Using Insulin Infusion Sets in CSII for Longer Than the Recommended Usage Time Leads to a High Risk for Adverse Events

Abstract

Background:

Infusion sets for use with insulin pumps are recommended for use for 2 to 3 days to avoid local skin reactions, for example, to the insulin formulation and preservatives like meta-cresol. However, many patients use the catheters longer for economic reasons. We performed this study to investigate the tolerability of 2-day use of infusion sets in comparison to 4-day use in a real-world setting.

Methods:

This prospective randomized controlled crossover study with 2 × 3-month observation periods was performed with 24 type 1 patients. At baseline, patients were trained on the use of the infusion system (Medtronic /Mio® or inset™ II) and randomized to any of the 2 treatment sequences. Observation parameters included glycemic control, frequency and nature of device-related, and procedure-related adverse events and patient preference.

Results:

The per-protocol analysis was performed with 22 patients (5 men, 17 women, age 39 ± 11 years, BMI 27.0 ± 3.5 kg/m2). The number of catheter related adverse events was 290 with 2-day use versus 495 with 4-day use (P < .05). The overall number of treatment related events was 750 with 2-day use versus 934 with 4-day use (P < .001). There was no difference in glycemic control between the treatment arms. Treatment satisfaction was higher with 2-day use (very high/high satisfaction: 90.4% versus 4 day-use: 77.3%, P < .05).

Conclusion:

Our results demonstrate that using the infusion sets for a longer usage period of 2-3 days resulted in a clinically relevant increase in treatment-related tolerability problems. Patients should be trained and encouraged not to use insulin pump infusion sets for a longer than the recommended time period.

Modern intensive insulin treatment with continuous subcutaneous insulin infusion (CSII) by means of insulin pump devices is the most optimal method to achieve near-normal blood glucose levels in patients with type 1 diabetes^1,2 and has also been used in patients with type 2 diabetes.^3,4 Despite the economic and technological challenges and the complexity of patient training associated with CSII, it is estimated that more than 500 000 patients worldwide are currently being treated with CSII.⁵ When applying CSII, the insulin is administered via an infusion set, which is composed of a polymer cannula or steel needle located in the subcutaneous tissue, and which is connected to the insulin container in the pump device by means of a plastic tube. To avoid skin irritation, infusion site reactions and other adverse effects, it is recommended that infusion should not occur for longer than 48 to 72 hours at the same tissue location. In clinical studies, it has been shown in different investigations that regular human insulin, insulin glulisine, insulin aspart, and insulin lispro can be safely applied in CSII, when the infusion sets are changed every 2 days.^6-10

In a first pilot study, we investigated the incidence of infusion-set-related events in correlation with the duration of catheter use in 12 patients with type 1 diabetes. While all patients could use the infusion sets without any problem for at least 2 days, multiple problems started to occur on the third day, partly requiring a change of the infusion set. There was no single specific event related to an extended catheter use but events of all kinds and nature occurred after day 3, including but not limited to pulling out, kinking, adhesive getting loose, leakage of insulin from the infusion set or the infusion site, and signals of skin irritation, such as bruising, redness, or swelling. Even when the patients locally tolerated a longer use of the infusion sets at the infusion site for up to 7 days, there appeared to be changes in the insulin absorption leading to a slow but steady loss in glycemic control with increased duration of use, as indicated by a steady increase in mean daily blood glucose concentrations, which is in line with a most recent observational study by Perrin et al.^11,12

In general, potential problems that may occur in adult and adolescent patients when using the infusion sets longer than the recommended 48-72 hours may be technical problems, such as catheter occlusions or kinking,^13,14 bacterial contaminations leading to skin inflammation,^15-17 potential changes in physicochemical delivery characteristics,^18,19 and local skin reactions to the adhesive, resulting in redness, itching, and other symptoms of skin irritation.¹²

The purpose of this study was to evaluate the safety and efficacy of using infusion sets for 2 days versus 4 days during a period of 3 months each in a prospective and randomized crossover protocol in patients with type 1 diabetes.

Patients and Methods

This prospective randomized crossover study was performed in accordance with the ethical and scientific principles governing clinical research as set out in the declaration of Helsinki and the applicable guidelines for good clinical practice (GCP), whichever provides the greater protection of the individual. The study was approved by the responsible ethics committee of the State of Rheinland Pfalz, and was conducted in accordance with all applicable German regulations. Participants signed informed consent prior to any study procedure.

Twenty-four patients were planned to be enrolled into this study protocol. Patients could be included into the study if they were diagnosed with type 1 diabetes, had an age between 18 and 75 years, and had experience with CSII for at least 3 months. Exclusion criteria were significantly raised laboratory safety parameters (2.5 times above the normal reference range of 1 of the following parameters: ALAT, ASAT, γ-GT, leukocytes, erythrocytes, platelets, haematocrit), clinically significant physical findings as well as mental or physical or legal incapacity jeopardizing compliance of patients in the judgment of the investigator, pregnancy, or breast feeding.

The primary study objective was to evaluate the impact of using insulin infusion sets for 2 days versus 4 days on patient safety (adverse events [AEs], catheter-related AEs). Further objectives were to evaluate the influence of the 2 treatment strategies on treatment efficacy (HbA1c, hypoglycemia frequency), treatment satisfaction, and technical treatment aspects (insulin dose, infusion site reactions). HbA1c was measured in a central laboratory.

The design of this prospective randomized controlled crossover study with 2 × 3-month observation periods is shown in Figure 1. At baseline, patients were trained on the use of the respective infusion system for their insulin pump (Medtronic /Mio® or inset™ II, both manufactured by Unomedical A/S, Lejre, Denmark) and randomized to any of the 2 treatment sequences. Observation parameters included glycemic control, frequency, and nature of device-related and procedure-related AEs, and patient preference. The patients received diaries to document their treatment data, blood glucose measurements, dates of infusion set changes, and other events, which were exchanged every 4 weeks during the study.

Figure 1.

Study design.

The patients visited the site every 4 weeks. The investigator inspected all new infusion sites for skin reactions and documented the findings in comparison to the patient reported findings in the patient diaries. For assessment of hypoglycemia (any reading <63 mg/dl or <3.5 mmol/l) and hyperglycemia (any reading >250 mg/dl or >13.8 mmol/l), blood glucose values were downloaded from the patient’s blood glucose meters and also recorded from the patient diaries. All infusion sets were to be returned to the site after use and were inspected by the investigator to identify technical issues, such as crimping/rippling, bending, occlusions, and so on.

The statistical evaluation was performed in accordance to the Good Statistical Practice in Clinical Research guideline for statisticians in the pharmaceutical industry and to the International Conference of Harmonization (ICH) guidelines E3 and E9. The safety information was analyzed with all patients who used at least 1 catheter in the study, and the patients completing the study per protocol were included into the efficacy analysis. Results were evaluated between the groups for mean values and for changes from baseline with descriptive (summary of continuous and categorical variables) and inferential statistical methods (2-sided (un)paired T-test (alternatively Mann-Whitney U-test) or Wilcoxon signed rank test (if a normal distribution was not met). In addition, the chi-square test of homogeneity was applied. A P value < .05 was considered to be statistically significant.

Results

Originally, 29 patients were screened and 25 could be enrolled into the study. Three patients withdrew consent in the first study period resulting in 22 eligible patients for the efficacy analysis (5 men, 17 women, age 39 ± 11 years, BMI: 27.0 ± 3.5 kg/m², 11 patients in each randomization arm). All remaining patients had Caucasian ethnicity and completed the study per protocol. The observation parameters at baseline and after the 2 treatment periods are provided in Table 1.

Table 1.

Patient Characteristics and Key Study Results.

Parameter	Baseline	4-day use	2-day use
HbA1c (%)	7.8 ± 1.4	7.6 ± 1.1	7.4 ± 1.2
Basal insulin dose (U)		20.2 ± 8.3	20.2 ± 8.2
Total daily insulin dose (U)		47.8 ± 19.9	45.5 ± 16.5
Hyperglycemic events (>250 mg/dL)		358	257*
Hypoglycemic events (<70 mg/dL)		450	458
Adverse events (AEs)
Infusion set related AEs		517	305*
Treatment related AEs		467	463
Other AEs		550	324*
Injection site reactions (patients)		59	33*
Injection site reactions Health Care Professionals (HCP)		27	11*

Results are provided as mean ± SD.

^*P < .05 vs 4-day use.

To elucidate the impact of using insulin infusion sets for 2 days versus 4 days on patient safety and treatment efficacy, the number of catheter and treatment-related AEs was observed as primary parameter. A total of 822 catheter-related AEs were documented during the entire observation period. The majority was observed with 4-day use (517 events, 62.9%) and a significantly smaller amount was seen with 2-day use (305, 37.1%, P < .05). There was no difference between the 2 randomization arms (2 days/4 days: 130/232 AEs; 4 days/2 days: 285/175 AEs). The majority of AEs were hyperglycemic events (615 AEs, 74.8%), followed by erythema (3.5%), rash (3.2%), pain (2.6%), injection site reddening (2.2%), skin irritation (2.1%), hemorrhage (bleeding, 3.5%), and several less frequent events. Skin irritations were reported by the patients or the investigator or both. The investigator documented 11 and 27 injection site reactions and the patients reported 33 and 59 reactions for the 2-day and the 4-day use time, respectively. The difference in the mean number of injection site reactions between the 2 infusion set usage time groups was statistically significant (2-day use: 1.7 ± 2.2 events/patient vs 4-day use: 7.5 ± 14.0 events/patient, P < .05). In addition, 930 treatment-related events (AEs not related to catheter problems including hypoglycemic events) were documented, which were equally distributed between the 2 treatment arms (2-day/4-day use: all AEs 463/467, ns). An overview about the frequency and the nature of catheter related events that were inspected and confirmed by the investigator is provided in Figure 2.

Figure 2.

Absolute number and nature of catheter related events as confirmed by the investigator.

Many more unscheduled infusion set changes were required during the 4-day use (143 changes/3 months) compared to the 2-day use (73 changes/3 months, P < .05). A summary about the frequency of different technical cannula/infusion set events is provided in Figure 2. Longer usage was associated with a higher frequency of problems with the adhesive of the infusion set, while the more frequent infusion set changes resulted in more cases of bending (all P < .05). Soft cannula crimping was seen 3 times during the 2-day use but never reported during the 4-day use time (ns). No difference was seen with respect to the other observed technical infusion set issues.

Glycemic control as indicated by HbA1c was significantly improved during 2-day use in the combined groups (from 7.7 ± 1.2% to 7.4 ± 1.2%, P < .05), while it remained stable during 4-day use (7.5 ± 1.2% vs 7.6 ± 1.1%, ns). In parallel, there were no differences with respect to the hypoglycemia event rate (2-day use: 458 vs 4-day use: 450, ns).

Slightly less bolus insulin was used during 2-day use versus 4-day use (25.3 ± 11.9 U vs 27.6 ± 14.9 U, ns), and there was no difference between the treatment arms with respect to the basal insulin doses (20.2 ± 8.2 U vs 20.2 ± 8.3 U, ns).

Treatment satisfaction was significantly better during 2-day use and patients ranked general handling of the infusion systems to be easier during the shorter usage treatment arm. The results of the patient preference questionnaire are provided in Table 2.

Table 2.

Answers to the Device Handling Questionnaire.

Question	2-day use	4-day use
How satisfied are you with your present treatment/ treatment regime?	1.8 ± 0.7	2.1 ± 0.8*
How often do you have the feeling that your blood glucose levels were recently unacceptably high? (1 = very rare)	2.7 ± 1.0	3.4 ± 1.1***
How often do you have the feeling that your blood glucose levels were recently unacceptably low? (1 = very rare)	2.7 ± 1.2	2.6 ± 1.2
How easy is the removal of the lid of the inset II/Mio insertion device?	1.7 ± 0.6	2.0 ± 1.0*
How easy is the unwinding of the tubing from the insertion device?	3.2 ± 1.3	3.3 ± 1.3
How easy is it to connect the inset II/Mio to your pump?	1.3 ± 0.5	1.4 ± 0.5
How easy is the removal of the adhesive backing paper?	1.9 ± 1.0	2.0 ± 0.8
How easy is the preparation of the inserter?	2.3 ± 0.9	2.4 ± 0.9
How easy is the removal of the needle guard?	1.5 ± 0.6	1.8 ± 1.0*
How easy is the insertion of the insert II/Mio cannula at the chosen insertion site?	2.1 ± 0.9	2.1 ± 0.9
How easy is the removal of the introducer needle?	1.9 ± 0.6	2.1 ± 1.0
How easy is the general handling of the insert II/Mio infusion set?	2.3 ± 0.8	2.6 ± 0.9*
Total score	2.1 ± 0.6	2.3 ± 06*

Results are given as mean ± SD of the responses given on a rank scale (ranking: 1 = very easy to 6 = not at all easy).

^*P < .05. ***P < .001 vs 2-day use.

Only 2 serious AEs were reported. One case of gastritis occurring in the 2-day use arm required hospitalization and was classified to be not related to the study device or treatment procedure. The second case was a case of ketoacidosis, which occurred in the prerandomization period between V1 and V2, when participants were using the infusion sets for 2 days only. The patient realized an increase in blood glucose levels despite bolus applications after changing the infusion set in the morning. In the evening of the same day, the glucose levels had gone up to 600 mg/dL and symptoms of ketoacidosis were apparent. When the catheter was finally removed, kinking of the tip was observed. This case was classified as being “definitely related” to the study device.

Discussion

The purpose of our study was to confirm previous findings from our group with respect to the tolerability of prolonged use of infusion systems in patients performing continuous subcutaneous insulin infusion therapy. In a pilot investigation, we have been able to show that a longer but 3 day use of insulin infusion sets was associated with an increasing number of skin problems and a decrease in treatment efficacy.¹¹ Our current study was therefore comparing 2-day use with 4-day use to explore the clinical importance of our previous experimental findings. Our present study confirms that type 1 diabetes patients on CSII treatment, when using the infusion systems for 2 days, had significantly less technology-related problems and less AEs, while experiencing an improvement in glycemic control with slightly lower bolus insulin dose requirements as compared to 4-day use. It can be seen in Figure 2 that the AEs occurring in a higher frequency during 2-day use appear to be related to the insertion procedure and potentially the use of the auto-inserter (kinking of the cannula, cannula leaking, infusion site leaking). The insertion was performed twice as frequently during 2-day use versus 4-day use. Therefore, the event rate per insertion set used can be considered comparable or even lower during 2-day use. In addition, the participants expressed a higher treatment satisfaction when using the infusions sets for 2 days only.

Our results are in line and support the conclusions from the few reports about this topic in the past literature. The instructions for use of infusion sets for use with insulin pumps recommend use of the sets for not more than 3 days. The initial guidelines to change an infusion set every 2-3 days were published for the first time in 1983 and were based on anecdotal case reports,²⁰ and only very few investigation reports are available in the literature on this topic. In a recent laboratory investigation, Kerr et al demonstrated with all 3 short-acting insulin analogs that early catheter occlusions (within 72 hours) are rare and independent of the choice of insulin analog. The authors concluded that for patients using insulin pump therapy, the importance of catheter change within 72 hours should be emphasized irrespective of the insulin used.¹³ A self-report questionnaire was used by Pickup et al to collect information about the nature of nonmetabolic complications of continuous subcutaneous insulin infusion, which was completed by 92 patients with a mean infusion set usage time of 3.2 ± 0.7 days (range 2-6 days).²¹ The most common infusion set problems were kinking (64.1% of subjects) and blockage (54.3%), which was associated with >3 days of use of infusion sets plus lispro insulin in the pump. The assessment method used in this report was a standardized patient questionnaire, while we collected the information from the diary or during the visits and by inspecting the used catheters. The most common infusion site problem was lipohypertrophy (26.1%), which occurred more often in those with long duration of CSII. Pump malfunction had occurred in 48% of subjects. The authors concluded that pump, infusion set, and infusion site problems are common with CSII, even with contemporary technology.²¹ Another group investigated dermatological changes associated with continuous subcutaneous insulin infusion therapy in adolescents with type 1 diabetes mellitus, and tried to assess their association with duration of CSII, age, adiposity, HbA1c, insulin dose, insulin brand, infusion set or site in a cross-sectional study with 50 participants. The authors reported the following skin reactions, which were increased with longer duration of use: scars <3 mm diameter, erythema not associated with nodules, subcutaneous nodules, and lipohypertrophy. Use of 90° infusion sets was associated with a lower incidence of skin problems. In general, dermatological changes were frequent, with increased severity associated with lower adiposity. These complications were not associated with glycemic control, nor did they prompt most users to consider stopping CSII.²² As mentioned in the introduction, it has been observed that longer use of insulin pump catheters may result in impairment of glycemic control.^11,12 In our study HbA1c was improved with slightly less prandial insulin requirements during 2-day use, while no such effect was seen during 4-day use. It can be stated that all current and past limited evidence votes for keeping and reinforcing the usage time for insulin pump infusion sets to be no longer than 3 days.

It has been reported by Patel and coauthors that early failure rate of polymer versus steel catheters is related to a 15% event rate of kinking during the insertion procedure.²³ Otherwise, they found no major differences between Teflon and steel catheters even in patients wearing the infusion sets for up to 7 days. This report is in line with our finding that most frequent failure event was kinking or bending of the soft polymer cannula most likely in the course of the insertion procedure. It is possible that the kinking or bending is a consequence to striking muscle fascia. It has been determined by ultrasound measurements that 6% of all 6-mm 90° insertions will be intramuscular in adults; however, these measurements also included arms and thighs, which on average had 2-6 mm less subcutaneous tissue compared with the abdomen or buttocks.²⁴

The small sample size may be considered a weakness of our study. However, a statistical power analysis based on the findings of our pilot study¹¹ indicated that only 18 analyzable subjects would be required to demonstrate the expected clinical difference in a statistically significant way. We therefore initially enrolled 25 participants to account for potential loss by premature study termination. With the remaining per-protocol analysis set of 22 participants, we were indeed able to confirm the initial clinical hypotheses. Further limitations of our study include self-reporting of infusion site problems by the patient, except for measurements obtained by study staff at the monthly visit.

When pump users were changing the infusion set every 4 days instead of every 2 days, they suffered from more AEs, more skin problems and more hyperglycemic episodes, had a less favorable blood glucose control despite using slightly more bolus insulin, and were overall less satisfied with the treatment. There may be individual patients that are capable to wear infusion sets for a longer period than 2 to 3 days without facing increased AE rates, but the risks of doing so need to be balanced versus the potential economic benefits of such a behavior.

In any case, our study underlines the need to emphasize the safety and efficacy of using insulin infusion sets for CSII for the recommended usage time of 2-3 days when training patients on this way of insulin administration.