Long-term survival analysis of HIV patients on antiretroviral therapy in Congo: a 14 years retrospective cohort analysis, 2003-2017

Background The long-term survival of patients on antiretroviral treatment in Congo remains less documented. Our study aimed to analyze the long-term survival of adults living with HIV on ART (Antiretroviral Therapy). Methods We conducted a historical cohort study on 2,309 adult PLHIV (People Living with HIV) followed between January 1, 2003 and December 31, 2017 whose viral load and date of initiation of ART were known. The Kaplan Meier method was used to estimate the probability of survival and the Cox regression model to identify factors associated with death. Results The median age was 49 years; the female sex was predominant with 68.56%. The probability of survival at 14 years was 83%, (95% CI (Confidence Interval) [78-87]). On the other hand, when the lost to follow-up died, it was 66% (95% CI [62-70]) in the worst scenario. Stratified cox regression analysis showed that: being male, AHR (Adjusted Hazard Ratio) = 1.65 (95% CI [1.26-2.17]) was significantly associated with death, p-value <0.0001. Furthermore, having a viral load> 1000 copies / ml, AHR = 2.56 (95% CI [1.93-3.40]), be in the advanced WHO clinical stage, in particular: stage II, AHR = 4.07 (95% CI [2.36-7.01]); stage III, AHR = 13.49 (95% CI [8.99-20.27]) and stage IV, AHR = 34.45 (95% CI [23.74-50]) were also significantly associated with death; p-value <0.0001. Conclusion The long-term survival of PLHIV is worrying despite the offer of ARVs.


Conclusion
The long-term survival of PLHIV is worrying despite the offer of ARVs.

Introduction
HIV / AIDS remains a major public health concern in the world, particularly in low and middleincome countries [1]. Since the start of the epidemic, 77. with HIV in 2017, is the most affected region. It also concentrates more than two-thirds of new infections occurring in the world with over 35 million deaths worldwide [3]. According to the WHO, between 2000 and 2016, 13.1 million lives were saved thanks to antiretroviral treatment (ART). Access to these treatments in sub-Saharan Africa remains very low due to many obstacles [4,5], such as the low number of physicians, limited access to viral load measurement and the low availability of basic tests such as measurement of TCD4 cells or biochemistry tests [6].
Faced with these difficulties, WHO has proposed a "lean" approach to support large-scale access to ART in countries with limited resources. [7]. These programs have shown results comparable to those obtained in northern countries in terms of survival, compliance, emergence of resistance, toxicity, virological, immunological and clinical efficacy [8,9]. The details are in Table 1.  15]. These conditions can increase the mortality of PLHIV.
In addition, few studies have been done to document these aspects. This study aims to fill these gaps by using the Eval-Co cohort database. It aimed to estimate the probability of survival at 14 years while identifying the prognostic factors associated with the death of PLHIV on ART between 2003 and 2017 in Congo.

Study type and period
We performed a historical cohort study, including patients on ART between 2003 and 2017, in the Republic of Congo.

Presentation of the data source: Eval-CO cohort
The data used for our study come from the historical cohort 2003-2017 of the EVAL-CO project (evaluation of the therapeutic management of HIV patients in Congo).
The primary objective of EVAL-CO was to assess the survival of HIV patients after initiation of treatment in Congo. Secondary objectives were to: 1) assess retention of patients on antiretroviral therapy, 2) measure the proportions of patients with CD4 cell counts below <350 cells/mm3 among those with baseline CD4 cell counts, 3) assess the virological efficacy of treatment regimens, and 4) measure the incidence of tuberculosis and hepatitis B among people living with . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted March 15, 2022. ; https://doi.org/10.1101/2022.03.15.22272252 doi: medRxiv preprint the first part of the Eval-co project on treatment failure, 6924 patients were included (reference n°13). We excluded patients under the age of 18 and those with no viral load data. For this second project, we excluded all patients who did not have a start date for antiretroviral therapy.
At the end, 2,309 patients with the available data have been included for this study. (Fig 1) shows the flowchart of the patients included in this study.

Presentation of study variables
The variables in our study included: socio-demographic data: sex (Male, Female); the patient's age (in years), level of education (None, Primary, Secondary, University) and marital status . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted March 15, 2022

Data analysis
We performed data analysis with Stata 15.0 software. The frequencies of socio-demographic characteristics (sex, age, marital status, level of education) were calculated. We used the Kaplan-Meier method to estimate the 14-year survival probabilities from the survival curve. The log-rank test was used to compare survival curves. The significance was set at 5% (p <0.05).
We used the Cox model to identify prognostic factors associated with patient death. In univariate analysis, each relevant explanatory variable identified in the review of the literature (sex, viral load, WHO clinical stage of the disease, CD4 number) was crossed with the explanatory variable "vital status" (1 = death; 0 = survival) [17][18][19]. The endpoint was death. The variables having a significant effect with a P value less than 0.05 were retained for the modeling. The significance level for integration into the final model was 0.05. We measured the strength of the association between the independent variable and the dependent variable by the hazard ratio, its 95% confidence interval, and the P-value.

Ethical considerations
The research protocol has been approved by the ethics committee of the Faculty of Health centralizes all the country's data through a national database. All data was anonymous. Only the team that participated in the research had access to the data. Permission to use the data was obtained from the Ministry of Health.   . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) 1 1

Worst case overall survival curve
As shown in (Fig 3), according to the worst-case scenario, if the lost to follow-up had died; the 14-year survival was 66.39% CI [62.14-70.29].

Overall survival curve by CD4T
( Fig 5)  . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted March 15, 2022    . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.   CI: Confidence Interval; P-value < 0.05: *; P-value < 0.01: **; P-value < 0.001: ***.

Discussion
Our work has shed light on the long-term survival of adult PLHIV on antiretroviral treatment between 2003-2017 in the Republic of Congo as well as the prognostic factors associated with death. In this retrospective cohort study, there were 212 deaths out of the 2,309 PLHIV included, for a death rate of 9.18%. Our descriptive results show a predominance of women compared to men (68.56% of women against 31.40% of men). This could be explain by the fact that women are more exposed to risky sexual practices given the financial precariousness that this layer faces.
This is the case of sex workers where a single woman is been exposing to a high frequency of sexual relations with several men. They are at the same time the most at risk layer in the Republic of Congo [20].
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) 1 5 The overall survival rate was 83% under the base case scenario and 66% under the worst scenario. Among the prognostic factors associated with death were: male sex advanced WHO clinical stage of the disease, and high viral load. The overall survival rate under the base case scenario was lower than that of the worst scenario by the simple fact that the lost to follow-up were considered to be dead. The fact that males were associated with the prognosis of death could be explained by several factors, the late diagnosis of the disease. This could lead to late management, an advanced clinical stage of the disease or a high viral load. Because in our study, the advanced clinical stage of the disease and the high viral load were also associated with the prognosis of death. Poor ARV adherence has also been a factor in treatment failure leading to a high rate of patient death in several African studies [21][22][23][24][25].
The probability of survival under the base scenario estimated to be 98% at one year, 97% at two years, 96% at three years, and 95% at 5 years. Our rates were slightly higher than those found in the DRC by LOANDO [27].
In the worst-case scenario, survival after two years estimated at 79%. In South Africa, Coetzee et al. measured survival on ARVs in adults at 86.3% at two years [28]. In Malawi, Zachariah et al.
In this study, according to the same scenario, the probability of survival at 5 years estimated at 76%. In Botswana, Bussmann H. and Collaborators, for the same duration found 79% (31). In Cameroon, Sieleunou and collaborators found 47% [30] against 75.4% in Senegal [31].
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) 1 6 The male gender was a prognostic factor for mortality with a risk almost double that of the female sex. Our results corroborate those found in Nepal [32].
The final model on predictors of survival consisted of male gender, viral load, and WHO advanced clinical stage of the disease. [33]. Similar results have been found in several African studies [34][35][36]. The limits of the study: We did not include some possible predictors in the study due to variable insufficiency. These include, for example, hemoglobin level, therapeutic line, level of adherence to antiretroviral treatment, Tuberculosis coinfection, other opportunistic infections, body mass index, anemia, prophylactic treatment with cotrimoxazole.

Conclusion
The long-term survival of PLHIV remains a concern according to the baseline scenario. These . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

Conflicts of interest
The authors do not present any conflicts of interest.

Acknowledgments
We would like to thank all the staff of the National Council for the Fight against AIDS and Epidemics who facilitated the realization of our study. We would also like to thank all the researchers from Marien University who contributed to the proofreading of this research work. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)