Higher prevalence of kidney function impairment among older people living with HIV in Uganda

Background People living with HIV (PLWH) are at risk of kidney function impairment due to HIV-related inflammation, antiretroviral therapy (ART), diabetes mellitus, and hypertension. Older persons may experience a higher burden of chronic kidney disease (CKD) as kidney function declines with increasing age. There is a paucity of data comparing the prevalence of kidney function impairment in older PLWH to that in HIV-uninfected people in sub-Saharan Africa. Methods We conducted a cross-sectional study among people aged ≥ 60 years living with and without HIV in Kampala, Uganda who were matched 1:1 by community location. We collected data on sociodemographics, comorbidities, and HIV-related clinical characteristics. We defined kidney function impairment as an estimated glomerular filtration rate(eGFR) < 60mls/min/1.73m2 with or without proteinuria. We constructed multivariable logistic regression models to study associations between participant characteristics and kidney function impairment. Results We enrolled 278 people (median age 66 years); 50% were PLWH, and 51.8% were female. Overall, the prevalence of kidney function impairment was 23.0% (95% CI:18.4%−28.4%); 33.1% (95% CI: 25.7%−41.4%) versus 12.9% (95% CI: 8.3%−19.7%) among people living with and without HIV (p-value < 0.01). The prevalence of proteinuria among PLWH versus people without HIV was 43.9% (95% CI:35.8%−52.3%) versus 19.4% (95% CI:13.6%−26.9%) p-value < 0.01. Living with HIV (OR = 3.89(95% CI: 2.04–7.41), p-value < 0.01), older age (OR = 1.13, (95% CI:1.07–1.20), p-value < 0.01), female sex (OR = 1.95, (95% CI:1.06–3.62), p-value = 0.03) and a prior diagnosis of hypertension (OR = 2.19(95% CI:1.02–4.67), p-value = 0.04) were significantly associated with kidney function impairment. Conclusions HIV infection is strongly associated with kidney function impairment among older PLWH. Prioritizing routine measurements of kidney function and proteinuria in older PLWH will enable early detection and institution of measures to reduce the progression of kidney disease.


Background
Globally, access to antiretroviral therapy (ART) has averted at least 20 million HIV-related deaths over the last two decades; therefore, more people are aging with HIV [1].Aging with HIV is associated with an increased risk of multimorbidity [2] such as noncommunicable diseases (NCDs) including chronic kidney disease (CKD) [3,4].Nephron senescence is a recognized age-related change that eventually leads to a decline in kidney function and increases the risk of CKD in older persons [5].In people living with HIV (PLWH); aging, chronic in ammation which persists during suppressive ART, and nephrotoxic ART such as tenofovir disoproxil fumarate (TDF) [5][6][7] converge to increase the risk of CKD by threefold [8].Moreover, traditional risk factors for CKD such as diabetes mellitus and hypertension are more prevalent among PLWH [2,9].Chronic kidney disease is a progressive disease that manifests with several indicators of kidney function impairment such as proteinuria [10].Kidney failure is associated with increased morbidity, poor quality of life, and increased risk of death among PLWH [11][12][13][14].In low-income countries (LICs) such as Uganda, there is limited access to life-saving dialysis and kidney transplants to manage end-stage kidney disease(ESKD)/kidney failure and even when available, the costs exceed the average income for most patients [15,16].Studies in the general population have documented the association of HIV with kidney disease in both low-and high-resource settings [17][18][19][20], but there is a paucity of data documenting the excess prevalence of kidney function impairment in older PLWH in Uganda and sub-Saharan Africa.We sought to determine the additional prevalence of kidney function impairment in older people living with HIV compared to those without HIV in Uganda.

Study design and setting
We conducted a cross-sectional study at the Infectious Diseases Institute (IDI) clinic in Mulago, Kampala, Uganda between April and August 2023.The IDI is an implementing partner for the President's Emergency Plan for AIDS Relief (PEPFAR) with clinics in 18 districts in the country.The IDI agship adult clinic at Mulago Hospital takes care of more than 8,000 HIV-infected patients and among these patients, more than 1000 are aged ≥ 60 years.
We consecutively enrolled PLWH aged ≥ 60 years who were attending the HIV clinic at IDI and people without HIV recruited within a 2-kilometer distance of the same communities where the participating PLWH lived.The participants were recruited on a 1:1 ratio.We de ned older people as those aged ≥ 60 years per the United Nations de nition [21,22].Participants were excluded if they could not provide blood or urine samples for study measurements.

Data collection tools and procedures
We collected data using an electronic structured questionnaire developed in Research Electronic Data Capture (REDCap) developed by Vanderbilt University [23].HIV-uninfected participants were counseled and tested for HIV before enrollment in the study.
For each participant, we collected sociodemographic data (sex, age in years, education level, income status), and anthropological measurements (height in meters and weight in kg) using calibrated scales and calculated the body mass index in kg/m 2 .We collected information on comorbidities (diabetes mellitus, hypertension, other cardiovascular disease, liver disease, and others), prescriptions and overthe-counter drugs, past or present history of smoking, history of current alcohol use, and history of HIV, which included duration of ART, antiretroviral drugs (previous and current regimens), history and duration of TDF use, most recent viral load and CD4 count.We measured blood pressure using a standardized automated Omron M2 basic blood pressure monitor with consistent positioning (patients were seated upright with back and arm support).We took three readings and calculated the average.We measured fasting blood sugar in mmol/l using a point-of-care calibrated on-call plus glucometer manufactured by ACON laboratories, USA.For participants with no prior history of diabetes mellitus, a diagnosis of diabetes mellitus was made if one had elevated blood glucose (nonfasting ≥ 11.1 mmol/L or fasting ≥ 7 mmol/L) in the presence of symptoms.If elevated values were found in someone asymptomatic; a repeat fasting blood sugar was performed on a subsequent day to con rm the diagnosis.
We collected an early morning urine sample from each patient for measurement of urine protein by dipstick urinalysis and a venous blood sample for measurement of serum creatinine.We de ned kidney function impairment as an estimated glomerular ltration rate (eGFR) < 60 mls/min per 1.73 m 2 with or without proteinuria on a dipstick urine test.This creatinine-based eGFR was calculated from the 2009 version of the chronic kidney disease epidemiology collaboration) (CKD-EPI) (CKD-EPI 2009) [24] without correcting for race [25,26].We used the CKD-EPI 2009 which is widely used in clinical practice and a recent study showed it to perform better than the race-free creatinine-based CKD-EPI 2021 equation in sub-Saharan Africa [27].We categorized the stages of kidney function impairment by eGFR as stage 1: ≥90, stage 2:≥60-<90, stage 3: ≥30-<60, stage 4:≥15-<30 and stage 5: <15mls/min/1.73m 2 [28].
We measured proteinuria using Siemens Multistix GP Urine Test Strips and we categorized it as negative if no protein was detected or positive if at least trace protein (≥ 30mg/dl) was detected.

Data Analysis
We exported the data to STATA version 14 (Stata Corp LLC, College Station, TX).
We reported the proportions of participants with kidney function impairment and proteinuria in both groups with 95% con dence interval (95% CI) and compared them with a chi-square test.
The additional prevalence of kidney function impairment due to HIV was reported as a difference between the proportions in the group of PLWH and people without HIV with its 95% CI and a p-value corresponding to the z-statistic for testing the signi cance of a difference in proportions between two groups.
We summarized other categorical variables as proportions and continuous variables as medians with interquartile ranges (IQR).Comparisons were based on a chi-square test for categorical variables otherwise Fisher's exact test was used if any of the cells had ≤ 5 observations.Comparisons for continuous variables were based on the Wilcoxon rank sum test.
We performed logistic regression with HIV as the main exposure and performed a strati ed analysis by HIV status to study associations between various participant characteristics and kidney function impairment.We considered the level of signi cance at a p-value ≤ 0.2 for bivariate analysis and p ≤ 0.05 for multivariate analysis.We assessed confounding factors by comparing unadjusted and adjusted models, interactions using the loglikelihood ratio test, and goodness of t of the model using the Hosmer-Lemeshow test.

Results
We enrolled 278 men and women aged ≥ 60 years; 50% were PLWH and 51.8% were female.
The prevalence of kidney function impairment among PLWH and people without HIV varied by disease stage.PLWH constituted the majority in stage 3 and stage 4 of kidney function impairment (Fig. 2).

Discussion
Our study shows that one-third of the aging Ugandan HIV population had kidney function impairment that was signi cantly higher than the prevalence in community-matched HIV-uninfected controls.Data documenting the burden of kidney function impairment in aging HIV populations come mostly from highincome countries [8, 19,29].Overall, we found a high prevalence of kidney function impairment among people aged ≥ 60 years living with and without HIV in Uganda.Our data agree with studies that have shown that PLWH experience a high burden of kidney disease [19,29,30].HIV infection signi cantly increases the odds of kidney function impairment [8,9,31,32] and various forms of HIV-associated nephropathy have been documented in previous research [9,33].However, our study demonstrated an excess burden of kidney function impairment among older PLWH compared to previous prevalences reported in the general population of PLWH in Uganda [34,35] and other LICs [18,30,36,37].These ndings imply that there is a need for research and clinical care programs to evolve and prioritize kidney disease detection in the population of older PLWH.
The prevalence of proteinuria in our study was at least two times higher in PLWH than in people without HIV.Other studies such as the AGE h IV cohort study [38] have shown that the burden of albuminuria is higher among people living with HIV than in people without HIV.However, our study reported higher estimates of proteinuria than what has been reported in non-African settings [38,39].Previous studies have pointed to genetic risks for kidney injury such as the presence of the APOL1 gene which is unique to African populations and may account for racial disparities in the burden of CKD [40][41][42][43].Proteinuria is a biomarker for kidney injury [28]; thus further research into context-speci c risk factors for kidney disease is needed to understand the cause of kidney injury in older PLWH.Kidney disease is progressive; thus these ndings of the excess prevalence of kidney function impairment and proteinuria may have implications for the health and survival of old PLWH in a country where there is limited access to life-saving therapies such as dialysis and renal transplants for kidney failure [15,16].
Increasing age and female sex were signi cantly associated with kidney function impairment in both PLWH and people without HIV.These ndings are consistent with previous studies from Uganda and Africa where increasing age and female sex were associated with kidney disease in both PLWH and people without HIV [18,31,44,45].
Among older PLWH, a prior diagnosis of hypertension signi cantly increased the odds of kidney function impairment.Hypertension is a known traditional risk factor for CKD [46] and hypertensive nephrosclerosis is a well-recognized pathological process that eventually leads to kidney failure [47].A high prevalence of hypertension has been reported among aging PLWH [2,48] and the association of such traditional risk factors with CKD is documented in previous studies [46,49].The implementation of blood pressure control programs in older PLWH and hypertension can bene t kidney health as the same approach has been shown to improve outcomes in other populations at risk of hypertension-related kidney injury such as people with cardiovascular diseases and diabetes mellitus [50].
In our study, HIV was the strongest signi cantly associated factor with kidney function impairment.Unlike other studies, we did not nd an association between diabetes mellitus [29,51] and kidney function impairment, smoking [9,52,53], body mass index (BMI), socioeconomic status (monthly household income) [44,54] or prior diagnosis of tuberculosis [55] [56] as other studies did.Unlike these previous studies, in our study, fewer participants reported a history of smoking, and there were fewer diabetic patients.The monthly household income was not different between PLWH and people without HIV or between people with and without kidney function impairment.Likewise, the BMI was not different between people with and without kidney function impairment and all the people who had a previous diagnosis of tuberculosis were among PLWH.
Among the PLWH in our study, only a small proportion, 2.2%, had CD4 less than 200 cells per ml and only 5.8% had detectable viral loads (> 50copies/ml).Almost all (95.7%) of our PLWH had ever used TDF.These proportions may explain the lack of association between these variables and kidney function impairment in our study although previous studies have shown a signi cant association [8,18,57,58].In our study, we observed higher prevalence of kidney function impairment in participants on abacavirbased regimens; most of the participants on abacavir-based regimens had previously been on TDF, but were switched over due to low eGFR.

Figures
Figures