This site is intended for health professionals only

Screening for diabetes in primary care

Rodger Charlton
MD FRCGP
GP
Associate Clinical Professor
Warwick Medical School

Mary Charlton
BSc MRCP
Associate Specialist
Heart of England Foundation Trust

Diabetes is very common and as practitioners we should all be thinking about looking after patients with diabetes, but also considering how many people we see with undiagnosed diabetes.

Why screen for diabetes?
Let's begin with some startling statistics:

  • Currently, 2.3 million people in the UK are known to have diabetes and another half a million or more have type 2 diabetes but don't yet know it.1
  • In around 15 years, the national figure is likely to have risen to 4 million – the equivalent of half the population of southeast England.1
  • Type 2 diabetes can go undiagnosed for up to seven years after it develops. This symptom-free period explains why one in five people with type 2 diabetes has retinopathy by the time their diabetes is picked up.
  • People are developing type 2 diabetes (previously known as maturity-onset diabetes) at younger and younger ages. In 2002, 1,041 pregnancies were in women who already had type 2 diabetes.2
  • Every five minutes, on average, someone in the UK is told they have diabetes.1

Does this leave you with an uneasy feeling you might be missing the diagnosis? If so, you're in good company. The charity Diabetes UK (www.diabetes.org.uk) has launched its biggest ever UK-wide advertising campaign that diabetes is serious and that it is a "silent assassin" as it so often goes undiagnosed.

Type 1 and type 2 diabetes
When talking about screening, we're talking about type 2 diabetes. In type 1 diabetes the hyperglycaemic patient is likely to be:

  • Younger.
  • Slimmer.
  • Losing weight.
  • Thirsty and dehydrated.
  • Often without a family history of diabetes.
  • Found to have ketones in urine.

It is unusual for type 1 diabetes to be picked up on screening because insulin secretion falls so much more rapidly than in type 2 diabetes. The patient quickly develops symptoms and people around them are likely to be alarmed by their weight loss.

This diagnosis needs to be made immediately as undiagnosed type 1 diabetes can rapidly lead to ketoacidosis with its associated high mortality. If the condition is suspected, a blood glucose and urine analysis for ketones is mandatory.

Fortunately, there is a relatively rare presentation to a GP or practice nurse, but it must always be considered when a young person presents unwell with a history of thirst, urinary frequency or weight loss. If suspected, type 1 diabetes should prompt immediate referral to a specialist team to assess whether ketoacidosis is present and to start insulin
at once.

Who should be screened?
For some conditions (carcinoma of the cervix, for example), blanket screening is worthwhile. In spite of the alarming statistics about the rising prevalence of diabetes, at present the benefits of population screening are not thought to outweigh the costs. This means we are left with two approaches:

  • Targeted screening.
  • Opportunistic screening.

Targeted screening: who is at risk?
Targeted screening means identifying a population or populations at increased risk of diabetes and offering them screening tests. Predicting these populations is sometimes obvious, sometimes less so.

Lifestyle, weight and age
We know that diabetes is on the rise as a result of us living longer and, more importantly, as a result of more of us being overweight thanks to readily-available calorie-rich food and a less active lifestyle. As we accumulate more fat, our body becomes insulin-resistant and then our insulin production begins slowly to deteriorate until it can no longer control our blood glucose.

How fat do we have to be to be at risk of type 2 diabetes? Diabetes UK identifies men with a waist measurement of 37 inches or more (or 35 inches or more for men of South Asian origin) and women with a waist of 31.5 inches as at risk. The definition of overweight is a body mass index (BMI) of 25 or more is reduced in people of South Asian origin. All the same, remember that 20% of people who develop type 2 diabetes are not overweight at the time of diagnosis.

The age criteria for screening for diabetes also vary with ethnicity. Although cut-offs are bound to be arbitrary, the consensus is to screen the white over-40s and the over-25s of non-European ethnicity, with one other risk factor (see Box 1).4

[[Box 1 diabs]]

Family history
Diabetes is particularly common in people with a family history of type 2 diabetes; if one twin has type 2 diabetes their twin is twice as likely to develop it if they're identical than if they're not.5 Genes, environment and nutrition in the womb are all-important.

Ethnicity
People from South Asian and Black backgrounds have an increased risk of developing type 2 diabetes, although the reasons for this are unclear. What's more, they develop it on average five years earlier than white people.1 People of South Asian origin seem to develop the unhealthy consequences of being overweight at a lower BMI and waist measurement than others. More than 300,000 people from those ethnic groups in the UK have already been diagnosed with diabetes.

Deprivation
Type 2 diabetes is an example of health inequality. The most deprived people in the UK have been found to be 2.5 times more likely to have diabetes.1 This may be a result of more obesity and less exercise but other factors such as less fresh food, stress and other illnesses may contribute.

Hypertension and arterial disease
People with diabetes are up to five times more likely to develop cardiovascular disease. South Asian people are at even higher risk.1

The flipside to this coin, bearing in mind that type 2 diabetes can be asymptomatic, is that people with arterial disease may have undiagnosed diabetes and should be regarded as a "high-risk" group.

People with hypertension are also often found to have or to develop type 2 diabetes because both conditions can arise from insulin resistance. If treated with a betablocker and a thiazide, this risk increases and this combination is generally avoided.

PCOS and/or a history of gestational diabetes
Polycystic ovary syndrome (PCOS) is a condition of insulin resistance and carries, as you would expect, a risk of diabetes. If it's being treated with metformin, it is advisable to stop it for around four weeks before screening for diabetes.

Women with a history of gestational diabetes have already been shown to have inadequate insulin reserve for the insulin resistance which develops in pregnancy, and are therefore at risk of developing diabetes if that insulin reserve continues to deteriorate and/or their own insulin resistance rises with weight gain and inactivity.

Mental health problems
As with hypertension, there is a drug-related element to the extra risk of type 2 diabetes in those with mental illness, especially schizophrenia; the newer antipsychotics such as olanzapine promote weight gain and
insulin resistance.

Targeted screening: you need a register
If targeted screening is to succeed (and there is no specific incentive in the Quality and Outcomes Framework) there needs to be a register of those "at risk". Many of these individuals will already be on a GP register for conditions such as hypertension, obesity or coronary artery disease. These would need to be collated into a "diabetes screening register" and supplemented with women with a history of gestational diabetes or PCOS, people with schizophrenia and people with a strong family history of diabetes. This register should then be used to recall people for screening, as discussed below.

Opportunistic screening
Professionals working in primary care are seeing a greater number of patients on their practice lists than ever before for urine and blood tests. With a patient's permission, these consultations are chances for "opportunistic" screening. This means that people in the "at-risk" groups are offered a screening test when they are seen primarily for review of another related condition, such as hypertension, or for an unrelated reason, such as intercurrent illness.
However, the Diabetes UK website encourages us to think about such screening in a more organised way such as that summarised in Box 2.

[[Box 2 diabs]]

Which test?
Once we've identified a group to screen, which test is most appropriate? The short answer is fasting plasma glucose, and here's why:

  • At present, an oral glucose tolerance test (OGTT) with plasma glucose measurements, fasting and two hours after a 75 g glucose oral load, is the gold standard for diagnosis. But as a screening test it's unwieldy (where do people sit for two hours?), expensive and unpopular (some people vomit, others need to fetch their children from nursery). It is also surprisingly inconsistent.
  • Urine testing may seem cheap and simple but misses anything between 20% and 80% of cases.
  • HbA1c or glycated haemoglobin is used to monitor blood glucose control in people with diabetes but as a screening tool it, too, seems to miss cases of diabetes. However, HbA1c measurements are becoming more standardised across laboratories and it may eventually have a role in screening for diabetes.
  • Which leaves us with blood sugar. Random blood glucose values are helpful if clearly diabetic (above 11) or clearly nondiabetic (below 5.5) but difficult to interpret in between.
  • Capillary ("fingerprick") testing is not regarded as sufficiently valid for diagnosis and misses cases that laboratory measurements of venous blood pick up. All the same, Diabetes UK recognises the convenience of capillary testing but if the finger-prick result is positive (see below) a laboratory test must be performed.

Fasting plasma glucose (venous blood in a fluoride tube) is, by consensus, the best available screening test (Figure 1). It introduces the inconvenience of an overnight fast for the person being tested. However, it is fairly consistent from day to day, making it reliable at least as an initial screen.

[[Fig 1 diabs]]

Interpreting the result of fasting plasma glucose

  • A fasting plasma glucose over 7 indicates diabetes.
    –     If the person has symptoms (thirst and excessive urination), the diagnosis is safe. If they're asymptomatic, the fasting glucose should be repeated and two readings over 7 are diagnostic of diabetes. 
  • Between 5.5 and 6.9 should be followed up with an OGTT.
  • If readings below 7 are ignored, some cases are missed. So where do we set our cut-off for a positive result? A big survey in Leicestershire showed that if we used 4.3 as our threshold for further testing then we wouldn't miss any diabetes but would end up sending nearly everyone for an OGTT.6 Diabetes UK recommends 6.1 as a threshold for further testing. However, 5.5 seems a reliable compromise; it misses only 10% of cases and yields only 20% false positives.
  • Below 5.5 is a negative result.

Interpretation and action
The following World Health Organization (WHO) criteria should be used:

  • Diabetes:
    –    Fasting plasma glucose of 7.0 or greater
        and/or
    –    Random glucose/two-hour glucose in an OGTT of 11.1 or greater.
  • Normal response to a 75 g OGTT is:
        –     –

If the screening and subsequent tests show diabetes then, assuming there are no features of type 1 diabetes, the person should be offered routine diabetes care, including a structured educational package, annual review and medication for cardiovascular protection. If the screening is negative, repeat screening should be offered in three years.

Other possible outcomes include:

  • Impaired fasting glucose (IFG). So far, the diagnosis has been relatively straightforward. But what about people whose fasting glucose is above 6.0, the upper limit of normal, but below 7.0 mmol/L? This is referred to as impaired fasting glycaemia (IFG) and the person should be actively managed with lifestyle advice. It is important, as 1–2% of these people will go on to develop diabetes in a year, so a system of recall needs to be put into place to screen them for diabetes again every 12 months. They also have an increased risk of cardiovascular problems and should be treated for hypertension and with statins.
  • Impaired glucose tolerance (IGT). To complicate matters further, if a person has an OGTT and their two-hour glucose result is 7.8 mmol/L or greater, but

Advantages and disadvantages of screening for diabetes
The pros
All this screening can be justified only if our interventions will benefit the person screened. The UK Prospective Diabetes Study established that improved glycaemic control reduced microvascular complications, such as retinopathy and blood pressure treatment improved outcome from cardiovascular disease in patients with clinically diagnosed diabetes.8 So it is worth treating people who are asymptomatic from diabetes.

Statins are important here; the Heart Protection Study showed that they reduce cardiovascular events in people with diabetes.9 A German study of 166 people diagnosed with diabetes in a screening programme showed after 32-years that people lived longer if their diabetes was found by screening rather than conventionally diagnosed diabetes.10
People with IGT have the added benefit of being able to use lifestyle interventions to reduce their risk of developing diabetes (see Box 3).

[[Box 3 diabs]]

The cons
Asking someone to participate in a screening programme is asking them to undergo the discomfort and inconvenience of fasting and blood tests, and people often find the oral glucose of an OGTT unpleasant or even emetic. If the test is positive, it brings a lifelong diagnosis with its attendant implications, both medical and social (travel and health insurance). If negative, then unnecessary anxiety has been introduced.

Salutary thoughts
Diagnosing and managing diabetes in a GP practice is akin to climbing Everest. The challenges posed in making the diagnosis are huge, not least the logistics of following up an ever-mobile population and combating the inverse law of care whereby those who most need screening are least likely to attend.

Screening is designed for the asymptomatic. Don't forget that people with the following rather vague symptoms should be tested for diabetes:

  • Tiredness.
  • Recurrent infections.
  • Oral, vaginal and penile thrush.
  • Unexplained weight loss.
  • A complication of diabetes.
  • Nocturia in an older man is not necessarily due to prostatism, so ensure a potential diagnosis of diabetes is excluded. And don't overlook those borderline glucose results.

Dr Tim Holt published an important paper in 2008 where he searched the computer records of 3.6 million patients from 480 GP practices.11 He found that the last blood glucose level in approximately 0.1% was indicative of undiagnosed diabetes and that in 0.9%, that their last blood glucose level was at best borderline, leaving many of them at risk of diabetes or having undiagnosed diabetes. The lesson from this paper - it is important that people with borderline glucose readings are entered on the GP practice computer system for recall.
 
Conclusion
Screening for diabetes should be:

  • Targeted at high-risk groups recorded on a practice register. An individual over 40 years (25 in higher risk ethnic groups) with at least one risk factor for diabetes would benefit from being screened.
  • Initially by fasting plasma glucose.
  • Followed up with a glucose tolerance test when the fasting glucose is above 5.5.
  • Repeated every three years if negative. People with IFG or IGT should be followed up annually.
  • Followed by intensive lifestyle advice and cardiovascular medication in all degrees of glucose intolerance and by diabetes follow-up in those with diabetes.
  • Opportunistic, especially among groups unlikely to attend a formal programme of recall.

Research suggests that the earlier a diagnosis is made and the person managed appropriately, the better the outcome in both their morbidity and mortality. An educational message needs to be delivered that leading a healthy lifestyle will help to reduce the likelihood of developing diabetes and that screening is available.

GPs and practice nurses have a pivotal role in this process that will vary according to their practice population and demographics, but it is a topic that should be reviewed regularly at team meetings locally, in collaboration with their primary care organisations.

References
1. Diabetes UK. Diabetes: Beware the Silent Assassin. London: Diabetes UK; 2008. Available from: http://www.diabetes.org.uk/Documents/Reports/Silent_assassin_press_repor...
2. Confidential Enquiry into Maternal and Child Health (CEMACH). Pregnancy in women with type 1 and type 2 diabetes in 2002-03, England, Wales and Northern Ireland. London: CEMACH; 2005. 
3. Tringham JR, Davies MJ. Screening for IGT and diabetes. British Journal of Diabetes and Vascular Disease 2004;4:254–58.
4. Diabetes UK. Position Statement: Early identification of people with Type 2 diabetes. London: Diabetes UK; 2006.
5. Kaprio J, Tuomilehto J, Koskenvuo M et al. Concordance for type 1 (insulin-dependent) and type 2 (non-insulin-dependent) diabetes mellitus in a population-based cohort of twins in Finland. Diabetologia 1992;35(11):1060–7.
6. Tringham JR, Davies MJ, Jarvis J, Thomson J, Farooqi A, Khunti K. Effectiveness of fasting plasma glucose and HbA1C as a screening test for type 2 diabetes in south Asians and white Europeans in
Leicestershire. Diabet Med 2004;21:A53.
7. World Health Organization (WHO). Screening for Type 2 Diabetes. Geneva: WHO; 2003.
8. UK Prospective Diabetes Study Group. Tight blood pressure control and risk of macrovascular and microvascular complications in type 2 diabetes: UKPDS 38. BMJ 1998;317(7160):703–13.
9. Collins R, Armitage J, Parish S, Sleigh P, Peto R. Heart Protection Study Collaborative Group. MRC/BHF Heart Protection Study of cholesterol-lowering with simvastatin in 5963 people with diabetes: a randomised placebo-controlled trial. Lancet 2003;361(9374):2005–16.
10. Schneider H, Ehrlich M, Lischinski M, Schneider H. Did the intensive glycosuria screening of the sixties and seventies in East Germany improve the survival prognosis of early detected diabetics? Diab Stoffw 1996;5:33-8
11. Holt RI. Diagnosis, epidemiology and pathogenesis of diabetes mellitus: an update for psychiatrists. Br J Psychiatry Suppl 2004;47:S55–63.