College of American Pathologists
Printable Version



cap today

Clamping down: How tight should glycemic control be?

March 2002
Anne Paxton

Like the Red Queen in Lewis Carroll’s Through the Looking Glass, clinicians are running as hard as they can just to stay in place when diagnosing diabetes mellitus. Of the 16 million Americans with diabetes, more than five million are unaware they have it. Clinicians now diagnose some 800,000 new cases of Type 2 diabetes annually, but with the incidence of diabetes having tripled in the last 30 years, they still may be falling behind.

The surge in diabetes prevalence has sharpened concern that guidelines for screening and managing this deadly and costly disease are not stringent enough. "We need more aggressive, complete, and cohesive standards," contends Rhoda Cobin, MD, president of the American Association of Clinical Endocrinologists.

The causes of the escalation in diabetes cases are fairly well known. "The frequency and severity of obesity are rising, and at the same time, the population is becoming more and more sedentary," says William E. Winter, MD, professor of pathology and pediatrics at the University of Florida. "This is causing a high frequency of the population to be affected with insulin resistance, a precursor of Type 2 diabetes."

The numbers show a dangerous lag time between when the disease is contracted and when it is diagnosed. Many newly diagnosed patients have had the disease, undetected, for 10 years, and half of them are experiencing complications by the time they are diagnosed.

Can these patients be detected earlier? And should there be changes in how they are monitored? Those were pivotal questions behind a consensus conference of world experts in diabetes convened by the American College of Endocrinology. Agreeing that more aggressive screening and monitoring are needed, the conference, held last August, set forth three recommendations to change guidelines for glycemic control in people with diabetes and put them in line with guidelines already in force in Europe:

  • Start screening people at high risk for the disease at age 30, which is 15 years earlier than is currently recommended by the American Diabetes Association.
  • Lower HbA1c (termed A1c in the recommendations) target levels for diabetes control from seven percent to 6.5 percent.
  • Lower target levels of preprandial blood glucose to 110 (ADA’s range, 90-130) and establish a target level of 140 for postprandial blood glucose.

The Association of Diabetes Educators has already endorsed the new guidelines. The cornerstone group, the ADA, says it is considering them, as is the National Institute of Diabetes and Digestive and Kidney Diseases, a key federal agency involved in diabetes.

In a presentation at the 2001 ASCP/CAP meeting, Dr. Winter confirmed that the latest refinement of diabetes testing guidelines is part of a continuing debate about the value and complexities of glycemic control. For example, there has not been unanimous support for the most recent lowering of the fasting plasma glucose cutoff, from 140 mg/dL to 126 mg/dL, a change the ADA endorsed in 1997.

"There’s been controversy since that time as to whether the cut point was actually too low," Dr. Winter says. "Some people believe that it should still be 140 mg/dL because many people who don’t yet have complications would still be defined as diabetic."

At the time, the ADA offered several arguments in favor of a lower cut point. "First," says Dr. Winter, "a sizable portion of Type 2 patients already had microvascular complications at the time of diagnosis. So if one waited to 140 mg/dL or above—the pre-1997 cutoff—one could actually delay or miss the diagnosis. Second, there’s no cut point below which there’s a 100 percent chance of not having complications. It’s possible to have complications even with an average blood glucose of less than 140 mg/dL."

Third, when researchers looked at the numbers of people being diagnosed with Type 2 diabetes and compared them with the population frequency of diabetes, they found the 140 mg/dL cut point was not as sensitive as the two-hour oral glucose tolerance test cutoff of 200 mg/dL. By lowering the cut point to 126 mg/dL, they would increase the sensitivity of the test and pick up a group that would otherwise have been missed.

The hard reality of diabetes diagnosis is that even these cutoffs may be too high, says Dr. Winter. Microvascular complications like retinopathy and nephropathy, which can cause blindness or kidney failure, are one thing. But serious macrovascular complications, including coronary artery disease, cerebrovascular disease, and peripheral vascular disease, may develop prior to microvascular. Fifty percent of Type 1 diabetes patients eventually die of macrovascular disease, as do 80 percent of Type 2 patients.

This, in part, explains the greater concern over false negatives than false positives in diagnosing diabetes. "The minimal requirement to document hyperglycemia on at least two occasions should protect patients from being misdiagnosed as diabetic," Dr. Winter says.

To test for diabetes, excluding situations where ketoacidosis or nonketotic hyperglycemic coma are present, outpatients should be in their general state of health without recent hospitalizations or changes in diet or exercise. That’s because diabetes, a state of chronic hyperglycemia, must be distinguished from transient hyperglycemia. Reversible pancreatic islet ß-cell dysfunction or dysfunctional autonomic control of pancreatic islets, for example, may cause transient hyperglycemia. "Chronic pancreatitis can clearly lead to diabetes, but with acute pancreatitis, you can have a transient dysfunction. So clinicians just need to be careful," Dr. Winter says.

Other potential causes of transient hyperglycemia are severe head trauma, increases in anti-insulin "stress" hormones—which could go along with recent hospitalization for acute myocardial infarction, burns, or severe stress—or iatrogenic etiologies such as use of diabetogenic medications, or excessive rates of parenteral glucose infusion from intravenous glucose infusions or hyperalimentation, he says.

"This is not to say that physicians should ignore hyperglycemia in hospitalized patients," notes Dr. Winter. "And they may very well need to treat a non-diabetic individual with insulin if they are sufficiently hyperglycemic. But if a physician is worried about a patient who may be diabetic, it would be prudent to follow up in a number of weeks. The take-home message is, if the patient recovers from the illness, the physician may need to retest for hyperglycemia."

World consensus on which test should be used to identify diabetes is split. The ADA, for example, recommends fasting plasma glucose as a screen, while the World Health Organization endorses oral glucose tolerance testing. OGTT is more sensitive in identifying diabetes, Dr. Winter says, "but in the U.S., people have found that general population screening with OGTT isn’t practical."

The ADA’s rationale for using fasting plasma glucose is that it’s relatively inexpensive; it doesn’t require glucose administration, which in some cases can cause stomach upset; it doesn’t require a second venipuncture; and the patient can have an FPG at the same time as a routine lipid profile. Dr. Winter maintains it is more reasonable to have a test that is simple, straightforward, minimally invasive, and affordable, and could arguably be done by most patients, if not all.

A more stable measure is the hemoglobin A1c test, which measures a patient’s average blood sugar level over the past three months. It is considered the gold standard for assessing and monitoring glycemic control and predicting a diabetic patient’s future risk of complications. All diabetes trials have found that for each one percent reduction in A1c, there is a 30 to 35 percent reduction in microvascular complications.

While some diabetes experts have promoted use of the hemoglobin A1c test for screening, most are against it, Dr. Winter says. "The ADA and most practitioners do not feel A1c should be used for screening or diagnosis, and there are several reasons why. Not everybody with diabetes necessarily has an elevated A1c—in fact, in Type 1 diabetes, A1c can be normal at diagnosis—and it’s not as reproducible an assay as plasma glucose. The variation of the assay can be in the neighborhood of 10 percent, so the precision is not as good as plasma glucose measurements."

Another issue is definitional. "Diabetes is defined as hyperglycemia, and elevated A1c is a consequence of diabetes," says Dr. Winter. "Diabetes is not defined in terms of elevated A1c. If we wanted to define it that way, then we could screen that way." But this would greatly hamper world comparisons of data about diabetes. Although A1c doesn’t vary with time of day, and one could argue that is a great advantage, the relative cost of this type of testing versus glucose is much higher. "If you think about the diagnosis of diabetes from an epidemiological point of view, only industrial countries are likely to have the resources to afford or even consider a screening program with A1c."

Regarding the A1c guideline endorsed by the American College of Endocrinology, Dr. Winter says: "It makes sense that if less than seven percent is good, less than 6.5 percent is better. We know that there is no single A1c level below which no complications occur." Dr. Winter points to data from the United Kingdom Prospective Diabetes Study, which supported a lowered A1c goal of less than 6.5 percent. That study found an elevated risk for all microvascular and macrovascular complications was shown to begin at 6.5 percent A1c. But Dr. Winter says that if the general goal is lowered, "we need to remember that goals must be individualized." Running a lower A1c will increase the risk of hypoglycemia, which can lead to coma and even death.

The Diabetes Control and Complications Trial, reported in 1993, compared intensive monitoring of 1,400 adolescents and young adults with Type 1 diabetes with conventional monitoring and found an astounding 76 percent reduction in the incidence of retinopathy in the intensively treated group. The correlation had been suspected before but not demonstrated, says Craig A. Alter, MD, assistant professor of pediatrics at The Children’s Hospital of Philadelphia, University of Pennsylvania School of Medicine. "The feeling from the endocrinology community became fairly strong that aiming for tighter control would lead to a reduction in long-term complications," he says.

In advocating a preprandial glucose treatment target of less than 110 mg/dL, the ACE notes that there is an increased risk of retinopathy clearly associated with fasting plasma glucose over 110 mg/dL, and the group argues that a similar value for preprandial plasma glucose seems reasonable. Nevertheless, ACE concedes that blood glucose values obtained through self-monitoring "may or may not reflect plasma glucose," and that patients and health providers should become familiar with what their individual meters measure.

Dr. Winter believes setting postprandial glucose goals is not yet justified. "There is reason to support postprandial monitoring in gestational or pre-existent diabetes during pregnancy; however, outside of that situation, there is little data that monitoring postprandial glucose levels leads to a better outcome," he says. "It may prove to be very important to monitor postprandial glucose levels if the patient is already at target for FPG and preprandial glucose levels and has not achieved an A1c of less than seven percent."

The ACE admits there is a relatively small body of evidence from which to draw conclusions about postprandial control guidelines and notes that many of the data have been obtained from studies using postglucose challenge rather than postprandial loads. But the consensus panel nevertheless recommends a targeted two-hour post-meal glucose of less than 140 mg/dL. Dr. Winter questions the wisdom of moving to a new target when the role of self-monitoring blood glucose has not been proven conclusively: "Why make the monitoring regimen more rigorous unless we know there is going to be a better outcome?"

He agrees, however, with the ACE recommendation to expand screening by starting it at age 30 for high-risk groups. "As I read it, the ACE statement defined 30 to 44 as the ages to screen for Type 2 diabetes mellitus when a risk factor was present. Type 2 diabetes is seen before age 40 with increasing frequency, although there are no evidence-based studies to prove earlier screening improves outcome."

The increase in Type 2 diabetes among adolescents and children as young as age seven is becoming a major problem at endocrine centers throughout the United States, Dr. Winter says. As best he can recall, when he was a fellow in pediatric endocrinology more than 20 years ago, he encountered only one pediatric patient who was considered to have Type 2 diabetes, a male Native American boy. "Now," he says, "10 to 30 percent of pediatric patients are Type 2—predominantly obese adolescent minority patients."

Dr. Winter stresses, however, that the ACE did not recommend routine Type 2 diabetes screening for the general population in the 30- to 44-year-old group. Screening always has tradeoffs. "It’s going to cost more to treat somebody for diabetes than not treat them, and it costs more in general to be in tight control than not in tight control," he says. "But those patients that aren’t in good control are more likely to suffer complications, and this increases morbidity and early mortality."

The ADA, he notes, does not recommend screening for diabetes in the context of community health fairs; rather, it recommends that people be screened on an individual basis. "One reason not to screen in community settings is that people attending health fairs usually are very aware of their health and really would not benefit from being tested," says Dr. Winter. "The other issue is that random fingerstick glucose is very difficult to interpret. If it’s less than 110 mg/dL, it’s normal. But with a value of 110 mg/dL or above, you really can’t say whether the patient is diabetic or not."

As closer glycemic control becomes a higher priority in diabetes care, evidence of yet more benefits from glucose regulation is emerging. In a provocative study reported recently in the New England Journal of Medicine (2001;345:1359- 1367), adults admitted to an intensive care unit and receiving mechanical ventilation were randomly assigned to receive intensive insulin therapy (maintenance of blood glucose at a level between 80 and 110 mg/dL) or conventional treatment (insulin infusion only if blood glucose level exceeded 215 mg/dL and maintenance of glucose between 180 and 200 mg/dL).

"What they asked was, If we maintain the glucose between 80 and 110 mg/dL, versus routine care that is maintaining a glucose at 210 or 220 mg/dL, is there any change in mortality? So the patients who gave permission in the study were randomized to receive routine or intensive therapy, and these people on admission were not diabetic," Dr. Winter says.

"There was a tremendous improvement in mortality," he adds. "Intensively treated patients had a mortality of 4.6 percent versus an eight percent mortality in conventionally treated patients." Overall in-hospital mortality fell by 34 percent, bloodstream infections by 46 percent, acute renal failure requiring dialysis or hemofiltration by 41 percent, median number of red cell transfusions by 50 percent, and critical-illness polyneuropathy by 44 percent.

"The authors were very clear in the article: They stated the study pertains to their study population. They didn’t study people who presented with cardiogenic shock or people not being ventilated who were in septic shock. So how broadly this applies is unknown," Dr. Winter says. "But it’s certainly a very intriguing study." In fact, the study had been designed to include more than 2,000 patients but was stopped after 1,500 patients were enrolled because the efficacy of intensive therapy was demonstrated statistically.

It indicates, Dr. Winter says, that tighter glycemic control may have therapeutic benefits far beyond diabetes.

Anne Paxton is a writer in Seattle.