ABSTRACT
Background: Persistent hyperkalemia (hyperK) and hyperphosphatemia (hyperP) despite renal replacement therapy (RRT) was anecdotally reported in COVID-19 and acute kidney injury (AKI) requiring RRT (CoV-AKI-RRT). However, observation bias could have accounted for the reports. Thus, we systematically examined the rate and severity of hyperK and hyperP in patients with CoV-AKI-RRT in comparison with the pre-COVID-19 era. Methods: We identified patients with CoV-AKI-RRT treated with sustained low-efficiency dialysis (SLED) for ≥2 days in March-April 2020. As pre-COVID-19 control, we included patients with AKI treated with SLED in December 2019. We examined the rates of hyperK (serum potassium [sK] ≥5.5 mEq/L), severe hyperK (sK ≥6.5 mEq/L), hyperP (serum phosphate [sP] ≥4.5 mg/dl), and moderate or severe hyperP (sP ≥7-10 and >10 mg/dl, respectively) as %SLED-days with an event. Results: Along the duration of SLED, the incidence of hyperK was greater in CoV-AKI-RRT (n=64; mean 19%±2% versus 14%±3% SLED-days, P=0.002) compared with control (n=60). The proportion of patients with one or more event of severe hyperK was greater in CoV-AKI (33% versus 7%, P<0.001). The incidence of hyperP was similar between groups (mean 56%±4% versus 53%±5% SLED-days, P=0.49). However, the proportion of patients with one or more event of moderate and severe hyperP was greater in CoV-AKI-RRT (86% versus 60%, P=0.001, and 50% versus 18%, P<0.001, respectively). Among those with CoV-AKI-RRT, sK and sP correlated with lactate dehydrogenase (LDH; r=0.31, P=0.04, and r=0.31, P=0.04, respectively), whereas hyperP also correlated with shorter SLED runs (hours/run; r=-0.27, P=0.05). Conclusions: Refractory hyperK and hyperP were more frequent in CoV-AKI-RRT compared with the pre-COVID-19 era. Because of the correlation of sK and sP with higher LDH and sP with shorter SLED runs, intracellular ion release from cell injury due to cytokine storm and RRT interruptions may account for the findings.