Effects of hypoglycaemia on working memory and regional cerebral blood flow in type 1 diabetes: a randomised, crossover trial

Aims/hypothesis The aim of this randomised, crossover trial was to compare cognitive functioning and associated brain activation patterns during hypoglycaemia (plasma glucose [PG] just below 3.1 mmol/l) and euglycaemia in individuals with type 1 diabetes mellitus. Methods In this patient-blinded, crossover study, 26 participants with type 1 diabetes mellitus attended two randomised experimental visits: one hypoglycaemic clamp (PG 2.8 ± 0.2 mmol/l, approximate duration 55 min) and one euglycaemic clamp (PG 5.5 mmol/l ± 10%). PG levels were maintained by hyperinsulinaemic glucose clamping. Cognitive functioning was assessed during hypoglycaemia and euglycaemia conditions using a modified version of the digit symbol substitution test (mDSST) and control DSST (cDSST). Simultaneously, regional cerebral blood flow (rCBF) was measured in pre-specified brain regions by six H2 15O-positron emission tomographies (PET) per session. Results Working memory was impaired during hypoglycaemia as indicated by a statistically significantly lower mDSST score (estimated treatment difference [ETD] −0.63 [95% CI −1.13, −0.14], p = 0.014) and a statistically significantly longer response time (ETD 2.86 s [7%] [95% CI 0.67, 5.05], p = 0.013) compared with euglycaemia. During hypoglycaemia, mDSST task performance was associated with increased activity in the frontal lobe regions, superior parietal lobe and thalamus, and decreased activity in the temporal lobe regions (p < 0.05). Working memory activation (mDSST − cDSST) statistically significantly increased blood flow in the striatum during hypoglycaemia (ETD 0.0374% [95% CI 0.0157, 0.0590], p = 0.002). Conclusions/interpretation During hypoglycaemia (mean PG 2.9 mmol/l), working memory performance was impaired. Altered performance was associated with significantly increased blood flow in the striatum, a part of the basal ganglia implicated in regulating motor functions, memory, language and emotion. Trial registration NCT01789593, clinicaltrials.gov Funding This study was funded by Novo Nordisk. Electronic supplementary material The online version of this article (10.1007/s00125-017-4502-1) contains peer-reviewed but unedited supplementary material, which is available to authorised users.

seconds. Activity decayed for 10 minutes (5 half-lives) before each new tomography session. To ensure that the participants were fully engaged in the cognitive task during acquisition of the PET data, a new cognitive session was initiated (excluding instructions) before the tracer injection. Radioactivity in radial arterial blood samples was recorded with an automatic counter (Allogg, Sweden) with a frequency of once per second during the imaging time of 3 minutes. All scans were performed between 9:00 and 12:00 hours to account for diurnal variation in cerebral blood flow. The participants were imaged in the supine position with their heads fixed in place. Head position was checked prior to each scan using laser marking.
Each participant underwent high-resolution MRI (GE Signa Excite HDx 3T; threedimensional (3D) fast spoiled grass sequence with IR preparation: echo time=3.0 ms, Inversion time=450 ms, flip angle=20°, slices = 120, slice thickness = 1.5 mm, field of view (FOV) = 240 mm, matrix = 2562) and PET (3D mode: slices = 47, slice thickness = 3.1 mm, FOV = 256 mm, matrix = 1282) imaging. PET images were reconstructed using filtered back-projection with a 0.5 cycles-1 ramp filter, followed by application of a 7-mm Gaussian filter. The regional cerebral blood flow (rCBF) quantification used a 500-MBq bolus injection of H2 15 O delivered intravenously. The CBF data were acquired in 21 frames (12 × 5 seconds, 6 × 10 seconds, 3 × 20 seconds) over 3 minutes and modelled using single tissue compartment with added vascular space, according to established procedures. Parameter estimation was performed in each participants' native PET space. A transformation between native PET and MRI spaces was calculated for each participant using a six-parameter rigidbody transformation with a mutual information cost function. MRI volumes were finally mapped to a common reference space (MNI 1 mm) with linear and non-linear registrations in BioImage Suite (www.bioimagesuite.com). In addition, MRI images were skull-stripped and segmented into grey/white/cerebral spinal fluid (CSF) space using FSL BET and FAST (fsl.fmrib.ox.ac.uk). Once in common space, averaged PET maps were computed for different parameters in Matlab (MathWorks, Natick, MA) using custom scripts. While alternative techniques, such as blood oxygenation level-dependent contrast and fMRI, have better spatial and temporal resolution compared to PET H2 15 O (as well as no radiation exposure), PET offers the advantages of being quantifiable, less confounded by movement, and allows for serial studies, making this methodology subtler to the scope of the study.

ESM Results
To explore the impact of the perception of hypoglycaemia during the glycaemic clamps (responding yes or no to, 'do you feel hypo') on the primary endpoint during the state of hypoglycaemia, we investigated the differences in mDSST score and

Inclusion criteria:
Eligible participants had to fulfil all the criteria below. 1 Informed consent obtained before any trial-related activities. Trial-related activities are any procedures that are carried out as part of the trial, including activities to determine suitability for the trial. 2 Male or female right-handed participants aged 18-64 years (both inclusive). 3 Type 1 diabetes mellitus (as diagnosed clinically) ≥ 12 months. 4 Treated with multiple daily insulin injections or continuous subcutaneous insulin infusion (CSII) ≥ 12 months. 5 Body mass index (BMI) 18.0-28.0 kg/m 2 (both inclusive). 6 Glycosylated haemoglobin (HbA1C) ≤ 9.0 % by local laboratory analysis. 7 Fasting C-peptide ≤ 0.2 nmol/L. Exclusion criteria: Eligible participants were not allowed to meet any of the exclusion criteria below 1 Previous participation in this trial. Participation is defined as screened. 2 Receipt of any investigational medicinal product within 3 month prior to screening in this trial. 3 Participants that have been exposed to ionising radiation or isotopes as part of diagnostic or experimental investigations within the past 5 years. 4 Haemoglobin < 8.0 mmol/L (male) or < 6.4 mmol/L (female), total leukocyte count < 3.0 × 10 9 /L, thrombocytes <100 × 10 9 /L, plasma creatinine levels ≥ 126 μmol/L (male) or ≥ 111 μmol/L (female), bilirubin > 3 × the upper limit of normal (ULN), alanine aminotransferase (ALAT) and alkaline phosphatase (ALP) > 2 × ULN. 5 Cardiac problems defined as decompensated heart failure (New York Heart Association [NYHA] class III and IV) at any time and/or angina pectoris within the last 12 months and/or acute myocardial infarction at any time. 6 Supine blood pressure at screening (after resting for 5 minutes) outside the range 90−140 mmHg for systolic or 50−90 mmHg for diastolic. This exclusion criterion also pertains to participants taking antihypertensives. 7 Clinically significant abnormal ECG at screening, as judged by the investigator. 8 Visual impairment (i.e. cataract, colour blindness etc.) or auditory impairment. 9 Known abnormalities of the central nervous system (CNS) or any endocrinological (with the exception of diabetes mellitus and euthyroid struma), haematological, neurological, psychiatric diseases or other major disorders that in the opinion of the investigator precludes compliance with the protocol, evaluation of the results or represent an unacceptable risk for the participant's safety. 10 Proliferative retinopathy (funduscopy/fundus photography performed within 12 weeks before the screening visit is acceptable) and/or severe neuropathy. 11 Contraindication to PET or magnetic resonance imaging (MRI) scanning (e.g. suffer from claustrophobia, some metallic implants etc.) 12 Structural brain abnormalities as evident from MRI scanning. 13 Participant who has donated any blood or plasma in the past month or more than 500 mL within 3 months prior to screening. 14 Surgery or trauma with significant blood loss (more than 500 mL) within the last 3 months prior to screening. 15 Current treatment with systemic (oral, intravenously (iv). or inhaled) corticosteroids, monoamine oxidase (MAO) inhibitors, prostaglandin blockers, systemic non-selective beta-blockers, growth hormone, and other drugs, which may interfere with glucose metabolism. 16 Not able or willing to refrain from any use of herbal products and non-routine vitamins within 1 week and routine vitamins within 48 hours prior to start of the experimental visit. 17 Significant history of alcoholism or drug/chemical abuse as per investigator's judgement or a positive result in the urine drug/alcohol screen at the screening visit. 18 Current tobacco user (any smoking or use of nicotinic products within 3 months prior to screening). 19 Severe hypoglycaemic event during the past 6 months or hospitalisation for diabetic ketoacidosis during the previous 6 months. 20 Hypoglycaemic unawareness. 21 Participant with mental incapacity or language barriers precluding adequate understanding or co-operation or who, in the opinion of the investigator or their general practitioner, should not participate in the trial. 22 Female of childbearing potential (menopause: at least 12 months without menstrual cycle) who is pregnant, breast-feeding or intend to become pregnant or are not using adequate contraceptive methods (adequate contraceptive measures include sterilisation, hormonal intrauterine devices, oral contraceptives, spiral, transdermal depot plaster and depot injection). 23 Any chronic disorder or severe disease which, in the opinion of the investigator might jeopardise participant's safety or compliance with the protocol. Experimental visit exclusion criteria 1 Consumption of alcohol within 24 hours prior to Day 1 of an experimental visit, or positive result of alcohol breath test. 2 Consumption of coffee, tea, chocolate or beverages such as cola and energy drinks containing methylxanthine (caffeine, theophylline or theobromine) within 12 hours prior to Day 1 of an experimental visit. 3 Physical exercise within 24 hours prior to Day 1 of an experimental visit, as judged by the investigator to interfere with trial results. 4 Presence of any medical condition that may confound the results of the trial or pose an unacceptable risk to the participant, as judged by the investigator. 5 Any use of systemic (oral, i.v. or inhaled) corticosteroids, MAO inhibitors, prostaglandin blockers, non-selective beta-blockers, growth hormone and other drugs, which may interfere with glucose metabolism. 6 Any use of herbal products and non-routine vitamins within 1 week prior to Day 1 of an experimental visit and routine vitamins within 48 hours prior to Day 1 of an experimental visit. 7 Hypoglycaemia (PG ≤ 3.9 mmol/L) within 48 hours prior to Day 1 of an experimental visit. 8 Injection of IDeg after 21:00 the evening before the experimental visit Day 1 and/or injection of long or intermediate-acting insulin products after 10:00 at the experimental visit Day 1 and/or injection or bolus infusion of insulin after 19:00 at the experimental visit Day 1. (Note: Participants using continuous subcutaneous insulin infusion (CSII) may arrive at the trial site with their basal rate running). Table 2. Individual mDSST scores, response times and baseline characteristics for participants with outlier values for mDSST score and/or response time.  ETD in rCBF by region of interest normalised to cerebral cortex. Completers: Participants in the full analysis set with available endpoints at both experimental visits. cDSST, control digit symbol substitution test; CI, confidence interval; ETD, estimated treatment difference; mDSST, modified digit symbol substitution test; rCBF, regional cerebral blood flow ESM Fig. 1. Trial design.
In the cDSST a, participants were to click with their right hand if they were presented with the letter H ('Højre' is Danish for right) with a symbol, or click with their left hand if they were presented the letter V ('Venstre' is Danish for left) with a symbol. During the mDSST b, participants were first shown a white fixation cross on a black screen for 1 second, after which they were shown a target digit-symbol combination for 1.2 s. Next, a white fixation cross was shown on a black background for 200 ms, after which a block of four digit-symbol combinations was shown for a maximum of 3.6 s. If the target digitsymbol combination was present among the four other combinations, participants had to click with