Use of the Estimated Average Glucose e AG

Use of the Estimated Average Glucose (e. AG) in Patient Care www. diabetes. org 1 -800 -DIABETES

A Typical Patient Encounter “So, Mrs. Smith, it looks like you do have diabetes. Your repeat fasting blood sugar was 178, and as you recall the first one was 187. Over 126 is diabetes. Also, your hemoglobin A 1 c was way too high at 8. 6%. Normal is less than 6%. We need to get it below 7%. ” www. diabetes. org 1 -800 -DIABETES

A Typical Patient Encounter “What’s a hemoglobin A…whatever you said? I remember my hemoglobin was low when I was pregnant. What were those other numbers? What do you mean, 7%. . . of what? ” www. diabetes. org 1 -800 -DIABETES

G G G www. diabetes. org G G GG = __% 1 -800 -DIABETES

Uh… ? ? ? www. diabetes. org 1 -800 -DIABETES

It’s Not Just Confusing for Newly Diagnosed Patients • High levels of testing of Hb. A 1 c for patients with known diabetes (> 90%). • Of patients with test in past 6 months: – 66% did not know result – 25% accurately reported within 1% range (< 7%, 7 -8%, 8 -9%) – 9% inaccurately reported within 1% range Heisler, Diabetes Care 28: 816, 2005 www. diabetes. org 1 -800 -DIABETES

The Clinical Dilemma • Hb. A 1 c: useful for research, risk prediction, target of therapy • Well standardized • HOWEVER, difficult to explain to patients • Concept of % is not intuitive • Glucose more familiar to patients from self -monitoring or from laboratory glucose results www. diabetes. org 1 -800 -DIABETES

The Concept of Average Glucose • We tell patients the Hb. A 1 c reflects their “average glucose over 2 -3 months” • But: do we know this for sure? www. diabetes. org 1 -800 -DIABETES

Year Cohort Study period (weeks) Number of glucose tests per patient per 1 -3 months Svendsen 1982 15 T 1 DM 5 200 -300 Nathan 1984 21 T 1 DM 8 200 -300 DCCT 2002 1439 T 1 DM 12 7 Hempe 2002 128 T 1 DM 4 80 Murata 2004 182 T 2 DM 8 180 Nathan 2007 22 T 1 DM 3 Normals 12 24, 000 (CGMS) Study www. diabetes. org 1 -800 -DIABETES

The A 1 C-Derived Average Glucose (ADAG) Study International study designed to: • Carefully look at relationship between Hb. A 1 c and average glucose • Determine the mathematical relationship between the two for reliable conversion • Establish that the relationship is valid across: - Diabetes types - A wide range of Hb. A 1 c levels and age - Different races/ethnicities Nathan et al, Diabetes Care 31: 1473, 2008 www. diabetes. org 1 -800 -DIABETES

ADAG Study Centers • Cameroon • Denmark • Italy • The Netherlands • United States – – Boston New York San Antonio Seattle • India (site dropped due to specimen handling issues) www. diabetes. org 1 -800 -DIABETES

Participants in ADAG • Goal was to recruit people with – Type 1 diabetes – Type 2 diabetes – No diabetes • With a range of – Ethnicity/race – Hb. A 1 c levels • Excluded those with conditions that would interfere with measurement/interpretation of Hb. A 1 c or glucose www. diabetes. org 1 -800 -DIABETES

Measures of Glycemia in ADAG Study • CGM (calibrated by 8 -point glucose profiles with Hemocue meter) for at least 48 hours at baseline and every month for 3 months • 7 -point glucose profiles for 3 days per week with One Touch Ultra meter • Hb. A 1 c at baseline and monthly X 3 months with DCCT-aligned assay in a central laboratory • Four measures of Hb. A 1 c to assure stable control, but only final value used for correlation with prior 3 months’ glucose readings www. diabetes. org 1 -800 -DIABETES

ADAG Study Flow Total Enrolled 661 Eliminated from analysis 154 (23%) - Dropped out or excluded during study 91 (14%) - Inadequate CGM 11 (2%) - Inadequate Hb. A 1 c samples 52 (8%) www. diabetes. org 1 -800 -DIABETES

Baseline Characteristics of ADAG Participants Type 1 Type 2 Non-DM Total 268 159 80 507 43 + 13 56 + 9 40 + 14 46 + 14 52% 50% 69% 54% 93% 71% 83% African/Af-Am 2% (5) 13% (21) 15% (12) 8% (38) Hispanic 6% (15) 8% (12) 15% (12) 8% (39) Number Age Gender (% F) Race/Ethnicity White Treatment Pump / ≥ 3 inject/day 47% / 53% Diet only/ 10% Oral agent only 52% Insulin only 19% Insulin & oral 19% www. diabetes. org 1 -800 -DIABETES

ADAG Study: Distribution of Baseline Hb. A 1 c Number of subjects 38% 44% Normal 18% D i a b e t i c Baseline Hb. A 1 c (%) www. diabetes. org 1 -800 -DIABETES

ADAG Study: Glucose Monitoring • CGM – mean of ~ 2, 400 measurements per participant • Life. Scan meter ~ mean of 300 measurements per participant – Mean of ~ 25 measurements per week – Goal was a minimum of 21 tests per week • Total ~ 2, 700 measurements/participant during 12 weeks www. diabetes. org 1 -800 -DIABETES

ADAG Study: Analyses • CGM results corrected upward by 5% to be consistent with BG • Each glucose measure weighted in proportion to the inverse of total number of measurements on that day (each day had equal weight) • Arithmetic mean glucose calculated for each participant • Linear regression model used to estimate relationship between average glucose and the 3 - month Hb. A 1 c www. diabetes. org 1 -800 -DIABETES

ADAG Study: Study Success Calc. AG (mmol/L) 90% of cohort values fall in this range Hb. A 1 c (%) 90% of values fell within +/- 15% www. diabetes. org 1 -800 -DIABETES

AG (mg/dl) ADAG Study: Correlation of AG With Hb. A 1 c AG (mg/dl) = 28. 7 x Hb. A 1 c – 46. 7 R 2 = 0. 84 P < 0. 0001 Hb. A 1 c (%) www. diabetes. org 1 -800 -DIABETES

ADAG Study: Correlation of AG with Hb. A 1 c: CGM data vs. Meter No difference in relationship (P=0. 18) whether Life. Scan or CGMS data used Hb. A 1 c (%) www. diabetes. org 1 -800 -DIABETES

ADAG Study: Other Factors Examined • Does the Hb. A 1 c-Average Glucose relationship differ by: - Type 1 or type 2 diabetes NO - Diabetes or no diabetes NO - Amount of glucose variability NO - Gender NO - Age NO - Ethnicity/Race NO (but trend toward higher Hb. A 1 c per AG in African and African-American participants vs. whites, P=0. 07) - Smoking NO www. diabetes. org 1 -800 -DIABETES

ADAG Study Excluded Known Sources of “Inaccuracy” of Hb. A 1 c • • • Hemoglobinopathy Anemia Pregnancy Hepatic or renal disease Etc. www. diabetes. org 1 -800 -DIABETES

ADAG Study Conclusion: Hb. A 1 c Correlates Highly With AG 450 400 AG (mg/dl) 350 AG (mg/dl) = 28. 7 x Hb. A 1 c – 46. 7 300 250 200 150 100 50 3 4 5 6 7 8 9 10 11 12 13 Measured Hb. A 1 c (%) www. diabetes. org 1 -800 -DIABETES

Implications • Tight correlation between Hb. A 1 c and AG allows us to translate Hb. A 1 c into an estimated Average Glucose (e. AG) • e. AG will apply to the majority of patients with diabetes – Barring “traditional” conditions interfering with the assay or the relationship between glycemia and Hb. A 1 c www. diabetes. org 1 -800 -DIABETES

ADAG Study: “Translation” of Hb. A 1 c into e. AG Hb. A 1 c (%) 5 6 7 8 9 10 www. diabetes. org e. AG (mg/dl) (mmol/l)_ 97 5. 4 126 7. 0 154 8. 6 183 10. 2 212 11. 8 240 13. 4 1 -800 -DIABETES

Note that the numbers are different Hb. A 1 c (%) 6 7 8 9 10 www. diabetes. org ADAG DCCT (mg/dl)______ 126 135 154 170 183 205 212 240 275 1 -800 -DIABETES

Consensus Statement FCC, EASD, IDF, ADA Sept 2007) • Hb. A 1 c assay to be standardized worldwide using the new IFCC standard and expressed as: – % as currently used (DCCT values) – IFCC units in mmol HBA 1 c/mol Hb. A – e. AG in mmol/l or mg/d. L (if ADAG study meets its data acceptability goals) • This paved the way for reporting both Hb. A 1 c and EAG on lab reports Diabetes Care and Diabetologia, 2007 www. diabetes. org 1 -800 -DIABETES

What Won’t Change…And What’s New • To a clinician, there is no change in the Hb. A 1 c assay • To clinical chemists, there is a new IFCC standard in the background • We have the potential for a valuable educational tool for patients www. diabetes. org 1 -800 -DIABETES

A Typical Patient Encounter “So, Mrs. Smith, it looks like you do have diabetes. Your average blood sugar is around 200. When people don’t have diabetes, this number is below 125. We need to work with you to try to get this number, the average glucose, down below 150 over the next few months with some weight loss, exercise, and a medication. Let’s talk some more about what you can do…” www. diabetes. org 1 -800 -DIABETES

A Typical Patient Encounter “Wow, I’m not happy to hear that…I know that diabetes can do some bad things. Tell me what I can do to get my average glucose down. ” www. diabetes. org 1 -800 -DIABETES

What is ADA Doing to Promote Use of e. AG in Patient Care? • Health care provider education – – ADA Scientific Sessions, June ’ 08 American Association of Clinical Chemists, August ‘ 08 AADE Annual Meeting, August ‘ 08 e. AG calculators (handheld and on professional. diabetes. org) • Patient education – – Website Diabetes Forecast magazine, books Pamphlets and brochures ADA will include term “average glucose” in all consumer pieces www. diabetes. org 1 -800 -DIABETES

What Can Clinicians and Educators Do? • Choose which term—A 1 C or Average Glucose— to use with each patient (some may already be used to A 1 C) • In verbal communications, no need to say “estimated” • We want to keep the A in A, B, Cs • Use updated table, calculator on www. diabetes. org, or other tools to convert A 1 C to average glucose • “Lobby” your lab to report both numbers www. diabetes. org 1 -800 -DIABETES

What Can Clinical Chemists Do? • Even with tools, most clinicians will not take the time to calculate conversions • Reporting both Hb. A 1 c (DCCT-aligned) AND e. AG on lab reports will do the most to promote wide use of the term • Professional and patient education may drive demand • Conversion is a simple regression equation www. diabetes. org 1 -800 -DIABETES

Average Glucose Blood pressure Cholesterol to help make the “A” understandable! www. diabetes. org 1 -800 -DIABETES