Continuous Glucose Monitoring in Patients with Type 1 Diabetes and Impaired Awareness of Hypoglycemia: Also Effective in Patients with Psychological Distress?

Abstract

The aim of this study was to evaluate whether psychological distress modifies the effect of continuous glucose monitoring (CGM) in patients with type 1 diabetes (T1D) and impaired awareness of hypoglycemia. Fifty-two patients with T1D and impaired awareness of hypoglycemia participated in an earlier reported randomized crossover trial with two 16-week intervention periods comparing CGM with self-monitoring of blood glucose (SMBG). During the CGM phase, time spent in euglycemia (4–10 mmol/L), the primary outcome, was 9.6% higher compared with the SMBG phase (P < 0.0001). Psychological distress was operationalized as low emotional well-being (World Health Organization Well-being Index 5 [WHO-5] < 50), high diabetes-related distress (Problem Areas in Diabetes 5 [PAID-5] ≥ 8), and/or high fear of hypoglycemia (Hypoglycemia Fear Survey [HFS] Worry > mean HFS Worry score +1 standard deviation). Modifying effects were assessed by analyzing psychological distress score × intervention—interaction effects. Results showed that both the low emotional well-being group and normal emotional well-being group had equal glycemic outcomes during the CGM phase. High diabetes distress and elevated fear of hypoglycemia did not result in significant interaction effects for glycemic outcomes. This study demonstrated that CGM is equally effective in terms of glycemic improvements in high versus low distressed patients with T1D and impaired awareness of hypoglycemia.

Introduction

Continuous glucose monitoring (CGM) can lower HbA1c¹ and reduce the risk of severe hypoglycemia in patients with type 1 diabetes (T1D) and impaired awareness of hypoglycemia.^2
–4 Impaired awareness of hypoglycemia affects ∼25% of adults with T1D and increases the risk of severe hypoglycemia substantially.^5,6 CGM may also help to reduce the adverse psychosocial effects of unpredictable recurrent hypoglycemia in T1D, but evidence is limited and inconsistent.^7
–9 In our recently reported crossover trial, CGM reduced the occurrence of severe hypoglycemia by 59% and significantly lowered worry of hypoglycemia.⁴ It is important to note that CGM can aid patients to self-manage their diabetes more precisely, provided they are capable of handling the device and data feedback adequately, with support of a diabetes healthcare team.^10,11 In this context, the question comes up whether CGM is suitable and beneficial for patients with an unfavorable psychological profile, for example, with high psychological distress. Psychological distress is common in diabetes, including low emotional well-being, high diabetes-related distress, and fear of hypoglycemia, negatively affecting patients' daily self-care and glycemic control.^12

–16 With CGM, patients are faced with real-time feedback on blood glucose variation and alarms that may be experienced as stressful and difficult to handle for those with pre-existing high levels of distress.^10,17,18 To the best of our knowledge, no trials investigating the effect of CGM have studied the modifying role of psychological distress on treatment outcomes.

We therefore aimed to investigate whether psychological distress, operationalized as low emotional well-being, high diabetes-related distress, and/or elevated fear of hypoglycemia, modifies the effect of CGM on glycemic outcomes in patients with T1D and impaired awareness of hypoglycemia.

Patients and Methods

The trial was a two-center, randomized, open-label crossover trial with a 12-week washout period in between 16-week intervention periods, comparing CGM with self-monitoring of blood glucose (SMBG).¹⁹ Fifty-two adults with T1D and impaired awareness of hypoglycemia as confirmed by a Gold score⁶ ≥ 4 and treated with either continuous subcutaneous insulin infusion (CSII) or multiple daily injections were randomized to the CGM/SMBG (n = 26) or SMBG/CGM (n = 26) sequence. During the CGM phase, patients were instructed to use the CGM device continuously. During the SMBG phase, patients were equipped with a masked CGM system, which does not display real-time glucose values. During both intervention periods, patients attended monthly follow-up visits to review glucose data, based on CGM data during the CGM phase and SMBG data during the SMBG phase, and make therapy changes accordingly. These visits were followed by telephone consultations 2 weeks after each follow-up visit comprising inquiry after adverse events, episodes of (severe) hypoglycemia, study device use and related technical issues, and check on medication changes.

Measurements

The primary endpoint of the trial was the difference in percentage of time spent in euglycemia (4–10 mmol/L) between the CGM and SMBG phases. Secondary outcomes included the percentage of time patients spent in hypoglycemia (≤3.9 mmol/L), average daily area under the curve ≤3.9 mmol/L (expressed as mmol/L × minutes), frequency (episodes per week) of CGM-derived hypoglycemic episodes ≤3.9 mmol/L and <2.8 mmol/L, duration (minutes per episode) of CGM-derived hypoglycemic episodes ≤3.9 mmol/L, severe hypoglycemia (requiring third-party assistance), and psychological distress scores (World Health Organization Well-being Index 5 [WHO-5], Problem Areas in Diabetes 5 [PAID-5], and Hypoglycemia Fear Survey [HFS] Worry). The primary and secondary outcomes of the trial have been reported previously.⁴

At baseline and endpoint of both intervention periods, the psychological status of patients was assessed, covering three indicators of emotional distress: emotional well-being, diabetes distress, and fear of hypoglycemia. Emotional well-being was assessed using the WHO-5 items questionnaire.²⁰ A score of <50 indicates clinically relevant low emotional well-being.²⁰ Diabetes-related distress was assessed using the PAID-5 Short Form.²¹ A score of ≥8 indicates high diabetes distress. Fear of hypoglycemia was assessed with the HFS.²² The HFS comprises a Worry and a Behavior subscale. In this study, we only used data from the Worry subscale, with higher scores indicating more hypoglycemia fear. A definite cutoff score for clinically meaningful fear of hypoglycemia in patients with T1D has yet to be established. Therefore, in line with Hajos et al.,²³ we used the mean baseline HFS Worry score (38.4) + 1 standard deviation (19.2) as a cutoff value for elevated fear of hypoglycemia.

Statistical methods

Using intention-to-treat analyses, we analyzed the effect of CGM on glycemic outcomes using linear mixed model analysis with the glycemic outcome as the dependent variable, the treatment group (CGM or SMBG) as a fixed factor, and the participant as a random factor. To demonstrate whether indicators of psychological distress modified the effect of CGM on glycemic outcomes, we included a two-way interaction term (psychological distress score × intervention [CGM or SMBG]) in the linear mixed model. If a significant interaction effect was seen, we conducted post hoc analyses on the subgroups, using similar linear mixed models. Statistical significance for interaction effects was set at 0.1 (2-tailed) and statistical significance for all other analyses was set at 0.05 (2-tailed). Patients were scored as having low emotional well-being if the WHO-5 score was <50 at baseline of the CGM phase or at baseline of the SMBG phase. Likewise, patients were scored as having high diabetes distress if the PAID-5 score was ≥8 at baseline of the CGM phase or at baseline of the SMBG phase. Elevated fear of hypoglycemia was reported in patients with an HFS Worry score >57.6 at baseline of the CGM phase or at baseline of the SMBG phase. All analyses were conducted using SPSS 22.0 for Windows (IBM SPSS, Inc., Chicago, IL).

Results

The mean age of the population was 48.6 ± 11.6 years, 24 (46.2%) were females, median duration of diabetes was 30.5 years (interquartile range 18.5–40.8), 23 patients (44.2%) were treated with CSII, mean HbA1c was 7.5% ± 0.8%, and all patients reported impaired awareness of hypoglycemia, with a mean Gold score of 5.4 ± 0.7. The intention-to-treat population consisted of 52 randomized patients.

Questionnaire completion was high at baseline and endpoint of both intervention periods (baseline CGM 50 patients [96%] and endpoint CGM 48 patients [92%]; baseline SMBG 49 patients [94%] and endpoint SMBG 47 patients [90%]). Thirteen patients (25%) reported low emotional well-being (WHO-5 score <50) indicative of depressive effect at baseline of the CGM phase or at baseline of the SMBG phase, 19 (36.5%) patients reported high diabetes distress (PAID-5 score ≥8) at baseline of the CGM phase or at baseline of the SMBG phase, and 13 patients (25%) reported elevated fear of hypoglycemia (HFS Worry score >57.6) at baseline of the CGM phase or at baseline of the SMBG phase. Mean PAID-5 score was 6.0 (95% confidence interval [CI] 5.0–7.1) at baseline of the CGM phase and 5.9 (95% CI 4.9–7.0) at baseline of the SMBG phase. Mean WHO-5 score was 65.9 (95% CI 60.6–71.2) at baseline of the CGM phase and 62.9 (95% CI 57.6–68.3) at baseline of the SMBG phase. Mean HFS Worry score was 38.9 (95% CI 33.5–44.2) at baseline of the CGM phase and 38.2 (95% CI 32.8–43.6) at baseline of the SMBG phase. The 16-week HFS Worry score was significantly lower after the CGM phase compared with the 16-week HFS Worry score after the SMBG phase (32.5 vs. 38.9; Δ 6.4 [95% CI 1.4–11.4], P = 0.014), but no significant differences were seen in the 16-week WHO-5 or PAID-5 scores comparing CGM with SMBG (data not shown).

The interaction term “low emotional well-being × intervention” demonstrated significant effects for the primary outcome of the trial (time spent in euglycemia, P = 0.024) and for the hypoglycemia-related outcomes (time spent in hypoglycemia, P = 0.009; frequency of hypoglycemic events ≤3.9 mmol/L, P = 0.098; frequency of hypoglycemic events <2.8 mmol/L, P = 0.019; duration of hypoglycemia, P = 0.079; and area under the curve in hypoglycemia, P = 0.007). Table 1 shows post hoc subgroup analyses of the outcomes with significant interaction effects (emotional well-being × intervention). Although still significant for the primary outcome (time spent in euglycemia), the effect sizes of CGM versus SMBG on glycemic outcomes were consistently lower in 13 patients with low emotional well-being compared with the 39 patients with normal emotional well-being. These lower effect sizes were mainly caused by a higher level of glycemic control during the SMBG phase in the low emotional well-being group rather than a lower level of glycemic control during the CGM phase in the low emotional well-being group (Table 1). The interaction term “high diabetes distress × intervention” resulted in one significant interaction effect for time spent in euglycemia (P = 0.05), but no significant interaction effects were seen for the other glycemic outcomes. Again, post hoc analyses showed no relevant difference in time spent in euglycemia during the CGM phase between patients with high diabetes distress compared with those without high diabetes distress (65.4% [95% CI 61.8–69.1] vs. 64.5% [61.5–67.4]). No significant interaction effects were found for the interaction term “elevated fear of hypoglycemia × intervention.”

Table 1.

Post Hoc Analyses of Items with Significant Emotional Well-Being × Intervention Interaction Effects

Glycemic outcome	CGM	SMBG	Mean difference (95% CI)^a	P
Low emotional well-being (WHO-5 < 50, n = 13)
Percent of time spent with glucose level in range
4.0–10 mmol/L	65.0 (60.2 to 69.8)	58.4 (53.6 to 63.2)	6.6 (2.5 to 10.7)	0.004
≤3.9 mmol/L	6.8 (3.7 to 9.9)	9.0 (5.9 to 12.1)	2.2 (−0.4 to 4.9)	0.091
Frequency of hypoglycemia ≤3.9 mmol/L (events/week)	9.2 (6.7 to 11.7)	9.2 (6.7 to 11.7)	0.1 (−2.2 to 2.1)	0.939
Frequency of hypoglycemia ≤2.8 mmol/L (events/week)	2.8 (1.2 to 4.5)	3.5 (1.8 to 5.2)	0.6 (−0.9 to 2.1)	0.380
Duration of CGM-derived hypoglycemic events (minutes/event)	64.9 (51.6 to 78.2)	94.5 (81.2 to 107.8)	29.6 (16.2 to 43.1)	0.001
AUC ≤3.9 mmol/L per 24 h (mmol/L × minutes)	69.3 (30.7 to 107.9)	59.8 (51.2 to 128.5)	20.5 (−12.8 to 53.8)	0.200
Normal emotional well-being (WHO-5 ≥ 50, n = 39)
Percent of time spent with glucose level in range
4.0–10 mmol/L	64.7 (62.1 to 67.3)	53.9 (51.3 to 56.5)	10.8 (9.1 to 12.5)	<0.0001
≤3.9 mmol/L	6.6 (4.7 to 8.5)	12.4 (10.5 to 14.3)	5.8 (45 to 7.1)	<0.0001
Frequency of hypoglycemia ≤3.9 mmol/L (events/week)	10.2 (8.5 to 11.9)	12.0 (10.3 to 13.7)	1.7 (0.7 to 2.7)	0.001
Frequency of hypoglycemia ≤2.8 mmol/L (events/week)	1.9 (1.1 to 2.7)	4.2 (3.4 to 5.0)	2.3 (1.6 to 2.9)	<0.0001
Duration of CGM-derived hypoglycemic events (minutes/event)	58.1 (51.3 to 64.8)	100.5 (93.7 to 10.7.2)	42.4 (34.4 to 50.4)	<0.0001
AUC ≤3.9 mmol/L per 24 h (mmol/L × minutes)	59.8 (38.5 to 81.3)	126.6 (105.1 to 148.0)	66.7 (50.7 to 82.8)	<0.0001

Data are mean (95% CI).

Result of linear mixed model analysis with glycemic outcome as the dependent variable, the treatment group (CGM or SMBG) as a fixed factor, and the participant as a random factor.

AUC, area under the curve; CGM, continuous glucose monitoring; CI, confidence interval; SMBG, self-monitoring of blood glucose; WHO-5, World Health Organization Well-being Index 5.

Discussion

We report on the possible modifying effect of psychological distress, specifically emotional well-being, diabetes-related distress, and fear of hypoglycemia, on CGM outcomes as recently reported in a trial in patients with T1D and impaired awareness of hypoglycemia, comparing CGM versus SMBG.⁴ We observed lower effect sizes in the low emotional well-being group compared with the normal emotional well-being group, which could suggest that CGM is less effective in patients with low emotional well-being. However, the effect of CGM on the primary outcome (time spent in euglycemia) remained significant also in patients with low emotional well-being. Furthermore, the lower effect sizes were mainly due to better glycemic control during the SMBG phase rather than poorer glycemic control during the CGM phase. No relevant or consistent interaction effects were observed between diabetes-related distress or fear of hypoglycemia and the glycemic outcomes.

With regard to the external validity, our study population was relatively small, well educated, and lacked ethnic minorities. Otherwise, the clinical profile of our study sample does not suggest selection bias. The mean age, HbA1c, and diabetes duration of our study population seem comparable with baseline characteristics of patients included in other trials in diabetes patients with impaired awareness of hypoglycemia or frequent severe hypoglycemia.^2,3,5,24 The mean total WHO-5 score and mean total PAID-5 score are comparable with WHO-5 and PAID-5 scores reported in previously published trials in type 1 and 2 diabetes patients without established impaired awareness of hypoglycemia.^7,25,26 Not unexpectedly, the mean HFS Worry score indicated that the patients in our study report that they worry more about hypoglycemia relative to T1D patients without established impaired awareness of hypoglycemia.^7,8,27 Further research in larger and more diverse T1D patient populations is warranted.

The crossover design is a limitation as the level of psychological distress can change during the course of a study period, although baseline scores at the two respective study periods were not dissimilar. Future studies should replicate our findings in studies not using a crossover design. We did not find an indication that psychological distress adversely affects CGM use in this patient population with impaired awareness of hypoglycemia. However, any current psychiatric disorder, including major depression, was an exclusion criterion limiting generalizability. Further research should help clarify if and to what extent GCM outcomes are negatively impacted when used in patients with more severe psychological distress. In addition, all patients had impaired awareness of hypoglycemia, limiting generalization to other patient populations. The level of support during the trial was relatively high compared with usual care. It could be hypothesized that patients with psychological distress require a higher level of support compared with those without psychological distress and that this higher level of support cannot be met in usual clinical practice.

In conclusion, low emotional well-being, high diabetes-related distress, and fear of hypoglycemia do not substantially modify the effect of CGM on glycemic outcomes in T1D patients with impaired awareness of hypoglycemia. We can therefore expect these more psychologically vulnerable patients to benefit from CGM as much as those with a more favorable psychological profile.

Footnotes

Acknowledgments

This research received funding from Eli Lilly and Sanofi. Medtronic provided continuous glucose monitoring devices. Through M.H.H.K. and M.D., the VU University Medical Center received research grants from AstraZeneca, Boehringer Ingelheim, Novo Nordisk, and Sanofi. J.H.D. or his institution received research support and speaker's fees from Abbott, Dexcom, Medtronic, Roche Diabetes Care, and Senseonics. P.H.G.-D. received speaker's fees from Abbott and Medtronic and is a member of the advisory board of Medtronic.

Authors' Contributions

M.D. and S.J.K. conceived and designed the trial. E.H.S. was the principal investigator in Amsterdam and was involved in the design of the protocol, coordination of the study, and drafting of the article. P.H.G. was the principal investigator in The Hague and was involved in the design of the protocol, coordination of the study, and drafting of the article. C.A.J.B., J.H.D., M.H.H.K., and F.J.S. participated in the study design. C.A.J.B. was the main study physician responsible for data gathering in both trial centers, data analysis, and drafting of the first version of the article. S.J.K. gathered a significant proportion of the data in the VU University Medical Center, Amsterdam, The Netherlands. P.H.G. supervised the study physicians C.A.J.B. and S.J.K. M.D. obtained funding for the trial. C.A.J.B. performed the data analysis and M.W., J.H.D., P.H.G., M.H.H.K., E.H.S., and F.J.S. made significant contributions to the data analysis. M.W., J.H.D., S.J.K., P.H.G., M.H.H.K., E.H.S., and F.J.S. reviewed and commented on article drafts.

Author Disclosure Statement

No competing financial interests exist.

References

Pickup

, Freeman

, Sutton

: Glycaemic control in type 1 diabetes during real time continuous glucose monitoring compared with self monitoring of blood glucose: meta-analysis of randomised controlled trials using individual patient data. BMJ, 2011; 343:d3805.

Choudhary

, Ramasamy

, Green

, et al.: Real-time continuous glucose monitoring significantly reduces severe hypoglycemia in hypoglycemia-unaware patients with type 1 diabetes. Diabetes Care, 2013; 36:4160–4162.

, Nicholas

, Retterath

, et al.: Effect of sensor-augmented insulin pump therapy and automated insulin suspension vs standard insulin pump therapy on hypoglycemia in patients with type 1 diabetes: a randomized clinical trial. JAMA, 2013; 310:1240–1247.

van Beers

, DeVries

, Kleijer

, et al.: Continuous glucose monitoring for patients with type 1 diabetes and impaired awareness of hypoglycaemia (IN CONTROL): a randomised, open-label, crossover trial. Lancet Diabetes Endocrinol, 2016; 4(11):893–902.

Geddes

, Schopman

, Zammitt

, et al.: Prevalence of impaired awareness of hypoglycaemia in adults with type 1 diabetes. Diabet Med, 2008; 25:501–504.

Gold

, MacLeod

, Frier

: Frequency of severe hypoglycemia in patients with type I diabetes with impaired awareness of hypoglycemia. Diabetes Care, 1994; 17:697–703.

Hermanides

, Norgaard

, Bruttomesso

, et al.: Sensor-augmented pump therapy lowers HbA(1c) in suboptimally controlled type 1 diabetes; a randomized controlled trial. Diabet Med, 2011; 28:1158–1167.

Beck

, Lawrence

, Laffel

, et al.: Quality-of-life measures in children and adults with type 1 diabetes: Juvenile Diabetes Research Foundation continuous glucose monitoring randomized trial. Diabetes Care, 2010; 33:2175–2177.

Polonsky

, Hessler

, Ruedy

, et al.: The impact of continuous glucose monitoring on markers of quality of life in adults with type 1 diabetes: further findings from the DIAMOND randomized clinical trial. Diabetes Care, 2017; 40(6):736–741.

10.

Ritholz

, Atakov-Castillo

, Beste

, et al.: Psychosocial factors associated with use of continuous glucose monitoring. Diabet Med, 2010; 27:1060–1065.

11.

Peters

, Ahmann

, Battelino

, et al.: Diabetes technology-continuous subcutaneous insulin infusion therapy and continuous glucose monitoring in adults: an endocrine society clinical practice guideline. J Clin Endocrinol Metab, 2016; 101:3922–3937.

12.

Rubin

, Peyrot

: Psychological issues and treatments for people with diabetes. J Clin Psychol, 2001; 57:457–478.

13.

Lustman

, Anderson

, Freedland

, et al.: Depression and poor glycemic control: a meta-analytic review of the literature. Diabetes Care, 2000; 23:934–942.

14.

Wild

, von

, Brohan

, et al.: A critical review of the literature on fear of hypoglycemia in diabetes: implications for diabetes management and patient education. Patient Educ Couns, 2007; 68:10–15.

15.

Barendse

, Singh

, Frier

, et al.: The impact of hypoglycaemia on quality of life and related patient-reported outcomes in type 2 diabetes: a narrative review. Diabet Med, 2012; 29:293–302.

16.

Fisher

, Hessler

, Polonsky

, et al.: Diabetes distress in adults with type 1 diabetes: prevalence, incidence and change over time. J Diabetes Complications, 2016; 30:1123–1128.

17.

Tanenbaum

, Hanes

, Miller

, et al.: Diabetes device use in adults with type 1 diabetes: barriers to uptake and potential intervention targets. Diabetes Care, 2017; 40:181–187.

18.

Pickup

, Ford

, Samsi

: Real-time continuous glucose monitoring in type 1 diabetes: a qualitative framework analysis of patient narratives. Diabetes Care, 2015; 38:544–550.

19.

van Beers

, Kleijer

, Serne

, et al.: Design and rationale of the IN CONTROL trial: the effects of real-time continuous glucose monitoring on glycemia and quality of life in patients with type 1 diabetes mellitus and impaired awareness of hypoglycemia. BMC Endocr Disord, 2015; 15:42.

20.

Hajos

, Pouwer

, Skovlund

, et al.: Psychometric and screening properties of the WHO-5 well-being index in adult outpatients with type 1 or type 2 diabetes mellitus. Diabet Med, 2013; 30:e63–e69.

21.

McGuire

, Morrison

, Hermanns

, et al.: Short-form measures of diabetes-related emotional distress: the Problem Areas in Diabetes Scale (PAID)-5 and PAID-1. Diabetologia, 2010; 53:66–69.

22.

Cox

, Irvine

, Gonder-Frederick

, et al.: Fear of hypoglycemia: quantification, validation, and utilization. Diabetes Care, 1987; 10:617–621.

23.

Hajos

, Polonsky

, Pouwer

, et al.: Toward defining a cutoff score for elevated fear of hypoglycemia on the Hypoglycemia Fear Survey worry subscale in patients with type 2 diabetes. Diabetes Care, 2014; 37:102–108.

24.

Rondags

, de Wit

, Twisk

, et al.: Effectiveness of HypoAware, a brief partly web-based psychoeducational intervention for adults with type 1 and insulin-treated type 2 diabetes and problematic hypoglycemia: a cluster randomized controlled trial. Diabetes Care, 2016; 39(12):2190–2196.

25.

Snoek

, Kersch

, Eldrup

, et al.: Monitoring of individual needs in diabetes (MIND): baseline data from the cross-national diabetes attitudes, wishes, and needs (DAWN) MIND study. Diabetes Care, 2011; 34:601–603.

26.

Nicolucci

, Kovacs

, Holt

, et al.: Diabetes attitudes, wishes and needs second study (DAWN2): cross-national benchmarking of diabetes-related psychosocial outcomes for people with diabetes. Diabet Med, 2013; 30:767–777.

27.

Rubin

, Peyrot

: Health-related quality of life and treatment satisfaction in the sensor-augmented pump therapy for A1C reduction 3 (STAR 3) trial. Diabetes Technol Ther, 2012; 14:143–151.