A randomised controlled study of high intensity exercise as a dishabituating stimulus to improve hypoglycaemia awareness in people with type 1 diabetes: a proof-of-concept study

Severe hypoglycaemia is an adverse effect of insulin therapy in diabetes that has a well-recognised morbidity and mortality [1]. Fear of hypoglycaemia often outweighs concerns about long-term consequences of chronic hyperglycaemia [2], and acts as a barrier to patients achieving optimal glycaemic control [3]. The major risk factor for severe hypoglycaemia is impaired awareness of hypoglycaemia (IAH), which increases the risk of an event six-fold [4]. People with IAH have a reduced ability to perceive the onset of acute hypoglycaemia [5]. It affects 25% of all people with type 1 diabetes and, of concern, the incidence of IAH has not changed in the last 2–3 decades despite the introduction of insulin analogues and improved insulin delivery systems [6]. IAH is an acquired abnormality that should be placed alongside chronic microvascular complications such as retinopathy, neuropathy and nephropathy as it can be just as serious and disabling.

The average individual with type 1 diabetes will experience about two episodes of symptomatic hypoglycaemia per week [7, 8]. Ordinarily, hypoglycaemia initiates a biological counterregulatory response that comprises a broad range of physiological, symptomatic and behavioural changes that together act to restore normal glucose homeostasis. Failure of this biological response leads to progressive hypoglycaemia and may result in individuals being unable to self-treat or to loss of consciousness (severe hypoglycaemia). In their seminal study in 1991, Heller and Cryer demonstrated that prior exposure to hypoglycaemia leads to suppression of both hormonal and symptomatic responses during hypoglycaemia of the same depth and duration induced the following day [9]. This finding was later confirmed in people with type 1 diabetes [10], and it was further established that increasing the frequency [11] and duration [12] of prior exposure to hypoglycaemia increases the extent of counterregulatory response suppression, while strict avoidance of hypoglycaemia restores counterregulatory responses [13]. Cumulatively, this exposure to recurrent hypoglycaemia leads to the development of IAH where individuals demonstrate a reduced magnitude and increased threshold (lower glucose) for sympathoadrenal activation and symptomatic experience of hypoglycaemia. Unfortunately, despite nearly three decades since Heller’s and Cryer’s seminal work, the biological mechanisms underpinning the development of IAH remain unknown.

Habituation is a form of adaptive memory that develops in many organisms in response to a repeated, often stressful stimulus [14]. Habituation has been defined as a reduction in the psychological, behavioural or physiological response to a stimulus as a result of recurrent or prolonged exposure [14, 15]. Individuals with type 1 diabetes who develop IAH demonstrate reduced physiological (hormonal and symptomatic), psychological (reduced anxiety) and behavioural (inappropriate responses to hypoglycaemia while driving) responses to hypoglycaemia [14‐18]. This led us to propose that IAH may represent a habituated response to repeated hypoglycaemia. To test this hypothesis, we recently examined in a rodent model whether IAH could be reversed, at least temporarily, by the introduction of a single novel stress stimulus, a phenomenon referred to as dishabituation. Rodents were exposed to non-severe hypoglycaemia three times a week for a period of 4 weeks (habituation), and subsequently randomised to a single burst of high intensity training (HIT; dishabituation) or control (no exercise) before undergoing a controlled experimental hypoglycaemia study 24 h later [19]. In comparison with control animals, those who had been exposed to the dishabituatory stimulus had a significantly greater counterregulatory hormonal response during the controlled hypoglycaemia study [19]. While this finding supported the hypothesis that IAH develops as a result of habituation, the rodents did not have type 1 diabetes and the counterregulatory response was assessed via hormonal measures only [19]. In the current proof-of-concept study, we sought to further test the habituation hypothesis by examining whether a single burst of HIT would act as a dishabituatory stimulus in people with long-duration type 1 diabetes who had IAH.

This was a single-centre, randomised crossover study carried out at Ninewells Hospital, Dundee, Scotland. Ethics approval was obtained from an independent research ethics committee (17/SS/0150) and the study was registered with the International Standard Randomised Controlled Trials Register (ISRCTN15236211). The study was conducted in accordance with the Declaration of Helsinki, and written informed consent was obtained from all participants before inclusion in the study.

Inclusion criteria were: diagnosis of type 1 diabetes for at least 5 years, treatment with multiple daily injections or continuous subcutaneous insulin infusion, HbA_1c <75 mmol/mol (9%), age of 18 to 55 years, and IAH (Gold score [4] ≥4, modified Clarke score [20] ≥4 or Dose Adjustment For Normal Eating [DAFNE] hypoglycaemia awareness rating [21] 2 or 3). Exclusion criteria were: history of significant cardiac, renal, respiratory or neurological disease, pregnancy or breastfeeding, treatment with beta-blockers, history of high-risk foot disease, competitive sportsman [22] or participation in HIT in the preceding 6 months, or physical disability that might limit exercise.

The randomisation relates to the order that the HIT or rest intervention was carried out by the participant. Randomisation was carried out in a good clinical practice (GCP)-compliant manner using http://​www.​randomization.​com/​.

Participants were identified using the Scottish Diabetes Research Network (SDRN) and from the diabetes outpatient clinics in NHS Tayside, Scotland. The study took place at the Clinical Research Centre, Ninewells Hospital and Medical School, Dundee.

After obtaining informed written consent and initial screening, which included confirmation of IAH, each participant attended the Clinical Research Centre on seven separate occasions (See Fig. 1). During visit 1 anthropometric measurements, physical examination and electrocardiogram were performed. In addition, participants undertook an incremental maximal to volitional exhaustion exercise test using on a cycle ergometer (Corival, Lode, Groningen, the Netherlands; Metamax 3B, CORTEX Biophysik, Leipzig, Germany) to determine their \( \dot{V}{\mathrm{O}}_2 \) and peak heart rate.

For this study we have defined HIT as episodic short bursts of high intensity exercise separated by short periods of recovery at lower intensity [23]. The HIT programme used in this study was 20 min in duration; 5 min of gentle warm up cycling at 50 W, 40–60 rev/min followed by 4 × 30 s cycle sprints [2 min recovery] at 150% of maximum wattage achieved during \( \dot{V}{\mathrm{O}}_{2\mathrm{peak}} \) assessment with an aim to achieve ≥90% peak heart rate; concluding with 5 min cool down at 50 W, 40–60 rev/min. Standardised advice [24] was given regarding blood glucose levels and insulin dose adjustment prior to and following HIT. The control intervention was 20 min of seated rest. Venous blood samples were taken for counterregulatory hormones (adrenaline, noradrenaline [norepinephrine], glucagon and lactate) at baseline of both the rest and HIT intervention visits with samples also taken after the third and fourth bouts of exercise during the HIT intervention. Each participant underwent a single HIT programme or controlled rest intervention the afternoon before undergoing a hyperinsulinaemic–hypoglycaemic clamp study the following day. The hyperinsulinaemic–hypoglycaemic clamp study was used to assess their hormonal counterregulatory response, symptom awareness and cognitive response. Hyperinsulinaemic–hypoglycaemic clamp study is considered the gold standard for comparing counterregulatory response to equivalent hypoglycaemia [25]. There was a 2 week cooling off period between the interventions to avoid any lasting effect prior to crossing in to the second arm of the study.

Habituation is a form of adaptive learning where the response to a stimulus is reduced or ceases altogether following repeated exposure to the stimulus. This is often considered in the context of innate behaviours [14, 15]. The gill-withdrawal (sensory–motor) reflex in Aplysia [31] is a classic model of habituation. When the gill-withdrawal reflex is repeatedly induced by a tactile external stimulus to the siphon, the scale of the response is markedly diminished (habituation). Hypoglycaemia can be considered an internal sensory stimulus, which leads in turn to a reflex counterregulatory (motor) response [32]. If organisms adapt to repeated hypoglycaemia through habituation then it would be predicted that repeated exposure to hypoglycaemia would lead to a markedly diminished response. This is indeed the case as has been shown in humans [9, 10] and rodents [33], and as with a habituated response, we see widespread and progressive suppression of physiological (e.g. suppressed counterregulatory hormonal and symptomatic responses), behavioural (e.g. reduced drive to feed, impairment when driving [17]) and psychological (e.g. reduced anxiety) responses to hypoglycaemia as a result of repeated exposure [16]. Furthermore, a cardinal feature of the habituated gill-withdrawal reflex in Aplysia is that it can then be restored, at least temporarily, by applying a strong tactile stimulus to another part of the animal (dishabituation) [31]. Our present findings, supported by our previous work in rodents [19], that hormonal, symptom and cognitive responses to hypoglycaemia can be at least partially restored following a single episode of HIT in individuals with long-standing type 1 diabetes who have been exposed to multiple episodes of hypoglycaemia, therefore provides robust support for the hypothesis that IAH develops through habituation to recurrent hypoglycaemia.

There are nine proposed criteria that define habituation [15]. As described, criterion 1: ‘Given that a particular stimulus elicits a response, repeated applications of that stimulus result in a decreased response (habituation)’ [15], is characteristic of recurrent hypoglycaemia [9, 10]. Similarly, criterion 2: ‘If the stimulus is withheld, the response tends to recover over time (spontaneous recovery)’ [15], and we know that hypoglycaemia avoidance can lead to recovery of counterregulatory responses to subsequent hypoglycaemia [13, 34]. In this study, we have addressed criterion 8: ‘Presentation of another (usually strong) stimulus results in recovery of the habituated response (dishabituation)’, and have shown that it is possible, at least temporarily, to recover hypoglycaemia counterregulation. Interestingly, criterion 7: ‘Habituation of response to a given stimulus exhibits stimulus generalization to other stimuli’, may explain why our results differ from previous work suggesting that exercise can further suppress counterregulatory responses to hypoglycaemia [35, 36]. It is possible that the mild–moderate exercise regimen, leads through stimulus ‘generalisation’, to suppression of the subsequent counterregulatory response to insulin-induced hypoglycaemia. The key differentiating feature between the studies being the intensity of the different stimulus.

Habituation is considered an adaptive survival response to repeated exposure of a physiological stressor in order to develop a degree of tolerance of that specific stressor. In contrast, the physiological response to repeated hypoglycaemia is currently considered maladaptive and is often referred to as hypoglycaemia-associated autonomic failure (HAAF), as originally proposed by Cryer [37]. The term HAAF is, however, misleading because, as mentioned, the suppressed counterregulatory response can be reversed in many people through strict avoidance of hypoglycaemia [13, 34]. This means the autonomic system does not ‘fail’, and instead is more likely to be adapting to the presence or absence of repeated hypoglycaemia. Pre-clinical studies are supportive of the habituation hypothesis in that recurrent hypoglycaemia triggers a series of changes at a cellular level that act to induce tolerance, facilitating organisms to cope better when subsequently deprived of energy. For example, prior recurrent hypoglycaemia protects against neuronal death during subsequent very severe hypoglycaemia [38]. Glucocorticoids [39] and urocortin [40], which are key to neuroprotection and initiation of adaptive cellular mechanisms [41], are also integral to the organism’s response to recurrent hypoglycaemia. Therefore, these systematic and cellular responses to recurrent hypoglycaemia also point to IAH developing as an adaptive, habituated response to repeated hypoglycaemic stress.

In the current study we employed HIT as the novel dishabituating stimulus. We confirmed the effectiveness of the HIT intervention as a novel stimulus by demonstrating that it evoked a marked rise in heart rate, adrenaline and lactate. The day following the HIT (or rest) intervention, participants underwent a controlled hypoglycaemia study where we demonstrated significant increases induced by preceding HIT in the hormonal counterregulatory response. There was also, and of relevance in clinical practice, an increase in total symptom scores (EHS [27]) in response to matched hypoglycaemia following HIT, indicating participants were more aware of hypoglycaemia. Finally, the 4CRT was prolonged during hypoglycaemia following HIT. Choice reaction time is used to examine psychomotor and attentional focus. Acute hypoglycaemia markedly impairs performance in almost all aspects of executive function in adults with type 1 diabetes, with time taken to complete cognitive testing taking considerably longer [42, 43]. Zammitt et al have shown that 4CRT is also prolonged during experimental hypoglycaemia in people with type 1 diabetes and intact awareness, but not in those with IAH [44]. Our finding of a significant slowing of 4CRT during hypoglycaemia in participants with IAH following HIT is therefore indicative that dishabituation has, at least partially, restored counterregulatory hormonal, symptom and cognitive responses towards that seen in individuals with intact awareness.

There are a number of limitations to our study. Although the hyperinsulinaemic–hypoglycaemic clamp is the gold standard technique for assessing counterregulatory responses to hypoglycaemia [25], the studies are conducted under strictly controlled conditions in a laboratory and, as such, the experience of hypoglycaemia may not fully reflect its real-world experience. We also used three different hypoglycaemia awareness questionnaires [4, 20, 21], which were not concordant for all participants. However, 11/12 participants scored 2 or 3 on the DAFNE hypoglycaemia awareness rating and these individuals show a fourfold increased risk of severe hypoglycaemia [21], while 10/12 participants had a modified Clarke score ≥4, which carries a 4.6-fold increased risk of severe hypoglycaemia [45].

Overall, the current study supports the hypothesis that IAH in type 1 diabetes develops as a result of habituation. This is not firmly established as additional studies will need to be performed to test the other features of habituation. However, consistent with our hypothesis, we have demonstrated that the introduction of HIT as a dishabituating stimulus, in people with long-standing type 1 diabetes and IAH, acutely restores a broad array of physiological, symptomatic and cognitive responses during subsequent experimental hypoglycaemia. If this effect is maintained following a more prolonged intervention with HIT, it may offer a novel, therapeutically attractive strategy to improve and restore hypoglycaemia awareness in people with type 1 diabetes and reduce the risk of severe hypoglycaemia.