Acceptability of and Barriers to Continuous Glucose Monitoring in Individuals with Type 2 Diabetes Mellitus Receiving Chronic Hemodialysis: Research Letter

Abstract

Continuous glucose monitoring (CGM) in individuals receiving hemodialysis is more accurate than hemoglobin A1c (HbA1c) and may improve diabetes management in this population, but little is known regarding patient perspectives on CGM technology and its use in clinical outpatient hemodialysis care. We aimed to explore the real-world acceptability of CGM use in nonhospitalized individuals receiving hemodialysis. This pragmatic prospective observational study was nested within a qualitative improvement initiative that embedded an endocrinologist providing care for people with type 2 diabetes mellitus and HbA1c greater than 8% within the hemodialysis unit. As part of this initiative, some people seen by the endocrinologist were prescribed CGMs. These individuals were approached to participate in the study and followed for 2 weeks during CGM wear. Reasons for CGM decline were collected and collated inductively using content analysis. Of 47 consults, only 22 (47%) people were prescribed CGM. Of these, six (27%) declined CGM use and only six (27%) consented to the study, but ultimately only three participants (13.6%) completed the study. Reasons for CGM decline included lack of comfort with technology, cost and poor cellphone, internet and data availability. Our findings suggest that further exploration of barriers to CGM use and solutions to these barriers is required prior to broad use of CGM in real-world outpatient hemodialysis settings.

Introduction

People with type 2 diabetes (T2DM) receiving hemodialysis have high mortality rates which correlate with glycemic variability.^1,2 Given the unreliability of hemoglobin A1c (HbA1c) levels in hemodialysis settings, alternative glycemic monitoring strategies are needed. ³ Glucose measurements derived from continuous glucose monitoring (CGM) in people receiving hemodialysis correlate closely with capillary glucose levels and are more accurate than HbA1c.^3,4 Continuous glucose monitoring has potential to personalize and improve diabetes management. ⁵ Yet, published studies of CGM in people receiving hemodialysis are small and frequently involve hospitalized individuals. ⁵

We aimed to explore the real-world acceptability of CGM use in nonhospitalized individuals with T2DM receiving maintenance hemodialysis.

Research Design and Methods

Our prospective, observational cohort study was embedded within the Can-SOLVE CKD Network Triple I—Hubs of Care Project in Winnipeg, Canada. This quality improvement initiative implemented an endocrinologist in the Seven Oaks General Hospital outpatient hemodialysis unit from January 3 to August 31, 2023. Clinical indications for individual referral to the endocrinologist during this quality initiative included a confirmed diagnosis of T2DM, HbA1c greater than 8%, and individual agreeable to referral and in-unit endocrinology consultation. Individuals fulfilling these criteria were referred to the endocrinologist at the discretion of the attending nephrologist in the unit.

One endocrinologist provided initial consultations and follow-up visits related to these referrals. During these consultations, FreeStyle Libre 2 sensors, a CGM system, were offered to select individuals with T2DM. The endocrinologist prescribed CGM to individuals on basal bolus insulin to obtain information about the pattern of their blood glucose levels, but also used CGM strategically in individuals who were not on insulin to gain insight into glycemic patterns, particularly in individuals who did not self-test at home (eg, an individual who had a high HbA1c, but reported “good” glucose levels). Only individuals who were offered and/or prescribed a FreeStyle Libre 2 sensor by the in-unit endocrinologist were approached to participate in this study. Additional eligibility for the CGM study included willingness to wear a CGM sensor and ability to communicate in English. There were no absolute exclusion criteria.

We adhered to ethical conditions outlined in the University of Manitoba Health Research Ethics Board certification (H2023:022) and the Declaration of Helsinki. ⁶ All study participants signed written informed consent.

At the time of the study, the cost of CGMs was not covered by the single payer universal healthcare system in Manitoba. Participants were provided FreeStyle Libre 2 sensors free of charge and received instructions for and help, if required, with sensor application and use from the study research coordinator and endocrinologist. Sensors were worn for two weeks and glycemic data was shared with healthcare providers through LibreView, a secure online data sharing platform. ⁷ Hemodialysis treatment records of consenting individuals were reviewed for demographic and clinical data. Reasons for CGM refusal and discontinuation were collected. We also planned to assess change in weekly interdialytic fluid gains four weeks before and after CGM wear (hemodialysis records) and glycemic index (LibreView), but final study sample size was less than six and this analysis was not performed.

Results

Of the 244 individuals with T2DM who dialyzed in the hemodialysis unit, only 63 (26%) had baseline HbA1c > 8 and were eligible for endocrinology referral. Of these, 47 (78%) received endocrinology consultation during the six months of the quality improvement project. Of those individuals who received an endocrinology consult 47% (n = 22) were offered/prescribed a CGM. Mean age of those who were offered a CGM was 63 (SD 12) years and 50% were male.

Sixteen individuals (72%) were agreeable to CGM use and six declined due to: concern wearing medical device on body (n = 1), reluctance to use a new technology (n = 1), no reason provided (n = 4).

Of those agreeable to CGM, one died and four declined to be contacted about the study.

The remaining 11 individuals were approached regarding study participation. Five became uninterested in CGM and/or study participation due to: confusion surrounding technology (n = 1), financial coverage (n = 1), refusal to sign consent (n = 1), and two individuals provided no specific reason.

In total, only six individuals (38%) willing to wear CGM consented to study participation. These participants had a mean age of 61 (SD 11) years, 50% were male (n = 3) and 50% were white (n = 3). Four participants (67%) had lived with diabetes for more than two years. Similarly, four participants (67%) had been receiving dialysis for more than five years. Of these consented participants, two did not have a cellphone with internet/data and one preferred not to use their cellphone for the application. These three individuals were provided with CGM readers. All had a computer with internet to access their LibreView account. Following receipt of CGM, one individual’s CGM sensor fell off after two days of use; they were not interested in replacing it and withdrew from the study. One individual’s CGM reader was misplaced/lost prior to data sharing; thus, data were inaccessible for the study. One individual withdrew their consent partway through the study as they no longer wanted to share glycemic data. Of those eligible for CGM, 13.6% (n = 3) complete the study with full data available. Study recruitment and participation is presented in Figure 1.

Figure 1.

Screening, recruitment, and follow-up.

Discussion

Continuous glucose monitoring systems hold promise to modernize diabetes care for individuals receiving hemodialysis. Our study raises concerns regarding the feasibility of CGM use for clinical care in the hemodialysis setting.

Continuous glucose monitoring cost was an expressed concern, despite all sensors and readers being provided free of charge in the context of the study. Technology was also a voiced barrier; there existed reluctance to learn a new glucose monitoring system and/or use a new application on personal electronic devices. Furthermore, some did not have a cellphone with internet/data which added complexity.

Wearing a medical device on the body was a concern for some, and one participant’s sensor fell off within two days of use. Interestingly, the concerns raised in this study were consistent with a previous report of CGM use in the non-hemodialysis population, where important barriers included cost, sensor wearing and system complexity. ⁵ Potential complexities unique to CGM sensor use in individuals on hemodialysis include adhesion of sensors in those with pruritis, change in adhesion due to change in volume status and the risk of skin and soft-tissue infection in an immunocompromised population.

This study has several limitations. Our study was embedded in a quality improvement project that evaluated outcomes (change in mean HbA1c, patient satisfaction and shared decision making) in aggregate at the unit level. No individual patient-level data were collected for the quality improvement project. This limits our ability to identify clinical and demographic differences between those individuals who were or were not prescribed a CGM and those who wore the CGM versus those who declined. However, compared with the median HbA1c of 6.5% (interquartile range [IQR] = 5.5-7.9) in all 244 individuals with T2DM in the hemodialysis unit, median HbA1c in those individuals who consented to participation in the CGM study was higher at 9.1% (IQR = 8.3-9.7). The small proportion of individuals who were referred to endocrinology as part of the quality improvement project could lead to selection bias in CGM prescription. However, this concern is somewhat mitigated when one considers that 78% of those eligible received an endocrinology consult. Finally, the small number of participants from a single hemodialysis center included in this study limit broad generalizability. Despite this, our findings highlight concerns regarding the real-world application of CGM and need for further investigation to barriers to CGM use and solutions to these barriers in the maintenance hemodialysis population.

Conclusions

Individuals with T2DM treated with hemodialysis have unique challenges to optimal diabetes care.^{1 -3} This study identified barriers to CGM use in maintenance hemodialysis in a real-world setting. Given the potential utility of CGM technology in this population, larger studies that explore and address barriers to its use in the hemodialysis setting are greatly needed.