This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results on this page are restricted to the results posted in the ClinicalTrials.gov record; calculations added by Clinical Biostats are labelled as such.
1. Trial at a Glance
This page covers the registry results for Post Approval Study of the Eversense® Continuous Glucose Monitoring (NCT03908125), a na na study. The primary outcome is The Long-term Safety of the Eversense CGM Systems Over Repeat Insertion and Removal Cycles in the Post-market Setting..
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT03908125 |
| Phase | NA |
| Condition | Diabetes Mellitus, Diabetes Mellitus, Type 1, Diabetes Mellitus, Type 2 |
| Allocation | Not applicable (single group) |
| Design model | Single Group |
| Masking | None (open label) |
| Enrollment | 273 |
| Arms in the primary outcome | n/a |
| Lead sponsor | Senseonics, Inc. |
| Results posted | Yes |
2. Clinical Question
Population
The registry lists 273 participants and the conditions Diabetes Mellitus, Diabetes Mellitus, Type 1, and Diabetes Mellitus, Type 2. Further eligibility details and the characteristics of the enrolled participants are not reported in the registry record.
Intervention
The intervention was the Eversense® Continuous Glucose Monitoring System, described as a commercial continuous glucose monitoring device. The study examined use over repeat sensor insertion and removal cycles in a post-market setting.
Comparator
No separate comparison group is listed. The study is recorded as single-group, with allocation listed as NA and masking as none. It therefore does not provide a randomized, like-for-like comparison against another device or usual care.
Primary question
The primary question concerned the long-term safety of the system over repeat insertion and removal cycles. The endpoint was a composite: infection, a secondary procedure to remove the sensor, or a procedure-related adverse event of at least moderate severity. The stated performance goal was < 4%.
The study also listed additional effectiveness and patient-reported outcomes, including average hours of use per day, change in HbA1c levels at each 6-month interval from baseline, satisfaction with CGM use, and diabetes distress. Procedure success and serious adverse events related to insertion and removal were also listed. These outcomes broaden the questions considered, but their presence in the registry does not establish what was observed.
The registry record does not report a result for the primary composite endpoint, an estimate of its incidence, or a statistical comparison with the performance goal. It also does not report statistical analyses for the listed outcomes. As a result, the clinical question is identifiable from the registered endpoint, but the posted record does not establish how the endpoint turned out.
3. Trial Design
This was a completed post-approval study of a continuous glucose monitoring device. The registry lists a single group, an enrollment of 273, no masking, and allocation as NA. The intervention was the commercial Eversense® Continuous Glucose Monitoring System. The stated purpose was to evaluate long-term safety and effectiveness over repeat insertion and removal cycles and demonstrate long-term safety in the post-market setting.
The registry dates the study start to 2019-03-19 and primary completion to 2024-02-01. The listed study status is completed. The study phase is recorded as NA. Further details about the study sites, participant eligibility criteria, follow-up schedule, or the number of sensor cycles contributed by participants are not reported in the registry record.
Because the design is single-group, there is no concurrent control group listed for comparison. The safety endpoint is framed against a performance goal, rather than as a randomized contrast between treatments. The goal is part of the endpoint description; the registry record does not report a primary endpoint estimate with which to assess it. A single-group design can describe outcomes among enrolled participants, but it cannot by itself separate device-related effects from other influences in the way a randomized comparison is designed to do.
The record lists 13 outcome measures. Alongside the primary safety endpoint, secondary outcomes include average hours of use per day, change in HbA1c at each 6-month interval from baseline, patient satisfaction, diabetes distress, and the success and serious adverse event rates for insertion and removal procedures. The registry record does not report statistical analyses for these outcomes. It also does not describe a protocol, a sample-size justification, or an interim-analysis plan.
4. Participant Flow and Accounting
The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.
Overall Study
| Stage | Continuous Glucose Monitoring Device |
|---|---|
| STARTED | 273 |
| COMPLETED | 173 |
| NOT COMPLETED | 100 |
Reasons for not completing this period:
| Reason | Continuous Glucose Monitoring Device |
|---|---|
| Withdrawal by Subject | 92 |
| Lost to Follow-up | 8 |
5. Baseline Characteristics
The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.
Population: Subjects enrolled and inserted with the Eversense CGM system
| Characteristic | Unit | Continuous Glucose Monitoring Device |
|---|---|---|
| Participants analysed | 273 | |
| Age, Continuous | years | 44.7 (SD 14.6) |
| Sex: Female, Male: Female | Participants | 123 |
| Sex: Female, Male: Male | Participants | 150 |
| Ethnicity (NIH/OMB): Hispanic or Latino | Participants | 14 |
| Ethnicity (NIH/OMB): Not Hispanic or Latino | Participants | 259 |
| Ethnicity (NIH/OMB): Unknown or Not Reported | Participants | 0 |
| Race (NIH/OMB): American Indian or Alaska Native | Participants | 1 |
| Race (NIH/OMB): Asian | Participants | 7 |
| Race (NIH/OMB): Native Hawaiian or Other Pacific Islander | Participants | 1 |
| Race (NIH/OMB): Black or African American | Participants | 23 |
| Race (NIH/OMB): White | Participants | 235 |
| Race (NIH/OMB): More than one race | Participants | 3 |
| Race (NIH/OMB): Unknown or Not Reported | Participants | 3 |
| Region of Enrollment: United States | participants | 273 |
| Diabetes Type: Type 1 | Participants | 218 |
| Diabetes Type: Type 2 | Participants | 55 |
| Years Since Diabetes Diagnosis | years | 19 (SD 13) |
| CGM Naive | Participants | 71 |
| Diabetes Therapy Regimen: Non-insulin medications,basal insulin, lifestyle | Participants | 10 |
| Diabetes Therapy Regimen: Intensive insulin therapy | Participants | 263 |
6. Primary Endpoint
The registry lists the primary outcome as The Long-term Safety of the Eversense CGM Systems Over Repeat Insertion and Removal Cycles in the Post-market Setting., measured over up to 27 months.
| Element | Registry entry |
|---|---|
| Result type | Count Of Units |
| Measure of spread | not applicable |
| Unit | Sensor cycle |
| Endpoint classification | not reduced to a single comparison |
The primary endpoint is a composite safety measure over sensor cycles, with a timeframe of up to 27 months. It counts the incidence of any of three kinds of event: infection, a secondary procedure to remove the sensor, or a procedure-related adverse event of at least moderate severity. The stated performance goal is < 4%.
A composite endpoint brings multiple event types together into a single outcome. This can capture a broader set of clinically relevant safety events than any one component alone. Its interpretation depends on knowing how often the composite occurred and, ideally, how much each component contributed. A combined result alone would not show whether one component accounted for most events or whether events were distributed across components.
The unit is sensor cycle, so the endpoint is framed around cycles rather than only around the number of participants. Repeated cycles from one participant may not be statistically independent: characteristics of the same participant can influence outcomes across cycles. In general, analyses of repeated observations should account for this dependence when appropriate. The registry record does not report the endpoint count, denominator, component results, or the method used to account for repeated cycles.
The record defines the endpoint and its goal, but does not report an observed endpoint result or a comparison with the goal. Therefore, the registered description explains what was intended to be measured; it does not establish the endpoint’s observed incidence.
7. Why No Re-analysis Is Shown
8. Statistical Interpretation
The registry record identifies the study as completed and reports enrollment of 273 participants. It describes a single-group evaluation of the Eversense® Continuous Glucose Monitoring System over repeat insertion and removal cycles, with a primary safety endpoint and additional effectiveness, procedure, and patient-reported outcomes. However, no statistical analyses are listed, and no result for the primary endpoint is reported in the registry record.
This distinction matters: an endpoint definition is not an outcome result. The primary endpoint specifies which events were to count, the unit of sensor cycle, a timeframe of up to 27 months, and a performance goal of < 4%. The registry does not give the number of qualifying events, the relevant denominator, an incidence estimate, or uncertainty around an estimate. It is therefore not possible to determine from this record whether the observed endpoint met the stated goal.
The registry’s adverse-event information gives counts of participants at risk and affected for serious and other adverse events. It reports 273 at risk and 21 affected for serious adverse events, and 273 at risk and 115 affected for other adverse events. These are safety summaries, not the primary composite endpoint result. The record does not connect these broad adverse-event categories to the primary endpoint’s specific components, and the numbers should not be treated as equivalent measures.
For the serious-event terms listed, the registry gives percentages for hypoglycemic episode, diabetic ketoacidosis, congestive heart failure exacerbation, depression requiring hospitalization, appendicitis, and fracture of tibia and fibula. These listings do not supply a complete analysis of event timing, duration, severity beyond the serious-event classification, or causal attribution. The registry reports 0 deaths in the group. These data describe what the record lists; they do not by themselves establish whether an event was caused by the device.
The study has no listed comparator group, so it cannot provide a randomized between-group estimate. A single-group result, even if fully reported, would describe observed outcomes in the enrolled group and would not isolate the effect of the device from all other possible explanations. Here, the absence of the primary endpoint result and statistical analyses further limits interpretation. The registry record supports describing the design, endpoint definitions, enrollment, and adverse-event summaries, but not drawing a conclusion about whether the primary safety goal was achieved or about comparative effectiveness.
9. Statistical Methods Explained
Statistical methods determine how observations are turned into estimates and how much uncertainty accompanies those estimates. The registry record identifies the study design, endpoints, and some adverse-event summaries, but does not report statistical analyses. The concepts below explain how these kinds of data are generally interpreted; they are not a description of analyses confirmed to have been used in this study.
What does a single-group design estimate?
In a single-group study, outcomes are observed among participants who receive the intervention, without a separate concurrent comparison group. A proportion or rate can describe how often an outcome occurred in that group, if the numerator and denominator are reported and defined. Such an estimate answers a descriptive question about the study group. It does not, on its own, estimate the difference that would have occurred under another treatment or in the absence of the intervention.
This design is distinct from a randomized comparison. Random assignment is intended to make treatment groups comparable, on average, with respect to measured and unmeasured factors. No randomized comparison is listed in this record. The primary endpoint instead has a stated performance goal. Comparing an estimate with a goal can be informative if the estimate, denominator, and uncertainty are reported, but the registry record gives no primary endpoint estimate or uncertainty interval.
Why does the endpoint unit matter?
The primary endpoint is expressed per sensor cycle, not solely per participant. A participant may contribute more than one cycle in a study designed around repeat insertion and removal. When observations from the same person are related, treating every cycle as independent can understate uncertainty. General approaches can account for clustering or repeated measurements, but the appropriate choice depends on the data structure and analysis plan. The registry record does not report the number of cycles, their distribution across participants, or an analysis method for repeated cycles.
For an incidence measure, the denominator must match the question. A participant-based proportion uses people as the unit, whereas a cycle-based measure uses cycles. These quantities can differ when participants contribute different numbers of cycles. The endpoint description identifies sensor cycle as the unit, but the registry record does not report the observed numerator or denominator for the primary endpoint.
What does a composite endpoint capture?
A composite endpoint classifies an observation as having the outcome if any included component occurs. This can make a broad safety question measurable through several event types. It also means that a single composite estimate may conceal differences among components. A less frequent but more severe component can have a different interpretation from a more frequent, less severe component. Component-specific counts help readers understand the composition of the composite. Here, the endpoint definition lists infection, a secondary removal procedure, and procedure-related adverse events of at least moderate severity, but the registry does not report their individual results.
The stated performance goal of < 4% is a benchmark in the endpoint description. In general, judging an estimate against a benchmark requires attention both to the point estimate and to its uncertainty. A confidence interval describes a range of values compatible with the data under a specified statistical procedure and assumptions; it does not guarantee that the true value lies in that range. No confidence interval or other uncertainty measure for the primary endpoint is reported in the registry record.
How should adverse-event counts be read?
Counts of participants affected are different from counts of event episodes. One participant may experience more than one event, and event summaries can be organized by participants, event episodes, or exposure time. The registry reports 21 participants affected by serious adverse events and 115 affected by other adverse events, with 273 at risk in each category. These counts should not be treated as event rates per sensor cycle or as results for the primary composite endpoint. The registry does not report event-level details or the method used to summarize repeated events.
General safety interpretation also distinguishes an event occurring during a study from an event caused by the intervention. Causal attribution requires more than observing an event after exposure; it depends on clinical assessment and study procedures. The listed serious-event terms and percentages do not, by themselves, establish causality. The registry record does not report the criteria or process used for attribution.
What can be inferred when analyses are absent?
Without an analysis result, readers cannot assess an estimated effect, its precision, or whether an inferential threshold was met. The registry lists 13 outcome measures but no statistical analyses. It does not report a hypothesis-testing procedure, confidence intervals, model specification, or handling of the repeated-cycle structure. Those details are not reported in the registry record, so no particular method should be assumed.
Sample-size planning is also different from observed evidence. A planning illustration describes how a future or hypothetical study might be sized under stated assumptions; it is not an observed result and is not evidence that the study had a particular probability of detecting an effect. The registry record does not report the study’s own sample-size justification. It also does not provide analysis results from which to judge the precision of the primary endpoint. Consequently, the registered design and event summaries can be described, but statistical conclusions about the primary safety question cannot be recovered from this record.
10. Multiplicity, Missing Data and Analysis Populations
The registry record lists 13 outcome measures, including the primary safety endpoint and several secondary effectiveness, procedure, and patient-reported outcomes. When multiple outcomes or repeated time points are examined, the chance of at least one apparently unusual result can increase if each is assessed separately. General statistical approaches to multiplicity may adjust significance thresholds, define a hierarchy of tests, or distinguish confirmatory from exploratory outcomes. The registry record does not report a multiplicity strategy, a testing hierarchy, or adjusted analyses.
Change in HbA1c is listed at each 6-month interval from baseline. Repeated assessments raise questions about how results across time points are summarized and whether the analysis accounts for within-participant correlation. The registry record does not report those methods or results. It also does not identify which outcomes, if any, were treated as confirmatory.
Missing-data handling is not reported in the registry record. This includes the extent of missing outcome data, reasons for missingness, and any imputation or sensitivity analyses. Without this information, readers cannot assess whether incomplete follow-up could affect the reported outcomes. The record also does not describe analysis populations, such as whether all enrolled participants or a defined subset would be included in each analysis.
No statistical analyses are listed. Accordingly, no conclusion can be drawn about how multiplicity, missing observations, or analysis populations were handled. These are important reporting details for interpreting outcomes, but their absence from the registry record does not establish which procedures were or were not used outside the record.
11. Safety Results
The registry reports serious adverse events as affected participants divided by participants at risk.
| Arm | Participants with serious adverse events | Participants at risk | Percentage |
|---|---|---|---|
| Continuous Glucose Monitoring Device | 21 | 273 | 7.7% |
Deaths
| Arm | Deaths (participants affected of at risk) |
|---|---|
| Continuous Glucose Monitoring Device | 0 of 273 |
Other (non-serious) adverse events
Participants with at least one non-serious event at or above the registry's 0% reporting threshold.
| Arm | Participants affected of at risk |
|---|---|
| Continuous Glucose Monitoring Device | 115 of 273 |
Most frequent serious adverse event terms
| Event term | Organ system | Continuous Glucose Monitoring Device |
|---|---|---|
| Hypoglycemic episode | Endocrine disorders | 5 of 273 |
| Diabetic Ketoacidosis | Endocrine disorders | 4 of 273 |
| Congestive heart failure exacerbation | Cardiac disorders | 2 of 273 |
| Depression requiring hospitalization | Psychiatric disorders | 1 of 273 |
| Appendicitis | Gastrointestinal disorders | 1 of 273 |
| Fracture of tibia and fibula | Musculoskeletal and connective tissue disorders | 1 of 273 |
| Hypoglycemia w/acute metabolic encephalopathy | Endocrine disorders | 1 of 273 |
| Acute stroke with ataxia | Nervous system disorders | 1 of 273 |
| Cellulitis, bilateral lower extremity | Skin and subcutaneous tissue disorders | 1 of 273 |
| Foot infection, left | Musculoskeletal and connective tissue disorders | 1 of 273 |
Most frequent non-serious adverse event terms
| Event term | Organ system | Continuous Glucose Monitoring Device |
|---|---|---|
| Unable to remove sensor on first attempt | Surgical and medical procedures | 20 of 273 |
| COVID-19 | Respiratory, thoracic and mediastinal disorders | 17 of 273 |
| Upper respiratory tract infection | Respiratory, thoracic and mediastinal disorders | 8 of 273 |
| Sinus infection | Respiratory, thoracic and mediastinal disorders | 6 of 273 |
| Pain, sensor location | Nervous system disorders | 5 of 273 |
| Skin discoloration, sensor site | Skin and subcutaneous tissue disorders | 5 of 273 |
| Infection, sensor insertion site | Skin and subcutaneous tissue disorders | 4 of 273 |
| Influenza | Respiratory, thoracic and mediastinal disorders | 4 of 273 |
| Skin atrophy, sensor site | Skin and subcutaneous tissue disorders | 4 of 273 |
| Skin irritation, adhesive patch location | Skin and subcutaneous tissue disorders | 4 of 273 |
Serious events with the highest arm-level rate: hypoglycemic episode (1.8%), diabetic ketoacidosis (1.5%), congestive heart failure exacerbation (0.7%), depression requiring hospitalization (0.4%), appendicitis (0.4%).
The registry lists 273 participants at risk for serious adverse events and 21 affected. For other adverse events, it lists 273 at risk and 115 affected. The record also reports 0 deaths. These figures describe participants affected, not necessarily the total number of event episodes. They are separate from the primary composite endpoint, which is defined around infection, a secondary procedure to remove the sensor, or a procedure-related adverse event of at least moderate severity.
The listed serious-event terms and percentages are hypoglycemic episode, 1.8%; diabetic ketoacidosis, 1.5%; congestive heart failure exacerbation, 0.7%; depression requiring hospitalization, 0.4%; appendicitis, 0.4%; and fracture of tibia and fibula, 0.4%. The registry record does not give further details about the denominator used for these percentages or their timing and attribution. They should be read as registry safety listings, not as proof that the device caused the events.
These summaries do not report the primary endpoint’s event count or incidence. They also do not establish how many events occurred per sensor cycle or how often individual participants experienced repeated events. Such distinctions are needed to compare broad safety summaries with a cycle-based endpoint.
12. What the Result Does, and Does Not, Mean
The registry record documents a completed, single-group post-approval study with enrollment of 273 participants. It states that the study evaluated the Eversense® Continuous Glucose Monitoring System over repeat insertion and removal cycles, and it defines a primary safety composite with a performance goal of < 4%. It also lists other outcomes covering use, HbA1c, patient-reported measures, procedure success, and serious events associated with insertion and removal.
Those details explain the questions the study was intended to address, but the record does not report the primary endpoint result or any statistical analyses. It therefore does not show whether the composite incidence was below, at, or above the stated goal. No confidence interval or other measure of precision for that endpoint is reported. The adverse-event summaries do not fill that gap: they describe affected participants and listed event terms, not the cycle-based primary composite.
The absence of a comparator is also important. A single-group study can describe outcomes in its enrolled participants, but it does not provide a randomized estimate of how those outcomes compare with another device or treatment. Even a reported outcome from this design would need to be interpreted in light of other factors that could influence observed events. The registry record does not permit a comparative effectiveness conclusion.
The adverse-event listings should likewise not be interpreted as evidence of causation. Events occurring during a study may have many possible explanations, and the record does not report the attribution methods. The reported 0 deaths is a descriptive entry for the group; it does not answer the primary composite question.
In short, the record establishes the design, endpoint definitions, enrollment, and selected adverse-event summaries. It does not establish the primary safety result, its precision, whether the performance goal was achieved, or comparative effectiveness. Those conclusions require outcome analyses that are not reported in the registry record.
13. Why This Trial Matters Statistically
This study raises statistical issues that are common in post-market device evaluations. Its primary safety endpoint combines multiple event types and uses sensor cycle as the unit. That framing makes the denominator consequential: the interpretation of an incidence depends on which cycles are counted and how repeated cycles from the same participant are handled. The registry record defines the unit but does not report the event count, denominator, or analysis method.
A stated performance goal gives a single-group study a benchmark, but the benchmark alone is not an outcome. An estimate and a measure of uncertainty are needed to understand how the observed data relate to the goal. Because neither is reported for the primary endpoint, the record cannot establish whether the goal was met or how precisely safety was estimated.
The study also lists multiple outcomes across device use, HbA1c, patient experience, procedure success, and adverse events. This breadth can give a fuller view of the questions considered, while also making analysis planning and transparent reporting important. Readers need to know which outcomes were prioritized, how repeated assessments were handled, and how missing data and multiple testing were addressed. These details are not reported in the registry record.
Finally, the single-group design limits the kind of inference available. Descriptive outcomes can characterize the enrolled group, but there is no randomized control group to support a like-for-like comparison. The registry’s design and endpoint definitions are therefore useful for understanding the intended evaluation, while the lack of reported statistical analyses sharply limits what can be concluded from the posted record.
14. Limitations
- The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
- Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.
The registry record does not report the primary endpoint result, including the number of qualifying events, the denominator of sensor cycles, or an estimate of incidence. It also does not report a confidence interval or other measure of uncertainty. Without these elements, the primary safety outcome cannot be assessed against its stated performance goal of < 4%.
The record lists a single group and no comparator. As a result, the design does not provide a randomized comparison with another device or care approach. The registry does not report participant eligibility details or baseline characteristics, limiting the ability to describe how the enrolled group represents the broader populations named in the condition list.
Although the study concerns repeat insertion and removal cycles, the number and distribution of cycles are not reported. The record does not describe how repeated observations within a participant were analyzed. Component-specific results for the primary composite are also not reported, so the contribution of infection, secondary removal procedures, and procedure-related adverse events cannot be distinguished.
The registry lists secondary outcomes and adverse-event summaries but does not report statistical analyses for them. Methods for multiplicity, missing data, analysis populations, event attribution, and sample-size justification are not reported in the registry record. These omissions restrict interpretation of both the primary question and the wider set of registered outcomes.
The serious and other adverse-event summaries are participant counts, and the registry does not provide the event-level detail needed to determine recurrence, timing, or relation to sensor exposure. The listed serious-event percentages also lack further denominator and attribution information in the record. These reporting limitations do not establish what analyses may have been conducted elsewhere; they define what cannot be concluded from the registry record.
15. Related Tutorials
16. Related Statistical Calculators
17. Sources
- ClinicalTrials.gov: Post Approval Study of the Eversense® Continuous Glucose Monitoring — NCT03908125.
- Registry record last updated 2025-06-12; status last verified 2025-03.
- Results first posted 2025-06-12.
- Data retrieved from ClinicalTrials.gov on 2026-11-09.
- Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.
Continue through the Clinical Biostats statistical pathway
Use this trial's endpoint as a starting point for deeper study of effect measures, confidence intervals, and sample-size planning.
18. Record Summary
NCT03908125 is a completed, single-group, unmasked post-approval study of the Eversense® Continuous Glucose Monitoring System. The registry lists enrollment of 273, study start on 2019-03-19, and primary completion on 2024-02-01. The stated evaluation covered repeat insertion and removal cycles, with a primary safety composite assessed over sensor cycles for up to 27 months and a performance goal of < 4%.
The record lists 13 outcome measures, including use, HbA1c, patient-reported outcomes, procedure success, and adverse events. It reports 273 participants at risk and 21 affected for serious adverse events, 273 at risk and 115 affected for other adverse events, and 0 deaths. It lists no statistical analyses and does not report a primary endpoint result. The registry record therefore describes the study design and selected safety information, but does not establish whether the primary performance goal was achieved.