Conclusive PrEP study in China paves the way for national provision for gay men

Michael Moloney/

A Chinese study, comparing HIV incidence in 1222 gay and bisexual men who took HIV pre-exposure prophylaxis (PrEP) with incidence in 507 men who chose not to take it, found five HIV infections amongst the PrEP users and 28 in the non-users. The study was presented at last month’s International AIDS Society Conference on HIV Science (IAS 2021).

This translates to an effectiveness in preventing infections of 87% – almost identical to that seen in the two pivotal European studies, PROUD and IPERGAY, which in 2015 set the stage for wider PrEP use in gay men, largely in higher-income countries.

Hopefully, the world’s most populous country will now join them. Dr Junjie Xu of the China Medical University in Shenyang told the conference that due to the conclusive study results, an expert consensus document was issued by the study researchers and other HIV experts, when the results first came out at the end of 2020, to guide wider PrEP implementation in China.


control group

A group of participants in a trial who receive standard treatment, or no treatment at all, rather than the experimental treatment which is being tested. Also known as a control arm.


In relation to pre-exposure prophylaxis (PrEP), 2-1-1 dosing (and 2+1+1 dosing) are alternative terms for event-based dosing. 

on demand

In relation to pre-exposure prophylaxis (PrEP), on-demand dosing is an alternative term for event-based dosing. See ‘event based’.


A patient’s agreement to take a test or a treatment. In medical ethics, an adult who has mental capacity always has the right to refuse. 

loss to follow up

In a research study, participants who drop out before the end of the study. In routine clinical care, patients who do not attend medical appointments and who cannot be contacted.

Due to China’s population, it has the largest number of HIV infections in the Asia-Pacific region, accounting for 29% of cases. The country's HIV epidemic started as one among injecting drug users and recipients of blood products but is now almost entirely sexually acquired, with 72% of infections among heterosexuals and 23% in gay and bisexual men (implying - because there are fewer of them - that gay men are at least ten times more vulnerable to HIV than heterosexuals).

The study

There had been a lack of local high-quality data on PrEP in China, Dr Xu said, until the CROPrEP (Chinese Real-world Oral PrEP) study. This started recruiting in December 2018, enrolled its last participant in October 2019, and ended in October 2020. It was conducted at general hospital-based voluntary counselling and testing centres in Shenyang, Beijing, Chongqing and Shenzen, supervised by the four local Centres for Disease Control, which in China are the local directorates of public health..

Its protocol and recruitment methods were described in a 2019 article in Biomed Infectious Diseases journal.

Participants were recruited both online and offline. Online recruitment was done via advertisements on Chinese social media apps such as Weibo and WeChat. Offline, clinics recruited existing service users, outreach recruitment was done in cruising areas, bars, clubs, and bathhouses, and existing enrollees were encouraged to recruit friends.

CROPrEP was not a randomised controlled study, in the sense that participants were randomly allocated to receive PrEP or not. It was a multicentre prospective cohort study that allocated participants to study arms by choice, rather than randomly. Gay or bisexual clinic attendees who, in the last six months, had had condomless anal sex and/or anal sex with at least two partners and/or a bacterial STI diagnosis, were initially asked if they wanted to be in a study and take PrEP.

It pre-specified the number of participants it wanted: 1500 were to be enrolled in total, with 1000 of them taking PrEP: 500 of them taking it daily and 500 of them taking it on demand (so called ‘2-1-1’ PrEP, in accordance with the IPERGAY regimen).

A group of 500 men who said they did not want PrEP were to be recruited to form the control arm. 

PrEP-taking participants were excluded if they turned out to have HIV, had chronic hepatitis B (which is treated with the PrEP drug tenofovir) or had poor kidney function or other contra-indicatory lab results. Those remaining were asked to choose their preferred PrEP regimen. Enrolment stopped when all three arms reached over 500 recruits.


Arms were ‘over-recruited’ to allow for exclusions. Originally 1222 men expressed willingness to take PrEP but 199 either dropped out or were excluded between screening and starting. Of these, 97 either had hepatitis B (13) or poor lab results. Fifty-seven either withdrew consent or were lost to follow-up and 40 dropped out for other reasons, such as moving. Five men were diagnosed with HIV.

Considerably more men in the control arm – 36 – had a positive HIV test. Although the researchers don’t speculate, presumably the reason some participants declined PrEP was that they already knew or suspected they had HIV but had not disclosed to clinic staff. Others may have not considered themselves to be at risk. In addition to men who tested HIV positive, 17 withdrew their consent or were lost to follow-up and 13 turned out to have no sexual risk within the original screening criteria, when this was re-assessed using a more formal questionnaire that quantified risk.

This left 520 who started daily PrEP, 503 who started on-demand PrEP and 507 in the control arm. Men in the control arm were older than PrEP users (average age 33 versus 29) and had significantly lower scores on the sexual risk measuring instrument (12 versus 18, out of a maximum of 30). There were no significant differences between daily and on-demand PrEP users.

PrEP users had a follow-up visit at one month, while all participants had three-monthly visits for HIV and STI testing and to fill in a behavioural questionnaire. In addition, all PrEP users were sent weekly SMS text reminders to take their pills and members of the control arm were sent monthly appointment reminders. (The SMS support was so important, researchers felt, that a new trial has been proposed linking SMS adherence support messages to electronic pill bottles that can sense if a dose is taken.)

Attendance, adherence and lockdown

Attendance at the three-monthly visits averaged 92% in daily PrEP users, 88% in 2-1-1 PrEP users, and 73% in the control arm, but over 90% of participants from all three arms attended the final 12-month visit.

Adherence, defined as the proportion of sex acts covered by PrEP, followed an interesting pattern. Sex act coverage by daily PrEP peaked at 81% at month three, and then declined somewhat to 72% by month 12.

Adherence to 2-1-1 PrEP started off lower at 57% and by month three had sunk to 46%. However, it then started to rise and by month 12 was actually higher than adherence to daily PrEP, at 78%.

Part of the reason for this initial dip in 2-1-1 PrEP use may be that the majority of participants were three to four months into the study around the time COVID-19 hit China.

A sub-study of trial participants done during the peak of lockdown in China (12 February to 8 March 2020 – see second reference below) found that, for this brief time, the proportion of men testing for HIV in the past month halved from 50% to 26%, the proportion using PrEP at all decreased from 98% to 64%, and the prevalence of poor PrEP adherence increased from 24% to 50%. Further investigation found that one reason participants stopped taking PrEP is that they were worried about doing it without having had an HIV test. The use of condoms with regular partners declined.

By month 12, 458, 452 and 448 men remained in the daily PrEP, 2-1-1 PrEP and control arms respectively, not counting participants who had tested positive for HIV.


There were five HIV infections in the daily PrEP arm, two in the 2-1-1 arm, and 28 in the control arm. This corresponds to an annual incidence of 0.90%, 0.37% and 5.1% respectively, or of 0.64% for both PrEP arms.

This means that there were 12.55% as many infections in the PrEP arms as in the control arm. This is the 'incidence rate ratio'. and subtracting this from 100% gives you the effectiveness of PrEP – about 87%. The effectiveness of 2-1-1 PrEP, versus no PrEP, was 93% and of daily PrEP 82%.

Because CROPrEP was not a randomised study and participants in the control arm differed somewhat from those in the PrEP arms, this efficacy figure cannot be seen as being as definitive as those from PROUD and IPERGAY. But given that participants taking PrEP had higher sexual risk scores than members of the control arm, it is likely to err in the direction of underestimating the number of HIV infections stopped by PrEP in this group of Chinese gay men.


Xu JJ et al. Effectiveness and medication adherence of daily and event-driven pre-exposure prophylaxis regimens among Chinese men who have sex with men: a real-world CROPrEP study. 11th IAS Conference on HIV Science, poster presentation no PECLB23, 2021.

Gao Y et al. Substantial increase in HIV risk behaviors and poor PrEP adherence among men who have sex with men during the COVID-19 pandemic: an observational longitudinal study in four cities of China. 11th IAS Conference on HIV Science, poster presentation no PEC288, 2021.

Update: Following the conference presentation, this study was published in a peer-reviewed journal:

Wang H et al. Association of HIV Preexposure Prophylaxis Use With HIV Incidence Among Men Who Have Sex With Men in China. JAMA Network Open, 5: e2148782, February 2022.