including an increase risk in bone fractures and HF.

258 TZDs are associated with

an approximately twofold increased risk of HF and thus are contraindicated in

NYHA Class III-IV HF.

185,187

In the long-term ADOPT trial, rosiglitazone was

associated with a mean weight gain of 4.8 kg, whereas weight decreased by 2.9 kg in

the metformin group; glyburide-treated patients had a mean weight gain of 1.6 kg

during the first year, and then remained stable.

257 The TZDs also have been found to

reduce bone density and increase fracture risk. Given the adverse effects associated

with TZDs, and the availability of other agents, their use has fallen out of favor.

Sitagliptin, saxagliptin, and exenatide, which are approved for monotherapy, are

often reserved for combination therapy. However, DPP-4 inhibitors may be useful as

monotherapy for patients in whom the use of other oral agents is precluded (such as

renal dysfunction with metformin or severe HF with pioglitazone). GLP-1RAs may

be particularly useful in obese patients.

7,39

It should be noted that AACE recommends

GLP-1RAs as first line if metformin can’t be initiated first.

39

Finally, one should not forget that until the newer agents became available,

sulfonylureas were used quite successfully to treat obese patients with Type 2

diabetes. Although these agents are relatively effective (average A1C reduction 1%–

1.5%), there are still several issues with this class of medications such as

unfavorable adverse effects (hypoglycemia and weight gain), durability issues, and

may worsen the lipid panel. Although the use of these agents has fallen out of favor

due to these effects, they are effective and often utilized when cost is an issue (Table

53-26).

Several products that combine two oral agents into one medication are available.

Although many of these products are approved as first-line therapy, thus skipping the

monotherapy step, consider reserving them for use in patients who are already

established on the two agents or whose medication regimens must be simplified to

enhance adherence or decrease cost or are already established on the two agents.

Based on this discussion, metformin is favored as initial therapy in L.H.

CASE 53-12

QUESTION 1: N.H. is a 46-year-old, obese (BMI 33 kg/m

2

) man with a history of Type 2 diabetes, and

hyperlipidemia. He presents with complaints of fatigue and nocturia. N.H. used to smoke two packs of

cigarettes/day for 15 years, but he quit after his diabetes diagnosis. He has a strong family history of CHD. At

diagnosis 3 months ago, an elevated A1C was documented (7.6%), confirming the diagnosis of Type 2 diabetes.

Tests of liver and renal function are within normal limits. Current medications include lisinopril and atorvastatin.

N.H. started metformin but stopped taking it because of GI symptoms (loose stools); he refuses to try this

medication again. What is a reasonable next monotherapy option for N.H.?

Unfortunately, N.H. did not tolerate metformin. Although this is less common if

patients take metformin with food and the dose is titrated slowly, some do not

overcome the GI adverse effects. N.H. is refusing to even consider a retrial. An A1C

goal of less than 7% may be appropriate for N.H., although given his young age and

long life expectancy, a target of less than 6.5% would be preferred. Therefore, the

A1C lowering required to get to goal is only 0.6% to 1.1%.

With a BMI of 33 kg/m2

, hypertension, and dyslipidemia, N.H. has many

components of the metabolic syndrome and insulin resistance (see Pathogenesis

section). Sulfonylureas or insulin would not be an optimal next choice of drug for

N.H. because they do not exert a favorable effect on plasma lipids and generally are

associated with weight gain and hypoglycemia. Also, his A1C is not high enough to

need insulin therapy yet. However, it must be acknowledged that some sulfonylureas

remain the least expensive oral antidiabetic agents, and this may be an important

factor in the initial selection of therapy for some patients. Acarbose as monotherapy

is not likely to attain near-normal plasma glucose concentrations. Given he is obese,

and likely insulin resistant, although pioglitazone improves peripheral target tissue

responses to insulin and lipid profiles, avoiding pioglitazone due to adverse effects

of weight gain is prudent (see Case 53-11, Question 8, and Case 53-15). A GLP-1

agonist, DPP-4 inhibitor, or SGLT-2 inhibitor are reasonable options for N.H.

because they are not associated with weight gain (GLP-1 agonists and SGLT-2

inhibitors may even promote weight loss).

For N.H., initiation of a GLP-1RA, such as liraglutide, is a reasonable option. A

GLP-1RA, along with lifestyle changes, should be able to get to his A1C goal. If

liraglutide is used, it should be initiated at a dose of 0.6 mg once daily for 1 week

and then increased to the therapeutic dose of 1.2 mg once daily thereafter. Some

patients may require further titration to 1.8 mg once daily in order to reach their

glycemic goals. N.H. should be counseled about the likelihood of GI side effects,

which should subside over time, as well as the rare risk of pancreatitis (and to stop

taking and notify his provider if he has persistent severe abdominal pain that may

radiate to the back and vomiting) and severe allergic reactions.

N.H. was successful in reducing his A1C to 6.4% through a combination of

lifestyle modifications and sitagliptin. He was able to lose 10 lb with walking and

reduction in fatty food.

N.H. quit smoking 3 months ago. Smoking has been shown to increase the risk of

IFG and Type 2 diabetes.

259,260 The mechanism for this is thought to be related to

increased insulin resistance and oxidative stress and reduced insulin secretion.

Smoking also increases the risk of microvascular and macrovascular disease in

people with diabetes.

261 Therefore, N.H. should be congratulated on his cessation,

and continued abstinence should be encouraged through follow-up visits (see Chapter

91, Tobacco Use and Dependence).

Failure of Antidiabetic Monotherapy

PATHOGENESIS

CASE 53-13

QUESTION 1: Q.R. is a 68-year-old, 5 feet 1 inch, 155-lb (BMI, 29.3 kg/m

2

) woman with an 8-year history

of Type 2 diabetes who has been treated with diet, exercise, and metformin. According to clinic records, she

was well controlled initially (A1C 6.7%–7.2%) for the first 5 years. When her glycemic control worsened

(FPG, 130–160 mg/dL; A1C, 7.5%–8.5%), her metformin dose was increased from an initial dose of 500 mg

BID to her current dose of 1,000 mg BID. Recent chart notes indicate that Q.R.’s chief complaints have

included loss of appetite and fatigue. Her current A1C is 8.4%. Other medical problems include hypertension

managed with hydrochlorothiazide 25 mg daily and mild peripheral neuropathy managed with acetaminophen

500 mg BID. Her eGFR is 70 mL/minute.

At this clinic visit, Q.R., who is well known to you, seems particularly listless and flat in her affect. Her BG

records, which are typically meticulous, are incomplete. AM fasting BG values consistently exceed 200 mg/dL

and range from 202 to 340 mg/dL. While taking her history, you discover that her husband passed away last

year and that one of her adult children has recently been diagnosed with a terminal illness. What factors may be

contributing to Q.R.’s poor glucose control?

p. 1135

p. 1136

Several factors may be contributing to Q.R.’s deteriorating BG control and

apparent lack of responsiveness to maximal dose of metformin during the past year

(as evidenced by her elevated BG and A1C, listlessness, and lack of appetite).

Monotherapy failure (also called secondary failure) is characterized by

progressively poor glucose control that occurs after a period of good response

(months to years). The cause of failure may be related to progressive pancreatic

failure; poor adherence to diet, exercise, or medications; and exogenous diabetogenic

factors such as increased weight, illness, or drugs that induce hyperglycemia (e.g.,

atypical antipsychotics or glucocorticoids).

Type 2 diabetes is a progressive condition, which is most likely to require

combination drug therapy. The UKPDS confirmed that monotherapy failure

represents a natural progression of Type 2 diabetes. In this study, only 16% and 19%

of the subjects achieved an FPG less than 108 mg/dL and A1C less than 7%,

respectively, after 3 years of dietary therapy alone.

262 By 9 years, only 9% were able

to maintain their glycemic goals using just diet therapy. The investigators also found

monotherapy failure occurred at the same rate regardless of the initial treatment

selected: glyburide, chlorpropamide, metformin, or insulin. In all the monotherapy

treatment groups, patients required additional therapies over the course of the study

duration.

262 At 3 years, fewer than 55% of patients randomly assigned to single

pharmacologic therapy could maintain an A1C less than 7%, and by 9 years this

dropped to about 25% of patients. In the ADOPT study, patients were able to

maintain A1C levels less than 7% using monotherapy for 60, 45, and 33 months with

rosiglitazone, metformin, and glyburide, respectively.

257 The reason for glycemic

deterioration is likely owing to the natural progression of Type 2 diabetes in which

β-cell function declines with increased duration of disease.

Q.R.’s deteriorating control on a therapeutic dose of metformin after 5 years of

reasonable response is consistent with the natural progression of Type 2 diabetes.

However, the stress and depression arising out of her life situation have no doubt

contributed to her poor control. The latter may have led to a change in her usual

adherence to diet, exercise, and medications, and should resolve with time and

appropriate management. Women with Type 2 diabetes are at increased risk of

depression. Evidence exists for a bidirectional association between diabetes and

depression in women.

263 Although hyperglycemia has been attributed to

hydrochlorothiazide, the dose prescribed for Q.R. has few adverse metabolic effects.

MANAGING MONOTHERAPY FAILURE

CASE 53-13, QUESTION 2: How should Q.R. be managed? What is an appropriate antidiabetic agent to

add?

Q.R. is exhibiting symptoms of depression (e.g., listlessness and flat affect). Her

depression likely started after her husband’s death. Every effort must be made to

address Q.R.’s depression because it is unlikely that she will be able to effectively

implement more aggressive treatment of her diabetes until her situation is improved.

Resources that may be used include her family, a therapist, and a social worker.

Treatment of monotherapy failure includes identifying and correcting any

diabetogenic factors and altering her drug therapy. When failure to any oral agent

occurs, one should always add another agent rather than switch to another, unless

adverse effects are intolerable or contraindications warrant the discontinuation of the

drug. This is supported by a study that evaluated the effect of metformin alone in a

population of patients who had failed oral sulfonylureas and the effect of metformin

plus the sulfonylureas. Substitution of metformin for glyburide did not produce any

significant change in glycemic control, but the addition of metformin to glyburide

therapy substantially improved glucose concentrations.

264 Many combinations of

antidiabetic agents can be used. The key is that they should have different

mechanisms of action. For example, it is not reasonable to combine a sulfonylurea

with a glinide (i.e., repaglinide and nateglinide) because they are both insulin

secretagogues. Options for Q.R. include adding one of the following to metformin:

insulin secretagogue (sulfonylurea or glinide), acarbose, TZD (pioglitazone), DPP-4

inhibitor (sitagliptin, linagliptin, alogliptin or saxagliptin), SGLT-2 inhibitors

(canagliflozin, dapagliflozin, empagliflozin), or a GLP-1 agonist (exenatide,

exentatide ER, albiglutide, dulaglutide, or liraglutide). SGLT-2 should be used

cautiously in this patient because this drug class is associated with genital mycotic

infections and urinary tract infections.

228

Table 53-25 summarizes the FDA-approved combination therapy indications. One

also could introduce insulin therapy at this time. However, because Q.R.’s A1C

should be effectively lowered to less than 7% with an additional noninsulin agent,

two noninsulin antidiabetic agents should be utilized.

In summary, patients such as Q.R. who are unresponsive to the maximally effective

dose of metformin should be initiated on combination therapy.

Combination Antidiabetic Therapy

CASE 53-13, QUESTION 3: As anticipated, Q.R. refuses to consider insulin therapy at this time. Which

combination of antidiabetic agents is preferred?

When agents from different antidiabetic classes are combined, their effects are

essentially additive. With the availability of many antidiabetic agents, there is no one

best combination therapy. As discussed, the choice of therapy should take into

account the patient’s organ function, amount of A1C lowering required to reach an

individual’s goal, possible adverse effects of a particular drug or drug combination,

cost, and patient preference.

A mixed-treatment comparison meta-analysis evaluated the efficacy of second-line

antidiabetic therapy added to patients on stable doses of metformin.

265 The risk of

weight gain and hypoglycemia was also determined. A total of 27 randomized

controlled trials were included in the analysis, with a mean study duration of 32

weeks. Table 53-27 summarizes these findings. In addition, the change in A1C

depending on the baseline level of A1C was evaluated and found, as previously

discussed (see Case 53-11, Question 7), to be greater when the baseline A1C was

8% or more.

Q.R. is concerned with the costs of her medications, and brand drugs have a higher

copay in her insurance plan. Q.R. would like to remain on oral medications if

possible. Because she is postmenopausal, avoid a TZD owing to the concerns of

increased risk of fractures. Given the significant GI side effects associated with the

α-glucosidase inhibitors, their use is often avoided all together. The DPP-4 inhibitors

are available by brand only, so initiating a generic is an option. Therefore, she is

started on glimepiride 2 mg PO daily.

Combination Antidiabetic Therapy

CASE 53-13, QUESTION 4: Q.R. is titrated to glimepiride 4 mg/day and continued on metformin 1,000 mg

BID. This improved her FPG and A1C modestly for approximately 12 months (FPG, 120–150 mg/dL; currently

A1C, 7.6%). She remains resistant to starting insulin, despite repeated counseling. She has heard about a new,

injectable class of diabetes medications that cause weight loss, and asks whether she is a good candidate to

start one. It quickly becomes apparent that Q.R. is referring to the GLP-1 agonists, exenatide, exenatide ER,

dulaglutide, albiglutide, and liraglutide; only exenatide is on her health plan’s formulary.

p. 1136

p. 1137

Table 53-27

Effects of Adding a Noninsulin Antidiabetic Agent in Patients with Type 2

Diabetes Already on Metformin

265

Effects of Adding a Noninsulin Agent Compared with Placebo in Patients Taking

Metformin on Change in A1C, Body Weight, and Hypoglycemia

Group vs.

Placebo

% Change in A1C

Change in Body Weight

(kg) Overall Hypoglycemia

No. of

Trials WMD (95% CI)

No. of

Trials WMD (95% CI)

No. of

Trials RR (95% CI)

All drugs 20 −0.79 (−0.90 to

−0.68)

12 −0.14 (−1.37 to

1.65)

19 1.43 (0.89 to 2.30)

SFUs 3 −0.79 (−1.15 to 2 −1.99 (0.86 to 3.12) 3 2.63 (0.76 to 9.13)

−0.43)

Glinides 2 −0.71 (−1.24 to

−0.18)

2 −0.91 (0.35 to 1.46) 2 7.92 (1.45 to

43.21)

TZDs 3 −1.00 (−1.62 to

−0.38)

1 −2.30 (1.70 to 2.90) 2 2.04 (0.50 to 8.23)

AGIs 2 −0.68 (−1.11 to

−0.19)

1 −1.80 (−2.83 to

−0.077)

2 0.60 (0.08 to 4.55)

DPP-4

inhibitors

8 −0.79 (−0.94 to

−0.63)

4 −0.09 (−0.47 to

0.30)

8 0.67 (0.30 to 1.50)

GLP-1

agonists

2 −0.99 (−1.19 to

−0.78)

2 −1.76 (−2.90 to

0.62)

2 0.94 (0.42 to 2.12)

Effects of Adding a Noninsulin Agent Compared with Placebo in Patients

Taking Metformin on Change in A1C, Depending on Baseline A1C

Group vs. Placebo

Baseline A1C

<8% WMD (95% CI) ≥8% WMD (95% CI)

SFUs −0.57 (−0.75 to −0.39) −0.97 (−1.35 to −0.62)

Glinides −0.44 (−0.85 to −0.04) −0.65 (−1.10 to −0.26)

TZDs −0.62 (−0.88 to −0.39) −1.02 (−1.39 to −0.69)

AGIs NR −0.65 (−1.07 to −0.24)

DPP-4 inhibitors −0.51 (−0.69 to −0.34) −0.89 (−1.11 to −0.68)

GLP-1 agonists NR −0.99 (−1.36 to −0.63)

Results are from a subgroup mixed-treatment comparison meta-analysis.

AGIs, α-glucosidase inhibitors; CI, confidence interval; DPP-4, dipeptidyl peptidase-4; GLP-1, glucagon-like

peptide-1; NR, not reported; RR, relative risk; SFUs, sulfonylureas; TZDs, thiazolidinediones; WMD, weighted

mean difference.

When the A1C is above goal despite the use of a combination of two agents,

practitioners often add a third agent before considering insulin. Although this is

tempting, depending on a patient’s current level of glycemic control, this practice

only delays insulin therapy, which is likely to be required to achieve A1C goals.

However, because Q.R. is close to an A1C level of less than 7%, it is reasonable to

try a third, noninsulin antidiabetic agent. Although glimepiride 8 mg/day is the

maximal dose, there is little difference in clinical efficacy compared with 4 mg/day.

Therefore, increasing glimepiride to 8 mg/day is not likely to achieve her glycemic

goals.

266

Exenatide is approved for use in patients as monotherapy and also for those taking

metformin, sulfonylurea, or a TZD alone, or a combination of metformin plus a TZD

or metformin plus a sulfonylurea. When added to patients on a sulfonylurea and

metformin, exenatide at 10 mcg BID resulted in a 0.8% A1C reduction compared

with baseline.

258 Thus, for Q.R., the addition of exenatide could result in an A1C of

less than 7%. The use of exenatide with a sulfonylurea is associated with an

increased risk of mild-to-moderate hypoglycemia (28% when used with sulfonylurea

and metformin and 36% when used with sulfonylurea alone).

258,267 Most practitioners

reduce the dose of the sulfonylurea on initiation of exenatide and then make

adjustments based on the patient’s response to exenatide.

Q.R. should be started on exenatide 5 mcg SC BID, taken within 60 minutes of the

two main meals of the day and at least 6 hours apart. She should be counseled about

nausea, which is the most frequent side effect; 44% of patients will experience

nausea, but there is only a 3% dropout rate in the clinical trials. The glimepiride

should be reduced to 2 (or 3) mg/day to avoid hypoglycemia. After 1 month, if she is

tolerating exenatide, the dose should be increased to 10 mcg SC BID. Her A1C

should be monitored within 3 months of using the 10-mcg dose. An advantage of

exenatide (versus DPP-4 inhibitors) is the potential for weight loss. In a 30-week

blinded study of exenatide added onto sulfonylurea and metformin, patients on the 10-

mcg dose had an average weight loss of 1.6 kg.

258

In the three open-label,

uncontrolled extension trials with exenatide, at 3 years, exenatide progressively

reduced (−5.3 kg) and sustained weight loss.

268

Q.R. should be counseled about the rare risk of pancreatitis. Also, if she finds that

nausea is significant enough to reduce her fluid intake, she should contact her

provider. The FDA has received case reports of altered renal function and renal

failure with exenatide, which are likely owing to dehydration from reduced fluid

intake because of the GI side effects (i.e., nausea, vomiting, and diarrhea) caused by

exenatide.

214

In patients with significant GI side effects, reduced fluid intake, or

preexisting renal dysfunction, the SCr should be monitored more closely. Exenatide

should be used cautiously with moderate renal impairment and is not recommended

in patients with ClCr of less than 30.

214

A small proportion of patients will form antiexenatide antibodies. At high titers,

these antibodies could result in failure to achieve adequate improvement in glycemic

control. If there is worsening glycemic control or failure to achieve targeted glycemic

p. 1137

p. 1138

control while on exenatide, the formation of blocking antibodies should be

considered as a reason. If this occurs, it may be reasonable to switch over to another

GLP-1 analog or exenatide ER. It is important to note that any of the available GLP-1

RAs are options for Q.R. for initial GLP-1 RA therapy. Discussions should be

regarding adherence, efficacy, side effects, and cost.

Combination Antidiabetic with Insulin Therapy

CASE 53-13, QUESTION 5: If Q.R. were willing to start insulin, why would it be reasonable to use insulin in

combination with other antidiabetic agents? How should it be combined with other agents? Is this combination

more effective than insulin alone?

Most patients with Type 2 diabetes eventually require insulin. The combined use

of insulin with a variety of oral agents has been extensively evaluated, but the studies

differ in their interventions. In some studies, single doses of an intermediate- or longacting insulin or once- or twice-daily doses of premixed insulin are added to a single

or combination of oral antidiabetic agents in poorly controlled patients. The primary

outcomes that have been evaluated include measures of glycemic control (e.g., A1C,

FPG) and the extent to which insulin doses have been decreased. DeWitt and Hirsch

published a comprehensive review of these studies, and readers are referred to this

article for this evidence.

93 The combined use of insulin and oral agents can be

considered based on the ACCE algorithm, particularly when the A1C is higher than

approximately 9%.

Before starting insulin, it is important to review the BG profile:

Fasting hyperglycemia, which may improve or persist throughout the day. This is the

more typical glucose profile for patients with Type 2 diabetes, in whom addition

of basal insulin at nighttime would be the most appropriate next step.

Fasting BG is at target, with daytime hyperglycemia. This is less common for

people with Type 2 diabetes, in whom addition of prandial insulin would be an

appropriate next step for therapy (see Case 53-13, Question 6).

In Q.R.’s case, it makes the most sense to add a single dose of basal insulin to

metformin and glimepiride, with the potential to discontinue glimepiride once the

insulin dose is optimized. Advantages attributed to adding insulin to an oral agent as

the next step after failure, as opposed to using insulin alone, include the following

93

:

Lower-insulin dosages can be used, and this minimizes weight gain and

hypoglycemia.

Simpler, single-dose insulin regimens are possible (vs. monotherapy with insulin).

Lowering the fasting glucose concentrations improves glucose control throughout the

day because glucose excursions related to meals are layered over lower values.

Furthermore, lower glucose values improve β-cell responsiveness to glucose and

enhance tissue responsiveness to insulin action.

However, lowering glucose concentrations is the first priority, not attempting to

use lower-insulin doses. Q.R. should be started on 10 units or 0.2 unit/kg of NPH, or

basal insulin glargine, detemir or degludec. This dose is based on empiric use of

insulin in a variety of studies (0.1–0.2 units/kg)

269 and a conservative estimate for

basal insulin of approximately 0.5 units/hour (Q.R. weighs 155 lb, or 70.5 kg). The

dose also takes into account the possibility that Q.R. is secreting some basal insulin

of her own and will have some residual stimulation of insulin secretion by glipizide.

The basal dose should be titrated upward based on the FPG for three consecutive

days. A common titration method used is the “treat-to-target” schedule.

92,270

FPG (mg/dL) Adjustment of Basal Insulin Dose (Units)

≥180 8

160–180 6

140–159 4

120–139 2

100–119 1

80–99 Maintain dose

60–79 −2

<60 −4 or more

Alternatively, the basal insulin dose can be increased by 2 units every 3 days until

the FPG is in a target range (80–130 mg/dL); if the FPG is greater than 140 to 180

mg/dL, larger increments can be used (e.g., by 4 units every 3 days).

269

If

hypoglycemia occurs or the FPG is less than 70 mg/dL, the dose should be reduced

by at least 4 units, or by 10% if the dose is more than 60 units. Of note, U300

glargine and U100 and U200 degludec should not be titrated sooner than every 3 to 4

days. If there is no improvement in glycemic control after 3 months, prandial insulin

or a GLP-1RA should be added (see Case 53-13, Questions 6 and 7). At that point,

the insulin secretagogue (in Q.R.’s case, glimepiride) is usually discontinued, but

metformin should be maintained.

An alternative for Q.R. is to discontinue the sulfonylurea and begin insulin

monotherapy using methods similar to those described for Type 1 diabetes patients.

This option also is rational based on the observation that patients like Q.R. are likely

to require insulin therapy because of progressive β-cell failure. Furthermore, insulin

monotherapy may be less expensive and easier to assess than combination oral agents

plus insulin therapy. Nevertheless, many clinicians use single doses of basal insulin

in combination with oral agents as a bridge to eventual insulin monotherapy,

especially for those patients unwilling to adhere to multiple daily insulin injections.

In the setting of insulin resistance, such as with Type 2 diabetes, continuing

metformin therapy should be a priority if at all possible.

Insulin Monotherapy in Patients with Type 2 Diabetes

INSULIN REGIMENS

CASE 53-13, QUESTION 6: Q.R.’s insulin glargine dose was eventually titrated to 25 units at bedtime. In

combination with glimepiride 4 mg/day and metformin 1,000 mg BID, her fasting glucose levels fell to the 110s

and 120s on most occasions; her A1C dropped to as low as 6.9%. However, after 1 year, she began to note a

gradual rise in glucose concentrations throughout the day. This resulted in a further increase in her bedtime

insulin glargine to 40 units (0.57 units/kg). Currently, her morning BG levels are 120 to 140 mg/dL, and BG

levels before or after meals range between 170 and 200 mg/dL. A recent A1C value was 8.5%. For the past 6

months or so, Q.R. has noted increasing fatigue, bouts of blurred vision, and recurrence of her monilial

infections. How should she be managed now?

The next step in managing Q.R.’s diabetes is to institute prandial insulin therapy,

which is needed as indicated by her daytime hyperglycemia. Like Type 1 diabetes

patients, those with long-standing Type 2 disease may require prandial insulin before

meals to minimize postprandial excursions. Insulin lispro has been

p. 1138

p. 1139

shown to decrease postprandial glucose concentrations to a greater extent than

regular insulin (30% lower at 1 hour and 53% lower at 2 hours) and was associated

with a lower rate of hypoglycemia, particularly between midnight and 6 AM (36%).

However, A1C levels were not significantly different after 6 months.

271 A similar

response with insulin aspart and glulisine would be expected.

Because Q.R. is on insulin glargine, the most appropriate next step is to add

prandial insulin using a rapid-acting insulin. Initiation of basal-bolus insulin

regimens are discussed in detail in the Type 1 Diabetes Insulin Therapy section (see

Case 53-2, Questions 4–6). An easy way to introduce prandial insulin is to initially

add only one mealtime dose. The premeal BG levels that are associated with the

most hyperglycemia are targeted. For example, if the prelunch BG levels are

elevated, a prandial dose at breakfast is added; or if the bedtime BG levels are

mainly elevated (owing to large dinners), a prandial dose at dinner is added. Once a

patient adjusts to this, prandial insulin is added to the other meals; at that point

glimepiride dose should be reduced or discontinued, altogether. Patients with Type 2

diabetes may require large insulin TDDs (>1 unit/kg) to reach A1C targets less than

7%. Although not technically insulin monotherapy, metformin is often continued to

assist in reducing insulin resistance and minimizing weight gain with insulin.

PREMIXED INSULINS

CASE 53-13, QUESTION 7: Q.R. has difficulty adhering to a basal-bolus insulin regimen. She is currently

taking insulin glargine 38 units at nighttime and approximately 7 units of insulin aspart before each meal (she

follows a high-sugar correction scale). What options are available for Q.R.?

Because people with Type 2 diabetes retain some pancreatic function, it often is

possible to achieve an acceptable level of control with twice-daily doses of

intermediate-acting insulin in combination with rapid- or short-acting insulins, which

are available as premixed insulins. These are referred to as split-mixed insulin

regimens. Although convenient, they do limit flexibility in dosing, and thus A1C

lowering ability without increasing hypoglycemia risk. In the United States, fixed

mixtures of NPH and regular insulin in a 70:30 ratio and a 50:50 ratio are available

(Table 53-6). Commercial combinations of the rapid-acting insulins plus an

intermediate-acting insulin also are available as Humalog Mix 75/25 (a mixture of

insulin lispro plus lispro protamine suspension), Humalog Mix 50/50 (a mixture of

insulin lispro plus lispro protamine suspension), and NovoLog Mix 70/30 (a mixture

of insulin aspart plus aspart protamine suspension). These fixed mixtures are

available in prefilled syringes, which can add to their flexibility and convenience for

administration.

The Treating to Target in Type 2 Diabetes (4-T) study assessed the efficacy of a

basal insulin (detemir), prandial insulin (insulin aspart), and biphasic insulin

(NovoLog Mix 70/30) regimen added to patients on metformin and a sulfonylurea.

272

After 3 years, a biphasic-based insulin regimen was found to be less effective in

achieving an A1C goal of less than 6.5% compared with prandial and basal insulin

regimens (31.9%, 44.7%, and 43.2%, respectively). Despite the findings of this large

multicenter study, use of twice-daily, premixed insulin remains a common insulin

therapy for patients with Type 2 diabetes.

Q.R. is currently receiving a TDD of insulin of 59 units, or 0.84 units/kg/day. To

convert her to a premixed insulin, one could begin with a conservative TDD of 0.5 to

0.6 units/kg/day, split equally before breakfast and dinner. An older method of twothirds in the morning and one-third in the evening is used less commonly now. When

starting an insulin-naïve Type 2 diabetes patient on premixed insulin, doses of 5 to 6

units twice daily (administered before breakfast and dinner) are often used, with the

doses being titrated before breakfast (to affect prelunch and predinner glucose

levels) and before dinner (to affect the bedtime and fasting glucose levels).

92,273

If the premixed insulin does not achieve adequate glycemic control, an option is to

mix the short- or rapid-acting insulin with NPH in the same syringe. By doing this,

each insulin dose can be individually adjusted. A disadvantage to this is the chance

for patients to make errors in measuring and mixing insulin, especially in the elderly

when they may have vision and/or dexterity issues. Alternatively, a prandial dose of

rapid-acting insulin (i.e., a third injection) can be added at lunch; the breakfast dose

of premixed insulin should be decreased as well.

93 Another option is to add a third

dose of the premixed insulin at lunch (so premix insulin TID); however the lunchtime

dose is smaller than the breakfast and dinner dose. An initial premixed insulin

lunchtime dose of 2 to 6 units, or 10% of the current TDD of premixed insulin, can be

started.

274

Q.R. should be started on a rapid-acting premixed insulin twice daily, as this has

the convenience of being administered right before eating (within 15 minutes). A

dose of 20 units SC BID represents a conservative starting dose. She should monitor

her fasting and predinner BG levels, at a minimum, to further adjust her insulin doses.

Adding Antidiabetic Agents to a Basal Insulin TherapyBased Regimen

CASE 53-14

QUESTION 1: M.A., a 62-year-old woman, has had Type 2 diabetes for 11 years. She is currently taking

metformin 500 mg PO TID and insulin glargine 47 units at bedtime. Her A1C is 8.2%. She tries to follow a

meal plan that a dietitian developed for her, but her BMI remains 31 kg/m

2

. Her physical activity is limited

because of an arthritic knee, for which she plans to have knee replacement surgery in the future. Other medical

problems include hypertension (hydrochlorothiazide 25 mg daily and benazepril 40 mg daily) and dyslipidemia

(atorvastatin 40 mg daily), which are both well controlled. Can any antidiabetic agents be added to her current

therapy?

The TZDs have been well studied in patients with Type 2 diabetes who are

already taking insulin. In a meta-analysis, pioglitazone lowered A1C by 0.58% when

added to insulin therapy; unfortunately, it was associated with weight gain (3 kg) and

increased peripheral edema and therefore should be avoided.

206

A GLP-1 agonist would be a logical agent to considering adding in obese patients.

Any GLP-1RA would be a good option. The agents should be chosen based on cost,

effectiveness, and adherence. For example, exenatide may be the only GLP-1RA on

formulary, or perhaps M.A. has trouble remembering to take once-weekly

medications, such as dulaglutide, and as such would be more successful with oncedaily liraglutide. Adding a DPP-4 inhibitor would be another option. Sitagliptin is

FDA-approved for use with insulin; however, the A1C lowering may be less than

with a GLP-1 agonist.

Therefore, addition of liraglutide may be a reasonable approach in M.A. Her A1C

can reach her target goal of less than 7%, with potential for weight loss and lowerinsulin glargine requirements. Liraglutide should be initiated at a dose of 0.6 mg

injected SC once daily for 1 week, after which the dose should be increased to 1.2

mg once daily. Some patients may require further titration to a maximum dose of 1.8

mg once daily to reach their glycemic targets.

p. 1139

p. 1140

Use of Antidiabetic Agents in Special Situations

HYPOGLYCEMIA

CASE 53-15

QUESTION 1: C.A. is a 73-year-old woman who has had a 20-year history of Type 2 diabetes and a 5-year

history of mild renal dysfunction (SCr, 1.2 mg/dL; eGFR, 47 mL/minute/1.73 m

2

; BUN, 22 mg/dL). Her son,

who lives with her, quickly called 9-1-1 when he noticed his mother appearing lethargic and sleepy, with her

eyes closed, as she was sitting on the couch. He assumed she was having a “low sugar reaction.” When the

paramedics arrived, her BG concentration was 46 mg/dL. C.A. was able to be aroused, and indicated she could

drink something, so 4 ounces of orange juice was administered. After 10 minutes, her BG level was 80 mg/dL,

so she was given another 4 ounces of juice. According to her son, C.A.’s diabetes has been well controlled for

the past several months with glyburide 10 mg BID and metformin 850 mg TID. For the past 3 days, C.A. has

been eating less and vomiting on association, with the “flu.” What was the likely cause of her hypoglycemic

episode? Were there any predisposing factors?

C.A. has experienced a hypoglycemic episode secondary to glyburide.

Hypoglycemia is the most common and potentially severe adverse effect of the

sulfonylureas. The incidence and severity of this effect increase with the duration of

action and potency of the agents.

Most sulfonylurea-induced hypoglycemia occurs in patients who are predisposed

to hypoglycemia in some way, and C.A. is no exception. She is an elderly woman

with renal impairment who was on relatively high doses of an agent, a portion of

which is excreted unchanged in the urine. Even in the face of decreased carbohydrate

intake (reduced appetite and vomiting), she continued to take her usual dose of

glyburide. Even though the stress of illness most often raises glucose levels, the

decreased food intake resulted in glyburide causing hypoglycemia. Because

glyburide has a long duration of action, hypoglycemia may last for several hours.

C.A. and her son should be educated about treatment of hypoglycemia. It is likely

that her son did not need to call 9-1-1, and he could have treated the hypoglycemia.

Glucagon should not be used to treat hypoglycemia caused by a sulfonylurea, as

glucagon can cause a paradoxical fall in the glucose levels.

275

RENAL DYSFUNCTION

CASE 53-15, QUESTION 2: C.A. has mild renal insufficiency (eGFR, 47 mL/minute/1.73 m

2

) and is taking

the maximal dose of metformin. What is the risk of lactic acidosis with metformin? How should age and kidney

function be taken into account with metformin use? Which agents should be avoided? Which agents could be

used?

Sulfonylurea compounds that are metabolized to active products that depend on the

kidney for elimination (e.g., acetohexamide, chlorpropamide glyburide, and

tolazamide) should be avoided in the elderly and in patients with decreased renal

function. Sulfonylureas that are completely metabolized to inactive or weakly active

products may be used (i.e., glipizide, glimepiride, or tolbutamide). Although

glyburide is unlikely to accumulate in patients with a ClCr greater than 30 mL/minute,

it should not be used in C.A because it caused a severe hypoglycemic reaction.

276

C.A. should be instructed to eat regularly because skipped meals may result in

recurrent hypoglycemia. Consideration to discontinuing the sulfonylurea and changing

adding an agent with a lower incidence of hypoglycemia should be considered.

The most notorious side effect associated with metformin—although extremely

rare—is lactic acidosis. The risk of lactic acidosis is increased with renal

insufficiency, which can result in accumulation of metformin because it is almost

exclusively clear unchanged by the kidneys. Lactic acidosis is a metabolic acidosis

characterized by a significant reduction in the arterial pH and an accumulation of

serum lactate, a product of anaerobic metabolism. It is a condition that is highly

lethal (50% mortality) and resistant to therapy. Lactic acidosis occurs when there is

an increased production of or decreased utilization of lactate. Decreased utilization

of lactate occurs when tissues are unable to oxidize lactate to pyruvate (these two

substances are normally present in the serum in a ratio of 10:1). Metformin might

predispose a patient to lactic acidosis by augmenting anaerobic metabolism or by

decreasing the kidney’s ability to handle an acid load. Other factors that might

contribute to lactic acidosis include severe cardiac or pulmonary disease (anoxia,

increased lactate production), septic shock, renal dysfunction (retention of metformin

and lactate), patients receiving contrast dye, and excessive alcohol intake (increased

lactate production and decreased utilization).

165–167

Signs and symptoms generally are acute in onset and commonly include nausea,

vomiting, diarrhea, and hyperventilation. Hypovolemia, hypotension, confusion, and

coma also may occur; death is usually secondary to cardiovascular collapse. Typical

laboratory findings include a low serum bicarbonate and PCO2

, a low arterial pH, an

elevated potassium, a normal or low serum chloride, elevated lactate and pyruvate

levels, an increased lactate to pyruvate ratio, and an anion gap of 30 mEq/L or

higher.

Although metformin rarely is associated with lactic acidosis, the manufacturer and

the FDA have taken extreme measures to prevent its improper use because another

biguanide, phenformin, which induced this life-threatening condition, was removed

from the market in 1977.

277 The estimated rate of phenformin-induced lactic acidosis

was 0.25 to 4 cases per 1,000 users versus 5 to 9 cases per 100,000 users for

metformin.

165–167 A group of clinicians from the FDA summarized 47 confirmed cases

of metformin-related lactic acidosis (lactate levels ≥5 mmol/L) that had been

reported to the FDA between May 1995 and June 1996.

261 Unfortunately, the

condition continues to be resistant to treatment; the mortality rate was 43%.

Importantly, 43 of the 47 cases (91%) had concurrent conditions that predisposed

them to lactic acidosis. These included cardiac disease (64%), decreased renal

function (28%), and chronic pulmonary disease (6%). Several patients (17%) were

older than 80 years of age and may have had decreased renal function despite normal

SCr concentrations. Interestingly, 38% of the patients had HF, and those who died

were more likely to be under treatment with digoxin and furosemide. The mean daily

dose of metformin was well within the therapeutic range and was not higher in the

group that succumbed (1,259 ± 648 mg in the group that died and 1,349 ± 598 mg in

the group that survived).

Metformin should be initiated with care in patients older than 80 years of age

because of the potential for a low GFR, even when SCr is normal.

164 Because C.A.

has moderate renal dysfunction (GFR <60 mL/minute, but >40 mL/minute), lower the

dose of metformin to 500 mg BID to minimize the potential for accumulation.

The TZDs are primarily metabolized by the liver and are not contraindicated in

patients with mild renal failure. The use of pioglitazone, beginning with low doses,

could be considered, as can acarbose, which is poorly absorbed from the GI tract.

The DPP-4 inhibitors could also be used, but their doses may require adjustment

depending on the degree of renal dysfunction (except for linagliptin). Exenatide can

be used in patients with ClCr greater than 30 mL/minute, and the dose does not need

to be adjusted.

p. 1140

p. 1141

Liraglutide should be used with caution in renal insufficiency, but does not require

dose adjustment. None of these agents cause hypoglycemia when used as

monotherapy.

HEPATIC DYSFUNCTION

CASE 53-16

QUESTION 1: B.R., a 60-year-old man with cirrhosis of the liver, is found to have Type 2 diabetes. Glipizide

10 mg/day is initiated. How will B.R.’s liver function affect the disposition of glipizide and his response to this

agent?

Most noninsulin antidiabetic agents should be avoided in patients with severe liver

disease, and insulin therapy is often the safest option. Because hepatic metabolism is

the primary route of elimination for most sulfonylureas, including glipizide, patients

with hepatic disease should be expected to have an exaggerated response to those

drugs metabolized to less active products. Liver disease can be a separate

predisposing factor for severe, prolonged hypoglycemia because glycogenolysis and

gluconeogenesis are impaired; thus, sulfonylureas are relatively contraindicated for

cirrhotic patients. If they are used, shorter-acting agents are preferred, and small

initial doses should be used. For B.R., glipizide could be initiated at a dose no

greater than 2.5 mg/day and increased if needed by 2.5-mg increments at no less than

weekly intervals. Other options are low doses of repaglinide (0.5 mg) or nateglinide

(60 mg) with meals because they are very short acting. Other good options for which

severe liver disease is not a specific contraindication include GLP-1RAs, DPP-4

inhibitors, and SGLT2 inhibitors, although there is little data regarding their use in

these patients. Basal-bolus or mixed insulin regimens are also good choices.

Diabetes in the Elderly

CLINICAL PRESENTATION

CASE 53-17

QUESTION 1: J.M. is a frail, 82-year-old, unresponsive man who is brought to the emergency department.

According to J.M.’s family, he has become increasingly confused, dizzy, and lethargic, with a recent weight loss

of 10 lb. J.M. lives by himself and has been generally healthy with the exception of mild-to-moderate chronic

obstructive pulmonary disease and arthritis. Fasting serum chemistry reveals the following:

Na, 128 mEq/L

Glucose, 798 mg/dL

Serum osmolality, 374 mOsm/L (normal, 280–295 mOsm/kg H2O)

Blood pH, 7.5

HCO3

, 22

His serum is negative for ketones. On physical examination, J.M. has poor skin turgor and dry mucous

membranes, and is responsive only to deep pain. His BP is 90/60 mm Hg with a pulse of 96 beats/minute. He

has rales at the left lower base of his lung, and a chest radiograph confirms pneumonia. Despite aggressive fluid

replacement, J.M.’s BG remains consistently greater than 250 mg/dL and his A1C is 11%. J.M. presents with

very high glucose concentrations, but has no history of diabetes mellitus. What special factors contribute to a

late and atypical presentation of diabetes in the elderly?

Diabetes in the elderly commonly is underdiagnosed and undertreated because it

often presents atypically.

278,279 Classic symptoms associated with diabetes mellitus

may be masked by other illnesses, entirely absent, or explained away by the normal

aging process. For example, polyuria is minimized by higher renal thresholds for

glucose, or it may be confounded by urinary incontinence or “prostate problems.”

Thirst is commonly blunted in elderly persons, increasing their risk of dehydration

and electrolyte imbalance. Hunger can be altered by medications or depression.

Fatigue often is discounted as “part of getting old,” and weight loss, although

sometimes profound, may be so gradual that it goes unnoticed for months to years.

(See Table 53-28 for a comparison of presenting symptoms for diabetes mellitus in

elderly patients compared with younger patients.)

Table 53-28

Presentation of Diabetes Mellitus in Elderly Patients Compared with Younger

Patients

Metabolic Abnormality Symptoms in Young Patients Symptoms in Elderly Patients

Serum osmolality Polydipsia Dehydration, confusion, delirium

Glucosuria Polyuria Incontinence

Catabolic state owing to insulin

deficiency

Polyphagia Weight loss, anorexia

HYPEROSMOLAR HYPERGLYCEMIC STATE

CASE 53-17, QUESTION 2: J.M. is diagnosed with hyperosmolar hyperglycemic state (HHS). Why are the

elderly predisposed to this condition, and what signs and symptoms are consistent with this diagnosis?

HHS is a condition characterized by extremely elevated plasma glucose

concentrations (>600 mg/mL) and high serum osmolality (>320 mOsm/L) without

ketoacidosis. Because patients with Type 2 diabetes have some residual insulin

production, they are usually protected against excessive lipolysis and ketone

production. Patients with Type 2 diabetes may exhibit HHS in later stages of diabetes

as loss of β-cells becomes advanced and residual insulin production continues to

drop.

156 Measurements of serum ketones and blood pH differentiate this condition

from DKA (see Case 53-10). The condition occurs when urinary fluid and electrolyte

losses secondary to glucosuria are inadequately replaced by oral fluid intake.

156

HHS primarily occurs in the elderly because several factors predispose this

population to hypodipsia. These include an inability to recognize thirst,

280 an inability

to ask for fluids (e.g., dementia, sedation, intubation), and an inability to get fluids on

demand (e.g., physical disabilities or restraints). Infections or other acute illnesses

(e.g., MI, GI bleeding, pancreatitis) that exacerbate diabetes can interact with the

hyperosmolar diuresis and hypodipsia to produce severe dehydration and

hyperglycemia. Drugs that increase plasma glucose concentrations (e.g.,

glucocorticoids), increase diuresis, or decrease mentation also can contribute to this

unfortunate situation.

J.M. presents with several symptoms of HHS dehydration, including osmolality

greater than 320 mOsm/L, plasma glucose greater than 600 mg/dL, pH > 7.3, elevated

bicarbonate, decreased skin turgor, hypotension, and the absence of serum ketones.

His pneumonia was probably the precipitating factor. Treatment involves rapid IV

hydration. Fluid replacement is provided in the same manner as with DKA. See Case

53-10, Question 2, and Table 53-24 for details.

156