Journal of Controlled Release 188 (2014) 61–66

Contents lists available at ScienceDirect

Journal of Controlled Release

j ourna l homepage: www.e lsev ie r .com/ locate / jconre l

On-demand antimicrobial release from a temperature-sensitive polymer— Comparison with ad libitum release from central venous catheters

Jelmer Sjollema a,⁎, Rene J.B. Dijkstra a, Caroline Abeln b, Henny C. van der Mei a,Dirk van Asseldonk b, Henk J. Busscher a

a University of Groningen and University Medical Center Groningen Department of Biomedical Engineering, Antonius Deusinglaan 1, 9713 AV Groningen, The Netherlandsb Dolphys Medical B.V., De Lismortel 31, 5612 AR Eindhoven, The Netherlands

⁎ Corresponding author. Tel.: +31 50 3633149; fax: +E-mail address: j.sjollema@umcg.nl (J. Sjollema).

http://dx.doi.org/10.1016/j.jconrel.2014.06.0130168-3659/© 2014 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 27 March 2014Accepted 10 June 2014Available online 17 June 2014

Keywords:Central venous cathetersChlorhexidineSilverBiofilmsInfection

Antimicrobial releasing biomaterial coatings have found application for instance in the fixation of orthopedic jointprostheses and central venous catheters. Most frequently, the release kinetics is such that antimicrobially-effectiveconcentrations are only reached within the first days to weeks after implantation, leaving no local antimicrobialrelease available when a biomaterial-associated infection occurs later. Here we compare the ad libitum release ofchlorhexidine and silver-sulfadiazine from a central venous catheterwith their release from a new, on-demand re-lease coating consisting of a temperature-sensitive copolymer of styrene and n-butyl (meth)acrylate. The copoly-mer can be loaded with an antimicrobial, which is released when the temperature is raised above its glasstransition temperature. Ad libitum release of chlorhexidine and silver-sulfadiazine from a commercially-purchased catheter and associated antimicrobial efficacy against Staphylococcus aureus was limited to 16 days.Consecutive temperature-triggers of our on-demand coating yielded little or no antimicrobial efficacy of silver-acetate release, but antimicrobially-effective chlorhexidine concentrations were observed over a time period of60–80 days. This attests to the clear advantage of on-demand coatings above ad libitum releasing coatings, thatmay have released their antimicrobial content before it is actually needed. Importantly, glass transition tempera-ture of chlorhexidine loaded copolymers was lower (48 °C) than of silver loaded ones (61 °C), facilitating theirclinical use.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

Antimicrobial releasing biomaterial coatings have found applicationfor instance in the fixation of orthopedic joint prostheses [1] and centralvenous catheters (CVCs). CVCs are widely applied to offer vascular ac-cess for e.g. blood dialysis or administration of chemotherapeutics, pa-rental nutrition or antibiotics. The use of CVCs however, is associatedwith a high incidence of infectious complications ranging from 0.5 to26%, dependent on the type of catheterization. CVC-associated infec-tions are expensive to treat [2] and moreover, extremely hazardous topatients, as they can easily evolve from a subcutaneous infection ofthe skin near the entry point into a CVC-associated bloodstream infec-tion. It is estimated that CVC-associated bloodstream infections occuraround 80,000 times per year in the United States alone [3]. CVC-associated infections can cause great clinical dilemmas as removal ofthe catheter is not an easy decision. The number of veins that can beused as entry point is limited (jugular, subclavian vein and femoralveins), while generally a once infected insertion site cannot be used

31 50 3633159.

for a second time. Thus removal of a CVC due CVC-associated infectionfor the administration of chemotherapeutics for instance, may disruptthe chemotherapy, while keeping the catheter in place may yield therisk of death due to sepsis of an already weakened patient.

Catheter-related infections are thought to arise mainly through twodifferent routes of infection [4,5]. The first route is through the insertionsite: bacteria, primarily from the skin, can migrate along the externalcatheter surface to cause subcutaneous infection and subsequent blood-stream infection. The second route is through the contamination of thecatheter hub, leading to intra-luminal bacterial colonization and migra-tion to the bloodstream. In order to reduce the incidence of CVC-associated infection, topical application of chlorhexidine gluconate asa skin disinfectant for catheter site care has proven to be effective [6].As another strategy to prevent CVC-associated bloodstream infections,catheters have been equipped with antimicrobial coatings. Specificallycatheters releasing chlorhexidine or a combination of chlorhexidineand silver have been demonstrated to reduce the number of CVC-associated bloodstream infections [7]. Most of the antimicrobials how-ever, are released immediately upon contactwith human tissue or fluidsin a so-called “burst release” followed by a low-level tail-release that isonly antimicrobially effective for a limited period of time [8]. As a conse-quence, whenever a CVC-associated infection develops in the later

62 J. Sjollema et al. / Journal of Controlled Release 188 (2014) 61–66

stages of catheterization, much of the antimicrobial content of thesecoatings may have been released at a time, when needed most.

These drawbacks in the use of antimicrobial-coatings can be avoidedthrough the use of on-demand release coatings. On-demand releasecoatings can be externally triggered to release their content, either ina single episode or after multiple triggers, which is the preferred modeof on-demand release as it will allow triggered release at different in-stances over prolonged periods of time. Different on-demand mecha-nisms have been reported [9–12]. One such mechanism is to bring thecarrier polymer in a more open state by increasing its temperature toor above its glass transition temperature to allow drug release, whilewhen below its glass transition temperature the polymer structure isclosed, impeding drug release [13]. Glass transition temperatures ofma-terials suitable to be applied for triggered release of drugs are around40–60 °C and clinically temperature-triggers in that temperaturerange can be given by infrared irradiation or the use of an electricalheating filament [13].

In this paper, a temperature-triggering mechanism based on bring-ing a material temporarily above its glass transition temperature, hasbeen applied for the first time to offer an on-demand release of antimi-crobials for potential use in CVCs to prevent catheter-associated infec-tions. To this end, a copolymer of styrene (PS) and n-butyl (meth)acrylate (BA) was synthesized (abbreviated in the following as PS-BA)with temperature-dependent release characteristics of chlorhexidineand silver. In order to prevent release due to incidental rises in bodytemperature, polymer materials were designed with glass transitiontemperatures considerably higher than 40 °C. The polymer is intendedto be used as a coating on catheters in the form of small beads loadedwith chlorhexidine or silver-acetate, overlaid with an aromaticpolyether-based thermoplastic polyurethane as a permeable, thermallyinsulating material (European Patent Application: EP 2732832 A2). It isthe aim of this paper to compare the antimicrobial efficacies of the PS-BA polymer loadedwith either chlorhexidine or silver over an extendedperiod of time up to 60 days with the one of a commercially available,antimicrobial catheter, loaded with a combination of chlorhexidineand silver.

2. Materials and methods

2.1. Preparation of temperature-sensitive polymer and antimicrobialloading

In order to prepare a temperature-sensitive polymer with enhancedantimicrobial release characteristics above its glass transition tempera-ture, Tg, a copolymer of styrene (PS) and n-butyl (meth)acrylate (BA)was synthesized (abbreviated PS-BA). For chlorhexidine loading, a solu-tion of 1.92 mol styrene and 0.82 mol BA (styrene 99%, n-butylacrylate99%, Sigma–Aldrich Chemie B.V., Zwijndrecht, The Netherlands) wasprepared (70/30 mol% PS-BA) in 420 mL butyl acetate (99%, Sigma–Aldrich) and mechanically stirred at 69 °C in a 3-necked, 2 L round bot-tom flask under argon at 23 °C. A degassed suspension of 2.7 gazobisisobutyronitrile (98%, Sigma–Aldrich) in 20 mL butyl acetatewas added three times at intervals of 2 h. For a silver loadedtemperature-sensitive polymer, initial solutions of 1.60 mol PS and0.40 mol BA (80/20 mol% PS-BA) were employed in the same amountof butyl acetate as for the chlorhexidine loaded polymer (420 mL).The reaction times were approximately 6.5 h and 40 h for polymersintended for chlorhexidine or silver loading, respectively. The solutionwas poured into 1 L methanol and stirred for 5 min, separated fromthe supernatant by decantation, covered with 500 mL ethanol, stirredovernight and subsequently for 1 h in 1.5 L ethanol. As a next step, dry-ingwas performed in vacuo, first at 23 °C for 5 h, then overnight at 80 °Cresulting in 105 g of a colorlessmaterial, whichwas dissolved in 550mLchloroform and re-precipitated in 2 L methanol by drop wise addition.Finally, the polymer was filtered, stirred in 1 L of methanol for 1 h

(filtering and stirring were performed three times) and dried in vacuofor 96 h at 40 °C.

For antimicrobial loading of the polymer, the appropriate polymerwas compounded in a custom-built twin-screw mini-compounder at110 °C either with 20 wt.% or 30 wt.% chlorhexidine diacetate (70/30mol% PS-BA) or with 2 wt.% Ag acetate (80/20mol% PS-BA polymer)to produce strands of 100mgwith a diameter of 3.5mmand a length of2 cm. Glass transition temperatures of the antimicrobially loadedpolymer strands were measured by differential scanning calorimetry(Perkin Elmer, Pyris Diamond, Waltham, Massachusetts, USA) on10 mg samples in a temperature range from 0 to 100 °C, yielding aglass transition temperature of 48 °C and61 °C for chlorhexidine and sil-ver releasing coatings, respectively. The concentration of antimicrobialloading had no effect on the glass transition temperatures.

For determination of antimicrobial efficacy, strands were compres-sion molded at 110 °C to prepare sample discs with a diameter of2 cm and a thickness of 1.5 mm and average weight of 400 mg. Priorto evaluation of antimicrobial efficacies, discs were sterilized by 3 minsonication in 70% ethanol followed by air-drying.

2.2. Chlorhexidine and silver release

The release of antimicrobials was assessed from 2 cm PS-BA strands.For comparison, 2 cm sections of a commercially available,chlorhexidine- and silver-sulfadiazine loaded CVCs (Arrowg + ardBlue Plus) were employed.

Samples were immersed in 50 mL sterile demineralized water inglass containers under mild shaking at 37 °C, i.e. well below the glasstransition temperature of PS-BA during a time period of 80 days. At reg-ular time intervals, the PS-BA strands and CVCs were removed andplaced for 1 h in a shaking bath at 60 °C. Aliquots (0.5 mL) were takendaily, but with a higher frequency immediately after a temperature-trigger, and the cumulative amounts of chlorhexidine or silver releasedweremeasured. Chlorhexidine concentrationswere determined byUV–VIS at 229 nm, using a calibration based on measuring the optical den-sity of a series of chlorhexidine solutions with known concentrations,while the concentration of Ag+ was determined by Graphite FurnaceAtomic Absorption Spectroscopy (GF-AAS) at 328 nm using a calibra-tion based on measuring a series of Ag+ solutions with knownconcentrations.

2.3. Antimicrobial efficacy

Staphylococcus aureus was selected because of its role in catheter-associated bloodstream infections [14]. S. aureus ATCC 49230 was cul-tured from cryopreservative beads (Protect Technical Surface Consul-tants Ltd., UK) onto blood agar plates at 37 °C in ambient air. Forexperiments, one colony was taken to inoculate 5 mL of tryptone soybroth (TSB, Oxoid, Basingstoke, Great Britain) at 37 °C for 24 h in ambi-ent air. This pre-culturewas diluted 1:20 in 200mL TSB and grown stat-ically for 16 h at 37 °C. Cultures were harvested by centrifugation for5 min at 5000 g and bacteria were re-suspended in 10 mL phosphatebuffered saline (PBS, 10 mL potassium phosphate and 150 mM NaCl,pH 7.0). Centrifugation was done twice in order to remove all traces ofgrowth medium after which staphylococci were re-suspended in PBSto a concentration of 2 × 104 bacteria/mL, as determined in a Bürker-Türk counting chamber. Viability of the staphylococci in suspensionamounted 90%, as determined using a Petrifilm Aerobic Count Plate®(3M Microbiology, St. Paul, MN, USA; see below). Next, 50 μL dropletsof demineralized water were deposited on a sample disk and atemperature-triggerwas applied by placing the sample discs in an incu-bator at 60 °C for 1 h. In order to limit evaporation and to retain constantconditions during antimicrobial release into the droplet, incubationwasexecuted under saturated vapor conditions (indicated as the “ON”-statein Fig. 1). Subsequently, the sample discs were cooled and left at 37 °Cfor 24 h, also under saturated vapor conditions during which small

Fig. 1.Assessment of the antimicrobial efficacy of temperature-triggered release of chlorhexidine or silver fromPS-BA discs. In theON-state, antimicrobial release occurs in thewater drop-let (50 μL) at 60 °C for 1 h, after which closure of the polymer matrix is ensured by keeping the sample discs at 37 °C for 24 h. Next, staphylococci (900 CFUs) are injected and thePetrifilmTM is closed for 48 h of incubation at 37 °C. A similar experiment was done in the subsequent “OFF”-state. Temperature-triggers were repeatedly given during a time periodof 63 days, while storing the samples in PBS at 37 °C in between triggers.

63J. Sjollema et al. / Journal of Controlled Release 188 (2014) 61–66

amounts of antimicrobials could still be released into the droplet (“tailrelease”). Next, sample discs with the droplets were carefully trans-ferred to a Petrifilm Aerobic Count Plate® after which 5 μL of the staph-ylococcal suspension was added, corresponding with a bacterialchallenge of 900 colony forming units (CFUs).

The Petrifilm system offers a water soluble agar gel top sheet withnutrients and coloring agents which, after “closing”, covered the top ofthe sample. To allowbacterial growth, the Petrifilm systemwas incubat-ed at 37 °C for 48 h, well below the glass transition temperature to en-sure shut-down of the polymer, after which the number of CFUs onthe gel was counted. When bacterial growth caused a homogeneousreddish film, the number of CFUs was too high to count.

Subsequently, sample discs were thoroughly rinsed by 3 min sonica-tion in 70% ethanol followed by air drying and the above procedure wasrepeated again, nowmaintaining the temperature at 37 °C during the en-tire experiment: no temperature triggerwas applied (indicated as “OFF”-state in Fig. 1). The entire experiment involved four temperature-triggers(including “ON”- and “OFF”-states) at days 1, 9, 18 and63 and antimicro-bial efficacies associated with each of the triggers were determined. Inbetween temperature-triggers, sample discs were kept in PBS at 37 °C.All experiments were carried out in three-fold with different samplesand separately grown bacterial cultures.

Similarly, the antimicrobial release and associated efficacy ofcommercially-purchased Arrowg+ard Blue Plus catheters (Teleflex In-corporated, Limerick, PA, USA) were evaluated as a function of timewith a slight modification in order to accommodate the use of cathetersections. Fifty microliters of the staphylococcal suspension was placedon the inside of the top sheet of a Petrifilm system and the sheet wassubsequently folded around a 2 cm section of the Arrowg + ard BluePlus catheter, after which the system was used as described above.Again antimicrobial efficacy was assessed four times all on the samecatheter section (days 1, 9, 18 and 63), while keeping the sections at

37 °C in PBS in between measurements. All experiments were carriedout in three-fold with different samples and separately grown bacterialcultures.

3. Results and discussion

3.1. On-demand and ad libitum release of chlorhexidine and silver

Fig. 2 shows an example of the temperature-triggered (Fig. 2a and b)release of chlorhexidine and silver from PS-BA strands over an 80 daytime period. Release rates for chlorhexidine amounted around 90 ng/min/cm2 to 180 ng/min/cm2, and were higher for 30 wt.% than for20 wt.% loaded polymer strands (Fig. 2c). Release rates remained rela-tively high during 4 triggers, but decreased to lower levels after 60 days.

In the shut-down phase (the “OFF”-state), chlorhexidine releaserates dropped to around 0.5 to 14 ng/min/cm2 for the 20 wt.% and30 wt.% loaded polymer strands, respectively. Silver release ratesamounted to 1 to 3 ng/min/cm2 (Fig. 2d) and opposite to observationsfor chlorhexidine remained constant over the different temperature-triggers. Cumulative amounts of chlorhexidine and silver released arepresented in Fig. 2e and f. Cumulative release of chlorhexidine uponsix temperature triggers over the 80 day experimental period amountsaround 180 μg/cm2 and 270 μg/cm2 from the 20wt.% and 30 wt.% load-ed polymer strands respectively, corresponding with 2% of the totalamount of chlorhexidine incorporated in the strands for both concen-trations. The cumulative amount of released silver upon ten tempera-ture triggers however, was much lower (2.7 μg/cm2, equivalent to0.3% of the total amount incorporated). Both observations indicatethatmost of the antimicrobials remain incorporatedwithin thematerialupon six to ten temperature triggers and as a consequence it can be ex-pected that more triggers will still yield antimicrobial release.

Fig. 2. Examples of the temperature triggers applied to PS-BA strands (a and b) and resulting release rates (c and d) and cumulative amounts released (e and f) of chlorhexidine and silverfrom PS-BA sample discs as a function of time.

64 J. Sjollema et al. / Journal of Controlled Release 188 (2014) 61–66

The ad libitum release of chlorhexidine and silver from Arrowg+ ardBlue Plus catheter sections is presented in Fig. 3 and shows a burst releasein the first ten days after immersion in PBS, while no further increase in

Fig. 3. Ad libitum release rates (a and b) and cumulative amounts released (c and d) of chlorhe

cumulative release of either antimicrobial is observed after 20 days. Themaximal cumulative release amounted to 310 μg/cm2 for chlorhexidineand 14 μg/cm2 for silver. These amounts are comparable with the

xidine and silver from sections of Arrowg + ard Blue Plus catheters as a function of time.

65J. Sjollema et al. / Journal of Controlled Release 188 (2014) 61–66

amounts released from the PS-BA strands upon six to ten temperaturetriggers.

3.2. Antimicrobial efficacy of on-demand and ad libitum releases

S. aureus growth was inhibited by 3 log-units upon temperature-triggered release of chlorhexidine from the 30 wt.% chlorhexidine sam-ples over the entire experimental period, whereas significant antimicro-bial efficacy differences between the “ON”-state and “OFF”-state wereobserved for all 4 triggers (p b 0.05based on Student's t-test,seeFig. 4). This implies that the 30 wt.% loaded polymer can be effectivelyactivated on-demand, even after 63 days and that the experimental pe-riod might have been extended to longer times or that the sample discsmight have been successfully triggered more often during the experi-mental period.

Antimicrobial efficacy was evaluated in a Petrifilm assay as it allowsrelease and evaluation in a small volume, such as between a CVC andsubcutaneous tissue. Accordingly, the concentrations of chlorhexidinereached in the water droplets (see Fig. 1) on the different sample discsduring the “ON” state (around 500 mg/L or 1 mM) were well abovetheminimal inhibitory concentration of chlorhexidine against staphylo-cocci (around 1 mg/L or 2 μM) [15], which explains the high antimicro-bial efficacy observed in the Petrifilm assay. It is likely that such highconcentrations are within reach in the subcutaneous volume surround-ing a CVC in the absence of significant fluid flow, but clearly not for theintravascular parts of a CVC. In case of a CVC-associated bloodstream in-fection however, the catheter section exposed to blood will be coveredwith biofilm and chlorhexidine release will be directly into the biofilm.Since biofilm is known to absorb antimicrobials [16], it is likely thatchlorhexidine concentrations high enough to yield antimicrobial effica-cy will accumulate directly in CVC biofilms in the blood stream.

For the 20 wt.% chlorhexidine loaded sample discs antimicrobialgrowth was inhibited by almost 2 log-units after the fourth trigger

Fig. 4. Number of staphylococcal CFUs per sample after temperature-triggered release ofchlorhexidine (Chex) from PS-BA with 20 wt.% or 30 wt.% chlorhexidine loading or 2wt.% silver-acetate (AgAc) loading. Bacterial growth inhibition was measured after the“ON”-state and during the “OFF”-state for each of four consecutive temperature-triggerson one and the same sample, for both conditions inoculated with 900 CFUs. As a control,growth inhibition was also measured on unloaded PS-BA materials. The days indicatethe time at which each of the consecutive experiments was executed. Error bars indicatestandard deviations over three experiments with different samples and different bacterialcultures. Asterisks indicate that bacterial growth caused a homogeneous reddish film andthe number of CFUs amounted maximally 900. † signs indicate significant differences (pb 0.05) between the “ON”-state and the “OFF”-state.

(Fig. 4) as compared to unloaded samples, but differences betweenthe “ON”-state and “OFF”-state were non-significant after the thirdand fourth triggers, probably because the release of chlorhexidine inthe “OFF”-state is not entirely zero, as can be observed from Fig. 2e. Be-cause unloaded sample discs showed nomeasurable growth inhibition,no other antimicrobial substances releasing from the polymer couldhave caused this growth inhibition during the “OFF”-state.

Silver loaded PS-BA demonstrated no growth inhibition beyond thefirst trigger, probably because the high minimal inhibitory concentra-tions of silver-acetate against staphylococci (around 10 mg/L or60 μM) [17] were at reach only during the first trigger with silver con-centrations around 100 μMwithin the 50 μL volumes of thewater drop-lets employed.

Staphylococcal growth inhibition by the ad libitum releasing catheterwas complete on day 1 and day 9 after initiating antimicrobial release,but already after 18 days growth inhibition was fully absent (Fig. 5).Since silver was largely ineffective as an antimicrobial from the on-demand coatings, and amounts of chlorhexidine were released in al-most similar amounts from our on-demand release coatings as fromthe ad libitum releasing CVC (compare Figs. 2 and 3), we here speculatethat the antimicrobial efficacy of the ad libitum releasing catheter ismainly due to chlorhexidine and not to silver. This speculation is in ac-cordance with literature, describing that silver releasing catheters areclinically ineffective in reducing the incidence of catheter-associatedbloodstream infections [7].

4. Conclusion

Summarizing, we have described the advantages of temperature-triggered release of antimicrobials from PS-BA polymers over their adlibitum release from coatings. Temperature-triggered release was in-duced by heating the PS-BA polymer to above its glass transition tem-perature. In the current study we elevated the temperature to wellabove the glass transition temperature of antimicrobially loaded PS-BA polymers to demonstrate the advantages of on-demand releaseabove ad libitum release. Glass transition temperatures were higher forsilver releasing coatings (61 °C) than for chlorhexidine releasing ones(48 °C), which may both be sufficiently high to prevent antimicrobialrelease due to incidental rises in body temperature coatings. However,such high trigger temperatures, especially of the silver loaded PS-BApolymer, bear the risk of causing tissue damage. The antimicrobial effi-cacy of silver loaded PS-BA polymers was extremely low as comparedwith chlorhexidine loaded ones, implying that for clinical applications

Fig. 5.Number of staphylococcal CFUsper sample for Arrowg+ard Blue Plus catheter sec-tions on one and the same sample, repeatedly inoculated with 900 CFUs at the days indi-cated. Results shown are averaged over three experiments with different samples anddifferent bacterial cultures. Standard deviationswere zero for all measurements. Asterisksindicate that bacterial growth caused a homogeneous reddish film and the number ofCFUs amounted maximally 900.

66 J. Sjollema et al. / Journal of Controlled Release 188 (2014) 61–66

polymers solely loaded with chlorhexidine and with a glass transitiontemperature of only 48 °C, will be sufficiently efficacious. Yet, exposureof subcutaneous tissue to this temperature is harmfulwhen applied lon-ger than 1–20 s [18]. Therefore, for clinical applications we foresee anoverlayer consisting of an aromatic polyether-based thermoplasticpolyurethane as a permeable, thermally insulating material (EuropeanPatent Application: EP 2732832 A2). Such an insulating overlayer maynot be needed in the intra-vascular section of a catheter, where due toblood flow, cells will be in contact with the heated part of the catheterfor less than 5 s. Reportedly [19], such short episodes of temperature in-creases from 51 °C to 53 °C do not yield increased hemolysis. In the finalapplication (European Patent Application: EP 2732832 A2), this willyield a coating with a thickness between 200 and 400 μm. Based onthe antimicrobial efficacy arising from the 1.5 mm thick chlorhexidinePS-BA strands, this coating thickness will yield sufficiently high antimi-crobial concentrations.

Acknowledgments

This research forms part of the Project P4.01 NANTICO of the researchprogram of the BioMedical Materials institute, co-funded by the DutchMinistry of Economic Affairs (P4.01 NANTIC). JS, RJBD, HCvdM and HJBare employed by UMCG, Groningen, The Netherlands. CA and DvA areemployed byDolphysMedical B.V. (De Lismortel 31, 5612AREindhoven,The Netherlands), while HJB is also a director of a consulting company,SASA BV (GN Schutterlaan 4, 9797 PC Thesinge, The Netherlands). Theauthors declare no potential conflicts of interest with respect to author-ship and/or publication of this article. Opinions and assertions containedherein are those of the authors and are not construed as necessarilyrepresenting views of the funding organization or their respectiveemployers.

References

[1] J. Klekamp, J. Dawson, D. Haas, D. DeBoer, M. Christie, The use of vancomycin andtobramycin in acrylic bone cement — biomechanical effects and elution kineticsfor use in joint arthroplasty, J. Arthroplasty 14 (1999) 339–346.

[2] D.C. McGee, M.K. Gould, Current concepts— preventing complications of central ve-nous catheterization, N. Engl. J. Med. 348 (2003) 1123–1133.

[3] K. Halton, N. Graves, Economic evaluation and catheter-related bloodstream infec-tions, Emerg. Infect. Dis. 13 (2007) 815–823.

[4] L.A. Mermel, What is the predominant source of intravascular catheter infections?Clin. Infect. Dis. 52 (2011) 211–212.

[5] N. Safdar, D.G. Maki, The pathogenesis of catheter-related bloodstream infectionwith noncuffed short-term central venous catheters, Intensive Care Med. 30(2004) 62–67.

[6] N. Chaiyakunapruk, D. Veenstra, B. Lipsky, S. Saint, Chlorhexidine compared withpovidone-iodine solution for vascular catheter-site care: a meta-analysis, Ann. In-tern. Med. 136 (2002) 792–801.

[7] A.L. Casey, L.A. Mermel, P. Nightingale, T.S.J. Elliott, Antimicrobial central venouscatheters in adults: a systematic review and meta-analysis, Lancet Infect. Dis. 8(2008) 763–776.

[8] V. Monzillo, S. Corona, P. Lanzarini, C. Dalla Valle, P. Marone, Chlorhexidine-silversulfadiazine-impregnated central venous catheters: in vitro antibacterial activityand impact on bacterial adhesion, New Microbiol. 35 (2012) 175–182.

[9] C.P. McCoy, C. Brady, J.F. Cowley, S.M. McGlinchey, N. McGoldrick, D.J. Kinnear, G.P.Andrews, D.S. Jones, Triggered drug delivery from biomaterials, Expert Opin. DrugDeliv. 7 (2010) 605–616.

[10] S. Mura, J. Nicolas, P. Couvreur, Stimuli-responsive nanocarriers for drug delivery,Nat. Mater. 12 (2013) 991–1003.

[11] D.E. Meyer, B.C. Shin, G.A. Kong, M.W. Dewhirst, A. Chilkoti, Drug targeting usingthermally responsive polymers and local hyperthermia, J. Control. Release 74(2001) 213–224.

[12] T. Hoare, J. Santamaria, G.F. Goya, S. Irusta, D. Lin, S. Lau, R. Padera, R. Langer, D.S.Kohane, A magnetically triggered composite membrane for on-demand drug deliv-ery, Nano Lett. 9 (2009) 3651–3657.

[13] J.T.F. Keurentjes, M.F. Kemmere, H. Bruinewoud, M.A.M.E. Vertommen, S.A. Rovers,R. Hoogenboom, L.F.S. Stemkens, F.L.A.M.A. Peters, N.J.C. Tielen, D.T.A. vanAsseldonk, A.F. Gabriel, E.A. Joosten, M.A.E. Marcus, Externally triggered glass transi-tion switch for localized on-demand drug delivery, Angew. Chem. Int. Ed. 48 (2009)9867–9870.

[14] A.I. Hidron, J.R. Edwards, J. Patel, T.C. Horan, D.M. Sievert, D.A. Pollock, S.K. Fridkin,Natl Healthcare Safety Network Tea, Participating Natl Healthcare Safe,Antimicrobial-resistant pathogens associated with healthcare-associated infections:annual summary of data reported to the National Healthcare Safety Network at thecenters for disease control and prevention, 2006–2007, Infect. Control Hosp.Epidemiol. 29 (2008) 996–1011.

[15] S.M. Tambe, L. Sampath, S.M. Modak, In vitro evaluation of the risk of developingbacterial resistance against antiseptics and antibiotics used in medical devices, J.Antimicrob. Chemother. 47 (2001) 589–598.

[16] M.P.T. Otten, H.J. Busscher, F. Abbas, H.C. van der Mei, C.G. van Hoogmoed, Plaque-left-behind after brushing: intra-oral reservoir for antibacterial toothpaste ingredi-ents, Clin. Oral Invest. 16 (2012) 1435–1442.

[17] J.M. Schierholz, L.J. Lucas, A. Rump, G. Pulverer, Efficacy of silver-coated medical de-vices, J. Hosp. Infect. 40 (1998) 257–262.

[18] D.M. Simanovskii, M.A. Mackanos, A.R. Irani, C.E. O'Connell-Rodwell, C.H. Contag, H.A. Schwettman, D.V. Palanker, Cellular tolerance to pulsed hyperthermia, Phys. Rev.E. 74 (2006) 011915.

[19] D. Herron, R. Grabowy, R. Connolly, S. Schwaitzberg, The limits of bloodwarming:maximally heating blood with an inline microwave bloodwarmer, J. Trauma 43(1997) 219–226.