Abstract
Much progress has been made in diabetes treatments since the first dose of insulin was administered in 1921. However, a truly transformational moment in diabetes care occurred when urine testing gave way to capillary blood home glucose monitoring. As improvements were made to these devices, continuous glucose monitoring (CGM) was introduced. The advantages of experiential learnings gleaned from seeing continuous real-time data have been borne out in numerous peer-reviewed journals. Limitations to use of CGM include patient's level of numeracy and literacy, development of alarm fatigue, interfering substances leading to erroneous readings, high rates of discontinuation, and poor reimbursement.
I
Since 1978 there have been dramatic improvements in blood glucose monitoring devices. Smaller, more intuitive devices with historical and predictive analytics and most recently meters embedded with intelligent dosing algorithms have crowded the pharmacy shelves. Incremental improvements in accuracy have further demonstrated their usefulness in identifying trends that help titrate insulin. Furthermore, T1D Exchange, a non-profit research organization with one of the largest type 1 diabetes (T1D) registries in the world housing data from over 27,000 well-characterized patients, demonstrated that the frequency of SMBG testing directly correlates with the hemoglobin A1c (HbA1c) level. 2
Despite these advancements in episodic glucose testing, the process of connecting the dots and establishing relatedness of blood sugar levels throughout the day is akin to the challenges of seeing a frame of a movie every 30 min and attempting to reconstruct the movie storyline. Patients often note “my blood sugar went from x to y” a few hours later, suggesting a straight line could be drawn between the two data points.
With the advent of real-time continuous glucose monitoring (CGM), the need to connect the episodic SMBG data points is no longer required. Experiential learning while wearing CGM devices has for many people resulted in improved glycemic control and quality of life. Many articles have been published in peer-reviewed journals that have demonstrated significant improvements in HbA1c, time in range, and decreased hypoglycemia in diverse populations, in multiple geographies, and across the spectrum of baseline HbA1c. 3 –15
Although CGM is currently labeled for adjunctive use to SMBG only, its accuracy now parallels or surpasses those of SMBG devices, and it is hoped that non-adjunctive use will soon be the standard. Improved accuracy and fidelity of new-generation sensors will undoubtedly potentiate the success of automated insulin delivery, as will more accurate insulin delivery devices.
Finally, I agree with many of my colleagues who when asked how they would advise someone with T1D stranded on a desert island when forced to choose between an insulin pump or a CGM device, they would almost universally recommend choosing CGM and multiple daily insulin injections. Insights gained from CGM can inform individuals on insulin therapy, regardless of the mode of delivery.
Despite the incontrovertible evidence of the safety and efficacy of CGM use, numerous barriers exist to adoption, adherence, and effectiveness in many patients and segments of the population. These considerations and limitations must be taken into account when designing an appropriate educational program and realistically level-setting expectations in the individual for whom CGM is recommended. In this issue, Picard et al. 16 report the results of an observational study demonstrating that a 6-month-long multidisciplinary education program on sensor-augmented pumps in patients with type 1 diabetes improved metabolic control with a high level of adherence and satisfaction. Their approach is one example of tailor-made education, albeit requiring resources and time not necessarily available in all clinics in all geographies.
Levels of literacy and numeracy in people with diabetes vary and can impact significantly on the ability to interpret trends in numbers and their significance. 17 Typical CGM devices generate a value in milligrams per deciliter every 5 min, or roughly 300 times a day. Without specific and thoughtful dialogue and an assessment of a patient's numeracy, potential remediation and educational opportunities can be missed. A cycle of frustration, disappointment, embarrassment, and failed adherence leading to discontinuation is a likely outcome. Inherent in the successful use of CGM above and beyond numeracy is the ability of the user to, in real time and almost intuitively, draw on experience, trial and error, memory recall, and integration of diet, activity and stress when interpreting those values either in retrospective download or in real time. However, it is important to note that although patients may successfully download their own data to review, there have been no studies to substantiate improvement in any outcomes.
CGM devices are intentionally designed to alert users to changes in glucose values. Separate alarms and alerts can be programmed to sound when absolute glucose thresholds are breached or when rate of change limits are exceeded. Most alarms can be silenced by the user, set to vibrate, or turned off entirely. When an individual experiences frequent alarms, he or she may silence the alarm, change the parameters in favor of higher and lower glucose values, or worse, ignore them entirely. So-called alarm fatigue is a well-recognized phenomenon that can lead to a “betrayal of trust” in the device, especially when the alarms are false-positives (nuisance alarms). This topic was reviewed by this author in a previous publication. 18
Unstable and therefore unreliable CGM signals can occur due to several factors. Maahs et al. 19 demonstrated falsely elevated CGM values compared with SMBG after ingestion of paracetamol in an outpatient setting. This is consistent with the known generation of a nonglucose signal by paracetamol's phenolic moiety oxidizing at the sensing electrode. 19 Ascorbic acid and salicylate ingestion, and other causes of in vivo measurement error include lag time between capillary blood glucose and the interstitial space, trauma at the site insertion, changes in perfusion at the sensor site, and physical pressure placed on the sensor. 20
Accuracy and discomfort are major causes of discontinuation of CGM use. T1D Exchange clinic registry published data on the results of a survey indicating 41% of participants who reported using CGM (defined as use of real-time CGM within the past 30 days) on enrollment into the registry reported discontinuing within a year primarily because of discomfort or technical problems with the device. Forty-two percent felt discomfort when wearing the sensor, 33% reported problems inserting the sensor, 30% had problems with the adhesive, 28% had problems with the sensor working properly, 27% experienced too many alarms, 25% were concerned with the accuracy of the CGM data, 18% complained of interference with sports, and skin reactions were reported in 18%. Not surprisingly they also found that patients could maximize their benefit by wearing the devices more frequently and using them in real time. 21
No discussion about limitations to usage of diabetes devices is complete without addressing the lack of adoption due to poor reimbursement. Although there have been positive coverage decisions for CGM devices from a number of payers, CGM remains non-reimbursed by Medicare. 22 Additional challenges include cessation of coverage in individuals already on CGM who cross into the Medicare age group or those patients who are on dialysis. These two groups of patients will essentially need to discontinue their use of CGM (Irl Hirsch, personal communication).
Although the evidence substantiating the effectiveness in safely reducing HbA1c levels in individuals with diabetes is incontrovertible, significant barriers exist that limit effectiveness, adherence, access, and reimbursement. Precision medicine has ushered in a new era that demands novel ways to approach and overcome these limits by moving us away from a one size fits all approach. Through iterative improvements in devices, further investment in education programs and decision support tools, some of these obstacles and limitations to use will be overcome. In the meanwhile, when I am asked “which CGM device do you most often recommend?”, my response is always “it's the one you will agree to wear.”