I have read that the best screening test for Diabetes is HbA1c

Vivek, you have already answered your question... if you are interested in new onset, 2-3 independent fasting blood glucose will do nicely (if really high, even one will be a good indicator). If you wish to identify diabetics who are not necessarily the part of new onset group, HbA1c is useful because it represents an overall state over ~3months period. If on the other hand, you are interested in identifying individuals on the borderline (pre-diabetics), they may still manage their fasting blood glucose level slightly under the cutoff limit but they will show a skewed glucose disposal in glucose tolerance tolerance test. 

On the level of large groups, OGTT is time consuming, relative expensive, but most useful if the purpose is to implement a plan of intervention which avoids or delays the clinical onset. In the long-run, benefits far out weigh the initial efforts.

@ Marianne, I am afraid, HbA1c is not the best test. It has poorest sensitivity to pick up new onset diabetes among all three tests. Read these lines given in ADA supplement of Jan 2015.

"The concordance between the FPG and 2-h PG tests is imperfect, as is the concordance between A1C and either glucose-based test. National Health and Nutrition Examination Survey (NHANES) data indicate that an A1C cut point of $6.5% identifies one-third fewer cases of undiagnosed diabetes than a fasting glucose cut point of $126 mg/dL (7.0 mmol/L) (9). Numerous studies have confirmed that, compared with these A1C and FPG cut points, the 2-h PG value diagnoses more people with diabetes."

  The problem that I want to highlight here is how many practitioners out there in the field know these facts.  I have not seen anybody doing 2-h PG in diabetes these days.

@ Asif, you have correctly stated my concern. To tell a person that he does not have a diabetes, I think we must have done all three tests recommended by ADA.

I have been practicing for almost 10 years now, still we all go for a single test and tell the person you are not diabetic, just because one test is negative.

I think this is wrong!

Vivek, as you probably already understand, diabetes is a progressive disease and the clinical diagnosis on the basis of a fixed number, though useful and well-reasoned is somewhat arbitrary.  If a person begins to show a gradual increase over in fasting blood glucose level >110mg/dl approaching 126 mg/dl, you do not tell them that they are clinically diabetic however, it is important to emphasize to them that they need to be monitored a little more attentively. It is important to get them useful information about better nutritional and health practices to delay further increase in fasting hyperglycemia (type 2 diabetes).

"Screening test,undiagnosed diabetes and rural setting"! The best test for such a setting from a public health point of view is FPG. It is "plasma" not serum or blood glucose. However it is the easiest, convenient, lower cost and most acceptable to patients. Since you require a screening test. It is the most sensitive and specific test for the diagnosis of diabetes from a public health point of view. No test can be 100% sensitive and 100% specific. As fasting capillary glucose is equivalent to fasting plasma glucose. However 2 hour post OGTT has to be "plasma" glucose and not serum or blood glucose. To have plasma you need a centrifuge machine. This is impractical in rural settings for day to day practice. This test is not acceptable to most of the patients. Yes for research it is mandatory.  Remember 2 hour OGTT test is not even being promoted by ADA for diagnosis. ADA says OGTT is not required if FPG is >126mg%.

@ Sujeet Raina

I found the following recommendation by WHO and IDF:

The oral glucose tolerance test (OGTT) should be retained as a diagnostic

test for the following reasons:

fasting plasma glucose alone fails to diagnose approximately 30% of

cases of previously undiagnosed diabetes,

fan OGTT is the only means of identifying people with IGT,

fan OGTT is frequently needed to confirm or exclude an abnormality

of glucose tolerance in asymptomatic people.

An OGTT should be used in individuals with fasting plasma glucose 6.1–

6.9mmol/l (110–125mg/dl) to determine glucose tolerance status.

This recommendation dates back to 2006.

I am not sure, though, regarding the current status of WHO recommendation.

I could not find a revised recommendation from WHO, which means this recommendation still stands.

but even WHO and IDF recommends "either"  FPG>126mg% or 2 hr post OGTT >200 for the diagnosis of Diabetes. They don't advocate both tests. It is true to pick more and more of IGT  you need 2hr OGTT. Remember both WHO and IDF say IFG at >110mg% but ADA is more strict at >100mg% 

Something to consider is to ask ourselves a question about the actual goal here:

• an early intervention and optimum management of health care of patients
• Doing nothing until an official label of cilinical diagnosis is slapped on the patient chart and only then dealing with the health care

Both positions can be justified, depending on what is viewed as more important, but one of them serves the patients better.

@Tausif I think you are getting my point.

My goal is  an early intervention and optimum management of health care of health care.

Even to diagnose diabetes by the criteria of ADA and IDF with FPG > 125, 30% undiagnosed diabetes is missed by FPG alone.

So, when I plan a study in the field to find out the prevalence of undiagnosed diabetes, can I rely on the FPG alone?

The diagnostic criteria say FPG or 2 hr PG or HbA1c ???

What does that mean?

does that mean we have to do all of these if we want to rule out diabetes?

"@Tausif I think you are getting my point."

That's not just your point, its an age old conundrum.

As for your last question, most progressive diseases are hard to "rule out". That's why practical limits are set to "rule in" the onset. The simple idea being that if left alone without intervention, there is no realistic possibility that disease will not progress forward causing long-term systemic damage, finally resulting in death.

Problem with "ruling out" is that one can argue that a person may be at an early stage of progression, falling within the normal range. If a person is likely to head on the path of onset of T2D and if life-style change is implemented for intervention, likelihood of better outcome are far greater at such an early stage.

But you know what they say (and they do say an awful lot ;) ) about life - life is a terminal disease with mortality rate of 100%. Ultimately, no one gets out of it alive!  ;) 

The question is screening for diabetes in a rural setting & this would entirely depend on funding . It would also depend on the number who come to the camp & how long they have to wait to get their test done . The test usually done would be a random blood sugar test in a single sample . At the time of taking the  sample , it could be determined whether it is fasting or not & the criteria for DM , IFG or IGT should be determined . The person should be advised to undergo further tests depending on the results . The screening should also include checking BP & BMI . DM is under diagnosed in rural areas & the screening is the only method to get the prevalence in rural areas . This is for poor rural people .

  Those who can afford , Master health check up is usually done which can include all the three tests ie FBS , PPBS & HbA1c .

In case you wish make one-step procedure for many people and are not thinking about money in the first place, my preference will be anyway combination of fasting glucose + HbA1c. Yes, you will miss some people. You will find though a big group with abnormality which otherwise will not find time to visit more time-consuming procedure

Yes, I agree with most of the things discussed here. Screening is very important in rural population in developing countries. However when it comes to mass screening, funding becomes a major problem. Getting every one fasting is another challenge. Therefore, compromise is to do spot random glucose with a good glucometer. Then pickup which patient will need further testing.

I would like to suggest the measurement of Fasting Glucose alongwith HbA1c for screening the rural population. Although this might be liitle expensive but following these tests we would not miss the diagnosis even during the screening process.

@ Singh Sivakumar Master Health Checkup :)

Very good commentaries. My suggestion is to know also who to screen...

Practically everyone be screened in the rural set-up.

It seems to me that blod glucose measure 1.5 - 2 hours after the last meal will be cheap, easy and informative enough as a first line test. If glycemia is higher than 150 mg

Made a list to the technical aspects of the study (blood collection, transport and stability of the sample). The test adapted for screening study would glicated hemoglobin (HbA1C) . Blood collection in tube of EDTA has easy preservation, transport and good stability. It is possible to standardize, inclusive, the technique with use of filter paper , which would make easy still more the transport and would dispense use of ice for conservation.

i liked the comment about thinking about the population you are screening. Is iron deficiency common? This may interfere with interpretation of HbA1c. How are you going to collect and transport your glucose samples for measurement of plasma glucose so that you have minimal pre-analytical error? Are you going to collect into citrate tubes to abolish pre-analytical glycoysis?  Depending on the setting, I think use of a fasting sample with a reliable glucose meter may turn out to be just as good as more expensive tests - but this answer does not fit in with IFCC or similar recommendations. 

Dear Dr. Vivek...

If there is one "GOLD STANDARD" to diagnose diabetes with minimal ambiguity, that is the Oral Glucose Tolerance Test (OGTT) with 75g oral glucose.

However, as you know,  the limitations are.. the test is relatively complex and that you may need to send the patient back home and come once again in the fasting state and do the GTT. He might not turn up again. In this case, it won't be comprehensive. 

You have mentioned this ....

//  What would you do if you do not want to leave any patient of diabetes undiagnosed?// 

If you do not want to leave any patient undiagnosed, check any one of the blood glucose tests at the " first point of contact with the patient ".  

If the patient is in the fasting state, please do an FBS....

If the patient has taken his meal and he has come to you about two hours after it... do a PPBS.

If you encounter him in a random state not fitting into either of the two, do RBS test. 

In this way, you'll comprehensively cover all. You would not need to send the patient back and ask him to come again. He might not come back to you.

So.. if any one of the above tests is abnormal..... you detect "dysglycaemia". 

You may  proceed further that and do the remaining tests (FBS/ PPBS/ RBS / HbA1C ) & further work-up and treat the patient accordingly.

In a country of high prevalence like India, you'll be able to detect a significant number of patients by this approach.

In case where the test you did turns out to be normal, then you would need to go for the second test.... 

For example,.... if the FBS is normal... you can do either a PPBS or an GTT. 

PPBS is simpler than the GTT. 

If the PPBS is more than 200 mg/dl, that is diabetes.

If the PPBS is between between 140 and 199, what will the diagnosis be ?

In simpler terms, many people call it an IGT.

However, it need not be so. A patient whose PPBS lies in the so-called "IGT range" may turn out to be a diabetic if a GTT is performed. 

Similarly, if the PPBS of a patient is < 140 mg/dl.... we won't be able to call it simply as normal. This is because, a GTT done in the same patient may show him to be in the IGT range. 

HbA1C, especially in the lower ranges, is not a very good test to diagnose either IGT or DM. However, it is a good prognostic indicator.

Therefore..... in a nutshell......... if you are in a position to do a GTT, do it.

You'll have a wholesome picture.

You may workup further with PPBS and HbA1C.

If you encounter a patient and you are unsure if he would turn up for a GTT, do any of the three tests FBS, PPBS or RBS that seems fitting to you at that moment. 

In my opinion.. at any instance and anywhere... it is always better to perform a simple and available test, at the outset.

Thank you.

I have been involved in similar research in Nigeria and i will comment based on my experience. In a rural setting in which everyone that shows up will be screened. i advise you use FBG for those that will come in the morning hours without eating, RBG for those that will come at anytime and have eaten. I will rule out 2PPGT because from my experience in rural screening research, most of the subjects can not tell the exact time they ate their last meal  and you know 2PPGT should be 2 hours after meal.

From my experience on pregnant women, pregnant women usually show dysglycemia. Most of them show high FBS and/or RBS but but when subjected further to OGTT, some of them will show normal glucose value with OGTT. i advise you use OGTT for pregnant women to completely rule out gestational diabetes. But you have to convince the pregnant women to stay at once and complete their OGTT, tell them the implications of uncontrolled DM on their fetus. From my experience, most pregnant women will stay provided the research is for the well being of their unborn children.

@ S Charles Bronson: I agree with your statements fully. Go for the test FBS if the patient is fasting, go for RBS if the patient is non-fasting. If they are negative, go for the other tests like A1c and 2 hr PPG.

What I am planning is a prevalence study in a rural area. My objective is to find the prevalence of undiagnosed diabetes in a rural area. My sample size would be near 1000 randomly selected subjects from a population of 50000. 

Now, I need a reliable approach fitting within the ADA and IDF guidelines so that I can estimate the true prevalence of undiagnosed diabetes in this population.

Unfortunately, use of glucometers won't fit within the guidelines.

Remember, I have sufficient funds with me! The answer that I am looking for is one of these:

1. Shall I call the selected subjects to the subcenter in the fasting state for testing and perform FPG only?

2. Shall I perform FPG and OGTT when they come?

3. Shall I perform FPG, OGTT, and HbA1c in the same sitting?

4. Shall I do only one of the above tests and call them again if the test is negative to perform the other tests? Will they come again if they know that the first test is negative?

4. Do I go to their house and perform random blood sugar + A1c first? will they come fasting for FPG and OGTT if they know their RBS and A1c are negative??

Doing research if rural areas is not easy. I have to contact 1000 people through my field staff. I should be very clear of my methodology before I contact the first person. If I have only one shot, which one of the above choices is the best to achieve my objective stated above i.e. prevalence of undiagnosed diabetes in the rural population? Remember! I am not hitting for the tip of the iceberg, I want to go for the whole iceberg submerged under the water :)).

Please give your comments...

If you have sufficient funds, personnel, and all the needed resources, have them come to your facility (or mobile unit) and do all three tests.

But first you have to carefully select the criteria for inclusion/exclusion. Random is seldom random enough by all possible criteria.

Mainly fasting glucose and post prandial glucose

Nicky

We should not confuse "screening" with "diagnosis". Yes, to diagnose diabetes you need a FPG and/or an OGTT and perhaps an HnA1c (if you have a certified assay and have excluded anaemia etc).  However, in a rural setting one is looking for a simple test that can be done inexpensively on patients as they arrive. Many may come from far away and cannot come back fasting just to be screened. You will need to kiss a lot of frogs to find a prince! If your purpose is to screen for diabetes (as opposed to diagnosing it) then a simple random finger-prick glucose at time of presentation will suffice.  Depending on when the patient last ate (ort didn't eat) you could then decide if necessary to ask them to come back fasting for a diagnostic test. This would be the least expensive, least resource-consuming and least time consuming approach. It might be different in urban America, but this pragmatic approach works in rural communities in developing countries. And by it's very nature screening for any condition does not pick up every case. If screening can detect 90% of undiagnosed cases you are doing very well. Again, we need to have a different mind set between "screening" and "diagnosis". 

That's fine in an ideal world - maybe a suburban clinic in Boston. And possibly for Vivek's research project where funding and resources seem to be available. I was referring to a generic screening programme for underdeveloped rural communities..

@ Tausif, I am planning to study 45-75 year population. I am going to exclude the known diabetics. Do you have some suggestions for the inclusion/exclusion criteria?

Vivek,

Are males and females going to receive equal weight?

And speaking of weight, are overweight people going to be preferentially included? If so, you will have to decide about the cutoff BMI (or whatever criteria you select). Of course, in India not all T2D patients are overweight so you have to be mindful of that as well.

I have seen some projections of T2D in India and initially I didn't believe those numbers. But clearly, the disease is on the rise at an alarming rate. One would assume that unlike in the other places, you may not have to have 20 misses for every hit - there may already be a larger proportion of prediabetics, hence the utilitarian value of OGTT (given unlimited resources).

I have to wonder though that if T2D is not quite as pervasive in rural areas as in the cities because rural communities may have not encountered the switch to more sedentary and overindulgent life-style due to excesses our systems seems to sell. But I may be wrong in assumption.

@ Tausif Type 2 diabetes is quite common in rural areas. I am going to include both, males and females and since diabetes is frequent in non-obese, I will not use BMI cutoff.

Sounds good!

Best wishes...

Thank you:)

Vivek,

Last year I finished  screnning for unknown (undetected) T2DM with opportunistic method in family practice (the best setting for screnning). Target population were- patients aged 45-70 in total  population in 23 GP'practice. First step, we excluded patients with previously DM (9.4%). From remain patients we derived those with at least one risk as prediction for undetected T2DM (51% population aged 45-70): treating  hypertension and hyperlipidemia with medication,  high body weight, patients with famiy hstory of DM in parents and sibling, with gestational DM... (IDF). Date were derived from electronic medical record (age, sex, hipert. and hyperlipid), date of usual care (crude estimation of high body weight, family histori....) and from self report of patients during consultation. In 13 mounth response rate was 67%.   Yeald of DM was 2.1%.

The best predictor was high body weight, in cumulation with family history...

Diagnostic test FPG, is the  best initialy  test in epidemilogic study. The least one risk with one elevated FPG detected 1 DM in 7 risk patients and one IFG in 5 patients.

It is need to improve EMR records in order to better collection data and reduction miss data.

Best wishes

Prof. Marija Vrca Botica Phd.

Medical School Zagreb, Department of family medicine

Zagreb Croatia

@Vivek Chauhan...

// 1. Shall I call the selected subjects to the subcenter in the fasting state for testing and perform FPG only? //

FPG alone may miss out diabetes easily.

// 2. Shall I perform FPG and OGTT when they come? //

Ideal. 

If you advertise as a 'screening programme' beforehand... then I think quite a good number of people would come in the fasting state and FPG and OGTT can be done.

// 3. Shall I perform FPG, OGTT, and HbA1c in the same sitting? //

Very ideal.

Comprehensive. Wouldn't miss out anything. 

You may consider the same approach as above.

// 4. Shall I do only one of the above tests and call them again if the test is negative to perform the other tests? Will they come again if they know that the first test is negative? //

When the funds are available, this approach would be less efficient.

// 5. Do I go to their house and perform random blood sugar + A1c first? will they come fasting for FPG and OGTT if they know their RBS and A1c are negative?? //

This is an alternate approach to all that is said above.

This is a "field-activity" per se. 

When you go to their houses, you are sure to cover the majority. Please do it on a holiday or in the evening hours so that most of the people would be available. (personal experience doing the other way round). 

If you are detecting a diabetic, it's fine. If you are getting a negative test, then as you said, you may invite them to your health-centre. Some will come. Some will not.

As Dr. LARRY DISTILLER  said... we must  'screen' many people to find a few people with the disease.

In fact... even if the study is aimed at  1000 people... you will need to approach more than thousand people to get these 1000 people and then screen them and find the prevalence. Ideally, those people screened must represent the composition of the population studied.

Then the results may be extrapolated to the population at large.

I think... if funds and manpower are available, then the approach can be 'Method-5 followed by Method-3 or 2.' 

You may invite not only the negative result group but also the positive result group to your centre for remaining tests. This will strengthen your database for the study.

Thank you.