Characterizing Severe and Level 2 Hypoglycemia and Associated Risk Factors in Adults with Type 1 Diabetes Using Hybrid Closed-Loop Insulin Pumps

Abstract

Background:

Despite advances in diabetes technologies, severe hypoglycemia (SH) and level 2 hypoglycemia (Lv2Hypo) persist among hybrid closed-loop (HCL) insulin pump users. This study assessed the relationship of impaired awareness of hypoglycemia (IAH) and other patient characteristics with SH and Lv2Hypo in HCL insulin pump users with type 1 diabetes.

Methods:

A cross-sectional survey assessed 6-month SH history, hemoglobin A1C, IAH (using the Hypoglycemia Awareness Questionnaire), and 30-day continuous glucose monitoring (CGM) data among adult HCL insulin pump users recruited from a national U.S. type 1 diabetes patient registry. Analyses included logistic regression, t-tests, and Chi-square tests.

Results:

Of 601 participants (female: 54%; mean age: 43), IAH and higher glucose coefficients of variation (CVs) were associated with both SH and spending ≥1% of time in Lv2Hypo (Ps < 0.05). Individuals with SH were further characterized as having spent more time with glucose levels >180 mg/dL and >250 mg/dL and having a lower education level (Ps < 0.05). CGM hypoglycemia measures were not associated with SH. Age and diabetes duration were not associated with experiencing SH or spending ≥1% of time in Lv2Hypo. Participants who both experienced SH and spent ≥1% of time in Lv2Hypo showed trends toward exhibiting the most severe IAH and the highest glucose CVs.

Conclusions:

Among adults with type 1 diabetes using HCL insulin pumps, IAH and higher glucose CVs are risk factors of experiencing SH and Lv2Hypo. Hyperglycemia and lower education level are also associated with a higher risk for experiencing SH.

Keywords

hypoglycemia impaired awareness of hypoglycemia type 1 diabetes hybrid closed-loop insulin pumps

Introduction

Impaired awareness of hypoglycemia (IAH) refers to the reduced ability to perceive hypoglycemic symptoms.¹ People with IAH often fail to recognize and manage hypoglycemic episodes in a timely manner. Thus, these individuals are at greater risk of developing severe hypoglycemia (SH; a state in which hypoglycemia-induced cognitive dysfunction prevents one from self-managing hypoglycemia).² IAH has also been associated with spending longer time in level 2 hypoglycemia (Lv2Hypo).³ Lv2Hypo, with glucose levels being less than 54 mg/dL, is particularly dangerous as people can experience acute cognitive dysfunction and cardiac arrhythmias at these low blood glucose levels.⁴

By linking insulin delivery with real-time continuous glucose monitoring (CGM) data, hybrid closed-loop (HCL) insulin pumps automatically adjust insulin doses to minimize hypoglycemia and hyperglycemia. These systems also provide real-time glucose information and alerts for hyperglycemia and hypoglycemia, allowing users to take additional corrective actions.^1,5 Clinical trial data have confirmed the efficacy of these technologies in reducing hypoglycemia.^6,7 Nonetheless, recent real-world evidence indicates that about a quarter of current diabetes technology users,^8,9 including 17% of HCL users,¹⁰ still experience SH. Sparse research has yet evaluated the relationship of IAH and other risk factors with experiencing SH and spending significant time in Lv2Hypo among HCL insulin pump users.

In this cross-sectional observational study, we tested the hypothesis that IAH remains associated with SH and spending ≥1% of time in Lv2Hypo in adults with type 1 diabetes using HCL insulin pumps. Additionally, we explored other risk factors of SH and spending ≥1% of time in Lv2Hypo and compared characteristics among HCL insulin pump users who experienced SH only (without spending ≥1% of time in Lv2Hypo), those who spent ≥1% of time in Lv2Hypo only (without experiencing SH), and those who presented with both SH experience and spending ≥1% of time in Lv2Hypo.

Materials and Methods

Study design

This cross-sectional survey study was conducted from July to December 2023. The study was approved by the University of Michigan Institutional Review Board (HUM#00232351). All participants provided informed consent before engaging in the study. Data were reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology guidelines.¹¹

Participants

Members of the T1D Exchange Registry¹² were approached for recruitment. The inclusion criteria were having a diagnosis of type 1 diabetes and being at least 18 years old; the exclusion criterion was not actively using an insulin pump with HCL features or not able or willing to provide CGM Lv2Hypo data. An email containing the study information was sent to T1D Exchange Registry members who met the inclusion criteria. Each registry member could open the survey weblink attached to the email, provide consent, and complete the survey. Datasets were then screened for the exclusion criteria; ineligible respondents were subsequently excluded from analysis. An online survey was deemed appropriate for this population, as all T1D Exchange Registry members have email addresses and regularly receive emails from the T1D Exchange.

Survey data collection

The survey website was developed using Research Electronic Data Capture (REDCap).¹³ Participant demographics, household income, education level, self-reported hemoglobin A1C (HbA1c), and HCL insulin pump use and brand were assessed via the adapted T1D Exchange Core Questionnaire.¹² Six-month SH history was evaluated based on the American Diabetes Association/Endocrine Society 2013 consensus statement (“in the past six months, how many times did you have low sugar levels requiring assistance of another person to actively administer carbohydrates, glucagon, or take other corrective actions?”).¹⁴ Participants’ hypoglycemia awareness was measured with the Hypoglycemia Awareness Questionnaire Impaired Awareness subscale (HypoA-Q IA),¹⁵ a 5-item instrument assessing a spectrum of hypoglycemia symptom and awareness experiences: symptom development during hypoglycemia, self-recognition of hypoglycemic episodes, others recognizing hypoglycemia before the respondent, and changes in hypoglycemia awareness and hypoglycemia symptoms over time. The scores generated by each item’s Likert scale sum up to a score of 20, with scores ≥12 considered indicative of IAH.¹⁶ We used the HypoA-Q IA to determine IAH given its high correlation with several other IAH measuring instruments and ability to more granularly characterize hypoglycemia awareness with a wider score range.¹⁷ Gold¹⁸ and minimally modified Clarke¹⁹ questionnaires were also used to measure hypoglycemia awareness for sensitivity analyses.

CGM data collection

Thirty-day CGM data,²⁰ with the last day being the day of survey response, were obtained from participants through either the CGM website or CGM reports. The following information was collected: CGM time spent with glucose >180 mg/dL and >250 mg/dL, time-in-range, time with glucose <70 mg/dL (which included the time spent with glucose <54 mg/dL), time with glucose <54 mg/dL, and the glucose coefficient of variation (CV). Glycemia risk index (GRI) hypoglycemia component²¹ was calculated with CGM glucose information for sensitivity analysis.

Statistical methods

Based on Cochrane’s formula modified for smaller populations,²² a sample of 377 participants would be representative of a pool of 18,000 adult members in the T1D Exchange Registry at a 95% confidence level with a 5% margin of error. As the number of responses from HCL insulin pump users who would be willing to provide Lv2Hypo data could not be fully predicted, we oversampled the population to collect sufficient responses. All incomplete and duplicate responses were discarded, and responses from non-HCL insulin pump users or respondents using HCL insulin pumps who yet did not provide Lv2Hypo data were not included in the analysis. Nonsensible data (e.g., illogical age and diabetes duration) were also removed and reported as missing data.

Descriptive analysis was conducted to summarize participants’ characteristics, which are presented as % or mean ± standard deviation (SD) unless otherwise noted. Student’s t-tests and Chi-square tests were performed to compare population characteristics. Logistic regression was employed to (1) evaluate the relationships of IAH with experiencing SH and spending ≥1% of time in Lv2Hypo and (2) explore the relationships of other participant characteristics and CGM measures with SH and spending ≥1% of time in Lv2Hypo including the relationship between SH and spending ≥1% of time in Lv2Hypo, without and with adjustments for age, gender, diabetes duration, and IAH status (when not the independent variable). Sensitivity analyses included investigating (1) the relationships of experiencing SH and spending ≥1% of time in Lv2Hypo with IAH assessed using different IAH-measuring instruments and (2) the relationship between SH and other CGM hypoglycemia measures with spending ≥4% of time with glucose <70 mg/dL, the percentage of time with Lv2Hypo and glucose <70 mg/dL, GRI hypoglycemia component, as well as the interaction of spending <1%, 1%–2%, and >2% of time in Lv2Hypo with the relationship between IAH and SH. Missing data were considered “missing completely at random.” All analyses were completed in IBM SPSS version 29.0.2. A P value <0.05 was considered statistically significant.

Results

Participant characteristics

Of the cohort of 601 eligible individuals retained for analysis (Supplementary Fig. S1), 54% were female, 94% were Caucasian, and 93% were non-Hispanic. The mean ± SD age was 43 ± 15 years, the mean ± SD diabetes duration was 24 ± 15 years, and the mean ± SD self-reported HbA1c was 6.5 ± 0.8% (48 ± 9 mmol/mol; see Supplementary Table S1 for participant characteristics). The mean ± SD HypoA-Q IA score was 7.2 ± 3.8, with a HypoA-Q-determined IAH prevalence of 14%. Slightly more than one-quarter of participants (26%) had developed SH in the past 6 months.

Risk factors for experiencing ≥1 SH episode in the past 6 months

In multivariable analyses with adjustments for age, gender, and diabetes duration, each incremental increase in HypoA-Q IA scores was related to 11% greater odds of having ≥1 SH episode(s) in the past 6 months (95% confidence interval [CI]: 1.06, 1.17; P < 0.001; Fig. 1A). IAH status was associated with 1.70-fold higher odds of SH (95% CI: 1.02, 2.82; P = 0.042; Table 1). A sensitivity analysis in which Gold scores (odds ratio [OR]: 1.22; 95% CI: 1.08, 1.38; P = 0.001) and Clarke scores (OR: 1.57; 95% CI: 1.40, 1.76; P < 0.001) served as measures of hypoglycemia awareness also demonstrated the relationship between lower hypoglycemia awareness and higher SH odds.

FIG. 1.

The relationship between HypoA-Q IA scores and the probability of (A) experiencing SH in the past 6 months and (B) spending ≥1% of time in Lv2Hypo. Multiple regression analyses were performed with adjustments for age, gender, and diabetes duration. Gray dots represent datapoints. Vertical dash lines indicate the cut-off indicating IAH. CI, confidence interval; HypoA-Q IA, Hypoglycemia Awareness Questionnaire Impaired Awareness subscale; IAH, impaired awareness of hypoglycemia; Lv2Hypo, level 2 hypoglycemia; OR, odds ratio; SH, severe hypoglycemia.

Table 1.

Participant Characteristics Associated with 6-Month SH Experience and Time Spent in Lv2Hypo in the Past 30 Days (n = 601)

	n (%) or mean ± SD		Odds ratio (95% CI)	P value	n (%) or mean ± SD		Odds ratio (95% CI)	P value
	Without SH (n = 447)	With SH (n = 154)	Odds ratio (95% CI)	P value	<1% of time in Lv2Hypo (n = 553)	≥1 of time in Lv2Hypo (n = 48)	Odds ratio (95% CI)	P value
Age, year	44 ± 15	43 ± 14	1.00 (0.98, 1.01)	0.568	43 ± 15	44 ± 14	0.99 (0.96, 1.02)	0.469
Gender
Male (Reference)	195 (44)	67 (44)	—	—	240 (43)	22 (46)	—	—
Female	246 (55)	79 (51)	0.91 (0.62, 1.34)	0.625	300 (54)	25 (52)	0.88 (0.47, 1.62)	0.669
Other	6 (1)	8 (5)	3.69 (1.20, 11.32)	0.023	13 (2)	1 (2)	0.79 (0.10, 6.60)	0.829
Race
Caucasian (reference)	423 (95)	144 (94)	—	—	522 (94)	45 (94)	—	—
Other	17 (4)	5 (3)	0.78 (0.28, 2.23)	0.646	20 (4)	2 (4)	1.10 (0.24, 5.01)	0.898
Multirace	7 (2)	5 (3)	1.70 (0.50, 5.61)	0.403	11 (2)	1 (2)	0.97 (0.12, 7.98)	0.975
Ethnicity
Non-Hispanic (reference)	420 (94)	140 (91)	—	—	514 (93)	46 (96)	—	—
Hispanic	27 (6)	14 (9)	1.51 (0.76, 3.01)	0.245	39 (7)	2 (4)	0.56 (0.13, 2.45)	0.444
Education level
Associate’s degree or below (reference)	116 (26)	55 (36)	—	—	157 (28)	14 (29)	—	—
Bachelor’s degree	179 (40)	67 (44)	0.83 (0.54, 1.28)	0.405	224 (41)	22 (46)	1.23 (0.60, 2.52)	0.570
Master’s degree or above	151 (34)	32 (21)	0.46 (0.28, 0.77)	0.003	171 (31)	12 (25)	0.86 (0.38, 1.96)	0.721
Annual household income			0.90 (0.82, 0.99)	0.030			0.98 (0.84, 1.14)	0.793
Less than $25,000	30 (7)	12 (8)			38 (7)	4 (8)
$25,000 to less than $35,000	12 (3)	11 (7)			21 (4)	2 (4)
$35,000 to less than $50,000	31 (7)	15 (10)			42 (8)	4 (8)
$50,000 to less than $75,000	58 (13)	24 (16)			75 (14)	7 (15)
$75,000 to less than $100,000	63 (14)	20 (13)			77 (14)	6 (13)
$100,000 to less than $200,000	151 (34)	42 (27)			179 (32)	14 (29)
$200,000 or more	68 (15)	19 (12)			79 (14)	8 (17)
Diabetes duration, year	24 ± 15	24 ± 15	1.00 (0.99, 1.02)	0.803	24 ± 15	26 ± 13	1.01 (0.98, 1.04)	0.478
Self-reported HbA1c			1.21 (0.96, 1.53)	0.111			0.27 (0.16, 0.46)	<0.001
%	6.4 ± 0.8	6.6 ± 0.8			6.5 ± 0.8	6.0 ± 0.7
mmol/mol	46 ± 9	49 ± 9			48 ± 9	42 ± 8
IAH, %	55 (12)	29 (19)	1.70 (1.02, 2.82)	0.042	71 (13)	13 (27)	2.52 (1.24, 5.11)	0.010
CGM glycemic characteristics
Time with glucose >250 mg/dL, %	5 ± 7	7 ± 9	1.03 (1.01, 1.06)	0.012	6 ± 8	3 ± 3	0.90 (0.83, 0.98)	0.011
Missing values	1	1			2	0
Time with glucose >180 mg/dL, %	22 ± 15	27 ± 16	1.02 (1.01, 1.03)	0.002	24 ± 15	15 ± 10	0.94 (0.92, 0.97)	<0.001
Missing values	9	2			11	0
≥4% time with glucose <70 mg/dL	48 (11)	17 (11)	1.02 (0.56, 1.85)	0.941	23 (4)	42 (88)	181 (66, 492)	<0.001
Missing values	1	3			4	0
≥1% time with glucose <54 mg/dL	33 (7)	15 (10)	1.26 (0.66, 2.43)	0.484	—	—	—	—
Missing value	1	3			4	0
Glucose coefficient of variation, %	31 ± 5	32 ± 6	1.05 (1.01, 1.08)	0.014	31 ± 5	35 ± 5	1.18 (1.11, 1.26)	<0.001
Missing values	1	1			2	0

Data analyzed with multiple logistic regression with adjustments for age, gender, diabetes duration, and IAH (when not the independent variable).

CGM, continuous glucose monitoring; CI, confidence interval; HbA1c, hemoglobin A1C; IAH, impaired awareness of hypoglycemia; Lv2Hypo, level 2 hypoglycemia; SD, standard deviation; SH, severe hypoglycemia.

With adjustments for age, gender, diabetes duration, and IAH status, other SH risk factors identified included spending more time with glucose levels >250 mg/dL (P = 0.012) and >180 mg/dL (P < 0.002) and having higher glucose CVs (P = 0.014). Self-identifying with a gender of “other” (P = 0.023) was associated with higher SH odds. Having a master’s degree or above (compared with having an associated degree or below; P = 0.003) and a higher household income (P = 0.030) was related to lower odds of SH. The relationships between SH and education level remained statistically significant when adjusting for time spent in hyperglycemia (P = 0.027). Conversely, SH was no longer related to household income after adjusting for time spent with glucose >180 mg/dL (P = 0.082).

Notably, spending ≥1% of time in Lv2Hypo in the past 30 days was not associated with experiencing SH in the past 6 months (P = 0.941). Similarly, sensitivity analyses suggested no significant associations between SH and spending ≥4% of time with glucose <70 mg/dL, the percentage of time in Lv2Hypo (P = 0.973) and with glucose <70 mg/dL (P = 0.693), as well as the hypoglycemia component of GRI (P = 0.758) based on 30-day CGM data. The percentage of time in Lv2Hypo also did not interact the relationship between IAH and SH (Ps > 0.25).

Age and diabetes duration were not associated with SH. No differences in the statistical significance between the multiple and simple regression analyses (Supplementary Table S2) were identified.

Risk factors for spending ≥1% of time in Lv2Hypo

Each incremental increase in HypoA-Q IA scores was related to approximately 10% greater odds of spending ≥1% of time in Lv2Hypo (95% CI: 1.02, 1.19; P = 0.020; Fig. 1B) with adjustments for age, gender, and diabetes duration. IAH corresponded to 2.52-fold higher odds of spending ≥1% of time in Lv2Hypo (95% CI: 1.24, 5.11; P = 0.010; Table 1). A sensitivity analysis with Gold scores (OR: 1.22; 95% CI: 1.01, 1.47; P = 0.036) and Clarke scores (OR: 1.33; 95% CI: 1.13, 1.56; P < 0.001) further reinforced the relationship between lower hypoglycemia awareness and higher odds of spending ≥1% of time in Lv2Hypo.

With adjustments for age, gender, diabetes duration, and IAH status, lower HbA1c, less time spent with glucose >250 mg/dL and >180 mg/dL, spending ≥4% of time with glucose <70 mg/dL, and higher glucose CVs were associated with ≥1% of time in Lv2Hypo (Ps < 0.05).

Age, gender, education level, household income, and diabetes duration were not related to spending ≥1% of time in Lv2Hypo. No differences in the statistical significance between the multiple and simple regression analyses were identified.

Characteristics in subgroups with SH only, ≥1% of time in Lv2Hypo only, and both

Among the subgroups with different hypoglycemia experiences (Fig. 2), 139 participants (23.1%) experienced SH only (i.e., without spending ≥1% of time in Lv2Hypo); 33 (5.4%) spent ≥1% of time in Lv2Hypo only (i.e., without experiencing SH); and 15 (2.5%) both had SH and spent ≥1% of time in Lv2Hypo. Taking the subgroup of 414 participants without SH and who spent <1% of time in Lv2Hypo as the control group, participants with only SH demonstrated more time with glucose >180 mg/dL (P = 0.003), higher glucose CVs (P = 0.033), and higher HypoA-Q IA scores (P < 0.001) (Table 2). By contrast, compared with the control group, participants spending ≥1% of time in Lv2Hypo exhibited lower HbA1c and less time with glucose >180 mg/dL; they also exhibited higher glucose CVs (Ps < 0.001 for all). Participants who both experienced SH and spent ≥1% of time in Lv2Hypo presented with lower HbA1c (P < 0.030) and higher glucose CVs and HypoA-Q IA scores (Ps < 0.001 for both) compared with the control group.

FIG. 2.

Radar plots of factors associated with SH and Lv2Hypo. In each of the three radar plots, there are four spokes, one for each of the four factors. Values are scaled to the center and vertex values as indicated in the table. For example, for the HbA1c spoke, the center of the square is 5.5%, and the vertex of the outer square labeled with HbA1c is 7%. Mean values of the factors for each group are plotted along the spokes of the square. Gray area in each of the radar plots indicates the profile for participants without SH and with <1% in Lv2Hypo (n = 414). CV, coefficient of variation; HbA1c, hemoglobin A1C.

Table 2.

Participants Grouped by Experiencing SH Only, ≥1% of Time in Lv2Hypo Only, and with Both and Neither of These Experiences

	Group characteristics				P value
	Group 1: Without SH and with <1% of time in Lv2Hypo (n = 414)	Group 2: With SH only (n = 139)	Group 3: ≥1% of time in Lv2Hypo only (n = 33)	Group 4: Both SH and ≥1% of time in Lv2Hypo (n = 15)	Group 1 versus 2	Group 1 versus 3	Group 1 versus 4	Group 2 versus 3	Group 2 versus 4	Group 3 versus 4
HbA1c, % (mmol/mol)	6.5 ± 0.8 (48 ± 8)	6.6 ± 0.8 (49 ± 9)	5.9 ± 0.7 (41 ± 8)	6.0 ± 0.7 (42 ± 8)	0.053	<0.001	0.030	0.001	0.008	0.352
Time with glucose >180 mg/dL	23 ± 15	27 ± 16	12 ± 9	20 ± 11	0.003	<0.001	0.315	<0.001	0.071	0.028
Glucose CV, %	31 ± 5	32 ± 6	35 ± 5	37 ± 5	0.033	<0.001	<0.001	0.004	0.001	0.277
Missing data	1	1	0	0
HypoA-Q IA score	6.8 ± 3.7	8.1 ± 3.7	7.7 ± 4.1	10.2 ± 3.6	<0.001	0.111	<0.001	0.612	0.039	0.053

Data analyzed with Student’s t-test.

CV, coefficient of variation; HypoA-Q IA, Hypoglycemia Awareness Questionnaire Impaired Awareness subscale.

In comparing subgroups with significant hypoglycemia experiences, participants with both SH and ≥1% of time in Lv2Hypo exhibited a greater degree of IAH (i.e., higher HypoA-Q IA scores; P = 0.039), higher glucose CVs (P = 0.001), and lower HbA1c (P = 0.008); they also exhibited trends toward less time with glucose >180 mg/dL (P = 0.071) compared with the SH-only group. The group with both SH and ≥1% of time in Lv2Hypo also spent more time with glucose >180 mg/dL (P = 0.028) and tended to have higher HypoA-Q IA scores (P = 0.053) compared with the group that spent ≥1% of time in Lv2Hypo only. The SH-only and ≥1% of time in Lv2Hypo-only subgroups differed in their HbA1c values (P = 0.001) and time spent with glucose >180 mg/dL (P < 0.001), with the SH-only subgroup’s values higher on these measures. Differences in glucose CVs among these two subgroups were also observed, with the subgroup who spent ≥1% of time in Lv2Hypo rating higher on this metric (P = 0.004).

Discussion

In our cohort of adults with type 1 diabetes using HCLs, IAH and higher glucose CVs were associated with experiencing SH and spending ≥1% of time in Lv2Hypo, despite adjustments for potential confounders. Additionally, more time spent in hyperglycemia and social factors such as lower education levels were associated with higher SH risk but not with Lv2Hypo. More time spent in Lv2Hypo in the past 30 days was not associated with SH experiences in the past 6 months. Age and diabetes duration were not identified as risk factors for spending ≥1% of time in Lv2Hypo in our exclusively HCL insulin pump–using cohort. Only a small proportion of participants both experienced SH and spent ≥1% of time in Lv2Hypo. Unique IAH and glycemic characteristics emerged among subgroups experiencing SH only, spending ≥1% of time in Lv2Hypo only, and having both of these hypoglycemia experiences. Compared with those spending ≥1% of time in Lv2Hypo only, participants experiencing SH only were characterized by more hyperglycemia and less glucose CVs.

Our study is one of the first to specify risk factors for SH and Lv2Hypo among HCL insulin pump users. Consistent with recent real-world studies,^10,23 our cohort shows that hypoglycemia continues to be problematic despite the use of contemporary diabetes technologies. Despite HCL use, a significant portion of participants continue to experience IAH (14% measured by HypoA-Q IA), and slightly more than one-quarter of participants (26%) reported at least one episode of SH in the past 6 months. These findings indicate that HCL insulin pump use alone does not fully restore awareness of hypoglycemia in all HCL insulin pump users, and such technology use does not completely eradicate SH events. Our research also identified a persistent effect of IAH on the risks of both SH and Lv2Hypo in a type 1 diabetes patient cohort who use HCL insulin pumps. In addition, our analysis suggests that HypoA-Q IA scores in the subgroup with both SH and spending ≥1% of time in Lv2Hypo tended to be the highest among all subgroups. The effect of IAH thus may exist on a continuum in HCL insulin pump users (i.e., more severe IAH leads to overall worse hypoglycemia experiences).

The reason for persistent episodes of SH and ≥1% of time in Lv2Hypo among HCL insulin pump users may be multifactorial. Although current HCL insulin pumps proactively reduce and suspend insulin infusion during impending or ongoing hypoglycemia, both automated and manually administered subcutaneous insulin boluses are active for at least 2 h²⁴ and the glucose lowering effects can linger. Also, the lag time of interstitial glucose value measured by CGM²⁵ may delay the hypoglycemic alert and HCL insulin suspension. Physiological and behavioral counterregulatory mechanisms to prevent further declines in blood glucose and accelerate recovery from hypoglycemia thus remain crucial. The loss of epinephrine response and the associated reduction in gluconeogenesis in individuals with IAH^26,27 could contribute to SH development and prolonged time spent in Lv2Hypo. In addition, despite the availability of CGM data, hypoglycemia symptoms continue to be important somatic prompts for self-management.^28,29 Interventions that focus on restoring these counterregulatory mechanisms^30–32 to enhance physiological and behavioral responses to hypoglycemia could thus be key for minimizing clinically significant hypoglycemia in HCL insulin pump users.

The lack of association between SH and extended time spent in Lv2Hypo as well as other CGM hypoglycemia measures investigated in our sensitivity analyses may suggest that HCL insulin pump use has altered this relationship that has been established in the pre-HCL era³³ and cohorts with primarily CGM-only users.³⁴ Our subgroup analysis revealed that the SH-only subgroup spent a longer time in hyperglycemia. Although the association between SH and hyperglycemia may initially seem counterintuitive, concerns about hyperglycemia and reactive compensation behaviors (e.g., excessive bolusing or stacking of insulin boluses) have been shown to contribute to the development of SH in individuals with access to continuous glucose information.^9,35 Nonetheless, our analyses are limited by the use of short-term (i.e., 30 days rather than six months) CGM data, as well as a small sample spending a significant time in hypoglycemia; additional research incorporating a larger population and long-term CGM data is needed to fully investigate the relationship between CGM hypoglycemia measures and SH in HCL insulin pump users.

In our sample, the subgroup spending ≥1% of time in Lv2Hypo only (i.e., without SH) did not present with statistically significant higher HypoA-Q IA scores. This finding could be due to a power issue. Alternatively, IAH could be a stronger risk factor for SH¹⁷ than for Lv2Hypo in HCL insulin pump users.

Despite age and diabetes duration traditionally being associated with hypoglycemia development, particularly in research preceding the contemporary diabetes technology era,^36,37 our study suggests that HCL insulin pumps might mitigate these factors. Our findings also indicate that socioeconomic disadvantages may have continued to negatively affect hypoglycemia outcomes despite HCL insulin pump usage. Evaluating how lower education–related health literacy³⁸ or psychological complications³⁹ mediate hypoglycemia in this population may reveal potential intervention targets to further alleviate hypoglycemia. Finally, although we identified a connection between self-identification as a gender other than male or female and SH risk, the small sample size (n = 14) limits our ability to definitively conclude that these variables are related.

Clinically, our data suggest that the co-occurrence of both SH and spending ≥1% of time in Lv2Hypo is uncommon among HCL insulin pump users. Therefore, time spent in Lv2Hypo from short-term CGM data should not be used as the sole screening criterion for SH in this population; health care providers should routinely inquire about a patient’s history of SH, even among those who spend minimal time in Lv2Hypo per CGM reports. Moreover, evaluating hypoglycemia awareness (including using HypoA-Q IA) and unhelpful hypoglycemia beliefs may guide treatment to reduce hypoglycemia in this population.

The strengths of our study include a nationally recruited cohort of individuals with type 1 diabetes and minimal missing data. In addition, remote recruitment may have eased concerns about social implications,⁴⁰ potentially making participants more comfortable in accurately disclosing their SH history. Our study also has several potential limitations. SH was self-reported and not confirmed by a third party, while, due to recall bias, SH history tends to be underreported.⁴¹ The exclusive recruitment from a patient registry—together with participants’ higher SH prevalence, overall favorable glycemic control, and potential self-selection bias—may limit findings’ generalizability. Furthermore, the sample size, particularly among the subgroups spending ≥1% of time in Lv2Hypo, was small. Finally, 30-day CGM data, while being commonly used in clinical practice, may not be sufficient to investigate long-term SH outcomes.

Conclusions

From a nationally recruited cohort of adult HCL insulin pump users with type 1 diabetes, IAH was associated with both SH and spending ≥1% of time in Lv2Hypo. Participants who presented with SH demonstrated more hyperglycemia, and those spending ≥1% of time in Lv2Hypo exhibited lower glucose levels and greater glucose CVs. Social factors, such as lower education levels, may also contribute to SH risk in HCL insulin pump users. A lack of association between SH and CGM hypoglycemia measures was observed, although confirmatory research is needed. A thorough history and assessment of hypoglycemia are crucial despite the availability of CGM reports. Behavioral interventions that improve hypoglycemia awareness and address unhelpful beliefs about hypoglycemia may help prevent clinically dangerous hypoglycemia in HCL insulin pump users with type 1 diabetes.

Authors’ Contributions

M.D., S.A.A., S.J.F., and Y.K.L. were involved in the conception, design, and interpretation of the results. E.H. and A.A. collected the data. Y.K.L. and W.Y. conducted the analysis. M.D. wrote the first draft of the article, and all authors edited, reviewed, and approved the final version of the article. Y.K.L. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Footnotes

Acknowledgments

We appreciate the support from the T1D Exchange and all the study participants, without whom this study would not have been possible.

Author Disclosure Statement

S.A.A. has served on an advisory board for Vertex Pharmaceuticals, chaired a symposium presentation for Vertex Pharmaceuticals, and spoken at a symposium sponsored by Sanofi. Other authors have no conflicts of interest relevant to this study to disclose.

Funding Information

This work was supported by the National Institute of Diabetes and Digestive and Kidney Diseases (K23DK129724, U01DK135111). M.D. is supported by National Center for Advancing Translational Sciences (NCATS) TL1TR001997. REDCap was supported by NCATS UM1TR004404.

Data Availability

Data are available on request to Yu Kuei Lin.

Supplemental Material

References

Holt

RIG

, DeVries

, Hess-Fischl

, et al. The management of Type 1 Diabetes in adults. a consensus report by the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD). Diabetes Care, 2021; 44(11):2589–2625; doi: 10.2337/dci21-0043

Cryer

. The barrier of hypoglycemia in diabetes. Diabetes, 2008; 57(12):3169–3176; doi: 10.2337/db08-1084

Bilal

, Yi

, Whitaker

, et al. Impaired awareness of hypoglycemia in older adults with Type 1 Diabetes: A post hoc analysis of the WISDM study. Diabetes Care, 2024; 47(7):1202–1210; doi: 10.2337/dc24-0201

International Hypoglycaemia Study Group. Glucose concentrations of less than 3.0 mmol/L (54 mg/dL) should be reported in clinical trials: A joint position statement of the American Diabetes Association and the European Association for the Study of Diabetes. Diabetes Care, 2017; 40(1):155–157; doi: 10.2337/dc16-2215

American Diabetes Association Professional Practice C. 7. Diabetes technology: Standards of care in diabetes—2024. Diabetes Care, 2023; 47(Supplement_1):S126–S144; doi: 10.2337/dc24-S007

Brown

, Forlenza

, Bode

, et al.; Omnipod 5 Research Group. Multicenter trial of a tubeless, on-body automated insulin delivery system with customizable glycemic targets in pediatric and adult participants with type 1 diabetes. Diabetes Care, 2021; 44(7):1630–1640; doi: 10.2337/dc21-0172

Choudhary

, Kolassa

, Keuthage

, et al.; ADAPT study Group. Advanced hybrid closed loop therapy versus conventional treatment in adults with type 1 diabetes (ADAPT): A randomised controlled study. Lancet Diabetes Endocrinol, 2022; 10(10):720–731; doi: 10.1016/s2213-8587(22)00212-1

Lin

, Hung

, Sharma

, et al. Impaired awareness of hypoglycemia continues to be a risk factor for severe hypoglycemia despite the use of continuous glucose monitoring systems in Type 1 diabetes. Endocr Pract, 2019; 25(6):517–525; doi: 10.4158/ep-2018-0527

Lin

, Richardson

, Dobrin

, et al. Beliefs around hypoglycemia and their impacts on hypoglycemia outcomes in individuals with type 1 diabetes and high risks for hypoglycemia despite using advanced diabetes technologies. Diabetes Care, 2022; 45(3):520–528; doi: 10.2337/dc21-1285

10.

Sherr

, Laffel

, Liu

, et al. Severe hypoglycemia and impaired awareness of hypoglycemia persist in people with type 1 diabetes despite use of diabetes technology: Results from a cross-sectional survey. Diabetes Care, 2024; 47(6):941–947; doi: 10.2337/dc23-1765

11.

Vandenbroucke

, von Elm

, Altman

, et al.; STROBE Initiative. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): Explanation and elaboration. PLoS Med, 2007; 4(10):e297; doi: 10.1371/journal.pmed.0040297

12.

Beck

, Tamborlane

, Bergenstal

, et al.; T1D Exchange Clinic Network. The T1D Exchange clinic registry. J Clin Endocrinol Metab, 2012; 97(12):4383–4389; doi: 10.1210/jc.2012-1561

13.

Harris

, Taylor

, Thielke

, et al. Research electronic data capture (REDCap)–a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform, 2009; 42(2):377–381; doi: 10.1016/j.jbi.2008.08.010

14.

Seaquist

, Anderson

, Childs

, et al. Hypoglycemia and diabetes: A report of a workgroup of the American Diabetes Association and the endocrine society. Diabetes Care, 2013; 36(5):1384–1395; doi: 10.2337/dc12-2480

15.

Speight

, Barendse

, Singh

, et al. Characterizing problematic hypoglycaemia: Iterative design and preliminary psychometric validation of the Hypoglycaemia Awareness Questionnaire (HypoA-Q). Diabet Med, 2016; 33(3):376–385; doi: 10.1111/dme.12824

16.

Matus

, Flatt

, Peleckis

, et al. Validating and establishing a diagnostic threshold for the hypoglycemia awareness questionnaire impaired awareness subscale. Endocr Pract, 2023; 29(10):762–769; doi: 10.1016/j.eprac.2023.08.004

17.

Lin

, Ye

, Hepworth

, et al. Characterising impaired awareness of hypoglycaemia and associated risks through HypoA-Q: Findings from a T1D Exchange cohort. Diabetologia, 2025; 68(2):433–443; doi: 10.1007/s00125-024-06310-5

18.

Gold

, Macleod

, Frier

. Frequency of severe hypoglycemia in patients with Type I diabetes with impaired awareness of hypoglycemia. Diabetes Care, 1994; 17(7):697–703; doi: 10.2337/diacare.17.7.697

19.

Sepúlveda

, Poínhos

, Nata

, et al. Differentiating hypoglycemia awareness status from hypoglycemia experience in tools for measuring impaired awareness of hypoglycemia. Diabetes Technol Ther, 2020; 22(7):541–545; doi: 10.1089/dia.2020.0034

20.

Camerlingo

, Vettoretti

, Facchinetti

, et al. An analytical approach to determine the optimal duration of continuous glucose monitoring data required to reliably estimate time in hypoglycemia. Sci Rep, 2020; 10(1):18180; doi: 10.1038/s41598-020-75079-5

21.

Klonoff

, Wang

, Rodbard

, et al. A Glycemia Risk Index (GRI) of hypoglycemia and hyperglycemia for continuous glucose monitoring validated by clinician ratings. J Diabetes Sci Technol, 2023; 17(5):1226–1242; doi: 10.1177/19322968221085273

22.

Kotrlik

, Higgins

. Organizational research: Determining appropriate sample size in survey research. Information Technology, Learning, and Performance J, 2001; 19(1):43.

23.

Hernar

, Strandberg

, Nilsen

, et al. Prevalence and risk factors for impaired awareness of hypoglycaemia: A registry-based study of 10,202 adults with type 1 diabetes in Norway. Diabet Med, 2025; 42(3):e15480; doi: 10.1111/dme.15480

24.

Attia

, Jones

, Holcombe

, et al. Comparison of human regular and lispro insulins after interruption of continuous subcutaneous insulin infusion and in the treatment of acutely decompensated IDDM. Diabetes Care, 1998; 21(5):817–821; doi: 10.2337/diacare.21.5.817

25.

Zaharieva

, Turksoy

, McGaugh

, et al. Lag time remains with newer real-time continuous glucose monitoring technology during aerobic exercise in adults living with Type 1 Diabetes. Diabetes Technol Ther, 2019; 21(6):313–321; doi: 10.1089/dia.2018.0364

26.

Dagogo-Jack

, Craft

, Cryer

. Hypoglycemia-associated autonomic failure in insulin-dependent diabetes mellitus. Recent antecedent hypoglycemia reduces autonomic responses to, symptoms of, and defense against subsequent hypoglycemia. J Clin Invest, 1993; 91(3):819–828; doi: 10.1172/jci116302

27.

Flatt

, Peleckis

, Dalton-Bakes

, et al. Automated insulin delivery for hypoglycemia avoidance and glucose counterregulation in long-standing type 1 diabetes with hypoglycemia unawareness. Diabetes Technol Ther, 2023; 25(5):302–314; doi: 10.1089/dia.2022.0506

28.

Lin

, Agni

, Chuisano

, et al. Facilitators and barriers to hypoglycemia self-management in adults living with type 1 diabetes and using advanced diabetes technologies. Diabetes Technol Ther, 2023.

29.

Lin

, Ye

, Hepworth

, et al. Impaired awareness and symptomatology of hypoglycemia in adults with Type 1 diabetes using continuous glucose monitoring. J Clin Endocrinol Metab, 2024:dgae846; doi: 10.1210/clinem/dgae846

30.

Cox

, Gonder-Frederick

, Polonsky

, et al. A multicenter evaluation of blood glucose awareness training-II. Diabetes Care, 1995; 18(4):523–528; doi: 10.2337/diacare.18.4.523

31.

Amiel

, Potts

, Goldsmith

, et al. A parallel randomised controlled trial of the Hypoglycaemia awareness restoration programme for adults with type 1 diabetes and problematic hypoglycaemia despite optimised self-care (HARPdoc). Nat Commun, 2022; 13(1):2229; doi: 10.1038/s41467-022-29488-x

32.

Lin

, Ye

, Rogers

, et al. Mitigating severe hypoglycemia in users of advanced diabetes technologies: Impaired awareness of hypoglycemia and unhelpful hypoglycemia beliefs as targets for interventions. Diabetes Technol Ther, 2024; 26(10):739–747; doi: 10.1089/dia.2024.0039

33.

Beck

, Bergenstal

, Riddlesworth

, et al. The association of biochemical hypoglycemia with the subsequent risk of a severe hypoglycemic event: Analysis of the DCCT data set. Diabetes Technol Ther, 2019; 21(1):1–5; doi: 10.1089/dia.2018.0362

34.

Deshmukh

, Wilmot

, Choudhary

, et al. Time below range and its influence on hypoglycemia awareness and severe hypoglycemia: Insights from the association of British clinical diabetologists study. Diabetes Care, 2025; 48(3):437–443; doi: 10.2337/dc24-1833

35.

Lin

, Agni

, Chuisano

, et al. ‘You have to use everything and come to some equilibrium’: a qualitative study on hypoglycemia self-management in users of continuous glucose monitor with diverse hypoglycemia experiences. BMJ Open Diabetes Res Care, 2023; 11(3):e003415; doi: 10.1136/bmjdrc-2023-003415

36.

Weinstock

, Xing

, Maahs

, et al.; T1D Exchange Clinic Network. Severe hypoglycemia and diabetic ketoacidosis in adults with type 1 diabetes: Results from the T1D Exchange clinic registry. J Clin Endocrinol Metab, 2013; 98(8):3411–3419; doi: 10.1210/jc.2013-1589

37.

Allen

, LeCaire

, Palta

, et al.; Wisconsin Diabetes Registry Project. Risk factors for frequent and severe hypoglycemia in Type 1 Diabetes. Diabetes Care, 2001; 24(11):1878–1881; doi: 10.2337/diacare.24.11.1878

38.

Jansen

, Rademakers

, Waverijn

, et al. The role of health literacy in explaining the association between educational attainment and the use of out-of-hours primary care services in chronically ill people: a survey study. BMC Health Serv Res, 2018; 18(1):394; doi: 10.1186/s12913-018-3197-4

39.

, Allebeck

, Burstöm

, et al. Educational level and the risk of mental disorders, substance use disorders and self-harm in different age-groups: A cohort study covering 1,6 million subjects in the Stockholm region. Int J Methods Psychiatr Res, 2023; 32(4):e1964; doi: 10.1002/mpr.1964

40.

Pedersen-Bjergaard

, Færch

, Allingbjerg

M-L

, et al. The influence of new European Union driver’s license legislation on reporting of severe hypoglycemia by patients with Type 1 diabetes. Diabetes Care, 2015; 38(1):29–33; doi: 10.2337/dc14-1417

41.

Jørgensen

, Pedersen-Bjergaard

, Rasmussen

, et al. The impact of severe hypoglycemia and impaired awareness of hypoglycemia on relatives of patients with type 1 diabetes. Diabetes Care, 2003; 26(4):1106–1109; doi: 10.2337/diacare.26.4.1106

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.10 MB