Visual Abstract
Keywords: acute kidney failure, AKI, risk factors, survival
Abstract
Key Points
Severe AKI is associated with reduced hospital survival, regardless of whether it occurs before or during extracorporeal membrane oxygenation (ECMO).
AKI onset during veno-venous ECMO is less because of ECMO-related factors than to recurrent septic episodes.
Background
AKI is a frequent concomitant organ failure during veno-venous extracorporeal membrane oxygenation (VV-ECMO). This study investigated the prevalence and the impact of AKI on survival to hospital discharge and up to 1 year after discharge and risk factors for developing AKI during VV-ECMO.
Methods
This is an observational retrospective study of 500 consecutive patients receiving VV-ECMO between November 2014 and December 2021. Patients were divided into three groups: (1) AKI onset before extracorporeal membrane oxygenation (ECMO), (2) AKI onset during ECMO, and (3) AKI onset before and new onset during ECMO. The Kidney Disease Improving Global Outcomes definition was used to define AKI. Follow-up was 1 year after hospital discharge. Propensity score matching was performed for patients without AKI and patients with AKI onset during ECMO to analyze risk factors for AKI onset during VV-ECMO.
Results
A total of 320 patients (64.0%) had AKI: 182 (36.4%) with onset before ECMO and 158 (31.6%) with onset during ECMO. At ECMO initiation, patients with AKI onset before VV-ECMO presented significantly higher inflammatory markers and higher NE dosage, whereas patients developing AKI during VV-ECMO did not differ from those without AKI. Survival to hospital discharge was 67.0% (AKI, 60.9%; no AKI, 77.8%; P < 0.001). Cox regression analysis revealed AKI Kidney Disease Improving Global Outcomes stage 3, independent from onset, as an independent risk factor for reduced survival to hospital discharge (hazard ratio, 2.15; 95% confidence interval, 1.37 to 3.37; P = 0.001). During follow-up, the survival was 92.5%; age was shown as the sole risk factor for reduced survival in hospital survivors in the multivariate logistic regression model. In the propensity score–matched cohort (41 patients in each group), the AKI group had lower mean arterial pressure and significantly higher C-reactive protein levels the days before AKI. Factors associated with VV-ECMO support (blood flow, cell-free hemoglobin) did not differ.
Conclusions
Severe AKI is associated with reduced hospital survival, regardless of whether it occurs before or during ECMO. AKI onset during VV-ECMO is less because of ECMO-related factors than to recurrent septic episodes.
Introduction
If conventional treatment options for acute respiratory distress syndrome fail, veno-venous extracorporeal membrane oxygenation (VV-ECMO) may be used to support patients as a bridging strategy.1 One of the most frequent complications during VV-ECMO is AKI.1 Limited data showed rates ranging from 57% to 70% in extracorporeal membrane oxygenation (ECMO) patients.2,3
Causes may be the underlying disease, sepsis with multiorgan failure, and preexisting renal dysfunction, reflected by higher intensive care unit scores, indicating the severity of the disease, in addition to inflammatory markers.4–6 Furthermore, ventilator settings and blood gases are controversially discussed to contribute to the development of AKI6–11 as is the potential contribution of ECMO treatment.12
Other retrospective but smaller studies reported the need of KRT in up to 55.8% of the patients on VV-ECMO.13 After AKI, up to 33.3% were found to have CKD,14 with a lower 1-year survival rate in patients with AKI during ECMO.14,15
However, questions concerning the risk of AKI in patients with acute respiratory failure on VV-ECMO, the risk associated with mechanical ventilation and hemodynamic changes, and risk factors and time course of AKI during ECMO treatment remain unanswered.4,16
Therefore, this observational retrospective single-center study examined the prevalence, survival, and risk factors for developing AKI in a large cohort of patients suffering from acute respiratory failure supported with VV-ECMO.
Methods
Patients
This is an observational retrospective study of 500 consecutive patients receiving VV-ECMO support at the University Hospital Regensburg (UKR) between November 2014 and December 2021 to examine the prevalence, survival, and risk factors regarding the development of AKI. The UKR is a tertiary referral hospital providing up to 200 ECMO runs a year.
Patients were excluded if treated with chronic hemodialysis before ECMO therapy or age younger than 18 years. Veno-arterial ECMO patients were excluded as were all patients with change in configuration to veno-arterial ECMO or hybrid forms.
ECMO Management
The UKR follows ECMO standard management. More detailed information and information about ECMO systems and cannulas are provided in Supplemental Figure 1, Supplemental Method 1, and Supplemental Tables 1 and 2. The criteria for indication of ECMO changed during the coronavirus disease 2019 pandemic; the criteria are shown in Supplemental Method 1 and the effects in Supplemental Table 3.
Definitions
AKI was defined according to the Kidney Disease Improving Global Outcomes (KDIGO) guidelines17 (Supplemental Table 4).
Survival was defined as survival to hospital discharge. Follow-up was defined as 1 year after hospital discharge. A new episode of AKI was defined as rise of creatinine (according to the KDIGO definition) after complete resolution of serum creatinine/kidney function after an episode of AKI.
According to the information above, we defined three subgroups of patients regarding the prevalence of AKI: (1) AKI onset before ECMO (AKI with onset before ECMO could also persist during ECMO); (2) AKI onset during ECMO; and (3) AKI onset before ECMO and new onset of AKI during ECMO after an initial recovery. For the analysis of risk factors for AKI, we excluded the group of patients who had AKI onset before and new onset of AKI during ECMO to allow a clear attribution.
The criteria for KRT initiation are provided in Supplemental Method 1.18 Dependency from chronic-KRT was defined as requirement for KRT at least 3 days per week and described chronic hemodialysis in patients with KDIGO CKD G5D. Sepsis was defined as an excessive immune response to an infectious stimulus leading to organ dysfunction.19 The latter was measured by the Sequential Organ Failure Assessment (SOFA) score at ECMO initiation (Horowitz index, a rise in creatinine, bilirubin, high catecholamine doses or low mean arterial pressure [MAP], low platelets, and a low Glasgow Coma Scale score).
Data Collection
Data were obtained from routine medical history recording and the patient data management system (PDMS). Owing to a change in electronic PDMS (MetaVision) in January 2018, detailed hemodynamic information during ECMO therapy was available only for 310 of 500 patients (baseline characteristics in Supplemental Table 5). Follow-up was performed for all patients who had information about their vital status from ECMO team checks (n=335). Creatinine levels before ECMO initiation were estimated to be normal, if not stated as chronic renal insufficiency in medical reports to categorize KDIGO stages before ECMO therapy. More information about data collection is provided in Supplement Method 1.
Statistical Analysis
Statistical analysis was performed with IBM SPSS 26.0 (Chicago, IL). Medians and percentiles (25th–75th) were computed for continuous variables. Categorical variables were described as number (n) and percentage (%). Mann–Whitney U test, chi-squared test, and Wilcoxon Test were used as indicated. The survival probability is calculated using the Kaplan–Meier method. P values below 0.05 were considered statistically significant. Cox regression analysis used only variables that were significant in univariate analyses.
Propensity Score Matching
To investigate risk factors for AKI onset during VV-ECMO, propensity score matching was performed in categories of AKI onset during ECMO and no AKI during ECMO. Eligible for propensity score matching were the 310 patients recorded in the new electronic PDMS. To exclude influences occurring before ECMO, the matched cohorts included only patients without AKI and patients with AKI onset during ECMO diagnosed 3 days or later after initiation of ECMO support (n=82). The 3-day cutoff was chosen to account for the 2-day delay in creatinine elevation, allowing to focus on risk factors for AKI that arise exclusively during ECMO therapy. Variables for matching were age, sex, body mass index (BMI), diabetes, arterial hypertension, chronic renal failure, creatinine pre ECMO, SOFA score without kidney, duration of ECMO treatment, and ECMO indications. Propensity score matching characteristics and propensity score matching baseline characteristics are shown in Supplemental Tables 6 and 7.
The day on ECMO on which AKI occurred and the 5 days before were compared with the corresponding same days on ECMO of the matched no AKI patients. For reasons of clarity, the corresponding day of AKI in no AKI patients was also referred to as day of AKI in the following.
Results
Study Population
From November 2014 to December 2021, 548 consecutive patients treated with VV-ECMO were screened for eligibility. Twenty-three patients were excluded because of temporary veno-arterial-ECMO/veno-veno-arterial-ECMO, 23 patients because they were younger than 18 years, and two patients because of chronic hemodialysis.
Baseline Characteristics
Baseline characteristics are shown in Table 1; additional baseline information is provided in Supplemental Table 8. Reasons for ECMO were pneumonia (bacterial 33%, viral [non-coronavirus disease (non-COVID)] 13.0%, coronavirus disease 2019 [COVID-19] 26.8%) and acute lung failure postsurgery (12.4%) or post-trauma (3.2%) and other reasons (11.6%, Supplemental Method 2 and Supplemental Table 8). ECMO support was provided for a median of 13 days (25th–75th: 7–23).
Table 1.
Baseline characteristics
| Variable | All (N=500) | Nonsurvivor (n=165) | Survivor (n=335) | P Value (Nonsurvivor versus Survivor) |
|---|---|---|---|---|
| Age (yr) | 55 (46–63) | 58 (51–65) | 54 (45–62) | <0.001a |
| Male, No. (%) | 352 (70) | 132 (80) | 220 (66) | 0.001a |
| BMI (kg/m2) | 27.8 (24.5–32.9) | 27.8 (24.2–31.2) | 27.8 (24.7–34.8) | 0.079 |
| SOFA (before ECMO) | 13 (12–15) | 13 (12–16) | 13 (12–14) | 0.006a |
| Days between admission and ECMO initiation | 6 (2–13) | 9 (2–17) | 5 (1–119) | <0.001a |
| Preexisting disease, No. (%) | ||||
| Diabetes | 97 (19) | 27 (16) | 70 (21) | 0.228 |
| Hypertension | 191 (38) | 66 (40) | 125 (37) | 0.561 |
| Chronic renal failure | 38 (8) | 17 (10) | 21 (6) | 0.109 |
| Rheumatic disease | 32 (6) | 10 (6) | 22 (7) | 0.828 |
| Malignancy | 53 (11) | 32 (19) | 21 (6) | <0.001a |
| Hematological predisease | 22 (4) | 4 (2) | 18 (5) | 0.131 |
| Hematological malignancy | 16 (3) | 10 (6) | 6 (2) | 0.011a |
| COPD | 49 (10) | 13 (8) | 36 (11) | 0.311 |
| Active smoking | 127 (25) | 39 (24) | 88 (26) | 0.525 |
| Organ transplantation | 11 (2) | 8 (5) | 3 (1) | 0.005a |
| Laboratory variables before ECMO initiation | ||||
| Hemoglobin (g/dl) | 10.5 (9–12.6) | 9.5 (8.4–12) | 10.5 (9.1–12.6) | <0.001a |
| LDH (U/L) | 392 (269–607) | 429 (293–675) | 378 (274–541) | 0.001a |
| ASAT (U/L) | 61 (36–111) | 60 (33–95) | 64 (37–100) | 0.152 |
| AP (U/L) | 79 (42–132) | 129 (71–188) | 83 (60–133) | 0.002a |
| Bilirubin (mg/dl) | 0.7 (0.4–1.3) | 0.7 (0.5–1.4) | 0.6 (0.4–1.2) | 0.043a |
| CRP (mg/L) | 144 (43–251) | 153 (54–255) | 165 (45–277) | 0.220 |
| aPTT (s) | 39 (32–48) | 41 (34–50) | 38 (32–46) | <0.001a |
| INR | 1.1 (1.0–1.3) | 1.2 (1–1.3) | 1.1 (1–1.2) | 0.020a |
| D-dimers (mg/L) | 5 (2–11) | 5 (3–11) | 4 (2–10) | 0.032a |
| Fibrinogen (mg/dl) | 527 (382–649) | 537 (386–637) | 577 (448–723) | 0.997 |
| Platelets (/nl) | 213 (140–311) | 220 (134–317) | 263 (162–318) | 0.001a |
| Cell-free hemoglobin (mg/L) | 39 (22–77) | 25 (13–58) | 31 (20–67) | 0.127 |
| Neutrophil-lymphocyte ratio | 12.9 (6.4–21.2) | 13.8 (8.2–21.2) | 13.5 (7.8–20.8) | 0.599 |
| Creatinine (mg/dl) | 1.1 (0.8–1.9) | 1.2 (0.8–2) | 1.0 (0.7–1.7) | 0.016a |
| Ventilator settings and blood gas analysis before ECMO initiation | ||||
| PEEP (mbar) | 15 (12–16) | 15 (12–16) | 15 (12–16) | 0.397 |
| Pmax (mbar) | 33 (30–36) | 32 (30–35) | 33 (30–36) | 0.153 |
| PaO2/FiO2 ratio (mm Hg) | 71 (60–96) | 71 (61–94) | 70 (60–92) | 0.670 |
| PaO2 (mm Hg) | 68 (60–83) | 68 (60–93) | 68 (59–83) | 0.873 |
| PaCO2 (mm Hg) | 61 (51–77) | 62 (54–78) | 62 (51–77) | 0.722 |
| pH | 7 (7–7.2) | 7 (7–7.2) | 7 (7–7.2) | 0.446 |
| Lactate (mg/dl) | 15 (10–26) | 17 (11–31) | 15 (10–26) | 0.115 |
| NE (µg/kg body wt per minute) | 0.2 (0.08–0.48) | 0.23 (0.09–0.54) | 0.19 (0.07–0.45) | 0.144 |
| MAP (mm Hg) | 68 (61–75) | 67 (61–74) | 68 (62–76) | 0.117 |
Table shows baseline characteristics of the entire cohort. Continuous variables are shown as medians and percentiles (25th–75th), compared with Mann–Whitney U test. Categorical variables are presented as number (No.) and percentage (%), compared with chi-squared test. AP, alkaline phosphatase; aPTT, activated partial thromboplastin time; ASAT, aspartate-aminotransferase; BMI, body mass index; COPD, chronic obstructive pulmonary disease; CRP, C-reactive protein; ECMO, extracorporeal membrane oxygenation; INR, international normalized ratio; LDH, lactate-dehydrogenase; MAP, mean arterial pressure; PEEP, positive end-expiratory pressure; SOFA, Sequential Organ Failure Assessment.
Statistically significant (P < 0.05).
Prevalence of AKI
In total, 320 patients (64%) met the KDIGO criteria for AKI. The incidence rate for the median ECMO support of 13 days was 5%. At ECMO initiation, 182 (36.4%) patients in total had AKI, whereas 158 patients (31.6%) in total developed AKI during ECMO. Of them, 13 patients had more than one episode of AKI onset during ECMO. The first episode of AKI (n=158) occurred in a median of 3 (25th–75th: 1–9) days after ECMO initiation, the second AKI (n=13) occurred after 15 (25th–75th: 7.5–27.5) days, and the third episode of AKI occurred (n=2) 6.5 days after the previous episode (Supplemental Figure 2). In patients with viral pneumonia, AKI occurred later than in patients with bacterial pneumonia (5.5 versus 2 days, Supplemental Table 9).
Twenty patients had AKI onset before and during ECMO (Figure 1). Excluding these 20 patients, 162 patients had AKI onset before ECMO. Of them, 124 patients had AKI stage 3, predominantly (95.2%) because of need of KRT. A total of 138 patients had AKI onset during ECMO including 84 (60.9%) patients with AKI stage 3, mainly (88.1%) because of need of KRT. In total, 224 patients had AKI KDIGO stage 3, including the 16 patients who had both AKI onset before and during ECMO of KDIGO stage 3. Those patients were mainly treated with KRT (92.9%).
Figure 1.
Flow chart of the distribution of AKI and KDIGO stages. AKI was defined according to KDIGO guidelines. We defined three subgroups of patients regarding the prevalence of AKI: (1) AKI onset before ECMO, AKI with onset before ECMO could also persist during ECMO; (2) AKI onset during ECMO; and (3) AKI onset before ECMO and new onset of AKI during ECMO after an initial recovery. ECMO, extracorporeal membrane oxygenation; KDIGO, Kidney Disease Improving Global Outcomes.
In total, 220 patients (44%) needed KRT during ECMO therapy. Although the need for KRT before ECMO was associated with need of KRT at hospital discharge (Supplemental Table 10), there were no differences in baseline characteristics between patients with AKI onset before ECMO or during ECMO, except for a higher prevalence of bacterial pneumonia in patients with AKI onset before ECMO (Supplemental Table 11).
KRT modes and vascular accesses are summarized in Supplemental Table 10. Regarding AKI onset during VV-ECMO, no ECMO system dependencies or association with cannula size were found (Supplemental Figure 1 and Supplemental Tables 1 and 2).
Baseline Characteristics of Patients with AKI
Patients with AKI onset before ECMO had a higher BMI (29.4 versus 27 kg/m2, P < 0.001) and higher SOFA score (15 versus 12, P < 0.001) compared with those without AKI (Table 2). Patients with AKI were older (56.7 versus 53.6 years, P = 0.010) and had shorter ECMO support (10 versus 12.5 days, P = 0.029). KRT was started at a median of 1 day after ECMO initiation (25th–75th: 0–3).
Table 2.
Baseline characteristics of AKI onset before and AKI onset during extracorporeal membrane oxygenation versus no AKI
| Variable | No AKI (n=180) | AKI Onset Before ECMO (n=162)a | P Value (AKI Onset Before ECMO versus No AKI) | AKI Onset during ECMO (n=138)a | P Value (AKI Onset during ECMO versus No AKI) |
|---|---|---|---|---|---|
| Age (yr) | 54 (41–62) | 57 (48–64) | 0.010b | 55 (46–63) | 0.141 |
| Male, No. (%) | 119 (66) | 122 (75) | 0.063 | 96 (70) | 0.514 |
| BMI (kg/m2) | 27 (23.8–30.1) | 29.4 (25.4–36.0) | <0.001b | 27.8 (24.5–34.0) | 0.008b |
| SOFA (before ECMO) | 12 (11–13) | 15 (13–17) | <0.001b | 13 (12–14) | <0.001b |
| SOFA without kidney variable (before ECMO) | 12 (11–12) | 13 (12–15) | <0.001b | 12 (12–13) | 0.002b |
| VV-ECMO support (d) | 13 (7–22) | 10 (6–16) | 0.029b | 15 (8–29) | 0.085 |
| Days between admission and ECMO initiation | 7.5 (2–16) | 5 (2–12) | 0.115 | 4.5 (1–10.3) | 0.002b |
| Indications of ECMO | |||||
| Pneumonia | |||||
| Bacterial | 52 (29) | 72 (44) | 0.003b | 36 (26) | 0.580 |
| Viral (all) | 82 (46) | 39 (24) | <0.001b | 65 (47) | 0.784 |
| Viral (non-COVID) | 13 (7) | 22 (14) | 0.053 | 25 (18) | 0.003b |
| COVID-19 | 69 (38) | 17 (11) | <0.001b | 40 (29) | 0.082 |
| Acute lung failure | |||||
| Postsurgery | 14 (8) | 23 (14) | 0.056 | 23 (17) | 0.014b |
| Post-trauma | 9 (5) | 3 (2) | 0.114 | 4 (3) | 0.348 |
| Other | 23 (13) | 25 (15) | 0.480 | 10 (7) | 0.109 |
| Laboratory variables before ECMO initiation | |||||
| Hemoglobin (g/dl) | 10.7 (9–12.6) | 10 (8.7–12.7) | 0.068 | 10.5 (9.1–12.2) | 0.905 |
| LDH (U/L) | 342 (246–513) | 423 (300–648) | 0.001b | 410 (283–586) | 0.003b |
| ASAT (U/L) | 47 (34–80) | 88 (54–168) | <0.001b | 51 (33–83) | 0.213 |
| Bilirubin (mg/dl) | 0.6 (0.4–0.9) | 1 (0.5–2.9) | <0.001b | 0.7 (0.4–1.1) | 0.055 |
| CRP (mg/L) | 110 (27–209) | 157 (61–270) | 0.001b | 148 (38–263) | 0.071 |
| aPTT (s) | 35 (31–41) | 46 (38–59) | <0.001b | 39 (34–51) | 0.024b |
| INR | 1.1 (1–1.2) | 1.2 (1.1–1.5) | 0.001b | 1.1 (1–1.3) | 0.234 |
| D-dimers (mg/L) | 4 (2–11) | 5 (3–15) | 0.041b | 5 (2–10) | 0.296 |
| Fibrinogen (mg/dl) | 521 (354–654) | 552 (422–712) | 0.928 | 584 (492–688) | 0.907 |
| Platelets (/nl) | 270 (191–371) | 192 (84–284) | <0.001b | 255 (160–334) | <0.001b |
| Cell-free hemoglobin (mg/L) | 33 (20–57) | 42 (23–80) | <0.001b | 26 (17–54) | 0.007b |
| Creatinine (mg/dl) | 0.8 (0.6–1) | 2.2 (1.6–3.3) | <0.001b | 1 (0.7–1.5) | <0.001b |
| Ventilation parameters before ECMO initiation | |||||
| MV (L/min) | 9.6 (7.1–11.5) | 10.1 (8.2–12.8) | 0.004b | 10 (8–12) | 0.125 |
| TV/IBW (ml/kg IBW) | 6.23 (4.88–7.35) | 6.60 (5.56–7.91) | 0.010b | 6.59 (5.41–7.42) | 0.122 |
| PEEP (mbar) | 14 (11–15) | 15 (12–16) | 0.002b | 15 (12–16) | 0.002b |
| Pmax (mbar) | 32 (29–35) | 33 (30–36) | 0.022b | 33 (30–36) | 0.061 |
| PaO2/FiO2 ratio (mm Hg) | 79 (60–99) | 68 (59–98) | 0.433 | 70 (61–86) | 0.067 |
| PaO2 (mm Hg) | 72 (58–87) | 67 (59–82) | 0.199 | 69 (60–80) | 0.127 |
| PaCO2 (mm Hg) | 65 (51–80) | 61 (52–77) | 0.243 | 60 (50–77) | 0.183 |
| pH | 7.23 (7.15–7.30) | 7.16 (7.10–7.23) | <0.001b | 7.22 (7.12–7.29) | 0.311 |
| Lactate (mg/dl) | 13 (10–20) | 19 (11–35) | <0.001b | 15 (10–27) | 0.034b |
| NE (µg/kg body wt per minute) | 0.15 (0.06–0.27) | 0.31 (0.15–0.67) | <0.001b | 0.17 (0.05–0.45) | 0.227 |
| MAP (mm Hg) | 68 (64–76) | 67 (61–72) | 0.052 | 70 (63–76) | 0.730 |
Table shows baseline characteristics of patients with AKI onset before and AKI only during extracorporeal membrane oxygenation compared with patients without AKI. AKI was defined according to Kidney Disease Improving Global Outcomes guidelines. We defined three subgroups of patients regarding the prevalence of AKI: (1) AKI onset before extracorporeal membrane oxygenation, AKI with onset before extracorporeal membrane oxygenation could also persist during extracorporeal membrane oxygenation; (2) AKI onset during extracorporeal membrane oxygenation; and (3) AKI onset before extracorporeal membrane oxygenation and new onset of AKI during extracorporeal membrane oxygenation after an initial recovery. For the analysis of risk factors for AKI, we excluded the group of patients who had AKI onset before and new onset of AKI during extracorporeal membrane oxygenation to allow a clear attribution. Additional information about preexisting diseases can be seen in Supplemental Table 12. Continuous variables are shown as medians and percentiles (25th–75th), compared with Mann–Whitney U test. Categorical variables are presented as numbers (No.) and percentage (%), compared with chi-squared test. aPTT, activated partial thromboplastin time; ASAT, aspartate-aminotransferase; BMI, body mass index; COVID, coronavirus disease; COVID-19, coronavirus disease 2019; CRP, C-reactive protein; ECMO, extracorporeal membrane oxygenation; IBW, ideal body weight; INR, international normalized ratio; LDH, lactate-dehydrogenase; MAP, mean arterial pressure; MV, minute ventilation; PEEP, positive end-expiratory pressure; SOFA, Sequential Organ Failure Assessment; TV, tidalvolume; VV-ECMO, veno-venous extracorporeal membrane oxygenation.
Patients with AKI onset before or onset during extracorporeal membrane oxygenation, without 20 patients who had both.
Statistically significant (P < 0.05).
Patients with AKI onset before ECMO suffered more frequently from arterial hypertension (P = 0.028), chronic renal insufficiency (P = 0.006), hematological malignancy (P = 0.003), chronic obstructive pulmonary disease (P = 0.002), active smoking (P = 0.003; Supplemental Table 12), and bacterial pneumonia (P = 0.003). They also had worse liver function and higher inflammatory markers as opposed to no AKI patients (Table 2) and received more NE (0.31 versus 0.15 µg/kg body wt per minute, P < 0.001) at ECMO initiation.
In patients who had AKI onset during ECMO, there were no relevant differences in terms of demographic variables compared with patients who never had AKI. KRT was started in this patient group at a median of 9 days after ECMO initiation (25th–75th: 2–16).
Patients with AKI onset during ECMO required ECMO support more frequently because of viral (non-COVID) pneumonia (18.1% versus 7.2%, P = 0.003) and acute lung failure postsurgery (16.7% versus 7.8%, P = 0.014) than no AKI patients.
Outcome at Hospital Discharge
The overall survival to hospital discharge was 67%. Survival rates to discharge depending on AKI are included in Figure 1 (additional information in Supplemental Table 13).
Among the 335 patients discharged from hospital, 32 patients (9.6%) were discharged on KRT. Twenty one patients with KRT at discharge had AKI onset before ECMO, ten patients had AKI onset during ECMO, and one patient had AKI onset before and during ECMO.
Three patients received KRT after ECMO therapy because of AKI onset after ECMO (0.9%), and none of them survived.
In the Cox regression analysis for risk factors associated with reduced hospital survival age, lower BMI, malignancy, and AKI KDIGO stage 3, independent from onset were independent risk factors for reduced survival in patients with VV-ECMO therapy (Table 3 and Supplemental Table 14).
Table 3.
Cox regression analysis for risk factors associated with reduced hospital survival
| Variables | HR (95% CI) | P Value |
|---|---|---|
| Male sex | 1.27 (0.90 to 1.77) | 0.173 |
| Age (yr) | 1.03 (1.01 to 1.04) | <0.001a |
| BMI (kg/m2) | 0.97 (0.95 to 1.00) | 0.018a |
| Days between admission and ECMO initiation | 1.00 (0.10 to 1.00) | 0.542 |
| SOFA without kidney variable (before ECMO) | 1.04 (0.97 to 1.12) | 0.273 |
| Pneumonia | ||
| Bacterial | 0.75 (0.52 to 1.08) | 0.126 |
| Viral (non-COVID) | 0.63 (0.39 to 1.03) | 0.063 |
| Viral (COVID) | 1.24 (0.83 to 1.86) | 0.302 |
| Preexisting disease | ||
| Diabetes mellitus | 0.77 (0.53 to 1.14) | 0.193 |
| Hypertension | 1.00 (0.73 to 1.36) | 0.999 |
| Chronic renal failure | 1.39 (0.81 to 2.40) | 0.230 |
| Malignancy | 2.70 (1.85 to 3.95) | <0.001a |
| COPD | 1.07 (0.66 to 1.71) | 0.794 |
| Organ transplantation | 1.30 (0.61 to 2.79) | 0.495 |
| AKI | 1.15 (0.72 to 1.82) | 0.566 |
| AKI with KDIGO stage 3 | 2.17 (1.38 to 3.40) | 0.001a |
| AKI with KDIGO stage 3 onset before or during ECMO | 1.18 (0.80 to 1.72) | 0.403 |
Table shows the Cox regression analysis (adjusted for male sex, age, body mass index, days between admission and extracorporeal membrane oxygenation initiation, sequential organ failure assessment before extracorporeal membrane oxygenation without kidney variable, bacterial and viral pneumonia, preexisting diseases, and onset of AKI for risk factors associated with reduced hospital survival. One hundred and sixty-five of 500 patients did not survive to hospital discharge. AKI was defined according to Kidney Disease Improving Global Outcomes guidelines. We defined three subgroups of patients regarding the prevalence of AKI: (1) AKI onset before extracorporeal membrane oxygenation, AKI with onset before extracorporeal membrane oxygenation could also persist during extracorporeal membrane oxygenation; (2) AKI onset during extracorporeal membrane oxygenation; and (3) AKI onset before extracorporeal membrane oxygenation and new onset of AKI during extracorporeal membrane oxygenation after an initial recovery. BMI, body mass index; COVID, coronavirus disease; COPD, chronic obstructive pulmonary disease; CI, confidence interval; ECMO, extracorporeal membrane oxygenation; HR, hazard ratio; KDIGO, Kidney Disease Improving Global Outcomes; SOFA, Sequential Organ Failure Assessment.
Statistically significant (P < 0.05).
The multivariate logistic regression analysis for risk factors associated with KRT at discharge displayed viral (non-COVID) pneumonia as an independent risk factor (odds ratio, 5.32; 95% confidence interval, 1.27 to 22.24; P = 0.022; Supplemental Table 15).
Outcome after 1 Year
In total, 335 of 500 (67%) patients survived to hospital discharge (195 patients with AKI [58.2%], 140 without AKI [41.8%]).
Twenty-five patients (7.5%) died within the 1 year follow-up after hospital discharge. Seven patients with AKI onset during ECMO and 11 patients with AKI onset before ECMO did not survive. One deceased patient had both AKI onset before and during ECMO. The other six patients had no AKI (Supplemental Table 16).
During follow-up, nonsurvivors were older (61.6 versus 53.4 years, P = 0.001), had predominantly bacterial pneumonia (P = 0.031), malignancy (P = 0.003), and chronic obstructive pulmonary disease (P = 0.004) as well as fewer cases of COVID-19 (P = 0.013).
Figure 2A shows the survival of all patients up to 1 year after hospital admission. Patients with AKI KDIGO stage 3 survived significantly less often (P < 0.001 by log-rank test) than patients with KDIGO stage 1 or 2 or patients without AKI. AKI KDIGO stage 3 was confirmed as an independent risk factor for reduced survival to hospital discharge in the Cox regression analysis (Table 3).
Figure 2.
Kaplan-Meier survival analysis. Kaplan–Meier survival curves of patients after (A) hospital admission and (B) hospital discharge undergoing ECMO. The end point is 365 days after (A) hospital admission whether still hospitalized or discharged or (B) hospital discharge. *Equals P < 0.05, **equals P < 0.01, and ***equals P < 0.001. Unless otherwise stated, no significant differences. (A) Survival function of all patients after hospital admission divided by no AKI, patients with AKI KDIGO stage 1/2, patients with AKI KDIGO stage 3, and all 500 patients. Significant difference between KDIGO stage 3 and KDIGO stage 1/2 as well as KDIGO stage 3 and no AKI (each P < 0.001 by log-rank test). (B) Survival function of 335 survivors after hospital discharge divided by no AKI, patients with AKI KDIGO stage 1/2, patients with AKI KDIGO stage 3 and all 335 patients. Significant difference between KDIGO stage 3 and no AKI (P = 0.021 by log-rank test) as well as KDIGO stage 1/2 and no AKI (P = 0.003 by log-rank test).
The univariate survival of discharged patients (Figure 2B) shows significantly lower 1-year survival in patients with AKI KDIGO stage 1/2 (P = 0.003 by log-rank test) and AKI KDIGO stage 3 (P = 0.021 by log-rank test) compared with patients without AKI. The multivariate logistic regression analysis of the hospital survivors demonstrates no influence of AKI or KRT on the 1-year survival after hospital discharge, whereas higher age was the only independent risk factor for reduced survival during follow-up (odds ratio, 1.06; 95% confidence interval, 1.01 to 1.11; P = 0.017; Supplemental Table 17).
Dependency of AKI from underlying diseases is summarized in Supplemental Table 18.
Risk Factors for AKI Onset during VV-ECMO
To evaluate risk factors for AKI onset during ECMO, propensity score matching was performed (see Methods section). The MAP of AKI patients was significantly lower the day before AKI than in no AKI patients (69 versus 76 mm Hg, P = 0.021, Figure 3A). AKI patients had a longer cumulative and over the days increasing time below MAP of 65 mm Hg (Figure 3B, hemodynamic parameters during ECMO course in Supplemental Figure 3). There was a significant increase in SOFA score in AKI patients before AKI (Figure 3C). Patients with AKI had higher C-reactive protein levels the days before AKI (day-2: 70 versus 34 mg/L, P = 0.002; day-1: 76 versus 23 mg/L, P < 0.001; AKI: 85 versus 18 mg/L, P < 0.001, Figure 3D). In addition, the number of AKI patients receiving vancomycin increased before AKI, but there was no difference in dosage between AKI and no AKI patients (Figure 3E). There was no difference in acyclovir and contrast media administration between AKI and no AKI patients (Figure 3, F and G).
Figure 3.
Comparison between matched no AKI patients and patients with AKI who developed AKI during ECMO. Displaying different variables on the 5 days before AKI and the day of AKI. The graphs are presented as boxplots. *Equals P < 0.05, **equals P < 0.01, and ***equals P < 0.001. Unless otherwise stated, no significant differences. (A) MAP. Comparing AKI and no AKI patients, the third and the last day before AKI show significant decreased MAPs in AKI patients (P = 0.040, P = 0.021). (B) Cumulative daily time below MAP of 65 mm Hg. (C) SOFA Score. Comparing AKI and no AKI patients, the fifth and fourth day before AKI and the day of AKI show significant elevated SOFA scores in AKI patients (P = 0.007, P = 0.026, P < 0.001). SOFA score increased significantly in AKI patients over the days. (D) CRP. Comparing AKI and no AKI patients, the second and last day before AKI and the day of AKI show significant elevated CRP-levels in AKI patients (P = 0.002, each P < 0.001). (E) Proportion of patients who received vancomycin. Comparing AKI and no AKI patients, the last day before AKI shows significant elevated administration of vancomycin in AKI patients (P = 0.037). Vancomycin administration increased significantly in AKI patients over the days. (F) Proportion of patients who received acyclovir. (G) Proportion of patients who received contrast media. CRP, C-reactive protein; MAP, mean arterial pressure; SOFA, Sequential Organ Failure Assessment.
Clinically relevant differences in cell-free hemoglobin, platelet count, ECMO flow rates, ventilator settings, or arterial blood gases between AKI and no AKI patients were not found (Supplemental Figure 4, 3A–J, Supplemental Table 19).
Discussion
To our knowledge, this study is the largest examining AKI in patients supported with VV-ECMO. It is the first study to elaborate risk factors for developing AKI during VV-ECMO using propensity score matching. Key findings are severe AKI being a strong predictor of reduced hospital survival and the attribution of AKI onset during VV-ECMO therapy not to ECMO-related factors, but to parameters describing sepsis or the severity of the underlying disease.
AKI occurred in 64% of the patients, which is in accordance with current literature (50.3% up to 76.7% AKI in VV-ECMO).13,14,20 The first episode of AKI during VV-ECMO occurred 3 days after ECMO-initiation, which is in line with other studies.21,22
Patients with AKI onset before ECMO had worse liver function, higher inflammatory markers, less platelets, higher NE dosage, and higher SOFA score, even after excluding the renal variable, prior ECMO therapy, all indicating more severe sepsis. Thus, the occurrence of AKI onset before ECMO may be attributed to the severity of the underlying disease. This is in accordance with existing literature in ECMO patients, as other6,23,24 identified systemic inflammation and sepsis as most prevalent clinical scenario associated with AKI.
A subgroup of patients had no AKI at ECMO initiation but developed AKI during ECMO. These patients had almost no clinically relevant differences concerning the laboratory and ventilatory variables in the beginning of ECMO therapy compared with patients without AKI. Therefore, AKI onset during ECMO must be related to factors occurring during ECMO support, which is discussed below.
Patients with AKI had a significantly lower survival rate than patients without AKI (P < 0.001), which is in line with other studies.14,15,25 In this context, the multivariable regression model showed another key finding of our study: AKI KDIGO stage 3, irrespective from onset, is an independent risk factor for reduced survival. This reveals that not the occurrence of AKI onset before or during ECMO is important regarding the outcome, but the severity of the AKI. A previous study1 has found a lower survival rate in patients on KRT before ECMO, too, but could not identify AKI requiring KRT during ECMO as independent risk factor for reduced survival. However, this result might be because of smaller sample size.
Considering the long-term outcome, we found a 92.5% (310/335) survival rate 1 year after discharge, which is in line with other findings showing a high survival rate after hospital discharge.26 Among the patients surviving to hospital discharge, univariate analysis showed an effect of AKI on long-term survival, whereas multivariate logistic regression analysis, adjusted to disease severity and comorbidities, showed no significant difference, and revealed age as the sole independent risk factor for reduced 1-year follow-up survival. By contrast, Pilarczyk et al.2 and Lumlertgul et al.14 stated that survival up to 5 years was significantly decreased in patients with moderate or severe AKI. This could be due to the longer follow-up period of 4–5 years compared with 1 year in our study and should be verified in future studies.
However, it is unclear whether ECMO itself, the underlying disease, or nephrotoxic events during ECMO support contribute to the development of AKI onset during ECMO. Therefore, our study evaluated risk factors of AKI occurrence on the day of AKI and the 5 days before using propensity score matching. Surrogate parameters of ECMO therapy (cell-free hemoglobin, platelet levels, ECMO flow rates, and blood gas analysis) were not significantly different between AKI and no AKI patients. However, SOFA score, C-reactive protein, and bilirubin levels were higher in AKI patients compared with no AKI patients. Together, this suggests the underlying disease or recurrent sepsis as a risk factor for AKI and not factors that are usually associated with ECMO therapy (e.g., cell-free hemoglobin). Mou et al.4 showed that a higher Acute Physiology, Age, Chronic Health Evaluation II and a higher SOFA score before ECMO are associated with a higher risk of AKI in ECMO patients, emphasizing that the underlying disease is a risk factor for the occurrence of AKI. Although it has been shown that ECMO can activate coagulation and inflammation,12,27,28 the magnitude of the underlying infectious process or sepsis is by far higher than the processes initiated by the ECMO itself and are the main factors to contribute to AKI.
In our study, MAP was lower in AKI patients despite higher dosages of vasopressors. This underlines the severity of the recurrent septic episode with increasing need of vasopressors to maintain MAP. Some current studies show that a higher MAP (80–85 mm Hg) might be protective of AKI occurrence or the need of KRT.29–31 None of these studies evaluated patients on ECMO. Whether an even higher dosage of vasopressors could have prevented AKI in our study cannot be answered but should be addressed in further studies.
This study is limited to its observational retrospective single-center design. Therefore, it cannot establish any causal relationships, and generalizability is impaired.
Furthermore, there was only limited data from the patients before the hospital stay. This includes creatinine levels, which were estimated to be normal if not stated chronic renal insufficiency in medical reports, to categorize KDIGO stages before ECMO therapy. The prevalence of AKI may have been influenced by using only creatinine for diagnosis and the unavailability of baseline creatinine (before the beginning of the actual disease leading to ECMO) and urine output. As a result of the change in PDMS, the number of patients for propensity score matching was reduced, which is why other possible influencing factors before 2018 may not be included. This applies to the higher severity of the disease that could be observed in the group of patients excluded from propensity score matching. The generalizability of these findings to other centers with different ECMO protocols may be limited. Follow-up was only 1 year. Unfortunately, the duration and need of KRT during follow-up were not recorded in the database and cannot be analyzed and reported.
In addition, because of the pandemic situation from 2020, the indication of ECMO in patients with AKI was made stricter. This means that the limited ECMO spots went primarily to the healthier COVID-19 patients without kidney failure so that fewer patients of our VV-ECMO cohort with COVID-19 had AKI in our observational retrospective study.
For survival analysis, we focused on AKI. Therefore, other important factors that could affect survival, such as ECMO complications or sepsis, were not included.
AKI is a frequent complication in patients supported with VV-ECMO. AKI onset before VV-ECMO therapy is linked to the severity of the underlying disease. AKI onset during VV-ECMO is less because of not only ECMO-related factors but also the severity of the disease and septic episodes. Severe AKI, i.e., KDIGO stage 3, is associated with reduced hospital survival, regardless of whether it occurs before or during VV-ECMO. One year survival after hospital discharge is not influenced by AKI. Future research should aim on earlier AKI diagnosis and nephroprotective strategies.
Supplementary Material
Acknowledgments
We would like to thank the nursing staff and doctors of the intensive care units at the UKR for their excellent patient care and all perfusionists for the daily, meticulous maintenance of the database. We thank Florian Zeman for statistical advice and support.
Footnotes
C.W. and F.F. contributed equally as first authors.
Disclosures
Disclosure forms, as provided by each author, are available with the online version of the article at http://links.lww.com/KN9/B177.
Author Contributions
Conceptualization: Franziska Fuchs, Matthias Lubnow, Clemens Wiest.
Data curation: Maik Foltan, Franziska Fuchs, Alois Philipp, Clemens Wiest.
Formal analysis: Franziska Fuchs, Clemens Wiest.
Investigation: Franziska Fuchs, Clemens Wiest.
Methodology: Franziska Fuchs, Matthias Lubnow, Clemens Wiest.
Project administration: Franziska Fuchs, Matthias Lubnow, Clemens Wiest.
Resources: Maik Foltan, Franziska Fuchs, Alois Philipp, Clemens Wiest.
Software: Clemens Wiest.
Supervision: Matthias Lubnow, Clemens Wiest.
Validation: Franziska Fuchs, Matthias Lubnow, Clemens Wiest.
Visualization: Franziska Fuchs, Matthias Lubnow, Clemens Wiest.
Writing – original draft: Franziska Fuchs, Clemens Wiest.
Writing – review & editing: Alexander Dietl, Christoph Fisser, Maik Foltan, Franziska Fuchs, Matthias Lubnow, Dirk Lunz, Thomas Müller, Alois Philipp, Roland Schneckenpointner, Clemens Wiest.
Funding
None.
Declarative Statements
This study includes clinical experimentation and received Institutional Review Board or Ethics Committee approval. The need to obtain informed patient consent was waived.
Data Availability Statements
Original data generated for the study will be made available on reasonable request to the corresponding author. Observational Data. Deidentified data are available on reasonable request to the corresponding author.
Supplemental Material
This article contains the following supplemental material online at http://links.lww.com/KN9/B178.
Supplemental Table 1. Characteristics about ECMO pumps and oxygenators of all patients (N=500).
Supplemental Table 2. Characteristics about ECMO cannulas of all patients (N=500). (A) Outflow cannulas. (B) Inflow cannulas.
Supplemental Figure 1. ECMO-related parameters of different ECMO pumps and oxygenators of all patients (N=500).
Supplemental Table 3. Prevalence of AKI before (n=312) and after the start (n=188) of COVID-pandemic.
Supplemental Table 4. Staging of AKI.
Supplemental Table 5. Baseline characteristics of patients with and without detailed hemodynamic information (N=500).
Supplemental Table 6. Summary of propensity score matching AKI onset during ECMO starting from day 3.
Supplemental Table 7. Baseline characteristics of propensity score–matched cohort (n=82) AKI onset during ECMO starting on day 3.
Supplemental Table 8. Additional information to baseline characteristics of all patients (n=500) in Table 1.
Supplemental Figure 2. Distribution of occurrence of AKI onset during ECMO therapy over time.
Supplemental Table 9. Distribution of occurrence of AKI onset during ECMO therapy over time regarding different ECMO indications (n=158).
Supplemental Table 10. Baseline characteristics about KRT of all patients (n=500).
Supplemental Table 11. Baseline characteristics of patients with AKI who were discharged on KRT (n=31) compared with AKI onset before and AKI onset during ECMO.
Supplemental Table 12. Additional information to baseline characteristics of AKI onset before (n=182) and AKI onset during ECMO (n=138) versus no AKI (n=180) in Table 2.
Supplemental Table 13. Survival rate compared between patients without AKI and with AKI, AKI onset before and AKI onset during ECMO.
Supplemental Table 14. Cox regression analysis and sensitivity analysis for risk factors associated with reduced hospital survival.
Supplemental Table 15. Multivariate logistic regression analysis for risk factors associated with KRT at discharge.
Supplemental Table 16. Death and survival in the follow-up after hospital discharge (n [survivors]=335, N [all]=500).
Supplemental Table 17. Multivariate logistic regression model about mortality to 1 year after hospital discharge (n [survivors]=335).
Supplemental Table 18. Added risk of AKI among patients with various aetiologies of acute respiratory failure in reference to bacterial pneumonia (all patients with AKI, n=320; patients with AKI onset before ECMO, n=162; patients with AKI onset during ECMO, n=138).
Supplemental Figure 3. Hemodynamic parameters on the time course of ECMO treatment in 3.1. Two hundred and eighteen patients (new PDMS and without AKI onset before ECMO), comparing AKI and no AKI patients. 3.2. Twenty one patients (new PDMS and without AKI onset before ECMO) with AKI stage 3 and survived to hospital discharge, comparing KRT and no-KRT at hospital discharge.
Supplemental Figure 4. Comparison between propensity score–matched no AKI patients and patients with AKI who developed AKI during ECMO.
Supplemental Table 18. Added risk of AKI among patients with various aetiologies of acute respiratory failure in reference to bacterial pneumonia (all patients with AKI, n=320; patients with AKI onset before ECMO, n=162; patients with AKI onset during ECMO, n=138).
Supplemental Table 19. Potential risk factors for AKI in propensity score–matched cohort (n=82) 5 days before and on the day of AKI (time matched for no AKI patients).
References
- 1.Haneya A Diez C Philipp A, et al. Impact of acute kidney injury on outcome in patients with severe acute respiratory failure receiving extracorporeal membrane oxygenation. Crit Care Med. 2015;43(9):1898–1906. doi: 10.1097/CCM.0000000000001141 [DOI] [PubMed] [Google Scholar]
- 2.Pilarczyk K Huenges K Bewig B, et al. Acute kidney injury in patients with severe ARDS requiring extracorporeal membrane oxygenation: incidence, prognostic impact and risk factors. J Clin Med. 2022;11(4):1079. doi: 10.3390/jcm11041079 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Chen H, Yu R-G, Yin N-N, Zhou J-X. Combination of extracorporeal membrane oxygenation and continuous renal replacement therapy in critically ill patients: a systematic review. Crit Care. 2014;18(6):675. doi: 10.1186/s13054-014-0675-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Mou Z, Guan T, Chen L. Risk factors of acute kidney injury in ECMO patients: a systematic review and meta-analysis. J Intensive Care Med. 2022;37(2):267–277. doi: 10.1177/08850666211003485 [DOI] [PubMed] [Google Scholar]
- 5.Kielstein JT Heiden AM Beutel G, et al. Renal function and survival in 200 patients undergoing ECMO therapy. Nephrol Dial Transplant. 2013;28(1):86–90. doi: 10.1093/ndt/gfs398 [DOI] [PubMed] [Google Scholar]
- 6.Ostermann M, Lumlertgul N. Acute kidney injury in ECMO patients. Crit Care. 2021;25(1):313. doi: 10.1186/s13054-021-03676-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Ranieri VM, Giunta F, Suter PM, Slutsky AS. Mechanical ventilation as a mediator of multisystem organ failure in acute respiratory distress syndrome. JAMA. 2000;284(1):43–44. doi: 10.1001/jama.284.1.43 [DOI] [PubMed] [Google Scholar]
- 8.Fanelli V Cantaluppi V Alessandri F, et al. Extracorporeal CO2 removal may improve renal function of patients with acute respiratory distress syndrome and acute kidney injury: an open-label, interventional clinical trial. Am J Respir Crit Care Med. 2018;198(5):687–690. doi: 10.1164/rccm.201712-2575LE [DOI] [PubMed] [Google Scholar]
- 9.Husain-Syed F, Slutsky AS, Ronco C. Lung-kidney cross-talk in the critically ill patient. Am J Respir Crit Care Med. 2016;194(4):402–414. doi: 10.1164/rccm.201602-0420CP [DOI] [PubMed] [Google Scholar]
- 10.Barrot L Asfar P Mauny F, et al. Liberal or conservative oxygen therapy for acute respiratory distress syndrome. N Engl J Med. 2020;382(11):999–1008. doi: 10.1056/NEJMoa1916431 [DOI] [PubMed] [Google Scholar]
- 11.van den Akker JPC, Egal M, Groeneveld ABJ, Groeneveld ABJ. Invasive mechanical ventilation as a risk factor for acute kidney injury in the critically ill: a systematic review and meta-analysis. Crit Care. 2013;17(3):R98. doi: 10.1186/cc12743 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Hatami S, Hefler J, Freed DH. Inflammation and oxidative stress in the context of extracorporeal cardiac and pulmonary support. Front Immunol. 2022;13:831930. doi: 10.3389/fimmu.2022.831930 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Deatrick KB Mazzeffi MA Galvagno SM JR, et al. Breathing life back into the kidney-continuous renal replacement therapy and veno-venous extracorporeal membrane oxygenation. ASAIO J. 2021;67(2):208–212. doi: 10.1097/MAT.0000000000001210 [DOI] [PubMed] [Google Scholar]
- 14.Lumlertgul N Wright R Hutson G, et al. Long-term outcomes in patients who received veno-venous extracorporeal membrane oxygenation and renal replacement therapy: a retrospective cohort study. Ann Intensive Care. 2022;12(1):70. doi: 10.1186/s13613-022-01046-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Devasagayaraj R, Cavarocchi NC, Hirose H. Does acute kidney injury affect survival in adults with acute respiratory distress syndrome requiring extracorporeal membrane oxygenation? Perfusion. 2018;33(5):375–382. doi: 10.1177/0267659118755272 [DOI] [PubMed] [Google Scholar]
- 16.Joannidis M Forni LG Klein SJ, et al. Lung-kidney interactions in critically ill patients: consensus report of the acute disease quality initiative (ADQI) 21 workgroup. Intensive Care Med. 2020;46(4):654–672. doi: 10.1007/s00134-019-05869-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Khwaja A. KDIGO clinical practice guidelines for acute kidney injury. Nephron Clin Pract. 2012;120(4):c179–c184. doi: 10.1159/000339789 [DOI] [PubMed] [Google Scholar]
- 18.Ostermann M Joannidis M Pani A, et al. Patient selection and timing of continuous renal replacement therapy. Blood Purif. 2016;42(3):224–237. doi: 10.1159/000448506 [DOI] [PubMed] [Google Scholar]
- 19.Singer M Deutschman CS Seymour CW, et al. The third international consensus definitions for sepsis and septic shock (Sepsis-3). JAMA. 2016;315(8):801–810. doi: 10.1001/jama.2016.0287 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Chen S-W Lu Y-A Lee C-C, et al. Long-term outcomes after extracorporeal membrane oxygenation in patients with dialysis-requiring acute kidney injury: a cohort study. PLoS One. 2019;14(3):e0212352. doi: 10.1371/journal.pone.0212352 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Vavullipathy NU, Prasannan B. WCN23-0290 acute kidney injury in patients on ECMO. Kidney Int Rep. 2023;8(3 suppl):S4. doi: 10.1016/j.ekir.2023.02.009 [DOI] [Google Scholar]
- 22.Delmas C Zapetskaia T Conil JM, et al. 3-month prognostic impact of severe acute renal failure under veno-venous ECMO support: importance of time of onset. J Crit Care. 2018;44:63–71. doi: 10.1016/j.jcrc.2017.10.022 [DOI] [PubMed] [Google Scholar]
- 23.White KC Serpa-Neto A Hurford R, et al. Sepsis-associated acute kidney injury in the intensive care unit: incidence, patient characteristics, timing, trajectory, treatment, and associated outcomes. A multicenter, observational study. Intensive Care Med. 2023;49(9):1079–1089. doi: 10.1007/s00134-023-07138-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Husain-Syed F Ricci Z Brodie D, et al. Extracorporeal organ support (ECOS) in critical illness and acute kidney injury: from native to artificial organ crosstalk. Intensive Care Med. 2018;44(9):1447–1459. doi: 10.1007/s00134-018-5329-z [DOI] [PubMed] [Google Scholar]
- 25.Kaushal M Schwartz J Gupta N, et al. Patient demographics and extracorporeal membranous oxygenation (ECMO)-related complications associated with survival to discharge or 30-day survival in adult patients receiving venoarterial (VA) and venovenous (VV) ECMO in a Quaternary care urban center. J Cardiothorac Vasc Anesth. 2019;33(4):910–917. doi: 10.1053/j.jvca.2018.08.193 [DOI] [PubMed] [Google Scholar]
- 26.Rilinger J Krötzsch K Bemtgen X, et al. Long-term survival and health-related quality of life in patients with severe acute respiratory distress syndrome and veno-venous extracorporeal membrane oxygenation support. Crit Care. 2021;25(1):410. doi: 10.1186/s13054-021-03821-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Burrell AJC Lubnow M Enger TB, et al. The impact of venovenous extracorporeal membrane oxygenation on cytokine levels in patients with severe acute respiratory distress syndrome: a prospective, observational study. Crit Care Resusc. 2017;19(suppl 1):37–44. [PubMed] [Google Scholar]
- 28.Malfertheiner MV Philipp A Lubnow M, et al. Hemostatic changes during extracorporeal membrane oxygenation: a prospective randomized clinical trial comparing three different extracorporeal membrane oxygenation systems. Crit Care Med. 2016;44(4):747–754. doi: 10.1097/CCM.0000000000001482 [DOI] [PubMed] [Google Scholar]
- 29.Asfar P Meziani F Hamel J-F, et al. High versus low blood-pressure target in patients with septic shock. N Engl J Med. 2014;370(17):1583–1593. doi: 10.1056/NEJMoa1312173 [DOI] [PubMed] [Google Scholar]
- 30.Corrêa TD, Vuda M, Takala J, Djafarzadeh S, Silva E, Jakob SM. Increasing mean arterial blood pressure in sepsis: effects on fluid balance, vasopressor load and renal function. Crit Care. 2013;17(1):R21. doi: 10.1186/cc12495 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Demiselle J Seegers V Lemerle M, et al. Prior exposure to angiotensin II receptor blockers in patients with septic shock to individualize mean arterial pressure target? A post hoc analysis of the sepsis and mean arterial pressure (SEPSISPAM) trial. Crit Care Med. 2021;49(4):e412–e422. doi: 10.1097/CCM.0000000000004872 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Original data generated for the study will be made available on reasonable request to the corresponding author. Observational Data. Deidentified data are available on reasonable request to the corresponding author.