Journal of Clinical Pharmacy and Therapeutics (1990) 15,53-58.

USE OF ENDOTOXIN RETENTIVE INTRAVENOUS FILTERS WITH PAEDIATRIC TOTAL PARENTERAL NUTRITION

SOLUTIONS

C. Richards and P. Thomas* Sterile Product Unit, Pharmacy Department, The University Hospital of Wales, Heath Park, Cardiff and

The Pharmaceutical Unit, St Mary’s Hospital, Penarth, South Glamorgan, Wales

SUMMARY

We studied the retention of endotoxin released in a clinically relevant form by senescent bacteria during filtration of amino acid-based total parenteral nutrition (TPN) solutions by the PALL Intravenous Set Saver Filter (ELD96).

The ELD96 was shown to be capable of providing an effluent free of detectable endotoxin ( < 0.125 endotoxin unit-equivalents/ml) when chal- lenged with lo6 total cells of Pseudomonas aeruginosa during simulated clinical T P N infusion. Parallel control experiments showed that endotoxin was released from the test organism and penetrated through conventional intravenous (i.v.) solution filters.

I N T R O D U C T I O N

Extrinsic microbial contamination of intravenous (i.v.) fluid lines is well documented (1,2). The extrinsic contamination can originate from the preparation of the i.v. fluids or at the administration stage.

It has been demonstrated that total parenteral nutrition (TPN) fluids will support the growth of bacteria and fungi (3, 4), although growth rates of individual organisms are variable and specific for each species. Scott et a1.(5) reported that, in an infant surgical unit where T P N was under the care of a nutrition team, samples of T P N fluids and giving sets were examined for microbiological contamination at the end of the 24-h administration period. Contamination was found in eight of 98 systems examined from eight patients.

The Special Care Baby Unit (SCBU), The University Hospital of Wales, Heath Park, Cardiff, is increasing its use of percutaneous silastic central venous lines (epicutaneo- cava-ketheter, Vygon) for the administration of T P N fluids. The contamination of these i.v. lines with micro-organisms can have serious consequences in the neonates. The use of an effective 0.2 pm-rated filter might therefore be of benefit during the adminstration of T P N fluids which do not contain lipid. To avoid potential hazards of bacterial penetration through the filter and to give endotoxin protection during continuous infusion, it has been recommended that the filter should be replaced daily (6).

53

54 C. Richards and P. Thomas

Recently, an end-line i.v. infusion filter (ELD96, Pall Biomedical, Portsmouth) became available which is claimed to provide 96-h patient protection from inadvertent microbial and associated endotoxin contamination. This filter contains a 0.2 pm-rated ‘Posidyne@” membrane which is intrinsically positively charged across a wide range of pH values and gives the filter its endotoxin-retentive capacity. As T P N fluids contain a high level of electrolytes and are at a low pH (approximately pH 5.5), it was considered necessary to test the ability of the ELD96 to retain endotoxins over a 96-h period with continuous T P N infusion before their introduction to the SCBU.

M A T E R I A L S A N D M E T H O D S Total parenteral nutrition solutions

University Hospital of Wales. The T P N formulation used (Table 1) was typical of those used on the SCBU,

Table 1. Formula for paediatric T P N solution (300 ml)

Constituent Concentration

Nitrogen 0.5 g Glucose 120 kcal Sodium 10-2 mmol Potassium 5-9 mmol Calcium 1.2 mmol Magnesium 0-3 mmol Phosphate 1-4 mmol Chloride 16-4 mmol Zinc 7-7 pmol Copper 0.5 pmol Vitamins 2 ml (Solivito N)

Filter The filter studied was the PALL Intravenous Set Saver (ELD96LL; Pall Biomedical

Limited, Portsmouth, Hampshire). The filter contains 0.2 pm-rated ‘Posidyne’ media and has a priming volume of about 2 ml.

Preparation of inoculum A suspension of Pseudomonas aeruginosa 8626 (Nation Collection of Industrial

Bacteria, Torry Research Station, Aberdeen) was prepared by inoculating 100 ml tryp- tone soya broth with a single colony from a tryptone soya agar plate and incubating this culture overnight in a shaking water bath at 37°C. Aliquots of the overnight culture were centrifuged (2,600 g for 10 min). The supernatant was discarded and the cell pellet resuspended in approximately 3 ml of the T P N solution to be tested. The washed cells

Endotoxin filters and TPN solutions 55

were collected by re-centrifugation (2600g for 10 min). The washed cell pellet was resuspended in sufficient T P N solution to give a cell suspension of approximately 1 x lo7 bacteria/ml.

The total bacterial count was assessed using a counting chamber (Neubauer haemo- cytometer-BDH Limited, Dagenham, Essex) and then the cell suspension diluted appropriately with T P N solution to give lo5 cells/ml. This represented the working challenge suspension.

Evaluation of ELD96 endotoxin retention A 300-ml bag of T P N solution was aseptically connected to an i.v. solution adminis-

tration set and an ELD96. The administration set and ELD96 were primed with T P N solution. The administration tubing was inserted into an i.v. solution pump (Ivac 531 infusion pump-Ivac U.K., Harrow) and the pump was set to deliver at a flow rate of 4 ml/h. The outlet tubing of the ELD96 was placed into a sterile, pyrogen-free con- tainer, the pump was started and approximately 2 ml of T P N solution were collected. The pump was stopped and the bag of T P N solution was inoculated with 3 ml of the bacterial challenge suspension, to give a final concentration of 1 x lo3 total cells/ml. The bag was inverted several times to ensure adequate mixing of the bacteria with the TI" solution. Flow through the ELD96 was initiated at 4 ml/h and the effluent from the filter was collected in a sterile, pyrogen-free container.

Immediately after initiating flow through the filter, a 1-ml sample was withdrawn from the bag of T P N solution. This sample was serially diluted and 0.1-ml aliquots of appropriate dilutions plated on to tryptone soya agar and incubated for 48 h at 30°C. The number of cfu/ml was counted after incubation. This represented the viable count of bacteria in the bag of T P N solution.

After a 24 h flow of 250 ml T P N solution through the ELD96, a sample of filtrate was collected in a sterile, pyrogen-free test tube. This sample was analysed for the presence of endotoxin using the LAL assay. In addition, a I-ml sample was withdrawn from the bag of T P N solution. This sample was serially diluted and 0.1 -ml aliquots of appropri- ate dilutions plated on to tryptone soya agar and incubated (48 h at 30°C). The number of cfu/ml was counted after incubation. This represented the viable count of bacteria in the bag of T P N solution after 24 h.

The above method was repeated daily for a total of 4 days, using the same ELD96. Control experiments were performed following the method outlined above using a

0.2-pm rated cellulose acetate membrane filter not manufactured by Pall. This was to confirm that the passage of endotoxins through a non-charged filter medium would be detectable using this procedure.

All endotoxin assays were performed using glassware depyrogenated by heating at 260°C for 4 h. The filtrate samples (0.1 ml) were mixed with 0.1 ml of Limulus amoebo- cyte lysate (Atlas Bioscan Limited, Bognor Regis, West Sussex) reconstituted with Pyrosol LAL reconstitution buffer (Atlas Bioscan Limited) and incubated in a vibration-free water bath at 37°C for 60 min. The formation of a clot indicated a pyro- genic reaction. Positive controls, using control-standard endotoxin (Atlas Bioscan Limited), were performed with each experiment to confirm the sensitivity of the lysate. This was 0.125 endotoxin units (EU)/ml, equivalent to 12.5 pg/ml Escherichza coli 0,55:B5 endotoxin. Negative controls were performed using the T P N solution. As Ps. aeruginosa was used as the challenge organism and as the endotoxin released is different

56 C . Richards and P. Thomas

Table 2. Daily viable counts of Ps. aeruginosa for a typical filter tested

Time Viable counts Day (h) (cfuiml)

1 0 24

2 0 24

3 0 24

4 0 24

2.9 x 103

1.2 x 103 6.0 x 10'

9.5 x 101 9.1 x lo2 3.4 x 102 2.5 x 10' 1.3 x lo2

Table 3. Total Ps. aeruginosa challenge for each filter tested

Total 2's. aeruginosa Filter number challenge (cfu)

4.2 x lo6 4.5 x 106 4.0 x lo6 2.5 x lo6 1.9 x lo6 1.9 x lo6 1.9 x lo6 2.3 x lo6 1.5 x lo6

Table 4. Endotoxin concentration determined downstream of 0.2-pm rated cellulose acetate filter

Endotoxin concentrations detected Time downstream of cellulose acetate (h) filter (EUE/ml)

0 24 48 72 96

0 1.25

400 3,200 6,400

Endotoxin filters and TI” solutions 57

from the E. coli055:B5 endotoxin standard used for the LAL assay, all values quoted are as endotoxin unit-equivalents (EVE).

R E S U L T S A N D D I S C U S S I O N S

Control experiments demonstrated that the T P N solution analysed neither inhibited nor enhanced the LAL assay used.

The data relating the challenge of the ELD96 with Ps. aeruginosa are shown in Tables 2 and 3. No endotoxin (<0.125 EUE/ml) was found downstream of ELD96 filters in nine separate experiments. Table 2 shows the daily bacterial challenge for a typical filter tested. It can be seen that the viable count decreased over each 24-h period indicating that a proportion of the bacteria inoculated was dying. The total bacterial challenge for each filter was approximately lo6 (Table 3) . This was calculated by multiplying the daily bacterial challenge/ml (approximately 1 x 103/ml) by the infusion volume (250 ml) by the number of days challenge (4).

Thus, the study showed clearly that the ELD96 is capable of retaining endotoxin (released from cells of Ps. aeruginosa) during simulated clinical infusion of paediatric T P N solutions.

The experiments, performed using a 0.2-pm rated cellulose acetate membrane filter, confirm that endotoxin was produced in the experimental model. Table 4 shows the levels of endotoxin measured downstream of the cellulose acetate filter over a 4-day challenge period. Thus, the absence of detectable amounts of endotoxin ( < 0.125 EEU/ ml) downstream of the ELD96 is a genuine reflection of endotoxin retention.

Ps. aeruginosa was used in this study because it was impractical to use E. coli as the challenge organism. Prelminary investigations using E. coli 055:B5 had shown that the growth of this organism in the T P N solutions studied was so rapid and varied as to make controlled experimental conditions impossible. The rate of bacterial growth ranged from log 1 to log 5 in 24 h. This is probably strain-related because such rapid growth of E. coli in T P N solutions has not been reported clinically. The strain of Ps. aeruginosa (NCIB 8626:ATCC 9027) used for this study was the same as that used by Rusmin et al. (7) in their study to determine the effectiveness of 0.22-pm and 0.45-pm rated mem- brane filters in retaining organisms during i.v. fluid adminstration over 72 h. The growth characteristics of this organism mirrored those reported in other studies examining in vitro the growth characteristics of bacteria in i.v. solutions (8).

The data reported here complements that obtained by other workers. Baumgartner et al. (9) demonstrated that the PALL ‘Intravenous Set Saver’ Filter was capable of retaining endotoxin from a challenge of E. coli (ATCC 12014) for 96 h when filtering 5 D-glucose. Spielberg (10) similarly demonstrated the ELD96’s endotoxin-retentive capability when filtering both 596 D-glucose and 0.9% saline. Ball and Saunders (11) demonstrated the ability of the ELD96 to retain endotoxins during changes in i.v. solution composition, including the infusion of negatively charged solutions and drugs, and during bolus injection of antibiotic.

As the results of the above studies demonstrated that the ELD96 filter was compatible with the T P N solutions used on the SCBU, University Hospital of Wales, all neonates who received parenteral nutrition supplied via percutaneous long lines or via peripheral lines now have a PALL ELD96 filter in-line. In addition to the improvement in patient care provided by such filters, the filters have proved to give considerable cost savings as

58 C. Richards and P . Thomas

only one administration set need be used during the 4-day life of the filter. Current practice is to use Ivac 565 infusion pumps with Ivac 953703 administration sets. It should be noted that when parenteral nutrition is being administered to a patient, the required regimen may change every 24 h. When a bag containing the new regimen replaces the old regimen, there will still be a volume of the old regimen in the adminis- tration set. This could represent 20”/b of the daily fluid requirments for a small preterm neonate. The old solution must therefore be flushed through so that the neonate receives all of that day’s regimen.

ACKNOWLEDGMENTS

The authors wish to express their appreciation to Dr Karen Wilkins and Mrs Elise Maynard, PALL Europe Corporate Services Limited, for their technical assistance during this study.

REFERENCES

1. Maki, D.G., Goldman, D.A. & Rhane, F.S. (1973) Infection control in intravenous therapy. Annals of Internal Medicine, 19,867-887.

2. Daschner, F.D., Frey, Wolff, G., Bauman, P.C. & Suter, P. (1982) Nosocomial infections in intensive care wards: a multi-centre prospective study. Intensive Care Medicine, 8,559.

3. Scheckelhoff, D.J., Mirtallo, J.M., Ayers, L.W. & Visconti, J.A. (1986) Growth of bacteria and fungi in total nutrient admixtures. American Journal of Hospital Pharmacy, 43,73-77.

4. Dunk-Richards, G. (1980) Ability of selected organisms to grow in total parenteral nutrition fluids and its relevance. AustralianJournal of Hospital Pharmacy, 10,7G73.

5. Scott, E.M., Gorman, S.P., Wyatt, T.O. & Magill, E. (1985) Growth of microorganisms in total parenteral nutrition mixtures and related clinical observations. JournaZ of Clinical and Hospital Pharmacy, 10,74-78.

6. Holmes, C.J., Kundsin, R.B., Ausman, R.K. & Walter, C.W. (1980) Potential hazards associ- ated with microbial contamination of in-line filters during intravenous therapy. Journal of Clinical Microbiology 12,725-731.

7. Rusmin, S., Althauser, M.B. & Deluca, P.P. (1975) Consequences of microbial contamination during extended intravenous therapy using in-line filters. American Journal of Hospital Pharmacy, 32, 373-377.

8. Wilkinson, W.R., Flores, L.L. & Pagones, J.N. (1973) Growth of microorganisms in parenteral nutrition fluids. Drug intelligence and Clinical Pharmacy, 7,226-231.

9. Baumgartner, T.G., Schmidt, G.L., Thakker, K.M., Sitren, H.S., Cerda, J.J., Mahaffey, S.M. & Copeland, E.M. (1986) Bacterial endotoxin retention by in-line intravenous filters. Americanrournal of Hospital Pharmacy, 43,68 1-684.

10. Spielberg, R. (1985) Evaluation of the endotoxin/bacterial retention capabilities of IV filters during simulated extended infusions. Pall Technical Report IV-I 001.

11. Ball, P.R. & Saunders, D. (1987) Endotoxin retention capabilities of the PALL Intravenous Set Saver. P A L L Technical Report BM2117.