Superior Effects of Antiretroviral Treatment among Men Who have Sex with Men Compared to Other HIV At-Risk Populations in a Large Cohort Study in Hunan, China

This study assesses association between CD4 level at initiation of antiretroviral treatment (ART) on subsequent treatment outcomes and mortality among people infected with HIV via various routes in Hunan province, China. Over a period of 10 years, a total of 7333 HIV-positive patients, including 553 (7.5%) MSM, 5484 (74.8%) heterosexuals, 1164 (15.9%) injection drug users (IDU) and 132 (1.8%) former plasma donors (FPD), were recruited. MSM substantially demonstrated higher initial CD4 cell level (242, IQR 167–298) than other populations (Heterosexuals: 144 IQR 40–242, IDU: 134 IQR 38–224, FPD: 86 IQR 36–181). During subsequent long-term follow up, the median CD4 level in all participants increased significantly from 151 cells/mm3 (IQR 43–246) to 265 cells/mm3 (IQR 162–380), whereas CD4 level in MSM remained at a high level between 242 and 361 cells/mm3. Consistently, both cumulative immunological and virological failure rates (10.4% and 26.4% in 48 months, respectively) were the lowest in MSM compared with other population groups. Survival analysis indicated that initial CD4 counts ≤200 cells/mm3 (AHR = 3.14; CI, 2.43–4.06) significantly contributed to HIV-related mortality during treatment. Timely diagnosis and treatment of HIV patients are vital for improving CD4 level and health outcomes.


Introduction
Antiretroviral therapy (ART) can significantly reduce mortality and prevent opportunistic infections in HIV-positive patients by strengthening immune response that suppress virus replication [1,2]. Timely diagnosis and treatment initiation have been recognized as an effective approach to prevent HIV transmission from people living with HIV (PLHIV) to the susceptible population [3,4]. In 2014, UNAIDS declared its ambitious 90-90-90 goals to enable 90% of people living with HIV to be diagnosed, 90% of diagnosed individuals to receive ART and 90% of those on ART have successful included in this study. The cohort started from 1 January 2003 and ended at the 31 December 2013. We only used the baseline, follow-up and treatment status information for the purpose of this study.

Clinical Laboratory Testing at Baseline
Demographic information, mode of infection, medical history and laboratory test results of patients were collected only once by the health workers at ART initiation, but CD4 count level was monitored after positive confirmation and throughout the course of treatment. Viral load was only recorded at baseline and followed up on a voluntary basis as the detection cost was high. Patients who have experienced immunological failure previously were recommended to receive viral load test, but recently more patients with strong health consciousness also requested for viral load tests. Disease progression stages were classified according to WHO guidelines [26]. Days from diagnosis to treatments was calculated as the difference between the date of diagnosis and ART initiation date.

Disease Progression Indicators
Three key outcome indicators were collected to evaluate the effects of ART: (1) CD4 and HIV viral levels. The median with interquartile range (IQR) of CD4 counts and viral load were calculated every three months in the first year then every six months over the remaining course of follow-up. (2) Treatment failure (including both immunological and virological failure). Immunological failure was defined as CD4 Cell count of the patient fell below 100 cells/mm 3 after receiving ART for three months without a concomitant diseases that may sharply decrease the CD4 counts [27]; virological failure was defined as plasma viral load above 1000 copies/mL based on two consecutive viral load measurements after 3 months of treatment [28].
Considering that the majority of participants (5411, 73.8%) initiated ART after 2010, only a few people have CD4 and viral load test more than 43 months of follow-up, we categorized records beyond 43 months as a single time interval.
(3) HIV-related mortality. All death cases were recorded until the end of the study (December 2013).
Mortality was compared across different CD4 count strata and transmission route categories. Factors associated with mortality were identified with Cox regression.

Statistical Analysis
All surveyed patients were stratified according to their transmission route (FPD, MSM, Heterosexual and IDU). All data were entered and analysed using statistical software SAS 9.2 (Statistical Analysis System). Descriptive and inferential statistical analyses were performed. The mean, median and IQR were used to summarise numerical variables; whereas frequencies and percentages were used to describe categorical variables. A linear regression model was constructed to verify whether the ascending trend of CD4 Cell counts during the follow-up was significant. The survival analysis was performed to compare the survival probability across transmission groups. A multivariable Cox proportional hazards model was used to identify the hazard factor contributing to HIV-related mortality, immunological failure and virological failure. A p-value of less than 0.05 was considered significant in the final model. The patients with missing data were included in the "unknown group".

Ethical Considerations
This study was reviewed and approved by the Monash University Human Research Ethics Committee (approval number CF15/4321-2015001862). No personal information about the patients were disclosed in this research. The collected data were analysed only for the purposes of this study. Besides the members of the experimental group, no other individual, group, or institution had access to this data. No further informed consent about this study was required.

Trend of CD4 and Viral Load Level
Median CD4 level among all ART patients increased from 148 cells/mm 3

Significant Factors Associated with Treatment Failure and Mortality
Our adjusted multivariable Cox proportional hazards analysis indicated that being infected through homosexual contacts, plasma donation, and initial ART CD4 counts ≥200 cells/mm 3 were protective factors against treatment failure. In particular, the rate of immunological failure in IDU, MSM and FPD during the 48-month follow-up was approximately 17%, 8% and 12%, respectively.

Significant Factors Associated with Treatment Failure and Mortality
Our adjusted multivariable Cox proportional hazards analysis indicated that being infected through homosexual contacts, plasma donation, and initial ART CD4 counts ě200 cells/mm 3 were protective factors against treatment failure. In particular, the rate of immunological failure in IDU, MSM and FPD during the 48-month follow-up was approximately 17%, 8% and 12%, respectively. These  Being female, being infected through homosexual contacts, initial ART CD4 counts ≥200 cells/mm 3 and younger age were protective factors against HIV-related death. Specifically, male patients had a 1.4 times greater risk of death (AHR = 1.40; 95% CI, 1.15-1.71) than female, and the Being female, being infected through homosexual contacts, initial ART CD4 counts ě200 cells/mm 3 and younger age were protective factors against HIV-related death. Specifically, male patients had a 1.4 times greater risk of death (AHR = 1.40; 95% CI, 1.15-1.71) than female, and the mortality risk in MSM was only about 20% (AHR = 0.23; CI, 0.09-0.55) of the heterosexuals while the risk is more than double in IDUs (AHR = 2.12; CI, 1.75-2.57). Patients with CD4 level below 200 cells/mm 3 had three times higher risk of death (AHR = 3.14; CI, 2.43-4.06) than otherwise. Aging patients have slightly elevated mortality risk (AHR = 1.01; CI, 1.00-1.02) (Figure 3c).

Discussion
This retrospective cohort study demonstrated that MSM in Hunan have higher median CD4 T-cell counts at baseline and retained better health outcomes than other populations throughout the course of ART treatment. MSM consistently demonstrated lower rates of immunological failure, virological failure and mortality compared with the patients infected via other routes. The robust CD4 response and viral suppression presented in MSM are comparable to findings of long-term ART patients in developed countries [29,30].
Our study confirmed that high CD4 cell level at treatment initiation among MSM can lead to improved treatment outcomes. On the contrary, previous MSM studies in other Chinese settings have consistently reported low testing coverage, which resulted in delayed diagnosis and treatment, and rapid progression to AIDS [19,31,32]. Improved health outcomes among MSM in our study may be due to multiple factors. In particular, the HIV screening program promoted by the Gates and Melinda Foundation, with the cooperative efforts from Hunan CDC under the national "Four Frees and One Care" policy, may have diagnosed a large population of HIV-infected MSM in the province and substantially improved HIV awareness during the study period [24]. Besides, HIV outbreaks in MSM in Hunan are reportedly much later than other coastal provinces in China [33], MSM in Hunan are also better prepared for the epidemic and adhered to ART than other parts of the country [34].
Maintaining high CD4 count at early phrase of treatment plays an important role for subsequent treatment outcomes. This is consistent with previous findings in China [35,36] and internationally [37]. These findings convey a unified message that timely treatment initiation is vitally important for improved treatment outcomes in the long term. The low treatment failure rate in MSM implies a relatively low drug resistance level in this population in Hunan, which concurs with a previous study [22]. In contrast, without targeted interventions, virological failure rates in IDU in Hunan are noticeably higher than that in other parts of China [36,38]. These substantial differences may be attributed to the later ART initiation (only 31% of IDUs initiated on time) and their poor treatment adherence. A separate study reported that 15% IDU patients skipped medication in the last month [35], and that treatment adherence among IDUs is significantly worse than that of non-injecting drug users [39,40]. Psychological distress, strained family relationship, unstable employment and concurrent drug use may all contribute to the inherence [41,42]. Notably, the markedly elevated immunological and virological failure rates in IDU may also relate to their high opportunistic infections rate, because of their high-risk injecting behaviours [43][44][45]. AIDS-defined complications may cause permanent damages to the immune system that cannot be fully recovered even with ART [46][47][48].
The analysis of HIV-related mortality demonstrated similar results. Participants with baseline CD4 <200 cells/mm 3 , mostly IDU, have the lowest survival probability. Similarly, FPD, with lowest median CD4 level at baseline, demonstrated similar survival probability as IDU. As HIV transmission through illegal blood transfusion occurred mostly during the early stage of HIV epidemics in China, FPD likely experienced delayed diagnosis and treatment due to the absence of a national ART program then [49,50]. This may have led to inadequate immune responses to sustain high enough CD4 level to combat against the HIV infection [51][52][53]. In IDUs, additional adverse effects associated with harm from drug addiction may add to the risk of mortality in the HIV+ IDUs. Further, addiction may also lead to sub-optimal treatment adherence in IDUs, thereby counteracting the benefits of ART [54,55]. The antagonism between ART regimens and opioid substitution therapy can also lead to reduced adherence to ART. Male participants are known to be more prone to risk behaviours and less aware of their own health status. In comparison, females are more self-aware and hence more likely to participate in HIV testing and comply with clinical advices to receive timely ART [56][57][58]. This finding concurs with studies conducted in other developing countries [59,60]. In addition, age has a moderate impact on mortality possibly because aging can further weaken the immune system. Although the influence is marginal, given that HIV-infected population is gradually ageing, further studies are still necessary to understand the complications associated with ageing on death in this population.
A number of limitations of this study should be noted. First, the routes of HIV infection were self-reported by the participants; as such, it is difficult to determine the accuracy of the information. In particular, MSM may choose to conceal their sexual identity due to social stigma. Second, most MSM started ART after 2007, so the record of follow-up was not as complete as that for IDU, FPD and heterosexuals. Besides, the record prior to 2007 may be less accurate than more recent data, owing to the immature reporting mechanism then. This may affect data quality of FPD and comparability between study populations. Third, the study design did not capture risk factors for HIV exposure or information on past interventions that participants might have received. This prevents us from investigating the reasons for their early or late diagnoses. Fourth, as a retrospective study, the recorded information is subjected to recall bias by participants. Fifth, only 228 (3.11%) patients received viral testing at baseline. Given the high out-of-pocket cost of the test, this may have led to overestimation of the actual virological failure rate.

Conclusions
Conclusively, our study indicated that HIV-positive MSM can achieve better treatment outcomes and reduced mortality than other at-risk populations over the long-term course of ART if they have high CD4 count at ART initiation. This conveys an important message, that diagnosis and early treatment are possible in an opaque population that is often stigmatised in a developing country setting.
has a moderate impact on mortality possibly because aging can further weaken the immune system. Although the influence is marginal, given that HIV-infected population is gradually ageing, further studies are still necessary to understand the complications associated with ageing on death in this population.
A number of limitations of this study should be noted. First, the routes of HIV infection were self-reported by the participants; as such, it is difficult to determine the accuracy of the information. In particular, MSM may choose to conceal their sexual identity due to social stigma. Second, most MSM started ART after 2007, so the record of follow-up was not as complete as that for IDU, FPD and heterosexuals. Besides, the record prior to 2007 may be less accurate than more recent data, owing to the immature reporting mechanism then. This may affect data quality of FPD and comparability between study populations. Third, the study design did not capture risk factors for HIV exposure or information on past interventions that participants might have received. This prevents us from investigating the reasons for their early or late diagnoses. Fourth, as a retrospective study, the recorded information is subjected to recall bias by participants. Fifth, only 228 (3.11%) patients received viral testing at baseline. Given the high out-of-pocket cost of the test, this may have led to overestimation of the actual virological failure rate.

Conclusions
Conclusively, our study indicated that HIV-positive MSM can achieve better treatment outcomes and reduced mortality than other at-risk populations over the long-term course of ART if they have high CD4 count at ART initiation. This conveys an important message, that diagnosis and early treatment are possible in an opaque population that is often stigmatised in a developing country setting. you