Supplemental Methods:
Study Design and Enrollment:
The study was conducted at the Level I Trauma Center of the University of Pennsylvania between September 30th, 2017 and October 24th, 2018. Patients identified by the clinical team who required emergent operation or admission to the ICU were enrolled and an initial blood sample was obtained in the ED. Of patients identified in the trauma bay, those who were either transferred to a non-ICU hospital unit within 6 hours or who expired within 12 hours were subsequently excluded. We obtained consent from patients or their representative at the earliest feasible opportunity. If the patient or representative declined participation, we offered the possibility of limiting involvement to samples already collected and clinical data from the medical record. If participation was declined all collected data and samples were discarded. Samples were centrifuged immediately after collection and plasma was placed into cryovial aliquots and frozen at -80C until DAMP measurement was performed at a later date.
DAMP quantification
We quantified nDNA and mtDNA in triplicate, as reported previously, using log-transformed copy number per microliter (uL) of plasma (copy number/uL) to ensure normality, using the following primer sequences: MT- ND1 (forward, 5′-ATACCCATGGCCAACCTCCT-3′ and reverse, 5′- GGGCCTTTGCGTAGTTGTAT-3′), and COXIV (forward, 5’- GAAAGTGTTGTGAAGAGCGAAGAC-3’ and reverse, 5’- GTGGTCACGCCGATCCAT-
‘3)(1). PCR standards for COXIV and NADH were amplified from DNA extracted from endothelial cell lysates and gel purified using QiaEX II Gel Extraction Kit (Qiagen). Copy number was derived from cycle
threshold using an online DNA copy number calculator. nDNA and mtDNA quantification demonstrated high reproducibility, with coefficients of variation <5% for all samples. We quantified plasma nucleosomes in duplicate using the Cell Death Detection PLUS ELISA (Roche, Basel, Switzerland). Nucleosome quantities are reported in arbitrary units (AU), consistent with previous investigations of nucleosomes in critical illness (2).
Coefficient of variation for nucleosome replicates was higher, but 89% of sample replicates had <15%
coefficient of variation.
Defining AKI:
We did not use urine output to define AKI because urinary catheter use was variable and hourly urine output data were unreliable. Use of Kidney Disease Improving Global Outcomes (KDIGO) criteria to define AKI was performed as a sensitivity analysis.
Supplemental Results:
Plasma samples were collected at all 5 time points in 39 patients (71%), at 4 time points in 10 (18%), and at 3 time points in 6 (11%). All patients had plasma obtained at presentation, and 54 of 55 patients (98%) had 6- hour plasma samples collected (Supplemental Tables 3 and 4, http://links.lww.com/CCX/A952). Inclusion of only patients with complete samples for all time points using complete case analysis did not change the associations of each nucleic acid DAMP with AKI (Supplemental Figure 3, http://links.lww.com/CCX/
A966).
Correlation of DAMP Concentrations
mtDNA, nDNA, and nucleosome levels showed weak to moderate positive correlations at each time point (Supplemental Table 1, http://links.lww.com/CCX/A952). Correlations between nDNA and nucleosomes were significant at all time points. The correlations between mtDNA and nDNA were significant only at 48 hours, while mtDNA and nucleosome concentrations were moderately correlated at presentation, 24 and 48 hours.
ISS was correlated with all nDNA concentrations,but not with mtDNA or nucleosome levels (Supplemental Table 2, http://links.lww.com/CCX/A952).
Association of Presentation nDNA with Transfusion.
No RBC units were transfused prior to presentation. Presentation nDNA was weakly but significantly correlated with RBC units transfused by 6 hours (rho=0.30, p=0.02) and 12 hours (rho=0.31, p=0.021). Transfusion of plasma and platelets were recorded and are highly collinear with RBC transfusion given 1:1:1 resuscitation protocols (6 hour RBC and FFP units transfused: rho =0.89, p<0.001, 6 hour RBC and platelet units
transfused: rho=0.75, p<0.001), making inclusion in logistic regression or linear mixed effects models in addition to RBC units transfused duplicative.
Association of Presentation nDNA with AKI Stage and RRT
Due to the small size of the study, we stratified stages of AKI into KDIGO stage 1 or KDIGO stage 2+3.
Presentation nDNA was associated with AKI stage, with an odds ratio of 2.26 (95% CI 1.35-3.77) per log
copies/uL to progress from no AKI to stage 1 or stage 1 to stage 2/3. Four patients required renal replacement therapy, starting at 32, 34, 35, and 73 hours after presentation. Presentation nDNA was associated with need for RRT but did not meet statistical significance due to the low number of patients who required RRT (4 patients, OR for log copies/uL nDNA at presentation 2.35, 95% CI 0.90-6.14, p=0.08).
Supplemental Figure Legend:
Supplemental Figure 1: Kaplan-Meier curve demonstrating cumulative incidence of AKI over the study period. Most AKI occurred within 24 hours of presentation. Abbreviations: AKI=acute kidney injury.
Supplemental Figure 2: Receiver operator characteristics (ROC) curves for the associations of ISS and presentation plasma nDNA concentration with subsequent AKI. ROC curves were generated from logistic regression models using ISS alone (AUC 0.62, 95% CI 0.45-0.78 ,Hosmer-Lemeshow goodness-of-fit χ2=6.35, p=0.608) presentation nDNA level alone (AUC 0.80, 95% CI 0.69-0.92, Hosmer-Lemeshow goodness-of-fit χ2=7.63, p=0.47), and both ISS and nDNA level (AUC 0.81, 95% CI 0.69-0.92, Hosmer-Lemeshow χ2=6.23, p=0.622) as explanatory variables for the outcome of AKI. ISS=injury severity score. nDNA=nuclear DNA.
AKI=acute kidney injury.
Supplemental Figure 3: Associations of mtDNA, nDNA and nucleosomes using complete case analysis. Association of mtDNA (panel A), nDNA (panel B), and nucleosomes (panel C) with AKI in only patients with available samples at all 5 time points (n=38 for mtDNA, n=39 for nDNA, and n=38 for nucleosomes), complete case analysis, CCA).
A. mtDNA concentration differences by AKI status demonstrated little early difference (𝛽 at presentation 0.27 (- 0.51-1.05), p=0.497, 6h: 0.46 (-0.35-1.27), p=0.267), 0.72 (-0.11-1.54), p=0.090 (12 hours), but diverged at 24 and 48 hours with subsequent higher levels in AKI patients (𝛽 at 24h 1.40 (0.56-2.24), p=0.001 and 48h: 1.21 (0.36-2.06), p=0.005).
B. The association of nDNA concentration with AKI showed no significant difference over time (interaction p=0.311), with nDNA levels consistently higher in those with AKI (overall 𝛽 = 1.08 (0.55-1.61), p<0.001. 0.29 (- 0.04-0.61), p=0.081.).
C. Nucleosome levels were marginally higher in AKI patients at early timepoints (𝛽 at presentation: 𝛽 = 0.29 (- 0.04-0.61), p=0.081, 6h: 0.26 (-0.08-0.60), p=0.130 and 12h: 0.27 (-0.08-0.61), p=0.205, but more markedly different at 24h: 0.37 (0.02-0.73), p=0.037 and 48 hours: 0.73 (0.38-1.09), p<0.001. No significant difference between primary analysis and CCA was noted. Abbreviations: mtDNA=mitochondrial DNA, nDNA=nuclear DNA.
