Patients with type 2 diabetes have very variable endogenous insulin secretion. While some patients have relatively preserved endogenous insulin with marked insulin resistance others may develop the very severe insulin deficiency seen in type 1 diabetes. The impact of this variation on hypoglycaemia risk and treatment response in type 2 diabetes is unclear. This project aims to determine the impact of residual endogenous insulin secretion on glucose variability, hypoglycaemia risk and treatment response in insulin-treated participants with a clinical diagnosis of type 2 diabetes. The investigators will recruit participants from existing cohorts known to have severe insulin deficiency despite classical clinical characteristics of type 2 diabetes. The investigators will recruit other participants with insulin-treated type 2 diabetes and retained endogenous insulin secretion matched for glycemia and gender. The investigators will assess glucose variability (using continuous glucose monitoring system (CGMS)) and treatment response to a single dose of the glucose lowering therapy vildagliptin and compare responses between groups. This study will allow us to assess the potential utility of measuring endogenous insulin secretion in insulin-treated type 2 diabetes as a marker of hypoglycaemia risk and in determining likely response to oral therapy.

In insulin-treated type 2 diabetes the relationship between endogenous insulin levels and both glucose variability and hypoglycaemia risk is unclear. An association between endogenous insulin levels and hypoglycaemia risk was reported in the United Kingdom Hypoglycaemia Study, and those with low endogenous insulin levels have increased post-meal glucose rise (a measure of glucose variability and therefore hypoglycaemia risk). However, in another large cross-sectional study there was no correlation between endogenous insulin levels and questionnaire-assessed hypoglycaemia risk.

C-peptide as a potential biomarker for predicting treatment response to combination therapies in those with type 2 diabetes

Other glucose-lowering therapies may be combined with insulin in those with type 2 diabetes, as combined therapy in this condition may achieve better blood glucose control than insulin alone. In contrast non-insulin therapy is usually ineffective in type 1 diabetes, likely due to differences in endogenous insulin secretion. Among the commonly used glucose-lowering treatments which are licensed for combined therapy with insulin in type 2 diabetes are the dipeptide peptidase 4 (DPPIV) inhibitors. They have multiple mechanisms of action that include potentiation of beta cell insulin secretion and suppression of glucagon. The relative importance of these mechanisms to treatment response is unclear.

Recent work by the investigators group has identified that preserved endogenous insulin secretion is important for the glucose-lowering effect of a related class of glucose lowering therapies called GLP-1 receptor agonists, with little response seen in those with severe endogenous insulin deficiency. The investigators hypothesise that potentiation of endogenous insulin secretion may be the major mechanism of glucose lowering by DPPIV inhibitors and therefore those who have type 2 diabetes, but very low residual insulin secretion, will have little change in glucose with these medications. The pharmacodynamics of these agents (>90% inhibition of the target enzyme occurring within 45 minutes (for vildagliptin) and marked glucose lowering effect seen within an hour would allow us to explore this idea by assessing the impact of residual beta cell function on the glucose-lowering effect of a one-off dose of a DPP IV-inhibitor .

Hypotheses: The investigators hypothesize that patients who have the clinical characteristics of type 2 diabetes but have severe loss of endogenous insulin secretion will have markedly increased glycaemic variability (and therefore hypoglycaemia risk), and decreased response to non-insulin glucose lowering treatment in comparison to patients with similar clinical characteristics but preserved endogenous insulin.

Aim:

To determine whether individual residual beta cell function influences glucose variability, hypoglycaemia, and response to DPP-IV therapy in patients with insulin-treated type 2 diabetes

Research questions:

In participants with insulin-treated type 2 diabetes:

1. Does individual residual endogenous insulin secretion influence glucose variability and hypoglycaemia risk?
2. Does individual residual endogenous insulin secretion influence glycaemic response to DPPIV inhibitor therapy?

Study design (Overview):

The investigators will recruit participants who have a clinical diagnosis of type 2 diabetes, are currently treated with insulin therapy and are known to have preserved or very low endogenous insulin secretion. Participants with and without severe endogenous insulin deficiency will be matched for clinical characteristics. Glucose variability, hypoglycaemic risk, and response to a single dose of a standard glucose lowering treatment will be assessed and compared between the groups. This will require 3 research visits over a 2 week period (one non-fasting, two fasting).

For the response to treatment aspect of the study, glucose response to a mixed meal will be assessed with and without a one-off dose of the DPPIV inhibitor vildagliptin, with these tests performed in random order.

Participants and recruitment: The investigators will recruit 54 participants. Potential participants will be identified from existing research cohorts or routine clinical practice where endogenous insulin status (C-peptide in blood or urine) has been previously assessed. Where patients are identified from existing cohorts who have given consent for future contact, a letter from that database administrator will be sent providing generic information on the study. Contact details of those interested in further information will be passed to the research team to continue the recruitment process. Where potential participants are identified in clinical practice they will be given generic information about the study by their clinician and if interested, and with their permission, their contact details will be passed to a member of the research team for further contact. The research team will be responsible for the recruitment process.

Participants identified as having severe insulin deficiency (n=27) will be matched (for gender, HbA1c) with participants who have retained insulin secretion (n=27).

Visit one (baseline, research facility or participant's home) Participants will attend non-fasting. Following written informed consent, a blood sample (approximately 30mls in total) will be taken for baseline measurements (to include C-peptide, glucose, HbA1c, creatinine, liver function) and saved serum and plasma. A urine sample will be collected and stored and a pregnancy test offered to pre-menopausal women. Baseline data will include: height, weight, age, duration of diabetes, age of diagnosis and ethnicity, and confirmation of treatment regimen. A standard hypoglycaemia questionnaire (the Modified Clarke/Edinburgh Hypoglycaemia questionnaire, incorporating the Gold question) will be completed by the participant. The participant's ability to self-monitor blood glucose will be confirmed (with further education provided if necessary). Participants will be instructed on the use of the CGM system, and requirements for self-monitoring of blood glucose. Whilst wearing the continuous glucose monitors participants will be requested to test their own blood glucose levels 4 times spread over the course of each day (for subsequent calibration purposes) and record this on a data collection sheet provided. They will also be asked to record details of meals, exercise or symptoms over the duration of monitoring. The automatic device will be set up for 4-6 days continuous glucose monitoring.

Visit two (4-6 days after visit 1, research facility) Participants will be asked to attend for a morning visit having omitted their morning insulin and not consumed anything (except water) from midnight the night before. If they feel unwell, or need to treat low blood glucose the visit will be postponed. A small cannula will be inserted to enable multiple blood samples to be taken with the minimum of discomfort. A fasting blood test (approximately 15mls) will be performed for glucose, c-peptide, glucagon and saved serum/plasma. Participants will be asked to take their normal morning insulin with a 10% reduction immediately prior to consuming 250mls of a meal replacement drink (Fortisip) with or without a DPPIV inhibitor (Vildagliptin 50mg) tablet. Blood for glucose, c-peptide, glucagon and saved serum/plasma will be taken at 30, 60, 90, 120, 150 and 180 minutes after they have completed the drink (volume of blood at each time point approximately 15 mls). A urine sample will be collected and saved at 180 minutes, and then the CGM device will be disconnected and removed. Participant's bedside blood glucose will be measured, and a light lunch provided together with advice from study clinicians regarding insulin adjustments if required.

Visit three (4-14 days after visit 2, research facility) This will be as visit 2 with participants given the alternative treatment (DPPIV inhibitor or no treatment). Participants will be asked to keep glucose-lowering therapy identical between visits 2 and 3, to attend the research department at a similar time of day as visit 2 and to take evening insulin prior to visit 3 at the same time and dose as prior to visit 2.

Randomisation All participants will be required to undertake a mixed meal test with/without the standard DPPIV inhibitor. The order of the mixed meal tests at visits 2 and 3 will be decided using a randomisation table (StatsDirect), overseen by the CRF statistician. The randomisation is included in order to minimize any theoretical order bias effects (e.g. glucose levels improving through the study, or participant drop-out affecting the third visit more).

Participants and clinical research staff will not be blinded to this, however the statistician will be blinded to the group randomisation allocation for the purposes of analysis Concomitant Therapies

All diabetes related medication will continue unaltered. Participants will be asked not to change diabetes treatment for the duration of the study. The study team will manage diabetes medication for the duration of the participant's involvement in the study. The General Practitioner (GP) or secondary care diabetes team will manage other aspects of diabetes care in line with standard clinical practice.

Data Collection and Recording All participants will be assigned a unique study identifier (ID). All data collected will be recorded and stored under this ID number. Data will be initially recorded onto a study specific data collection form (DCF). Data will then be recorded onto a study specific database. Hard copies will be stored in the study specific site file. As part of the ECRF data quality procedures, data collected and recorded will be screened and reviewed for discrepancies and missing data prior to analysis.

Permission will be obtained to access medical notes should this be required for diabetes data relevant to taking part in the project.

Confidentiality:

All participant data will be held in a link-anonymised format, with personal identifiable data only accessible to personnel with training in data protection who require this information to perform their duties. Participants' research and sample data will be identified by unique study ID numbers and all data will be held on password-protected computers. Only the Chief Investigator (CI) or designated research team member will have access to personal identifiable data.

Data Storage and Archiving:

All consent and paper data collection forms will be scanned onto discs and stored in locked filing cabinets within the controlled access ECRF for the duration of the project. Paper copies will be shredded. Where permission has been obtained, samples and data will be transferred for safe keeping to the Exeter Tissue Bank (ETB) Where no such permission has been obtained at the end of the project the study documentation will be archived and stored following the standard R&D archiving procedures and biological samples will be destroyed following standard clinical guidelines for sample destruction.

Sample analysis and storage:

All saved serum and plasma samples will be stored under the study ID, with the file linking the study code to personal identifiable information held securely by the clinical research facility data manager and accessible only to personnel with training in data protection who require this information to perform their duties.

Analysis will be undertaken by the biochemistry laboratory of the Royal Devon and Exeter Hospital. HbA1c, lipid, renal and liver profile will be sent for immediate processing and results copied to participants clinicians (with participants consent). Other analysis will be performed batched samples at the end of the study, with samples stored at -80oc prior to analysis. Blood samples for DNA extraction will be processed and stored by the Royal Devon & Exeter National Health Service Foundation Trust Molecular Genetics Laboratory.

Consent will be sought to transfer samples remaining at the end of the study to the ethically approved Peninsula Research Bank (PRB) for safe storage for future use.

Power Calculation and statistical analysis The investigators will recruit 54 participants (27 each group). This will give us 80% power to detect an 0.8 Standard Deviation (SD) change in mean average glucose excursions (MAGE) (1.44 moll/L [35]), and an 18% absolute difference in glucose reduction with vildagliptin at the 5% level of significance between those with and without substantial endogenous insulin secretion.

The CRF statistician will be blinded to the group randomisation allocation for the purposes of analysis. The investigators will compare MAGE, time spent in hypoglycaemia (hours/day), and Clarke scores in those with and without severe insulin deficiency using unpaired T-tests. The investigators will compare response to DPPIV inhibitor (incremental area under the curve glucose with mixed meal test with vildagliptin - incremental area under the curve glucose test without vildagliptin) between groups using unpaired T tests. The investigators will assess whether each outcome variable is normally distributed prior to analysis and log transform or use equivalent non-parametric analysis where necessary.

Study Management:

The study will be managed by the National Institute of Health Research (NIHR) Exeter Clinical Research Facility. All contact with volunteers and sample collection will be facilitated within the R D&E NHS Foundation Trust setting.

Timescale:

It is anticipated the project objectives will be met over a period of 2 years

Budget Summary and Costing

All costs will be met by existing research funding as outlined below:

• Service support costs will be will be funded by the NIHR Exeter Clinical Research Facility core funding.
• All other research costs will be funded by Professor Andrew Hattersley's Wellcome Trust Senior Investigator Award

Project development and user involvement The study team will have access to the user representative group of the NIHR Exeter Clinical Research Facility (ECRF). In keeping with the National Health Service (NHS) Patient Carer and Public Involvement (PCPI) strategy the ECRF invites user representatives to contribute to the development of various projects within its portfolio. These individuals have agreed to maintain contact and regular meetings have been established at which researchers discuss the development of current projects within the ECRF.

Adverse Event Recording and Reporting This is a low risk study and it is not anticipated that participants involved in this project will be subject to adverse effects. The CI will be informed of any adverse effects within 24 hours and they will be reported following local NHS R&D SOPs with a copy of any adverse event form stored in the project site file.

Participant Feedback:

Should any biochemistry result potentially impact on clinical care, the results will be initially discussed by the CI and the clinical members of the research team and a decision will be made to contact that individual participant's General Practitioner or healthcare team. A statement to this effect is included in the consent form.

Dissemination/implementation of research Results will be written up and submitted for publication in a peer-reviewed journal. Abstracts will be submitted to national and international conferences. Results will be presented to clinical colleagues at regular in-house meetings. Written information in the form of a letter outlining the key findings of the study will be sent to all participants.

Potential impact and benefit of the proposed research:

In the short term the investigators hope to gain an improved understanding of the impact of residual beta cell function on hypoglycemic risk and treatment response in patients with type 2 diabetes. In the longer term the availability of a predictive biomarker for treatment response and hypoglycaemia risk would enable improved management of patients with insulin-treated diabetes.

End of Study The study will finish when data and sample collection has been completed on all participants, and analysis of data has been undertaken.

Inclusion Criteria:

Clinical diagnosis of type 2 diabetes Diagnosis of diabetes ≥ age 35 Treated with insulin Time from diagnosis to requiring insulin therapy of ≥36 months HbA1c ≥53mmol/mol (7%) and ≤100mmol/mol (11.3%)

Presence or absence of severe endogenous insulin deficiency:

• Severe deficiency = fasting blood C-peptide ≤0.08nmol/L or stimulated C-peptide ≤0.2nmol/L or post meal urine C-peptide creatinine ratio ≤0.2nmol/mmol
• Retained endogenous insulin secretion = fasting blood C-peptide ≥0.25nmol/L or stimulated C-peptide ≥0.6nmol/L or post meal urine C-peptide creatinine ratio ≥0.6nmol/mmol

Exclusion Criteria:

Pregnancy or breastfeeding Moderate renal impairment (EGFR<30 mls/min/1.73m2) Hepatic impairment (ALT >3 times upper limit of the normal range) Concurrent treatment with GLP-1 receptor analogue or DPPIV inhibitor glucose-lowering therapy.

Planned changes to glucose lowering therapy during the study duration Unable to self-monitor blood glucose Unable to give informed consent

access to data may be allowed via application to the study Chief Investigator