Is the Arthrex Angel System a Lemon or Have They Made Lemonade? Part Two

In the previous post, I referred to comments that surgeons had shared with me at the AOSSM conference in early July about the Arthrex Angel system, which Arthrex markets to produce hematocrit (Hct)-specific PRP and BMC preparations. None of the surgeons I asked was able to provide any evidence that specific Hct levels were of clinical value in generating PRP or BMC. That left me wondering why Arthrex was taking the approach of marketing Hct-based PRP and BMC preparations. There are processing implications of selecting Hct as an output characteristic, as I pointed out in the previous post. Moreover, the potential quirkiness of the Arthrex Angel-Hct approach was revealed in a paper by Oliver, et al. (2015; see previous post for the link). I had summarized data taken from Table 7 in the paper in which I had calculated the fold-decrease in cellular components when the BMC was made with a 15% or a 2% Hct setting on the Angel system, as follows:

Fold-decrease in going from 15% Hct to 2% Hct for BMC
Prepared by the Angel System (Oliver, et al.; 2015)

Hct      7.5x

Plt        7.1x

WBC    27.8x

NE       107x

LY        19.2x

MO      20x

RBC    37.8x

TNC    25.7x

HPC    146x

As is evident from the values listed in going from a higher level of Hct to a lower value, which is achieved in the Angel system by selecting the desired target Hct level, there are serious variations in recoveries of important components like HPCs (a designation of CD34+ progenitor cells, which clinically is used as a marker for hematopoietic stem cells).

What does this mean? My interpretation of the fold-decrease values is that the response of the Angel’s LEDs and the programed algorithm for collecting the BMC is highly dependent on the Hct of the sample being processed. For example, consider the behavior of nucleated cells: while the fold-reduction for TNC (total nucleated cells) and WBC were close, there was a huge reduction in the HPC level—a type of progenitor cell, which is a nucleated cell of the monocyte cell type. And there isn’t just a little variation in recovery of HPCs, their number dropped by a whopping 146-fold when going from a 15% Hct to a 2% Hct, while the WBC component only dropped by 28-fold. Clearly, the Oliver et al. data suggests that the various cellular components in the bone marrow being processed are not treated “equally” in the Angel system. I consider this disparity as evidence of non-linearity in the way the Arthrex Angel system processes whole blood or bone marrow aspirate.

Given the wild variations in fold-reduction for the various important components in BMC, I think it is fair to say the Angel system very well could behave in an unpredictable manner patient-to-patient. The non-linearity of the Angel system suggested in the Oliver, et al. paper also means that a physician can’t reliably predict how many progenitor cells will be present in BMCs produced from patient to patient. After seeing this data, I think I understand why Arthrex’s marketing approach is to have orthopedic surgeons focus on Hct, since setting an Hct target value is easy on the Angel system.

While the Oliver, et al. publication revealed evidence of the non-linear operational behavior of the Angel system, the data reported in Table 7 reveals another unpleasant surprise for orthopedic surgeons: the number of HPCs are seriously low in terms of absolute numbers of HPCs recovered. HPCs (hematopoietic stem and progenitor cells) are cells that have been reported to correlate to CD34+ cells, but without the need to use flow cytometry (Buttarello and Plebani, 2008). From Table 7, the average number of HPCs reported in the 15% Hct level BMC samples was 29,200 cells/mL of BMC, and the total number of HPCs on average for a 7 mL BMC volume (the upper limit of the variable volume output reported by Oliver, et al. for the set of BMCs studied) comes to 204,400 HPCs. As indicated in the Buttarello and Plebani article, the HPC measurement was sensitive to delays in measurement of approximately 50%. Even if you double the HPC value to 408,800, since the samples were sent out for hemoanalysis, you still are woefully short of this type of stem cell when compared with the CD34+ cell data for BMC samples reported in the Pettine, et al. paper on treating degenerative disc disease with BMC (just a heads up—I am a coauthor on the paper). Pettine, et al. reported that there was an average of 1.69 x 106 CD34+ cells per mL of BMC (approximately 7 mL), which gives a total number of 11.8 x 106 CD34+ cells. Not sure where the stem cells went in the Angel-processed BMCs, but there clearly is a substantial under-collection of these critical cells as reported in the Oliver, et al. paper.

There is a lack of data published on the Arthrex Angel system in terms of hemoanalysis of PRP and BMC preparations made with it, so I will cover in the next post additional details about component analysis of the BMCs reported in the Oliver, et al. paper, as well as introduce data from the only other paper I have found that reported on the hemoanalysis of PRP preparations made from whole blood from seven unrelated blood donors processed in several different device-based technologies, including the Angel system at 2% and 7% Hct levels.

Leave a Reply