Biomarkers of acute kidney injury and associations with short- and long-term outcomes

Cardiac surgery

Cardiac surgery is one of the most popular settings for the evaluation of biomarkers of AKI, partially because there is a clear point of injury. AKI after cardiac surgery has been commonly ascribed to an inflammatory cascade, although thus far interventions to quell the inflammatory cascade have been unsuccessful¹⁰. Biomarkers of AKI could help identify patients at higher risk of adverse outcomes.

Short-term outcomes. One of the most carefully studied cohorts for biomarkers, Translational Research Investigating Biomarker Endpoints of AKI (TRIBE-AKI), featured over 1200 patients who underwent cardiac surgery. Levels of post-operative tubular injury biomarkers, such as urinary interleukin (IL)-18, neutrophil gelatinase-associated lipocalin (NGAL), kidney injury molecule-1 (KIM-1), and liver fatty acid binding protein (L-FABP), were associated with increased risk of in-hospital death, increased length of stay, and need for RRT (Table 1)^30,31. Patients with elevated urinary cystatin C, another tubular injury biomarker, additionally had increased risk of requiring RRT³².

Another cohort of cardiac surgery patients, Southern Acute Kidney Injury Network (SAKInet), were evaluated to determine whether combinations of biomarkers could predict if patients with stage I AKI (AKIN) after cardiac surgery progressed to stage III AKI or death. Urinary IL-18 combined with percentage change in serum creatinine or urinary KIM-1 had the best discriminative ability to identify patients at high risk for progressing to higher stages of AKI or death within 30 days³³. The SAKInet cohort study additionally evaluated urinary angiotensinogen and renin, biomarkers of intra-renal renin-angiotensin system activity, and found that they were associated with severe AKI or death within 30 days³⁴.

Changes in urinary biomarkers have been associated with improved patient outcomes. One study examined biomarkers that induce cell cycle arrest in cardiac surgery by examining the product of tissue inhibitor metalloproteinase 2 (TIMP-2) and insulin-like growth factor binding protein 7 (IGFBP-7). Differences in the decline of these biomarkers within 24 hours after surgery discriminated between patients who recovered and those who failed to recover from AKI³⁵.

Long-term outcomes. In the TRIBE-AKI study, patients with AKI as defined by serum creatinine and elevated peak post-operative tubular injury markers, including urinary IL-18, NGAL, KIM-1, and L-FABP, had an increased risk of long-term mortality²⁸. Patients without AKI as defined by serum creatinine, elevated urinary IL-18, and KIM-1 also had an increase in long-term mortality of about 1.5-fold. These patients may have experienced subclinical AKI that was not detected by serum creatinine, leading to increased mortality, or it may be that urinary IL-18 or KIM-1 is capturing a systematic inflammatory process which increases the risk of death independently of AKI.

Biomarkers that probe inflammatory pathways involved in AKI have additionally been investigated in TRIBE-AKI. Elevated first post-operative serum IL-10, which is known to attenuate the inflammatory cascade, was found to decrease the risk of long-term mortality³⁶. Interestingly, this effect was isolated to patients with elevated urinary markers of tubular injury, which suggests that IL-10 protected those patients who were experiencing proximal tubular cell damage.

Critical illness

AKI is a common complication in critical illness and is consistently associated with high mortality, although this is a challenging setting in which to study biomarkers of AKI¹. There may be multiple renal insults, and the timing of renal insults is not always clear. Regardless, biomarker data have identified subphenotypes for complex intensive care unit (ICU) syndromes, such as acute respiratory distress syndrome (ARDS), which benefit from different treatment strategies³⁷. The incorporation of prognostic information from AKI biomarkers could eventually lead to similar benefit. Multiple urinary biomarkers have been studied in critically ill patients and are associated with adverse outcomes, although the data are strongest for short-term outcomes (Table 2).

Short-term outcomes. One ICU cohort study, Validation of Biomarkers for Acute Lung Injury Diagnosis (VALID), comprising critically ill patients in both medical and surgical ICU settings, collected urinary biomarker specimens on admission and 48 hours after admission in over 400 patients. In VALID, admission levels of urinary IL-18 were independently associated with the composite outcome of 28-day mortality or dialysis even after adjustment for APACHE-II score, serum creatinine, and sepsis criteria³⁸. Despite the relationship between urinary IL-18 and the composite outcome, other tubular injury markers, such as urinary NGAL and L-FABP, were associated only with the initiation of dialysis at 28 days and were not associated with the composite outcome^39,40. This persisted despite the authors excluding patients with unidentified CKD or AKI upon enrollment into the study³⁹. Similar results for urinary IL-18 and NGAL have been found in other cohorts of critically ill patients. A post hoc case-control study of 132 patients in the ARDS Network additionally found that increasing levels of urinary IL-18 on admission were associated with increased 1-month mortality⁴¹. Another prospective cohort of 339 patients in a mixed ICU setting showed that admission urinary L-FABP, IL-18, and NGAL could discriminate between survivors and non-survivors, and a combination of urinary NGAL and L-FABP was best able to discriminate between survivors and non-survivors⁴². Additionally, increasing levels of urinary cystatin C, a biomarker of glomerular filtration, in a group of 444 ICU patients were associated with increased risk of 1-month mortality⁴³.

As in cardiac surgery, a decline in tubular injury biomarkers is also associated with improved patient outcomes. A post hoc analysis of 76 patients from the Acute Renal Failure Trial Network (ATN) trial, which evaluated the effect of dialysis dose on critically ill patients, investigated whether urinary biomarkers were associated with recovery from AKI and found that decreasing levels of urinary NGAL and hepatocyte growth factor (HGF) over a 2-week period predicted an increased chance of recovery of renal function 60 days after ICU admission⁴⁴.

Long-term outcomes. While biomarkers of tubular injury are studied most carefully, another cohort comprising 744 ICU patients (Sapphire) found that urinary biomarkers of cell cycle arrest (TIMP-2*IGFBP-7) were associated with a twofold increased risk of RRT and 9-month mortality, although this relationship was found only in patients with AKI as defined by serum creatinine⁴⁵.

Cardiorenal syndrome

Heart failure patients commonly suffer from AKI, which is strongly associated with increased mortality⁴⁶. Cardiorenal syndrome, or the co-existence of AKI and heart failure, is often ascribed to hemodynamic perturbations or venous congestion. This is likely an over-simplification, as patients with cardiorenal syndrome respond differently to standard heart failure therapies, such as diuresis and inotropic support. Biomarker research has attempted to untangle the numerous phenotypes that exist within cardiorenal syndrome⁴⁷.

Long-term outcomes. There is extensive literature surrounding long-term prognosis and biomarkers of tubular injury in heart failure patients. Tubular injury biomarkers, including urinary KIM-1, NGAL, and N-acetyl-β-D-glucosaminidase (NAG), were associated with long-term mortality and recurrent heart failure hospitalization even after accounting for glomerular filtration rate (GFR) in multiple case-control studies and cohorts (Table 3)^48–50. KIM-1 and NAG were significantly correlated with New York Heart Association class, suggesting that tubular injury may contribute to the pathophysiology of cardiorenal syndrome⁵⁰. While urinary biomarkers of tubular injury have been consistently associated with mortality and other adverse outcomes, plasma NGAL has not demonstrated a consistent relationship, particularly after adjustment for renal function^51–53. This has also been found in cardiac surgery and is likely because plasma NGAL is renally excreted, effectively rendering it another marker of GFR⁵⁴.

Kidney transplantation

Novel urinary biomarkers have also been investigated for prognostic utility in the kidney transplant setting and can have unique applications in the transplant field. Given the persistent shortage of organs and the increasing number of patients on the transplant list, biomarkers of AKI can potentially be used to decide if a deceased donor organ is acceptable for transplant as well as improve prediction for adverse recipient outcomes, such as delayed graft function (DGF) and graft failure (Table 4).

Short-term outcomes. In the perioperative period, the presence of urinary IL-18 and NGAL immediately after transplant was associated with an increased risk of DGF^55,56. While the relationship between recipient kidney injury biomarkers and outcomes is relatively clear, the relationship between donor kidney injury biomarkers and recipient outcomes is more complex. One large multi-center cohort found that elevated donor urinary NGAL had a modest association with DGF in the recipient, although other tubular injury biomarkers were not significantly associated after multivariable adjustment⁵⁷.

Long-term outcomes. Single-center studies of transplant recipients found that elevated urinary tubular injury biomarkers, such as urinary IL-18, NGAL, NAG, and KIM-1, are associated with an increased risk of graft failure or death (Table 4)^58,59. However, urinary donor biomarkers, including urinary IL-18, NGAL, and L-FABP, modified the relationship between DGF and 6-month recipient GFR. For recipients with DGF, higher levels of tubular injury biomarkers were associated with higher 6-month GFR, while for recipients who did not experience DGF, higher urinary donor biomarkers were associated with lower 6-month GFR⁵⁷.

Heterogeneous populations: hospitalized patients and the emergency department

While the most extensive evidence for the association between biomarkers of AKI and prognosis is in specific patient populations, biomarkers of AKI also have prognostic utility for short-term outcomes in heterogeneous populations. Multiple single-center cohorts of patients requiring nephrology consultation found that biomarkers of tubular injury were associated with an increased risk of in-hospital RRT or death (Table 5)^60–63. Alternatively, elevated plasma NGAL, a marker of glomerular filtration, in patients hospitalized with pneumonia and severe AKI meant they were twice as likely to fail to recover renal function by discharge in adjusted analysis⁶⁴. In another cohort of hospitalized patients requiring nephrology consultation owing to presumed acute tubular necrosis, elevated urinary epidermal growth factor levels allowed the authors to distinguish between patients who recovered renal function within 1 week of renal consult and those who did not. Additionally, elevated inflammatory cytokines identified patients who failed to recover renal function within 1 week of renal consult⁶⁵. Urinary NGAL and KIM-1 measured at the time of emergency room visit were both independently associated with in-hospital RRT initiation or death after accounting for clinical parameters⁶⁶.

Next-generation biomarkers

The bioinformatics revolution has improved discovery techniques for the identification of next-generation biomarkers. Biomarker research has primarily focused on biomarkers that are considered part of the proteome. Recent studies have begun examining other layers of cellular data, including microRNA and mitochondrial DNA^67–70. While these biomarkers are in the early stages of investigation, they hold great promise, as they provide opportunities to probe other components of cellular pathways.

Limitations for clinical application and future directions

Several barriers remain for the clinical application of biomarkers despite the extensive evidence that they are associated with prognosis. Biomarkers of AKI often provide modest improvement in discriminative ability compared to traditional clinical models and have different trajectories and time scales depending upon the patient population and biomarker being studied. This will pose a challenge for busy clinicians attempting to incorporate a cornucopia of biomarkers into clinical practice. Moreover, data are lacking to support whether biomarkers are associated with organ-specific long-term outcomes, such as CKD or cardiovascular disease. Technical challenges also exist, as developing, standardizing, and validating biomarker assays is a laborious process⁷¹. Most importantly, researchers have yet to identify therapeutic changes to institute once patients are identified as being at high risk for adverse outcomes based on biomarkers of AKI.

Given the heterogeneity and dynamic nature of AKI, one biomarker will not accurately predict all adverse outcomes across several settings, nor will it accurately identify the various stages in the natural history of AKI. One important task is to definitively identify the natural stages of AKI as demarcated by biomarkers: injury, maintenance, and repair. This could help create a more balanced picture of prognostic outcomes for patients. For example, patients may have different outcomes depending on the magnitude of “injury” biomarkers balanced by the level of “repair” biomarkers.

While challenges remain, biomarkers can offer several potential benefits to the diagnosis, management, and treatment of AKI in the foreseeable future. Perhaps the most immediate application is to identify patients who are experiencing renal injury from nephrotoxic agents, prolonged surgeries, cardiorenal syndrome, or hepatorenal syndrome. Patients with elevated injury or fibrosis markers could be considered for alternative therapy instead of waiting for these patients to develop AKI or CKD as defined by serum creatinine. However, for biomarkers to become the “standard of care” for the management of these patients, prospective trials should be conducted to identify cut-points for biomarker levels which will identify patients at a sufficiently high risk of renal complications to warrant providing a therapeutic intervention or adjusting surgical management.

Biomarkers of AKI can have two potential applications in clinical trials of putative drug treatments for AKI. First, they will allow us to identify patients who have ongoing injury so that therapies may be provided while the injury process is occurring. Second, the greatest potential for biomarkers is if they can identify patients at high risk of adverse outcomes. In such a scenario, treatment can then be personalized for high-risk subgroups of patients. Trials in AKI have been hindered by low event rates and are likely enrolling patients with multiple phenotypes of AKI. Given the wealth of information showing that biomarkers are related to important patient outcomes, such as RRT and death, future clinical trials for AKI should enroll patients with elevated biomarkers of AKI in order to increase the likelihood that patients would actually benefit from the putative treatment. The TOPCAT trial, which examined the benefit of spironolactone in heart failure patients with preserved ejection fraction, found that treatment benefit was isolated to patients with elevated natriuretic peptide levels⁷². Future AKI trials should employ similar enrollment strategies, as treatment benefit may be isolated to particular AKI phenotypes. The Assessment, Serial Evaluation and Subsequent Sequelae of Acute Kidney Injury study is ongoing to determine if patients with AKI experience increased risk of long-term complications, such as cardiovascular events and CKD²⁷. This study should help ascertain if it is worthwhile enrolling patients with AKI as defined by biomarkers in lieu of serum creatinine in future clinical trials.