Selection of fluoropolymer for FILP coating
First, we set out to select suitable fluoropolymers which can support a layer of LP coating against shear force and the resulting surface would slide off of a 5 µL water droplet when the surface is tilted at 5-degree angle, a characteristic property well known for slippery omniphobic coatings. Commercially available 125-micron thick films of different fluoropolymers e.g. polytetrafluoroethylene (PTFE), fluorinated ethylene propylene (FEP), perfluoroalkoxy alkane (PFA), polyvinylidene fluoride (PVDF) and ethylene tetrafluoro ethylene (ETFE) were attached adjacent to each other on a 20 × 25 cm aluminum plate, for direct comparison. A Pebax film of comparable thickness (negative control) and an F-SAM deposited Pebax film (positive control, shows slippery omniphobic properties) produced by the previously reported CVD technique24 were also attached to the same aluminum plate. All polymer films were lubricated with perfluorotributyl amine (LP) using a paint brush, and the excess LP was removed by vertically submerging the plate in a water bath. All LP coated films were tested for the slippery hydrophobic property by placing five water droplets (5 µL each) per film using a multichannel pipette while the plate was secured at a 5-degree angle. We defined failure when more than two droplets failed to slide off the surface. Only, LP immobilized on PTFE, FEP, PFA and F-SAM deposited Pebax (LP/F-SAM) films allowed all water droplets to slide off, whereas PVDF, ETFE and Pebax retained all five water droplets. In the next round of testing, we selectively compared PTFE, FEP, PFA and F-SAM deposited Pebax films after they were freshly lubricated with perfluorotributyl amine (LP). This time the plate was dipped in water bath consecutively in sets of 5, 10, 15, 20, 25 and 30 dippings, and after each set of dipping the water sliding ability was tested in similar manner, i.e., 5µL water droplets were placed on the films which were tilted at 5-degree angle. After dipping for five times, only PTFE, FEP and PFA allowed all five droplets to slide off, whereas more than three droplets failed to slide off the F-SAM deposited Pebax surface. The results are summarized in Fig. S-1 (supporting information). We repeated the experiment with heparinized sheep blood instead of water droplets and the blood droplet (20 µL) also slid off the surface at 5-degree slide angle (Video S-1). We also confirmed that PTFE, FEP or PFA does not show this low slide property with blood or water droplets in the absence of the LP layer, or if we allow the LP layer to evaporate off these fluoropolymer surfaces by exposing them to air flow for 30 min before the sliding test. The LP immobilized on PTFE, FEP and PFA surfaces survived dipping up to 30 times, indicating the durability of the omniphobic coating, whereas the LP/F-SAM deposited surfaces failed to pass the slide test after five dipping in all three replicated tests.
Slide angle and contact angle measurement
Hydrophobicity and slippery omniphobicity of the surfaces were evaluated by measuring water contact angles and slide angles on fluoropolymer surfaces, before and after the application of a LP layer. Three different LPs, perfluorodecalin, perfluorotributylamine, and perfluorotripentylamine of different viscosities and surface tensions (Table S-3, supporting information), were tested and the contact angle data is summarized in table S1 (supporting information). Static water contact angles on PFA, FEP, ETFE, and PVDF increased upon the addition of either of the three LP layers (Fig. S-2, supporting information) on the fluoropolymer surface. Perfluorodecalin showed an increase (~ 15°) in contact angle, when applied on both PFA and FEP surfaces. Contact angles on PTFE did not show any significant change upon the addition of either of the three LPs. In a sharp contrast, F-SAM deposited Pebax showed a reduction of more than 15° in contact angle upon the addition of either of the three LP layers. Sliding angle was measured as the minimum tilting angle required for a 5 µL water droplet to start sliding off the surface. Sliding angle expressed in logarithmic scale is summarized in Fig. S-3 (supporting information). The pure fluoropolymer surfaces showed > 45° sliding angle which dramatically decreased to 3°, 1° and 2.5° for PTFE, FEP and PFA, respectively, when these fluoropolymer surfaces were coated with the LP; but no significant change was observed on PVDF or ETFE surfaces upon the addition of any LP layer. A similar decrease in sliding angle was noted on F-SAM deposited Pebax surface as reported before24. Since perfluorotributylamine has the lowest surface tension, low volatility and showed a sliding angle comparable to perfluorodecalin which has the lowest sliding angle among all three LP tested, we decided to use it as the LP layer for the following experiments.
Producing FILP PICC
Perfluorinated plastics e.g. PTFE, FEP, PFA are hard28 and non-elastomeric materials, hence, PICC lines produced with pure fluoropolymers lack the mechanical flexibility required for use as an indwelling catheter. To maintain the flexibility of FILP coated PICC we decided to incorporate a thin layer of perfluoropolymer (PTFE, FEP or PFA) in both inner lumen and on the outer surface of the commercially available thermoplastic polyurethane (TPU) PICC, yielding a three-layer perfluoropolymer-TPU-perfluoropolymer composite structure. The process of modifying a BARD PICC with the fluoropolymer layers is shown in Fig. 1. For the inner liner, a commercially available PTFE sleeve was selected with a 25 ± 12 micron wall thickness and a 0.89 mm outer diameter (OD), slightly less than the inner diameter (ID, 1.02 mm) of commercially available 18 Ga PICC catheters. PTFE liners, similar to those used herein, are also used for the manufacturing of coronary guide catheters29 and are available from fluoropolymer tubing manufacturers around the world. Outer surface of these commercially available PTFE liners is chemically etched (as supplied) for stronger adherence to the thermoplastic polyurethane. To obtain the three-layer structure shown in Fig. 1, a stainless-steel mandrel was used as the guide mandrel and inserted inside the PTFE liner. Subsequently the PTFE lined mandrel was inserted into a commercially available 18 Ga, 4Fr single lumen PICC. Finally, a commercially available 50 ± 12 micron thick, heat shrinkable FEP (FEP-HS) outer cover was sealed over the entire assembly, to construct the three-layer PTFE-TPU-FEP structure. The 1.52 mm expanded inner diameter of the FEP-HS outer cover allowed us to conveniently slide it over the PICC-PTFE lined mandrel. The entire assembly was then heated in a convection oven at 89 ºC for 15 min to soften the TPU middle layer sandwiched between the PTFE and FEP layers. The heating causes uniform redistribution of the TPU middle layer between the inner PTFE liner and outer FEP-HS cover, which simultaneously undergoes thermal shrinkage to form uniform lamination on the soft polyurethane middle layer. The inner diameter of the heat shrunk FEP cover measures 1.24 mm, closely matching the 1.33 mm outer diameter of the 4Fr PICC catheter, allowing the FEP-HS to tightly fit onto the middle layer constructed by the 4Fr PICC line. Finally, the assembly was allowed to cool down to room temperature and the stainless-steel mandrel guide was removed to obtain the three-layer structure shown in Fig. 1, where both inner and outer surfaces of the PICC line were laminated with perfluoropolymer layers. The resulting PICC catheter showed mechanical flexibility comparable to a TPU PICC (video link S2, supporting information).
X-ray photoelectron spectroscopic (XPS) analysis was performed to determine the elemental composition of pure PTFE, FEP, PFA and PVDF surfaces as well as PICC line surfaces modified either by lamination with fluoropolymer (FEP & PTFE) layers or by F-Silane grafting (F-SAM/PICC). Elemental composition of pure fluoropolymer (FEP, PTFE, PFA, ETFE and PVDF) surfaces showed 61, 64, 65, 55 and 39 at.% F, respectively, which closely matched the molecular formula of those fluoropolymers, within the accuracy range (10%) of XPS analysis30. The results are summarized in Fig. 2. Elemental compositions of fluoropolymer (FEP and PTFE) laminated PICC line surfaces were consistent with the elemental composition of the pure fluoropolymers, which showed significantly high (> 60 atom%) fluorine compared to the elemental concentration (44 at.%) of fluorine on the F-SAM deposited PICC line surface.
Comparing in vitro thrombogenicity of FILP PICC, LP/F-SAM PICC with commercially available BARD PICC and BioFlo PICC
In a clinical setting, implanted PICC lines are typically exposed to a 1–10 cm/s intravenous blood flow velocity31. Following a standardized test protocol32, thrombogenicity of surface coated PICCs, namely LP/F-SAM PICC and FILP PICC was compared with two commercially available PICC lines: BARD PICC and BioFlo PICC under physiological flow conditions. For this test each PICC line was exposed to heparin stabilized (1 IU/mL) fresh ovine blood flow at 500 mL/min, similar to jugular blood flow rate33. Sheep blood used for this study was collected from 1–4 years old, healthy donor sheep of mixed breed, ranging 70–146 kg in weight. All donors were free from any anticoagulant medication e.g. aspirin, ibuprofen, acetaminophen, heparin or coumadin for two weeks before the blood was drawn for the thrombogenicity test. No blood was drawn from the donors during this two-week period. Porcine heparin was added during the venipuncture and collection process to yield a final concentration of 1 IU/mL. The heparin concentration used is comparable with other in vitro thrombogenicity study models published before34. The minimally heparinized fresh ovine blood was then analyzed for activated clotting time (ACT) which is within the therapeutic range (240–180 s) when it began circulating through a closed loop created with a 140 cm long blood perfusion tube. All catheter samples were inserted into the blood containing loop towards the direction of blood flow, using a peel away introducer and needle. This insertion procedure is clinically relevant. All PICC samples were collected and rinsed with saline at the end of the four-hour period and used for visual assessment to quantify the relative percentage of thrombus formation on the catheter surface. Since all catheter samples tested are identical cylinders with comparable diameter throughout the length, the total coverage of an identified area was calculated by taking the length of the thrombus formed multiplied by a visual estimate of the percentage of the circumference of the study article covered within the length. These segments were summed for the entire catheter, and this area was divided by the total implanted length and converted to a percentage. The thrombogenicity assessment thus obtained, as shown in Fig. 3, consistently established the absolute thromboresistance of FILP coated PICC with a 0% thrombus occupied catheter area, in all nine replicates analyzed over three repeated trials. Whereas both BARD PICC and BioFlo PICC had 57% and 34% thrombus covered surfaces, respectively. The LP infused F-SAM deposited PICC surface also showed a 35% thrombus occupied area which reflects its poor durability on the outer surface of the catheter. Digital images of a FILP coated catheter showed high contact angles of water droplets which indicates that no blood component was deposited on the catheter surface, even under the high blood flow conditions used for this test.
Resistance to CLABSI causing bacterial colonization
Efficacy of FILP PICCs in resisting bacteria colonization was compared with BARD PICCs by exposing both catheter samples to cell culture media containing 0.5% glucose and either S. epi or S. aureus, both bacteria known to cause CLABSI. After 48 h of incubation at 37 °C, the S. epi biofilm adhered to the catheter surface was dislodged by sonication and a viable cell count assay was done on the dislodged biofilm by the quantitative culture of bacterial species using tryptic soy ager plates. The S. aureus biofilm was allowed to grow for 96 h at room temperature, under 1 cm/sec flow of the cell culture media and then isolated and quantified following the same protocol. The bacterial colonization, measured as the colony forming unit (CFU) observed per centimeter length of each catheter tube are shown in Fig. 4G,H. Freshly coated FILP PICCs showed 87-fold reduction of S. epi biofouling and 37-fold reduction of S. aureus biofouling in comparison with BARD PICC. Both are statistically significant (p < 0.00001) reductions. SEM images of PICC after exposure to bacteria solution clearly show the presence of a bacterial biofilm on the BARD PICC whereas only discrete bacterial cells, instead of a compact biofilm, was observed on FILP PICC lines, as shown in Fig. 4A–F.
For its intended use, the FILP coating needs to continuously resist biofilm formation and blood stream infections at least for the first month, when bacteria from patient’s cutaneous microflora and care giver’s hands are believed to travel through the catheter wall, hubs and internal routes, leading to hematogenous dissemination and bacteremia35,36. To test the durability of anti-biofouling property of FILP PICCs under physiological flow conditions37, FILP PICCs were inserted into a LivaNova perfusion pack (0.95 cm ID) tubing loop, previously filled with phosphate buffer saline (PBS) which was circulated to mimic a 9.5 cm/s venous flow. Catheter samples were collected at 7, 14, 21, 30, and 60-day time points and further tested as follows. Uncoated BARD PICC as positive control, freshly coated FILP PICC as negative control and the catheter samples, harvested at various time points indicated above were challenged with S. epi in nutrient broth at 37 °C. After 48 h of incubation of the catheter samples in the bacterial culture broth, any S. epi biofilm formed was dislodged and simultaneously quantified by quantitative culture of bacterial species on tryptic soy ager plates. The durability of FILP coating, shown in Fig. 4G, was evaluated in terms of antibiofouling effectiveness of FILP PICC, which is quantified as biofilm formation per centimeter length of the catheter. FILP PICC showed significant (p < 0.00001) reduction of S. epi colonization throughout the period of 60 days when compared with BARD PICC. During the first 21 days, under physiological flow, FILP PICC did not show any significant difference from freshly coated FILP PICC. Although a statistically significant reduction of antibiofouling property was observed at the 30 (p = 0.0027) and 60-day (p = 0.00618) time points, FILP PICC still showed significant (p > 0.00001) improvement in resisting pathogen colonization throughout the entire 60 day period, when compared to BARD PICC. The perfluorotributylamine used as the LP layer of the FILP coating does not have any cytotoxic effect on pathogens, rather it resists bacterial colonization due to the slippery character of a molecularly smooth surface22. Since the slippery omniphobic surface of FILP coatings resists pathogen attachment we anticipate that it will resist CLABSI pathogens from traveling through the catheter surface into the blood stream and will also resist biofilm formation on indwelling catheter surfaces. Bacteria colonies in a mature biofilm eventually develop resistance against systemic antibiotics6,13,38 leading to significant complications in 400,000 cancer patients suffering from PICC associated CLABSI every year39.