According to the Joint United Nations Program on HIV/AIDS (UNAIDS), sexual transmission remains the primary route of HIV transmission, accounting for 70% of new cases of people living with HIV in 2021. In 2021, 1.5 million people have been infected with HIV worldwide (UNAIDS, 2022).

The risk of HIV infection is directly related to the HIV plasma viral load of people living with HIV (LeMessurier et al., 2018). Evidence demonstrates that HIV transmission is very low among people living with HIV taking antiretroviral therapy (ART) and with undetectable plasma viral load for at least 6 months (Bavinton et al., 2019; Rodger et al., 2016).

People living with HIV taking ART experience low levels of virus circulation, which eventually interrupts the chain of HIV transmission. Consequently, offering ART to people living with HIV is a primary strategy to prevent HIV transmission (Montaner et al., 2014; The Lancet HIV, 2017). The early use of ART among people living with HIV regardless of their CD4 levels has been recommended by the World Health Organization and more than 750 other organizations across the globe (Centers for Disease Control and Prevention [CDC], 2018).

The occurrence of other sexually transmitted infections (STIs; Cohen et al., 2019), suboptimal treatment adherence (Getachew et al., 2020), HIV viral resistance to therapy, and stage of HIV infection (Getachew et al., 2020; Melo et al., 2019) are some of the factors that can influence HIV viral load detection in plasma or in genital secretions despite the effectiveness of ART. For instance, untreated STIs may enhance the risk of HIV transmission due to the local inflammatory process of ulcerative lesions, increasing cellular expression of the main HIV-1 entry receptors (Bertram et al., 2019; Gray et al., 2020; Politch et al., 2014). This process interferes with the suppressive effect of antiretrovirals (Baggaley et al., 2010; Fiscus et al., 2013), resulting in the increase of HIV replication in the genital tract (Mayer & Venkatesh, 2011; Williams-Wietzikoski et al., 2019) even when plasmatic viral load is suppressed (Fiscus et al., 2013; Melo et al., 2019; Ouedraogo et al., 2006). Consequently, diagnosis of STIs among people living with HIV has been associated with a higher risk of HIV transmission (Fiscus et al., 2013; Gray et al., 2020; Melo et al., 2019).

Consistent condom use has been acknowledged as a relevant HIV prevention strategy, mainly among serodiscordant couples whose partner living with HIV has unsuppressed viral load (Davari et al., 2020; LeMessurier et al., 2018) or another untreated STI (Fiscus et al., 2013; Melo et al., 2019). Nevertheless, regular condom use remains challenging, despite the robust evidence of its effectiveness against STIs, including HIV infection (Analogbei et al., 2020).

People living with HIV taking ART who have an unsuppressed viral load, regular condomless intercourse, or present co-occurrence of other STIs are at greater risk of transmitting HIV. Successful HIV prevention strategies, such as treatment as prevention (TasP), require the early identification of individuals who are more likely to transmit HIV, including those with amplified HIV transmission risk (Magidson et al., 2016).

Previous studies have shown relevant variations in the frequency of people living with HIV with amplified HIV transmission risk due to use of different definitions of risk (Horvath et al., 2020). Moreover, few studies have considered the characteristics associated with condomless anal and/or vaginal intercourse with partners with a detectable HIV viral load or with sexually transmitted co-infections, defined throughout as “amplified HIV transmission risk” (Magidson et al., 2016). The aim of this study was to investigate the relationship of sociodemographic characteristics, sexual behaviors, professional counseling, serological HIV status, HIV-related clinical data, and beliefs about risk of HIV transmission with amplified risk of HIV transmission among people living with HIV from a city in southeast Brazil.

Methods

This was a cross-sectional study carried out in health care referral units for treatment of infectious diseases in a city in southeast Brazil. The method of this study was described previously (Antonini et al., 2021; Sousa et al., 2021).

Sample Size Calculation

The sample size of the study was initially calculated to estimate the prevalence of people with active sex life living with HIV in the city (n = 10,000), considering an expected prevalence of 62% (Antonini et al., 2021; Sousa et al., 2021). A prevalence study with a precision estimate of 0.05 and 95% confidence interval (CI) would result in a sample size of 339 participants. The sample was increased to 397 due to a nonresponse rate of 17%. Of them, 97 were excluded due to missing clinical data. The final sample of 300 participants would lend a power of 85% to detect statistically significant effect sizes of at least 0.05 for a multiple regression model with six predictors and 5% level of significance.

Recruitment and Data Collection

People living with HIV were recruited in five referral health care units for prevention and treatment of infectious diseases from May 2016 to May 2017. Included participants were those aged 18 years or older, under clinical outpatient follow-up, who had an active sex life and had a sexual partner in the last 6 months regardless of their partners' HIV status. After signing a written consent, participants responded a standardized questionnaire through face-to-face interviews in private rooms at the health care units conducted by trained nurses.

The questionnaire was developed by the researchers and was evaluated by five health care workers with expertise in HIV infection. After that, some changes were made after the experts' suggestions. The final version of the questionnaire was composed of 64 items. Of these, 24 items were used in this study, which are presented as Supplemental Digital Content 1 (Supplementary File 1, https://links.lww.com/JNC/A41). Information on amplified risk of HIV transmission, sociodemographic characteristics, sexual behaviors, use of professional counseling, serological HIV status, HIV-related clinical data, and beliefs about risk of HIV transmission were self-reported. Laboratory and clinical data were confirmed using patient's medical records.

Measures Outcome

Amplified risk of HIV transmission was the outcome variable, which was composed of the following four measures reported in the past 12 months: (a) detectable HIV viral load, (b) condomless sex, (c) symptoms of STIs, and (d) diagnosis of STIs (Magidson et al., 2016). Detectable HIV viral load information was collected from the medical records (0 = undetectable [HIV RNA plasmatic viral load ≤ 40 copies/ml], 1 = detectable [HIV RNA > 40 copies/ml]). Condom use (0 = yes, 1 = no), symptoms of STIs (e.g., discharge, warts, blisters, and genital ulcers; 0 = no, 1 = yes), and if they have been diagnosed with any STIs during the past 12 months were recorded (0 = no, 1 = yes). Amplified risk of HIV transmission was measured considering the number of risk factors according to the following groups: 0, 1, 2, and 3–4.

Sociodemographic characteristics

Demographic and socioeconomic characteristics were age, sex (female or male), skin color (white, yellow/indigenous, brown, black), years of schooling (≤11 or >11 years), and employment status (employed, unemployed, housewife, or retired).

Sexual behaviors

Sexual behaviors in the past 6 months assessed sexual orientation (heterosexual women, heterosexual men, or men who have sex with men), number of sexual partners (one or multiple), and type of sexual partners (stable, casual, or both). Participants were asked whether they have used alcohol (no or yes) or illicit drugs (no or yes) before having sex in the last 6 months.

Professional counseling

Receiving counseling about HIV sexual transmission from health professionals was also collected (no or yes).

Serological HIV status

Information regarding disclosure of HIV status to their partners (yes or no), whether the partner has ever tested for HIV (yes or no/do not know), and partner's serological HIV status (negative/unknown or positive) was also collected.

HIV-related clinical data

Clinical data collected were time since HIV diagnosis (<5 or ≥5 years) and use of ART (no or yes).

Beliefs about risk of HIV transmission

Participants' health beliefs on the transmission of HIV were assessed using the following questions: “Do you think there is a risk of transmitting HIV to his/her partner if a person is under ART and with undetectable viral load?,” “Does low viral load decrease the risk of HIV transmission?,” “Do use and adherence to ART reduce the risk of HIV transmission?”, “Does having genital ulcer increase the risk of HIV transmission?,” “Does a person living with HIV transmit HIV to a partner who does not live with HIV?.” The items of infectiousness beliefs were adapted from previous research (Kalichman et al., 2010). The response options were agree, disagree, or I do not know.

Data Analysis

All variables were computed for each participant and then according to amplified risk of HIV transmission groups. The distribution of continuous and categorical variables was presented through M (SDs) and frequencies, respectively, for the total sample and according to amplified HIV transmission risk groups.

Hierarchical ordered multinomial logistic was used to test the association between independent variables and amplified risk of HIV transmission. The latter was a four-level ordinal variable, namely 0, 1, 2, and 3–4. Ordered logit models were used to estimate the cumulative distribution of probabilities of the response category. The reference group was “0.” Coefficients estimated on the statistical models indicated likelihood of moving into a higher category of amplified risk of HIV transmission. Variables that presented p < .20 in the unadjusted analysis were considered for hierarchical multivariate analysis for adjustments. The stepwise forward selection of predictors in different blocks was used in accordance with the theoretical model hypothesizing the relationship between variables (Figure 1). The results are presented as odds ratios (ORs) with 95% CIs. Five models were tested. The association between sociodemographic characteristics and amplified risk of HIV transmission was tested in Model 1. Sexual behaviors and professional counseling were inserted in Model 2, sequentially, serological HIV status and HIV-related clinical data in Model 3, and HIV transmission beliefs in Model 4. The significance of additional variables was tested in each model. Nonsignificant variables (p > .05) were removed to reach an economic model with relatively few parameters. The final parsimonious model (Model 5) included only variables that remained statistically significant in Model 4. The SPSS version 28.0 (IBM Corporation, Armonk, NY) was used for data analyses.

Figure 1.:

Theoretical model of the hypothesized relationships of sociodemographic characteristics, sexual behaviors, serological HIV status, HIV-related clinical data, and beliefs about risk of HIV transmission with amplified HIV transmission risk.

Results

A total of 397 people living with HIV were interviewed. Of these, 97 were excluded due to missing data in one or more variables related to clinical data and outcome measures. The final sample included 300 participants with complete data. Two hundred and six participants (68.6%) had one or more factors of amplified risk of HIV transmission. The most common characteristics of amplified risk of HIV transmission were STI diagnosis (34.3%), followed by condomless intercourse (31.0%), detectable HIV viral load (25.3%), and symptoms of STIs (14.7%).

Sociodemographic characteristics and sexual behaviors of the participants and the association between these variables and the amplified risk of transmitting HIV are presented in Table 1. The mean age of the participants was 41.2 years (SD = 11). Most participants were men (63.7%), had white skin color (50.0%), with 11 years of education or less (58.0%), and were employed (60.7%). Of the participants, 36.3% were heterosexual women, 69% informed one sexual partner and 64.7% had stable sexual partners. Alcohol consumption and drug using before having sex, and receiving professional counseling about HIV transmission was reported by 37.0%, 16.3%, and 81.3% of the participants, respectively. In the unadjusted analysis, lower age, having multiple sexual partners, use of alcohol consumption before having sex, use of illicit drugs before having sex, and not receiving professional counseling were associated with higher odds of amplified risk of HIV transmission.

Table 1. - Distribution of Sociodemographic Characteristics and Sexual Behaviors and Estimated Unadjusted Odds Ratio (OR) for Amplified HIV Transmission Risk Groups (n = 300) Variables Total Amplified HIV Transmission Risk ORa CI 95% p-Value 0 1 2 3–4 Sociodemographic characteristics Age, M (SD) 41.2 (11.0) 44.9 (11.1) 39.9 (10.8) 40.4 (10.4) 34.9 (9.0) 0.96 0.94–0.98 <.001 Sex, n (%)  Female 109 (36.3) 38 (40.4) 41 (34.2) 24 (36.9) 6 (28.6) 1  Male 191 (63.7) 56 (59.6) 79 (65.8) 41 (63.1) 15 (71.4) 1.21 0.79–1.87 .380 Skin color, n (%)  White 150 (50.0) 45 (47.9) 64 (53.3) 31 (47.7) 10 (47.6) 1  Yellow/indigenous 12 (4.0) 5 (5.3) 5 (4.2) 2 (3.0) 0 (0.0) 0.58 0.19–1.73 .326  Brown 99 (33.0) 35 (37.2) 36 (30.0) 20 (30.8) 8 (38.1) 0.91 0.57–1.44 .678  Black 39 (13.0) 9 (9.6) 15 (12.5) 12 (18.5) 3 (14.3) 1.49 0.78–2.84 .222 Schooling, n (%)  >11 years 126 (42.0) 38 (40.4) 59 (49.2) 21 (32.3) 8 (38.1) 1  ≤11 years 174 (58.0) 56 (59.6) 61 (50.8) 44 (67.7) 13 (61.9) 1.18 0.77–1.79 .449 Employment status, n (%)  Employed 182 (60.7) 60 (63.8) 69 (57.5) 38 (58.5) 15 (71.4) 1  Unemployed 72 (24.0) 17 (18.1) 35 (29.2) 17 (26.2) 3 (14.3) 1.16 0.70–1.91 .338  Housewife 14 (4.6) 3 (3.2) 5 (4.2) 4 (6.2) 2 (9.5) 1.90 0.71–5.11 .205  Retired 32 (10.7) 14 (14.9) 11 (9.1) 6 (9.2) 1 (4.8) 0.62 0.31–1.25 .184 Sexual behaviors Sexual orientation, n (%)  Women heterosexual 109 (36.3) 38 (40.4) 41 (34.2) 24 (36.9) 6 (28.6) 1  Man heterosexual 98 (32.7) 30 (31.9) 40 (33.3) 23 (35.4) 5 (23.8) 1.31 0.79–2.18 .294  Man who has sex with man 93 (31.0) 26 (27.7) 39 (32.5) 18 (27.7) 10 (47.6) 1.13 0.68–1.86 .636 Number of sexual partners, n (%)  One 207 (69.0) 75 (79.8) 83 (69.2) 39 (60.0) 10 (47.6) 1  Multiple 93 (31.0) 19 (20.2) 37 (30.8) 26 (40.0) 11 (52.4) 2.21 1.40–3.48 .001 Type of sexual partners, n (%)  Stable 194 (64.7) 68 (72.3) 73 (60.8) 41 (63.1) 12 (57.1) 1  Casual 96 (32.0) 25 (26.6) 44 (36.7) 20 (30.8) 7 (33.3) 1.28 0.81–2.00 .289  Both 10 (3.3) 1 (1.1) 3 (2.5) 4 (6.2) 2 (9.6) 4.20 1.31–13.50 .016 Alcohol consumption before having sex, n (%)  No 189 (63.0) 70 (74.5) 78 (65.0) 32 (49.2) 9 (42.9) 1  Yes 111 (37.0) 24 (25.5) 42 (35.0) 33 (50.8) 12 (57.1) 2.31 1.49–3.58 <.001 Illicit drugs use before having sex, n (%)  No 251 (83.7) 89 (94.7) 101 (84.2) 47 (72.3) 14 (66.7) 1  Yes 49 (16.3) 5 (5.3) 19 (15.8) 18 (27.7) 7 (33.3) 3.46 1.95–6.14 <.001 Professional counseling, n (%)  Yes 244 (81.3) 84 (89.4) 93 (77.5) 53 (81.5) 14 (66.7) 1  No 56 (18.7) 10 (10.6) 27 (22.5) 12 (18.5) 7 (33.3) 1.79 1.05–3.04 .048

The coefficients estimated indicated likelihood of moving into a higher category of amplified HIV transmission risk.

aORs were estimated using ordered multinomial cumulative logit model. The reference group was “amplified HIV transmission risk = 0.”

Table 2 shows the HIV serological status, HIV-related clinical data, and beliefs about risk of HIV transmission of the sample. Nearly 70% of the participants were aware of the HIV status of their partners and had their partners tested for HIV. Most of the participants' partners were people not living with HIV (72%), 63% received HIV diagnosis 5 years ago or more, and around 94% of participants were under ART. Regarding beliefs about HIV risk transmission, most of the participants believed that HIV is transmissible when people living with HIV are taking ART and have an undetectable viral load (77%), have low viral load (50.7%) and are using ART (63.0%). Only 6% of participants believed that genital ulcer increases the risk of HIV transmission and almost 17% believed there is a risk of HIV transmission to a serodiscordant partner. The unadjusted associations reported that participants who had more than 5 years of HIV diagnosis, those who are under ART, and those who believe that a person living with HIV may transmit HIV to a serodiscordant partner showed lower odds of amplified risk of HIV transmission.

Table 2. - Distribution of Serological HIV Status, HIV-Related Clinical Data, and Beliefs About Risk of HIV Transmission and Estimated Unadjusted Odds Ratio (OR) for Amplified HIV Transmission Risk Groups (n = 300) Variables Total Amplified HIV Transmission Risk ORa CI 95% p-Value 0 1 2 3–4 Serological HIV status Disclosure HIV partner status, n (%)  Yes 209 (69.7) 69 (73.4) 79 (65.8) 45 (69.2) 16 (76.2) 1  No 91 (30.3) 25 (26.6) 41 (34.2) 20 (30.8) 5 (23.8) 1.08 0.69–1.69 .746 Partner tested for HIV, n (%)  Yes 207 (69.0) 64 (68.1) 82 (68.3) 46 (70.8) 15 (71.4) 1  No 93 (31.0) 30 (31.9) 38 (31.7) 19 (29.2) 6 (28.6) 0.91 0.58–1.43 .685 Partner's HIV status, n (%)  Negative 216 (72.0) 72 (76.6) 89 (74.2) 41 (63.1) 14 (66.7) 1  Positive 84 (28.0) 22 (23.4) 31 (25.8) 24 (36.9) 7 (33.3) 1.54 0.97–2.44 .069 HIV-related clinical data Time since HIV diagnosis, n (%)  <5 years 110 (36.7) 21 (22.3) 48 (40.0) 28 (43.1) 13 (61.9) 1  ≥5 years 190 (63.3) 73 (77.7) 72 (60.0) 37 (56.9) 8 (38.1) 0.44 0.28–0.68 <.001 Use of ART, n (%)