Costs After Initiating Saxagliptin, Sulfonylurea, or Sitagliptin in Patients With T2DM

Controlling for baseline characteristics, saxagliptin patients had lower medical costs compared with sulfonylurea patients, and lower medical and total costs than sitagliptin patients.
Published Online: June 20, 2014
Anna Kaltenboeck, MA; Jasmina Ivanova, MS; Howard Birnbaum, PhD; Yana Yushkina, BA; Frances Schwiep, BA; Kelly Bell, PharmD; and Nina Thomas, MPH
Objectives: Compare healthcare resource use and costs in patients with type 2 diabetes mellitus (T2DM) treated with saxagliptin, sulfonylurea, or sitagliptin.

Study Design: Patients with T2DM who were treated with saxagliptin (N = 13,929), sulfonylurea (N = 117,756), or sitagliptin (N = 36,813) in 2010 or later were identified in a database of commercially insured beneficiaries. Patients were required to have no claims of index drug in the 6 months prior to treatment initiation, at least 6 months of continuous eligibility before and after initiation (baseline and study period, respectively), and be 18 years or older.

Methods: Demographic characteristics and comorbidities were evaluated in the baseline period. χ² and Wilcoxon rank sum tests assessed differences in all-cause and diabetes-related resource use rates and costs. Generalized linear model regression with a log link and gamma distribution estimated the treatment effect on study period costs, adjusting for baseline characteristics and costs.

Results: During the study period, saxagliptin patients experienced lower rates of all-cause hospitalization, emergency department visits, and other medical visits than sitagliptin and sulfonylurea patients (7.2% vs 10.6% and 12.1%; 14.1% vs 17.5% and 18.5%; 67.0% vs 70.4% and 70.3%, respectively; P <.001). All-cause and diabetes-related medical costs were lower for saxagliptin patients when compared with sitagliptin patients and sulfonylurea patients even after adjusting for baseline characteristics (all-cause: $5073 vs $5535 and $4984 vs $6002, respectively; diabetes related: $1149 vs. $1387 and $1143 vs $1558, respectively; all P <.001).

Conclusions: Patients initiating treatment with saxagliptin had lower resource use and medical costs in a 6-month follow-up period than patients initiating with sitagliptin or sulfonylurea.

Am J Pharm Benefits. 2014;6(3):e60-e69
Type 2 diabetes mellitus (T2DM) is a growing contributor to mortality and health services utilization and costs.1 In 2011, the prevalence of diabetes in United States was estimated at 25.8 million cases (8.3% of the population), of which 18.8 million were actually diagnosed.2 Both type 1 and type 2 diabetes add a substantial burden on healthcare resources. A 2008 study based on claims data estimated that mean annual healthcare costs for patients with T2DM were $8070, compared with $3853 for healthy control patients.3 A 2012 analysis found that patients with T2DM incurred $9677 in medical costs annually, and that each undiagnosed case of diabetes accounted for $2864 in annual costs.4 The SHIELD study, which focused on a representative sample of 22,001 adult individuals who had diabetes mellitus (type 1 or type 2) or cardiometabolic risk factors related to diabetes, reported a monthly average of $108 in out-of-pocket medical expenses for patients with T2DM, which presents a heavy financial burden and may prevent investment in other necessary medications.5 Such costs have large aggregate impacts: a 2010 study estimated that in 2008, over 7.7 million hospitalizations were incurred by US patients with diabetes. The same study found that the average cost for 1 such hospitalization was $10,937, compared with $8746 for patients who did not have diabetes.6 Authors of another analysis predicted that a 0.5% increase in the prevalence of the disease could result in an additional 1.1 million hospitalizations by 2015.7 Together, these factors contribute to a substantial economic burden. In the most recent estimate available, the national direct costs of diabetes were $174 billion in 2007, a sum which would be even greater if the cost of intangibles such as pain and suffering, care provided by unpaid caregivers, excess medical costs associated with undiagnosed diabetes, and diabetes-attributed healthcare expenditures were incorporated.8,9

T2DM is associated with serious comorbidities that impact mortality, quality of life, and healthcare resource use. Commonly associated conditions include heart disease, stroke, atherosclerosis, retinopathy, hypertension, nephropathy, diabetic foot problems, neurological complications, and dental diseases.2,10-14 Studies suggest an association between glycemic control, disease duration, and such complications of T2DM15-17; several studies have demonstrated substantial increases in healthcare resource use and costs in association with the comorbidities of diabetes.4,8,18,19 Even comorbidities that initially incur low costs can contribute significantly to overall cost over time due to disease progression. For example, microalbuminuria may result in progressive and costly renal complications and disease.20

Treatment options for the control of T2DM include various classes of antidiabetic agents, such as dipeptidyl peptidase-4 (DPP-4) inhibitors, glucagon-like peptide- 1 (GLP-1) analogues, sulfonylureas (SU), metformin, meglitinides, and thiazolidinediones (TZDs). Treatment guidelines for T2DM generally recommend metformin as first-line therapy, noting its extensive study history, high effectiveness, low cost, and comparatively mild side effect profile.21-24 For patients who do not achieve desired blood glucose control with metformin monotherapy, treatment augmentation with an additional agent from a different class—for example, a GLP-1 receptor agonist or DPP-4 inhibitor—should be considered.25 In patients who do not tolerate metformin, including those with renal impairment, guidelines recommend using an alternative agent.23,25,26 The CDC estimates that more than 35% of patients 20 years and older with diabetes suffer from chronic renal disease.27

DPP-4 inhibitors, including sitagliptin (approved by FDA in 2006), saxagliptin (approved by FDA in 2009), and linagliptin (approved by FDA in 2011), present a well-tolerated and effective option for patients with T2DM.28-30 Multiple doses are available for saxagliptin (2.5 mg and 5 mg) and sitagliptin (25 mg, 50 mg, and 100 mg), the lower doses of which are for patients with renal impairment.28,29 Linagliptin is approved for use at a single dose of 5 mg.30

While there is some literature on the cost effectiveness of saxagliptin in Europe, and an adaptation of such a model for the United States, little is published on real-world resource use and costs in the United States due to its recent entry to the market.31-34 This retrospective claims data analysis estimates the healthcare resource use and cost outcomes in the 6 months following treatment initiation among patients with T2DM who (1) initiated therapy with saxagliptin, compared with those who initiated an SU; and 2) initiated therapy with saxagliptin compared with those who initiated sitagliptin.


The Truven MarketScan administrative claims database, spanning from Q1 2009 to Q2 2011, was used for this study. This deidentified database included health plan enrollment and demographic information, as well as medical and pharmacy claims for all beneficiaries. A total of 4,778,738 patients with a diagnosis of T2DM (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] codes 250.x0 or 250.x2) in 2009 or later were identified. Patients were required to: (1) have initiated treatment with saxagliptin, an SU, or sitagliptin in 2010 or later; (2) have made no index drug claims in the 6 months prior to initiation; (3) have at least 6 months of continuous eligibility prior to and following initiation (baseline and study period, respectively); and (4) be 18 years or older at initiation (index date).


Baseline characteristics including demographic characteristics, Charlson Comorbidity Index (CCI)35 comorbidities, and healthcare resource use and costs were described for each cohort. Healthcare resource use and costs were also estimated during the 6-month study period.

All-cause resource use was calculated based on all medical and pharmacy claims. Diabetes-related resource use was calculated based on claims having primary or secondary ICD-9-CM diagnosis codes 250.x0 or 250.x2, along with antidiabetic prescription drug claims. Resource use type was classified as inpatient stay, emergency department (ED) visit, outpatient (ambulatory) visit, other visit, and the use of prescription medications indicated for the treatment of T2DM.

Costs were based on the third-party payer paid amount recorded on claims during the baseline and study periods. All-cause medical costs included all medical claims while diabetes-related medical costs included those having primary or secondary ICD-9-CM diagnosis codes 250.x0 or 250.x2. All-cause prescription drug costs included all prescription drug claims, while diabetes-related prescription drug costs were based on prescription drug claims for antidiabetic agents. Total costs were the sum of medical and prescription drug costs. All estimates were inflated to 2011 dollars using the Consumer Price Index for Medical Care.36

Results were compared descriptively during the study period between the saxagliptin and the SU and sitagliptin cohorts, respectively. χ ² tests were used to assess differences in resources use rates; Wilcoxon rank sum tests were used to calculate differences in number of visits and in costs.

Risk-adjusted 6-month study period medical and total costs were estimated using generalized linear models with a log link and gamma distribution for the error term controlling for effects of index therapy, patient demographics, geographic region, and baseline comorbidities with a prevalence of at least 5% in either cohort, and log of baseline costs. SAS version 9.3 (SAS Institute, Inc, Cary, North Carolina) was used for all analyses (statistical significance defined as a P <.05).


Baseline Characteristics

Following the application of sample selection criteria, 13,929 patients were selected into the saxagliptin cohort, 117,756 were selected into the SU cohort, and 36,813 were selected into the sitagliptin cohort (eAppendix).

Saxagliptin Versus SU

On average, saxagliptin patients were approximately 1 year younger than SU patients (58 years vs 59 years), and slightly more likely to be female (47.0% vs 45.1% male). Significantly more saxagliptin patients were in the South Atlantic census region of the United States (52.5% vs 40.1%). Saxagliptin patients had a significantly lower mean CCI (0.48 vs 0.57) and significantly higher rates of dyslipidemia and hypertension (44.8% vs 36.1% and 51.8% vs 45.7%, respectively) (Table 1). Use of prescription antidiabetic drug classes, excluding DPP-4 inhibitors, was greater among saxagliptin patients (P <.001).

Saxagliptin patients were significantly less likely than SU patients to have all-cause inpatient hospitalizations or ED visits during the baseline (6.6% vs 12.0% and 14.9% vs 20.6%, respectively) and more likely to have all-cause outpatient visits (97.4% vs 90.8%). Total all-cause costs were $1511 lower for saxagliptin patients than for SU patients ($5939 vs $7450). Total diabetes-related costs were also significantly lower among saxagliptin patients ($1476 and $1769, respectively; P <.001) (Table 2).

Saxagliptin Versus Sitagliptin

Diabetes Compendium
Oncology Compendium