Acute Kidney Injury - Basic Research and Clinical Practice. Группа авторов

a range of conditions. For example, urinary and plasma NGAL predicting AKI after cardiopulmonary bypass has been studied in over 1,000 adults and over 300 children. In children, urinary and plasma NGAL levels peak within 6 h, whereas SCr continues to rise for up to 48 h post insult. Higher levels of urinary NGAL are associated with both a fivefold increase in the odds of developing AKI as well as improved risk prediction for those who went on to develop severe AKI. No association was found for urinary NGAL with mild AKI (stage 1). Plasma NGAL had a weak association with mild AKI, and no association with severe AKI. The adult study yielded similar results; however, post-operative plasma NGAL levels were also shown to have an association with severe AKI. In adults, the area under the receiver operator curve (AUROC) for the detection of AKI was 0.67 for urinary NGAL, and 0.7 for plasma NGAL. In children, the AUROC was 0.71 and 0.56 respectively [35, 36]. A value of 0.72 has been derived for the prediction of AKI by urinary NGAL from meta-analysis. With regard to kinetics, a study of 196 children undergoing cardiopulmonary bypass revealed that urinary NGAL levels rose markedly within 2 h of surgery with the 2-h value correlating with clinical outcomes such as length of hospital stay and need for haemofiltration [37]. Of note, in non-cardiac surgical patients, NGAL levels were elevated post operatively only in patients who had sustained rises in SCr, or in those with irreversible injury to the kidney [37]. Although the performance of predicting AKI using NGAL is somewhat variable, it may be that the application of NGAL as part of a panel of biomarkers may aid in discrimination and under such conditions, may prove to be a reasonable and a promising tool for AKI risk prediction and diagnosis.

      Interleukin 18

Interleukin 18 (IL-18) is a pro-inflammatory cytokine. Similar to NGAL, it is produced in the renal tubule cells and is detectable in urine. As observed with NGAL, several studies have demonstrated that levels of urinary IL-18 are raised at 4–6 h post CPB, significantly earlier than SCr and that IL-18 may have potential in the prediction of severity of AKI [35, 36]. However, despite early promise, several meta-analyses have derived AUCs ranging from 0.66 to 0.77 [38, 39]. In keeping with NGAL, raised levels are also known to correlate with other co-morbidities including cardiac and lung injury, and urinary tract infection [40]. This is unsurprising, given that IL-18 is a non-specific marker of infection and as such may be elevated for reasons other than AKI.

      Urinary KIM-1

KIM-1 is a type 1 transmembrane glycoprotein, which may be measured at the point of care. KIM-1 expression is upregulated in response to ischaemic injury to the kidneys and may have a role in the remodelling of kidneys through signalling of apoptotic renal tubular cells and their subsequent phagocytosis [41]. While KIM-1 is a marker of renal injury, it also exhibits the ability to predict AKI. Interestingly, in a study on 123 patients, KIM-1 was found to be elevated pre-operatively in patients who subsequently developed AKI and elevated urinary KIM-1 has also been shown to be associated with a higher risk of mortality [42, 43]. Meta-analysis of 11 studies (4 of which followed patients undergoing CPB) estimated an AUROC of 0.86 for AKI prediction using KIM-1 values 2–6 h post insult. In keeping with NGAL, KIM-1 appears to perform better in infants and children rather than in adults, as data derived from an adult surgical cohort demonstrates where the derived AUROC was 0.72 [38, 44].

      Liver Type Fatty Acid Binding Protein

      Liver type fatty acid binding protein (L-FABP) facilitates long chain fatty acid transport and the reduction of oxidative stress. Believed to have a reno-protective role renal ischaemia reduces the reabsorption of L-FABP in the proximal tubule as evidenced by urinary L-FABP levels being strongly correlated with renal cold-ischaemic time in renal transplant surgery [35]. Similar AUROC values are quoted for L-FABP of approximately 0.72 for AKI prediction [38].

      IGFBP-7 and TIMP-2

IGFBP-7 and TIMP-2 are proteins that are implicated in G1 cell cycle arrest as a response to renal insult and unlike other candidate molecules do not exhibit non-specific elevations [45, 46]. In postoperative patients a pre-planned post hoc analysis from both trials including cardiothoracic and emergency surgery demonstrated an AUROC of 0.84 in terms of predicting severe AKI within a 12-h window. Subsequent studies in cardiothoracic patients demonstrated interesting kinetics of these biomarkers of renal stress. Levels of TIMP-2/IGFBP-7 were significantly increased on the first postoperative day (AUROC 0.7), but on the day of surgery, the values were significantly lower [47].

A subsequent study on 107 major general surgical, non-cardiac patients (i.e. trauma, hepatobiliary and vascular) performed postoperatively on the day of surgery delivered an AUROC of 0.85 with high TIMP-2/IGFBP-7 levels being associated with an increased risk of requiring renal replacement therapy and overall mortality [48].

      Subclinical AKI: Dysfunction or Injury?

Clearly there is now a significant body of literature that has investigated AKI biomarkers in a variety of clinical settings and in diverse patient groups. However, there remains an almost slavish addiction to highlighting the performance of AKI biomarkers to predict AKI defined by the classical parameters. Although this is clearly of interest, one must also question what message these biomarkers are telling us if any. Figure 2 highlights some theoretical scenarios where a biomarker associated with AKI may inform us above and beyond that of the SCr. For example, in a previously healthy individual with “normal” renal function, a significant tubular injury may occur undetected if one relies solely on the SCr as a marker of AKI, whereas employing urine output criteria in isolation may result in misclassification of AKI. Under both conditions, a biomarker of AKI may inform beyond that of the SCr and urine output particularly where renal dysfunction rather than true damage has occurred. Therefore, AKI biomarkers may allow a diagnosis of AKI to be made earlier but also indicate where renal injury has occurred even in the absence of subsequent dysfunction [49]. This concept of sub-clinical or silent AKI implies that although AKI may not clinically manifest, it has occurred but has escaped detection by conventional means. This requires a shift in approach from a predictive role of the AKI biomarker to that of a diagnostic or indeed prognostic tool. Coupled with this would be the acceptance that acknowledging such “biomarker positivity” or subclinical (silent) AKI would reflect on the epidemiology of AKI as well as affect the management of patients and perhaps result in better outcomes. However, this approach implies that the presence of a positive biomarker may result in poorer outcomes. Indeed there is evidence that this is the case. For example, a pooled analysis on over 2,000 patients with cardiorenal syndrome examined both plasma and urinary NGAL. Patients