Sensor-Augmented Pump Therapy at 36 Months

Introduction

S ensor-augmented pump (SAP) therapy—the combined use of an insulin pump and continuous glucose monitoring (CGM)—can improve glycemic control in adults with suboptimally controlled type 1 diabetes, thereby reducing the risk of micro- and macrovascular complications. ^1,2 This has been demonstrated in several randomized studies of 26–52 weeks in duration, ^{3 –10} and persisting effects of SAP after 18 months of sustained use have recently been reported. ¹¹ Furthermore, SAP has the potential to improve diabetes treatment satisfaction and to reduce perceived magnitude of diabetes-related problems, hypoglycemia fear, and hypoglycemia-avoidant behavior, which are all critical to treatment adherence. ^9,12 The long-term effects of this relatively new treatment regimen, however, are yet to be determined. Here we present 36-month follow-up data—the longest SAP follow-up published to date—from the Danish subcohort of patients who participated in the Eurythmics Trial. ⁹

Subjects and Methods

From April 2007 to January 2009, 78 adults with type 1 diabetes completed the 26-week open-label European multicenter randomized controlled Eurythmics Trial comparing the effects of SAP with multiple daily injections therapy. Twenty-four patients (31% of the study cohort) with long-standing poor metabolic control were recruited from Copenhagen University Hospital, Hvidovre, Hvidovre, Denmark. Of these 24 patients, 13 were randomized to receive the study intervention, SAP (Paradigm^® REAL-Time System; Medtronic, Northridge, CA), while 11 were randomized to the control group continuing multiple daily injections therapy. After completion of the trial, the 11 Danish patients in the control group (EUR-CON) started SAP, and the 13 patients in the intervention group (EUR-INT) continued their use of SAP with quarterly routine follow-up appointments at the diabetes clinic. The costs of devices and consumables were covered by the public health services. This 36-month follow-up was not a part of the original Eurythmics protocol and included only the Danish subgroup of participants.

For this study, we invited all 24 patients from the Danish subgroup of Eurythmics participants. The aim of the study was to assess SAP use, metabolic control, and psychosocial status 36 months after initiation of SAP therapy. Baseline was defined as the time of SAP initiation. Accordingly, EUR-INT patients were assessed 36 months after entry into the Eurythmics Trial, and EUR-CON patients were assessed 36 months after completion of the Eurythmics Trial. We sent out invitations 27 months after SAP start. Twenty-three patients gave consent to participation; one patient had left our clinic and did not respond to the invitation. Seventeen patients replied that they were still using SAP and were included in the study: nine from EUR-INT and eight from EUR-CON. After having given consent to participation but before completion of the 36-month observation period, a SAP user from EUR-INT died from a condition unrelated to diabetes. ¹³ We asked the patients to record episodes of severe hypoglycemia (blood glucose ≤50 mg/dL and assistance needed from another person) and ketoacidosis from 30 to 36 months after SAP initiation corresponding to the last 6 months of observation. Otherwise, patients were encouraged to continue their diabetes care, including insulin pump and sensor use, regardless of study participation. Clinic visits were arranged at 30, 33, and 36 months. Glycated hemoglobin (HbA1c) was routinely measured locally (normal range, 4.1–6.4%) at all visits to the clinic. Historic HbA1c levels obtained biannually 36 months prior to SAP start and at 0, 3, 6, 12, 18, 24, and 27 months (±1 month) of SAP use were retrieved from medical records. We downloaded data from insulin pumps at 33 and 36 months, thus obtaining data from 6 months, and administered questionnaires at 36 months. The questionnaires included the Diabetes Treatment Satisfaction Questionnaire (DTSQs), the Hypoglycemia Fear Survey (HFS), and the Problem Areas In Diabetes (PAID) questionnaire, which had also been distributed at entry and at 26 weeks in the Eurythmics Trial. The study was approved by The Regional Committee on Biomedical Research Ethics (protocol number H-3-2009-150).

Results

Characteristics of the 24 patients at SAP start were as follows: age, 39.0±11.9 years; female sex, 50%; white, 95.8%; HbA1c, 8.7±0.9%; and diabetes' duration, 20.3±10.3 years. Insulin pump downloads showed that at 36 months, 14 patients were using CGM on a daily basis (88.9±10.3% of the time), and two patients were using CGM intermittently (mean, 20.6±0.3% of the time). Figure 1 shows mean HbA1c values 3 years prior to and 3 years after SAP start.

OPEN IN VIEWER

FIG. 1.

Glycated hemoglobin (HbA1c) 36 months before and after sensor-augmented pump (SAP) therapy start (0 month). Data are mean±SEM (bars) values.

The initial glycemic effect of SAP (baseline to 6 months) in all 24 patients was a reduction of 1.6% (95% confidence interval 1.3–1.9%; P<0.0001) in HbA1c, from 8.7% to 7.0%. From 6 to 36 months, HbA1c in these same patients increased slightly by 0.3% (P=0.042) but remained significantly different from baseline HbA1c (P<0.0001).

Baseline HbA1c was 8.9% in EUR-INT and 8.4% in EUR-CON. The change in HbA1c from baseline to 6 months in the two groups was 1.9% and 1.3%, respectively (P=0.030). However, there was no difference in change in HbA1c between patients in EUR-INT and EUR-CON (P=0.097) when adjusted for differences in baseline HbA1c values.

The 16 patients who were using SAP at 36 months demonstrated a mean reduction in HbA1c of 1.7% (95% confidence interval 1.3–2.1%, P<0.0001), from 8.8% to 7.0%, during the first 6 months of SAP use. From 6 to 36 months an insignificant 0.3% increase in HbA1c was observed (P=0.065), but compared with baseline values the improvement in HbA1c obtained remained highly significant (P<0.0001). The two patients who were using CGM intermittently (included in the analyses above), both from EUR-CON, had baseline HbA1c values of 8.3% and 8.2% and 36-month HbA1c values of 7.5% and 7.7%, respectively. One of the six patients who ceased SAP therapy returned to multiple daily injections therapy. This patient's baseline and 36-month HbA1c values were 9.3% and 8.7%, respectively. The remaining five patients had chosen to terminate CGM and continue insulin pump therapy only. One of these was a pregnant woman with an HbA1c level of 6.3%. In the four nonpregnant former SAP users, HbA1c decreased from 8.2±0.3% at baseline to 7.1±0.3% at 36 months. The duration of SAP usage in the six former SAP users was 13.0±9.3 months (patient-reported data).

From SAP start to 36 months the mean DTSQs score increased from 23.1±6.1 to 32.1±4.5 (95% confidence interval 5.5–12.5; P<0.0001) in the 16 patients who were still using SAP. PAID questionnaire scores improved with a reduction from 27.0±21.2 to 16.2±12.4 (95% confidence interval −19.0 to −2.7; P=0.013). HFS scores were reduced from 25.8±19.2 to 20.3±14.7 (95% confidence interval, −13.3 to −2.3; P=0.152). None of the changes in patient-reported outcomes differed between EUR-INT and EUR-CON.

One SAP user with hypoglycemia unawareness experienced three episodes of severe hypoglycemia during the 30–36-month interval. No other cases of hypoglycemia or ketoacidosis were reported. Three episodes in 16 patients in 6 months correspond to a hypoglycemia rate of 37.5 episodes/100 person-years. This rate did not differ significantly from the hypoglycemia rate reported in the year prior to SAP start: 31.3 episodes/100 person-years (five patient-reported episodes in four patients in 1 year) (P=1.000).

Discussion

This follow-up study documents that the positive effects of SAP observed in the Eurythmics Trial are maintained 3 years after SAP initiation in the Danish sub-cohort of Eurythmics Trial participants. ⁹ The effects include clinically important and statistically significant reductions in HbA1c as well as improvements in treatment satisfaction, magnitude of perceived diabetes-related problems, and fear of hypoglycemia compared with baseline values. These 36-month data are in line with the 18-month data reported by the STAR3 study group. ¹¹ The study also demonstrates that the same effects can be achieved regardless of whether SAP is initiated during a clinical trial (EUR-INT) or in daily clinical practice (EUR-CON). After participating in the Eurythmics Trial, patients returned to quarterly follow-up visits at the outpatient clinic, and the majority of the results presented here have been obtained during routine clinical practice. However, it might be that some of the patients modified their behavior toward the end of the 36-month period when informed about the upcoming monitoring for the study. We have provided the reader with descriptive information about the patients who chose to cease SAP and continue insulin pump therapy only. The sample size does not allow statistical testing; nevertheless, it seems that these patients maintained improvements in HbA1c similar to the SAP users. One could speculate whether continued CGM is needed to achieve the observed improvements in glycemic control, compared with continued insulin pump use following a period of prior CGM use. Another explanation could be that shorter-term use of SAP mediates treatment optimization that is sustained even after CGM termination. The two individuals who were using CGM intermittently experienced below average reductions in HbA1c, which is in line with previous studies suggesting that CGM should be used the majority of time to be beneficial. ^{3 –8,14} Our data, however, cannot support or reject these hypotheses because of sample size and differing duration of device use before CGM cessation.

The changes in questionnaire scores from baseline to 36 months in the 16 patients still using SAP at 36 months are comparable to the changes observed during the first 6 months of SAP use in the Eurythmics trial, ^9,12 documenting a sustained beneficial effect of SAP also on psychosocial outcomes. Although the scores improved in all three questionnaires used, only changes in DTSQs and PAID questionnaire scores reached the statistical significance level, and, correspondingly, only the changes recorded in these questionnaire scores were clinically important (i.e., mean change >0.5 baseline SD). ¹⁵ Fear of hypoglycemia was, however, significantly reduced 12 months after SAP start in the STAR3 study population, and the diverging results may depend on the small sample size in our follow-up study. ¹²

Six patients made the informed choice to discontinue SAP usage despite having experienced beneficial metabolic effects from SAP. To explore mechanisms associated with SAP adherence we conducted a focus group study including these patients. ¹⁶ We found that patient motivation, body image, and strategies for coping with sensor alarms and inaccuracies were factors affecting SAP adherence.

The small sample size is an obvious limitation of our study. The results of larger long-term studies comparing SAP with insulin pump therapy are still needed to determine the long-term effects of SAP. One barrier to long-term use, however, is the cost of sensors. SAP use in the Danish Eurythmics subcohort was not limited by economic considerations as sensor costs were covered by the public health services.

In summary, this 36-month follow-up study documents sustained beneficial effects of SAP on both metabolic and psychosocial outcomes.