Evidence-based practice in type 2 diabetes

The NMC Code requires all nurses to follow the principles of evidence-based practice, but how confident are you in applying ‘evidence’ to your daily practice? We look at this question in the context of type 2 diabetes

The usefulness of medical evidence depends (among other things) on asking the right questions. The reality is that most research is driven by specialists and academics: it is a shame that not more research is done by workers in primary care as they are the people who know what primary care needs, and are able to ask the right questions.

DOES DIABETES EXIST?

When somebody breaks a leg, it is clear that there is something wrong and deserving of treatment. Diabetes is rather different. Diabetes is defined by what are considered to be abnormally high levels of blood glucose.¹ Abnormally high blood glucose can cause symptoms, and you will all have been taught that diabetes mellitus is characterised by weight loss, thirst, excessive urination, tiredness, and possibly clouding of consciousness if things are really out of hand. However, most people with type 2 diabetes (T2D) have none of these symptoms, and many would actively welcome a bit of weight loss.

So T2D is difficult to define as a disease on the basis that it causes illness. Perhaps therefore T2D should be defined by what it might lead to, the complications. These are indeed important and numerous (see Box 1), and much GPN time in clinic is devoted to checking for or monitoring the progress of the various complications of T2D. But as a way of defining T2D this is not that useful either. The United Kingdom Prospective Diabetes Study (UKPDS) found that at diagnosis of T2D a significant proportion of patients already had complications, and that on follow-up there was apparently a continuous correlation between blood glucose levels and complications.² This means that there is no clear threshold blood glucose level below which you do not have T2D and can relax, and above which you do have T2D with all the attendant risks of complications and care needed.

So the question: ‘does diabetes exist’ is not quite so dumb after all. As far as blood glucose is concerned, the evidence says that lower is better. If complications rise gradually with glucose levels there is hardly any justification in using a rigid blood glucose threshold (except that it makes life more convenient for clinicians, but probably not for patients). The published blood glucose levels to make the diagnosis are a product of inspired guesswork based on the blood glucose levels achieved in the published research.

This is not at all very neat, and has led to interest in pre-diabetes: that in-between group who do not officially have diabetes, but on the other hand are not quite normal either. The main risk of having pre-diabetes is that it progresses to full-blown T2D. Yet people with pre-diabetes have more than their share of kidney damage, and are prone to unrecognised heart attacks,3 so they already have some of the complications of T2D. The number quoted by NICE/CKS for a level at which T2D is diagnosed is a persistent HbA1c of 48 mmol/mol or greater,4 but this apparent certitude does not really mean that if your HbA1c is 47 mmol/mol then all is sunny and bright and you can forget about it, but if it drifts up to 48 mmol/mol then all the Swordsmen of the Apocalypse will descend on your head. Besides anything else, any pathology laboratory in their right mind would not guarantee the accuracy of their results to this degree.

It is important that Evidence-Based Nursing Practice (EBNP) is driven by the available research, but this can sometimes lead to a false sense of security, or indeed insecurity.⁵ There is still scope for looking at the patient in front of you, and asking the question: ‘What is right for him?’

IS DIABETES WORTH RESEARCHING?

Research tends to be done on the things that interest clinicians, or politicians, or else as a means of selling drugs. The coughs and colds, the bread-and-butter of general practice, the things which make a lot of people slightly miserable, are not the stuff of ground-breaking or cutting-edge research. Diabetes has, on the other hand, generated a lot of research over a lot of years, and is important for patients, for society and for primary and secondary care.

According to Diabetes UK, over the last 20 years the number of people diagnosed with diabetes in the UK has doubled.⁶ In 2014 there were an estimated 3.5 million diagnosed cases in the UK, with a further 1.1 million people with diabetes but not yet diagnosed. This represents 700 new diagnoses a day or one every 2 minutes. Ninety per cent of people with diabetes have T2D.

There are also people out there who have pre-diabetes and so are at increased risk of developing T2D in the future. The Health Survey for England7 found that nearly 11% of the UK population already have a slightly raised HbA1c of 42-46 mmol/mol. The Health and Social Care Information Centre calculates that, based on Body Mass Index and waist circumference data, the UK population at-risk of T2D is nearly 12 million, or approaching 20% of the population.⁸

The NHS cost of diabetes (types 1 and 2) was estimated in 2012 to be £10 billion a year, or 10% of the entire NHS budget. If you add in non NHS costs – social care, lost production, sickness benefits etc. – the total national cost is approaching £24 billion a year.⁹ Diabetes is a very big deal, both for the people who have it, and for the country and its healthcare services.

Until about the 1970s, diabetes was considered to be a ‘hospital’ condition, with all newly diagnosed patients being automatically referred to outpatients. This is no longer feasible because of the number of patients, nor necessary – a lot of diabetes specialists were perturbed to find that primary care was just as capable of looking after people with diabetes as they were themselves.¹⁰ Diabetes is now an overwhelmingly general practice condition. And when faced with a population with serious ongoing healthcare needs, GPs did what they always do in such situations – entrust the job to GPNs who do it so very much better. Diabetes was one of the very first diseases to generate specific clinics in primary care, and indeed for some time such clinics were regarded as practice quality markers: if a practice did its diabetes clinic well, then there was a good chance that it did all its care well.

HOW GOOD IS THE RESEARCH INTO DIABETES?

When critically appraising a piece of research, the general rules are that more patients is better, and a longer duration of the study is better. But more patients and a longer duration means that doing the research is more expensive, and there is a longer delay in getting the results, and so a longer delay in implementing any care improvements that may be suggested by the research.

The perfect piece of medical research has never been published. There is always something which can be criticised as leading to bias or misinterpretation. Even when presented with an authoritative paper, you still have to form an opinion about whether it is to be believed, about whether the strengths outweigh the weaknesses. An additional barrier is that when published, some papers do not offer enough information about the study population, drop-out rates, randomisation process etc., and all the other things which might be sources of bias: the research may in reality be superb, but without this information you are unable to confirm the fact. And, of course, some research is started but never finished, and some is completed but never published, so any useful findings remain hidden.

For the people who write the clinical guidelines, this struggle with the quality of the evidence is a constant challenge. However, constructing the guidelines means that something has to be written down. NICE Clinical Knowledge Summaries (CKS)⁴ give estimates of the quality of the evidence used, but you really have to drill-down into the website to find it.¹¹ The Scottish Intercollegiate Guidelines Network is rather better in this respect.¹² As far as nursing practice is concerned, Polit and Beck13 recommend the use of a hierarchy of evidence. Level-one evidence is the best. Personal anecdotes, or what the bloke in the pub thinks, do not appear on this list – see Box 2.

IS TREATING BLOOD GLUCOSE WORTHWHILE?

Logically, if raised blood glucose is to be characterised as a disease, then reducing blood glucose ought to confer some benefits. The benefits to be looked for are reductions in symptoms, complications and death. However, some early research was not very encouraging. The University Group Diabetes Program was run in America in the 1960s to find out if the (then) new tablets to reduce blood glucose did patients any good. The trial was stopped early because the death rate was not improved by insulin compared to placebo, and the death rate from the tablet group (tolbutamide, still available in the UK) went up threefold.¹⁴ This generated consternation in the diabetes community. The result did not really make sense (’not making sense’ is not a bad yardstick for assessing the quality of research, but on the other hand the research may be genuinely ground-breaking and likely to alter nursing practice).

A bigger and longer trial was required, and a group at Oxford (UK) took up the baton. The trial had to look at several important outcomes; had to have a placebo group; had to be properly randomised (taking into account any other know risk-factors that might bias an assessment of the effect of reducing the glucose, such as age, smoking and blood pressure); had to undertake meticulous follow-up to detect end-points (and to make sure too many people did not drop out); had to be long enough so that slow-onset outcomes would be picked up; and had to have sufficient numbers of subjects so that the confidence limits were not too wide. The result was the United Kingdom Prospective Diabetes Study.

UKPDS recruited over 5,000 newly diagnosed patients with T2D and followed them up over 10 years. It was found that reducing HbA1c reduced the risk of microvascular complications (particularly nephropathy and retinopathy), and also a trend – not statistically significant – that the risk of heart attack was reduced. During the study they also started a blood-pressure arm to see if control of blood pressure also helped microvascular and macrovascular complications, and it did (indeed it was concluded that controlling blood pressure is probably more important than controlling blood glucose as a way of reducing the complications of diabetes).

Once they got the wind in their sails there has been no stopping the UKPDS group: they have their own website,¹⁵ and their work has already generated over 80 published papers. For the obsessive critical appraiser my only caveat is that the full details of the methodology used are not readily available from the UKPDS website for scrutiny. However, the general informed consensus appears to be that UKPDS is high quality research,^11,12 and so is the ‘truth’ until a better truth comes along.

WHICH ARE THE BEST DRUGS?

Some people will not wish to take drugs for their diabetes, but most people with T2D at some stage will need medication.¹⁶ If you Google ‘type 2 diabetes treatment without medication’ you get over 2 million hits, with many sites claiming a cure. The lifestyle changes recommended are invariably arduous and unpleasant. This almost suggests that someone with T2D who has to take drugs is weak and feckless and a failure. For a condition with lifelong implications, this is hardly a positive frame of mind to encourage.

To add to the quandary, people with T2D who take medication for it do not have to pay prescription charges, not only for the diabetes medication itself but for any medication, prescribed for whatever reason. People with diabetes controlled just with diet and lifestyle modification continue to be liable for prescription charges. (If you are able to see any logic in this, please let me know.)

A trial looking at the benefits of treating T2D with glucose-lowering or blood-pressure-lowering drugs also needs to be of sufficient size and duration to pick up rare complications (of the drugs and of the T2D): it has to be ‘sufficiently powered’ (in the jargon) to detect treatment differences. Again there must be consideration of other risk factors, including the severity of any pre-existing microvascular or macrovascular disease. And it must be placebo-controlled, properly randomised, and if possible single or double blinded so that ideally neither the researchers nor the trial subjects know if they are being treated or not, since if they do know it affects the results.

Research done comparing different drugs needs to be of a ‘head-to-head’ nature with subjects randomly assigned to the different drugs being investigated. This sort of trial is very popular with drug companies. In order to get a licence for use a new drug only has to prove that it works, but evidence that it might work better or has fewer side effects than its competitors is understandably welcome.

The Cochrane Library contains numerous systematic reviews of T2D drugs, and is freely available on-line to anyone in the UK. In the unlikely event that you find yourself with spare time in the working day you can happily spend an hour trawling their website.¹⁷

NICE/CKS guidance⁴ suggests that metformin is the first choice blood glucose lowering drug for T2D. However, it is not a good idea in people at risk of lactic acidosis, a rare but possibly fatal event, and the at-risk group includes people with reduced renal function. Evidence revealing this has come from case reports and observational studies: a trial dicing with such a serious side effect would be unlikely to get ethical approval or many consenting recruits.

The four other available oral therapeutic groups are considered equally effective: gliptins (such as sitagliptin); sulfonylureas (e.g. gliclazide); pioglitazone; and sodium-glucose cotransporter 2 inhibitors (SGLT-2i) of which three are available in the UK and all have the suffix ‘liflozin’. It is apparent that what matters as far as T2D side effects are concerned is the reduction of blood glucose, and not how it is done.

According to the evidence reducing systolic blood pressure to 140mmHg (if necessary) in people with T2D is a good idea, but that pushing it lower results in more deaths. This was revealed by a meta-analysis of 49 pieces of research including 73,738 patients.¹⁸ It does not matter how you do it – getting the blood pressure to target is what matters.⁴

CONCLUSION

Looking after patients with diabetes is a big part of GPN life. The diabetes itself requires management and monitoring, and the sort of lifestyle changes that T2D requires will need constant motivation, with often disappointing progress. If your patients could easily bring about the necessary lifestyle adjustments they probably would not have T2D in the first place. Some GPNs will also be starting and making changes to medication, so it is as well to know that such adjustments will probably help. And, of course, as well as the diabetes many patients will have associated problems such as hypertension and CHD, making the visit to your clinic even more complicated.

There is a lot of good quality evidence to support the use of glucose- and blood-pressure-lowering drugs in diabetes. There are also several sets of guidelines to assist, of which probably the most important is from NICE/CKS. As well as basic research, there are lots of examples of systematic reviews (see the Cochrane Library), which give highly reliable evidence to support best clinical practice.

UKPDS made it pretty clear that getting glucose and blood pressure levels down (with a possible emphasis on blood pressure) are more important than how the levels are brought down. Start with metformin at first, but if this can’t be used or else is not sufficient, there are now several other medications that can confidently be added on or can replace the metformin. However, diabetes cannot be cured, and even if without drugs blood glucose can be brought back to normal levels (a very occasional occurrence), there still must be ongoing monitoring as the problem has a nasty habit of returning. Even with perfect glucose and blood pressure control, complications are minimised but not eliminated.

REFERENCES

1. Mayo Clinic. Diabetes: Definition. http://www.mayoclinic.org/diseases-conditions/diabetes/basics/definition/con-20033091

2. King P, Peacock I, and Donnelly R. The UK Prospective Diabetes Study (UKPDS): clinical and therapeutic implications for type 2 diabetes

Br J Clin Pharmacol. 1999 Nov; 48(5): 643–648.

3. Prediabetes. Mayo Clinic http://www.mayoclinic.org/diseases-conditions/prediabetes/symptoms-causes/dxc-20270026

4. NICE/CKS. Diabetes – type 2. https://cks.nice.org.uk/diabetes-type-2

5. Warren E. Evidence-based practice. Practice Nurse. December 2015;45(12):27-32.

6. Diabetes UK. Facts and stats. https://www.diabetes.org.uk/Documents/Position%20statements/DiabetesUK_Facts_Stats_Oct16.pdf

7. Public Health England (2015) NHS Diabetes Prevention Programme (NHS DPP) Non-diabetic hyperglycaemia. Produced by: National Cardiovascular Intelligence Network (NCVIN) https://www.gov.uk/government/publications/nhs-diabetes-prevention-programme-non-diabetic-hyperglycaemia

8. HSCIC. Statistics on Obesity, Physical Activity and Diet: England 2014. http://www.content.digital.nhs.uk/catalogue/PUB13648/Obes-phys-acti-diet-eng-2014-rep.pdf

9. Hex, N., et al. Estimating the current and future costs of Type 1 and Type 2 diabetes in the United Kingdom, including direct health costs and indirect societal and productivity costs. Diabetic Medicine. 2012;29 (7) 855-862

10. Ismail H, Wright J, Rhodes P, Scally A. Quality of care in diabetic patients attending routine primary care clinics compared with those attending GP specialist clinics. Diabet Med. 2006 Aug;23(8):851-6.

11. NICE/CKS. Type 2 diabetes in adults: management. GRADE tables and meta-analysis results. https://www.nice.org.uk/guidance/ng28/evidence/appendix-d-grade-tables-and-metaanalysis-results-pdf-2185320353

12. Health Improvement Scotland/SIGN. 116 Management of diabetes. http://www.sign.ac.uk/assets/sign116.pdf

13. Polit DF, Beck CT. Nursing research: generating and assessing evidence for nursing practice Eighth Edition, 2008. Lippincott Williams &Wilkins; Philadelphia, Pa.

14. The Diapedia Collective. The University Group Diabetes Program. https://www.diapedia.org/introduction-to-diabetes-mellitus/1104938831/the-university-group-diabetes-program

15. University of Oxford. UK Prospective Diabetes Study. https://www.dtu.ox.ac.uk/UKPDS/

16. Understanding medication – Type 2 diabetes – NHS.UK. https://beta.nhs.uk/conditions/type-2-diabetes/understanding-medication

17. Cochrane Library. Cochrane Database of Systematic Reviews. http://www.cochranelibrary.com/cochrane-database-of-systematic-reviews/

18. Brunström M, Carlberg B. Effect of antihypertensive treatment at different blood pressure levels in patients with diabetes mellitus: systematic review and meta-analyses. BMJ 2016;352:i717