Low-cost blood plasma separation method using salt functionalized paper

Abstract

This study describes an extremely low-cost method for separating plasma in a sample of whole human blood on salt functionalized paper by means of osmotic pressure. When a sample of whole blood was introduced onto the salt functionalized paper, plasma dissolves the salt and places the red blood cells (RBCs) in a hypertonic medium. This leads to the generation of osmotic pressure across the cells membrane, and also the crenation of RBCs. The effect of different concentrations of salt on RBC deformation and crenation has been monitored using confocal microscopy. Depending upon the salt concentration, RBCs deform into various shapes under osmotic pressure. At high salt concentration, RBCs turn into deflated thin disks. This increases the RBCs' contact with one another and with fibres in the paper as well. Besides, the counter ion valency charge of the Na⁺ suppresses the thickness of the charged double layer of RBC. Subsequently, aggregation of the deflated RBCs occurs. The aggregates are large enough to be separated chromatographically from the plasma phase of the wicking front. Our results show that 0.5 μL addition of 0.68 M (4% w/v) saline solution (NaCl) can provide sufficient plasma separation on a filter paper for diagnostic applications. A colorimetric blood glucose concentration assay is employed to demonstrate the efficiency of this plasma separation method on paper. The experimental investigation indicates that although the crenation of RBCs forced a small amount of water into plasma, this method is suitable for performing glucose assay in human blood on paper. Our method can enhance bioassays performed on microfluidic paper-based analytical devices (μPADs) by combining the separation and testing of plasma into a single device with no significant additional cost.

---

Introduction

Paper and alternative materials have been proven to be a suitable substrate for the fabrication of microfluidic devices.^1–8 One benefit of using these materials is their low-cost; making them accessible for developing and remote areas.^9,10 To date, microfluidics paper-based analytical devices (μPADs) have mostly been used for colorimetric detection of biomarkers existing in human urine and blood samples.¹¹ Urine typically has no strong colour and is suitable for colorimetric analysis. However, haemoglobin pigments in red blood cells give blood a strong red colour, rendering whole blood samples incompatible with colorimetric diagnostic tests. Therefore, colorimetric assays using blood samples often require the separation of plasma from blood. Plasma is a major part of human blood. It consists of water, sugars, proteins and other substances. It directly correlates with the condition of the body, and can be used as the subject of sensitive biological assays and medical diagnostics.¹² When a sample of whole blood is placed on a paper (i.e. filter or chromatography paper), plasma separation does not occur. This is because typical papers are not able to filter out RBCs from the plasma phase, and blood penetrates the fibre network and stains the paper.¹³ There are some types of paper (i.e. Whatman LF1, MF1, VF1 and VF2) that are designed specifically to separate plasma from sample of whole blood.^14–16 The downside is that they are comparatively expensive (∼100 times of Whatman no. 1 and no. 4 filter papers), so there is good reason for more research to be done to find cheaper methods for plasma separation.

Several groups have developed various methods¹⁷ for human blood separation such as: egg beater centrifugation¹⁸ membrane-based sedimentation¹⁹ and the application of negative dielectrophoretic force.²⁰ However, these methods require the use of additional equipment and time for sample preparation, which limit their application for on-site biosensing.

Recently, a few groups have proposed using antibody trigged RBC haemagglutination to separate plasma from whole human blood samples on paper.^13,21,22 The basic principle of this method is based on a strong specific antibody–antigen interaction, leading to the agglutination of RBCs inside the fibre network of paper. The large lumps formed inside the fibre network as a result are not able to move in the network. However, the plasma phase still moves forward, separating from agglutinated RBCs.^13,21,22 Since the strong reaction of specific antibody–antigens (A, B and Rh) is crucial, this method works well for blood samples of types A, B, AB and O Rh positive. However, its application for O Rh negative, a phenotype that has a significantly high frequency in human population,¹² and weak ABO subgroups^12,23 have not been demonstrated.²⁴ Although some researchers suggested that anti-H would be able to separate type O negative RBCs from plasma phase on paper,¹³ there is no reported experimental evidence supporting the idea. In addition, this method also requires preparatory blood typing to know whether antibody–antigen interactions are able to induce blood agglutination. Furthermore, antibodies stored on paper for a prolonged time require extra treatment to maintain their bioactivity; this will significantly increase the material and equipment costs for producing such devices.²⁵

Most recently, Li et al. used calcium chloride solution to aggregate RBCs in rabbit blood samples.²⁶ Calcium ions are able to destabilise the suspension of RBCs in blood by suppressing the electric double layer on the RBCs surfaces, reducing the charge repulsion between RBCs.²⁷ Although, they showed that aggregation of RBCs occur in high salt concentration,²⁷ the effect of osmotic pressure produced upon addition of the salt solutions was not discussed.

As μPADs have been mainly designed for medical diagnostic applications for use in developing regions, the development of low-cost and sensitive methods for in situ blood analysis on paper is particularly advantageous. In this paper we describe an extremely simple and low-cost method for separating plasma from a sample of whole human blood on μPADs. Salt solutions with different concentrations have been utilized to generate a hypertonic medium for RBCs. The magnitude of osmotic pressure causes the RBCs to change shape, resulting in close packing of RBCs and the formation of RBC aggregates in fibre networks and being immobilized. The plasma phase, however, still wicks forward in the fibre network; it can therefore be separated from the aggregated RBCs and collected in the detection zone. The underlying mechanism and the effect of salt concentration on RBC morphology have been investigated using confocal microscopy. The increase in salt concentration not only increases the osmotic pressure across the cell membrane, but also compresses the cell surface electric double layer. Contributions of these two factors to plasma separation on paper were semiquantitatively evaluated using salts containing metal ions of different valency. Finally, the efficiency of this method of plasma separation was verified by performing colorimetric determination of glucose concentration in different blood samples.

Experimental section

Materials, reagents and blood samples

Whatman no. 4 filter paper was obtained from Sigma-Aldrich. Paraffin wax (55 FRG) was purchased from Dussek Campbell, Australia and used to pattern paper, creating a simple device with a sample inlet and detection area, by a previously reported wax patterning method.^28–31 Blood samples from adult donors were obtained from Red Cross, Australia and stored at 4 °C in Vacutainer® test tubes containing lithium-heparin anticoagulant and used within 5 days of collection. A hand-held Glucometer (Accu-Check Mobile) was obtained from a commercial store in Melbourne, Australia. Physiological saline solution (PSS) and phosphate-buffered saline (PBS) solution were prepared from analytical grade (AR, from Sigma-Aldrich) NaCl, KCl, Na₂HPO₄, and KH₂PO₄. Anti-H and anti-A,B IgM blood typing antibodies were purchased from ALBA, UK and used as received without dilution. Glucose assay reagent was prepared by dissolving 5 mg of 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid (ABTS)) reagent into horseradish glucose oxidase–peroxidase solution (500 units of enzyme activity and buffer salts in 39.2 mL deionized water).³² Fluorescein isothiocyanate (FITC, isomer I) and anhydrous dimethylsulphoxide (DMSO) were obtained from Sigma-Aldrich and MERCK Chemicals Ltd, Australia, respectively. ID-CellStab red cell stabilization solution was purchased from BioRad, Australia. The confocal images were captured by a Nikon Ai1Rsi confocal microscope.

Methods

Effect of salt concentration on RBC morphology and plasma separation. The effect of different concentrations of salt solution on the morphology of RBCs was investigated. Saline solution was used as a model salt to examine RBC behaviour under different osmotic pressures. Confocal microscopy was used to observe cell morphology changes under different hypotonic conditions. The stock saline solution (1.54 M or 9% w/v) was prepared by dissolving 9 g of sodium chloride in deionized water, to a total volume of 100 mL. It was then diluted with water to obtain serially diluted solutions with lower concentrations (1 (6% w/v), 0.68 (4% w/v), 0.34 (2% w/v) and 0.154 M (0.9% w/v)). The Whatman no. 4 filter paper was functionalized by the serially diluted saline solution prior to introducing 5 μL of FITC-labelled blood sample for confocal observation. RBCs were stained with FITC using a previously reported fluorescent-labelling method.³³ Samples of whole blood between 1 mL and 1.5 mL were centrifuged in 2 mL micro-tubes at 1300g (R.C.F.) for 5 minutes to separate the RBCs from the plasma fraction and buffy coat (white cells and platelets). The top two layers were then removed carefully using a disposable pipette (2.5 mL, Sigma Aldrich). RBCs were re-suspended to 40% haematocrit using PBS. Following another centrifugation and removal of the liquid phase, the RBCs were re-suspended to 40% haematocrit using a cell stabilising solution (ID-CellStab) containing the FITC stain (0.8 mg mL⁻¹) and incubated for 2.5 hours, with gentle mixing every 20 minutes to maintain suspension. Once staining was complete, cells were washed repeatedly to remove any FITC that was not bound to the RBCs to allow them to be clearly observed during confocal microscopy. Cells were washed in PSS a total of 10 times, centrifuging 3 minutes between washes at 1300g. After the final wash, cells were re-suspended to 40% haematocrit with ID-CellStab solution for conservation of the samples.

The effect of salt concentration on the plasma separation method was also studied. A 5 μL of whole blood sample was gently dosed on the inlet area of patterned papers, each functionalized with 0.5 μL of one of the serially diluted saline solutions before drying for around 3 min. The flow resistance of RBCs changed in different salt concentration as they aggregated, leading to appreciable changes in extracted plasma volume. The separation distance of plasma from the RBC front (defined as the wicking distance of separated plasma from the nearest RBC front) on patterned paper was measured using ImageJ by finding edge of region of interest and calculating the separated plasma length (see Fig. 1SA†). The data was transferred into Microsoft Excel to obtain a calibration curve (the separation distance of plasma versus salt concentration).

Since there is no reported study of H antigen anti-H reaction on paper sensors for purpose of plasma separation, we performed such assays on paper to test the feasibility of using anti-H to separate O group RBCs from their plasma phase, which was proposed recently in the literature.¹³ Different volumes of blood typing anti-H, and anti-A,B antibodies (2, 3, 4, 5 and 7 μL) were deposited on Whatman no. 4 filter paper and left to dry at room temperature for 10 min. Afterwards, 5 μL of whole blood sample was introduced onto the antibody-treated zone. The experiment has also been repeated on glass slides. The results are presented in the ESI (Table 1S, Fig. 2S and 3S†).

Effect of metal counter ion valency on plasma separation. RBCs carry a weak negative charge,¹² which contributes to the stability of RBC suspension in plasma. Whilst salts generate osmotic pressure across the RBC membrane, they also destabilize the RBC suspension through suppressing the electric double layer on the surface of RBCs. In order to understand the contributions of these two factors to plasma separation, salts of different metal ion valencies were used to investigate the plasma separation efficiency on paper. We used two salt solutions made of saline (NaCl) and magnesium chloride (MgCl₂), to functionalize the patterned paper, while keeping the osmotic pressure constant. Since osmotic pressure is a colligative property of the solutions, it is dependent upon the ratio of the number of solute molecules to the number of solvent molecules, and not on the species of the solute molecules. Ideally, 0.67 molar of MgCl₂ can generate the same osmotic pressure as 1 molar of NaCl. Therefore, magnesium and sodium chloride solutions have been prepared with concentration of (0.1, 0.22, 0.45 and 0.67 molar) and (0.154, 0.34, 0.68 and 1 molar), respectively. The sample inlet area of the patterned paper was functionalized by depositing 0.5 microlitres of one of the serially diluted salt solutions onto its surface, and allowed to dry for 3 min. A 5 μL of whole blood sample was then introduced into the salt functionalized area of the patterned paper. The separation distance of plasma was measured using ImageJ software as mentioned before (Fig. 1SB†) and the data were transferred into Microsoft Excel to plot a calibration curve (the separation distance of plasma versus salt concentration). The error bars represent the standard deviation of at least three different measurements.

Device fabrication and operation. Whatman no. 4 filter paper was used as a substrate for fabricating two-dimensional μPAD using wax.^28,29 In short, hydrophobic wax was transferred onto the paper through wax-impregnated paper with a soldering iron. An aliquot of 1 μL of colorimetric glucose indicator was deposited into the detection zone. The device was then functionalized by depositing 0.5 μL of 0.68 M (4% w/v) saline solution onto the sample inlet area, and left to dry for around 6 min at room temperature. The fabrication and patterning method cost < %0.01 as we used normal Whatman filter paper and simple wax patterning method to create the microfluidic channels on paper.¹ The functionalization of the device using sodium chloride solution is also low-cost, even lower than filter paper, which is negligible (calculation based on the Sigma-Aldrich price shows the cost of salt for each device being 10⁻⁵ cent). A 5 μL of whole blood sample was gently applied into the sample inlet of the device. As RBC aggregates formed, they were unable to travel through the fibre network of paper with the plasma phase, while plasma wicked forward into the fibre network, separating from the RBC aggregates.

The effect of the initial blood sample volume on the separation distance of plasma was also investigated¹³ by placing various volumes of blood (3, 5, 7, 10 μL) onto the 0.68 M saline functionalized μPAD. The separation distance of plasma was measured using ImageJ software (Fig. 4S†) and the data transferred into Microsoft Excel to plot a calibration curve (the separation distance of plasma versus salt concentration).¹³

Quantification of glucose in separated plasma. Glucose concentration in whole blood samples was measured spectrophotometrically using a Glucometer according to the instructions of the manufacturer and used for glucose assay using the salt functionalized device.¹³ Plasma in the same blood samples was also separated using a lab scale centrifuge and used for determining the glucose level on μPADs. A 3.5 μL of blood plasma, which is enough to fill the device, was introduced into each sample inlet area. The centrifuged plasma was wicked into detection area of μPAD and reacted with deposited glucose indicator. In the presence of glucose, a green colour produced which is correlated with the concentration of glucose in plasma. The results were then compared with the colorimetric results achieved using our separated plasma method for glucose detection in blood. Through this comparison, it is expected to examine the suitability of the separated plasma and effect of released water from inside RBCs to the plasma in glucose assay.

Results and discussion

Effect of salt concentration on RBC morphology and plasma separation

μPAD was functionalized by depositing and drying a half microliter of saline solution in room temperature. A 5 μL of whole blood sample was then introduced onto the device. The blood plasma dissolved the salt from paper and diluted it in whole blood by ten times. This means that the functionalized paper using i.e., 0.68 M saline solution, only added 0.068 M of the salt to the whole blood sample, assuming perfect mixing. However, it was found that the addition of this low volume of salt solution can assist separation of plasma from sample of whole blood. In order to understand the underlying mechanism of the plasma separation, the morphological structure of RBCs on salt functionalized paper was determined from the confocal images, as illustrated in Fig. 1. To ensure that the blood staining protocol required for confocal microscopy does not strongly affect the RBCs, we first thought that it is desirable to show them under isotonic condition on glass slide and filter paper. The conventional glass slide method was performed (see Fig. 1A) to capture the image using the confocal microscope. As shown in Fig. 1A, RBCs have perfect donut shape as they should. But, isotonic condition is different on paper (Fig. 1B). This is because that the cellulose fibres tend to absorb water and swell. Besides, water can evaporate through both sides of the paper. In this circumstance, paper easily becomes dried and salt concentration increases. Consequently, the medium changes from isotonic to hypertonic condition. The comparison between Fig. 1B and C indicates that the morphological changes of RBCs in hypertonic medium are dependent upon the salt concentration which generates the corresponding level of osmotic pressure across the cell membrane. In paper functionalized with saline solution, concentration of 0.154–0.34 M or 0.9–2% w/v, RBCs lose water due to the osmotic pressure and show crenation, followed by an increase in RBC viscosity.³⁴ However, we found very thin discs were formed when the salt concentration was increased to 0.68 M or 4% w/v. As cells deflated, aggregations of RBCs formed which were unable to wick with plasma in the fibre network (Fig. 2). As shown in Fig. 2A, the wicking distance of plasma was also increased when the salt concentration increased.

Fig. 1 Confocal micrographs of RBCs on (A) glass slide under the isotonic condition, (B) and (C) Whatman no. 4 filter papers functionalized with saline solution having different salt concentration; (B) 0.154 M (0.9% w/v or isotonic condition) and (C) 0.68 M (4% w/v), respectively. As shown in the images, morphological structure of the RBCs changed on salt functionalized paper.

Fig. 2 Effect of salt concentration and cation valence on plasma separation distance: Whatman no. 4 filter paper functionalized using different concentrations of (A) saline and (B) magnesium chloride solutions.

Regarding the anti-H and RBC interaction test, it was found that the H antigen anti-H interaction is weak and produced much smaller agglutinated RBCs lumps than the anti-A,B and A antigen interaction. The former can still travel with plasma phase within the paper network; no strong plasma-RBC separation could be observed (see Fig. 2S, 3S and Table 1S†). Such limited plasma separation on paper caused by anti-H is much weaker than that caused by anti-A,B with their corresponding RBCs, and could not be used for practical sensing applications.

Effect of salt valency on the plasma separation method

It is expected that salt causes RBC separation from the plasma phase through two mechanisms: osmotic pressure and RBC surface charge neutralization. In order to determine the dominant mechanism of plasma separation, we evaluated surface charge neutralization using cations of different valence to suppress the electric double layer on the surface of RBCs, while keeping the osmotic pressure constant. According to the Schulze-Hardy rule, the critical coagulation concentration (CCC) varies, and is inversely proportionate to the sixth power of the counter ion valence (CCC ∝ 1/z⁶, where z is the valence of the counter ion). If the electrical charge repulsion between RBCs plays a significant role, the magnesium salt would be expected to cause much stronger RBC aggregation than sodium salt. As shown in Fig. 2, magnesium and sodium salts do not show significant difference in their capacity for plasma separation on paper. Therefore, the effect of electric double layer on plasma separation could only be secondary to the effect of osmotic pressure. In other words, besides the salt-induced decrease in charge repulsion,²⁶ salt added on paper also changes the morphology of RBCs introduced on paper. The interactions between the highly deflated RBCs allow them to aggregate into a more compact.³⁵ Consequently, an increase in the salt concentration reduces the aggregation time and shortens RBCs advancement with the plasma phase resulting stronger RBC and plasma separation. Based on our finding, in paper-based device design, we used sodium chloride solution (0.5 μL of 0.68 M) to functionalize paper for simplicity and cost efficiency.

Plasma separation on paper

The schematic diagram shown in Fig. 3, illustrates the use of salt functionalized μPAD for efficient separation of plasma from whole blood samples through osmotic pressure. As expected, RBC aggregation does not occur in an isotonic medium and whole blood samples could easily penetrate through paper networks by capillary driving force as a single fluid. However, when blood samples are introduced into a μPAD that is functionalized with a high concentration of sodium chloride solution, RBCs will be exposed to a hypertonic medium; RBCs lose water and shrink in size under osmotic pressure. In this circumstance, RBCs aggregate. The strongly deflated, disc-like RBCs form closely packed lumps and are separated from the plasma phase. The separated plasma phase continues to wick forward into the detection zone of the paper device (Fig. 3).

Fig. 3 Schematic illustration of the plasma separation method on saline functionalized μPADs: (A) device functionalization; (B) addition of whole blood sample to the device; (C) plasma separation. As shown, RBCs lose their water in a hypertonic medium and stay behind the capillary front, while plasma collects in the detection zones for biomedical diagnostics.

The plasma separation efficiency was also studied based on the ratio of the separation distance of plasma to the initial volume of blood sample dosed on the functionalized paper.¹³ As shown in Fig. 4 and 3S,† the wicking distance of the separated plasma front was up to 2.5 mm, depending on the initial volume of the blood sample applied. Such a separation is sufficient for the performance of low sample volume blood analysis.

Fig. 4 The effect of initial blood sample volume in the separation distance of plasma. The sample inlet point of the device was functionalized using 0.5 microlitres of 0.68 M saline solution.

Quantification of glucose concentration in whole blood using the plasma separation method

Glucose was used as a model analyte sample to evaluate the suitability of the separated plasma for diagnostic assays on μPAD. The physiological range of glucose concentration in sample of human blood is 3.5–5.5 mM which has been used in this study.^38,39 Glucose level in each sample was determined by introducing 5 μL of whole blood sample on the salt functionalized inlet point of each device. Blood plasma was separated from RBC aggregates and filled the detection zone, previously treated with colorimetric glucose indicator, within 1 min. In the presence of atmospheric oxygen, the horseradish oxidase catalyzes glucose in the plasma to gluconic acid and hydrogen peroxide.^36,37 The horseradish peroxidase then catalyzes the reaction of hydrogen peroxide with ABTS to form a green colour within 20 min (Fig. 5). Blood glucose level has been quantified by colour intensity measurement of the detection zones. As shown in Fig. 5, the intensity of the green colour in each detection zone was plotted as a function of glucose concentration in whole blood sample. According to the data presented in Fig. 5, the colour intensities measured from different levels of glucose by our plasma separation method are lower than the colour intensities obtained from pure plasma assays performed using the same blood samples. Potential reason for lower colour intensity and the higher standard deviation of our test is due to hematocrit dependent dilution of the sample. Within a wide physiological range of hematocrits (35–55%) the change in color intensity is at most 10%, which would not affect the capability of this test to effectively distinguish the majority of blood glucose pathologies as they often cause higher fluctuations in blood glucose levels. For example diabetic blood glucose concentrations are on average ∼30% higher than that of healthy patients.⁴⁰

Fig. 5 Semi-quantitative measurement of glucose concentration using pure plasma and whole blood sample on μPADs.

Finally, we calculated R² value of both calibration curves obtained from the glucose assays using pure plasma and whole blood sample on the salt functionalized paper. The results indicate that our method shows a linear relationship to glucose concentrations measured with standard glucose measurement method (an approved commercial Glucose Meter). Such performance suffices the requirement for determination of glucose, across a range of physiologically relevant glucose levels in human blood.

Conclusions

This paper demonstrates a simple and low-cost paper-based plasma separation method using salt functionalized μPADs. The addition of salt solution to μPAD, before introducing blood sample, leads to the suppression of the electric double layer on the surface of RBCs and also generation of osmotic pressure in the sample of whole blood. The underlying mechanism has been investigated using confocal microscopy showing that the salt concentration affects the morphological structure of RBCs. At high salt concentrations in blood, strongly deflated and crenated RBCs form closely packed lumps, which are unable to travel in a fibre network with the plasma phase by capillary wicking, and are left behind the wicking front of the plasma phase. Blood plasma separation using our method does not require special equipment or any chemicals other than salt solution, and is well suited for use in resource-limited setting. A diagnostic application has been demonstrated by performing glucose assay in the separated plasma showing its suitability for small sample volume blood analysis. The fabrication and operation cost of our method will allow designing more functional μPADs for use in developing and underdeveloped areas.

Acknowledgements

This work is supported by the Australian Research Council Grant (DP1094179) and (LP1120973). A.N. gratefully acknowledges the research scholarships provided by Monash University and the Department of Chemical Engineering. The authors also would like to thank Dr John Zhu of the MCN for his guidance in the use of the confocal microscope. The help of Mr Hansen Shen from Monash University for proof reading is also acknowledged.

References

1. A. W. Martinez, S. T. Phillips, G. M. Whitesides and E. Carrilho, Anal. Chem., 2010, 82, 3–10.
2. D. R. Ballerini, X. Li and W. Shen, Microfluid. Nanofluid., 2012, 13, 769–787.
3. X. Li, D. R. Ballerini and W. Shen, Biomicrofluidics, 2012, 6, 011301–011313.
4. X. Li, J. Tian, G. Garnier and W. Shen, Colloids Surf., B, 2010, 76, 564–570.
5. X. Li, J. Tian, T. Nguyen and W. Shen, Anal. Chem., 2008, 80, 9131–9134.
6. X. Li, J. Tian and W. Shen, Anal. Bioanal. Chem., 2010, 396, 495–501.
7. A. Nilghaz, D. R. Ballerini, X. Y. Fang and W. Shen, Sens. Actuators, B, 2013, 191, 586–594.
8. A. Nilghaz, L. Zhang, M. Li, D. R. Ballerini and W. Shen, ACS Appl. Mater. Interfaces, 2014, 6, 22209–22215.
9. W. Dungchai, O. Chailapakul and C. S. Henry, Anal. Chim. Acta, 2010, 674, 227–233.
10. M. Webster and V. S. Kumar, Clin. Chem., 2012, 58, 956–958.
11. A. W. Martinez, S. T. Phillips and G. M. Whitesides, Proc. Natl. Acad. Sci. U. S. A., 2008, 105, 19606–19611.
12. L. Dean, National Center for Biotechnology Information, 2005.
13. X. Yang, O. Forouzan, T. P. Brown and S. S. Shevkoplyas, Lab Chip, 2012, 12, 274–280.
14. T. Songjaroen, W. Dungchai, O. Chailapakul, C. S. Henry and W. Laiwattanapaisal, Lab Chip, 2012, 12, 3392–3398.
15. M. M. Gong, B. D. MacDonald, T. V. Nguyen, K. Van Nguyen and D. Sinton, Lab Chip, 2014, 14, 957–963.
16. J. Noiphung, T. Songjaroen, W. Dungchai, C. S. Henry, O. Chailapakul and W. Laiwattanapaisal, Anal. Chim. Acta, 2013, 788, 39–45.
17. M. Kersaudy-Kerhoas and E. Sollier, Lab Chip, 2013, 13, 3323–3346.
18. A. P. Wong, M. Gupta, S. S. Shevkoplyas and G. M. Whitesides, Lab Chip, 2008, 8, 2032–2037.
19. C. Liu, M. Mauk, R. Gross, F. D. Bushman, P. H. Edelstein, R. G. Collman and H. H. Bau, Anal. Chem., 2013, 85, 10463–10470.
20. Y. Nakashima, S. Hata and T. Yasuda, Sens. Actuators, B, 2010, 145, 561–569.
21. L. Ge, S. Wang, X. Song, S. Ge and J. Yu, Lab Chip, 2012, 12, 3150–3158.
22. M. S. Khan, G. Thouas, W. Shen, G. Whyte and G. Garnier, Anal. Chem., 2010, 82, 4158–4164.
23. B. Thakral, K. Saluja, M. Bajpai, R. R. Sharma and N. Marwaha, Science, 2005, 36, 32–34.
24. A. K. Yetisen, M. S. Akram and C. R. Lowe, Lab Chip, 2013, 13, 2210–2251.
25. L. Guan, R. Cao, J. Tian, H. McLiesh, G. Garnier and W. Shen, Cellulose, 2014, 21, 717–727.
26. H. Li, D. Han, G. M. Pauletti and A. J. Steckl, Lab Chip, 2014, 14, 4035–4041.
27. F. I. Ataullakhanov, A. V. Pohilko, E. I. Sinauridze and R. I. Volkova, Thromb. Res., 1994, 75, 383–394.
28. A. Nilghaz, D. H. B. Wicaksono, D. Gustiono, F. A. Abdul Majid, E. Supriyanto and M. R. Abdul Kadir, Lab Chip, 2012, 12, 209–218.
29. A. Nilghaz, D. R. Ballerini and W. Shen, Biomicrofluidics, 2013, 7, 051501–051516.
30. A. Nilghaz, D. H. B. Wicaksono and E. Supriyanto, 2nd International Conference on Instrumentation Control and Automation, Bandung, Indonesia, 2011.
31. A. Nilghaz, D. H. B. Wicaksono and F. A. Abdul Majid, 2nd International Conference on Instrumentation Control and Automation, Bandung, Indonesia, 2011.
32. C. A. Chaplan, H. T. Mitchell and A. W. Martinez, Anal. Methods, 2014, 6, 1296–1300.
33. M. Sauer, J. Hofkens and J. Enderlein, in Handbook of Fluorescence Spectroscopy and Imaging, Wiley-VCH Verlag GmbH & Co. KGaA, 2011, pp. I–IX.
34. W. H. Reinhart and S. Chien, Am. J. Physiol., 1985, 248, C473–C479.
35. L. Li, J. Tian, D. Ballerini, M. Li and W. Shen, Analyst, 2013, 138, 4933–4940.
36. E. Hohenberger and H. Kimling, Compendium Urinalysis with test strips, Roche Diagnostics, 2004.
37. M. Reches, K. A. Mirica, R. Dasgupta, M. D. Dickey, M. J. Butte and G. M. Whitesides, ACS Appl. Mater. Interfaces, 2010, 2, 1722–1728.
38. A. Nilghaz, S. Bagherbaigi, C. Lam, S. Mousavi, E. Córcoles and D. B. Wicaksono, Microfluid. Nanofluid., 2015, 1–17.
39. Z. Nie, C. A. Nijhuis, J. Gong, X. Chen, A. Kumachev, A. W. Martinez, M. Narovlyansky and G. M. Whitesides, Lab Chip, 2010, 10, 477–483.
40. Diabetes.co.uk, the global diabetes community, http://www.diabetes.co.uk/diabetes_care/blood-sugar-level-ranges.html, accessed 21 March, 2015.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra01468a

---