Management of hypoglycemia in older adults with type 2 diabetes

ABSTRACT

Treatment of older adults with type 2 diabetes (T2D) is complex because they represent a heterogeneous group with a broad range of comorbidities, functional abilities, socioeconomic status, and life expectancy. Older adults with T2D are at high risk of recurring hypoglycemia, a condition associated with marked morbidity and mortality, because their counter-regulatory mechanism to hypoglycemia is attenuated, and recurring hypoglycemic episodes can lead to hypoglycemia unawareness. In addition, polypharmacy, a result of multiple chronic comorbidities (including heart disease, stroke, and chronic kidney disease), can increase the risk of severe hypoglycemia, especially when patients are taking sulfonylureas or insulin. Often the signs of hypoglycemia are nonspecific (sweating, dizziness, confusion, visual disturbances) and are mistaken for neurological symptoms or dementia. Consequences of hypoglycemia include acute and long-term cognitive changes, cardiac arrhythmia and myocardial infarction, serious falls, frailty, and death, often resulting in hospitalization, which come at a high economic cost. The American Diabetes Association has recently added three new recommendations regarding hypoglycemia in the elderly, highlighting individualized pharmacotherapy with glucose-lowering agents with a low risk of hypoglycemia and proven cardiovascular safety, avoidance of overtreatment, and simplifying treatment regimens while maintaining HbA1c targets. Thus, glycemic goals can be relaxed in the older population as part of individualized care, and physicians must make treatment decisions that best serve their patients’ circumstances. This article highlights the issues faced by older people with T2D, the risk factors for hypoglycemia in this population, and the challenges faced by health care providers regarding glycemic management in this patient group.

KEYWORDS:

• Hypoglycemia
• type 2 diabetes
• elderly
• polypharmacy

Introduction

There were 49 million elderly people (i.e. those aged ≥65 years) in the United States in 2016, and this number is predicted to increase to almost 95 million by 2060 [1]. Approximately 25% of Americans aged ≥65 years have diabetes mellitus (diagnosed and undiagnosed), the vast majority of whom have type 2 diabetes (T2D) [2,3]; thus, the number of older people with diabetes is expected to increase as the population continues to age. Around 61% of all US health care expenditure attributed to diabetes is used by people aged ≥65 years (i.e. $146 billion of the $237 billion total), most of which is obtained via the Medicare program [4]. Older adults with diabetes are at increased risk of micro- and macrovascular complications, such as heart disease and heart failure (HF), peripheral vascular disease, stroke, and chronic kidney disease (CKD) [5]. However, improvements in glycemic control have coincided with an increased prevalence of hypoglycemia, and hypoglycemia is one of the most common nonfatal complications in older patients with diabetes [6] and comes at a high economic cost [7]. The direct medical costs of hypoglycemia hospitalizations in the United States are substantial and have increased in recent years [8], with hospitalizations for hypoglycemia now surpassing those for hyperglycemia [9]. Data from the Healthcare Cost and Utilization Project National Inpatient Sample indicated that an estimated 1.5 million patients were admitted to the hospital for hypoglycemia between 1999 and 2011. In 2011 the annual cost of hypoglycemia hospitalization was $1.6 billion, up from $1.2 billion in 2001, with an average cost per hospitalization of $10,139 [9].

The exact incidence of hypoglycemia in older patients with diabetes is difficult to estimate because of a limited number of clinical trials in this age group, the heterogeneity of this population (i.e. independent individual living in the community vs. frail elderly living in nursing homes), and a lack of standardization in the diagnosis of hypoglycemia [10]. Nevertheless, older adults with diabetes have a greater risk of hypoglycemia and suffer greater adverse consequences from such events than younger individuals; thus, clinical management must balance optimizing glycemic control with minimizing hypoglycemia risk. The aim of this article is to highlight the issues faced by older people with T2D, the risk factors for hypoglycemia in this population, and the challenges faced by health care providers regarding glycemic management in this patient group.

Definition and classification of hypoglycemia

The glycemic threshold for hypoglycemia and for counter-regulatory responses can vary markedly among patients with diabetes, especially those receiving insulin or insulin secretagogues, but also within the same patient depending on their glycated hemoglobin (HbA1c) levels and previous hypoglycemic episodes. The glycemic threshold for hypoglycemia is higher in those with poor glycemic control and infrequent hypoglycemic episodes, and is lower in those with good glycemic control and more frequent hypoglycemic episodes [11]. Thus, it is insufficient to use one specific glucose level to define hypoglycemia; consequently, the American Diabetes Association (ADA) defined iatrogenic hypoglycemia as ‘all episodes of an abnormally low plasma glucose that expose the individual to potential harm’ [12]. More recently, a joint position statement from the International Hypoglycaemia Study Group of the ADA and the European Association of the Study of Diabetes (EASD) proposed revised definitions for reporting hypoglycemia in clinical trials of glucose-lowering agents [11]. These new recommendations establish three levels for blood glucose concentration: level 1, ≤70 mg/dL (hypoglycemia alert); level 2, <54 mg/dL (clinically significant hypoglycemia); and level 3, no specific glucose level (severe hypoglycemia, associated with severe cognitive impairment requiring external assistance for recovery). These categories are consistent with a Consensus Report from the ADA, the American Association of Clinical Endocrinologists (AACE), and others in the diabetes community [13], and are also reflected in the updated ADA standards of medical care in diabetes released in April 2018 [14]. Measurement of hypoglycemia episodes using accepted standard diagnostic criteria would allow greater accuracy of reporting during clinical trials for all patients with diabetes, including the elderly.

Challenges of accurately measuring glucose levels

Self-monitoring blood glucose (SMBG) meters are the mainstay of glucose monitoring in diabetes patients, including the elderly. The major drawbacks of these devices include the inability to show trends in glucose levels and the lack of an alarm for high or low glucose values, which would be particularly useful for older patients with reduced hypoglycemia awareness.

These devices also have limited accuracy, especially at glucose levels <75 mg/dL. International Organization for Standardization (ISO) and US Food and Drug Administration (FDA) standards require accuracy within 20% of the actual value in 95% of samples with glucose levels ≥75 mg/dL and ±15 mg/dL for glucose levels <75 mg/dL [12]. A recent analysis by the Diabetes Technology Society Blood Glucose Monitor System (BGMS) Surveillance Program revealed that only 6 of 18 commercially available SMBG meters met those standards [15]. In contrast, continuous glucose monitoring (CGM) devices take interstitial glucose readings every 5 minutes and provide a trend analysis. Newer models show accuracy to ±10% [16].

The original CGM devices were intended for short-term use and were blinded (i.e. patients were unable to see their glucose values and had to visit their physician for retrospective analysis), making them less than ideal for patient self-management. Newer models are unblinded and allow patients to see their glucose levels in real time, allowing a better understanding of how daily activities, meals, and medications affect blood glucose levels.

Real-time (RT) CGM has been shown to improve glycemic control and reduce hypoglycemia in adults with type 1 diabetes (T1D) and T2D [17–19], and is endorsed as part of the standard of care by AACE [20], the ADA [21], and the Endocrine Society [22]. Access to CGM devices for older patients has improved with the recent coverage of these devices by Medicare [23]. A recent survey assessed the impact of RT-CGM on the occurrence of self-reported episodes of hypoglycemia and quality of life in patients aged ≥65 years, and suggests that an RT-CGM device may be of significant value for adults aged ≥65 years with T1D or T2D [24].

Similarly, intermittently viewed glucose monitoring (iCGM), also known as ‘flash’ monitoring, has been shown to address several limitations of SMBG. This system provides the current glucose value as well as retrospective glucose data for a specified time period when scanned. In an open-label randomized clinical study of patients with T2D treated with intensive insulin therapy, the use of flash sensor-based technology resulted in less time spent in a hypoglycemic state, especially at night, and was particularly significant in those aged ≥65 years [25]. Drawbacks can include lack of alarms for high or low glucose and visualization of data only upon the patient’s choosing [26]. More recent developments in glucose monitoring technology include the use of biosensors that are integrated into wearable platforms, and measurement of glucose concentrations via noninvasive methods by using physiological sources other than blood, such as breath, saliva, and ocular fluid (e.g. via smart contact lenses) [27]. The ultimate cost and availability of such new devices, particularly to patients using Medicare, as well as their ease of use by the elderly, are yet to be determined.

Pathophysiology and risk factors for hypoglycemia

The counter-regulatory mechanism in response to hypoglycemia in healthy individuals is shown in Figure 1 [28]. When blood glucose levels fall, insulin release is suppressed, and counter-regulatory hormones, such as glucagon and epinephrine, are activated to stimulate hepatic glucose production and inhibit peripheral glucose uptake until euglycemia is restored. In adults with diabetes this counter-regulation is defective: there is a loss of ability to decrease insulin production and to increase glucagon release, combined with an attenuated increase in epinephrine [12]. The response by older adults to a drop in glucose appears to depend on how frequently they experience hypoglycemia [12]. Recurrent hypoglycemia reduces the glucose level at which the counter-regulatory response occurs; consequently, patients with frequent hypoglycemia experience the symptoms from the adrenergic response to a decrease in blood glucose at lower and lower glucose levels. In some instances the glucose level that triggers the counter-regulatory response is below the glucose level associated with neuroglycopenia, also called hypoglycemia unawareness, and results from attenuated increases in sympathoadrenal activity. The combination of a defective counter-regulation with hypoglycemia unawareness is called hypoglycemia-associated autonomic failure, and is caused most often by prior hypoglycemic episodes [12].

Figure 1. Glycemic thresholds for secretion of counter-regulatory hormones and onset of symptoms in response to hypoglycemia [28]. Republished with permission of John Wiley and Sons, Ltd., from Hypoglycaemia in Clinical Diabetes, Frier & Fisher 1999, permission conveyed through Copyright Clearance Center, Inc.

In addition to older age and prior hypoglycemic episodes, there are several other risk factors for hypoglycemia in older adults with diabetes. Polypharmacy, which is more common in the elderly due to the greater occurrence of multiple chronic clinical conditions, may increase the risk of severe hypoglycemia, especially when patients are on sulfonylureas or insulin [5,10]. Additionally, age-related changes in pharmacokinetics and pharmacodynamics have the potential of increasing the adverse effects of polypharmacy in this patient population [29]. Older adults are also at an increased risk of hypoglycemia due to numerous clinical conditions, including decreased hormonal regulation and counter-regulation, suboptimal intake of water and/or food, decreased intestinal absorption, and cognitive impairment [14]. The elderly tend to present with neuroglycopenic symptoms (i.e. dizziness, visual disturbances, increased agitation, and/or confusion) rather than adrenergic symptoms (i.e. palpitations, sweating, tremors) [30], and the former symptoms may be caused by other conditions, particularly in the elderly [5,30]. Finally, both cardiovascular (CV) disease [31,32] and CKD [29] have been shown to be strong risk factors for hypoglycemia in older adults. An observational cohort study of >900,000 adults from the SUPREME-DM network assessed severe hypoglycemia rates from 2005 to 2011 [31]. Annual rates of severe hypoglycemia were higher among adults with CKD, congestive HF, and CV disease (CVD), among others, and increased over time [31]. Similarly, in a prospective cohort study from 2001 to 2012, the development of severe hypoglycemia was associated with a history of CVD in 624 patients with T2D after adjusting for sex, age, diabetes duration, hypertension, HbA1c, diabetic complications, CV autonomic neuropathy, and insulin use [32].

The challenge of glycemic management in older adults

Consequences of hypoglycemia in the elderly can include acute and long-term cognitive changes, serious falls, cardiac arrhythmia and myocardial infarction (MI), frailty, and possibly death [5]. Treatment of older adults with diabetes is complex because they are a heterogeneous group with a broad range of comorbidities (e.g. CVD, visual impairment, arthritis), functional disabilities (e.g. mobility, eating, personal hygiene), socioeconomic influences (e.g. education, income, living accommodations), and life expectancy. Although the overall treatment goals for diabetes management in older adults are similar to those in younger people [33], glycemic targets need to be more flexible for the former population.

The majority of evidence regarding treating to intensive versus standardized glycemic targets comes from four randomized controlled clinical trials (ADVANCE [34], UKPDS [35], ACCORD [36], VADT [37]) that included few older patients. All four of these clinical trials demonstrated that intensive glucose control increases the risk of severe hypoglycemia [34–37]. The ACCORD trial also demonstrated an increased risk of mortality in the intensive treatment arm [36]. In the ACCORD trial, the risk of severe hypoglycemia increased with age (hazard ratio [HR], 1.03; 95% confidence interval [CI], 1.03–1.07 per 1-year increase in age) and with the duration of diabetes [38]. In addition, mortality was three times higher for patients who had severe hypoglycemia in either the conventional or intensive treatment groups compared with those patients without severe hypoglycemia [39]. Similarly, in VADT, more than one episode of severe hypoglycemia led to an 88% increase in the relative risk of sudden death [40]. It has been suggested that the increase in mortality is not directly linked to lower HbA1c levels, but that hypoglycemia could be an important factor, especially in older patients with lower HbA1c levels who may be at higher risk of mortality due to poor nutrition and frailty [10]. Of note, the aforementioned studies were all conducted in the United States. A recent meta-analysis investigated use of intensive versus standard glycemic control in North America compared with the rest of the world [41]. For trials conducted in North America, intensive therapy compared with standard glycemic control resulted in significantly higher all-cause mortality (odds ratio [OR], 1.21; 95% CI, 1.05–1.40) and CV mortality (OR, 1.41; 95% CI, 1.05–1.90) than trials conducted in the rest of the world (all-cause mortality OR, 0.93; 95% CI, 0.85–1.03; interaction p = 0.006; CV mortality OR, 0.89; 95% CI, 0.79–1.00; interaction p = 0.007). The risk of severe hypoglycemia was significantly higher in trials of intensive therapy in North America (OR, 3.52; 95% CI, 3.07–4.03) compared with the rest of the world (OR, 1.45; 95% CI, 0.85–2.47; interaction p = 0.001).

A retrospective cohort study of Veterans Affairs (VA) data and Medicare claims from 15,880 veterans aged ≥65 years with T2D and dementia assessed their degree of glycemic control and, in those with tight control (HbA1c <7%), use of medications with a high risk of hypoglycemia [42]. Overall, 52% (n = 8276) of veterans had tight glycemic control, of whom 75% used sulfonylureas and/or insulin, with a higher risk of hypoglycemia occurring in those aged 75–84 years (OR, 1.28; 95% CI, 1.13–1.45; p < 0.001) and aged ≥85 years (OR, 1.60; 95% CI, 1.37–1.88; p < 0.001). This study demonstrated that many older veterans are at an increased risk of hypoglycemia based on the intensive diabetes treatment they are receiving [42]. Importantly, a bidirectional interaction between hypoglycemia and cognitive impairment in older patients receiving glucose-lowering agents has been shown [43]. A meta-analysis of five studies showed a significantly increased risk of dementia in patients who experienced hypoglycemic episodes (OR, 1.68; 95% CI, 1.45–1.95). Additionally, an increased risk of hypoglycemia was found in patients with dementia (OR, 1.61; 95% CI, 1.25–2.06). Given the nonspecific symptoms that can disguise hypoglycemia in older people (e.g. sweating, dizziness, visual disturbances, nightmares, confusion), there is opportunity for physicians to uncover hidden hypoglycemia in their patients by asking about these symptoms and potentially considering a CGM device.

The ADA has added three new recommendations regarding hypoglycemia in the elderly that highlight the importance of the following: (1) individualizing pharmacologic therapy in this patient population to reduce the risk of hypoglycemia, (2) avoiding overtreatment, and (3) simplifying complex regimens (if possible) while maintaining the HbA1c target [14]. Thus, glycemic goals can be relaxed as part of individualized care, but symptomatic hyperglycemia must be avoided. Physicians and their older patients with diabetes must make treatment decisions that best serve the individual’s circumstances. Estimated life expectancy, the duration of diabetes, the need for insulin, the use of sulfonylureas (particularly glyburide), and the presence of other comorbidities, including cognitive impairment, all factor into the pros and cons of intensive glycemic control and the risk of hypoglycemia [44]. Treatment goals include HbA1c <7.5% for healthy adults with few comorbidities and good cognitive function, and HbA1c <8% to 8.5% for adults with multiple comorbidities and cognitive impairment [14]. Physicians also need to match treatment complexity with the self-management abilities of the patient, and simplify complex treatment regimens if individualized HbA1c targets can be maintained. To aid physicians in individualizing treatments for their patients with T2D, an algorithm was developed via a worldwide survey conducted among leading diabetologists [45]. The algorithm is based on five objective parameters, of which the risk of hypoglycemia was deemed the most important [45].

Treatment options for older adults with T2D

Classes of glucose-lowering agents with demonstrated CV safety and a low risk of hypoglycemia are preferred for this patient population; these include metformin, dipeptidyl peptidase-4 (DPP)-4 inhibitors (linagliptin, sitagliptin; note potential HF risk with saxagliptin, alogliptin) [46], glucagon-like peptide-1 (GLP)-1 receptor agonists, and sodium-glucose co-transporter 2 (SGLT2) inhibitors (Table 1). If treatment costs are a major issue and the patient does not have atherosclerotic CVD or CKD, a thiazolidinedione (pioglitazone) may be chosen as an add-on to metformin [44,46,47]. However, thiazolidinediones should be used cautiously in older patients due to their risk of fluid retention, HF, macular edema, and bone fractures [48].

Table 1. Considerations for glucose-lowering therapy in older adults [21,44,49,50].

Therapy	Expected HbA1c reduction (%) [44]	CV effects – ASCVD	CV effects – CHF	Additional considerations	Considerations for older adults	Monthly cost
Metformin	1–2	Potential benefit	Neutral	Gastrointestinal side effects common Potential B12 deficiency	Recommended initial pharmacotherapy after lifestyle Low risk of hypoglycemia Contraindicated with eGFR <30	Low
SGLT2 inhibitors	0.5–0.7	Benefit: Canagliflozin Empagliflozin*	Benefit: Canagliflozin Empagliflozin	FDA black box: Risk of amputation (canagliflozin) Risk of bone fractures (canagliflozin) Genitourinary infections Volume depletion DKA risk (all agents, rare in T2D) Increased LDL cholesterol Risk of Fournier’s gangrene	Convenient oral route Causes some weight loss Causes slight reduction in BP Canagliflozin: not recommended for eGFR <45 Dapagliflozin: not recommended with eGFR <60; contraindicated with eGFR <30 Empagliflozin: contraindicated with eGFR <30	High
DPP-4 inhibitors	0.5–0.8	Neutral	Potential risk: Saxagliptin Alogliptin	Potential risk of acute pancreatitis Joint pain	Convenient oral route Renal dose adjustment required except for linagliptin; can be used in renal impairment	High
GLP-1 RAs	1–1.5	Neutral: Lixisenatide Exenatide ER Benefit: Liraglutide*	Neutral	FDA black box: risk of thyroid C-cell tumors (liraglutide, abiglutide, dulaglutide, exenatide ER) Gastrointestinal side effects common Acute pancreatitis risk Injection site reactions	Injection route requires good visual acuity and motor skills Weight loss may be undesirable Exenatide: not indicated with eGFR <30 Lixisenatide: caution with eGFR <30 Increased risk of side effect in patients with renal impairment	High
TZDs	1–2	Potential benefit: Pioglitazone	Increased risk	FDA black box: CHF (pioglitazone, rosiglitazone) Benefit in NASH Fluid retention (edema, CHF) Bladder cancer (pioglitazone) Increased LDL cholesterol (rosiglitazone) Risk of bone fractures	Generally not recommended in renal impairment due to potential for fluid retention Use with caution in those at risk for falls/fractures	Low
Sulfonylureas (Second generation)	1–2	Neutral	Neutral	Risk of hypoglycemia FDA Special Warning on increased risk of CV mortality based on studies of an older SU (tolbutamide)	Use with caution Glyburide not recommended Glipizide and glimepiride: initiate conservatively to avoid hypoglycemia	Low
Insulin	1.5–3.5	Neutral	Neutral	Risk of hypoglycemia Weight gain	Injection route requires good visual acuity and motor skills Lower insulin doses required with decrease in eGFR; titrate per clinical experience	Low (human insulin) High (analogs)

*FDA approved for CVD benefit.

ASCVD = atherosclerotic cardiovascular disease; BP = blood pressure; CHF = congestive heart failure; CV = cardiovascular; DKA = diabetic ketoacidosis; DPP-4 = dipeptidyl peptidase-4; eGFR = estimated glomerular filtration rate (mL/min/1.73 m²); ER = extended release; FDA = US Food and Drug Administration; GLP-1 RA = glucagon-like peptide-1 receptor agonist; HbA1c = glycated hemoglobin; LDL = low-density lipoprotein; NASH = non-alcoholic steatohepatitis; SGLT2 = sodium-glucose co-transporter 2; SU = sulfonylurea; T2D = type 2 diabetes; TZD = thiazolidinedione.

The recent Consensus Report for the Management of Hyperglycemia in Type 2 Diabetes issued by the ADA/EASD has taken into account new evidence from CV outcomes trials (CVOTs) demonstrating that SGLT2 inhibitors and GLP-1 receptor agonists improve CV outcomes as well as secondary outcomes such as HF and progression of renal disease in patients with established CVD or CKD [47]. Consequently, SGLT2 inhibitors and GLP-1 receptor agonists are now recommended as part of glycemic management in patients with established atherosclerotic CVD after metformin therapy [47]. Metformin is the first-choice agent for older patients with estimated glomerular filtration rate (eGFR) ≥30 mL/min/1.73 m² [14]. Weight loss and gastrointestinal side effects may limit the use of this agent in this patient population, and it is contraindicated in those with advanced renal insufficiency and should be used with caution in those with impaired hepatic function or congestive HF.

Oral DPP-4 inhibitors are weight-neutral and cause minimal hypoglycemia. Neutral effects on major CV events have been shown in four CVOTs: TECOS [51], EXAMINE [52], SAVOR-TIMI 53 [53], and CARMELINA [54]. However, a significant increase in the risk for hospitalization for heart failure (HHF) was reported for saxagliptin [55] and a nonsignificant risk was reported for alogliptin [56], resulting in a warning in the prescribing information for both agents [57,58]. Since then the FDA has extended a class-labeling warning for HF to all marketed DPP-4 inhibitors in the United States. A post hoc analysis from the SAVOR-TIMI trial revealed that the use of saxagliptin in the elderly population demonstrated a safety and efficacy profile similar to that observed in the overall population [59]. Similarly, an analysis of data from the TECOS trial in patients aged ≥75 years [60] and data on alogliptin use in those aged ≥65 years [61] raised no safety concerns. Other smaller studies have reported the renal safety of all DPP-4 inhibitors [62], although dose adjustments in renal impairment are necessary for all DPP-4 inhibitors, with the exception of linagliptin. The recently published CARMELINA trial, which evaluated the effects of linagliptin on CV and kidney outcomes in patients with T2D at high risk of CV and kidney events, demonstrated a noninferior risk of the primary composite CV outcome (major adverse cardiac event [MACE]: CV death, nonfatal MI, or nonfatal stroke; HR, 1.02; 95% CI, 0.89–1.17; p < 0.001 for noninferiority) and no increased risk for HHF (HR, 0.90; 95% CI, 0.74–1.08; p = 0.26), for linagliptin versus placebo over a median 2.2 years [54]. The ongoing CAROLINA trial is comparing the effects of linagliptin and glimepiride, in addition to standard of care, on CV events in patients with early T2D and increased CV risk or established CV complications. The trial is expected to complete in 2019 [63,64] and will provide additional evidence on the CV safety of these two agents.

The use of injectable GLP-1 receptor agonists requires a certain amount of visual, motor, and cognitive skills, which may be difficult to master for some older individuals. Moreover, these agents can be associated with nausea, vomiting, and diarrhea, and the ensuing weight loss may not be desirable in some older people. However, clinical studies have demonstrated CV benefits for liraglutide and semaglutide. The LEADER trial showed a significant 13% reduction in the first occurrence of CV death, nonfatal MI, or nonfatal stroke, a 22% reduction in CV death, and a 15% reduction in total mortality with liraglutide versus placebo [65]; liraglutide is indicated to reduce the risk of MACE in adults with T2D and established CVD [66]. The SUSTAIN-6 trial showed a significant 26% reduction to the first occurrence of CV death, nonfatal MI, or nonfatal stroke, a significant 39% decrease in nonfatal stroke and a nonsignificant 26% decrease in nonfatal MI with semaglutide versus placebo [67]; semaglutide is not indicated in the United States for the reduction of macrovascular events. Finally, results from the REWIND trial, which assesses the effects of once-weekly dulaglutide on the 3-point MACE outcome of CV death, nonfatal MI, or nonfatal stroke are expected in 2019 [68].

Emerging evidence from preclinical studies suggests that GLP-1 receptor agonists may provide neuroprotection in several neurodegenerative diseases, including Parkinson’s and Alzheimer’s disease [69]. Improvements in motor and cognitive assessments have been demonstrated in a clinical trial of exenatide in patients with Parkinson’s disease [70], and other trials assessing exenatide and liraglutide in both disease states are currently ongoing. GLP-1 receptor agonists may thus represent a sound therapeutic option for older patients with T2D and cognitive impairment or early Alzheimer’s disease.

SGLT2 inhibitors, including canagliflozin, dapagliflozin, and empagliflozin are orally administered, lead to modest weight loss, and have demonstrated CV benefits in clinical studies. Dose adjustment and restrictions are in place for those with renal impairment. Three CVOTs have been completed for the SGLT2 inhibitors. The pivotal empagliflozin CVOT, EMPA-REG OUTCOME, showed a 14% relative risk reduction in MACE, which was driven by a marked (and unexpected) 38% reduction in CV mortality, and a 35% relative risk reduction in HHF in patients receiving empagliflozin versus placebo, in addition to standard of care [71]. The composite of HHF or CV death was reduced by 34% (HR, 0.66; 95% CI, 0.55–0.79; p < 0.001), and the HF and mortality outcomes were reduced consistently irrespective of the presence of baseline HF or whether patients had a history of atherothrombotic events (stroke or MI) [72,73]. Moreover, the addition of empagliflozin to standard of care was associated with a lower risk of progression of kidney disease (HR, 0.61; 95% CI, 0.53–0.70; p < 0.001) [74]. Empagliflozin was not associated with an increased risk of lower limb amputations versus placebo (HR, 1.00; 95% CI, 0.70–1.44) [75]. Empagliflozin is indicated to reduce the risk of CV death in adult patients with T2D and established CVD [76].

The CANVAS Program (an integrated analysis of data from the CANVAS and CANVAS-R trials) [77] showed a significant 14% reduction in the risk of MACE for canagliflozin groups compared with placebo; in addition, HHF was reduced by 33% (HR, 0.67; 95% CI, 0.52–0.87; p = 0.002), and serious declines in renal function (composite of reduction in eGFR, requirement for renal replacement therapy, or death from renal causes) was reduced by 40% (HR, 0.60; 95% CI, 0.47–0.77), although the renal outcomes were not statistically significant. The CANVAS Program showed an increased risk for amputations (HR, 1.97; 95% CI, 1.41–2.75) [77] and bone fractures (HR 1.26; 95% CI, 1.04–1.52) [78]. The increase in fractures was driven by CANVAS patients who were older, had a history/risk of CVD, and had lower baseline eGFR and higher baseline diuretic use in comparison to patients in the CANVAS-R study [78]. The FDA has added a boxed warning about the amputation risk and a warning about the bone fracture risk to the canagliflozin prescribing information [79]. A new analysis compared the CV benefits of canagliflozin in a primary prevention cohort (those ≥50 years of age with two or more CV risk factors but no prior CV events) and a secondary prevention cohort (those ≥30 years of age with a prior CV event) [80]. Although patients with T2D and prior CV events had higher rates of CV outcomes than the primary prevention cohort, canagliflozin reduced CV and renal outcomes across both treatment groups without statistical evidence of heterogeneity [80].

The DECLARE-TIMI 58 trial is the most recent CVOT for SGLT2 inhibitors and the largest trial in this class, with more than 17,000 patients. Results showed a positive effect of dapagliflozin on the coprimary endpoint of CV death and HHF, with a significant reduction for dapagliflozin versus placebo (4.9% vs. 5.8%; HR, 0.83; 95% CI, 0.73–0.95; p < 0.005). This finding was driven by a 27% reduction in HHF (HR, 0.73; 95% CI, 0.61–0.88) with no effect on CV mortality [81]. Dapagliflozin was noninferior, but not statistically superior, for 3-point MACE compared with placebo (8.8% vs. 9.4%; HR, 0.93; 95% CI, 0.84–1.03; p = 0.17). There was no increased risk for stroke, fracture, or amputation. Finally, a meta-analysis of the three aforementioned CVOTs evaluated MACE, the composite of CV death or HHF, and progression of renal disease [82]. Efficacy outcomes were stratified by baseline presence of atherosclerotic CVD, HF, and degree of renal function. SGLT2 inhibitors had a moderate effect on MACE (11% reduction), but showed solid benefits on CV death or HHF (23% reduction) and risk of progression of renal disease (45% reduction) regardless of the presence of atherosclerotic CVD or history of HF [82].

Although not a traditional CVOT, CVD-REAL 2, an observational study, examined a broad range of CV outcomes in patients taking SGLT2 inhibitors versus other glucose-lowering agents [83]. In this study, use of SGLT2 inhibitors was associated with a lower risk of death (HR, 0.51; 95% CI, 0.37–0.7; p < 0.001), HHF (HR, 0.64; 95% CI, 0.5–0.82; p = 0.001), MI (HR, 0.81; 95% CI, 0.74–0.88; p < 0.001), and stroke (HR, 0.68; 95% CI, 0.55–0.84; p < 0.001) [83].

A newer generation sulfonylurea may be an appropriate choice for older patients with T2D and without established CVD or kidney disease when the prescription cost is a consideration [47]. The second-generation sulfonylureas glyburide (glibenclamide) and glimepiride have a long half-life and are associated with a high risk of hypoglycemia, especially in vulnerable patient populations, such as the elderly and those with renal impairment [84]. Guidelines from the American Geriatric Society state that glyburide generally should not be prescribed to older patients with T2D, and chlorpropamide (a first-generation sulfonylurea) should be avoided altogether, because of the increased risk of hypoglycemia [85]. Similarly, the 2018 World Health Organization guidelines on use of second- and third-line medications in adults with T2D recommends to avoid glyburide in those aged ≥60 years [86]. If sulfonylurea therapy is chosen, short-acting agents such as gliclazide (if outside the United States) [86] and glipizide are preferable in an elderly patient population, as they may have a lower risk of hypoglycemia than other sulfonylureas [14,47].

Like sulfonylureas, insulin is associated with a risk of hypoglycemia [14] and should be used with caution in older adults. However, newer insulin analogs, such as insulin degludec and insulin glargine 300, have demonstrated good efficacy and safety with low rates of documented hypoglycemia in this patient population [87–90]. In the DEVOTE trial, the efficacy and safety of insulin degludec versus insulin glargine 100 was assessed in 7637 patients with T2D at high risk of CV events [87]. Prespecified adjudicated severe hypoglycemia occurred in 4.9% and 6.6% of patients in the degludec and glargine groups, respectively, an absolute difference of 1.7% [87]. The SENIOR trial was an open-label parallel-group study in which older patients (mean age, 71 years) were randomized to either insulin glargine 100 or 300 [88]. Rates of documented symptomatic hypoglycemia were lower with insulin glargine 300, and this was more apparent in the subgroup of patients aged ≥75 years versus the overall population (annualized rate of documented symptomatic hypoglycemia, glargine 300, 1.12; glargine 100, 2.71; rate ratio, 0.45) [88]. Similarly, two patient-level meta-analyses from the EDITION 1, 2, and 3 trials support these findings, demonstrating that insulin glargine 300 resulted in consistently less nocturnal hypoglycemia than insulin glargine 100 [89]. These newer insulin analogs may thus be viable options for treating older adults with T2D.

Of note, the administration of insulin therapy requires good visual and motor skills, as well as cognitive ability, in the patient and/or their caregiver. Although once-daily basal insulin injections are associated with minimal side effects and may be reasonable options for elderly patients, multiple daily injections may be too complex for older patients with diabetes complications, coexisting chronic conditions, and/or limited functional status.

Hospitalization for hypoglycemia in older adults and their subsequent transition home

In patients with diabetes and who are aged ≥80 years, severe hypoglycemia accounted for up to one in six hospital admissions [91]. The US Centers for Disease Control and Prevention (CDC) reported that in 2014 there were 245,000 emergency department visits for hypoglycemia in adults aged ≥18 years (11.2 for 1000 persons with diabetes) [3]. A retrospective study of 33.9 million Medicare beneficiaries aged ≥65 years showed that from 1999 to 2011 hospital admissions for hypoglycemia were two-fold higher for older patients (aged ≥75 years) compared with younger patients (aged 65–74 years) [8]; additionally, 1-year mortality rates for those with hypoglycemia during that 12-year period ranged from 18% to 23% [8]. A further retrospective US study examined the association between hypoglycemia, length of hospital stay, and mortality in approximately 107,000 insulin-treated patients [92], and found that inpatient mortality occurred in 6.5% of hospitalizations with hypoglycemia versus 3.8% of those without (p < 0.001), and in 7.6% of hospitalizations for those with severe hypoglycemia. After adjusting for age, gender, and selected comorbidities, hypoglycemia and severe hypoglycemia were associated with a significant increase in inpatient mortality risk (OR, 1.66 and 1.44, respectively) [92]. Length of stay was also increased in hospitalizations with hypoglycemia versus in those without (median 8.2 days vs. 5.2 days; p < 0.0001) [93].

Based on CDC data from 2014, there were a total of 7.2 million hospital discharges for patients with a diabetes diagnosis [3]. Irrespective of the reason for hospitalization, the transition of an older patient back to his/her home can be difficult [94]. Older patients in particular may have cognitive problems, functional limitations, pain, and/or a complex treatment regimen, all of which can result in inadequate symptom management and poor transition outcomes [94]. The risk for medication errors is five to six times greater at home than in an acute care setting, and mismanagement of hypoglycemic agents at home is a common problem for patients with diabetes [94].

The Transitional Care Model, designed by the University of Pennsylvania, is a widely recognized and proven model that helps transition older patients from the hospital to their home, even in cases of an interim stay at another nursing facility [93]. The model features a nurse-led multidisciplinary provider approach that includes patients and their caregivers as part of a team. Another care coordination model, the Patient-Centered Medical Home (PCMH), comes from the Agency for Healthcare Research and Quality [95]. The PCMH is a model of primary care that organizes and delivers primary health care through five core functions, one of which is coordinated care. The PCMH coordinates patient care across the health care system, including specialists, hospitals, home health care, and community services. This is particularly critical during times of transition between different sites, such as hospital discharges. These models could be applicable to elderly patients with diabetes who have been hospitalized for hypoglycemia.

Improving patient outcomes and reducing health care cost are both important goals in managing any chronic disease, including diabetes. The pay-for-performance model provides financial incentives to physicians for achieving better health outcomes, which is in contrast to the traditional fee-for-service model. Some studies have shown that pay-for-performance can improve physician continuity and patient outcomes [96,97]. However, a recent systematic review indicated that pay-for-performance programs may be associated with improved processes of care in ambulatory settings, but that consistently positive associations with improved health outcomes have yet to be demonstrated [98].

Conclusions

Risk factors for hypoglycemia such as renal impairment, CVD, and polypharmacy all increase with advancing age in adults with T2D. Treating older adults with T2D centers on finding the right medication for each patient’s particular circumstances and maximizing optimal glycemic control, while minimizing risk of hypoglycemia. This can be achieved through individualized management approaches that take into account the patient’s clinical characteristics (including body weight, comorbidities, and life expectancy), social circumstances (living in the community alone/with family, resident in a nursing home), personal preferences, and level of health insurance coverage.

Declaration of interest

Dr Freeman has served as a speaker for Boehringer Ingelheim, Eli Lilly, Novo Nordisk, AstraZeneca, Amgen, and Valeritas. The author has no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties. Writing support was provided by Linda Merkel, PhD, and Marissa Buttaro, MPH, RPh, of Envision Scientific Solutions, which was contracted and compensated by Boehringer Ingelheim Pharmaceuticals, Inc. for these services. Peer reviewers on this manuscript have no relevant financial relationships to disclose.

Additional information

Funding

This manuscript was funded by Boehringer Ingelheim. Writing support was provided by Linda Merkel, PhD, and Marissa Buttaro, MPH, RPh, of Envision Scientific Solutions, which was contracted and compensated by Boehringer Ingelheim Pharmaceuticals, Inc. for these services.