Diabetes UK Professional Conference 2013

March 13-15, 2013: Manchester, England Day #3 – Draft

Executive Highlights

The final day of Diabetes UK, Day #3, was some good learning mixed in with some good review. The biggest news of the day was a new trial in the works to assess whether metformin is cancer protective. The trial will be nearly 12,000 people with prediabetes and will take six years. Dr. Holman said that in this trial, everyone will get diet and exercise advice, and can take any agent. They will be treated to best usual practice. With this in mind, presumably the GLP-1 and pancreatitis issue may also be assessed (we are checking on this). 

More good learning came in discussing CVD and type 1 diabetes, if not reassuring commentary. The Diabetes UK cohort consists of nearly 24,000 patients with type 1 diabetes diagnosed at less than 30 years of age in the years 1972-1993. The cohort was followed out to 2000. The relative risk of death (compared to the general population) was 4.0 for females and 2.7 for men. After 13.4 years follow up, there were 949 deaths, which seems pretty high! Said Dr. Miles Fisher (University of Glasgow, Scotland), in discussing risk factors, “We can do risk reduction with statins, weight, blood pressure. But we have a laissez faire approach at the moment.” Not so reassuring from a patient perspective. 

On the GLP-1 front, Dr. Stephen Gough talked about whether GLP-1 is cardioprotective – there are currently eight trials investigating cardiovascular outcomes with GLP-1 and he emphasized that some valuable results should certainly emerge over the next few years. There is some evidence, he said, that GLP-1 causes vasodilation in the endothelium and increases LV ejection fraction and he noted that there are GLP-1 receptors in the heart. In animal models, moreover, there are reduced aortic lesions. In a randomized controlled trial of n=387 patients treated for their first heart attack, exenatide improved the myocardial salvage index by 15-20%. We believe the interest in the CVD outcomes will grow larger over time; we just hope that trials don’t have to be delayed due to lower than expected events. 

On the new therapy front, Dr. Clifford Bailey talked about new drugs – his discussion was more than theoretical, since Lyxumia, Forxiga and Tresiba have all been launched in the UK within the last few weeks. Dr. James Shaw discussed prevention of recurrent severe hypoglycaemia in type 1, showing that people with consistent hypoglycemia (53 minutes/day at baseline) were able to reduce severe hypoglycemia by 90% after education with a tool called My Hypo Compass. 

Finally, Dr. Rury Holman, the legendary trialist, turned up at Diabetes UK ostensibly to discuss  metformin and cancer, but like the magician he is, he pulled a gleaming new trial (called GLINT) out of his hat. There is controversy because although observational trials strongly suggest that metformin has a protective effect against cancer, a meta-analysis of randomized controlled trials shows precisely zero effect, said Dr. Holman. A new trial is required to lay the matter to rest. The GLINT trial takes people with prediabetes, randomized to metformin or placebo, and tracks cardiovascular outcomes and cancer. The trial will be powered for cardiovascular endpoints, but cancer rates are expected to be higher, making this feasible. GLINT will recruit nearly 12,000 people and take six years.



Detailed Discussion and Commentary

Hot Topics 1


Melanie J Davies MB ChB, MRCP, MD, FRCP (University Hospitals of Leicester, UK) 

The recent ORIGIN trial was a large and ambitious trial, investigating the effect of early insulin on cardiovascular risk. The trial showed that early insulin resulted in stable glucose control with good safety, and the result was neutral with respect to cardiovascular outcomes and cancer. Participants with prediabetes had a 20% reduction in new onset diabetes with insulin, but it seems likely that standards of care for diabetes and prediabetes won’t change as a result of ORIGIN. 

  • The ORIGIN trial, designed over a decade ago, was intended to answer the question: “in high risk patients, does early insulin replacement therapy targeting normal FPG reduce cardiovascular risk?” The trial was international (40 countries), and included ~12,500 people with over six years follow up. Most participants had type 2 diabetes, but some had impaired glucose tolerance (IGT). There were two groups – early insulin intervention and standard care. The intervention group was treated with insulin glargine (Lantus, Sanofi) with fasting plasma glucose titrated to <95mg/dl. 
  • The ORIGIN study showed no benefit in cardiovascular events or mortality – it was a completely neutral trial. However, there was a 20% reduction in new onset diabetes in patients with prediabetes. There was no real signal in secondary outcomes either – early insulin showed no benefit, but also no harm. The intervention group lowered A1c quickly and maintained it for seven years. Insulin dose was also steady 0.3-0.4 U/kg. Adherence to glargine was 84% after seven years. Rates of hypoglycemia were low in comparison to other major trials, but nonetheless represented real risks.
  • The implication of ORIGIN is that we should not delay insulin, but it’s unlikely that we will use it for prevention. Dr. Davies emphasized that the field won’t be changing standards of care either for diabetes or prediabetes. 
  • In ORIGIN there was no increase in overall cancer, cancer mortality or any particular cancer in the intervention group. In recent years, there had been some controversy regarding insulin and cancer. Dr. Davies commented that the cancer debate ‘certainly been informed and I would think offers some good reassurance’. The use of metformin was slightly lower in the glargine group (47% versus 60%) and some commentators are concerned that metformin has blunted the effect of cancer, that the insulin dose was relatively low or that the study was too short for carcinogenesis. Dr. Davies did not seem to be one of those people. 
  • For more details on ORIGIN, see our ADA 2012 piece at  https://closeconcerns.box.com/s/f3fec2377804f65c834c. 


Questions and Answers 

Q: What are the implications for diabetes prevention? 

A: There was a 20% risk reduction, but there was a 2 kg (4.4 pound) weight difference with insulin and an increased rate of hypoglycemia. So this is a proof of concept and nice data, but in clinical practice we are unlikely to use insulin to prevent diabetes.

Q: I am wondering if the wrong group is there. If they have cardiovascular disease and don’t have diabetes. maybe they are not the people we should be looking at - since diabetes is probably not the cause of their heart disease. 

A: 80% did have diabetes. 11% were recruited with IGR. It would have taken tens if not hundreds of 
thousands of participants to power the study if we recruited a lower risk group… but point is well made. 


Miles B Fisher MD, (University of Glasgow, Scotland) 

The relative risk for cardiovascular disease in people with type 1 diabetes has declined relative to earlier studies, but is still very high, particularly in younger women. Real world data shows that glycemic control is sub-optimal in many patients. There is an urgent need to control risk factors – particularly the use of statins. 

  • The Diabetes UK cohort consists of nearly 24,000 patients with type 1 diabetes diagnosed at less than 30 years of age in the years 1972-1993. The cohort was followed out to 2000. 
  • The relative risk of death (compared to the general population) was 4.0 for females and 2.7 for men. That’s high. It was even 5.7 for females aged 20-29 years, although the absolute risk is lower. Below 30 (at age of diagnosis), the leading causes of death were related to metabolic complications (hypoglycemia and DKA) and above 30 they were related to cardiovascular disease. After 13.4 years follow up, there were 949 deaths. However, risk factors for ischemic heart disease were considerably higher in younger people than in the general population, leading to the need to treat coronary risk factors in young patients. 
  • In a study from the Scottish Diabetes Research Network (SDRN), the risk of cardiovascular disease for people with type 1 diabetes is roughly double the general population at most age groups. There is not too much difference between males and females – males have slightly higher risk at older ages and lower risk when younger. The SDRN is a very powerful resource for epidemiological research, since 99.5% of patients in Scotland are in a single diabetes database. This study considered roughly 19,000 people with type 1 diabetes over 20 years of age between January 2005 and May 2008. Again, the most common cause of death for those under 40 years was diabetes related (hypoglycemia or DKA) and was cardiovascular disease for those over 40. 
  • In the SDRN study, 37% of patients over 40 had moderately high or high blood pressure, median total cholesterol was 175 mg/dl, and 39% were not on a statin. A1c was 8.4% for males and 8.5% for females and only 13% had A1c <7.0% (while 37% had A1c >9%). Smoking and obesity were similar to general population. These struck us as quite depressing statistics. 

Questions and Answers 

Q: In other work, people who remain non-albuminuric over 20 years have same risks as the general population. It’s only those showing signs of microvascular disease that have higher risks.

A: Yes, we need to make this data available for our database. 

Q: You call for attacking risk factors. But we need to model how much it would help in the real world. The presentation of cardiovascular disease is death (not angina). The reduction is really intervention, not prevention. We need to intervene earlier. 

A: That’s another hour lecture. We can do risk reduction with statins, weight, blood pressure. But we have a laissez faire approach at the moment. 


Mike E. J. Lean, MB, BChir, MD, MRCP, FRCPS, FRCP (University of Glasgow, Scotland) 

Dr. Lean, an expert on physical activity, nutrition and diabetes prevention, gave an outsider’s  perspective on the Look AHEAD trial. The Look AHEAD trial was designed to investigate the effect of  intensive lifestyle intervention and weight reduction on cardiovascular outcomes. The intervention was halted at end of 2012, because there was no difference in cardiovascular events. At that point, the intervention group had achieved about a 5% weight reduction compared to the control group, although this gap had been steadily declining over time. Dr. Lean suggested that the trial was not necessarily a failure, and didn’t prove that diet and exercise doesn’t work. However, he said that the trial asked too many questions, the intervention was unsustainable, and in any case he felt that using a randomized controlled trial to develop an effective service is not a good idea. 

  • The Look AHEAD trial was possibly the most expensive trial ever conducted in the USA for any purpose. It was designed to investigate the effect of intensive lifestyle intervention (designed to reduce weight) on cardiovascular outcomes. The trial included over 5,000 overweight people (BMI ~36 kg/m2) with type 2 diabetes. It was powered to find an 18% difference in the composite outcome over a five-year period. The intervention goal was 10% weight loss, obtained via diet and physical activity, encouraged via constant exposure to coaching via group sessions (once a week at the beginning of the trial). Participants followed a low fat diet (<30%) including ready meals and meal replacement. 
  • The intervention was halted at end of 2012, because there was no difference in cardiovascular events, and event rates were dramatically smaller than anticipated (only a quarter of the expected rate). The participants are still being followed. 
  • Weight loss was 8.6% after the first year, in the intervention group, but then steadily dropped to about 6.2% after four years. There was little difference in measured fitness, A1c, blood pressure, or lipids between the groups. At one year, A1c reduction was 0.7% in the intervention group, but this dwindled to almost nothing after four years, with about half of both groups achieving A1c <7% at this time. 
  • Look AHEAD was portrayed by the media as ‘diet and exercise doesn’t work’. This is not true. The trial didn’t find an impact on cardiovascular disease, but there are proven health benefits of weight loss and we have a rich data set. For example, diabetes reversal is significantly improved with weight loss. 16% of patients with >6.5% weight loss experienced more than one year remission of type 2 diabetes. 
  • Dr. Lean made three main observations about the value of the Look AHEAD trial: 
    • In hindsight, the trial asked too many questions simultaneously, such as: Does weight loss reduce cardiovascular risk? What intervention is the best? How can it best be implemented? It’s hard to generalize the result. 
    • Furthermore, the intervention appeared to be unsustainable. Also, the control group experienced good benefits, implying that they were not the routine patients who don’t turn up for clinics. 
    • Medicine is the only field that attempts to develop an effective service using randomized controlled trials. In the world of business, for example, the best technique is a Continuous Improvement Methodology (feedback, evaluate, improve). 

Questions and Answers 

Q: Recent data on the Mediterranean diet shows a 30% reduction in cardiovascular disease. Weight loss is not mentioned. Shouldn’t the focus be on the quality of the diet, rather than weight loss or physical activity…? 

A: Well, I never say anything against physical activity. Achieving weight loss in people with type 2 diabetes is the dominant factor in remission. Some elements of the Mediterranean diet would be difficult to replicate in the USA e.g. the consumption of large amount of fish. 

Q: In improving lifestyle we should first tackle diet and then bring in physical activity. Look AHEAD did lots of things at once that people couldn’t sustain. Shouldn’t we do things step wise? 

A: But it’s hard to intervene twice when a patient’s lifestyle is upended the first time. We need to hit 
weight very hard at the very beginning of the disease, to get out of the diabetes category. But by and large, studies show that a “big bang” works less well than a step-wise approach… 

Q: So is lifestyle a waste of time? 

A: Lifestyle is really important, and it’s really important to enjoy it. A fun environment is better than being bullied. 



Dean Pooley (Diabetes Development Market Manager, Roche Diagnostics, Brighton, UK) 

Mr. Pooley presented some sobering market research work regarding driving habits of people with diabetes. Although his methodology was gently criticized for bias during the Q&A session, the work was very intriguing, because it suggests that people with diabetes are not paying enough attention to their driving safety, nor are aware of the rules relating to driving with diabetes. Undoubtedly, driving safety issues due to hypoglycemia are being under-reported. 

  • In the UK, the DVLA (Driving and Vehicle Licensing Authority) has strict rules for driving with diabetes. The DVLA must be informed of one of more episode of severe hypoglycemia in a 12-month period, and licenses can be revoked. 
  • Patients must check their blood glucose before driving, and test every 2 hours on the road. They should not drive if blood glucose is <4 mmol/l (72 mg/dl). If blood glucose is 4-5mmol/l (72-90mg/dl) they should take a snack and wait before driving. 
  • Roche Diagnostics surveyed ~16,000 drivers who had diabetes and found that 34% were not fully aware of their statutory requirements, and 44% were not full aware of the dangers of hypoglycemia. Roche sent 168,000 survey invitations to the Accu-Chek database, preselected based on propensity for hypoglycemia. Response rate was around 9.5%. Of the drivers that used insulin, 26% were not fully aware of their statutory requirements and 33% were not fully aware of the dangers of hypoglycemia. 
  • According to survey, 53% of insulin users don’t always test before driving, and 59% of insulin users don’t always take a break to test after two hours of driving. 22% of all respondents said they never test before driving. 
  • 13% of insulin users reported having a hypoglycemic event behind the wheel, and 1.3% of insulin users reported that they had had an accident due to hypoglycemia while driving. Of the group reporting an accident, 74% have hypos at least monthly and 94% were insulin users. 21% of all respondents said they have regular hypoglycemic episodes. 

Questions and Answers 

Q: You can’t say that doctors and nurses have to tell patients to do better. This doesn’t work. 

A: I agree. We send our patients literature regularly, and ret and educate them about responsible driving, among other things. 

Q: How many tests and car trips in your sample per month? Was strip availability an issue? 
A: This was a quick and dirty study, only 17 questions and we didn’t get to that. 

Q: Is a blood glucose test at breakfast time 30 minutes before driving acceptable? 
A: Yes, that would be acceptable. The rules say ‘before driving’ but they are not specific and you are eating. 

Comment: This is not acceptable. The issue is clarity. The rules say before driving (implied that it should be immediately before). 

Comment: I suspect that only the people who tend to test before driving have responded to your survey. I wonder if anyone ever tests? Only a very few people. Non-response bias demands that we do a randomized trial. 


Hot Topics 2


Stephen Gough, MD FRCP (University of Oxford, UK) 

Dr Gough presented a tidal wave of information in very short order – reviewing the literature and current state of play with GLP-1. He noted that GLP-1 has many beneficial effects, as well as A1c lowering and weight loss. In particular, GLP-1 may be cardioprotective, it may have positive CNS  effects over and above satiety/appetite, and it improves liver function (indirectly). In response to recent news from the FDA, Dr. Gough also considered the case for GLP-1 causing pancreatitis, which he felt was not at all conclusive. A number of randomized controlled trials in process should shed light on the topic. 

  • Pancreatic effects are the best-known aspects of GLP-1, but it also works in the liver, brain and cardiovascular system. There is growing evidence that GLP-1 in the brain starts a neuroendocrine cascade that in turn affects other organs. 
  • Commercially available GLP-1 agonists are exenatide (Byetta/Bydureon, Amylin), liraglutide (Victoza, Novo Nordisk), lixisenatide (Lyxumia, Sanofi). Exenatide twice daily and lixisenatide are shorter acting, and exenatide once weekly and liraglutide are longer acting. Waiting in the wings are albiglutide (Eperzan, GlaxSmithKline), dulaglutide (Eli Lilly), and semaglutide (Novo Nordisk). 
  •  Head to head studies show some differences between the compounds, but all show good A1c lowering, weight loss, plus no increase in hypoglycemia. Head to head studies include LEAD-6, DURATION-6 and HARMONY-7. 
  • The most exciting future use in type 2 diabetes (and possibly in type 1 diabetes) is the combination of GLP-1 agonists with insulin. This allows lower A1c with no increased risk of hypoglycemia and less weight gain. 
  • GLP-1 may be cardioprotective – there are currently eight trials investigating cardiovascular outcomes with GLP-1 and we will get some good results over the next few years. There is some evidence that GLP-1 causes vasodilation in the endothelium and increases LV ejection fraction. There are GLP-1 receptors in the heart. In animal models there are reduced aortic lesions. In a randomized controlled trial of n=387 patients treated for their first heart attack, exenatide improved the myocardial salvage index by 15-20%. 
  • GLP-1 has CNS effects such as satiety and lower appetite. It may potentially have other effects - it also protects the brain against ischemic injury and prevents degenerative processes in a mouse model of Alzheimer’s disease. GLP-1 also decreases brain glucose during hyperglycemia, which is considered to be neuroprotective. 
  • Liver function effects have been reported with GLP-1, but they may not be an independent effect. A meta-analysis of the LEAD program showed that ALT (alanine transaminase) clearly improves with the use of liraglutide, but the effect is not significant after correcting for weight and A1c. So we don’t have data to support an independent effect, although it is a treatment effect of GLP-1. 
  • GLP-1 is being investigated for links to pancreatitis. The evidence is not at all clear, but a can of worms has been opened. In February 2013, a JAMA article showed an increase in risk of acute pancreatitis with GLP-1. But the patients who were taking GLP-1 also had known risk factors for pancreatitis. In contrast, a recent study of 1 million patients in a US medical claims database found no associations at all. This week, the FDA stated that it is investigating reports of increased pancreatitis, and Dr. Gough believes that the statement is in response to the JAMA article, although the FDA cites unpublished findings. The ADA/AACE perspective is that evidence from retrospective databases is not as robust as randomized controlled trials and we should wait for the trial results. Apparently, there are nine ongoing trials that can shed light on this issue, although subjects in these trials will have many other risk factors. 


Questions and Answers 

Q: What’s your view on renal impairment with GLP-1. 

A: We don’t have adequate data. There are reports, but the affected patients seem to have a predisposition and then possibly GLP-1 makes it worse. So for now, let’s be careful in patients with pre-existing renal disease. 

Q: There may not be an independent effect on the liver, but who cares? 

A: Yes, it doesn’t matter at all. The fact is that patients lose weight and liver function gets better with GLP-1. 

Q: What data are the FDA investigating? 

A: The FDA call was based on the JAMA paper. They have had lots of databases thrust in front of them, some more reliable than others. But I understand that it was this paper. 


Clifford J. Bailey PhD, FRCP, FRCPath (Aston University, Birmingham, England) 

Dr. Bailey gave us a veritable shopping list of exciting new drugs for diabetes, including linagliptin  (Tradjenta, BI/Lilly), exenatide once weekly (Bydureon, Amylin), lixisenatide (Lyxumia, Sanofi), dapagliflozin (Forxiga, BMS/AZ) and insulin degludec (Tresiba, Novo Nordisk). It’s notable that Lyxumia, Forxiga and Tresiba have all been launched in the UK within the last few weeks. 

  • Linagliptin (Tradjenta, BI/Lilly): Is the fourth of the available DPP-4 inhibitors. Alogliptin (Nesina, Takeda) is available in Japan, has been approved by the FDA, and is currently with EU regulators. Linagliptin is disposed of in bile rather than via the kidney. So the difference with linagliptin is that it doesn’t require dose reduction in renally impaired patients. 
  • Exenatide once weekly (Bydureon, Amylin): Good A1c and weight reduction. In a head to head against liraglutide it showed consistent results in A1c and weight lowering. Dr. Bailey thinks that it’s almost impossible to make a direct comparison in GLP-1 head to head trials, because of variation in dosing and in titrating up to full dose. 
  • Lixisenatide (Lyxumia, Sanofi): Lixisenatide is a little bit different, Dr. Bailey said. Liraglutide is very homologous to human GLP-1, while lixisenatide is an exenatide like molecule, with extra protection to last a little longer in circulation. The GETGOAL phase 3 trials of lixisenatide showed a clear and consistent lowering of A1c (in the range 0.7 -1.0%) at 24 weeks, no matter what it’s paired with. For weight lowering, we see a 2-3kg reduction again from monotherapy to all the add-on combinations. A study shows that we can use lixisenatide to knock out the glucose excursion of the first main meal after injection, while liraglutide has a much longer time of action. 
  • There are many SGLT-2 inhibitors in the pipeline. Dapagliflozin (Forxiga, BMS/AZ) is now approved. Also in line are empagliflozin (BI/Lilly), canagliflozin (Invokana, Janssen), ipragliflozin (Astellas), and tofogliflozin (Chugai). SGLT-2 inhibition enables the kidney to divert 50-80g/day of glucose to the urine, which equates to 200-300kcal/day. Dapagliflozin can give A1c lowering of 0.8% from baseline, it’s durable over two years, and it’s not insulin dependent. SGLT-2 inhibitors also reduce weight, with no risk of hypoglycemia. That’s because the kidney has the safety net of SGLT-1, which will pick up all the glucose when blood glucose gets low. With this class, there is an increase in genital infections, but in trials, most were self-treated. Against sulfonylureas, dapagliflozin has a consistent effect, yielding about a 0.5% A1c reduction from baseline after one year. Sulfonylureas have a larger A1c reduction to start with, but quickly regress. A1c reductions are identical after one year, and dapagliflozin has superior A1c reduction after two years, plus consistent weight reduction as oppose to weight gain and no hypoglycemia vs. frequent hypoglycemia. 
  • Degludec (Tresiba, Novo Nordisk): is a newly available insulin. It’s an ultra-long acting basal insulin. It works by forming soluable multi-hexamers after subcutaneous injection that monomerize slowly. It can act for about 40 hours. Dosing is once daily, or possibly less frequent. It has been shown to be non-inferior to insulin glargine (Lantus, Sanofi) with a meaningful reduction in nocturnal hypoglycemia. 

Questions and Answers 

Q: What is the hypoglycemia risk if we add an SGLT-2 to sulfonylurea and insulin? 

A: When we added dapagliflozin on top of a sulfonylurea, there was a 2% increase in hypoglycemia 
reporting and in 800 patients there were 3 more serious events than with the sulfonylurea alone. This is because the sulfonylurea itself has the propensity for hypoglycemia. So if you are fairly near to target, reduce the sulfonylurea dose when adding SGLT-2. You will get a bigger A1c reduction with less risk. We are still learning about adding to insulin. You can lower the insulin dose to get the same A1c reduction. Or even go lower on A1c. 

Q: What about SGLT-1 inhibition? 

A: A little SGLT-1 inhibition would be good, but too much in either the kidney or the intestine would be very bad. So we would have to be cautious. 

Q. Are we testing SGLT-2 inhibitors for type 1 diabetes? For people who want to take less insulin? 

A: There are trials underway, but I have no idea how they are panning out. 


Rury R Holman FRCP FMedSci (University of Oxford, UK) 

Dr. Holman, the legendary trialist, turned up at Diabetes UK ostensibly to discuss metformin and cancer, but like the magician he is, he pulled a gleaming new trial (called GLINT) out of his hat. There is controversy because although observational trials strongly suggest that metformin has a protective effect against cancer, our meta-analysis of randomized controlled trials shows precisely zero effect, said Dr. Holman. A new trial is required to lay the matter to rest. The GLINT trial takes people with prediabetes, randomized to metformin or placebo, and tracks cardiovascular outcomes and cancer. The trial will be powered for cardiovascular endpoints, but cancer rates are expected to be higher, making this feasible. GLINT will recruit nearly 12,000 people and take six years. 

  • Observational data consistently suggests that metformin protects against cancer. Type 2 diabetes carries an increased risk of pancreatic, liver, and endometrial cancers. In a 2006 study, the hazard ratio (HR) for cancer was 30% higher with sulfonylurea compared to metformin. In a 12,000 person DARTs study, metformin had an adjusted risk reduction of 23% for cancer. In a Scottish observational cohort study of 4,085 people with type 2 diabetes and matched comparators, there was an early and consistent benefit to metformin, with a HR for cancer of 0.63. 
  • By contrast, a meta-analysis of randomized controlled trial data, suggests that metformin has no effect. Metformin is common in diabetes trials. The trials aren’t cancer trials, but data is available. After reviewing the universe of diabetes trials, fourteen studies were selected in which metformin was compared to either an active comparator or placebo. There were 398 cancers and over 52,000 person years. The headline result is a HR of 1.02 (i.e. no different to the comparator). 
  • Because all that glitters is not gold, Dr. Holman introduced the GLINT trial, which should provide some illumination. There has been a signal from observational studies and everyone believes it, so proof is required. The GLINT trial (Glucose Lowering In Non-diabetic hyperglycaemia Trial) studies people with prediabetes, since we can’t randomize people with diabetes to metformin as it is now the standard first-line therapy. Funding comes from NIHR, with medication from Merck Serono. The primary endpoint is time to a composite cardiovascular outcome, and secondary endpoints include cancer. The plan is to recruit 11,834 participants (after a 500-person feasibility trial) with a ten year CV risk of >20%. The trial will take a minimum of six years. 

Questions and Answers 

Q: I believe there is data to suggest that there are major flaws in the observational studies. It’s to do with misallocation of time at risk. So I don’t see any controversy here. 

A: The problem with database retrospective searches is you can’t be sure you got all the confounders. The people who undertake observational analyses often attract a lot of media attention but we need 
randomized controlled trials to be certain about treatment risks and benefits. For the metformin and 
cancer story we have genuine uncertainty and this new trial is the only way we can lay it to rest. 

Q: Where will you recruit such a large number of people with prediabetes from? 

A: We are twinned with the Cambridge to perform this study and will use their ADDITION study 
methodology. We are also partnering with the NHS vascular checks program. The feasibility phase of the study is to test, can we find them? will they take the tablet? will they stay on it? 

Q: Can we offer lifestyle advice to patients in the study? 

A: There will be no preclusions, it’s an “on top of usual care” study. We will have a very light touch, and collect data via electronic systems. Everyone will get diet and exercise advice, but can take any agent. They will be treated to best usual practice. 


James A. M. Shaw MD, PhD, (University of Newcastle, Newcastle-Upon-Tyne, England) 

Severe hypoglycemia is the limiting factor for glucose control in type 1 diabetes. Dr. Shaw showed the staggering result that people with consistent hypoglycemia (53 minutes/day at baseline) were able to reduce severe hypoglycemia by 90% after education with a tool called My Hypo Compass. Dr. Shaw noted that all technology (e.g. CGM) can only work with education. The education forms part of the HypoCOMPaSS study, which will investigate success in reducing hypoglycemia and impaired awareness, against the use of pumps/MDI and SMBG/CGM. The trial is over, but data is still being obtained. 

  • Impaired awareness of hypoglycemia affects 25% of those with type 1 diabetes and it confers a six-fold risk of severe hypoglycemia. But it is reversible, even in longstanding diabetes by avoiding hypoglycemia for a few weeks. Educational awareness can help too. 
  • There are many unsatisfactory aspects of studying hypoglycemia in typical randomized controlled trials. For example, in treat to target trials, we tend to select participants who don’t have hypoglycemia. Pump trials imply better hypoglycemia, but the pump arms get more attention, and the MDI arms don’t always use analogs. 
  • The HypoCOMPaSS study is 24 week randomized controlled trial with two year follow up, to avoid low glucose without relaxing glucose control using an educational tool. It has a 2x2 factorial design. The four groups are MDI/pumps and SMBG/CGM. Primary outcome is the difference in IAH (impaired awareness of hypoglycemia) score after 24 weeks. 
  • To measure hypoglycemia unawareness, the team developed the HypoA-Q questionnaire. This builds from the one question Gold Score - “do you know when your hypos are commencing?” and the eight-question Clarke questionnaire. HypoA-Q takes 5-7 minutes and has 18 questions. It includes recalling hypoglycemia when awake and asleep, need for emergency services, blood glucose thresholds for symptoms, symptom awareness and altered awareness. 
  • The team also developed “My Hypo Compass”, an education tool to optimize symptom recognition and action. The concept is based on points of the compass. North = Now (act immediately), E = early warning (understand times of increased risk), S = subtle symptoms, and W = check (to know for sure they are low). It only takes an afternoon to train someone with this tool. 
  • Full results from HypoCOMPaSS are not in, but across the sample there was an enormous reduction in severe hypoglycemia (by 90%), mild/moderate hypoglycemia and hypoglycemia unawareness – while A1c remained constant. N=96 participants were selected to have high hypoglycemia (53 minutes/day at baseline). 92% experienced severe hypoglycemia before the trial, and only 19% during the trial. HypoA-Q score reduced from 14 to 9. A1c was 6.5%. 


Tim A Holt PhD, MRCP, FRCGP (University of Oxford, UK) 

A number of GLP-1 agonist trials have reported a small increase in heart rate, which is usually written off as not clinically significant. A systematic review of all trials and subsequent meta-analysis shows a statistically significant increase in heart rate of 1.9bpm with GLP-1. This effect was more evident for long-acting GLP-1s than for shorter acting GLP-1s. 

  • GLP-1 trials usually report A1c as the primary outcome, but a number of trials have also reported a small increase in heart rate, which is usually written off as not clinically significant. 
  • A systematic review of all trials and meta-analysis shows a statistically significant increase in heart rate of 1.9bpm with GLP-1. This effect was more evident for liraglutide and exenatide long-acting release than for exenatide twice daily. 521 articles were whittled down to n=32 studies in the final meta-analysis. There was also a decrease in systolic blood pressure by 1.8 mmH (vs. placebo) and 2.4 (vs. active comparator). Body weight decreased by 3.3 kg compared to active control, but by only 1.2 kg compared to placebo. The heterogeneity across studies was larger than anticipated.
  • A study by Gill et al is the only trial where heart rate is an outcome. They used a 24 hour heart rate monitor, and reported a rise in heart rate of 2.8 bpm vs. placebo, which was not significant in this small pilot study. 
  • GLP-1 increases heart rate by a statistically significant amount, but the clinical significance needs to be understood. This effect might offset the beneficial results of glycemia, body weight and blood pressure. 

Questions and Answers 

Q: Can you speculate on cause and effect? 

A: A lot of work has been done on animal models, with conflicting results. And animal models are not reliable for results in humans. In the meantime we need more research to investigate how heart rate varies across subgroups. Some groups might be at more risk than others. 

Q: We note that lixisenatide doesn’t have a heart rate change. Why? 

A: I can’t comment on the basic science behind it. But be careful on relying too heavily on early data. 

Q: If there is a small increase in heart rate and a little drop in blood pressure they could be related. So this might be part of a normal physiological process and nothing to worry about. 

A: That’s quite possible.

-- by John and Kelly Close