How Many Sexual Partners Are Too Many for Elderly Men From Rural China Who Visit Female Sex Workers Becoming HIV-Positive?

Yi Yang (  thehanyang@163.com ) Chengdu University of Traditional Chinese Medicine ShuangFeng Fan Chengdu Center for Disease Control and Prevention Shu Liang Sichuan Center for Disease Control and Prevention Yang Liu Chengdu Center for Disease Control and Prevention Yuan Li Chengdu University of Traditional Chinese Medicine Jing Xi Chengdu Center for Disease Control and Prevention Dan Yuan Sichuan Center for Disease Control and Prevention Jie Xiao Chengdu University of Traditional Chinese Medicine


Introduction
China is facing big challenges to ending the AIDS epidemic as a public health threat [1,2] . One of the biggest challenges is that HIV prevalence of elderly (≥50 years old) men has been steadily increasing [3][4][5] , much higher than the general population [6] . Due to immune functional decline and increased vulnerability to infections [7] , multiple morbidity [8] , HIV-positive elderly men quickly progress into AIDS even deaths [3,7] .
Compared with other age groups, elderly men have minimal knowledge about AIDS [9] , condom use among them is even lower [10,11] . On one hand, HIV-positive elderly men could be a newly infected group.
On the other hand, current elderly men are from previous sexually active 20-50-year-old group who might be HIV-positive but were not detected.
Sexual behaviors of elderly men in China are far from well-studied. Elderly men in China, who are considered as sexually inactive [12] and had been excluded from the priority of HIV prevention programs [13][14][15][16] , are not recognized and listed as a key prevention group until 2017 for the rst time [17] .
Due to convention [18] and stigma [19] , Chinese, especially elderly people, do not talk about sex. Previous studies of sexual behaviors in China mainly focus on 20-50-year-old group [10,[20][21][22][23][24][25][26][27] . Based on data from national HIV Sentinel Surveillance System(NHSSS) [3][4][5] , heterosexual transmission has been widely accepted as the major route for HIV epidemics in China. Another key concern is men who have sex with men (MSM). It is estimated that MSM account for 1.73% of 18-64-year-old men in China [28] , and HIV infection among them continue to expand just as in most countries [29] .
The contributions of heterosexual behavior and homosexual behavior in HIV infection among elderly men in China is not clear at population level. It is not clear which kind of heterosexual behavior play the most important role. Whether it is commercial heterosexual behavior, non-commercial extramarital sexual behavior, or casual sexual behavior? More importantly, how many sexual partners is too many for elderly men becoming HIV-positive? Are there certain methods to nd out the cut-off points for "high-risk" groups?
Receiver-operating characteristics (ROC) curves have been applied to evaluate the accuracy (sensitivity and speci city) of parameters in predicting risks obtaining such as HIV infection, and identify the most suitable cut-off points for prediction [30,31].

Methods
This study was a case-control study with 99 cases and 88 controls who visited female sex workers (FSW) in their lifetime, and the protocol was approved.

Study Site
Chengdu, the capital city of Sichuan province, is a one of the unique suitable study sites in China. One of fth people living with HIV(PLHIV) in Sichuan are from Chengdu. Located in southwest of China, Sichuan province is one of the few provinces in China heavily affected by the HIV epidemic, though successful comprehensive HIV prevention strategies has been applied [1] . There are two major characteristics of HIV epidemics in Sichuan: HIV prevalence consistently increasing among elderly population and high prevalence of HIV infection among MSM in Chengdu. Elderly people in Sichuan accounted for less than 5% of PLHIV in 2005, and over 40% in 2017, and 59.2% in 2019 [1,5] , meanwhile the proportion of elderly men in the population is less than 40% [32] . The overall HIV prevalence among MSM in Chengdu between 2009 and 2014 was 15.5% [33] . Between 2012 and 2018, the HIV incidence density among MSM decreased annually, but the total incidence density was as high as 5.95 (95% CI: 5.37-6.56)/100 person-years [34] . Moreover, in 2013, 17.9% of MSM reported that they had sex with both men and women(MSM/W), and living in Chengdu was found out as risk factor [35] .
With the coverage of sentinel surveillance improving [1] and around 800,000 population [36] , rural County A in Chengdu ranked one of the top ve counties in the number of PLHIV, mainly from male group.
During April 2019 and October2020, 115 elderly men from 8 townships in County A were con rmed as HIV (+), and 114 participated the study. In their lifetime, 106 of 114(92.98%) in case group reported visiting FSW. We obtained 99 cases' numbers of different types of accumulated sexual partners and sexual partners in the last three years in detail.
During June to July 2019, a cross-sectional study from the same townships of case was conducted, 802 men were recruited, and 797(99.38%) questionnaires were reliable, and among them 795 were HIV (-) [37] . 88 of 795(11.07%) HIV (-) respondents admitted visiting FSW in their lifetime.

Data Collection
A written informed consent was completed before the investigation. Local slangs were used to refer sexual behaviors. An in-depth interview with one case was conducted by one skilled medical staff from County A center for disease prevention and control (CDC). In total, two medical staff conducted the surveys in separate rooms in CDC, county infectious hospital, or township health centers according to cases' convenience. A 30-minute face-to-face structured interview with a control was conducted in a separate room at the village health centers to make sure the respondents felt comfortable to talk about their sexual behaviors by one well-trained male interviewer.

Measures
Basic information demographic characteristics, migration experiences.
Biology sexual desire changes after the age of 50(no sexual need/decline/no change/increase).

Sexual behavior
Numbers of different types of accumulated sexual partners and sexual partners in the last three years were measured, including xed sexual partners, extramarital sexual partners, casual sexual partners and commercial sexual partners. In order to make respondents feel comfortable about the survey, homosexual behavior was measured only whether they had homosexual behaviors or not without asking the speci c number.

Health Services
Whether they had heard of AIDS before the investigation for HIV (-) ones was asked. Then, HIV-related health education (HRHE) was measured in detail, including whether they got HRHE before, what kind of health educator provided HRHE before their HIV con rmation for cases and before the investigation for controls, including health workers from CDC, health workers from township health center/village doctors, doctors from other hospitals, their children/grandchildren's teachers(primary school and high schools), township civil servants, village cadres, pharmacy staff, HIV-related service volunteers, and others.

ROC
Receiver-operating characteristics (ROC) curves were plotted to establish cut-off values of numbers of different sexual partners associated with HIV infection. The validity of the model was measured by the area under the curve (AUC). The more AUC reach 1, the more the data is well modelled. Based on Youden's indexes (Sensitivity +Speci city-1), the most suitable cut-off values were identi ed. The more Youden's index is close to 1, the more accurately the cut-off points can distinguish the difference.

Data analysis
Frequencies for nominal variables, mean and standard deviation for interval variables were assessed. T test and Chi-square tests/ sher's exact test were used to examine the relationships between HIV infection and independent variables. Binary logistic regression with backward selection was applied to examine factors associated with HIV infection, including all factors with p<0.05 in bivariate analyses. Adjusted odds ratio (AOR) and 95% con dence intervals were calculated. Factors with AOR greater than one was categorized as risk factors, and less than one as protective factors.

Basic Information
The age was 63.28±8.62 for cases, 61.63±7.08 for controls, the difference was not statistically signi cant (t=1.43, P=0.16), similar with the differences of types of residency, education levels, marital statuses, monthly expenses for entertainment and migration experiences between two groups(P>0.05). In terms of latest occupations, types of living with, monthly income, the differences between two groups were statistically signi cant(P<0.05), details are showed in Table1.

Number of Sexual Partners
Cases had more sexual partners in their lifetime, mainly commercial sexual partners than controls(P<0.01), similarly in the last three years(P<0.01). Details are showed in Table 3.  Table 4 and gure1.   1-11.64)). Details are showed in Table 5.

Discussion
As a case-control study, this study design is well organized, and proves that having more than 52 and above sexual partners in their lifetime and visiting 5 and above FSW in the last three years are too many for elderly men from rural China becoming HIV-positive. Becoming HIV-positive among elderly men is not related with sexual desire change and sexual identity. As groups visiting FSW in their lifetime, risks of HIV infection do not come from their unmarried statuses, neither from whom they live with, even not from migration experiences.
It is not suddenly that elderly men become HIV-positive, but rather delayed detections. In this study, around 80 percent of the respondents from both groups had migration experience. Back to 1980s, they were 20-50-year-old, sexually active, and migrated alone, easy to be involved in high-risk sexual behaviors such as visiting FSW and having sex with men to ful l their emotional and sexual needs [10,18,[20][21][22][23][24]27] .
Without doubt, HRHE helps elderly men avoid becoming HIV-positive. Rather than from informal channel(other), getting HRHE from formal channels (health workers from CDC, health workers from township health center/village doctors) helps elderly men avoid HIV infection. However, comparing with having multiple sexual partners and visiting FSW, the protective role of HRHE is small.
In this study, 2 of 99 cases (2.02%) and 2 of 88 controls(2.27%) were MSM, a little higher than national level(1.73%) [28] . 2 of 2 MSM cases self-reported as heterosexual, and 2 of 2 controls self-reported as bisexual. All of them were MSM/W. MSM in China are facing stigma and family pressure to get married and have children that make HIV(+) MSM transmit HIV infection from their homosexual partners [29,38] to their wives [39] .

Limitations
The current study should be noted self-report behavior information. Face-to-face interviews may heighten socially desirable responses such as low report of high-risk sexual behaviors. In order to confront the problems, our interviewers were well trained, interviews were conducted in separate rooms, and local slangs were used.
Moreover, in-depth interviews were conducted by skilled medical staffs from County A CDC for case group, but by well-trained young male interviewers for control group. Informational bias due to different types of interviewers was inevitable, and would overestimate the difference between two groups. In order to minimize the bias, interviewers were trained by same trainers, only well interviewers conducted the indepth interview, and interviewers followed the same in-depth procedure.
Though we conduct census among HIV (+) elderly male population, due to limited sample size, we can't con rm our suspicion about the role of homosexual behavior in HIV infection among elderly men. We will continue to conduct researches to con rm our suspicion. The institutional review board (IRB) from a liated hospital of Chengdu University of Traditional Chinese Medicine approved the protocol (reference number: 2019KL-008). All methods were performed in accordance with the relevant guidelines and regulations including a statement. A written informed consent was completed before the investigation, and respondents agreed to participate the study and agreed that their information can be published anonymously if needed.

Consent for publication
A written informed consent was completed before the interviews, and respondents agreed that their information can be published anonymously if needed.

Availability of data and materials
The datasets analyzed during the current study are available from the corresponding author on reasonable request. Figure 1 ROC curve and AUCs for number of sexual partners