Metabolic disturbances potentially attributable to clogging during continuous renal replacement therapy

Abstract

Background: Clogging is characterized by a progressive impairment of transmembrane patency in renal replacement devices and occurs due to obstruction of pores by unknown molecules. If citrate-based anti-coagulation is used, clogging can manifest as a metabolic alkalosis accompanied by hypernatremia and hypercalcemia, primarily a consequence of Na₃Citrate infusion. An increased incidence of clogging has been observed during the COVID-19 pandemic. However, precise factors contributing to the formation remain uncertain. This investigation aimed to analyze its incidence and assessed time-varying trajectories of associated factors in critically ill patients on continuous renal replacement therapy (CRRT).

Methods: In this retrospective, single-center data analysis, we evaluated COVID-19 patients undergoing CRRT and admitted to critical care between March 2020 and December 2021. We assessed the proportional incidence of clogging surrogates in the overall population and subgroups based on the specific CRRT devices employed at our institution, including multiFiltrate (Fresenius Medical Care) and Prismaflex System (Baxter). Moderate and severe clogging were defined as Na > 145 or ≥ 150 mmol/l and HCO₃^- > 28.0 or ≥ 30 mmol/l, respectively, with a total albumin-corrected calcium > 2.54 mmol/l. A mixed effect model was introduced to investigate factors associated with development of clogging.

Results: Fifty-three patients with 240 CRRT runs were analyzed. Moderate and severe clogging occurred in 15% (8/53) and 19% (10/53) of patients, respectively. Twenty-seven percent (37/136) of CRRTs conducted with a multiFiltrate device met the criteria for clogging, whereas no clogging could be observed in patients dialyzed with the Prismaflex System. Occurrence of clogging was associated with elevated triglyceride plasma levels at filter start (p = 0.013), amount of enteral nutrition (p = 0.002) and an increasing white blood cell count over time (p = 0.002).

Conclusions: Clogging seems to be a frequently observed phenomenon in critically ill COVID-19 patients. The presence of hypertriglyceridemia, combined with systemic inflammation, may facilitate the development of an impermeable secondary membrane within filters, thereby contributing to compromised membrane patency.

Keywords: COVID-19; CRRT; Dialysis; Hypertriglyceridemia; ICU.

Fig. 1

Mechanism of clogging. Representation of a patent hemodialysis filter A with clearance up to 60% of calcium–citrate complexes. B Occurrence of clogging leads throughout formation of an impermeable secondary membrane to an impairment of clearance capacity of the filter with consecutive rise of sodium, bicarbonate and calcium in the patient

Fig. 2

Screening procedure. GC = general consent for the use of health-related date for research purposes

Fig. 3

Patients and individually conducted CRRT runs. Numeration on the left side represent patients, each square is a conducted CRRT run without clogging (yellow), transient metabolic disturbances indicative for clogging (turquoise) and persistent signs of clogging until end of filtration (blue). CRRTs with metabolic disturbances already in the first measurements were not included in the final analysis, since no distinguishment between new rapid onset of clogging and pre-existing metabolic alterations without new clogging could be made (red boarders)

Fig. 4

Time-event analysis for clogging of two studied CRRT devices. Patients submitted to CRRT with an mFT device and CVVHD modality had an increased risk to develop clogging during duration of treatment (solid red line). No signs of clogging could be observed in patients treated with CVVHDF using PFS devices (solid blue line). Competing events (discontinuation of CRRT and death) are represented as dashed and dotted lines

Fig. 5

Time-varying trajectories of metabolic alterations. The figure summarizes the course of metabolic markers over time for every conducted filter run. Only values were included, which were measured during a CRRT run. Maximal run time was 72 h as per institutional protocol. Clogging filters (blue line) showed persistently higher sodium concentrations in the plasma A with increasing bicarbonate levels over time B. Corrected_Alb Ca was only slightly increased in filters allocated to the clogging group C. Whereas calcium substitution (CaSR) did not differ between overall clogging and non-clogging filters at the beginning of CRRT, a decline of substitution rate could be observed in filters with first onset of clogging (first-clogging, blue line in D)

Fig. 6

Factors associated with clogging formation—overall CRRT runs. Trajectories over time of factors associated with the formation of clogging (blue line) in comparison with filters without clogging (red line) in all assessed CRRT runs. The association of clogging with increasing WBC A without any concomitant increase in CRP levels B indicates an early inflammatory response phase. Triglyceride levels were remarkable elevated during CRRT runs with clogging C

Fig. 7

Factors associated with clogging formation in the first-clogging subgroup. While clogging was found to be associated with consistently elevated triglyceride plasma levels in the overall cohort, subgroup analysis of the first-clogging subgroup revealed an association with decreasing plasma levels over time A. In addition, first-clogging was attributed to reduced propofol exposure B and an increased enteral nutritional intake C. The quantity of enteral nutrition and propofol administered was measured as the cumulative dose received within the preceding 24 h, assessed at the initiation of the filter (T0), as well as at 24 h (T24) and 48 h (T48) thereafter