Secondary Logo

Do We Need Noninvasive Biomarkers for Delayed Graft Function After Kidney Transplantation?

Van Loon, Elisabet, MD1,2; Naesens, Maarten, MD, PhD1,2

doi: 10.1097/TP.0000000000002473

1 Department of Microbiology and Immunology, KU Leuven, Leuven, Belgium.

2 Department of Nephrology and Renal Transplantation, University Hospitals Leuven, Leuven, Belgium.

Received 27 September 2018.

Accepted 2 October 2018.

The authors declare no funding or conflicts of interest.

Both authors contributed equally to the writing of this manuscript.

M.N. is supported by the Seventh Framework Programme (FP7) of the European Commission, in the HEALTH.2012.1.4-1 theme (grant number 305499); The Research Foundation Flanders (F.W.O.) and Flanders Innovation and Entrepreneurship (Agentschap Innoveren en Ondernemen) of the Flemish Government (grant numbers IWT.2015.0199 and T004417N) and KU Leuven (grant number C32/17/049). E.V.L. holds a fellowship grant (1143919N) from The Research Foundation Flanders (F.W.O.) and M.N. is senior clinical investigator of The Research Foundation Flanders (F.W.O.) (1844019N).

Correspondence: Maarten Naesens, MD, PhD, Department of Nephrology and Renal Transplantation, University Hospitals Leuven Herestraat 49, 3000 Leuven, Belgium. (

Delayed graft function (DGF) after kidney transplantation associates with impaired short- and long-term graft outcomes, early rejection, and prolonged postoperative hospitalization.1 Despite introduction of less nephrotoxic strategies, the incidence of DGF has not changed in recent years, as more kidneys from extended criteria donors and from donors after circulatory death (DCD) are transplanted to overcome the relative shortage of organs.

Over the last decade, many noninvasive biomarkers have been proposed in the field of kidney transplantation,2 primarily for acute rejection. Most of these markers, however, lack sufficient validation and specificity for disease processes, impeding their implementation in clinical practice. Many proof-of-concept studies target only 1 or few candidate markers while ignoring other promising markers floating around in the literature. The constant quest for innovation and the ensuing lack of thorough validation of previously proposed markers pose a major hurdle to advancement of the field.

Back to Top | Article Outline

Urinary TIMP-2 as a Biomarker for DGF

In this issue, Bank et al3 report on urinary TIMP-2 for monitoring of DGF in kidneys from DCD, essentially validating previous literature on this molecule in this setting (Table 1).4-6 Bank et al found that higher TIMP-2 levels in urine, corrected for urine osmolality (TIMP-2/mOsm), have excellent diagnostic accuracy (area under the receiver operator curve [ROC AUC], 0.91) for DGF, as early as day 1 posttransplant. When measured at day 10, higher TIMP-2/mOsm levels predicted prolonged DGF (>21 days), with retaining of good accuracy (ROC AUC, 0.80). An independent cohort for validation was not available, but partially accounted for by internal cross-validation. Finally, and perhaps most interestingly, trajectories of consecutive urinary TIMP-2/mOsm measurements over time after transplantation, in 3 patients, suggested that decreasing TIMP-2 levels preceded estimated glomerular filtration rate (eGFR) increase and resolution of DGF.



The increasing literature on urinary TIMP-2 and IGFBP7 (Table 1)3-6 builds on the promise of these molecules as biomarkers for acute kidney injury in mixed intensive care populations and different postoperative settings. Measurement of TIMP-2 and IGFBP7 has even been elaborated as a Food and Drug Administration–approved test for acute kidney injury (“NephroCheck,” Astute Medical, San Diego, CA).7 TIMP-2 and IGFBP7 are G1 cell cycle arrest markers that increase in response to stress such as ischemia, toxins, oxidative stress, or inflammation.8

A captivating approach of the present study is the monitoring of TIMP-2 over time, in an attempt to draw trajectories of graft function. Especially the finding that decreasing TIMP-2 levels decrease prior to eGFR improvement is alluring for management of patients with DGF in clinical practice, although the fact that only few patients were followed over time makes it necessary to confirm this finding in a larger population with the ability for appropriate statistical analysis.

It should be noted that in the published literature on TIMP-2 in kidney transplantation (Table 1), definitions of DGF are heterogeneous as well as donor characteristics, with repercussions on DGF prevalence. The higher prevalence of DGF in this study by Bank et al is most probably explained by the use of an extended definition for DGF, namely, “functional DGF” and the restriction to DCD kidneys. The general validity of the findings by Bank et al is therefore not established, but taken together with the previous literature, this biomarker merits further discussion.

Back to Top | Article Outline

Potential Clinical Relevance of TIMP-2 as Biomarker

Given the potential value of TIMP-2 as biomarker for DGF in the initial phase after transplantation of DCD kidneys (Table 1),3-6 the question arises what could be the next steps in the development of this candidate marker toward clinical use.

Back to Top | Article Outline

Validation in Large Unselected Populations

Replication of the suggestions that TIMP-2 could be a biomarker for DGF after kidney transplantation in a larger multicenter study is a clear necessity. Obviously, to increase power and opportunity to draw firm conclusions, and also to assess in detail the diagnostic performance. Indeed, an essential step to validation of a biomarker is assessment of diagnostic performance in an unselected population, with real-life disease prevalence. Before a biomarker can be further elaborated for application in clinical practice, information on parameters like sensitivity, specificity, positive predictive value, and negative predictive value is essential. The assessment of these parameters can only be done in a representative unselected population, since positive predictive value and negative predictive value are highly dependent on prevalence.

Back to Top | Article Outline

Comparison With Already Available and Routinely Used Markers

Another factor of major importance in biomarker development and validation is comparison of innovative biomarkers with already available and established markers. Bank et al described that urinary 24-hour creatinine excretion has at least comparable or even better performance to urinary TIMP-2/mOsm (ROC AUC = 0.98 at day 1, ROC AUC = 0.82 at day 10) but suggest that daily 24-hour urine collections and confounding factors (like residual renal function and hemodynamic factors) might limit the reliability of this universally available marker. However, if the diagnostic performance of a readily available and cheap biomarker is this high, one can question whether it is worth investing in further validation of innovative biomarkers for early detection of DGF, that have at best a similar diagnostic and predictive performance.

Back to Top | Article Outline

Clinical Relevance of a Biomarker for DGF

Perhaps the most important question in biomarker development relates to the potential clinical place of a biomarker. The added value of a biomarker to diagnose DGF is intuitively limited, as DGF is defined on serum creatinine or dialysis need and thus “detection” of DGF is evident from clinical evaluation and does not need an additional diagnostic biomarker.

However, as proposed by Bank et al, TIMP-2 might become useful as a very early predictor of future DGF stratifying patients before DGF diagnosis and allowing more timely measures to avoid DGF, whatever these may be. It should be noted that TIMP-2 is not specific for the underlying disease process causing DGF (rejection, calcineurin inhibitor toxicity, ischemia reperfusion injury, etc.), and thus targeted therapeutic implications upon changes in TIMP-2 levels seem illusive. Moreover, whether TIMP-2 is a better predictor for DGF than trajectories of serum creatinine, urine output, or 24-hour creatinine excretion needs further study.

Finally, the illustration of time evolution data of TIMP-2 presented by Bank et al in a few cases suggests that the duration of DGF could potentially be estimated by use of TIMP-2/mOsm levels, which, if validated, could be of clinical relevance. Nevertheless, also this prognostic capacity will need to be confronted with trajectories of readily available markers like serum creatinine, urine output, or 24-hour creatinine excretion.

Back to Top | Article Outline


In conclusion, Bank et al illustrate that urinary TIMP-2/mOsm is a noninvasive biomarker with excellent performance for diagnosis of DGF and potentially also prognostic capacities. Most studies on biomarkers present combinations of multiple biomarkers, so the excellent performance of a single biomarker is an alluring accomplishment. However, further work is necessary to elucidate the full diagnostic accuracy of TIMP-2/mOsm. We need larger multicenter trials and comparison with readily available biomarkers (serum creatinine, urine output, and urinary creatinine excretion) for early prediction of DGF and prediction of recovery of kidney transplant function in patients with DGF.

Let us not throw out our dirty water until we get in fresh.

Back to Top | Article Outline


1. Schröppel B, Legendre CDelayed kidney graft function: from mechanism to translation. Kidney Int. 2014;86:251–258.
2. Naesens M, Anglicheau DPrecision transplant medicine: biomarkers to the rescue. J Am Soc Nephrol. 2018;29:24–34.
3. Bank JR, Ruhaak R, Soonawala D, et alUrinary TIMP-2 predicts the presence and duration of delayed graft function in donation after circulatory death kidney transplant recipients. Transplantation. 2018; 102:S181.
4. Pianta TJ, Peake PW, Pickering JW, et alEvaluation of biomarkers of cell cycle arrest and inflammation in prediction of dialysis or recovery after kidney transplantation. Transpl Int. 2015;28:1392–1404.
5. Yang J, Lim SY, Kim M-G, et alUrinary tissue inhibitor of metalloproteinase and insulin-like growth factor-7 as early biomarkers of delayed graft function after kidney transplantation. Transplant Proc. 2017;49:2050–2054.
6. Schmitt FCF, Salgado E, Friebe J, et alCell cycle arrest and cell death correlate with the extent of ischaemia and reperfusion injury in patients following kidney transplantation - results of an observational pilot study. Transpl Int. 2018;31:751–760.
7. Bihorac A, Chawla LS, Shaw AD, et alValidation of cell-cycle arrest biomarkers for acute kidney injury using clinical adjudication. Am J Respir Crit Care Med. 2014;189:932–939.
8. Endre ZH, Pickering JWCell cycle arrest biomarkers win race for AKI diagnosis. Nat Rev Nephrol. 2014;10:683–685.
Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.