Rapid Fabrication of Sterile Medical Nasopharyngeal Swabs by Stereolithography for Widespread Testing in a Pandemic

2.1 Conventional Swabbing Tools

This section offers a comparison of the different approved methods for the detection of the SARS-Cov-2 virus. Background information about each method is available in Supporting Information. Virus detection from NA samples was shown to be superior to NPS samples.^[16] However, aspirate methods require additional suction equipment that does not make this tool cost effective, scalable, and easily implementable for widespread testing, unlike the range of different NS and NPS which are easy and safe and can be done anywhere without the need for additional devices.^[17] One study showed that patients found the nasal wash (NW) more comfortable, in addition to being a more effective method than NPS in pathogen detection by culture, but with reverse transcription polymerase chain reaction (RT-PCR), both methods were comparable.^[18-20] On the contrary, another comparison between self-administered foam NS versus staff-collected NW showed that the former was comfortable, easy, and preferred by patients over the latter.^[20] The adequate detection sensitivity when using NS for a wide range of respiratory viruses has been proven in previous clinical work,^[17] as well as for NPS,^[21] where it was shown that NPS sampling is more effective than OPS sampling, but NW is more sensitive than NPS sampling. Another study observed that although NPS was more reliable than OPS, it was not optimal and that the combination of the two had a better detection yield.^[22, 23] However, data sensitivity must be balanced by the cost of additional swab fabrication, patient comfort, and compliance, hence why the additional use of OPS is unnecessary. A solution is to combine swab collection with saline spray, which showed to provide more sensitive results and was recommended as a better alternative to the high sensitivity of NW for conducting respiratory virus inspection.^[20]

Midturbinate swabs are designed to come into contact with a larger nasal surface area and for self-collection, as shown in Figure 2. Compared with other swabs, midturbinate swabs have an additional collar at the handle to guide the maximum insertion depths of 55 mm, tolerated for adults. The study showed that self-collected flocked midturbinate NSs were equivalent to staff-collected NS or NPS and more efficient than rayon NPS.^[24] Midturbinate swabs have shown to be less invasive and better suited for self-collection due to the safety collar, as well as similarly comfortable compared with NSs and at least as sensitive as standard NPSs.^[25, 26] The longer conical flocked head maximizes mucus contact, enabling better collection of specimens while offering a safe method of respiratory cell collection.^[27] The major drawback of the collar in midturbinate swabs is in its rapid fabrication. Its structure takes more volumetric space, thus reducing the number of swabs that can be printed in a batch when using stereolithography (SLA). Under noncrisis conditions, this is not an issue, but under current conditions, fast-paced production and item count are of the utmost priority.

Self-collected swabs have been proven to be a reliable alternative to healthcare worker-collected swabs, because self-swabbing is feasible without any necessary prior training.^{[20, 24, 25]} Specifically, self-administered collection has proven to be useful in community-oriented research.^[28] This could be a key facilitator of widespread testing of quarantined individuals during a pandemic outbreak and allow healthcare workers to allocate time to urgent treatment rather than testing. Furthermore, patients seem to prefer self-collection.^[25] In a comparison of flocked nylon fiber swabs versus polyurethane foam swabs, as shown in Figure 1, it was found that the former had a larger interacting surface area, held more fluid volume, and could release the fluid more readily into the testing medium than the latter. However, the latter is better in performance when it comes to anterior nares swabbing.^[29] Flocked nylon fiber swabs have been shown to collect more respiratory epithelial cells from the posterior nasopharynx.^[30] Flocked swabs also have the advantage of preserving samples at its extremities, allowing for easy sample separation during specimen testing preparation.^[25] Another study showed that different swab material and head or bristle designs impact sampling differently based on the collection site. This same study claimed however that the design of the swab seemed to matter less than the location of the virus in the respiratory tract.^[23]

This history of swab-type comparisons provides benchmarks for specimen volume collection, respiratory virus detection, comfort of use, flexibility, fabrication and scalability, cost effectiveness, and widespread implementation, as shown in Table 1. A variety of materials, beyond the polymers discussed here, have been used for swab designs in the past, including wood,^[31] which suggests that using nontraditional materials for swab fabrication, such as dental photopolymers, is not far-fetched. Although dental photopolymers are not cost effective due to the chemical complexity of the polymer composite, their rapid production ability using affordable widespread precise equipment, in addition to their comparable functionality and efficiency, renders them an attractive possibility.

Sterile Nasal Swab Nylon

2.2 Open-Source Collaborative Designs

Countless additive manufacturing companies across the world have joined the COVID-19 response efforts through the fabrication of biomedical equipment such as swabs, ventilators, face shields, and more. Many universities have transitioned their courses online, and research has been limited to essential work, leading to the emergence of new collaborations between laboratories, manufacturers, and healthcare institutions, in part, yielding important considerations for swab design, fabrication methods, and validation protocols. These, combined with those that have resulted from our team's collaborators—namely, Indiana University (IU) Protolab, IU Health, Eskenazi Health, and Deaconess Health—and previous assessments of regular manufactured swabs discussed in the previous section, have been listed in terms of constraints and objectives in Table 1.

2.3 Design Requirements

Designs must adhere strictly to requirements and avoid mass production printing errors, as the nasopharyngeal space is very sensitive and prone to tissue damage, bleeding, and irritation. Institutions seeking to develop their own designs should use these parameters as a reference for their work and should seek local partners for production and testing. Methods for assembling swabs with more conventional polymers, such as polypropylene (PP), polytetrafluoroethylene (PTFE), polydimethylsiloxane (PDMS), and more, safe adhesives, and wicking substrates have been reviewed.^[15] However, this fabrication process is hardly scalable for high throughput and requires significant labor if not automated, meaning that this solution should be left to contributors with existing industrial capabilities to that effect. The design fabrication should not depend on the labor of untrained students, researchers, and other members of the local community, even given the crisis conditions. While some have suggested that only immunocompromised patients need sterilized swabs, and that all other swabs can simply be disinfected,^[15] this is too low of a standard, especially when dealing with such a novel and dangerous virus. Further, as swabs are produced in mass, making sterilization the base requirement, as opposed to making separate batches of sterilized and disinfected swabs, is more efficient and eliminates the risk of caregivers, giving immunocompromised patients the wrong swab. Safety and efficiency are not only important to those producing swabs, but also to the frontline healthcare workers dealing with an already overwhelming workload.

Sterile Nasal Swab Nylon

2.4 Design Selection for Testing

At the Fibers and Additive Manufacturing Laboratory (FAMES Lab), we developed a variety of designs that meet the constraints and objectives of Table 1, as shown in Figure 1. These were designed to optimize specimen collection. They are printed out of biocompatible dental resin by SLA and their assessment is covered in the next sections. We include here designs X and XIV from Abiogenix and the University of South Florida, respectively. These parallel design efforts have been picked due to their additional evaluation. The Abiogenix design (X) picked was one of three preferred by Stanford Health in their feedback.^[15] The USF Health design (XIV) has gone through a series of validation processes and rapid clinical trials at Northwell Health and Tampa General Hospital, showing promising results in comparison with standard swabs. Figure 2 shows the accuracy level of the printing results of the selected designs and the FAMES Lab designs. FAMES Lab designs revolve around the inclusion of a hollow core to serve as a reservoir for carrying collected mucus when withdrawing a swab from the nasopharynx. In addition, FAMES Lab designs explore mechanisms and geometries such as those found in bubble wands, Archimedes screws, and negative Poisson ratios to optimize mucus collection and retention. We provide information here on swab prototype testing from open-source collaborative work in an attempt to set a general protocol for swab prototyping and fabrication through additive manufacturing technology.

3.1 Stereolithography

SLA is one of the oldest methods of 3D printing, invented by Hideo Kodama in 1981 as an automated fabrication of 3D objects through layers of hardened polymers by ultraviolet exposure.^[32] Rapid prototyping through SLA enables us to 3D print objects with arbitrarily complex geometries and with critical demands on precision and accuracy. SLA is a 3D vat polymerization printing process that involves the change of properties of light-activated resin polymers when exposed to ultraviolet or visible light, namely photopolymerization. The polymer composition involves a variety of chemical components such as photoinitiators, absorbers, precursors, additives, and fillers. The hardened resin is a collection of low-molecular-weight monomers bonded by covalent bonds into a solid crosslinked unit. As polymerization is an ongoing reaction that is never fully complete, it often involves required postprocessing steps in the process. Photopolymers are slow to dissolve and instead swell and soften when absorbing a solvent.^[33-35] Among light-activated composite resin types, dental resins have a history of use in dentistry, making it well suited to invasive human specimen sampling.

3.2 Autoclavable Dental Resins

In dentistry, the use of biocompatible photopolymers as an oral restorative biomaterial dates back to 1960s.^[36, 37] Dental resins are generally based off vinyl, polystyrene, or acrylic resins to be used as relining, die, impression, crown, and with other oral care applications.^[38] Although biocompatibility resists precise definition in areas like tissue engineering,^[39] in dentistry, an extended history of biocompatibility tests and protocols have been designed for local (mucosal and pulpal toxicity) and systemic (allergic, estrogenic, mutagenicity, and more) adverse reactions in vitro, in animals, in clinical studies, as well as occupational cyclic exposure.^[40-44] One notable biocompatible acute toxicity assessment test is performed on fish embryos, where their stages of development are observed for chemical effects from exposure to the material over time.^[45] Furthermore, the effects of the photo-hardening transition from liquid to solid resin have also been studied for a variety of conditions, such as shrinkage, irradiation, or cure depth effects.^[46-49]

Dental SG Resin (Formlabs, Inc.) as a photoreactive resin is a light-yellow translucent liquid and contains toxic elements. It has a viscosity between 800 and 1500 mPa s and an approximate composition of 90% methacrylic oligomers,^[50] a standard monomer group used in dental resins,^[41] and 3% phosphine oxides,^[50] used for material color stability.^[51] As the liquid resin holds toxic properties—causing possible irritation if ingested, inhaled, or through skin and eye contact—its curing process has to be followed strictly, especially for clinical applications.^[52] Dental SG Resin has a type D durometer hardness value of 80 according to standard ISO 868:2003 and a Carpy impact strength of 12–14 kJ m⁻² according to standard ISO 20795-1:2013. In terms of flexibility, the material has a flexural strength greater or equal to 50 MPa and a flexural modulus greater or equal to 1500 MPa, according to standard ISO 20795-1:2013.^[53] However, material properties can vary depending on the print geometry, temperature, and orientation. To address this concern, three-point bending tests done on swabs by Formlabs will be discussed in a later section of this paper.

The autoclave was first invented by Denis Papin in 1679 and is now a standard for sterilization used in biological and medical research and industry, hospitals, mortuaries, and waste disposal. Its cycle of operation combines saturated steam loaded with latent heat, elevated moist heat up to 134 °C or dry heat up to 180 °C, and pressure and clean water to create a germicidal environment and induce protein denaturation, within its sealed chamber. This standard process must be used for cleaning surgical tools and any implantable or intrusive medical equipment, such as swabs. The cured dental resin must therefore be stable both against heat and humidity in the autoclave process. Dental SG Resin has been designed in its proprietary chemical composition for that purpose, unlike Dental LT Resin by Formlabs. However, it has been observed that once autoclaved, the brittleness of the material increases requiring a further look into the properties of autoclaved dental resins.

4.1 Artificial Mucus

Xanthan gum is a commonly used biopolymer, acting as a thickening agent for applications such as food, pharmaceutical, and cosmetic formulations.^[54] The xanthan gum solution prepared from Xanthomonas campestris (Sigma-Aldrich, G1253) to mimic mucus with different viscosity properties will be hereinafter referred to as artificial mucus. Human mucus viscosity varies based on many factors, including the specific disease a patient is affected with, such as respiratory track diseases like rhinitis, chronic sinusitis, and chronic bronchitis.^[55] A wide viscosity range was achieved by altering the percentage weight to volume of the xanthan gum powder, as shown in Figure 4, effectively covering a broad range of viscosity behaviors. Homogeneity of the artificial mucus highly depends on mixing method. Highly viscous artificial mucus which is above 1% W/V needs more mixing time until the gum is dissolved homogenously without any clumping. Xanthan gum powder should be mixed directly when exposed to Milli Q water. Xanthan gum powder may stick to the bottom of the beaker, though this can be prevented by a glass stirrer to collect all agglutinate powders from the bottom. Mixing with a magnetic stirrer caused air bubbles in the solution which were eliminated by the desiccator before the viscosity measurement of each concentration of the artificial mucus to obtain accurate viscosity data.

4.2 Mucus Retention Dipping Test

Printed NPS were dipped 30 mm deep according to design requirements in 50 mL conical Falcon tubes. During each test, swabs were introduced without touching the inside of the Falcon tube, which was necessary to keep the contact area with the artificial mucus consistently sterile for each trial. Shaft design and the head of the volume did not affect the retention rate because the dipping motion was directed vertically, straight down within the confinement of the diameter of the 50 mL falcon tube, which was wide enough for the manipulation of the head volume of all NPS designs. This consistency is shown in general by the amount of small error bars (see Supporting Information) for each of the concentrations of the artificial mucus, resulting in a more coherent visual comparison of the different selected NPS designs.

Five trials of the 11 dipping tests for the 10 different NPS swab designs were conducted across 7 different concentrations of artificial mucus, as shown in Figure 4. When overall results from all the artificial mucus concentrations were considered with the normalization by the total head volume of the NSP, the highest retention was found in design II and design VIII. These have a similar design that heavily relies on the surface tension of the collected liquid to hold itself in the hollow reservoir of the head. Design II has higher head volume than design VIII, enabling it to hold a greater volume of sputum. A dramatic increase in retention was observed with design VI and VII past a threshold off 2% W/V concentration of artificial mucus, as the latter's viscosity increases. The dimension of the honeycomb structure affected the amount of artificial mucus collected by the NPS head in design VI and VII, suggesting that the larger honeycomb gaps result in better sample collection. NPS design II was designed alike Archimedes screw involving a specific rotational direction which helps withdraw sputum as the head is rotated during sample collection. Due to the direction-specific nature of this design, two dipping tests were administered clockwise and counterclockwise for design III. The results show that there is a slightly better retention rate when rotating the swab in the counterclockwise direction as expected.

We have compared the performance of 3D-printed designs to that of standard swabs (FLOQSwabs 503CS01, Copan Diagnostics). The result of those trials is shown in Figure 4, labeled as design XXI, where the flocked swab performed fine for low viscosities but surprisingly bad for higher viscosities, in terms of both average retention and its random variation, reflected by the larger error bars (see Supporting Information).

4.3 Nasal Model for Swab Behavior Analysis

The nasal model consists of a 3D-printed support structure, a 10 mL narrow mouth Erlenmeyer flask, and a plastic tube with an inner diameter of 6.35 mm, which was used to mimic the confined nose canal of the nasal cavity, as shown in Figure 5. The model simulated the nasal cavity, wherein the NPS has to be inserted to reach the nasopharynx, and was based on computed tomography measurements done on an asymptomatic adult-sized nasal air space.^[56] NPS designs with different shafts and different head volumes and the viscous behaviors of the artificial mucus both affected retention rates significantly. Due to the slight bend in the nasal pathway featured in the model, NSP with a rigid shaft design (designs V–VIII, X, and XIV) had a tendency to break while turning inside the 10 mL narrow-mouth Erlenmeyer flask. The flexible shaft performed much better with the nasal model. Large NPS heads (designs I–V) were not very efficient in the confined space of the tube. During the swab withdrawal step, a significant amount of the collected artificial mucus, especially the mucus with low viscous behavior, was lost. This loss invited instability in retention measurements for each trial with artificial mucus with same viscosity.

Equation (1): Volumetric flow of the withdrawn head with collected specimen, to which swab performances in the nasal model are normalized.

It is interesting to point out that in this test the clockwise direction of swab III performed better despite expectations. Design IV was created with a negative Poisson ratio, such that it would radially contract upon application of compressive stress along the head axis to maximize the patient's comfort during insertion and would radially expand upon application of tensile stress along the head axis to maximize sample retention during extraction. Indeed, the geometric shape of the head provided better retention among most of the other NSP designs. To address the lack of rigid shaft performance in the bent confined model, a comparison between the performance of a flexible handle and a rigid handle for the same head design II and VIII was conducted, as shown in Figure 5. The results show that the difficulty in maneuverability of the shaft barely affects the retention capabilities of the head designs.

The trials ran for the flocked swab, design XXI, shown in Figure 5, had a similar outcome as the dipping test. The flexibility of the industrial swab made the collection as easy as it was for designs I–IV, which have a flexible shaft. However, its retention of artificial mucus with high viscosity was suboptimal, unlike its efficiency with lower viscosity behaviors which were adequate. All performances for both the dipping and nasal model tests are shown in Table 2.