Individuals with type 1 diabetes face unique glycemic challenges during exercise, and glucose responses can vary depending on the type, intensity, and duration of activity (1). Automated insulin delivery (AID) systems can help by adjusting insulin delivery as glucose changes occur during exercise (2).
Swimming poses additional challenges for glycemic management with continuous glucose monitoring (CGM) and AID devices. Waterproof reliability is crucial, but accumulated cracks or leaks can compromise water-resistant integrity over time. The Omnipod 5 Automated Insulin Delivery System addresses this with a waterproof “Pod” (ingress protection rating 28 [IP28], up to 25 feet for 60 min), which is replaced every 3 days. However, submersion in water also attenuates Bluetooth signal strength, even with robust antennas (3,4).
In the Omnipod 5 pivotal trial (5), some participants reported continued system function in the water. To investigate in-water system functionality, a single-center observational substudy was conducted from September 2021 until April 2022 among individuals in the Omnipod 5 pivotal extension study at Stanford University (5). Individuals eligible for substudy participation were ≥6 years old and swimming ≥2 times per month for ≥30 min per swim.
On swim days, participants used a study-provided Garmin Swim 2 activity tracker to log swims and completed an electronic postswim survey on REDCap (Vanderbilt University, Nashville, TN). The survey collected information related to swims, including measured interdevice distance and device placement on the body. Participants followed their usual device placement and glucose management practices. Swim times were recorded using the activity tracker and swim logs. Insulet provided Pod data for each swim. Only swims with ≥70% active CGM data on Dexcom Clarity were analyzed. Using a two-tailed t test, the coefficient of determination (R2) and P value were calculated from the trend line of each swim’s percent successful CGM-to-Pod communication.
Seven eligible pivotal trial participants were enrolled in the substudy. Four (57%) participants were female, six (86%) identified as non-Hispanic White, and one (14%) identified as Hispanic. Of 137 recorded swimming events, 124 (91%) met CGM data requirements for analysis. The median swim duration was 62 min (interquartile range [IQR] 48–86 min). While all participants contributed swims, data from two participants accounted for most of the observations (103/124 swims).
From the 124 swims, the median successful CGM-to-Pod communication rate was 9.1% (IQR 0–27.4%). Median interdevice distance was 33 cm (IQR 23–38 cm). Figure 1A plots CGM-to-Pod communication success rate against interdevice distance for the 88 (71%) swims with completed swim logs that reported <10% of time spent out of the water. The six (7%) swims with closest interdevice distances (2.5–10 cm) had communication success rates of ≥71%. The swim with the next closest device distance (14 cm) had a success rate of 57%.
Interdevice distance versus percent successful CGM-to-Pod communication in the water. A: Interdevice communication rates from swims in the observational substudy of frequent swimmers in the Omnipod 5 pivotal trial (NCT04196140). Data show higher percent successful communication at closer device distances (N = 88 swims). Six swims were excluded due to ≥10% of recorded swim time spent out of the water. Overall percent successful communication across all 88 swims was 19.1 ± 25.4%. The six (7%) swims with closest interdevice distances (2.5–10 cm) had communication success rates of ≥71%. The swim with the next closest distance (14 cm) had a success rate of 57%. R2 and P values are calculated for trend line (red) from the percent successful CGM-to-Pod communication at each of the 88 swims using two-tailed t test. B: Interdevice measurement strategy. In both the observational substudy and the in vivo communication test, all device measurements were taken from the device edges, as shown. The figure also depicts how the devices were positioned in the in vivo test to ensure that their antennas were facing each other. Device orientations were not reported or controlled in the observational substudy. C: Underwater interdevice communication rates from on-body in vivo testing. The graph depicts improved success rates (>50%) at distances ≤13 cm. Data are presented as percent successful CGM-to-Pod communication out of total CGM-to-Pod communication opportunities. Each distance represents N = 16 opportunities. R2 and P values are calculated from the cumulative percentages at each of the eight distances using a two-tailed t test. D: Water measurements taken during the in vivo testing. These data represent typical chlorinated pool conditions. Ca, calcium hardness.
Interdevice distance versus percent successful CGM-to-Pod communication in the water. A: Interdevice communication rates from swims in the observational substudy of frequent swimmers in the Omnipod 5 pivotal trial (NCT04196140). Data show higher percent successful communication at closer device distances (N = 88 swims). Six swims were excluded due to ≥10% of recorded swim time spent out of the water. Overall percent successful communication across all 88 swims was 19.1 ± 25.4%. The six (7%) swims with closest interdevice distances (2.5–10 cm) had communication success rates of ≥71%. The swim with the next closest distance (14 cm) had a success rate of 57%. R2 and P values are calculated for trend line (red) from the percent successful CGM-to-Pod communication at each of the 88 swims using two-tailed t test. B: Interdevice measurement strategy. In both the observational substudy and the in vivo communication test, all device measurements were taken from the device edges, as shown. The figure also depicts how the devices were positioned in the in vivo test to ensure that their antennas were facing each other. Device orientations were not reported or controlled in the observational substudy. C: Underwater interdevice communication rates from on-body in vivo testing. The graph depicts improved success rates (>50%) at distances ≤13 cm. Data are presented as percent successful CGM-to-Pod communication out of total CGM-to-Pod communication opportunities. Each distance represents N = 16 opportunities. R2 and P values are calculated from the cumulative percentages at each of the eight distances using a two-tailed t test. D: Water measurements taken during the in vivo testing. These data represent typical chlorinated pool conditions. Ca, calcium hardness.
A subsequent study used in vivo, on-body testing to determine the optimal line-of-sight interdevice distance for underwater CGM-to-Pod communication. A Dexcom G6 CGM (Dexcom, Inc., San Diego, CA) was inserted on the right thigh. A Pod was activated off-body and secured against the skin at distances ranging from 10–17 cm from the CGM in 1-cm increments. The thigh was kept stationary and parallel to the water surface at a depth of 33–35 cm (median 33 cm). Communication was evaluated at each distance in four 20-min intervals, resulting in 80 min per distance (16 CGM-to-Pod communication opportunities) and 24 total hours of cumulative communication data. Testing was conducted in still, chlorinated water at a controlled temperature of 33.9°C (93°F) and a pH of 7.4–7.5 on four separate days.
The communication rate was 100% at 10 cm and remained >68% up to a 13-cm distance. However, at 14 cm, the rate decreased to <44%, and at 17 cm there was no communication (Fig. 1C).
The Omnipod 5 AID system’s waterproof features may benefit individuals with type 1 diabetes who swim regularly by providing effective automated insulin delivery and glycemic management during exercise. This substudy examined interdevice communication in the water and found a median CGM-to-Pod communication success rate of 9.1% with usual Pod and CGM wear, noting higher success rates (>50%) occurred at the closest interdevice distances. On-body testing showed that communication success rates dropped off quickly with interdevice distances >13 cm. However, improved connectivity does not fully address glycemic control, which is also influenced by initial glucose at exercise start, circulating insulin levels, carbohydrate intake, and exercise intensity. Nonetheless, those who wish to obtain optimal interdevice communication in the water should attempt to place the CGM and Pod within a 13-cm (5.1-inch) line-of-sight distance, with 10 cm being optimal.
Article Information
Funding. This study was performed as a substudy of the Pivotal Omnipod Horizon Automated Glucose Control System trial (NCT04196140) and was supported by additional gift funding from Insulet Corporation. M.S.H. received support from the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) of the National Institutes of Health under award numbers 5T32 DK007217 and 5K12 DK122550. R.A.L. is supported by a Diabetes, Endocrinology, and Metabolism Career Development Grant (1K23 DK122017 and P30 DK116074) from the NIDDK and has additional research support from JDRF. D.P.Z. is supported by the ISPAD-JDRF Research Fellowship and the Leona M. and Harry B. Helmsley Charitable Trust.
Duality of Interest. M.S.H. has consulted for Dexcom, Inc. D.P.Z. has received honoraria for speaking engagements from Ascensia Diabetes, Insulet Canada, and Medtronic Diabetes and is on an advisory board for Dexcom. R.A.L. has consulted for Abbott Diabetes Care, Biolinq, Capillary Biomedical, Deep Valley Laboratories, Morgan Stanley, Gluroo, ProventionBio, PhysioLogic Devices, and Tidepool. B.A.B. has been on advisory boards for Medtronic, Lilly, and Novo Nordisk and has received research grant support from Medtronic, Tandem, Insulet, JDRF, and the NIDDK. No other potential conflicts of interest relevant to this article were reported.
Author Contributions. M.S.H. and D.P.Z. researched data, conducted statistical analysis, and wrote and edited the manuscript. R.S.K. and L.H. coordinated study participation, managed institutional review board approval, and reviewed and edited the manuscript. B.A.B. and R.A.L. researched data and reviewed and edited the manuscript. M.S.H. and D.P.Z. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Prior Presentation. Portions of these data were previously presented as a poster (782-P, https://doi.org/10.2337/db22-782-P) at the American Diabetes Association 82nd Scientific Sessions, 3–7 June 2022, New Orleans, LA.
