Secondary Logo

Journal Logo

Patient Safety: Research Reports

Should In-line Filters Be Used in Peripheral Intravenous Catheters to Prevent Infusion-Related Phlebitis? A Systematic Review of Randomized Controlled Trials

Niël-Weise, Barbara S. MD (B)*; Stijnen, Theo PhD; van den Broek, Peterhans J. MD, PhD

Author Information
doi: 10.1213/ANE.0b013e3181da8342

Phlebitis is the most frequent complication associated with peripheral IV infusions, occurring in up to 96% of all patients.1 Phlebitis is the clinical finding of pain, tenderness, swelling, induration, warmth, and palpable cord-like veins caused by inflammation. Because it is painful, phlebitis requires removal of the catheter and insertion of a new catheter at a different site; this often requires local treatment and an analgesic. Frequent catheter changes cause patient distress and are time consuming for both nurses and physicians. Repeated episodes can lead to venous access difficulties and more invasive procedures such as placement of central venous access devices. Fortunately, peripheral catheters have rarely been associated with bloodstream infection or deep vein thrombosis.

Major etiologic factors of infusion-related phlebitis include the composition of the infusion set and catheter material, anatomic location of the catheter, duration of catheterization, pH and osmolality of the infused fluid, and contaminants in the infusion solutions. Infusion fluids may be contaminated with particulates, bacteria, endotoxins, precipitates, large lipids, and air.2,3 These contaminants might cause inflammation. Different types of particulates have been described, i.e., glass fragments resulting from the opening of glass ampules,4 particles from rubber stoppers and IV equipment,5 or particles from plastic syringes.6 Some chemicals remain undissolved as particles in the infusion fluids, for example, antibiotics. In-line IV filters could be an effective approach to remove these contaminants from IV solutions, thus reducing the rate of phlebitis.7,8

In-line filters may also have unwanted effects. Friedland3 reported that certain solutions cause clogging and reduced flow rates. He also reported that some drugs bind to the filter itself or they are retained by the filter resulting in a potential decrease in the amount of available drug in the circulation. Furthermore, in-line filters add cost.

The British Pharmaceutical Nutrition group recommends the use of in-line filters.2 The Centers for Disease Control and Prevention recommend that filters not be used routinely for infection control purposes.9 However, a narrative review published in 2003 concluded that there was strong evidence for the routine use of in-line filters.7 Because these recommendations are not based on the findings of a systematic review, we decided to summarize the evidence of the efficacy of in-line filters in reducing infusion-related phlebitis in a systematic review of the literature. We assessed the impact of the baseline risk of phlebitis on the effect of in-line filters. It was not within the scope of this review to assess the effect of in-line filters on either IV solutions flow rate or drug binding to in-line filters.

METHODS

Search

Publications were retrieved by a search of MEDLINE and the Cochrane Library up to August 10, 2009. Search terms included randomized controlled trial (RCT), phlebitis, thrombophlebitis, and in-line filter. The complete search strategy can be found in the online Appendix, which also includes a table (see Supplemental Digital Content 1, https://links.lww.com/AA/A115). In addition, the lists of references of all identified trials were checked for more trials.

Selection

We included studies that were planned as a randomized trial or quasi-randomized trial and were also published as an article. The studies had to include the definition of phlebitis that was used and had to present sufficient data to be able to calculate the risks of phlebitis in both the treatment and the control group. No language restrictions were applied. Two reviewers assessed all titles and abstracts independently to confirm the eligibility of the selected trials. Disagreements were resolved by consensus.

Assessment of Trial Quality

Two reviewers assessed trial quality independently by evaluating each study to determine concealment of treatment allocation, double blinding, completeness of follow-up, use of intention-to-treat analysis, selective reporting of events, and premature discontinuation of the trial due to benefit. Central randomization, sealed envelopes, or a similar method was assumed to yield adequate randomization. The description of dropouts was considered adequate if the number of patients lost and reasons why patients were lost were reported according to allocation to treatment. Disagreements were resolved by consensus.

Data Extraction, Analysis, and Quality of Evidence

Data on study population, interventions, and outcomes were independently extracted and crosschecked by both reviewers. Only trial data that were related to the question posed in the review were considered. For the dichotomous outcome phlebitis, we calculated the overall relative risk (RR) with a 95% confidence interval (CI) by means of Review Manager (version 4.2.7), using the standard random effects method of DerSimonian and Laird.10 When appropriate, meta-analyses were performed using a random-effects model to calculate pooled estimates and their 95% CIs. Subgroups were not defined a priori. The baseline risk was defined as the risk of phlebitis in the control arm of the individual trials, i.e., the risk of phlebitis without treatment (formula: total number of patients with phlebitis in the control group/total number of patients in the control group). Publication bias was examined by using a funnel plot. The quality of the available evidence for phlebitis was assessed by the GRADE method.11

RESULTS

Selection

Initially, 35 potentially relevant studies were identified by our search. The flow diagram showing the steps that we followed to identify the RCTs fulfilling the inclusion criteria of our systematic review can be found in the online Appendix (see Supplemental Digital Content 1, https://links.lww.com/AA/A115). After reading the titles and abstracts, 21 studies seemed to fulfill the selection criteria. Of the 21 studies, 10 articles were excluded after reading the entire article.3,1220 The reasons for exclusion can be found in the online Appendix (see Supplemental Digital Content 1, https://links.lww.com/AA/A115). Ultimately, 11 trials were included in the review.1,2130

Quality Assessment of Trials

In 7 of the 11 selected trials, an adequate method of concealment was used; all were double-blind trials.1,2123,25,26,28 Description of dropouts was adequate in 2 trials.22,26 None of the trials stated clearly that the analysis took place according to the intention-to-treat principle. None of the trials was stopped earlier than planned because of treatment benefit or reported events in a selective manner.

Data Extraction, Analysis, and Body of Evidence

Data on study populations, interventions, and outcome definitions are shown in Table 1. Study populations consisted of surgical patients,1,23,26,27,29 surgical and medical patients,21,28 general hospital patients,22 pediatric oncology patients,25 and cardiac patients.30 In 1 trial, the patients were not described.24

T1-18
Table 1:
Study Population, Interventions, and Outcome Definitions

In 2 trials, patients received long peripheral catheters (7.5–20 cm), i.e., Intracath or E-Z Cath type.24,30 In all other trials, although not always explicitly described, we assumed that the patients received short peripheral catheters (<7.5 cm).

Different types of IV solutions were administered: hyperalimentation,28 isotonic glucose,30 cephalothin,23 clear fluids,24 buffered solutions,23 and “various kinds of fluids and additives.”1,2126,29

Three types of in-line filters were studied: a 0.22-μm filter,1,2123,26,28 a 0.45-μm filter,23,24,27,29 and a 0.5-μm filter.25 Participants in the control groups received a “dummy” filter1,2123,25,26,28,30 or no filter.24,27,29

Phlebitis was diagnosed by clinical findings ranging from just a single finding, such as pain,21,29 to at least 3 findings, such as erythema, induration, and palpable IV cord >2.5 cm in length.28 The baseline risks for phlebitis across trials varied from 23%29 to 96%.1 The majority of trials did not report the mean duration of IV catheterization.

The overall RR for all trials was statistically borderline significant in favor of the in-line filter group (RR, 0.66; 95% CI, 0.43–1.00) (Fig. 1). However, the meta-analysis showed marked heterogeneity (P < 0.00001) (Fig. 1). The I2 value was 90.4%, indicating that 90% of the observed variability in treatment effects was attributable to systematic variation among trials beyond chance. Seven studies showed a neutral RR of approximately 1, whereas 4 studies indicated a strong treatment effect with RR <0.5. Subgroup meta-analysis was performed according to catheter types, filter types, and study quality. However, the marked heterogeneity remained. Meta-regression analysis of RR on the baseline risk did not show evidence for a relationship between the treatment effect and the underlying risk.31

F1-18
Figure 1:
Summary estimates of association between peripheral catheters with in-line filters and peripheral catheters without in-line filters expressed as relative risk (RR) and 95% confidence interval (CI) using a random-effects model.

Although there was no funnel plot asymmetry, publication bias could not be excluded because of the small number of included studies.32

According to the GRADE method,11 the quality of the available evidence for phlebitis was determined by considering the quality of the trials, consistency, indirectness, imprecision, and publication bias. The grade of the evidence decreased by 1 level because the RCTs exhibited serious limitations in study methods and execution, as described above; it decreased by 2 levels because of the marked unexplained statistical heterogeneity.

DISCUSSION

The pooled RR for the 11 trials identified (1633 peripheral catheters) showed that the use of in-line filters (826 peripheral catheters) reduced the occurrence of phlebitis in hospitalized patients. However, as indicated by the GRADE method, the evidence that in-line filters reduce the rate of phlebitis by 34% is very uncertain because of the serious methodological shortcomings and the marked heterogeneity, which remained unexplained despite attempts to identify the causes of the heterogeneity.

Although overall the effect supports the use of in-line filters and the magnitude of the effect is clinically relevant, the considerable variation between trials (as seen in Fig. 1, and the large I2 value of 90.4%) remains unexplained. Seven studies showed treatment effects close to 1, whereas 4 studies revealed a large in-line filter effect. Unfortunately, we were not able to identify any factor common to these 4 studies but not common to the other 7 studies. Regression analysis did not show that the estimated benefit of in-line filters depended on the baseline risk. Subgroup meta-analysis according to catheter types, filter types, and study quality could not solve the problem of statistical heterogeneity. Subgroup meta-analysis according to the type of IV infusion could not be performed, because each study used unique infusion fluids. From the literature, it is known that the type and composition of the IV solutions and the presence of additives in the infusions are important causes of infusion-related phlebitis.3 It is conceivable that the heterogeneity of the IV solutions administered during the individual trials might have been an important source of the statistical heterogeneity.

It is possible that methodological limitations in study quality may have introduced a bias. In 4 trials, selection, performance, and detection bias could not be excluded because they had unclear or inadequate allocation concealment, and they were not reported as double-blind trials. There was an inadequate description of dropout rates in 9 trials, and all trials had an unclear or inadequate handling of attrition.

Our finding of a very uncertain benefit of in-line filters does not agree with the narrative review that states that there is strong evidence for routine use of in-line filters7 and does not support the experts' recommendations to use in-line filters routinely to prevent infusion-related phlebitis.2,7,8

In this review, the effects of in-line filters did not apply to patients receiving blood products. The pore sizes of the filters investigated in the individual trials were too small for administration of these fluids and, therefore, nearly all trials excluded patients with indications for blood or blood products administration.

CONCLUSION

This systematic review shows that in-line filters in peripheral IV catheters cannot be recommended for routine use, because evidence of their benefit is uncertain.

There is a need for large, high-quality RCTs to assess the benefit of in-line filters, to determine which in-line filter is most effective in reducing phlebitis, and to assess their cost-effectiveness. Future trials should use appropriate methods, i.e., concealment of randomization, placebo control, intention-to-treat analysis, and description of the study end point. Future trials should be large enough to allow effects to be explored in subgroups according to catheter types, filter types, and especially types of infusion fluids. They should also include direct measures of clinical outcomes and express risks as incidence per unit time such as 100 catheter days.

REFERENCES

1. Falchuk KH, Peterson L, McNeil BJ. Microparticulate-induced phlebitis. Its prevention by in-line filtration. N Engl J Med 1985;312:78–82
2. Bethune K, Allwood M, Grainger C, Wormleighton C. Use of filters during the preparation and administration of parenteral nutrition: position paper and guidelines prepared by a British pharmaceutical nutrition group working party. Nutrition 2001;17:403–8
3. Friedland G. Infusion-related phlebitis—is the in-line filter the solution? N Engl J Med 1985;312:113–5
4. Shaw NJ, Lyall EG. Hazards of glass ampoules. Br Med J (Clin Res Ed) 1985;291:1390
5. Kirkpatrick C. Particulate matter in intravenous fluids: the importance for medicine. Krankenhauspharmazie 1988;9:487–90
6. Mehrkens HH, Klaus E, Schmitz JE. Possibilities of material contamination due to additional injections [in German]. Klin Anasthesiol Intensivther 1977:106–13
7. Ball PA. Intravenous in-line filters: filtering the evidence. Curr Opin Clin Nutr Metab Care 2003;6:319–25
8. Kunac D, Ball P, Broadbent R. In-line intravenous filtration in neonates-help not hindrance. Aust J Hosp Pharm 1999;29:321–7
9. O'Grady NP, Alexander M, Dellinger EP, Gerberding JL, Heard SO, Maki DG, Masur H, McCormick RD, Mermel LA, Pearson ML, Raad II, Randolph A, Weinstein RA. Guidelines for the prevention of intravascular catheter-related infections. Infect Control Hosp Epidemiol 2002;23:759–69
10. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986;7:177–88
11. Atkins D, Best D, Briss PA, Eccles M, Falck-Ytter Y, Flottorp S, Guyatt GH, Harbour RT, Haugh MC, Henry D, Hill S, Jaeschke R, Leng G, Liberati A, Magrini N, Mason J, Middleton P, Mrukowicz J, O'Connell D, Oxman AD, Phillips B, Schunemann HJ, Edejer TT, Varonen H, Vist GE, Williams JW Jr, Zaza S; GRADE Working Group. Grading quality of evidence and strength of recommendations. BMJ 2004;328:1490–4
12. Bach A, Böttiger BW. Microfilters within the scope of infusion therapy—possibilities and problems in retention of microbial and particle contaminants [in German]. Zentralbl Chir 1994;119:268–75
13. DeLuca PP, Rapp RP, Bivins B, McKean HE, Griffen WO. Filtration and infusion phlebitis: a double-blind prospective clinical study. Am J Hosp Pharm 1975;32:1001–7
14. Foster J, Richards R, Showell M. Intravenous in-line filters for preventing morbidity and mortality in neonates. Cochrane Database Syst Rev 2006:CD005248
15. Frazer IH, Eke N, Laing MS. Is infusion phlebitis preventable? Br Med J 1977;2:232
16. Maddox RR, Rush DR, Rapp RP, Foster TS, Mazella V, McKean HE. Double-blind study to investigate methods to prevent cephalothin-induced phlebitis. Am J Hosp Pharm 1977;34:29–34
17. Roberts GW, Holmes MD, Staugas RE, Day RA, Finlay CF, Pitcher A. Peripheral intravenous line survival and phlebitis prevention in patients receiving intravenous antibiotics: heparin/hydrocortisone versus in-line filters. Ann Pharmacother 1994;28:11–6
18. Rusho WJ, Bair JN. Effect of filtration on complications of postoperative intravenous therapy. Am J Hosp Pharm 1979; 36:1355–6
19. Thomas ET, Evers W, Racz GB. Postinfusion phlebitis. Anesth Analg 1970;49:150–9
20. van Lingen RA, Baerts W, Marquering AC, Ruijs GJ. The use of in-line intravenous filters in sick newborn infants. Acta Paediatr 2004;93:658–62
21. Adams SD, Killien M, Larson E. In-line filtration and infusion phlebitis. Heart Lung 1986;15:134–40
22. Allcutt DA, Lort D, McCollum CN. Final inline filtration for intravenous infusions: a prospective hospital study. Br J Surg 1983;70:111–3
23. Bivins BA, Rapp RP, DeLuca PP, McKean H, Griffen WO Jr. Final inline filtration: a means of decreasing the incidence of infusion phlebitis. Surgery 1979;85:388–94
24. Collin J, Tweedle DE, Venables CW, Constable FL, Johnston ID. Effect of a Millipore filter on complications of intravenous infusions: a prospective clinical trial. Br Med J 1973;4:456–8
25. Evans WE, Barker LF, Simone JV. Double-blind evaluation of 5-mum final filtration to reduce postinfusion phlebitis. Am J Hosp Pharm 1976;33:1160–3
26. Maddox RR, John JF Jr, Brown LL, Smith CE. Effect of inline filtration on postinfusion phlebitis. Clin Pharm 1983;2:58–61
27. Ryan PB, Rapp RP, DeLuca PP, Griffen WO Jr, Clark JD, Cloys D. In-line final filtration—a method of minimizing contamination in intravenous therapy. Bull Parenter Drug Assoc 1973;27:1–14
28. Rypins EB, Johnson BH, Reder B, Sarfeh IJ, Shimoda K. Three-phase study of phlebitis in patients receiving peripheral intravenous hyperalimentation. Am J Surg 1990;159:222–5
29. Swift RG, Searcy DM, Pickard AS. The effect of in-line final filtration on occurrence of phlebitis. Drug Intell Clin Pharm 1975;9:76–9
30. Thayssen P, Kortegaard N, Winding O. Postinfusion phlebitis and in-line terminal membrane filtration. Dan Med Bull 1977;24:160–2
31. van Houwelingen HC, Arends LR, Stijnen T. Advanced methods in meta-analysis: multivariate approach and meta-regression. Stat Med 2002;21:589–624
32. Lau J, Ioannidis JP, Terrin N, Schmid CH, Olkin I. The case of the misleading funnel plot. BMJ 2006;333:597–600

Supplemental Digital Content

© 2010 International Anesthesia Research Society