Randomised trials, most notably the DCCT, have provided substantial data on the epidemiology of hypoglycaemia in adults with type 1 diabetes and, in particular, on the impact of intensive insulin therapy.
The Diabetes Control and Complications Trial (DCCT)
The DDCT was a landmark study and provided diabetes specialists with the long-awaited proof that strict glycaemic control limited the incidence and severity of microvascular complications in people with type 1 diabetes. A total of 1441 patients with type 1 diabetes
Table 3.3 Risk factors for severe hypoglycaemia in adults with type 1 diabetes
Low Hb AI
C-pep tide negative
Muhlhauser et al., 1985 Muhlhauser et al, 1987 MacLeod et al., 1993 Gold et al., 1994 EURODIAB, 1994 MacLeod et al., 1994 SDIS, 1994 Clarke et al., 1995 Pampanelli, et al., 1996 Bolt et al., 1997 DCCT, 1997 Gold et al., 1997 Muhlhauser et al, 1998 Pedersen-Bjergaard et al., 2001 ter Braak et al., 2000 Leese et al., 2003 Pedersen-Bjergaard et al., 2003 Pedersen-Bjergaard et al., 2004 Leckie et al., 2005 UK Hypoglycaemia Study Group, 2007
This table examines the risk factors for severe hypoglycaemia that have been most commonly examined in adults with type 1 diabetes. Studies examining predominantly mild hypoglycaemia were not included. + = positive association between risk factor and severe hypoglycaemia; = no association between risk factor and severe hypoglycaemia. * = only a risk factor if awareness of hypoglycaemia not included. M = severe hypoglycaemia more common in men; F = severe hypoglycaemia more common in women.
were randomly assigned either to intensive insulin therapy (based on multiple injection insulin regimens or continuous subcutaneous insulin infusion therapy) or to conventional insulin therapy (one or two insulin injections daily) (The Diabetes Control and Complications Trial Research Group, 1993). In the conventional group, patients did not generally perform home blood glucose monitoring, clinical reviews were undertaken every three months and patients were not informed about their HbA1c result, unless it was in excess of 13%. By contrast, in the intensive therapy group, subjects performed frequent home blood glucose monitoring, had monthly visits with the study team and also had frequent telephone contact to achieve as strict glycaemic control as possible. Over a mean follow-up of 6.5 years, the average HbA1c concentration in the intensive group was approximately 7.0% and in the conventional group approximately 8.8% (The Diabetes Control and Complications Trial Research Group, 1993).
The subjects recruited to this study were not typical of the wider population of people with type 1 diabetes. The participants had greater motivation and individuals were not permitted to take part if, in the preceding two years, they had experienced more than one episode of severe neurological impairment without warning symptoms of hypoglycaemia, or more than two episodes of seizure or coma, regardless of attributed cause. Moreover, during the study itself, the occurrence of an episode of severe hypoglycaemia in an individual patient prompted a review of conventional risk factors and, in instances where probable causes were identified, corrective actions such as re-educating the patient were undertaken (The Diabetes Control and Complications Trial Research Group, 1997).
Despite all these factors, 3788 episodes of severe hypoglycaemia occurred in the 1441 patients over the course of the study and, of these, 1027 were associated with coma and/or seizure (The Diabetes Control and Complications Trial Research Group, 1997). The rate of severe hypoglycaemia in the intensively-treated patients was 0.61 per patient per year, while that in the conventionally-treated group was 0.19 per patient per year, i.e., a three-fold difference. Over the mean of 6.5 years of follow-up, 65% of the intensive group experienced at least one episode of severe hypoglycaemia, compared with 35% of the individuals in the conventional group.
The Stockholm Diabetes Intervention Study (SDIS)
The SDIS was a much smaller study than the DCCT, but had similar aims and objectives. A group of 102 patients with type 1 diabetes were recruited and 89 remained after 7.5 years of follow-up (Reichard and Pihl, 1994). The mean HbA1c was 7.1% in the intensive group and 8.5% in the conventional group. Severe hypoglycaemia, defined as episodes requiring third party assistance, occurred in 80% of subjects in the intensive group and 58% of those in the conventional group over the follow-up period. Overall, the rate of severe hypoglycaemia was 1.1 per patient per year in the intensive group and 0.4 per patient per year in the intensive group. Other risk factors for severe hypoglycaemia were not reported.
The DCCT was a multicentre study; 27 out of the 29 centres reported that intensive therapy was associated with an increased risk of severe hypoglycaemia, but in two centres no increased risk was observed (The Diabetes Control and Complications Trial Research Group, 1997). This led some investigators to claim that with appropriate education and training intensive insulin therapy need not necessarily be associated with an increased risk of hypo-glycaemia (Plank et al., 2004). In the Bucharest-Dusseldorf Study, 300 individuals with type 1 diabetes were randomised to: (a) conventional therapy for one year and then to one year of intensive therapy; (b) two years of intensive therapy; or (c) a four-day in-patient group teaching programme with conventional insulin therapy for one year (Muhlhauser et al., 1987). Glycated haemoglobin (HbA1), remained unchanged at around 12-13% during conventional therapy, but fell to ~9.5% during intensive therapy. In the first year of the study, severe hypoglycaemia occurred in 6% of the intensively-treated patients and 12% of the conventionally-treated patients, and in year two the proportion of patients experiencing severe hypoglycaemia in both intensive groups fell to 3-4%.
The Dusseldorf team reported observational data on the impact of intensive therapy on the frequency of severe hypoglycaemia (Bott et al., 1997). A total of 636 people with type 1 diabetes who had participated in a structured five-day in-patient treatment and teaching programme for intensification of insulin therapy in one of ten different hospitals in Germany, were re-examined at intervals over six years. The mean HbA1c fell from 8.3% to 7.6%. Severe hypoglycaemia, defined as episodes treated with intramuscular glucagon or intravenous glucose, decreased from 0.28 episodes per patient per year in the year preceding the programme to 0.17 episodes per patient per year afterwards (Bott et al., 1997). The variation in incidence of severe hypoglycaemia between different centres ranged from 0.05-0.27 episodes per patient per year.
Pampanelli and colleagues (1996) from Perugia reported retrospective data on 112 individuals who had been commenced on intensive insulin therapy (preprandial soluble insulin and bedtime isophane insulin) at diagnosis of diabetes and who were attending clinic at least four times per year. Mean HbAlc was 7.2% and mean duration of diabetes was 7.8 years. Frequency of mild hypoglycaemia was estimated by a review of the patients' blood glucose monitoring diaries in the nine months prior to study, and the overall rate was 35.6 episodes per patient per year. Severe hypoglycaemia, defined as episodes requiring third party assistance, was assessed retrospectively over the duration of diabetes and was recalled by only six patients (representing an overall rate of 0.001 episodes per patient per year).
Thus, in the DCCT and the SDIS, intensive insulin therapy was associated with a nearly three-fold increase in the risk of severe hypoglycaemia. By contrast, in other studies, the incidence of severe hypoglycaemia fell when intensive therapy was coupled with a detailed education programme. The teams from Dusseldorf and Perugia would argue that the high quality of their education programmes resulted in the low frequencies of observed severe hypoglycaemia. Although this may be the case, the importance of patient selection in all of these studies cannot be over-emphasised. In their study of 1076 adults with type 1 diabetes from the United Kingdom and Denmark, Pedersen-Bjergaard et al. (2004) examined a subset of 230 individuals whose clinical characteristics were similar to patients enrolled in the DCCT. This sub-group accounted for only 5.4% of all reported episodes of severe hypoglycaemia, with an overall rate of 0.35 episodes per patient per year, i.e., approximately one quarter that of the entire group. Risk factors for severe hypoglycaemia in this group (impaired awareness and retinopathy) were different from that of the study population as a whole (see Table 3.3). Thus, the DCCT patients were not representative of people with type 1 diabetes, whereas the education programme undertaken in Dusseldorf is not something that is widely replicated in mainstream diabetes practice. Risk factors for hypoglycaemia may differ according to the specific characteristics of individuals being studied.
Thus, in conclusion, although most specialists would accept that severe hypoglycaemia is more common in patients receiving intensive insulin therapy, it is not inevitable and better patient education may actually reduce the incidence.
Strict glycaemic control, as evidenced by glycated haemoglobin concentrations that approach the upper end of the non-diabetic range, is closely linked to intensive insulin therapy. Since the DCCT results were published, glycated haemoglobin concentrations at or around this level have become the usual target for many people with type 1 diabetes. In the DCCT, there was a quadratic relationship between HbA1c and risk of severe hypoglycaemia (Figure 3.5), with the risk of hypoglycaemia increasing as HbA1c decreased (The Diabetes Control and Complications Trial Research Group, 1997). However, the attained HbA1c did not account for all the difference in risk of severe hypoglycaemia between the two arms of the study, as subjects in the intensive group still had an excess risk of severe hypoglycaemia after
HbA1c (%) During Study
Figure 3.5 Risk of severe hypoglycaemia as a function of monthly updated HbA1c in the Diabetes Control and Complications Trial. The circles represent data from the intensive group and asterisks data from the conventional group. The bold solid and bold dashed lines represent the regression plots for each group, and the non-bold dashed lines on either side show the upper and lower 95% confidence bands; DCCT (1997). Copyright © 1997 American Diabetes Association. Reprinted with permission from The American Diabetes Association
HbA1c (%) During Study
Figure 3.5 Risk of severe hypoglycaemia as a function of monthly updated HbA1c in the Diabetes Control and Complications Trial. The circles represent data from the intensive group and asterisks data from the conventional group. The bold solid and bold dashed lines represent the regression plots for each group, and the non-bold dashed lines on either side show the upper and lower 95% confidence bands; DCCT (1997). Copyright © 1997 American Diabetes Association. Reprinted with permission from The American Diabetes Association statistical adjustment for HbA1c concentration. Indeed, in the intensive group, only about 5% of the variation in frequency of severe hypoglycaemia could be explained by the glycated haemoglobin concentration (The Diabetes Control and Complications Trial Research Group, 1997).
Other studies have reported variable relationships between glycaemic control and frequency of severe hypoglycaemia. In the study by Bott et al., (1997) a lower mean HbA1c was associated with severe hypoglycaemia, but there was no linear or quadratic relationship between HbA1c and severe hypoglycaemia. In the EURODIAB IDDM Complications Study, severe hypoglycaemia (defined as episodes requiring third party assistance) occurred in 32% of individuals over 12 months. A clear relationship to glycated haemoglobin was evident, in that 40% of individuals with an HbA1c < 5.4% were affected compared with 24% of individuals with a HbA1c > 7.8% (The EURODIAB IDDM Complications Study Group, 1994). In the study of workplace hypoglycaemia, recurrent severe hypoglycaemia was associated with strict glycaemic control, but no link was found in individuals who had experienced only one episode (Leckie et al., 2005). In several other studies, no relationship was observed between severe hypoglycaemia and glycated haemoglobin (Muhlhauser et al., 1985; Muhlhauser et al., 1987; MacLeod et al., 1993; Gold et al., 1997; Muhlhauser et al., 1998; ter Braak et al., 2000; Pedersen-Bjergaard et al., 2001; Leese et al., 2003; Pedersen-Bjergaard et al., 2003a; Pedersen-Bjergaard et al., 2004) after adjustment for other risk factors.
Glycated haemoglobin does not, of course, provide the entire picture about an individual's glycaemic control and although HbA1c may not predict risk of hypoglycaemia, low mean home blood glucose concentrations and excessive variability in blood glucose can identify individuals more prone to hypoglycaemia (Cox et al., 1994; Janssen et al., 2000).
Thus, a straightforward relationship does not exist between severe hypoglycaemia and glycaemic control. For an episode of mild hypoglycaemia to progress to one that causes significant neuroglycopenia and impairs consciousness, other factors must operate, which negate the normal symptomatic and hormonal responses to hypoglycaemia.
Within each treatment group of the DCCT, the number of previous episodes of severe hypoglycaemia was the strongest predictor of risk of future episodes (The Diabetes Control and Complications Trial Research Group, 1997). Moreover, approximately 30% of patients in each group experienced a second episode of severe hypoglycaemia within four months following an initial episode. The importance of a previous history of severe hypoglycaemia in predicting future risk has been replicated in several other studies (MacLeod et al., 1993; Bott et al., 1997; Gold et al., 1997; Muhlhauser et al., 1998). Moreover, as demonstrated in Table 3.3, many studies have also linked increased duration of diabetes with an increased risk of severe hypoglycaemia.
Cryer has suggested that the integrity of the glucose counterregulatory system may be a pivotal factor in determining whether the relative or absolute hyperinsulinism that frequently occurs in insulin-treated diabetes ultimately results in the development of hypoglycaemia (Cryer, 1994; Cryer et al., 2003). Three acquired hypoglycaemia syndromes are associated with an increased risk of severe hypoglycaemia in people with type 1 diabetes. These are considered in greater detail in Chapters 6 and 7.
Hypoglycaemia-induced secretion of glucagon declines in most patients within five years of developing type 1 diabetes (Gerich et al., 1973; Bolli et al., 1983). A defective epinephrine response to hypoglycaemia may develop some years later (Bolli et al., 1983; Hirsch and Shamoon, 1987; Dagogo-Jack et al., 1993). As with the glucagon response, the impaired epinephrine response is hypoglycaemia-specific but, in contrast to glucagon, it exhibits a threshold effect - i.e., an epinephrine response can still be elicited by hypoglycaemia, but only at a lower blood glucose concentration (Dagogo-Jack et al., 1993). If hypoglycaemia develops in patients who have this combined counterregulatory hormonal deficiency, glucose recovery may be severely compromised (see Chapter 6). Subjecting such patients to intensified insulin therapy increased the risk of severe hypoglycaemia by 25 times, compared with subjects who had an intact epinephrine response (White et al., 1983).
In many patients with insulin-treated diabetes, the hypoglycaemia symptom profile alters with time, resulting in impaired perception of the onset of hypoglycaemia (see Chapter 7). Commonly, autonomic warning symptoms are diminished and neuroglycopenic symptoms predominate. Around 25% of people with type 1 diabetes develop impaired awareness of hypoglycaemia and the prevalence of this problem increases with the duration of insulin treatment (Hepburn et al., 1990; Gerich et al., 1991; Pramming et al., 1991). Prospective studies have demonstrated that the frequency of severe hypoglycaemia is increased up to six-fold in patients with impaired awareness compared to those with normal hypoglycaemia awareness (Gold et al., 1994; Clarke et al., 1995).
The above acquired hypoglycaemia syndromes tend to segregate together clinically. Patients with glycated haemoglobin concentrations close to the non-diabetic range are at greater risk of developing impaired awareness (Boyle et al., 1995; Kinsley et al., 1995; Pampanelli et al., 1996), while the glycaemic thresholds for the onset of symptoms and responses are altered in patients with impaired awareness (Grimaldi et al., 1990; Hepburn et al., 1991; Mokan et al., 1994; Bacatselos et al., 1995). Cryer has suggested that these acquired abnormalities represent a form of central 'Hypoglycaemia-Associated Autonomic Failure' (HAAF) in type 1 diabetes, speculating that recurrent severe hypoglycaemia is the primary cause (Cryer, 1992; Cryer et al., 2003). If hypoglycaemia is the precipitant, then it is possible to see how a vicious cycle may become established with the development of the acquired hypoglycaemia syndromes promoting further episodes of severe hypoglycaemia.
There is a well-known adage that 'hypoglycaemia begets hypoglycaemia'. People who have experienced one episode of severe hypoglycaemia are much more likely to develop further episodes, and the greatest risk occurs in the weeks and months after the index event. Severe hypoglycaemia becomes a more common problem in people with long-standing type 1
diabetes (UK Hypoglycaemia Group, 2007). This is the legacy of the burden of hypoglycaemia that such individuals have experienced over many years with diabetes. The impaired symptomatic and counterregulatory responses to hypoglycaemia dramatically increase the likelihood that an episode of mild hypoglycaemia will progress to a more severe event.
Most of the precipitants and risk factors for hypoglycaemia have been known about for many years. In 2001, Pedersen-Bjergaard et al. (2001) reported a novel risk factor for severe hypoglycaemia in adults with type 1 diabetes and raised the notion that some individuals may have an inherent genetic susceptibility to hypoglycaemia. The Danish investigators noted the similarity between endurance exercise and hypoglycaemia in that both are states of limited metabolic fuel availability. Previous studies had linked exercise performance to a particular polymorphism of the angiotensin-converting enzyme (ACE) gene. Specifically, the insertion (I) allele, which resulted in low serum ACE activity, was associated with superior performance capacity compared with the deletion (D) allele. In an initial retrospective survey of 207 adults with type 1 diabetes, patients with the DD genotype had a 3.2-fold increased risk of severe hypoglycaemia in the preceding two years, compared with individuals with the II genotype. There was also a significant relationship between serum ACE activity, with a 1.4 increment in risk of severe hypoglycaemia for every 10U/l rise in serum ACE concentration (Figure 3.6). The serum ACE activity was directly linked to ACE genotype, and it remained a significant risk factor even after adjustment for conventional risk factors. Moreover, serum ACE activity was a stronger risk factor for severe hypoglycaemia in C-peptide negative individuals with impaired awareness, than in other groups (relative risk 1.7 per 10U/l; Figure 3.7). No significant relationship was observed between serum ACE activity or genotype and frequency of mild hypoglycaemia.
m O 20 40 60 80 100 120 Serum ACE activity (U/L)
Figure 3.6 Risk of severe hypoglycaemia according to serum ACE activity in 207 patients with type 1 diabetes, untreated with ACE inhibitors or angiotensin-2 receptor antagonists. Broken lines represent 95% confidence intervals. Reprinted from The Lancet, Pedersen-Bjergaard et al. (2001) with permission from Elsevier m 1 1 1 1 1 1 1
m O 20 40 60 80 100 120 Serum ACE activity (U/L)
Figure 3.6 Risk of severe hypoglycaemia according to serum ACE activity in 207 patients with type 1 diabetes, untreated with ACE inhibitors or angiotensin-2 receptor antagonists. Broken lines represent 95% confidence intervals. Reprinted from The Lancet, Pedersen-Bjergaard et al. (2001) with permission from Elsevier
C-peptide-negative, impaired awareness (n=47) C-peptide-positive, impaired awareness (n=63) C-peptide-negative, normal awareness (n=32) C-peptide-positive, normal awareness (n=60)
C-peptide-negative, impaired awareness (n=47) C-peptide-positive, impaired awareness (n=63) C-peptide-negative, normal awareness (n=32) C-peptide-positive, normal awareness (n=60)
0 20 40 6 0 80 100 120 Serum ACE activity (U/L)
Figure 3.7 Association between severe hypoglycaemia and serum ACE activity according to C-peptide status and self-estimated awareness of hypoglycaemia. Reprinted from The Lancet, Pedersen-Bjergaard et al. (2001) with permission from Elsevier lO-i
0 20 40 6 0 80 100 120 Serum ACE activity (U/L)
Figure 3.7 Association between severe hypoglycaemia and serum ACE activity according to C-peptide status and self-estimated awareness of hypoglycaemia. Reprinted from The Lancet, Pedersen-Bjergaard et al. (2001) with permission from Elsevier
These findings were replicated in a prospective study for one year in 107 adults (Pedersen-Bjergaard et al., 2003a). Serum ACE activity in the fourth quartile was associated with a 2.7-fold increased risk of severe hypoglycaemia compared to activity in the lowest quartile. Compared to subjects with the II genotype, individuals with the DD genotype had a 1.8-fold increased risk of severe hypoglycaemia, although this did not reach statistical significance. Higher serum ACE concentrations were also associated with an increased risk of severe hypoglycaemia in Swedish children and adolescents with type 1 diabetes (Nordfeldt and Samuelsson, 2003).
The Danish group have put forward a number of possible mechanisms to explain their observations (Pedersen-Bjergaard et al., 2001; Pedersen-Bjergaard et al., 2003a). They speculate that low levels of serum ACE may be associated with less cognitive deterioration during acute hypoglycaemia, thereby increasing the likelihood that remedial action to correct hypoglycaemia is taken. Alternatively, low serum ACE activity might enhance counterregulation or reduce production of toxic substances, e.g. reactive oxygen species, during hypogly-caemia. All these putative mechanisms remain highly speculative, but the authors also raise one other intriguing possibility: namely that ACE inhibition might reduce the frequency of hypoglycaemia. This may seem counterintuitive, because previous population-based studies suggested an association between severe hypoglycaemia and use of ACE inhibitors (Herings et al., 1996; Morris et al., 1997). However, these studies are flawed and a re-examination of the role of ACE inhibitors in ameliorating the impact of hypoglycaemia seems warranted.
Recent data from one centre in Scotland have not demonstrated such a strong link between serum ACE and severe hypoglycaemia in adults with type 1 diabetes (Zammitt et al., 2007), nor has a study of children and adolescents with type 1 diabetes in Australia (Bulsara et al., 2007). This should, therefore, serve as a reminder that genetic susceptibility to any biological variable may not be the same in different populations and reinforces the need for the Danish observations to be examined in other countries and ethnic groups. However, the studies by Pedersen-Bjergaard and colleagues raise the intriguing prospect that there may be yet other genetic factors that influence susceptibility to hypoglycaemia. Identification of these may help stratify risk in individuals with type 1 diabetes and may ultimately lead to the development of novel therapeutic interventions to prevent or ameliorate the impact of hypoglycaemia.
Several studies have demonstrated the importance of endogenous insulin secretion in defining risk of hypoglycaemia. Individuals who are C-peptide negative, i.e., who have no endogenous insulin production, have an approximately two- to fourfold increased risk of severe hypoglycaemia compared to people with detectable C-peptide (Bott et al., 1997; The Diabetes Control and Complications Trial Research Group, 1997; Muhlhauser et al., 1998; Pedersen-Bjergaard et al., 2001). These data mirror the clinical experience that severe hypoglycaemia is rare in the 12 months after the diagnosis of type 1 diabetes (Davis et al., 1997), when significant concentrations of endogenous insulin can be measured. The presumption is that the ability of endogenous insulin levels to fall in the face of a declining blood glucose concentration provides an additional layer of protection in the defences against hypoglycaemia and thus reduces overall risk.
Few diabetes specialists who were practising in the mid-1980s will forget the controversy that surrounded the introduction of human insulin. A substantial and vocal minority of patients with type 1 diabetes claimed that the change from animal-derived to human insulin was associated with a loss of warning symptoms of hypoglycaemia and thus an increased risk. These claims were subject to considerable scientific scrutiny and ultimately a systematic review of the evidence found no evidence to support the premise that treatment with human, as opposed to animal, insulin was associated with an increased risk of hypoglycaemia (Airey et al., 2000).
Since that time, several insulin analogues have been developed and their introduction into clinical practice has been accompanied by the publication of studies that purport to demonstrate that use of the insulin analogues is associated with a lower frequency of hypoglycaemia, in the face of stable or improved glycaemic control (Anderson et al., 1997; Garg et al., 2004; Hermansen et al., 2004). However, recent systematic reviews suggest that neither short-acting (Siebenhofer et al., 2004) nor long-acting (Warren et al., 2004) insulin analogues are associated with clinically significant lower rates of hypoglycaemia. The usual caveats of such clinical trials apply in terms of patient characteristics and recording of hypoglycaemia. In one observational study from Colorado, the frequency of severe hypoglycaemia rose in clinic patients in the immediate aftermath of the DCCT as efforts to intensify insulin therapy were instituted, but from 1996 onwards, the rates of severe hypoglycaemia declined and the authors linked this to the introduction of insulin lispro (Chase et al., 2001). This study was subject to the effects of numerous confounders, but further data from routine clinical practice are required to help clarify the impact of insulin analogues on risk of hypoglycaemia. The introduction of inhaled insulin (Exubera) has not been associated with a lower risk of hypoglycaemia in either type 1 or type 2 diabetes; when compared with insulin administered by the subcutaneous route, rates of severe hypoglycaemia were either equivalent (Hollander et al., 2004) or higher (Skyler et al., 2005).
Higher doses of insulin have also been associated with an increased risk of hypoglycaemia in some studies (The Diabetes Control and Complications Trial Research Group, 1997; ter Braak et al., 2000), but not in others (Table 3.3). Several possible explanations may account for this relationship - high insulin doses may be a sign of less endogenous insulin production, or of efforts to achieve strict glycaemic control. It may also reflect sub-optimal compliance with insulin therapy and so indicate a pattern of behaviour that predisposes to marked fluctuations in glucose control.
As has already been discussed in the section on asymptomatic hypoglycaemia, nocturnal hypoglycaemia is very common (Vervoort et al., 1996). In the DCCT, 43% of episodes of severe hypoglycaemia occurred between midnight and 8.00 a.m., and 55% of episodes occurred when individuals were asleep (The DCCT Research Group, 1991). The potential for nocturnal hypoglycaemia engenders significant anxiety among patients, particularly in individuals who live alone. Patients worry that they will not awaken when hypoglycaemia occurs and that they may be left incapacitated or die as a consequence. Many individuals, as a result, maintain higher blood glucose concentrations at bedtime to reduce the risk of nocturnal hypoglycaemia. Hypoglycaemia appears to be more common at night because counterregulatory hormone responses are blunted during sleep in people with type 1 diabetes (Jones et al., 1998; Banarer and Cryer, 2003). Moreover, hypoglycaemia awareness is also reduced during sleep (Banarer and Cryer, 2003), and so ultimately sleep impairs both the physiological and behavioural responses to hypoglycaemia. Unsurprisingly, considerable effort has been directed at developing strategies to reduce the risk of nocturnal hypoglycaemia and these are discussed in Chapter 4.
In a retrospective study of 44 patients with type 1 diabetes with impaired renal function (serum creatinine > 133umol/l and proteinuria), the incidence of severe hypoglycaemia was five times higher than in matched subjects with normal renal function (Muhlhauser et al., 1991). In other studies, severe hypoglycaemia was linked with nephropathy (ter Braak et al., 2000), peripheral neuropathy and retinopathy (Pedersen-Bjergaard et al., 2004). Although it is generally recognised that insulin requirement declines in advanced renal disease, with reduced clearance of insulin, this association between hypoglycaemia and nephropathy (and other microvascular complications) could be confounded by many other factors, e.g. concomitant drug therapy (ter Braak et al., 2000) and the co-existence of acquired hypoglycaemia syndromes which, like microvascular complications, are linked with increasing duration of diabetes.
The role of peripheral autonomic neuropathy in increasing the risk of severe hypogly-caemia has been considered in several studies and deserves mention. In the EURODIAB IDDM Complications Study, the presence of abnormal cardiovascular reflexes was associated with a 1.7-fold increased risk of severe hypoglycaemia (Stephenson et al., 1996). Gold et al. (1997) also demonstrated that autonomic neuropathy was associated with a small increased risk of severe hypoglycaemia in 60 patients with type 1 diabetes, but no such relationship was demonstrated in the DCCT (The DCCT Research Group, 1991). The mechanism underlying this association remains to be fully elucidated. Peripheral autonomic neuropathy often coexists with impaired hypoglycaemia awareness in patients with type 1 diabetes, presumably because both conditions are associated with diabetes of long duration (Hepburn et al., 1990), but impaired awareness can readily occur in the absence of peripheral autonomic neuropathy (Hepburn et al., 1990; Ryder et al., 1990; Bacatselos et al., 1995). It is well established that severe autonomic neuropathy is associated with 'gastroparesis diabeticorum', which can cause marked swings in blood glucose concentration. Although such severe gastroparesis is now rare, delayed gastric emptying is relatively common (Kong et al., 1996) and may explain at least part of the association between hypoglycaemia and autonomic neuropathy.
Psychological factors clearly play a crucial role in determining an individual's likelihood of developing severe hypoglycaemia. Low mood (Gonder-Frederick and Cox, 1997), emotional coping (Bott et al., 1997) and socio-economic status (Muhlhauser et al., 1998; Leese et al., 2003) have been linked to risk of severe hypoglycaemia and so too have other more straightforward behavioural factors such an individual's propensity to carry a supply of carbohydrate for emergency use (Bott et al., 1997) and their determination to achieve normoglycaemia (Muhlhauser et al., 1998).
Since the 1980s, Cox and colleagues at the University of Virginia, USA, have carried out seminal investigations to explore the psychological impact of hypoglycaemia on people with type 1 diabetes. They developed a Fear of Hypoglycaemia scale, which sought to quantify the anxieties that people with type 1 diabetes have with respect to hypoglycaemia and the extent to which individuals take steps to avoid experiencing such episodes (Cox et al., 1987). Unsurprisingly, there is a close association between fear of hypoglycaemia and perceived risk of future severe hypoglycaemia (Gonder-Frederick et al., 1997). In many instances this is appropriate, in that 'fear' ratings are often high in people who have impaired awareness of hypoglycaemia and/or have experienced multiple episodes in the past (Gold et al., 1996). However, in other people, fear of hypoglycaemia may be high while absolute risk is low -such individuals often display high levels of trait anxiety or have had a traumatic previous experience of hypoglycaemia. It is often extremely difficult to persuade such individuals to maintain strict glycaemic control. Conversely, there are people who have a very low fear of hypoglycaemia, despite a propensity to recurrent episodes. Such individuals may fail to take appropriate precautionary measures, thereby putting themselves and others at risk if hypoglycaemia occurs, for example, when that individual is driving a car.
Endocrine disorders, such as Addison's disease and hypopituitarism, which are associated with a deficiency of counterregulatory hormones, can be associated with an increased risk of hypoglycaemia in adults with type 1 diabetes. These are uncommon in everyday diabetes practice, but clinicians should maintain a high index of suspicion particularly in the patient who simultaneously demonstrates a marked and otherwise unexplained decline in insulin requirements.
The list of other possible risk factors for hypoglycaemia is a long one. Smoking is a relatively novel marker of increased risk (Pedersen-Bjergaard et al., 2004), but may be confounded by other diabetes-related and lifestyle factors, e.g. regular use of alcohol (ter Braak et al., 2000). In the DCCT, men and adolescents were at increased risk (The Diabetes Control and Complications Trial Research Group, 1997), but these associations have not been replicated with any consistency in other studies (Table 3.3). Risks associated with pregnancy are addressed in Chapter 10.
Causes and Risk Factors for Hypoglycaemia: Summary and Conclusions
Many of the risk factors described above are inter-related and any given individual may have more than one, which clearly will increase their overall risk. However, at its most fundamental level, hypoglycaemia in adults with type 1 diabetes is a consequence of the inability of exogenous insulin levels to fall in response to a declining blood glucose concentration. In people who have had type 1 diabetes for several years, the situation is exacerbated because of a failure of the normal physiological counterregulatory defence mechanisms, which in non-diabetic individuals serve to increase exogenous glucose production and generate warning symptoms. This counterregulatory failure worsens with increasing duration of diabetes, particularly if glycaemic control is strict and the individual has experienced recurrent episodes of severe hypoglycaemia. Other factors may come in to play, for example particular behavioural patterns and the time-action profiles of the exogenous insulin. Recent data also raise the possibility that there may be inherent genetic susceptibility to hypoglycaemia, but this needs to be affirmed in wider populations and the underlying mechanisms more clearly dissected. Novel strategies to reduce overall risk of hypoglycaemia are urgently required.
• There are no definitive criteria for what constitutes hypoglycaemia, but most specialists distinguish mild from severe episodes depending on whether or not the individual is able to self-treat.
• The 'average' adult with type 1 diabetes will experience many thousands of episodes of mild hypoglycaemia over a lifetime, with a typical frequency of one to two episodes per week.
• Severe hypoglycaemia is less common, and on average occurs once or twice every year, with an annual prevalence of around 30%. However, the distribution is heavily skewed, such that many individuals are unaffected over a calendar year, while a small number experience recurrent episodes.
• Severe hypoglycaemia requiring treatment with intramuscular glucagon and/or intravenous glucose is even less common, and the majority of all episodes of hypoglycaemia are managed in the community by the patient and/or relatives and friends, without recourse to emergency services.
• Hypoglycaemia occurs when there is an imbalance between insulin-mediated glucose disposal and glucose influx into the circulation from the liver and exogenous carbohydrate. Typical precipitants include patient error in insulin dosage, alcohol and exercise.
• Plasma concentrations of exogenous insulin cannot decline in response to falling blood glucose and the time action profiles of current insulins do not accurately mimic the normal physiological variation in insulin. In a high proportion of cases, no underlying precipitant of a given episode of hypoglycaemia can be identified.
• The DCCT and other intervention studies have provided substantial information of the epidemiology of severe hypoglycaemia, but individuals participating in such studies may not be representative of the wider population who have type 1 diabetes.
• A major determinant of increased risk of severe hypoglycaemia is Hypoglycaemia-Associated Autonomic Failure, which is a consequence of exposure to recurrent episodes of hypoglycaemia. This is a feature of individuals with long-standing diabetes, intensive insulin therapy and strict glycaemic control.
• Recent data suggest that there may be specific genetic factors that predispose to an increased risk of hypoglycaemia.
Was this article helpful?