Slow recruitment and sufficient data available to analyse

It is well established that bariatric (weightloss) surgery affords considerable improvement in glycaemic control (control of blood sugar), and in many cases may lead to a complete resolution of type 2 diabetes. However, the mechanisms underlying these changes are yet to be elucidated and no research project to date has attempted to characterise changes in glycaemic control sooner than 3 days post surgery.

The primary objective of this study is to characterise changes in glycaemic control in individuals immediately following such surgery for a period of five days. Participants will be fitted with a continuous blood glucose measurement system (CGMS) prior to leaving theatre, which electronically records their blood glucose concentration every minute for up to five days. Upon returning the device each participant will undergo a standard meal test and have a small blood sample taken at 30 minute intervals (0-120 minutes) for the quantification of incretins (gut hormones involved in medium term control of blood sugar) insulin, glucose and appetite hormones.

These measurements will be compared to those collected at the baseline session, three weeks prior to the patient's surgery. Additional baseline visit measurements include: fasting lipid profile, insulin concentration, blood glucose concentration, HBA1C (long term blood glucose measurement), blood pressure, height, weight, waist circumference, and an oral glucose tolerance test (OGTT) and medical and family history.

All patients will be followed up twelve weeks post surgery, during which, all baseline measurements will be repeated.

The primary objective of the proposed study is to track glycaemic control in real-time using a state of the art continuous blood glucose monitoring system (CGMS) immediately after surgery to pin-point the exact timing of improvements in glycaemic control. The main secondary objective will be the association between change in glycaemic control and change in the levels of incretin hormones (GLP-1 and GIP).

Study Design

This is a prospective pilot study investigating the effects of bariatric surgery on measures of glycaemic control and incretin levels.

Aims

We aim to:

i) temporally characterise changes in glycaemic control in obese people undergoing bariatric surgery using a state of the art continuous blood glucose monitoring system (CGMS) ii) explore the putative relationship between incretins and improved glycaemic control iii) explore changes in subjective appetite using visual analogue scales (VAS) for appetite and palatability

Data Collection Baseline (V 1) Return of CGMS Day of surgery (V 2) + 8 weeks (V 3) Return of CGMS CGMS (5 day) √ √ √ Standard meal test plus associated fasting and postprandial bloods √ √ Appetite & palatability* (VAS) *5 √ √ (only appetite VAS) √ Exercise questionnaire IPAQ √ √ BP √ √ Waist, hip, weight and height √ √ Family & medical history √ Table 1: Data to be collected at each of the study time points. V = visit number. * visits 1 and 3 only for the palatability assessment i.e. conducted in conjunction with the standard meal test.

Visit 1: The baseline pre-operative visit (- 3 weeks) Informed written consent will be obtained either at the preoperative assessment visit or at an appointment at the Leicester Diabetes Centre scheduled before surgery. Informed consent will be followed by the baseline measures including fasting blood samples for analysis of lipid profile, insulin, HBA1c, and glucose levels. Blood pressure, body weight, height and waist circumference will be taken and recorded in the participants case report file. Each participant will complete two visual analogue scales, one for appetite and one for palatability. Medical and family history will additionally be recorded (table 1).

Patient's glycaemic status will be assessed by HbA1c at the time of referral. According to the 2011 World Health Organisation guidelines HbA1c can be used as a diagnostic test for diabetes, provided that stringent quality assurance tests are in place and assays are standardised to criteria aligned to the international reference values [20]. The advised HbA1c cut-point of ≥6.5% will be employed for this study.

Each participant will additionally undergo 5 days' CGMS detailed below. The participant will return to the either the LRI or diabetes department for removal of the CGMS where the data will be uploaded onto a password protected PC. Upon returning the monitor each participant will undergo a standard meal test followed by 30 minute blood sampling (0, 30, 60, 90 and 120 minutes) for the purpose of profiling GLP-1, GIP, PYY, leptin, ghrelin, insulin and glucose. At time-point 0 a sample will also be taken for analysis of vitamin C levels to provide an indication on the subject's diet i.e. level of fruit and vegetable intake. A sample will also be taken to measure vitamin D, a wide range of plasma proteins and metabolites will also be carried out on this sample using proteomic and metabolimic analysis.

If you participant prefers, or if timelines demand it due to the imminence of surgery, these visits will take place on the same day. The CGMS will be fitted and then the standard meal test will be conducted.

Visit 2: Post-operative day of surgery A fasting blood sample will be obtained on the morning of surgery for the quantification of glucose, insulin and a full lipid profile.

Immediately after surgery and before the participant leaves theatre CGMS recording will be repeated. CGMS will not be conducted during the patient's surgery due to the metal components of the glucose sensor potentially having diathermic properties i.e. electrically induced heating of the surrounding tissues which could lead to serious injury.

Visit 3: post-operative + 8weeks At 8 weeks post surgery the patient will be on a stable 900 - 1200 kcal diet i.e. as part of post-operative routine care from +8 weeks following surgery patients without complications reintroduce solids into their diets. At this point all baseline data including HbA1c will be collected and each participant will have 5 day CGMS fitted which will again be removed either at the LRI or the diabetes department by a member of the research team where the data will be uploaded to a password-protected PC. Again when returning the monitor each participant will undergo a standard meal test followed by 30 minute blood sampling (0, 30, 60, 90 and 120minutes) for the purpose of profiling GLP-1, GIP, PYY, Ghrelin, leptin, insulin and glucose. Again samples will be obtained at time-point 0 for the quantification of Vitamin C, vitamin D and plasma proteins and metabolites.

Standardising consumption pre-meal test. The gut hormones that are being investigated are exceptionally sensitive to caffeine, smoking and differential nutritional and caloric loads [30]. Therefore in order to standardise this test so that any changes observed between the visits are not compromised by the subjects prior eating habits we will ask them to abstain from smoking and consuming any caffeine up to 48 hours before the meal test. Furthermore, we will provide each subject with a meal replacement drink (Resource® 2.0 Fibre) to consume in place of their evening meal the night before their appointment and request that they abstain from ingesting anything more than water until after their appointment the following morning. This 200ml meal replacement contains calories 400 kcal and comprises 42.8g carbohydrate, 17.4g fat and 5g of fibre and therefore provides sufficient calorific and nutritional content for replacing a main evening meal.

Study Outcomes Primary Outcome The primary outcome is time taken for a glycaemic shift to be observed, as measured by CGMS.

The mean proportion (%) of time spent either above (≥10mmol/l), below (≤3.1mmol/l) or within (3.2 -9.9mmol/l) our predefined glycaemic ranges will be calculated for each participant over the recording period and adjusted for 24 hours. A glycaemic shift in this context is defined as a statistically (p<0.05) significant reduction or increase in the proportion of time spent within a glycaemic range i.e. shift from hyperglycaemic to normoglycaemic range.

It is argued that blood glucose levels of <3.9mmol/l can be defined as hypoglycaemic however a lower threshold of <3.1mmol/l has been selected for the hypoglycaemic cut point for this study. This is because although physiological responses are evident when levels are <3.9mmol/l, including a reduction in endogenous insulin and an increase in pancreatic glucagon secretion [21], actually blood glucose levels drop to <3.9 mmol/l in healthy individuals, without any clinically significant outcomes, particularly in women. Thus this definition potentially leads to an over estimation of clinically significant hypoglycaemic events. Blood glucose levels of <3.1 mmol/l result in autonomic (sweating or shaking) and neuroglycopenic symptoms (confusion, incoordination, speech difficulties) [22,23]. A number of studies have used this definition for determining hypoglycaemic events in people with diabetes mellitus [24-26].

The threshold of ≥ 10 mmol/l has been selected for the hyperglycaemic range because blood glucose levels at this magnitude are associated with T2DM complications thus the main objective in diabetes management is to keep blood glucose levels below this threshold with the aim to maintain blood glucose levels around 7 mmol/l in such patients.

Secondary Outcomes

There are a number of secondary outcomes for this study which are listed below:

1. The time taken for a change in the number of hyperglycaemic events (≥ 10.0 mmol/l for a duration of ≥ 10 minutes) to be observed (adjusted for 24 hours)
2. The time taken for a change in the duration of hyperglycaemic events (minutes) to be observed
3. The time taken for a change in the number of hypoglycaemic events (≤ 3.1 mmol/l for a duration of ≥ 10 minutes) to be observed (adjusted for 24 hours)
4. The time taken for a change in the duration of hypoglycaemic event (minutes) to be observed
5. The change in the area under the curve of the GLP-1, GIP, PYY, ghrelin, insulin, glucose and leptin profile between visit 1 and 2 3
6. The change in the fasting levels GIP between visit 1 - 2 and 2 - 3
7. The change in the fasting levels of insulin between visit 1 - 2 and 2 - 3
8. The change in Hba1c between visit 1 and 3
9. The change in fasting glucose levels between visit 1 and 3
10. The change in mean amplitude of glycaemic excursions (MAGE) between visit 1 and 3
11. The change in subjective appetite between visit 1 - 3 and over the course of the hospital stay
12. The change in subjective palatability of meal test between visit 1 - 3
13. The change in levels of vitamin C between visit 1 - 3
14. The change in levels of vitamin D between visit 1 - 3
15. The change in plasma proteins between visit 1, 2 and 3
16. The change in plasma metabolites between visit 1,2 and 3

NB: 1 - 4 will be calculated as change between visit 1 - 2, visit 1 - 3 and between 2 - 3.

HbA1c test will be conducted on visit 1 and visit 3 to confirm presence/remission of pre- or overt diabetes. We will additionally collect standard demographic data such as ethnicity, age, family history of T2DM and CHD/CVD, BMI, body weight, waist and hip circumference and systolic and diastolic blood pressure.

We will seek consent from each participant to access their medical notes to record any data that may be relevant to this study.

Sample Size We aim to recruit 20 obese patients who are awaiting either a gastric bypass or sleeve gastrectomy at the Leicester Royal Infirmary. This is a pilot study and therefore does not require a formal power calculation. However, a very recently published trial assessed glycaemic control in 10 obese individuals with T2DM 7 days after surgery using a standard liquid meal followed by blood sampling at 0, 30,60,90 and 120 minutes [19]. This group were able to detect improvements in glycaemic control despite the small sample size. With 90 surgeries taking place a the Leicester Royal Infirmary in 2010 and with approximately 65 currently on the waiting list we feel it achievable and relevant in light of existing evidence to recruit a target number of 20 patients.

Statistical Analysis The analysis will be of a descriptive nature, we will comprehensively phenotype the sample rather than comparing groups. Therefore, statistical tests will not be performed. For the primary outcome the mean time to glycaemic shift with 95% confidence intervals will be calculated. For secondary outcomes each participant will act as their own control and data will be analysed using the change from baseline. Results will be reported as mean (%% CI) for continuous measures or count (percentages) for categorical. For hypothesis generation the analysis will be repeated by type of surgery.

Standard operating procedures We will collect all study measures in line with the Leicester Diabetes Research Group standard operating procedures.

Inclusion Criteria:

• 18 years of age or older
• No active psychotic illness
• On the waiting list for bariatric surgery at the Leicester Royal Infirmary and thus meeting the local eligibility criteria for this procedure

Exclusion Criteria:

• < 18 years of age
• Pregnant
• Active psychotic illness
• Receiving either GLP1 analogue or DPPIV inhibitor therapy
• History of dug or alcohol dependancy
• History of poorly controlled/severe mental health problems
• Presence of any comorbidities contraindicative of abdominal surgery or anaesthesia (American Society of Anaesthesiology grade 4, certain grade 3)

20 successive obese patients who are awaiting either a gastric bypass or sleeve gastrectomy