Analysis of “Randomized Cross-Over Study Comparing Two Infusion Sets for CSII in Daily Life”

Abstract

In an article in Journal of Diabetes Science and Technology, Freckmann et al report an evaluation of 2 infusion sets used for CSII in daily life, both from Roche. Because the 2 catheters differed at the soft cannula level in terms of geometry and introducer needle, focus was made on pain at insertion. While no significant difference was noted on this primary endpoint, unplanned catheter changes occurred in close to 20% of cases also with both catheter models, mainly driven by apparent insulin delivery issues. Since kinked cannulas were not frequent, there is a clear need for better understanding of “unexplained hyperglycemia” as the main reason for reduction of catheter lifetime. Fixing this weak aspect of CSII becomes crucial while moving toward closed-loop insulin delivery.

Treatment of type 1 diabetes by insulin pumps has been documented as associated with multiple benefits, including improved rate of reaching HbA1c target, reduced occurrence of severe hypoglycemia, increased flexibility of insulin delivery allowing better quality of life.^1,2 Although not allowing the most physiological way of insulin delivery according to metabolic needs, subcutaneous infusion has been widely adopted for thanks of simplicity.³ Nevertheless, the increased risk of lack of insulin leading to quicker ketoacidosis in case of delivery issue has been identified quite early in the implementation of CSII use.⁴ This has been further highlighted in the Diabetes Control and Complications Trial during which patients treated by CSII in the intensive diabetes management group presented a significantly higher occurrence of ketoacidosis events.⁵ Patient training for quick identification of insulin delivery issues while under CSII, including those related to catheter failures, has been underscored as an essential component of education to CSII use.⁶

In contrast with this early identified risk associated with CSII, very limited research and development has been dedicated to the improvement of infusion sets during the last decades.⁷ More attention has been paid to the improved reliability of insulin pumps following reports of frequent pump failures which might have led to their market withdrawal.⁸ Recent surveys pointed to the crucial role of infusion sets in problems of under-delivery met while using insulin pumps.⁹ This prompted analyses of the published reports on this topic and reviews of the current practices on the management of infusion sets in various countries, followed by recommendations for the choice and the education on use of infusion sets and for dedicated research aiming at a better knowledge of the root causes and the epidemiology of catheter issues and complications.¹⁰ Scientific assessment of insulin infusion sets has been scarce during the past decade, and the reported trial by Freckmann et al¹¹ illustrates the proper way for getting valuable data thanks to an appropriate methodology.

Because pain at catheter insertion is a perceived reason for patient adopting or declining CSII use, efforts have been made to ease the insertion procedure through the development of cannula inserters and/or the improvement of cannula geometry and introducer needle in view of minimizing pain, as assessed as primary endpoint in the study reported by Freckmann et al.¹¹ Study results show no significant difference on pain at insertion with the 2 assessed infusion sets, but included patients seemed to have no baseline complaints on this aspect which could minimize the detection of benefit related to the FlexLink Plus compared to the FlexLink (Roche, Mannheim, Germany). Selecting patients with reported pain at insertion could have resulted in showing differences.

Interestingly, this study provides valuable data on the occurrence of unplanned catheter changes since the protocol requested a thorough collection of information in a diary for each catheter replacement. Most unplanned changes did not occur early after catheter insertion which is in agreement with a previously reported study we conducted with FlexLink that showed a dramatic reduction of initial failures when using this infusion set model compared to others: 3% versus 9%.¹² Unplanned changes occurred on average between 2 and 3 days after insertion in 16 % of FlexLink catheters and 20% of FlexLink Plus catheters. This occurrence is high when compared to that of around 10% previously reported in our FlexLink study.¹² The fact that most patients included in the present study were experienced pump users compared to those included in our study (mean experience of CSII of 12 vs 4 years) may have contributed to make them more reactive to catheter change in case of persistent hyperglycemia. The main reasons for unplanned changes were kinked catheters and occlusion alarms. However, cannula kinking was absent with any of the 2 assessed infusion sets in 81.3% of patients in contrast with a previous report by Pickup et al where kinking was reported by 64.1% of patients, including 12% with frequent kinking,¹³ but most of them used other catheter models. Resistance to cannula kinking may then be a positive feature of the FlexLink/FlexLink Plus cannulas or their mode of insertion. Hence, cannula occlusion appears as the suspected main cause of premature catheter change.

Cannula occlusion remains a difficult diagnosis to establish since uncommonly documented by a visualized obstacle at catheter removal. Pump occlusion alarms and more often unexplained hyperglycemia which cannot be solved by insulin bolus usually lead to the suspected diagnosis of cannula occlusion. Among factors promoting occlusions, duration of use longer than 3 days has been identified by Pickup et al.¹³ Neither using soft cannulas versus steel ones nor the nature of fast-acting analogue appeared as significant factors in their analysis. Patel et al more recently pointed to individual factors as involved in catheter lifetime.¹⁴ Among these factors, the presence of subcutaneous lipohypertrophy at infusion sites has been well identified as a frequent cause of impaired insulin infusion.¹³ Whether individual impaired tolerance to the materials of cannulas or to infused insulin or both could play a role as a determining cause of cannula occlusion has not been investigated thoroughly so far. From the reports on investigations dedicated to occluded intraperitoneal catheters used for insulin delivery, these hypotheses can be relevant. Indeed, although peritoneal space is somewhat different from subcutaneous tissue, granulomas including macrophages and giant cells as well as amyloid deposits stained by anti-insulin antibodies have been identified around occluded peritoneal catheters.¹⁵ Such inflammatory reactions could also be involved in the occlusion of SC cannulas through a promotion of fibrin and/or insulin aggregates.

In conclusion Freckmann et al’s article points to several interesting matters. First of all, clinical investigations of infusion sets according to a scientific methodology are needed and should be developed, at best before launch on the market. Second, these investigations can provide valuable characterization of studied infusion sets that will help in the choice of infusion sets for individual cases as recommended by recent position statements.¹⁰ Moreover, this study highlights the need for more research on the root causes of catheter occlusions. When taking into account the deleterious outcomes of catheter occlusions, the cost associated with more frequent catheter changes and the need for a high reliability of insulin infusion in forthcoming artificial pancreas to be used in real life, a call for more attention dedicated to the improvement of infusion sets appears obvious.