It is known that hypoglycaemia affects various domains of cognitive function. One aspect of cognitive function is 'social cognition', which is our ability to interpret facial expressions, gestures and speech. It is an approach to understanding human judgement and behaviour. There is anecdotal evidence for negative behavioural responses such as aggressiveness and argumentativeness during hypoglycaemia and while research has shown that hypoglycaemia can cause significant changes in mood, little information exists regarding its effect on social cognition. It is therefore not known whether hypoglycaemia affects this aspect of cognitive function but, if it did, it could explain why people with low blood sugar due to insulin treatment are often resistant to offers of help (for example, they may misinterpret a friendly gesture as being threatening). Knowledge of this effect of hypoglycaemia can be used to educate relatives and carers of people with diabetes who may suffer this problem.

Hypoglycaemia is also known to have an effect on the electrical rhythm of the heart. This is thought to be secondary to adrenaline secretion during hypoglycaemia which provokes a fall in the blood level of potassium, a type of electrolyte. Other electrolyte imbalances are known to predispose to heart rhythm abnormalities or arrhythmias in other situations and it is not known if the levels of these other electrolytes are affected during hypoglycaemia.

Using specific tests of social cognition and continuous electrocardiographic (ECG) monitoring, this study aims to find out whether social cognition is affected by an hour of hypoglycaemia and how hypoglycaemia affects blood electrolyte levels and the electrophysiology of the heart.

Hypoglycaemia is a very common side-effect of insulin treatment in people with diabetes. Anglo-Danish studies have confirmed the high frequency of hypoglycaemia affecting people with type 1 diabetes. The rate of severe hypoglycaemia is as high as 0.35 episodes per patient-year.

The human brain requires a constant supply of glucose for normal functioning, which is impaired by hypoglycaemia. As a consequence, cognitive and motor function can deteriorate rapidly with significant consequences. This has been studied extensively and it is known that hypoglycaemia affects various domains of cognitive function including short-term and working memory, attention, non-verbal intelligence, visual and auditory information processing, psychomotor function, spatial awareness, and executive functioning. However, to our knowledge, the effect of acute hypoglycaemia on social cognition has not been studied previously.

Social cognition examines the higher cognitive functions involved in interpreting facial expressions, gestures or speech. It is an approach to understanding human judgement and behaviour. There is anecdotal evidence for negative behavioural responses such as aggressiveness and argumentativeness during hypoglycaemia and while research has shown that hypoglycaemia can cause significant changes in mood, little information exists regarding its effect on social cognition. It is therefore not known whether hypoglycaemia affects this aspect of cognitive function but, if it did, it could explain why people who are hypoglycaemic are often resistant to offers of help (for example, they may misinterpret a friendly gesture as being threatening). Knowledge of this effect of hypoglycaemia can be used to educate relatives and carers of people with diabetes who may suffer this problem.

Hypoglycaemia is known to have an effect on the electrophysiology of the heart. It can cause prolongation of the QT interval(represents the interval between the Q wave and T wave in an ECG), demonstrable with electrocardiography which may predispose some individuals to an increased risk of cardiac arrhythmias. Studies have shown this QT interval prolongation to be associated with adrenaline secretion secondary to sympathoadrenal activation during an acute episode of hypoglycaemia and this provokes a fall in plasma potassium. Electrolyte imbalances are known to predispose to arrhythmias and it is not known if other electrolyte disturbances during hypoglycaemia may also affect the QT interval.

Using the hyperinsulinaemic glucose clamp technique, tests of social cognition will be administered during one hour of hypoglycaemia in 20 adults with type 1 diabetes. The study participants will act as their own controls by performing the same tests during euglycaemia on a separate study session separated by at least two weeks, using the same hyperinsulinaemic glucose clamp technique.

A Holter monitor will be used to perform continuous ECG monitoring throughout the duration of the study. Blood will be taken to measure plasma potassium, magnesium and calcium concentrations at various time-points during the study to seek any acute changes in these electrolytes during hypoglycaemia.

Study Aims

1. To determine the effect of hypoglycaemia on social cognition in adults with type 1 diabetes
2. To determine the effect of hypoglycaemia on blood electrolyte levels and QT interval on ECG

Methods

Subjects will undergo modified hyperinsulinaemic glucose clamps and initially will be maintained at euglycaemia, followed by an experimental condition of either hypoglycaemia or euglycaemia for 60 minutes. Euglycaemia will be restored before the clamp is discontinued.

Tests of social cognition will be administered within one hour of the experimental phase of the study. In addition, continuous ECG monitoring will be undertaken and plasma electrolyte concentrations will be measured at different time points during the baseline, experimental and recovery phases of the study.

Study design and procedures Each participant will take part in two sessions separated by at least two weeks: the control and experimental conditions.

Each session will commence at 08.00 hours after an overnight fast. Participants will be asked not to administer their usual morning insulin dose on study days. Female participants will be asked to have a pregnancy test to exclude any pregnant individual. Studies will take place in the Clinical Research Facility of the Royal Infirmary of Edinburgh.

The participant will be asked whether they have experienced or recorded any episode of hypoglycaemia in the preceding 24 hours. This is because antecedent hypoglycaemia is known to modify the counterregulatory and symptomatic responses to subsequent hypoglycaemia. The study will be postponed if any hypoglycaemia (symptomatic or biochemical) has occurred.

The participant will be seated/reclined on a hospital bed with a table in front of them which will be used for the various computer and paper based tests of social cognition.

A Holter monitor, which will continuously record a 3-lead ECG during the study, will be attached once the participant is seated. A baseline 12-lead ECG will also be made.

An intravenous cannula will be inserted into the non-dominant hand vein for regular sampling of blood glucose. Samples will be arterialised by inserting the cannula in a retrograde manner and a heated blanket will be used for warming the hand as is standard practice. A second cannula will be inserted into a vein in the antecubital fossa of the same arm for intravenous infusion of glucose and insulin. A modified hyperinsulinaemic glucose clamp will be performed. After a run-in period, arterialised blood glucose will be maintained at 4.5 mmol/L for 30 minutes. Blood glucose will then either be maintained at 4.5 mmol/L throughout (euglycaemia), or lowered over 20 minutes to 2.5 mmol/L (hypoglycaemia), and maintained at this level for 60 minutes. At the end of this experimental phase, euglycaemia (4.5 mmol/L) will be restored and maintained for a further period of 20 minutes. The glucose clamp will then be discontinued and the subjects will be given a meal. Subjects will be advised regarding adjustment of doses and times of insulin administration for the rest of the day by the study doctor who is experienced in the management of insulin-treated diabetes.

The subjects will not be informed which condition of the study is being performed, and will undergo the control (i.e. euglycaemic) and experimental (i.e. hypoglycaemic) conditions in a randomised and counter-balanced fashion. Alternative forms of the social cognition test will also be performed in a counter-balanced fashion. Every participant will receive the same intervention and randomisation is limited to the order in which these interventions are received.

Tests of social cognition: These will be performed once during the experimental phase of the study, with the exception of the National Adult Reading Test which will be performed at baseline as a measure of pre-morbid intelligence.

1. Faux pas test: A faux pas occurs when someone says something they should not have said, not realising that they should not say it. In this test of faux pas detection, 10 very short stories will be read out to the participant. They will also be presented with the story in written form so they can read along while it is being read aloud to them. Out of the 10 stories, 5 contain a faux pas. After each story, the participant will be asked questions based on the story in order to help determine if they have recognised the faux pas or not. The remaining 5 stories work as controls as they do not contain any faux pas. This test will take up to 10 minutes to deliver.

   Since the participants act as their own controls, a different set of 10 stories; 5 containing faux pas and 5 controls, will be used in the same manner as described above during the second visit.

2. Reading the mind in the eyes: This is a validated test which identifies subtle impairments in social intelligence in otherwise intelligent adults. In this test the participant is presented with a series of 18 photographs (printed on paper) of the eye-region of the face of different actors, and is asked to choose which one word describes best what the person in the photograph is feeling. The participant is given a choice of 4 words. The participant is provided with a glossary of all terms that they will come across during this test and is asked to read through the glossary and the meanings of any words that they may not be familiar with during the baseline phase of the study. The participant is also advised that this glossary remains available to them for reference throughout this test.

   Since participants act as their own controls, a different set of 18 photographs is shown during the second visit with the same instructions.

3. The awareness of social inference test (TASIT): This test is designed to assess the ability to "read" individual social cues and to assimilate them and make judgements about the speaker's intention, attitude, feeling, etc. The participant will be shown multiple short videos on a laptop, each lasting 15-60 seconds and then asked questions about the actor's intention, attitude and meaning. Each video will be played and will be followed by questions relating to the scene that has just been played in the video. The test contains 2 parts:

   1. Social inference (minimal) The participant is required to determine the speaker's meaning and intentions based on the dialogue, emotional expression and other paralinguistic clues of each of the performers in the video. The participant is tested on the ability to differentiate between a sincere exchange and a sarcastic exchange in order to look for any changes in social inference during the hypoglycaemic period.
   2. Social inference (enriched) This uses a series of vignettes, all of which contain a message that is literally untrue. Half of these represent sarcastic exchanges. However the other half use the same script as the sarcasm vignettes but in each one the speaker is lying, i.e., he/she intends the literal message to be accepted but it is contrary to what he/she believes.

Since each participant acts as his or her own control, a similar number of recorded vignettes will be played during the second visit with the same instructions.

Edinburgh Hypoglycaemia Symptom Scale: This validated questionnaire is used to assess the hypoglycaemic symptoms. This will be applied at baseline, at the start and end of the experimental phase.

General tests of cognitive function: The following tests of cognitive function have repeatedly been shown to be affected by hypoglycaemia and are quick to administer. They will therefore be used to confirm that the glucose nadir has been sufficient to impair cognitive function. They will be administered at the same time points as the Edinburgh Hypoglycaemia Symptom Scale.

Trail Making B. This test assesses visual conceptual and visual motor tracking. The subject is given a hand held tablet with a stylus and is presented with a grid containing randomly positioned numbers and letters, which are to be connected in alternating numerical and alphabetical order by them tapping with the stylus (ie 1 - A - 2 - B - 3 - C and so on.) The test score is the time taken to tap all the letters and numbers on the screen in the correct order.

Digit Symbol Test. This test measures sustained attention, response speed and visuo-motor co-ordination. Rows of blank squares are displayed on a piece of paper. Each blank square is paired with a number from one to nine. A printed key pairs each number with a different symbol, and the participant fills the blank squares with the symbols that match the numbers. The score is the number of successful codings within the 120 second time limit.

Choice reaction time. Subjects are presented with a laptop and are asked to press one of four keys as quickly as they can in response to the appearance of 4 'X' symbols on the computer screen.

Holter monitoring and ECG recording: In addition to looking for changes in social cognition, continuous ECG monitoring using a 3 lead ECG Holter monitor will be used to look at QT interval changes. The monitor will be attached to the participant at the start of the study and removed at the end of the 20 minute recovery phase.

In addition to Holter monitoring, 12-lead ECGs will be taken at baselines, 30 minutes into the experimental phase, at the end of the experimental phase and end of the clamp on both occasions. If there are any persisting QT changes at the end of the clamp, the participant will be invited to attend for a repeat ECG the next day to ensure that any ECG changes have resolved. There is a single case report of persisting QT interval prolongation for up to 48 hours after a period of hypoglycaemia in a 75 year old lady with end-stage renal disease in the context of acute myocardial ischaemia. Since we are recruiting from a younger population with no significant co-morbidities, we would not anticipate any lasting ECG changes.

The QT interval will be determined by the tangent or slope-intersect method. This will be an automated measurement produced by the Holter machine itself. If U waves are identified at any point, then a rhythm strip will be printed out and the QT interval determined manually by the researcher. This will be from the start of the QRS complex(represents the 'QRS complex' on the ECG which corresponds to the depolarization of the ventricles of the heart) till the nadir between the T and U wave. The QT interval will also be measured manually once (taken from the start of the QRS complex till the end of the T wave) on the baseline ECG. QTc (corrected QT interval) will be calculated using Bazett's formula which is calculated by dividing the QT interval with the square root of the RR interval. While there are a number of formulas to calculate QTc, there is no recommended standard formula for this. Bazett's formula is the most commonly used clinically. It is often criticised as it is thought to be inaccurate in extremes of heart rate, in particular tachycardia. However, since it is widely used in clinical practice to calculate QTc, and since we do not expect our subjects to exhibit extremes of bradycardia or tachycardia during the study as the blood glucose will not be allowed to fall below 2.5 mmol/L, we have chosen to use Bazett's formula.

Measurement of electrolytes: Blood samples for measurement of plasma electrolyte concentrations (potassium, magnesium, calcium) will be taken at baseline; at the beginning, middle (30 minutes and end (60 minutes) of the experimental phase; at the point when euglycaemia is restored and after euglycaemia has been maintained for 20 minutes. The blood samples will be spun down immediately after obtaining them.

Statistical Methods

The effects of hypoglycaemia on cognitive function will be assessed by repeated measures analysis of variance, with experimental condition as a within-subjects factor, and order of session (euglycaemia-hypoglycaemia or hypoglycaemia-euglycaemia) as a between-subjects factor. This means that performance on cognitive tests will be compared for each patient between their hypoglycaemia and euglycaemia study sessions. Statistical analysis will be performed using SPSS (Statistical Package for the Social Sciences). A p value of less that 0.05 will be considered significant.

Deterioration in test scores of 0.5 standard deviations or more between the euglycaemia and hypoglycaemia study sessions would be considered a clinically relevant change. With 20 subjects, the power of the study to detect a 0.5 standard deviation change in any test (assuming α=0.05, reliability of test=0.8) is 94.

Inclusion Criteria:

• Type 1 diabetes (insulin-treated) for at least two years duration.
• HbA1c values between 42 mmol/mol and 97 mmol/mol within the preceding 6 months.
• Aged 18-40 years.
• Body mass index 18-35 kg/m2.

Exclusion Criteria:

Participants will be excluded from the study if they meet any of the following criteria:

• Significant co-existent systemic disease or malignancy.
• Past history of severe hypoglycaemia.
• Past history of cardiovascular disease
• Past history of cerebral injury, seizure, chronic alcoholism or psychiatric disorder.
• Pregnancy. (Pregnancy testing will be carried out on all prospective female subjects).
• Impaired awareness of hypoglycaemia (defined as a score of 4 or more on the Gold scale)
• Use of drugs that affect heart rate and QT interval within the week before the study.

People with type 1 diabetes (i.e. insulin treated diabetes from the outset) will be recruited for this study.