Clinical and biological profiles associated with the time of occurrence of citrate accumulation in patients receiving continuous renal replacement therapy

Abstract

Background: Citrate accumulation (CA) is a feared complication in critically ill patients undergoing regional citrate anticoagulation (RCA) for continuous renal replacement therapy (CRRT). This study aimed to describe the characteristics of patients presenting CA within a large cohort of unselected critically ill patients receiving RCA-CRRT depending on the time of occurrence of CA after CRRT initiation.

Methods: This retrospective, multicenter observational study performed in nine intensive care units (ICU) in Lyon, France, included patients treated with RCA-CRRT between January 2020 and January 2022. CA was defined by a total to ionized calcium ratio (tCa/iCa) ≥ 2.3 associated with hypocalcemia and metabolic acidosis.

Results: Among 2080 patients, 76 (3.7%) developed CA: 69 (91%) experienced CA within 24 h after CRRT initiation (initial CA) and 7 (9%) after 24 h (late-onset CA). Only lactate levels at CRRT initiation differed between patients with initial CA and those with late-onset CA (10 mmol/l [4.6-16] vs 1.4 mmol/l [1-3.8], p = 0.006 respectively). In the initial CA group, 39 (57%) exhibited signs of CA within 6 h or less (immediate CA) and 30 (43%) showed signs of CA between 6 and 24 h. Over the first 24 h, patients with initial CA presented a marked increase in lactate levels, worsening norepinephrine requirements, persistent elevation of the tCa/iCa ratio, decrease in prothrombin time, and increase in transaminases. Patients with immediate CA showed higher lactate concentration and more severe metabolic acidosis at CRRT initiation compared to patients with early CA whereas other markers did not differ significantly between the two groups. The area under the Receiver Operating Characteristic curve of lactate and pH for predicting immediate citrate accumulation were 0.75 [0.62 - 0.87] and 0.74 [0.62 - 0.85] respectively, with optimal cutoff values of 10.6 mmol/L and 7.14 respectively. The ICU mortality rate among patients with CA was 97% compared to 55% in the whole cohort.

Conclusions: CA is a rare phenomenon in patients under RCA-CRRT. Severe metabolic acidosis with hyperlactatemia at CRRT initiation is the most relevant marker to identify patients at risk of immediate CA and should encourage close monitoring of tCa/iCa ratio.

Keywords: Citrate accumulation; Continuous renal replacement therapy; Lactate; Regional citrate anticoagulation.

Fig. 1

Study flow chart. Criteria defined by Khadzhynov et al.: (1) decrease of systemic iCa (< 1.1 mmol/L), (2) concomitant increase of tCa concentration and, thus, increase of tCa/iCa ratio ≥ 2.3, (3) relevant metabolic acidosis (pH < 7.2 and/or BE < − 5 mmol/L) or (4) an increase of anion gap (> 12 mmol/L) RCA-CRRT: Regional citrate anticoagulation-continuous renal replacement therapy

Fig. 2

Changes in markers of shock, calcium homeostasis, and liver failure during the first 24 h after CRRT initiation a. Lactate concentration (mmol/L) b. pH c. Norepinephrine dose (µg/Kg/min) d. Ionized calcium (mmol/L) e. Total to ionized calcium ratio f. Calcium substitution (mmol/L of blood) for patients under CVVHD g. Alanine aminotransferase (IU/L) h. Aspartate aminotransferase (IU/L) i. Prothrombin time (%) j. Bilirubin (µmol/L). Results are represented as mean values along with their 95% confidence intervals displaying the predicted marginal means for each time-point calculated from linear mixed model analysis. Statistically significant differences in changes from baseline (CRRT initiation, H0) are indicated by *p < 0.05, **p < 0.01, *** p < 0.001.

Fig. 3

Changes in markers of interest in patients with immediate and early citrate accumulation a. Lactate concentration (mmol/L) b. pH. Results are represented as mean values along with their 95% confidence intervals displaying the predicted marginal means for each time-point calculated from linear mixed model (LMM) analysis. Pairwise comparison tests were performed between groups and time-points using estimated marginal means derived from the LMM. Adjustments for multiple comparisons were made using the Bonferroni correction method. Statistically significant differences between the"early CA"and"immediate CA"groups are indicated by *p < 0.05, **p < 0.01, *** p < 0.001. c. ROC curve of lactate at CRRT initiation for prediction of immediate CA. d. ROC curve of pH at CRRT initiation for prediction of immediate CA. CA: citrate accumulation. CRRT: continuous renal replacement therapy. Se: Sensibility. Spe: Specificity