Impact of Simplification Strategies on Postmeal Glucose Excursions in Older Adults with Type 1 Diabetes and Hypoglycemia

Abstract

Continuous glucose monitoring with simplification strategies reduces hypoglycemia in older adults with type 1 diabetes (T1D), however the impact on postmeal glycemia is not known. A post-hoc analysis of older adults with T1D randomized to intervention with mealtime simplification strategies, or control, assessed weekly postmeal hypoglycemia and hyperglycemia. At baseline, 88 older adults with T1D (71 ± 5 years) in intervention (n = 47) and control (n = 41) had similar number of episodes of postmeal hypo- and hyperglycemia. The mean decrease from baseline to 6 months in episodes of postmeal hypoglycemia was: after breakfast (−0.77 vs. −0.32; P = 0.02), lunch (−0.80 vs. −0.32; P = 0.05), and dinner (−0.73 vs. −0.22; P = 0.04); and the mean change in episodes of postmeal hyperglycemia was: after breakfast (−2.05 vs. −1; P = 0.04), lunch (−1.23 vs. −0.87; P = 0.09), and dinner (−1.45 vs. −1.66; P = 0.33), respectively in intervention and control. Simplification strategies in older adults with T1D resulted in fewer episodes of postmeal hypoglycemia without worsening episodes of postmeal hyperglycemia.

Introduction

Diabetes self-management is challenging in persons with type 1 diabetes (T1D), particularly around meals,^1,2 as the amount and timing of insulin are critical to maintain euglycemia while avoiding hyper- and hypoglycemia. Older adults with T1D may experience a decline in functional and cognitive status while aging, potentially reducing their ability to perform tasks related to diabetes self-management, and increasing the risk for hypoglycemia.^3–5 In older adults, hypoglycemia can result in adverse consequences, including cardiovascular events, falls with fractures, and hospital admissions.^6–9 Due to the risks associated with hypoglycemia, and the lack of benefit of tight glycemic control, guidelines ³ recommend integrating geriatric principles, like adapting glycemic goals based on health status and simplifying insulin regimens for older adults with diabetes. However, in older adults with T1D, there is a paucity of information on how to adapt these geriatric principles for complex tasks such as carbohydrate counting and calculation of insulin doses for meals or correction insulin doses. In addition, there is a concern among clinicians that such strategies, while reducing hypoglycemia risk, may increase the time spent in hyperglycemia. ¹⁰

Initiation of continuous glucose monitoring (CGM) in older adults with T1D has been shown to reduce hypoglycemia. ¹¹ In the recently published TANGO (Technological Advances in Glucose Management in Older Adults) study, we showed that in a heterogeneous cohort of older adults with T1D and hypoglycemia, as well as multiple comorbidities, the use of CGM with the integration of geriatric principles can reduce hypoglycemia without worsening glycemic control in both CGM naïve and CGM users. ¹²

In this post-hoc analysis of the TANGO study, we looked at the impact of the simplification strategies on postmeal hypo- and hyperglycemia.

Methods

This is a post-hoc analysis of the TANGO study (ClinicalTrials.gov identifier NCT03078491). ¹² The study was approved by the local Institutional Board Review.

Community-living older adults (≥65 years) with a clinical diagnosis of T1D with hypoglycemia (≥2 episodes of <70 mg/dL for ≥20 min) on either multiple daily insulin injections or insulin pump, either naïve to CGM or CGM users, were enrolled. Participants were randomized 1:1 to the intervention or control group. This analysis includes participants for whom 2 consecutive weeks of CGM raw data (.CSV file) were available at baseline and 6 months (N = 88 of the 131 original randomized study cohort).

Control participants continued to receive usual care by their endocrinologist, and they received attention control time to match the time spent with the intervention group.

Participants randomized to the intervention had an initial visit to individualize glycemic goals and recommendations for simplified strategies based on their clinical and CGM characteristics (Table 1).

Table 1.

Simplification: Recommendations and Strategies

Recommendations	Strategies
Avoid problem-solving	Avoid carbohydrate counting, sliding scales, or correction factor calculations, and provide fixed doses for small or large meals.
	Provide simplified fixed correction dose for use only with mealtime boluses. Pattern management can be done by clinicians or caregivers instead of patients.
Decrease the treatment burden	No correction doses between meals, for snacks, or at bedtime.
	Check availability of caregivers for assisting with monitoring, injections, or reminders for meals and insulin doses.
Emphasize “non-reactive behavior”	Change CGM alarm settings to increase threshold of low alarm to avoid even mild hypoglycemia and reactive eating.
	Turn off high alarm to avoid reactive insulin dosing.

Presence of comorbidities, cognitive and functional decline, and stress perceived by the participant and/or caregivers due to the complexity of the treatment regimens were assessed. CGM patterns were analyzed for recurrent episodes of hypoglycemia (<70 mg/dL for ≥20 min); wide glycemic excursions (coefficient of variation [CV%] >33%); and episodes of hyperglycemia followed by hypoglycemia and again hyperglycemia—suggestive of a mismatch between insulin and food intake or overcorrection of high and low glucose levels.

Overall, simplification strategies were recommended to address 1) avoiding problem solving (i.e., avoiding carbohydrate counting, sliding scales, or correction factor calculations; and providing fixed doses for small or large meals), 2) decreasing treatment burden (i.e., no correction doses between the meals, for snacks or at bedtime), and 3) emphasizing “nonreactive behavior” (i.e., change CGM alarm settings to increase threshold of low alarm to avoid even mild hypoglycemia and reactive eating, or turn off high alarm to avoid reactive insulin dosing). More specifically, we used CGM patterns to identify potential challenges with mealtime insulin dosing, such as episodes of prolonged hypoglycemia or hyperglycemia close to mealtime, suggestive of challenges with problem-solving tasks involved in carbohydrate counting or using sliding scale. We, instead, provided fixed insulin doses based on meal size (i.e., small and large meals), and we instructed the participants to avoid any additional insulin boluses in between meals. To reduce the risk of reactive behaviors triggered by low or high glucose readings, we increased the CGM thresholds. Low alerts were increased to a higher value (i.e., 90 mg/dL). High alerts were either completely disabled or increased to at least >250 mg/dL. The increased low alert threshold would facilitate prompt treatment of impending hypoglycemia; balancing the risk of overtreatment, which may result in hyperglycemia, and under or delayed treatment, which may result in hypoglycemia. The change in high alert threshold would avoid the alert going off in between meals and overnight, which could trigger insulin correction, potentially resulting in hypoglycemia. These strategies were reinforced over the 6 months of the study period (Table 1).

Outcomes

Data were analyzed using a mathematical algorithm that assessed time windows for breakfast (6 am to 11 am), lunch (11 am to 4 pm), and dinner (4 pm to 10 pm). During the mealtime windows, a meal glucose excursion was defined as a rise in sensor glucose of >50 mg/dL within 30 min, preceded by stable sensor glucose readings. Stable sensor glucose readings were defined as sensor glucose readings of ≥95 mg/dL for at least the prior 30 min, without a glucose increase ≥50 mg/dL. An episode of postmeal hypoglycemia was defined as sensor glucose reading <70 mg/dL for ≥15 consecutive minutes within 3 h from the defined mealtime excursion. An episode of postmeal hyperglycemia was defined as sensor glucose >180 mg/dL for ≥15 consecutive minutes between 3–5 h from meal time excursion. ¹³ The number of postmeal episodes for breakfast, lunch, and dinner are reported as number of episodes per week.

Statistical analysis

Continuous variables are presented as mean ± standard deviation (SD) or median and interquartile range, while categorical variables are presented as number and percentage (n, %).

Changes in the number of episodes of postmeal hypoglycemia and hyperglycemia from baseline to 6 months were evaluated. Multivariable linear regression models were constructed to assess whether changes in episodes of postmeal hypoglycemia or hyperglycemia differed significantly between control and intervention groups, while adjusting for age, baseline number of episodes, and duration of diabetes. These models provided β-coefficients representing the adjusted difference in the mean change in number of episodes per week between intervention and control groups, after controlling for age, diabetes duration, and baseline number of episodes. Negative β values indicate greater reductions in the intervention group compared with the control.

Statistical significance was set at P < 0.05 for all analyses. All statistical procedures were performed using STATA statistical software (StataCorp LLC, College Station, TX).

Results

We evaluated 88 older adults with T1D (47 in the intervention group and 41 in the control group) out of the original study cohort of 131 randomized participants. Overall, at baseline the age of the cohort was 71 ± 5 years; diabetes duration 41 ± 17 years, HbA1c was 7.2 ± 0.7%; 65% were pump users and 43% were CGM users. Twenty-two percent were living alone and 45% had cognitive impairment (Montreal Cognitive assessment [MoCA] < 26). ¹⁴ At baseline, there were no differences between the intervention and control groups in demographic and clinical characteristics, as well as CGM metrics.

At baseline, the number of episodes of postmeal hypoglycemia per week was similar in the intervention and the control group: after breakfast (1.2 ± 1.4 vs. 1.3 ± 1.6; P = ns), after lunch (1.4 ± 1.3 vs. 1.4 ± 1.2; P = ns), and after dinner (1.4 ± 1.3 vs. 1.4 ± 1.5; P = ns). The mean decrease in the number of episodes of postmeal hypoglycemia per week from baseline to 6 months was: after breakfast −0.77 ± 1.37 vs. −0.32 ± 1.58 (P = 0.02; β = −0.6); after lunch −0.80 ± 1.52 vs. −0.32 ± 1.46 (P = 0.04; β = −0.44); and after dinner −0.73 ± 1.25 vs. −0.22 ± 1.38 (P = 0.05; β = −0.48), respectively, in the intervention and control groups, after adjusting for age, duration of diabetes, and baseline number of episodes (Fig. 1A).

FIG. 1.

A; B: Mean changes in number of episodes of postmeal hypoglycemia per week (panel A) and number of episodes of postmeal hyperglycemia per week (panel B) in the intervention and control group from baseline to 6 months. *Adjusted for age, duration of diabetes, and baseline number of episodes.

In addition, at baseline, there were no differences in the number of episodes of postmeal hypoglycemia or hyperglycemia at any meal between the intervention and control groups when evaluating the original study cohort of 131 randomized participants.

Next, we looked at the number of episodes of postmeal hyperglycemia per week. At baseline, there was no difference in postmeal hyperglycemia episodes between the intervention and the control group: after breakfast (6.6 ± 4.2 vs. 7.2 ± 4.2; P = ns), after lunch (5.5 ± 3.4 vs. 6.4 ± 4.3; P = ns), and after dinner (6.2 ± 3.7 vs. −7.0 ± 4.2; P = ns). The mean change in number of episodes of postmeal hyperglycemic per week from baseline to 6 months in the intervention compared with the control group was: after breakfast –2.05 ± 4.55 vs. –1.0 ± 4.56 (P = 0.04; β = −1.4), after lunch 1.23 ± 4.03 vs. −0.87 ± 4.99 (P = 0.09; β = −1.2), and after dinner −1.66 ± 3.52 vs. −1.45 ± 3.81 (P = 0.33; β = −0.57), respectively, in the intervention and control groups, after adjusting for age, duration of diabetes, and baseline number of episodes (Fig. 1B).

Discussion

In this post-hoc analysis of the TANGO study, ¹² we show that the application of simplification strategies around meal-time boluses did not result in worsening of postmeal hypoglycemic or hyperglycemic episodes. In fact, the use of simplification strategies reduced the number of postmeal episodes of hypoglycemia, consistent with the primary results, ¹² without worsening the number of episodes of postmeal hyperglycemia. These results support and underscore the importance of simplification of strategies in older adults with T1D experiencing hypoglycemia.

In this cohort of older adults, at baseline, episodes of postmeal hypoglycemia occurred once per week, while episodes of postmeal hyperglycemia occurred almost daily. These results highlight the challenges of maintaining postmeal glycemic control, ¹⁵ and are consistent with data in the literature where episodes of postmeal hypo- and hyperglycemia can occur almost 50% of the time in people with diabetes on complex insulin regimens. ¹³

Current guidelines recommend simplifying diabetes self-management and de-intensifying insulin regimens in older adults with diabetes to reduce hypoglycemia, ³ as supported by our prior work in older adults with type 2 diabetes.^3,16 However, in older adults with T1D guidance on how and when to simplify insulin regimens is lacking. ³ In our recent TANGO study, we adapted the simplification strategies used in older adults with T2D and implemented it for older adults with T1D. ¹² The simplification strategies used aimed to reduce problem-solving tasks, reactive behaviors, and burden of treatment. These strategies were implemented and adapted over the study period, depending on the challenges with diabetes self-management tasks encountered by each participant. ¹² The simplified tasks consisted of more directive instructions, such as providing fixed insulin doses for meals and correction, avoiding correction in between meals, and recommending conservative insulin doses at bedtime (Table 1), as discussed in the methods section. The implementation of some, or all, of these strategies in this cohort of older adults with T1D and hypoglycemia resulted in an improvement of glycemic excursions, irrespective of age, diabetes duration, insulin delivery methods, and CGM use status at baseline.

The results of this post-hoc analysis are the first to our knowledge to show that the application of simplification strategies in older adults with T1D and hypoglycemia decreased episodes of postmeal hypoglycemia and did not increase postmeal hyperglycemia.

While clinicians recognize the value of simplification strategies to reduce the risk of hypoglycemia, ³ it is rarely implemented in clinical practices, even in older adults with tight glycemic control. ¹⁷ In a national survey, most clinicians reported reluctance to de-intensify or switch hypoglycemia-causing medications due to concerns of hyperglycemia. ¹⁸ In older adults with T1D, simplifying insulin regimens can be a complex clinical decision with limited information on how to adjust multi-daily insulin regimens and their impact on glycemia. Our study results fill this important knowledge and care gap. The suggested simplification strategies (Table 1) offer a practical framework for clinicians, which is particularly relevant, as most older adults with T1D in the United States are cared for by primary care clinicians. ¹⁹

Study limitations included the post-hoc nature of the analysis, in a single-center, with predominantly white and highly educated participants with T1D and hypoglycemia. Thus, simplification strategies may not apply in older adults with T1D without hypoglycemia. However, challenges with complex insulin regimens likely affect all age groups and especially older adults with diabetes with less education and/or fewer resources.

Conclusions

In older adults with T1D and hypoglycemia, simplifying mealtime strategies lowered overall hypoglycemia and the number of episodes of postmeal hypoglycemia, without worsening postmeal episodes of hyperglycemia. These findings support using CGM-guided geriatric principles to help clinicians manage older adults with T1D and hypoglycemia.

Authors’ Contributions

E.T.: Conceptualization, methodology, software, formal analysis, investigation, writing—original draft, writing—reviewing and editing, visualization, supervision, funding acquisition. M.S.: Software, formal analysis, data curation, visualization. C.C.: Software, data curation. N.K.: Data curation. C.S.: Investigation, data curation, writing—original draft, writing—reviewing and editing, visualization, project administration. A.A.: Formal analysis, data curation, visualization. M.M.: Conceptualization, methodology, investigation, writing—original draft, writing—reviewing and editing, supervision, funding acquisition.