Diagnostics for non-blood-associated pathogens often use swabs as a specimen-collecting tool. For example, swabs are used to collect throat specimens for Group A Streptococcus [1]; nasal and nasopharyngeal specimens for Staphylococcus aureus [2], [3], Bordetella pertussis, influenza virus [4], [5], and respiratory syncytial virus (RSV) [6], [7]; female endocervical or male urethral specimens for Neisseria gonorrhea [8] and Chlamydia trachomatis [9]; and fecal swabs for viral gastroenteritis [10]. Depending on the source of collection, swabs should have shaft properties (flexibility, length) and tip size/shape appropriate for the sampling site, and the swab tip material and microstructure should provide efficient sample capture and target release in the presence of sample matrix components (e.g., human cells, body fluids, and other contaminants). Commercially available swabs are currently utilized with a variety of swab tip materials (e.g., nylon, rayon, cotton, polyester, polyurethane, and alginate polymer) and microstructures (e.g., tightly wound, knitted, flocked fiber, and reticulated). In laboratory settings, swabs are typically agitated by vortex mixing to release organisms into a transfer fluid [11]–[15] that is analyzed by culture, immunoassays (ELISA), or nucleic acid tests (PCR).

Swab sampling and fluid transfer are also used in lateral flow tests (LFTs) intended for point-of-care (POC) testing in non-laboratory settings. Commercial LFTs that are being used with swabs worldwide include rapid streptococcal antigen assays, respiratory syncytial virus (RSV) assays: BinaxNOW RSV Lateral Flow (Alere Inc., Waltham, MA), RSV Respi-Strip (Coris Bioconcept, Namur, Belgium); and influenza detection tests: BinaxNOW Influenza A&B Card (Alere Inc., Waltham, MA), QuickVue Influenza Test (Quidel Corp., San Diego, CA). Typical instructions require the user to dip the swab into transfer fluid (∼0.7 mL to 1 mL), manually agitate the swab for a specified time [16], and transfer a fraction of the fluid (∼100 µL) to the device [17]. The low fluid capacity of LFTs results in most of the sample being discarded, and manual agitation may be less effective than vigorous vortex mixing used in laboratory settings. Since the sensitivity of LFTs is typically lower than laboratory-based tests, there is a need to maximize transfer of sample from the swab to the device.

Studies evaluating swab transfer use widely varied definitions of transfer efficiency, often focus on a specific clinical application, and typically rely on qualitative analysis techniques. Thus, we built on previous work to develop quantitative analysis methods and definitions for swab transfer efficiency that can be applied to a variety of swab types, sampling applications, and transfer methods. We present methods to quantify swab transfer efficiency, discuss potential pitfalls that could bias quantitative analysis, and evaluate transfer efficiency for a range of swab types, sample properties, and manual agitation methods that meet the unique needs for POC applications. The results and discussion can aid researchers or test developers in rational selection of swabs and transfer methods for diagnostics development in both POC and laboratory settings.

A variety of swab types were evaluated under different conditions, as illustrated in Figure 1. We selected a set of commercially-available swabs representing a range of materials, microstructure, and size: low absorbent foam (reticulated polyurethane swab – PUR and knitted-pattern PES swabs – PES), low absorbent fiber (MT nylon flocked swabs – MT nylon, and regular-tip nylon flocked swabs), high absorbent fiber (tightly wound cotton and rayon swabs), and dissolvable swabs (calcium alginate swabs). Figure S3 shows scanning electron microscopy images of the nylon (flocked), PUR (reticulated), PES (knitted), and rayon (wound) swab tips; this set exhibits a wide variety of material structures and pore sizes. We focused on the manual “insert and twirl” agitation method common for LFTs and used a reduced fluid transfer volume (∼100 µL) appropriate for typical LFTs. Swabs were evaluated in a series of tests, with subsets of swabs chosen for each test to illustrate key differences and analysis methods. Swabs were first tested for fluid retention (loss of fluid sample to the swab), and swabs with low retention volume were carried forward to test organism recovery. We used Staphylococcus aureus as a model system, with recovery quantified by quantitative polymerase chain reaction (qPCR). Organisms were applied to swabs under various conditions to represent a range of sampling conditions from dry to wet: a) low-volume fluid sample (less than swab saturation), b) high-volume fluid sample (beyond swab saturation), and c) sample dried on a surface. For the low-volume sample condition, recovery was tested for a wide range of concentrations (100 to 10⁶ organisms) [20], and a simulated human nasal matrix was used as an example of a complex sample. Pre-wet swabs were compared to dry swabs for the cases of excess sample volume and dry sample collection. We also tested variations on manual agitation to identify sensitivity to user operation, and demonstrated an engineered manual agitation method to improve recovery for swab-sample combinations that performed poorly. The data from all experiments are summarized in Tables S1 and S2.

In this study, we characterized several important properties of swabs currently utilized world-wide in many POC diagnostic tests. The experimental design had three overall aims: 1) to provide quantitative analysis and definitions of swab transfer efficiency that can be used to evaluate swabs and transfer methods, 2) to identify differences in swab transfer efficiency for a selection of common swabs and sample conditions, and 3) to identify commercially-available swabs that perform well for sample conditions and manual agitation methods typical of POC testing. The results demonstrate that the choice of swab type for POC testing may be critical to achieving a sensitive and reliable test.

We built on previous work to develop analysis methods and definitions for quantitative evaluation of swab transfer efficiency. Some studies have assumed that the volume absorbed is an accurate representation of organisms picked up by the swab [14], [21], [22]. This assumption applies only when organism densities are the same inside and outside the swab; it will be inaccurate if organisms accumulate in the swab during agitation or if organisms do not flow freely into the swab. Our results show that swab transfer efficiency cannot be predicted based on fluid volume recovery alone, rather it requires analysis of organism recovery. The method for measuring the number of organisms in a sample also varies across studies. For bacteria, common measures include optical density at 600 nm (OD₆₀₀), colony counts (as colony forming units, CFU) [14] [23]–[25], and quantitative polymerase chain reaction (qPCR). However, OD is only applicable to high organism density. In colony counting, the transfer procedure itself can affect organism vitality (e.g., due to fluid composition, physical damage by vortexing or agitation) [26]. For clustering or biofilm-forming organisms [27], such as S. aureus, fragmentation during the transfer procedure can artificially increase the apparent recovery [26]. Thus, colony counting can bias results when comparing transfer methods or using a control to calculate percent organism recovery since each sample experiences different conditions. In contrast, qPCR provides consistent results independent of cell viability or clustering [28] [29], and it allows calculation of transfer efficiency by normalizing organism recovery to an input control sample analyzed by the same method. Our results demonstrate that qPCR and input controls allow quantitative measurement of swab transfer efficiency that can be applied across studies.

The scenario of low-volume samples showed the largest difference between swab types. For example, the sample beaded on the hydrophobic tip of the PUR foam swab and was promptly released into the lysis buffer; this scenario is not a realistic sampling case since the swab resists collection of the small fluid volume, and the high recovery should be interpreted with caution (although we note that PUR swabs also gave high recovery for high-volume samples and dry samples). In less hydrophobic materials such as PES and MT nylon, a thin film coating of bacterial solution was created surrounding the swab tip resulting in intermediate recovery. The hydrophilic tip of the rayon swab [14] fully absorbed the small sample solution, and absorption of transfer fluid may have further driven sample into the swab interior resulting in poor recovery during agitation. Thus, the wettability of the swab itself may impact specimen recovery.

Intuitively, swabs with larger volume capacity will collect more organisms when excess fluid sample is available. We hypothesized that, in addition to the swab volume capacity, swab chemical composition and structure can impact organism pick up and release. To investigate this, we normalized the number of organisms recovered by the estimated number of organisms that would have been picked up if there were no impact of material composition and structure. The nylon, PES, and rayon swabs all transferred near the expected number of organisms based on their volume capacity, but the normalized organism recovery for the PUR swab was more than 2.5 fold higher than the other swabs (Fig. 4C). This result suggests that swabs may accumulate organisms during agitation. Surprisingly, the PUR and rayon swabs maintained their recovery under pre-wet conditions. For the PUR swab, the large open pore structure may have allowed effective exchange of pre-wetting fluid with sample during collection. For the rayon swab, the large tip and dense structure may have prevented access to sample in the swab interior (as in Figure 3) for both dry and pre-wet conditions; the good recovery would then imply that rayon also accumulated organisms on its outer surfaces during collection. The results suggest that swab composition and structure can have a significant impact on collection and release efficiency and should be evaluated for the specific organism and sample type.

Swabs can be used to collect samples from dry surfaces, such as for environmental testing and sampling from skin or dry nasal passages, and pre-wet swabs are sometimes used to increase collection efficiency for dry samples [24]. The improved performance of rayon swabs with dry samples (Fig. 5B) compared to the low-volume case (Fig. 3B) suggests that organisms collected from a dry surface were more accessible to manual agitation. This would be expected since dry collection deposits sample on the exterior surface of the swab, whereas liquid samples wick into the swab interior where they may be inaccessible by manual agitation. Similarly, the effect of pre-wetting was more pronounced for rayon and regular nylon (Fig. 5B), which have relatively high volume absorption compared to PES and PUR (Fig. 2B). The pre-wetting step filled the interior volume of swabs prior to sampling, and presumably this allowed dried sample to remain on the exterior surface of swabs where they were more effectively released during manual agitation.

Biological fluids found in clinical specimens can affect organism recovery. Complex matrices can affect physical properties (e.g., viscosity) or chemical properties (e.g., binding to swab materials or passivating swab surfaces). For example, mucin has been found to reduce non-specific binding of protein and has been used to coat biomaterials to create non-fouling surfaces [30] and repel other negatively charged molecules [31] (e.g., DNA). However, the increased viscosity of complex matrices may reduce the effectiveness of agitation. As an example of a sample matrix, we measured recovery in the presence of simulated nasal matrix (SNM) and found that it had little effect on organism recovery (Fig. 3C). The effect of sample matrix will be highly dependent on the sample type and should be evaluated for each application; the experimental design used here for SNM can be applied to quantitatively measure the effects of sample matrix on swab transfer efficiency.

User variations in swab transfer procedure could affect test sensitivity and reproducibility. Commercial LFTs typically include instructions to agitate the swab for a given time but do not specify the method. Our results with selected swabs showed that variations in agitation method and time had modest impact on organism recovery for the case of low-volume samples. Robustness to user variation will be especially important to maintain sensitivity and reproducibility for POC tests performed by untrained users and should be tested for all applications.

We demonstrated that an engineered swab transfer method can increase recovery from poorly performing swabs. However, the improved organism recovery was still significantly lower compared to vortexing (gold standard). Other engineered methods could include buffer-filled swab shafts that push fluid from the swab interior or methods that compress the swab to remove absorbed fluid. This finding has direct application to future developments in POC testing.

The PUR swab was the best performing swab across all sample conditions. This may be of clinical interest as similar PUR swabs have been found to be sensitive for the detection of respiratory viruses in immunocompetent and immunocompromised human subjects [32]. In addition, smaller swab size may be applicable to clinical use in subjects of diverse size and in sampling of diverse structure. The small volume of the PUR swab and reproducible recovery for variations in agitation methods also make it well-suited for application in POC devices.

Limitations of this work included the use of analytical samples instead of clinical samples, and the use of a single organism, Staphyloccocus aureus, as a model due to its relevance as a human pathogen. However, these choices allowed us to create replicable cultured samples with known numbers of organisms and to create various sample types representative of swab collection sites (from dry to wet samples, in the presence or absence of other biological components). Other limitations of this work include the evaluation of a limited number of swab types, although they represent some of the most common commercially-available swab types used in clinical testing today. Finally, although the number of organisms transferred from swabs involves both collection and release, swab collection efficiency is highly dependent on the target pathogen and details of the sampling site; our work only focused on swab transfer release efficiency.

We have built on previous work to develop a quantitative method to evaluate and compare swab transfer performance. We evaluated a variety of swabs under manual agitation conditions appropriate for POC testing. By selecting a set of commercially-available swabs representing a variety of tip sizes, shapes, and materials, and utilizing qPCR as a direct measure of target quantity, we were able to quantitatively measure the transfer efficiency of a model organism. Our data show that swab size, structure, or composition affects swab release performance under different sampling conditions (low-volume, excess volume, or dry samples). Variations in manual agitation method and time had modest impact on swab transfer efficiency for three swabs tested, which is encouraging considering the likelihood of user variation in POC tests. For cases when a test is constrained to a swab with poor transfer efficiency, we demonstrated how forced-flow transfer methods could be used to improve transfer efficiency. The results and discussion presented here highlight key factors that should be considered in selecting swabs for POC applications. The quantitative evaluation methods developed here can be applied to other swab types in the future, both for POC applications or laboratory tests.