Remote Pharmacist-Assisted Flash Continuous Glucose Monitoring Improves Glycemic Outcomes in Patients With Poorly Controlled Diabetes: A Retrospective Case Series

In February 2019, an international consensus panel met to establish glycemic metrics and parameters for assessing and using CGM data in clinical practice. The percentage of time in range (TIR), time below range (TBR), and time above range (TAR) were identified as key metrics (Table 1) (15).

As reported by Beck et al. (16), each incremental 10% increase in TIR equates to a reduction of ∼0.5% in A1C. To mitigate the risk for hypoglycemia, the panel advised that TIR and TBR be used as a composite metric, focusing primarily on reducing TBR while increasing %TIR, which would, by default, decrease TAR. An additional metric was glycemic variability, presented as the percentage of coefficient of variation (%CV), with the target being ≤36%. To accurately interpret the data, it is recommended that the CGM should be active ≥70% of the time, over a minimum time of at least 14 days (15,17–19).

Another new metric identified by the panel was the glucose management indicator (GMI), to replace the previously used eA1C (estimated A1C) metric. Although the GMI value strongly correlates with laboratory A1C values in ∼19% of patients, the values can differ by 0.3% in 51% of patients and up to 0.5% in 28% of patients (20). However, because the difference between A1C and GMI values remains stable in each patient, clinicians can compensate by adjusting each patient’s target A1C accordingly (20). For example, if a patient’s A1C target is 7.5% and the GMI is consistently 7.9%, it may be advisable to lower the patient’s target A1C (e.g., to 7.2%) to minimize excessive hyperglycemia.

The consensus panel group also adopted a one-page standardized report known as the ambulatory glucose profile (AGP). The AGP report incorporates all of the core CGM metrics, presented in both numerical and graphical formats. Originally created by Mazze et al. (21) and further developed by the International Diabetes Center (22), the AGP template has been integrated in the data-downloading software provided by all CGM system manufacturers.

The following cases illustrate how the AGP component of the LibreView software program (Abbott Diabetes Care, Alameda, CA) was used to assess glycemic status and adjust therapy in patients with type 2 diabetes seen in our pharmacy practice. The software is a secure Cloud-based diabetes management system that is used to download and analyze data from the flash CGM systems.

A.B. is a 61-year-old man with an 18-year history of type 2 diabetes and an A1C of 13.4% despite his multiple daily injection (MDI) insulin regimen and performance of BGM four times daily. His medications included insulin degludec 30 units twice daily, insulin lispro 20 units three times daily and 25–30 units at night as needed, ramipril 5 mg twice daily, amlodipine 5 mg daily, and nadolol 80 mg twice daily. He reported performing BGM at least three times daily, with glucose levels of 190 to >595 mg/dL (10.6 to >33 mmol/L) before breakfast, 185–538 mg/dL (10.3–29.9 mmol/L) before lunch, and 259–570 mg/dL (14.4–31.7 mmol/L) before supper.

His history revealed two myocardial infarctions, nonalcoholic fatty liver disease, esophageal varices, obesity, hypertension, hyperlipidemia, depression, and renal disease. His cardiovascular issues were complicated by mild to moderate kidney disease characterized by an albumin-to-creatinine ratio of 180 mg/dL and an estimated glomerular filtration rate of 49 mL/min/1.73 m². A.B. is physically disabled and reported a sedentary lifestyle.

During the initial visit on 3 December 2019, it was decided that the goal of therapy would be to achieve an A1C <7.0% within 3–6 months without undue risk of hypoglycemia. However, considering A.B.’s comorbidities—particularly diabetic kidney disease and established cardiovascular disease—the decision was made to initiate the cardioprotective glucagon-like peptide 1 (GLP-1) semaglutide initially, followed by a cardioprotective sodium–glucose cotransporter 2 agent at a later date. This strategy would be challenging given A.B.’s current state of glycemic variability, as evidenced by his BGM values, erratic diet and lifestyle habits, and use of an MDI insulin regimen.

We created an action plan that was simple and facilitated safe and effective titration of insulin, while adding both cardioprotective agents in a step-wise manner. A.B.’s 30-unit twice-daily dose of insulin degludec was changed to 60-unit once-daily administration, a more appropriate dosing schedule based on the pharmacokinetics of this insulin formulation. Once-weekly semaglutide was added to his regimen without an initial reduction in insulin dose because of his current state of persistent hyperglycemia without any hypoglycemia. Concurrently, he was educated on the proper dosing of prandial insulin and was provided with a sliding scale to allow for corrections of higher preprandial glucose levels.

Of course, insulin adjustments in someone who is unwilling to adhere to dietary advice is not without significant risk of hypoglycemia. A.B. had been performing BGM regularly, but he was asked to increase his BGM frequency to four times daily and to keep a detailed log of his dietary choices. Although he was adherent despite the significant burden of performing BGM so often, it became clear that both A.B. and his health care provider were still not able to visualize the true state of his daily glycemic profile. On 30 December 2019, the decision was made to provide A.B. with a 2-week trial of the flash CGM. This trial would allow the clinician to follow A.B.’s glycemic data remotely via the flash CGM system’s downloading software and would also provide A.B. with significantly more information regarding the effects of his dietary choices on his blood glucose levels.

Within 2 weeks of starting flash CGM, A.B. already started to show significant improvement of his glycemia, which was a result of dietary changes he implemented after visualizing his blood levels, as his insulin dosing was not changed during the initial 2-week period of flash CGM. He reported that he would avoid snacking when his glucose was already elevated and that he had tried to eat less at meals. His initial AGP report (Figure 1A) revealed that although only 30% of his glucose values were within his target range, and 70% of values were above his target range (a), there was still no evidence of hypoglycemia (b). His GMI value of 8.4% (c) indicated an estimated A1C that was quite a dramatic improvement from his initial A1C of 13.4%. His glucose was fairly stable, with relatively low glycemic variability (d), but he still had persistent hyperglycemia throughout the day and night (e).

We continued his titration schedule of semaglutide and adjusted both his basal and bolus insulin accordingly using the CGM data. On 7 February 2020, A.B. reached the full dose of 1.0 mg semaglutide, and canagliflozin 100 mg was added to his existing regimen to provide both renal and cardiovascular benefit, as well as modest glucose lowering.

In May 2020, a newly measured laboratory A1C was 7.7% within only about 3 months of the medication changes. We were able to review our AGP report data for the past 90 days to more effectively examine the utility of his reported A1C (Figure 1B). With the use of these data, we were able to determine that his glucose had been reduced safely, as evidenced by 0% TBR (a), and although his TIR was 47%, there were significant improvements in almost all glycemic markers, including his GMI (b) and glycemic variability (c), with a relatively flat 24-hour glucose profile (d). We were able to successfully adjust his insulin doses effectively and add two important cardioprotective glycemic medications while achieving a remarkable 5.7% reduction in A1C without any hypoglycemia.

A.B.’s continued use of flash CGM in conjunction with remote monitoring led to the safe achievement of an A1C of 7.2% on 25 August 2020, which was a very positive outcome.

Use of flash CGM allowed us to make more informed decisions about therapy adjustments for A.B. Remote monitoring via the data downloading software facilitated more timely intervention and counseling. Importantly, with the improved ability to visualize his changing glucose in response to meals and snacks, A.B. gained valuable insights into how his dietary choices affected his glycemic control. Moreover, he experienced a sense of empowerment and engagement with his self-management regimen as his glycemic data improved over time.

N.H., a 42-year-old man with a BMI of 35.5 kg/m², was referred to our pharmacy with newly diagnosed type 2 diabetes, with signs of metabolic decompensation, including obesity, hypertension, and hyperlipidemia. He presented with an initial A1C of 10.2%.

To address his acute glucose toxicity, we initiated treatment with insulin glargine (10 units daily) and metformin (500 mg twice daily), according to current clinical guidelines (23). Taking into consideration his age, elevated BMI, and short duration of diabetes, N.H. agreed to a therapeutic A1C goal of<6.5%, while striving for a target weight loss of 5–10% within the next 3–6 months with minimal hypoglycemia risk. However, he was overwhelmed with his diabetes diagnosis and very apprehensive about starting insulin because of a fear of hypoglycemia.

In response to his concern about potential hypoglycemia, he was offered the immediate use of flash CGM and was assured that his glucose could be monitored as frequently as he would like and that the pharmacy would be able to remotely monitor his glucose levels to ensure his safety as he up-titrated his insulin dose, which was done in increments of 1 unit per day until he reached a fasting blood glucose goal of 99 mg/dL (5.5 mmol/L). Because he was overwhelmed during our initial visit, we concentrated on familiarizing him with the flash CGM system and provided him with instructions on how to administer his insulin properly. His first dose was administered in the pharmacy to help alleviate his concerns about self-administering injections, and he was asked to scan the sensor to check his glucose levels at least once every 8 hours to ensure close to 100% data capture. To avoid “information overload,” no specific dietary advice was offered at the initial visit beyond reminding him of the impact of sugary drinks, which he had been having to help him deal with his hyperglycemia side effects.

By reviewing flash CGM data within the first 2 weeks, we were not only able to assess the effects of the new insulin regimen, but also to see the patient’s dietary response to higher blood glucose values after meals (Figure 2A). The 5th and 95th percentiles illustrated in the glucose profile were used to detect significant hypoglycemic events (a). Remote monitoring indicated that N.H. made significant dietary changes within 8 days of initiating flash CGM, which led to significant reductions in his post-lunch glucose spikes (b).

Within the first 14 days of flash CGM use, the prescribed insulin regimen and patient-driven dietary changes had a dramatic impact on overall glycemic control, with 92% of the glucose values within the target range (c), 0% below the target range (d), and excellent control of glycemic variability (%CV 25.2%) (e). The only detected hypoglycemia occurred on day 14 when N.H. skipped his usual lunch and had only a snack (f). Considering our original composite goal of achieving optimal glycemic control while striving for weight loss with minimal risk of hypoglycemia, we discontinued the insulin and transitioned him to once-weekly semaglutide (starting dose 0.25 mg). With this adjustment, N.H. managed to surpass his original glycemic target safely and effectively (Figure 2B). Within 3 months of starting flash CGM, he had achieved a laboratory-measured A1C of 6.0%, which was identical to his GMI value (a) and continued to reduce his glycemic variability (b), which was clearly illustrated in the glucose profile (c). An improvement in A1C of 4.2% is usually associated with weight gain, particularly when initiating insulin. However, N.H. was able not only to exceed his glycemic target, but also to lose 5 lb. GLP-1 receptor agonist therapy was only initiated 1 month before his laboratory A1C result of 6.0%. By the time he reached his full 1.0-mg weekly dose, he had lost a total of 15 lb (6.8%), while achieving his A1C goal of <6.5% with no hypoglycemia.

Because the usual method of titrating basal insulin can lead to overbasalization if the basal insulin dose is adjusted based on fasting blood glucose values, use of remote monitoring and frequent telemedicine visits allowed us to closely monitor this patient’s progress and keep him within his target range without hypoglycemia. Using additional metrics within the AGP report, we validated that this level of glycemic control was safe, with a TBR of 1%, and highly consistent, with a TIR of 98% and an impressive AGP over the 3-month period.

This approach also helped reassure N.H. provided valuable insights into the role of dietary choices and lifestyle modification in modulating glycemic activity. These benefits allowed N.H. to focus on weight loss and continue to pursue healthy behaviors. Moreover, the decision to initiate flash CGM immediately eliminated the additional burden of performing frequent BGM. We believe this is particularly useful in patients who are still dealing with the emotional turmoil of their diabetes diagnosis.

S.M., a 54-year-old woman who smokes and has a BMI of 47 kg/m², presented with a 4-year history of type 2 diabetes complicated by comorbid conditions, including hypertension, hyperlipidemia, and fibromyalgia. Her A1C was 10.0% despite the use of a long-acting basal insulin (35 units daily), semaglutide (0.5 mg once weekly), and dapagliflozin/metformin (5/1,000 mg twice daily). She was referred to the diabetes educator pharmacist for potential initiation of prandial insulin because of the ineffectiveness of her current therapy regimen.

It was noted that significant weight loss efforts were ongoing, with the recent initiation of both bupropion/naltrexone and once-weekly semaglutide. S.M. admittedly did not perform BGM, but she agreed to bring four daily blood glucose tests (before meals and at bedtime) for 3 days to her initial appointment. Her baseline glucose levels from these BGM results ranged from 283 to 396 mg/dL (15.7 to 22.6 mmol/L) during the day. S.M. made a point of mentioning that she would not continue with BGM because of the pain and discomfort associated with it. She also said she did not think that BGM mattered because she was not making any therapeutic changes related to her blood glucose values.

Based on her initial glucose values, her basal insulin was increased conservatively to 42 units. Prandial insulin was not initiated because, although it could certainly alleviate her current hyperglycemia, it would undoubtedly have a negative impact on her weight loss efforts. Moreover, our inability to ensure that appropriate BGM would occur made the risk of hypoglycemia too great. The contribution of her dietary intake to her overall glycemia was likely significant, and considering her weight loss journey, she needed to be made aware of the impact of her eating habits on her diabetes management.

Flash CGM was implemented to not only evaluate her glycemic profile, but also allow S.M. to experiment with diet and lifestyle changes to determine whether she was able to improve her blood glucose without the addition of any other pharmacologic agents. She was advised to closely monitor and minimize postprandial glucose excursions through portion control and healthier food options. In our follow-up consultation, S.M. showed strong engagement by scanning frequently (∼23 scans on day 2). Much of her scanning occurred around eating times, which revealed the glycemic effects of food intake and thus enhanced her understanding of the glycemic impact of meal composition and portion size. Based on this insight, S.M. modified her eating pattern to include more nonstarchy vegetables and refrained from eating heavy late-night meals.

The improvement in her glycemic profile was both impressive and immediate, as seen with her first day of using flash CGM (Figure 3). Once she started using the flash CGM system, she made changes to her diet immediately. She had been in the habit of eating snacks throughout the day, but was able to see immediately that she did not need snacks despite being on basal insulin to avoid hypoglycemia.

Her initial 2-week AGP report showed notable improvement (Figure 4). Remarkably, this improvement was achieved almost entirely as a result of her increased awareness of the effects of diet and lifestyle behaviors. S.M. achieved an impressive TIR of 81% within 14 days of starting flash CGM (a) and steady improvement in daily glucose profiles (b) by increasing her physical activity and changing her food intake. She had an improved GMI (c) and reduced glycemic variability (d), indicating a potential need for reduction in her basal insulin dose. This result was quite surprising given that she had been referred initially for the possible addition of prandial insulin.

Despite occasional bouts of relapse due to poor dietary habits, pharmacist-assisted remote monitoring provided encouragement and advice as needed. S.M. continued using flash CGM and having periodic virtual follow-ups, which led to a remarkable improvement in her glycemic parameters. Down-titration of insulin, along with weight reduction and continued attention to the glycemic impact of meals, helped her attain an A1C of 7.6% on 26 January 2021. Although she had yet to achieve her A1C goal of <7.0%, she was able to lose 36 lb (14% weight loss) and was able to discontinue her basal insulin.

Although S.M. had not been successful in achieving an A1C level <7.0% within 6 months of intervention, we were encouraged from examination of her AGP profile for the 4 weeks preceding the date of the A1C collection. Unlike A1C, which only measures the degree of glycation of red blood cells over the past ∼3 months, the AGP report can be used to assess a specific period to give a more acute assessment of a person’s current level of glycemic control. At our follow-up consultation in February 2021, the GMI on S.M.’s AGP report (Figure 5) showed that she was doing far better than her most recent A1C suggested. In fact, it showed that she was achieving all the glucose metric goals, with a GMI of 6.9%, TIR of 81%, TBR of 0%, and glycemic variability of 27.3%. This assessment of her most current glucose control assured both the patient and the health care professional that she was, in fact, reaching the glycemic goals we had set, even if her A1C value had yet to demonstrate the full impact of all of her hard work.

Although S.M. had occasional bouts of dietary “relapse,” her frequent scanning of the flash CGM sensor (up to 23 times per day) indicated that she was very engaged in her self-management. Moreover, our use of remote monitoring provided encouragement and advice as needed. She was finally able to gain control of her diabetes and, at the same time, lose a significant amount of weight. This experience was empowering to her, as the majority of her improvement in glycemic control and health was the result of her own actions.

As demonstrated in the case studies presented above, the use of flash CGM can be foundational in developing and implementing diabetes care plans that meet the individual needs and concerns of each patient. This approach facilitates collaborative goal-setting and person-centered diabetes management by providing comprehensive, actionable data to inform therapy decision-making, support healthy lifestyle modification, and enhance patients’ understanding of their disease.

The establishment of CGM metrics and the standardization of data reports has enabled clinicians to quickly identify a patient’s frequency and magnitude of glucose excursions and initiate appropriate therapy changes to address problematic glycemia, resulting in greater efficiency and improved quality of care. For patients, the ability to obtain immediate feedback on how their medications and lifestyle behaviors affect their glucose enhances understanding of their disease, resulting in greater treatment satisfaction and adherence to the prescribed regimen. Studies have shown strong associations among optimal treatment adherence, improved glycemic control, and decreased health care resource utilization (24,25).

Another significant benefit of CGM functionality is the ability to transmit data to clinicians, who can remotely monitor patients’ glycemic status, determine whether therapy adjustments are needed, and then provide feedback by communicating their recommendations to patients via telehealth consultations. Numerous studies have demonstrated the efficacy of remote monitoring and virtual visits in improving glycemic control (26–31) and enhancing treatment adherence (32). Once considered to be more futuristic than practical, use of remotely monitored CGM data in conjuntion with telemedicine consultations is now emerging as a new standard of diabetes care (33). The successful use of these technologies throughout the coronavirus disease 2019 pandemic further highlights their value and clinical utility (34–38).

In summary, these case studies underscore the importance of understanding and using glycemic metrics for a comprehensive evaluation of glycemic activity, which cannot be determined using BGM or A1C alone. Use of flash CGM in combination with lifestyle modification, patient education, and appropriate medications provides clinicians the tools they need to develop individualized treatment plans that help patients improve glycemic outcomes and facilitate diabetes self-management.

The author thanks Christopher G. Parkin, MS, of CGParkin Communications, Inc., for editorial assistance in writing the manuscript.

Funding for the development of this article was provided by Abbott Diabetes Care.

No potential conflicts of interest relevant to this article were reported.

As sole author, R.S. is the guarantor of this work and takes responsibility for the integrity of the case studies presented and the overall accuracy of the content.