Abstract
Background:
Prediction of risk of severe hypoglycemia (SH) in patients with type 1 diabetes is important to prevent future episodes, but it is unknown if it is possible to predict the long-term risk of SH. The aim of the study is to assess if long-term prediction of SH is possible in type 1 diabetes.
Methods:
A follow-up study was performed with 98 patients with type 1 diabetes. At baseline and at follow-up, the patients filled in a questionnaire about diabetes history and complications, number of SH in the preceding year and state of awareness, and HbA1c and C-peptide levels were measured.
Results:
During the 12 years of follow-up, there was a decrease in HbA1c, C-peptide levels, and incidence of SH (1.1 to 0.4 episodes per patient-year; P < .001). At baseline, the relative rate of SH was 3.6 (P = .001) and 10.9 (P < .0001) in patients with impaired awareness and unawareness of hypoglycemia, respectively, as compared to patients with normal awareness. At follow-up, patients with unawareness at baseline tended to have maintained an increased rate of SH (RR = 3.1; P = .07). Impaired awareness, HbA1c and C-peptide determined at baseline did not correspond with an increased rate of SH at follow-up.
Conclusions:
Long-term prediction of severe hypoglycemia in type 1 diabetes was not possible, although baseline hypoglycemia unawareness tended to remain a predictor for risk of SH at follow-up. Therefore, it is important repeatedly to assess the different risk factors of SH to determine the actual risk.
Keywords: prediction, severe hypoglycemia, type 1 diabetes, hypoglycemia awareness
Severe hypoglycemia is still a major clinical problem in the treatment of type 1 diabetes that impairs metabolic control,1,2 despite improvement of the diabetes management including the introduction of different hypoglycemia detecting and alarm devices, for example, continuous glucose monitoring systems.3 Therefore, the ability to predict risk of severe hypoglycemia in patients with type 1 diabetes is important to prevent future episodes. Several risk factors for severe hypoglycemia are known and described in previous studies, for example, impaired hypoglycemia awareness, long duration of diabetes, strict glycemic control and loss of endogenous insulin production (undetectable C-peptide).4-7 However, the majority of these studies are short-term studies lasting 1 year or less except 2 older studies8,9 with a follow-up of 3-14 years. Patients’ state of awareness is crucial to perceive and react to hypoglycemia and remains relatively constant during a 1-year period. However, hypoglycemia awareness may be impaired during periods with (too) tight glycemic control or—opposite—improved by interventions successful in strict avoidance of hypoglycemia.10,11 Likewise, endogenous insulin production that is a prerequisite for mounting a glucagon response during hypoglycemia disappears in most patients with time.1 At present, it is unknown if these important risk factors can predict the risk of severe hypoglycemia over a longer period than 1 year.
The aim of the study was to describe the changes in severe hypoglycemia in a cohort of type 1 diabetes and to assess if long-term prediction of severe hypoglycemia is possible in type 1 diabetes.
Methods
Patients and Design
From 1999 to 2001, 269 outpatients with type 1 diabetes followed at Nordsjællands Hospital Hillerød were consecutively enrolled in a cross-sectional study. In 2012, it was possible to reexamine 98 of these patients (during follow-up, 39 patients had died, 60 patients had ended their control at our outpatient clinic and 71 patients declined to participate in the follow-up study in 2012). At entry, included patients were adults (>18 years) with type 1 diabetes for more than 2 years. Pregnant women, patients with end-stage renal disease or with concomitant malignant disease were excluded from participation.
In this study, we defined short-term prediction of hypoglycemia as prediction of hypoglycemia over a period of 1 year or less and long-term prediction as the ability to predict the risk of severe hypoglycemia over a period of more than 1 year.
The Regional Committee on Biomedical Research Ethics approved the study.
Questionnaires
At baseline and at follow-up in 2012, the patients filled in a questionnaire about number of severe hypoglycemic episodes in the preceding year and self-estimated state of hypoglycemia awareness. The questionnaire has been described in previous studies.5 Information about history of diabetes, treatment and late complications were extracted from the patients’ medical records at baseline and follow-up.
Severe hypoglycemia was defined as an episode requiring assistance of another person actively to correct the plasma glucose to normal.12 The primary end-point was the rate of severe hypoglycemic episodes in the preceding year. One-year recall of rate of severe hypoglycemia is a validated measure as shown previously in this cohort.13 State of awareness was assessed by the validated Hillerød method based on the question: “Do you recognize symptoms, when you have a hypo?” According to the answer, patients are classified as having normal or impaired awareness or hypoglycemia unawareness.13,14 Thus, this method divides hypoglycemia awareness into 3 groups.
Laboratory Analyses
HbA1c and C-peptide were measured at baseline and at follow-up. If C-peptide levels were undetectable (<10 pmol/l), the patient was classified as C-peptide negative.
Statistical Analyses
Standard descriptive statistics were used to characterize patients at baseline and follow-up. Comparisons between variables were made by parametric (chi-square test, paired-samples t-test) and nonparametric (Wilcoxon signed rank test) methods as appropriate. The influence of well-known risk factors on number of severe hypoglycemic events were assessed by a log-linear negative binomial model, which takes into account the skewed distribution of events of severe hypoglycemia. The risk factors included in the analysis were state of hypoglycemia awareness, duration of diabetes, HbA1c and C-peptide status. Calculations were done by using SPSS software package (Version 22.0). Level of statistical significance was chosen as P value < .05 (2-sided).
Results
Baseline characteristics of the patients are shown in Table 1. At baseline, there were no significant differences between the 98 participants and 171 nonparticipants according to age, diabetes duration, HbA1c, or awareness status but peripheral neuropathy and macrovascular complications were more frequent in nonparticipants. Of the participants, 57% were male, and at baseline the mean age was 44 years and the mean duration of diabetes was 20 years. In 2012, 98 patients participated in the follow-up with a median follow-up of 12 years.
Table 1.
Baseline and Follow-Up Characteristics of the Participants and Nonparticipants With Type 1 Diabetes.
| Baseline |
Follow-up | P a | ||
|---|---|---|---|---|
| Total | Participants | |||
| Number of subjects, n (%) | 269 (100) | 98 (36) | 98 (36) | |
| Age (years) | 45 ± 13 | 44 ± 10 | 56 ± 10 | |
| Duration of diabetes (years) | 20 ± 12 | 20 ± 11 | 32 ± 11 | |
| C-peptide negative (%)b | 40 | 42 | 91 | <.01 |
| Body mass index (kg/m2) | 25 ± 4 | 25 ± 4 | 26 ± 4 | <.01 |
| Total insulin dose (units/day) | 52 ± 19 | 50 ± 17 | 51 ± 19 | .87 |
| Insulin analogues (%) | 5 | 5 | 58 | <.01 |
| Insulin pump (%) | 0.5 | 1 | 21 | <.01 |
| HbA1c | ||||
| - % | 8.6 ± 1.2 | 8.7 ± 1.2 | 8.0 ± 0.9 | <.01 |
| - mmol/mol | 70 ± 11 | 72 ± 11 | 63 ± 11 | |
| Diabetic complications (%) | ||||
| - Retinopathy | 56 | 57 | ||
| - Nephropathy | 30 | 29 | ||
| - Peripheral neuropathy | 35 | 25c | ||
| - Autonomic neuropathy | 22 | 10 | ||
| - Macrovascular complications | 8 | 3c | ||
| Awareness (%) | ||||
| Hillerød method (aware/impaired/unaware) | 44/44/12 | 41/47/11 | 36/53/11 | .68 |
| Episodes of severe hypoglycemia (per patient, last year) | 1.1 ± 0.2 | 1.1 ± 0.3 | 0.4 ± 0.2 | <.01 |
Values are mean ± SD or percentage unless otherwise indicated.
Paired P value comparing change in characteristics for participants from baseline to follow-up.
Undetectable: <10 pmol/l.
Significant difference between participants and nonparticipants at baseline (P < .05); data not shown.
From baseline to follow-up, the percentage of C-peptide-negative patients increased from 42% at baseline to 91% at follow-up (P < .001) while the average HbA1c level decreased from 72 mmol/mol to 63 mmol/mol (P < .001) (Table 1). Only few patients were treated with insulin analogs or insulin pumps at baseline while at follow-up 58% and 21% of the patients used these treatment modalities, respectively (P < .001). The daily dose of insulin did not change in the follow-up period.
The yearly prevalence of severe hypoglycemia decreased from 38% at baseline to 22% at follow-up. The incidence of severe hypoglycemia decreased from 1.1 episodes per patient-year at baseline to 0.4 (P < .001) episodes per patient-year at follow-up. Although the prevalence of impaired awareness and unawareness of hypoglycemia was unaltered during follow-up, a significant portion of the patients (39%) changed state of awareness from baseline to follow-up (P < .001), including 17% with improved and 21% with decreased hypoglycemia awareness after 12 years of follow-up (Table 2).
Table 2.
Change in Self-Estimated State of Hypoglycemia Awareness From Baseline (2000) to Follow-Up (2012) in 98 Patients With Type 1 Diabetes.
| State of hypoglycemia awareness | ||||
|---|---|---|---|---|
| Baseline | Follow-up |
Total | ||
| Normal | Impaired | Unaware | ||
| Normal | 24 (25) | 14 (14) | 2 (2) | 40 (41) |
| Impaired | 10 (10) | 31 (33) | 5 (5) | 46 (48) |
| Unaware | 1 (1) | 6 (6) | 4 (4) | 11 (11) |
| Total | 35 (36) | 51 (53) | 11 (11) | 97 (100) |
Only patients participating both at baseline and follow-up are included in the table. Values are n (%).
Severe Hypoglycemia and State of Awareness
At baseline, patients classified as having impaired awareness and unawareness of hypoglycemia had a 3.6-fold (95% CI 1.7-7.6; P = .001) and 10.9-fold (95% CI 4.4-27.3; P < .0001) increased relative rate of severe hypoglycemia, respectively, as compared to patients with normal awareness of hypoglycemia (Table 3A). State of awareness at baseline was not associated with rate of severe hypoglycemia at follow-up with relative rates of 0.8 (95% CI 0.4-1.5) and 1.1 (95% CI 0.4-3.0) in patients with impaired awareness and unawareness, respectively, as compared to patients with normal hypoglycemia awareness. However, at follow-up, 1 single patient classified as having normal awareness accounted for a third of the total number of severe hypoglycemic events. If excluding this patient from the analyses, state of awareness at baseline corresponded to a 1.4-fold (95% CI 0.5-3.7) increased rate of severe hypoglycemia for impaired awareness and a 3.1-fold (95% CI 0.9-10.5; P = .07) increased relative rate for unaware patients as compared to patients with normal awareness of hypoglycemia. Patients classified with impaired awareness and unawareness at follow-up had a relative rate of 3.3 (95% CI 0.9-12.2) and 10.3 (95% CI 2.3-44.3; P = .002), respectively, as compared to patients with normal awareness of hypoglycemia.
Table 3.
Rate of Severe Hypoglycemia and Its Relation to Different Well-Known Risk Factors for Severe Hypoglycemia in 98 Patients With Type 1 Diabetes.
| A. The relation according to state of hypoglycemia awareness | ||||
|---|---|---|---|---|
| Severe hypoglycemia | State of hypoglycemia awareness | Relative ratea of severe hypoglycemia (95% confidence interval) | P | |
| Baseline | Baseline | Impaired | 3.6 (1.7-7.6) | .001 |
| Unaware | 10.9 (4.4-27.3) | <.0001 | ||
| Follow-upb | Baseline | Impaired | 1.4 (0.5-3.7) | .52 |
| Unaware | 3.1 (0.9-10.5) | .07 | ||
| Follow-upb | Follow-up | Impaired | 3.3 (0.9-12.2) | .08 |
| Unaware | 10.3 (2.3-44.3) | .002 | ||
| B. The relation according to other risk factors for severe hypoglycemia | ||||
| Severe hypoglycemia | Relative ratec of severe hypoglycemia (95% confidence interval) | P | ||
| C-peptide status | ||||
| Baseline | Baseline | 2.6 (1.5-4.6) | .001 | |
| Baseline | Follow-up | 1.0 (0.4-2.4) | .96 | |
| Follow-up | Follow-up | 1.9 (0.2-16.3) | .55 | |
| Duration of diabetes | ||||
| Baseline | Baseline | 3.2 (1.2-12.5) | .096 | |
| Baseline | Follow-up | 1.5 (0.5-4.4) | .47 | |
| Follow-up | Follow-up | 1.9 (0.5-7.0) | .34 | |
| HbA1c | ||||
| Baseline | Baseline | 0.8 (0.5-1.3) | .40 | |
| Baseline | Follow-up | 1.0 (0.7-1.4) | .81 | |
| Follow-up | Follow-up | 0.9 (0.6-1.5) | .71 | |
The relation is assessed by univariate analyses.
State of hypoglycemia awareness as compared to normal awareness.
1 patient is excluded as described in the text.
C-peptide status: C-peptide-negative patients as compared to patients with detectable C-peptide levels, duration of diabetes: per 10 year’s increment, HbA1c: per 1% increment.
Severe Hypoglycemia and Other Risk Factors
As shown in Table 3B, C-peptide-negative patients had a 2.6-fold (95% CI 1.5-4.6; P = .001) increased rate of severe hypoglycemia at baseline as compared to patients with a detectable level of C-peptide. The level of C-peptide measured at baseline was not associated with the rate of severe hypoglycemia at follow-up (relative rate 1.0; 95% CI 0.4-2.4). Neither at baseline nor at follow-up was HbA1c level, long duration of diabetes, or gender associated with the rate of severe hypoglycemia.
Discussion
In this cohort of patients with type 1 diabetes we were unable to predict severe hypoglycemia from well-known risk factors assessed 12 years earlier despite the fact that the expected associations between the risk factors and severe hypoglycemia were observed with expected and comparable strength both at baseline and at the follow-up. At baseline, there was a correlation between state of awareness and C-peptide status and rate of severe hypoglycemia. This is in line with earlier studies also reporting these factors to be short-term risk factors for severe hypoglycemia.4-6 Especially, state of awareness was a strong predictor for the short-term risk of severe hypoglycemia. Our study confirms that the increased short-term risk is replicable over time as an approximately 10-fold higher risk of severe hypoglycemia for patients with hypoglycemia unawareness was found both at baseline and at follow-up. At follow-up, there was a trend toward a 3.1-fold increased rate of severe hypoglycemia in those with unawareness at baseline. This suggests that although unawareness predicts severe hypoglycemia over a period of at least 12 years, the effect is declining with time. The fact, that only patients with the most severely impaired awareness of hypoglycemia (ie, unawareness assessed by the Hillerød-method, normally comprising 10-15% of the type 1 diabetes population14) at baseline tended to have a persisting increased risk of severe hypoglycemia at follow-up may be explained by a significant number of patients altering state of awareness from baseline to follow-up. Thus state of awareness is not stable over time but can both be regained or deteriorate, for example due to avoidance of hypoglycemia or tight glycemic control, respectively. There was no association between HbA1c or C-peptide levels determined at baseline and rate of severe hypoglycemia at follow-up, that is, HbA1c and C-peptide levels did not predict risk of severe hypoglycemia over a longer period than 1 year in this cohort. This may likewise be explained by the significant alterations in both HbA1c and C-peptide status during the follow-up time, for example, 91% of the patients were C-peptide-negative at follow-up, which may have resulted in loss of power.
During the 12 years of follow-up in our study, we observed a significant decrease in the incidence of severe hypoglycemia from 1.1 episodes per patient-year to 0.4 episodes per patient-year, corresponding to an approximately 50% decrease in the prevalence of severe hypoglycemic events. The incidence of 0.4 episodes of severe hypoglycemia per patient-year is much lower than in previous studies reporting an incidence of severe hypoglycemia of 1-1.5 episodes per patient-year.2,13,15,16
Despite the decreasing rate of severe hypoglycemia, we showed a simultaneous decrease in both HbA1c and detectable C-peptide levels in our cohort. The prevalence of impaired awareness and unawareness of hypoglycemia did not differ from baseline to follow-up and therefore the lower rate of severe hypoglycemia is not due to a regained awareness of hypoglycemia. Moreover, it is not due to a selection bias, as the nonparticipants in the baseline cohort did not differ from the participants according to baseline C-peptide status, duration of diabetes or HbA1c. Only few studies explore the rate of severe hypoglycemia and risk factors over a longer period than 1 year.8,9 These long-term studies find unaltered9 or increased8 rates of severe hypoglycemia in contrast to the declining rate in our study. We speculate that the reduced rate in our cohort is explained by the constant focus on problematic hypoglycemia in our diabetes outpatient clinic during the study period. Moreover, improvements in the diabetes management implemented after ending follow-up in the other long-term studies,8,9 for example, insulin analogs and insulin pumps, may be of influence. Thus, during the 12 years of follow-up in our study, use of insulin analogs has become much more frequent in the management of type 1 diabetes and insulin analogs are shown to reduce the risk of hypoglycemia.17-19 The increase in the use of analogs in our study from 5% at baseline to 58% at the end of follow-up in 2012 may explain some of the decline in rate of severe hypoglycemia in our study as compared to the earlier long-term studies. However, the implementation of insulin analogs and pumps were not limited to patients prone to severe hypoglycemia or with reduced awareness of hypoglycemia at baseline, that is, the intervention during follow-up were balanced. This may emphasize the importance of a constant focus on problematic hypoglycemic episodes and their prevention in outpatient clinics to reduce the rate of severe hypoglycemia. Finally, in 2012, the driver’s license legislation has been tightened up in Denmark and in the European Union, which may have influenced the patients’ willingness to report severe hypoglycemic events at follow-up.20,21
Conclusion
Long-term prediction of severe hypoglycemia in type 1 diabetes is difficult. In this study, we demonstrate that prediction of the long-term risk of severe hypoglycemia is not possible based on well-known risk factors for severe hypoglycemia. However, the strongest risk predictor of severe hypoglycemia, that is, self-estimated unawareness of hypoglycemia, tended to predict an increased long-term rate of severe hypoglycemia, even though the effect weakened over time. Therefore, it is important repeatedly to assess risk factors of severe hypoglycemia in type 1 diabetes, for example, state of hypoglycemia awareness and C-peptide status. By repeating assessment of the well-known risk factors, it may be possible to estimate the present risk of severe hypoglycemia in patients with type 1 diabetes.
Acknowledgments
We thank research nurses P. Banck and T. Larsen for their skillful handling of patients and data.
Footnotes
Abbreviations: CI, confidence interval; RR, relative rate; SH, severe hypoglycemia.
Declaration of Conflicting Interests: The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Louise Færch currently works at Novo Nordisk A/S.
Funding: The author(s) received no financial support for the research, authorship, and/or publication of this article.
References
- 1. Cryer PE. Mechanisms of hypoglycemia-associated autonomic failure in diabetes. N Engl J Med. 2013;369(4):362-372. [DOI] [PubMed] [Google Scholar]
- 2. Kristensen PL, Hansen LS, Jespersen MJ, et al. Insulin analogues and severe hypoglycaemia in type 1 diabetes. Diabetes Res Clin Pract. 2012;96(1):17-23. [DOI] [PubMed] [Google Scholar]
- 3. Cichosz SL, Frystyk J, Tarnow L, Fleischer J. Combining information of autonomic modulation and CGM measurements enables prediction and improves detection of spontaneous hypoglycemic events. J Diabetes Sci Technol. 2015;9(1):132-137. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Hypoglycemia in the Diabetes Control and Complications Trial. The Diabetes Control and Complications Trial research group. Diabetes. 1997;46(2):271-286. [PubMed] [Google Scholar]
- 5. Pedersen-Bjergaard U, Agerholm-Larsen B, Pramming S, Hougaard P, Thorsteinsson B. Activity of angiotensin-converting enzyme and risk of severe hypoglycaemia in type 1 diabetes mellitus. Lancet. 2001;357(9264):1248-1253. [DOI] [PubMed] [Google Scholar]
- 6. UK Hypoglycaemia Study Group. Risk of hypoglycaemia in types 1 and 2 diabetes: effects of treatment modalities and their duration. Diabetologia. 2007;50(6):1140-1147. [DOI] [PubMed] [Google Scholar]
- 7. Pedersen-Bjergaard U, Pramming S, Heller SR, et al. Severe hypoglycaemia in 1076 adult patients with type 1 diabetes: influence of risk markers and selection. Diabetes Metab Res Rev. 2004;20(6):479-486. [DOI] [PubMed] [Google Scholar]
- 8. Bragd J, Adamson U, Lins PE, Wredling R, Oskarsson P. A repeated cross-sectional survey of severe hypoglycaemia in 178 type 1 diabetes mellitus patients performed in 1984 and 1998. Diabet Med. 2003;20(3):216-219. [DOI] [PubMed] [Google Scholar]
- 9. Faerch L, Pedersen-Bjergaard U, Thorsteinsson B. High serum ACE activity predicts severe hypoglycaemia over time in patients with type 1 diabetes. Scand J Clin Lab Invest. 2011;71(7):620-624. [DOI] [PubMed] [Google Scholar]
- 10. Cranston I, Lomas J, Maran A, Macdonald I, Amiel SA. Restoration of hypoglycaemia awareness in patients with long-duration insulin-dependent diabetes. Lancet. 1994;344(8918):283-287. [DOI] [PubMed] [Google Scholar]
- 11. Fanelli CG, Epifano L, Rambotti AM, et al. Meticulous prevention of hypoglycemia normalizes the glycemic thresholds and magnitude of most of neuroendocrine responses to, symptoms of, and cognitive function during hypoglycemia in intensively treated patients with short-term IDDM. Diabetes. 1993;42(11):1683-1689. [DOI] [PubMed] [Google Scholar]
- 12. Seaquist ER, Anderson J, Childs B, 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] [PMC free article] [PubMed] [Google Scholar]
- 13. Pedersen-Bjergaard U, Pramming S, Thorsteinsson B. Recall of severe hypoglycaemia and self-estimated state of awareness in type 1 diabetes. Diabetes Metab Res Rev. 2003;19(3):232-240. [DOI] [PubMed] [Google Scholar]
- 14. Hoi-Hansen T, Pedersen-Bjergaard U, Thorsteinsson B. Classification of hypoglycemia awareness in people with type 1 diabetes in clinical practice. J Diabetes Complications. 2010;24(6):392-397. [DOI] [PubMed] [Google Scholar]
- 15. Donnelly LA, Morris AD, Frier BM, et al. Frequency and predictors of hypoglycaemia in type 1 and insulin-treated type 2 diabetes: a population-based study. Diabet Med. 2005;22(6):749-755. [DOI] [PubMed] [Google Scholar]
- 16. ter Braak EW, Appelman AM, van de Laak M, Stolk RP, van Haeften TW, Erkelens DW. Clinical characteristics of type 1 diabetic patients with and without severe hypoglycemia. Diabetes Care. 2000;23(10):1467-1471. [DOI] [PubMed] [Google Scholar]
- 17. Pedersen-Bjergaard U, Kristensen PL, Beck-Nielsen H, et al. Effect of insulin analogues on risk of severe hypoglycaemia in patients with type 1 diabetes prone to recurrent severe hypoglycaemia (HypoAna trial): a prospective, randomised, open-label, blinded-endpoint crossover trial. Lancet Diabetes Endocrinol. 2014;2:553-561. [DOI] [PubMed] [Google Scholar]
- 18. Siebenhofer A, Plank J, Berghold A, et al. Short acting insulin analogues versus regular human insulin in patients with diabetes mellitus. Cochrane Database Syst Rev. 2006;2(2):CD003287. [DOI] [PubMed] [Google Scholar]
- 19. Monami M, Marchionni N, Mannucci E. Long-acting insulin analogues vs. NPH human insulin in type 1 diabetes. A meta-analysis. Diabetes Obes Metab. 2009;11(4):372-378. [DOI] [PubMed] [Google Scholar]
- 20. Pedersen-Bjergaard U, Faerch L, Allingbjerg ML, Agesen R, Thorsteinsson B. 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] [PubMed] [Google Scholar]
- 21. Domgaard M, Bagger M, Rhee NA, Burton CM, Thorsteinsson B. Individual and societal consequences of hypoglycemia: a cross-sectional survey. Postgrad Med. 2015;127(5):438-445. [DOI] [PubMed] [Google Scholar]