A Randomized Trial of the Effect of Continuous Glucose Monitoring (CGM) in Individuals With Type 1 Diabetes Treated With Multiple Daily Insulin Injections (MDI)
A Randomized Trial of the Effect of Continuous Glucose Monitoring (CGM) in Individuals With Type 1 Diabetes Treated With Multiple Daily Insulin Injections (MDI)
Breif summary
A keystone in preventing diabetic complications in patients with type 1 diabetes is good glycaemic control. Frequent self-measurements of blood glucose (SMBG) levels has been an essential part of insulin dosing before meals. However, in recent years continuous glucose monitoring (CGM) has become a treatment option for notifying the patient on trends in glucose levels and warning when these are estimated to be too high and too low.
In some countries today, Sweden among others, CGM is reimbursed in combination with continuous subcutaneous insulin infusions (CSII) in patients with very poor glycaemic control or a history of repeated severe hypoglycaemia in adult type 1 diabetic patients. This is based on existing clinical trials showing a beneficial effect on HbA1c by combining CGM with CSII. However, the majority of adult type 1 diabetic patients are treated with multiple daily insulin injections (MDI). Clinical trial data are sparse on the effect of CGM in adult type 1 diabetic patients treated with MDI, and there are no clinical trial data including only patients on MDI.
The aim of the current study is to evaluate effectiveness, safety and treatment satisfaction among adult type 1 diabetic patients on CGM treated with MDI. The design is a 69-week, cross-over clinical trial, including 26 weeks treatment with CGM, 26 weeks treatment with conventional SMBG and a wash-out period of 17 weeks. In total 120 patients will be included at 8 sites in Sweden. The study will have 80% power to detect a 3 mmol/mol (0.3 percentage unit) change in HbA1c resulting from CGM.
Background
A keystone in preventing diabetic complications in patients with type 1 diabetes is good glycaemic control (1). Today, intensive glycaemic treatment is generally achieved through multiple daily insulin injections (MDI) or an insulin pump, also termed continuous subcutaneous insulin infusion (CSII, [2]). Regular capillary self-measured blood glucose values have been most crucial in obtaining good glycaemic control and guiding the patient on insulin doses (3, 4, 5).
During recent years continuous glucose monitoring (CGM) has become a treatment option for guiding the patient on insulin dosage and other activities (6). CGM has the advantage of informing the patient on estimated glucose values continuously, not the least important of which is to illustrate trends on increases or decreases in glucose levels.
Data from several clinical trials on CGM has shown divergent results on its glycaemic control effects (7). In some clinical trials, only patients on CSII have been included or have initiated CGM and CSII simultaneously as an intervention. In other trials both patients with MDI and CSII have been included, and post hoc analyses have also resulted in divergent findings whether the effect on glycaemic control potentially differs when combining CGM with MDI or CSII (8, 9, 10). Although the absolute majority of adult type 1 diabetic patients are treated with MDI, clinical trials initiating CGM in a pure MDI-treated group are absent.
The current trial is a cross-over design and 69 weeks in duration, where patients will be randomized patients to CGM-treatment for 26 weeks, conventional therapy for 26 weeks and a wash-out period for 17 weeks. The primary endpoint is the effect on HbA1c.
Purpose/aim
The aim of this study is to analyse the effect of CGM on glycaemic control measured by A1C, high and low glucose levels measured by CGM, and quality of life in patients with type 1 diabetes treated with MDI.
Treatment
The studied intervention will be CGM (Dexcom G4, Dexcom Corporation) which will be compared to conventional therapy using only self-measurements of blood glucose levels (SMBG) for guiding the dosage of insulin.
Randomization
After a maximum run-in period of six weeks patients will be randomized to either CGM or continued conventional therapy. During the run-in period blinded CGM will be performed during two weeks. After the blinded CGM period patients that do not believe they will wear a CGM sensor more than 80% of the study time during the period of randomization to CGM, or patients who did not perform adequate calibrations during the run-in period (on average at least 12 of 14 during a 7-day period), will not be randomized. The patient will be shown an example picture of glucose curves (not their own curves) with trend arrows, explained by the physician/diabetic educator to give the patient a better chance to judge how often they will use the sensor. Consenting patients will be randomized to CGM or conventional therapy for 26 weeks and conventional therapy for 26 weeks, with an intermittent wash-out period for 17 weeks.
Patients will be initially randomized 1:1, stratified by site, to CGM or conventional therapy. A centralised web system will be used for randomisation. Each patient will be assigned a unique and anonymous Subject ID at randomisation.
Duration
The expected study duration for each participant is 72 weeks, including an assumed mean run-in period of 3 weeks. The total study period is expected to be 84 weeks, including a recruitment period of 12 weeks.
Selection and withdrawal of subjects
Patients fulfilling all inclusion and no exclusion criteria will have their HbA1c levels, fasting C-peptide and creatinine analysed by the central laboratory.
The study is planned to include 120 patients randomized 1:1, stratified by site, to CGM or conventional therapy. Treatment will be for 26 weeks for each group, with a wash-out period of 17 weeks between treatments. An expected drop-out rate is assumed to be 5%-10% without replacement.
Rescreening
Rescreening of patients is possible in the study, but maximally at one time. There is no time limit for rescreening. It can be performed at any interval from previous screening. Rescreening can be performed for any inclusion/exclusion criterion that did not fit the inclusion criteria or fulfilled any exclusion criteria at the previous screening. However, there should be a possibility that this criterion can fit this inclusion criterion or not fit the exclusion criterion at the rescreening visit; e.g. that the glycaemic control has worsened since last screening and HbA1c did not fit inclusion criteria at the previous screening visit.
Treatment procedures
During the run-in period all patients will have blinded CGM during two weeks. If the patient believes after performing blinded CGM that he/she will not be able to wear the sensor and use the CGM-system during the majority of the study period (more than 80% of the time) when randomized to CGM, he/she will be excluded from randomization. In addition, subjects not adherent with calibration procedures (require on average > 12 out of 14 calibrations over a 7 days period). There should be at least 10 days with valid blinded CGM data before randomization.
All patients in the trial will be instructed regarding basic information on insulin dosing, such as bolus correction, types of food elevating glucose levels and the effect of physical activity on glucose control. This information will be provided at the same level as in clinical practice for patients with type 1 diabetes, i.e. to guarantee that all patients have basic skills for dosing insulin. All patients will also be educated on the proportion of rapid acting insulin analogues remaining at various time points after injection.
At clinical visits the care-giver will discuss glucose levels measured by SMBG and CGM data with the patient for possible improvements in the diabetes care. This will be performed in correspondence with intensive therapy used in clinical practice. All patients will have the possibility to contact the responsible staff for the trial at each site for additional support between the visits if needed, e.g. technical problems with SMBG meters or the DexCom 4G system, but extra visits will not be planned with the aim of improving the glycaemic control.
During at least the first week of randomization to CGM there will be no alarm levels set on the CGM, other than a constantly active acute alarm to low glucose levels. The reason is that the patient shall be taught to be active in judging trends of CGM and not only reacting at certain levels for alarms.
Alarm settings will be introduced 2 weeks after randomization, at the latest. At each visit the patient will be motivated to be active using the information from the CGM at least every 1-2 hours during daytime. In correspondence, patients will be motivated in measuring blood glucose levels when randomized to conventional therapy in accordance with guidelines, i.e. at least 4 times a day. At the 2 initial visits of each treatment period patients will be checked for general skills adopted on dosing insulin, types of foods that elevate glucose levels and the influence of physical activity on glucose levels. BG-values will be evaluated at the visits for patients receiving conventional therapy for possible improvements in dosing insulin, food intake and physical activity. In correspondence, CGM-curves will be analysed as well as algorithms for dosage of insulin from CGM-data and physical activity and influence of eating habits on glucose levels. When randomized to CGM patients will be instructed at the start point of the treatment phase and the two consecutive visits on a predefined algorithm for adjusting insulin. In each treatment phase visits will take place at starting point, weeks 2, 4, 13 and 26.
Blinded CGM will be performed for all participants during two weeks before baseline and two weeks before the starting point of the second treatment phase. Participants randomized to conventional therapy will also have CGM during 2 of the 4 last weeks of each treatment period (performed 23-26 and 66-69 respectively).
HbA1c will be recorded at the starting point of each treatment period and all subsequent study visits except week 2 in each treatment phase. At starting point and at the end of each treatment phase extended blood samples will be taken including biobank samples.
At all visits SMBG and CGM data will be downloaded for randomized patients, and a diabetes educator or physician will discuss potential improvements for optimising glycaemic control with the patient. SMBG-data will be downloaded also for patients with CGM-treatment.
Rescue criteria
If the clinician or diabetic educator determines that CGM use is associated with severe risks, e.g., severe hypoglycaemia, CGM treatment shall be stopped and the patient will receive conventional treatment.
Treatment satisfaction and quality of life
Patients will fill in questionnaires before blinded CGM of each treatment phase and at the end of each treatment phase. The DTSQ has been used in many diabetes therapy clinical trials and is a validated questionnaire consisting of 8 questions. Two versions are used, the DTSQs and DTSQc, where the DTSQs is used for recording the current treatment satisfaction and the DTSQc for patients to retrospectively compare various treatments.
SWE-HFS consists of 23 questions concerning actions to prevent hypoglycaemia and fears about hypoglycaemia. The Swedish translation has been well validated (12).
SWE-PAID-20 consists of 20 questions regarding situations about diabetes that may be a problem to the individual.
WHO-5 consists of 5 questions assessing patient well-being.
IPAQ consists of 4 questions of various levels of physical activity during the last 7 days. (13)
Hypoglycaemia confidence questionnaire consists of 9 questions regarding how confident the patients are regarding handling of hypoglycaemia
The questionnaires will be completed at the study site. The patients will be allowed to individually complete the questionnaires in a reasonably quiet environment. It will be emphasized that patients complete the questionnaires prior to clinical measurements and before meeting a doctor. Questionnaires should be answered by the patient alone; however, the nurse/assistant will be informed to help patients complete the questionnaires, if necessary, but without influencing patients' responses. Only the anonymous Subject ID will be used to identify questionnaires to ensure patient confidentiality. Study nurses/assistants should check questionnaires for completeness. The PI shall ensure that appropriate study training is provided.
Hypoglycaemia
Periods of hypoglycaemia will be compared using blinded CGM versus randomization to open CGM during the corresponding time period. The regular definitions of hypoglycaemia using SMBG will be difficult to compare since patients using CGM will detect asymptomatic hypoglycaemia due to CGM and be more alert to symptoms of hypoglycaemia. The number of severe hypoglycaemic events, defined as unconsciousness due to hypoglycaemia or need of assistance from another person to resolve hypoglycaemia, will be recorded. The time with low glucose values will be analysed by comparing active treatment with CGM with blinded CGM for the corresponding time period.
Inclusion Criteria:
Exclusion Criteria: