Hypoglycemia in patients with type 2 diabetes using concomitant exenatide BID and long-acting insulin therapy

Abstract

Objective:

The objective of this study was to examine the frequency of hypoglycemia among patients with type 2 diabetes who had concomitantly used exenatide BID (exenatide) and long-acting insulin and continued this combination vs those who continued long-acting insulin alone.

Methods:

Retrospective analyses, using a large managed care database, were used to estimate the frequency of hypoglycemia (episodes/patient/6 months) for patients who concomitantly used exenatide and long-acting insulin during a 6-month follow-up period.

Results:

From among 2082 patients on concomitant exenatide and long-acting insulin, those who continued this combination (n = 472) had a lower frequency of hypoglycemia compared to those who remained on long-acting insulin alone (n = 312) (0.03 ± 1.9 vs 0.10 ± 1.01 [episodes/patient/6 months]; p < 0.0001).

Limitations:

Only hypoglycemia that required medical intervention (coded for hypoglycemia) was captured. The study could not evaluate any association between insulin dose titration and hypoglycemia or examine other outcomes such as HbA1c, weight, and body mass index, due to lack of data availability.

Conclusions:

Patients who concomitantly used exenatide BID and long-acting insulin experienced a lower rate of hypoglycemia.

Key words::

• Hypoglycemia
• Type 2 diabetes
• Exenatide BID
• Insulin therapy

Introduction

Exenatide, a glucagon-like peptide-1 (GLP-1) receptor agonist, improves glycemic control when used twice daily as monotherapy or in combination with oral anti-hyperglycemic medication therapy in patients with type 2 diabetes1–3. Concomitant use of exenatide BID (exenatide) and insulin is not an approved indication4. Observational studies from clinical settings and randomized clinical trials (RCTs) reported that adding exenatide to existing insulin regimens resulted in additional HbA1c reduction, weight loss, and lower risk of hypoglycemia2,5–8. Hypoglycemia concerns are considered a key barrier for physicians and patients to achieve glycemic goals, especially with insulin therapy9,10. The primary objective of this study was to examine the frequency of hypoglycemia (episodes/patient/6 months) over a 6-month period in patients with type 2 diabetes who were concomitantly using exenatide BID and long-acting insulin in real-world clinical practice settings. We were most interested in examining patients that continued both exenatide BID and insulin and those who continued insulin, but discontinued exenatide.

Methods

The HealthCore Integrated Research database was used for analyses. This study included patients with: (1) a claim of exenatide between 5/1/2005 and 10/1/2007, (2) at least 12 months pre- and 6 months post-index continuous eligibility (index date defined as the date of receiving first exenatide BID (exenatide) prescription), (3) diagnosis of type 2 diabetes, and (4) long-acting insulin use was defined as a claim for NPH or basal analog insulin (detemir or glargine) within 100 days before to 15 days following the index date. Patients <18 years old with a diagnosis of type 1 or gestational diabetes were excluded. Persistence on either or both therapies was defined as a <60-day gap between subsequent prescription claims11. This study identified that 99% of claims for insulin had days supply of ≤30 days. Hence, the Sikka et al.11 method reported for measurement of persistence to medications was deemed appropriate for this study.

Four cohorts were defined based on use of insulin and exenatide in the post-index period. Patients on both therapies were defined based upon the persistency to both the therapies (i.e., a <60-day gap between subsequent prescription claims for both the medications concurrently). The cohorts were (1) patients continuing on both therapies (n = 472 [22.7%]), (2) patients continuing insulin and discontinuing exenatide (n = 312 [15.0%]), (3) patients continuing exenatide and discontinuing insulin (n = 498 [23.9%]), and (4) patients discontinuing both therapies (n = 800 [38.4%]). The current study focuses on hypoglycemia comparisons between only the first two cohorts. Hypoglycemic events were identified by the International Classification of Diseases, Ninth Revision (ICD-9) codes 250.8, 251.0, 251.1, and 251.212. The frequency of hypoglycemia was assessed using unadjusted Poisson regression.

Results

A total of 2082 patients met the study criteria of concomitantly using exenatide and long-acting insulin at baseline. Mean (standard deviation) age was 54 (9) years and 45% were females. Common comorbidities included hypertension (88–91%), dyslipidemia (84–90%), and any cardiovascular disease (21–26%). No significant differences were observed in demographic or clinical characteristics among the four cohorts except for dyslipidemia (p = 0.037). Demographic and clinical characteristics data are summarized in Table 1.

Table 1.  Demographic and clinical characteristics of study population.

	Continued both Exenatide and Insulin, n = 472 (22.7%)	Continued Insulin, discontinued Exenatide, n = 312 (15.0%)	Continued Exenatide, discontinued Insulin, n = 498 (23.9%)	Discontinued both Exenatide and Insulin, n = 800 (38.4%)	p-value
Age, M ± SD (at index)	55 ± 8	54 ± 9	54 ± 9	54 ± 10	0.895
18–34	5 (1.1%)	7 (2.2%)	11 (2.2%)	25 (3.1%)
35–44	50 (10.6%)	39 (12.5%)	57 (11.5%)	96 (12.0%)
45–54	161 (34.1%)	109 (34.9%)	173 (34.7%)	267 (33.4%)
55–64	213 (45.1%)	123 (39.4%)	211 (42.4%)	319 (39.9%)
≥65	43 (9.1%)	34 (10.9%)	46 (9.2%)	93 (11.6%)
Female, n (%)	235 (49.8%)	146 (46.8%)	225 (45.2%)	367 (45.9%)	0.477
Pre-index* comorbidity
Hypertension	428 (90.7%)	284 (91.0%)	438 (88.0%)	713 (89.1%)	0.415
Dyslipidemia	412 (87.3%)	280 (89.7%)	430 (86.4%)	668 (83.5%)	0.037
Obesity	62 (13.1%)	38 (12.2%)	50 (10.0%)	98 (12.3%)	0.485
Nephropathy	56 (11.9%)	35 (11.2%)	66 (13.3%)	91 (11.4%)	0.749
Neuropathy	79 (16.7%)	57 (18.3%)	71 (14.3%)	122 (15.3%)	0.421
Retinopathy	54 (11.4%)	37 (11.9%)	47 (9.4%)	71 (8.9%)	0.312
Congestive heart failure	31 (6.6%)	20 (6.4%)	30 (6.0%)	38 (4.8%)	0.496
Peripheral vascular disease	28 (5.9%)	18 (5.8%)	25 (5.0%)	30 (3.8%)	0.273
Neoplasm, malignant and benign	80 (17.0%)	57 (18.3%)	102 (20.5%)	139 (17.4%)	0.456
Deyo Charlson Comorbidity  Index Score13, M ± SD (at index)	2.8 ± 1.9	3.0 ± 2.5	2.8 ± 2.3	2.6 ± 2.1	0.124

Results presented as n (%) unless otherwise noted.

*Pre-index defined as 365 days prior to index (any time in the pre-index period).

Among patients who were concomitantly using exenatide and long-acting insulin, 22.7% continued both treatments while 15% continued insulin but discontinued exenatide in the post-index period (Table 1). Patients who continued on both therapies had a lower frequency of hypoglycemia compared to those who remained on long-acting insulin alone (0.03 [1.9] vs 0.10 [1.01] episodes/patient/6 months, p < 0.0001). There was a decrease in use of sulfonylureas (SUs) in both cohorts. Subsequent SU use was lower in the insulin-only group compared to the exenatide plus insulin (38.7% vs 45.9%) group. The frequency of hypoglycemic events for patients who continued on exenatide alone and for those who discontinued both therapies (groups not considered for comparison in the current study) are shown in Figure 1.

Figure 1.  Hypoglycemic events (episodes/patient/6 months) during the 6-month post-index period.

Discussion

The current study of frequency of hypoglycemia in 784 (out of the original 2082) patients using concomitant exenatide and long-acting insulin is the largest reported observational study of a US managed-care patient population receiving this treatment combination in real-world settings. This study reinforces the findings of previous clinical studies reporting that concomitant use of exenatide and long-acting insulin was associated with a lower incidence of hypoglycemia1,2,5,6.

When metformin, thiazolidinediones, or exenatide are used either alone or in combination, hypoglycemia risk is lower than the risk with SUs and/or insulin, when used either alone or in combination with the above agents10. In this study, the use of SUs was lower in patients using insulin alone (38.7%) than in patients using exenatide and insulin (45.9%). Thus, SU use does not seem to explain the difference in hypoglycemia between the two study groups. The observational data presented here reinforce the concept that concomitant exenatide and insulin therapy was associated with a lower risk for hypoglycemia than continuing insulin therapy without exenatide. However, due to descriptive study design we cannot make any inferences about causal relationship between drug use and hypoglycemia. This study reports associations observed in the real world patient population using claims database.

Two retrospective analyses from selected clinical practice settings have found that use of exenatide combined with insulin was associated with further reductions in HbA1c and lower rates of hypoglycemia4,5. Additionally, two RCTs of exenatide added to titrated basal insulin have been completed. Buse et al.7 reported a greater reduction in HbA1c with no significant differences in hypoglycemia rates in patients treated with exenatide and titrated insulin glargine compared to titrated insulin glargine with placebo (1.4 vs 1.2 episodes/patient/year). In a smaller trial, Riddle et al.8 also showed no significant differences in number of hypoglycemia episodes and greater efficacy with insulin plus exenatide compared to insulin plus placebo.

Limitations of claims databases should be noted. Only hypoglycemia that required medical intervention (coded for hypoglycemia) was captured. This study was of short duration. A hypoglycemic claim suggests that an episode was of clinical impact sufficient to bill for services; however, hypoglycemic events may have been under-reported. Hypoglycemia frequency was recorded only for a 6-month period and time-to-event analyses were not performed. The study could not evaluate any association between insulin dose titration and hypoglycemia or examine other outcomes such as HbA1c levels, weight, and body mass index, due to lack of data availability. Hence, results may not be generalizable to various patient populations. We could not identify total duration of insulin therapy prior to initiating exenatide as it was not reported in the claims database. Another limitation of this study is the availability of NPH insulin over-the-counter without a prescription; therefore, if patients were using this medication there would have been no claim(s).

Conclusions

This study reinforces the clinical and RCT observations that concomitant use of exenatide BID and long-acting insulin is associated with a lower frequency of hypoglycemia in a large, diverse, type 2 diabetes patient population from real-world clinical practice settings. Further study is still necessary to determine any association between the lower risk of hypoglycemia seen with continuing exenatide BID with insulin vs discontinuing exenatide BID and possible adverse outcomes or mortality.

Transparency

Declaration of funding

This research was funded by Amylin Pharmaceuticals, Inc., San Diego, CA and by Eli Lilly and Company, Indianapolis, IN.

Declaration of financial/other relationships

Amy Blickensderfer, PharmD, is an employee of Amylin Pharmaceuticals, Inc., and owns stock in Amylin Pharmaceuticals. Ralph Quimbo is an employee of HealthCore, Inc. Rolin Wade, RPh, is an employee of Cerner Corporation. Byron J. Hoogwerf, MD, and Manjiri Pawaskar, PhD, are employed by Eli Lilly and Company and own stock in Eli Lilly and Company.

Acknowledgments

The authors thank Rebecca McCracken of i3 for her writing contributions to the preparation of the manuscript. The authors also thank Derek Misurski of GSK, US Health Outcomes and Rosalind Fabunmi formerly of Amylin Pharmaceuticals, Inc. for their contribution in the concept development and study design, and Likun Hou of HealthCore, Inc. for her contribution in the analysis.