ABSTRACT
INTRODUCTION: Acute kidney injury requiring renal replacement therapy (AKI-RRT) is associated with high mortality, especially in the setting of COVID-19. During the peak of the delta wave in New Mexico in late 2021, crisis standards of care were declared and strategies to ration care were explored. Our hypothesis is that a simplified SOFA score in patients with COVID-19 and AKI-RRT may predict short-term mortality. METHOD(S): We retrospectively analyzed all COVID-19 patients started on CRRT for AKI in the medical ICU at our center between April 2020 and July 2021. A 4-organ SOFA score (4OSS), with renal and neurologic sub-scores excluded, was calculated at the time of CRRT initiation. Neurologic sub-score was excluded because it is subjective, inconsistently documented, and confounded by the frequent use of sedation and paralysis in severe COVID-19. ECMO patients were included and assigned the maximum respiratory sub-score. Patients started on RRT at an outside hospital, found to be incidentally COVID-positive, or on chronic dialysis were excluded. P values were obtained using 1-sided Mann-Whitney U tests. RESULT(S): 63 total COVID-19 patients on CRRT were identified with 73% 30-day mortality and 83% in-hospital mortality. The median 4OSS was 8 in both in-hospital survivors and non-survivors with interquartile range [IQR] of 4-9 and 7-9.75, respectively (difference between groups non-significant, p = 0.075). The median 4OSS was 7 [5.5- 8.5] and 8 [7-10] in 30-day survivors and non-survivors, respectively (p = 0.018). Those with 4OSS of >=10 (n=13, 20.6%) had 100% in-hospital mortality. CONCLUSION(S): Similar to other analyses of SOFA score in COVID-19, 4OSS at CRRT initiation in patients with COVID-19 and AKI-RRT appears to have limited prognostic ability, with substantial overlap in scores between survivors and non-survivors. However, while additional multicenter studies are needed, 4OSS of >=10 may identify a group of about 20% of COVID-19 patients with AKI-RRT and mortality approaching 100%. Given the absence of a superior validated metric, a 4OSS of >=10 may be a reasonable tool for triage of CRRT in the setting of crisis standards of care and CRRT machine or supply shortages. At a minimum, 4OSS could inform goals of care discussions prior to CRRT initiation in patients with COVID-19 complicated by AKI-RRT.
ABSTRACT
RATIONALE Acute kidney injury requiring renal replacement therapy (AKI-RRT) in the intensive care unit (ICU) is associated with significant mortality, with short-term death rates often exceeding 50% in modern cohtorts.1 Similarly high mortality with AKI-RRT has been reported in multiple U.S. cohorts of patients with coronavirus disease 2019 (COVID-19)2-4, but none have specifically focused on the outcomes of AKI treated with continuous RRT (CRRT) in the ICU or compared the outcomes of AKI-CRRT to COVID-negative controls. METHODS We carried out a retrospective review of all patients admitted to the University of New Mexico Hospital and initiated on CRRT in January to October 2020 and compared outcomes between those with and without symptomatic COVID-19. Patients felt to be incidentally infected with COVID-19 and those with end-stage kidney disease (ESKD) were excluded. Crude death rates in AKI-CRRT patients with and without COVID-19 were compared by chisquared test. Patients discharged before 30 days were assumed alive at 30 days. RESULTSA total of 102 patients were treated with 103 CRRT treatments over the 10-month period. Of these, two felt to be incidentally infected were excluded. Ten with ESKD, including three with COVID-19, were also excluded. Of the remaining 90 with AKI-CRRT, 30 were treated for symptomatic COVID-19 starting in April 2020 and had 30-day and in-hospital mortality rates of 67.7% and 80.0%, respectively. Of the 60 COVID-19-negative patients with AKI-CRRT, the 30-day and in-hospital mortality rates were 58.3 and 63.3%, respectively (p = 0.44 and = 0.11, respectively, versus COVID-positive patients). When broken into pre-pandemic and post-pandemic groups, the 30-day and in-hospital death rates for AKI-CRRT in COVID-negative patients were 56.5% and 60.9% in January to March and 59.5% and 64.9% in April to October, respectively (p = >0.05 for both comparisons). CONCLUSIONS These data confirm the high mortality associated with AKI-CRRT in the setting of severe COVID-19. Though not statistically significant in this limited sample, the trend for higher in-hospital mortality in COVID-19 patients suggests the mortality of AKI-CRRT in this setting may be higher than other ICU patients. Notably, the mortality of AKI-CRRT in COVIDnegative patients did not significantly differ before and after the start of the pandemic. Overall, while conclusions about the independent effect of COVID-19 are limited with these unadjusted data, awareness of the high mortality of AKI-CRRT in the setting of COVID-19 may be useful in discussing prognosis and goals of care in these patients.