Control of albuminuria in overt diabetic nephropathy: durability counts

Abnormal urinary excretion of albumin (albuminuria) is widely regarded as a potentially modifiable risk factor for the development of cardiovascular (CV) as well as renal events, such as slowing progression of chronic kidney disease (CKD) to its end stages [1]. Indeed, albuminuria has been incorporated into prognosis-stratified classifications of CKD used globally for its definition and classification [2]. Albuminuria associates with the risk of CV disease, even at levels barely above the normal range [3], although the directionality of this association remains in dispute [4]. Progression of CKD to end-stage disease is more directly linked to the presence of albuminuria, although the absolute thresholds of the magnitude of albuminuria associated with disease progression can vary greatly from disease to disease [5]. A remission of high-grade albuminuria or proteinuria as assessed by time averaged or time-variable measurements, whether spontaneous or therapeutically induced, is accepted as portending a more favorable long-term evolution in non-diabetic glomerular disease [6]. Persistent albuminuria is also a well-recognized marker for underlying pathology in diabetic kidney disease, but progression of CKD can occur even without albuminuria in some subjects with type 2 diabetes mellitus [7]. In both type 1 and type 2 diabetes lower levels of albuminuria (30–300 mg/day; incipient nephropathy) can ‘spontaneously’ remit, especially with improved glycemic or blood pressure control [8, 9]. At higher levels of albuminuria (>300 mg/day; overt nephropathy), complete remissions of albuminuria are much less common with intensified management, but reductions in urinary albumin excretion rate, albeit not to normal levels, can alter the progression of diabetic kidney disease, at least to some degree [10, 11]. Both the magnitude and the duration of the reduction in abnormal albuminuria need to be taken into account in determining the overall impact of changes in urinary albumin excretion on the subsequent rate of decline in renal function in CKD, including that due to diabetic kidney disease. While short-term (e.g. 3 months) reductions in albumin excretion rate induced by better control of blood pressure and the use of agents affecting the renin–angiotensin–aldosterone system (RAAS inhibitors) predict the degree of both renal and CV protection in diabetic kidney disease, the persistence of the antialbuminuric effects of these interventions must also be considered in estimating their impact on outcomes. Thus, the durability of a response is an important element in the design of trials aimed at evaluation of new renoprotective agents in CKD of various causes. In order to examine this issue, Krὃpelin et al. [12] examined the combined results of both the Reduction in End Points in NIDDM with the Angiotensin II Antagonist Losartan and Irbesartan Diabetic Nephropathy Trial randomized controlled studies of angiotensin receptor blockers (ARBs) in patients with type 2 diabetes mellitus and overt nephropathy (baseline serum creatinine levels of 1–3 mg/dL and albuminuria >900 mg/day). The objective of this post hoc composite analysis was to evaluate the durability (between 3 and 12 months) of the antialbuminuric effects of the interventions (and placebo) on longer-term (36 month) outcomes, including doubling of serum creatinine from baseline and development of treated ESRD, hospitalization for congestive heart failure and all-cause mortality. Among the ARB-treated subjects, an initial response (>30% decline in albuminuria from baseline by 3 months) was observed in 36% of the subjects; at 12 months, 45% of these ‘initial responders' had experienced a further decline in albuminuria (enhanced responders), 32% had a stable albuminuric response (sustained responders) and 24% exhibited an increase in albuminuria (escaped responders). Thus, in the initial responder group, 76% experienced a sustained or enhanced response when followed for an additional 9 months. An enhanced response was associated with improved renal outcome with follow-up from 12 to 36 months compared with a sustained or an escaped response. The total exposure to albuminuria (time-averaged albuminuria) over the 12-month period predicted renal outcomes better than the 3-month evaluation. The covariate-adjusted event rate for ESRD or doubling of serum creatinine among sustained responders (28%) was similar to the event rate (29%) for sustained nonresponders, but much less for enhanced responders (13%) and similar (26%) to that of escaped responders. The favorable renal outcomes of initial responders were driven largely by those that experienced an enhanced response between 3 and 12 months of treatment. Similarly, the poor outcome of the initial nonresponders was driven largely by those that experienced an enhanced nonresponse between 3 and 12 months of treatment. Placebo-treated subjects experienced considerable ‘spontaneous variability’ albuminuria over time, but the time-averaged albuminuria in these subjects also showed greater risk for renal events in those with higher albumin excretion rates. The message of this post hoc analysis is very clear. The durability of a decline in albuminuria induced by ARB therapy after the initial 3-month observation period is a potentially crucial determinant of long-term outcome. While the post hoc character of the analysis raises the issue of hidden confounding biases, its conclusions tend to negate the recommendations of the American College of Physicians that patients with type 2 diabetes mellitus and overt nephropathy under treatment with RAAS inhibitors do not need to be monitored with sequential estimates of albumin excretion [13]. The impact of the durability of a decline in albuminuria on hospitalization for congestive heart failure was similar to that observed for renal outcomes, but no impact was seen on all-cause mortality, perhaps because of the short follow-up period (36 months) and low power to detect such a connection. The study did not detect any baseline characteristics that were capable of predicting the patterns of response following the initial changes in albuminuria, including baseline sodium excretion, hemoglobin A1c levels, BMI or estimated glomerular filtration rate (eGFR). However, initial responders who later escaped (escaped responders) tended to have higher blood pressure during the 3–12-month period of observation, raising questions about compliance with treatment regimens or changes in diet (such as sodium chloride intake). This might also explain a higher rate of admission for congestive heart failure. Several caveats must be considered for extrapolation of these findings to the general population of patients with type 2 diabetes mellitus and overt nephropathy. First, the trial protocols required a fixed dose of ARB and the need to control blood pressure to the same extent in the placebo and treated groups. A strategy of up-titration of ARB dose according to the response of albuminuria, or even the sequential use of combined ARB and angiotensin-converting enzyme inhibitors, would have undoubtedly resulted in markedly different findings. Also, such dual RAAS inhibition and greater use of aldosterone receptor antagonists for the escaped responders is now becoming more practical with the advent of new oral potassium binding resins to forestall the undesirable side effect of hyperkalemia, which can limit the dosage of RAAS inhibitors [14]. Second, the effect of variations in overall control of phosphate and sodium homeostasis on the patterns of albuminuria response were not assessed, and both of these perturbations can apparently have a marked influence on the effectiveness of RAAS inhibition on albuminuria and renal outcomes [15, 16]. Third, blood pressure control was assessed by casual office/clinic measurements only. The contribution of poor 24-hour control of blood pressure on the patterns of albuminuria (escape responders and enhanced nonresponders) might be significant and contribute to both the congestive heart failure as well as the adverse renal outcomes. Finally, the role of noncompliance as a determinant of the patterns of albuminuria cannot be ignored. Despite these caveats, this analysis sheds new light on the considerable long-term variability of albumin excretion rate in patients with type 2 diabetes mellitus and overt nephropathy, both with and without ARB treatment. The origins of this variability cannot be fully explored by this study design, but they are likely to be multifactorial, involving compliance, diet, hormonal adjustments and altered vascular reactivity, to name a few. The crucial point, deserving of emphasis, is that the enhancement as well as the durability of a decline in albuminuria with ARB treatment is an important surrogate for favorable longer-term renal outcomes. Systematic augmentation of efforts to continually reduce the magnitude of albuminuria over extended time periods would seem to be an appropriate conclusion of this post hoc analysis, but this hypothesis will have to be tested in prospective randomized trials. The findings of this study also will likely have an impact on the design of future clinical trials evaluating the efficacy and safety of new agents on the natural history of type 2 diabetes mellitus and overt nephropathy by requiring regular postintervention monitoring of albuminuria to assess the durability of an initial response. This invited paper has not been submitted to any other journal. The author has no conflicts of interest to disclose. (See related article by Felix Kröpelin et al. Individual long-term albuminuria exposure during angiotensin receptor blocker therapy is the optimal predictor for renal outcome. Nephrol Dial Transplant 2016; 31: 1471–1477)