iconstar paper   HIV Articles  
Back grey arrow rt.gif
 
 
(MACS) Long-term kidney function, proteinuria, and associated risks among HIV-infected and uninfected men
 
 
  - proteinuria, HIV, diabetes, hypertension, ARTs
 
Download the PDF here
 
"The prevalence of hypertension (56 hiv- men, 74 HIV+ HAART, and 88% HIV+, respectively; P < 0.001 for trend) and of diabetes (9, 11, and 27%, respectively; P < 0.001 for trend)
increased while current smoking prevalence decreased (40, 23, and 18%, respectively; P < 0.001 for trend) incrementally with lower baseline eGFR strata (≥90, 60-89 and <60)."
 
o These observations support current clinical practices that include aggressive control of both HIV infection and any comorbid conditions that are known to result in impaired renal function over time. Given the high and growing prevalence of traditional risk factors for chronic kidney disease among aging HAART-treated HIV+ persons, our findings support current recommendations advising that screening for kidney disease in HIV-infected persons should include routine assessments of both eGFR and proteinuria, with the latter often providing the first indication that kidney dysfunction is present. Furthermore, our findings indicate that, among HIV+ persons with proteinuria, cumulative use of specific antiretrovirals, notably TDF and atazanavir may contribute to GFR decline. The extent to which these observations should inform antiretroviral selection, particularly among aging HIV+ persons with competing risks for kidney disease, and whether the current widespread use of newer antiretrovirals such as tenofovir alafenamide that have fewer adverse renal effects or the use of NRTI-sparing HAART will ameliorate the burden of HIV-related renal disease over time, requires further evaluation and will be the subject of future work from our group.
 
o In this well characterized, prospectively followed cohort of HIV-infected and uninfected men, we observed that declines in eGFR over an 11-year period were not clearly associated with HIV serostatus (though we studied few HIV+ persons not receiving HAART) per se but were positively associated with HAART use in general. Although HIV-infected HAART-recipients more often had low baseline kidney function (eGFR < 60) than HIV-uninfected men, proteinuria was nearly seven times more common in HAART-treated HIV-infected than HIV-uninfected men. In addition, proteinuria was more common among HIV-infected persons than were eGFR declines and was concurrent with recent eGFR declines and highly predictive of subsequent eGFR declines. Proteinuria was also associated with historical exposure to didanosine, saquinavir, and/or nelfinavir use, and positively correlated with systemic inflammation. Not unexpectedly, proteinuria was likewise more common in men with classic risk factors for kidney disease, such as hypertension, diabetes mellitus, and tobacco smoking. Longitudinal measurements of eGFR and concurrent serial assessments of proteinuria were available and, among HIV-infected persons on HAART, in analytic models that included adjustment for the presence of proteinuria, our observed associations of cumulative use of TDF, emtricitabine, ritonavir, atazanavir, any protease inhibitors, or fluconazole with eGFR declines in the following year, were noteworthy......support the hypothesis that proteinuria is concurrent with and predictive of kidney disease among HIV+ persons....We found that proteinuria was strongly predictive of subsequent eGFR decline over the next year, a finding that has been identified in HIV-uninfected populations.....In the overall cohort, we noted anticipated associations between eGFR declines, proteinuria, and older age over the 11-year time period of observation represented in this report. Age-associated declines in eGFR have become more apparent in analyses of cohorts with longer periods of follow-up.
 
o Proteinuria was detected in 1613 (7%) of 22 177 person-visit pairs: 243 (2%) of person-visit pairs from 64 HIV- men, 1337 (15%) of person-visit pairs from 260 HIV+ HAART-users, and 33 (6%) of person-visit-pairs from nine HIV+ HAART-naive men. As shown in table 3, factors independently associated with proteinuria were a baseline eGFR 60-90 or less than 60, vs. more than 90; at least 3% annual decline in eGFR; HIV infection with HAART use; age at least 50 vs. less than 40 at the visit prior to proteinuria detection [V (i-1)]; presence of any of the following at the V (i-1) visit: a history of diabetes or hypertension, current smoking, HCV infection, or a CD4+ cell count 200 cells/μl or less or 201-500 (vs. >500 cells/μl). Among HAART-users, use of didanosine, saquinavir, or nelfinavir were additionally positively associated with proteinuria, while lamivudine use appeared protective.
 
o The median eGFR decline was significantly (P < 0.001) greater following confirmed proteinuria (-2.0% per year) than observed among those without proteinuria (-0.70% per year) table 4.
 
o For the 12 608 UPCR person-visit-pairs at which C-reactive protein (CRP) [inflammation marker] was measured, adding CRP (categorized as <1, 1-3, ≥3) to the final model revealed adjusted ORs for proteinuria of 1.70 (95% CI: 1.12-2.59) and 1.97 (95% CI: 1.28-3.02) for men with CRP levels of 1-3 and ≥3 vs. <1 mg/l, respectively. Similarly, men with a CRP level at least 3 (vs. <1) were more likely to have a subsequent eGFR decline of more than 3% in the next year after adjusting for proteinuria and other factors in the final model (OR = 1.13 (95% CI: 1.02-1.26).....Our finding of a positive association between plasma CRP levels and proteinuria suggests that the latter is positively linked to the presence of systemic inflammation. The increased likelihood of proteinuria observed among persons in our cohort who had coexisting chronic systemic illnesses associated with increased inflammation such as diabetes mellitus, as well as its positive association with tobacco smoking, support this hypothesis. Furthermore, our finding that, even in the presence of these comorbid conditions, more advanced HIV infection remained independently associated with proteinuria and declines in kidney function, also supports an etiologic link between chronic systemic inflammation and proteinuria
 
------------------------
 
Long-term kidney function, proteinuria, and associated risks among HIV-infected and uninfected men
 
AIDS June 19 2018 - Palella, Frank, J., Jra; Li, Xiuhongb; Gupta, Samir, K.c; Estrella, Michelle, M.d; Phair, John, P.a; Margolick, Joseph, B.b; Detels, Rogere; Kingsley, Lawrencef; Jacobson, Lisa, P.b
 
Background: Factors affecting kidney function and proteinuria among HIV-positive (HIV+) and HIV-negative (HIV-) persons need better characterization.
 
Methods: We evaluated estimated glomerular filtration rate (eGFR, ml/min per 1.73 m2) changes, proteinuria prevalence (a urine protein-to-creatinine ratio of ≥0.2 at two consecutive visits) and associated factors among HIV+ and HIV- men.
 
Results: There were 917 HIV+ men receiving HAART, 159 HIV+ men not receiving HAART, and 1305 HIV- men seen from October 2003 to September 2014.
 
Median annual eGFR change was -0.5, -0.8% for HIV+ and -0.3% for HIV- men (P < 0.001). Factors significantly (P < 0.05) associated with more than 3% annual eGFR decline were HAART receipt (but no specific antiretroviral drug), age more than 50, hypertension, diabetes, current smoking.
 
Proteinuria existed in 14.9% of visit-pairs among HAART recipients, 5.8% among non-HAART recipients, and 1.9% among HIV- men, and was associated with subsequent annual more than 3% eGFR decline (odds ratio 1.80, P < 0.001).
 
Proteinuria-associated factors also included HAART use (vs. HIV-), age at least 50 (vs. <40), diabetes, hypertension, current smoking, hepatitis C virus-infection (all P < 0.05) and, among HIV+ men, lower CD4+ cell count, didanosine, saquinavir, or nelfinavir use (all P < 0.05). After adjusting for proteinuria, among HAART users, having a detectable HIV RNA, cumulative use of tenofovir disoproxil fumarate, emtricitabine, ritonavir, atazanavir, any protease inhibitor, or fluconazole were associated with more than 3% annual eGFR decline.
 
Conclusion: Longitudinal kidney function decline was associated with HAART use but no individual antiretroviral drug, and traditional kidney disease risks. Proteinuria was nearly seven times more common in HAART-treated men than HIV- men, reflected recent eGFR decline and predicted subsequent eGFR decline.
 
Background
 
Kidney disease is an important cause of morbidity among HIV-positive (HIV+) persons . Although HIV itself is a known kidney pathogen and accounted for much kidney disease among HIV+ persons before the advent of HAART , it is now clear that diverse multifactorial processes can impact kidney function in HAART-treated HIV+ persons . Nuanced longitudinal characterization of kidney function over time and of factors impacting it among HIV+ persons receiving HAART is needed. Using an appropriate control group of HIV-negative (HIV-) but at-risk persons allows differentiation of HIV-specific risk factors from those that are not unique to HIV+ individuals, such as hypertension and diabetes mellitus. Likewise, the role of systemic inflammation in kidney disease development, particularly among HIV+ persons, requires better evaluation . These all are important to evaluate among HAART-treated virally suppressed HIV+ persons, for whom survival has been extended and in whom chronic comorbidities, aging, antiretroviral therapies, and systemic inflammation may contribute to the development and progression of kidney disease .
 
Estimated glomerular filtration rate (eGFR) and urine protein-to-creatinine ratios (UPCR) are complementary indicators of kidney function and damage widely used for kidney disease screening and monitoring in the general population . Although GFR estimates are standard monitoring tools in the management of HIV+ persons, quantitative assessments of proteinuria are less common, and concurrent examination of both are uncommon in this population.
 
In this report, we assessed factors impacting longitudinal kidney function and proteinuria among HIV+ and demographically similar HIV- men followed in the Multicenter AIDS Cohort Study (MACS).
 
Methods
 
The cohort

 
The MACS is an ongoing multicenter, prospective study of MSM. MACS sites in Baltimore, Maryland; Washington, District of Columbia; Chicago, Illinois; Los Angeles, California; and Pittsburgh, Pennsylvania, USA have enrolled a total of 6972 men in 1984-1985, 1987-1990, and 2001-2003. The MACS prospectively schedules participants for semi-annual study visits where medical and therapy histories, physical and behavioral assessments, and biological specimens are collected under standardized protocols. Individuals eligible for the current study were HIV+ and HIV- participants who contributed at least three serum creatinine (SCr) measurements from visit 40 (2003), when SCr was first measured in the study, through visit 61 (2014).
 
Hepatitis B virus (HBV) coinfection was determined by the presence of a positive hepatitis B surface antigen. Hepatitis C virus (HCV) coinfection was defined by a positive HCV RNA. HAART was defined in accordance with US Department of Health and Human Services treatment guidelines (www.aidsinfo.gov; November 2014 version) as use of three or more antiretroviral (ART) drugs including one or more of protease inhibitors, nucleoside reverse transcriptase inhibitors (NRTI), a non-NRTI, integrase inhibitor, or entry inhibitor (including fusion inhibitors). Information regarding the presence of traditional risk factors for chronic kidney disease (e.g. hypertension, diabetes) were based on a combination of participant self-report; study-obtained measurements; medication usage. Informed consent from each participant and approval from each local institutional review board were obtained.
 
Outcomes and analytic methods
 
The baseline visit was defined as the first visit with SCr data. Participants with incident HIV infection (n = 93) or those HIV+ men who initiated HAART during the study period (n = 124) were included as HIV- before HIV infection and HIV+ but not receiving HAART before HAART initiation, respectively. GFR was estimated in ml/min per 1.73 m2 using the SCr-based Chronic Kidney Disease Epidemiology Collaboration equation . Baseline characteristics of HIV- and HIV+ subject groups were compared using the Wilcoxon test or the chi-square test, when appropriate. Cochran-Armitage Trend Test was used to examine trends in the prevalence of hypertension, diabetes mellitus, and current smoking by baseline eGFR strata (≥90, 60-89, and <60).
 
We first evaluated longitudinal eGFR changes over an 11-year study period. For each person, a simple linear regression of their log-transformed eGFR with years of follow-up from baseline as a predictor was conducted. After examining the linear trend of eGFR over time, the annual eGFR percentage change was calculated using this formula: [exp(log gfrslp) - 1] x 100%, where log gfrslp was the estimate of change in log-transformed eGFR per year obtained from the linear regression. Individuals were then classified according to whether the annual eGFR declined by more than 3%. Multivariable logistic regression models were constructed to evaluate the association of potential risk factors with the odds of annual eGFR decline more than 3%. Potential risk factors examined were baseline characteristics: HIV/HAART status [three groups: HIV- (reference), HIV+ no-HAART, HIV+ HAART], eGFR categories [<60, 60-89, ≥90 (reference)], age [<40 (reference), 40-49, 50-59, ≥60], race [White (reference), African-American, and Hispanic or other race], a history of diabetes, a history of hypertension, cigarette smoking status [current, former, never (reference)], HBV and/or HCV infection. Cohort entry status [pre-2001 (reference) vs. 2001-2003 cohort] was included to control for additional differences in the cohorts, and the number of eGFR measurements contributed per person. Among HIV+ men, a history of a clinical AIDS diagnosis, baseline CD4+ T-cell counts [≤200, 201-350, 351-500, and >500 cells/μl (reference)] and detectable plasma HIV RNA (≥50 copies/ml) levels were also evaluated in association with eGFR decline. Among HIV+ HAART-users, years from HAART initiation to baseline and cumulative years on individual antiretroviral drugs from HAART initiation to baseline were also examined in separate models. Baseline eGFR levels and age were fixed in the final model; other factors with a P value less than 0.1 were kept in the final multivariable model.
 
We then measured the prevalence of confirmed proteinuria, defined as a UPCR at least 0.2 at two consecutive visits (no more than 1-year apart), and evaluated factors associated with confirmed proteinuria among HIV+ and HIV- participants from April 2006 (when UPCR measurements were added to the protocol) through September 2014. This definition of proteinuria was used to minimize the likelihood that the presence of proteinuria was due to a transient process not related to chronic kidney disease. Multivariable logistic regression models were used to determine factors associated with proteinuria. In this analysis, except for HIV/HAART status and baseline eGFR, other exposure variables were time-varying and included those taken from the visit [V (i-1)] prior to the first visit of the visit-pairs at which proteinuria was assessed. Potential risk factors considered in these models included those present in the eGFR analyses, as well as the presence of annual eGFR decline more than 3% from baseline to V (i). Separate models were constructed for HAART-users when examining specific antiretroviral drugs. Generalized estimating equations were used to analyze longitudinal data with repeated measures.
 
We also examined whether proteinuria predicted eGFR change. The key exposure variable here was proteinuria, and the outcome variable was annual eGFR decline more than 3% in the next year [i.e. the calculation was based on three eGFRs: GFR at the current visit (V (i) when proteinuriawas determined) and the two visits that followed]. Other exposure variables adjusted in multivariable logistic regression models were similar to the previous analysis of proteinuria, but were taken at V (i).
 
Multiple imputation was performed for missing risk factor data in multivariable models. For each outcome, the multiple imputation included all predictors and the outcome. Missing values were imputed five times on the basis of the distribution of covariates by using a Markov chain Monte Carlo method assuming multivariable normality . Missing values were imputed for multiple regression analyses for smoking status (2% of all person-visits data for analysis), CD4+ cell count (<1% among HIV+), HIV RNA (<1% among HIV+), HCV-infection (<1%), and HBV-infection (<1%). All analyses were done by using SAS 9.2 (SAS Institute, Cary, North Carolina, USA).
 
Evaluation of high sensitivity C-reactive protein
 
We cross-sectionally evaluated high-sensitivity C-reactive protein serum levels which were available from specimens drawn at routine visits during April to September 2006, to assess associations of systemic inflammation with eGFR and proteinuria.
 
Results
 
Baseline characteristics

 
We analyzed eGFR changes among 2381 men. At baseline, 1305 (54%) men were HIV-, 917 (39%) were HIV+ and had initiated HAART, and 159 (7%) were HIV+ and HAART-naive (). Among the 917 HAART users in our study, 60 were pre-HAART ART-experienced. shows differences in baseline characteristics in these patient groups with regard to hypertension, diabetes, HBV and HCV serostatus, and HIV clinical indicators.
 
Risk factors and estimated glomerular filtration rate at baseline
 
Median baseline eGFR was 90 for HIV- men, 94 for HIV+ HAART-users, and 97 for HIV+ HAART-naive men (P = 0.004 for HIV- vs. HIV+ HAART-users, P < 0.001 for HIV- vs. HIV+ HAART-naive men). The proportion of men with baseline eGFR < 60 was higher among HIV+ men receiving HAART (6%) than HIV- men (4%, P < 0.001).
 
The prevalence of hypertension (56, 74, and 88%, respectively; P < 0.001 for trend) and of diabetes (9, 11, and 27%, respectively; P < 0.001 for trend) increased while current smoking prevalence decreased (40, 23, and 18%, respectively; P < 0.001 for trend) incrementally with lower baseline eGFR strata (≥90, 60-89 and <60).
 
Association of baseline clinical factors with longitudinal estimated glomerular filtration rate
 
The median number of eGFR assessments performed per person during the 11-year follow-up period was 19 for HIV- men, 19 for HIV+ HAART-users, and seven for HIV+ HAART-naive men. Median annual eGFR change was -0.5% overall, -0.8% for HIV+ men and -0.3% for HIV- men (P < 0.001). A 3% annual decline in eGFR represented approximately the lowest 15th percentile of the range of observed changes.
 
For examining change in eGFR, men aged 50-59 and at least 60 years at baseline demonstrated similar trends, so were combined in the final multivariable model. Factors independently associated with more than 3% annual eGFR decline included: HIV+ HAART+, age at least 50, current smoking, a history of diabetes or hypertension, HCV infection, and MACS enrollment in 2001-2003 (vs. earlier, ). Men with baseline eGFR 60-89 (compared with ≥90), Hispanic or other race (compared with White) and men with more eGFR measurements were less likely to have more than 3% annual eGFR decline. In this model, HBV infection, a history of AIDS, CD4+ cell count, and detectable HIV RNA were not independently associated with eGFR decline. Among HAART-users, no specific antiretroviral drug use before baseline was independently associated with eGFR decline over the 11-year study period after adjusting for factors included in previous model.
 
Proteinuria analysis
 
Proteinuria was detected in 1613 (7%) of 22 177 person-visit pairs: 243 (2%) of person-visit pairs from 64 HIV- men, 1337 (15%) of person-visit pairs from 260 HIV+ HAART-users, and 33 (6%) of person-visit-pairs from nine HIV+ HAART-naive men. Proteinuria was observed in 522 person-visit-pairs (36%) of the eGFR less than 60 group, 694 (7%) of the eGFR 60-90 group, and 397 (4%) of the eGFR more than 90 group (). Within each eGFR stratum, proteinuria occurred more frequently among HIV+ HAART-users than HIV- men (all P < 0.01). Among 2059 men with at least three eGFR measurements and at least two urine UPCR measurements performed at two consecutive visits over the 8-year period of follow-up, men with annual eGFR declines of more than 3% more often exhibited proteinuria in at least one visit pair (45%) than their counterparts [12%, odds ratio (OR) = 5.8, P < 0.001].
 
As shown in , factors independently associated with proteinuria were a baseline eGFR 60-90 or less than 60, vs. more than 90; at least 3% annual decline in eGFR; HIV infection with HAART use; age at least 50 vs. less than 40 at the visit prior to proteinuria detection [V (i-1)]; presence of any of the following at the V (i-1) visit: a history of diabetes or hypertension, current smoking, HCV infection, or a CD4+ cell count 200 cells/μl or less or 201-500 (vs. >500 cells/μl). Among HAART-users, use of didanosine, saquinavir, or nelfinavir were additionally positively associated with proteinuria, while lamivudine use appeared protective.
 
The median eGFR decline was significantly (P < 0.001) greater following confirmed proteinuria (-2.0% per year) than observed among those without proteinuria (-0.70% per year). The unadjusted ORs of having a subsequent annual eGFR decline of more than 3% among persons with proteinuria vs. without proteinuria were 1.74 [95% confidence interval (CI): 1.42-2.13, P < 0.001], 2.09 (95% CI: 1.79-2.46, P < 0.001), and 2.38 (95% CI: 1.84-3.07, P < 0.001) for men with GFR at least 90, 60-89, and less than 60 at V (i), respectively. However, of the 1613 person-visit pairs with proteinuria, 39% had eGFR annual decline more than 3% before proteinuria was measured, while only 12% had prior eGFR annual decline more than 3% of 20 564 person-visit pairs without proteinuria. illustrates that proteinuria was significantly associated with subsequent eGFR decline (>3%) in the following year, after adjusting for other factors significantly associated with eGFR decline: a beginning eGFR more than 90, HIV infection, HAART receipt, age more than 40, cohort entered study in 2001-2003, and a history of diabetes or hypertension. Inclusion of CD4+ cell count in the multivariable model did not attenuate the association between proteinuria and subsequent eGFR decline (OR = 1.77, 95% CI: 1.56-2.01). In addition, among HAART-users, a detectable HIV RNA, cumulative use of tenofovir disoproxil fumarate (TDF), emtricitabine, ritonavir, atazanavir, any protease inhibitors, or fluconazole were positively associated with a subsequent eGFR annual decline of more than 3% in separate models adjusting for proteinuria, HIV/HAART status, baseline eGFR level, age, race, cohort entry, and a history of diabetes or hypertension. The association of proteinuria with subsequent eGFR decline more than 3% tended to be stronger in men with eGFR less than 90 ().
 
C-reactive protein subanalysis
 
For the 12 608 UPCR person-visit-pairs at which C-reactive protein (CRP) was measured, adding CRP (categorized as <1, 1-3, ≥3) to the final model revealed adjusted ORs for proteinuria of 1.70 (95% CI: 1.12-2.59) and 1.97 (95% CI: 1.28-3.02) for men with CRP levels of 1-3 and ≥3 vs. <1 mg/l, respectively. Similarly, men with a CRP level at least 3 (vs. <1) were more likely to have a subsequent eGFR decline of more than 3% in the next year after adjusting for proteinuria and other factors in the final model (OR = 1.13 (95% CI: 1.02-1.26).
 
Discussion
 
In this well characterized, prospectively followed cohort of HIV-infected and uninfected men, we observed that declines in eGFR over an 11-year period were not clearly associated with HIV serostatus (though we studied few HIV+ persons not receiving HAART) per se but were positively associated with HAART use in general. Although HIV-infected HAART-recipients more often had low baseline kidney function (eGFR < 60) than HIV-uninfected men, proteinuria was nearly seven times more common in HAART-treated HIV-infected than HIV-uninfected men. In addition, proteinuriawas more common among HIV-infected persons than were eGFR declines and was concurrent with recent eGFR declines and highly predictive of subsequent eGFR declines. Proteinuria was also associated with historical exposure to didanosine, saquinavir, and/or nelfinavir use, and positively correlated with systemic inflammation. Not unexpectedly, proteinuria was likewise more common in men with classic risk factors for kidney disease, such as hypertension, diabetes mellitus, and tobacco smoking.
 
Longitudinal measurements of eGFR and concurrent serial assessments of proteinuria were available and, among HIV-infected persons on HAART, in analytic models that included adjustment for the presence of proteinuria, our observed associations of cumulative use of TDF, emtricitabine, ritonavir, atazanavir, any protease inhibitors, or fluconazole with eGFR declines in the following year, were noteworthy. Associations between TDF use and GFR declines have been well described in other cohorts . Such declines are not always reversible , may be cumulative, and may result in drug discontinuation. During the 11-year study period, HIV-infected participants often switched their HAART regimen over time (data not presented), which could have affected their subsequent eGFR levels. This may partly explain why no associations were seen between individual antiretroviral drugs used before baseline and the eGFR decline over the 11-year period. Recent data from other cohorts have indicated associations between cumulative atazanavir use and kidney function decline . Although our observed association between ritonavir use and eGFR decline was not unanticipated because of the drug's inhibition of renal tubular creatinine secretion, eGFR declines may also have occurred because of ritonavir's intended effects of increasing levels of atazanavir and of TDF, both drugs that can exhibit dose-dependent nephrotoxicity . The association of emtricitabine use with eGFR decline likely reflects the fact that it was virtually always coprescribed with TDF during the time period included in this study, and vice-versa, rather than emtricitabine-induced adverse kidney effects. Although our observed association between fluconazole use and eGFR decline was not expected, it may be related to the fact that fluconazole use (both in the general HIV+ population as well as in our cohort) has been most often associated with having more advanced HIV infection (data not shown), higher plasma HIV viremia and lower CD4+cell count, all factors associated historically with worse kidney function among HIV+ persons. The same principle may be have been operative in our finding of the association between historical exposure to didanosine, saquinavir, or nelfinavir with proteinuria; all of these were antiretroviral drugs used primarily in the early HAART era among patients who in general had more advanced HIV infection than more recent HAART-initiators. It is possible, though, that unmeasured indices of overall health may have influenced our observed associations between use of specific antiretroviral drugs with proteinuria and/or eGFR declines, despite having controlled for stage of HIV disease and the presence of important comorbid diseases .
 
The strong associations observed between proteinuria and HIV infection, HAART use, and more rapid declines in eGFR, were less pronounced but still apparent even among persons with baseline eGFRs in the normal range and support the hypothesis that proteinuria is concurrent with and predictive of kidney disease among HIV+ persons . We found that proteinuria was strongly predictive of subsequent eGFR decline over the next year, a finding that has been identified in HIV-uninfected populations . Furthermore, the observed associations of known non-HIV causes of kidney disease (such as hypertension and diabetes) and proteinuria in our cohort are consonant with observations from other cohorts of HIV+ persons , attesting to the sensitivity of proteinuria as a marker of incipient decline in GFR . The extent to which and how HIV infection, specific antiretroviral drug use, and these other comorbidities may interact to hasten the onset of proteinuria among HIV+ persons with other risk factors for kidney disease, is unclear. Nevertheless, the identification of proteinuria as a marker of eGFR decline may have implications for future research seeking to evaluate whether or not, among HIV-infected persons, specific therapeutic drug interventions for hypertension (such as angiotensin-converting enzyme-inhibitors/AT1 blockers) can ameliorate proteinuria-associated declines in kidney function, as has been observed in some subgroups of the proteinuric general population. Our finding of a positive association between plasma CRP levels and proteinuria suggests that the latter is positively linked to the presence of systemic inflammation. The increased likelihood of proteinuria observed among persons in our cohort who had coexisting chronic systemic illnesses associated with increased inflammation such as diabetes mellitus, as well as its positive association with tobacco smoking, support this hypothesis. Furthermore, our finding that, even in the presence of these comorbid conditions, more advanced HIV infection remained independently associated with proteinuria and declines in kidney function, also supports an etiologic link between chronic systemic inflammation and proteinuria. Other recent work from our group and others supports this . The proportional contributions of HIV infection, specific antiretroviral drug use, and the presence of inciting comorbid diseases toward the development of kidney disease (as evidenced by either proteinuria and eGFR decline) probably varies from individual to individual and likely has much to do with the severity and chronicity of the underlying systemic diseases as well as the success of medical management of each of the component illnesses.
 
In the overall cohort, we noted anticipated associations between eGFR declines, proteinuria, and older age over the 11-year time period of observation represented in this report. Age-associated declines in eGFR have become more apparent in analyses of cohorts with longer periods of follow-up. Furthermore, the association between higher baseline eGFR (>90) with increased likelihood for subsequent declines in eGFR () at first glance seems counterintuitive, but may reflect fact that persons with higher GFRs have greater room for decline, particularly since, in general, we did not find differentially higher rates of non-HIV risks for kidney disease among persons with higher baseline eGFRs, and may possibly be related to the phenomenon of hyperfiltration, which has been linked to more rapid subsequent kidney function decline among HIV+ persons .
 
Limitations to this analysis exist. There may have been factors influencing eGFR decline that we did not or could not directly measure. Also, determinations of relationships between specific antiretroviral use and eGFR declines or proteinuria can be complex. Most of our measurements were made while persons were already receiving HAART; most of these regimens did not constitute a given patient's first exposure to antiretrovirals. Previously received antiretrovirals were not taken into account, and we did not evaluate pre-HAART eGFRs systematically. This analysis reflects persons who entered care prior to 2003 and stayed in care for 11 years and may not be generalizable to persons who more newly entered care. Also, our report is subject to limitations that are inherent to analyses of any observational cohort. Another potential limitation is the inherent inaccuracy of eGFRs, even those based upon the most modern estimating equations, particularly among persons with higher true GFRs .
 
These observations support current clinical practices that include aggressive control of both HIV infection and any comorbid conditions that are known to result in impaired renal function over time. Given the high and growing prevalence of traditional risk factors for chronic kidney disease among aging HAART-treated HIV+ persons, our findings support current recommendations advising that screening for kidney disease in HIV-infected persons should include routine assessments of both eGFR and proteinuria, with the latter often providing the first indication that kidney dysfunction is present. Furthermore, our findings indicate that, among HIV+ persons with proteinuria, cumulative use of specific antiretrovirals, notably TDF and atazanavir may contribute to GFR decline. The extent to which these observations should inform antiretroviral selection, particularly among aging HIV+ persons with competing risks for kidney disease, and whether the current widespread use of newer antiretrovirals such as tenofovir alafenamide that have fewer adverse renal effects or the use of NRTI-sparing HAART will ameliorate the burden of HIV-related renal disease over time, requires further evaluation and will be the subject of future work from our group.

 
 
 
 
  iconpaperstack View Older Articles   Back to Top   www.natap.org