Nephrocheck®: Checkmate or reality check?

Acute kidney injury (AKI) is a sudden deterioration in kidney function that occurs in 15% of hospitalized patients and can lead to serious complications, including irreversible kidney damage and death. ¹ It is defined by the Kidney Disease Improving Global Outcomes (KDIGO) 2012 guidelines as a rise in serum creatinine or fall in urine output. ² However, urine output is impractical and therefore rarely measured in clinical practice, while serum creatinine is known to be a slow riser, taking up to 24–40 h to rise after kidney injury. ³ As a result, there is strong interest in finding new biomarkers that can serve as kidney ‘troponins’, rising early after the onset of kidney injury while being highly specific to the kidneys.

Over the last two decades, a handful of biomarkers have emerged with the potential to fulfill that need, but only three have so far been approved for clinical use in various countries: Neutrophil gelatinase-associated lipocalin (NGAL), Liver-type fatty acid-binding protein (L-FABP), and the combined Tissue Inhibitor of Metalloproteinases-2 (TIMP-2) and Insulin-like Growth Factor-Binding Protein 7 (IGFBP7) (commonly marketed as NephroCheck® by Astute Medical, San Diego, CA – now part of BioMérieux, Inc., Lyon, France). ⁴ In studies comparing the performance of all three markers for the early detection of AKI, the combined urinary [TIMP-2] × [IGFBP7] performed best. However, barriers still exist to its widespread clinical adoption.

In this issue of Annals of Clinical Biochemistry, Ilaria et al.⁵ review the performance of urinary [TIMP-2] × [IGFBP7] for the early detection of AKI in clinical studies. As the authors point out that both TIMP-2 and IGFBP7 are cell cycle arrest biomarkers and are expected to increase in response to stress or injury to the kidneys in order to prevent proliferation of injured cells. Both markers are measured in urine using fluorescence immunoassay on the ASTUTE140® meter or a chemiluminescent immunoassay on VITROS®, with the product of both divided by 1000 reported as AKIRisk™ in (ng/mL)²/1000. Early validation studies evaluated the performance of two different cut-offs for the identification of patients at risk of developing AKI stage 2 or 3 within 12 h: a sensitive one at 0.3 (ng/mL)²/1000 suggesting moderate risk of AKI development, and a specific one at 2.0 (ng/mL)²/1000 suggesting high risk of AKI development. In comparison with other AKI biomarkers, it had the best performance, averaging an area-under-the-curve (AUC) of around 0.80. This performance can be improved further when the test is used in conjunction with clinical prediction models or AKI e-alerts, as outlined by the authors’ own experience from the nephrology rapid response team (NRRT) protocol at San Bortolo Hospital in Vincenza.

Clinical adoption of urinary [TIMP-2]×[IGFBP7] is gaining momentum, especially since it was recently recommended for use in the US by the Enhanced Recovery After Surgery (ERAS) Cardiac Society guidelines for perioperative care of cardiac surgery patients. ⁶ However, laboratory medicine and nephrology expert groups have not officially endorsed its use. In fact, the National Institute for Health and Care Excellence (NICE) in the UK and the American Association for Clinical Chemistry’s AACC Academy in the US (guideline in press), both recently reviewed the evidence and do not recommend urinary [TIMP-2] × [IGFBP7] for clinical use. ⁷ It is important to understand why these organizations have not endorsed this biomarker and what is required in order for that to happen.

Urinary [TIMP-2] × [IGFBP7] has so far been studied in over 1800 critically ill patients in different settings with promising results (as reported by AUCs). ⁸ However, there have been no reported improvements in clinical outcomes associated with its use (as measured by, for example, initiation of renal replacement therapy, mortality, length of stay and/or major adverse kidney events at hospital discharge), and its performance has varied widely depending on the cut-off used, population tested and testing time.^8–10 This variability may be explained in part by the fact that the reference interval for urinary [TIMP-2] × [IGFBP7] was reported to be 0.04 to 2.22 (ng/mL)²/1000, spanning both recommended clinical cut-offs (0.3 and 2.0). ¹¹ In addition, the cut-off at 0.3 (ng/mL)²/1000 falls near the middle of the reference interval. ¹¹ So, at least half of the patients who will not go on to develop AKI will falsely test positive at the 0.3 (ng/mL)²/1000 cut-off. As a result, this biomarker will perform poorly by itself in populations with low prevalence of AKI (i.e. where the proportion of individuals who will falsely test positive is higher). One potential solution to this problem would be to risk-stratify patients upfront (prior to testing) in an effort to increase the pretest probability, as was done using the NRRT protocol at San Bartolo Hospital. ¹² An alternative solution may be to evaluate further the performance of this urinary marker after normalizing its values to urine creatinine or urine osmolality to account for a patient’s hydration status; this is commonly done when measuring analytes in urine.^13,14 This would require re-establishing reference intervals and examining if greater separation is achieved in the values of urinary [TIMP-2] × [IGFBP7] between patients who do not develop AKI and those who do. Finally, as the authors also point out in their review, the timing and frequency of biomarker measurement continues to be an issue and future studies are needed to understand how the marker changes in critically ill patients. Along similar lines, there are no reported biological variability studies to help us understand how urinary [TIMP-2] × [IGFBP7] changes throughout the day in healthy individuals, an essential step to determining reference change values.

In summary, urinary [TIMP-2] × [IGFBP7] has the potential to transform how we detect and treat AKI, especially when implemented as part of care bundles. However, urinary [TIMP-2] × [IGFBP7] is not ready for clinical implementation due to its poor specificity. Thus, more research is needed to better characterize and improve its performance, either by combining it with other risk stratification methods or normalizing its values to urine creatinine or urine osmolality. Ultimately, it will all come down to its ability to improve clinical outcomes. Given recent reports on machine learning algorithms used to detect AKI, urinary [TIMP-2] × [IGFBP7] may also face competition from an unlikely yet intelligent source. ¹⁵