A Preanalytic Blood Separation and Metering System for Qualitative and Quantitative Lateral Flow BiosensorsA Preanalytic Blood Separation and Metering System for Qualitative and Quantitative Lateral Flow Biosensors
This paper describes a simple, cost-effective, and practical blood separation and metering system that is well suited for use in conventional lateral flow strips.
June 6, 2013

The concentration quantification of an analyte (e.g., glucose, cholesterol, antigens, enzymes) in a physiological sample (e.g., saliva, urine, blood) plays a prominent role in the diagnosis and management of a variety of health conditions. Examples of such tests, including glucose monitoring, cholesterol monitoring, pregnancy testing, and many others, have become routine in both clinical and home settings.1–7
Table I. Development goals for a preanalytic blood sampling system for both qualitative and quantitative lateral flow tests.
Blood sample size | 1–5 μl of blood from a finger or an alternative site (e.g., ATS). 2–3 μl are realistic targets. |
Blood sampling | Spot sampling, no need for user to meter blood volume and then do blood transfer. |
RBC separation | Complete (or at least >95%) separation of red blood cells after “Spot” sampling at several microliter levels. |
Plasma metering and transfer | Plasma and filtered red blood cells will automatically be in different zones after separation of red blood cells. No extra step to remove separated red blood cells. |
Detection methods | Qualitative as well as quantitative lateral flow. |
Cost/manufacturing | Existing commercial available materials and avoiding sophisticated microfluidics. |
To determine if an analyte is present or at the right levels for a certain health condition, a physiological sample must first be obtained. Depending on the testing method, test procedure, and sample sizes, obtaining an appropriate sample in some circumstances can be cumbersome, complicated, painful, and may also involve multiple steps requiring sophisticated devices. For example, a blood sample is often taken through a multiple-step process using first a skin-piercing mechanism, such as a needle or lancet, followed by a collection mechanism, such as a capillary tube. After the collection, the blood sample must be transferred to a testing device (e.g., a test strip or the like). Often, the test strip is then transferred to a measuring device such as a meter. Accordingly, the steps of accessing the sample, collecting the sample, transferring the sample to a biosensor, and measuring the analyte concentration in the sample are generally performed as separate, consecutive steps with various devices and instrumentation.
In addition to the multiple steps mentioned, blood-based diagnostic tests must additionally perform two major preanalytical tasks before running the assay: the removal of red blood cells and the application of a known volume of plasma or serum to the detection device. These two extra steps must be performed because they compensate for hematocrit (i.e., fractional volume of red blood cells contained within the whole blood sample) variations within the normal population. These steps also remove potential impediments and color interferences associated with the red blood cells, which often prevent obtaining accurate and quantitative detection results. The metering of the sample into known testing volumes allows for precise measurements, particularly in the case of quantitative assays.
Separating red blood cells, particularly at sample sizes at submicroliter to several microliter levels, is challenging because it is difficult to handle and transfer small liquid volumes. When the detection does not require the separation of red blood cells, the diagnostic industry has effectively solved blood sample collection and subsequent detection by direct spot technology. For example, various glucose meters on the market use prefabricated sampling channels to spot blood samples as small as 0.3 μl.6 However, spot sampling to a traditional lateral flow strip’s sampling zone has not yet been realized for quantitative tests because it is hard to uniformly distribute blood onto the sample pad of a traditional lateral flow strip and complete the separation of red blood cells.
This paper describes a simple, cost-effective, and practical blood separation and metering system that is well suited for use in conventional lateral flow strips. The blood sample sizes for the described metering system are generally at 1–5 μl. The design and development of this metering device stemmed from existing lateral flow technology, and it can be produced in large scale by the same lateral flow manufacturing processes. This metering system provides an interesting option for overcoming technological challenges that currently limit the expansion of lateral flow technology into quantitative diagnostic applications with alternative site testing and red blood cell separation capabilities.1–2 Thus a traditional lateral flow test that can handle blood samples like various glucose monitoring strips is highly desirable. A lateral flow testing strip with spot blood sampling and red blood cell separation capabilities will significantly expand lateral flow technology’s reach from mainly qualitative testing to fully quantitative measurement of specific protein biomarkers in plasma that are more indicative of the state of health. The significant advantage to using plasma for molecular diagnostics is that many cancers and other serious diseases can be discovered early by detecting the presence of shed biomarkers in the plasma. Drawing blood and separating the plasma allows biomarkers to be directly accessed without the need for more costly and invasive surgical techniques (e.g., biopsies). Additionally, this will expand lateral flow technology into alternate site (AST) testing with small sample sizes because consumers perceive it as nearly painless as compared with traditional finger sticks. Although AST provides consumer benefits, it presents technical challenges for diagnostic tests because considerably smaller samples are presented with AST as compared with traditional methods (e.g., 1–3 μl versus 10–30 μl, respectively).
The Inception of the Concept
Using a channel-sized membrane or foam to replace the channel led us to look at the traditional lateral flow strip itself as a blood separation and metering device, as shown in Figure 1. We wanted to find out if it is possible to develop a miniaturized lateral flow strip (Figure 1A) with a blood separation membrane at one end to filter off red blood cells while the nitrocellulose membrane is used as a plasma or serum storage meter (Figure 1B). This concept can potentially provide all desired performance attributes as described in Table I if a fixed area of plasma or serum on the nitrocellulose membrane in Figure 1B can be isolated and the plasma stored in the isolated area can be efficiently transferred out for quantitation (Figure 1C).1,2 Initial technology scouting found that fine fluidic channels, as exemplified by TheraSense’s Freestyle glucose meters, are attractive for their abilities to rapidly take in small and fixed amounts of fresh blood (e.g., ~0.3 μl as the standard) with pain-free alternative site spot sampling.6 Attempts were made to look at the possibility of building similar types of channels in a lateral flow strip’s sampling area, but this proved impractical because a channel disrupts the continuity of the flow path of the diluent and makes it almost impossible to separate red blood cells. To compensate for the flow-path disruption by the channel, two options were proposed and tried. The first used a bridging mechanism to guide diluent to pass over the channel with the hope that diluent can bring the blood sample from the channel by capillary to the detection zone.3 The second used a channel-sized porous membrane or foam to replace the channel (e.g., a porous membrane or foam to hold blood).1,2 This paper only discusses the development of the second option.The blood sample meter described in this paper is simple and practical but has a clear set of tough development goals (Table I).
The concept to have a miniaturized lateral flow strip (e.g., a blood separation laminate (BSL)) is a practical spot sampling idea for several reasons. First, nitrocellulose membranes, as well as other types of similar membranes, have already been used extensively in the lateral flow industry as the test matrix for lateral flow immunoassays. Second, lateral flow test strips with nitrocellulose membranes are routinely produced with sufficient tolerances such that the material variations are insignificant within a lot or across the billet of material. Third, it is known within the diagnostic industry that a bibulous matrix possessing a predefined volume laminated to a size exclusion matrix can separate the plasma from a blood sample. When a whole blood sample is applied to this laminate, the size exclusion screen preferentially allows only the plasma component to penetrate the bibulous material. This method has been implemented in at least one commercial test (i.e., Cholestrak total cholesterol test), which claims to attain 97% accuracy (although this system is not amenable to our development goals as described in Table I, particularly in consideration of sample volume and the lateral flow platform).
Additionally, this concept has the potential to address the variations of blood sample sizes by postulating that a fixed length of the membrane with the stored plasma can be physically isolated for quantitation (Figure 1C). This isolation consideration is very important if spot, pain-free sampling is the chosen method for taking blood from a user’s finger or other alternative sites for performing quantitative measurements. Furthermore, the concept can use all existing lateral flow technology materials and manufacturing methods, which would allow manufacturers to use same manufacturing facilities to make both metering and detection strips without too much added cost.
Initial Prototyping of the Metering System
Miniaturization of a regular lateral flow strip to a blood sample metering device was first carried out by focusing only on the nitrocellulose membrane portion of the strip. The concern is that nitrocellulose membrane may or may not be the correct matrix as a plasma meter because we don’t know if plasma can be washed out by a diluent or if proteins in the plasma will be partitioned by membrane pore structures into different protein bands according to protein sizes and molecular weights. As such, early versions of these strips were made (by using a slitter such as a Kinematic 2360 from Kinematic Automation, the exact slitter for cutting regular lateral flow strips) without blood separation membrane attached. The primary goals were to look at the limitations of such miniaturized strips for fluid wicking consistency as well as fluid travel distances under different cutting widths, as shown in Figure 2. With a colored-dye PBS buffer solution, the ministrips (made from Millipore HF 120, 75, or others) were found to respond well for ~1–2 μl of solution at different cutting widths (Figure 2A) with excellent wicking consistency (Figure 2B) and travel distances (Figure 2C). The distance that 1–2 μl of solution can travel on a 1–1.25-mm strip was surprising to see, as it clearly demonstrated the potential of such miniaturized strips for being used as meters for small liquid samples not limited to blood.
To probe if nitrocellulose membranes in our proposed metering system could induce protein size exclusion banding, experiments were performed by first applying a plasma sample to one end of the nitrocellulose membrane and letting the capillary forces pull the plasma along the strip to the other end (as indicated in the bottom of Figure 3). After the plasma flow stopped, a sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) analysis of plasma extracted from contiguous regions of the miniaturized lateral flow strip was used to evaluate potential molecular partitioning of proteins (central panel in Figure 3). The SDS-PAGE data were also analyzed by densitometry to quantify the signal density for bands from the gel. A slight trend toward lower signal density was observed, but was not perceived as a significant effect. No molecular partitioning was observed along the length of the miniaturized lateral flow strip. This supports the notion that the sample transferred from any section of the BSL strip is representative of the original sample applied to the strip.
Table II. Miniaturized lateral flow strip width consistency.
1 | 2.00 | 2.00 | 100.01 |
2 | 2.03 | 2.00 | 101.60 |
3 | 1.84 | 2.00 | 92.08 |
4 | 2.02 | 2.00 | 100.81 |
5 | 1.95 | 2.00 | 97.63 |
6 | 1.98 | 2.00 | 99.22 |
Average | 1.97 | ||
St.Dev. | 0.06 |
Slitter Performance
As the width of the strips decrease, the cutting knife’s slitting quality (i.e., the width of the knife’s cutting line) will have a significant impact on the strip consistency. The direct measurement of the miniaturized strip width by a micrometer or an imaging system was done to evaluate the slitter’s cutting quality. Table II contains data from a random sampling of 2 mm wide metering strips. Based on this data and similar measurements, we concluded that current commercial slitters like the Kinematic 2360 from Kinematic Automation are good enough to make metering strips at widths wider than 1 mm. The slitting consistency can be conceivably improved if the cutting knife’s quality can be tailored to cut miniaturized lateral flow strips and a strict registration system can be implemented during the slitter’s cutting operation.
Another consideration for reliable metering quality is the lamination of a blood separation membrane onto the nitrocellulose membrane. From our observation, strips wider than 1 mm can generally provide a solid overlap interaction between the blood separation and nitrocellulose membranes to ensure the plasma transfer while strips below 1 mm width may require additional reinforcement for the lamination.
Blood Separation
After we demonstrated the feasibility of miniaturized strips for metering, we screened a variety of membranes including nonwoven fabrics to identify a fast and effective blood separation filtration material. From this exercise, the vertical blood separation membrane from PALL Inc. identified as BTS SP 300 came out as the top choice. BTS SP 300 offers fast blood intake and almost complete separation of red blood cells. The BTS SP 300 works by allowing only plasma into the nitrocellulose matrix, while the red blood cells are totally filtered by size exclusion in the z-axis direction (i.e., top of BTS plane to the bottom plane). The filtered plasma traverses through the nitrocellulose until the capillary force is no longer sufficient to continue pulling the plasma within the matrix. On a typical whole blood sample, the plasma travels a distance of roughly 12 mm on a 1 mm wide laminate. The exact distance traveled appears to be a function of nitrocellulose flow rate and blood hematocrit. Therefore, this can be a simple method for measuring the hematocrit of a blood sample in addition to metering.
In a typical BSL fabrication, a Millipore nitrocellulose HF 75 or HF 120 was first laminated onto a transparent or regular card material that serves as the backing strip shown in Figure 4A (a transparent backing) and 4B (a regular backing). A separate piece of BTS SP 300 blood separation membrane was then laminated onto the backing strip with the desired overlap with the nitrocellulose. The overlap of the two membranes can be between ~1–3 mm, depending on the blood sample size and transfer speed required. The backing card with laminated membranes can then be processed through a Kinematic slitter from Kinematic Automation Inc. to cut the assembled card into BSL strips with a desired width dimension (e.g., 1 mm, 2 mm).
It should be readily appreciated that economical mass production of the sample meters is not only possible but also practical. In Figure 4C, a representative meter is shown after first placing blood on the BTS section and then letting the separation finish (picture was taken after 30 min to show the colors of partially oxidized plasma on the nitrocellulose membrane).
It should be noted here that there are other ways to make BSL strips, particularly strips that are 1 mm wide or thinner. One example is to use laser burning by forming an isolation pattern (or totally burning off unwanted nitrocellulose) along the desired metering features on the strip, as shown in Figure 5A. For these small metering strips, it is fairly easy to demonstrate almost complete red blood separation from 1 and 2 μl blood samples with BTS blood separation membranes (1 μl for 5B and 2 μl for 5C).
Isolation of a Predefined Area in BSL for Quantitation by a Scraping System
To provide a precisely determined volume of the test sample to the main detection test strip, different internal scraping mechanisms, as shown in Figure 6 as one example, were designed and tested in our laboratory. These scraping mechanisms were configured to score and scrape the sample meter so that a well-defined region of the BSL strips is isolated (Figure 6C) and presented to the collection region of the main test strip. The scraping mechanism works by scoring the meter’s membrane to the backing substrate to create a physical barrier that prevents further fluid communication between the isolated zone and the rest of the strip. The length of the isolated section may be, for example, 5 mm or any other desired length. The section is saturated with the test sample fluid and, thus, based on the known saturation volume of the isolated area, a precisely determined amount of the test sample fluid is known and presented to the collection region of the test strip.
There are several methods to achieve this membrane isolation, but one example of a feasible scraping mechanism is shown in Figure 6A and Figure 6B (the blue metering strip line as Point A in Figure 6A is perpendicular to the detection strip Point B in 6B) and includes a pair of blades spaced apart and movably mounted in the device.1,2 In the original position illustrated in
Table III. The Scraper performance.
Trial | Average Length (mm) | St. Dev. | % CV |
1 | 5.26 | 0.09 | 1.7 |
2 | 5.38 | 0.16 | 3.0 |
3 | 5.28 | 0.12 | 2.2 |
4 | 5.23 | 0.07 | 1.3 |
5 | 5.25 | 0.07 | 1.3 |
Average | 5.27 | 0.06 | 1.1 |
The reproducibility of the scraping mechanism was assessed by using the exact scraping system that will be built into the final housing. The results are shown in Table III. The tabulated data (six different scrapers for one trial data point) as well as the actual scraping inspection (Figure 7) clearly demonstrated a highly consistent process, with the length and width spreads of only ~1% and 3% respectively for ~5.27 mm of isolated length. This leads to a compounded area spread at less than 4%. In comparison to the slitter performance data, the scraper performance is actually smaller and can surely be improved further.Figure 6A, the blades (Point C) are positioned below the membrane surface and are spaced apart a distance that defines the length of the the section to be isolated. In a second actuated position illustrated in Figure 6B, the blades (Point D) contact and score the membrane side of the sample metering strip to create a defined, isolated region. The blades may be mounted relative to the common longitudinal axis on opposite longitudinal sides of the test strip, as particularly illustrated in Figure 6A and Figure 6B, and configured to rotate away from the test strip in the second actuated position of the blades.
Integration of PAS to a Main Detection Strip in Housing
Integration of the preanalytical blood separation metering strip (PAS) described in this paper with a lateral flow test is highly contingent upon the design and configurations of the specific testing kit. An exemplary integration developed in our laboratory was to place the metering strip at the sampling zone of the detection lateral flow strip in a perpendicular orientation, as shown in Figure 8. The metering strip is not in contact with the nitrocellulose membrane of the lateral flow strip in the testing kit before the detection step. The metering strip in Figure 8 will first take in blood samples by spot sampling (or with a pipet, syringe, or related sampling device), complete the separation of red blood cells, and then be pushed down by a prebuilt scraper system for establishing fluid communication with the main detection strip’s nitrocellulose membrane. The scraper system is built to do three things simultaneously. First, it isolates a portion of the plasma storage membrane to have a fixed sample area for quantitative measurement. Second, it triggers the release of the diluent for washing out the analytes from the isolated plasma metering area. And third, it ensures the smooth integration of the isolated metering membrane with the main strip’s nitrocellulose membrane. The full details of the integration of a preanalytical bloob separation (PAS) system with a low-cost quantitative lateral flow test with built-in optics and reader have been developed and will be the subject of another publication.
Plasma Transfer Efficiency from PAS to Main Detection Strip
The volumetric capacity of the PAS and the kinetics of plasma transfer from the PAS to the main detection lateral flow strip are critical considerations for proper functioning of the diagnostic device. In our testing, PAS strips with widths of 1 mm, 1.5 mm, and 2 mm were tested for the amount of plasma that the excised portion of the PAS can meter. This was accomplished by extracting the plasma from the PAS into a phosphate buffered saline wash solution (100 μl) and then measuring the absorbance at 280 nm of the wash solution. From these experiments, some fragmentation of the PAS nitrocellulose was observed when strip widths are below 1.5 mm, which may be a source of interference leading to relatively large standard deviations. The comparison of washing samples to the pure plasma indicated that washing essentially removed all proteins from the PAS strips.
Efficiency of plasma transfer was first measured by examining the plasma that remained on the PAS after transfer by SDS-PAGE, as shown in Figure 9. A known amount of plasma, without exposure to the PAS or transfer, is the benchmark profile on the SDS-PAGE gel (lane marked “1ul plasma”). An empty (no plasma) excised portion of PAS was analyzed and showed no detectable signal by SDS-PAGE (lane marked “empty PAS”). PAS saturated with serum but without transfer (lane marked “PAS with plasma”) shows a profile identical to plasma alone (“1ul plasma”). Dry transfer, an attempted transfer to the lateral flow strip without activating the diluent buffer reservoir, results in little transfer with much of the plasma being retained on the PAS (lane marked “PAS with serum dry transfer”). Significantly, under normal transfer conditions (lane marked “PAS with serum wet transfer”), no detectable plasma remained on the PAS after a 4 min transfer time. This indicates that transfer of plasma from the PAS is an efficient process; transfer occurs with >99% efficiency, from qualitative assessment of the SDS-PAGE data.
With combined use of absorbance measurements and SDS-PAGE analysis, the kinetics of plasma transfer from the PAS to the lateral flow strip were measured (Figure 10). Plasma transfer was found to occur with >99% efficiency, near complete plasma transfer occurs by ~20 sec from test activation.
Compression Force for Sample Transfer
Ensuring intimate contact between the PAS and main detection strips after performing the scraping procedure is one of the key requirements for a complete plasma transfer from PAS membrane to main detection strip membrane. The plasma transfer will be inconsistent if the two membranes are pressed together too loosely or too tightly. If the contact is too loose, only diluent will pass through the main detection strip’s nitrocellulose membrane. If the contact is too tight, the diluent flow between the two membranes may be impeded. To understand this, experiments were conducted to determine the optimal amount of compression force required between the PAS and the test strip to obtain maximum sample transfer. Accordingly, sample lateral flow assays were run inside a specially designed breadboard device with different compression springs to achieve different amounts of force. SDS-PAGE analysis as described for the transfer efficiency experiment was used to determine the amount of protein sample left on the isolated portion of the PAS and the results were shown in Figure 11. These results indicate that complete sample transfer (>99%) is achieved with a minimum spring constant of approximately 1 lb/in. and the range of acceptable spring constants was determined to be between 1 and 4 lb/in.
Conclusions
The potential of using a miniaturized lateral flow strip as a blood separation and metering device has been demonstrated in this paper.1–2 With this development, lateral flow technology is now potentially feasible for spot blood sampling at alternative sites. As such, biomarkers in plasma can be directly accessed without the need for more costly and invasive surgical techniques (biopsies). In comparison to classical plasma preparation technologies (e.g., 5–10 ml tube and a centrifuge to recover 30–40% of the volume for proteomic analysis), a lateral flow diagnostic device with spot blood sampling can expand lateral flow technology into early diagnosis and continued monitoring of a variety of potentially serious health conditions such as cancers from a small volume blood sample.
References
1.Kaiyuan Yang Shawn R. Feaster, Ning Wei, Rosann M. Kaylor, Chibueze O. Chidebelu-Eze, “Metering Strip and Method for Lateral Flow Assay Devices,” U.S. Patent 7,618,810 B2, November 17, 2009.
2.Shawn R. Feaster, Kaiyuan Yang, Ning Wei, Chibueze O. Chidebelu-Eze, James M. Takeuchi, Rosann M. Kaylor, Enrico L. DiGiammarino, Jeffrey E. Fish, U.S. Patent Application 20080145272A1, June 19, 2008.
3.James M. Takeuchi, Xuedong Song, Kaiyuan Yang, Ning Wei, Shawn R. Feaster, “Metering Technique for Lateral Flow Assay Devices,” U.S. Patent 7279136 B2, October 9, 2007.
4.Charles T. Liamos, Joseph A. Vivolo, Fredric C. Colman, “Small Volume In Vitro Analyte Sensor and Methods,” U.S. Patent 6616819 B1, September 9, 2003.
5.Jacques Wallach, Interpretation of Diagnostic Tests, 6th Ed., (Philadelphia: Luppincott-Raven Publishers, 1996).
6.Aurora F. Castro, Jeseph W. Fraser, Janice L. Shultz, Surendra K. Gupta, “Blood Separation Filter Assembly and Method,” U.S. Patent 5139685 B2, August 18, 1992.
7.Robert S. Hillman, Ian Gibbons, “Blood Separation Device Comprising a Filter and a Capillary Flow Pathway Exiting the Filter,” U.S. Patent 5135719 B2, August 4, 1992.
Kaiyuan Yang is currently a senior research scientist in Kimberly-Clark’s corporate research and engineering department. Yang has a Ph.D. degree in organic/organometallic chemistry and he is an author/inventor with 50 peer-reviewed research articles and 31 issued U.S. patents. He can be reached at [email protected].
James Takeuchi is currently a project scientist in Kimberly-Clark’s global healthcare sector. Takeuchi has a M.S. in biomedical engineering and is an inventor with nine issued U.S. patents. He can be reached at [email protected].
You May Also Like