Swimming Sessions Revealed Predictable Glucose Drops Under Omnipod 5
Swimming with type 1 diabetes creates a clinical problem that is easy to describe and hard to manage: water exercise can drive glucose down quickly, yet carbohydrate dosing before a swim can also push glucose higher than intended. An observational analysis nested within the Omnipod 5 pivotal trial offers a rare look at what actually happened during real swims in children using an automated insulin delivery system. The cohort was small, but the data are unusually practical because they capture a setting where sensor-driven insulin adjustment meets intermittent, often intense aerobic effort.
Seven swimmers, aged 11 ± 2 years and 57% female, contributed information on 124 swims lasting 23 to 146 minutes. Participants logged their swims with an activity tracker and completed online surveys afterward, creating a picture of day-to-day glycemic behavior rather than a controlled laboratory experiment. Median starting glucose was 133 mg/dL, with an interquartile range of 93 to 178 mg/dL. During swims, glucose fell by a median of 17 mg/dL, although the spread was wide, ranging from a decrease of 61 mg/dL to a slight increase of 7 mg/dL.
Swimming was not metabolically uniform.
Higher Starting Glucose Produced the Sharpest In-Session Falls
The most striking pattern emerged when starting glucose was already elevated. When participants began swimming at 181 mg/dL or higher, glucose fell by a median of 86 mg/dL, with an interquartile range of 111 to 25 mg/dL. In practical terms, the greatest downward shifts occurred in the very group that might appear safest at the outset. Aerobic exercise increases skeletal muscle glucose uptake and insulin sensitivity, so a higher starting glucose does not guarantee protection once the swim begins. In children using AID, the insulin already onboard may continue to exert its effect while the muscles consume circulating glucose at an accelerated pace.
Carbohydrate intake before swimming was modest overall. The median amount consumed in the 2 hours before a swim was 20 grams, with an interquartile range of 0 to 32 grams. That pattern suggests families were not systematically overcorrecting with large pre-exercise carbohydrate loads, which may have helped limit unnecessary hyperglycemia before activity. It also means the observed glucose declines likely reflect the combined effect of exercise physiology and system behavior rather than a simple carbohydrate overshoot.
Activity Mode Use Was Common, Yet Not Universal Across Swims
The Omnipod 5 activity feature was used in 87% of swims, indicating that most sessions were entered with a deliberate attempt to soften insulin delivery around exercise. In an AID framework, that setting is clinically meaningful because it is designed to reduce insulin exposure during periods when glucose is expected to fall. Swimming adds another layer of complexity: duration is variable, energy expenditure can be substantial, and water-based activity may make corrective action less immediate than on dry land. The study does not claim the feature prevented every decline, but it does show that families and clinicians were actively using it in the majority of sessions.
The data fit a familiar physiologic pattern. When insulin delivery remains aligned with normal physiologic needs, glucose tends to drift downward during sustained aerobic activity. When the starting glucose is modestly elevated, there is still room for a large fall because exercise can unmask the underlying insulin effect already present in the body. That is especially relevant for pediatric swimmers, where the goal is not simply to avoid hypoglycemia but to keep glucose stable enough to finish a practice or recreational session without interruption.
Small Cohort, Real-World Conditions, and a Narrow Clinical Message
This investigation was observational, limited to seven participants, and built from self-reported swim sessions rather than a controlled crossover design. Those constraints leave major questions unanswered, including how the system performs across different swim intensities, water temperatures, race conditions, or longer training blocks. No one should read the results as a universal dosing rule. Still, the signal is clinically coherent: swimming while using Omnipod 5 was generally compatible with manageable glucose shifts, and the largest drops occurred when the pre-swim glucose level was already high.
The age of the participants matters as well. Children and young adolescents do not approach exercise with the same insulin reserve, carbohydrate planning, or symptom awareness seen in adults with long-standing diabetes. A median pre-swim glucose of 133 mg/dL suggests families were starting in a range that many clinicians would consider acceptable, yet the median in-session decrease of 17 mg/dL shows how quickly the balance can move. The wide interquartile spread also argues for individualized planning rather than one-size-fits-all advice.
What the Swim Data Mean for Bedside Advice
For clinicians counseling children with type 1 diabetes who swim, this study supports a straightforward message: pre-exercise glucose should be reviewed with the expectation that levels may fall once the swimmer is in the pool, even when the starting value appears safe. The AID activity feature was used in most swims, which reinforces its role as a practical tool rather than a theoretical setting. Families still need to pair that tool with observation, because the system cannot fully neutralize the glucose-lowering force of sustained aerobic exercise.
The physiologic lesson here is simple but important. Swimming can generate a meaningful glucose decline despite automated insulin modulation, and elevated starting glucose does not necessarily protect against a steep fall. In pediatric diabetes care, that means the conversation before a swim should focus less on a single target number and more on the entire context: recent carbohydrate intake, anticipated duration, prior insulin exposure, and the likelihood that the child will keep moving long enough for glucose to drift lower.
The study ends with a narrow but useful observation: real-world swimming was feasible under Omnipod 5, yet glucose declines remained common enough to warrant respect from clinicians, families, and coaches. Larger studies will be needed to determine how often activity mode should be activated, whether carbohydrate strategies should differ by swim length, and whether certain starting glucose bands offer a better safety margin for children in the pool.