3106 PART 12 Endocrinology and Metabolism

emotional burdens…in having to manage a chronic disease like diabetes,” should be recognized and may require the care of a mental health

specialist. Emotional stress may provoke a change in behavior so that

individuals no longer adhere to a dietary, exercise, or therapeutic regimen. Eating disorders, including binge eating disorders, bulimia, and

anorexia nervosa, appear to occur more frequently in individuals with

type 1 or type 2 DM.

■ MONITORING THE LEVEL OF

GLYCEMIC CONTROL

Optimal monitoring of glycemic control involves glucose measurements by the patient and an assessment of long-term control by the

providers on the diabetes management team (measurement of hemoglobin A1c [HbA1c] and review of the patient’s SMBG and/or CGM).

These measurements are complementary: the patient’s measurements

provide a picture of short-term glycemic control, whereas the HbA1c

reflects average glycemic control over the previous 2–3 months. Most

measurements should be performed prior to a meal and supplemented

with postprandial measurements to assist in reaching glucose targets

(Table 404-2). By combining glucose measurements with diet and exercise history, the diabetes management team and patient can improve

glycemic control. Clinical practice is changing rapidly with CGM

replacing SMBG is many patients, especially those with T1 DM.

Self-Monitoring of Blood Glucose In SMBG, a small drop of

blood (3–10 μL) and an enzymatic reaction allow rapid and accurate

measurement of the capillary blood glucose by glucose monitors (calibrated to provide plasma glucose value even though blood glucose

is measured). The blood is obtained from the fingertip; alternative

testing sites (e.g., forearm) are less reliable. The frequency of SMBG

measurements should be individualized. Individuals with type 1 DM

or individuals with type 2 DM taking multiple insulin injections each

day should measure their blood glucose >3 times/day (some measure

>10 times/day). Most individuals with type 2 DM require less frequent

monitoring, although the optimal frequency of SMBG has not been

clearly defined. Individuals with type 2 DM who are taking insulin

should use SMBG more frequently than those on oral agents. Individuals with type 2 DM who are on oral medications should use SMBG

as a means of assessing the efficacy of their medication and the impact

of dietary choices and exercise. Because glucose levels fluctuate less in

these individuals, one or two SMBG measurements per day may be

sufficient.

Continuous Glucose Monitoring CGM technology utilizes a

sensor or electrode to detect interstitial glucose, which is in equilibrium with blood glucose, but may lag behind when the blood glucose

is changing. In one CGM approach, the interstitial glucose is detected

and reported essentially continuously while in another approach, the

sensor is in place, but the glucose is only recorded when a detector is

placed over the sensor. The glucose sensors are placed subcutaneously

and are replaced every 3–14 days. Some CGM requires calibration

by SMBG. CGM provides unlimited glucose datapoints that can be

used to define a time in a glycemic range (TIR, or time in range), the

ambulatory glucose profile, the amount of time in the hypoglycemic

range, and the glucose management indicator (GMI), which correlates

with A1C (Fig. 404-1; Table 404-2). TIR and GMI are useful metrics

but CGM also allows the patient to monitor the rate of glucose change

and glucose trends that can be used to avoid predicted hyper- or hypoglycemia. CGM in type 1 DM especially in those with hypoglycemia

unawareness can decrease the frequency of serious hypoglycemia

(especially nocturnal hypoglycemia). The combination of an insulin-infusion device (discussed below) and a CGM can now automate

insulin delivery with either predictive suspension of insulin delivery to

avoid hypoglycemia or closed-loop control that automatically adjusts

insulin delivery by a predictive algorithm (Fig. 404-1).

Assessment of Long-Term Glycemic Control Measurement

of glycated hemoglobin (HbA1c) is the standard method for assessing

long-term glycemic control. When plasma glucose is consistently elevated, there is an increase in nonenzymatic glycation of hemoglobin; this

alteration reflects the glycemic history over the previous 2–3 months,

because erythrocytes have an average life span of 120 days (glycemic

A

B

Type 1 & Type 2

Diabetes

Target

<5%

<25%

<250 mg/dL

(13.9 mmol/L)

>180 mg/dL

(10.0 mmol/L)

Target Range:

70–180 mg/dL

(3.9–10.0 mmol/L)

<70 mg/dL (3.9 mmol/L)

<54 mg/dL (3.0 mmol/L)

C

>70%

<4%

<1%

0

Rapid (aspart, lispro, glulisine, inhaled human insulin)

Short (regular U-100)

Mixed short/intermediate (regular U-500)

Intermediate (NPH)

Long (determir)

Long (U-100 glargine)

Ultra-long (glargine U-300)

Ultra-long (degludec)

Plasma Insulin Levels

2468 10 12 14 16 18

Time (hr)

20 22 24 26 28 30 32 34 36

D

FIGURE 404-1 Glycemic monitoring and insulin administration options for treatment of diabetes. A. CGM profile and delivery of rapid-acting insulin analog by

continuous subcutaneous insulin infusion pump involves a basal rate (light purple line) and prandial and correction boluses (purple circles) based on estimated

carbohydrate intake (orange squares) and an insulin sensitivity factor. B. CGM profile with sensor-communicating insulin pump that automates insulin delivery by

suspending delivery for predicted hypoglycemia and increasing basal delivery for predicted hyperglycemia (light purple curves) while still requiring user input for

estimated carbohydrate intake (orange squares) to provide prandial insulin boluses (purple circles). C. CGM profile is used to generate an estimate of time-in-range

with glycemic goal shown on the left side of the bar and target % time in that glycemic range shown on the right side of the bar. D. Pharmacokinetic profile of individual

insulin products. (C. Reproduced with permission from T Battelino et al: Clinical targets for continuous glucose monitoring data interpretation: Recommendations from

the International Consensus on Time in Range. Diabetes Care 42:1593, 2019; D. Adapted with permission from JJ Neumiller: Insulin update: New and emerging insulins.

American Diabetes Association, 2018.)

3107 Diabetes Mellitus: Management and Therapies CHAPTER 404

level in the preceding month contributes about 50% to the HbA1c

value). Measurement of HbA1c at the “point of care” allows for more

rapid feedback and may therefore assist in adjustment of therapy.

HbA1c should be measured in all individuals with DM during their

initial evaluation and as part of their comprehensive diabetes care.

As the primary predictor of long-term complications of DM, the HbA1c

should mirror, to a certain extent, the short-term measurement by

SMBG or CGM. Measurements of HbA1c and actual glucose levels are

complementary in that recent intercurrent illnesses may impact SMBG

or CGM measurements but not the HbA1c. The HbA1c may reflect

postprandial or nocturnal hyperglycemia not detected by SMBG of

fasting and preprandial capillary blood glucose. However, it does not

detect interprandial or nocturnal hypoglycemia—these require very

frequent SMBG or CGM for detection. The HbA1c is an “average” and

thus does not detect glycemic variability in the way SMBG and CGM

can. In standardized assays, the HbA1c approximates the following

mean plasma glucose values: an HbA1c of 6% = 7.0 mmol/L (126 mg/

dL), 7% = 8.6 mmol/L (154 mg/dL), 8% = 10.2 mmol/L (183 mg/dL),

9% = 11.8 mmol/L (212 mg/dL), 10% = 13.4 mmol/L (240 mg/dL),

11% = 14.9 mmol/L (269 mg/dL), and 12% = 16.5 mmol/L (298 mg/

dL). However, there is interindividual variability in the HbA1c to mean

glucose relationship, and in blacks the HbA1c is on average 0.4% higher

than in whites for the same mean glucose. Clinical conditions leading

to abnormal RBC parameters such as hemoglobinopathies, anemias,

reticulocytosis, transfusions, and uremia may alter the HbA1c result.

In patients achieving their glycemic goal, the ADA recommends measurement of the HbA1c at least twice per year. More frequent testing

(every 3 months) is warranted when glycemic control is inadequate

or when therapy has changed. Laboratory standards for the HbA1c test

have been established and should be correlated to the reference assay

of the Diabetes Control and Complications Trial (DCCT). The degree

of glycation of other proteins, such as albumin, or measurement of

1,5-anhydroglucitol can be used as an alternative, shorter-term indicator of glycemic control when the HbA1c is inaccurate. The fructosamine

assay (measuring glycated albumin) reflects the glycemic status over

the prior 2 weeks.

PHARMACOLOGIC TREATMENT

OF DIABETES

Comprehensive care of type 1 and type 2 DM requires an emphasis on

nutrition, exercise, and monitoring of glycemic control but also usually

involves glucose-lowering medication(s). This chapter discusses classes

of such medications but does not describe every glucose-lowering

agent available worldwide. The initial step is to select an individualized,

glycemic goal for the patient.

■ ESTABLISHMENT OF TARGET LEVEL OF

GLYCEMIC CONTROL

Because the complications of DM are related to glycemic control,

normoglycemia or near-normoglycemia is the desired, but often elusive, goal for most patients. Normalization or near-normalization of

the plasma glucose for long periods of time is extremely difficult, as

demonstrated by the DCCT and United Kingdom Prospective Diabetes

Study (UKPDS). Regardless of the level of hyperglycemia, improvement in glycemic control will lower the risk of diabetes-specific complications, most notably the microvascular complications (Chap. 405).

The target for glycemic control (as reflected by the HbA1c) must be

individualized, and the goals of therapy should be developed in consultation with the patient after considering a number of medical, social,

and lifestyle issues. The ADA calls this a patient-centered approach, and

other organizations such as the IDF and American Association of Clinical Endocrinologists (AACE) also suggest an individualized glycemic

goal. Important factors to consider include the patient’s age and ability

to understand and implement a complex treatment regimen, presence

and severity of complications of diabetes, known CVD, ability to recognize hypoglycemic symptoms, presence of other medical conditions or

treatments that might affect survival or the response to therapy, lifestyle

and occupation (e.g., possible consequences of experiencing hypoglycemia on the job), and level of support available from family and friends.

In general, the ADA suggests that the goal is to achieve an HbA1c as

close to normal as possible without significant hypoglycemia. In most

individuals, the target HbA1c should be <7% (Table 404-2) with a more

stringent (≤6.5%) target for some patients. With modern implementation of intensive insulin therapy for type 1 DM, the level of HbA1c is no

longer inversely related to the frequency and severity of hypoglycemia as

seen in the DCCT; nevertheless, it may still be appropriate to set a higher

HbA1c target <7.5 or 8% for patients with impaired awareness of hypoglycemia. A higher HbA1c goal may also be appropriate for the very young

or old or in individuals with limited life span or comorbid conditions.

For individuals using CGM, maximizing time-in-range 70–180 mg/dL,

representing normoglycemia, while minimizing time-below-range <70

mg/dL, representing hypoglycemia, are shorter-term targets of therapy.

More stringent glycemic control (HbA1c ≤6%) is not beneficial, and

may be detrimental, in patients with type 2 DM and a high risk of

CVD. Large clinical trials (UKPDS, Action to Control Cardiovascular

Risk in Diabetes [ACCORD], Action in Diabetes and Vascular Disease:

Preterax and Diamicron MR Controlled Evaluation [ADVANCE],

Veterans Affairs Diabetes Trial [VADT]; Chap. 405) examined glycemic control in type 2 DM in individuals with low risk of CVD, with

high risk of CVD, or with established CVD and have found that more

intense glycemic control is not beneficial and, in some patient populations, may have a negative impact on some outcomes. These divergent

outcomes stress the need for individualized glycemic goals based on

the following general guidelines: (1) early in the course of type 2 diabetes when the CVD risk is lower, improved glycemic control likely

leads to improved cardiovascular outcome, but this benefit may occur

more than a decade after the period of improved glycemic control; (2)

intense glycemic control in individuals with established CVD or at

high risk for CVD is not advantageous, and may be deleterious, over a

follow-up of 3–5 years; (3) hypoglycemia in such high-risk populations

(elderly, CVD) should be avoided; and (4) improved glycemic control

reduces microvascular complications of diabetes (Chap. 405) even if it

does not improve macrovascular complications like CVD.

■ TYPE 1 DIABETES MELLITUS

General Aspects The ADA recommendations for glycemic goals

and HbA1c targets are summarized in Table 404-2. The goal is to

design and implement insulin regimens that mimic physiologic insulin

secretion. Because individuals with type 1 DM partially or completely

lack endogenous insulin production, administration of basal insulin

is essential for regulating glycogen breakdown, gluconeogenesis,

lipolysis, and ketogenesis (i.e., largely fine-tuning hepatic and adipose

metabolism). Likewise, insulin replacement for meals should be appropriate for the carbohydrate intake and insulin sensitivity, promoting

normal glucose utilization and storage.

Intensive Management Intensive insulin therapy has the goal

of achieving near-normal glycemia. This approach requires multiple

resources, including thorough and continuing patient education, comprehensive recording of plasma glucose measurements and nutrition

intake by the patient, and a variable insulin regimen that matches

carbohydrate intake and insulin dose. Insulin regimens include multiple-component insulin regimens, multiple daily injections (MDIs), or

continuous subcutaneous (SC) insulin infusion (CSII).

The benefits of intensive insulin therapy and improved glycemic control include a reduction in the acute metabolic and chronic microvascular complications of DM. From a psychological standpoint, the patient

experiences greater control over his or her diabetes and often notes an

improved sense of well-being, greater flexibility in the timing and content of meals, and the capability to alter insulin dosing with exercise.

In addition, intensive insulin therapy prior to and during pregnancy

reduces the risk of fetal malformations and morbidity. Intensive insulin

therapy is encouraged in newly diagnosed patients with type 1 DM

because it may prolong the period of C-peptide production, which may

result in better glycemic control and a reduced risk of serious hypoglycemia. Although intensive management confers impressive benefits,

it is also accompanied by significant personal and financial costs and

therefore may not be appropriate at all times for all individuals.

3108 PART 12 Endocrinology and Metabolism

TABLE 404-4 Properties of Insulin Preparationsa

PREPARATION

TIME OF ACTION

ONSET, h PEAK, h

EFFECTIVE

DURATION, h

Short-acting

Aspartb <0.25 0.5–1.5 3–5

Glulisine <0.25 0.5–1.5 3–5

Lisproc <0.25 0.5–1.5 3–5

Regulard 0.5–1.0 2–3 4–8

Inhaled human insulin <0.5 1–2 3

Long-acting

Degludec 1–9 —e 42f

Detemir 1–4 —e 12–24f

Glargineg 2–4 —e 20–24

NPH 2–4 4–10 10–16

Examples of insulin combinationsh

 75/25–75% protamine lispro,

25% lispro

<0.25 Duali 10–16

 70/30–70% protamine aspart,

30% aspart

<0.25 Duali 15–18

 50/50–50% protamine lispro, 50%

lispro

<0.25 Duali 10–16

70/30–70% NPH, 30% regular 0.5–1 Duali 10–16

Combination of long-acting insulin

and GLP-1 receptor agonist

See text

a

Injectable insulin preparations (with exception of inhaled formulation) available in

the United States; others are available in the United Kingdom and Europe. Standard

formulations are U-100 (100 units of insulin per mL solution). b

Formulation with

niacinamide (vitamin B3) has a slightly more rapid onset and offset. c

Lispro-aabc

formulation has a slightly more rapid onset and offset; both formulations are also

available in U-200 concentration. d

Formulation also available in U-500 concentration

with delayed onset and offset. e

Degludec, determir, and glargine have minimal

peak activity. d

Duration is dose-dependent. g

Formulation also available in U-300

concentration with delayed onset and offset. h

Other insulin combinations are

available. i

Dual: two peaks—one at 2–3 h and the second one several hours later.

Insulin Preparations Current insulin preparations are generated by recombinant DNA technology and consist of the amino acid

sequence of human insulin or variations thereof. In the United States,

most insulin is formulated as U-100 (100 units/mL); short-acting

insulin formulated as U-200 (200 units/mL; lispro) and long-acting as

U-300 (300 units/mL; glargine) are available in order to limit injection

volumes for patients with high insulin requirements. Regular insulin

formulated as U-500 (500 units/mL) is sometimes used in patients with

severe insulin resistance. Human insulin has been formulated with distinctive pharmacokinetics (regular and neutral protamine Hagedorn

[NPH] insulin have the native insulin amino acid sequence) or genetically modified to alter insulin absorption and hence insulin action.

Insulins can be classified as short-acting or long-acting (Table 404-4;

Figure 404-1D). For example, one short-acting insulin formulation,

insulin lispro, is an insulin analogue in which the 28th and 29th amino

acids (lysine and proline) on the insulin B chain have been reversed by

recombinant DNA technology. Insulin aspart and insulin glulisine are

genetically modified insulin analogues with properties similar to lispro.

A biosimilar version of lispro has been approved. These insulin analogues have full biologic activity but less tendency for self-aggregation,

resulting in more rapid absorption and onset of action and a shorter

duration of action. These characteristics are particularly advantageous

for allowing entrainment of insulin injection and action to rising

plasma glucose levels following meals. The shorter duration of action

also appears to be associated with a decreased number of hypoglycemic

episodes, primarily because the decay of insulin action corresponds to

the decline in plasma glucose after a meal. Thus, insulin aspart, lispro,

or glulisine is preferred over regular insulin for prandial coverage in

many patients. Insulin glargine is a long-acting biosynthetic human

insulin that differs from normal insulin in that asparagine is replaced

by glycine at amino acid 21, and two arginine residues are added to

the C terminus of the B chain, leading to the formation of microprecipitates at physiologic pH in subcutaneous tissue. Compared to NPH

insulin, the onset of insulin glargine action is later, the duration of

action is longer (~24 h), and there is a less pronounced peak. A lower

incidence of hypoglycemia, especially at night, has been reported with

insulin glargine when compared to NPH insulin. A biosimilar version

is available. Insulin detemir has a fatty acid side chain that reversibly

binds to albumin and prolongs its action by slowing absorption and

catabolism, but its duration of action may only reach 12–20 h. Twicedaily injections of glargine, or especially detemir, are sometimes

required to provide optimal 24-h basal insulin coverage. Because of

modification and extension of the carboxy-terminus of the B chain,

insulin degludec forms multihexamers in subcutaneous tissue and

binds albumin, prolonging its duration of action (>42 h); it provides

similar glycemic control as glargine but with less frequent nocturnal

and severe hypoglycemia. Other modified insulins, such as one with a

duration of action of several days, are in development.

Basal insulin requirements are provided by long-acting insulin formulations (NPH insulin, insulin glargine, insulin detemir, or insulin

degludec) (Fig. 404-1D; Table 404-4). These are usually prescribed

with short-acting insulin in an attempt to mimic physiologic insulin

release with meals. Although mixing of NPH and short-acting insulin

formulations is common practice, this mixing may alter the insulin

absorption profile (especially the short-acting insulins). For example,

lispro absorption is delayed by mixing with NPH. The alteration in

insulin absorption when the patient mixes different insulin formulations should not prevent mixing insulins. However, the following

guidelines should be followed: (1) mix the different insulin formulations in the syringe immediately before injection (inject within 2 min

after mixing); (2) do not store insulin as a mixture; (3) follow the same

routine in terms of insulin mixing and administration to standardize

the physiologic response to injected insulin; and (4) do not mix insulin

glargine, detemir, or degludec with other insulins. The miscibility of

some insulins allows for the production of combination insulins that

contain 70% NPH and 30% regular (70/30), or equal mixtures of NPH

and regular (50/50). By including the insulin analogue mixed with

protamine, several additional combinations have a short-acting and

long-acting profile (Table 404-4; Fig. 404-1D). Although more convenient for the patient (only two injections/day), combination insulin

formulations do not allow independent adjustment of short-acting

and long-acting activity. Several insulin formulations are available as

insulin “pens,” which are more convenient for some patients. Insulin

delivery by inhalation to provide meal-time insulin is approved, but not

widely used. Prior to its use, the forced expiratory volume in 1 second

(FEV1

) should be measured. Inhaled insulin can cause bronchospasm

and cough and should not be used by individuals with lung disease

or those who smoke. Long-acting insulin/glucagon-like peptide-1

(GLP-1) receptor agonist combinations in fixed doses (degludec +

liraglutide or glargine + lixisenatide) are effective, and are associated

with less weight gain.

Insulin Regimens There is considerable patient-to-patient variation in the peak and duration. In all regimens, long-acting insulins

(NPH, glargine, detemir, or degludec) supply basal insulin, whereas

regular, insulin aspart, glulisine, or lispro provide prandial insulin

(Fig. 404-1D; Table 404-4). Short-acting insulin analogues should be

injected just before (<10 min) and regular insulin 30–45 min prior to a

meal. Sometimes short-acting insulin analogues are injected just after

a meal (gastroparesis, unpredictable food intake).

A shortcoming of current insulin regimens is that injected insulin

immediately enters the systemic circulation, whereas endogenous insulin is secreted into the portal venous system. Thus, exogenous insulin

administration exposes the liver to subphysiologic insulin levels. No

current insulin regimen reproduces the precise insulin secretory pattern of the pancreatic islet. However, the most physiologic regimens

entail more frequent insulin injections, greater reliance on short-acting

insulin, and more frequent SMBG and/or CGM. In general, individuals

with type 1 DM require 0.3-0.7 units/kg per day of insulin divided into

3109 Diabetes Mellitus: Management and Therapies CHAPTER 404

multiple doses, with approximately 50% of daily insulin given as basal

insulin and 50% as prandial insulin.

MDI regimens refer to the combination of basal insulin and bolus

insulin (preprandial short-acting insulin). The timing and dose of

short-acting, preprandial insulin are altered to accommodate the

SMBG or CGM results, anticipated food intake, and physical activity.

Such regimens offer the patient with type 1 DM more flexibility in

terms of lifestyle and the best chance for achieving near normoglycemia. Most often basal insulin with glargine, detemir, or degludec is

used in conjunction with preprandial lispro, glulisine, or insulin aspart.

The insulin aspart, glulisine, or lispro dose is based on individualized

algorithms that integrate the preprandial glucose and the anticipated

carbohydrate intake. To determine the meal component of the preprandial insulin dose, the patient uses an insulin-to-carbohydrate ratio (a

common ratio for type 1 DM is 1 unit/10–15 g of carbohydrate, but this

must be determined for each individual). To this insulin dose is added

the supplemental or correcting insulin based on the preprandial blood

glucose (one formula uses 1 unit of insulin for every 1.6–3.3 mmol/L

[30–60 mg/dL] over the preprandial glucose target; this correction

factor can be estimated from 1500/[total daily insulin dose]). Such

calculations must be adjusted based on each individual’s sensitivity to

insulin. Other variations of this regimen use twice daily NPH as basal

insulin but have the disadvantage that NPH has a significant peak,

making hypoglycemia more common. Frequent SMBG (≥4 times per

day) or CGM is essential for these types of insulin regimens.

CSII is a very effective insulin regimen for the patient with type 1

DM (Fig. 404-1). To the basal insulin infusion, a preprandial insulin

(“bolus”) is delivered by the insulin infusion device based on instructions from the patient, who uses an individualized algorithm incorporating the preprandial plasma glucose and anticipated carbohydrate

intake. These sophisticated devices can accurately deliver small doses

of insulin (microliters per hour) and have several advantages: (1) multiple basal infusion rates can be programmed to accommodate nocturnal

versus daytime basal insulin requirement; (2) basal infusion rates can

be altered during periods of exercise; (3) different waveforms of insulin infusion with meal-related bolus allow better matching of insulin

depending on meal composition; and (4) programmed algorithms

consider ongoing action of prior insulin administration and blood

glucose values in calculating the insulin dose. These devices require

instruction by a health professional with considerable experience with

insulin infusion devices and frequent patient interactions with the diabetes management team. Insulin infusion devices may present unique

challenges, such as infection at the infusion site, unexplained hyperglycemia because the infusion set becomes obstructed, or diabetic ketoacidosis (DKA) if the insulin infusion device becomes disconnected.

Because most physicians use lispro, glulisine, or insulin aspart in CSII,

the extremely short half-life of these insulins quickly leads to insulin

deficiency if the delivery system is interrupted. Essential to the safe

use of infusion devices is thorough patient education, frequent SMBG

and/or CGM, and a backup plan for injecting long- and/or rapidacting insulins in the event of insulin infusion device failure. CGM

sensor-augmented insulin infusion devices integrate the information

from the CGM to inform insulin delivery (Fig. 404-1). Currently,

sensor communicating functions can interrupt basal insulin delivery

during hypoglycemia (threshold suspension) or when hypoglycemia is

anticipated (predictive suspension), which may be particularly useful

for addressing nocturnal hypoglycemia. Hybrid closed-loop systems

have recently become available that combine patient-directed preprandial boluses with automated adjustment of between-meal and basal

insulin delivery based on CGM. Clinical experience with closed-loop

systems is rapidly increasing and expanding. Bihormonal infusion

devices that deliver both insulin and glucagon are under development.

Other Agents That Improve Glucose Control The role of

amylin, a 37-amino-acid peptide co-secreted with insulin from pancreatic beta cells, in normal glucose homeostasis is uncertain. However,

based on the rationale that patients who are insulin deficient are also

amylin deficient, an analogue of amylin (pramlintide) was created and

found to reduce postprandial glycemic excursions in individuals with

type 1 or type 2 DM taking insulin. Pramlintide injected just before

a meal slows gastric emptying and suppresses glucagon but does not

alter insulin levels. Pramlintide is approved for insulin-treated patients

with type 1 or type 2 DM. Addition of pramlintide produces a modest

reduction in the HbA1c and seems to dampen meal-related glucose

excursions. In type 1 DM, pramlintide is started as a 15-μg SC injection

before each meal and titrated up to a maximum of 30–60 μg as tolerated. In type 2 DM, pramlintide is started as a 60-μg SC injection before

each meal and may be titrated up to a maximum of 120 μg. The major

side effects are nausea and vomiting, and dose escalations should be

slow to limit these side effects. Because pramlintide slows gastric emptying, it may influence absorption of other medications and should not

be used in combination with other drugs that slow gastrointestinal (GI)

motility. The short-acting insulin given before the meal should initially

be reduced to avoid hypoglycemia and then titrated as the effects of

the pramlintide become evident. Because pramlintide suppresses glucagon, it may worsen hypoglycemia recovery and should not be used

in patients with hypoglycemia unawareness.

■ TYPE 2 DIABETES MELLITUS

General Aspects The goals of glycemia-controlling therapy for

type 2 DM are similar to those in type 1 DM. Whereas glycemic control

tends to dominate the management of type 1 DM, the care of individuals with type 2 DM must also include attention to the treatment

of conditions associated with type 2 DM (e.g., obesity, hypertension,

dyslipidemia, CVD) and detection/management of DM-related complications (Fig. 404-2; Chap. 405). Reduction in cardiovascular risk is

of paramount importance because this is the leading cause of mortality

in these individuals.

Type 2 DM management should begin with MNT (discussed above).

An exercise regimen to increase insulin sensitivity and promote weight

loss should also be instituted. Pharmacologic approaches to the management of type 2 DM include oral glucose-lowering agents, insulin,

and other agents that improve glucose control; most physicians and

patients prefer oral glucose-lowering agents as the initial choice. Any

therapy that improves glycemic control reduces “glucose toxicity” to

beta cells and may improve endogenous insulin secretion. However,

type 2 DM is a progressive disorder and ultimately requires multiple

therapeutic agents and often insulin in most patients.

Glucose-Lowering Agents Advances in the therapy of type 2

DM have generated oral glucose-lowering agents that target different

pathophysiologic processes in type 2 DM. Based on their mechanisms

of action, glucose-lowering agents are subdivided into agents that

increase insulin secretion, reduce glucose production, increase insulin

sensitivity, enhance GLP-1 action, or promote urinary excretion of glucose (Table 404-5). Glucose-lowering agents other than insulin (with

the exception of amylin analogue) are ineffective in type 1 DM and

should not be used for glucose management of severely ill individuals

with type 2 DM. Insulin is sometimes the initial glucose-lowering agent

in type 2 DM.

Individualized

glycemic control

 • Diet/lifestyle

 • Exercise

 • Medication

Treat associated

conditions

 • Dyslipidemia

 • Hypertension

 • Obesity

Screen for/manage

complications

of diabetes

• Retinopathy

• Nephropathy

• Neuropathy

• Cardiovascular

 disease

• Other complications

Management of

Type 2 Diabetes

FIGURE 404-2 Essential elements in comprehensive care of type 2 diabetes.

3110 PART 12 Endocrinology and Metabolism

BIGUANIDES Metformin, representative of this class of agents, reduces

hepatic glucose production and improves peripheral glucose utilization

slightly (Table 404-5). Metformin activates AMP-dependent protein

kinase and enters cells through organic cation transporters (polymorphisms of these may influence the response to metformin). Metformin

acts in multiple tissues, but its mechanism of action remains undefined. There is evidence for reducing hepatic glucose production by

antagonizing cAMP generation in hepatocytes as well as for actions in

the gut. Metformin reduces fasting plasma glucose (FPG) and insulin

levels, improves the lipid profile, and promotes modest weight loss. An

extended-release form is available and may have fewer GI side effects

(diarrhea, anorexia, nausea, metallic taste). Because of its metformin’s

relatively slow onset of action and GI symptoms with higher doses,

the initial dose should be low and then escalated every 1–2 weeks to

a maximally tolerated dose of 2000 mg daily. Metformin is effective as

monotherapy and can be used in combination with other oral agents

or with insulin. Long-term use is associated with reduced micro- and

macrovascular complications. The major toxicity of metformin, lactic

acidosis, is very rare and can be prevented by careful patient selection.

Vitamin B12 levels are lower during metformin treatment and should be

monitored. Metformin should not be used in patients with moderate

renal insufficiency (glomerular filtration rate [GFR] <30 mL/min), any

TABLE 404-5 Agents Used for Treatment of Type 1 or Type 2 Diabetes

MECHANISM OF

ACTION EXAMPLESa

HBA1C REDUCTION

(%)b

AGENT-SPECIFIC

ADVANTAGES

AGENT-SPECIFIC

DISADVANTAGES CONTRAINDICATIONS

Oral

Biguanidesc* ↓ Hepatic glucose

production, ↑ insulin

sensitivity, influence

gut function

Metformin 1–2 Weight neutral, do not

cause hypoglycemia,

inexpensive,

extensive experience,

↓ CV events

Diarrhea, nausea,

lactic acidosis,

vitamin B12

deficiency

Renal insufficiency

(see text for GFR

<30 mL/min), CHF,

radiographic contrast

studies, hospitalized

patients, acidosis

α-Glucosidase

inhibitorsc**

↓ GI glucose

absorption

Acarbose, miglitol,

voglibose

0.5–0.8 Reduce postprandial

glycemia

GI flatulence,

elevated liver function

tests

Renal/liver insufficiency

Dipeptidyl peptidase

IV inhibitorsc***

Prolong endogenous

GLP-1 action; ↑

Insulin, ↓ glucagon

Alogliptin, linagliptin,

saxagliptin, sitagliptin,

vildagliptin

0.5–0.8 Well tolerated, do not

cause hypoglycemia

Angioedema/urticarial

and immune-mediated

dermatologic effects

Reduced dose with renal

insufficiency

Insulin

secretagogues:

Sulfonylureasc*

↑ Insulin secretion Glibornuride,

gliclazide,

glimepiride, glipizide,

gliquidone, glyburide,

glyclopyramide

1–2 Short onset of action,

lower postprandial

glucose, inexpensive

Hypoglycemia, weight

gain

Renal/liver insufficiency

Insulin

secretagogues:

Nonsulfonylureasc***

↑ Insulin secretion Mitiglinide,

nateglinide,

repaglinide

0.5–1.0 Short onset of action,

lower postprandial

glucose

Hypoglycemia Renal/liver insufficiency

(except repaglinide)

Sodium-glucose

cotransporter 2

inhibitorsc***

↑ Renal glucose

excretion

Canagliflozin,

dapagliflozin,

empagliflozin,

ertugliflozin

0.5–1.0 Do not cause

hypoglycemia, ↓

weight and BP, renal

protective, ↓ CV

events

Urinary and genital

infections, polyuria,

dehydration,

exacerbate tendency

to hyperkalemia and

DKA; see text

Moderate renal

insufficiency, insulindeficient DMf

Thiazolidinedionesc*** ↓ Insulin resistance,

↑ glucose utilization

Pioglitazone,

rosiglitazone

0.5–1.4 Lower insulin

requirements

Peripheral edema,

CHF, weight gain,

fractures, macular

edema

CHF, renal/liver

insufficiency

Parenteral/Oral

GLP-1 receptor

agonistsc***

↑ Insulin, ↓ glucagon,

slow gastric

emptying, satiety

Dulaglutide,

exenatide, liraglutide,

lixisenatide,

semaglutide (oral

formulation available)

0.5–1.0 Weight loss, do not

cause hypoglycemia

(unless combined

with another insulin

secretagogue or

insulin); ↓ CV events

Injection, nausea,

pancreatitise

Renal disease, agents

that also slow GI motility;

medullary carcinoma

of thyroid, pancreatic

disease

Parenteral

Amylin agonistsc,d*** Slow gastric

emptying, ↓ glucagon

Pramlintide 0.25–0.5 Reduce postprandial

glycemia, weight loss

Injection, nausea, ↑

risk of hypoglycemia

with insulin

Agents that also

slow GI motility

Insulinc,d**** ↑ Glucose utilization,

↓ hepatic glucose

production, and other

anabolic actions

See text and

Table 404-4

Not limited Known safety profile Injection, weight gain,

hypoglycemia

None

Medical nutrition

therapy and physical

activityc*

↓ Insulin resistance,

↑ insulin secretion

Low-calorie,

carbohydratecontrolled diet,

exercise

1–3 Other health benefits Compliance difficult,

long-term success

low

None

a

Examples are approved for use in the United States; others are available in other countries. Examples may not include all agents in the class. b

HbA1c reduction (absolute)

depends partly on starting HbA1c. c

Used for treatment of type 2 diabetes. d

Used in conjunction with insulin for treatment of type 1 diabetes. Cost of agent in the United

States: *

low, **moderate, ***high, ****variable. e

Degree of risk uncertain, avoid in individuals with risk factors for pancreatitis. f

Risk of euglycemic DKA) in patients with insulin

deficiency (e.g., type 1 diabetes).

Note: Some agents used to treat type 2 diabetes are not included in table (see text).

Abbreviations: CHF, congestive heart failure; CV, cardiovascular; GI, gastrointestinal; HbA1c, glycosylated hemoglobin A1c.

3111 Diabetes Mellitus: Management and Therapies CHAPTER 404

form of acidosis, unstable congestive heart failure (CHF), liver disease,

or severe hypoxemia. Metformin should be discontinued in hospitalized patients, in patients who can take nothing orally, and in those

receiving radiographic contrast material. Insulin should be used until

metformin can be restarted.

INSULIN SECRETAGOGUES—AGENTS THAT AFFECT THE ATP-SENSITIVE

K+ CHANNEL Insulin secretagogues stimulate insulin secretion by

interacting with the ATP-sensitive potassium channel on the beta cell

(Chap. 403). These drugs are most effective in individuals with type 2

DM of relatively recent onset (<5 years) who have residual endogenous

insulin production. First-generation sulfonylureas (chlorpropamide,

tolazamide, tolbutamide) have a longer half-life, a greater incidence

of hypoglycemia, and more frequent drug interactions, and are no

longer used. Second-generation sulfonylureas have a more rapid onset

of action and better coverage of the postprandial glucose rise, but the

shorter half-life of some agents may require more than once-a-day

dosing. Sulfonylureas reduce both fasting and postprandial glucose

and should be initiated at low doses and increased at 1- to 2-week

intervals based on SMBG. In general, sulfonylureas increase insulin

acutely and thus should be taken shortly before a meal; with chronic

therapy, though, the insulin release is more sustained. Long-term use

is associated with reduced micro- and macrovascular complications.

Glimepiride and glipizide can be given in a single daily dose and are

preferred over glyburide, especially in the elderly. Repaglinide, nateglinide, and mitiglinide are not sulfonylureas but also interact with

the ATP-sensitive potassium channel. Because of their short half-life,

these glinide agents are given immediately before each meal to reduce

meal-related glucose excursions.

Insulin secretagogues, especially the longer acting ones, have the

potential to cause hypoglycemia, especially in elderly individuals.

Hypoglycemia is usually related to delayed meals, increased physical

activity, alcohol intake, or renal insufficiency. Individuals who ingest

an overdose of some agents develop prolonged and serious hypoglycemia and should be monitored closely in the hospital (Chap. 406).

Most sulfonylureas are metabolized in the liver to compounds (some

of which are active, such as those of glyburide and the glinide nateglinide) that are cleared by the kidney. Thus, their use in individuals

with significant hepatic or renal dysfunction is not advisable. For

patients with chronic kidney disease requiring an insulin secretagogue, the shorter-acting sulfonylureas glimepiride or glipizide or the

glinide repaglinide may be used with caution. Weight gain, a common

side effect of sulfonylurea therapy, results from the increased insulin

levels and improvement in glycemic control. Some sulfonylureas have

significant drug interactions with alcohol and some medications

including warfarin, aspirin, ketoconazole, α-glucosidase inhibitors,

and fluconazole. A related isoform of ATP-sensitive potassium channels is present in the myocardium and the brain. All of these agents

except glyburide have a low affinity for this isoform. Despite concerns

that this agent might affect the myocardial response to ischemia and

observational studies suggesting that sulfonylureas increase cardiovascular risk, studies have not shown an increased cardiac mortality with

glyburide or other agents in this class.

INSULIN SECRETAGOGUES—AGENTS THAT ENHANCE GLP-1 RECEPTOR

SIGNALING “Incretins” amplify glucose-stimulated insulin secretion

(Chap. 403). Agents that either act as a GLP-1 receptor agonist or

enhance endogenous GLP-1 activity are approved for the treatment of

type 2 DM (Table 404-5). Agents in this class do not cause hypoglycemia because of the glucose-dependent nature of incretin-stimulated

insulin secretion (unless there is concomitant use of an agent that can

lead to hypoglycemia—sulfonylureas, etc.). GLP-1 receptor agonists

increase glucose-stimulated insulin secretion, suppress glucagon, and

slow gastric emptying. These agents do not promote weight gain; in

fact, most patients experience modest weight loss and appetite suppression. Short-acting GLP-1 receptor agonists are exenatide twice daily,

liraglutide daily, and lixisenatide daily. Long-acting GLP-1 receptor

agonists include sustained-release exenatide, dulaglutide, lixisenatide,

and semaglutide, all administered weekly. Short-acting GLP-1 receptor

agonists provide mostly postprandial coverage whereas the long-acting

GLP-1 receptor agonists reduce both the postprandial and fasting glucose. Daily oral semaglutide is now available that depends on gastric

absorption to avoid proteolytic degradation in the small intestine. All

are modified to avoid enzymatic inactivation by dipeptidyl peptidase

IV (DPP-IV) in the circulation.

For example, exenatide, a synthetic version of a peptide initially

identified in the saliva of the Gila monster (exendin-4), is an analogue

of GLP-1. Unlike native GLP-1, which has a half-life of ~2 min, differences in the exenatide amino acid sequence render it resistant to

DPP-IV. Thus, exenatide has prolonged GLP-1-like action. Liraglutide,

another GLP-1 receptor agonist, is almost identical to native GLP-1

except for an amino acid substitution and addition of a fatty acyl

group (coupled with a γ-glutamic acid spacer) that promote binding

to albumin and plasma proteins and prolong its half-life. Higher doses

of liraglutide and semaglutide than used for glucose-lowering effects

are effective for weight-loss therapy for obesity. Liraglutide treatment

has also been associated with a decrease in CVD events in patients

with type 2 DM and established CVD and with lower rates of diabetic

kidney disease. In similar patient populations, semaglutide treatment

has been associated with fewer CVD events and reduced diabetic

kidney disease, but with an increased rate of retinopathy-related complications, while dulaglutide treatment has been associated with both a

reduction in CVD events and in composite microvascular retinopathy

and nephropathy-related complications primarily driven by prevention of renal events. Similar reductions in CVD events have not been

observed with exenatide once weekly or lixisenatide. Treatment with

GLP-1 receptor agonists should start at a low dose to minimize initial

side effects (nausea being the limiting one). GLP-1 receptor agonists

can be used as combination therapy with metformin, sulfonylureas,

and thiazolidinediones. Some patients taking insulin secretagogues

may require a reduction in those agents to prevent hypoglycemia. The

major side effects are nausea, vomiting, and diarrhea. Some formulations carry a black box warning from the FDA because of an increased

risk of thyroid C-cell tumors in rodents and are contraindicated in

individuals with medullary carcinoma of the thyroid or multiple

endocrine neoplasia. Because GLP-1 receptor agonists slow gastric

emptying, they may influence the absorption of other drugs. Whether

GLP-1 receptor agonists enhance beta cell survival or promote beta cell

proliferation in humans as in rodents is not known, but these agents do

not appear to alter the natural history of type 2 DM.

DPP-IV inhibitors inhibit degradation of native GLP-1 and thus

enhance the incretin effect. DPP-IV, which is widely expressed on the

cell surface of endothelial cells and some lymphocytes, degrades a wide

range of peptides (not GLP-1 specific). DPP-IV inhibitors promote

insulin secretion in the absence of hypoglycemia or weight gain and

appear to have a preferential effect on postprandial blood glucose.

The levels of GLP-1 action in the patient are greater with the GLP-1

receptor agonists than with DPP-IV inhibitors. DPP-IV inhibitors are

used either alone or in combination with other oral agents in type 2

DM. Reduced doses should be given to patients with renal insufficiency. Allergy, including rash, hypersensitivity reactions (including

anaphylaxis, angioedema, and Stevens-Johnson syndrome), and severe

joint pain have been reported in association with DPP-IV inhibitors.

There is evidence concerning a potentially increased risk for acute

pancreatitis with GLP-1 receptor agonists and less so with DPP-IV

inhibitors. For now, it is reasonable to avoid these agents in patients

with pancreatic disease or with other significant risk factors for acute

pancreatitis (e.g., heavy alcohol use, severely elevated serum triglycerides, hypercalcemia).

α-GLUCOSIDASE INHIBITORS α-Glucosidase inhibitors reduce postprandial hyperglycemia by delaying glucose absorption; they do not

affect glucose utilization or insulin secretion (Table 404-5). Postprandial hyperglycemia, secondary to impaired hepatic and peripheral

glucose disposal, contributes significantly to the hyperglycemic state in

type 2 DM. These drugs, taken just before each meal, reduce glucose

absorption by inhibiting the enzyme that cleaves oligosaccharides into

simple sugars in the intestinal lumen. Therapy should be initiated at a

3112 PART 12 Endocrinology and Metabolism

low dose with the evening meal and increased to a maximal dose over

weeks to months. The major side effects (diarrhea, flatulence, abdominal distention) are related to increased delivery of oligosaccharides

to the large bowel and can be reduced somewhat by gradual upward

dose titration. α-Glucosidase inhibitors may increase levels of sulfonylureas and increase the incidence of hypoglycemia. Simultaneous

treatment with bile acid resins and antacids should be avoided. These

agents should not be used in individuals with inflammatory bowel

disease, gastroparesis, or a serum creatinine >177 μmol/L (2 mg/dL).

This class of agents is not as potent as other oral agents in lowering the

HbA1c but is unique because it reduces the postprandial glucose rise. If

hypoglycemia from other diabetes treatments occurs while taking these

agents, the patient should consume glucose because the degradation

and absorption of complex carbohydrates will be retarded.

THIAZOLIDINEDIONES Thiazolidinediones (Table 404-5) reduce

insulin resistance by binding to the peroxisome proliferator-activated

receptor γ (PPAR-γ) nuclear receptor (which forms a heterodimer

with the retinoid X receptor). The PPAR-γ receptor is found at highest levels in adipocytes but is expressed at lower levels in many other

tissues. Agonists of this receptor regulate a large number of genes,

promote adipocyte differentiation, reduce hepatic fat accumulation, and promote fatty acid storage. Thiazolidinediones promote a

redistribution of fat from central to peripheral locations. Circulating

insulin levels decrease with use of the thiazolidinediones, indicating a reduction in insulin resistance. Although direct comparisons

are not available, the two currently available thiazolidinediones

appear to have similar efficacy. The prototype of this class of drugs,

troglitazone, was withdrawn from the U.S. market after reports of

hepatotoxicity and an association with an idiosyncratic liver reaction

that sometimes led to hepatic failure. Although rosiglitazone and

pioglitazone do not appear to induce the liver abnormalities seen

with troglitazone, the FDA recommends measurement of liver function tests prior to initiating therapy. Modestly increased transaminase

levels related to underlying fatty liver disease should not preclude

treatment as these levels may improve with thiazolidinediones due to

a reduction in hepatic fat content.

Rosiglitazone raises low-density lipoprotein (LDL), high-density

lipoprotein (HDL), and triglycerides slightly. Pioglitazone raises HDL

to a greater degree and LDL a lesser degree but lowers triglycerides.

The clinical significance of the lipid changes with these agents is not

known and may be difficult to ascertain because most patients with

type 2 DM are also treated with a statin.

Thiazolidinediones are associated with weight gain (2–3 kg), a small

reduction in the hematocrit, and a mild increase in plasma volume.

Peripheral edema and CHF are more common in individuals treated

with these agents. These agents are contraindicated in patients with

hepatic insufficiency or CHF (class III or IV). The FDA has issued an

alert that rare patients taking these agents may experience a worsening

of diabetic macular edema. An increased risk of fractures has been

noted in postmenopausal women taking these agents. Thiazolidinediones have been shown to induce ovulation in premenopausal women

with polycystic ovary syndrome. Women should be warned about the

risk of pregnancy because the safety of thiazolidinediones in pregnancy

is not established.

Concerns about increased cardiovascular risk associated with

rosiglitazone led to considerable restrictions on its use and to the FDA

issuing a black box warning in 2007. However, based on new information, the FDA has revised its guidelines and categorizes rosiglitazone

similar to other drugs for type 2 DM. According to an FDA review,

pioglitazone may be associated with an increased risk of bladder cancer. In one study, pioglitazone lowered the risk for recurrent stroke or

myocardial infarction in insulin-resistant individuals without diabetes

who had a prior stroke or transient ischemic attack.

Sodium-Glucose Co-Transporter 2 (SGLT2) Inhibitors These

agents (Table 404-5) lower the blood glucose by selectively inhibiting

this co-transporter, which is expressed almost exclusively in the

proximal, convoluted tubule in the kidney. This inhibits glucose

reabsorption, lowers the renal threshold for glucose, and leads to

increased urinary glucose excretion. Thus, the glucose-lowering effect

is insulin independent and not related to changes in insulin sensitivity

or secretion. The loss of urinary glucose may promote modest weight

reduction. Since these agents also impair proximal reabsorption of

sodium, their use is associated with a diuretic effect and 3–6 mmHg

reduction in systolic blood pressure. Due to the increased urinary

glucose, urinary and genital mycotic infections are more common

in both men and women, and the diuretic effect can lead to reduced

intravascular volume and acutely impaired kidney function. Inhibition

of SGLT2 may lead to increased glucagon and consequently liver production of glucose and ketones. Euglycemic DKA may occur during

illness or when ongoing glucosuria masks stress-induced requirements

for insulin. These agents should not be prescribed for patients with

type 1 DM or pancreatogenic forms of DM associated with insulin

deficiency. Empagliflozin or canagliflozin reduces CVD events and

all cause cardiovascular mortality in patients with type 2 DM and

established CVD. All SGLT2 inhibitors may reduce hospitalization

for CHF. Empagliflozin, canagliflozin, and dapagliflozin have all been

shown to reduce progression of diabetic kidney disease but should not

be initiated in patients with stage 3b CKD (eGFR <45 mL/min per

1.73 m2

) and should not be used with stage 4 CKD (eGFR <30 mL/min

per 1.73 m2

). A possible increased risk of bladder cancer has been seen

with dapagliflozin.

OTHER THERAPIES FOR TYPE 2 DM • Bile Acid–Binding Resins

Evidence indicates that bile acids, by signaling through nuclear

receptors, may have a role in metabolism. Bile acid metabolism is

abnormal in type 2 DM. The bile acid–binding resin colesevelam has

been approved for the treatment of type 2 DM (already approved for

treatment of hypercholesterolemia). Because bile acid–binding resins

are minimally absorbed into the systemic circulation, how bile acid–

binding resins lower blood glucose is not known. The most common

side effects are GI (constipation, abdominal pain, and nausea). Bile

acid–binding resins can increase plasma triglycerides and should be

used cautiously in patients with a tendency for hypertriglyceridemia.

The role of this class of drugs in the treatment of type 2 DM is not yet

defined.

Bromocriptine A formulation of the dopamine receptor agonist bromocriptine (Cycloset) has been approved by the FDA for the treatment

of type 2 DM. However, its role in the treatment of type 2 DM is

uncertain.

INSULIN THERAPY IN TYPE 2 DM Insulin should be considered as part

of the initial therapy in type 2 DM, particularly in lean individuals

or those with severe weight loss, in individuals with underlying renal

or hepatic disease that precludes oral glucose-lowering agents, or in

individuals who are hospitalized or acutely ill. Insulin therapy is ultimately required by a substantial number of individuals with type 2 DM

because of the progressive nature of the disorder and the relative insulin deficiency that develops in patients with long-standing diabetes.

Both physician and patient reluctance often delay the initiation of insulin therapy, but glucose control and patient well-being are improved

by insulin therapy in patients who have not reached glycemic targets.

Because endogenous insulin secretion continues and is capable of

providing some coverage of mealtime caloric intake, insulin is usually

initiated in a single dose of long-acting insulin (0.1–0.4 U/kg per day),

given in the evening or just before bedtime (NPH, glargine, detemir,

or degludec). Because fasting hyperglycemia and increased hepatic

glucose production are prominent features of type 2 DM, bedtime

insulin is more effective in clinical trials than a single dose of morning

insulin. Glargine given at bedtime has less nocturnal hypoglycemia

than NPH insulin. Some physicians prefer a relatively low, fixed starting dose of long-acting insulin (5–15 units) or a weight-based dose (0.1

units/kg). The insulin dose may then be adjusted in 10–20% increments as dictated by SMBG results. Both morning and bedtime longacting insulin may be used in combination with oral glucose-lowering

agents. Initially, basal insulin may be sufficient, but often prandial

insulin coverage with multiple insulin injections is needed as diabetes

progresses (see insulin regimens used for type 1 DM). Other insulin

formulations that have a combination of short-acting and long-acting

3113 Diabetes Mellitus: Management and Therapies CHAPTER 404

insulin (Table 404-4) are sometimes used in patients with type 2 DM

because of convenience but do not allow independent adjustment of

short-acting and long-acting insulin dose and often do not achieve the

same degree of glycemic control as basal/bolus regimens. In selected

patients with insulin-deficient type 2 DM, insulin infusion devices may

be considered.

CHOICE OF INITIAL GLUCOSE-LOWERING AGENT The level of hyperglycemia and the patient’s individualized goal (see “Establishment of

Target Level of Glycemic Control”) should influence the initial choice

of therapy. Patients with mild hyperglycemia (FPG <7.0–11.0 mmol/L

[126–199 mg/dL]) often respond well to a single, oral glucose-lowering

agent, while those with moderate hyperglycemia (FPG 11.1–13.9 mmol/L

[200–250 mg/dL]) will usually require more than one oral agent or

insulin. Patients with more severe hyperglycemia (FPG >13.9 mmol/L

[250 mg/dL]) may respond partially but are unlikely to achieve normoglycemia with oral therapy. Insulin can be used as initial therapy

in individuals with severe hyperglycemia (FPG <13.9–16.7 mmol/L

[250–300 mg/dL]) or in those who are symptomatic from the hyperglycemia. This approach is based on the rationale that more rapid glycemic control will reduce “glucose toxicity” to the islet cells, improve

endogenous insulin secretion, and possibly allow oral glucoselowering agents to be more effective. If this occurs, the insulin may

be discontinued.

Insulin secretagogues, biguanides, α-glucosidase inhibitors, thiazolidinediones, GLP-1 receptor agonists, DPP-IV inhibitors, SLGT2

inhibitors, and insulin are approved for monotherapy of type 2 DM.

Although each class of oral glucose-lowering agents has advantages and

disadvantages (Table 404-5), certain generalizations apply: (1) insulin

secretagogues, biguanides, GLP-1 receptor agonists, and thiazolidinediones improve glycemic control to a similar degree (1–2% reduction in HbA1c) and are more effective than α-glucosidase inhibitors,

DPP-IV inhibitors, and SLGT2 inhibitors; (2) insulin secretagogues,

GLP-1 receptor agonists, DPP-IV inhibitors, α-glucosidase inhibitors,

and SLGT2 inhibitors begin to lower the plasma glucose immediately,

whereas the glucose-lowering effects of the biguanides and thiazolidinediones are delayed by weeks; (3) not all agents are effective in all

individuals with type 2 DM; (4) biguanides, α-glucosidase inhibitors,

GLP-1 receptor agonists, DPP-IV inhibitors, thiazolidinediones, and

SLGT2 inhibitors do not directly cause hypoglycemia; (5) most individuals will eventually require treatment with more than one class

of oral glucose-lowering agents or insulin, reflecting the progressive

nature of type 2 DM; and (6) durability of glycemic control is slightly

less for sulfonylureas compared to metformin or thiazolidinediones.

Considerable clinical experience exists with metformin and sulfonylureas because they have been available for several decades. It is

assumed that the α-glucosidase inhibitors, GLP-1 receptor agonists,

DPP-IV inhibitors, thiazolidinediones, and SLGT2 inhibitors will

reduce DM-related complications by improving glycemic control.

Pioglitazone may reduce CVD events through targeting a fundamental

abnormality in type 2 DM, namely insulin resistance. A reduction in

CVD events and in progression of diabetic kidney disease seen with

some GLP-1 agonists and SGLT2 inhibitors may also operate through

glucose-independent mechanisms (Chap. 405).

Treatment algorithms by several professional societies (ADA/

European Association for the Study of Diabetes [EASD], IDF, AACE)

suggest metformin as initial therapy because of its efficacy, known

side-effect profile, and low cost (Fig. 404-3). Initiation of pharmacologic therapy should be accompanied by an emphasis on lifestyle

modification (e.g., MNT, increased physical activity, and weight loss).

Metformin’s advantages are that it promotes mild weight loss, lowers

insulin levels, and improves the lipid profile slightly. Based on SMBG

results and the HbA1c, the dose of metformin should be increased until

the glycemic target is achieved or maximum dose is reached.

COMBINATION THERAPY WITH GLUCOSE-LOWERING AGENTS A

number of combinations of therapeutic agents are successful in type

2 DM: metformin + second oral agent, metformin + GLP-1 receptor

agonist, metformin + insulin, or combinations of a long-acting insulin

and a GLP-1 receptor agonist. Because mechanisms of action of the

first and second agents should be different, the effect on glycemic

control is usually additive. There are little data to support the choice

of one combination over another combination. Recent results from

the NIH-funded Glycemia Reduction Approaches in Diabetes: A

Comparative Effectiveness Study (GRADE) indicated that addition

of liraglutide or basal insulin to metformin leads to better glycemic

control than glimepiride or sitagliptin (SLGT2 inhibitors were not

studied). Based on recent demonstrations of a beneficial cardiovascular

effect in certain individuals with type 2 DM and CVD, or at high risk

of CVD, a GLP-1 receptor agonist or a SGLT2 inhibitor should now be

considered in these populations. Medication costs vary considerably

(Table 404-5), and this often factors into medication choice. Several

fixed-dose combinations of oral agents are available, but evidence that

they are superior to titration of a single agent to a maximum dose and

then addition of a second agent is lacking. If adequate control is not

achieved with the combination of two agents (based on reassessment of

the HbA1c every 3 months), a third oral agent, GLP-1 receptor agonist,

or basal insulin should be added (Fig. 404-3). Treatment approaches

vary considerably from country to country. For example, α-glucosidase

inhibitors are used commonly in South Asian patients (Indian), but

infrequently in the United States or Europe. Whether this reflects an

underlying difference in the disease or physician preference is not clear.

Treatment with insulin often becomes necessary as type 2 DM

enters the phase of relative insulin deficiency and is signaled by inadequate glycemic control with one or two oral glucose-lowering agents.

Insulin alone or in combination should be used in patients who fail

to reach glycemic targets. For example, a single dose of long-acting

insulin at bedtime is often effective in combination with metformin.

As endogenous insulin production falls further, multiple injections of

long-acting together with short-acting insulin are necessary to control

postprandial glucose excursions. These insulin regimens are identical

to the long-acting and short-acting combination regimens discussed

above for type 1 DM, although usually at higher doses given insulin

resistance. Weight gain and hypoglycemia are the major adverse effects

Insulin + metformin

Reassess HbA1c

Reassess HbA1c

Reassess HbA1c

Patient with type 2 diabetes

Individualized glycemic goal

Medical nutrition therapy,

increased physical activity,

weight loss

+

metformin

Combination therapy

metformin + second agent

Combination therapy

metformin + two

other agents

FIGURE 404-3 Glycemic management of type 2 diabetes. See text for discussion

of treatment of severe hyperglycemia or symptomatic hyperglycemia. Agents that

can be combined with metformin include insulin secretagogues, thiazolidinediones,

α-glucosidase inhibitors, DPP-IV inhibitors, GLP-1 receptor agonists, SLGT2

inhibitors, and insulin. HbA1c, hemoglobin HbA1c.

3114 PART 12 Endocrinology and Metabolism

of insulin therapy. The daily insulin dose required can become quite

large (1–2 units/kg per day) as endogenous insulin production falls and

insulin resistance persists. Individuals who require >1 unit/kg per day

of long-acting insulin should be considered for combination therapy

with metformin a GLP-1 receptor agonist, or a thiazolidinedione as

these can reduce insulin requirements in some individuals with type

2 DM. Insulin plus a thiazolidinedione promotes weight gain and may

be associated with peripheral edema. Addition of a thiazolidinedione

to a patient’s insulin regimen may necessitate a reduction in the insulin

dose to avoid hypoglycemia. Patients requiring large doses of insulin

(>200 units/day) can be treated with a more concentrated form of

insulin.

■ OTHER THERAPIES FOR DIABETES

Metabolic (also referred to as bariatric) surgery for obese individuals

with type 2 DM has shown considerable promise, sometimes with

dramatic resolution of the diabetes or major reductions in the needed

dose of glucose-lowering therapies (Chap. 402). Several large, nonrandomized clinical trials have demonstrated a much greater efficacy of

metabolic surgery compared to medical management in the treatment

of type 2 DM and may be considered in individuals with type 2 DM

and a BMI >35 kg/m2

. The ADA clinical guidelines state that metabolic

surgery should be considered in individuals with type 2 DM and a

body mass index >30 kg/m2

 if hyperglycemia is inadequately controlled

despite optimal medical therapy.

Short-term intense caloric restriction (very-low-calorie diet, typically 800–1000 calories/day) can dramatically improve type 2 DM,

sometimes leading to resolution of the diabetes. Such an approach is

more effective in recent-onset type 2 DM and should be supervised

by a provider with expertise and should be followed by a long-term,

weight-maintenance program.

Whole-pancreas transplantation can normalize glucose control in

type 1 DM and when performed simultaneously with or after kidney transplantation can prolong the life of the kidney transplant by

offering protection against recurrent diabetic nephropathy. Pancreatic

islet transplantation is available as a less invasive form of beta-cell

replacement therapy for type 1 DM, but it remains investigational in

the United States. Due to the risks associated with chronic immunosuppression, whole-pancreas and pancreatic islet transplantation may

be considered for patients with severe metabolic instability or already

requiring immunosuppression in support of a kidney or other organ

transplant. Patients with chronic pancreatitis and preserved islet function

who require pancreatectomy for pain relief may benefit from autologous

islet transplantation as this may prevent or ameliorate postsurgical DM.

■ EMERGING THERAPIES

Many individuals with long-standing type 1 DM still produce very

small amounts of insulin or have insulin-positive cells within the

pancreas. This suggests that beta cells may slowly regenerate but are

quickly destroyed by the autoimmune process. Particularly early in

the disease course, efforts to suppress the autoimmune process, for

example with anti-CD3 monoclonal antibodies that target T lymphocytes, are being tested at the time of diagnosis of type 1 DM, and for

prevention in autoantibody-positive individuals at stages 1 and 2 of

type 1 DM (Chap. 403). Agents that target thioredoxin-interacting

protein (TXNIP), especially Ca++ channel blockers, have some promise in recent-onset T1D and in rodent models of diabetes. Closed-loop

insulin infusion devices that automate insulin delivery in response to

changing glucose levels are progressing rapidly. New therapies under

evaluation or development for type 2 DM include activators of glucokinase, inhibitors of 11 β-hydroxysteroid dehydrogenase-1, GPR40

agonists, dual agonists targeting the glucose-dependent insulinotropic

polypeptide receptor and the GLP1-receptor, combined SLGT1 and

SLGT2 inhibitors, and agents that may reduce inflammation, for example by inhibiting IL-1β.

Because whole-pancreas and pancreatic islet transplantation are

both limited by organ availability from deceased donors, stem cell–

derived islet cells and xenogeneic sources of islets may eventually allow

for a limitless supply of insulin-producing cells for transplantation.

ADVERSE EFFECTS OF THERAPY FOR DM

As with any therapy, the benefits of efforts directed toward glycemic

control must be balanced against the risks of treatment (Table 404-5).

Side effects of intensive treatment include an increased frequency of

serious hypoglycemia, weight gain, increased economic costs, and

greater demands on the patient. In the DCCT, quality of life was very

similar in the intensive and standard therapy groups. The most serious

complication of therapy for DM is hypoglycemia, and its treatment

with oral glucose or glucagon injection is discussed in Chap. 406.

Severe, recurrent, or unexplained hypoglycemia warrants examination

of treatment regimen and glycemic goal for the individual patient.

Weight gain occurs with most (insulin, insulin secretagogues, thiazolidinediones) but not all (metformin, α-glucosidase inhibitors, GLP-1

receptor agonists, DPP-IV inhibitors) therapies. The weight gain is

partially due to the anabolic effects of insulin and the reduction in

glucosuria.

ACUTE DISORDERS RELATED TO

SEVERE HYPERGLYCEMIA

Individuals with type 1 or type 2 DM and severe hyperglycemia

(>13.9 mmol/L [250 mg/dL]) should be assessed for clinical stability,

including mentation and hydration. Depending on the patient and

the rapidity and duration of the severe hyperglycemia, an individual

may require more intense and rapid therapy to lower the blood glucose. However, many patients with poorly controlled diabetes and

hyperglycemia have few symptoms. The physician should assess if

the patient is stable or if DKA or a hyperglycemic hyperosmolar state

(HHS) should be considered. Ketones, an indicator of DKA, should be

measured in individuals with type 1 DM when the plasma glucose is

persistently >13.9 mmol/L (250 mg/dL), during a concurrent illness,

or with symptoms such as nausea, vomiting, or abdominal pain. Blood

measurement of β-hydroxybutyrate is preferred over urine testing with

nitroprusside-based assays that measure only acetoacetate and acetone.

DKA and HHS are acute, severe disorders directly related to diabetes. DKA was formerly considered a hallmark of type 1 DM, but it also

occurs in individuals with type 2 DM who can sometimes subsequently

be treated with oral glucose-lowering agents (frequently in individuals

of Hispanic or African-American descent). HHS is primarily seen in

individuals with type 2 DM. Both disorders are associated with absolute or relative insulin deficiency, volume depletion, and acid–base

abnormalities. DKA and HHS exist along a continuum of hyperglycemia, with or without ketosis. The metabolic similarities and differences

in DKA and HHS are highlighted in Table 404-6. Both disorders are

associated with potentially serious complications if not promptly diagnosed and carefully treated.

■ DIABETIC KETOACIDOSIS

Clinical Features The symptoms and physical signs of DKA are

listed in Table 404-7 and usually develop over 24 h. DKA may be the

initial symptom complex that leads to a diagnosis of type 1 DM, but

more frequently, it occurs in individuals with established diabetes.

Nausea and vomiting are often prominent, and their presence in an

individual with diabetes warrants laboratory evaluation for DKA.

Abdominal pain may be severe and can resemble acute pancreatitis or

ruptured viscus. Hyperglycemia leads to glucosuria, volume depletion,

and tachycardia. Hypotension can occur because of volume depletion

in combination with peripheral vasodilatation. Kussmaul respirations

and a fruity odor on the patient’s breath (secondary to metabolic acidosis and increased acetone) are classic signs of the disorder. Lethargy

and central nervous system depression may evolve into coma with

severe DKA but should also prompt evaluation for other reasons for

altered mental status (e.g., infection, hypoxemia). Cerebral edema, an

extremely serious complication of DKA, is seen most frequently in

children. Signs of infection, which may precipitate DKA, should be

sought on physical examination, even in the absence of fever. Failure

to augment insulin therapy during physiologic stress often compounds

the problem. Tissue ischemia (heart, brain) can also be a precipitating

factor. Omission of insulin because of an infusion pump delivery site