Ghost Cells as a Two-Phase Blood Analog Fluid -Fluorescent Mechanical Hemolysis Detection

Abstract

Background: This study investigated fluorescent hemolysis detection as an optical method to detect local hemolysis in mechanical circulatory support systems, addressing the limitations of standard hemolysis tests and current simulation methods. Standard tests, per ASTM1841-19, quantify general hemolysis but do not localize it.

Methods: We employ a two-phase blood analog fluid composed of calcium-loaded ghost cells and phosphate-buffered saline. Ghost cells are hemoglobin-depleted red blood cells, allowing for optical measurements. A calcium-sensitive fluorescent indicator (Cal590 potassium salt, AAT Bioquest, Pleasanton, USA), activated by calcium released upon ghost cell hemolysis, enables fluorescent hemolysis detection. Hemolysis tests were conducted using porcine whole blood and the blood analog fluid, confirming that both undergo mechanical hemolysis in the Food and Drug Administration pump model.

Results: The results revealed increased fluorescence intensity in response to hemolysis, with a quantitative fluorescence increase of 8.85/min at 3500 rpm and 2.5 L/min, indicating hemolysis, particularly at the rotor tip. Through image processing of fluorescence images, local hemolysis was visualized.

Conclusion: This study is the first to use fluorescent hemolysis detection for local detection of mechanical hemolysis. Further refinement may enhance the design of mechanical circulatory support systems and bridge simulation limitations with experimental, localized hemolysis detection.

Keywords: fluorescent mechanical hemolysis detection; resealed ghost cells; translucent two‐phase blood analog fluid.

FIGURE 1

EDTA concentration‐dependent fluorescence intensity. The fluorescence intensity was measured in a 96‐well microplate reader (Tecan Spark, Tecan Trading AG, Switzerland) at different time points after mixing 100 μL of GC with 100 μL of PBS with increasing EDTA concentrations. [Color figure can be viewed at wileyonlinelibrary.com]

FIGURE 2

Time‐dependent hemolysis of the hemolysis tests is demonstrated as free plasma hemoglobin in the hemolysis test with a blood analog fluid made from ghost cells (GCs) and porcine whole blood. Normalized free plasma hemoglobin accounts for the different hemoglobin concentrations in blood and GCs. It increases significantly (p < 0.05) faster for GCs compared with blood. [Color figure can be viewed at wileyonlinelibrary.com]

FIGURE 3

Normalized extracellular calcium of the hemolysis tests. The calcium concentration was measured via optical emission spectrometry with inductively coupled plasma (ICP–OES, Spectroblue, Spectro Analytical Instruments GmbH, Germany). [Color figure can be viewed at wileyonlinelibrary.com]

FIGURE 4

Fluorescence intensity of the hemolysis tests. The fluorescence was measured with 200 μL of sample in a 96‐well microplate reader (Tecan Spark, Tecan Trading AG, Switzerland). [Color figure can be viewed at wileyonlinelibrary.com]

FIGURE 5

Images for fluorescent hemolysis detection before adding EDTA, before hemolysis on the FDA pump, and after mechanical hemolysis at a pump speed of 3500 rpm. [Color figure can be viewed at wileyonlinelibrary.com]

FIGURE 6

Mean value of fluorescence images taken with fluorescent hemolysis detection. The image mean values for 1198 images are calculated from the added double‐frame images. The mean of multiple images is then calculated from the last 50 images of one series. For this run at 3500 rpm, an increase in fluorescence of 8.85/min was measured. [Color figure can be viewed at wileyonlinelibrary.com]

FIGURE 7

Fluorescence hemolysis detection on single images with mean‐based z‐standardization. Panels (A–G) show selected single images that visually demonstrate expected FHD signal patterns, with localized brightness increases at the rotor tip or within the diffuser region, consistent with hemolysis hotspots. Panels (H–J) display similar but less pronounced signal intensities. In contrast, panels (K–O) highlight the limitations of single‐image analysis: Variations in laser intensity and noise result in inconsistent regions being falsely highlighted. [Color figure can be viewed at wileyonlinelibrary.com]

FIGURE 8

Fluorescence hemolysis detection resolved locally after image processing. A total of 6761 images were processed by reducing them to a 74 × 74 raster, dividing each image by its own mean value, z‐standardizing the images with the same EDTA concentration, and adding all image values per raster together. The results for each individual EDTA concentration are added in the Appendix S1. [Color figure can be viewed at wileyonlinelibrary.com]