Novel Biomarkers of AKI: the challenges of progress ‘Amid the noise and the haste’
Sean M. Bagshaw, Michael Zappitelli and Lakhmir S. Chawla
Correspondence and offprint requests to: Sean M Bagshaw; E-mail:bagshaw@ualberta.ca
Abstract
The clinical integration of novel biomarkers specific for kidney damage have brought the promise of a new era in our understanding of and care for those patients susceptible to or suffering from acute kidney injury (AKI) and has consistently been viewed as a top research priority. The expectations are clearly high; however, as with many promises, there are often accompanying challenges and a degree of pessimism. In this issue of Nephrology Dialysis Transplantation, Van Massenhove et al. offer their ‘Devil's advocacy’ view in a narrative review focused on the state of novel biomarkers for the diagnosis of AKI. While AKI biomarkers would appear to clearly have value, in particular for informing on the pathobiology of AKI, the question of how to optimally utilize them remains unresolved. Their performance is influenced by patient case-mix, comorbid illness, inciting kidney injury event, timing of measurement, the specific biomarker being investigated and the selected thresholds for diagnosis, not to mention factors related to study design, methodology and how to best translate to the bedside. The challenge as the field moves forward is to fully and appropriately utilize and interpret information from AKI biomarker studies in order to understand and evaluate how to optimally utilize these novel biomarkers (or panel of biomarkers) in the susceptible patient across a spectrum of clinical settings to improve and better inform our clinical decision-making.
Previous SectionNext SectionAcute kidney injury (AKI) is defined by an ‘abrupt and sustained’ decrease in kidney function, represents a complex syndrome that occurs across a broad spectrum of clinical settings and can manifest from mild increases in serum creatinine (sCr) to overt oligoanuric kidney failure, prompting the initiation of renal replacement therapy (RRT) [1]. AKI remains a common and challenging clinical problem, particularly for critically ill patients, with a dearth of interventions to specifically modify or attenuate injury or recovery once established [2]. Moreover, it is frequently iatrogenic, necessitates increased intensity of care, consistently portends a negative impact on survival and recovery of kidney function and burdens our health systems with added expenditures [3].
Historically, inferences on the epidemiologic burden of AKI have been imprecise due to the lack of agreement on an acceptable clinical definition [4]. However, consensus criteria, based largely on the detection of changes to sCr, urine output and/or the provision of RRT, have been validated and are now widely viewed as an important advance in the field [5, 6]. Yet, we recognize that these criteria, emphasizing sCr as the driving marker for AKI, are clearly inadequate, such that our current paradigm for diagnosing AKI has not essentially changed in several decades and remains susceptible to not only delay in diagnosis but also missed episodes of important declines to glomerular filtration.
The discovery, characterization and validation of a number of novel biomarkers specific for kidney damage have brought the promise of a new era in our understanding of and care for those susceptible to or suffering from AKI. The American Society of Nephrology Renal Research Report assigned the discovery and translational validation of novel AKI biomarkers as having the highest research priority and this is largely consistent with the National Institutes of Health roadmap for Medical Research (New Pathways to Discovery) [7].
What exactly is the promise of novel biomarkers over conventional metrics of AKI?
i.Improved diagnostics: biomarkers will enable a ‘proactive’ and timely diagnosis of AKI through better reflection of real-time ‘damage’ to kidney tissue, provide considerable ‘lead-time’ in the detection of impeding AKI and will enable discrimination of the underlying etiology and severity of AKI (i.e. ‘pre-renal’ from more established AKI from CKD).
ii.Improved prognostics: biomarkers will enable incremental risk identification for worsening AKI, overt kidney failure and poor clinical outcome.
iii.Improved decision-making: biomarkers will better enable risk stratification to ‘renal protective’ strategies, timing of initiation and/or discontinuation of interventions and aid in assigning priorities for triage disposition.
iv.Improved innovation: biomarkers will better enable risk stratification, triage and enrollment into future clinical studies aimed at advancing the field of acute care nephrology.
The expectations are clearly high with the potential for AKI biomarkers to move the field forward, as we draw from an analogy of how the discipline of cardiology has introduced novel diagnostic biomarkers to advance in the quality of care for acute coronary syndrome (ACS). As with many promises, however, there are often accompanying challenges and a degree of pessimism. In this issue of Nephrology Dialysis Transplantation, Van Massenhove et al. offer a ‘Devil's advocacy’ view of the current state of novel biomarkers for the diagnosis of AKI [8]. The authors have performed a narrative review. In this review, they have cataloged the range of diagnostic performance of an array of biomarkers [predominantly focused on eight biomarkers: neutrophil gelatinase-associated lipocalin (NGAL); kidney injury molecule-1 (KIM-1); interleukin-6 (IL-6); interleukin-18 (IL-18); cystatin c (CystC); N-acetyl-glucosaminidase (NAG); liver fatty acid binding protein (L-FABP); glutathione transferases (GSTs)] measured in either the blood or urine across five clinical settings (i.e. pediatrics, emergency department, intensive care, cardiac surgery and contrast-induced nephropathy) for the diagnosis of AKI or prediction of the need for RRT. Their search of the literature yielded 87 publications, representing 74 unique clinical studies; however, important publications were unfortunately omitted [9, 10]. Data were extracted and summarized in tables, followed by a largely narrative account consistent with a commentary. The authors did not summarize and/or stratify data across additional clinically important subgroups (i.e. study design, study quality indicators, definition used for AKI) that would have enabled aggregate and/or pooled data analysis. The author's view on the diagnostic performance of AKI biomarkers could be summarized as rather pessimistic and glib. Regrettably, we believe that the method in which the results of their review are presented and interpreted fail to adequately depict the current state of AKI biomarkers.
The authors found that, despite consensus recommendations for the diagnosis of AKI (i.e. RIFLE, AKIN or KDIGO classification schemes), the included biomarker studies utilized a wide variety of operational definitions to ascertain the diagnosis of AKI, not only in terms of biochemical thresholds (i.e. absolute or relative changes in SCr and urine output), but also the temporal duration over which these changes were to occur (i.e. 24, 48, 72, 120 h, etc.). Few studies evaluated the predictive performance for worsening AKI or need for RRT. This finding is clearly problematic and no doubt contributes to undermining the comparability and generalizability of included studies of biomarker performance. The authors, however, omitted addressing the obvious question: why is there still a lack of consistency in the literature on the application of definitions for AKI? These data would imply that the currently proposed consensus definitions for AKI are not so easily applicable across all settings of investigation. For example, the use of urine output criteria, as part of the consensus definition for AKI, may be hindered and problematic in selected settings where patients do not routinely have urinary catheters (i.e. contrast nephropathy versus pediatrics versus emergency department). The extent of how variable utilization of the ‘gold standard’ definition of AKI against which biomarkers are measured unduly impacts their performance in validation studies is still unclear.
Second, numerous aspects of reviewed studies were heterogeneous, which the authors acknowledge on several occasions. This included disparities in study design (i.e. single center, small cohort size), discrepancy in event rates that directly contribute to wide confidence intervals across diagnostic discriminatory estimates (i.e. many studies had event rates <5%; again, contributed by variable AKI definitions used) and clinical heterogeneity in patient cohorts being evaluated across five clinical settings, whereby consistency in the risk profile (i.e. age, comorbidities, diabetic status, chronic kidney disease), susceptibility and inciting events for AKI are difficult to account for and may independently influence biomarker values (i.e. sepsis and CKD) [11]. Perhaps most importantly, the authors have descriptively summarized data from an array of 25 different biomarkers, which clearly have fundamental differences in pathobiology [i.e. filtered biomarker (CysC) versus upregulated biomarker (NGAL, KIM-1) versus preformed released biomarker (GGT, GST)]. Finally, the authors have not specifically addressed whether there is potential for the timing of measurement and temporal trends in biomarker detection to further improve their diagnostic accuracy [9]. Hence, the variation in diagnostic performance of biomarkers for AKI from such a descriptive review of heterogeneous studies is not surprising, and in fact probably could have been predicted. Unfortunately, as aforementioned, the authors did not perform an evidence synthesis and meta-analysis whereby data were pooled across identifiable subgroups, in particular, for studies within specific settings and with similar diagnostic criteria for AKI (i.e. in 12 of 20 listed pediatric studies, AKI was defined as a 1.5× increase in SCr) so as to evaluate whether there was any improvement in discrimination. In addition, despite their sharp criticism of the heterogeneity across AKI biomarker studies, the authors grouped all pediatric studies into a single table. This perhaps fails to account for the important nuanced clinical settings, whereby these studies were performed. Ultimately, the question remains whether or not the presence of study heterogeneity in the AKI literature is such a dire issue. Should biomarkers not be evaluated across a range of clinical settings? The fact that this variation is present in the literature at minimum conveys the awareness and emphasis of researchers that biomarkers need to be validated in a variety of settings. Accordingly, we believe this is a good thing. However, simply drawing attention to the presence of ‘low AUCs’ in the biomarker literature should be counterbalanced with the fact that the AKI community is appropriately publishing both positive and negative studies. Essentially, the suggested view presented by the authors might be to ‘throw out the baby with the bathwater’ on AKI biomarkers. However, their method of assessment (simply listing the wide ranges of diagnostic performance without accounting for the heterogeneity factors they mention) was likely inadequate to justify their conclusions. Unfortunately, it appears that the authors may have already made their ‘judgment’ on AKI biomarkers before having performed their limited analysis.
Third, the authors suggest currently the most reliable scenario for AKI biomarkers are in pediatric settings where confounding by comorbid conditions is minimized and in settings when there is well-defined single injury to the kidney such as cardiac surgery or contrast exposure. The authors conclude by suggesting that there are limited generalizable conclusions on the diagnostic value of biomarkers in AKI, in particular, for any one biomarker in any one clinical setting. This seems like an obvious conclusion; why would biomarker results in one setting be generalizable to other diagnostic settings? We also respectfully disagree with the author's statement that current use of AKI biomarkers in clinical practice remains ‘cumbersome’. particularly in settings where the timing of injury is unclear. Rather, we support that an area of further investigation should identify when to preferably measure biomarkers across variable settings. For example, AKI is an early event in pediatric critical illness, generally occurring very soon after ICU admission; thus early admission or pre-admission AKI biomarkers should undergo ‘ideally’ further evaluation.
Importantly, we need to focus our attention on the concept that any biomarker ‘alone’, however ideal, will not likely lead to informative decision-making. For example, cardiac-specific troponin-I (cTnI) is an excellent biomarker and currently the gold standard for the diagnosis of ACS; however, its diagnostic performance deteriorates when used outside this context. Multiple studies have shown that when cTnI is measured in heterogeneous cohorts of critically ill patients (one standard to which we currently hold AKI biomarkers), the diagnostic performance for ACS is mediocre to poor [12, 13].
Accordingly, we believe it is essential that clinical context be incorporated into the use of AKI biomarkers to better enable their ‘ideal’ utilization (i.e. right patient, right context) and to maximize the incremental information gained in order to best inform decision-making. Notably, many of the studies included in this review either failed to integrate and/or were statistically incapable of evaluating for ‘incremental discrimination’ for prediction or diagnosis of AKI, which may have been appropriate for early-phase clinical biomarker studies. The integration of biomarkers should enable incremental diagnostic and/or prognostic information beyond that found by clinical evaluation and/or conventional measures of kidney function alone. Clearly, there are instances when the conventional measures of kidney function are sufficient and the incremental values of AKI biomarkers are questionable, such as when clinically and biochemically confirmed AKI are evident at presentation [14–16]. The use of biomarkers in these circumstances may be uninformative. Additional studies have focused on whether AKI biomarkers can provide insights into incremental risk for worsening AKI [17], need for RRT [18] and recovery of kidney function [19, 20]; however, few studies have prospectively evaluated how biomarker integration directly informs decision making. This would establish greater value for AKI biomarkers; however, currently represents one of the key barriers to bedside translation. For example, in the GALLANT study, the measurement of serum NGAL at the time of discharge from hospital of patients admitted with acute decompensate heart failure was incrementally beneficial and superior than sCr for predicting death or readmission within 30 days [21]. In another example, recognizing the challenges in AKI due to the lack of symptoms and early physical findings, the concept of ‘renal angina’ has been proposed, whereby a clinical risk score is integrated to help direct the appropriate use of AKI biomarkers [22].
AKI biomarkers have also enabled the characterization of a subgroup of patients at risk for conventionally defined AKI; however, who only have detectable elevations in biomarkers termed ‘subclinical AKI’ [23]. A recent pooled analysis of cohort studies found that 19.2% of patients had elevations in NGAL but no detectable increase in SCr [24]. These patients were found to have a worse clinical outcome compared with patients with no increase in either NGAL or SCr characterized by higher mortality, greater utilization of RRT and longer duration of hospitalization. There remains uncertainty of whether these findings represent subclinical kidney damage rather than a surrogate for overall illness severity [25, 26]. However, the novel cohort identified by AKI biomarkers alone appears to be at increased risk for less favorable outcomes and clearly warrants further characterization. Arguably, this has prompted innovative clinical trial designs where biomarkers are utilized in decision support algorithms (ClinicalTrials.gov Identifier: NCT 01416298).
In summary, AKI biomarkers would appear to have value, in particular for informing on the pathobiology of AKI; however, their performance is influenced by patient characteristics, comorbid illness, inciting kidney injury event, timing of measurement and selected thresholds for diagnosis. The challenge as we move forward is to fully and appropriately utilize and interpret information from previously performed studies in order to understand and evaluate how best to utilize these novel biomarkers (or panel of biomarkers) in the susceptible patient across a spectrum of clinical settings to improve and better inform and improve our clinical decision-making. The simple determination of high probability of development of AKI can certainly provide decision support by adopting a strategy to mitigate further harm.
copyright©金宝搏网站登录技巧 版权所有,未经许可不得复制、转载或镜像
京ICP证120392号 京公网安备110105007198 京ICP备10215607号-1 (京)网药械信息备字(2022)第00160号