In this diverse, contemporary cohort of HIV-infected patients, we found that patients receiving ART continued to lose kidney function despite achieving durable viral suppression. Consistent with prior studies, we found that untreated poorly controlled HIV infection was associated with ongoing loss of kidney function and provision of ART had a beneficial effect on kidney function over time [15
]. However, effective ART did not fully reverse the risk of eGFR loss. Although traditional risk factors for kidney disease such as diabetes and hypertension were found to be important risk factors for continued kidney function decline in this group, episodes of low-level viremia of below 500 copies/ml (‘blips’) were stronger predictors of eGFR loss. Overall, these findings suggest that ART is beneficial to kidney function, but fully suppressive treatment may be needed to curb eGFR decline.
Our results build on the observations of previous studies [15
] by demonstrating the benefit of ART in a relatively unselected population of patients (i.e., heterogeneous in terms of race and underlying kidney disease). However, it is important to note that patients receiving ART continued to lose kidney function and even small increases in the viral load were strongly associated with eGFR loss. The observed effect of active viral replication on kidney function is consistent with laboratory studies [18
] that have demonstrated that HIV directly infects all cell types in the kidney and actively replicates there in a wide variety of patients, including those with and without HIV-associated nephropathy (HIVAN) or clinically overt kidney disease, and in patients of nonblack race. It is possible that breakthrough viremia (or ‘blips’) provide reasons for the continued decline in kidney function among HIV patients receiving ART, and suggest that complete and durable viral suppression may be needed to minimize risk.
The impact of HIV disease, its treatment, and immune function on kidney function is complex. Although most prior studies have indicated that ART is beneficial for the treatment of HIV-related kidney diseases, more recent long-term data from the SMART continuation study demonstrate that the relationship between ART and kidney function may be more complicated than originally thought [2
]. In the SMART trial, ART-associated renal benefit was suggested by the observation that participants who interrupted therapy had higher rates of ESRD than those receiving continuous therapy. However in contrast, a more recent analysis [23
] from this study reported that after re-initiation of ART, more ESRD events were observed in the group that received continuous versus interrupted therapy. Similarly, a study [24
] of patients with biopsy-proven kidney disease found that only individuals with HIVAN benefited from ART, whereas the ART effect in those without HIVAN suggested possibility of harm. In our study, among patients effectively controlling viral replication with ART compared with those who maintained low levels of viremia via host mechanisms (treated versus untreated controllers), we found that those exposed to ART manifested higher rates of kidney function decline.
Taken together, these findings suggest that factors related to race, underlying kidney disease, immune status, or viral replication may interact with ART to produce contrasting effects. Additional studies are needed to identify HIV-infected persons who may be susceptible to declines in kidney function as a result of ART and to determine whether declines in kidney function are related to complications of treatment, such as diabetes or hypertension, or nephrotoxicity associated with the drugs themselves. This information will be relevant to clinical practice, as today most HIV-infected patients receive or will receive ART and lifetime exposure to ART may increase with earlier initiation of therapy [25
]. Clinicians will need to counsel patients on the risk of kidney disease with long-term therapy.
Prior studies [3
] of HIV-related kidney disease have conflicted in their assessment of risk factors such as CD4 cell count. These discrepancies may be in part due to study design, as most have analyzed CD4 cell count and viral load measured at a single time point as fixed baseline covariates [27
], use dichotomous versions of these variables (e.g. CD4 cell count <200 cells/μl) [3
], or analyze CD4 cell count or viral load in isolation [17
]. We found that CD4 cell count was not prognostic of kidney function decline in any of the groups studied. Our results are consistent with Post et al.
] who reported no difference in CD4 cell counts above 200 cells/μl between those who developed ESRD and those who maintained stable renal function in a cohort of HIVAN patients. Similarly, Gardner et al.
] found that viral load, but not CD4 cell count, was associated with changes in creatinine clearance in a cohort of women. On the basis of these studies and the results reported here in treated controllers, it appears that virus-specific mechanisms may be more important than host immunity in the development of HIV-related kidney disease, and the benefit of ART is more closely tied to a reduction in viremia versus improvements in immune status. Alternatively, mechanisms of the host response (such as inflammation), not captured by the CD4 cell count, may influence the development of kidney disease [34
]. Novel measures, such as T-cell activation, may be needed to further investigate the role of host immunity in HIV-related kidney diseases [36
Our study has several limitations that deserve comment. First, the determination of treatment effects from observational data is difficult because hidden confounding may bias results. Similarly, our comparison of treated and untreated controllers may be affected by factors not available in the SCOPE data set. However, it is reassuring that these analyses were consistent in their conclusions and with previous studies [2
] of ART and kidney function. Second, although eGFR has not been validated against the gold standard of measured GFR in HIV-infected persons or in the upper range of kidney function in the general population, change in eGFR over time is a valid and clinically relevant outcome measure [37
]. In addition, although the magnitude of eGFR loss in this study was substantial, it was consistent with prior studies in HIV-infected persons [16
]. Third, we did not have histopathologic data or specific kidney disease diagnoses available for this analysis. Therefore, it is possible that certain disease entities found among HIV-infected persons may behave differently than what we found in our study. Fourth, we were unable to analyze the contribution of several important factors including signs of kidney damage (e.g. proteinuria or hematuria), vital signs (e.g. blood pressure measurements), or antihypertensive medication exposure (e.g. angiotensin-converting enzyme inhibitors) because these data were not systematically collected in SCOPE. Finally, we were under-powered to analyze the impact of individual antiretroviral drugs, which may have been responsible for eGFR loss in treated patients. Specifically, tenofovir and indinavir have been identified as predictors of kidney disease in other studies, but could not be reliably analyzed in this study because these drugs were poorly represented in our cohort. Further study is needed to replicate our results and prospectively examine the effects of these key predictors in a larger cohort of treated persons with HIV.
We found that ART is associated with improvements in kidney function over time; however, overall rates of eGFR decline were still significant among those successfully treated with ART. Among the risk factors studied, viremia appears to be most strongly associated with loss of kidney function and may contribute to continued rates of kidney function decline among those who are successfully treated with ART.