Skip to main content

The updated Pedersen-Bjergaard method for assessment of awareness of hypoglycaemia in type 1 diabetes

Ulrik Pedersen-Bjergaard1, 2, Louise Færch1 & Birger Thorsteinsson1, 2

25. aug. 2022
14 min.



Loss of awareness of hypoglycaemia is by far the most important known risk factor for severe hypoglycaemia in type 1 diabetes [1]. Identifying persons with loss of awareness is essential to instigating measures to reduce the consequences or restore awareness. Several clinical tools have been developed for assessment of awareness and in type 1 diabetes, three tools have been validated and used more widely in research and clinical practice [2-4].

The Gold score from 1994 was the first published method classifying people with type 1 diabetes by their answer to a single question (“Do you know when your hypos are commencing?”) as having a normal or an impaired awareness [2]. Patients with impaired awareness have an almost six-fold higher risk of severe hypoglycaemia than the group with normal awareness.

The Clarke method, published in 1995, uses an eight-item questionnaire (including previous experience of severe hypoglycaemia) that classifies patients as having a normal or a reduced awareness [3]. A gray-zone score leaves patients as unclassifiable, accounting for 10-20% in cohort studies. Patients with a reduced awareness have a three-fold higher risk of severe hypoglycaemia than those with a normal awareness.

The third method developed by our group (and referred to as the Pedersen-Bjergaard or the Hillerød method) was published in 2003 [4]. The rationale for developing an additional tool was the general clinical experience that awareness is not an all-or-none phenomenon, i.e., many patients seem to develop partial reduction of awareness, even if this does not result in an increased risk of severe hypoglycaemia as compared to the general population with type 1 diabetes. Our method is based on a single question “Can you feel when your blood sugar is low?” with the possible answers “always”, “usually”, “occasionally” or “never” (and “don’t know”). The method classifies people into three groups: normal awareness (“always”), impaired awareness (“usually”) and unawareness (“occasionally” or “never”). In our original paper, those with impaired awareness prospectively reported a more than five-fold and those with unawareness an almost ten-fold higher rate of severe hypoglycaemia than those in the group with normal awareness [4]. Our method has subsequently been used in many studies on people with type 1 diabetes in various countries and languages [5-8], basically confirming this result; and the Pedersen-Bjergaard method is recommended in the Danish National Guidelines on Treatment of Type 1 Diabetes [9].

During the years, confusion has surrounded the interpretation of our awareness classes [10, 11]. The reasons for this are that our classification operates with three classes in contrast to the others which are bimodal, and that we use of the term “impaired awareness” to describe the intermediate group. Thus, comparisons have, in some instances, been made between our pooled impaired and unawareness groups and the impaired/reduced groups of the Gold/Clarke methods [10]. The relevant comparison, however, is with our unaware class [8]. We have, therefore, updated the Pedersen-Bjergaard classification by renaming our intermediate group to “intermediate” instead of “impaired”.

Here, we present a comparison of our updated classification with the Gold and the Clarke classifications in a type 1 diabetic cohort in terms of prevalence of loss of awareness and rates of severe hypoglycaemia.


A total of 416 outpatients at Nordsjællands Hospital Hillerød, Denmark, were invited to participate in the study. Following written informed consent, 325 adult people with type 1 diabetes for more than one year were recruited. Type 1 diabetes was defined by insulin treatment from the time of diagnosis and a random C-peptide level below 300 pmol/l at a venous blood glucose concentration ≤ 12 mmol/l or a C-peptide level below 600 pmol/l at a venous blood glucose concentration >12 mmol/l. Pregnant women and individuals with significant comorbidities were excluded from participation. Data on history of diabetes, insulin therapy, concomitant medication and late diabetic complications were extracted from the medical records (Table 1). Baseline blood samples were drawn in the non-fasted condition for measurement of C-peptide, HbA1c and routine biochemistry. Patients completed a questionnaire based on previously used and validated questions. Hypoglycaemia awareness was self-estimated using the three methods: Gold, Clarke and Pedersen-Bjergaard scores [2, 3, 4]. The Gold and Clarke methods were applied as described in the original publications [2, 3]. The Pedersen-Bjergaard classification was updated as described in the introduction. The rate of severe hypoglycaemia (defined by need for third party assistance to actively administer carbohydrate or glucagon) was based on one-year recall [12]. The study was approved by the Committee on Biomedical Ethics of the Capital Region of Denmark (H-4-2011-043) and the Danish Data Protection Agency (2012-58-0004).

Table 1

Laboratory tests

HbA1c was measured by a Tosoh G8 HPLC Analyzer, Tosoh Bioscience Inc., San Francisco, CA, and C-peptide concentration by AutoDELFIA, Wallac Oy, Turku, Finland. Patients were classified as C-peptide-negative when the C-peptide concentration was below the detection limit of 10 pmol/l.


Standard descriptive and comparative statistics were used. The term “loss of awareness” was used for the comparison between the unawareness group according to the Pedersen-Bjergaard classification, the impaired awareness group according to the Gold classification and the reduced awareness group according to the Clarke classification.

Trial registration: not relevant.


Prevalence of awareness classes

A total of 43%, 76% and 74% of patients were classified with normal awareness according to the Pedersen-Bjergaard method, the Gold method, and the Clarke method, respectively (Table 2), whereas 14%, 24% and 17% were classified as having loss of awareness with our method, the Gold method, and the Clarke method, respectively. A total of 43% were classified with intermediate awareness with our method and 9% were unclassifiable according to the Clarke method.

Table 2

Agreement between methods

35% of patients classified with loss of awareness by at least one method had loss of awareness according to all three methods, and a further 24% agreed between two of the methods, leaving 41% of the cases to be classified with loss of awareness only by one method (Figure 1). Good agreement was recorded between our and the other methods (Table 3 and Figure 1). Thus, 64% of patients classified with unawareness according to the Pedersen-Bjergaard method had loss of awareness according to both the Gold and the Clarke methods (Figure 1). In comparison, 38% and 50% of those classified with impaired/reduced awareness with the Gold and the Clarke methods, respectively, had loss of awareness by all methods.

Figure 1
Table 3

Only one patient classified as unaware by the Pedersen-Bjergaard method was not impaired/reduced by at least one of the other methods (Figure 1). For comparison, 29 patients classified with impaired awareness by the Gold method were not classified as unaware/reduced according to our method or the Clarke method, whereas 13 patients classified with reduced awareness by the Clarke method were not classified as unaware/impaired according to our method or the Gold method.

Risk of severe hypoglycaemia

Severe hypoglycaemia in the preceding year was reported by 23.7% of the patients with a mean rate of 0.62 episodes per patient year (Table 1). All three classifications identified groups of patients with a clinically significantly increased prevalence and rate of hypoglycaemia as compared with those with normal awareness (Table 2). No significant differences were found between the methods in prevalence of severe hypoglycaemia in the groups with loss of awareness. The relative rates of severe hypoglycaemia compared with normal awareness was 1.9 and 5.7 in the groups with intermediate and unawareness by the Pedersen-Bjergaard method, respectively. A relative rate of 2.9 was found in patients with impaired awareness according to the Gold method and of 4.8 in patients with reduced awareness according to the Clarke method, respectively, when compared with normal awareness.

The prevalence of severe hypoglycaemia was lower in the group with normal awareness according to the Pedersen-Bjergaard method than according to the Gold method (χ2 = 5.0; p = 0.024), but not according to the Clarke score (χ2 = 1.9; p = 0.17).


This study shows that the updated version of the Pedersen-Bjergaard classification of awareness of hypoglycaemia works in terms of agreement with the Clarke and Gold classifications in identification of patients with loss of awareness and of patients at increased risk of severe hypoglycaemia. Both regarding the fraction of patients identified with loss of awareness and risk of severe hypoglycaemia, the Pedersen-Bjergaard classification was closer to the Clarke classification than to the Gold method. The Gold method identified more patients with loss of awareness, and the rate of severe hypoglycaemia in this group was numerically lower than with the Pedersen-Bjergaard and the Clarke classifications, suggesting a lower specificity of the Gold classification.

The important difference between the Pedersen-Bjergaard method and the Gold and Clarke method remains the three classes, reflecting that loss of awareness probably develops gradually and may become clinically important already at an earlier disease stage, just like many other diabetic complications, thereby enabling further differentiation in risk of severe hypoglycaemia. Thus, previous studies using our classification in unselected type 1 diabetic cohorts have consistently shown that 40-50% of patients are classified as having a normal awareness with a very low risk (0.1-0.4 episodes of severe hypoglycaemia per patient-year), 40-50% as having an intermediate (previously termed impaired) awareness with an intermediate (but already two- to eight-fold increased) risk and 10-15% as having unawareness with a high (seven- to 21-fold increased) risk of severe hypoglycaemia [5-7, 9]. Furthermore, the rate of severe hypoglycaemia in the intermediate group has consistently fallen close to the mean rate in the cohorts, which was also the case in the present study. The intermediate group in the Pedersen-Bjergaard classification also explains why the risk of hypoglycaemia in the group with normal awareness tended to be lower than in the other classifications.

The three different awareness classes in the Pedersen-Bjergaard classification provide benefit both in daily clinical practice and in research. Clinically, patients at an increased risk of severe hypoglycaemia who are in need of special attention can be identified and offered educational interventions [13, 14] aimed at restoring awareness or provision of artificial awareness in terms of continuous glucose monitoring (CGM) systems with alarm features [15]. Ultimately, hybrid closed-loop systems that can trigger an alarm and automatically reduce insulin infusion when hypoglycaemia is predicted may help these people [16]. Also, patients at a particularly low risk of severe hypoglycaemia who may be candidates for tighter glycaemic control if needed to avoid progression of microvascular complications can be identified. In a research context, studies exploring, e.g., the pathophysiology of loss of awareness may gain considerable additional power using a control group with the opposite phenotype – best possible awareness.

The Gold and the Pedersen-Bjergaard classifications are both based on the answer to a single question and as such they are easy to implement in clinical practice. This contrasts with the eight-item Clarke score. However, the Clarke score does not seem to provide more information than the other methods. Even the fact that one of the items in the Clarke score is occurrence of severe hypoglycaemia in the preceding year did not provide an advantage for the Clarke score in terms of association with severe hypoglycaemia in the preceding year.

All three methods have been validated in hyperinsulinaemic hypoglycaemic clamp studies showing that patients with loss of awareness have reduced sympathoadrenal counterregulatory responses and blunted warning symptoms compared with patients with normal awareness [17]. Patients with loss of awareness according to all three methods less frequently reported autonomic symptoms as their main warning symptoms [8] and a lower glucose level as the threshold for perception of warning symptoms [8]. Furthermore, studies using CGM have shown that patients with loss of awareness are more prone to asymptomatic hypoglycaemic events than patients who declare normal awareness [18].

The strengths of this study are the use of validated questionnaires and a high inclusion rate. The major weakness is the lack of an objective gold standard measurement of awareness. Except for the very cumbersome recording of warning symptoms generated during a hypoglycaemic clamp, such a method has still to be developed. Furthermore, the real value of an awareness classification is prediction of future risk of severe hypoglycaemia, which we cannot identify from this retrospective study. However, the general impression from the literature is that retrospective associations with and prospective prediction of severe hypoglycaemia are rather consistent between the methods.


The revised version of the Pedersen-Bjergaard classification for hypoglycaemia awareness and the Gold and Clarke classifications – all based on a self-reported perception of the ability to feel hypoglycaemic warning symptoms – identified people with loss of awareness with reasonably high degree of concordance. The Pedersen-Bjergaard method, in contrast to the others, identified groups with a normal and an intermediate awareness with clinically significantly different risks of severe hypoglycaemia.

Correspondence Ulrik Pedersen-Bjergaard. E-mail:
Accepted 22 June 2022
Conflicts of interest Potential conflicts of interest have been declared. Disclosure forms provided by the authors are available with the article at
References can be found with the article at
Cite this as Dan Med J 2022;69(9):A01220041



  1. Frier BM. Hypoglycaemia in diabetes mellitus: epidemiology and clinical implications. Nat Rev Endocrinol. 2014;10(12):711-22.

  2. Gold AE, MacLeod KM, Frier BM. Frequency of severe hypoglycemia in patients with type I diabetes with impaired awareness of hypoglycemia. Diab Care. 1994;17(7):697-703.

  3. Clarke WL, Cox DJ, Gonder-Frederick LA et al. Reduced awareness of hypoglycemia in adults with IDDM: a prospective study of hypoglycemic frequency and associated symptoms. Diab Care. 1995;18(4):517-22.

  4. Pedersen-Bjergaard U, Pramming S, Thorsteinsson B. Recall of severe hypoglycaemia and self-estimated state of awareness in type 1 diabetes. Diab Metab Res Rev. 2003;19(3):232-40.

  5. Pedersen-Bjergaard U, Pramming S, Heller SR et al. Severe hypoglycaemia in 1076 patients with type 1 diabetes: influence of risk markers and selection. Diab Metab Res Rev. 2004;20(6):479-86.

  6. Kristensen PL, Hansen LS, Jespersen MJ et al. Insulin analogues and severe hypoglycaemia in type 1 diabetes. Diab Res Clin Pract. 2012;96(1):17-23.

  7. Östenson CG, Geelhoed-Duijvestijn P, Lahtela J et al. Self-reported non-severe hypoglycaemic events in Europe. Diab Med. 2014;31(1):92-101.

  8. Høi-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):392-7.

  9. Type 1 diabetes. (1 May 2022).

  10. Geddes J, Wright RJ, Zammitt NN et al. An evaluation of methods of assessing impaired awareness of hypoglycaemia in type 1 diabetes. Diab Care. 2007;30(7):1868-70.

  11. Agesen RM, Kristensen PL, Beck-Nielsen H et al. Effect of insulin analogs on frequency of non-severe hypoglycemia in patients with type 1 diabetes prone to severe hypoglycemia: much higher rates detected by continuous glucose monitoring than by self-monitoring of blood glucose – the HypoAna trial. Diab Technol Ther. 2018;20(3):247-56.

  12. Rokamp KZ, Olsen NV, Færch L et al. Impact of genetic polymorphism in the β₂-receptor gene on the risk of severe hypoglycemia in patients with type 1 diabetes. J Clin Endocrinol Metab. 2018;103(8):2901-8.

  13. Cox DJ, Gonder-Frederick L, Polonsky W et al. Blood glucose awareness training (BGAT-2): long-term benefits. Diabetes Care. 2001;24(4):637-42.

  14. Little SA, Leelarathna L, Walkinshaw E et al. Recovery of hypoglycemia awareness in long-standing type 1 diabetes: a multicenter 2 × 2 factorial randomized controlled trial comparing insulin pump with multiple daily injections and continuous with conventional glucose self-monitoring (HypoCOMPaSS). Diab Care. 2014;37(8):2114-22.

  15. Rickels MR, Peleckis AJ, Dalton-Bakes C et al. Continuous glucose monitoring for hypoglycemia avoidance and glucose counterregulation in long-standing type 1 diabetes. J Clin Endocrinol Metab. 2018;103(1):105-114.

  16. Malone SK, Peleckis AJ, Grunin L et al. Characterizing glycemic control and sleep in adults with long-standing type 1 diabetes and hypoglycemia unawareness initiating hybrid closed loop insulin delivery. J Diab Res. 2021;2021:6611064.

  17. Sejling A-S, Kjaer TW, Pedersen-Bjergaard U et al. Hypoglycemia-associated changes in the electroencephalogram in patients with type 1 diabetes and normal hypoglycemia awareness or unawareness. Diabetes. 2015;64(5):1760-9.

  18. Henriksen MM, Andersen HU, Thorsteinsson B, Pedersen-Bjergaard U. Hypoglycemic exposure and risk of asymptomatic hypoglycemia in type 1 diabetes assessed by continuous glucose monitoring. J Clin Endocrinol Metab. 2018;103(6):2329-2335.