TIMP-2*IGFBP7 as an auxiliary identi cation of successful discontinuation CRRT and prediction of renal recovery in critically ill patients: a case control study

yuanyuan xie (  xie_yuanyuan516@126.com ) Renji Hospital, School of Medicine, Shanghai Jiaotong University https://orcid.org/0000-0001-62701935 Alexander Zarbock department of anesthesiology, intensive care and pain medicine, university hosital munster Alessandra Brendolan Ospedale San Bortolo di Vicenza Francesca Martino Ospedale San Bortolo di Vicenza Sara Samoni Ospedale San Bortolo di Vicenza Nicola Marchionna Ospedale San Bortolo di Vicenza Weixuan fan internation renal research institute of vicenza Silvia De Rosa Ospedale San Bortolo di Vicenza Davide Giavarina Ospedale San Bortolo di Vicenza Gregorio Romero-González clinica universidad de navarra Claudio Ronco Ospedale San Bortolo di Vicenza


Introduction
Acute kidney injury (AKI) is a common complication in critically ill patients and patients who require renal replacement therapy (RRT) have an increased mortality even when adjusted for severity of disease [1,2].
Several studies have investigated the optimal timing to start continuous renal replacement therapy (CRRT) [3,4,5] without de nitively answering the question. Some studies proposed that biomarkers can be used to predict the development of a severe AKI and need for RRT [6,7]. However, only a few studies investigated the use of biomarkers to predict successful discontinuation of CRRT [8][9][10][11]. Newer biomarkers, TIMP-2*IGFBP7, have been shown to be able to predict long term outcome and renal recovery [12,13].
In daily practice, physicians terminate CRRT because the urine output increases, kidney function recovers, or the demand on the kidney is reduced due to improvement of other organ functions. Previous studies showed that different factors are associated with renal recovery, including less severe organ failure, shorter duration of RRT, lower age, higher urine output during CRRT, and decreasing plasma NGAL on the rst day of RIFLE-F [8-10, [14][15][16][17][18][19][20]. However, none of these factors is able to predict successful discontinuation of CRRT. Although it has been shown that biomarkers can predict AKI progression [21,22], it is unknown whether biomarkers can predict successful discontinuation of CRRT The prediction of the successful discontinuation of CRRT in critically ill patient might prevent harmful under-and overtreatment and subsequently improve patients' outcome and rational use of health care resources. As none of the clinical variables can reliably forecast successful weaning from CRRT, we investigated in this study whether TIMP-2*IGFBP7 in addition to renal and non-renal parameters can predict successful CRRT weaning.

Study design and participants
This is a single-center prospective observational study. All patients (age≥18 years old) admitted to ICU requiring CRRT from October 2017 to April 2019 at San Bortolo Hospital (Vicenza, Italy) were enrolled in the study. Medical and surgical ICU patients were enrolled. The initiation and cessation of CRRT were decided by nephrologist. Ethical Commite of San Bortolo Hospital of Vicenza approved this study (protocol number 03/17). The consent to participate is pursuant to Italian laws and the written informed consent was obtained from all participants or their legal representative.
Patients with dialysis requirement before ICU admission, end stage renal disease (ESRD), and patients with chronic kidney disease (CKD) stage 4 were excluded. The primary endpoint was the independence from RRT, IHD (intermittent hemodialysis) or restart of CRRT, for at least 7 days after CRRT discontinuation. Restart of CRRT were decided by nephrologist depending on patients' internal environment and capacity load. Renal recovery, which was de ned as serum creatinine (SCr) level <1.5 times the baseline value and independence RRT at ICU discharge or day 28 (the treatment time of CRRT did not exceed or reach 28 days), was the secondary endpoint [23]. TIMP-2*IGFBP7 lasting positive was de ned as TIMP-2*IGFBP7 concentration >2(ng/ml) 2 /1000 at enrollment and TIMP-2*IGFBP7 concentration >0.3(ng/ml) 2 /1000 at discontinuation of CRRT.
Blood and urine samples for measuring SCr and urinary TIMP-2*IGFBP7 levels were collected immediately upon enrollment and at the moment when CRRT was stopped. TIMP-2*IGFBP7 results were blinded to clinicians. Blood samples for measuring SCr concentration were also monitored 7 days after discontinuation and at ICU discharge or at day 28, whatever occurred rst. All clinical data were collected from hospital records, including demographic data, medical history, preadmission creatinine (baseline SCr, de ned as creatinine up to six months prior to admission), urine output, preadmission estimated glomerular ltration rate (eGFR) (calculated with CKD-EPI formula), previous kidney disease, reason for ICU admission (post-operative, sepsis, respiratory failure, post cardiac arrest, cardiac failure), cause of AKI (de ned as toxic, sepsis, primary renal disease, ischemic/other) and other variables. In addition, data on RRT, CRRT restart, ICU discharge, and death were recorded. SCr was measured by enzymatic method with an automatic analyzer (Dimension Vista, Siemens healthcare, Tarrytown, NY, USA). TIMP-2*IGFBP7 levels in the urine were measured by using NephroCheck® (ASTUTE140 meter, Astute Medical, San Diego, CA, USA) approved by the Food and Drug Administration (FDA). TIMP-2*IGFBP7 concentration > 0.3 (ng/ml) 2 /1000 is positive [24].
Estimated glomerular ltration rate (eGFR) was determined by using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation. Baseline SCr was determined by the concentration obtained up to six months before hospital admission or at ICU admission in critically ill patients when previous values were not available.

Sample size calculation
Estimation of sample size is based on preliminary data obtained in a pilot study performed in 77 patients with CRRT requirements in ICU. Estimations performed by using sample size calculators for designing clinical research (www.sample-size.net), assuming an error alpha of 0.05 and a power of 90%, rendered a sample size of 67 patients.
Statistical analysis SPSS 20.0 statistical software IBM (NY, USA) was used. Variables were tested for normal distribution using the Kolmogorov-Smirnov test. Normally distributed variables are expressed as mean (standard deviation), non-normally distributed variables as interquartile ranges [P25, P75], and categorical data as number and percentage. Unpaired Student's t-test, Mann-Whitney-U test or Chi-square test was used, where appropriate. Statistical signi cance was de ned as p < 0.05.
Mann-Whitney test was used for two-group comparison. Categorical variables were compared between groups using Fisher exact test or the chi-square test. The sensitivity and speci city of TIMP-2*IGFBP7 concentrations at enrollment and discontinuation of CRRT for successfully discontinue CRRT were evaluated using the receiver operating characteristic (ROC) curve and area under the curve. Logistic regression analysis was used to analyze related risk factors that in uenced discontinuation CRRT. Cox multifactor regression analysis was used to analyze the relationship between all risk factors and successful discontinuation CRRT. Multicollinearity was checked with a maximum variance in ation factor (VIF) of 10. The Kaplan-Meier curves were plotted to assess the renal recovery. Finally, all P-values were two-sided and a P-value<0.05 was considered statistically signi cant.

General data and outcomes
From October 2017 to April 2019, a total of 136 patients were screened and 73 patients were enrolled in the study (Figure 1). Baseline characteristics of the included patients are listed in Table 1. Admission diagnoses of the patients are summarized in Table 1. There are no differences regarding TIMP-2*IGFBP7 levels at enrollment and at CRRT discontinuation between the patients of different pathogeny. 45 patients were successfully weaned from CRRT (61.6%), and 34 patients had renal recovery (46.6%) (Figure 1). 20.5% of cohort died during their ICU stay.
Identi cation of successful CRRT discontinuation There were no differences regarding different clinical and laboratory variables, including mean arterial blood pressure, PCT, and lactate, between the successful and unsuccessful CRRT discontinuation groups (Table 2). However, the TIMP-2*IGFBP7 levels at enrollment and at CRRT discontinuation were signi cantly lower in patients in which CRRT discontinuation was successful, and urine output was signi cantly increased in this patient group compared to patients in which CRRT discontinuation failed ( Table 2). The CRRT treatment duration was signi cantly longer in patients who had a TIMP-2*IGFBP7 concentration > 2 (ng/ml) 2 /1000 at enrollment (9.93±1.59 days versus 5.43±0.66 days, P=0.013).
There was no multiple collinearity in our model by calculating the variance in ation factor (VIF) cut-off below 10.
Among the 29 patients, who had an urine output less than 400ml at the moment when CRRT was stopped, successful CRRT discontinuation rate increased form 0% in biomarker-positive at CRRT discontinuation to 88.9% in biomarker-negative patients (P=0.000).

Identi cation of renal recovery
Patients in the renal recovery group had signi cantly lower TIMP-2*IGFBP7 levels at enrollment and at CRRT discontinuation as well as a signi cantly increased urine output compared to the patients in the non-recovery groups (Table 4). There were no differences regarding the other parameters between the recovery and non-recovery group (Table 4).
TIMP-2*IGFBP7 values were signi cantly higher in non-renal recovery patients compared to patients with renal recovery at enrollment and at the discontinuation of CRRT (Table 4).

Discussion
The TIMP-2*IGFBP7 has been validated in previous cohorts as a potential biomarker in the detection of AKI risk [24][25][26][27]. To our knowledge, this is the rst study that describes the role of TIMP-2*IGFBP7 in identifying effective discontinuation of CRRT and renal recovery in critically ill patients.
Over the last decades, there has been considerable progress in the development of biomarker utilization for AKI, including risk assessment, diagnosis, prognosis and management. Urinary TIMP-2 and IGFBP7 are novel biomarkers of cell cycle arrest, which may detect the stress of renal epithelial cells and shutdown function without permanent injury. Urinary TIMP-2*IGFBP7 values can be used as diagnostic test to predict AKI in various clinical settings [28-30] and to predict long-term adverse outcome in AKI patients [12,31]. In these study [12,31], all-cause mortality and the need of RRT were evaluated. High TIMP-2*IGFBP7 levels were signi cantly associated with death or RRT in AKI patients. Most clinical variables cannot predict renal recovery and successful weaning from RRT. In a recently published study, urine output, serum and urine creatinine levels were identi ed as variables that might determine the optimal timing of therapy discontinuation [36]. Urine output is the clinical variable that has the best performance for predicting the successful discontinuation of CRRT [37, 38]. However, urine output can be affected by different factors, including ultra ltration and diuretics. Therefore, new variables need to be found which can predict renal recovery and successful discontinuation of RRT.
Here, we have found that renal recovery after AKI prior to ICU discharge or 28 days signi cantly increased in patients who had a TIMP-2*IGFBP7 concentration <2 (ng/ml) 2 /1000 at the enrollment and were TIMP-2*IGFBP7-negative at the time CRRT was discontinued. Both TIMP-2*IGFBP7 concentration >2 (ng/ml) 2 /1000 at enrollment and TIMP-2*IGFBP7 lasting positive are risk of non-renal recovery. The risk of non-renal recovery was 6 times higher in patients with TIMP-2*IGFBP7 concentration >2 (ng/ml) 2 /1000 at enrollment. The decrease of the urinary TIMP-2*IGFBP7 levels during the course of the disease could potentially predict renal recovery after AKI.
This study has also some limitations. Firstly, TIMP-2*IGFBP7 is a biomarker to predict the risk of AKI within 12 to 24 hours. While we measured urinary TIMP-2*IGFBP7 concentrations immediately upon enrollment, some patients started CRRT after 24 hours or more, and we did not monitor the changes of marker. Secondly, during the CRRT treatment we did not observe the changes. A persistent rise or fall to be negative may be more sensitive. Building on these reasons, we have implemented sequential measurement of TIMP-2*IGFBP7 and established a nephrology rapid response team to care and manage the process of AKI [39].

Conclusions
Despite of the limitations, the study provided important knowledge. TIMP-2*IGFBP7 was a potential auxiliary tool to identify effective CRRT discontinuation and predict renal recovery after AKI. This nding further con rmed a biologic link between the marker and clinical disease process. Measurement of these biomarkers may improve the ability of management and caring for critically ill patients under CRRT treatment. Measurement of the levels of tubular damager biomarkers complements the information of predicting renal recovery and disease prognosis. Declarations Ethics approval and consent to participate The protocol was approved by the medical ethics committee of the San Bortolo Hospital and the local ethical committees. Written informed consent was obtained from all participants or their legal representative.

Consent for publication
All authors read and approved the nal manuscript.
Availability of data and materials All relevant data are within the paper and its supporting information les. All data are fully available without restriction.

Competing interests
None.

Funding
None.
Authors' contributions R.C. and G.G. designed the study; B.A., S.S., N.M., X.Y. and F.W. acquired the data; X.Y. and R.C. interpreted the data; X.Y. and M.F. did statistical analysis; X.Y. and Z.A. drafted the manuscript; R.C., B.A., R.S. and G.D. supervised and mentored the study. All authors read and approved the nal manuscript.