Introduction
AKI occurs in 7%–18% of hospital admissions (1). It is associated with a high economic burden (2) and increased risk of long-term adverse outcomes, including mortality and development of CKD (1,3). The latter may progress to ESKD requiring RRT and increases cardiovascular risk (3).
AKI is defined on the basis of the two markers of kidney function, serum creatinine (sCr) and urine output (2,4). Due to the rather low sensitivity and specificity of these markers of kidney function (2,5), multiple early biomarkers predicting AKI have been described (6). However, biomarkers that can aid patient stratification on the basis of the need for post-AKI follow-up to minimize the sequelae of AKI and that can be used as markers of intervention in post-AKI care are still lacking (3,7).
The pathologic processes that occur after an episode of AKI, including the maladaptive repair processes resulting in fibrogenesis, resemble those that drive CKD progression (7). Fibrosis is characterized by an imbalance between the formation and degradation of extracellular matrix components, such as collagen. Collagen type VI is found in the interface between the basement membrane and the interstitial matrix in the kidney (8,9). During formation of collagen type VI, the signaling fragment endotrophin (ETP) is released from mature collagen secreted from fibroblasts (10,11). The PRO-C6 ELISA measures levels of ETP and, therefore, also measures the formation of collagen type VI (12,13). Plasma levels of ETP have previously been shown to directly reflect the extent of fibrosis in the kidneys (14,15) and are associated with adverse outcomes in patients with CKD (12,1617–18).
In this study, for the first time, we investigated the potential of ETP as a marker of CKD in patients who previously sustained AKI. In addition, we studied the prognostic ability of ETP for longer-term kidney disease progression and mortality and compared its performance with eGFR.
Materials and Methods
Patient Population
The AKI Risk in Derby (ARID) study (ISRCTN25405995) is a prospective cohort study designed to report long-term outcomes after AKI (19). Between May 2013 and May 2016, the study recruited two cohorts of people who had been hospitalized at the Royal Derby Hospital (Derby, United Kingdom) and had survived to at least 90 days after hospital admission. One cohort consisted of people who had sustained AKI during hospital admission (AKI group), and a second cohort who had not (control group). After recruitment, AKI and control patients were matched 1:1 for baseline eGFR stage (eGFR >60 ml/min per 1.73 m2, eGFR stage 3A, 3B, or 4), age (±5 years), and presence of diabetes. A total of 866 participants were matched, with 433 participants in each of the AKI and control groups. For this analysis, we studied those patients (n=801) who were successfully matched, alive, and under active follow-up at 1 year because we aimed to measure sustained fibrotic processes at this time point, i.e., separate from the changes that happened at time of AKI and immediately afterwards (20). We also measured samples from a subset of patients with available plasma collected 3 months after the episode of AKI (n=280 in the AKI group and n=305 in the control group). The presence of AKI was determined according to sCr components of the Kidney Disease Improving Global Outcomes criteria (21). The baseline creatinine value was taken as the most recent stable sCr before hospital admission. Patients without a baseline sCr value in the 12 months preceding hospital admission were not eligible to participate. Urine output was not used due to its inaccurate recording in a general hospitalized population. Other exclusion criteria were total or partial nephrectomy during index admission, preexisting CKD stage 5, receiving palliative care, or CKD after renal transplantation. Approval for the study was obtained from Derbyshire Research Ethics Committee and the National Information Governance Board. All patients provided written informed consent. The study was conducted in compliance with the Declaration of Helsinki.
Data Collection
sCr, eGFR (calculated using the Chronic Kidney Disease Epidemiology Collaboration equation [21]), proteinuria-creatinine ratio (PCR), and urinary albumin-creatinine ratio (ACR) were measured at 1 year and 3 years after AKI onset. Kidney disease progression from month 3 or year 1–3 was defined as a decline in eGFR of ≥25% associated with a change in CKD stage (3). Participants were asked not to eat meat for at least 12 hours before giving a blood sample and to provide an early morning urine specimen. sCr was measured using an enzymatic assay, and urinary albumin was measured using an immunoturbidimetric assay (Tina-quant Albumin Generation 2), both on the Roche Cobas 702 module (Roche Diagnostics Limited, Burgess Hill, United Kingdom). Blood samples were collected in BD Vacutainer EDTA tubes, transported to the laboratory at ambient temperature, and centrifuged within 6 hours of collection at 3000 rpm/1508 RCF for 10 minutes. Plasma was then removed, aliquoted, and stored at −80°C.
Hospital admission data, Charlson score, inpatient laboratory test results, coded comorbidities, and mortality were extracted from the hospital electronic medical record. Commencement of long-term RRT was tracked by crossreferencing with the local renal database.
ETP Measurements
Plasma levels of ETP were measured at year 1 and month 3 using a competitive ELISA, PRO-C6, according to manufacturer instructions (Nordic Bioscience, Herlev, Denmark). The mAb used in the assay specifically detects the last ten amino acids of the α3 chain of collagen VI (3168′KPGVISVMGT′3177) (12,13).
Statistical Analysis
A Mann–Whitney test or chi-squared test was used to analyze the differences between groups. Spearman rank correlation was used to analyze the correlations between ETP and different variables. The ability of eGFR and ETP to discriminate between those with and without kidney disease progression and mortality occurring between month 3 or year 1 and year 3 were investigated using the receiver operating characteristic (ROC). Comparison of C-statistics was used to compare the ROC results of ETP and eGFR. Logistic regression for kidney disease progression and Cox proportional hazards regression for mortality were used to evaluate different univariate and multivariate models. Comparison of Kaplan–Meier curves for ETP tertiles was done using the Mantel–Cox test. Statistical analysis was performed using R software (version 3.6.2; R Development Core Team) and GraphPad Prism (version 8.4.3; GraphPad Software, LLC). P values <0.05 were considered significant.
Results
Characteristics of the Study Cohort
The study cohort consisted of 801 patients with plasma available at 1 year after the AKI episode: 393 in the AKI group and 408 in the control group. The baseline characteristics of the AKI and the control groups are shown in Table 1. In the AKI group, 114 of 393 patients (29%) had CKD before index hospitalization, versus 117 of 408 (29%) in the control group (P=0.92). A total of 176 patients had diabetes at baseline, with similar proportions in each group (85 [22%] in the AKI group and 91 [22%] in the control group; P=0.82).
Table 1. -
Patient characteristics
| Characteristic |
Baseline |
Year 1 |
Year 3 |
| AKI Group |
Control Group |
P Value |
AKI Group |
Control Group |
P Value |
AKI Group |
Control Group |
P Value |
|
N
|
393 |
408 |
0.60 |
393 |
408 |
0.60 |
313 |
351 |
0.14 |
| Sex, % female |
44 |
50 |
0.12 |
44 |
50 |
0.12 |
44 |
51 |
0.19 |
| Age |
71 (64–78) |
71 (64–76) |
0.43 |
72 (66–79) |
72 (65–78) |
0.44 |
74 (67–80) |
74 (67–80) |
0.74 |
| sCr, µmol/L |
89 (76–106) |
87 (74–105) |
0.24 |
100 (83–122) |
84 (72–105) |
<0.001a |
100 (83–123) |
86 (71–102) |
<0.001a |
| eGFR, ml/min per 1.73 m2
|
70 (57–82) |
72 (56–87) |
0.30 |
60 (45–74) |
73 (58–86) |
<0.001a |
60 (45–74) |
71 (57–84) |
<0.001a |
| PCR, ng/mmol |
NA |
NA |
— |
11 (7–20) |
9 (6–14) |
<0.001a |
12 (8–18) |
9 (6–14) |
<0.001a |
| ACR, mg/mmol |
NA |
NA |
— |
1.4 (0.5–6.0) |
0.8 (0.0–3.2) |
<0.001a |
1.3 (0.3–5.1) |
0.6 (0.0–2.3) |
<0.001a |
| CRP, mg/L |
NA |
NA |
— |
3.0 (1.6–7.0) |
3.0 (1.1–6.0) |
<0.05a |
3.0 (1.4–7.0) |
2.0 (1.0–6.0) |
0.06 |
| DM at BL, % yes |
22 |
22 |
0.82 |
NA |
NA |
— |
NA |
NA |
— |
| CKD at BL, % yes |
29 |
29 |
0.92 |
NA |
NA |
— |
NA |
NA |
— |
| CKD stage 1, 2, 3A, 3B, 4, 5, % |
16, 55, 20, 7, 2, 0 |
17, 54, 19, 7, 3, 0 |
0.79 |
8, 43, 25, 18, 5, 1 |
16, 56, 18, 7, 3, 0 |
<0.001a |
6, 44, 25, 18, 7, 0 |
17, 54, 18, 6, 5, 0 |
<0.001a |
| AKI stage 1, 2, 3, % |
60, 24, 16 |
NA |
— |
NA |
NA |
— |
NA |
NA |
— |
| Recurrent AKI, % |
— |
— |
— |
12 |
2 |
<0.001a |
13 |
7 |
0.01a |
Data are presented as median (interquartile range). Recurrent AKI for year 1 is from BL to year 1, and recurrent AKI for year 3 is from year 1 to year 3. Statistical difference between the AKI and the control group was assessed by Mann–Whitney or chi-squared tests. sCr, serum creatinine; PCR, proteinuria-creatinine ratio; NA, not available; ACR, albuminuria-creatinine ratio; CRP, C-reactive protein; DM, diabetes mellitus; BL, baseline.
aP<0.05.
Follow-up data were available for all 801 patients at year 1. At the year 1 follow-up, patients in the AKI group had lower eGFRs and higher levels of PCR and ACR compared with the control group. At the year 3 follow-up, patients in the AKI group still had significantly lower eGFR and higher PCR and ACR levels (Table 1).
Correlations and Group Distribution of ETP
Levels of ETP, measured using the PRO-C6 ELISA, at year 1 in both the AKI and the control groups correlated positively with age, sCr, PCR, ACR, and C-reactive protein, and negatively with eGFR measured at year 1 (all P<0.05; Table 2). There were no correlations between ETP measured at year 1 and the change in sCr, PCR, ACR, and C-reactive protein from year 1 to year 3 (Table 2). ETP levels at year 1 increased with increasing CKD stages in both the AKI and control groups (Figure 1).
Table 2. -
Correlations of endotropin Y1 with variables measured at Y1, Y3, and the change from Y1 to Y3
| Variables |
Endotropin Y1 and Variables Y1 |
Endotropin Y1 and Variables Y3 |
Endotropin Y1 and Δ Variables |
| AKI Group |
Control Group |
AKI Group |
Control Group |
AKI Group |
Control Group |
|
ρ
|
P value |
ρ
|
P Value |
ρ
|
P Value |
ρ
|
P Value |
ρ
|
P Value |
ρ
|
P Value |
| Age, yr |
0.15 |
<0.01a |
0.14 |
<0.01a |
0.13 |
<0.05a |
0.12 |
<0.05a |
0.04 |
0.49 |
−0.01 |
0.85 |
| sCr |
0.42 |
<0.001a |
0.38 |
<0.001a |
0.36 |
<0.001a |
0.35 |
<0.001a |
0.00 |
0.96 |
0.07 |
0.16 |
| eGFR |
−0.45 |
<0.001a |
−0.41 |
<0.001a |
−0.41 |
<0.001a |
−0.38 |
<0.001a |
0.03 |
0.62 |
−0.03 |
0.62 |
| PCR |
0.31 |
<0.001a |
0.18 |
<0.001a |
0.20 |
<0.01a |
0.14 |
<0.05a |
−0.05 |
0.34 |
0.02 |
0.70 |
| ACR |
0.23 |
<0.001a |
0.10 |
<0.05a |
0.19 |
<0.01a |
0.07 |
0.18 |
0.01 |
0.92 |
0.03 |
0.64 |
| CRP |
0.18 |
<0.001a |
0.21 |
<0.001a |
0.17 |
<0.05a |
0.13 |
<0.05a |
0.01 |
0.81 |
−0.05 |
0.38 |
Spearman rank correlation. Δ Variables were defined as measurements at year 3 minus measurements at year 1. Y1, year 1; Y3, year 3; Δ, change; sCr, serum creatinine; PCR, proteinuria-creatinine ratio; ACR, albuminuria-creatinine ratio; CRP, C-reactive protein.
aP<0.05.
;)
Figure 1.: Endotropin (ETP) levels in different CKD stages. Endotrophin (ETP) levels measured at year 1, divided into CKD stages, in (A) the AKI group and (B) in the control group. Data are presented on a log10-scale as median with 95% CI. Statistical differences between groups were assessed by Mann–Whitney tests. The dotted grid line indicates y-values of 10. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.
Patients in the AKI group had significantly higher ETP levels at year 1 (median [interquartile range (IQR)]=10.85 [8.41–14.99] ng/ml) compared with patients in the control group (median [IQR]=9.23 [6.94–12.35] ng/ml; P<0.001; Figure 2A). Patients with both AKI and preexisting CKD had the highest levels (median [IQR]=14.20 [10.60–18.85] ng/ml), whereas patients with neither AKI nor preexisting CKD had the lowest levels of ETP (median [IQR]=8.43 [6.53–11.21] ng/ml; P<0.0001) measured at year 1 (Figure 2B). When stratified by diabetes, patients with both AKI and diabetes at baseline had the highest levels (median [IQR]=12.53 [9.48–18.15] ng/ml), and patients with neither AKI nor diabetes at baseline had the lowest levels of ETP (median [IQR]=8.99 [6.80–12.08] ng/ml; P<0.0001) at year 1 (Figure 2C).
Figure 2.: Endotropin (ETP) levels in patients with AKI and controls with or without CKD and diabetes mellitus (DM). ETP levels measured at year 1 in (A) patients with and without AKI, (B) patients with AKI and preexisting CKD, and (C) patients with AKI and baseline diabetes mellitus (DM). Data are presented on a log10-scale as median with 95% CI. Statistical differences between groups were assessed by Mann–Whitney tests. The dotted grid line indicates y values of 10. ***P<0.001, ****P<0.0001. ns, not significant.
Prognostic Value of ETP for Kidney Disease Progression
Kidney disease progression was defined as a decline in eGFR of ≥25% in combination with a change in CKD stage. eGFR values from year 3 from 95 of the 801 patients were missing; therefore, kidney disease progression was only evaluated in 706 patients. There were 43 of 706 patients (6%) who had kidney disease progression from year 1 to year 3; 26 of 337 in the AKI group (8%) and 17 of 369 in the control group (5%; P=0.08). On the basis of ROC curve analysis, ETP measured at year 1 could discriminate patients with kidney disease progression at year 3 in the AKI group (area under the ROC curve [AUC]=0.67; P<0.01) but not in the control group (AUC=0.53; P=0.35), whereas eGFR at year 1 could not discriminate patients with kidney disease progression in either the AKI (AUC=0.51; P=0.57) or in the control group (AUC=0.61; P=0.94; Table 3). In the AKI group, ETP was a significantly better discriminator than year 1 eGFR (comparison of ROC curves, P<0.01; Table 3).
Table 3. -
Receiver operating characteristic curve analysis for kidney disease progression and mortality from Y1 to Y3
| Y1 Variables |
AKI Group |
Control Group |
| Area Under Curve (95% Confidence Interval) |
P Value |
Area Under Curve (95% Confidence Interval) |
P Value |
|
Kidney disease progression
|
| eGFR |
0.51 (0.41 to 0.61) |
0.57 |
0.61 (0.46 to 0.76) |
0.94 |
| ETP |
0.67 (0.55 to 0.79) |
<0.01a |
0.53 (0.36 to 0.70) |
0.35 |
| eGFR versus ETP |
— |
<0.01a |
— |
0.37 |
|
Mortality
|
| eGFR |
0.64 (0.59 to 0.69) |
<0.01a |
0.66 (0.61 to 0.71) |
<0.01a |
| ETP |
0.64 (0.59 to 0.68) |
<0.01a |
0.57 (0.52 to 0.62) |
0.24 |
| eGFR versus ETP |
— |
0.89 |
— |
0.08 |
Kidney disease progression was defined as ≥25% decline in eGFR and a decline in CKD stage. Number of events for kidney disease progression: 26 of 337 patients in the AKI group and 17 of 369 patients in the control group. Number of events for mortality: 43 of 393 in the AKI group and 27 of 408 in the control group. Y1, year 1; Y3, year 3; ETP, endotrophin.
aP<0.05.
The prognostic value of ETP for kidney disease progression was also assessed with univariate and multivariate logistic regression. In the AKI group, the univariate model with ETP was associated with kidney disease progression at year 3 (odds ratio [OR]=1.08; 95% CI, 1.03 to 1.14; P<0.01), but eGFR was not (OR=1.00; 95% CI, 0.98 to 1.02; P=0.90; Table 4). In a multivariate model including ETP, eGFR, sex, age, ACR, baseline CKD, and presence of diabetes, ETP (OR=1.10; 95% CI, 1.03 to 1.17; P<0.01) was independently associated with kidney disease progression in the AKI group (Table 4). No association of ETP or eGFR with kidney disease progression was observed in the control group (Table 4).
Table 4. -
Logistic regression for kidney disease progression from Y1 to Y3
| Y1 Variables |
AKI Group |
Control Group |
| Odds Ratio (95% Confidence Interval) |
P Value |
Odds Ratio (95% Confidence Interval) |
P Value |
|
Univariate models
|
| Sex, F |
0.91 (0.39 to 2.03) |
0.81 |
1.38 (0.52 to 3.88) |
0.52 |
| Age |
1.05 (1.00 to 1.11) |
<0.05a |
1.04 (0.98 to 1.11) |
0.16 |
| ACR |
1.01 (1.00 to 1.02) |
<0.05a |
1.03 (1.01 to 1.05) |
<0.001a |
| BL DM, Yes |
1.78 (0.70 to 4.17) |
0.21 |
2.82 (0.99 to 7.61) |
0.05 |
| BL CKD, Yes |
2.35 (1.03 to 5.31) |
<0.05a |
3.09 (1.15 to 8.45) |
<0.05a |
| eGFR |
1.00 (0.98 to 1.02) |
0.90 |
0.98 (0.95 to 1.00) |
0.06 |
| ETP |
1.08 (1.03 to 1.14) |
<0.01a |
1.04 (0.96 to 1.12) |
0.29 |
|
Multivariate full model
|
| Sex, F |
1.02 (0.42 to 2.44) |
0.96 |
1.93 (0.65 to 6.43) |
0.24 |
| Age |
1.07 (1.02 to 1.14) |
<0.01a |
1.03 (0.97 to 1.11) |
0.32 |
| ACR |
1.01 (1.00 to 1.02) |
0.05 |
1.02 (1.01 to 1.05) |
<0.001a |
| BL DM, Yes |
1.12 (0.36 to 3.08) |
0.83 |
1.88 (0.56 to 5.74) |
0.29 |
| BL CKD, Yes |
3.96 (1.21 to 13.68) |
<0.05a |
1.59 (0.34 to 7.09) |
0.55 |
| eGFR |
1.05 (1.02 to 1.09) |
<0.01a |
1.00 (0.96 to 1.05) |
0.87 |
| ETP |
1.10 (1.03 to 1.17) |
<0.01a |
1.02 (0.92 to 1.12) |
0.65 |
|
Multivariate model with BE
|
| Sex, F |
Not retained |
Not retained |
| Age |
1.07 (1.02 to 1.14) |
<0.01a |
Not retained |
| ACR |
1.01 (1.00 to 1.02) |
<0.05a |
1.03 (1.01 to 1.06) |
<0.001a |
| BL DM, Yes |
Not retained |
Not retained |
| BL CKD, Yes |
4.01 (1.24 to 13.67) |
<0.05a |
Not retained |
| eGFR |
1.05 (1.02 to 1.09) |
<0.01a |
Not retained |
| ETP |
1.10 (1.03 to 1.17) |
<0.01a |
Not retained |
Kidney disease progression was defined as ≥25% decline in eGFR and a decline in CKD stage. Number of events: 26 of 337 patients in the AKI group and 17 of 369 patients in the control group. Y1, year 1; Y3, year 3; F, female; ACR, albuminuria-creatinine ratio; BL, baseline; DM, diabetes mellitus; ETP, endotrophin; BE, backward elimination.
aP<0.05.
Association of ETP with Mortality in AKI
In the AKI group, 43 of 393 patients (11%) died between year 1 and year 3, whereas 27 of 408 patients (7%) died in the control group (chi-squared test, P=0.03). ETP levels measured at year 1 were significantly lower in the patients with AKI that were still alive at year 3 (median [IQR]=10.68 [8.31–14.49] ng/ml) compared with patients with AKI who died (median [IQR]=12.66 [9.58–22.01] ng/ml; P<0.01; Figure 3A). There was no significant difference in ETP levels between survivors (median [IQR]=9.21 [6.91–12.26] ng/ml) and nonsurvivors (median [IQR]=9.97 [7.24–13.25] ng/ml; P=0.23) in the control group (Figure 3A).
Figure 3.: Endogropin (ETP) levels in survivors and non-survivors in the AKI group and in the control group. ETP levels measured at year 1 in (A) survivors versus nonsurvivors in the AKI and in the control groups, and (B) survival curves for tertiles of ETP measured at year 1 in patients with AKI and (C) controls. (A) ETP data are presented on a log10-scale as median with 95% CI. Statistical differences between groups were assessed by Mann–Whitney tests. The dotted grid line indicates y values of 10. Survival data are on the basis of tertiles of ETP (B) in the AKI group and (C) in the control group. Statistical differences between curves were assessed by Mantel–Cox test. **P<0.01. T1, tertile 1; ns, not significant.
On the basis of the ROC curve analysis, ETP measured at year 1 could discriminate survivors from nonsurvivors in the AKI group (AUC=0.64; P<0.01) but not in the control group (AUC=0.57; P=0.24), whereas eGFR could discriminate survivors in both the AKI (AUC=0.64; P<0.01) and control groups (AUC=0.66; P<0.01; Table 3). Patients were stratified into tertiles on the basis of ETP levels, and patients in the highest tertile were more likely to die in the AKI group (tertile 1 (T1), [3.92-9.32] ng/mL; T2, (9.32-12.80] ng/mL; T3, (12.80-84.9] ng/mL; P<0.05; Figure 3B), but not in the control group (T1, [1.92-7.70] ng/mL; T2, (7.70-10.8] ng/mL; T3, (10.80-41.7] ng/mL; P=0.22; Figure 3C).
The association of ETP with mortality was also investigated using Cox proportional hazards regression. To assess the adjusted association of variables of interest with mortality in the AKI group, we included ETP, eGFR, sex, age, urinary ACR, baseline CKD, and presence of diabetes in a full model. In the full model, ETP (hazard ratio [HR]=1.05; 95% CI, 1.02 to 1.07; P<0.001) was independently associated with mortality, whereas eGFR (HR=0.98; 95% CI, 0.96 to 1.00; P=0.06) was not (Table 5). In the full AKI model with backward elimination, ETP (HR=1.05; 95% CI, 1.03 to 1.07; P<0.001) and age (HR=1.06; 95% CI, 1.02 to 1.11; P<0.01) were the only variables retained in the model. In the control group, ETP was not associated with mortality in any of the models (P=0.43–0.83; Table 5).
Table 5. -
Cox proportional hazards regression for mortality from Y1 to Y3
| Y1 Variables |
AKI Group |
Control Group |
| Hazard Ratio (95% Confidence Interval) |
P Value |
Hazard Ratio (95% Confidence Interval) |
P Value |
|
Univariate models
|
| Sex, F |
1.02 (0.56 to 1.86) |
0.96 |
0.42 (0.19 to 0.97) |
<0.05a |
| Age |
1.07 (1.03 to 1.12) |
<0.001a |
1.06 (1.01 to 1.11) |
<0.05a |
| ACR |
1.00 (1.00 to 1.01) |
0.72 |
1.00 (0.99 to 1.01) |
0.62 |
| BL DM, Yes |
1.40 (0.72 to 2.73) |
0.32 |
1.77 (0.80 to 3.95) |
0.16 |
| BL CKD, Yes |
1.19 (0.63 to 2.25) |
0.59 |
1.71 (0.80 to 3.69) |
0.17 |
| eGFR |
0.98 (0.96 to 0.99) |
<0.01a |
0.98 (0.96 to 0.99) |
<0.01a |
| ETP |
1.05 (1.03 to 1.07) |
<0.001a |
1.03 (0.96 to 1.09) |
0.43 |
|
Multivariate full model
|
| Sex, F |
1.06 (0.58 to 1.95) |
0.84 |
0.47 (0.20 to 1.08) |
0.08 |
| Age |
1.06 (1.01 to 1.10) |
<0.05a |
1.04 (0.98 to 1.09) |
0.20 |
| ACR |
1.00 (0.99 to 1.00) |
0.46 |
1.00 (0.99 to 1.01) |
0.96 |
| BL DM, Yes |
1.19 (0.58 to 2.45) |
0.63 |
1.30 (0.55 to 3.07) |
0.54 |
| BL CKD, Yes |
0.44 (0.19 to 1.01) |
0.05 |
0.62 (0.22 to 1.78) |
0.38 |
| eGFR |
0.98 (0.96 to 1.00) |
0.06 |
0.97 (0.95 to 1.00) |
0.08 |
| ETP |
1.05 (1.02 to 1.07) |
<0.001a |
0.99 (0.92 to 1.07) |
0.83 |
|
Multivariate model with BE
|
| Sex, F |
Not retained |
Not retained |
| Age |
1.06 (1.02 to 1.11) |
<0.01a |
Not retained |
| ACR |
Not retained |
Not retained |
| BL DM, Yes |
Not retained |
Not retained |
| BL CKD, Yes |
Not retained |
Not retained |
| eGFR |
Not retained |
0.98 (0.96 to 0.99) |
<0.01a |
| ETP |
1.05 (1.03 to 1.07) |
<0.001a |
Not retained |
Number of events: 43 of 393 patients in the AKI group and 27 of 408 patients in the control group. Y1, year 1; Y3, year 3; HR, hazard ratio; F, female; ACR, albuminuria-creatinine ratio; BL, baseline; DM, diabetes mellitus; ETP, endotrophin; BE, backward elimination.
aP<0.05.
Post Hoc Analysis 3 Months after AKI
Although the primary analysis focused on ETP levels at 1 year after AKI, we also performed post hoc analysis in 585 patients who had ETP levels available at month 3 (280 in the AKI group and 305 in the control group) (see Supplemental Tables 1–4). Baseline characteristics were not different from the main study population (see Supplemental Table 1). ETP levels were higher in the AKI group at month 3 (median [IQR]=13.40 [9.68–18.13] ng/ml) compared with the control group (median [IQR]=9.90 [7.70–12.60] ng/ml; P<0.001). At year 1, a similar difference between the AKI and control groups was observed, although ETP levels at year 1 had fallen slightly in both groups (median [IQR]=11.14 [8.39–15.449 ng/ml and 8.91 [6.90–11.88] ng/ml, respectively; P<0.001). These levels and differences between groups were very similar to the year 1 results in the main study population. ETP measured 3 months after the AKI event was prognostic for kidney disease progression and mortality at year 3 (AUC=0.61 [P<0.05] and AUC=0.68 [P<0.001], respectively; Supplemental Table 2). In a multivariate model, ETP measured at 3 months was retained in the model for mortality in the AKI group (HR=1.04; 95% CI, 1.02 to 1.06; P<0.001; Table 4).
Discussion
The balance between adaptive and maladaptive repair is essential for the AKI to CKD transition. In preclinical models, it has been shown that fibrosis is a key process from AKI to CKD (20), but this is more difficult to study in humans, highlighting the importance of using a marker after AKI that directly reflects structural changes that take place during fibrogenesis. Although there are many biomarkers of early injury in AKI, only a few have been evaluated in terms of long-term outcomes (22). ETP, measured using the PRO-C6 ELISA, is a structural damage biomarker measuring the changes that take place during fibrogenesis in the kidney. In this study, ETP was measured 1 year after an episode of AKI, because we aimed to assess sustained fibrotic processes at later time points after the acute event to achieve clear separation from the effects of the acute episode. The main findings of this study were as follows: (1) in patients with AKI, ETP can predict long-term mortality and has a stronger association with mortality than eGFR; and (2) in patients with AKI, ETP can predict kidney disease progression at later time points and is a superior prognostic marker than eGFR.
On the basis of ROC curve analysis, ETP measured at year 1 could discriminate survivors from nonsurvivors at year 3 in the AKI group but not in the control group. In comparison, eGFR could discriminate survivors from nonsurvivors in both the AKI and the control group. This could indicate that ETP release is triggered by processes that follow the episode of AKI, whereas a change in eGFR might be the result of a range of processes. This is also supported by the Cox proportional hazards analysis for mortality, where ETP and age were retained in the final model in the AKI group, whereas only eGFR was retained in the final model in the control group. Finally, only ETP (and not eGFR) could discriminate patients with kidney disease progression at year 1 in the AKI group, whereas neither of the biomarkers could discriminate patients with disease progression in the control group, which may also reflect different pathologic processes in the two groups of patients and the lower number of events in the controls. A number of studies have shown the potential of the collagen type VI–derived signaling molecule ETP as a molecule triggering proinflammatory and profibrotic processes (232425–26). Therefore, the difference in disease progression between the AKI and the control groups may, in part, be explained by the ability of ETP to trigger disease progression itself.
Although the primary analysis focused on ETP levels 1 year after AKI, we also observed the prognostic potential of ETP in a subgroup of 585 patients with ETP levels available at month 3. The results largely confirmed the prognostic potential observed for ETP at year 1, both for kidney disease progression and mortality. Because the measurements were only performed in a subset of patients at month 3, we consider the results at year 1 more reliable.
Although a number of biomarkers monitoring patients subsequent to AKI have recently been evaluated, progress in this area is urgently required. As described in a 2020 recommendation on AKI biomarkers based on the literature and expert consensus for practicing clinicians and researchers (2), it is recommended to merge biomarkers predictive of CKD staging and progression into post-AKI assessment. ETP, measured by PRO-C6, is a promising and novel post-AKI biomarker that needs to be validated in larger AKI cohorts, ideally with kidney biopsy specimens available to confirm the biomarker directly reflects fibrosis (although this may be challenging to achieve). However, this has already been confirmed in patients with IgA nephropathy and kidney transplant recipients (14,15). It is also possible that there may be utility in identifying patients across the spectrum of fibrosis that have a rapidly progressing disease.
The limitations associated with this study are that circulating biomarkers may be affected by factors outside of the kidney, the single-center design of this study limits generalizability, and we are unable to provide causal evidence for the association of ETP with the outcomes. Also, it has been shown that ETP levels can be affected by cofounding factors not available in this study, such as body mass index, measures of body composition, and data on mobility/immobilization (12,13). A major strength is that this was a large study with prospective data collection and few missing data. In addition, the study includes a large, well-matched control cohort of patients enrolled from the same site as the AKI group.
To conclude, ETP levels measuring the bioactive molecule ETP predict kidney disease progression and mortality at 1 year after AKI. Because ETP is a profibrotic molecule, we suggest that ETP identifies patients with active fibrogenesis after AKI, suggestive of long-term renal remodeling, which is associated with patient outcomes. Future work to establish whether ETP could be used as a novel post-AKI biomarker is warranted.
Disclosures
F. Genovese, M.A. Karsdal, and D.G.K. Rasmussen report being full-time employees at, and holding stock in, Nordic Bioscience. N. Sparding reports being a full-time employee at Nordic Bioscience and the University of Copenhagen. All remaining authors have nothing to disclose.
Funding
None.
Acknowledgments
We would like to thank the patients participating in the ARID study and Giti Akbatani for technical support.
The results presented in this article have not been published previously in whole or in part, except in abstract format.
Nordic Bioscience is a privately owned, small to medium–sized enterprise partly focused on the development of biomarkers. None of the authors received fees, bonuses, or other benefits for the work described in this article. The funder provided support in the form of salaries for authors N. Sparding, F. Genovese, D.G.K. Rasmussen, and M.A. Karsdal, but did not have any additional role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.
The patent for the PRO-C6 ELISA used in this work is owned by Nordic Bioscience.
Author Contributions
B. Feldt-Rasmussen, F. Genovese, M. Hornum, and D.G.K. Rasmussen provided supervision; F. Genovese was responsible for resources; F. Genovese, R. Packington, N.M. Selby, and N. Sparding conceptualized the study; R. Packington, D.G.K. Rasmussen, N.M. Selby, and N. Sparding were responsible for investigation; R. Packington and N.M. Selby were responsible for project administration; R. Packington, N.M. Selby, and N. Sparding were responsible for data curation, formal analysis, and methodology; N.M. Selby and N. Sparding wrote the original draft; N. Sparding was responsible for visualization; and all authors reviewed and edited the manuscript.
Supplemental Material
This article contains the following supplemental material online at http://kidney360.asnjournals.org/lookup/suppl/doi:10.34067/KID.0000422021/-/DCSupplemental.
Supplemental Table 1. Patient characteristics at M3.
Supplemental Table 2. ROC curve analysis for kidney disease progression and mortality based on M3 variables.
Supplemental Table 3. Logistic regression for kidney disease progression based on M3 variables.
Supplemental Table 4. Cox proportional hazards regression for mortality based on M3 variables.
References
1. Lewington AJP, Cerdá J, Mehta RL: Raising awareness of acute kidney injury: A global perspective of a silent killer. Kidney Int 84: 457–467, 2013
https://doi.org/10.1038/ki.2013.153
2. Ostermann M, Zarbock A, Goldstein S, Kashani K, Macedo E, Murugan R, Bell M, Forni L, Guzzi L, Joannidis M, Kane-Gill SL, Legrand M, Mehta R, Murray PT, Pickkers P, Plebani M, Prowle J, Ricci Z, Rimmelé T, Rosner M, Shaw AD, Kellum JA, Ronco C: Recommendations on acute kidney injury biomarkers from the Acute Disease Quality Initiative Consensus Conference: A consensus statement. JAMA Netw Open 3: e2019209, 2020
https://doi.org/10.1001/jamanetworkopen.2020.19209
3. Horne KL, Packington R, Monaghan J, Reilly T, McIntyre CW, Selby NM: The effects of acute kidney injury on long-term renal function and proteinuria in a general hospitalised population. Nephron Clin Pract 128: 192–200, 2014
https://doi.org/10.1159/000368243
4. National Kidney Foundation. KDIGO clinical practice guideline for acute kidney injury (AKI). Available at
https://kdigo.org/guidelines/acute-kidney-injury/. Accessed July 1, 2021
5. Devarajan P: The use of targeted biomarkers for chronic kidney disease. Adv Chronic Kidney Dis 17: 469–479, 2010
https://doi.org/10.1053/j.ackd.2010.09.002
6. Oh DJ: A long journey for acute kidney injury biomarkers. Ren Fail 42: 154–165, 2020
https://doi.org/10.1080/0886022X.2020.1721300
7. Horne KL, Shardlow A, Taal MW, Selby NM: Long term outcomes after acute kidney injury: Lessons from the ARID Study. Nephron 131: 102–106, 2015
https://doi.org/10.1159/000439066
8. Katsuda S, Okada Y, Minamoto T, Oda Y, Matsui Y, Nakanishi I: Collagens in human atherosclerosis. Immunohistochemical analysis using collagen type-specific antibodies. Arterioscler Thromb 12: 494–502, 1992
https://doi.org/10.1161/01.ATV.12.4.494
9. Groulx JF, Gagné D, Benoit YD, Martel D, Basora N, Beaulieu JF: Collagen VI is a basement membrane component that regulates epithelial cell-fibronectin interactions. Matrix Biol 30: 195–206, 2011
https://doi.org/10.1016/j.matbio.2011.03.002
10. Aigner T, Hambach L, Söder S, Schlötzer-Schrehardt U, Pöschl E: The C5 domain of Col6A3 is cleaved off from the Col6 fibrils immediately after secretion. Biochem Biophys Res Commun 290: 743–748, 2002
https://doi.org/10.1006/bbrc.2001.6227
11. Heumüller SE, Talantikite M, Napoli M, Armengaud J, Mörgelin M, Hartmann U, Sengle G, Paulsson M, Moali C, Wagener R: C-terminal proteolysis of the collagen VI α3 chain by BMP-1 and proprotein convertase(s) releases endotrophin in fragments of different sizes. J Biol Chem 294: 13769–13780, 2019
https://doi.org/10.1074/jbc.RA119.008641
12. Fenton A, Jesky MD, Ferro CJ, Sørensen J, Karsdal MA, Cockwell P, Genovese F: Serum endotrophin, a type VI collagen cleavage product, is associated with increased mortality in chronic kidney disease. PLoS One 12: e0175200, 2017
https://doi.org/10.1371/journal.pone.0175200
13. Sun S, Henriksen K, Karsdal MA, Byrjalsen I, Rittweger J, Armbrecht G, Belavy DL, Felsenberg D, Nedergaard AF: Collagen type III and VI turnover in response to long-term immobilization. PLoS One 10: e0144525, 2015
https://doi.org/10.1371/journal.pone.0144525
14. Yepes-Calderón M, Sotomayor CG, Rasmussen DGK, Hijmans RS, Te Velde-Keyzer CA, van Londen M, van Dijk M, Diepstra A, Berger SP, Karsdal MA, Bemelman FJ, de Fijter JW, Kers J, Florquin S, Genovese F, Bakker SJL, Sanders JS, Van Den Born J: Biopsy-controlled non-invasive quantification of collagen type VI in kidney transplant recipients: A post-hoc analysis of the MECANO Trial. J Clin Med 9: 3216, 2020
https://doi.org/10.3390/jcm9103216
15. Sparding N, Genovese F, Rasmussen DGK, Karsdal MA, Neprasova M, Maixnerova D, Satrapova V, Frausova D, Hornum M, Bartonova L, Honsova E, Kollar M, Koprivova H, Hruskova Z, Tesar V: Endotrophin, a collagen type VI-derived matrikine, reflects the degree of renal fibrosis in patients with IgA nephropathy and in patients with ANCA-associated vasculitis [published online ahead of print April 29, 2021]. Nephrol Dial Transplant
10.1093/ndt/gfab163
16. Rasmussen DGK, Fenton A, Jesky M, Ferro C, Boor P, Tepel M, Karsdal MA, Genovese F, Cockwell P: Urinary endotrophin predicts disease progression in patients with chronic kidney disease. Sci Rep 7: 17328, 2017
https://doi.org/10.1038/s41598-017-17470-3
17. Rasmussen DGK, Hansen TW, von Scholten BJ, Nielsen SH, Reinhard H, Parving HH, Tepel M, Karsdal MA, Jacobsen PK, Genovese F, Rossing P: Higher collagen VI formation is associated with all-cause mortality in patients with type 2 diabetes and microalbuminuria. Diabetes Care 41: 1493–1500, 2018
https://doi.org/10.2337/dc17-2392
18. Pilemann-Lyberg S, Rasmussen DGK, Hansen TW, Tofte N, Winther SA, Holm Nielsen S, Theilade S, Karsdal MA, Genovese F, Rossing P: Markers of collagen formation and degradation reflect renal function and predict adverse outcomes in patients with type 1 diabetes. Diabetes Care 42: 1760–1768, 2019
https://doi.org/10.2337/dc18-2599
19. Horne KL, Packington R, Monaghan J, Reilly T, Selby NM: Three-year outcomes after acute kidney injury: Results of a prospective parallel group cohort study. BMJ Open 7: e015316, 2017
https://doi.org/10.1136/bmjopen-2016-015316
20. Basile DP, Bonventre JV, Mehta R, Nangaku M, Unwin R, Rosner MH, Kellum JA, Ronco C; ADQI XIII Work Group: Progression after AKI: Understanding maladaptive repair processes to predict and identify therapeutic treatments. J Am Soc Nephrol 27: 687–697, 2016
https://doi.org/10.1681/ASN.2015030309
21. National Kidney Foundation. KDIGO clinical practice guideline for for the evaluation and management of chronic kidney disease (CKD). Available at
https://kdigo.org/guidelines/acute-kidney-injury/https://kdigo.org/guidelines/ckd-evaluation-and-management/. Accessed July 1, 2021
22. Wilson M, Packington R, Sewell H, Bartle R, McCole E, Kurth MJ, Richardson C, Shaw S, Akani A, Banks RE, Selby NM: Biomarkers during recovery from AKI and prediction of long-term reductions in estimated GFR [published online ahead of print October 12, 2021]. Am J Kidney Dis
10.1053/j.ajkd.2021.08.017
23. Park J, Scherer PE: Adipocyte-derived endotrophin promotes malignant tumor progression. J Clin Invest 122: 4243–4256, 2012
https://doi.org/10.1172/JCI63930
24. Karousou E, D’Angelo ML, Kouvidi K, Vigetti D, Viola M, Nikitovic D, De Luca G, Passi A: Collagen VI and hyaluronan: The common role in breast cancer. BioMed Res Int 2014: 606458, 2014
https://doi.org/10.1155/2014/606458
25. Zhao Y, Gu X, Zhang N, Kolonin MG, An Z, Sun K: Divergent functions of endotrophin on different cell populations in adipose tissue. Am J Physiol Endocrinol Metab 311: E952–E963, 2016
https://doi.org/10.1152/ajpendo.00314.2016
26. Funcke JB, Scherer PE: Beyond adiponectin and leptin: Adipose tissue-derived mediators of inter-organ communication. J Lipid Res 60: 1648–1684, 2019
https://doi.org/10.1194/jlr.R094060
