Insulin depot formation in subcutaneoue tissue

Abstract

Background: The size and geometry of an insulin depot that is formed during subcutaneous administration by an insulin pump is evaluated. A novel method is used to visualize accurately the depot formation for small volumes of insulin (of the order of 10-100 µl) at a given point in time. Conventional visualization methods such as magnetic resonance imaging are unable to provide such accurate measurements because of their coarse imaging resolution and long measurement time.

Methods: The described method consists of subcutaneously infusing dyed insulin into porcine tissue and subsequently shock freezing it with liquid nitrogen. The frozen sample is then sliced into thin layers using a cryomicrotome. A digital image of each layer is taken and then processed with proprietary software, which identifies the dyed areas on each layer and reconstructs a three-dimensional model of the insulin depot with a planar resolution of 30 × 30 µm(2) and a depth resolution of 100 µm. Since this process is not viable for living organisms, porcine tissue was used immediately following slaughter of the animal.

Results: To date, it is most often assumed that the insulin depot takes the shape of a sphere around the tip of the cannula (e.g., 50 µl insulin equates to a spherical radius of 2.3 mm). However, in practice, such a depot form is never observed. Instead, the insulin depot initially spreads laterally (i.e., parallel) to the skin surface and in the collagen matrix that binds the adipose cells together. The depot outreach increases with larger infused volumes, e.g., maximum outreach measured at 5.0/5.7/7.1 mm (quartiles, n = 17) for 50 µl of infused insulin. Beyond a given infused volume (approximately 100 µl), the insulin also starts to spread perpendicular to the skin surface.

Conclusions: It is concluded that formation of the insulin depot depends on the opening of channels at the boundaries between adipose cells. Hence the insulin follows a path of least resistance and depot formation is determined by the local structure of the subcutaneous tissue.

Figure 1

Dyed insulin infused by (A) an insulin pump through (B) an infusion set into (C) a porcine tissue sample. After infusion, the tissue sample is shock frozen and sliced (20 µm thick). The top surface of the sample containing dyed areas was (D) digitally photographed and then (E) the images were processed to reconstruct the 3D model of the insulin depot.

Figure 2

Superimposed projections of top, front, and side view of 50 and 150 µl (normal and massaged) depots (n = 6). Cannula length is 8 mm. Length A defines the maximum extension of the insulin parallel to the skin surface, length B is perpendicular to skin surface up along the needle tip.

Figure 3

Infused dyed insulin forming channels along the boundaries of adipocyte cells (approximately 80 µm diameter). (A) Channel branch around adipose cell clump (channel’s edge indicated with red dots), and (B) adipose cells separated from each other by the dyed insulin.

Figure 4

Reconstructed 3D depots of (A)–(C) 50 µl and (D)–(F) 150 µl. Cannula length is 8 mm.

Figure 5

Quartiles of reconstructed amount of dyed volume per height for n samples: (A) 20 µl (n = 6); (B) 50 µl (n = 17); (C) 100 µl (n = 7); and (D) 150 µl (n = 6).

Figure 6

Infused volume versus reconstructed volume resulting with set color segmentation parameters.

Figure 7

Quartiles of detected outreach of dyed area per height for n samples: (A) 20 µl (n = 6); (B) 50 µl (n = 17); (C) 100 µl (n = 7); and (D) 150 µl (n = 6).

Figure 8

Comparison of medians for reconstructed amount of dyed volume per height for different infusion volumes.