Monitoring hemolysis continuously in real time
MMonitoring hemolysis continuously in real time
Tyler Van Buren, ∗ Gilad Arwatz, Alexander J. Smits Mechanical and Aerospace Engineering Department, Princeton UniversityPrinceton, NJ 08544, USA Instrumems Inc.Sunnyvale, CA 94805, USA ∗ To whom correspondence should be addressed; E-mail: [email protected].
Blood damage (hemolysis) can occur during clinical procedures, e.g. dialysis,due to human error or faulty equipment, and it can cause significant harm tothe patient or even death. We propose a simple technique to monitor changesin hemolysis levels accurately, continuously, and in real time. As red bloodcells rupture, the overall conductivity of the blood increases. Here, we demon-strate that small changes in porcine blood hemolysis can be detected accuratelythrough a simple resistance measurement.
Introduction
Hemolysis (the rupture of red blood cells) can occur in medical procedures where blood isremoved from the body (
1, 2 ). For example, passing blood through a faulty dialysis machinecan potentially risk the life of the patient (
3, 4 ), and even drawing blood too quickly througha needle can lead to defective laboratory samples ( ). Preventing hemolysis is therefore animportant design constraint for medical pumps, prosthetic organs, hypodermic needles, and1 a r X i v : . [ phy s i c s . m e d - ph ] J u l igure 1: Simplified illustration of blood flow to show how hemolysis leads to a higher bloodconductivity. The fluid becomes more conductive left-to-right as red blood cells rupture andrelease their hemoglobin.blood extraction procedures ( ).The cytoplasm of red blood cells is rich in hemoglobin, an iron-containing biomolecule thatcan bind oxygen and is responsible for the red color of the cells. As red blood cells rupture,they release their hemoglobin into the plasma— which is mostly water— changing the plasmafrom being relatively colorless to having a red tint. The degree of hemolysis can be mea-sured by separating the plasma from the red blood cells and analyzing the amount of cell-freehemoglobin ( ) using a spectrophotometer, which measures how much light of a given wave-length is absorbed by the sample. Spectrophotometry is considered the most accurate methodfor measuring hemolysis, but there are a number of other possible methods that do not requireextracting red blood cells or using chemical analysis. For example, Tarasev ( ) suggested ablood hemolysis analyzer that can measure the amount of cell-free hemoglobin by using two ormore wavelengths of light and comparing the different levels of light absorption. Karlsson ( )and Lee ( ) both describe devices to identify hemolysis by the naked eye through the changein plasma color. None of these methods, however, allow clinicians to monitor hemolysis during2 procedure and provide immediate information on the level of blood damage.Recently, Zhou et al. ( ) proposed a complex method that could actively measure hemol-ysis by combining nanofilters, which actively filter the plasma from the red blood cells, andoptofluidic sensors for evanescent absorption detection. The process is similar to spectropho-tometry, in that it is analyzing light absorption and relating it to hemolysis level, with the addedbenefit of eliminating the need for specialty sample preparation.We propose a much simpler technique that can detect hemolysis continuously in real-timeby measuring the electrical resistance of the blood. Blood is naturally conductive, but the outerlipid bilayer of the red blood cell is insulating and so healthy blood cells do not contributemuch to the overall conductivity (
15, 16 ). As red blood cells rupture, however, they release theirhemoglobin and raise the conductivity of the entire fluid (as illustrated in figure 1). Hence, thechange in blood conductivity can be related directly to the level of hemolysis. Our proposedtechnique does not require external sources of light, the separation of the blood cells from theplasma, or specific chemical detection, yet it can determine the progression of hemolysis in realtime.
Results and discussion
The method uses a test cell that consists of a small converging/diverging channel equipped withtop and bottom electrodes. The cell was tested by inserting it in a laminar flow loop driven bya peristaltic pump, as shown in figure 2. The conductivity was measured using a high-qualityinductance-capacitance-resistance (LCR) meter for continuous sampling, and a conventionalconductivity probe for periodic sampling.The system was first tested using KCl saline solutions to achieve mass concentrations of0.5%, 1%, 2%, and 4% in deionized water. This range of salinity was chosen to give changesin conductivity similar to that expected under moderate levels of hemolysis. Figure 3 shows the3igure 2: Experiment schematic. Flow path: (1) magnetic stirrer with open reservoir; (2) peri-staltic pump; (3) 3D printed channel with a conductive floor and ceiling in the test section.variation in time of the saline resistivity, ρ = R/h , where R is the resistance in ohms and h isthe channel height (note that conductivity is defined as /ρ ). As salt is added to the reservoir,there is an initial step change in the resistivity followed by more gradual asymptotic behavioras the salt dissolves. The conductivity probe measurements were made just before each stage ofsalt addition, and we see that they agree well with the LCR meter readings. The measurementsalso match well with the known conductivity of KCl solutions ( ).The system was then tested using porcine blood. Measuring the absolute resistivity of bloodcan be challenging, in that it can act as a dielectric ( ), and also the flow shear can align theorientation of the red blood cells and make the electrical properties of blood anisotropic ( ).Since we are only interested in measuring relative changes in blood resistivity for a given flowcondition, these effects can be ignored. To control the level of damage, the blood was splitinto two separate 500 mL samples. One sample was left pristine while the other sample wasmechanically damaged using an immersion blender (77% hemolyzed). The undamaged sample4igure 3: Saline resistivity, as measured continuously with the LCR meter and periodically withthe conductivity probe, compared with the corresponding theoretical values.was used as the starting fluid in the flow loop, and then every 10 min 50 mL of the damagedblood sample was added into the stirring reservoir to increase the hemolysis levels in gradualsteps. Samples of 1.5 mL were extracted twice at each blood damage level for direct hemolysismeasurements using the spectrophotometer as described in the methods section.Figure 4.a shows the blood resistivity measured by the LCR meter (left axis) and hemolysispercentage measured using the spectrophotometer (right axis) over the 90 minute test period.Each addition of damaged blood causes a step change in the blood resistance followed by aslower asymptotic behavior as the mixture homogenizes. Figure 4.b shows a direct correlationbetween the blood conductivity level ( /ρ ) and the total hemolysis percentage.We see that a simple conductivity cell can be used to immediately detect changes in hemol-ysis. The measurement is continuous, in real time, and easy to implement in clinical practice.In dialysis, for instance, the test cell can simply be incorporated in the blood flow loop and5a)(b)Figure 4: (a) Blood resistivity (left axis) and hemolysis percentage (right axis) as a function oftime. (b) Blood conductivity as a function of hemolysis percentage.6llow for immediate detection of blood damage. With further development, the test cell can ob-viously be made smaller and constructed using biocompatible materials, and incorporate stable,high-accuracy resistance measurement electronics. Materials and methods
Measurements were made in the recirculating flow facility shown in figure 2. Fluid was drivenby a peristaltic pump (Cobe 043600-000) at 41.67 cm /s through a test channel and an openreservoir equipped with a magnetic stirrer (Sargent-Welch). The channel was custom designedand 3-D printed from a waterproof photopolymer (Watershed 11122XC) at the W.M. KeckCenter for 3D Innovation at the University of Texas El Paso. The channel test section hadlength l = 66 mm, height h = 7 . mm, and width w = 25 . mm and was equipped withan electrically conductive brass floor and ceiling for flow resistance measurements. The openreservoir station allowed for direct conductivity measurements of the recirculating fluid, sampleextraction, and mixing in other materials.During the tests, the resistance of the fluid was continuously measured with an inductance-capacitance-resistance meter, or LCR meter (Keysight Technologies E4980AL) which can con-tinuously read resistance with 0.1% accuracy. Measurements were recorded via LabView at 2Hz, and tests were conducted for up to 90 minutes. Direct conductivity measurements werealso made with a more conventional conductivity probe (Hach HQ14D), accurate to 0.5%, tovalidate the LCR meter readings.Two working fluids were used for this study. First, a potassium chloride (KCl) basedsaline solution was used to validate our resistance measurements, which was made by mix-ing a known mass of KCl (EMD PX1405-1) measured via a precision scale (VWR 1002E) intoroom-temperature deionized water. Second, we used room-temperature porcine blood for test-ing hemolysis. The blood was purchased fresh through Lampire Biological Laboratories where7t was obtained from healthy adult animals of unspecified gender with an added anticoagulantheparin. Hemolysis levels were determined by measuring the relative levels of free hemoglobinin the blood plasma using a spectrophotometer (Beckman Coulter DU730).The procedure to obtain hemolysis percentage measurements of blood samples using a spec-trophotometer is as follows.[1] Prepare a Drabkin’s Solution by combining Drabkin’s Reagent (Sigma Aldrich D5941) to 1L of deionized water.[2] To lyse the red blood cells, prepare a separate solution containing 100 mL of the Drabkin’ssolution from Step 1 and 0.05 mL of 30% Brij 35 Solution (Sigma Aldrich B4184).[3] Set aside two 1.5 mL samples of undamaged blood into centrifuge tubes (samples measuredvia Eppendorf 5 mL adjustable volume pipette). Then, acquire and similarly store bloodsamples during the experiment.[4] Centrifuge the blood samples for 3 minutes at 6000 RPM to separate the red blood cellsfrom the plasma. Leave one of the two 1.5 mL undamaged blood samples uncentrifuged.[5] In spectrophotometer cuvettes, mix 2 mL of the Drabkin’s Solution with 8 µ L of plasmafrom the centrifuged blood samples (extracted with Eppendorf 10 µ L adjustable volumepipette).[6] Mix 8 µ L of the undamaged/uncentrifuged blood sample from Step 4 with 2 mL of theDrabkin’s + Brij 35 Solution from Step 2 into a spectrophotometer cuvette.[7] Ensure that all cuvette samples are well mixed and allow to rest for 15 minutes.[8] Zero the spectrophotometer using only the Drabkin’s Solution (2 mL) in a cuvette, thisserves as the “blank” sample.[9] Using the spectrophotometer, measure and record the baseline reference case of the cuvettewith the original undamaged blood sample from Step 5. Denote as A , where A is the8bsorbance at a wavelength of nm.[10] Similarly, measure and record the fully damaged reference case from the lysed blood samplemade in Step 6. Denote as A ∞ .[11] Lastly, measure and record the spectrophotometer readings from all of the experimentalsamples from Step 5.[12] The relative hemolysis of a given sample is given by ( A − A ) / ( A ∞ − A ) . References
1. S. P. Sutera,
Circulation research , 2 (1977).2. G. Arwatz, A. Smits, Biorheology , 45 (2013).3. H. Polaschegg, Seminars in dialysis (Wiley Online Library, 2009), vol. 22, pp. 524–531.4. J. Yoon, S. Thapa, R. D. Chow, B. G. Jaar,
BMC research notes , 475 (2014).5. P. Carraro, G. Servidio, M. Plebani, Clinical chemistry , 306 (2000).6. J. W. Mulholland, J. C. Shelton, X. Y. Luo, Journal of fluids and structures , 129 (2005).7. M. Giersiepen, L. J. Wurzinger, R. Opitz, H. Reul, The International journal of artificialorgans , 300 (1990).8. M. K. Sharp, S. F. Mohammad, Annals of biomedical engineering , 788 (1998).9. M. S. Grant, Journal of Emergency Nursing , 116 (2003).10. V. Han, K. Serrano, D. V. Devine, Vox sanguinis , 116 (2010).11. M. Tarasev, Blood hemolysis analyzer (2010). US Patent 7,790,464.92. M. Karlsson, Arrangement for detection of hemolysis (2014). US Patent App. 14/363,368.13. K. H. Lee, Hemolysis detector (1994). US Patent 5,330,420.14. C. Zhou, et al. , ACS sensors , 784 (2018).15. F. G. Hirsch, et al. , Blood , 1017 (1950).16. K. R. Visser, Medical and Biological Engineering and Computing , 636 (1992).17. R. C. Weast, M. J. Astle, W. H. Beyer, et al. , CRC handbook of chemistry and physics , vol.1990 (CRC press, Boca raton FL, 1989).