|Research Connecting Organ Damage with Blood Sugar Level|
The studies you will read below, some of which are not cited in the AACE guidelines, make a cogent case that post-meal blood sugars of 140 mg/dl (7.8 mmol/L) and higher and fasting blood sugars over 100 mg/dl (5.6 mmol/L) when found in association with those higher than normal post-meal blood sugars, cause both permanent organ damage and the worsening of diabetes. Some of this data also suggests that maintaining an A1c of 5.7% to 6% is much safer for people with diabetes who wish to avoid developng diabetic complications.
NOTE: All blood sugar levels discussed on these pages refer to plasma calibrated meter readings, which are the ones today's meters use.
Nerve Damage Occurs when Blood Sugars Rise over 140 mg/dl (7.8 mmol/L) on Glucose Tolerance TestsA study conducted in the neurology clinic at the University of Utah examined patients who came to the clinic complaining of peripheral neuropathy of unknown origin. Peripheral neuropathy is the medical term for a kind of nerve damage which causes pain, tingling, "pins and needles," numbness or burning sensations in the hands and feet.
The University of Utah neurologists found that patients who were not known to be diabetic, but who registered 140/mg or higher on the 2-hour sample taken during a glucose tolerance test were much more likely to have a diabetic form of neuropathy than those who had lower blood sugars. Even more telling, the researchers found that the length of time a patient had experienced this nerve pain correlated with how high their blood sugar had risen over 140 mg/dl on the 2-hour glucose tolerance test reading.
It is important to note that this study also showed that only the glucose tolerance test results corresponded to the incidence of neuropathy in these patients, not their fasting blood sugar levels or their results on the HbA1c test This is significant because most American doctors do not offer their patients glucose tolerance tests, only the fasting glucose and HbA1c tests that fail to diagnose these obviously damaging post-meal blood sugars.
Increased prevalence of impaired glucose tolerance in patients with painful sensory neuropathy. Singleton, JR Smith AG, Bromberg, MB Diabetes Care 24 (8) 1448-1453 2001.
A second study performed by neurologists at Johns Hopkins confirmed these findings. Fifty-six percent of their patients who had neuropathy of unknown origin were found to have abnormal results on their oral glucose tolerance tests. When nerve conduction studies were run, doctors sorted their patients into those whose glucose tolerance test results were normal, those whose 2-hour glucose tolerance test readings fell between 140 mg/dl and 200 mg/dl (7.8 mmol/L and 11.1 mmol/L), who were diagnosed with impaired glucose tolerance (prediabetes), and those diagnosed with diabetes because their two hour test results were over 200 mg/dl (11.1 mmol/L)
They found that patients who had impaired glucose tolerance had experienced damage to their small nerve fibers, while those with diabetic-level test results had more involvement of the large nerve fibers.
The spectrum of neuropathy in diabetes and impaired glucose tolerance. C.J. Sumner, MD, S. Sheth, MBBS MPH, J.W. Griffin, MD, D.R. Cornblath, MD and M. Polydefkis, MD; Neurology 2003;60:108-111.
Yet another study, conducted at the Mayo Clinic in Scottsdale, AZ and published in August 2006 confirmed these results. In addition this study documented that the fasting plasma glucose test identified only half as many people with abnormal blood sugars as did the glucose tolerance test.
Value of the Oral Glucose Tolerance Test in the Evaluation of Chronic Idiopathic Axonal Polyneuropathy. Charlene Hoffman-Snyder; Benn E. Smith; Mark A. Ross; Jose Hernandez; E. Peter Bosch. Arch Neurol. 2006;63:1075-1079.
Another study published in 2008 which analyzed data from the MONICA/KORA study of a large population in the city of Augsburg, Germany, cast more light on how blood sugar levels relate to neuropathy. The researchers in this study gave glucose tolerance tests to the a group "normal people" selected from a non-diabetic control group. They found that 23.2% of these "normal" people had Impaired Glucose Tolerance and 35.9%)had Impaired Fasting Glucose. Only 41% had truly normal blood sugars even when using the very lax ADA diagnostic criteria.
In the study population as a whole (normal and diabetic combined) neuropathy was found in:
28.0% in the diabetic subjects
13.0% in those with Impaired Glucose Tolerance (defined as 2 hr glucose tolerance test results greater or equal to 140 mg/dl (7.8 mmol/L)
11.3% in those with IFG (Fasting glucose > 100 mg/dl (5.6 mmol/L))
7.4% in those with Normal Glucose Tolerance (2 hr OGTT < 140 mg/dl (7.8 mmol/L).
This study also found that having peripheral artery disease (PAD) added to the likelihood of a person with diabetes developing neuropathy
Prevalence of Polyneuropathy in Pre-Diabetes and Diabetes Is Associated With Abdominal Obesity and Macroangiopathy Dan Ziegler et al. Diabetes Care 31:464-469, 2008
NOTE: It is important to keep in mind that the readings people get on glucose tolerance tests may be lower at two hours after consuming the glucose than the readings they get while eating meals containing carbohydrates that require digestion. The speedy absorption of the glucose that occurs during a glucose tolerance test often causes reactive hypoglycemia in people with prediabetes, giving them what look like normal numbers. But in daily life they may be experiencing blood sugars well above normal for an hour or two after slower-digesting meals. Unless you live on a diet of pure glucose, this is an important finding. Because it is the number of hours your nerves spend exposed to high blood sugars that appears to damage them.
Sadly, there are no studies where subjects with neuropathy were given meal tests rather than the highly artificial oral glucose tolerance test. However, anecdotal reports from people with neuropathy who have lowered their post meal blood sugars suggest that keeping blood sugars under 140 mg/dl (7.8 mmol/L) at one hour after eating meals can slowly reverse neuropathy.
Unfortunately, since most doctors are not aware that it is possible reverse neuropathy by lowering blood sugars to normal levels or believe it is impossible for people with diabetes to lower their blood sugars to normal levels at all, few doctors suggest that patients with neuropathy treat it by striving to achieve normal blood sugars. Instead, they prescribe the very dangerous brain drugs, gabapenin (Neurontin) and pregabalin (Lyrica) which may relieve neurological pain to some extent but don't in any way heal damaged nerves. These drugs can have devastating side effects, so before you try one, try lowering your blood sugars to normal levels for a few months. You can learn how to do that HERE
Keeping Blood Sugars Below 140 mg/dl at All Times Improves Survival of Seriously Ill PatientsDr. Krinsley decreased the mortality of a group of critically ill patients by 29.3% simply by keeping their blood sugar below 140 mg/dl at all times using insulin. This intervention also cut down the incidence of kidney failure and shortened the patients' stay in the ICU. In numbers this means that 45 people out of a group of 800 left the hospital alive who would have died had they adhered to the ADA's standards for blood sugar control.
Krinsley, James, Effect of an Intensive Glucose Management Protocol on the Mortality of Critically Ill Adult Patients. Mayo Clinic Proc. Jan 2004, p. 992-1000.
Beta Cell Dysfunction Begins at 2-hr OGTT Test Readings Over 100 mg/dl (5.6 mmol/L)When a team of Italian researchers led by A Gastardelli started examining beta cell response to glucose in people with normal blood sugars, they discovered that a small amount of beta cell dysfunction began to be detectable in people whose blood sugar rose only slightly over 100 mg/dl on a 2-hour glucose tolerance test. The beta cells are the cells in the pancreas that produce the insulin your body uses to control your blood sugar.
Analyzing their data further, they found that with every small increase in the 2-hour glucose tolerance test result, there was a corresponding increase in how much beta cell failure was detectable.
Beta-cell dysfunction and glucose intolerance: results from the San Antonio metabolism (SAM) study. Gastaldelli A; Ferrannini E; Miyazaki Y; Matsuda M; De Fronzo RA;Diabetologia 2004 Jan;47(1):31-9
Beta Cells Die Off in People Whose Fasting Blood Sugar is Over 110 mg/dl (6.1 mmol/L)An intriguing study shows the severe organ damage experienced by people whose blood sugar falls into a range most doctors consider to be near-normal. A team of researchers autopsied the pancreases of deceased patients who were known to have had fasting blood sugars that tested between 110 mg/dl and 125 mg/dl within two years of their deaths. The researchers found that these patients, whose blood sugar was not high enough for them to be diagnosed as diabetic, had already lost, on average, 40% of their insulin-producing beta cells.
Since the American Diabetes Association believes that a fasting blood sugar level of 100 mg/dl to 125 mg/dl corresponds to a 2-hour glucose tolerance levels of 140 mg/dl to 199 mg/dl, this suggests that patients whose post-meal blood sugars rise only to the prediabetic level may be well on the way to losing as much as 40% of their beta cell mass.
However, it is important to remember that this study only measured fasting blood sugar. But when we look at the relationship of blood sugar and organ damage what we usually find is that it isn't the fasting blood sugars that are doing the damage when they are under that important 140 mg/dl level.
The reason slightly elevated fasting blood sugars correlate with beta cell dysfunction and/or destruction is almost certainly that people with slightly elevated fasting blood sugars who eat high carbohydrate meals are experiencing high, and often long lasting, blood sugar spikes after each meal they eat. A 210 lb person whose fasting blood sugar is 110 needs to eat only 12 grams of carbohydrate to raise their blood sugar to 150, and most of them are likely to be eating 50 to 60 grams of carbohydrate per meal, ensuring that their blood sugars are well over 150 for several hours after each meal.
It is almost certainly those high post meal readings that go along with elevated fasting levels that cause the glucose toxicity that damages organs and causes complications.
This conclusion is backed up by the experience of those of us with diabetes who have kept our post-meal blood sugars under 140 mg/dl after meals for a decade or more. We generally find that our diabetes does not progress and that we do not develop the classic diabetic complications.
Beta-cell deficit and increased beta-cell apoptosis in humans with type 2 diabetes. Butler AE, Janson J, Bonner-Weir S, Ritzel R, Rizza RA, Butler PC.Diabetes. 2003;52:102-110.
Blood Sugars Over 150 mg/dl (8.3 mmol/L) Kill Transplanted Beta Cells in MiceResearchers working with mice receiving beta cell transplants showed that beta cell death was much lower in groups of transplanted mice whose blood sugar was kept under 150 mg/dl than it was in those who were allowed prolonged exposure to blood sugars higher than 150 mg/dl. (Normal and diabetic blood sugar levels in rodents are the same as those in people.)
ß-Cell Death and Mass in Syngeneically Transplanted Islets Exposed to Short- and Long-Term Hyperglycemia. Montserrat Biarnés, Marta Montolio, Victor Nacher, Mercè Raurell, Joan Soler, and Eduard Montanya. Diabetes 51:66-72, 2002
Prolonged Exposure to Blood Sugars Over 140 mg/dl (7.8 mmol/L) Kills Human Beta CellsAnother series of experiments on beta cells grown in culture showed that there is a threshold over which the damage to beta cells caused by exposure to elevated blood sugars becomes irreversible. It found that amount of damage to genes in cells involved with producing insulin depended on the concentration of glucose the cells were exposed to. The effect was continuous, not a threshold effect--meaning that the more glucose the cell was grown in, the more function it lost.
In a second experiment, the same researchers took cells damaged by exposure to high blood sugars and moved them to media that had a lower concentration of blood sugar. They found the cells could survive and recover after being moved to a growth medium containing a much lower concentration of glucose, but only if the switch was made before a certain amount of time had passed (unfortunately, the time was not specified in hours or days.) Once the cells had been exposed to glucose for that fatal time period, they could no longer be revived.
In an email to me, R. Paul Robertson, one of authors of this study wrote, "I think the glucose toxic effects begin when blood glucose gets above 140 and probably earlier." However, he also explained that while studies with diabetic rats could better quantify the blood sugar levels at which this irreversible damage occurs, these rats cost $200 apiece and a lot of rats would be required. So such a project is not planned for any time soon."
Determinants of glucose toxicity and its reversibility in pancreatic islet Beta-cell line, HIT-T15.Catherine E. Gleason, Michael Gonzalez, Jamie S. Harmon, and R. Paul Robertson.Am J Physiol Endocrinol Metab 279: E997-E1002, 2000.
One Hour OGTT Result over 155 mg/dl(8.6 mmol/L) Correlates with Markers for Cardiovascular DiseaseA study published in November of 2009 linked blood sugar readings one hour after ingesting glucose with high fibrinogen and leukocytes count (WBC), which point to subclinical inflammation, and with abnormal lipid ratios, and insulin sensitivity in a population of 1062 participants with normal glucose tolerance or prediabetes. "Normal" glucose tolerance as defined by doctors and researchers means a 2 hour blood sugar reading of under 140 mg/dl on an oral glucose tolerance test.
This study found
Elevated 1hPG [one hour plasma glucose] in NGT [people with normal glucose tolerance] and pre-DM [prediabetic] subjects is associated to subclinical inflammation, high lipid ratios and insulin resistance. Therefore, 1hPG >155 mg/dl could be considered a new 'marker' for cardiovascular risk.This strongly supports the message I have been emphasizing on this web site since 2004 that it is essential to keep one hour blood sugar reading after meals under 140 mg/dl.
The reliance of doctors on two hour glucose tolerance test results allows people to live for years with blood sugars high enough to promote complications long before they are diagnosed even pre-diabetes.
Inflammation markers and metabolic characteristics of subjects with one-hour plasma glucose levels. Gianluca Bardini et al. Diabetes Care Published online before print November 16, 2009, doi: 10.2337/dc09-134
Heart Disease, A1c, and Post-Meal Blood Sugars Correlate TightlyLearn how tightly heart disease and blood sugar levels are connected, even at ranges below those used to diagnose diabetes on this page:
A1c and Post-Meal Blood Sugars Predict Heart Attack
Diabetic Retinopathy Develops at "Prediabetic" Blood Sugar LevelsThe first conclusive evidence for this finding was reported based on Diabetes Prevention Project data presented at the ADA 65th Annual Scientific Sessions. It found that:
Over three years of subsequent follow-up of 302 DPP participants, 13% of prediabetics who developed type 2 disease were found to have diabetic retinopathy, and nearly 8% of "prediabetics" also were found to have diabetic retinopathy, though they never developed diabetes per diagnostic criteria.The diagnostic criteria used by the DPP to define "prediabetic" was a fasting plasma glucose test result that ranged between 100 and 125 mg/dl (5.5-6.9 mmol/l and/or a glucose tolerance test result showing impaired glucose tolerance: glucose tolerance test values at two hours that fell between 150 and 199 mg/dl (8.3-11 mmol/l).
This finding would suggest that prolonged exposure to post-challenge blood sugars over 150 mg/dl are highly dangerous to your retina, with or without a diabetes diagnosis, but even more so with a type 2 diabetes diagnosis.
ADA Press Release: Diabetic Retinopathy Occurs in Pre-Diabetes.
Even more worrisome is the meta analysis published by Tien Y. Wong, that found that "More than 60% of retinopathy cases were among patients with fasting plasma glucose levels below 7.0 mmol/L (126 mg/dL). This review of many published studies found that the three studies [with over 11,000 participants] together, 7.4% to 13.4% of participants had retinopathy at fasting glucose levels below 5.6 mmol/L (100 mg/dL). These were undoubtedly those who like many people with prediabetes only experience prediabetic blood sugars after eating.
Relation between fasting glucose and retinopathy for diagnosis of diabetes: three population-based cross-sectional studies Wong TY, et al Lancet 2008; 371: 736-743.
That means something like one in ten people who were most likely to have only abnormalities of post-meal blood sugars, since the most common pattern of developing diabetes is for fasting blood sugars to remain near normal for years while post-meal blood sugars rise into the prediabetic and even diabetic range. You can read more about patterns in which diabetes develops on this page on this web site: The Patterns in which Diabetes Develops
Further confirmation that diabetic retinopathy occurs at blood sugar levels well below those defined as "diabetic" by the Amerian Diabetes Association was provided by an analysis of 2005-2006 NHANES data which was published in October 2009.
This study looked at 1,066 individuals with or without diabetes aged 40 years or more. "A1C, FPG, and 45° color digital retinal images were assessed."
This study found a prevalence of retinopathy of 11% in the whole group and 36% in those with diabetes diagnoses.
Most significantly, this study found that:
The steepest increase in retinopathy prevalence occurs among individuals with A1C equal to or greater than 5.5% and FPG equal to or greater than 5.8 mmol/l [104 mg/dl]. A1C discriminates prevalence of retinopathy better than FPG.Association of A1C and Fasting Plasma Glucose Levels With Diabetic Retinopathy Prevalence in the U.S. Population: Implications for diabetes diagnostic thresholds Yiling J. Cheng et al. Diabetes Care November 2009 vol. 32 no. 11 2027-2032. doi: 10.2337/dc09-0440
An even larger metastudy, published in late 2010 based on he records of "44,623 participants aged 20 to 79 years with gradable retinal photographs" examined the correlations between signs of retinopathy and the subjects' fasting, 2 hour glucose tolerance test, and A1c results.
Glycemic Thresholds for Diabetes-Specific Retinopathy: Implications for Diagnostic Criteria for Diabetes:The DETECT-2 Collaboration Writing group. Stephen Colagiuri et al. Diabetes Care Published online before print October 26, 2010, doi: 10.2337/dc10-1206
The conclusion of the 2010 study was this:
A narrow threshold range for diabetes-specific retinopathy was identified for FPG and HbA1c but not for 2-h PG. The combined analyses suggest that the current diabetes diagnostic level for FPG could be lowered to 6.5 mmol/L [117 mg/dl] and that an HbA1c of 6.5% is a suitable alternative diagnostic criterion. This was because
glycemic thresholds for diabetes-specific retinopathy were observed over the range 6.4-6.8 mmol/L [115 - 122 mg/dl] for F[asting]P[lasma]G[lucose] 9.8-10.6 mmol/L [176.4 - 191 mg/dl] for 2-h PG, and 6.3-6.7% for HbA1c.It is likely that the 2-hour result in the oral glucose tolerance test was not predictive because it misses so many people who experience prediabetic post-meal numbers after slow digesting meals but not after consuming a huge glass of glucose which often provokes reactive hypoglycemia in such people.
The French DESIR study of people diagnosed with pre-diabetes confirmed what the studies cited above have found. Subjects who developed retinopathy over a 9 year period had an average fasting blood sugar of 130 mg/dl and an average A1c of 6.4%. Those who did not develop retinopathy over this period had an average fasting blood sugar of 108 mg/dl and an average A1c of 5.7%.
Hemoglobin A1c and Fasting Plasma Glucose Levels as Predictors of Retinopathy at 10 Years: The French DESIR Study. Massin P. et al. Arch Ophthalmol.2011 Feb;129(2):188-195.
Multiple Studies Link Heart Failure to Blood Sugars in the so-called "Prediabetic" RangeAs reported in the Diabetes In Control Newsletter in November 2006, "Two studies presented at ...[the] American Heart Association 2006 Scientific Sessions suggested that, two common measures of insulin resistance are associated with poor outcomes in a "dose-response" fashion in nondiabetic patients with heart failure."
These studies of populations consisting of both diabetic and "non-diabetic" patients, i.e. those who did not meet the ADA's diagnostic criterion of Fasting Blood Glucose of 125 mg/dl (7 mmol/L or greater) and A1c greater than 7%, found that poor outcomes (i.e. death) from heart failure correlated in a straight line fashion with both fasting plasma glucose in the range below the ADA-defined diabetic range and increasing A1c.
The conclusion of one of the researchers is that "Definitions of diabetes are based on degree of glycemia, but patients with glycemic indices below the arbitrary diagnostic thresholds can still be at increased risk for worsening heart failure, according to Held. 'I look at glucose levels as I do cholesterol or blood pressure. It seems to be a continuous risk variable." "The lower the better, to a certain limit, of course.'"
Unfortunately, the Diabetes in Control article does not give the actual numbers, but the risk was described as rising in a "dose response" manner,
Fasting Glucose Predicts Heart Failure, Death Rate, MI and Stroke, Independent of Diabetes
Rates of Cancer Rise Significantly with "Mildly" impaired Blood SugarsA Swedish study that followed 64,597 people for 10 years discovered that there was a very strong increase in the risk of cancer for those participants, no matter what their weight who had fasting blood sugars over 110 mg/dl (6.1 mmol/L) or who scored over 160 mg/dl (8.9 mmol/L) two hours after a glucose tolerance test.
The risk continued to grow as participants moved into the diabetic category, but not by the same increment as it did when they moved from normal to what most doctors consider "mildly" impaired.
The cancers most responsive to the higher blood sugars appear to those of the pancreas, endometrium, urinary tract, and malignant melanoma.
Prospective Study of Hyperglycemia and Cancer Risk.Pär Stattin, MD, PHD, Ove Björ, BSC, Pietro Ferrari, BSC, Annekatrin Lukanova, MD, PHD, Per Lenner, MD, PHD, Bernt Lindahl, MD, PHD, Göran Hallmans, MD, PHD and Rudolf Kaaks, PHD. Diabetes Care 30:561-567, 2007
Blood Sugar Fluctuations Cause More Damage to Kidney Cells than Steady State High Blood SugarAn in vitro study (i.e. a study done on tissue, not live people) found that exposing kidney cells to blood sugar levels that fluctuated between 135 mg/dl and 270 mg/dl did more damage to those cell--in terms of causing growth of fibrous tissue--than constant exposure to high blood sugars. The explanation given involved osmotic changes caused by the fluctuations as well as the suppression of some corrective genetic changes that occurred with the steady state.
Of note in this study, too, and of importance to the topic of this page, are the many graphs showing that a steady exposure to blood sugars of only 135 mg/dl was almost the same as exposure to a steady normal blood sugar level of 90 mg/dl in terms of spurring the growth of fibrous tissue. It is fibrous tissue that destroys kidney function.
The authors conclude,
These results support the recommendation for postprandial monitoring of glucose levels in patients with diabetes mellitus and imply that important differences in end organ damage could occur in individuals with similar HbA1c but different postprandial glucose levels. They go on to explain that their findings prove that it is not glycation (the attachment of glucose molecules to proteins) that destroys the kidney tissue so much as the effect of the fluctuations of blood sugar on gene expression.
They urge more attention to eliminating spikes.
Short-term peaks in glucose promote renal fibrogenesis independently of total glucose exposure.T. S. Polhill, S. Saad P. Poronnik,G. R. Fulcher, and C. A. Pollock. Am J Physiol Renal Physiol 287: F268-F273, 2004.
Risk of Developing Chronic Kidney Disease Rises Significantly in a Straight-Line Fashion as A1c rises above 6.0%A study that followed 1871 adults with diabetes for 11 years as part of the Atherosclerosis Risk in Communities (ARIC) Study, found that risk of serious kidney disease rose significantly as soon as A1c crossed over the 6.0% A1c level and increased in a straight line manner as A1c climbed higher.
The study found,
Risk of CKD was higher in individuals with albuminuria [protein in urine at start of study] and retinopathy, and the association between HbA1c concentration and incident CKD was observed even in participants without either abnormality.The 6.0% A1c supposedly corresponds to an average blood sugar of 126 mg/dl (7 mmol/L). You can often lower your blood sugar to a level that corresponds to an A1c below 6.0% by keeping your post-meal blood sugars under 140 mg/dl (7.7 mmol/L) after meals.
Poor Glycemic Control in Diabetes and the Risk of Incident Chronic Kidney Disease Even in the Absence of Albuminuria and Retinopathy: Atherosclerosis Risk in Communities (ARIC) Study. Lori D. Bash et al. Arch Intern Med. Vol. 168 No. 22, Dec 8/22, 2008
Risk of Death from Stroke Increases 27% with 18 mg/dl Rise in 2-hr GTT ResultA study that tracked 19,019 men for 38 years took capillary blood samples 2 h after consumption of a glucose preparation equivalent to 50 g of anhydrous dextrose. Study participants were then followed for mortality for a maximum of 38 years.
The study found that as subjects' 2 hour glucose tolerance test (GTT) result rose above a reading of 83 mg/dl at 2 hours, their risk of stroke rose in a linear fashion (i.e. the higher the blood sugar at 2 hours, the more risk of stroke.)
A 1 mmol/l [18 mg/dl] increase in blood glucose after this point was associated with a 27% increase in risk of death from stroke. .. This increase in risk was partially attenuated by adjustment for covariates but remained statistically significant at conventional levels. This last sentence means that adjusting for other health issues lowered the correlation, but it was still strong,though it’s important to remember that the calculations used to estimate “risk” inflates the resulting percentage. The actual prevalence of fatal stroke in this total population over the 38 years was only 6.5%.
However, since those "other" health issues may be heart disease and high blood pressure which may be caused by high blood sugars, this is still a very important finding.
Note also that the GTT in this study used only 50 grams, not the 75 g that has become standard over the last several decades. So subjects with high readings in this study would have probably seen even higher readings on a standard glucose tolerance test. However, it is possible that their results were more similar to what would be seen in the more accurate meal test.
Post-challenge blood glucose concentration and stroke mortality rates in non-diabetic men in London: 38-year follow-up of the original Whitehall prospective cohort study G. D. Batty et al.Diabetologia.Vol 51, Num, July, 2008.