Albuminuria as a risk factor for acute kidney injury: what is the evidence?

The discovery of the link between proteinuria and kidney disease is firmly established as having been made by Richard Bright in 1827. But in a 2003 review, Cameron [1] points out that protein in the urine of patients with dropsy and anasarca was described even before, in the 16th century by Paracelsus, and then many others, much before Richard Bright. The presence of proteinuria is used to define both the presence and the severity of chronic kidney disease (CKD). Efforts are now afoot to have it accepted as a sufficient surrogate marker for clinical trials and drug development [2]. In contrast to this well-established association between proteinuria and CKD, the relation of proteinuria with acute kidney injury (AKI) remains much less well defined, but a growing body of evidence supports this link (Table 1). In a 2008 case–control study from California, Hsu et al. [3] reported that pre-existing proteinuria, as measured with a dipstick, was a risk factor for hospital-acquired AKI requiring dialysis. A more detailed assessment of this link came from the prospective Atherosclerosis Risk In Communities cohort study of Grams et al. [4], which reported that albuminuria was associated with AKI after adjusting for kidney function and comorbid conditions. Moreover, a dose–response relation was reported in this study, with the adjusted relative hazard being 1.9, 2.2 and 4.8 for subclinical albuminuria, microalbuminuria and macroalbuminuria, respectively, compared with no albuminuria. A larger study using the provincial database in Alberta, Canada from James et al. [5] included just under one million individuals, of whom 6520 developed AKI. It found a similar graded association according to dipstick proteinuria, with an adjusted rate of 2.5 times for mild proteinuria (trace or 1+) and 4.4 times for those with heavy proteinuria (≥2+) relative to those with none (negative). Several other studies, mostly from cardiac surgery, also report on this association and are summarized in Table 1 [3–12]. In particular, the systematic review by James et al. [13] synthesizes the data from eight cohorts, reporting a greater risk of AKI with albuminuria, both in the presence and absence of diabetes, and with an increased risk in stratified analysis with decreasing baseline kidney function. Major studies reporting proteinuria and AKI association n = 1746 Pre-existing proteinuria, with dipstick, associated with higher risk of AKI requiring dialysis n = 11 200 Prospective cohort, reported a graded association of albuminuria with subsequent AKI n = 920 985 Cohort study using provincial database, reported graded association according to dipstick proteinuria or albuminuria and subsequent AKI n = 1052 Pre-existing proteinuria, with dipstick, associated with higher risk of AKI in a cohort of patients undergoing cardiac surgery n = 1198 Prospective cohort, showing albuminuria hours after cardiac surgery precedes and is associated with development of AKI n = 294 children No association of preoperative albuminuria with post-cardiac surgery AKI in this paediatric population n = 8 cohorts, 1 285 045 participants Higher albuminuria associated with greater risk of AKI, in the presence or absence of diabetes, and with a greater risk in stratified analysis with decreasing GFR n = 153 767 Preoperative proteinuria associated with AKI even after non-cardiac surgery n = 5347 Pre-admission proteinuria is associated with non-recovery after AKI requiring dialysis n = 1538 Albuminuria measured post-AKI is associated with faster progression of kidney disease n = 5168 Graded association of proteinuria with post-operative AKI in patients with non-cardiac surgery n = 1746 Pre-existing proteinuria, with dipstick, associated with higher risk of AKI requiring dialysis n = 11 200 Prospective cohort, reported a graded association of albuminuria with subsequent AKI n = 920 985 Cohort study using provincial database, reported graded association according to dipstick proteinuria or albuminuria and subsequent AKI n = 1052 Pre-existing proteinuria, with dipstick, associated with higher risk of AKI in a cohort of patients undergoing cardiac surgery n = 1198 Prospective cohort, showing albuminuria hours after cardiac surgery precedes and is associated with development of AKI n = 294 children No association of preoperative albuminuria with post-cardiac surgery AKI in this paediatric population n = 8 cohorts, 1 285 045 participants Higher albuminuria associated with greater risk of AKI, in the presence or absence of diabetes, and with a greater risk in stratified analysis with decreasing GFR n = 153 767 Preoperative proteinuria associated with AKI even after non-cardiac surgery n = 5347 Pre-admission proteinuria is associated with non-recovery after AKI requiring dialysis n = 1538 Albuminuria measured post-AKI is associated with faster progression of kidney disease n = 5168 Graded association of proteinuria with post-operative AKI in patients with non-cardiac surgery ASSESS-AKI: Assessment, Serial Evaluation, and Subsequent Sequelae in Acute Kidney Injury. Major studies reporting proteinuria and AKI association n = 1746 Pre-existing proteinuria, with dipstick, associated with higher risk of AKI requiring dialysis n = 11 200 Prospective cohort, reported a graded association of albuminuria with subsequent AKI n = 920 985 Cohort study using provincial database, reported graded association according to dipstick proteinuria or albuminuria and subsequent AKI n = 1052 Pre-existing proteinuria, with dipstick, associated with higher risk of AKI in a cohort of patients undergoing cardiac surgery n = 1198 Prospective cohort, showing albuminuria hours after cardiac surgery precedes and is associated with development of AKI n = 294 children No association of preoperative albuminuria with post-cardiac surgery AKI in this paediatric population n = 8 cohorts, 1 285 045 participants Higher albuminuria associated with greater risk of AKI, in the presence or absence of diabetes, and with a greater risk in stratified analysis with decreasing GFR n = 153 767 Preoperative proteinuria associated with AKI even after non-cardiac surgery n = 5347 Pre-admission proteinuria is associated with non-recovery after AKI requiring dialysis n = 1538 Albuminuria measured post-AKI is associated with faster progression of kidney disease n = 5168 Graded association of proteinuria with post-operative AKI in patients with non-cardiac surgery n = 1746 Pre-existing proteinuria, with dipstick, associated with higher risk of AKI requiring dialysis n = 11 200 Prospective cohort, reported a graded association of albuminuria with subsequent AKI n = 920 985 Cohort study using provincial database, reported graded association according to dipstick proteinuria or albuminuria and subsequent AKI n = 1052 Pre-existing proteinuria, with dipstick, associated with higher risk of AKI in a cohort of patients undergoing cardiac surgery n = 1198 Prospective cohort, showing albuminuria hours after cardiac surgery precedes and is associated with development of AKI n = 294 children No association of preoperative albuminuria with post-cardiac surgery AKI in this paediatric population n = 8 cohorts, 1 285 045 participants Higher albuminuria associated with greater risk of AKI, in the presence or absence of diabetes, and with a greater risk in stratified analysis with decreasing GFR n = 153 767 Preoperative proteinuria associated with AKI even after non-cardiac surgery n = 5347 Pre-admission proteinuria is associated with non-recovery after AKI requiring dialysis n = 1538 Albuminuria measured post-AKI is associated with faster progression of kidney disease n = 5168 Graded association of proteinuria with post-operative AKI in patients with non-cardiac surgery ASSESS-AKI: Assessment, Serial Evaluation, and Subsequent Sequelae in Acute Kidney Injury. In this issue of Nephrology Dialysis Transplantation, Nishimoto et al. [14] report on the NARA acute kidney injury (NARA-AKI) cohort study, a single-centre study of 5168 individuals who underwent non-cardiac surgery. In keeping with prior literature, the overall risk of AKI was lower than for cardiac surgery (6%), but again, a graded association between proteinuria and AKI was reported. Preoperative proteinuria was measured using a dipstick (most recent value within 1 month prior to surgery) and was associated with an unadjusted odds ratio (OR) of 3.44 with AKI, which was attenuated somewhat to 1.92 after adjustment for factors including kidney function and comorbid conditions and to 1.80 after addition of medications [including renin–angiotensin system (RAS) blockers] and intra-operative haemodynamics to the model. Similar to other studies, a gradient was observed with adjusted ORs ranging from 1.14 (trace proteinuria), to 1.24 (1+ proteinuria), to 2.75 (2+ proteinuria) to 3.95 (3+ proteinuria). This is consistent with results previously reported in the non-cardiac surgery population. In a database study from the US Department of Veterans Affairs, of 153 767 surgeries (which included orthopaedic, general and vascular surgery procedures), preoperative dipstick proteinuria was found 44% of the time and AKI occurred 11% of the time [11]. A similar dose-dependent association of proteinuria and AKI, independent of preoperative estimated glomerular filtration rate (GFR), was reported. The authors of the present study suggest that these results should prompt us to assess preoperatively for proteinuria and monitor these patients more closely for development of AKI post-operatively. Reasonable as these suggestions seem, further exploration and understanding of these findings should precede pre-emptive planning and practice change. What could explain the persistent reports of the association of proteinuria and AKI? Could proteinuria truly be on the causal pathway in the development of AKI? The consistency of the association in the literature, coupled with the repeated reports of a biological gradient, are reassuring that the association is robust; however, it does not prove causation. Despite the sophisticated modelling from this body of literature, underlying residual confounding remains a distinct possibility. As an example, no association between albuminuria and post-cardiac surgery was found in a study of 294 children who did not have the usual suspects such as diabetes and vascular disease that travel together with albuminuria in adults [12]. Additionally, in the NARA-AKI study, the individuals with proteinuria were different not just by being older with a lower kidney function, but were different in almost every variable examined, including atrial fibrillation and peripheral arterial disease [14]. Interestingly enough, they also had a greater drop in intra-operative blood pressure and developed a greater net intra-operative fluid balance gain. These are difficult to explain as a consequence of proteinuria and highlight the potential for substantial residual confounding due to unmeasured variables. Sicker people have proteinuria, and sicker people do develop AKI, with proteinuria being a possible marker of the extent of cardiovascular disease. Another related issue is that proteinuria is, by itself, a sign of intrinsic kidney disease, even if it does not always manifest as decreased kidney function. In some cases (e.g. in the patients with 3+ proteinuria), the presence of proteinuria may just reflect underlying undiagnosed glomerular disease. Proteinuria in this context, and to some extent across the spectrum of proteinuria, may, as suggested by the NARA-AKI study, reflect susceptibility to AKI when an additional insult occurs. Nonetheless, it should also be considered that increased urinary protein excretion is a frequent manifestation of many different causes of AKI, including acute tubular necrosis (ATN). In ATN, proteinuria is due to excretion of cellular debris (i.e. tubular proteinuria), as well as an element of impaired reabsorption of normally filtered proteins that can manifest as albuminuria that is detectable on dipstick testing. In a prospective cohort study of 1198 cardiac surgery patients, in patients who developed post-operative AKI, albuminuria was found to increase in the hours after surgery and this presaged the increase in serum creatinine [9]. Mechanistic studies that report proteinuria in the causal pathway of AKI would go a long way towards clarifying the relationship between proteinuria and AKI. As our knowledge about the pathophysiology of AKI expands, we might come to understand this relation better. Another issue common to most of these studies relates to how patients came to have their urine tested in the first place. Although we would desire measurement of albuminuria to be universal, only a selected proportion of patients have this performed, which may be an informed decision on the part of the ordering physician. Hence it may be that the presence of a proteinuria measurement in the patient record represents someone at higher risk, regardless of the actual value of the test. This ascertainment bias can make the strength of the relation between proteinuria and AKI appear stronger than it actually is. In terms of prognostic stratification, awareness of the increased risk of AKI would be especially useful if we had any arrows in our therapeutic quiver to reduce this higher risk of AKI in patients with proteinuria. Unfortunately we have nothing. A variety of prophylactic agents have fallen by the wayside after being unable to demonstrate efficacy, notably in the setting of contrast administration, where, similar to elective surgery, the timing is amenable to testing interventions [15, 16]. The standard strategy of ‘withholding nephrotoxic’ agents also has a poor evidence base, even in the setting of surgery and sick days [17, 18]. Moreover, drugs that are potentially perceived as nephrotoxins, such as RAS blockade, may have a nephroprotective role. Withholding RAS blockade will increase proteinuria, which may be counterproductive if proteinuria itself is on the AKI causation pathway. RAS blockers, by reducing efferent arteriolar vasoconstriction, increases medullary blood flow and may potentially restore adequate oxygenation to the tubular cells in times of kidney stress, lowering the risk of AKI [19]. Other agents like sodium–glucose cotransporter-2 inhibitors (SGLT2is) are associated with a lower risk of AKI [20]. This lower risk of AKI has been reported not just in clinical trials, but also in real-world evidence from large observational studies [21]. The somewhat paradoxical observation of the consistently reported reduction in AKI occurs despite the initial decrease in GFR after initiation of these drugs and their natriuretic effects, which should theoretically be associated with a higher risk of AKI. SGLT2is also decrease albuminuria, raising the possibility that this could be a mediator in the SGLT2i–AKI relation and strengthening the putative role of albuminuria on the pathway towards kidney injury [20]. Lastly, of greater interest are the longer-term outcomes with proteinuria in patients who develop AKI. In a study of 5347 adults with AKI requiring dialysis, 3601 did not recover, with slightly less than half of them dying while on dialysis or within 28 days of stopping dialysis [8]. Pre-admission proteinuria (i.e. predating admission for AKI) was associated with non-recovery of AKI, even after adjustment for comorbid conditions, including baseline kidney function. Even more interesting aspects relate to the assessment of albuminuria after AKI. In a retrospective study, 90 614 patients with AKI were matched to the same number of patients without AKI. Over the next 12 months of follow-up, the odds of having ≥1+ proteinuria on dipstick was significantly higher (OR range 1.20–1.39), regardless of underlying diabetes or RAS blocker use [10]. Another prospective matched cohort study enrolled 1538 participants, half of whom had AKI requiring hospitalization, but who either did not require dialysis or recovered from needing dialysis [6]. In this cohort, urine albuminuria in the 3 months following the sentinel AKI hospitalization event was associated with progression of kidney disease, measured as a 50% decrease in kidney function or development of end-stage kidney disease. Whether this post-AKI albuminuria represents a sequelae of kidney injury during the AKI process or perhaps a measure of underlying susceptibility to AKI remains to be seen. Overall, whether proteinuria is on the causal pathway or not, the recurrent and robust association has not been adjusted away and seems here to stay. Specifically, proteinuria is associated with higher rates of AKI, more severe AKI, lower rates of recovery after AKI and greater development of subsequent CKD (see graphical abstract). Just as albuminuria has been added to CKD staging, is it time to add it to AKI staging? Widespread measurement of albuminuria should definitely be endorsed to allow a greater understanding and awareness. That albuminuria assessment might help in prognostic assessment in the perioperative AKI setting is also undoubtedly true. Whether it might serve as a potential therapeutic target remains to be seen. None declared. (See related article by Nishimoto et al. Pre-operative proteinuria and post-operative acute kidney injury in noncardiac surgery: the NARA-Acute Kidney Injury cohort study. Nephrol Dial Transplant 2020; 35: 2111--2116) No specific funding was received for this manuscript. SH and EC receive research salary support from the Department of Medicine, University of Ottawa, Ottawa, ON, Canada.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.