Eliminating viscosity bias in lateral flow tests

Lateral flow tests are widely used across sample types such as whole blood, serum, urine, saliva, and milk. Flow in LFTs is driven by capillarity through porous membranes with functionalized test lines where analyte-detection molecule binding occurs. For certain biomarkers, quantitative readouts are required, but variations in sample viscosity change wicking speed and flow rate, altering incubation times at binding sites and biasing signal intensity and analyte quantitation. Biological samples can exhibit large viscosity ranges (e.g., saliva 1.5–23 mPas; plasma 1.50–1.72 mPas at 25 °C; urine 0.635–0.797 mPas at 37 °C), and diseases can further alter plasma viscosity. Predilution with viscosity-matching buffers can equalize viscosity but reduces sensitivity and adds handling steps. Conventional capillary flow controls typically slow flow by increasing resistance or surface modifications, but they do not remove viscosity dependence. Previous approaches in capillaries introduced air-flow resistance or used a pump liquid to achieve viscosity-independent inflow, but translating these to standard LFTs is challenging due to sealing and adjustability issues. Centrifugal microfluidics enables precise, externally adjustable flow by tuning rotational speed and integrating sample preparation, yet prior centrifugal LFT solutions retained viscosity dependence. The study proposes a centrifugal cassette employing pneumatic flow control to balance sample inflow with air outflow, rendering the sample flow rate independent of sample viscosity while allowing external flow-rate tuning.

Prior work addressed viscosity effects by predilution with buffers of defined viscosity, which can impair sensitivity. Capillary-based solutions balanced sample flow with air outflow or used a pump liquid to decouple inflow from sample viscosity and surface energy, but these approaches face challenges when interfaced with standard LFT membranes, particularly achieving leak-tight seals and limited adjustability of flow rates. Centrifugal microfluidic platforms have demonstrated controllable flow via rotation and integration of preprocessing steps such as plasma separation and dilution; however, previously reported centrifugal LFT implementations still showed viscosity-dependent flow. This work builds on pneumatic balancing concepts demonstrated in capillary tubes and adapts them to conventional nitrocellulose LFTs within a centrifugal cassette to achieve viscosity-independent operation.

Device concept: A thermoformed centrifugal foil cassette comprises a vented inlet chamber connected via a low-resistance transfer channel (with a built-in siphon to prevent air backflow) to a pneumatic chamber housing the nitrocellulose membrane. The pneumatic chamber is vented only through a defined venting resistance channel. Under rotation, sample moves from the inlet through the transfer channel into the pneumatic chamber. The restricted venting induces an overpressure that opposes further inflow until equilibrium is reached, where sample inflow equals air outflow through the venting resistance channel. Flow-rate setting is achieved by the venting channel geometry and rotational speed; dependence is on air viscosity rather than sample viscosity. Bypass prevention is ensured by limiting inflow to or below the membrane’s maximum throughput at the set frequency based on the design guideline Q_M,max = rho_s * r_in,M * omega^2 * A_M * kappa / eta_s. The centrifugal pressure is given by Delta p_cent = rho_s * omega^2 (r2^2 - r1^2) / 2. The viscosity-independent flow rate relates to rotational frequency and vent geometry as Q_s^2 = (rho_s * omega^2 (r2^2 - r1^2)) / (2 eta_a * l_r / (A_r * C_g)). Fabrication: The cassette integrates three identical pneumatic flow-control structures, each with an inlet chamber, transfer channel, pneumatic chamber containing the membrane, and a venting resistance channel; optional vents were present but not used. Experimental fluids: Phosphate-buffered saline-based buffers were mixed with PEG1000 to yield viscosities of 1.1 mPas, 2.3 mPas, and 24 mPas. Measured densities were 1.021 g/ml, 1.037 g/ml, and 1.092 g/ml, respectively. Operation: Cassettes were run at constant rotational frequency of 15 Hz. Flow characterization measured the transferred volume during acceleration and steady-state flow rates for each viscosity. Theoretical expected flow rates were computed from the above relations and compared to observed values; membrane maximum flow limits Q_M,max were also estimated. Immunoassay: A model competitive human IgG LFT was used. The nitrocellulose membrane carried a human IgG test line. Samples were mixed with gold nanoparticles conjugated with anti-human IgG antibodies. In the absence of analyte, conjugates bind to the test line producing a strong signal; in the presence of human IgG analyte, conjugates bind in solution, reducing binding at the test line and lowering signal. The assay was performed as a conventional dipstick with 1.1 mPas and 2.3 mPas buffers and, in parallel, within the centrifugal cassette employing pneumatic flow control. Test lines were scanned and peak intensities quantified across analyte concentrations to compare viscosity-induced signal shifts between formats. Statistics: Flow-rate comparisons across viscosities used significance testing; p values were reported.

• The cassette achieved viscosity-independent flow across a >20-fold viscosity range (1.1–24 mPas) at 15 Hz: overall mean 3.01 ± 0.18 μl/min (±6%). Individual means: 1.1 mPas, 2.98 ± 0.14 μl/min (CV 5.0%); 2.3 mPas, 3.00 ± 0.15 μl/min (CV 5.0%); 24 mPas, 3.11 ± 0.29 μl/min (CV 9.5%). No significant differences between viscosities (p = 0.198). - During acceleration, 3.84 ± 0.24 μl of sample entered the pneumatic chamber to establish overpressure. - Theoretical expected flow rates from pneumatic balance at 15 Hz matched observations: approximately 3.02 μl/min (1.1 mPas), 3.07 μl/min (2.3 mPas), and 3.22 μl/min (24 mPas), all below membrane maximum flow limits to avoid bypass (Q_M,max about 6.43 μl/min for 1.1 mPas and 3.12 μl/min for 2.3 mPas at 15 Hz). - In the dipstick format, doubling sample viscosity from 1.1 to 2.3 mPas increased test-line signal by an average of 38 percentage points, evidencing viscosity bias due to longer incubation times. - In the centrifugal cassette with pneumatic flow control, the viscosity-induced signal shift between 1.1 and 2.3 mPas buffers was reduced to about 6 percentage points on average, representing more than an 84% reduction in viscosity bias. - No bypass was observed in the cassette under the tested conditions.

The study addresses the challenge that variable sample viscosity alters capillary wicking speeds in LFTs, changing incubation times and biasing signal intensities. By using centrifugal-driven pneumatic flow control that balances sample inflow with air outflow through a designed venting resistance, the cassette decouples flow from sample viscosity and ties it to air viscosity and centrifugal pressure. This yields consistent flow rates and incubation times without predilution or modification of porous materials. The observed 84% reduction in viscosity-induced signal shift in a competitive human IgG assay demonstrates that controlling flow via pneumatic balancing can substantially improve quantitative reliability. The residual 6 percentage point increase at higher viscosity likely stems from PEG1000 effects on binding rather than flow, suggesting the method can also help isolate buffer chemistry influences because incubation time is held constant. The approach allows external tuning of assay flow via rotational speed and vent geometry and can be combined with other centrifugal sample preparation steps while maintaining quantitative control.

A centrifugal microfluidic cassette implementing pneumatic flow control eliminates viscosity bias in conventional nitrocellulose LFTs by matching sample inflow to air outflow through a defined resistance. The system delivers a stable, viscosity-independent flow rate of about 3 μl/min over a broad viscosity range and reduces viscosity-induced signal shifts by more than 84% in a model human IgG assay, without requiring sample dilution or changes to the membrane. The method enables accurate, externally adjustable control of assay flow and incubation times and offers a practical route to more quantitative LFTs. Potential future work includes validating the approach with diverse biomarkers and real clinical matrices, exploring a wider range of rotational speeds and vent geometries, integrating upstream sample preparation, and investigating buffer component effects on binding independent of flow.

Experiments were performed at a single nominal rotational frequency (15 Hz) and with model buffers whose viscosities were adjusted using PEG1000, which may affect binding chemistry. The immunoassay demonstration used a single competitive human IgG model rather than multiple clinical targets or real patient samples. The high-viscosity condition had a smaller sample size (n = 8). Maximum flow is constrained by membrane permeability and sample viscosity (Q_M,max), requiring careful design to avoid bypass at higher set flow rates or different membranes. Publication details do not report broader environmental or temperature variations beyond the stated conditions.