DIABETES TECHNOLOGY & THERAPEUTICS
`Volume 20, Number 5, 2018
`ª Mary Ann Liebert, Inc.
`DOI: 10.1089/dia.2018.0106
`
`COMMENTARY
`
`Glucose Sensing in the Subcutaneous Tissue:
`Attempting to Correlate the Immune Response
`with Continuous Glucose Monitoring Accuracy
`
`Jeffrey I Joseph, DO, Gabriella Eisler, BS, David Diaz, PhD, Abdurizzagh Khalf, PhD,
`Channy Loeum, and Marc C. Torjman, PhD
`
`S afe and effective blood glucose (BG) control using a
`
`closed-loop artificial pancreas system requires a continu-
`ous glucose monitoring (CGM) system that is reliable, accu-
`rate, timely, and easy to use.1–5 The real-world performance of
`commercial needle-type subcutaneous tissue glucose sensors
`has significantly improved over the past 20 years, due to op-
`timization of their mechanical design, porous membranes,
`enzyme electrochemistry, signal processing, automated intro-
`ducer, and method of manufacture.6–8
`Despite these technological advances, the accuracy and
`longevity of current CGM are limited due to the subcutane-
`ous tissue’s cellular and humoral immune response to glucose
`sensor insertion and maintenance.9–14 The tissue’s reaction to
`short-term sensor implantation (1–7 days), medium-term im-
`plantation (7–14 days), and long-term implantation (>14 days)
`includes tissue injury, blood–biomaterial interaction, provi-
`sional matrix formation, acute inflammation, chronic inflam-
`mation, granulation tissue development, foreign body reaction,
`and fibrosis/fibrous capsule formation.15 The duration of each
`sequence and degree of immune reaction depend upon the
`tissue type, the amount of initial tissue trauma, the bioma-
`terial’s size, shape, composition, and surface texture, addi-
`tional tissue trauma due to macro-/micromotion, chemical
`leaching, electric current leaching, and the age and immune
`competency of the individual patient.11,14–22
`Insertion of a glucose sensor’s needle and electrode
`through the epidermis and dermis into the subcutaneous tis-
`sue damages cells, connective tissue, and extracellular ma-
`trix.9,11,12 The sensor’s thin and flexible electrode remains
`within this region of subcutaneous tissue after removal of the
`introducer needle. Injured arterioles, capillaries, and ve-
`nuoles release plasma, red blood cells, white blood cells, and
`activated platelets into the surrounding tissue.
`Arterioles may initially constrict to limit blood loss but
`will eventually dilate to increase local blood flow for nutrient
`delivery and waste product removal. Activation of the co-
`agulation cascade and the release of cytokines cause capil-
`
`laries to leak protein-rich plasma into the wound that coats
`the electrode with a layer of proteins and thrombus.15 A
`chemical gradient guides the migration of acute inflamma-
`tory cells from adjacent vascular tissue to the local environ-
`ment surrounding the implanted electrode.24
`Neutrophils and monocytes/macrophages infiltrate the
`surrounding tissue and actively phagocytize bacteria, injured
`cells, and debris, and release protease enzymes and reactive
`oxygen species. Fibroblasts, lymphocytes, eosinophils, and
`mast cells also migrate from adjacent capillaries into the
`wound and secrete a wide variety of cytokines, chemokines,
`and growth factors that modulate the acute inflammatory and
`wound healing process.11,14,15,17,22–24
`Glucose sensor performance (accuracy, time lag, and sig-
`nal stability) can be highly variable for 1–12 h after insertion
`of the electrode into the subcutaneous tissue. The reasons for
`this variability are multifactorial, complex, and poorly un-
`derstood.25,26 Loss of functional capillary vessels and local
`vasoconstriction can decrease the delivery of glucose, oxy-
`gen, and other nutrients, and decrease the removal of cellular
`waste products. Injured lymphatic vessels can limit the re-
`moval of edema fluid, cellular waste products, and phago-
`cytized debris. The tissue surrounding the glucose sensor
`electrode may become acidotic due to the accumulation of
`carbon dioxide, lactic acid, and protons, an environment that
`may cause variable enzyme function (glucose oxidase/de-
`hydrogenase) or function of the electrochemistry.11,27 The
`large number and high metabolic activity of the neutrophils
`and macrophages that surround the implanted electrode may
`further decrease the local concentration of oxygen and in-
`crease interstitial fluid acidosis.
`A layer of acute inflammatory tissue (containing damaged
`cells, connective tissue, edema fluid, and immune cells) may
`surround the working electrode to become a mechanical or
`physical barrier that significantly inhibits/slows the inward
`diffusion of glucose and oxygen. Metabolically active cells
`adjacent to the electrode (red blood cells, macrophages, and
`
`Department of Anesthesiology, Jefferson Artificial Pancreas Center, Sidney Kimmel Medical College, Thomas Jefferson University,
`Philadelphia, Pennsylvania.
`
`321
`
`ABBOTT Exhibit 2044
`Dexcom v. Abbott Diabetes Care, IPR2023-01409
`
`

`

`322
`
`JOSEPH ET AL.
`
`neutrophils) actively consume glucose, significantly de-
`creasing the number of glucose molecules reaching the
`working electrode. Macrophages have been identified as the
`major cell type producing a ‘‘Cell-Based Metabolic Barrier’’
`that limits the diffusion of glucose from the adjacent inter-
`stitial fluid to the sensor’s electrodes, causing an artificially
`low sensor output signal.10,12,28 Thus, performance of an
`enzyme-based electrochemical glucose sensor may be sig-
`nificantly affected by the dynamically changing local tissue
`environment immediately adjacent to the working and ref-
`erence electrodes.16,25
`The cellular, humoral, and chemical environment sur-
`rounding a CGM electrode will start to stabilize within sev-
`eral hours and become more stable within 12 h of
`implantation, depending upon the amount of initial tissue
`trauma, ongoing tissue trauma due to body movement, and
`the degree of immune response produced by the individual
`patient. Thrombus will undergo fibrinolysis, neutrophils and
`macrophages will continue to phagocytize debris, and cap-
`illary vessels will regain their vasomotor control and no
`longer release protein-rich fluid into the wound.15,27
`Neutrophil numbers will decrease after a few days if the
`wound does not become contaminated/infected with bacteria.
`Macrophage numbers will increase over time to further
`phagocytize debris and modulate the immune response. Fi-
`broblasts will transform into myofibroblasts in 2–5 days and
`synthesize extracellular matrix and collagen with fibrils that
`may parallel the surface of the electrode in an attempt to
`confine or wall off the foreign body.19–21 A dense layer of
`macrophages may surround the electrode and combine to
`produce foreign body giant cells.15,22 A layer of inflamma-
`tory cells and fibrous tissue that becomes thick, dense, and
`continuous may significantly affect sensor performance and
`limit longevity.10,17,22 The formation of granulation tissue
`with new normally functioning capillary and lymphatic
`vessels may take several weeks to develop and probably will
`not form in the environment immediately adjacent to the
`CGM’s electrodes.29,30
`Early generation subcutaneous tissue glucose sensors were
`Food and Drug Administration (FDA) approved for only
`3 days of use and required frequent recalibration multiple
`times per day to maintain sufficient accuracy for real-time
`monitoring, but not good enough for dosing insulin without a
`confirmatory BG measurement. Improvements in electrode
`insertion, size, softness, flexibility, membrane biomaterial,
`and electrochemistry have decreased the degree of initial and
`ongoing tissue damage, leading to enhance sensor stability/
`performance.
`These enhancements along with improvements in signal
`averaging and filtering have improved accuracy enough for
`the FDA to approve the use of real-time CGM data to dose
`insulin without requiring a confirmatory BG measurement.4–8,31
`Recent clinical trials using current CGM in closed-loop and
`hybrid closed-loop artificial pancreas systems have been
`extremely promising.1–3 Despite these advances, the accu-
`racy, lag time, reliability, longevity, and overall clinical
`performance of current commercial CGM systems are limited
`by the variable subcutaneous tissue’s immune response to
`needle/electrode insertion and maintenance.
`In a recent issue of the journal Diabetes Technology &
`Therapeutics, Rigla et al. describe a human clinical trial that
`correlated the subcutaneous tissue’s immune response with a
`
`metric of CGM accuracy.32 The authors should be com-
`mended for completing a systemic study in ambulatory hu-
`mans that
`included a qualitative/quantitative analysis of
`tissue histology surrounding implanted glucose sensors for
`1 and 7 days. These data are important because the number
`of CD68 positive macrophages/0.01 mm2 surrounding the
`electrodes was significantly higher in those CGM that had a
`mean absolute relative difference (MARD) >10%, whereas
`the CGM electrodes surrounded by significantly fewer mac-
`rophages had a lower MARD% (better accuracy). This as-
`sociation between macrophage location/density on CGM
`performance and other histology findings is consistent with
`prior in vitro and animal study results.
`Unfortunately, the current human study methods have
`many limitations that may seriously affect the conclusions.
`Ò
`2 CGM system was used with a second-generation
`The iPro
`Ò
`glucose sensor and CareLink iPro software (Med-
`Enlite
`tronic MiniMed, Northridge, CA) to determine MARD for
`the 7-day study.33,34 This system records sensor signals for
`6 days and requires a retrospective calibration using two
`reference BG measurements at start up and at least three
`reference BG measurements per 24 h, no longer than 12 h
`apart.35 The number of BG measurements used to calculate
`MARD was limited to one to four per day because the 12 study
`subjects self-monitored their blood glucose (SMBG) using a
`commercial glucose meter and test strips only 5.7 – 1.6 times
`per day.
`In addition, the methods do not describe which SMBG
`measurements were used for CGM calibration or used for
`calculating MARD. SMBG measurements used for calibra-
`tion should be obtained during a period of glucose stability to
`minimize calibration error.34,36 Since five of the study sub-
`jects were not diabetic and seven of the subjects had type 2
`diabetes mellitus (T2DM), the overall mean BG level was
`113 – 17 mg/dL with only an 87.6–158 mg/dL range. Dia-
`betes is known to affect the immune response to tissue injury
`and wound repair. The results do not report whether diabetes
`status affected per day MARD or histology results, and
`whether BG range or rate of change during SMBG mea-
`surements used for calibration affected MARD. A quality
`CGM accuracy study requires a much larger number of BG
`measurements per day using a reference analyzer over a
`wider range of BG values and rates of change.4,5,8,34,37–39
`Of note, the number of macrophages and tissue histology
`results obtained on day 7 were correlated with MARD results
`from CGM data recorded for the entire 7-day period. This
`correlation should have been limited to CGM data recorded
`from day 7 only or perhaps days 6 and 7 due to the dynamically
`changing environment surrounding the CGM electrodes. There
`may have been few macrophages located adjacent to the
`electrode for several days after implantation of some sensors.
`In addition, the histology analysis should have quantified the
`thickness, density, and continuity of the surrounding layer of
`inflammatory/fibrous tissue to determine whether these me-
`chanical parameters correlated with MARD.
`In conclusion, Rigla et al. successfully demonstrated a
`significant correlation between the number of macrophages
`surrounding the CGM electrodes implanted in humans for
`7 days and overall glucose sensor accuracy. These data
`support the importance of the local tissue environment im-
`mediately adjacent to the CGM electrodes on glucose sen-
`sor performance, especially the location and number of
`
`ABBOTT Exhibit 2044
`Dexcom v. Abbott Diabetes Care, IPR2023-01409
`
`

`

`GLUCOSE SENSING IN THE SUBCUTANEOUS TISSUE
`
`323
`
`metabolically active macrophages. The authors are encour-
`aged to continue this important area of research in patients
`with type 1 diabetes mellitus (T1DM) and T2DM using real-
`time CGM systems over a wide range of glucose values,
`understanding the difficulty in recruiting patients scheduled
`to undergo abdominoplasty surgery. Study of the subcuta-
`neous tissue’s acute inflammatory reaction to CGM implan-
`tation over a 14-day period would facilitate the development
`of next-generation glucose sensors with enhanced accuracy
`and longevity.
`
`Author Disclosure Statement
`
`J.IJ. is a cofounder and equity owner of Capillary Biomedical,
`Inc., a company dedicated to developing glucose sensor and
`insulin delivery systems for people with diabetes mellitus. He is
`also the Chairman of Capillary Biomedical’s Scientific and
`Clinical Advisory Boards. J.IJ. is also an equity owner and
`member of the Scientific Advisory Board of Thermalin Diabetes,
`Inc., a company dedicated to developing novel insulin formu-
`lations for people with diabetes. M.C.T. is an equity owner and
`member of the Scientific Advisory Board of Capillary Biome-
`dical. Research at the Jefferson Artificial Pancreas Center related
`to insulin delivery has been funded by NIH-NIDDK, The Ju-
`venile Diabetes Research Foundation (JDRF), Frederick Bant-
`ing Foundation, Capillary Biomedical, and Thermalin Diabetes.
`G.E., D.D., A.K., and C.L. have nothing to disclose.
`
`References
`
`1. Thabit H, Tauschmann M, Allen JM, et al.: Home use of an
`artificial beta cell in type 1 diabetes. N Engl J Med 2015;
`373:2129–2140.
`2. Cobelli C, Renard E, Kovatchev B: Artificial pancreas:
`past, present, future. Diabetes 2011;60:2672–2682.
`3. Garg SK, Weinzimer SA, Tamborlane WV, et al.: Glucose
`outcomes with the in-home use of a hybrid closed-loop
`insulin delivery system in adolescents and adults with type
`1 diabetes. Diabetes Technol Ther 2017;19:155–163.
`4. Andelin M, Kropff J, Matuleviviene V, et al.: Assessing the
`accuracy of continuous glucose monitoring (CGM) cali-
`brated with capillary values using capillary or venous glu-
`cose levels as a reference. J Diabetes Sci Technol 2016;10:
`876–884.
`5. Kovatchev BP, Patek SD, Ortiz EA, Breton MD: Assessing
`sensor accuracy for non-adjunct use of continuous glucose
`monitoring. Diabetes Technol Ther 2015;17:177–186.
`6. Garcia A, Rack-Gomer AL, Bhavaraju NC, et al.: an ad-
`vanced continuous glucose monitor for the artificial pan-
`creas. J Diabetes Sci Technol 2013;7:1436–1445.
`7. Christiansen M, Bailey T, Watkins E, et al.: A new-
`generation continuous glucose monitoring system: improved
`accuracy and reliability compared with a previous-generation
`system. Diabetes Technol. Ther 2013;15:881–888.
`8. Bailey TS, Chang A, Christiansen M: Clinical accuracy of a
`continuous glucose monitoring system with an advanced
`algorithm. J Diabetes Sci Technol 2015;9:209–214.
`9. Wang Y, Vaddiraju S, Gu B, et al.: Foreign body reaction
`to implantable biosensors: effects of tissue trauma and
`implant size. J Diabetes Sci Technol 2015;9:966–977.
`10. Wisniewski N, Klitzman B, Miller B, Reichert WM: De-
`creased analyte transport through implanted membranes:
`differentiation of biofouling from tissue effects. J Biomed
`Mater Res 2001;57:513–521.
`
`11. Joseph JI, Torjman MC: Implantable Glucose Sensors,
`Encyclopedia of Biomaterial and Biomedical Engineering,
`2nd Ed., Vol 2; edited by G. Wnek and G. Bowlino Informa
`Healthcare Inc.; 2008; pp 1174–1181.
`12. Klueh U: Analysis: on the path to overcoming glucose-
`sensor-induced reactions. J Diabetes Sci Technol 2013;7:
`452–454.
`13. Klueh U, Liu Z, Feldman, B, et al.: Metabolic biofouling of
`glucose sensors in vivo: role of tissue microhemorrhages. J
`Diabetes Sci Technol 2011;5:583–595.
`14. Kvist PH, Iburg T, Bielecki M, et al.: Biocompatibility of
`electrochemical glucose sensors implanted in the subcutis
`of pigs. Diabetes Technol Ther 2006;8:463–475.
`15. Anderson J, Rodriguez A, Chang D: Foreign body reaction
`to biomaterials. Semin Immunol 2008;20:86–100.
`16. Dungel P, Long N, Yu B, et al.: Study of the effects of
`tissue reactions on the function of implanted glucose sen-
`sors. J Biomed Mater Res 2007;85A:699–706.
`17. Ward KW: A review of the foreign-body response to
`subcutaneously-implanted devices: the role of macrophages
`and cytokines in biofouling and fibrosis. J Diabetes Sci
`Technol 2008;2:768–777.
`18. Helton K, Ratner B, Wisniewski N: Biomechanics of the sensor
`tissue interface-effects of motion, pressure, and design on
`sensor performance and the foreign body response-part I: the-
`oretical framework. J Diabetes Sci Technol 2011;5:632–646.
`19. Sharkawy AA, Klitzman B, Truskey GA, Reichert WM:
`Engineering the tissue which encapsulates subcutaneous
`implants. I. Diffusion properties. J Biomed Mater Res 1997;
`37:401–412.
`20. Sharkawy AA, Klitzman B, Truskey GA, Reichert WM:
`Engineering the tissue which encapsulates subcutaneous
`implants. II. Plasma-tissue exchange properties. J Biomed
`Mater Res 1998;40:586–597.
`21. Sharkawy AA, Klitzman B, Truskey GA, Reichert WM:
`Engineering the tissue which encapsulates subcutaneous
`implants. III. Effective tissue response times. J Biomed
`Mater Res 1998;40:598–605.
`22. Woodward SC: How fibroblasts and giant cells encapsulate
`implants: considerations in design of glucose sensors.
`Diabetes Care 1982;5:278–280.
`23. Gerritsen M, Jansen JA, Kros A, et al.: Influence of inflam-
`matory cells and serum on the performance of implantable
`glucose sensors. J Biomed Mater Res 2001;54:69–75.
`24. Klueh U, Czajkowski C, Ludzinska I, et al.: Impact of
`CCL2 and CCR2 chemokine/receptor deficiencies on
`macrophage recruitment and continuous glucose monitor-
`ing in vivo. Biosens Bioelectron 2016;86:262–269.
`25. Novak MT, Reichert WM: Modeling the physiological
`factors affecting glucose sensor function in vivo. J Diabetes
`Sci Technol 2015;9:993–998.
`26. Kamath A, Mahalingam A, Brauker J: Analysis of time lags
`and other sources of error of the DexCom SEVEN continuous
`glucose monitor. Diabetes Technol Ther 2009;11:689–695.
`27. Hunt TK: The physiology of wound healing. Ann Emerg
`Med 1988;17:1265–1273.
`28. Klueh U, Frailey JT, Qiao Y, et al.: Cell based metabolic
`barriers to glucose diffusion: macrophages and continuous
`glucose monitoring. Biomaterials 2014;35:3145–3153.
`29. Brauker JH, Carr-Brendel VE, Martinson LA, et al.: Neo-
`vascularization of synthetic membranes directed by membrane
`microarchitecture. J Biomed Mater Res 1995;29:1517–1524.
`30. Klueh U, Antar O, Qiao Y, Kreutzer DL: Role of vascular
`networks in extending glucose sensor function: impact of
`
`ABBOTT Exhibit 2044
`Dexcom v. Abbott Diabetes Care, IPR2023-01409
`
`

`

`324
`
`JOSEPH ET AL.
`
`angiogenesis and lymphangiogenesis on continuous glu-
`cose monitoring in vivo. J Biomed Mater Res A 2014;102:
`3512–3522.
`31. King C, Anderson SM, Breton M, et al.: Modeling of cali-
`bration effectiveness and blood-to-interstitial glucose dy-
`namics as potential confounders of the accuracy of continuous
`glucose sensors during hyperinsulinemic clamp. J Diabetes
`Sci Technol 2007;1:317–322.
`32. Rigla M, Pons B, Rebasa P, et al.: Human subcutaneous
`tissue response to glucose sensors: macrophages accumu-
`lation impact on sensor accuracy. Diabetes Technol Ther
`2018;20:296–302.
`33. Bailey TS, Ahmann A, Brazg R, et al.: Accuracy and ac-
`ceptability of the 6-day Enlite continuous subcutaneous
`glucose sensor. Diabetes Technol Ther 2014;16:277–283.
`34. Biagi L, Ramkissoon MC, Facchinetti A, et al.: Modeling
`the error of the Medtronic Paradigm Veo Enlite glucose
`sensor. Sensors 2017;17:pii: E1361.
`35. Medtronic: Carelink iPro: Therapy Management Software
`for Diabetes User Guide. Medtronic MiniMed 2016. www
`.medtronicdiabetes.com (accessed March 30, 2018).
`36. Facchinetti A, Del Favero S, Sparacino G, et al.: Modeling
`the glucose sensor error. IEEE Trans Biomed Eng 2014;61:
`620–629.
`
`37. Kirchsteiger H, Heinemann L, Freckmann G, et al.: Per-
`formance comparison of CGM systems: MARD values are
`not always a reliable indicator of CGM system accuracy. J
`Diabetes Sci Technol 2015;9:1030–1040.
`38. Reiterer F, Polterauer P, Schoemaker M, et al.: Significance
`and reliability of MARD for the accuracy of CGM systems.
`J Diabetes Sci Technol 2017;11:59–67.
`39. Kollman C, Wilson DM, Wysocki T, et al.: Limitations of
`statistical measures of error in assessing the accuracy of
`continuous glucose sensors. Diabetes Technol Ther 2005;7:
`665–672; discussion 673–674.
`
`Address correspondence to:
`Jeffrey I Joseph, DO
`Department of Anesthesiology
`Jefferson Artificial Pancreas Center
`Sidney Kimmel Medical College
`Thomas Jefferson University
`Jefferson Alumni Hall # 565
`1020 Locust Street
`Philadelphia, PA 19107
`
`E-mail: jeffrey.joseph@jefferson.edu
`
`ABBOTT Exhibit 2044
`Dexcom v. Abbott Diabetes Care, IPR2023-01409
`
`