June 2011: Volume 37, Number 6
NACB Guideline Update Captures Latest Evidence, Thought
By Genna Rollins
The field of diabetology has been one of the most fertile in all of medicine during the past decade, ripe with diagnostic, treatment, and analytical advances. Along the way, major milestones have been set. For the first time, professional organizations recommended use of HbA1c levels as a means of diagnosing the disease, and a laboratory working group developed a master equation to harmonize HbA1c results reporting. Underlying the decade’s progress has been lab’s essential role in improving diabetes-related analytics and educating clinicians—and even the public—about testing issues, themes emphasized in the National Academy of Clinical Biochemistry’s (NACB) 2011 update of Guidelines and Recommendations for Laboratory Analysis in the Diagnosis and Management of Diabetes Mellitus (LMPG).
“Coordination within the medical team is essential to improve the management of patients with diabetes, and since the lab plays such an important role, we hope the guidelines will be used not only by those performing lab testing, but also by those who will be using test results,” explained guidelines committee chair David Sacks, MD, senior investigator and chief of the clinical chemistry department at the National Institutes of Health in Bethesda, Md. “Our goal is to have labs and clinicians use the guidelines to provide clinicians with results that are as accurate as possible, thereby enabling clinicians to manage patients better. The broader goal is to reduce the terrible complications of diabetes and improve patient care.”
The Development Process
Sacks, who also led NACB’s 2002 LMPG on diabetes, intentionally held off on updating the new document until evidence grew around several key issues. “I made the conscious decision, because there were very substantive changes going on, that it didn’t make sense to put out a document that would be obsolete the day it came out,” he explained. A draft of the guidelines was available online during a comment period between November 3 and December 3, 2010, and the final document will be published in June, 2011 in both Clinical Chemistry and Diabetes Care.
Reflecting the multidisciplinary nature of diabetes, Sacks asked leading medical experts to serve on the 2011 NACB LMPG committee and address specific aspects of the disease that impact lab testing. “Most of the people on the committee are not clinical chemists. All the authors are well-recognized experts in their fields,” he said.
Under the stewardship of committee member A. Rita Horvath, MD, PhD, an expert in evidence-based medicine, the guidelines also feature a novel evidence grading system for diagnostic testing, a paradigm that may serve as a template for future lab medicine guidelines. “We adopted other systems in a pragmatic manner to create a grading system that is more suited to addressing lab-related rather than therapeutic questions,” she indicated. “The strengths of our system are its explicitness, transparency, and the reproducibility of the process, which probably will be more important when the next update of this guideline occurs.” Horvath is clinical director of the SEALS department of clinical chemistry at Prince of Wales Hospital in Sydney, Australia. For more information about the guidelines development process see the Online Extra box below.
A New Paradigm for Guideline Development
A self-described perfectionist, A. Rita Horvath, MD, PhD, is quick to point out what she sees as shortcomings in the process used to develop the National Academy of Clinical Biochemistry’s (NACB) 2011 update of Guidelines and Recommendations for Laboratory Analysis in the Diagnosis and Management of Diabetes Mellitus (LMPG). But in point of fact, Horvath and her eight colleagues who served on the guideline committee essentially created a new approach to identifying key areas of change in diabetology and assessing the quality of evidence for guidelines that have to do with laboratory medicine.
“It’s not perfect, but it’s a step forward and I’m hoping it will contribute to guideline development in the future,” said Horvath, who is clinical director of the SEALS department of clinical chemistry at Prince of Wales Hospital in Sydney, Australia.
The challenge was to assess evidence that relates specifically to lab testing for screening, diagnosing, and monitoring diabetes, rather than clinical management of the disease, the subject of most existing guidelines. For example, in the 2002 NACB LMPG on Diabetes, “we just accepted the American Diabetes Association grading criteria, which are designed to evaluate clinical information,” recalled guidelines committee chair David Sacks, MD, senior investigator and chief of the clinical chemistry department at the National Institutes of Health in Bethesda, Md. “This time we decided to develop a new grading system because what we had wasn’t a suitable grading system to evaluate lab tests.”
Horvath, an expert on evidence-based medicine, used the international Grades of Recommendation Assessment, Development and Evaluation (GRADE) system for rating guidelines as a resource, but modified this approach to deal with particular issues related to grading diagnostic testing. “We were a little stuck in how to overcome the problem that we didn’t have a perfect system available to use. So I had to develop a kind of hybrid system using internationally accepted grading rules,” she explained. “What we have is a system adapted from existing systems in a pragmatic manner, which was more suited for this groups’ expertise.”
In the end, Horvath paced her colleagues through a course of analyzing evidence for each of seven chapters which mirrored their professional expertise. For each guideline recommendation Horvath asked them to prioritize key questions that were most relevant in terms of outcomes. The committee member then explained why a modification to an existing NACB LMPG recommendation was needed, the key references supporting the new recommendation, and the types of study designs used in those references. Committee members also were asked to rank both the level and quality of evidence for each recommendation, and to provide any explanatory comments.
Though Horvath would have liked an even more thorough process –for instance, reviewing every single study that supported each recommendation—she still is satisfied with the result. “The strength of our process is its explicitness, transparency, and reproducibility. This will be important whenever the next update of this guideline occurs. Readers can look back to our evidence tables and see what was behind each recommendation.”
A full description of the methodology, prioritization criteria, and evidence tables is available online.
A New Diagnostic Standard
One of the most significant changes in the diabetes field reflected in the 2011 NACB LMPG is the recommendation to use HbA1c as a means of diagnosing diabetes in addition to glucose-based methods, such as fasting plasma glucose (FPG) and oral glucose tolerance testing (OGTT) (See Table, below). “The decision to use HbA1c levels for diagnosis probably represents a signal point,” said LMPG committee member David Nathan, MD. “It’s been discussed for decades, but we concluded there was enough data to use this assay, which has evolved over time and is now based on a chemical standard.”
Making the Diabetes Diagnosis
For the first time, authors of the 2011 NACB LMPG on Diabetes recommend using HbAlc as one of four methods of diagnosing diabetes. Individuals who meet any one of these criteria would be considered a diabetic.
- HbA1c ≥6.5% (48 mmol/mol)
- FPG* ≥7.0 mmol/L (126 mg/dL)
- 2-h plasma glucose ≥11.1 mmol/L (200 mg/dL) during an OGTT**
- Symptoms of hyperglycemia and casual plasma glucose ≥11.1 mmol/L (200 mg/dL)
In the absence of unequivocal hyperglycemia, these criteria should be confirmed by repeat testing.
*Fasting plasma glucose
**Oral glucose tolerance test
Source: Clin Chem 2011;57:e1-47.
Nathan, who is director of the Diabetes Center at Massachusetts General Hospital and professor of medicine at Harvard Medical School in Boston, also chaired an international expert committee convened by the ADA, International Diabetes Federation, and European Association for the Study of Diabetes that examined the pros and cons of using HbA1c to diagnose diabetes. Both that committee and the NACB LMPG committee concluded that there now is sufficient evidence about not only the diagnostic significance of HbA1c but also the test’s analytical characteristics.
Clinicians have used HbA1c levels extensively since the early 1990s as a barometer of patients’ diabetes control, but the assays had not been considered robust enough to use for diagnosis. Starting in the mid-1990s the National Glycohemoglobin Standardization Program (NGSP) led efforts to standardize HbA1c analysis, but did not use pure HbA1c for calibration—now defined as glycation of the N-terminal valine of the hemoglobin A β-chain—and therefore was not a true reference method. To address that concern, the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) developed a reference method using mass spectrometry. Now, virtually all labs that measure HbA1c in the U.S. use NGSP-certified methods.
Another boost to improved HbA1c analytics came with tightened proficiency testing standards. For example, the College of American Pathologists’ (CAP) HbA1c proficiency testing program began using accuracy grading in 2007. CAP has lowered the acceptable variance over time, from ±15% of the target value in 2007 to ±7% in 2011. The 2002 NACB LMPG already recommended that labs use NGSP-certified assay methods and that they participate in proficiency testing programs like CAP’s. The 2011 update also calls for HbA1c assay manufacturers to show traceability to the IFCC reference method.
Committee member David Bruns, MD emphasized that the new diagnostic standard and improved HbA1c analytics makes it incumbent on labs to maintain the highest accuracy possible with HbA1c testing. “If your method suddenly shifts so that your results go up by two-tenths of one percent, then anyone whose hemoglobin A1c is 6.3 percent would have a reported result of 6.5 percent, and he or she would be diagnosed as a diabetic,” he observed. “Applying that kind of error to the general population, there’d be millions of people incorrectly diagnosed if our assays went off by that amount. But people are not in the mode of getting worked up because they see their hemoglobin A1c quality control results go up by two-tenths of one percent. We need to change that.” Bruns is professor of pathology, director of clinical chemistry, and associate director of the molecular diagnostics laboratory at the University of Virginia School of Medicine in Charlottesville.
In addition to these analytical advances, Nathan explained that evidence has mounted such that the committee was able to recommend a diagnostic cutpoint, with HbA1c levels ≥6.5 % indicating diabetes. Although published research presents a complicated picture, the committee felt that within the past few years in particular, new studies linking HbA1c levels with microvascular complications of diabetes made it possible to go with that cutpoint. “Over the past decade there’s been a reassessment and clarification of metabolic goals we strive for in diabetes care, and greater justification for those goals. Although selecting greater than or equal to 6.5 percent as a diagnostic cutpoint for diabetes, based on its association with microvascular complications was relatively easy, glucose impairment runs on a continuum, making selection of a specific value to define persons at high risk for diabetes somewhat arbitrary. However, those persons with hemoglobin A1c levels close to 6.5 percent are clearly at higher risk,” he said.
On a related note, the guidelines discuss recommended changes in HbA1c reporting made in the past decade. When IFCC developed its reference method, it suggested that labs report HbA1c in units of mmol/mol rather than as a percent, the format recommended by NGSP. However, instruments calibrated to the IFCC reference standard produce results 1.5–2% lower than those calibrated to the NGSP standard. A study published in 2008 documented a linear relationship between the two methods—the master equation—and in 2010 IFCC and the major diabetes organizations recommended that labs report both NGSP and IFCC units, but left to the discretion of individual countries and labs the decision to report a calculated estimate of average glucose (eAG). The NACB LMPG does not take a stand about reporting eAG—which remains controversial—but notes that “implementation of the reporting recommendations needs to be carried out with education of healthcare providers and patients.”
“Personally, I don’t think it’s the role of clinical laboratorians to tell clinicians how to communicate with their patients,” said Sacks about the eAG reporting controversy. “Our role is to give them information and help them interpret as much information as they need to help them manage their patients. It comes down to the way individuals communicate with each patient. For some providers and some patients, it will be useful, and for others it won’t.”
Redefining Gestational Diabetes
The guidelines also spell-out several major changes involving gestational diabetes. Solid evidence about the condition and its complications emerged within the past 5 years, enabling the guidelines committee to make new recommendations. For the first time, the 2011 NACB LMPG advises that all pregnant women not previously known to have diabetes undergo testing for gestational diabetes when they’re between 24–28 weeks’ gestation. The NACB LMPG also provides new gestational diabetes diagnostic cutpoints for FPG, and 1- and 2-hour plasma glucose levels based on a 75 g OGTT (See Table, below).
New Diagnostic Criteria for Gestational Diabetes
Based on evidence from the landmark Hyperglycemia and Adverse Pregnancy Outcomes study, authors of the 2011 NACB LMPG on Diabetes recommend new thresholds for diagnosing gestational diabetes. One or more of these values from a 75-g oral glucose tolerance test must be equaled or exceeded for the diagnosis of gestational diabetes.
Glucose Test Type
|Fasting Plasma Glucose
|1-hr Plasma Glucose
|2-hr Plasma Glucose
|Source: Clin Chem 2011;57:e1-47|
These recommendations emanate from the landmark Hyperglycemia and Adverse Pregnancy Outcome (HAPO) study of approximately 25,000 pregnant women that demonstrated a strong, graded relationship between maternal glycemia and adverse pregnancy outcomes (CLN, May 2010). The International Association of Diabetes and Pregnancy Study Groups (IADPSG) issued recommendations based on the HAPO results that either have already been endorsed or are under review by various professional and public health organizations, according to IADPSG chair, Boyd Metzger, MD, who also served on the NACB LMPG committee. “In the past, the various professional and health organizations had cited the lack of evidence regarding the relationship between hyperglycemia to outcome in pregnancy. But with the HAPO data showing a strong, graded, predominantly linear, and continuous association between maternal glycemia and primary study outcomes, they’ll definitely move away from that,” he said. Metzger is Tom D. Spies professor of metabolism and nutrition at the Northwestern University Feinberg School of Medicine in Chicago.
As word spreads about the new guidelines, labs will be on the spot to maintain robust glucose testing accuracy and communicate closely with clinicians, Metzger predicted. “The evidence indicates that the continuation of risk across the whole range of glucose levels is striking, especially for fasting glucose results. Therefore, high accuracy in measurement is important,” he said. “That’s something labs recognize for any test they perform, but this is a situation where it’s particularly important.”
The communication angle comes in overlaying different obstetrical practice patterns with three means of diagnosing gestational diabetes. “This is a good example of how people can work together. There’s not one solution that fits every organization,” Metzger explained. “Here at Northwestern, we have about 13,000 deliveries each year coming from a highly diverse source of practitioners, so it’s difficult to come up with one strategy that fits them all. However, in another place where most of the testing is done in one central location, there’d be a different approach.”
The guidelines also clarify some pre-analytical issues that labs need to attend to, particularly how to deal with glycolysis. The 2002 NACB LMPG indicated that labs should separate plasma for glucose testing within 60 minutes, and when this is not possible, to use tubes with glycolytic inhibitors. The 2011 update emphasizes that samples should be placed in an ice slurry immediately, and if plasma can’t be separated within 30 minutes, labs should use tubes with a rapidly effective glycolytic inhibitor such as citrate buffer.
“This is a big challenge for clinical labs. If you leave the sample sitting around, you’re going to miss some people who would have been diagnosed because the glycolysis continues after the blood is in the tube. Even though fluoride in the tube stabilizes glucose in the long-term, it has little-to-no effect in the first several hours,” Sacks explained. “We recommended the optimum way of handling this, but there’s a realization that it’s very difficult to accomplish this in practice. Each hospital and clinical site will have to work with their clinical labs to figure out how to optimize this in their environment. It’s going to be different in each place.”
The Latest on Albuminuria
The NACB LMPG also stakes out new territory based on emerging evidence vis-à-vis diabetes and kidney function. One change that labs will need to educate clinicians about involves retirement of the term microalbuminuria, based on new nomenclature from Kidney Disease: Improving Global Outcomes, an international foundation dedicated to improving patient care and outcomes in kidney disease (See Table, below). “The term microalbuminuria is confusing. Many people think microalbumin is a unique molecule that has a special assay and is somehow divorced from the other albumins,” explained LMPG committee member George Bakris, MD. “From now on, albuminuria will be looked at as optimal, high, and very high, all on a continuum.” Bakris is professor of medicine and director of the hypertensive diseases unit at the University of Chicago Pritzker School of Medicine.
The 2011 NACB LMPG on Diabetes reflects new nomenclature and staging for albuminuria proposed by Kidney Disease: Improving Global Outcomes, an international foundation dedicated to improving patient care and outcomes in kidney disease.
||Unit of Measure|
|Very High and Nephrotic
|Source: Clin Chem 2011;56:e1-47|
As was the case with the 2002 NACB LMPG, the 2011 update recommends annual testing for low levels of albuminuria in type 1 diabetics starting 5 years after diagnosis, and at the time type 2 diabetics are diagnosed. However, the 2011 update also notes that while urine albumin to creatinine ratio (ACR) <30 mg/g is considered normal, patients with these levels should be reassessed annually. When patients’ ACR values are ≥30 mg/g, they should be rechecked after 6–12 months if the individual is on antihypertensive therapy, and if not, annually.
Based on recent research, the 2011 NACB LMPG update also suggests for the first time that currently available urine dipstick tests are not sensitive enough to detect small increases in urinary albumin excretion. In addition, the guidelines make a new recommendation that spot urine collection ideally should occur in the early morning and with the patient having fasted for at least 2 hours. To minimize variability, all collections should take place at the same time of day.
Extensive research over the past decade led to a better understanding of what urine albumin means, namely that it has more to do with endothelial dysfunction and vasculature than kidney function per se. The 2011 NACB LMPG reflects this evolution in thought, in that the guidelines for the first time note that albuminuria is a continuous risk marker for cardiovascular events, starting at 20 µg/min creatinine excretion.
“We know good glycemic and blood pressure control will improve albuminuria, and in some cases, normalize it. Some would question why we measure it, if it’s just a marker that’s not going to change anything,” Bakris explained. “The reason to measure it is because it is a marker. You may think you’re doing everything right, but if a patient’s albuminuria is increasing in the absence of blood pressure, lipid, or glucose abnormalities, then there’s some underlying pathology that you’re missing.”
The Meter Controversy
For some time, the lab community has been aware of accuracy issues with blood glucose meters (BGM), but the expanded use of these devices in point-of-care settings and other uses for which they were not originally intended has galvanized attention on this issue. In March 2010, the Food and Drug Administration (FDA) held a conference regarding the accuracy of and regulations for BGMs, but has yet to publish guidance on the topic (CLN, May 2010). The 2011 NACB LMPG update recommends overall accuracy ≤15% at glucose concentrations ≥100 mg/dL (5.6 mmol/L), which is tighter than the current ±20% standard codified in the ISO standard used by FDA to clear BGMs for the market. The guidelines also stress that lower total error would be desirable and may prove necessary in tight glucose control protocols.
Bruns, who has been a leading advocate for more accurate BGMs, urged labs to stay on top of this issue. “It’s a real problem, and not one clinicians had been aware of until recently. Now they’re more aware of it, but we need to continue to point out that there are real implications for people with diabetes that flow from errors in glucose meters,” he said. “The FDA is acutely aware that meters have not been cleared for the kind of uses they’re being put to.”
A Nod to the Future
In concert with increased use of continuous glucose monitoring during the past decade, additional evidence about this technology has come to the fore. This enabled the NACB LMPG committee to revise its 2002 decision to not recommend noninvasive glucose analysis as a replacement for either self monitoring of blood glucose or glucose measurements by an accredited lab. The 2011 update modified this statement to advise that real-time continuous glucose monitoring in conjunction with intensive insulin regimens can be a useful tool to lower HbA1c in selected adults with type 1 diabetes. However, the guidelines point out that while research in this area continues to advance, major technological hurdles remain before noninvasive sensing technology will be sufficiently reliable to replace existing portable meters or other minimally invasive technologies.
The chapters on genetic and autoimmune markers for diabetes make clear that while there’s been considerable research on both fronts, much more evidence will be needed before these methods make it into mainstream clinical use. The NACB LMPG points out that there are more than 60 distinct genetic disorders associated with glucose intolerance or frank diabetes, but that there are substantial, complex genetic factors that contribute to risk of type 2 diabetes. Still, advances in molecular diagnostics ultimately will aid in family counseling, prognostic information, and optimal treatments for diabetes, the guidelines predict. On February 24 AACC presented a webinar about novel biomarkers for the diagnosis and management of diabetes. A CD-ROM of this session is available through the AACC Store website.
Given the substantive changes in diabetes care during the past decade and the still active research environment, the 2011 NACB LMPG provides a good framework for labs and clinicians to address care processes jointly, with the goal of improving patient outcomes, Metzger contends. “What’s called for is a collaborative role, and I don’t believe that’s been the case to an optimal extent. We’ve tended to operate too much in our own worlds and not enough together,” he said. “But the working of this committee is a good example of the way we can work together. There’s not one solution that fits every situation, and by collaborating we can find the best strategy that works for our own institutions.”