Perioperative Care of Patients Using Wearable Diabetes Devices

The prevalence of diabetes continues to increase worldwide, including in the United States. Half of the 37 million individuals with diabetes in the United States will undergo a surgical procedure in their lifetime.¹ This is noteworthy, as hyperglycemia is associated with worse surgical outcomes.^2,3 Guidelines have been developed for the perioperative management of patients with diabetes that address insulin, noninsulin medications, and perioperative glycemic control^1,4–6 Though landmark clinical trials have demonstrated that long-term intensive glycemic control reduces microvascular complications in patients with type 1 diabetes (T1D) and type 2 diabetes (T2D),^7,8 optimal glucose targets in the perioperative period have been more controversial. Based on current evidence, several societies advocate a hyperglycemic treatment threshold of 180 mg/dL,² or even 250 mg/dL for minimally invasive ambulatory procedures of short duration.⁵

Advances in diabetes management technology have translated to improved glycemic control in adults with T1D, leading to the international recommendation to offer automated insulin delivery (AID) systems to most individuals with T1D.⁹ The landscape of diabetes technology, particularly of AID systems, is rapidly evolving with high uptake among the T1D community and increasing use in individuals with T2D, as insulin remains a cornerstone of therapy in some patients with T2D.

Wearable diabetes devices include continuous glucose monitors (CGMs) and insulin pumps that infuse insulin into the subcutaneous tissue. Patients with diabetes presenting for surgery or other procedures may use a CGM device, an insulin pump or both. Patients now commonly use integrated AID systems (also called “hybrid closed loop”) where glucose values from a CGM are transmitted via Bluetooth technology to an insulin pump that autonomously delivers insulin guided by an algorithm; these systems are designed to avoid hyper- and hypoglycemia and inherently target more strict glycemic control than is advocated for the perioperative period by published guidelines.

While recommendations for appropriate perioperative glucose target ranges and perioperative diabetes medication adjustments use are now readily available, the perioperative use of these increasingly prevalent wearable glucose monitoring and insulin delivery devices remains understudied with little expert guidance available, especially for newer autonomous systems; this is important as these more complex devices engender additional safety considerations and viable options for temporary perioperative modulation are limited. Management guidance is urgently needed, as the alternative of routine discontinuation of these devices before procedures also carries risk.

In this manuscript, we review evidence regarding wearable diabetes management devices in procedural patients with a focus on CGMs and AID systems. We present practical approaches for safe perioperative care in patients who use these rapidly evolving treatment devices.

INSULIN USE AND DEPENDENCY AMONG PATIENTS WITH DIABETES

Approximately 30% of Americans with diabetes are treated with insulin,¹⁰ which translates to approximately 6 million individuals, and the use of wearable treatment devices is expanding rapidly in patients with both T1D and T2D. The percentage of persons with T1D in the United States using insulin pumps increased from 1% in 1995 to 53% by 2017. Within this population, the prevalence of CGM usage increased from 3% in 2006 to 38% in 2017.¹¹ AID system usage is also rising rapidly, with recent data from Medtronic showing 123,000 patients with T1D using the MiniMed 670G AID system in the United States,¹² and registry data showing 33% of patients with T1D using AID systems in 2022.¹³ AID use has been shown to be safe and effective in patients with T2D, and use is expected to increase in subpopulations of patients with T2D requiring insulin therapy pending additional data.¹⁴

A critical first step in perioperative diabetes management is recognizing patients at risk of total or near-total insulin deficiency. By far, T2D is the most prevalent diabetes type, representing 90% to 95% of all cases in the population. In T2D, individuals have an insulin secretory defect, commonly on the background of insulin resistance. Though the proportion of patients with T2D using insulin is currently decreasing with the availability of of newer insulin therapies, >22% of US patients with T2D require insulin, and worldwide insulin use is expected to increase by 20% by 2030.¹⁵ Of critical importance, people with T1D, pancreatic diabetes, and some with long-standing T2D have beta-cell loss causing total or near-total insulin deficiency. People with insulin-deficient forms of diabetes require continuous exogenous insulin to prevent severe hyperglycemia, ketosis, and ketoacidosis, which can occur perioperatively if basal insulin is withheld. Ketoacidosis in T1D occurs within a few hours in the absence of insulin and regardless of glucose concentration through a lipogenesis/ketogenesis pathway. In practice, this means that individuals with T1D and other forms of insulin deficiency, such as cystic fibrosis and pancreatic diabetes, require the continuation of basal insulin by one of the following during all phases of perioperative care: (1) a long-acting insulin analog or intermediate-acting insulin neutral protamine hagedorn (NPH) given by subcutaneous injection, (2) subcutaneous insulin infusion via an insulin pump that is continued during the surgery or procedure, or (3) an intravenous (IV) insulin infusion. An IV insulin infusion can be used for the most precise intraoperative glucose management and is recommended for some procedures, especially those associated with greater duration or physiologic stress.¹⁶ Importantly, these patients might require dextrose supplementation to facilitate continuous insulin therapy while avoiding hypoglycemia.¹⁶

CGMs IN THE PERIOPERATIVE PERIOD

Most patients with diabetes monitor their glucose levels using a capillary blood glucose meter or increasingly a CGM. CGMs have become the standard of care for glucose monitoring in individuals managed with insulin therapy.¹⁷ These devices may be able to be used adjunctively with point-of-care (POC) glucose testing at surgical facilities, depending on the clinical context and institutional policies; adjunctive use implies that the CGM glucose values are not used for treatment decisions without POC glucose verification.¹⁸ Of even greater importance, CGMs are also an integral part of AID systems, which are commonly used in health care facilities depending on institutional policy.¹⁸

Preoperative Considerations

As the importance of hypoglycemia and hyperglycemia avoidance is heightened in the perioperative period, patients often partner with treating clinicians to improve glycemic control as they prepare for surgical care and CGMs can enhance the titration of normoglycemic therapies. CGMs can report an estimate of hemoglobin A1C known as a glucose management indicator (GMI), which is most precise if calculated using at least 10 to 14 days of glucose monitoring.¹⁹ Both CGM data and the GMI metric can be advantageous for preoperative optimization as recent glycemic control can be appraised much sooner than with an A1C.

Depending on the procedure site and patient positioning, individuals wearing CGM devices should be advised in advance whether they will be able to keep the device in place, and in what location. CGMs should be kept in place during the procedural period whenever possible for several reasons discussed further below, but this is particularly relevant for patients using AID systems. The majority of CGMs are US Food and Drug Administration (FDA) approved for external placement on the abdomen and/or posterior upper arm. An implantable CGM, the Eversense E3, is embedded in the upper arm with a removable transmitter attached by adhesive over the site.²⁰ Patients may be able to place CGM devices elsewhere, like the thigh, to avoid incompatibility with the surgical field, patient positioning, or potential sources of electromagnetic interference (eg, electrocautery). Placement of a CGM in an unapproved FDA location should be discussed among the patient and managing clinicians since the accuracy of the CGM may not be assured.

Intraoperative Considerations

Avoid Removing CGM Devices When Possible

Perioperative clinicians should recognize CGM devices and avoid removing them whenever possible, as removal will prevent AID systems from operating in AID mode, may disrupt postprocedural care, and may increase patient expense. Most CGMs are a single component, but some devices include 2 components: a sensor that attaches to the body and a transmitter that is clipped onto the sensor and used for 90 days. The sensor portion or combined unit is typically replaced every 10 to 14 days. CGM sensors cannot be reinserted after being removed and patients may not have access to an early replacement.²¹ Due to limitations in insurance coverage, loss of even 1 sensor can significantly disrupt care.^21,22 If a CGM transmitter is removed or inadvertently gets dislodged, it should be returned to the owner since transmitter components are reused and expensive to replace.

Avoid Compression of the CGM Device

Temporary accuracy errors due to physical compression of CGM devices have been extensively studied. Typically, compression of a CGM device leads to falsely low glucose readings. This has been observed in pediatric CGM data as there is more leniency in where these devices are placed.²³ Studies show that prolonged biomechanical pressure, as is seen during sleep, results in a physiologic response in affected tissue which alters sensor readings.²⁴ Patient positioning and padding for operative procedures or use of blood pressure cuffs may contribute to pressure on CGM sensors and adjacent tissues, resulting in glucose monitoring errors. Inaccurate CGM values from compression in the procedural period are most pertinent to patients with AID devices, as this will lead to inappropriate insulin dosing.

Interference by Fluoroscopy and Electrocautery Use

CGM devices may be unreliable in the setting of radiation exposure, specifically fluoroscopy, and generally in the operating room setting. In a small observational study of fifteen subjects undergoing coronary artery bypass surgery, glycemic data was collected both by Dexcom G6 CGMs located in the lower abdomen and POC glucose testing.²⁵ The CGMs exhibited intermittent signal loss frequently during the operative time period, hypothetically related to electrocautery interference, which was associated with a loss of sensor accuracy in those with longer signal gaps.²⁵ A second study followed 76 patients where Freestyle Libre 2.0 and Dexcom G6 CGMs were placed in the preoperative area and data compared to POC glucose values.²⁶ Sensor data loss occurred in up to 20% of CGMs; this was identifiable at the conclusion of the 1- or 2-hour sensor warm-up, was persistent, and could not be clearly attributed to an intraoperative cause.²⁶ Data obtained from the remaining patients suggested sufficient accuracy to guide intraoperative insulin therapies.²⁶ Intraoperative fluoroscopy may temporarily affect the accuracy of CGM devices although there is a paucity of studies addressing this issue. Only certain CGM models are magnetic resonance imaging (MRI) compatible.²⁷ These data highlight potential concerns about the reliability of CGM technology in the operating room and emphasize the continued need for simultaneous POC glucose testing during procedures at the present time, for the dual purposes of standard glycemic monitoring and as a verification that AID systems are registering accurate glucose data. Future studies evaluating CGM sensor location in relation to the procedural site, electrocautery return electrodes, and other intraoperative interference sources may clarify the impact of electrocautery and other intraprocedural confounders on these devices. If intraoperative reliability of these devices is demonstrated, their future use may positively influence the management of both hyperglycemia and hypoglycemia during surgery or procedures.

Limitations to CGM Data Access

Additional considerations for CGM devices in the operating room include the display and recording of derived data, though this data should not supplant POC glucose monitoring at present. In the outpatient setting, CGM data are transmitted to an integrated insulin pump, a compatible smartphone, or other receiving device. This presents a problem for some devices, since both phones and receivers are the property of the patient and frequently not allowed in procedural suites for a variety of reasons, including concerns about patient confidentiality (phone access), property loss, or device damage. Glucose telemetry, in which CGM data are received on a separate, hospital-owned platform has been piloted but is not currently available.^28,29 Innovative solutions for accessing patient CGM data on approved hospital devices while in the operating suite are an opportunity for further research.

Postoperative Considerations

Postoperatively, CGM use for integrated AID systems or for glucose surveillance may be resumed by the patient; however, continued periodic capillary glucose testing is indicated until there are assurances that the CGM is providing valid glucose monitoring data. For patients who will be admitted postoperatively, any use will need to align with institutional policies. Intraoperative loss of CGM pairing will produce a monitoring delay in the postoperative setting until the CGM reconnects. If the CGM was removed or dislodged during the procedure, the patient will need to continue capillary glucose checks until the device can be replaced. Capillary glucose testing for insulin dosing decisions in hospital settings remains the current standard of care,^30–32 meaning that hyper- or hypoglycemia detected by CGM devices must be confirmed by POC or laboratory glucose tests before insulin or IV dextrose is administered.

STANDALONE INSULIN PUMPS AND AID DEVICES IN THE PERIOPERATIVE PERIOD

Insulin pumps are small, computerized devices that deliver insulin to the subcutaneous space. Although the proportion of patients with T2D using insulin pumps is smaller compared to patients with T1D, it is rising. Perioperative clinicians are encountering more patients who use insulin pumps, and the proportion of these that are AID systems is rapidly increasing. Given their efficacy in achieving glycemic control in the outpatient setting, patients commonly express interest in using insulin pumps during procedural care.

Insulin pumps may be categorized as tubed or tubeless, and as standalone insulin pumps or integrated AID systems. With tubed pumps, insulin infuses from a reservoir or cartridge within the pump through tubing to a cannula or needle that is inserted subcutaneously. Tubeless pumps are worn directly on the body, with the cannula inserted into the subcutaneous tissue directly below the pod. Most insulin pump manufacturers recommend using insertion sites on the abdomen, flank, lower back, upper buttocks, outer thigh or back of the arm. Tubed pumps can be controlled with the device itself or a smartphone. Tubeless pumps are controlled by an external device or smartphone. Though these systems will often continue to function without the presence of the external device (eg, a smartphone), the external device may not be available in the procedure suite which may prevent visualization of device parameters. Regardless of whether the pump is tubed or tubeless, the user gives bolus doses by entering anticipated carbohydrate information, glucose values, or both on the pump. Table 1 lists currently available insulin pumps and AID systems.

Basal Insulin Delivery

Standalone insulin pumps use preset basal rate programming whereas AID systems use algorithmic dosing responses to CGM input. Both standalone insulin pumps and AID systems deliver basal insulin as a slow continuous infusion at rates as low as 0.1 units per hour and as high as 3.5 units per hour. Most patients use U100 insulin (100 units per ml) in pumps, but some receiving higher doses use concentrated insulin off-label (U200 and U500). AID systems incorporate glucose values from a CGM to modulate basal insulin infusion rates and in some cases provide small auto-correction doses. It is important to recognize that AID systems revert to pre-set basal rates if CGM pairing is lost or glucose readings are not registered, thus transitioning to function as a standalone pump. If this occurs, backup programmed basal rates may not reflect current insulin requirements for a patient. Therefore, conversion of AID systems to non-AID mode in the periprocedural period, whether intentionally or by unanticipated loss of paired CGM function, may promote significant dysglycemia.

Perioperative management principles for standalone insulin pumps have been previously published.^5,6,33–35 While the function of AID systems and possible perioperative challenges have been described,³⁶ published reports of AID systems used during intraprocedural care are limited.³⁷ Herzig et al³⁸ published a single-center study of 45 adult surgical patients with diabetes, with T1D excluded, undergoing surgeries longer than 2 hours and assigned them to AID management using the CamAPS HX AID system versus standard diabetic management, finding no adverse events and improved time in target glucose range with AID treatment. Kim et al³⁹ reported on 7 pediatric patients undergoing 8 short surgeries, where activity or exercise AID modes were used and where patients did not require clinician hypoglycemic or hyperglycemic intervention. Seget et al⁴⁰ describe 2 pediatric patients with T1D undergoing endoscopy and umbilical herniorrhaphy with AID mode continued in a temporarily higher target range; neither patient experienced adverse events or AID mode exit.

There is an urgent need for additional well-designed clinical trials evaluating the safety and efficacy of intraprocedural AID use. Comprehensive guidance for patients using these devices has not been published even though these more advanced devices now predominate. In the following sections, we update management guidance for patients using less common standalone insulin pumps and additionally provide novel management guidance for patients using AID systems (also see Figures 1 and 2) using published data relating to the unique perioperative concerns with these devices as well as management protocols in place at the authors’ institutions.

Preoperative Considerations

When preparing for surgery or procedures, a fundamental decision is whether an insulin pump will be continued without interruption or whether it would be preferable to remove the pump and switch to an alternative regimen providing basal insulin, either an IV insulin infusion,² or scheduled subcutaneous injectable insulin. Certain procedural characteristics may be problematic for insulin pump continuation such as more invasive procedures, expected use of vasopressors, or anticipated hypothermia which may compromise subcutaneous blood flow;³⁶ physiologically stressful procedures that may engender pronounced hyperglycemic responses; procedures where satisfactory device positioning may not be feasible; procedures where the device cannot be protected from contraindicated imaging/energies; or procedures where early postoperative device management by the patient will be compromised. The following is an actionable list of contraindicating procedural characteristics reflecting these factors: (1) procedures expected to last more than 2 hours,⁵ (2) procedures where current or potential device position (insulin pump or CGM) will interfere with the surgical field or patient positioning, (3) the use of electromagnetic interference sources near a pump or CGM such as electrocautery,^5,33,36 (4) electrocautery use in patients using a metal infusion cannula for a tubed pump,⁴¹ (5) expected intraoperative imaging (fluoroscopy and especially MRI),³³ or (6) concern about patient inability to manage the pump in the early postoperative period (also see Table 2).

Continuation of AID systems during surgery brings further considerations beyond those of standalone insulin pumps and is a particularly understudied area. In our experience, AID systems can be continued in most instances where maintaining a standalone insulin pump is appropriate if additional unique considerations are understood and accounted for. An important additional consideration is whether the AID will be continued in AID mode, requiring accurate and integrated CGM function, or whether it would be prudent to switch to non-AID mode (not closed loop). As discussed in the above section, the reliability of CGM input may be impacted by intraoperative factors. Changing an AID system to non-AID mode may be preferable if: (1) there are significant concerns about accurate CGM function during the procedure, (2) the CGM must be removed due to procedural site incompatibility, (3) the CGM device will be exposed to intraoperative fluoroscopy or nearby electrocautery, or (4) safe use of the AID system is dependent on an external third component (smartphone or controller) which cannot be present in the procedural area. Consulting with the patient’s diabetes clinician is strongly advised for such programming changes, especially since non-AID mode settings may not have been updated to reflect current insulin requirements. After surgery, the AID pump system user can return to the AID mode after verifying accuracy of the CGM.²⁵ It is noteworthy that while the first AID system released in 2017 was modified after FDA input to have more frequent AID mode exits to enhance safety, real-world outpatient data on the subsequent generation demonstrated that the AID mode is both safe and efficacious, requiring fewer exits from its proprietary autonomous algorithm.^42–44

When continuing an insulin pump during a procedure is considered, we suggest a checklist of prerequisites (see Table 2). For patient and procedural circumstances that do not meet eligibility for continuing an insulin pump through the procedure, the insulin pump should be substituted by an alternative insulin regimen which includes basal insulin administration (see Table 3) because many patients using insulin pumps are dependent on them for basal insulin and insulin nondependency is difficult to reliably confirm.

The American Diabetes Association recommends a target glucose range for the perioperative period of 100 to 180 mg/dL,³¹ with other societies advocating similar hyperglycemia treatment thresholds,² while the recent Society for Ambulatory Anesthesia consensus statement advocates for allowing a more permissive approach to hyperglycemia in minimally invasive ambulatory procedures of short duration with a threshold of 250 mg/dL before initiating insulin treatment.⁵ AID systems are designed to achieve near-normoglycemia, and thus these systems target glucose values ranging 100 to 160 mg/dL, depending on the brand and shared decision-making between patient and clinician. Most AID systems have an activity or exercise mode that targets higher glucose concentrations (see Table 1), which may be beneficial in the periprocedural period to minimize the potential for hypoglycemia. While smaller ambulatory procedures are the most appropriate context for intraprocedural AID continuation, currently available AID settings will target glucose values well below 250 mg/dL, even when more lenient activity or exercise modes are active (Table 1). Promoting glucose values approaching 250 mg/dL in patients using AID systems could be problematic or even dangerous, not only because this fundamentally counteracts AID algorithms but also because it may entail insulin restriction, leading to diabetic ketoacidosis in at-risk individuals.⁴⁶ Recommendations for preprocedural planning are outlined in Figure 1.

Blood glucose should be checked hourly starting on facility admission to guard against hypoglycemia from caloric restrictions and masking of dysglycemic symptoms when sedating drugs are administered. Correcting high or low glucose may be needed in the preoperative care unit on admission to achieve acceptable levels. If admission glucose is 70 to 100 mg/dL, patients with standalone pumps can often decrease the basal rate (eg, by 20%) using a temporary basal rate, if not already in place. However, for individuals with an AID system in AID mode, insulin delivery will be autonomously adjusted by the pump algorithm and will generally rectify hypoglycemia in this range. If the preoperative glucose is less than 70 mg/dL, the patient should receive IV dextrose, and the glucose monitored more frequently. Smaller doses of dextrose may be appropriate in patients with AID systems, as closed-loop insulin dose adjustments will contribute to normalizing blood glucose. If the blood glucose is higher than 250 mg/dL in the setting of a standalone insulin pump or AID pump in non-AID mode, a small dose of insulin can be administered subcutaneously or intravenously followed by more frequent monitoring. Administration of exogenous insulin to patients with AID systems running in AID mode is generally discouraged as the pump self-regulates insulin delivery and cannot recognize or account for insulin delivery by alternative routes. Regardless of whether the patient has a standalone insulin pump or an AID system, hyperglycemia of 250 mg/dL or greater should prompt concern that insulin delivery via the pump is compromised and an assessment for diabetic ketoacidosis may be necessary depending on the clinical context. Finally, before proceeding with the procedure, the insulin pump and CGM should be inspected for any evidence of kinking or disruption.

Intraoperative Considerations

When insulin pumps are continued into the intraprocedural period, the procedural team will need to be aware of component positions to ensure intraprocedural accessibility and to avoid pressure, tube kinking, or dislodgement during the procedure. If dexamethasone is used for prophylaxis of postoperative nausea and vomiting, a lower dose of 4 mg is suggested to minimize related hyperglycemia.⁵ POC glucose should be monitored hourly,⁴¹whether the patient is receiving insulin pump therapy or a replacement regimen. Standalone insulin pumps administer a constant dose and will not adapt to intraprocedural dysglycemia, necessitating regular monitoring and clinician-administered correction treatments as needed. Though AID systems will adapt to detected dysglycemia, appropriate AID dosing is dependent on accurate CGM function. Intraprocedural factors influencing CGM function are discussed previously and are critically pertinent to intraprocedural AID use. These considerations underscore the necessity of continued regular POC glucose monitoring during intraprocedural care of patients who use standalone insulin pumps or AIDs.

For patients with standalone insulin pumps or AIDs operating in a non-AID mode, subcutaneous or IV insulin should be considered for glucose 180 to 250 mg/dL and the patient monitored for suspicion of insulin pump malfunction. We recommend against perioperative clinicians administering bolus insulin doses via an insulin pump during the intraprocedural period (without active patient participation) because of potential dosing errors. For patients with AIDs operating in AID mode, the system will adapt to detected hyperglycemia and clinicians should not administer additional corrective therapy. However, hyperglycemia detected with POC glucose monitoring (eg, >250 mg/dL) should raise suspicion for inaccurate CGM data or AID system dysfunction, which would require transitioning to an alternative insulin regimen.

Intraoperative scenarios may arise that mandate insulin pump discontinuation and switching to an IV insulin infusion or subcutaneous regimen (see Figure 2) and patients should be informed of this potential, such as marked or persistent dysglycemia, requirement of high doses of vasoactive agents, hypothermia, shock, or unplanned component disruption.³⁶ Protocols that guide alternative basal insulin therapy (see Table 3) should be in place for such scenarios (accompanied by treatment of any hypoglycemia with IV dextrose as needed). Simple and effective methods of discontinuing insulin pump therapy in urgent scenarios are listed in Table 1.

Postoperative Considerations

While the patient is recovering in the postanesthesia care unit (PACU), POC glucose measurements should be obtained hourly until the patient is awake and able to resume management of his or her insulin pump. Any temporary basal rates or alternative glucose targets can also be reset once concerns about perioperative hypoglycemia have abated. Similar to other phases of care, uptrending blood glucose levels, especially >250 mg/dL, should prompt concern about pump system integrity. In this scenario, triaging and treating dysglycemia with the patient is appropriate if he or she is capable. If the patient is unable to participate in device interrogation, then transitioning to an alternative of insulin delivery such as scheduled injectable insulin is preferred.

When inpatient hospitalization is anticipated after a surgical procedure, we recommend an inpatient consultation with a diabetes consult team to assist with pump management when available. Hospital policies often govern the use of CGMs, standalone insulin pumps, and AIDs.¹⁸ Conditions for inpatient pump use vary between institutions, but typically include the following: (1) the patient is not cognitively or otherwise impaired, (2) the patient is capable of operating the insulin pump, (3) the patient has supplies to continue pump use, and (4) the patient is not in a critically ill state where pump use is contraindicated. If an intensive care level of care is required, the insulin pump should be discontinued, and an IV insulin infusion should typically be initiated.³¹

CONCLUSIONS

CGMs and insulin pumps with AID algorithms are revolutionizing diabetes management for persons requiring insulin therapy, with AID systems becoming predominant, and related algorithms and devices evolving rapidly. While these devices improve glucose control in the outpatient setting, they may be unsuitable for use during some procedures. Perioperative clinicians and endocrinologists should be aware of potential problems limiting device accuracy or safety and, when AID systems or standalone pumps will not be continued through surgery, make alternative plans for meeting basal insulin requirements. When patients and perioperative clinicians agree that these devices can safely remain in place during procedures, advance planning should take place, including compatible site placement(s) and the potential use of temporary glucose targets. Currently, patients with CGM devices or who use insulin pumps should continue to have POC glucose monitoring performed throughout the procedural period due to lingering concerns about intraprocedural CGM accuracy and connectivity. Intraoperative disruption or dysfunction of CGM devices and insulin pumps may occur and the perioperative team should monitor for such problems and be prepared to institute replacement insulin therapies.

In conclusion, using a systematic approach based on procedural factors and the phases of perioperative care can allow the safe perioperative continuation of CGMs, standalone insulin pumps, and AIDs for many shorter procedures. Awareness of potential pitfalls that may affect these patients and the availability of contingency insulin delivery therapies is important to avoid the complications of ketoacidosis, hyperglycemia, and hypoglycemia. There is an urgent need for further clinical studies reporting the outcomes of patients undergoing procedures with AID systems, particularly since these newer devices have already become standard practice for patients on insulin pump therapy.

REFERENCES