In an era where highly active antiretroviral therapy (ART) is becoming more widely available in Africa, much attention has been focused on the potential need for assessing and monitoring renal function to identify those at risk for chronic kidney disease (CKD) in resource-limited settings. According to 2005 guidelines from the Infectious Diseases Society of America,1 HIV-infected patients should be screened for pre-existing kidney disease with a urinalysis for proteinuria and estimated glomerular filtration rate (eGFR) as both HIV infection and certain antiretrovirals can cause severe kidney dysfunction. Because HIV-associated nephropathy occurs almost exclusively in patients of black African descent,2 and North American studies have indicated black race as a significant risk factor for both CKD and rapid progression to end-stage renal disease in HIV-infected individuals,3,4 there had been concern that high rates of kidney disease may be seen in Africa. Current studies of kidney disease in HIV-infected patients in Africa have reported substantially varying rates of decreased kidney function before ART initiation, ranging from 0.7% in areas of Uganda5 to 20% in Nigeria,6 with the largest studies reporting rates of 0.7%,5 7%,7 and 9%.8 Earlier studies were limited by smaller sample size, and with the exception of the Development of ART (DART) study,9 which was performed in Uganda and Zimbabwe, all studies have been performed in single-country cohorts.
More recent World Health Organization (WHO) 2010 guidelines10 do not promote routine universal assessment of CrCl in all HIV-infected individuals before ART initiation or at entry into care. However, tenofovir-containing regimens are now recommended as first-line ART as is calculation of creatinine clearance (CrCl) before its initiation, if feasible, given its nephrotoxic potential.11-13 In addition, CrCl monitoring is recommended with tenofovir use for high-risk groups including older persons, those with low body weight or other renal risk factors such as diabetes or hypertension, and those taking concomitant nephrotoxic medications.10 These guidelines reflect the fact that routine assessment and monitoring of renal function may not always be feasible in many resource-limited settings such as Africa. It is, therefore, important to balance the need for averting poor clinical outcomes and preserving already scarce health care resources. The purpose of this analysis was to describe the prevalence of clinically significant reductions in kidney function in a large clinical cohort of HIV-infected individuals across 7 sub-Saharan African countries and to identify characteristics of persons at high risk for reduced CrCl to prioritize laboratory evaluations of kidney function before ART initiation.
The MTCT-Plus Initiative in sub-Saharan Africa provided support to 13 clinical programs in Cameroon, Cote d'Ivoire, Kenya, Mozambique, Rwanda, South Africa, Uganda, and Zambia to implement HIV/AIDS care and treatment to HIV-infected women identified through perinatal HIV prevention services and their families, including male partners.14 At program entry, all participants underwent a physical examination, WHO clinical staging, and laboratory testing for CD4+ cell count. The CD4+ cell count was repeated every 6 months. Patients eligible for ART were recommended to have more extensive laboratory tests, including serum creatinine before ART initiation.
ART eligibility was determined based on WHO and local guidelines. In 2003-2004, adults were considered eligible for ART if they met the following criteria: WHO stage 4, CD4+ cell count ≤200 cells per cubic millimeter or WHO stage 2 or 3 and CD4+ cell count ≤350 cells per cubic millimeter. In January 2005, eligibility criteria were modified to exclude WHO stage 2 with CD4+ cell count ≤350 cells per cubic millimeter.
The current analysis includes African participants >15 years of age who initiated ART at enrollment or follow-up and had a serum creatinine value within 30 days of ART initiation. If multiple creatinine values were available for a patient, the last result before ART initiation was selected for this analysis. Those not eligible for ART were excluded from the analysis. Pregnant women and those within 4 weeks postpartum at the time of creatinine laboratory testing were excluded from this analysis due to physiologic increases in glomerular filtration rate (GFR), which render serum creatinine measurements unreliable for the estimation of kidney function. All patients with missing creatinine values, creatinine dates, or delivery dates were excluded from analysis. Data from the Rwanda MTCT-Plus Initiative site were excluded due to concerns regarding the quality of creatinine data.
The Institutional Review Board of Columbia University approved the MTCT-Plus Initiative as a service delivery program, and approved the use of retrospective clinical data for research purposes.
We evaluated the prevalence of clinically relevant reductions in kidney function before ART initiation using routinely collected serum creatinine data. Kidney function was estimated using 3 methods since no single equation has been validated in African populations: (1) Cockcroft-Gault (CG) equation for creatinine clearance,15 (2) 4-variable Modification of Diet in Renal Disease equation (MDRD)16 and, (3) CKD Epidemiology Collaboration (CKD-EPI) equation.17 Primary analyses were performed applying the CG equation, as current guidelines1 recommend its use for drug dosing in HIV patients. Because the majority of renally cleared ART agents require dose reduction for CrCl ≤50 mL/min, this was chosen as the threshold to define clinically significant kidney disease.
Continuous, categorical, and primary outcome variables were compared using 2-sample t test, χ2, and Wilcoxon tests. Crude and adjusted odds ratios (ORs) for the probability of having CrCl ≤50 mL/min were calculated in logistic regression modeling to identify factors independently associated with clinically significant kidney disease. In multivariate modeling, we tested CD4+ cell count as a categorical and continuous variable based on evaluation of Quantile-Quantile (QQ) plot, which showed a reasonable linear distribution. Variables found not significantly associated with the outcome measures (eg, P > 0.05) in the bivariate analysis were excluded from the final model. All statistical analyses were performed using SAS 9.2 (SAS Institute, Cary, NC).
Between 2003-2007, 8127 HIV-infected individuals initiated ART; 4404 patients had available creatinine values for analysis. After excluding pregnant patients and all data from the Rwanda site, 2619 patients remained. An additional 122 who began ART >30 days before obtaining creatinine, and 2 patients with missing weight information were excluded, leaving 2495 patients for analysis (Fig. 1).
Demographic and clinical characteristics of the study population (n = 2495) are shown in Table 1. The median age was 30 years [interquartile range (IQR): 27-35]; 70% were female. Median CD4+ cell count was 295 (IQR: 173-450) cells per cubic millimeter, and 78% were WHO stage 1/2. Median body mass index (BMI) was 22.8 (IQR: 20.4-25.6) kg/m2; 18.6% were taking trimethoprim prophylaxis before ART initiation and at the time of the creatinine testing. The largest number of patients was enrolled at sites in Uganda and South Africa.
Comparison of the study sample to the MTCT-Plus patients without creatinine values revealed a gender difference between the 2 groups where men were more likely to have creatinine results [OR = 3.1, confidence interval (CI) = 2.7 to 3.5]. This was not surprising as pregnant women were excluded from the sample. On average, patients with creatinine results were 1.5 years older than ones without (P = 0.0182). In terms of baseline clinical characteristics, patients with WHO stages 3/4 were more likely to receive creatinine testing (OR = 2.1, CI = 1.8 to 2.4) than ones with stages 1/2. The mean baseline CD4+ cell count (335 cells/mm3) of the study sample patients was also lower than the patients with no creatinine (484 cells/mm3).
Median estimated kidney function by CG CrCl was 94.8 mL/min (IQR: 77.8-116.8). The overall prevalence of CrCl ≤50 mL/min was 3.4% (95% CI= 2.7% to 4.1%) (Table 2) with notable variability between countries (South Africa: 1.2%; Côte d'Ivoire: 2.1%; Mozambique: 3.1%; Zambia: 3.4%; Kenya: 3.8%; Cameroon: 4.1%; Uganda: 5.6%). Median estimated kidney function varied somewhat when using the 4-variable MDRD and CKD-Epi equations (110 mL/min and 116 mL/min, respectively). However, overall prevalence rates of eGFR ≤50 mL/min were similar using these alternative formulas (Table 2). Persons with CrCl ≤50 mL/min had a higher median age and were more likely to have BMI <18, WHO clinical stage 3 versus stage 1 and 2, and CD4+ ≤50 cells per cubic millimeter compared with adults with CrCl >50 mL/min (Table 1).
In multivariate analysis adjusting for age, gender, CD4+ cell count, and country of enrollment, age >30 (adjusted OR = 2.06, 95% CI = 1.23 to 3.45), and decreased CD4+ cell count (adjusted OR = 5.43 for CD4+ cell count ≤ 50, 95% CI = 2.48 to 11.87) were independently associated with CrCl ≤50 mL/min (Table 3).
In this large sample of ambulatory HIV-infected adults representing multiple sub-Saharan African countries, 3.4% were found to have clinically significant kidney dysfunction based on CG CrCl ≤50 mL/min and may have required dose reductions in ART and further monitoring. This prevalence rate is lower than that compared with other studies in Africa. Those with CrCl ≤50 mL/min tended to be >30 years of age and with more advanced HIV disease. Of those >30 years old within the MTCT-Plus cohort, 53% were female. Despite some variation in prevalence rates by country and method of CrCl calculation, this overall low prevalence rate supports current WHO guidelines to initiate ART without routine prescreening of renal function. Our findings also suggest that targeting CrCl testing toward those >30 years of age or those with more immunologically advanced HIV disease may be a more judicious use of resources.
Evaluations of other smaller clinic-based cohorts in Kenya and Uganda reported higher prevalence rates of 4.8% and 20%, respectively, of CrCl ≤50 mL/min.18,19 Many previous studies of kidney disease in HIV-infected African patients have chosen to characterize CKD according to the United States National Kidney Foundation Kidney Disease Outcome Quality Initiative20 threshold of CrCl <60 mL/min, which primarily identifies those at greatest risk for cardiovascular disease, death, and hospitalization.21 We chose to define clinically significant kidney disease as a CrCl ≤50 mL/min because this threshold is used to guide clinical decision-making, including dose adjustments of medications.
One explanation for the low prevalence of kidney dysfunction, we observed may be the smaller proportion of MTCT-Plus Initiative participants with advanced HIV disease. The MTCT-Plus Initiative enrolled pregnant and recently postpartum women from PMTCT programs and was a largely young female population in relatively good health compared with other studies which often include adults presenting to care with evidence of advanced HIV infection.7,8,19 In contrast, the Ugandan cohort with a 20% prevalence of CrC l ≤50 mL/min had a mean baseline CD4+ count of 122 cells per cubic millimeter.19 Similarly, DART trial participants where the prevalence of CrCl <60 mL/min was 7% had notably advanced HIV disease as evidenced by a median CD4+ cell count of 86 cells per cubic millimeter.7 Other large studies reporting a 9% prevalence of CrCl < 60 mL/min have shown similar characteristics in severity of HIV disease.8 The association between advanced disease and kidney dysfunction was also demonstrated among MTCT-Plus Initiative participants: multivariate modeling showed that persons with CD4+ cell count ≤50 cells per cubic millimeter were more than 5 times more likely to have CrCl ≤50 mL/min than persons with CD4+ cell count >350 cells per cubic millimeter.
Another key difference between our study and others is the inclusion of multiple African countries. Most kidney disease prevalence studies among HIV-infected individuals in Africa have either been country-specific or included, at most, 2 countries. In contrast, our study spanned 7 sub-Saharan African countries, yielding a broader representation of overall kidney disease prevalence in ambulatory HIV-infected individuals in Africa. Though not statistically significant, variability was noted in prevalence rates of CrCl ≤50 mL/min across countries, ranging from 1.2% to 5.6%. These differences may signal differences in genetic constitution, geographic and environmental factors such as nutrition, and comorbid conditions such as hypertension and diabetes, which all play a role in the epidemiology of CKD in Africa and although clearly beyond the scope of this study, may be an area of further research.
Both in our study and in the DART trial, estimates of kidney function varied depending on different methods of calculation. Our results were consistent with the DART study,9 where the CG equation yielded higher prevalence rates for decreased GFR than 4-variable MDRD and with a large study in Ghana, where mean eGFR by 4-variable MDRD and CKD-Epi equations were higher than that by CG.22 Moreover, they were consistent with this latter study and a small South African study23 showing that elimination of the race coefficient in the MDRD and CKD-Epi equations produced more similar results to the CG. Diet, protein intake, weight, and muscle mass all affect serum creatinine, and, therefore, estimates of GFR or CrCl. Individuals in Africa may consume less creatinine-generating foods or have decreased muscle mass compared with those in Western countries. It is also possible that inclusion of the race coefficient in the MDRD equation leads to underestimation or overestimation of GFR in African individuals.23 Further studies should be done to validate equations for estimating kidney function in African populations.
One limitation of our study was the large number of patients (46%) in the MTCT-Plus Initiative who initiated ART but did not have available creatinine values for analysis. As a result, the study sample may have been biased as there were differences in gender, age, and clinical status between our study sample and the MTCT-Plus patients without creatinine laboratory values. This could be explained, in part, by the fact that sicker patients would have undergone more extensive tests before ART initiation. Therefore, we may have overestimated the prevalence of CKD. Another limitation of our study is the lack of data on proteinuria and the underlying etiology of kidney disease in our study population. Proteinuria data may have added information on early kidney disease. In addition, all MTCT-Plus laboratory testing was performed by local labs as the MTCT-Plus Initiative was a service delivery program and not as a clinical trial. Though the majority of sites used the Jaffe alkaline picrate reaction method for creatinine assays, we were unable to confirm the methodology used at each site during the entire observation period. Therefore, the lack of a standardized method for serum creatinine testing across all sites may have contributed to the variability in the prevalence rates of CrCl ≤50mL/min between countries. Last, prevalence estimates in our study were based on a single serum creatinine measurement, which may lead to overestimation or underestimation of the true prevalence of CrCl ≤50mL/min.
Today progressively more African countries are offering free and widespread ART to eligible HIV-infected individuals. The escalating number of ART eligible patients in a setting of still limited resources across much of the continent highlights the need to prioritize laboratory testing including renal function assessment. Our study demonstrates that in a largely ambulatory HIV-infected population, rates of clinically significant kidney disease remain low in many parts of Africa and should reassure ART programs that, in similar populations, one can safely follow current 2010 WHO guidelines,10 which recommend CrCl assessment and monitoring only with tenofovir-containing regimens and if feasible. A lack of resources for universal screening and monitoring of kidney disease should not preclude ART initiation in resource-limited settings. Because our study revealed CD4+ count ≤50 cells per cubic millimeter and age >30 years as factors associated with clinically significant kidney disease, selective screening of patients >30 years of age or those with advanced HIV disease for kidney disease may be a more strategic way of utilizing the already sparse resources available in sub-Saharan Africa.
We thank all patients and participating staff at the African sites of the MTCT-Plus Initiative of ICAP-Columbia.
1. Gupta SK, Eustace JA, Winston JA, et al. Guidelines for the management of chronic kidney disease in HIV-infected patients: recommendations of the HIV Medicine Association of the Infectious Diseases Society of America. Clin Infect Dis
2. Wyatt CM, Klotman PE. HIV-associated nephropathy in the era of antiretroviral therapy. Am J Med
3. Choi AI, Rodriguez RA, Bacchetti P, et al. Racial differences in end-stage renal disease rates in HIV infection versus diabetes. J Am Soc Nephrol
4. Lucas GM, Lau B, Atta MG, et al. Chronic kidney disease incidence, and progression to end-stage renal disease, in HIV-infected individuals: a tale of two races. J Infect Dis
5. Lucas GM, Clarke W, Kagaayi J, et al. Decreased kidney function in a community-based cohort of HIV-infected and HIV-negative individuals in Rakai, Uganda. J Acquir Immune Defic Syndr
6. Emem CP, Arogundade F, Sanusi A, et al. Renal disease in HIV-seropositive patients in Nigeria: an assessment of prevalence, clinical features and risk factors. Nephrol Dial Transplant
7. Reid A, Stohr W, Walker AS, et al. Severe renal dysfunction and risk factors associated with renal impairment in HIV-infected adults in Africa initiating antiretroviral therapy. Clin Infect Dis
8. Mulenga LB, Kruse G, Lakhi S, et al. Baseline renal insufficiency and risk of death among HIV-infected adults on antiretroviral therapy in Lusaka, Zambia. AIDS
9. Stohr W, Walker AS, Munderi P, et al. Estimating glomerular filtration rate in HIV-infected adults in Africa: comparison of Cockcroft-Gault and Modification of Diet in Renal Disease formulae. Antivir Ther
11. Gitman MD, Hirschwerk D, Baskin CH, et al. Tenofovir-induced kidney injury. Expert Opin Drug Saf
12. Fux CA, Simcock M, Wolbers M, et al. Tenofovir use is associated with a reduction in calculated glomerular filtration rates in the Swiss HIV Cohort Study. Antivir Ther
13. Young B, Buchacz K, Baker RK, et al. Renal function in tenofovir-exposed and tenofovir-unexposed patients receiving highly active antiretroviral therapy in the HIV outpatient study. J Int Assoc Physicians AIDS Care (Chic Ill)
14. Myer L, Rabkin M, Abrams EJ, et al. Focus on women: linking HIV care and treatment with reproductive health services in the MTCT-Plus Initiative. Reprod Health Matters
15. Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron
16. Levey AS, Bosch JP, Lewis JB, et al. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med
17. Levey AS, Stevens LA, Schmid CH, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med
18. Wools-Kaloustian K, Gupta SK, Muloma E, et al. Renal disease in an antiretroviral-naive HIV-infected outpatient population in Western Kenya. Nephrol Dial Transplant
19. Peters PJ, Moore DM, Mermin J, et al. Antiretroviral therapy improves renal function among HIV-infected Ugandans. Kidney Int
20. National Kidney Foundation. National Kidney Foundation Kidney Disease Outcome Quality Initiative
. 2002. Available at: www.kdoqi.org
. Accessed August 1, 2010.
21. Go AS, Chertow GM, Fan D, et al. Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization. N Engl J Med
22. Eastwood JB, Kerry SM, Plange-Rhule J, et al. Assessment of GFR by four methods in adults in Ashanti, Ghana: the need for an eGFR equation for lean African populations. Nephrol Dial Transplant
23. van Deventer HE, George JA, Paiker JE, et al. Estimating glomerular filtration rate in black South Africans by use of the modification of diet in renal disease and Cockcroft-Gault equations. Clin Chem
APPENDIX SECTION A: MTCT-PLUS INITIATIVE PARTICIPATING SITES
ACONDA FSU and Abobo clinics, Cote d'Ivoire (Dr Siaka Toure); Moi University College of Health Sciences Clinics, Kenya (Drs Robert Eintez and Joseph Mamlin); Nyanza Provincial General Hospital Clinic, Kenya (Dr Juliana Otieno); Treatment and Research AIDS Center, Rwanda (Dr Anita Assimwe); Cato Manor Clinic of UKZN, South Africa (Dr Anna Coutsoudous); Langa Health Clinic of Western Cape, South Africa (Dr Ivan Toms); Perinatal HIV Research Unit of University of Witswatersrand, South Africa (Dr James McIntyre); Thai Red Cross Clinic, Thailand (Dr Praphan Phanuphak); MU-JHU Cares Clinic, Uganda (Dr Philippa Musoke); St. Francis Hospital Clinic, Uganda (Dr Pius Okong); Mtendere and Chelstone Health Clinics, Zambia (Dr Elizabeth Stringer).