134 Diabetes Spectrum Volume 21, Number 2, 2008

Nutrition FYINutrition FYI

After evaluating numerous patient glucose records, clinicians have remarked that if they could simply take food out of the equation, diabetes management would be so much easier. Achieving optimal postprandial glycemic control has proven to be a major challenge for clinicians and patients.

Wide glucose excursions are potent activators of oxidative stress, one of the main contributors to vascular complications.1 Current nutrition recommendations for individuals with diabetes stress the importance of implementing mealtime interventions that reduce postprandial glucose (PPG) excur-sions.2 Yet doing so is challenging with available tools. Patients are frequently encouraged to check blood glucose levels preprandially and again 2 hours after the start of the meal to determine their glycemic response to the meal. This method often underestimates the peak glucose value and fails to identify the duration of the PPG excursion.3

Continuous glucose monitoring (CGM) has opened up a new world of opportunity to evaluate the glycemic effect of food, activity, medication, stress, and other fac-tors. With the introduction of the MiniMed CGMS in 1999, a 3-day retrospective sensor, continuous monitoring data began to validate some of the theories on the glycemic effects of various foods and pro-vided some unexpected discoveries. Today, real-time CGM systems display glucose readings every few minutes throughout the day and night, giving clinicians and patients a glimpse of information previously unavailable or difficult to ascertain

through episodic self-monitoring of blood glucose (SMBG).

CGM systems provide the opportunity for immediate feedback on the glycemic response of food choices. Glucose values, trend arrows, line graphs, and alarms viewed on the device screen provide real-time perspective. Patients and clinicians can also evaluate glycemic responses to foods and meals retrospectively by analyzing glucose trends and patterns from data management reports. Initially, clinicians were concerned that the extra data provided by CGM might be intimidating to patients, but studies have found that the use of CGM data actually makes treatment and insulin dose decisions easier for patients.4

Factors Affecting Postprandial GlycemiaEvaluating the glycemic response to a meal is difficult not only because of the composition of the meal, but also because of a number of other variables. Postprandial glycemia is based on how quickly carbohydrate is digested and absorbed and how quickly glucose is removed from the circulation.5 Carbohydrate is the major determinant of PPG levels, with the amount of carbohydrate being the primary contributor.2 The type of carbohydrate and other characteristics of the food, such as its physical form, ripeness, type of starch, style of preparation (cooking method, temperature, time, and amount of moisture used), and the degree of processing all influence the PPG response.2,5

For patients treated with insulin, the timing and dosage of prandial

insulin have major effects on PPG response. CGM allows patients to observe the pharmacodynamics of insulin and better match meal timing with insulin action. In addition, premeal glucose level, available insulin, insulin resistance, other diabetes medications, physical activ-ity, and stress must be taken into consideration in evaluating the PPG response.2

Evaluating Personal Glycemic Responses to FoodsMany carbohydrate foods have been classified based on their glycemic response; however, outside of research settings, patients have unique and varied responses to carbohydrates. To determine indi-vidual responses to commonly eaten carbohydrate foods, patients can experiment with measured amounts of a variety of carbohydrates. For example, a patient can compare the response to a breakfast of steel-cut oats to a breakfast of dry cereal eaten at the same time of day with the same amount of carbohydrate. The patient can track the real-time

The Use of Continuous Glucose Monitoring to Evaluate the Glycemic Response to FoodJanine Freeman, RD, LD, CDE, and Lynne Lyons, MPH, RD, CDE

Figure 1. Receiver display illustrat-ing real-time glucose value, trend arrow, and alarm. (Courtesy of Ab-bott Diabetes Care)

Nutrition FYI

Diabetes Spectrum Volume 21, Number 2, 2008 135

Nutrition FYI Nutrition FYI

glucose responses on the device screen by viewing the glucose value, trend arrow, or line graph (Figure 1). In addition, reports that overlay data from multiple days on one line graph (Modal Day/Sensor Daily Overlay) can be accessed from data management programs offered by the CGM manufacturers (Figure 2).

To evaluate the effect of a specific food or meal on PPG, other variables that influence PPG should be kept as constant as possible. The time of day, premeal glucose, timing and amount of bolus insulin, and physical activity can be electronically logged in some CGM devices or recorded on a paper log and considered when analyzing the data. A patient’s ability to estimate carbohydrate servings should also be frequently assessed because underestimating carbohydrate is a frequent occurrence that can significantly affect the insulin-to-carbohydrate ratio (ICR). A major benefit of CGM, however, is that when patients make an error in calculation or a poor decision, they can be alerted to change the course before a significant event occurs.

Evaluating PPG Response to Mixed MealsIn addition to the many factors that affect the glycemic impact of carbohydrate foods, the presence of other nutrients adds even more complexity to the task of predicting the glycemic effect of a mixed meal. Research has demonstrated that the presence of fat in a food or meal delays the peak glycemic response, and protein usually has little or no effect on blood glucose levels.2 CGM allows clinicians and patients to evaluate more easily the effect of various combinations of nutrients. In a pediatric clinical trial with CGM, Gandrud et al. suggested including protein and fat in the children’s typical breakfast of highly refined carbohydrate foods as one alternative to reduce the severity of glucose excursions.6

To evaluate their own glycemic responses, patients can begin by using CGM as a learning tool to discover how different meals affect their glucose levels, without altering the type or amount of a favorite meal. Which meals have the highest glucose peak and why? How long does it take for the glucose to peak? Which meals have the longest PPG

duration and why? Restaurant foods are particularly challenging because of the difficulty in assessing nutrient composition and portion sizes.

It is important for patients to understand how to interpret CGM results without drawing inaccurate conclusions. Patients need to keep in mind the following points:

Specific mixed meals must be evaluated on more than one occasion. The effects of a specific type of meal can vary at different times of the day.Portion sizes must be consistent and estimated accurately (i.e., use of household measuring utensils or use of prepackaged meals that have Nutrition Facts labels). PPG information provided by

CGM reveals the extent and dura-tion of hyperglycemia associated with the particular type of meal; this feedback helps patients plan a more effective strategy on the next occa-sion. For example, a high-fat meal that results in prolonged postpran-dial hyperglycemia might induce patients on insulin pump therapy to try a combination or dual-wave bolus or patients on multiple daily injections of insulin to split their prandial bolus the next time they choose a similar meal (Figure 3).

Prandial Insulin Dosing

Timing of meal bolus CGM results have demonstrated the difficulty in controlling PPG excursions because of the pharma-codynamics and pharmacokinetics of currently available prandial insu-lins.7 The practice of dosing rapid-acting insulin at mealtime has come into question based on experience with CGM. Many clinicians are now recommending that patients dose premeal insulin earlier to prevent large PPG excursions.8,9 Observing the downward trend arrows or change of slope in the line graph can help determine when premeal insulin begins to take effect and can be validated with SMBG. When glucose is in a rapid downward trend, the change in blood glucose will gener-ally precede the change observed by the CGM device. Coordination of

•

•

•

Figure 2. Sensor Daily Overlay by Meal report (Medtronic CareLink Per-sonal Therapy Management Software for Diabetes). This report reveals the effectiveness of the prandial insulin for various meals. This example illustrates a pattern of elevated PPG. (Provided courtesy of Jen Block, BS, RN, CDE)

136 Diabetes Spectrum Volume 21, Number 2, 2008

Nutrition FYI

bolus insulin timing with ingestion of food corrects one variable in PPG management.

Types of bolusesCGM provides information to help patients on insulin pump therapy determine the type and amount of insulin boluses that adequately cover a particular type of meal. Evaluating the glycemic response to high-fat meals can help determine the appro-priate type and duration of a bolus to minimize the glycemic excursions. A study using three different bolus regimens with a consistent pizza meal determined that a dual-wave (combination) bolus over 8 hours was most effective in controlling PPG and preventing hypoglycemia during the 12 hours.10 Because glucose information is constantly available with CGM, patients can readily determine the duration of hyperglycemia from the meal and fine-tune the bolus for that particu-lar type of meal on future occasions.

Correction bolusesA frequent concern with real-time CGM is the risk for excessive postprandial bolusing.9 It is impera-tive that patients be educated and trained in how to respond appropri-ately to postprandial hyperglycemia to avoid the risk of subsequent hypoglycemia. Some clinicians sug-gest that patients wait a designated

time period (i.e., 2 hours) after administering premeal insulin before taking a correction dose to avoid “insulin stacking,” the practice of taking correction dose insulin before a previous dose of prandial insulin has had its full effect.11 Patients must consider the insulin remaining from the premeal bolus, referred to as “insulin on board,” the direction of the trend arrow, the type of carbohydrate, the meal composition, and the level of physical activity before taking extra insulin to treat postprandial hyperglycemia.9

Algorithms for responding to trend arrows that have been used in clinical studies with real-time CGM include the following recommendations:9,11

If glucose is rising at > 2 mg/dl per minute, the bolus should be increased by 20%.If glucose is rising at 1–2 mg/dl per minute, the bolus should be increased by 10%.If glucose is decreasing by > 2 mg/dl per minute, the bolus should be decreased by 20%.If glucose is decreasing by 1–2 mg/dl per minute, the bolus should be decreased by 10%.

Determining the ICR and Insulin Sensitivity Using CGMExperience with CGM has rein-forced the fact that greater glycemic

•

•

•

•

excursions occur after breakfast than after other meals of the day, an observation that patients usually make when determining their ICRs.6,11 Several data management reports provide the opportunity to evaluate ICR over a period of time. In addition to identifying PPG patterns on line graphs, patients and clinicians can identify meals with inadequate insulin coverage from pie charts, bar graphs, and statistics reports that show the percentage of pre- and postmeal glucose readings below, within, and above target range. Once it is determined that basal insulin is appropriate and a patient is correctly estimating carbohydrates, the ICRs for specific meals can be adjusted. Reports that provide information on the percent-age of meals that require correction doses can also help to evaluate ICRs.

ConclusionsCGM has opened up a new world in the area of nutrition and diabetes. Clinicians and patients now have the opportunity to more effectively and easily evaluate their glycemic response to various types of foods and meals. This information provides patients the ability to more effectively adjust prandial insulin and lifestyle therapy based on their food choices. With the additional information available, clinicians have the responsibility to train patients in how to interpret the data and make appropriate decisions. The information will also facilitate researchers’ analysis of the PPG response to foods that may influence nutrition guidelines for individuals with diabetes.

AcknowledgmentsThe authors express appreciation to Jen Block, BS, RN, CDE, and Keri Weindel, MS, RD, CDE, for contributing CGM data manage-ment reports.

References:1Monnier L, Mas E, Ginet C, Michel F, Villon L, Cristol JP, Colette C: Activation of oxida-tive stress by acute glucose fluctuations com-pared with sustained chronic hyperglyce-mia in patients with type 2 diabetes. JAMA 295:1681–1687, 2006

Figure 3. Glycemic response to a pizza meal illustrated on a line graph (DexCom Data Manager 2 Software). At 8:00 p.m., the patient began eat-ing three slices of cheese, chicken, and garlic pizza at the time of the insulin injection. Peak glucose occurred 3.5 hours later, and a correction bolus was given. Full recovery from postprandial hyperglycemia occurred 5.5 hours after start of the meal. (Courtesy of DexCom)

Diabetes Spectrum Volume 21, Number 2, 2008 137

Nutrition FYI

2American Diabetes Association: Nutrition recommendations and interventions for dia-betes [Position Statement]. Diabetes Care 31 (Suppl. 1):S61–S78, 20083Daenen S, Sola A, Larger E, Elgrably F, Slama G: Use of the CGMS to assess the op-timal time to measure postprandial glucose [Abstract]. Diabetes 55 (Suppl. 1):17A, 20064Diabetes Research in Children Network (DirecNet) Study Group: Continuous glucose monitoring in children with type 1 diabetes. J Pediatr 151:388–393, 20075Sheard NF, Clark NG, Brand-Miller JC, Franz MJ, Pi-Sunyer FX, Mayer-Davis E, Kulkarni K, Geil P: Dietary carbohydrate (amount and type) in the prevention and management of diabetes [American Diabetes Association Statement]. Diabetes Care 27:2266–2271, 20046Grandrud L, Xing D, Kollman C, Block J, Kunselman B, Wilson D, Buckingham B: The Medtronic Gold Continuous Glucose

Monitoring System: an effective means to discover hypo- and hyperglycemia in chil-dren under 7 years of age. Diabetes Technol Therapeut 9:310–316, 20077Hovorka R: CGM and closed-loop systems. Diabet Med 23:1–12, 20058Buckingham B, Caswell K, Wilson D: Real time continuous glucose monitors. Curr Opin Endocrinol Diabetes Obes 14:288–295, 20079Wolpert HA: The nuts and bolts of achiev-ing end points with real-time continuous glu-cose monitoring. Diabetes Care 31:S146–S149, 200810Jones S, Quarry J, Caldwell-McMillan M, Mauger D, Gabbay R: Optimal insulin pump dosing and postprandial glycemia following a pizza meal using continuous glucose mon-itoring system. Diabetes Technol Therapeut 7:233–239, 200511Diabetes Research in Children Network (DirecNet) Study Group: Use of the DirecNet

Applied Treatment Algorithm (DATA) for di-abetes management with a real-time con-tinuous glucose monitor (the FreeStyle Navigator). Pediatr Diabetes. Published on-line 24 January 2008. Available from www.blackwell-synergy.com.

Janine Freeman, RD, LD, CDE, is a clinical science manager for Abbott Diabetes Care in Atlanta, Ga. Lynne Lyons, MPH, RD, CDE, is a clinical science manager for Abbott Diabetes Care in San Francisco, Calif.

Note of disclosure: Ms. Freeman and Ms. Lyons are employees of and stock shareholders in Abbott Diabetes Care, which manufactures a CGM device.