Monitoring The Blood Glucose levels

Monitoring The Blood Glucose levels

Dr. S.M.Sadikot.
Hon. Endocrinologist,
Jaslok Hospital and Research Centre,
Mumbai 400026

One of the major aspects of a good diabetic management is that the blood glucose levels should be optimally controlled. Whilst there are a few who may still question the role that hyperglycemia plays in the pathogenesis of the dreaded long term complications, most of the authorities are of the view that an optimal control of the blood glucose level will definitely help in retarding the progression of these complications even if one is not able to completely avoid the problems. By optimally, I would mean that they should be as close to normal as is possible without exposing the patient to the "peaks" of hyperglycemia as well as the "troughs" of hypoglycemia.

But how are we to judge the adequacy, or otherwise, of the control of the glucose levels? What are the parameters available which enable us to get a correct idea of the control?

Unfortunately, many of the so-called parameters which have traditionally been used to judge this control have so many shortcomings associated with them, that they should have no place in the modern management of diabetes. At the same time, one cannot accept all new methods of evaluating blood glucose control without inspecting their feasibility in our context. Having made this point, what are the methods that we commonly use to judge blood glucose control.

Presently, the following methods are in vogue :

  1. Testing the urine for the presence of glucose;
  2. Occassional blood glucose test done in an laboratory ;
  3. Estimation of Glycosylated Hemoglobin and serum Fructosamine levels ;
  4. Self monitoring of the blood glucose levels.

This is still the most common method used for estimating the blood glucose control ! I feel that using this estimation is in today's context mostly unacceptable, especially when other, better, parameters are easily available. Urine tests are associated with too many shortcomings to give any reasonable answer about the glucose control. At the same time, it would be worthwhile to examine the reason why urine testing for glucose was advocated as a means of evaluating blood glucose control as a time when other methods were not routinely available.

When blood flows through the kidneys, glucose that is present in it is filtered by the glomerulii. When this filtrate flows through the tubules, this glucose is reabsorbed back into the bloodstream and consequently, under normal circumstances, no glucose is found in the urine. The capacity of the tubules to reabsorb glucose is limited and if the amount of filtered glucose is more than this capacity, the excess glucose would be found in the urine, which would then test positive for the presence of glucose. As the amount of glucose found in the glomerular filtrate is dependent on the amount of glucose present in the blood, it follows that in cases where the blood glucose levels are increased, as in diabetes, more than normal amounts of glucose would be filtered out and if this overwhelms the reabsorptive capacity of the tubules, glucose would be found in the urine.

This has lead to the concept of "renal threshold". This is the blood glucose level beyond which so much glucose would be filtered out by the glomerulii that the urine would test positive for the presence for glucose. The textbooks mention the renal threshold for glucose to be 180mg%. In other words, if the blood glucose levels were to increase above 180mg% then the urine would test positive for the presence of glucose.

In view of this, it is easy to understand how testing the urine for the presence of glucose came to be used as a parameter to judge the blood glucose levels in the olden days. In the absence of wide availability to test for the blood glucose levels, it was felt that if the urine showed the presence of glucose, then it could be surmised that the blood glucose were above 180%. This is where the first fallacy creeps in. One must realise that the figure of 180% is just an average approximation and that the threshold differs markedly in every individual patient. Thus, there are those non-diabetics who will show the presence of glucose in their urine, even when the corresponding blood glucose level may be only 100mg% ! Conversely, some known diabetics may not show the presence of glucose in the urine, although their blood glucose levels may have reached as high as 300mg%. Most of the others would come in between these two extremes. Therefore, if urine testing is to have any relevance to judging the blood glucose control, then the renal threshold for every individual patient will have to be charted ! Unfortunately, the method for estimating the renal threshold is quite cumbersome and one should not be under the misconception that it just a simple matter of checking the urine for the presence of glucose and estimating the corresponding blood glucose levels. This "simple" method will invariably give wrong idea of the renal threshold as we shall discuss later.

Even if we were to take the trouble to estimate the renal threshold for every individual patient, we are then faced with the problem that this value does not remain constant for that individual for all times. The threshold value changes with age, pregnancy, any kidney disease including diabetic nephropathy. More importantly, and this is not too well known, changes in the blood glucose level themselves can effect a change in the renal threshold values ! It is now well accepted that when the blood glucose levels are very high, the renal threshold values tend to be lower and this then increases gradually as the blood glucose levels are brought under control. In other words, the very parameter that we use to judge blood glucose control, itself undergoes changes as the blood glucose levels change !

It should be obvious that such a labile renal threshold value should not have any important place to play in judging the blood glucose control in the modern context.

There are many more reasons why we should not use urine glucose tests to evaluate blood glucose control. For the purposes of discussion, let us take a theoretical patient and accept that his renal threshold value for glucose is 180mg%. What this implies is that if the patients blood glucose values are below 180mg%, then no glucose would be found in the urine. Often, we test the urine in the fasting stage and feel gratified that is shows no glucose. But are we justified in accepting that this signifies good control ? It is possible that the blood glucose levels may be 165mg% and this value in the fasting stage would definitely be unacceptably high ! Looking at this from a slightly different angle, when the urine shows the absence of glucose, it would mean that the corresponding blood glucose level could range, in theory, from 0mg% to 180mg%, a complete range from the absurdly low to the unacceptably high. How is it possible to accept such a parameter to judge blood glucose control, leave alone making treatment changes !

Thus, testing the urine for the presence of glucose as an indirect parameter to judge blood glucose control is, at best, a crude method and should be accepted as such.

Even if we are to accept that urine tests are only crude method and give us only a rough idea of the blood glucose levels, we would be erring. It should be realised that urine test for the presence of glucose does NOT give an estimate of the blood glucose levels at the very time of testing the urine. In simple terms, we test the urine for glucose, say, 2 hour after a meal, as patients are often asked to do. Let us for the sake of discussion accept that the renal theshold of the patient is 180mg%, 1 + urine sugar corresponds to 200mg%, a 2 + to 250mg% and so on. The patient tests his urine after two hours of a meal and finds that his urine shows a 2 + presence of sugar in the test. Can this be taken to mean that his blood glucose level 2 hours after the meal is 250mg% ? This could be completely off the mark. In this case, one may accept the blood glucose level at that time to be 250mg%, only if the urine that we are testing is that which is passed by the kidneys precisely 2 hours after the meal. But the urine that we are, in reality, testing is NOT the freshly passed urine but a mix of all the urine that has accumulated in the urinary bladder since the last time that the patient had voided. In other words, urine testing cannot give us a true idea of the blood glucose levels at the time that we test the urine. This point, though of utmost importance, is often forgotten by many of us.

One way out of this problem would be to ask the patient to completely empty his bladder, say, 5 minutes before the test and discard this urine. He would then pass a fresh sample of urine after about 5 minutes and this could then be construed to be akin to a freshly passed specimen from the kidneys and this may give an idea of the corresponding blood glucose levels in an ideal situation. To do this, that is to completely empty the bladder and then pass some more urine within 5-10 minutes is more easily said than done. Most, if no all patients find this extremely difficult (not to say inconvenient) and many will refuse to do the test at all. I always ask doctors who advise their patients to do this, if they have ever tried to do it themselves. Only then they will realise how troublesome these instructions are ! Often, the patients are advised that after they have discarded the first urine completely, they should drink lots of water. This is done in the hope that it will help them offer the second sample easily. Unfortunately, the very fact that the patient drinks a lot of water, makes him pass dilute urine, and this dilution in itself, changes the renal threshold !

More commonly, many diabetics have some amount of autonomic nerve involvement and this leads to an inability to completely empty the urinary bladder. Such patients, and there are quite a few of them, would find it impossible to offer a second, fresh sample of urine, as any urine that they pass would invariable have been mixed with some of the urine that has remained in the bladder due to incomplete emptying caused by the neuropathy. This inability to completely empty the urinary bladder caused by the neuropathy, leads to some urine stagnating in the bladder and this urine is particularly prone to get infected especially in diabetics with a high blood glucose levels who would pass an increased amount of glucose in the urine. The infecting bacteria will utilise the glucose that is present in the urine and when we test the urine, under such circumstances, one would find an absence of sugars. This would lead to a completely erroneous idea that the blood glucose control is quite acceptable. In reality, the blood glucose may be very high but this would not be reflected in the urine tests just at the time when it would be important to correctly assess the diabetic control so that adequate measures can be taken to manage the blood glucose levels and help in the eradication of the urinary tract infection.

Finally, let us briefly discuss the methods that are widely used in many parts of our country. The two most commonly used methods are the older Benedict's test and the "stick" test. The Benedicts test is the older test in which a certain amount of urine is added to a measured quantity of the blue copper sulfate solution, and the mixture is boiled after which the change in colour of the solution and precipitate, if any, is noted. This change in the color is supposed to indicate varying levels of sugars in the urine.

This test is still very widely used here, the main reason being its cheapness. But, it is also very non-specific and will show a positive reaction with many a reducing substance besides glucose. To give a common example, breast feeding mothers excrete lactose in the urine and the Benedicts test will show a positive result even though the blood glucose of the mother are completely normal and she is not spilling any glucose in the urine. There are numerous drugs and medications that also interfere with the correct result as we shall discuss later.

The "stick" has a special paper attached to one end. This paper is impregnated with certain enzymes and when it is dipped into the urine, the change in the color of the paper will denote the presence of, and the varying amounts, of glucose present in the urine. It has the advantage that it is quite specific for glucose and also does not require all the paraphernalia that one needs for the Benedicts test like test tubes, dropper, a source for boiling the solution etc. At the same time, it is costlier than Benedicts test, although the cost can be halved by cutting the strip longitudinally so that one can use each stick twice.

It, too, has drawbacks, the chief one being that there are many drugs and medications which interfere with the correct result. These drugs may interfere with false positive or false negative results. Just to show how common the problem can be, two rountinely used drugs which can interfere with the results are aspirin and Vitamin C. Thus, doing urine tests for the presence of sugar when the patient is on any of these tablets would give a completely wrong result. These drugs are so commonly used even by patients themselves that doctors are often unaware that the patient is on these medications. I do not think that it would be possible to find a single diabetic patient who is not taking a vitamin tablet and aspirin is a common household remedy for aches and pains and also now routinely prescribed as a preventive measure against atherosclerosis.

Does this imply that testing the urine for the presence of sugars has no role to play in the modern management of diabetes? Whilst, it would seem to be so, especially when other better methods are now available to judge blood glucose control, it should also be remembered that many of these methods are not easily accessible or affordable to many of our patients. They also have their inherent drawbacks. Therefore, I feel, that in spite of all the problems associated with the use of this mode of testing, it will continue to be used by many patients. One of the main reasons is the relative cheapness of the method, but more importantly, due to the fact that most of the patients are unaware of the useless results that one can get from such tests.

Urine testing will continue to be important to look for the presence of ketones and albumin in so far as diabetes management is concerned as well as to rule out mild, asymptomatic, urinary tract infection. Some authorities feel that urine testing for glucose may be an adequate parameter to judge blood glucose control in the elderly diabetic who does not have any other complication. The reasoning behind this is that elderly diabetics tend to have a higher threshold and are also quite averse to having the blood tested frequently. Under such circumstances, if the urine shows the presence of glucose, then it could be surmised that the blood glucose levels must be quite high and need to be brought down. Conversely, many young diabetics have a low threshold and a consistent absence of glucose from their urine should be a pointer to the possibility that they many be undergoing subclinical hypoglycemia. Frankly, I am not convinced about either of these aspects. From a practical compromise view point, a newly diagnosed patient with a blood glucose of 400mg% and a urine glucose of 4+, may be monitored by testing the urine glucose till it comes down slightly, to say 2+, and then one must shift to monitoring the blood glucose levels.

Finally, if one is going to utilise these tests to evaluate diabetes control it would be better to use the "stick" method, as this is somewhat specific for glucose. At the same time, it is essential that one be aware of the pitfalls associated with these urine tests so that one is not lulled into a false sense of complacency about the glycemic control when the tests for the presence of sugars in the urine are negative. Conversely, drastic changes in the treatment schedule should be avoided just on the basis of urine test results

Urine testing, is at best, a very crude method for evaluating blood glucose control and should be accepted as such.

This is the parameter that is most commonly used to judge blood glucose control. Patients are usually seen at 2-3 monthly intervals and the fasting and the 2 hour post lunch blood glucose levels are estimated. On the basis of these levels, it is presumed that one can judge blood glucose control, compare this with previous values and judge whether the control is better or has deteriorated and, also make adjustments in the treatment regimen of the patient. Such occasional blood glucose estimations cannot give adequate information to allow us to rationally make these suppositions.

The results of any blood glucose estimation are subject to quite a few variable factors which may interfere with the correct estimation and give a variation on the same sample of blood to the tune of 30-50mg%! Firstly, it should be clear that, at present times in our country, there is no accepted standard method to estimate the blood glucose levels. Many laboratories and even some hospitals still continue to use the outdated Folin-Wu method for the estimation of the blood sugar levels. This is the oldest method that was used to estimation of the blood sugar levels and its real place should be in the archives or a museum! It should never in the present time, be used to estimate the blood glucose levels. Like the Benedicts test for urine sugar, it is not very specific and estimates not only glucose but evaluates the presence and amount of any reducing substance in the blood. Therefore, besides, many drugs and medications, even substances which are normally present in the blood stream will give rise to abnormal results.

The Somogyi-Nelson and the Hoffman methods although better that the Folin-Wu method should also be discarded and only the methods which will give the true glucose values like those using the Glucose oxidase system should be used. When the patient goes for his blood glucose one must make sure that the laboratory does utilize the oxidase method.

Until we reach a stage where the method of estimating the blood glucose levels becomes unversally standardised, we would have to take the method by which any blood glucose is reported into consideration whilst interpreting a report especially if the report is to be compared with any previous results. I have often come across patients who come with two or three previous results and it becomes impossible to compare results because the methods by which the blood glucose have been estimated each, time are entirely different! One of the reasons why this occurs, besides the absence of a standard method, is that many of our patients do not have a fixed place where they get themselves tested.

Often they go to the nearest place available to where they are on the day the test has to be done. Thus one of the reports may be from the hospital where the patient may be seen, the next may be from a laboratory near the house of the patient and the third, especially if it is a post meal report, would be from a place of work of the patient. It is possible that these places may use different methods.

Even if the same method is used, other variables come into play. Some laboratories estimate the glucose levels in whole blood whilst others may do the estimation on plasma or even serum. It is quite well known that the glucose levels estimated in the plasma can be about 15% (not 15mg%) more than the glucose levels that would have been reported if the whole blood had been used instead of plasma. Therefore, not only is the method of importance, but also the material on which the estimation is done. Unfortunately, if you check your reports, how many of these note down whether the glucose was estimated on whole blood, plasma or serum ? With such a significant discrepancy of 15% between plasma glucose and whole blood glucose, even in the same sample, is it really possible to get a correct idea of the true blood glucose levels, leave alone any worthwhile comparison with previous values.

Even if the methods and the material on which the glucose is estimated are the same, the problem is still compounded by other factors. Did you know that if the blood is collected from the back of your hand, the glucose values will be about 10-15mg% more than if the blood has been collected from the front of your elbow?

Some laboratories have started using the pinprick method to collect a drop of blood from the fingertips and this capillary blood is known to show glucose values which are again anywhere between 10-20mg% higher than what would be estimated in a simultaneously collected venous sample! At the same time, it should be added that the newer glucose meters come calibrated to show the equivalent venous plasma glucose levels. One would have to make sure that the meter which one is using is properly calibrated and know what reading it gives in order to correctly analyse the test results.

The time that elapses between the collection of the blood sample and its estimation also affects the results. The greater the interval between the time of collection of the sample and the estimation, the lower will be the glucose result. This is in spite of whatever preservative that many laboratories add. It is obvious that there is no standard time interval at which laboratories do the glucose estimation.

In view of such variability in the test results, it is not possible to make any valid comparison between the latest and any previous report. Thus, it is not possible to judge whether the blood glucose levels of the patient are better or worse than before. More importantly, I do not feel that it is feasible to adjust the treatment on the basis of such variable, and occasional, blood glucose reports. We have seen that due to the presence of factors that cause so much variability in the blood glucose estimation, the same sample may show a discrepancy of around 30-50mg% ! Even if we are to take a conservative estimate and accept that the variation in the result is around 25mg% , what does this imply in clinical terms ? Let us suppose that we have estimated a fasting blood glucose level and this has been reported as 125mg%. If we accept a variation of even 25mg%, either way, it means that the true blood glucose could range from 100mg% to upto 150mg% ! Whilst a figure of 100mg% or even 125mg% in the fasting stage is acceptable, 150mg% isn't. So what adjustment in the management is to be made ? Are we to accept the result as acceptable and continue the same management ? If we were to do this, and the real blood glucose value is around 150mg%, then it is obvious that we are under treating the patient ! If we try and reduce the 125mg% still further and if the true blood glucose values were to be around 100mg% then we may be exposing the patient to the hazards of nocturnal hypoglycemia ! We would then be over treating the patient. Therefore, for the same figure of 125mg%, any decision that we make to treat the patient, may lead either to over treatment or under treatment of the patient.

Such, occasional blood glucose estimates cannot be used to correctly adjust the treatment and these estimates should really be called guesstimates (as the Americans with their penchant for joining two words call it), and accepted as such. Any change in the treatment made on the basis of these random blood glucose reports can, at best be based on intelligent guess work, and not on solid clinical grounds.

I would also like to point out the fallacy of presuming that one can judge blood glucose control on the basis of these occasional blood glucose reports. There is no way that such reports can give us an idea of the control. After all, what does a blood glucose report tell us? Say that the sample has been collected on 1/10/2002 at 3 p.m. The test report, even if it were to give a truly correct estimate, only allows us to judge the precise blood glucose levels at the precise time that the sample was collected. It does not reflect the blood glucose levels even 15 minutes previously or the blood glucose levels 15 minutes after the blood has been collected. Let us say that the patient may give the blood sample and go out and eat a lot of sweets. But the report may show, possibly, an acceptable level. Does this imply that the patient's control is adequate ?

When we talk of control we mean the average blood glucose levels at all times and not only at the time of the test ! Let us say that three months have passed since the time of the last test for blood glucose levels. A particular patient does not keep to any diet, takes his medicines only when he feels like it and generally in poor control in the sense that his blood glucose levels at most times during these three months are unacceptably high . Two or three days before he knows that he has to go for the test, he keeps to a very strict diet, takes a heavier dose of the medications and somehow brings his blood glucose down. I know of patients who go for a post lunch test without having any lunch ! I also have come across patients who when they have to go for the fasting blood glucose estimation will inject themselves with some regular insulin early in the morning and then go for the test after about 2 hours of the injection. With all these manipulations, often the blood glucose results come in the acceptable control, we would obviously be mistaken as the patient has been in poor control throughout the three months and has just brought down the blood glucose levels prior to the time of the test !

Conversely, we take the example of a patient who has kept to his diet, exercise and medications in the months since the last test and has been in acceptable control in he sense that his blood glucose levels have been in a good range throughout the three months. A few days prior to the test, he may have come down with flu and this will show up in the test as a high blood glucose level. Are we then to accept that this patient has been in poor control for the three months since the time of the previous test ?

Blood glucose control is to be judged over a span of time and it is obvious that occasional blood glucose test will NOT give us any idea about the true control. Thus, such blood glucose reports do not have any role to play in judging blood glucose control over any span of time, and it may not even be feasible to correctly adjust treatment on the basis of such random reports.

An estimation of the glycosylated hemoglobin (HbA1c) levels allows us to assess the blood glucose control over the previous two months. In common with many other proteins, some of the amino acids in the hemoglobin molecule can bind the glucose molecules to which they are exposed. In very simple terms, let us say that for every five glucose molecules, one would get attached to an amino acid. If ten glucose molecules are present, then it would follow that two amino acids would get glucose attached to them, then it would mean that the amino acids had been exposed to a greater than normal amount of glucose!

Therefore, if one were to estimate the amount of amino acids (protein) to which glucose has become attached, one would be able to gauge the amount of glucose to which the protein had been exposed. This is the basic principle behind the use of glycosylated hemoglobin levels in judge blood glucose control. Since the hemoglobin is found in the circulating red blood cells, the glucose to which it would be exposed would be the glucose present in the blood. If more than the normal amount of hemoglobin becomes glycosylated, this would imply that if had been exposed to more than normal amounts of glucose and therefore, the blood glucose levels had been higher than normal !

How is the figure is 60 day or 8 weeks arrived at? A red blood cell has an average life span of 120 days, and therefore, more red blood cells are being constantly made and released into the bloodstream. If we were to analyse the red blood cells taken from any blood sample, it would follow that some of the red blood cells would have been in the circulation for almost 120 days and others would be newly released cells. The other red blood cells would have been in the circulation for periods in between these two extremes. The longer the red blood cells has been in circulation, the greater has it been exposed to the blood glucose and the greater would be the amount of glucose that is attached to it. Conversely, newly released red blood cells would not have any glucose attached to them. In order to average out these extremes, it is presumed that the average bed blood cell has been in the circulation for about 60 days (the midpoint of 0-120 days), and therefore the amount of the hemoglobin that has become glycosylated would give an idea of the blood glucose levels over these 60 days. If the average blood glucose levels has been normal, than the estimated amounts of glycosylated hemoglobin would be within normal limits. If hyperglycemia has been present, than the degree of this hyperglycemia would be reflected in the increased amount of the hemoglobin that would become glycosylated. In other words, the levels of the estimated HbA1c reflect the prevailing blood glucose levels over a period of 60 days, or 8 weeks.

If all this sounds confusing, another way of looking at this would be to imagine that a needle has been inserted into the blood stream and this needle leads into a computer. This computer estimates the blood glucose levels every second and all the results are stored in the memory of the computer. At the end of 60 days, all the results are added up and the computer gives an average figure for the prevalent blood glucose levels over this time. HbA1c levels are akin to this average blood glucose levels.

Estimation of the HbA1c levels has some definite advantages over the occasional testing for blood glucose levels at intervals of 2-3 months, that we have previously discussed. It was seen that even if the numerous drawbacks associated with the estimation of the blood glucose levels are overcome, such occasional test for the blood glucose levels will only tell us about the blood glucose levels and glycemic control at the precise time of the testing! When we talk about control of the levels we mean that the blood glucose levels would under acceptable control at all times and not only at the time of the test. By giving an idea of the average blood glucose levels over a period of 60 days, HbA1c determination is a better indicator of blood glucose control.

Let us take the same two examples that we considered to show the fallacy of trying to judge blood glucose control by the occassional test for blood glucose levels. We took the examples of a patient who does not keep to his diet, does not exercise and takes his medications only when he feels like it. Two or three days before he knows that he has to go for the test, he keeps to a very strict diet, exercises heavily and also takes a larger than necessary dose of medications. With these manipulations, he manages to bring his blood glucose levels down so that when the blood is tested for the glucose levels, they may be found to be in the acceptable range. Before we could estimate the HbA1c levels, this would imply that the control of the patient was optimal although this is far from true. If in such a patient, the glycosylated hemoglobin levels were to be estimated, these would have been seen to be much higher than normal as they reflect the blood glucose levels over the previous 60 days when the patient had obviously been in poor control and all the wiles of the patient would not allow him to mask the true state of his control !

Conversely, the patient who has been in good control, but due to some reason like suffering from a 'flu' attack, shows a higher than acceptable blood glucose level would manifest a basically an acceptable HbA1c level as the deterioration in the blood glucose values in the past couple of days would not significantly alter the levels. This would enable us to understand that the patient is basically in good control but possibly for some reason presently has a high blood glucose levels. This would make us search for the cause of the deterioration rather than make a blind increase in the medications which may expose the patient to a risk of hypoglycemia.

The estimation of HbA1c levels is also more convenient for the patient. The blood for the estimation can be collected at any time and the patient does not have to be in a fasting stage or at any fixed interval after a meal. Thus, the blood collection can be done even when the patient comes for his routine checkup at any convenient time.

You must make sure that what is estimated is the HBA1c, which is the accepted norm today.

It is possible to get a fairly good idea of the average blood glucose levels over the past two months based on a correctly carried out HbA1c reading.

Estimation of the Glycosylated Hemoglobin Levels

Although, HbA1c levels do give an excellent idea of the blood glucose control over the previous 60 days, there is one other, clinical aspect that needs to be considered. Usually, the patient would come to the doctor after a period of two months with blood glucose tests and also HbA1c levels. If these levels are seen to be in the high range, then it is obvious that the patient has been in poor control over the previous 60 days. One can then only take corrective measures and even when these are taken, changes in the HBA1c levels would only be reflected after a further period of 60 days. Thus, at least four months would pass before the patient was known to be back in the good control. In case, as often happens, the HbA1c level at the second visit is not yet normal, then a further period of two months would be lost! In my opinion, if we are to talk of the correct, modern management of diabetes, then one cannot allow the patient to remain without optimal control for such a long time just because we are waiting for the glycosylated hemoglobin levels to be normalised. Rather than take corrective measures, it would be better to take immediate measures to optimise the blood glucose control if this is seen to be poor, rather than wait for a few months for changes in the HbA1c levels.

One way to overcome this problem is to estimate serum fructosamine levels. This is basically the same as HbA1c estimations, but estimates the attachment of glucose molecules to albumin in the blood. Due to the relatively shorter half life in the blood of albumin, serum fructosamine estimations give an idea of the average control over the past fortnight. Unfortunately, this estimation is not widely available and should preferably be done in special situations such as pregnancy where one wants a tight control throughout, without having the luxury to wait for 2 months to decide if the patient is in good control as would be seen with HbA1c estimations!

The other, and definitely better alternative is for the patients to self monitor their blood glucose levels (SMBG).

SMBG by the patient himself has, in my opinion, completely revolutionised the modern management of diabetes.

Such self monitoring tests are extremely convenient for the patient. He can carry out the test at home, work or even elsewhere without having to take the trouble to go to any laboratory. Thus, these tests can be done at varying times of the day and night so that one can get an idea of the blood glucose profile throughout the day. As the test can be done so conveniently and frequently, one can take corrective measures to optimise the blood glucose levels almost as soon as they occur without having to wait two months for the HbA1c levels or take the recourse either having to go a laboratory often and at odd hours or even get admitted to a hospital!

Ideally, all patients should self monitor their blood glucose levels, using test trips and meters which are now available.

If it is not feasible, SMBG should be carried out by :

  1. All patients on insulin therapy, especially those on multiple dose regimens;
  2. patients with widely fluctuating blood glucose levels;
  3. patients prone to severe ketosis or recurrent hypoglycemia;
  4. those manifesting hypoglycemia "unawareness";
  5. patients in whom a "tight" control is essential, i.e. pregnancy, etc.; during acute illness;
  6. in the perioperative period;
  7. those with abnormal renal thresholds

It should be clear that in many these conditions one would not have the luxury of waiting for HBA1c levels, it may not be possible to go to the laboratory at odd times and so frequently, and in many of such conditions, the urine tests are fraught with too many errors for them to be of any use.

I am often asked as to how often a patient should monitor the blood glucose levels. I feel that patients on multiple insulin injection regimens, patients with critical problems such as those with sight threatening macular edema and infections, and pregnant patients in whom a tight control is mandatory should preferably monitor their blood glucose two to three times daily. They should be taught how to adjust their doses depending on the levels. If in spite of this, the blood glucose remain high, them they should consult their doctor at once. It is clear from this that poor control would be detected early and corrective measures can be taken at the earliest, which is as it should be.

Patients who may not fall in these "critical" categories, are asked to test the blood glucose daily at different times for about 1-2 weeks or until they are confident about the method of testing and are optimally controlled. Once this is done, they are asked to test the blood twice or thrice a week just to make sure that the control remains at an optimal level. In case, they find that the control has deteriorated, then they should revert back to a more intensive frequency of testing and take the corrective measures until the control is back to the original optimal levels. Thus, here again, one would be fast off the mark in initiating corrective measures before the hyperglycemia has persisted for any significant time. At the same time, frequent testing for the blood glucose levels would also help in preventing any hypoglycemic attack in our quest for optimal control. Once the control has been optimised, the frequency of testing can be reverted back to twice or three times a week.

All doctors treating patients who have diabetes must use a glucose meter as a part of their medical equipment, just as they carry a stethoscope and a BP apparatus. Not only will this allow you to diagnose diabetes in your patients at an early stage before complications have set in, but even in your patients who are known to have diabetes, you would be in a unique position to test their blood glucose levels at different times of the day. This would allow you to get a profile of the blood glucose values, allowing a more rational adjustment of their treatment.

Carrying a meter with you would also be of help in certain emergencies which you may attend. Let us say that you are asked to see a diabetes patient at home, who is comatose. Is this hypoglycemia or is this caused by a hyperglycemic state? Correct and early treatment can save the person's life and this differential diagnosis can easily be made if you have a glucose meter with you. One should also remember that hypoglycemia can mimic many other critical states such as strokes etc., If you find a low blood glucose in such a patient and are able to administer glucose, the patient may show a " miraculous" recovery of the paralysis was due to hypoglycemia!

One problem with the use of glucose meters and strips is that the strips may spoil rapidly in the high ambient temperatures and the degree of humidity that occurs in our country. This is in spite of the desiccating materials that are incorporated in the bottles. Although manufacturer's advise that the strips can be used for about 3-4 months after the bottle is opened, in my experience, the strips spoil within 2-3 weeks. I hope that the manufacturers of the strips will make the individually packed strips available here soon. This should not be problem as they do market them in such individually wrapped foils abroad. Although there may be some additional cost, patients would rather pay this little extra rather than have their money wasted on spoilt strips. More important than the wastage of the money, is the fact that if the strips are spoilt even a bit, they may register low readings and their may be a false sense of complacency that the blood glucose levels are under good control.

Patients especially if they have to test themselves often complain about the pain of pinpricks. This usually occurs when the older all metal triangular lancets are used. A number of fine point, plastic coated lancets are available which are definitely easier to use and less painful especially if used with a companion lancet plunger. Even if the lancets are not.used, disposable needles will do fine. I find that the thinner needles, like No.23 give a much better drop of blood as compared to the thicker needles. Pricking on the lateral aspect of the middle three fingers further to the distal finger joint often causes less pain.

One fear that has been expressed with the routine use of self monitoring is that one may make patients neurotic about their blood glucose control. Such patients have been reported and have been called as "blood glucose junkies". A rational discussion about the need and the importance of self monitoring, the need not be to be very upset about a high reading but to calmly carry out the corrective treatment and a more humane and friendly approach by the doctor can go a long way to protect against this. There are some who will get very emotional about testing, but these are patients who will invariably have an over reaction not to the self monitoring of the blood glucose levels, but to the fact that they have diabetes per se. I do not think that self monitoring should be blamed for this tendency. The benefits of self monitoring far outweigh the occasional patient who may over react to this method for judging blood glucose control.

In conclusion, self monitoring of the blood glucose levels by the patient along with an estimation of the HbA1c level at 2-3 monthly intervals allows us to offer the most optimal method to judge blood glucose control at present.

Most importantly, it must be realized that monitoring blood glucose levels is not the same as monitoring diabetes.

Besides glycemic control, optimal monitoring in diabetes, implies optimising weight, blood pressure, lipid abnormalities, and importantly, the diagnosis of the presence of long term complications in their early, initial stages.