Analysis of “A New Optimized Percutaneous Access System for CIPII”

Abstract

While only used initially in cases with resistance to subcutaneous insulin therapy, intraperitoneal insulin therapy provides an overall more stable glucose control than subcutaneous insulin therapy thanks to its pharmacokinetics as pointed by Garcia-Verdugo et al from the experience of implantable insulin pumps. The expansion of these devices has been limited by underdelivery issues and high cost. The availability of a new percutaneous access to intraperitoneal route could allow a similar glucose control with less constraints of follow-up and expected lower cost. Currently reported clinical experience does not allow a reliable assessment of its main risk of infection which could impair its sustained usability. Because intraperitoneal insulin could allow a fully automated closed-loop insulin delivery, a specific interest for its means of performance is relevant.

Insulin therapy has been among the highest achievements in medicine in the twentieth century. After it was understood insulin could not be given orally due its destruction by gut enzymes, the easiest parenteral mode to deliver insulin was through subcutaneous (SC) insulin injections. In some people, local depots of SC insulin resulted in fibro-inflammatory reactions and immunogenic consequences such as anti-insulin antibody production. ¹ Local reactions against SC insulin depots could include insulin destruction by proteases in uncommon cases. ² Patients affected by this so-called ‘subcutaneous insulin resistance’ could only have their diabetes controlled by intravenous (IV) insulin infusion although the evidence for a defect of SC insulin action remained controversial. ³ Because intraperitoneal (IP) insulin delivery had been used in diabetes patients with renal failure treated by peritoneal dialysis and shown to be effective and convenient in these subjects, first cases of IP insulin delivery from portable pumps and via a catheter that was indwelled into the peritoneal cavity through the abdominal wall were reported in the early 1980s and indicated effectiveness of this therapeutic mode in brittle diabetes after failures to achieve glucose control with SC, intramuscular and sometimes IV insulin. ⁴

Meanwhile, studies were performed in animal models to assess the characteristics of peritoneal insulin absorption. ⁵ Interestingly, insulin was first distributed to the portal vein, before entering the systemic circulation. A positive gradient between portal and systemic blood insulin was further demonstrated that differed from intramuscular insulin injection but mimicked endogenous insulin secretion. ⁶ In humans, IP insulin bolus delivery was shown to allow quicker insulin peaks than SC route: 70 versus 120 min, and plasma insulin levels returned to baseline after 165 min as in normal subjects. ⁷ While the rate of systemic appearance of IP infused insulin was lower than SC infused insulin at steady-state, the percentage increase in the rate of systemic appearance after an increase of infusion rate was higher with IP route. ⁸ As a result, a lower peripheral insulinemia, close to physiological levels, was obtained in steady-state conditions with IP route than with SC insulin delivery. In spite of these demonstrations showing how IP insulin delivery was better mimicking physiology than SC insulin, IP insulin therapy remained anecdotal.

The Diabetes Control and Complications Trial (DCCT) clearly showed that near-normoglycemia had to be the treatment goal of type 1 diabetes mellitus to reduce the incidence of retinal, renal and peripheral neuropathic diabetic complications. ⁹ Although intensive subcutaneous (SC) insulin treatment using multiple daily injections (MDI) or continuous SC insulin infusion (CSII) could lead to sustained HbA1c levels close to 7% and effective prevention of diabetic complications, the high incidence of severe hypoglycemia appeared as a significant side effect that limited the benefit of this route of insulin delivery. Poorly reproducible and predictable insulin effects with SC route are the main cause of the variability of blood glucose levels, and hence of hypoglycemic events. ¹⁰

The currently available insulin analogues do improve the benefit/risk ratio of SC insulin by allowing at least similar HbA1c levels as regular and NPH insulins but with a lowered risk of hypoglycemia. ¹¹ Moreover, fast-acting analogues enhance the efficacy of CSII. ¹² CSII combination with continuous glucose monitoring (CGM) may further improve glucose control by allowing early patient reactions to glucose changes. Nevertheless, closed-loop trials based on SC insulin delivery have shown the need for hybrid use, that is, premeal patient-activated bolus according to carbohydrate intakes, due to the still slow and delayed insulin action when increased insulin delivery is ordered by the algorithm according to the CGM signal. ¹³ Fully automated closed-loop insulin delivery with SC infusion of fast-acting analogues is currently not feasible except at the cost of postmeal highs and secondary lows.

Garcia-Verdugo et al have nicely reviewed and summarized in their paper the reported clinical benefits of IP insulin delivery compared to SC insulin both on average glucose levels and glucose excursions including hypoglycemia. ¹⁴ Further interesting hormonal changes on the GH/IGF-1 axis and better quality of life have also been reported while using IP insulin. The only detrimental outcome of IP insulin that may occur in a subset of patients is an increased production of anti-insulin antibodies that may become deleterious when these antibodies have low/medium affinity for insulin. ¹⁵ In such cases, insulin binding to antibodies when insulin delivery increases after meals and insulin release at nighttime when insulin infusion decreases may indeed induce unpredictable highs and lows.

After underscoring the reported advantages of IP insulin, the authors address the key-question about the means of IP infusion. From the 1990s, most of IP insulin experience has come from the use of implantable insulin pumps. ¹⁶ While these devices were safe and effective on glucose control, their expansion has been mainly limited by insulin underdelivery issues. The most worrying causes are IP catheter occlusions because they need surgical reinterventions for catheter replacement. The most frequent reason for underdelivery is related to the formation of insulin aggregates in the pumping mechanism which induce gradual slowdown of insulin infusion due to backflows of insulin toward the pump reservoir. This adverse event can be fixed by a rinse procedure of the pump by NaOH solution and is reduced by limiting the time during which insulin spends in the reservoir, that is, 6 weeks instead of the initially scheduled 3 months. Moreover, systematic pump rinses with NaOH every 6-9 months further minimize underdelivery issues related to insulin aggregation. Interestingly, more frequent pump refills and rinses also reduce the occurrence of catheter blockages suggesting that insulin aggregates from the pump may promote reactions at the catheter tip. Altogether, implantable pumps using IP route provide more stable and close-to-normal glucose control than SC insulin at the cost of time-consuming burdens for the medical team. Hence implantable pumps have been approved for diabetes therapy only in the European Union and used so far in a few countries (France, the Netherlands, Sweden, and Belgium) whereas they never got approval for therapy in United States. ¹⁷ The limited market for these devices as well as for the specific U400 insulin formulation used in them resulted in very high costs: around €35 000 for a 7- to 9-year usable pump and €500 to €750 for 10 to 15 ml insulin per pump refill.

Hence, Garcia-Verdugo et al present the potential interest of a new generation of percutaneous access to the peritoneal route. While it also needs surgical implantation, the management of IP insulin is easier since through a connected infusion set from a CSII pump. Patient self-manages the system after education and follow-up is close to that of CSII-treated patients. Reported data show similar efficacy on glucose control as implantable pumps. ¹⁸ Catheter occlusions may still occur albeit fixed by an ambulatory nonsurgical catheter replacement through the percutaneous access. However, the most common adverse events are cutaneous infections around the percutaneous port. The new design of the percutaneous access to IP route, including a Dacron cuff around the port, which may constitute a barrier to infection and offer a better stabilization of the port at implantation site, could minimize the incidence of serious or persistent infections prompting for system removal. Whether this will be confirmed in clinical use is currently under investigation. Due to this still pending data from the clinical field, some of the conclusions of the authors look premature. Nevertheless, previously reported study data on the use of IP insulin delivery systems as part of an artificial pancreas justifies a specific interest. ¹⁹ Indeed, IP insulin infusion could pave the road toward a fully automated closed-loop system with neither need for meal announcement nor need of glucagon as a rescue option. The pharmacokinetics of IP insulin and the reported restoration of glucagon secretion while under IP insulin ²⁰ could both represent major advantages compared to hybrid single- or dual-hormone AP models developed so far.