnosis and treatment of people with TB disease, those with TB infection can go unnoticed and, eventually, progress to the active form, 20 sustaining the TB transmission chain among people living with HIV.
In addition to territorial disparities, there was a significant increase in cases of TB-HIV coinfection in the population aged 18 to 34.In the Brazilian context, young people have accounted for the majority of cases of HIV infection, mainly because of risky health practices, such as (i) starting one's sex life early, (ii) inconsistent use of prevention methods, such as condoms and PrEP, (iii) low level of education, (iv) intercourse with multiple partners and (v) use of alcohol and drugs, among others. 24rthermore, issues of sex, gender identity and sexual orientation must be considered, which may be related to the risk of HIV infection in Brazil, causing gay men and other men who have sex with men (MSM) to be considered as key populations for addressing the epidemic. 25dded to this, there is the issue of male vulnerability to TB: they are the most affected by TB infection in Brazil, accounting for 70% of new cases registered between 2020 and 2022. 26is intersection highlights the need for HIV and TB programs to promote strategies that take into account the social issues linked to infection.For example, the Interministerial Committee for the Elimination of Tuberculosis and Other Socially Determined Diseases (Comitê Interministerial para a Eliminação da Tuberculose e de Outras Doenças Determinadas Socialmente -CIEDDS) has been set up in Brazil, by Decree No. 11494, dated April 17, 2023.The CIEDDS aims to promote intersectorial actions to eliminate TB and other socially determined diseases, such as HIV.
In addition to the strategies adopted to control TB-HIV coinfection in Brazil, the aim is to develop specific actions, from the perspective of holistic care: organization of the line of care, for timely initiation and adherence to ART; dual testing and reduction of late diagnosis of infections; access, linkage and retention of affected individuals, for follow-up in health services; and detection and treatment of TB among the general population and among those living with HIV -including preventive treatment.
In Latin America and the Caribbean, diagnosis and treatment policies, such as those described above, have been adopted in more than 80% of countries; however, there are flaws that weaken the monitoring of TB-HIV coinfection, such as the lack of simultaneous integration of case notification, which affects the quality of care provided. 27Nevertheless, recognition must be given to the effort made by the Ministry of Health in carrying out periodic health information system linkage to improve information at the national level.
T h e COVI D -1 9 p a n d e m i c m a y h a ve accentuated the obstacles faced by the SUS, as it hampered access to diagnosis and compromised surveillance actions, resulting in a drop in TB-HIV coinfection incidence rates, as identified in this study.The pandemic also made TB monitoring and treatment difficult and interrupted follow-up activities for people living with HIV in Brazil, 28,29 which would explain the falling trends seen at the end of the time series (2019 to 2021).
It should also be noted that interruptions in the provision of TB and HIV services, resulting from the emergence of the COVID-19 pandemic, could result in significant increases in morbidity and mortality associated with infection in the coming years. 30As such, in addition to underdetection and/or underreporting of cases of dual infection in the Brazilian scenario, another warning is highlighted in this report, given the possibility of there being recorded a substantial number of deaths and years of potential life lost as a result of TB and/or HIV. 30 It is important to point out that this research has limitations.Firstly, the use of secondary data may be subject to incorrect and/or incomplete f illing out, in addition to different levels of underreporting in the territories, so the findings presented here may be underestimated in certain locations.Other limitations of this study would be the fact that (i) the models were not adjusted for confounding factors, which could influence the trends, (ii) the non-inclusion of cases f rom the entire Brazilian population, being restricted only to people aged 18 to 59, and (iii) the data being aggregated at the state level, preventing the understanding of dynamics at the municipal level.
Moreover, it should be noted that, although linkage between the SINAN, SIM, SISCEL and SICLOM databases results in an annual increase of around a thousand cases of TB-HIV coinfection, in relation to data produced by the SINAN alone, 5 the greater or lesser involvement of local surveillance services in notifying cases, both TB and HIV, may influence the downward or upward trends found in this study.All of this conf irms the importance of adequate recording and timely notification of cases of TB-HIV coinfection to ensure quality information.
In short, demographic and territorial disparities were evident in the trends of TB-HIV coinfection incidence in Brazil.Rising trends were seen, especially in the North and Northeast regions, among males and in the population aged 18 to 34 years.There was a reduction in incidence rates for most FU between 2019 and 2021, which points to the possible effects of the COVID-19 pandemic on diagnosis of TB and HIV.
Without disregarding the possible impact of the pandemic on the progress achieved so far, at a national level, the findings of this study can contribute to the planning of actions to control TB-HIV coinfection in the most affected territories and groups in Brazil.Therefore, the information presented can support the implementation or readjustment of state and national public policies, with a view to reversing the epidemiological scenario and achieving better conditions in Brazilian public health.