Associations Between Hypoglycemia Awareness Status and Symptoms of Hypoglycemia Among Adults with Type 1 or Insulin-Treated Type 2 Diabetes Using the Hypo-METRICS Smartphone Application

Introduction

Hypoglycemia is the most common side effect of insulin therapy in diabetes management. ¹ The fall in glucose levels occurring during a hypoglycemic episode leads to physiological changes, including the development of an array of symptoms which prompt corrective action such as carbohydrate ingestion to prevent severe hypoglycemia. ^2,3 Hypoglycemia symptoms are often grouped into autonomic symptoms such as heart palpitations, shaking, and sweating, resulting from the activation of the autonomic nervous system ^4,5 and neuroglycopenic symptoms, including headache, confusion, and dizziness, typically considered to manifest as a result of a significant reduction in glucose supply to the cerebral cortex. ⁶

It is known that the experience of hypoglycemia symptoms varies between and within individuals. ^5,7 This can reflect underlying factors such as type of diabetes, awareness of hypoglycemia, actual blood glucose level, age or diabetes duration, as well as the situation during which the hypoglycemic episode occurs. ^6,8 For example, people with impaired awareness of hypoglycemia (IAH) typically report hypoglycemia symptoms less frequently ⁹ and at lower glucose concentrations ^10,11 than those with normal awareness of hypoglycemia (NAH). However, such studies reporting on hypoglycemia symptoms have commonly been limited by the use of controlled experimental conditions to induce hypoglycemia ¹¹ or retrospective methods based on participants' recollection of a “typical” episode of hypoglycemia. ^{10,12 –14} While the use of ecological momentary assessment and smartphone technologies could address those limitations, ¹⁵ to the best of our knowledge, no such methods have been commonly used to record hypoglycemia symptoms in the hypoglycemia field.

In that respect, the Hypoglycaemia—MEasurement, ThResholds and ImpaCtS (Hypo-METRICS) smartphone application (app) is novel and overcomes previous limitations by allowing users to record episodes of hypoglycemia experienced in daily life and describe the glucose range and intensity of eight hypoglycemia symptoms at, or shortly after, an episode. ¹⁶ Recent quantitative evidence has demonstrated promising results in terms of its acceptability and validity with high completion rates (91%) and adequate psychometric properties (e.g., internal consistency, test-retest reliability) among participants with type 1 diabetes (T1D) or insulin-treated type 2 diabetes (T2D). ¹⁷

Qualitative evidence further supports this as the app was found to be “engaging,” “well designed,” and “easy to use” by study participants, with hypoglycemia-related questions and content deemed relevant and comprehensive. ¹⁸ While this is promising evidence, as the Hypo-METRICS app is a novel tool, it is essential to further validate it by examining whether differences in hypoglycemia symptoms experiences detected among people with diabetes using the app are consistent with previous evidence.

This study examined the validity of the Hypo-METRICS app by assessing the relationships between hypoglycemia awareness status, measured with the Gold questionnaire, and hypoglycemia symptoms, reported using the app, among adults with T1D or insulin-treated T2D.

Methods

The Hypo-METRICS app: hypoglycemia symptom tracker

Hypo-METRICS participants were instructed to record episodes of hypoglycemia at, or shortly after the event, using the app hypoglycemia symptom tracker. This was any episode that participants identified as hypoglycemia where they had to take corrective action. This was defined as symptomatic episodes that resolved on carbohydrate ingestion and/or a measured glucose below 72 mg/dL. While using the app and blinded study CGM, participants continued their usual form and frequency of glucose monitoring.

The symptom tracker consists of a flower motif with 10 petals as shown in Figure 1. Each petal offered different options to indicate (1) the timing of the hypoglycemic episode (“Now,” “15 minutes ago,” “30 minutes ago,” “1 hour ago,” or “>1 hour ago”), (2) the glucose range at which participants were experiencing a hypoglycemic episode (“<36 mg/dL,” “36–53 mg/dL,” “54–70 mg/dL” or “ >70 mg/dL,” and (3) the intensity (“Not at all,” “A little bit,” “Somewhat,” “Quite a bit” or “Very much”) of eight hypoglycemic symptoms, described in the next section. Blood glucose range was further categorized into three groups: >70 mg/dL, 54–70 mg/dL, and <54 mg/dL. The <54 mg/dL category included events reported at 36–53 mg/dL and <36 mg/dL, which were collapsed due to the small proportion of hypoglycemic episodes reported at the lowest glucose range (2% of episodes reported <36 mg/dL).

OPEN IN VIEWER

FIG. 1.

The Hypo-METRICS symptom tracker.

Results

There were 602 participants recruited in the Hypo-METRICS study. Five hundred and fifty-two participants submitted at least one episode of hypoglycemia using the Hypo-METRICS app symptom tracker. Of those, 531 reported a glucose range for their hypoglycemia episodes and were included in the present analyses (Fig. 2).

OPEN IN VIEWER

FIG. 2.

Flowchart of study participation.

Discussion

The Hypo-METRICS app allowed adults with insulin-treated diabetes to provide information about glucose values and symptoms experienced during hypoglycemia in real time. This study has shown that most episodes reported using the app were symptomatic and that autonomic symptoms were the most frequently reported in both T1D and T2D. Participants with IAH were significantly less likely to report autonomic symptoms (shaking and hunger in T1D and shaking and sweating in T2D) compared to those with NAH. Moreover, when autonomic symptoms were reported, participants with T1D and IAH were more likely to report them at a lower glucose concentration than those with NAH. No differences were found in neuroglycopenic symptoms reporting according to hypoglycemia awareness in either T1D or T2D.

This study shows that the vast majority of episodes of hypoglycemia recorded using the Hypo-METRICS app symptom tracker were symptomatic, even among participants with IAH. This may reflect the fact that app data were collected when participants knew they were experiencing an episode of hypoglycemia and, by collecting data at the time of episodes, this is likely to enhance reporting of symptomatic episodes. Nevertheless, results of the present study are in line with previous evidence that hypoglycemia awareness is a continuum and that total loss of symptomatic response is rather uncommon. ^10,23,24

The most commonly reported symptoms in the overall sample were autonomic: hunger and shaking followed by sweating and heart palpitations. The predominance of autonomic symptoms is consistent with previous research in T1D ^5,25 and T2D. ^13,26 This aligns with the evidence that autonomic symptoms are typically the first ones to manifest as early warning signs of hypoglycemia, before the onset of neuroglycopenic symptoms, ^3,23 thus more likely to be recognized first by people with diabetes and reported more commonly than their neuroglycopenic counterparts.

Data from the app demonstrate that participants with IAH were less likely to report autonomic symptoms than those with NAH, with no differences in neuroglycopenic symptoms in both T1D and T2D. This finding is consistent with previous studies that have shown a lower proportion of hypoglycemic episodes with autonomic symptoms in IAH than NAH, ⁹ lower mean autonomic symptom intensity score in IAH than NAH, ^27,28 and no differences in neuroglycopenic symptoms between groups. ^{27 –29} However, those studies used overall autonomic and neuroglycopenic scores or percentages and did not examine the odds of individual symptoms nor did they investigate the impact of glucose monitoring method.

The present study therefore adds to the understanding by showing that differences in autonomic symptoms between IAH and NAH may be driven by differences in shaking and hunger for T1D and shaking and sweating for T2D, as suggested by the analysis of individual symptoms. The results also suggest that in T2D, the relationships between Flash/CGM use and hypoglycemia symptom reporting vary according to awareness status.

Finally, it was shown that, relative to participants with T1D and NAH, those with IAH were more likely to report lower glucose concentrations for their autonomic symptoms. No such significant differences in the glucose ranges were observed in participants with T2D, which may be reflective of the lower number of hypoglycemic episodes in that group. Nevertheless, our results in adults with T1D are in keeping with the findings of a retrospective questionnaire study conducted by Hendrieckx et al., ¹⁰ showing that 30% of participants with T1D and IAH recognized autonomic symptoms when glucose was <45 mg/dL while 47% of those with NAH recognized autonomic symptoms when glucose was ≥63 mg/dL. Overall, our findings not only align with previous research but also with the pathophysiology of IAH involving a decreased autonomic response and a shift toward lower glucose concentrations before symptoms are recognized. ^2,30

The main strength of this study is that the symptoms examined were based on clinical episodes of hypoglycemia reported at, or shortly after the episode, using the Hypo-METRICS app, thus minimizing the risk of recall bias and reflecting more accurately real-time clinical hypoglycemia experience. Further, the results were based on robust data as an overall large number of episodes of hypoglycemia were reported (n = 11,011) over an extensive study duration (n = 10 weeks) and in a large study sample (n = 531). Finally, the use of a multilevel design for the present study allowed to account for the fact that a participant could report multiple hypoglycemic episodes during the study and thus account for within-individual variability while adjusting for method of glucose monitoring and diabetes duration.

This study is, however, limited by the fact that only eight symptoms could be reported using the app, and while those were carefully selected in accordance with a patient advisory committee, previous literature, and expert consensus, they may not represent the full range of hypoglycemia symptoms that adults with diabetes experience in daily life. Future versions of the app will, however, allow users to have a greater variety of symptoms to select from. In addition, although the Hypo-METRICS study was conducted across five different countries, the predominant ethnic group was white, which may limit the generalizability of the results to other ethnic groups. Our study may also be limited by the fact that it was partly conducted during the COVID-19 pandemic, which may have impacted on daily life patterns and glycemic management and thus, the usage of the app and frequency of hypoglycemia episodes and symptoms reported.

To conclude, the Hypo-METRICS app, which allows real-time interactive recording of hypoglycemia episodes experienced in daily life, is sensitive to differences in hypoglycemia symptoms reporting according to hypoglycemia awareness and blood glucose range. The detected differences in hypoglycemia symptoms experiences were consistent with the pathophysiology of IAH and previous research evidence in people with IAH. This study thus further validates the use of the Hypo-METRICS app as a tool with potential research, educational and clinical applications.