Point-of-care isothermal nucleic acid amplification tests: progress and bottlenecks for extraction-free sample collection and preparation

Abstract

Introduction

Suitable sample collection and preparation methods are essential to enable nucleic acid amplification testing at the point of care (POC). Strategies that allow direct isothermal nucleic acid amplification testing (iNAAT) of crude sample lysate without the need for nucleic acid extraction minimize time to result as well as the need for operator expertise and costly infrastructure.

Areas Covered

We review research to understand how sample matrix and preparation affect the design and performance of POC iNAATs. We focus on approaches where samples are directly combined with liquid reagents for preparation and amplification via iNAAT strategies. We review factors related to the type and method of sample collection, storage buffers, and lysis strategies. Finally, we discuss RNA targets and relevant regulatory considerations.

Expert Opinion

Limitations in sample preparation methods are a significant technical barrier preventing implementation of nucleic acid testing at the POC. We propose a framework for co-designing sample preparation and amplification steps for optimal performance with an extraction-free paradigm by considering a sample matrix and lytic strategy prior to an amplification assay and readout. In the next five years, we anticipate increasing priority on the co-design of sample preparation and iNAATs.

Introduction

Nucleic acid amplification tests (NAATs) enable highly sensitive and specific detection of nucleic acid targets. The workflow of most NAAT tests begins with sample collection from a patient, followed by sample preparation steps, and then subsequent amplification. Inside central diagnostic labs, NAATs generally rely on nucleic acid extraction followed by polymerase chain reactions (PCR). Nucleic acid extraction consists of the isolation, purification, and concentration of nucleic acids [1]. As depicted in Figure 1, one common method of nucleic acid extraction used in central diagnostic labs is column-based extraction kits which require significant time, expertise, and infrastructure.

Figure 1.

Comparison of laboratory-based and extraction-free workflows for sample preparation preceding iNAAT. Laboratory-based extraction involves sample collection, sample conditioning, cellular lysis, and column-based extraction and concentration of target. Ideal extraction-free sample preparation proceeds directly to cellular lysis without conditioning or extraction, reducing the need for laboratory expertise, and infrastructure, and shortening time to result. Created with BioRender.com.

There are significant efforts underway to deploy NAATs at the point of care (POC). For the purposes of this review, we employ the Lisby and Schneider definition that considers diagnostic testing performed outside a centralized laboratory to be the POC [2]. To be deployed at the POC, diagnostic tests should meet the REASSURED criteria [3] as defined by the World Health Organization; these criteria specify that a test should provide real-time connectivity, ease of specimen collection, be affordable, sensitive, specific, user-friendly, rapid and robust, equipment free or simple, and deliverable to end-users. In the case of NAAT testing, the steps of sample collection, preparation and amplification should all meet the REASSURED criteria.

There has been significant progress to develop NAAT assays that meet the REASSURED criteria. In particular, isothermal nucleic acid amplification tests (iNAATs) are particularly promising for use at the POC. iNAATs such as loop mediated isothermal amplification (LAMP) and recombinase polymerase amplification (RPA) have been widely investigated. The mechanisms of amplification vary widely between iNAAT technologies, and have been reviewed and compared recently by Glokler et al. [4], Oliveira et al [5], and Kang et al [6]. iNAAT amplification methods operate at a single temperature, reducing the time, expertise, and infrastructure requirements for nucleic acid amplification compared to laboratory-based PCR, making iNAATs favorable to the REASSURED criteria and appealing for use at the POC. Further, there has been significant progress in designing devices to deploy iNAATs at the POC, recently reviewed in Wang et al [7], Maffert et al [8], Pumford et al. [9], and Zasada et al [10]. To ensure an entire assay meets the REASSURED criteria and is suitable for use at the POC, iNAAT amplification steps must be integrated with equally accessible sample collection and preparation strategies. Extraction-free sample preparation is one promising approach to prepare samples in a way that adheres to the REASSURED criteria, and is the scope of this review. We define extraction-free as the collection of a crude sample from a patient in a fluid matrix, addition of any chemical agents, application of heat and/or mechanical force for lysis and/or stability, followed by the addition of the crude sample to the iNAAT reagents for amplification. This process, as outlined in Figure 1, reduces time as well as the need for equipment and user knowledge compared to traditional extraction methods, satisfying the REASSURED criteria for POC testing. For this review, we consider fluid-based samples, excluding dried blood spots and similar solid-sample methods, because they often require some sort of extraction and more resources. Progress in methods of nucleic acid extraction that can be performed at the POC has been reviewed elsewhere [11] and thus, we focus on extraction-free sample preparation. Hereafter, we review the progress that has been made to develop extraction-free sample preparation modalities for iNAAT assays, comment on priority areas for further investigation and highlight successful examples of extraction-free sample preparation. Where experimental data in the context of iNAAT assay development is not readily available, we discuss data gathered in other contexts, such as PCR assay development, that can be informative for iNAAT test developers.

Frequently, iNAAT-based assays suffer a loss in sensitivity and reliability when combined with extraction-free sample preparation methods [12,13]. For amplification to proceed, there must be target available, in sufficient concentration, without interference from conditions or substances that can inhibit amplification or detection of amplicons. As shown in Figure 2, extraction-free sample preparation methods must simultaneously achieve two goals: 1) maximize available target and 2) minimize inhibition and/or interference from contaminants and sample residuals. An amplification inhibitor can act through a variety of mechanisms. Inhibitors can degrade or sequester the nucleic acid target, inhibit the polymerase used for amplification, chelate co-factors needed for polymerase activity, or interfere with assay readout [14–16]. Thus, the deleterious effects of potential inhibitors are partially specific to the amplification strategy.

Figure 2.

Top row: Factors that must be balanced when optimizing an extraction free sample preparation workflow. Factors including lytic additives, cellular debris and non-cellular components can negatively impact assay performance, leading to higher limits of detection. These concerns must be considered simultaneously with factors that impact the amount of target available for amplification, such as efficacy of lysis, preservation of target, or volume of sample. Bottom row: Prioritizing only the amount of target available (left) can lead to poor assay sensitivity, while prioritizing only assay performance (middle) can lead to insufficient target. Thus, in an extraction free paradigm a balance must be struck to ensure sufficient target and assay sensitivity (right). Created with BioRender.com

In this review, we begin by discussing the sample preparation considerations dictated by the sample matrix and storage buffer. Then, we explore various mechanisms of lysis, their known efficacy, and their potential impact on downstream amplification of nucleic acids. Next, we discuss extraction-free sample preparation considerations particular to RNA targets. We provide a brief commentary on the regulatory framework that assay designers should consider when incorporating extraction-free sample preparation methods in iNAAT assay workflows. We conclude by detailing some successful examples of extraction-free sample preparation for iNAAT assays.

While our analysis of the impact of potential inhibitors largely focuses on data of tolerance in RPA and LAMP, there are a number of other iNAAT technologies that have been described including Nucleic Acid Sequence-based amplification (NASBA) [17], Strand Displacement Amplification (SDA) [18], Rolling Circle Amplification (RCA) [19], Helicase-dependent Amplification (HDA) [20], and Hybridization Chain Reaction (HCR) [21]. While all of these have amplification strategies have their advantages, all are less represented in literature than RPA or LAMP.

Some efforts have been made to integrate POC sample preparation with each of these iNAAT technologies, mostly through integration of nucleic acid extraction on custom microfluidic chips. NASBA is uniquely suitable to RNA targets, but is expensive relative to other iNAAT assays and susceptible to false positives [5]. Dimov et al. and Reinholt et al. successfully integrated NASBA with sample preparation by combining RNA purification from cell lysate and NASBA amplification into a microfluidic chip [22,23]. RCA is particularly useful for generating long fragments for whole genome amplification [5]. POC tests using RCA have relied on purified sample and/or capture bead technologies for sample preparation. Schopf et al. used magnetic beads for viral DNA capture and RCA amplification [24]; later, Shopf et al used activated sepharose beads to capture bacterial DNA for RCA [25]. Recently, Phillips et al. designed ssDNA probes to capture RNA targets, both synthetic and in a clinical sample matrix, and subsequently amplify them with RCA and a lateral flow readout [26]. Characterizations of POC tests using HCR are proof-of-concept relying mostly on purified nucleic acids [5,27]. Xu et al. developed HCR device for amplifying detecting circulating tumor DNA from whole blood [28].

There are examples of SDA and HDA utilizing extraction-free sample preparation. HDA is based on the mechanism by which human DNA is replicated in vivo and has been described with integrated sample preparation [5,29]. The AmpliVue Group A Streptococcus (GAS) HDA assay is a sample-to-answer platform in which GAS is successfully detected from wound swabs prepared using a simple method of adding the swab to a lysis buffer followed by a dilution buffer, and then amplified using HAD and detected on a lateral flow readout, as evaluated by Faron et al [30]. SDA uses two enzymes to amplify heat denatured double-stranded DNA and is highly efficient, but has a long time to detection making it infeasible for many POC settings [5]. Notably, Toley et al. developed a SDA reaction capable of detecting purified nucleic acids in less than 30 minutes, and Shah et al. designed a mobile phone-based strategy capable of real-time SDA amplification in paper, both of which make this technology more potentially accessible for use at the POC [31,32]. The latter was later incorporated with on-chip bacterial lysis and both SDA and LAMP amplification in a fully sample-to-answer device [32]. Zhang et al. also described SDA using simplified sample preparation protocols such as the two-step chemical lysis and centrifugation method [33]. While these examples demonstrate that extraction-free sample preparation with these technologies is feasible, there are gaps in the literature. Data regarding the impact of various inhibitors on these amplification strategies is needed to expedite development of extraction-free sample preparation strategies for additional indications.

1. Sample Collection

Targets of interest for POC iNAATs can be contained within biological structures including protozoa, viruses, bacteria, and human cells. For infectious targets, the amount of target in a sample is dependent in part on the type of sample collected. In contrast, for a human genomic target, the target is available in all tissue types, enabling greater choice in sample matrix selection. Further, at the POC sample collection must be relatively non-invasive. In an extraction-free paradigm, assays should utilize sample matrices with highly concentrated target and a minimum of potential inhibitors, while prioritizing patient comfort.

Without concentration steps, the chosen sample type determines the maximum amount of target that can be made available for amplification. For instance, in whole blood the concentration of genomic DNA ranges from 10–15 ug/mL [34]. Similarly, in saliva, Quinque et al found the concentration of genomic DNA averages 11.4 ug/mL [34]. This means that in an extraction free paradigm, it is not feasible to obtain higher concentrations of genomic DNA. In contrast, Quinque et al found an average of 8.29 ug genomic DNA could be obtained from cheek swabs [34]. If aliquoted into a volume less than 1 mL, more concentrated DNA can be obtained from a cheek swab. Thus, for targets that can be detected in multiple sample types, choosing a matrix that can yield a higher target concentration can be beneficial.

Regarding sample matrix inhibition of amplification, the majority of literature to date concerns PCR, but increasingly data are being acquired for isothermal amplification technologies. Due in part to their different polymerases, RPA and LAMP have been shown to be more tolerant of a variety of sample matrix components than PCR [35,36]. Polymerases used in different amplification technologies have distinct inhibitory profiles, necessitating data regarding potential inhibitors to be gathered on each polymerase, and data regarding inhibition of PCR cannot necessarily be extrapolated to isothermal amplification reactions like RPA and LAMP. In contrast, PCR inhibitors that interact with the polymerase co-factor (e.g., Mg²⁺) or target are likely to exhibit similar inhibitory effects across various amplification strategies. In these cases, higher concentrations of co-factor can be added and maximization of available target should be balanced with the sample’s associated inhibitors. An additional strategy is to use higher volume reactions which while increasing reagent costs, can ensure sufficient input copies while using sample concentrations that are not inhibitory. For instance, Wang et al. recently demonstrated their LAMP-based HIV assay could detect the same input copy number (24) in 20uL and 500uL reaction volumes, but with greater tolerance for plasma at the higher reaction volume [37].

1.1. Whole Blood

Whole blood has many components that can inhibit amplification assays. Specifically, hemoglobin, lactoferrin, and immunoglobin have all been shown to inhibit Taq polymerase in PCR [38,39], but more research is needed to elucidate the impact of individual blood components on various iNAAT amplification technologies. In one example, Kersting et al. found that hemoglobin, even at a very high concentration of 50 g/L, had a minimal (10-fold) impact on the LOD of an RPA and lateral flow assay [35]. Deoxy-hemoglobin has a strong absorbance peak at 556 nm [40] and overlaps with the emission peaks for common fluorophores including FAM (518 nm) and SYBR Green (522 nm) [41]. Thus, blood and therefore, hemoglobin, has been used at lower concentrations in iNAAT assays with fluorescent optical detection systems [42] than Kersting et al utilized on a lateral flow strip. This highlights the importance of considering sample preparation when designing an assay readout. If a target is most suitable for detection in blood, readout strategies not impacted by heme should be considered.

Although the effect of the individual components of blood on iNAAT assays have not been thoroughly analyzed, iNAAT assays have been shown to be more tolerant of whole blood than PCR. For instance, Ortiz et al. achieved successful RPA amplification and gel readout with samples containing up to 33% heat-treated whole blood in the reaction volume [43]. In contrast, Kersting et al. found that using unprocessed blood at 20% reaction volume inhibited their RPA and lateral flow assay [35]. As for LAMP, Suzuki et al. found that at 20% reaction volume, heat-treated whole blood did not inhibit amplification with turbidity readout [44]. This suggests that heat treatment of blood can make it less inhibitory to iNAATs. Overall, iNAAT assays can tolerate greater volumes of whole blood than PCR, and tolerance is increased with pretreatment by heat.

For blood collection, the most POC-friendly collection method is finger prick samples because it is easy to implement and is relatively low-cost. However, because finger pricks yield volumes of blood generally less than 100uL from adults, [45], the small volume, and hence small amount of target, may require particularly sensitive iNAAT assays. Many studies have shown successful sample-to-amplification integration with small volumes of whole blood and extremely sensitive iNAAT assays [46]. Larger volumes can be obtained with venous blood draws, but the process requires trained personnel and more expensive supplies, and is thus less optimal for POC testing.

1.2. Saliva

Saliva is another sample matrix suitable for extraction-free sample preparation for iNAAT amplification. The quality of saliva samples can be affected by external sources such as oral hygiene products, food and drinks, and other environmental contaminants [46], which should be considered in sample preparation protocols. Saliva poses an additional challenge: it has higher viscosity relative to blood or urine. Viscous samples can affect fluid properties and change the allowed movement, proximity, and interactions between reagents. This is particularly important in RPA, which uses polyethylene glycol to increase the reaction viscosity and physically bring together reagents for amplification to occur [47]. To account for this, certain additives can be used to maintain a desired viscosity. For example, reducing agents such as TCEP and EDTA or semi-alkaline proteinase [48] can lower the viscosity of saliva and mucus samples [16]. Although saliva poses challenges for robust extraction-free sample preparation protocols, it has the significant advantage of being simple to collect and suitable for the POC.

The method of saliva collection can significantly impact the amount of target obtained. As stated earlier, Quinque et al. found that an average 11.4 ug/mL genomic DNA could be obtained from whole saliva [34]. In contrast, Rogers et al found that oral rinse yielded an average of 54.74 ug of DNA per 10 mL of mouthwash, or about 5.5 ug/mL, substantially less DNA than obtained with saliva [49]. Presently, there are several collection devices that can be used effectively for the collection of saliva. In the context of SARS-CoV-2 testing, one study compared the experience of participants self-collecting saliva with four different devices: a P1000 pipette tip, a Salimetrics Saliva Collection Aid (Salimetrics LLC, Pennsylvania, USA), a funnel, and a bulb pipette [50]. They found that saliva that naturally pools in the mouth could be effectively collected in sufficient quantities with all four devices. In this study, the sample collection devices analyzed were selected for their presumed interchangeability, which would be practical in the case of supply chain disruptions. However, the authors cautioned that results should not be extrapolated to other saliva collection devices. In general, for iNAAT assays to be integrated with saliva samples, if a high DNA yield is desired, whole saliva, regardless of the device used to obtain it, is preferable to oral rinse.

1.3. Swabs

Other upper respiratory samples are also suitable to collection at the POC. Ideal sampling of upper respiratory tract infections may be disease-specific, but studies that arose from the COVID-19 pandemic provide interesting examples of comparative sampling methods. A review article compared the diagnostic performance obtained from the gold standard nasopharyngeal swab with nasal swabs, throat swabs, saliva samples [51]. They concluded that all methods except throat swabs were clinically acceptable. Test sensitivity was much lower using throat swabs (68%), which was deemed unacceptable. No significant differences were found in test performance for self-collected and provider-collected nasal swabs. Suitable upper respiratory sample location is disease specific, and when multiple sample options yield sufficient target, the easiest to collect and lyse at the POC should be prioritized.

Swabs made of different materials can have a significant impact on the amount of sample obtained. For example, a study for human papillomavirus screening compared the number of cells obtained during vaginal self-sample collection with two devices: cotton swabs and nylon flocked swabs [52]. The authors found that the mean number of cells per milliliter obtained with flocked swabs was 4.6 times greater than with cotton swabs. Thus, extraction-free iNAAT assay test developers should select collection mechanisms with sufficient amounts of target but minimal amounts of inhibitors when possible.

1.4. Urine

Urine has the advantage of being a sample type that is relatively easy to collect; however, urine contains a well-known inhibitor of nucleic acid amplification: urea. Urea inhibits amplification by destabilizing primers, templates, and polymerase interactions [46]. However, LAMP has been shown to be more tolerant to the presence of urea than PCR. Edwards et al. spiked a range of urea concentrations in PCR and LAMP reactions for the same target [53]. The LAMP reaction tolerated up to 1.8 M urea, whereas the PCR reaction was inhibited by concentrations as low as 100 mM. While tolerance to urea is promising for iNAAT assay integration, there are limited examples attempting extraction free sample to answer integration from urine with iNAAT assays.

Importantly, the concentration and quality of DNA in urine can be highly variable. El Bali et al. found a large concentration range of 2–439ng/mL of total DNA extracted from urine samples [54]. They also found varying total DNA yields and amounts of fragmented DNA from urine collected at different times throughout the day from the same patients. Therefore, if the target is genomic, other sample matrices would likely be preferable. On the other hand, if the target is pathogenic, collecting urine may be more comfortable than alternative methods, but the amplification assay used must be highly sensitive in order to detect DNA from any patient.

1.5. Sample Storage Buffers

Clinical samples are routinely collected into a specific transportation or preservation media compatible with standard laboratory testing. Because these media are meant to preserve samples, they contain chemicals, salts, or other additives that may impact the availability of nucleic acids for amplification, or inhibit the polymerase, co-factors, or readout of RPA or LAMP assays. While using samples without storage buffers immediately following collection may be the desired workflow for iNAAT assays at the POC, if a newly developed sample preparation method is incompatible with routinely used collection media, clinical uptake may be limited. Therefore, despite potential inhibitors and resultant added burden to optimization, designing sample preparation protocols compatible with existing sample storage buffers may increase translation.

Few studies directly assess the impact of sample collection media on assay efficiency, but those that have demonstrate media-dependent differences. Viral transport media can vary in composition depending on manufacturer, and composition is generally proprietary, but the standard operating procedure published the Centers for Disease Control for producing an alternative to commercial VTM during the Covid-19 pandemic identifies the ingredients Hanks Balanced Salt Solution (HBSS), phenol red for a pH indicator, heat-inactivated fetal bovine serum, and Gentamicin sulfate (antibiotic) and Amphotericin B (antifungal) [55]. HBSS contains calcium ions which are known to compete with the co-factors of amplification reactions, and several salts which are known to sequester nucleic acids [14]. Anahtar et al., collected nasopharyngeal samples in either viral VTM or saline solution and added the samples directly to a colorimetric RT-LAMP reaction for COVID-19 [13]. As LAMP amplification occurs, protons are produced, changing the pH of the reaction. Colorimetric LAMP reactions rely on a pH-dependent dye that turns the reaction from red to yellow as amplification occurs [16]. The authors reported complete inhibition of colorimetric readout with 3 μL (12% total reaction volume) of the VTM sample in the reaction, whereas the saline sample had little effect. Because VTM contains buffer and phenol red, in this case, it likely specifically inhibited the readout method and not amplification. This conclusion is further bolstered by the work of Ganguli et al, who also utilized VTM for sample elution, but with fluorescent readout [56]. They found that the assay LOD was unaffected with up to 50% VTM in total reaction volume, but that time to amplification was slower (>40 min) than with 12.5% VTM in total reaction volume (20–40 min) [56]. This example highlights the importance of decoupling the effects of sample collection media on amplification versus readout method.

Many samples collected for cytology are eluted into PreservCyt, an alcohol-based fixative [57]. While the formulation of PreservCyt is proprietary, the safety data sheet provided by the manufacturer indicates PreservCyt contains 35–55% methanol [58]. While methanol enables PreservCyt to preserve cells over time, this same preservation function makes it more difficult to lyse the cells [57]. Additionally, methanol is known to denature DNA and can denature enzymes, which can inhibit amplification. Studies relying on samples stored in PreservCyt have overcome this by utilizing a buffer exchange process, but this this process does increase the sample preparation steps necessary [12]. Other cytology collection buffers, such as SurePath used for cervical samples, contain formalin, which produces cross-linking between nucleic acids and proteins, making nucleic acids less available for amplification [57]. Although cross-linking can be reversed with enzymatic digestion or heat, these additional steps may not be POC-appropriate.

For routine blood testing, samples are often collected into tubes containing different types of anticoagulants like EDTA, citrate, and heparin, which are necessary for preservation and for specific biochemical assays [59] but can impact lysis and amplification. For instance, Kersting et al. showed that 0.5 U of heparin did not affect an RPA reaction with lateral flow readout [35]. This represents a marked improvement over PCR in which as little as 0.0016–0.1U heparin has been shown to be inhibitory [60]. For comparison, BD Vacutainers with heparin contain 17U heparin per mL of blood [61]. An assay that utilizes 10uL heparin-containing blood per reaction would contain 0.17U heparin, above the limited tolerated by PCR, but within that tolerated by RPA.

2. Sample Lysis

Most nucleic acid targets are contained within cells or viral capsids, which must be lysed prior to amplification. Lysis strategies rely on four main mechanisms of action, as summarized in Table I: chemical, enzymatic, mechanical, and thermal. Most lysis strategies for extraction-free iNAAT POC assays combine multiple mechanisms.

Table I:

Overview of Lytic Mechanisms

	Mechanism of Action	Advantages	Limitations	Appropriate use cases
Chemical	Disrupts hydrophobic/hydrophilic interactions	Straightforward to consistently formulate (highly reproducible); No infrastructure requirements	Dilutes target; can inhibit amplification	Alkaline agents – useful in samples significant amounts of nucleases (blood, saliva, etc.) Reducing agents – can lower viscosity (saliva); valuable in protecting RNA from degradation
Enzymatic	Breaks specific peptide bonds	Can be heat inactivated, essentially “removing” from downstream amplification reactions	Requires additional heat inactivation step	ACP – Gram-positive bacteria Proteinase K – suitable to a variety of sample types
Mechanical	Shear force disrupts cell membrane structure	No additives diluting target or potentially inhibiting downstream amplification	Susceptible to user inconsistencies; requires infrastructure	Not suitable to samples with RNA targets
Thermal	External force disrupts cell membrane structure	No additives diluting target or potentially inhibiting downstream amplification	Requires infrastructure	Inactivate samples containing pathogens; eliminate enzymatic inhibitors in blood or samples undergoing enzymatic lysis;

In POC workflows, chemical lytic agents have the advantage of producing highly reproducible lysis without equipment or infrastructure but the disadvantages of necessarily diluting the sample and possibly inhibiting amplification. There are three main types of chemical agents commonly used for lysis – alkaline, surfactant, and chaotropic, which disrupt intermolecular forces that maintain the integrity of cellular membranes. However, these chemical additives can also inhibit the assay by impacting the polymerase, co-factors, or assay readout. Many studies only assess the overall impact of a chemical agent on an iNAAT assay without differentiating the efficacy of lysis from assay inhibition. Future studies should assess these effects separately. Such a body of research will allow test developers to rationally choose a chemical lytic agent that balances obtaining sufficient target against minimizing assay inhibition as discussed in Figure 2.

Alkaline agents can be very effective at lysing cells, dissolving DNA, and denaturing nucleases by increasing the sample pH [62], but an increased pH can also inhibit downstream amplification [63]. The most frequently used alkaline agent for sample preparation in iNAAT assays is sodium hydroxide (NaOH). NaOH has been used for lysis of blood samples between concentrations of 33.3 mM and 200 mM [15,42], and for the preparation of upper respiratory samples (nasopharyngeal swabs, saliva) at concentrations of 11 or 57.5 mM [13,16,64]. Within these studies, NaOH has been added directly to LAMP (0.44–11. 5mM) RPA (0.4 mM) [13,15,16,42,64]), and a single nucleotide polymorphism detection system that the authors called Smart Amplification Process version 2 (SMAP-2) (1.33 mM) [65]. Soejima et al. and Natoli et al., utilized NaOH exclusively for sample preparation, making their work useful in understanding the impact of NaOH on iNAAT amplification [15,42]. Soejima et al. developed a LAMP assay with a turbidity readout and determined that concentrations of NaOH at or above 8 mM interfered with the turbidity readout [15]. They also compared NaOH alone to NaOH with a 5 min 95°C heat step and unexpectedly found decreased reaction efficiency with the addition of a heat step [15]. Natoli et al. utilized a two-step lysis strategy to achieve significant lysis with a high concentration of NaOH (200 mM) [42]. Then, they diluted the lysate so that the final concentration of NaOH (0.4 mM) would be compatible with amplification [42]. Because this method dilutes the original blood sample 1:200 [42], it is only suitable for very sensitive amplification reactions or targets present at high concentrations. Although this two-dilution step strategy can complicate the workflow, it is valuable in that it enables a highly alkaline environment for lysis that would not otherwise be compatible with downstream amplification, making it promising for some applications. Therefore, adapting the post-lysis solution to the requirements imposed by amplification assay should be considered when harsh alkaline agents are used.

Surfactants disrupt the hydrophobic-hydrophilic interactions that maintain membrane integrity, lysing cells [66], but this same function causes adverse effects in amplification reactions. For example, surfactants can sequester or degrade nucleic acids, degrade polymerases, [14] and interfere with the readout of fluorescent assays [16]. Triton X-100, Tween20 have been successfully incorporated at very low concentrations (<0.1% v/v) into fluorescence RT-LAMP assays [59,67,68]. SDBS (sodium dodecylbenzenesulfonate) has been incorporated in LAMP with turbidity readout at 0.05% of reaction volume [69]. Rabe et al. analyzed the effects of higher concentrations of these surfactants on fluorescence RT-LAMP. At 1% v/v Triton X-100 and 1.5% v/v Tween20, the fluorescence spectrum was significantly dampened, but time to amplification was still discernible and unchanged. At higher concentrations, fluorescence intensity was so reduced that the time to amplification could not be reliably discerned [16]. In this example, it is unclear whether higher concentrations of detergents impacted the amplification itself or only the readout. This example highlights the importance of detangling the effects of detergent on amplification versus readout. If readout is primarily impacted, then readout methods other than fluorescence are often possible.

Chaotropic agents disrupt hydrogen bond networks, enabling them to denature proteins disrupting membrane bilayers. Examples include guanidine hydrochloride, guanidinium thiocyanate, and sodium iodine. Rabe et al. demonstrated their colorimetric LAMP assay can tolerate up to 50 mM of either guanidinium thiocynate or sodium iodine [16]. Zhang et al. demonstrated that at 40–50 mM guanidine hydrochloride (GuHCl), colorimetric and fluorescence RT-LAMP reaction efficiency is increased, but at higher concentrations, GuHCl begins to inhibit amplification [70]. While the reason for this increased efficiency is unknown, the authors hypothesize that GuHCl improves primer-target interactions [70]. This finding highlights the potential of considering sample preparation and amplification not as separate processes but as potentially complementary aspects of iNAAT assay development.

Enzymatic agents function by cleaving specific peptide bonds, thereby degrading the cell wall and/or membrane. After lysis, enzymes can be inactivated using heat, preventing the lysis enzyme from impacting downstream amplification [71]. Two of the most utilized enzymatic lysis agents for iNAAT assays are achromopeptidase (ACP) and proteinase K (Pro K). ACP, in particular, has shown efficacy at lysing Gram-positive bacteria [71]. Enzymatic agents have been used with a wide variety of target types, including the SARS-CoV-2 virus, Gram-positive bacteria, and human cervical cells [12,59,67,72]. Heiniger et al. systematically compared lysis methods for the Gram-positive bacterium S. aureus and found 90% lysis using ACP, far greater than the 12% lysis achieved with heat alone [71]. In contrast, ACP exhibited little lysis above that achieved with heat alone for the Gram-negative bacterium B. pertussis [71]. Gram-positive bacteria have significantly greater layers of peptidoglycan than Gram-negative bacteria, making the former generally more difficult to lyse [62]. These results indicate that enzymatic agents provide little benefit for some targets over heat treatment alone. Most other studies using enzymatic lysis have not compared the degree of lysis due to the enzyme to that due to the associated heat treatment. In the absence of these data, it is possible the lysis observed is due to heat step and the enzyme is superfluous.

Mechanical mechanisms use shear force to disrupt cell membrane structures [66] and have the advantages of not requiring any dilution of the sample target nor impacting downstream amplification. Unfortunately, mechanical mechanisms are limited by susceptibility to user inconsistencies and infrastructure requirements. Vortexing is one mechanical step sometimes used in POC extraction-free studies [64,73]. While typically framed as a mixing step, vortexing can cause sample lysis. The degree of lysis produced by vortexing is typically not quantified when developing sample preparation protocols for assays, but it is recommended whenever possible to minimize reliance on inconsistent methods like vortexing or to standardize and characterize the vortex step.

Thermal methods use heat to disrupt cell membrane structure [66] and, like mechanical mechanisms, require no fluid addition to samples. While thermal methods necessitate heating infrastructure, the development of inexpensive, battery operated heaters [12] [74] enables heat lysis at the POC. Many extraction-free sample preparation workflows for iNAAT assays incorporate a heat step, most often in conjunction with chemical agents [13,16,64,65], or to inactivate an enzymatic lytic agent [12,32,67,73,75,76]. Some studies employ heat alone, including Heiniger et al. who found a 3-minute boiling step sufficient to achieve 79% lysis for the Gram-negative bacterium B. pertussis [71], and Modak et al. used heat at 90°C for 5 minutes to release target from malaria parasites in whole blood [77].

Thermal methods are also often used to inactivate samples containing pathogens. Many pathogens like viruses must be inactivated for safety considerations before further handling [66]. When heating with the primary aim of pathogen inactivation, time required for complete inactivation depends on the incubation temperature. For example, SARS-CoV-2 can be inactivated in less than 30 minutes, 15 minutes and 3 minutes at 56°C, 65°C and 95°C, respectively [78]. For POC sample preparation, shorter times are preferable, but the need for higher temperatures can increase infrastructure complexity. For safety, it is essential that inactivation be complete, and inactivation requirements can differ between species, even viruses within the same family. For example, amongst members of the the Togaviridae family, Ross River virus, o’nyong-nyong virus and Barmah Forest virus required greater than 120 minutes, 60 minutes and 20 minutes for inactivation when incubated at 56°C [79]. The inactivation step usually contributes to lysis and should be incorporated into extraction-free iNAAT sample preparation protocols.

3. RNA Considerations

There are increased requirements for extraction-free sample preparation for iNAAT assays targeting RNA. First, RNA is especially prone to degradation due to endogenous ribonucleases (Rnases) – molecules endogenous to all organisms that catalyze hydrolysis of RNA molecules. Therefore, sample preparation for RNA requires strategies to protect target from degradation [80]. RNase activity can be decreased using chelating additives such as 0.1–1 mM EDTA, reducing agents such as 5 mM TCEP or 1–50 mM diothiothreitol (DTT), or through enzymatic RNAse inhibitors [81–84]. 5–10 mg/mL polyvinyl sulfonic acid has also been shown to inhibit RNA cleavage in the presence of RNases, suggesting its potential use as another broad RNase inhibitor [85]. Second, the RNA molecule is less stable than DNA. RNA is structurally easier for hydrolyzing enzymes to attack than DNA due to its single helix structure and larger grooves. Additionally, RNA also has a reactive 2’ hydroxyl group that mediates chemical hydrolysis [86]. Due to its instability, RNA is also prone to degradation under conditions that DNA typically can withstand; for example, the integrity of RNA is disrupted at temperatures above 65°C and by vortexing [87]. Thus, extraction free sample preparation for RNA targets must devise means of inactivating RNases and avoid conditions that degrade RNA.

Some DNA viruses integrate into the host genome. Where clinically significant, iNAAT assays can be designed to target either viral DNA or host-produced mRNA, which require different sample preparation strategies. To specifically detect mRNA, viral DNA must be removed from the sample because it often has the same sequence as the mRNA molecule it encodes. As yet this process has not been done in an extraction-free manner, but has relied on hybridizing mRNA with a complementary oligonucleotide on a magnetic microparticle, and isolating mRNA with use of a magnet [88]. For other RNA targets, such as RNA viruses like SARS-CoV-2 and HIV, this distinction is not an issue as RNA is the only target present, or there is a much less clinically significant difference between detectable DNA and RNA [89,90]. For these targets, removing cellular DNA is unnecessary, simplifying the sample preparation process. While most work developing methods of extraction-free sample preparation have focused on DNA targets, developing extraction-free methods for preparation of RNA, and particularly mRNA, would significantly improve the range of clinical use cases of iNAAT testing that can be performed at the POC.

4. Regulatory Landscape

While there is increasing literature regarding extraction-free sample preparation for iNAAT assays, the ultimate success will be when these assays achieve regulatory approval and widespread use. In the United States, in vitro diagnostics (IVDs) require approval for use by the FDA [91]. The FDA 510(k) submissions require demonstrations of “substantial equivalence” – that a test matches or exceeds safety and efficacy of existing, approved tests [91]. Substantial equivalence concerns the analytic performance – sensitivity, specificity, accuracy – not usability or setting for which the diagnostic is suitable. As discussed earlier, extraction-free sample preparation, even with significant optimization, can negatively impact assay sensitivity [12,13]. While usability is not specifically mentioned, ease of use can affect the safety and efficacy of the device, which will determine if the IVD is approved for a specific intended use (e.g. home use vs. health care professional use) [92]. Additionally, the FDA determines the complexity of a test under the Clinical Laboratory Improvement Amendments, which will determine the level of laboratory in which tests can be performed [93]. This determination depends on the complexity of every step of the test process, including pre-analytic concerns such as sample preparation [94]. Tests may be approved for high or moderate-complexity laboratories, CLIA-waived laboratories (subject to minimal regulation), or for home use.

Extraction-free sample preparation can significantly improve the setting for which a diagnostic test is suitable, but this value can be overlooked in the traditional premarket approval process. For instance, POC NAAT tests with lower analytical sensitivity than their laboratory-based counterparts may be still be beneficial if they can increase population coverage and testing frequency [95,96]. These valuations are not accounted for in an evaluation of substantial equivalence, leading to challenges in the successful market introduction of POC-friendly iNAATs with extraction-free sample preparation that could improve health.

To date, few iNAATs have yet been approved for low-complexity locations. As of November 6, 2023, the FDA listed 276 EUA-approved SARS-CoV-2 molecular diagnostics - only six of these tests (four manufacturers) were approved for home use settings [97]. Twenty-five of the 276 tests were approved for CLIA-waived settings. Similarly, for SARS-CoV-2 molecular diagnostic tests approved through traditional premarket review pathways, one has been approved for Home Use, four for CLIA-waived settings, and the remaining nine only for high or medium-complexity laboratories [98]. The paucity of Home/Waived NAAT tests applies to not only SARS-CoV-2, but other test indications as well. For instance, the WHO list of essential in vitro diagnostics lists only two POC NAATs (one for Influenza and one for HIV) considered suitable for community health settings [99]. This reflects the limitations of effective sample preparation strategies suitable to the POC. The WHO list of essential in vitro diagnostics includes only tests that utilize unprocessed specimens like blood, urine, and swabs [99]. While pre-analytic concerns like sample preparation may not prevent a test from receiving regulatory approval, the lack of sample preparation strategies suitable for POC and near POC settings is a significant impediment to the availability of NAATs. Evaluating POC iNAATs with extraction-free sample preparation procedures against laboratory-based NAATs utilizing extracted nucleic acid samples has contributed to the under realization of the significant potential of POC iNAATs.

Fortunately, increasing attention is being paid to the potential value of POC iNAATs, and strategies are being developed to evaluate them appropriately. For instance, the POC Key Evidence Tool (POCKET) incorporates not only test performance but also utility, usability, cost-effectiveness, and patient experience into the framework for test evaluation [100]. To develop frameworks for evaluation of POC NAATs is significant, but the underlying evidence needed to establish benchmarks like utility and cost-effectiveness is necessary, hard to obtain, and generally lacking to utilize such frameworks in the evaluation of a particular POC [101].

Investments with the aim of developing POC NAATs and gathering the data necessary to evaluate them are increasing. In 2007, the National Institute of Biomedical Imaging and Bioengineering (NIBIB) created the Point-of-Care Technology Research Network (POCTRN) to support and accelerate the development of point-of-care technologies by supporting academic Centers to focus on particular needs [102]. In 2022 POCTRN Rapid Acceleration of Diagnostics (RADx) released guidelines for designing accessible COVID-19 tests for home use [103,104]. Many of these guidelines can be applicable not only to COVID-19 but also to extraction-free sample preparation for many other NAAT indications. For instance, to be user-friendly, fluid vials should have attached caps, internal ribbing to help agitate the sample, and be see through and easy to squeeze [104]. Further, fluids should not be transferred by pouring, but utilize prefilled vials or dropper caps [104]. POC iNAATs and associated use of extraction-free sample preparation constitute a significant departure from more established in vitro diagnostic technologies and the standards used to evaluate them.

Despite the associated challenges, there is significant value in POC iNAATs that can reach patients and settings without access to existing diagnostic modalities. Thus, evaluating POC iNAATs should take this potential into account and requires a wholistic analysis of test utility beyond that provided by traditional premarket pathways, and significant support to generate the evidence to demonstrate utility.

5. Promising Examples

Table II summarizes the results from recent studies that have made significant progress in developing and validating extraction-free fluid-based sample-to-answer iNAAT assays for different targets. Ways of calculating steps and amounts of additives differ, and so in preparing this table these have been normalized for comparison and easy reference. The number of processing steps was calculated as the number of distinct processes that require user input following sample collection through until addition of sample to the amplification reagents. Sample preparation time is reported as the sum of all timed steps in the paper, without including user-dependent steps like pipetting from container to container. Using the information provided in each paper, the concentration of each chemical additive in the lysate, regardless of amount of sample or lysate volume was calculated. Then, the concentration of each chemical additive that is in the iNAAT reaction, given the amount of sample added to the amplification reagents, was determined.

Table II:

Examples of Successful Extraction-free Fluid-based Sample-to-Amplification Integration for iNAAT Assays

Paper	Target Type	Sample matrix	Sample storage buffer	Amplification method	Number of sample processing steps	Sample prep time (min)	Chemical steps		Enzymatic steps	Mechanical steps	Thermal steps		Assay Performance with extraction-free sample preparation
							Concentration in lysate	Concentration in amplification			Temperature (°C)	Time (min)
Kundrod et al. 2022 [64]	RNA virus	Saliva Nasopharyngeal swab Nasal swab	X	RT-LAMP fluorescence	2	5	NaOH 57.5mM GuHCl 200mM TCEP 12.5mM EDTA 5mM	NaOH 11.5mM GuHCl 40mM TCEP 2.5mM EDTA 1mM	X	20 sec vortex	95	5	Positive agreement 91% with nasopharyngeal swabs, 69% for nasal swabs and 69% with saliva samples
Anahtar et al. 2020 [13]	RNA virus	Nasopharyngeal swabs	VTM or saline	RT-LAMP colorimetric	2	5	NaOH 11mM TCEP 2.5mM EDTA 1mM	NaOH 0.44mM or 2.2mM TCEP 0.1mM or 0.5mM EDTA 0.04mM or 0.2mM	X	X	95	5	87.5% sensitivity (extraction-based preparation improved LOD 10 fold per amount initial sample)
Jones et al. 2021 [67]	RNA virus	Nasopharyngeal swabs	X	RT-LAMP fluorescence	2	25	Triton X-100 1%	Triton X-100 0.08%	Pro K 0.04U/uL	X	95	10	98% positive agreement; 87% negative agreement
Hayashida et al. 2015 [68]	Protozoa	Whole blood	heparin; EDTA	LAMP fluorescence	2	nominal	Triton X-100 0.09%	Triton X-100 0.096%	X	X	X		LOD unchanged in presence of blood with spiked parasite
Natoli et al. 2021 [42]	Genomic	Whole blood	EDTA	RPA-exo	2	5	NaOH 200mM	NaOH 0.4mM	X	X	X		Amplification observed in 3/3 healthy volunteer samples
Mitani et al. 2007 [65]	Genomic	Whole blood	X	SMAP-2	2	3	NaOH 33.3mM	NaOH 1.33mM	X	X	98	3	63/63 clinical samples showed concordance with gold standard
Soejima et al. 2011 [15]	Genomic	Whole blood	X	LAMP turbidity	1	nominal	NaOH 49.5mM	NaOH 3.96mM	X	X	X		100% concordance with gold standard
Rabe et al. 2020 [16]	RNA virus	Nasopharyngeal swab	Saline or 1XPBS	RT-LAMP colorimetric	2	5	NaOH 11 mM TCEP-HCl 2.5mM EDTA 1mM	NaOH 2.2 mM TCEP-HCl 0.5mM EDTA 0.2mM	X	X	95	5	LOD of 50 spiked copies/uL sample in nasal or throat swabs (compared to LOD of 1 spiked copies/uL sample with extraction-based preparation)
Lafleur et al. 2016 [76]	Bacteria	Nasal swab	BD Liquid Amies kit buffer	iSDA lateral flow	No manual processing	7	X		ACP 0.5U/uL	X	95	5	7 of 11 samples agreed with gold standard
Kundrod et al. 2023 [12]	Integrated DNA virus	Cervical swab	1X TE Buffer; PreservCyt	RPA-nfo lateral flow	3	15	X		ACP 0.5U/uL	X	75/95	5	All provider-collected samples in 1X TE with target above assay LOD identified correctly; with self-collected PreservCyt samples, sensitivity 80% for HPV16
Ben-Assa et al. 2020 [75]	RNA virus	Nasal swabs Throat swabs Saliva	UTM	RT-LAMP colorimetric	3	20	X		Pro K 1.22mg/mL	X	95	5	True positive rate of 93% swab samples
Schmidt et al. 2021 [73]	RNA virus	Nasopharyngeal swab	X	RT-LAMP fluorescence	5	11	X		Pro K 2.5mg/mL	1 min vortex	95	5	100% sensitivity; 96.2% specificity (100% sensitivity 100% specificity with extraction-based preparation)
Modak et al. 2016 [77]	Protozoa	Whole blood	X	LAMP florescence	2	5	X		X	X	90	5	LOD 1.5 parasites/uL blood

RNA, ribonucleic acid; RT-LAMP, reverse transcriptase loop mediated isothermal amplification; NaOH, sodium hydroxide; GuHCl guanidine hydrochloride; TCEP, tris(2-carboxyethyl)phosphine; EDTA, ethylenediaminete traacetic acid; VTM, viral transport medium; Pro K, proteinase K; RPA-exo, recombinase polymerase amplification exonuclease; SMAP-2, smart amplification process version 2; PBS, phosphate buffered saline; iSDA, isothermal strand displacement amplification; ACP, achromopeptidase; RPA-nfo, recombinase polymerase amplification endonuclease; TE, tris-EDTA; UTM, universal transport medium; LOD, limit of detection.

This process enables direct comparison of studies that calculated additive concentrations at different points in their workflow and described the steps of their sample preparation process differently. These studies demonstrate that sample to iNAAT amplification integration can be achieved without extraction in ways suitable to the POC. They represent a variety of target types, sample matrices, and lytic strategies and their sample preparation methodologies can be used as a starting point for assay designers seeking to optimize extraction-free sample preparation with a variety of iNAAT amplification mechanisms.

In analyzing these studies together, several generalizations can be made about what has been successful in extraction-free sample preparation for iNAAT assays. First, most of the studies include some manual processing steps following sample collection, but can limit this to 1–3 steps [12,13,15,16,42,64,67,68,75,77]. The one reported assays with no manual processing steps both use lyophilized enzyme in a microfluidic cartridge for lysis [76]. Second, the time required for sample preparation is highly variable, but studies using chemical lytic methods have shorter sample processing times (0–5 minutes) [13,15,16,42,64,68] than those relying on enzymatic lysis (7–25 minutes) [12,67,73,75,76]. Third, most of the studies employing chemical lytic agents used a higher concentration of chemical for lysis than in the amplification reaction [13,15,16,42,64,65]. To employ these lytic strategies in cases where all the sample encompasses all the reaction volume (e.g. the reagents have been lyophilized) an additional dilution step may be required for sample preparation. Finally, note that the studies included amplifying RNA samples were confined to SARS-CoV-2 detection [13,16,64,67,73,75], reflecting the challenge of preserving most target RNA in an extraction-free paradigm.

While the methods utilized in these studies to evaluate assay performance are not consistent, we included in Table II the reported performance of the assays with the developed extraction-free sample preparation strategy and compared this to the performance of the assay with extraction-based sample preparation when reported. Given the differences in methods utilized to evaluate performance, it is not feasible to make exact comparisons between these studies, but some significant observations can be made. First, Kundrod et al. 2023 found a loss in sensitivity with self-collected samples that could be attributed to increased cellular debris relative to target concentration [12]. This finding highlights the increased need for assay designers working in an extraction-free paradigm to account for the heterogeneous nature of crude samples. Second, Anahtar et al. and Rabe et al. observed a higher assay LOD in target per original sample of one order of magnitude [13,16]. Anahtar et al. attributed this to the concentration step of the extraction process [13]. The impact on assay LOD can easily be estimated for sample preparation strategies that do not allow for target concentration prior to optimization, and we recommend investigators do so to benchmark if extraction-free sample preparation will impact clinical utility of an assay. Further, differences in the calculated and observed changes in LOD can help determine if assay inhibition is occurring. Third, conclusions regarding assay performance must be tied to clinical utility. For instance, Kundrod et al. 2022 despite obtaining a more desirable LOD using nasopharyngeal swabs than saliva (20 v. 93 virons/reaction), decided to move forward with saliva-based testing, as the assay was being designed for surveillance testing in which suitability for frequent testing was a higher priority than sensitivity [64]. Taken together, these results demonstrate that it is possible to maintain high overall assay performance with extraction-free sample preparation.

Conclusions

The challenges to sample preparation without extraction are immense and should be treated as an equally important part of iNAAT test development alongside amplification rather than as a last step. The investment in sample preparation for SARS-CoV-2 detection demonstrated the functionality of extraction-free workflows for rapid, at-scale testing. However, strategies that work for SARS-CoV-2 assays are not universal. Currently, there is a lack of data on appropriate sample preparation parameters for many promising iNAAT assay targets – research to address this gap is essential to ensure that iNAAT assays can be implemented at the POC.

There are a number of experimental gaps that should be prioritized. First, evaluation of the impact of potential inhibitors on iNAAT technologies is needed, especially with attention to studies that differentiate inhibition of amplification from interference with readout. Second, studies to differentiate the impact of chemical agents on lytic efficiency from inhibition of amplification are needed to understand what chemical agents are appropriate for what sample types. For studies employing both enzymatic lysis and heat inactivation, heat-only controls are valuable to avoid developing sample preparation protocols with superfluous enzyme addition. Freeze-thaw of samples should be avoided as freeze-thaw cycles can cause cell lysis and result in the over-estimation of efficacy of candidate sample preparation protocols. Caution should be employed when using spiked samples, as they do not mirror the complexities of actual samples.

In this article, we reviewed the literature relevant to developing extraction-free sample preparation methods for fluid-based iNAAT assays. Despite the myriad of challenges associated with designing extraction-free fluid-based sample preparation protocols for iNAAT assays, doing so can generate accessible, POC-friendly assays, minimizing time, infrastructure, and personnel requirements that fill essential gaps in diagnostic access.

Expert Opinion

Access to high-quality diagnostics is foundational to delivering healthcare and to public health systems. In part, due to the unequal distribution of laboratory infrastructure, currently an estimated 35–62% of the world’s population lacks access to even essential diagnostics [105]. Developing iNAATs in conjunction with extraction-free sample preparation constitutes a workflow that meets the REASSURED criteria and is thus suitable to the POC. This has significant potential to rapidly increase access to diagnostics through nucleic acid testing in the absence of laboratory infrastructure.

The lack of suitable sample preparation strategies and the data needed to develop them is one of the most significant barriers to the development of extraction-free sample preparation and implementation of iNAAT testing at the POC. Most currently available literature on iNAAT assays does not address sample preparation. When extraction-free sample preparation is attempted, studies typically first design an amplification assay, validate the assay with extracted samples, and then finally consider sample preparation. In considering sample preparation as a final step, assay performance is often diminished. Because of limitations in sample preparation, many otherwise promising iNAAT assays remain limited to use in a laboratory.

It is time for a paradigm shift in how iNAAT assay development for use at the POC is approached, in which development of amplification and of sample preparation methodologies should be treated as equally important and interdependent processes. Figure 3 outlines a framework of guiding questions advised for sample-to-answer iNAAT assay development in an extraction-free paradigm. Optimal sample preparation methods are target- and sample-specific, so first, researchers should begin by determining what sample matrix or matrices the eventual assay needs to incorporate. This depends on whether the target is genomic or pathogenic, the concentration of target in the matrix, and the clinical suitability of the matrix. Then, a lytic strategy should be developed. This may include a pathogen inactivation step and a choice between prioritizing speed of the sample preparation process or minimizing user steps may be necessary. For chemical lysis, concentration is an important factor, whereas for enzymatic lysis, the need and role of heat inactivation should be considered. Only after the sample preparation criteria have been established, should an amplification assay and readout be developed – selecting strategies likely to be compatible with the matrix and lytic strategy utilized. After the amplification assay has been developed and evaluated with the matrix and lytic strategy, optimization of the reaction can increase performance. This may include changing the concentration of reaction components and/or including additives to maintain viscosity, pH, or other parameters. This framework prioritizes co-designing sample preparation and assay development as interdependent processes.

Figure 3.

Framework of guiding questions for developing sample-to-answer POC iNAAT assays in an extraction-free paradigm. Sample preparation methods and amplification should be designed interdependently. First, the matrix selection should be considered followed by the lytic strategy development. Next, amplification and readout should be selected, and lastly, the integrated assay should be optimized, if needed. Created with BioRender.com

In the next five years, we anticipate increasing priority being placed on the role of sample preparation in assay development. Extraction-free sample preparation workflows are well suited to the POC, making investment in them worthwhile. Additionally, we expect increasing demand for rigorous sample preparation experiments by researchers, funders, and publishers. Most significantly, we anticipate that variations on existing iNAAT technologies like RPA and LAMP will be developed with the objective of increasing compatibility with desirable sample matrices, targets and their optimal lytic strategies.

Article Highlights.

• Point of care isothermal nucleic acid amplification tests (iNAATs) require effective sample preparation workflows that can also be implemented at the POC.

• Workflows that are extraction-free – excluding isolation, concentration, or purification of nucleic acids can minimize requirements of infrastructure, personnel, and time.

• Choosing a matrix with a high concentration of target, designing a compatible preparation strategy, and designing a readout strategy to maximize likelihood of compatibility enable assay performance.

• Following strategic design, increasing co-factor concentration, buffering pH, employing nucleic acid protectors, and adjusting reaction viscosity provide further opportunities for integrated assay optimization.

• Future work should treat sample preparation as an integral aspect of assay design equal in importance to amplification assay design.