Simplified regional citrate anticoagulation protocol for CVVH, CVVHDF and SLED focused on the prevention of KRT-related hypophosphatemia while optimizing acid-base balance

Abstract

Background: Hypophosphatemia is a common electrolyte disorder in critically ill patients undergoing prolonged kidney replacement therapy (KRT). We evaluated the efficacy and safety of a simplified regional citrate anticoagulation (RCA) protocol for continuous venovenous hemofiltration (CVVH), continuous venovenous hemodiafiltration (CVVHDF) and sustained low-efficiency dialysis filtration (SLED-f). We aimed at preventing KRT-related hypophosphatemia while optimizing acid-base equilibrium.

Methods: KRT was performed by the Prismax system (Baxter) and polyacrylonitrile AN69 filters (ST 150, 1.5 m2, Baxter), combining a 18 mmol/L pre-dilution citrate solution (Regiocit 18/0, Baxter) with a phosphate-containing solution (HPO42- 1.0 mmol/L, HCO3- 22.0 mmol/L; Biphozyl, Baxter). When needed, phosphate loss was replaced with sodium glycerophosphate pentahydrate (Glycophos™ 20 mmol/20 mL, Fresenius Kabi Norge AS, Halden, Norway). Serum citrate measurements were scheduled during each treatment. We analyzed data from three consecutive daily 8-h SLED-f sessions, as well as single 72-h CVVH or 72-h CVVHDF sessions. We used analysis of variance (ANOVA) for repeated measures to evaluate differences in variables means (i.e. serum phosphate, citrate). Because some patients received phosphate supplementation, we performed analysis of covariance (ANCOVA) for repeated measures modelling phosphate supplementation as a covariate.

Results: Forty-seven patients with acute kidney injury (AKI) or end stage kidney disease (ESKD) requiring KRT were included [11 CVVH, 11 CVVHDF and 25 SLED-f sessions; mean Acute Physiology and Chronic Health Evaluation II (APACHE II) score 25 ± 7.0]. Interruptions for irreversible filter clotting were negligible. The overall incidence of hypophosphatemia (s-P levels <2.5 mg/dL) was 6.6%, and s-P levels were kept in the normality range irrespective of baseline values and the KRT modality. The acid-base balance was preserved, with no episode of citrate accumulation.

Conclusions: Our data obtained with a new simplified RCA protocol suggest that it is effective and safe for CVVH, CVVHDF and SLED, allowing to prevent KRT-related hypophosphatemia and maintain the acid-base balance without citrate accumulation.

Trial registration: NCT03976440 (registered 6 June 2019).

Keywords: acute kidney injury; continuous venovenous hemofiltration; kidney replacement therapy; regional citrate anticoagulation; sustained low-efficiency dialysis.

Graphical Abstract

Figure 1:

A simplified KRT circuit with RCA for CVVH, CVVHDF and SLED. The low concentration citrate solution is infused in pre-dilution in proportion to Qb, aiming at a target estimated citrate concentration in the hemofilter of 2.5–3 mmol/L. In CKRT modalities, a calcium-free and bicarbonate and phosphate containing solution acting as post-dilution (A) and/or dialysis fluid (B) was adopted and set to prescribed dialysis dose of 30–35 mL/kg/h. In PIKRT, as SLED-f, the same solution was used as dialysis fluid and combined with a standard KRT solution for post-dilution replacement fluid (C) for a prescribed effluent volume of 100 mL/min. In all KRT modalities, an infusion of calcium chloride (10% solution) in a central venous line was scheduled to maintain the systemic ionized calcium (s-Ca²⁺) within the normal range. Qb, blood flow rate.

Figure 2:

Trend of serum citrate levels in critically ill patients with different MELD scores undergoing CVVH, CVVHDF and SLED. Serum citrate levels at 0, 4, 8, 24 and 48 h of KRT in patients with a MELD score ≤25 and ≥25.

Figure 3:

Serum phosphate levels trend during KRT treatment period in all patients and in subgroups of patients stratified according to baseline values. Trend of serum phosphate levels in all patients undergoing CVVH, CVVHDF and SLED (A), in baseline hyperphosphatemic patients (B) and in hypophosphatemic patients (C). Data are reported as mean and SD.