Fluids for Continuous Renal Replacement Therapy – an Evaluation of Microbial Integrity

Introduction

The potential for clinically significant trans-membrane transfer of pyrogens is well recognized in chronic hemodialysis (HD) and has led to the establishment of strict standards for fluid purity in this setting. Little attention had been paid to the scale of the problem in continuous renal replacement therapy (CRRT), although preliminary evidence has indicated the potential for contamination of the bicarbonate-based dialysis fluid circuit (1). We have subsequently confirmed frequent breaches of microbial integrity in industry-standard, bicarbonate-based, continuous veno-venous hemofiltration (CVVH), with widespread evidence of both bacterial growth and biofilm formation in the replacement fluid circuit (2).

These findings are of particular concern given the need for systemic infusion of replacement fluid into vulnerable, critically-ill patients. The optimal CRRT technique for minimizing microbial contamination remains unclear. It is also unclear whether alternative CRRT fluids might help mitigate the problem.

We compared the effects of commonly used CRRT fluids on in vitro bacterial growth and biofilm formation.

Materials and Methods

We compared the effect of 5, commonly-used CRRT fluids (Monosol S - lactate-based solution (Baxter Healthcare, Deerfield, Illinois, USA), Accusol 35 - bicarbonate-based solution (Baxter Healthcare), Prismocitrate - citrate-based anticoagulant (Gambro, Lund, Sweden), 4% tri-sodium citrate (Fresenius, Bad Homburg, Germany) and Ci-Ca K2 - bicarbonate-based, calcium-free dialysis fluid (Fresenius) on bacterial growth and biofilm formation. Their chemical compositions and intended uses are shown in Table I. All of the experiments described below were performed in duplicate on separate occasions to ensure that results were reproducible.

OPEN IN VIEWER

Table I

Chemical composition and intended use of continuous renal replacement therapy solutions

	Prismocitrate 2	Monosol S 1	Accusol 35 1	Tri-sodium citrate 3	Ci-Ca K2 3
Intended Use	Regional Citrate Anticoagulant	Lactate-based Dialysate/Replacement Solution	Bicarbonate-based Dialysate/Replacement Solution	Regional Citrate Anticoagulant	Bicarbonate-based, Calcium-Free Dialysate-used with regional citrate anticoagulant
Sodium	-	140 mmol/l	140 mmol/l		133 mmol/l
Potassium	-	-	2.0 mmol/l	-	2.0 mmol/l
Calcium	-	1.75 mmol/l	1.75 mmol/l	-	0.0 mmol/l
Magnesium	-	0.75 mmol/l	0.50 mmol/l	-	0.75 mmol/l
Chloride	-	115 mmol/l	111.5 mmol/l	-	116.5 mmol/l
Bicarbonate	-	-	35 mmol/l	-	20.0 mmol/l
Lactate	-	30 mmol/l	-	-	-
Glucose	-	5.55 mmol/l	5.55 mmol/l	-	-
Osmolarity	-	293 mOsm/l	296 mOsm/l	-	-
Citrate	10 mmol/l	-	-	136 mmol/l	-
Citric Acid	2 mmol/l	-	-	-	-

1

Baxter Healthcare, Deerfield, Illinois, USA;

2

Gambro, Lund, Sweden;

3

Fresenius, Bad Homburg, Germany.

Results

A sterility assessment of at least 4 bags of each type of CRRT fluid found no microbial contamination after 10 days incubation of culture media. For growth experiments, viable counts confirmed the high inoculum as a final concentration of bacterial cells within the range: 3.5 – 6 × 10⁵ CFU/ml. Low inocula were within the range: 60–180 CFU/ml.

Fluids without nutrient supplementation did not support bacterial growth or biofilm formation during 72 h incubation at either 22°C or 37°C, even when a relatively high inoculum was used.

The incorporation of nutrients allowed insights into the ability of CRRT fluids to affect growth and biofilm formation. All strains grew well after 24 h incubation at 37°C in control wells (i.e., IsoSensitest broth or tryptone soya broth reconstituted with deionized water rather than CRRT fluids).

Figure 1 shows the increase in spectrophotometric absorbance at 48 h incubation with higher values, indicating greater bacterial growth. Results are presented as the averages for each duplicate experiment.

OPEN IN VIEWER

Fig. 1

Growth of various bacteria after incubation at 37°C for 48 h in various continuous renal replacement therapy fluids supplemented with IsoSensitest broth.

The CRRT solutions showed varying degrees of bacterial growth inhibition. Accusol 35 severely restricted the growth of B. cepacia, S. maltophilia, P. fluorescens and S. aureus, whereas Ci-Ca K2 completely inhibited the growth of B. cepacia and S. aureus, and tri-sodium citrate completely inhibited the growth of both staphylococcal strains that were used. Monosol S and Prismocitrate showed the least evidence of bacterial inhibition. At 22°C, the growth of several bacteria was severely limited and no firm conclusions on the impact of CRRT fluids could be made (data not shown).

All 7 strains of bacteria used were able to form biofilms and were classified as showing weak, moderate, or strong adherence according to the criteria of Stepanovic (3). Table II shows the effects of the CRRT solutions on biofilm formation. E. coli was strongly adherent and retained moderate to strong adherence in the presence of all CRRT fluids. All other strains were affected to varying degrees by the presence of CRRT fluids. For example, the weak biofilm-forming ability of both staphylococci was abolished in the presence of Prismocitrate, Accusol 35, and tri-sodium citrate but was unaffected by Monosol S. Prismocitrate resulted in greater suppression of biofilm formation across the range of tested bacterial strains than other fluids.

OPEN IN VIEWER

Table II

Biofilm formation by various bacteria after 24 h incubation at 37°c with various supplemented continuous renal replacement therapy fluids

Bacterial strain	Broth only	Prismocitrate	Monosol S	Accusol 35	Trisodium citrate	Ci-Ca K2
Burkholderia cepacia	Weak	-	Weak	-	-	-
Stenotrophomonas maltophilia	Moderate	-	Moderate	-	-	Weak
Pseudomonas fluorescens	Weak	-	Weak	-	Moderate	-
Escherichia coli	Strong	Moderate	Strong	Strong	Strong	Strong
Pseudomonas aeruginosa	Moderate	-	Moderate	Weak	-	Weak
Staphylococcus epidermidis	Weak	-	Weak	-	-	Weak
Staphylococcus aureus	Weak	-	Weak	-	-	-
No bacteria	-	-	-	-	-	-

Discussion

Bacterial contamination of dialysis fluid has been well recognized in chronic HD for decades (4). The potential for clinically significant cross-dialyzer transfer of pyrogens has led to the establishment of strict standards for water and dialysis fluid purity (4). The International Organization for Standardization (ISO) requires that standard dialysis fluid contains a total viable microbial count of <100 CFU/ml and an endotoxin concentration <0.25 EU/ml, while the more stringent requirements for ultrapure fluid for convective therapies are <0.1 CFU/ml and <0.03 EU/ml, respectively (5), given the need for systemic infusion of the preparation.

Maintenance of optimal water quality now requires considerable infrastructural investment by renal services, whether supply is from a central water plant or from portable reverse osmosis units used for treatment remote from the main dialysis unit – for instance, in critical care areas. Although readily available, these standards have yet to be applied to CRRT – at least, in part, because the scale and clinical relevance of fluid contamination remains incompletely defined in this setting. The first indication that a problem existed was in an evaluation of a fractional dialysis fluid quantification method for continuous veno-venous HD (CVVHD) (6). Four of fifteen dialysis fluid collections, cultured for the presence of urease-producing bacteria, revealed heavy contamination with P. aeruginosa.

In a subsequent study, the bacterial integrity of this same, bicarbonate-based, CVVHD technique was specifically assessed (1). Fluid sampled from source dialysis fluid bags showed no clear evidence of contamination. However, 6 of the 18 dialysis fluid circuits studied during the course of treatment showed heavy microbial growths, most commonly of the types of environmental bacteria involved in chronic HD contamination. These were detected after an average of around 50 h of therapy, although in two cases, they were detected within 24 h of treatment initiation. Contamination appeared to spread in an antegrade fashion and, once present, positive cultures persisted until the end of therapy. The main caveat to this study was that therapy could have been regarded as “non-standard” with the use of multiple-component circuitry and with bicarbonate dialysis fluid manufactured in-house.

We therefore went on to assess the microbiological integrity of CVVH performed using fully integrated, industry-standard circuitry and industry-manufactured bicarbonate replacement fluid (2). Of the replacement fluid samples taken at the end of the 24 studied therapies (the lack of available ports precluded this, during treatment), 9 cultures (using highly sensitive techniques) and 1 endotoxin assay, breached standards for ultrapure water (4). Electron microscopy revealed the luminal surfaces of 13 harvested tubing samples to be contaminated with biofilm. Only 7 circuits proved to be entirely free from microbial contamination.

The results of these studies raises a number of questions, namely:

Bacterial detection techniques used in our previous work (2) were highly sensitive and compatible with recommendations for assurance of ultrapure water quality. The techniques used in the present study (including nutrient supplementation) had a different purpose – to provide insights into the relative impact on bacterial activity of different fluids that might not have been differentiated at the lower (but nevertheless potentially clinically relevant) levels of contamination seen in our previous work. In addition, although the compositions of the regional citrate anticoagulants tested (Prismocitrate and tri-sodium citrate) would not have reflected solute concentrations in the final circulating fluid product, their inclusion, we felt, allowed insights into any discrete local effects on bacterial activity.

Our results revealed that none of the three dialysis fluid/ replacement solutions (Monosol S, Accusol 35, Ci-Ca K2) showed consistent suppression of bacterial growth across tested species. Given that evidence of contamination in our previous studies had occurred in the presence of bicarbonate-buffered solutions, the fact that Accusol 35 seemed to result in greatest overall suppression of bacterial growth raises concerns that alternative, non-bicarbonate solutions may provide a more favorable environment for bacterial activity in clinical practice.

In addition, we speculated upon the possibility of at least a local inhibitory effect of Prismocitrate and tri-sodium citrate on bacterial activity before dilution to low concentrations within the final fluid product. We found neither to have a consistent impact on either bacterial growth or biofilm formation.

Conclusions

Different CRRT fluids appear to have different impacts on bacterial growth and biofilm formation. However, all remain susceptible to extrinsic contamination, which must therefore be a consideration in future CRRT technology and practice.