American Diabetes Association Guide to Nutrition Therapy for Diabetes. Marion J. Franz

improvements in blood glucose were the result of physiological insulin replacement combined with self-management of insulin dose adjustment in conjunction with a “flexible” meal-planning approach such as carbohydrate counting. Excellent reviews are available to guide insulin initiation and management of insulin therapy to achieve optimal glycemic control; therefore, this topic will not be extensively reviewed here (DeWitt 2003; Mooradian 2006).

      The chapter begins with a review of the literature pertaining to insulin dose algorithms, especially as they relate to insulin-to-carbohydrate ratios, timing of prandial insulin doses, and factors contributing to postprandial hyperglycemia. Nutrition therapy using carbohydrate-counting meal-planning approaches is reviewed and incorporates research published after 1 September 2009.

INSULIN THERAPY

      Insulin-Dosing Algorithms

      Since the advent of basal-bolus insulin therapy and SMBG, there has been interest in developing insulin-dosing algorithms for the self-adjustment of insulin therapy, and a variety of formulas or “rules” have been developed (Skyler 1981; Palerm 2007). Prandial dosing algorithms are commonly referred to as an insulin-to-carbohydrate ratio (ICR), and blood glucose correction algorithms are frequently referred as an insulin sensitivity factor (ISF). Guidelines also exist for estimating basal insulin as a fraction of the total daily dose (TDD) (see Table 5.1 for definitions of basal-bolus insulin terms).

      Table 5.1 Basal-Bolus Insulin Therapy Terms

      • Physiological insulin replacement: The insulin secretion of normally functioning b-cells is mimicked with prandial (mealtime) or “bolus” insulin and long-acting or “basal” insulin.

      • Basal insulin: Basal insulin represents the level of insulin always present in fasting and postmeal states. Basal insulin suppresses excess gluconeogenesis and the release of free fatty acids and enables the glucose transport in the fasting state. Insulin glargine and detemir are examples of “basal” insulin and are essentially peakless. Neutral protamine Hagedorn (NPH) can also be used as a basal insulin, but it has a pronounced peak. Basal insulin delivery can also be delivered via an insulin pump using either rapid- or short-acting insulin (see examples listed under “Bolus insulin” below). Basal insulin typically represents ~50% of the TDD insulin required each day.

      • Bolus insulin: Rapid-acting insulin (lispro, aspart, and glulisine) or short-acting insulin (regular) can be used at mealtimes and/or to correct an elevated blood glucose level. In the case of rapid-acting insulin, onset is typically 5–15 min and peak action is at 30–90 min after injection. The duration of rapid-acting insulin is ~4–6 h in most people. This is the “other” 50% of the TDD insulin required each day.

      • Total daily dose (TDD): TDD is the total number of units taken each day to achieve desirable blood glucose control. This number includes basal insulin plus bolus insulin used to “cover” the carbohydrates consumed at meals and the blood glucose correction needs.

      PRANDIAL INSULIN: • ICR or carbohydrate factor (CarbF): This ratio indicates how many grams of carbohydrate are “covered” or “matched” by 1 unit rapid- or short-acting insulin. The ICR is related to an individual’s insulin sensitivity and body size. An ICR for an adult with type 1 diabetes is typically 1:10 (1 unit of rapid-acting insulin is needed to “match” 10 g carbohydrate), whereas an obese person might require an ICR of 1:5 (1 unit of rapid-acting insulin to “match” 5 g carbohydrate due to insulin resistance).

      GLUCOSE CORRECTION INSULIN: • ISF or glucose correction factor (CorrF or CF): The ISF can be defined as the estimated drop in blood glucose (mg/dL) expected from the administration of 1 unit of rapid- or short-acting insulin. ISF is also related to the individual’s insulin sensitivity and body size. A typical ISF for an adult with type 1 diabetes is typically 1:50 mg/dL (1 unit of rapid-acting insulin will drop the blood glucose level 50 mg/dL). However, an overweight/insulin-resistant person may have an ISF of 1:20 mg/dL (1 unit rapid-acting insulin will drop the blood glucose level 20 mg/dL).

      The “1500 rule” for blood glucose correction was an original formula based on informal clinical observations using regular insulin by Davidson and his colleagues in 1982 (Table 5.2) (Davidson 2008). In 1998, Davidson and colleagues developed a formula for the ICR, and in 2008, they published a paper on how their mathematical models for basal-bolus insulin-dosing guidelines in patients with type 1 diabetes were derived from a retrospective controlled study (Davidson 2008). The goal of their analysis was to determine how best to prescribe insulin for use in continuous subcutaneous insulin infusion (CSII) pump therapy in their large endocrine practice. These formulae were statistically correlated and are referred to as the Accurate Insulin Management (AIM) formulae (Davidson 2002; Davidson 2003). The AIM system guidelines are based on the TDD. If the TDD is not available, the AIM system includes an additional formula to estimate TDD based on body weight in pounds (Table 5.2).

      Table 5.2 Insulin Dosing Formula or “Rules”

      In 1990, Howorka and colleagues also developed prandial algorithms (meal-related and correctional insulin doses for blood glucose increases induced by carbohydrate) for personalized adjustment with flexible insulin dosing (Table 5.3) (Howorka 1990). These formulae are still used by some researchers outside the U.S. In an observational study of 35 patients with type 1 diabetes using the Howorka flexible intensive insulin therapy algorithms, the authors concluded that use of these individualized parameters permitted fast and accurate adjustment of insulin doses (Franc 2009). Howorka also proposed additional algorithms for protein/fat in meals low in carbohydrate. However, the protein/fat algorithms have not been extensively studied, and patients are not being taught to use these currently in clinical practice. Another researcher proposed the addition of 1 unit of insulin per 20 g protein (Sachon 2003).

      Table 5.3 Howorka Algorithms for Flexible Insulin Therapy

      Other researchers also determined ICR through use of the hyperinsulinemic-euglycemic clamp with a meal challenge to determine a prandial dosing algorithm in people with type 1 diabetes (Bevier 2007). In 1994, based on clinical experience, a carbohydrate factor formula was introduced: a “450 rule” for multiple daily injection and CSII patients (Walsh 1994). An updated carbohydrate factor formula and a glucose correction factor formula were published in the consumer book Using Insulin (Table 5.2) (Walsh 2003). The authors have further refined their formula after analyzing anonymous consecutive downloads from 1,020 pumps collected during a routine pump software upgrade in 2007 (Walsh 2010).

      It should be noted that a majority of these popular ICR and ISF formulae were derived from patient populations using CSII with well-controlled blood glucose levels, not multiple daily injections. The authors of the AIM system recommend use of their formula for patients with type 1 diabetes, noting that only one person with insulin-requiring type 2 diabetes was included in their data analysis (Davidson 2008).

      Timing of Prandial Insulin Dose

      DeWitt and Hirsch described the term “lag time” in their reviews of outpatient insulin therapy in type 1 and type 2 diabetes (Hirsch 2005a; DeWitt 2003). Lag time is defined as the amount of time that elapses between the prandial insulin injection and a meal; lag time is critical in the control of postprandial hyperglycemia and in later risk of hypoglycemia. Given the pharmacodynamics of insulin analogs, sufficient lag time helps to decrease postprandial hyperglycemia (Rassam 1999). The use of lag time for rapid-acting insulin and improved matching action with carbohydrate absorption explains their clinical advantage (DeWitt 2003). In the pre-analog era, it was recommended to administer regular insulin ~20–30 min before eating a meal. With the introduction of rapid-acting insulin analogs, many physicians recommended injection of the prandial dose just before eating, or even after eating. However, these recommendations are not supported by the insulin action times reported by the insulin manufacturers (Hirsch 2005a). The