Determinants and Incidence of Chronic Kidney Disease on Tenofovir-Based Antiretroviral Therapy Regimens: A Cohort Study in HIV-Infected Adults in South China

Background: To evaluate the incidence and risk factors of stage 3 chronic kidney disease (CKD) and rapid kidney function decline (RKFD) among Chinese HIV-1 infected patients starting with tenofovir (TDF)-based regimen. Methods: We enrolled in 797 TDF-initiated HIV-1-infected patients in a Chinese cohort. Kidney dysfunction were dened as stage 3 CKD (eGFR < 60 mL/min/1.73 m 2 during follow-up) and RKFD (eGFR decline > 10 mL/min/1.73 m 2 /year). A linear mixed-effects model was used to quantify the average eGFR change per 48 weeks. A generalized estimating equation regression analysis was conducted to determine the risk factors associated with renal dysfunction. The method of multiple imputation was used to reduce bias caused by missing data. Results: In this retrospective study, 14 (2%) patients experienced stage 3 CKD, and 272 (34%) individuals experienced RKFD during a median of 26 (IQR, 4-78; maximum 325) weeks follow-up period. The mean loss in eGFR per 48 weeks increased consistently over time, from -2.59 mL/min/1.73 m 2 before 48 weeks to -17.61 mL/min/1.73 m 2 after 288 weeks. For every 10 mL/min/1.73 m 2 increase of eGFR, the risk of RKFD increased by 29% (95%CI: 18%, 40%). Each 10 years older and every 10 mL/min/1.73 m 2 higher in baseline eGFR, the risk of stage 3 CKD increased to 1.56 (95% CI: 1.00, 2.43) and decreased by 65% (95% CI: 48%, 76%), respectively. Anemia and higher viral load were signicantly associated with RKFD. The results were robust across a range of multiple imputation analysis. Conclusions: TDF-associated CKD is rare in HIV-1 infected Chinese adults. Longer TDF-exposed patients are more likely to have renal dysfunction, especially those with older age, anemia, lower baseline eGFR and higher viral load. primary results provided important clinical and research illumination. Treatment with TDF-containing regimen has been widely used as rst-line antiretroviral therapy in developing countries; ours is the rst largest and well-powered study, which utilized rigorous methodological approaches, to show the risks of TDF-related CKD on a follow-up duration of 6.5 years in China, to our knowledge. the rate of stage 3 CKD was low among individuals with TDF exposure, in a country like China with a high quantity of HIV-1 infected patients (1.25 which increasing year by year, the numbers of TDF-related CKD is expected to be high. Therefore, to assess renal function management and monitoring approaches a relevant either preventing nephrotoxicity in these settings; studies like ours underline an urgent need for this we identied age and baseline eGFR as important predictors for CKD among HIV-1 infected South Chinese patients receiving TDF. Future studies should focus on whether or not prediction models combined with the two above factors and other variables, such as weight, can effectively help clinicians identify patients at high risk of TDF-related nephrotoxicity, which may alter and improve the therapeutic program.


Introduction
The widespread use of combination antiretroviral therapy (cART) has essentially improved the life expectancy of the human immunode ciency virus (HIV)-positive individuals [1]. The need to manage noncommunicable diseases (NCDs) becomes evident with the advent of ART [2]. Chronic kidney disease (CKD) is one of the important NCDs associated with morbidity and mortality [3,4]. Since HIV-1 infection is currently incurable and requires lifelong treatment, maintaining kidney function is especially important for HIV-1 infected patients. In addition, kidney function has been reported to deteriorate faster in HIV-1 infected patients than in the general population [5,6].
Tenofovir disoproxil fumarate (TDF), a reverse transcriptase inhibitor, which was recommended as one of the antiretrovirals (ARVs) at rst-line therapy by World Health Organization (WHO), is currently the most valuable drug against HIV. However, TDF has clinically signi cant nephrotoxicity [7,8]. Patients in Asia are more susceptible to nephrotoxic ARVs, because of their lower body weight than whites or blacks [9,10]. HIV-1 infected patients in China live longer due to more and more abundant clinical resources, which further underscores the importance of maintaining these patients' kidney function. At present, only one study in China that included comorbidity status and information on kidney function at baseline but limited to patients < 50 years of age, has assessed the prevalence of and factors associated with renal dysfunction in TDF users infected with HIV-1 [11]. However, the small sample size (269 patients) and short-term observation (168 weeks) in this report limits its generalizability.
From the above points, we conducted a retrospective study in a Chinese cohort of treatment naïve HIV-1positive individuals who initiated TDF-containing regimens between Jan., 2012 and Jun., 2018. The objective of this study was to describe the incidence and risk factors of renal dysfunction in HIV-1 infected patients receiving TDF in a "real world" clinical setting in the largest infectious disease hospital in South China.

Study Population
This is a retrospective, observational cohort study conducted at the Infectious Diseases Department (IDD) at Xixi Hospital (Hangzhou, Zhejiang, South China). All treatment naïve HIV-1-positive patients with records of TDF initiation between January 10, 2012 and June 25, 2018 were screened for eligibility. A total of 900 patients were enrolled as of December 30, 2018. This study was approved by the Institutional Review Board of Xixi Hospital.

Data Collection and Inclusion Criteria
Data extracted from the IDD database included demographic parameters, date of TDF initiation, details of the cART regimens, route of HIV-1 transmission, comorbidities, AIDS diagnosis days and laboratory variables (HIV-1 RNA viral load, CD4 count, and serum creatinine [SCr]) at 2 weeks, 1 month, 2 months, 3 months, and every 3 months thereafter until December, 2018. Baseline was de ned as the date of starting cART. Each enrolled patient was 17 years old or more, had at least one additional eGFR measurement since January, 2012. The owchart was detailed in Fig.1.

Statistical Analyses
A linear mixed-effects regression model based on a Generalized Estimate Equation (GEE), with random intercepts, was applied to quantify the average change per 48 weeks of eGFR during different periods on cART. GEE logistic regression was used to analyze the relationship of baseline characteristics with RKFD and with stage 3 CKD. Time to the incident of stage 3 CKD was calculated by the Kaplan-Meier survival curve. We constructed two analyses; one was adjusted for age, gender and body mass index (BMI) to estimate the crude associations of each baseline characteristic with RKFD and with stage 3 CKD; another was performed on a range of multivariable analyses, to determine variables independently correlated with RKFD or associated with stage 3 CKD.
Data on HIV-1 RNA viral load were not available in up to 40% of patients, so a missing value category was used in the main analyses [14]. In sensitivity analyses, 5 imputed datasets, established by multiple imputation with chained equations, which can reduce bias caused by the exclusion of individuals with missing data of HIV-1 RNA viral load at baseline, were developed and run separately, and the results were combined using Rubin's method [15,16] (Additional le 1: Table S1-S3).
All analyses were performed using the R software, version 3.3.1 (http://www.R-project.org). A result was considered statistically signi cant when the two-tailed P value was below 0.05. The paper is reported according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement [17].

Quanti cation of Mean Loss in eGFR
The adjusted mean loss in eGFR increased consistently over time, from -2.59 mL/min/1.73 m 2 before 48 weeks to -17.61 mL/min/1.73 m 2 after 288 weeks. The results were similar in the unadjusted model and sensitivity analysis (Additional le 1: Table S1).   4 shows the factors associated with stage 3 CKD. Baseline eGFR and age were signi cantly associated with the development of stage 3 CKD in the adjusted analysis. For every 10 mL/min/1.73 m 2 increase in baseline eGFR, the risk of stage 3 CKD decreased by 65% (95% CI: 48%, 76%). Each 10 years older age was associated with a 1.56-fold increased risk of developing stage 3 CKD (aOR, 1.56; 95% CI: 1.00, 2.43). Sensitivity analyses that were conducted in ve imputed datasets did not alter the trends of our study Additional le 1: Table S2, S3).

Discussion
In this observational cohort study of HIV-1-infected south Chinese patients who initiated a TDFcontaining cART, we quanti ed the mean loss of eGFR over time and evaluated the determinants of renal dysfunction among patients who initiated TDF-based ART.
The slope of eGFR was negative in most patients, and more than 34% of individuals developed RKFD, but only 2% of patients experienced stage 3 CKD. The effect of TDF on the decrement of eGFR was evident, continuously increased during the 6-years observation period (from − 2.59 mL/min/1.73 m 2 before 48 weeks to -17.61 mL/min/1.73 m 2 after 288 weeks). The independent factors associated with a greater likelihood of RKFD was higher eGFR and HIV-RNA viral loads at baseline. Patients with anemia had an increased risk of RKFD. Older age and lower baseline eGFR were signi cantly associated with an increased risk of stage 3 CKD.
Previous studies in Sub-Saharan Africa have shown that renal dysfunction is highly prevalent in HIV-1 infected patients receiving cART. The incidence of stage 3 CKD was estimated to be 5.2-25% [3,18], which is higher than the incidence of 2% in our results. However, our ndings in general decline on renal function related to TDF exposure were based on a population in which the most patients had preserved kidney function, the pre-existing renal impairment (eGFR < 60 mL/min/1.73 m 2 at baseline) patients in the two studies were 5.2% and 25.1% respectively, which are both much higher than 0.1% of our research.
Of note, a study carried out by Qing Tan et al. revealed that the overall eGFR improved over 24 months following TDF initiation among 269 patients in Chengdu, South China [11]. But, the median eGFR at baseline in their research was higher than that of our study (124.0 vs. 108.2 mL/min/1.73 m 2 ). Also, Bygrave et al. in South Africa demonstrated that renal function improved in most patients, in which 176 (18.9%) had a baseline creatinine clearance < 50 ml/min [19]. However, their study was highly limited by missing data and lack of key covariates. While, the results of our sensitivity analyses were similar, which were conducted in 5 imputed datasets to reduce bias caused by missing data in our study, making the conclusion of that the overall eGFR decreased throughout the observation period following TDF initiation more robust. Moreover, the decreased trend of eGFR among TDF users was consistent with a previous study conducted among 650 HIV-positive patients with a median eGFR of 108 (IQR: 91-126) mL/min/1.73 m 2 at baseline, which is equivalent to our baseline eGFR level [20].
As in other studies, we found a negative association between stage 3 CKD and baseline eGFR, which was consistent with previous data from South Africa [3]. This cohort focused on individuals with TDF exposure for 48-months revealed that patients with mild (eGFR: 60-89 ml/min; HR 4.8, 95%CI: 1.5-15.2) or moderate (eGFR: 30-59ml/min; HR 15.0, 95%CI: 3.4-66.5) renal dysfunction presented a higher risk of nephrotoxicity by 48-months compared to normal renal function. However, the de nition of nephrotoxicity in this study was indistinct, which is de ned as the occurrence of any decline in kidney function from baseline (acute or chronic) was observed since TDF initiation, which resulted in a lack of precision for the relationship between baseline eGFR and renal dysfunction. Instead, we avoided obscuring associations between the two factors by the clear de nition of kidney dysfunction, and the quanti cation model of the two variables may help clinicians to assess TDF-related CKD risk according to baseline renal function, accurately.
In contrast, the risk of RKFD showed a positive association with baseline eGFR, as also reported by Takeshi and colleagues in Japan [10]. The de nition of RKFD in this study was utterly consistent with us. One possible reason for the positive association is that patients with normal renal function show faster responses for TDF exposure compared to patients with renal dysfunction. In our study, the rarely small proportion of patients with kidney impairment conform to this speculation. As in other studies, patients with anemia and higher viral load at baseline were signi cantly associated with RKFD [21-23]. Such patients deserve particular focus on renal monitoring. Consistent with previous studies, older age increased the incidence of stage 3 CKD [11,18,20]. These ndings may re ect the impact of age on TDF circulating levels. Unlike previous studies, we found no evidence that lower body mass index (BMI) was independently associated with a higher risk of renal dysfunction. A study that baseline eGFR is equivalent to ours has shown similar ndings to ours [20]. These ndings may result from unmeasured confounding, although we enrolled in more covariates than others [10] and adjusted for the cohort to account for some of these differences.
Our primary results provided important clinical and research illumination. Treatment with TDF-containing regimen has been widely used as rst-line antiretroviral therapy in developing countries; ours is the rst largest and well-powered study, which utilized rigorous methodological approaches, to show the risks of TDF-related CKD on a follow-up duration of 6.5 years in China, to our knowledge. Although the rate of stage 3 CKD was low among individuals with TDF exposure, in a country like China with a high quantity of HIV-1 infected patients (1.25 million), which is increasing year by year, the numbers of TDF-related CKD is expected to be high. Therefore, to assess renal function management and monitoring approaches would de nitely provide a relevant bene t either for preventing or salvaging TDF nephrotoxicity in these settings; studies like ours underline an urgent need for this topic. Eventually, we identi ed age and baseline eGFR as important predictors for CKD among HIV-1 infected South Chinese patients receiving TDF. Future studies should focus on whether or not prediction models combined with the two above factors and other variables, such as weight, can effectively help clinicians identify patients at high risk of TDF-related nephrotoxicity, which may alter and improve the therapeutic program.
Other than the limitations mentioned above, the duration of follow-up in this research was short. Longer follow-up will be necessary to assess the long-term impact of TDF on kidney outcomes [24]. The inherent shortcomings of retrospective observational single-center study limited its generality in other populations, and thus further validations from different races are warranted.
To conclude, the present study suggested that a longer duration of TDF using was associated with faster renal function decline, and compared to the low rate of stage 3 CKD in our study, patients infected with HIV-1 mainly experienced RKFD in South China. With the expanded use of TDF, prior to screening for baseline eGFR and taking account of age are necessary to reduce TDF-related CKD before ART initiation. In addition, anemia patients and those with unsuppressed viral load deserve particular focus on renal monitoring. Our results were consistent across a range of sensitivity analyses, and the large sample size makes the results more powerful for HIV-1 positive patients exposed to TDF in China.

Declarations
Ethics approval and consent to participate This study was approved by the Institutional Review Board of Xixi Hospital. All data were anonymized to comply with the provisions of personal data protection legislation. Due to the retrospective nature of this study and due to the fact that only historical medical data were collected, written informed consent was not required.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.

Funding
The study is funded by Hangzhou Agricultural and Social Development Research Project (20201203B191).

Authors' contributions
LF and JFB conceived, designed, and organized the study, interpreted the results, and drafted the manuscript. JSH helped supervise the study. The other authors contributed to collect the data on site.    Figure 1 Study selection ow Flowchart showing the study selection among TDF users infected HIV-1.

Figure 3
Determinants of reduced kidney function: multivariate analysis using Generalized Estimate Equation