In the previous two posts (Part One and Part Two), I have reviewed aspects of the performance of the Arthrex Angel system, which is being marketed to orthopedic surgeons to produce PRP and BMC for therapeutic treatments. I have covered the basics of the system and reviewed data from a publication that compared the cellular composition of six bone marrow aspirates processed with a 15% hematocrit (Hct) and five bone marrow aspirates processed with a 2% Hct. The resulting BMCs showed dramatically different yields of components like RBCs, Total Nucleated Cells (TNCs), WBCs and HPCs (hematopoietic progenitor cells; thought to be equivalent to CD34+ cells). I have pointed out that the Angel system has some operational quirkiness when creating BMC preparations, since the fold decrease in Hct between 15% and 2% is 7.5x, but RBCs changed by almost 40x, WBCs by 28x, and HPCs by an incredible 146x (see Part One and Part Two for the gory details).
At the end of the last post, I had compared the absolute yield of HPCs from the Angel system at 15% Hct as reported in the Oliver, et al. publication with the level of CD34+ cells in a publication by Pettine, et al., who had evaluated the number of CD34+ cells in BMCs prepared by a competitive centrifugation-based device platform. There was a shockingly low recovery of the hematopoietic progenitor cells in the Angel system with a reduction of 29x over the yield reported for the BMCs produced in the competitive device technology.
How does the Arthrex Angel system do in the recovery of Plts in BMCs prepared at 15% and 2% Hct? As indicated in Table 7 of the Oliver, et al. paper, there were just 145,000 Plt/mL recovered when the BMCs were collected at 2% Hct, which jumped up to a respectable 1,030,000 Plt/mL for the 15% Hct level. What these results suggest is that orthopedic surgeons who think they might want to reduce the Hct levels for their BMC preparations—for a reason that escapes me, but I am the other kind of doc—need to be careful with how low they go. If Plts in BMCs contribute at all in providing a therapeutic benefit, an orthopedic surgeon wouldn’t want to drop the Hct level down by very much or else Plts will join the hematopoietic progenitor/stem cells in the waste bag.
While the Oliver, et al. paper focused on BMC preparations, what about platelet recovery levels from whole blood used to produced PRP? As it happens, there is a publication that looked at the yield of Plts, RBCs, WBCs and Neutrophils for seven unrelated blood donors when aliquots were processed in a variety of centrifuge-based device technologies, including 7% and 2% Hct levels for the Arthrex Angel system. The Degen, et al. paper showed that the yield of Plts in going from a Hct level of 7% to a 2% Hct decreased by just 1.1x, while the Hct dropped by a theoretical 3.5x. I say theoretical, because Degen, et al. also measured the Hct directly, reporting that the “7%” Hct-specified level yielded a measured Hct of 8.8% on average, while the “2%” Hct-specified level yielded a measured Hct of 1.1% on average. Thus, the actual measured fold-decrease in Hct was 8x, not the theoretical 3.5x.
The variation reported in Degen, et al. between the selected Hct value and the measured outcome suggests that even the Hct of the resulting preparation will vary in an unpredictable way from the selected Hct. The “7%” Hct preparations yielded an average of 8.8% Hct, a +26% deviation, whereas the “2%” Hct preparations yielded an average of 1.1% Hct, a -45% deviation. These variations in selected and measured Hct are somewhat at odds with what I was told by orthopedic surgeons at AOSSM, who thought the Hct levels of their preparations matched pretty closely to the selected Hct. It might be that the orthopedic surgeons don’t select low Hct values, where deviations might be much higher, compared to higher values of targeted Hct that might more closely match the measured Hct. Nonetheless, the Degen, et al. paper on PRP also would seem to point to non-linear behavior of the Angel system in view of the variation in the measured Hct from the physician-selected Hct.
So, let’s summarize what we can infer from the data reported in the Oliver, et al. and Degen, et al. papers and from physician experiences shared with me at AOSSM when the Arthrex Angel system is used with whole blood or bone marrow aspirate:
- Physicians report that there is noticeable variation in the volume of the PRP or BMC produced when Hct is specified in the Angel system.
- Physicians are under the false impression that Hct is an important parameter to consider when treating patients with PRP and BMC.
- The Angel system doesn’t perform any analysis of the resulting PRP or BMC, but it can collect preparations at specified Hct levels, an outcome which is reported by physicians to be accurate, but which isn’t supported by the results in the Degen, et al. paper, especially at the 2% Hct level.
- The degree of reduction in critical components like RBC, WBC, and HPC (a measure of stem cells that are CD34+) when comparing component analysis of BMCs collected with 15% or 2% Hct varies widely: RBC levels decreased by 38x, WBC levels by 28x and HPC levels by 146x.
- Collection of BMC in the Angel system at 15% Hct shows a serious reduction of 29x in the absolute number of CD34+ progenitor/stem cells (measured as HPC) present when compared to a published report of CD34+ cells in BMCs obtained with a competitive centrifugation-based device.
- One possible explanation for the behavior of the Angel system is that it appears to be highly dependent on the Hct of the patient’s sample.
- There is no publicly available clinical data supporting the use of Hct as a guide to preparing PRP and BMC.
- There is no publicly available clinical data supporting the use of any specific Hct value to treat musculoskeletal conditions.
I stated in the first post of this series that I wasn’t aware of anything that was precise about the Arthrex Angel system, but I stand corrected, sort of. According to several physicians with whom I interacted at AOSSM, they measured the Hct values of PRP or BMC preparations and found them to be close to the Hct that was selected, except this apparent agreement might depend on the actual level of the Hct selected, as shown in the Degen, et al. paper. More importantly, I still don’t understand what Hct has to do with producing a good therapeutic preparation of PRP or BMC. All in all, in view of the non-linear processing of patient samples, I believe the Arthrex Angel system remains a lemon despite Arthrex’s marketing efforts to make it into lemonade.