Renal markers for monitoring acute kidney injury transition to chronic kidney disease after COVID-19

The association of renal markers with renal recovery after acute kidney injury (AKI) and progression to chronic kidney disease (CKD) after coronavirus disease 2019 (COVID-19) is apparently unexplored. Tubular epithelial cells are the cells most affected in COVID-19-associated AKI, in line with other AKI forms in critical illness [1, 2]. Maladaptive repair of injured tubules post-AKI may trigger persistent profibrotic signalling and delayed resolution of inflammation along multiple pathways that drive renal tubulointerstitial fibrosis and thereby progressive CKD [3]. Urinary dickkopf-3 (uDKK3) is a recently described novel CKD progression marker in various causes of CKD [4, 5]. DKK3 is a stress-induced, renal tubular epithelia–derived secreted glycoprotein that induces tubulointerstitial fibrosis through its action on the canonical Wnt–β-catenin signalling pathway [5]. The urinary protein:creatinine ratio, albumin:creatinine ratio (ACR) and α1-microglobulin:creatinine ratio (α1MGCR) are commonly evaluated glomerular dysfunction/damage and tubule cell dysfunction markers in CKD [6]; higher levels have also been associated with adverse outcomes in COVID-19 [2]. Accordingly, we evaluated the role of renal markers for monitoring subsequent estimated glomerular filtration rate (eGFR) decline in hospitalized COVID-19 patients. This is the second report of an ongoing prospective, observational, single-centre study investigating the temporal progression of renal markers in hospitalized COVID-19 patients. In the present study, patients were included if they consented to the linkage with administrative data for long-term follow-up. The protocol was approved by the local ethics committee (AZ 58/20) and complied with the Declaration of Helsinki. The participants or their legally authorized representatives provided written informed consent. The study was registered at ClinicalTrials.gov (NCT04353583). A decrease in eGFR (using the Chronic Kidney Disease Epidemiology Collaboration equation) 6 months post-admission was defined as an eGFR decline <5 and ≥5 mL/min/1.73 m2 from baseline based on theKidney Disease: Improving Global Outcomes CKD guidelines [6]. Baseline serum creatinine (SCr) was the most recent outpatient SCr taken 7–365 days pre-admission. Other baseline renal parameters beyond SCr were not available. Spot urine samples were collected daily on hospital admission [uDKK3 and urinary interleukin (IL)-6 were collected thrice weekly] up to discharge day and at outpatient follow-up 3 and 6 months post-admission. Detailed methods (including rationale for cut-off values) are described in the Supplementary data. Of 66 patients up to 9 November 2020, renal parameters up to 6 months post-admission (sample size numbers for key time points provided in Supplementary data, Table S1) were available for 55 patients (83.3%; the remaining 11 patients died in hospital). Patients with an eGFR decline ≥5 mL/min/1.73 m2 were more likely to be older and to have hypertension, diabetes, CKD, coronary artery disease and higher disease severity scores at admission (Supplementary data, Table S2). Patients with an eGFR decline ≥5 mL/min/1.73 m2 were more likely to be admitted to the intensive care unit, develop AKI (63.2% versus 5.6%), require invasive mechanical ventilation during the hospital stay, receive piperacillin/tazobactam or vancomycin and have a longer hospital stay (Supplementary data, Table S3). The temporal profile of renal parameters between both groups was markedly different (Table 1). Patients with subsequent higher versus lower eGFR decline had higher admission SCr values compared with baseline, likely owing to a higher proportion of patients presenting with AKI at admission (36.8% versus 2.8%). At Day 31 post-admission, both groups exhibited decreased SCr compared with baseline, likely reflecting decreased muscle mass associated with critical illness, causing eGFR overestimation. Cystatin C is considered a more accurate marker for estimating the GFR, but the relatively high levels observed in comparison with SCr may have been influenced, at least partially, by the high-dose corticosteroid medication on Days 7 and 31 post-admission (Table 1). It should be noted that 40 patients (72.7%) were recruited before the World Health Organization guidance on corticosteroids for COVID-19 was released, which is reflected in the low overall prescription. Notably, while the higher eGFR decline patients had sustained greater damage and higher functional marker levels, patients with lower eGFR decline showed rapid normalization after Day 7 post-admission. Furthermore, the former group exhibited a distinct biphasic pattern of uDKK3 and urinary IL-6 levels, with greater separation at Day 31 (Figure 1). (A) Albumin:creatinine ratio, (B) α1MGCR, (C) DKK3:creatinine ratio and (D) IL-6 excretion levels in urine stratified by eGFR decline 6 months after hospital admission. The figure illustrates the biomarker levels in patients with COVID-19 on Days 0 (day of hospital admission), 7, 31 and 180 post-admission stratified by eGFR decline (blue line <5 mL/min/1.73 m2; red line ≥5 mL/min/1.73 m2). eGFR decline was defined as the difference in eGFR values 6 months post-COVID-19 and premorbid ‘baseline’ values. Points are geometric means; error bars show the 95% confidence intervals. Values up to Day 31 were obtained from the time-course model and values for Day 180 are simple means (not considering values from earlier or later measurements). The dashed horizontal lines indicate the levels of non-physiologic concentrations [4, 6]. Note the use of a logarithmic scale on the y-axis. Due to the log scale, IL-6 values of zero are not displayed. Temporal profile of renal and inflammatory parameters of the cohort stratified by eGFR decline 6 months after hospital admissiona Values are expressed as the mean (95% confidence interval). Values from Day 0 up to Day 31 were obtained from the time-course model, premorbid values and values for Day 90 and Day 180 are simple means (not considering values from earlier or later measurements (see also Supplementary Material). eGFR decline was defined as the difference in eGFR values 6 months after COVID-19 and premorbid ‘baseline’ values. In the one patient who received KRT, all variables except hsCRP and urinary IL-6 were censored at the time of KRT due to the removal of creatinine, urea and uric acid. To convert the values for serum creatinine to μmol/l, multiply by 88.4. Eight patients (22.2%) received 6 mg dexamethasone to treat COVID-19. Twelve patients (63.2%) received 6 mg dexamethasone to treat COVID-19. Four patients (21.1%) received 8 mg dexamethasone equivalent to treat cryptogenic organizing pneumonia. To convert the value for urea to blood urea nitrogen, multiply by 0.467. hsCRP, high-sensitivity C-reactive protein; KRT, kidney replacement therapy; NA, not available. At 6 months post-admission, only the mean α1MGCR and uDKK3 remained elevated in the higher eGFR decline group. Furthermore, while the mean uDKK3 in the lower eGFR decline group was in the lower detection limit range, that in the higher eGFR decline group remained elevated and was comparable to that of other populations with progressive CKD [4]. The spline plots of the renal and inflammatory parameters during the observational period are shown in Supplementary data, Figure S1. Among COVID-19 survivors, short-term progression of renal function was associated with sustained greater damage and higher functional marker levels with a predominantly tubular pattern in the initial phase (up to 7 days post-admission). However, the uDKK3 and urinary IL-6 biphasic patterns in patients with higher eGFR decline suggest the development of a superimposed ‘second-hit’ injury on unresolved AKI, potentially contributing to sustained and/or progressive tubular stress and renal inflammation, resulting in maladaptive repair and AKI–CKD transition. Sustained uDKK3 elevation at 6 months may indicate the disease chronification of acute tubular damage; commonly used urinary markers (e.g. α1MGCR, ACR) might be less suitable for distinguishing patients with progressive CKD post-COVID-19. Study strengths are the prospective design, including patients post-AKI assessed for specific kidney damage, and evaluation of a biomarker panel to distinguish progression post-AKI. Study limitations are the single-centre design, small sample size and the duration of follow-up. Therefore the results must be considered as hypothesis generating. Another study limitation is that the number of samples obtained at Day 31 post-admission was significantly lower compared with the other key time points, since >50% of the patients were discharged earlier and were not available for outpatient visits on that day (e.g. due to rehabilitation). The observed uDKK3 and urinary IL-6 biphasic patterns show—in line with previous literature [3]—that a second wave of injury superimposed on unresolved AKI may represent a pathophysiological means of AKI–CKD transition. The aetiology of the biphasic pattern is unknown but apparently involves factors beyond systemic inflammation, as high-sensitivity C-reactive protein did not display such a pattern. Wnt–β-catenin signalling activation promotes CKD progression after experimental AKI through renin–angiotensin–aldosterone activation, stress-induced cytokine (e.g. IL-6 and IL-8) expression in tubular epithelial cells, uncontrolled fibroblast activation and extracellular matrix deposition [7, 8]. Although sustained activation of Wnt–β-catenin signalling is considered detrimental and promotes tubulointerstitial fibrosis, transient activation is thought to be beneficial by mitigating initial injury and accelerating subsequent recovery after AKI [9, 10]. Thus persistently elevated uDKK3 levels may indicate ongoing tubular ‘stress’ leading to progressive kidney fibrosis independent from the type of CKD. Notably, the extent to which the COVID-19-mediated renin–angiotensin–aldosterone system imbalance may contribute to its activation is unknown. Larger studies with long follow-up periods are needed to clarify the role of uDKK3 for monitoring the AKI–CKD transition post-COVID-19. Clinical Trial Registration Information: http://www.clinicaltrials.gov (NCT04353583). Supplementary data are available at ndt online. The authors thank the nursing staff of the medical and surgical intensive care units and outpatient departments for their hard work and commitment to patient well-being. Without their support, this work would not have been possible. This study did not receive any external funding. The authors shared responsibility for study design, data collection, data analyses and data interpretation, as well as preparation, review and approval of the manuscript. The corresponding author had full access to all study data and had final responsibility for the decision to submit for publication. H.-W.B. and C.R. are the senior authors of the work. F.H.-S., W.S., H.-W.B. and C.R. designed the study. G.V., S.S., U.M., I.V., K.T., B.J., H.D., J.T., S.K., B.A., H.R., M.S., S.H., W.S., S.J.S. and T.S. were responsible for literature research and clinical advice. F.H.-S., G.V., J.W., S.S., U.M., I.V., K.T., B.J., H.D., J.T., S.K., B.A., H.R., M.S., S.H., W.S., S.J.S., T.S., H.-W.B. and C.R. acquired, analysed and interpreted the data. S.S., H.-W.B. and C.R. were responsible for adjudication of renal function. F.H.-S., G.V., J.W., S.S., H.-W.B. and C.R. drafted the manuscript. F.H.-S., G.V., J.W., S.S., U.M., I.V., K.T., B.J., H.D., J.T., S.K., B.A., H.R., M.S., S.H., W.S., S.J.S., T.S., H.-W.B. and C.R. performed critical revision of the manuscript for important intellectual content. F.H.-S. and J.W. were responsible for statistical analyses. W.S., H.-W.B. and C.R. were involved in study supervision. K.T. received speaking fees from Actelion and Bayer outside the submitted work. W.S. received speaker/consultancy fees from Pfizer and Bayer Pharma outside the submitted work. C.R. reports support for acting as an advisory board member for ASAHI, Baxter, GE, Jafron and Medtronic and speaker’s fees from Astute, bioMérieux, B. Braun, Cytosorbents, ESTOR, Fresenius Medical Care and Toray, all unrelated to this work. None of the other authors declare any competing interests. The authors declare that the results presented in this article have not been published previously in whole or part. The data and R scripts used for analysis are available from the corresponding author upon reasonable request.