t:slim X2 insulin pump for managing blood glucose levels in type 1 diabetes

Study size, design and location

Multicentre randomised controlled trial over 6 months in 168 people with type 1 diabetes in the US.

Intervention and comparator(s)

Intervention: closed-loop system comprised of a t:slim X2 insulin pump with Control‑IQ and a Dexcom G6 continuous glucose monitor (CGM).

Comparator: sensor-augmented pump (t:slim X2 insulin pump without Control‑IQ in patients new to insulin pump use) and a CGM.

Key outcomes

The percentage of time glucose levels were within the target range increased from 61% (plus or minus 17%) at baseline to 71% (plus or minus 12%) in the closed-loop group but remained unchanged in the control group (59%, plus or minus 14%). There were lower glycated haemoglobin levels at 26 weeks in the closed-loop group than the control group (mean difference -0.33%; 95% confidence interval [CI], -0.53 to -0.13; p=0.001). There were 17 adverse events reported (including 1 instance of diabetic ketoacidosis) in the closed-loop group compared with 2 in the control group (p=0.05).

Strengths and limitations

The study randomised people in a 2:1 ratio to the closed-loop and control group, respectively. Both groups had similar patient demographics and baseline blood glucose control. No exclusions were made based on glycated haemoglobin level. There were also 100% of people who stayed in the study for the whole duration, and a high level of adherence to the assigned devices in both groups.

Study size, design and location

Randomised controlled multicentre trial of 109 people with type 1 diabetes (aged 14 to 72) over 3 months in the US. This is an extension study to Brown et al. (2019).

Intervention and comparator(s)

Intervention: closed-loop system comprised of a t:slim X2 insulin pump with Control‑IQ and a Dexcom G6 CGM.

Comparator: predictive low-glucose suspend system comprised of a sensor-augmented pump and a CGM.

Key outcomes

People who continued to use the closed-loop system from the first trial could maintain time in range (70 to 180 mg/dl) for this 3‑month extension period (71.1% plus or minus 11.2% at baseline and 67.6% plus or minus 12.6% after the extension period; baseline measurements calculated as the last 13 weeks of the initial trial). This is compared with the control group, which went from 70.0% plus or minus 13.6% at baseline to 60.4% plus or minus 17.1% at the end of the extension period. HbA1c at the end of this study was lower in the intervention than the control group (7.2% compared with 7.5%; p=0.0035).

Strengths and limitations

There was no remote monitoring and only 1 clinic visit in this study to reflect real-world device usage. Because this was an extension study, the device users were more familiar with the technology potentially leading to better glucose control. Control-IQ was suspended for 4 weeks during this study.

Study size, design and location

Multicentre randomised controlled trial in 24 school-aged children (6 years to 12 years) with type 1 diabetes in the US.

Intervention and comparator(s)

Intervention: closed-loop system comprised of a t:slim X2 insulin pump with Control‑IQ and a Dexcom G6 CGM as well as a Dexcom G5 CGM (because remote monitoring for Dexcom G6 was not yet available).

Comparator: sensor-augmented pump (user's own) and a Dexcom G5 CGM.

Key outcomes

In the 3 days of home device usage, time in target blood glucose range (70 mg/dl to 180 mg/dl) was improved in the closed-loop group compared with the control (closed loop: 71.0% plus or minus 6.6%; control: 52.8% plus or minus 13.5%; p=0.001). The mean glucose was 152.2 mg/dl (plus or minus 13.8 mg/dl) in the closed-loop group compared with 180.2 mg/dl (plus or minus 23.1 mg/dl; p=0.003) in the control. There was a reduction in hyperglycaemia exposure in the closed-loop group compared with the control group and no differences in hypoglycaemia exposure. A survey of device usability found that the closed-loop system was viewed favourably by the users.

Strengths and limitations

There were equal demographics and baseline diabetes control between the groups. Power calculations deemed the sample size to be sufficient to detect a true difference. However, this study was funded by the company and was only done over a 3‑day period.

Study size, design and location

Multicentre 6-week randomised crossover trial of 103 people with type 1 diabetes in the US. The trial consisted of 2 × 3‑week periods when people were randomly assigned to use the predictive low‑glucose version of the pump during 1 period and the non-predictive low‑glucose version during the other.

Intervention and comparator(s)

Intervention: closed-loop system comprised of t:slim X2 pump with Basal‑IQ integrated with a Dexcom G5 sensor and predictive low‑glucose suspend technology.

Comparator: sensor-augmented pump comprised of t:slim X2 pump with Basal‑IQ integrated with a Dexcom G5 sensor without predictive low‑glucose suspend technology.

Key outcomes

Mean time in range (70 mg/dl to 180 mg/dl) increased in the intervention arm compared with the control arm (64% at baseline, 65% with intervention, and 63% with control; p<0.001). There was a reduction of time in hypoglycaemia (p<0.001) in the intervention arm compared with comparator. No one in the trial experienced severe hypoglycaemia or diabetic ketoacidosis during the intervention period. There was 1 severe hypoglycaemic event reported in the control period. A system usability survey stated that people would like to continue using the device with predictive low-glucose functionality and stated that it was straightforward to use.

Strengths and limitations

This was a well-powered randomised crossover study. However it only compared the predictive low-glucose suspend technology rather than comparing the t:slim X2 pump with a comparator device. This study was funded by the company.

Study size, design and location

Retrospective analysis of 6,170 t:slim X2 pump users before and after installing the Basal-IQ software update with predictive low glucose suspend system in the US.

Intervention and comparator(s)

Closed-loop system comprised of t:slim X2 pump with a compatible CGM before and after using Basal‑IQ with the predictive low-glucose suspend system.

Key outcomes

Time in range (70 mg/dl to 180 mg/dl) increased from 57.8% to 58.5% (0.64, 95% CI 0.04 to 1.24; p<0.001) and time in which blood glucose was less than 70 mg/dl decreased from 2.14% to 1.18% (-1.01, 95% CI -0.97 to -1.05; p<0.001) after installation of the Basal‑IQ software. Of the 1,301 subjects in which 9 months of follow-up data was available, time in which blood glucose was less than 70 mg/dl remained at 1.2% and time in range was 62.6% (plus or minus 16.6%).

Strengths and limitations

This was a retrospective study using data submitted to an online diabetes management platform. Although there was a large study group, this analysis was funded by the company and verified independently.

Study size, design and location

Randomised controlled trial of 48 children and young people (aged 6 years to 18 years) participating in a 48-hour ski camp in the US.

Intervention and comparator(s)

Intervention: closed-loop system comprised of a t:slim X2 insulin pump with Control‑IQ and both a Dexcom G6 and G5 CGM.

Comparator: sensor-augmented pumps with low‑glucose suspend technologies deactivated and a Dexcom G5 CGM.

Key outcomes

The Control‑IQ system led to a 12% increase in time in range (70 mg/dl to 180 mg/dl) from 53.9% (plus or minus 24.8%) in the control group to 66.4% (plus or minus 16.4%) in the Control‑IQ group (p=0.01). There was a 15% reduction of time spent in hyperglycaemia (above 180 mg/dl) in Control‑IQ group compared with the control group (31.4%, plus or minus 17.6%, compared with 43%, plus or minus 24.5%; p=0.015). There was no difference in the time in hypoglycaemia or the amount of carbohydrate treatments for hypoglycaemia between the groups. There were no adverse events in either group.

Strengths and limitations

This randomised controlled trial, although only done over 48 hours, studied the efficacy of the pumps with prolonged intense activity. Remote monitoring by clinicians could have led to improved glycaemic control compared with real-world usage. The study was partially funded by the company.