Performance Qualification of a Novel Subcutaneous Insulin Infusion Set using Medical Imaging Natasha G. Bolick, MS1, Diane Sutter, cCRA1, Brian Pflug, PharmD2, Laurence Hirsch, MD2, Bruce Bode, MD3, Ronald J. Pettis, PhD1

1BD Technologies, RTP, NC, USA; 2BD, Franklin Lakes, NJ, USA; 3Atlanta Diabetes Assoc., Atlanta, GA

Background

Results and Conclusions

Preclinical Swine Fluoroscopic Imaging

Clinical Human Magnetic Resonance Imaging

References

Acknowledgements

The authors wish to thank E. Serfer, R. Carrero, M. Zhuk, and C. Woodhouse of Elite Imaging, Aventura FL, for their considerable expertise, assistance, and collaboration in the preparation and execution of this imaging study; the BD FlowSmartTM project team members for their design, development & manufacture of the BD FlowSmartTM device; the BD Technologies Parenteral Sciences department for their development & conduct of preclinical imaging studies; and BDT/CPDT vivarium staff for their assistance in conducting studies.

1.Heinemann, L. et al, 2012. Insulin infusion set: The Achilles Heel of CSII. J. Diab Sci & Tech, 6: 954-964.

2.Ponder, S. et al, 2008. Unexplained hyperglycemia in CSII. Diabetes Educator, 34: 327-333.

3.Gibney et al, 2015. CSII Sets – Reduced Flow Interruptions with a Novel Investigational Catheter Infusion Set. Diab Tech Ther, 17(S1):A-8.

4.Bolick, N.G. et al, 2015. Reduction of silent occlusion occurrence during continuous subcutaneous insulin infusion. Diab Tech Ther, 17(S1):A-35.

• An investigational flow-stabilizing ported set for reducing infusion pressure3,4 and other marketed SC infusion sets were evaluated using multiple medical imaging modalities.

• Catheter porting creates an auxiliary tissue flow pathway that may reduce occlusion and stabilize flow.

• In vivo swine fluoroscopy of contrast media infusion provided good high resolution visualization of device placement, bolus depot patterning, and delivery faults such as incomplete insertion and leakage. Contrast-filled polymer catheters were visible in situ.

• Diffuse SC depot patterns characterized by more depot layers were observed for the novel 28G, 6mm, ported polymer catheter set, potentially owing to side-port delivery.

• A human MRI study of the investigational set in 8 non-diabetic adults demonstrated effective SC placement and placebo delivery across multiple bolus volumes (1-30U) and various routine insulin infusion sites (arm, thigh, abdomen, and gluteus).

• Clinical depots were similar to preclinical observations with increased diffusivity in contrast to denser depots observed from marketed SC sets (Figure 4). Flow from both terminal outlet and side port appeared to be visible at lower delivery volumes, but side port visualization was obscured at larger bolus volumes (Figure 5).

• Variations in delivery depth owing to differences in catheter length were readily detected (Figure 7), as was catheter bending owing to intentional poor insertion technique (Figure 6); however this did not adversely impact delivery.

• Both fluoroscopic and MR imaging techniques provide effective CSII set characterization at relevant anatomical sites and may enable better comparative device performance assessment.

Subcutaneous (SC) insulin infusion sets have been characterized as the Achilles

heel of insulin pump therapy1 and are associated with hyperglycemia2, insulin

leakage, pump occlusion alarms, and “silent” occlusions where insulin flow is

impeded but no alarm occurs. An investigational continuous subcutaneous

insulin infusion (CSII) set (BD FlowSmart™, BD, Franklin Lakes, NJ) with a

novel side-ported 28G, 6mm polymer catheter designed to potentially reduce

flow interruptions and stabilize insulin delivery is under development.3,4

Evaluating CSII set functional performance is challenging due to the slow

response time and intrinsic variability of blood glucose (BG) responses, inability

to quantify delivered insulin in real time, and impracticality of site observation

with sets in place. Various medical imaging techniques allow visualization of

sets and delivery in situ to examine device placement, depot patterns, and

incomplete insertion or leakage; however imaging of CSII sets has not been

previously performed.

In the current studies, preclinical fluoroscopic imaging and clinical magnetic

resonance (MR) imaging were used to assess system performance of the

investigational set and other commercial CSII sets in animal and human studies.

Fluoroscopy Materials and Methods:

• A Glenbrook Technologies Labscope™ fluoroscopy unit was used to image

CSII sets (FS= BD FlowSmart™, BD; QS= Quick-set® 6mm, Medtronic,

Northridge, CA) in anesthetized Yorkshire swine (~25-40 kg weight range)

(Figure 1).

• Animals were positioned with the skin surface and CSII set perpendicular to

the radiation source for imaging. Skin surface at the set insertion site was

highlighted with radiopaque metallic paint to enhance tissue/device contrast.

• Infusions (100 uL) of radiopaque low viscosity (2.3 cP at 25C) iodinated

contrast media (NOVAPLUS® Omnipaque™ 140, GE Healthcare) were

infused via commercial insulin pumps to mimic 10U bolus insulin delivery.

• Still images were captured pre and post-infusion, with continuous dynamic

video capture during the infusion process. Image analysis included tissue

characteristics (dermal layer thickness, deformation at infusion site), device

function (ability to insert, performance, occlusion, leakage), and infusate

deposition characteristics (tissue location, size, depth).

Figure 1. In vivo swine fluoroscopy using iodinated contrast

media infusion A) In vivo swine fluoroscopy using Iohexol

contrast media infusion.

B) Insertion and (C) 10U depot delivery for BD FlowSmart™

(FS) investigational catheter showing diffuse multi-depot pattern

D) Insertion and (E) 10U depot delivery for Quick-set® (QS) 6mm

catheter showing singular depot. Note some QS devices also

demonstrate diffuse patterns owing to diffusion through tissue

strata.

F) Histogram showing increased tendency for multi-depot

patterns in the FS investigational set

Perc

en

t (%

)

MR Materials and Methods:

• An observational IRB-approved imaging study of SC placebo infusion across various CSII sets and dosing conditions was performed in N=8 healthy adult subjects (age 34.5±13.1 YO; n=3 Male/5 Female; BMI 26.2±4.0).

• Imaging was conducted using a Siemens Magnetom Trio 3 Tesla MRI unit (Figure 2), using T1 weighted Gradient Echo, T1 Flash w/ & w/out Fat Suppression sequences in the transverse plane. Anatomical sites for device insertion included the abdomen, thigh, posterior upper arm and gluteal regions (Figure 3). Fish oil caplets were used to mark areas lateral to device placements and register sites within the imaging field. External surface imaging coils were used to capture images of the upper arm.

• Placebo infusion solutions included dilute gadolinium contrast agent in saline (1:300 v/v Multihance®, Bracco) and 0.9% saline alone. Saline alone showed no degradation of image quality and was used for the majority of infusions.

• CSII sets were placed both manually and with an inserter device (Quick-serter®, Medtronic). Some manual placements also included intentionally poor technique (slow and/or angular insertion) to force cannula faults. Side ports of BD FlowSmart™ devices were oriented parallel to the image slice in order to see diffusion across the tissue planes. Both 6mm and 9mm commercial polymer sets (Medtronic Quick-set®) were used as comparators. (Figure 4-7)

• Animas® OneTouch ® Ping™ insulin pumps were used for rapid bolus delivery to minimize depot diffusion during the imaging process. Catheters sets were serially connected with microbore extension sets to keep pumps outside the MR magnetic field. Images were sequentially acquired before and after cannula fill, and at incremental stepwise volume increases up to 30U bolus.

Figure 6. Good vs. Poor Insertion. (A) A 10U bolus after insertion of the FS investigational catheter with a mechanical inserter in the left arm showing good perpendicularity to the skin surface. (B) FS in the left thigh of the subject after an intentionally poor manual insertion applied off-angle to the skin surface. Arrows are parallel to catheter direction. The bolus successfully delivered despite the obvious bent catheter under the skin.

Figure 7. Importance of device length. A) Variation in delivery depth of FS, QS 6mm, QS 9mm after a 20U bolus infusion in the right thigh. All sites remain in SC adipose layer, but variations in depth of depot location are clearly visible. B) QS 9mm, 10U bolus in the thigh showing deposition almost into the muscle layer.

Figure 4. (L to R) Sequential infusions including Cannula Fill (CF) and 5/10/30U total bolus from manually inserted QS 6mm in the right posterior thigh (upper row) and a FS in the gluteal region (lower row). Catheter is barely visible in tissue during catheter filling/priming. Subsequent depots exhibit increased size and density with increasing bolus volumes up to 30U.

Figure 2. Adjustment of the arm surface imaging coil in preparation for imaging. The insulin pump (visible on far right) is outside the high gauss magnetic field and connected by serial microbore set extensions to the infusion set.

Figure 3. Representative trans- verse images of standard infusion locations including A) upper arm, B) abdomen and posterior upper gluteal region, and C) thigh. Registration markers are visible as circles on the skin surface (arrows), and relevant anatomical markers are labeled.

Helping all people Live healthy lives

Figure 5. Human MR transverse

images of SC infusion from 6mm

CSII sets (L to R: FS/QS/FS) in the

upper gluteal region at increasing

bolus volumes (CF, 2, 5, 10, 20U).

Arrow highlights potential side-port

delivery on FS device at lower

volumes; blue circles highlight

depots at 2-20U. FS devices show

increased diffusion.

Upper Arm Abdomen

Thigh