Bas Schouwenberg

It was first noted in the 1940s that many patients with type 1 diabetes have lost autonomic symptoms as early warning signs of hypoglycaemia, and proceed directly to neuroglycopenic manifestations (1). We now call this phenomenon impaired awareness of hypoglycaemia (IAH) and know that it results from a process of habituation to recurrent hypoglycaemia (Chapter 1). Long diabetes duration, low HbA1c and unmeasurable C-peptide levels, all of which increase the risk of hypoglycaemia, are therefore also risk factors for IAH. As outlined in Chapter 1, classical risk factors cannot explain the total burden of IAH among the population with type 1 diabetes. In this thesis, we estimated the prevalence of IAH over time, investigated the potential impact of genetic predisposition for its occurrence and analysed whether IAH affects mortality.

For most patients with type 1 diabetes, hypoglycaemia remains the most important limitation in the pursuit of optimal glycaemic control and many patients are prone to counter regulatory defects including IAH (2). In the past 30 years, the prevalence of IAH was estimated in various type 1 diabetes cohorts around the world at 20-30% (3-6). Since then, basal-bolus insulin regimens with insulin analogues, continuous subcutaneous insulin infusion (CSII), frequent glucose monitoring and structured educational programs have become standard therapy for patients with type 1 diabetes (7). Since all of these developments are associated with lower risks of hypoglycaemia, we were interested in the current prevalence of IAH and also whether it had changed over time (Chapter 2). Our first analysis among 485 consecutive patients with type 1 diabetes disclosed a prevalence of IAH of 32.5%, certainly not lower than previous estimates performed elsewhere. Moreover, although a low HbA1c-level was a strong risk factor for IAH, high HbA1c levels certainly did not protect against its occurrence. Other clinical factors associated with IAH included male gender, age, duration of diabetes and impaired renal function (GFR<30ml/min.). Remarkably, when we repeated the questionnaire several years later, the vast majority of patients with IAH still had IAH and the overall prevalence of IAH had not changed. The more or less stable prevalence of IAH over time and across various type 1 diabetes cohorts suggests the contribution of a genetic propensity for the development of IAH. In Chapter 3, we therefore studied the role of polymorphisms in the beta2-adrenergic receptor in the occurrence of IAH in a population of type 1 diabetes. Since homozygosity Gly16 of the beta2-adrenergic receptor makes this receptor particularly prone to a process of desensitization upon repeated stimulation (8, 9), such as may occur during hypoglycaemia, we posited that patients with type 1 diabetes with this SNP would be at high risk of developing IAH. This hypothesis was first tested in a relatively small cohort of unselected patients with type 1 diabetes, described in chapter 3. In this study, we found that patients homozygous gly16 for the beta2-adrenergic receptor were at an approximately threefold greater risk of having been diagnosed with IAH than patients without this polymorphism. A limitation of this study was that its small sample size precluded an analysis of linkage disequilibrium, which is known to exist between two frequently occurring SNPs in the beta2-adrenergic receptor. To examine the role of beta2-adrenergic receptor in the development of IAH in more depth, we investigated the effect of antecedent hypoglycaemia as a model of IAH on next-day beta2-adrenergic sensitivity in healthy volunteers. Beta2-adrenergic sensitivity was assessed by measuring the forearm blood flow response to intra-arterial infusion with the beta2-adrenergic receptor agonist, salbutamol, and we enrolled subjects who were either homozygous Arg16 or homozygous Gly16 for the beta2-adrenergic receptor (Chapter 4). Somewhat surprisingly, antecedent hypoglycaemia increased rather than reduced next-day beta2-adrenergic sensitivity in the whole group. This effect appeared to be driven solely by the homozygous Arg16 subjects, who showed a significant increase in beta2-adrenergic sensitivity, whereas this effect was blunted in the gly16 homozygous subjects. These data argue against a direct role for reduced beta2-adrenergic sensitivity in the pathogenesis of IAH. Instead, the enhanced beta2-adrenergic sensitivity after hypoglycaemia in Arg16 homozygous subjects potentially serves as a protective mechanism to retain or even improve the detection of future hypoglycaemic events. Such a protective response appears to be absent in subjects homozygous for gly16, possibly explaining their propensity for developing IAH. It could also be speculated that exposure to many more hypoglycaemic events, as is the case in patients with type 1 diabetes and IAH (10-17), are required to eventually reduce sensitivity of the beta2-adrenergic receptor. In Chapter 5 we compared hemodynamic responses to systemic adrenaline infusion between men and women. This study revealed intriguing differences between de sexes, with men showing predominately beta2- and possibly beta1-adrenergic effects and women showing mainly alpha-adrenergic effects following adrenaline infusion that resulted in similar plasma levels as obtained during hypoglycaemia. Males have been reported to be at slightly higher risk of developing IAH than females (18-21). Since the adrenergic part of autonomic symptom responses to hypoglycaemia is mediated through beta-adrenergic receptor stimulation, this greater susceptibility in men for IAH may be explained partly by such differences in the sensitivity to adrenaline. In an effort to confirm our results on the relation between beta2-adrenergic receptor polymorphisms and IAH and to put these into context of previously reported genetic factors associated with either the occurrence of severe hypoglycaemia or hypoglycaemia (22, 23), we performed a larger cohort study. Patients from our outpatient department were asked about their awareness status using the same validated questionnaire and blood was drawn for DNA analysis. In this study, discussed in Chapter 6, we were able to reproduce our previous results by showing a greater prevalence of IAH in subjects homozygous gly16 than in subjects without this SNP. Haplotype analysis, performed because of linkage disequilibrium between the SNPs at codons 16 and 27, disclosed that Gly16Gly patients who were heterozygous Gln27Glu were at the highest risk of IAH. Neither SNP of the beta2-adrenergic receptor, or one of the other genes studied were a risk factor for severe hypoglycaemia and none of the other genetic factors was associated with IAH. The magnitude of the association was comparable to that of clinical risk factors for IAH, including HbA1c, age, duration of diabetes and gender, correction for which only slightly modified the relationship between beta2-adrenergic receptor polymorphisms and IAH. There is now substantial evidence that links severe hypoglycaemia to mortality, both in people with type 1 or type 2 diabetes and in patients treated with insulin on the ICU. Given that IAH is the most important risk factor for severe hypoglycaemia in people with type 1 diabetes, we aimed to explore the relationship between IAH and mortality. For this purpose, we collaborated with a research team from Denmark to combine their cohort of 269 patients (follow-up, 12 years) with ours of 482 patients (follow-up, 6.5 years). In neither cohort, IAH was associated with increased (or decreased) mortality (Chapter 7). In the Dutch cohort, there was a borderline significant association between IAH and cardiovascular mortality, which should be further explored. One possibility is that this was the result of fatal arrhythmia, since (nocturnal) hypoglycaemia has been shown to cause QTc-lengthening (24, 25), a well-known predictor of ventricular tachycardia syndromes. However, the association between IAH and cardiovascular death in our cohort was based on very low numbers (n=9) and was not mirrored in the Danish cohort, so that the play of chance cannot be excluded. General discussion and future perspectives Based on the research results presented in this thesis, several conclusions can be drawn. First of all, our results clearly show that IAH is still highly prevalent among patients with type 1 diabetes, despite the introduction of new insulin analogues and insulin delivering devices, all aimed at reducing the burden of hypoglycaemia. We also show that IAH is quite stable over time: three out of four patients affected at the first assessment retained IAH when questioned several years later. Because classical clinical risk factors, including low HbA1c, loss of residual beta-cell function and duration of diabetes, do not explain the full extent of IAH, we hypothesized involvement of genetic predisposition. Indeed, genetic variation in the beta2-adrenergic receptor appears to play a role in the pathogenesis of IAH, with patients with homozygosity for Gly16 being slightly more prone for developing IAH, especially when combined with the Glu27Gln genotype. However, the underlying mechanism explaining this genetic predisposition remains to be revealed, especially since hypoglycaemia did not reduce beta2-adrenergic sensitivity. Other researchers earlier reported other polymorphisms, including that of the ACE gene, to predispose for increased risk of severe hypoglycaemia. None of these genetic factors, however, appeared to be related to IAH or to the occurrence of severe hypoglycaemia in our cohort. Genetic research in much larger cohorts of patients with type 1 diabetes is needed to examine the potential role of these genotypes in the development of IAH and may uncover new candidate genes related to IAH. Since IAH also affects up to 10% of patients with type 2 diabetes on insulin, such genetic studies are also needed in the type 2 diabetes population, especially since the number of patients at risk of hypoglycaemia increases at alarming speed. Finally, we are the first to investigate the consequences of IAH on long-term clinical outcomes. We did not identify a link between IAH (or SH) and increased (cardiovascular) mortality. This is important and reassuring news for patients with type 1 diabetes, their relatives and health-care providers, but certainly requires confirmation in larger cohorts. Our studies were performed before widespread use of glucose monitoring devices or flash glucose meters. The evidence that severe hypoglycaemic events can be avoided with these devices is now slowly emerging, but there is little evidence that these devices are able to reverse IAH. Whether genetics are involved in this apparent discrepancy should be the focus of further research. Case vignette (continued) The woman in the case vignette in Chapter 1 had asked her health care provider several questions about hypoglycaemia and the altered perception of these events. She is particularly interested in why she no longer feels hypoglycaemic events like she used to do and whether hypoglycaemia and her reduced ability to perceive hypoglycaemic symptoms are harmful. We can tell her that she clearly has IAH, which has probably contributed to the severe hypoglycaemic event that she recently encountered. Risk factors for IAH for her include her low HbA1c – reflecting tight glucose control – and long duration of diabetes, possibly indicating loss of residual beta-cell function. It is possible that genetic factors also play a role, but we cannot recommend genetic testing on the basis of this thesis. We can reassure her that there is no indication that IAH increases mortality. However, precautionary action is required to avert the risk of severe hypoglycaemia. Therapies targeting the beta2-adrenergic receptor will probably not help her regain awareness for hypoglycaemia.