If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Pharmacy, Mie University Hospital, Mie, JapanDepartment of Clinical Pharmaceutics, Division of Clinical Medical Science, Mie University Graduate School of Medicine, Mie, JapanDepartment of Clinical Infectious Diseases, Aichi Medical University, Aichi, Japan
Department of Pharmacy, Mie University Hospital, Mie, JapanDepartment of Clinical Pharmaceutics, Division of Clinical Medical Science, Mie University Graduate School of Medicine, Mie, Japan
Department of Clinical Infectious Diseases, Aichi Medical University, Aichi, JapanDepartment of Molecular Epidemiology and Biomedical Sciences, Aichi Medical University Hospital, Aichi, Japan
Department of Pharmacy, Mie University Hospital, Mie, JapanDepartment of Clinical Pharmaceutics, Division of Clinical Medical Science, Mie University Graduate School of Medicine, Mie, Japan
In-line filters in peripheral and central venous catheters are used to remove bacterial cells mechanically. A recent study indicated an extension of the use of infusion sets to 7 days. There is no evidence regarding replacement intervals for in-line filters.
Aim
To test in-line filters that were used continuously for 7 days in order to investigate their ability to remove bacteria and assess the flow rate.
Methods
Three different in-line filters were attached to an ELNEOPA-NF No. 2 premixed infusion bag of intravenous hyperalimentation, into which Staphylococcus epidermidis ATCC12228 or Escherichia coli ATCC25922 was inoculated. These experiments were compared with a control infusion. The infusion was dropped at a flow rate of 40 mL/h and replaced at 24-h intervals for 7 days. Samples were collected 24 h after drop initiation.
Findings
S. epidermidis was not detected in droplets between Days 1 and 6, but In-line filters 1 and 2 showed droplets containing 6–10 colony-forming units/mL on Day 7. E. coli was not detected in any of the filters after 7 days of continuous use. Flow rates <40 mL/h were observed on Day 7 for In-line filter 3 in studies of S. epidermidis, and on Days 4 and 3 for In-line filters 2 and 3, respectively, in studies of E. coli.
Conclusion
This study revealed differences in bacterial removal and flow rates under high inoculation between the three in-line filters tested. It is suggested that in-line filters can be used continuously for a maximum of 6 days, and reductions in flow rate after 48 h of continuous use should be noted carefully.
Catheter-related bloodstream infections (CRBSIs) account for approximately 15% of healthcare-associated infections and 25–43% of cases of nosocomial bacteraemia [
]. This is due to hand hygiene, use of chlorhexidine alcohol solution for skin antisepsis, full barrier precautions, and daily observation of the need for catheterization [
]. However, once patients develop CLABSIs, hospitalization is prolonged, and medical costs and morbidity/mortality rates increase. Therefore, minimizing the likelihood of infection in patients with a venous catheter is a critical challenge in regard to healthcare burden and patient outcomes.
In-line filters in peripheral and central venous catheters remove large particles, precipitates, bacteria, fungi, large lipid globules and air mechanically. Recent studies have reported a reduction in phlebitis or systemic inflammatory response syndrome with the use of in-line filters [
]. This has led to increasing support for the adoption of in-line filters as a routine standard of care for infusion therapy. However, in the revised 2011 guidelines from the US Centers for Disease Control and Prevention, all comments regarding in-line filters were removed [
]. No data support the efficacy of in-line filters in the prevention of CRBSIs due to limited incidence. Therefore, definitive guidelines around the use of in-line filters remain controversial.
Catheters are accessed repeatedly each day and are often used for extended periods [
]. Nevertheless, no optimal time interval for infusion set replacement has been identified to date. In 2021, Rickard et al. indicated that a 7-day (vs 4-day) replacement of infusion sets was safe [
]. In 2022, the guidelines of the Society for Healthcare Epidemiology of America for central-line-associated bloodstream infection prevention in hospitals recommended routine replacement of infusion sets not used for blood, blood products or lipid formulations at intervals of up to 7 days [
] However, the optimal timing for the replacement of in-line filters is not mentioned in the updated guidelines. This is because, in the USA, all infusions are filtered during preparation. However, infusions are not filtered in other countries, such as Japan and the UK. Moreover, with the development of various premixed infusion bags, infusions do not always need to be filtered during preparation. Therefore, the frequency of using in-line filters may not be reduced. The optimal timing for replacing in-line filters must be provided to hospitals.
As such, this study investigated replacement intervals for in-line filters that were used continuously for 7 days. The aim was to provide basic evidence to inform clinical practice guidelines for maintaining patient safety and for using in-line filters wisely.
Methods
Chemical and in-line filters
ELNEOPA-NF No. 2 (Otsuka Pharmaceutical Factory, Inc., Tokushima, Japan), a premixed infusion bag of intravenous hyperalimentation containing commercial glucose, electrolytes, amino acids, vitamins and micro-elements, was used in this study [
]. The composition of ELNEOPA-NF No. 2 is shown in Table I. Three different in-line filters were used throughout the study (In-line filter 1, 22813 SISA-2-FR-H, TOP, Japan; In-line filter 2, TF-SW231H, TERUMO, Japan; In-line filter 3, ELD96LLC, PALL, Japan). The characteristics of each in-line filter are shown in Table II.
Table IComposition of ELNEOPA-NF No. 2 (premixed infusion bag of intravenous hyperalimentation)
Composition per 1000 mL
Glucose
175
g
Fe
10
μmol
Thiamine
3.84
mg
Mn
0.5
μmol
Riboflavin
2.3
mg
Zn
30
μmol
Pyridoxine
3.675
mg
Cu
2.5
μmol
Cyanocobalamin
2.5
μg
I
0.5
μmol
Nicotinic-acid amide
20
mg
L-leucine
8.4
g
Panthenol
7
mg
L-isoleucine
4.8
g
Folic acid
0.3
mg
L-valine
4.8
g
Biotin
30
μg
L-lysine
6.29
g
Ascorbic acid
100
mg
L-threonine
3.42
g
Vitamin A
1650
unit
L-tryptophan
1.20
g
Cholecalciferol
2.5
μg
L-methionine
2.34
g
Tocopherol
5
mg
L-cysteine
0.60
g
Phytonadione
0.075
mg
L-phenylalanine
4.20
g
Na+
50
mEq
L-tyrosine
0.30
g
K+
27
mEq
L-arginine
6.30
g
Mg2+
5
mEq
L-histidine
3.00
g
Ca2+
5
mEq
L-alanine
4.50
g
Cl-
50
mEq
L-proline
3.00
g
SO42-
5
mEq
L-serine
1.80
g
Acetate-
48
mEq
Glycine
3.54
g
L-lactate-
14
mEq
L-aspartic acid
0.60
g
Citrate3-
12
mEq
L-glutamic acid
0.60
g
P
6
mmol
Characteristics
pH
5.4
OPR
6
OPR, osmotic pressure ratio to physiological saline.
No ethical approval was required for the current study as this was not human subject research.
Micro-organisms
Standard strains were used for each micro-organism, namely Staphylococcus epidermidis ATCC12228 and Escherichia coli ATCC25922. The bacteria were stored at -70 °C in skimmed milk, and were subcultured twice on to trypticase soy agar plates with 5% sheep blood (Becton, Dickinson & Co., Sparks, MD, USA) and grown for 24 h at 37 °C under 5% CO2.
Simulated infusion system with in-line filter
A simulated infusion system was adopted in this study, as reported previously [
]. Bacterial suspensions were prepared using the second subculture of each pathogen. The suspension was injected into a 1-L bag of ELNEOPA-NF No. 2 to obtain a total concentration of 106 colony-forming units (CFU)/mL. As shown in Figure 1, the primary administration set was connected to ELNEOPA-NF No. 2, and the in-line filters were attached to the distal end of the administration set. Although the beaker was not sealed, all operations were performed aseptically. The contamination was not confirmed in whole experiments.
Figure 1Simulated infusion system with an in-line filter.
ELNEOPA-NF No. 2 was added dropwise at a constant rate of 40 mL/h, and replaced at 24-h intervals for 7 days. To evaluate bacterial counts, a simulated infusion system without an in-line filter was used as a control. A simulated infusion system which did not inoculate bacteria was used as a control when evaluating the flow rate. of the number of drops. All experiments were performed three times at 25 °C.
Measurement of bacterial colony
Droplet samples were collected 5 min after drop initiation on Day 1 (baseline), and 24 h after drop initiation from Days 2–7. All solution samples were serially diluted with normal saline, and fixed amounts of these solutions were placed on the plates. The number of microbial pathogen CFUs in each plate was counted after 24 h of incubation at 35 °C. The results are presented as CFU/mL values in semilogarithmic graphs.
Measurement of flow rate by spontaneous drop
The number of drops in 1 min under a fully open clamp was recorded 5 min after drop initiation on Day 1, as well as 24 h after drop initiation from Day 2 to Day 7. The flow rate was calculated using the following equation: flow rate (mL/h) = drops/20 × 60 min (20 drops = 1 mL). A flow rate <40 mL/h showed that 1000 mL/day could not be administered.
Results
Time-dependent change in bacterial counts
Bacterial counts in the simulated infusion system of S. epidermidis and E. coli are shown in Figure 2 (S. epidermidis, Figure 2A; E. coli, Figure 2B). During this study, the bacterial counts of S. epidermidis and E. coli ranged from 6.41 to 6.76 log10 CFU/mL in the inoculated samples and from 6.34 to 6.61 log10 CFU/mL in the control group, respectively. Only In-line filter 3 was able to produce S. epidermidis-free droplets continuously for 7 days. Although S. epidermidis was not detected in the in-line filtered droplets between Days 1 and 6 in in-line filters 1 and 2, droplets containing 6–10 CFU/mL of S. epidermidis were observed in these in-line filters on Day 7. In studies using E. coli, all in-line filters produced sterile droplets continuously for 7 days.
Figure 2Time-dependent change in bacterial count in fluid filtered with each in-line filter. (A) Staphylococcus aureus, (B) Escherichia coli. Data are presented as mean ± standard deviation. N=3 in each group. CFU, colony-forming units.
The flow rates in the simulated infusion system of S. epidermidis and E. coli are shown in Table III. The control group was admitted with no change in flow rate for 7 days, with flow rate >40 mL/h. The study of S. epidermidis showed a decrease in flow rate in all in-line filters. In-line filters 1 and 2 maintained flow rates of >40 mL/h, while the flow rate of In-line filter 3 dropped <40 mL/h on Day 7. Similarly, the study of E. coli showed a reduction in flow rate in all in-line filters, and only In-line filter 1 maintained a flow rate of >40 mL/h. The flow rates of In-line filters 2 and 3 dropped below 40 mL/h after Days 4 and 3, respectively.
Table IIITime-dependent change in flow rates for each in-line filter
Day 0
Day 1
Day2
Day 3
Day 4
Day 5
Day 6
Day 7
Control (In-line filter 3; mL/h)
553±11
528±8
531±12
580±9
506±18
539±7
583±8
496±5
Staphylococcus epidermidis
In-line filter 1 (mL/h)
832±20
842±32
836±42
604±22
402±18
291±11
172±3
100±5
In-line filter 2 (mL/h)
797±15
748±32
578±20
313±15
145±2
74±9
60±5
45±2
In-line filter 3 (mL/h)
541±22
393±3
405±3
304±10
134±2
79±2
42±2
17±2
Escherichia coli
In-line filter 1 (mL/h)
653±18
606±13
552±6
462±8
234±3
147±1
108±1
79±2
In-line filter 2 (mL/h)
560±17
512±20
116±2
52±2
29±2
21±1
15±1
15±1
In-line filter 3 (mL/h)
563±20
303±3
47±2
19±2
17±2
15±1
12±1
9±1
All data are shown as average ± standard deviation (N=3).
To date, neither US nor European pharmacopoeias, or any other guidance, has identified scientific data to support evidence-based recommendations on the duration of use of in-line filters [
]. Thus, the in-line filters used in clinical practice during intravenous and central venous infusion were selected based on the general information mentioned in the respective package insert or Summary of Product Characteristics. This study found that continuous use of in-line filters for up to 6 days avoided the potential hazards of bacterial penetration, and that sufficient flow rates could not be maintained depending on the type of in-line filters and bacteria.
Bacteria are contaminants that can survive and grow during intravenous hyperalimentation. In particular, S. epidermidis and E. coli are the main causative pathogens of CRBSIs [
]. The diameters of S. epidermidis and E. coli are 0.5–0.6 μm and 1.0–3.0 μm, respectively, both of which are larger than the filter pore size for infusion [
], which is consistent with the present results. On the other hand, the present study confirmed the existence of S. epidermidis in the effluents from in-line filters 1 and 2 on Day 7. In other words, S. epidermidis passes through in-line filters with asymmetric membranes. Similar findings have been reported in previous studies of Candida albicans. It was thought previously that C. albicans did not penetrate in-line filters because its cell size was larger than the pore size of the in-line filters. However, several studies have shown that C. albicans passes through in-line filters with asymmetric membranes within 2 days, while it was shown not to penetrate in-line filters with symmetric membranes for 7 days [
]. Asymmetric membranes have a certain thickness, and capture bacteria by collision and absorption to the filter media while bacteria pass through the layer. The filter pore size was not constant, and ranged between 0.2 and 1.0 μm in this study. Therefore, S. epidermidis may have passed through the layer with 0.6–1.0-μm pores. To determine the cause, further investigation observing the inside of the in-line filter is necessary.
The ability to remove bacteria may be influenced by the in-line filter medium and filter structures other than the shape of the membrane. A previous study reported the passage of endotoxins released by E. coli that were trapped on various in-line filter media [
], in-line filter media, such as cellulose ester, polyacrylate, polyethylene and polypropylene, were unable to retain endotoxins; only the posidyne nylon 66 medium was capable of retaining endotoxins for up to 7 days [
Although all in-line filters captured E. coli continuously for 7 days in this study, In-line filters 2 and 3 may not retain endotoxins based on previous data [
]. However, Rusmin et al. reported that effluents from in-line filters contaminated with Gram-negative bacteria may not be endotoxin-positive as a result of spontaneous release, but endotoxins from non-viable Gram-negative bacteria trapped on the in-line filter after the administration of antibiotics that are subsequently flushed through the in-line filters may be released [
]. Therefore, in-line filters with filter media, except for posidyne nylon 66, should be replaced before antibiotics are administered through the same route, regardless of the duration of use. However, there are no data on the efficacy of different filter structures. No differences in filter structures were observed in this study.
An adequate flow rate is necessary to meet daily requirements. However, when an in-line filter retains many bacteria, an asymmetric membrane can maintain a sufficient flow rate, whereas a symmetric membrane can decrease the flow rate. This finding is consistent with the results of the present study. In particular, the present study showed that In-line filter 3 (with a symmetric membrane) was unable to establish a flow rate of 1 L/day after Day 3. The package insert of In-line filter 3 recommends replacement within 96 h [
]. Therefore, this reduction in flow rate must be noted in the case of continuous use of In-line filter 3 for 96 h.
One of the limitations of this study was the level of contamination, which was chosen to provide a stringent challenge to the media. This allowed clear differences in bacterial retention to be evaluated. Inadvertent contamination can result in a bacterial count of 106 CFU/mL. A previous study showed that initial contamination by as little as 100 CFU/mL produced a concentration of 106 CFU/mL in a 24-h period [
]. In fact, the pre-test of the present study showed that inoculation with 100 CFU/mL did not lead to the detection of S. epidermidis or E. coli (data not shown). However, the experimental model would be thought to mimic the in-line filter used in clinical practice, unlike previously reported models.
In conclusion, this study revealed that there were differences in bacterial removal ability and flow rate under high inoculation among the three in-line filters currently available on the market. The authors suggest that in-line filters with symmetric membranes should be used continuously for 6 days. Reductions in the flow rate after 48 h of continuous use should be noted carefully. These findings may help in the selection of in-line filters for continuous use during intravenous hyperalimentation.
This study was designed as an in-vitro study, and in-vivo studies are needed to compare the bacterial removal efficacies of these filters. These in-vivo studies will assist in determining whether sufficient quantities of pathogens can pass through the in-line filters to cause bloodstream infection in these settings.
Acknowledgments
The authors wish to thank all the co-authors who assisted in this project, and PALL International for the provision of filters. H.K. gratefully acknowledges financial support from the OKASAN-KATO Foundation (OKF-22-1-49).
Conflict of interest statement
None declared.
Funding sources
This work was supported by OKASAN-KATO Foundation (Grant No. OKF-22-1-49). The funder had no role in the design, conduct or reporting of this work. None of the authors have any financial relationships with any commercial entity with an interest in this subject area.
References
Zarb P.
Coignard B.
Griskeviciene J.
Muller A.
Vankerckhoven V.
Weist K.
et al.
The European Centre for Disease Prevention and Control (ECDC) pilot point prevalence survey of healthcare-associated infections and antimicrobial use.
00033-6&id=gr1.jpg){kind=link}
00033-6&id=gr2.jpg){kind=link}