Advertisement

Real-World Effectiveness of Once-Weekly Semaglutide From a US Commercially Insured and Medicare Advantage Population

Open AccessPublished:March 28, 2021DOI:https://doi.org/10.1016/j.clinthera.2021.03.003

      Abstract

      Purpose

      Patients managing type 2 diabetes mellitus (T2DM) often require combination therapy to meet their blood glucose control targets. With limited real-world evidence focused on the use of glucagon-like peptide 1 receptor agonist (GLP-1RA) therapies, the objective of this study was to describe the association between semaglutide once weekly (OW) initiation and changes in hemoglobin A1c (A1c) levels.

      Methods

      This retrospective, descriptive cohort study used the HealthCore Integrated Research Environment (HIRE) to examine commercially insured and Medicare Advantage patients who had T2DM while taking semaglutide OW from December 1, 2017, to April 30, 2019. The first semaglutide OW prescription fill was defined as the study index date. Changes in mean A1c levels and A1c target attainment were evaluated among an intention-to-treat (ITT) population (overall group). Furthermore, a persistent population (PP) analysis on the same outcomes was performed that was limited to ITT patients who were observed to continue to use semaglutide OW at the time of the postindex A1c measurement. In addition, these outcomes were explored in patients stratified based on prior use of GLP-1RA therapy (experienced vs naive) and baseline A1c values >9% (75 mmol/mol).

      Findings

      A total of 1888 patients were identified in the overall ITT group. The mean change in the overall ITT group between preindex and postindex A1c values was −0.9% percentage points (−9.8 mmol/mol) (mean preindex A1c, 8.2% [66.1 mmol/mol]) and −1.1 percentage points (−12.0 mmol/mol) (mean preindex A1c, 8.2% [66.1 mmol/mol]) in the PP subgroup (all P < 0.001). Among the subgroup of patients with a baseline A1c value >9% (75 mmol/mol), percentage point changes in A1c were −2.2 (−24.0 mmol/mol) and −2.4 (−26.2 mmol/mol) (all P < 0.001). When accounting for prior GLP-1RA use, the GLP-1RA–naive stratum had double the mean reduction in A1c compared with the GLP-1RA–experienced stratum (−1.2 [−13.1 mmol/mol] vs −0.6 [−6.6 mmol/mol] percentage points).

      Implications

      Semaglutide OW is associated with clinically and statistically significant A1C reduction and increases in reaching A1 targets using real-world data, even in GLP-1RA–experienced patients, despite more frequent use of insulin or sodium glucose transport protein 2 inhibitors in this group. Target A1c attainment significantly increased overall and within all subgroups and strata, with approximately half of all patients attaining an A1c value <7% (53 mmol/mol) and three-quarters attaining an A1c value <8% (64 mmol/mol).

      Keywords

      Introduction

      Diabetes is a widespread, complex, chronic metabolic disorder. Every year 1.5 million individuals 18 years or older are newly diagnosed with diabetes in the United States, with 90% to 95% having type 2 diabetes mellitus (T2DM).
      Centers for Disease Control and Prevention
      National Diabetes Statistics Report 2020: Estimates of Diabetes and Its Burden in the United States.
      The prevalence of T2DM in the United States ranges from 8.5% to 13.0% or an estimated 34.1 million individuals.
      Centers for Disease Control and Prevention
      National Diabetes Statistics Report 2020: Estimates of Diabetes and Its Burden in the United States.
      • Xu G
      • Liu B
      • Sun Y
      • et al.
      Prevalence of diagnosed type 1 and type 2 diabetes among US adults in 2016 and 2017: population based study.

      National Health and Nutrition Examination Survey. Diabetes prevalence and glycemic control among adults aged 20 and over, by sex, age, and race and Hispanic origin. 2018. Accessed April 2020. URL: https://www.cdc.gov/nchs/data/hus/2018/014.pdf.

      Control of blood glucose is a key aspect of managing T2DM because lowering elevated blood glucose levels has been associated with reduced risk of microvascular complications associated with T2DM.
      Glycemic Targets: Standards of Medical Care in Diabetes—2019.
      Pharmacologic Approaches to Glycemic Treatment: Standards of Medical Care in Diabetes—2020.
      • Qaseem A
      • Wilt TJ
      • Kansagara D
      • Horwitch C
      • Barry MJ
      • Forciea MA.
      Hemoglobin A 1c Targets for Glycemic Control With Pharmacologic Therapy for Nonpregnant Adults With Type 2 Diabetes Mellitus: A Guidance Statement Update From the American College of Physicians.
      In addition, recent studies have found beneficial outcomes in patients with T2DM and cardiovascular disease, heart failure, or chronic kidney disease for certain T2DM medications.
      • Marso SP
      • Daniels GH
      • Brown-Frandsen K
      LEADER Steering Committee; LEADER Trial Investigators. Liraglutide and cardiovascular outcomes in type 2 diabetes.
      • Gerstein HC
      • Colhoun HM
      • Gagenais GR
      • et al.
      Dulaglutide and Cardiovascular Outcomes in Type 2 Diabetes (REWIND): a Double-Blind, Randomised Placebo-Controlled Trial.
      • Zinman B
      • Wanner C
      • Lachin JM
      • et al.
      EMPAREG OUTCOME Investigators. Empagliflozin, cardiovascular outcomes, and mortality in type 2 diabetes.
      • Neal B
      • Perkovic V
      • Mahaffey KW
      • et al.
      CANVAS Program Collaborative Group Canagliflozin and cardiovascular and renal events in type 2 diabetes.
      • Marso SP
      • Bain SC
      • Consoli A
      • et al.
      Semaglutide and Cardiovascular Outcomes in Patients with Type 2 Diabetes.
      The American Diabetes Association (ADA) recommends metformin for initial pharmacologic therapy of T2DM in conjunction with lifestyle modifications.
      Pharmacologic Approaches to Glycemic Treatment: Standards of Medical Care in Diabetes—2020.
      Because of the independent variation in glucose levels and the progressive nature of the disease, patients managing T2DM often require combination therapy to meet their blood glucose control targets if glycosylated hemoglobin (A1c) levels remain above an individual's target as well as to maximize reduction of vascular complications. The 2020 standards of medical care in diabetes from the ADA state it is preferable to initiate a glucagon-like peptide 1 receptor agonist (GLP-1RA) in individuals with T2DM and atherosclerotic vascular disease independent of A1C level or individualized A1c target.
      Pharmacologic Approaches to Glycemic Treatment: Standards of Medical Care in Diabetes—2020.
      In addition, initiation of a GLP-1RA should be considered in individuals with T2DM and heart failure or chronic kidney disease who cannot tolerate or have contraindications to sodium glucose transport protein 2 (SGLT2) inhibitors.
      Pharmacologic Approaches to Glycemic Treatment: Standards of Medical Care in Diabetes—2020.
      Semaglutide once weekly (OW) is a subcutaneous GLP-1 analogue first approved in the United States in December 2017, with an extended t½ of approximately 1 week.
      • Lau J
      • Bloch P
      • Schäffer L
      • et al.
      Discovery of the Once-Weekly Glucagon-Like Peptide-1 (GLP-1) Analogue Semaglutide.
      ,

      Hai MTT. Center for Drug Evaluation and Research Approval Package for: Application Number: 209637Orig1s000. Food and Drug Administration. Silver Spring;2017. Accessed April 2020. URL: https://www.accessdata.fda.gov/drugsatfda_docs/nda/2017/209637Orig1s000Approv.pdf.

      A number of Phase III clinical trials were performed with semaglutide OW as part of the Semaglutide Unabated Sustainability in Treatment of T2DM (SUSTAIN) clinical trial program.
      • Marso SP
      • Bain SC
      • Consoli A
      • et al.
      Semaglutide and Cardiovascular Outcomes in Patients with Type 2 Diabetes.
      ,
      • Pratley RE
      • Aroda VR
      • Lingvay I
      • et al.
      Semaglutide versus dulaglutide once weekly in patients with type 2 diabetes (SUSTAIN 7): a randomised, open-label, phase 3b trial.
      • Ahrén B
      • Masmiquel L
      • Kumar H
      • et al.
      Efficacy and safety of once-weekly semaglutide versus once-daily sitagliptin as an add-on to metformin, thiazolidinediones, or both, in patients with type 2 diabetes (SUSTAIN 2): a 56-week, double-blind, phase 3a, randomised trial.
      • Aroda VR
      • Bain SC
      • Cariou B
      • et al.
      Efficacy and safety of once-weekly semaglutide versus once-daily insulin glargine as add-on to metformin (with or without sulfonylureas) in insulin-naive patients with type 2 diabetes (SUSTAIN 4): a randomised, open-label, parallel-group, multicentre, mul.
      • Lingvay I
      • Catarig A-M
      • Frias JP
      • et al.
      Efficacy and safety of once-weekly semaglutide versus daily canagliflozin as add-on to metformin in patients with type 2 diabetes (SUSTAIN 8): a double-blind, phase 3b, randomised controlled trial.
      These studies were conducted within a variety of populations, against differing comparators and with varying timeframes and end points. For example, SUSTAIN-7 found that semaglutide OW was superior to dulaglutide in improving glycemic control.
      • Pratley RE
      • Aroda VR
      • Lingvay I
      • et al.
      Semaglutide versus dulaglutide once weekly in patients with type 2 diabetes (SUSTAIN 7): a randomised, open-label, phase 3b trial.
      In the trial to evaluate cardiovascular and other long-term outcomes with semaglutide in patients with T2DM (SUSTAIN-6), semaglutide OW users had a 26% lower risk of the primary composite outcome of death from cardiovascular causes, nonfatal myocardial infarction, or nonfatal stroke than those receiving placebo (P < 0.001).
      • Marso SP
      • Bain SC
      • Consoli A
      • et al.
      Semaglutide and Cardiovascular Outcomes in Patients with Type 2 Diabetes.
      To complement the data generated through the SUSTAIN clinical trial program, there is a need for real-world evidence on GLP-1RAs generally and semaglutide OW in particular. Evidence on the real-world effectiveness of semaglutide OW is scarce.
      • Brown RE
      • Liu AR
      • Mahbubani R
      • Aronson R.
      Semaglutide in patients with Type 2 diabetes: Real-world analysis in the Canadian LMC Diabetes Registry: The SPARE study.
      ,
      • Visaria J
      • Dang-Tan T
      • Petraro PV
      • Nepal BK
      • Willey VJ.
      Real-World Effectiveness of Semaglutide in Early Users from a U.S. Commercially Insured (CI) and Medicare Advantage (MA) Population.
      An analysis conducted on users of semaglutide OW would provide real-world evidence on the effectiveness of semaglutide OW and valuable insights to practitioners, payers, and patients. Therefore, the purposes of this study were to describe a cohort of US commercially insured and Medicare Advantage patients initiating semaglutide OW therapy and to evaluate the real-world effectiveness associated with its use. This study aimed to update results from the first real-world analysis of early semaglutide OW users in the United States
      • Visaria J
      • Dang-Tan T
      • Petraro PV
      • Nepal BK
      • Willey VJ.
      Real-World Effectiveness of Semaglutide in Early Users from a U.S. Commercially Insured (CI) and Medicare Advantage (MA) Population.
      by following up a larger cohort of patients initiating semaglutide OW therapy during the first 1½ years of market availability during a longer follow-up period.

      Methods

       Study Design

      This study used a claims-based, observational, retrospective cohort design to follow up patients initiating semaglutide OW therapy between December 1, 2017, and April 30, 2019 (study intake period).

       Data Source

      The data source for this study was integrated medical and pharmacy administrative claims data and clinical laboratory test results from the HealthCore Integrated Research Environment (HIRESM) for the period December 1, 2016, to July 31, 2019. The HIRE contains enrollment information and medical and pharmacy administrative claims for approximately 48.9 million enrollees in a large US private health plan, with approximately 40% having integrated medical laboratory test results. The researchers accessed and examined a limited data set, described in the Health Insurance Portability and Accountability Act Privacy Rule, because the objectives of this observational study did not require patient identifiers. All study materials were handled in strict compliance with applicable federal rules, and the privacy and security of all personal health information were safeguarded throughout the study.

       Study Population

      Individuals with commercial or Medicare Advantage insurance with ≥1 pharmacy claim for semaglutide OW between December 1, 2017, and April 30, 2019, were included. The first observed semaglutide OW prescription fill date during the study intake period was designated the study index date (Figure 1). Of these, patients 18 years or older on the index date with ≥12-month continuous enrollment before the study index date were retained. Patients were also required to have ≥1 claim with a diagnosis for T2DM (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] codes 250.x0 and 250.x2; International Classification of Diseases, Tenth Revision, Clinical Modification [ICD-10-CM] codes E11.% and E13.%) during the 12 months before the index date and ≥1 valid A1c result in the 90-day period before the index date. Finally, patients were also required to have ≥1 valid A1c measurement at least 90 days after the index date. This intention-to-treat (ITT) group consisted of all patients who initiated semaglutide OW therapy, irrespective of whether they were taking semaglutide OW as of the date of the postindex A1c measurement.
      Figure 1
      Figure 1Study schema. A1c = glycosylated hemoglobin; OW = once weekly; PP = persistent population; SC = subcutaneous.

       A1c Measurement

      Representative preindex and postindex A1c values were selected for each patient. The preindex A1c was value defined as a valid A1c laboratory test result during the 90 days before the index date. The postindex A1c value was defined as a valid A1c laboratory test between 90 days after the index date and the end of either the continuous postindex follow-up or July 31, 2019, whichever of these was earlier (Figure 1). If patients had multiple valid preindex or postindex A1c values, the one closest to the index date in the preindex period and the one farthest from the index date (most recent result) in the postindex period were considered for analysis. If multiple A1c values were reported on this date, then the highest value was considered as the representative preindex or postindex A1c value. To assess the robustness of the results to the selection of the postindex A1c values, sensitivity analyses were performed by using the earliest postindex A1c value and the lowest postindex A1C value.
      The preindex and postindex A1c values were analyzed as continuous variables and dichotomized at target A1c levels of A1c <7% (53 mmol/mol) and <8% (64 mmol/mol).

       PP Subgroup

      A persistent population (PP) analysis was also performed that was limited to the subgroup of ITT patients who were observed to continue to use semaglutide OW at the time of the postindex A1c measurement. Persistent use of semaglutide OW was defined as fulfillment of 1 of the following 2 criteria: the overlap of a semaglutide OW fill with the postindex A1c measurement date (date of semaglutide OW fill before the post-index A1c measurement date before the depletion date of semaglutide OW fill based on days’ supply) or the depletion of a semaglutide OW fill within a reasonable refill delay time before the postindex A1c measurement date, where reasonably delayed was defined as 1.5 times the days’ supply of the semaglutide OW fill. Among patients in the PP subgroup, continuous preindex and postindex A1c values as well as categorical target goal attainment of preindex and postindex A1c values were measured identically to the ITT cohort.

       Strata of Interest

      Both the ITT group and PP subgroup were further stratified based on preindex non-semaglutide GLP-1RA use and were defined as GLP-1RA–naive patients (no claims for GLP-1RA during the 12-month preindex period) and GLP-1RA–experienced patients (≥1 claims for GLP-1RA during the 12-month preindex period). Both the ITT and PP patients with a preindex A1c value >9% (75 mmol/mol) were also evaluated as separate strata.

       Patient Characteristics

      Patient characteristics, including age on index date, sex, region of residence, and prescriber specialty, associated with the index semaglutide OW claim were described. Baseline clinical characteristics that were reported included the number and proportion of patients with ≥1 claim for selected antidiabetic medications during the 12-month continuous preindex eligibility period, the number and proportion of patients with ≥1 claim for selected comorbidities, and other selected medications during the 12-month continuous preindex eligibility period.

       Statistical Analysis

      Descriptive statistics, including means (SD) and relative frequencies for continuous and categorical data, respectively, were reported for patient characteristics in the ITT cohort, in the overall group, and by GLP-1RA strata. Mean postindex and preindex A1c values were compared using paired t tests. These comparisons were made within the ITT cohort and PP subgroup, as well as within the GLP-1RA–naive and GLP-1RA–experienced strata of ITT and PP patients. Similar comparisons of mean preindex and postindex A1c were made among patients in the preindex A1c >9% (75 mmol/mol) stratum of the ITT cohort and PP subgroup. The proportion of patients who achieved A1c target goals of <7% (53 mmol/mol) and <8% (64 mmol/mol) were each compared using the McNemar test. Comparisons were between the preindex and postindex period within the ITT group as well as the preindex A1c >9% (75 mmol/mol) stratum of the ITT group. The level of statistical significance was set to α = 0.05 a priori. All statistical analyses were performed using SAS Enterprise Guide, version 5.1 (SAS Institute Inc, Cary, North Carolina).

      Results

      A total of 4990 adults who initiated semaglutide OW therapy during the intake period had ≥1 T2DM diagnosis within ≥12 months of continuous preindex eligibility and ≥1 A1c laboratory test result in the 90 days before semaglutide OW therapy initiation. Of these adults, 2641 (52.9%) also had 1 A1c test result at any time during postindex continuous eligibility, 1888 (71.5%) of whom had ≥1 A1c test result ≥90 days after the date of semaglutide OW therapy initiation and comprised the ITT cohort. More than half of all ITT patients (n = 1084 [57.4%]) were in possession of semaglutide OW at or near the time of their postindex A1c measurement and comprised the PP subgroup.
      The ITT group was composed of 1092 GLP-1RA–naive patients (57.8% of all ITT patients), 796 GLP-1RA–experienced patients (42.2% of all ITT patients), and 506 patients with a baseline A1c value >9% (75 mmol/mol) (26.8% of all ITT patients). The PP subgroup was composed of 595 GLP-1RA–naive patients (54.9% of all PP patients), 489 GLP-1RA–experienced patients (45.1% of all PP patients), and 279 patients with a baseline A1c value >9% (75 mmol/mol) (25.7% of all PP patients).

       Demographic and Clinical Characteristics

      The mean (SD) age of the overall ITT group was 54.3 (9.1) years, with the GLP-1RA–naive and GLP-1RA–experienced strata at 53.9 (9.3) years and 54.8 (8.8) years, respectively (Table 1). Approximately half of the overall ITT group's patients were female (50.5%), with similar proportions of female patients in the GLP-1RA–naive and GLP-1RA–experienced strata (51.4% and 49.4%, respectively). The GLP-1RA–experienced stratum had a higher proportion of endocrinology prescribers associated with the index semaglutide OW claim (41.0%) compared with the GLP-1RA–naive (29.6%) subgroup. The GLP-1RA–naive subgroup had a higher proportion of primary care physician prescribers associated with the index semaglutide OW claim (40.8%) compared with the GLP-1RA–experienced (32.3%) subgroup.
      Table 1Baseline demographic and clinical characteristics for the ITT group.
      Data are presented as number (percentage) of patients unless otherwise indicated.
      CharacteristicAll ITT patients (N = 1888)GLP-1RA–naive ITT patients
      Patients with no claims for a GLP-1RA during the 12-month preindex period (GLP-1RA–naive subset).
      (n = 1092)
      GLP-1RA–experienced ITT patients
      Patients with claims for a GLP-1RA during the 12-month preindex period (GLP-1RA–experienced subset).
      (n = 796)
      Demographic characteristics
      Age at index, mean (SD, median), y54.3 (9.1, 55)53.9 (9.3, 55)54.8 (8.8, 56)
      Age group, y
       18-39118 (6.3)73 (6.7)45 (5.7)
       40-641601 (84.8)926 (84.8)675 (84.8)
       65-74151 (8)83 (7.6)68 (8.5)
       ≥7518 (1)10 (0.9)8 (1)
      Female954 (50.5)561 (51.4)393 (49.4)
      Geographic region of residence of patient
       Northeast253 (13.4)141 (12.9)112 (14.1)
       Midwest226 (12)150 (13.7)76 (9.5)
       South1109 (58.7)633 (58)476 (59.8)
       West277 (14.7)160 (14.7)117 (14.7)
       Missing or unknown23 (1.2)8 (0.7)15 (1.9)
      Specialty of the prescribing physician specialty associated with index semaglutide OW claim
       Endocrinology649 (34.4)323 (29.6)326 (41.0)
       Nonphysician clinician (eg, nurse practitioner or physician assistant)409 (21.7)239 (21.9)170 (21.4)
       Primary care physician702 (37.2)445 (40.8)257 (32.3)
       Other
      Addiction medicine, cardiology, emergency medicine, hematology/oncology, infectious disease, nephrology, ophthalmology, physical medicine and rehabilitation, preventive medicine, radiology, and surgery.
      99 (5.2)66 (6.0)33 (4.1)
       Missing or unknown29 (1.5)19 (1.7)10 (1.3)
      Baseline (1 year before index date) CCI comorbidity profile
      Patients with ≥1 claim for diagnosis of interest
       Dyslipidemia1623 (86)914 (83.7)709 (89.1)
       Hypertension1536 (81.4)881 (80.7)655 (82.3)
       Obesity751 (39.8)413 (37.8)338 (42.5)
       Sleep apnea459 (24.3)256 (23.4)203 (25.5)
       Neuropathy343 (18.2)191 (17.5)152 (19.1)
       Retinopathy295 (15.6)145 (13.3)150 (18.8)
       Ischemic heart disease225 (11.9)124 (11.4)101 (12.7)
       Nephropathy205 (10.9)100 (9.2)105 (13.2)
       Cerebrovascular disease80 (4.2)46 (4.2)34 (4.3)
       Congestive heart failure71 (3.8)43 (3.9)28 (3.5)
       Myocardial infarction29 (1.5)16 (1.5)13 (1.6)
       Peripheral vascular disease6 (0.3)6 (0.5)0 (0)
      Baseline (1 year before index date) medication use profile
      Patients with ≥1 claim for medication of interest
      Antidiabetic medications (composite, insulin, or noninsulin)1,888 (100)1,092 (100)796 (100)
       Insulin (composite, any insulin)709 (37.6)363 (33.2)346 (43.5)
       Basal insulin (composite, any basal insulin)660 (35)341 (31.2)319 (40.1)
       Nonbasal insulin (composite, any non-basal insulin)312 (16.5)159 (14.6)153 (19.2)
       Noninsulin antidiabetic medication classes (composite, any antidiabetic)1888 (100)1092 (100)796 (100)
       Biguanides (metformin)1493 (79.1)858 (78.6)635 (79.8)
       Thiazolidinediones153 (8.1)70 (6.4)83 (10.4)
       Sulfonylureas507 (26.9)320 (29.3)187 (23.5)
       DPP-IV inhibitors440 (23.3)332 (30.4)108 (13.6)
       SGLT2 inhibitors766 (40.6)380 (34.8)386 (48.5)
       Meglitinide analogs23 (1.2)10 (0.9)13 (1.6)
       α-Glucosidase inhibitors8 (0.4)5 (0.5)3 (0.4)
       Amylin analogs1 (0.1)0 (0)1 (0.1)
       GLP-1RA, excluding semaglutide OW796 (42.2)0 (0)796 (100)
      Nonantidiabetic medications (analyzed at medication class level)
       Antihypertensive medication classes (any)1532 (81.1)861 (78.8)671 (84.3)
       ACE inhibitors774 (41.0)437 (40.0)337 (42.3)
       ARBs660 (35.0)370 (33.9)290 (36.4)
       β-Blockers547 (29.0)304 (27.8)243 (30.5)
       Thiazide diuretics582 (30.8)321 (29.4)261 (32.8)
       Calcium channel blockers398 (21.1)237 (21.7)161 (20.2)
       Other antihypertensives69 (3.7)40 (3.7)29 (3.6)
       Lipid-lowering therapy1430 (75.7)781 (71.5)649 (81.5)
      Abbreviations: ACE = angiotensin-converting enzyme; ARB = angiotensin receptor blocker; CCI = Charlson Comorbidity Index; DPP-IV = dipeptidyl peptidase; GLP-1RA = glucagon-like peptide 1 receptor agonist; OW = once weekly; SGLT2 = sodium/glucose cotransporter 2.
      low asterisk Data are presented as number (percentage) of patients unless otherwise indicated.
      Patients with no claims for a GLP-1RA during the 12-month preindex period (GLP-1RA–naive subset).
      Patients with claims for a GLP-1RA during the 12-month preindex period (GLP-1RA–experienced subset).
      § Addiction medicine, cardiology, emergency medicine, hematology/oncology, infectious disease, nephrology, ophthalmology, physical medicine and rehabilitation, preventive medicine, radiology, and surgery.
      The top 3 baseline comorbidities in the overall ITT group during the 1-year baseline before the index date were dyslipidemia (86.0%), hypertension (81.4%), and obesity (39.8%). Nearly a quarter of ITT patients also had sleep apnea (24.3%). Within the overall ITT group, the most frequently used preindex noninsulin antidiabetic drug classes were biguanides (79.1%), followed by non-semaglutide GLP-1RA analogues (42.2%) and SGLT-2 inhibitors (40.6%). Insulin was used by 37.6% of patients, which was predominantly basal insulin (35.0%). Patient characteristics in the PP subgroup were qualitatively similar to the ITT group, and are presented in the Supplemental Table.
      On the index date, 30.4% of the ITT patients had a claim for the 1-mg dose of semaglutide OW; approximately a quarter (23.2%) of the GLP-1RA–naive patients and 40.3% of GLP-1RA–experienced patients received the 1-mg dose of semaglutide OW. Furthermore, on the index date, 31.3% of the PP patients had a claim for the 1-mg dose of semaglutide OW; approximately a quarter (22.4%) of the GLP-1RA–naive patients and 42.1% of the GLP-1RA–experienced patients received doses of 1 mg of semaglutide OW.
      The median time leading up to semaglutide OW initiation after the preindex A1c measurement date (maximum, 90 days) in the overall ITT group was 26 days (interquartile range [IQR], 8–52 days). Within the PP subgroup, the median time leading up to semaglutide OW therapy initiation after the preindex A1c measurement date (maximum, 90 days) was 27 days (IQR, 9–51 days).
      The median time between semaglutide OW therapy initiation and the postindex A1c measurement date (minimum, 90 days) was 197 days (IQR, 135–284 days). Among the PP subgroup, the median time between semaglutide OW initiation and post-index A1C measurement date (minimum 90 days) was 218 days (162-298). All strata within the overall ITT group and the PP subgroup, respectively, had qualitatively similar time frames for preindex and postindex A1c measurement. The median duration of continuous postindex follow-up for the overall ITT group was 280 days (IQR, 202–370 days) (minimum 90 days). All strata within the overall ITT group and the PP subgroup had qualitatively similar postindex follow-up.

       Change in Mean A1c Values

      The mean (SD) percentage point changes between the preindex and postindex A1c values were −0.9 (16) (−9.8 mmol/mol; 95% CI, −0.96 to −0.82; mean preindex A1c, 8.2% [66.1 mmol/mol]) in the overall ITT group and −1.1 (1.6) (−12.0 mmol/mol; 95% CI, −1.15 to −0.96; mean preindex A1c, 8.2% [66.1 mmol/mol]) in the PP subgroup. Within the GLP-1RA–naive stratum, the mean percentage point change between the preindex and postindex A1c values was −1.2 (1.7) (−13.1 mmol/mol; 95% CI, −1.26 to −1.06; mean preindex A1c, 8.5% [69.4 mmol/mol]) in the overall ITT group and −1.3 (1.7) (−14.2 mmol/mol; 95% CI, −1.53 to −1.26; mean preindex A1c, 8.4% [68.3 mmol/mol]) in the PP subgroup. Within the GLP-1RA–experienced stratum, the mean (SD) percentage point change between the preindex and postindex A1c values was −0.6 (1.4) (−6.6 mmol/mol; 95% CI, −0.61 to −0.42; mean preindex A1c, 7.9% [62.8 mmol/mol]) in the overall ITT group and −0.6 (1.3) (−6.6 mmol/mol; 95% CI, −0.76 to −0.53; mean preindex A1c, 7.8% [61.8 mmol/mol]) in the PP subgroup. In all these cases, the change in A1c was statistically significant (all P < 0.001) (Figure 2).
      Figure 2
      Figure 2Mean change in glycosylated hemoglobin (A1c) level for intention-to-treat group and persistent population subgroup, overall and within glucagon-like peptide 1 receptor agonist (GLP-1RA)–naive and GLP-1RA–experienced strata. Data presented above paired bars represent differences in mean A1c values with 95% CIs. Index date is defined as the first semaglutide once-weekly fill date. A and B represent the same results for A1c in "%" and "mmol/mol" units, respectively.
      Larger reductions in A1c values were observed within the stratum of patients with a baseline A1c value >9% (75 mmol/mol). The mean (SD) percentage point change between preindex and postindex A1c value was −2.2 (2.0) (−24.0 mmol/mol; 95% CI, −2.40 to −2.05; mean preindex A1c, 10.5% [91.3 mmol/mol]) in the overall ITT group and −2.4 (2.0) (−26.2 mmol/mol; 95% CI, −2.70 to −2.24; mean pre-index A1C 10.4% (90.2 mmol/mol)]in the PP subgroup. Within the GLP-1RA–naive stratum with baseline A1c values >9% (75 mmol/mol), the mean (SD) percentage point change between preindex and postindex A1c values was −2.5 (2.0) (−27.3 mmol/mol; 95% CI, −2.66 to −2.24; mean preindex A1c, 10.6% [92.4 mmol/mol]) in the overall ITT group and −2.8 (1.9) (−30.6 mmol/mol; 95% CI, −3.05 to −2.49; mean preindex A1c, 10.5% [91.3 mmol/mol]) in the PP subgroup. Within the GLP-1RA–experienced stratum with baseline A1c values >9% (75 mmol/mol), the mean (SD) percentage point change between the preindex and postindex A1c values was −1.8 (1.9) (−19.7 mmol/mol; 95% CI, −2.04 to −1.45; mean preindex A1c, 10.3% [89.1 mmol/mol]) in the overall ITT group and −1.9 (1.8) (−20.8 mmol/mol; 95% CI, −2.25 to −1.49; mean preindex A1c, 10.3% [89.1 mmol/mol]) in the PP subgroup. In all these cases, the change in A1c value was statistically significant (all P < 0.001) (Figure 3).
      Figure 3
      Figure 3Mean change in glycosylated hemoglobin (A1c) level for intention-to-treat group with baseline A1c values >9% and persistent population subgroup with baseline A1c >9%, overall and within glucagon-like peptide 1 receptor agonist (GLP-1RA)–naive and GLP-1RA–experienced strata. Data presented above paired bars represent differences in mean A1c values with 95% CIs. Index date is defined as the first semaglutide once-weekly fill date. A and B represent the same results for A1c in "%" and "mmol/mol" units, respectively.
      Sensitivity analyses found that A1c change results were not meaningfully different based on the algorithm for identifying representative postindex values. Within the overall ITT group, using the earliest as opposed to the latest postindex A1c value was associated with a similar, statistically significant mean (SD) percentage point change in A1c of −0.9 (1.6) (−9.8 mmol/mol; mean baseline A1c, 8.2% [66.1 mmol/mol]). In another sensitivity analysis, the lowest A1c instead of the latest A1c value was used as the representative postindex A1c value. Within the overall ITT group, using this lowest A1c value was associated with a similar if not slightly larger, statistically significant mean (SD) percentage point change in A1c of −1.0 (1.6) (−10.9 mmol/mol; mean baseline A1c, 8.2% [66.1 mmol/mol]).

       A1c Target Attainment

      The absolute increase in the proportion of ITT patients who attained an A1c target of <7% (53 mmol/mol) between the preindex and postindex period was 25.3 percentage points (23.3% in the preindex period to 48.6% in the postindex period). Among ITT patients with baseline A1c values >9% (75 mmol/mol), 26.9% of patients attained an A1c target of <7% (53 mmol/mol) in the postindex period (0% in the preindex period). Similar results were observed at a target threshold of A1c <8% (64 mmol/mol). The absolute increase in the proportion of ITT patients who attained an A1c target of <8% (64 mmol/mol) between the preindex and postindex period was 23.4 percentage points (50.6% in the preindex period to 74.0% in the postindex period). Among ITT patients with baseline A1c values >9% (75 mmol/mol), 49.2% attained an A1c target of <8% (64 mmol/mol) in the postindex period. In all these cases, the increase in the proportion of patients achieving target A1c goals was statistically significant (all P < 0.001) (Figure 4).
      Figure 4
      Figure 4Target glycosylated hemoglobin A1c level attainment for intention-to-treat group and intention-to-treat subgroup with baseline A1c values >9%. Index date is defined as the first semaglutide once-weekly fill date.
      Absolute increases in A1c target attainment were also observed in the GLP-1RA–naive stratum of the overall ITT group at the A1c target of <7% (53 mmol/mol) of 27.2 percentage points (23.3% in the preindex period to 50.5% in the postindex period, P < 0.001) and at the A1c target of <8% (64 mmol/mol) of 31.1 percentage points (44.2% in the preindex period to 75.3% in the postindex period, P < 0.001). Increases in A1c target attainment were also observed in the GLP-1RA–experienced stratum of the overall ITT group, in which the level of target A1c attainment was higher than that among GLP-1RA–naive patients at baseline. At the A1c target of <7% (53 mmol/mol), an absolute increase of 18.2 percentage points (27.9% in the preindex period to 46.1% in the postindex period, P < 0.001) was observed, whereas at the A1c target of <8% (64 mmol/mol), an absolute increase of 15.1 percentage points (59.3% in the preindex period to 74.4% in the postindex period, P < 0.001) was observed.

      Discussion

      This real-world study observed clinically meaningful changes in A1c across a variety of patient populations using semaglutide OW. Overall, a 0.9 percentage point absolute reduction (−9.8 mmol/mol) in mean A1c values was observed after initiation of semaglutide OW therapy, with a 1.1 percentage point reduction (−12.0 mmol/mol) in patients taking semaglutide OW at the time of the postindex A1c measurement. Increases in attainment of various A1c targets were also observed after initiation of semaglutide OW therapy, with approximately one-quarter more patients attaining A1c values <7% (53 mmol/mol) (25.3%) and <8% (64 mmol/mol) (23.4%) overall. When accounting for prior GLP-1RA use, the GLP-1RA–naive stratum had twice the mean absolute reduction in A1c compared with the GLP-1RA–experienced stratum (−1.2 [−13.1 mmol/mol] vs −0.6 [−6.6 mmol/mol]). Of note, the GLP-1RA–experienced patients were observed to achieve a significant 0.6 percentage point reduction (−6.6 mmol/mol) in mean A1c value, despite their mean baseline A1c value being relatively well controlled at 7.9% (62.8 mmol/mol). In addition, the GLP-1RA–experienced stratum with a baseline A1c value >9% (75 mmol/mol) had a robust 1.8 percentage point reduction (−19.7 mmol/mol) in their mean A1c value, even though they had previously been taking a GLP-1RA and possibly other diabetes medications.
      In providing guidance to pharmacologic approaches to glycemic treatment, the ADA notes that many patients will require combination therapy to achieve their target A1c level.
      Pharmacologic Approaches to Glycemic Treatment: Standards of Medical Care in Diabetes—2020.
      Metformin, the recommended first-line therapy, was used by most patients initiating semaglutide therapy. However, for a variety of reasons, metformin monotherapy may not be adequate for many individuals to achieve their A1c target.
      Pharmacologic Approaches to Glycemic Treatment: Standards of Medical Care in Diabetes—2020.
      In addition, the 2020 ADA guidelines preferentially recommend GLP-1RA use in individuals with T2DM and atherosclerotic vascular disease independent of A1c level or individualized A1c target.
      Pharmacologic Approaches to Glycemic Treatment: Standards of Medical Care in Diabetes—2020.
      Other diabetes therapy prescribing patterns of note in the study ITT group included that GLP-1RA–experienced patients more commonly received insulin compared with the GLP-1RA–naive patients (44% vs 33%), and SGLT2 inhibitors were more commonly used by the GLP-1RA experienced subcohort (49%) compared with the GLP-1RA–naive stratum (35%). More use of prior diabetes medication may be indicative of more advanced T2DM disease or difficulties achieving glycemic control in the GLP-1RA–experienced stratum compared with the GLP-1RA–naive stratum.
      Our results complement the randomized controlled trials performed as part of the SUSTAIN clinical program that compared semaglutide with a variety of other second-line diabetes medications. The SUSTAIN trials (SUSTAIN-2, SUSTAIN-4, SUSTAIN-7, and SUSTAIN-8) found a significant reduction in A1c for semaglutide OW compared with sitagliptin, insulin, dulaglutide, and canagliflozin, respectively.
      • Pratley RE
      • Aroda VR
      • Lingvay I
      • et al.
      Semaglutide versus dulaglutide once weekly in patients with type 2 diabetes (SUSTAIN 7): a randomised, open-label, phase 3b trial.
      • Ahrén B
      • Masmiquel L
      • Kumar H
      • et al.
      Efficacy and safety of once-weekly semaglutide versus once-daily sitagliptin as an add-on to metformin, thiazolidinediones, or both, in patients with type 2 diabetes (SUSTAIN 2): a 56-week, double-blind, phase 3a, randomised trial.
      • Aroda VR
      • Bain SC
      • Cariou B
      • et al.
      Efficacy and safety of once-weekly semaglutide versus once-daily insulin glargine as add-on to metformin (with or without sulfonylureas) in insulin-naive patients with type 2 diabetes (SUSTAIN 4): a randomised, open-label, parallel-group, multicentre, mul.
      • Lingvay I
      • Catarig A-M
      • Frias JP
      • et al.
      Efficacy and safety of once-weekly semaglutide versus daily canagliflozin as add-on to metformin in patients with type 2 diabetes (SUSTAIN 8): a double-blind, phase 3b, randomised controlled trial.
      • Brown RE
      • Liu AR
      • Mahbubani R
      • Aronson R.
      Semaglutide in patients with Type 2 diabetes: Real-world analysis in the Canadian LMC Diabetes Registry: The SPARE study.
      SUSTAIN-7 found that semaglutide OW 0.5 and 1.0 mg was superior to dulaglutide in improving glycemic control (mean percentage A1c reduction with 0.5 mg semaglutide OW, 1.5% [16.4 mmol/mol]; 1.0 mg semaglutide OW: 1.8% [19.7 mmol/mol]; both P < 0.0001) and weight reduction in a 40 week-long study.
      • Pratley RE
      • Aroda VR
      • Lingvay I
      • et al.
      Semaglutide versus dulaglutide once weekly in patients with type 2 diabetes (SUSTAIN 7): a randomised, open-label, phase 3b trial.
      The mean age among our study patients was comparable with that in SUSTAIN-7 (54.3 vs 55.5 years).
      • Pratley RE
      • Aroda VR
      • Lingvay I
      • et al.
      Semaglutide versus dulaglutide once weekly in patients with type 2 diabetes (SUSTAIN 7): a randomised, open-label, phase 3b trial.
      In addition, the timeframes of SUSTAIN-7 and this analysis were comparable (40 vs 41.3 weeks). One difference between the studies was with regard to the baseline diabetic medication that was allowed because our study allowed for use of all classes of diabetic medications, whereas SUSTAIN-7 included only patients with stable metformin use. The combination of other diabetic medications may have driven the differences observed in A1c reductions, especially in the GLP-1RA–experienced subcohort from our real-world analysis, because those individuals were taking multiple diabetic medications, which already had an impact on A1c levels and may have indicated a more severe and complex patient population. Finally, clinical trials may have higher adherence to study medications compared with real-world patient management.
      • Pratley RE
      • Aroda VR
      • Lingvay I
      • et al.
      Semaglutide versus dulaglutide once weekly in patients with type 2 diabetes (SUSTAIN 7): a randomised, open-label, phase 3b trial.
      Microvascular and macrovascular disease outcomes were prospectively explored in the SUSTAIN-6 trial. Although beneficial major acute coronary event outcomes were observed, there were mixed results with regard to microvascular outcomes. For example, nephropathy occurred less frequently in the semaglutide vs placebo group (3.8% vs 6.1%; hazard ratio, 0.64; 95% CI, 0.46–0.88; P = 0.005).
      • Marso SP
      • Bain SC
      • Consoli A
      • et al.
      Semaglutide and Cardiovascular Outcomes in Patients with Type 2 Diabetes.
      Although the overall number of retinopathy events was low, there was an unexpected higher rate of diabetic retinopathy complications (eg, vitreous hemorrhage, blindness, or the need for treatment with an intravitreal agent or photocoagulation) in the semaglutide group (3.0% vs 1.8; hazard ratio, 1.76; 95% CI, 1.11–2.78; P = 0.02). In general, rapid improvement in glucose control has been associated with a temporary worsening of diabetic retinopathy.

      Ozempic® (semaglutide) [package insert]. Plainsboro, NJ: Novo Nordisk Inc.2017.

      SUSTAIN-6 was the only study in which diabetic retinopathy complications were proactively collected, and the progression of retinopathy was not measured in this or other studies of the SUSTAIN program. Semaglutide's long-term effects on diabetic eye disease are currently being evaluated in a randomized, placebo-controlled study (the FOCUS study [A Research Study to Look at How Semaglutide Compared to Placebo Affects Diabetic Eye Disease in People With Type 2 Diabetes]).
      • Nordisk Novo
      A Research Study to Look at How Semaglutide Compared to Placebo Affects Diabetic Eye Disease in People With Type 2 Diabetes (FOCUS).
      In contrast to randomized controlled trials, the study by Mody et al
      • Mody R
      • Huang Q
      • Yu M
      • et al.
      Clinical and economic outcomes among injection-naïve patients with type 2 diabetes initiating dulaglutide compared with basal insulin in a US real-world setting: the DISPEL Study.
      explored the clinical and economic outcomes in the real-world setting associated with dulaglutide compared with insulin. Mody et al
      • Mody R
      • Huang Q
      • Yu M
      • et al.
      Clinical and economic outcomes among injection-naïve patients with type 2 diabetes initiating dulaglutide compared with basal insulin in a US real-world setting: the DISPEL Study.
      observed a 1.1% reduction (−12.0 mmol/mol) in A1c values associated with dulaglutide, with a 2.1% reduction (−22.9 mmol/mol) observed in patients with a baseline A1c value >9% (75 mmol/mol). The study by Mody et al
      • Mody R
      • Huang Q
      • Yu M
      • et al.
      Clinical and economic outcomes among injection-naïve patients with type 2 diabetes initiating dulaglutide compared with basal insulin in a US real-world setting: the DISPEL Study.
      excluded patients who had any antidiabetic injectable medication in the 6 months before index medication use. This finding is most similar to our GLP-1RA–naive stratum, in which we observed a mean absolute A1c reduction of −1.2% (−13.1 mmol/mol) overall and a −2.5% reduction (−27.3 mmol/mol) among those with a baseline A1c value >9% (75 mmol/mol). To further put our results in context with the study by Mody et al,
      • Mody R
      • Huang Q
      • Yu M
      • et al.
      Clinical and economic outcomes among injection-naïve patients with type 2 diabetes initiating dulaglutide compared with basal insulin in a US real-world setting: the DISPEL Study.
      one-third of patients in our study's GLP-1RA–naive stratum had baseline insulin use in the last year, which may have blunted some of the A1c-lowering effect observed in these patients. Additional study of concomitant GLP-1RA and insulin use and other combinations of antidiabetic medications is needed to further understand the outcomes associated with their use.
      Initiation of semaglutide OW therapy was associated with clinically and statistically significant reductions in mean A1c values and increases in A1c goal attainment in an earlier real-world analysis in a US population.
      • Visaria J
      • Dang-Tan T
      • Petraro PV
      • Nepal BK
      • Willey VJ.
      Real-World Effectiveness of Semaglutide in Early Users from a U.S. Commercially Insured (CI) and Medicare Advantage (MA) Population.
      The initial cohort of individuals with T2DM initiating semaglutide OW therapy had a significant absolute reduction in mean A1c value in all ITT patients (−1.3% [−14.2 mmol/mol]), GLP-1RA–naive subcohort (−2.0% [−21.9 mmol/mol]), and A1c value >9% (75 mmol/mol) GLP-1RA–naive subcohort (−2.9% [−31.7 mmol/mol]) patients (all P < 0.001).
      • Visaria J
      • Dang-Tan T
      • Petraro PV
      • Nepal BK
      • Willey VJ.
      Real-World Effectiveness of Semaglutide in Early Users from a U.S. Commercially Insured (CI) and Medicare Advantage (MA) Population.
      Similar to our study, the above referenced study was also associated with statistically significant reductions in A1c values across all cohorts and strata, and larger A1c reductions in the earlier real-world analysis may be attributable to higher baseline A1c values and shorter follow-up time, among other factors.
      • Visaria J
      • Dang-Tan T
      • Petraro PV
      • Nepal BK
      • Willey VJ.
      Real-World Effectiveness of Semaglutide in Early Users from a U.S. Commercially Insured (CI) and Medicare Advantage (MA) Population.
      This study provides additional real-world evidence in a broader population with longer follow-up.
      Our study has several limitations. The analysis was limited to a US commercially insured and Medicare Advantage population, and as such, results may not be generalizable to individuals with other types of health insurance or those uninsured or underinsured persons who may not have adequate access to health care. Medical care and pharmacy claims may have had coding errors or omissions may have occurred, resulting in misclassification of some patients and potential underreporting for this population. Although laboratory data were available for a significant subset of the overall population (40%), A1c values in those individuals in whom laboratory test results were not available may have differed. Race/ethnicity or socioeconomic status data were not available, which would be informative with regard to application of these results to other populations. The time difference between semaglutide OW prescription fill and A1c measurement varied among patients. As a result, variability in A1c values may have been reflective, in part, because of being recorded at other time points. Although increased from prior real-world studies of semaglutide OW,
      • Visaria J
      • Dang-Tan T
      • Petraro PV
      • Nepal BK
      • Willey VJ.
      Real-World Effectiveness of Semaglutide in Early Users from a U.S. Commercially Insured (CI) and Medicare Advantage (MA) Population.
      postindex follow-up time was still relatively short, and additional research is required to assess the persistence and durability of its effect on A1c value, weight, and cardiovascular outcomes in real-world settings.
      • Visaria J
      • Dang-Tan T
      • Petraro PV
      • Nepal BK
      • Willey VJ.
      Real-World Effectiveness of Semaglutide in Early Users from a U.S. Commercially Insured (CI) and Medicare Advantage (MA) Population.

      Conclusions

      Semaglutide OW is associated with clinically and statistically significant A1c reduction as well as increases in the proportions of patients attaining A1c targets in this real-world study. Significant reductions in A1c associated with semaglutide OW were observed in all study subcohorts, even in GLP-1RA–experienced patients in whom significant reductions in A1c would be least expected. Target A1c attainment significantly increased overall and in all subgroups and strata, with approximately half of all patients attaining an A1c value <7% (53 mmol/mol) and three-quarters attaining an A1c value <8% (64 mmol/mol). This real-world study provides complementary evidence to that generated from randomized controlled trials for guidance on the use of semaglutide OW in clinical practice.

      Disclosures

      Chioma Uzoigwe, Caroline Swift, Tam Dang-Tan, and Yurek Paprocki report being employees of Novo Nordisk during the conduct of this study. In addition, Chioma Uzoigwe and Yurek Paprocki report being stockholders with Novo Nordisk. Jay Visaria and Vincent Willey report being employees of HealthCore, Inc during the conduct of this study. Authors from NovoNordisk contributed to the study by participating in study design, data analysis and interpretation, manuscript writing and revisions, and approval for publication. HealthCore Inc receives funding from numerous external entities to perform research projects. The authors have indicated that they have no other conflicts of interest regarding the content of this article.

      Acknowledgements

      Bal Nepal, PhD, MPH, HealthCore Inc, served as data programmer and analyst for the study. Patrick Callahan, PharmD, MS, Genesis Research, and Nicole Safran, MPH, Genesis Research, helped draft the original manuscript. Author contributions are as follows: Jay Visaria: conceptualization, formal analysis, investigation, writing–original draft; Chioma Uzoigwe-Smith: conceptualization, writing–review and editing; Caroline Swift: conceptualization, writing–review and editing; Tam Dang Tan: conceptualization, writing–review and editing; Yurek Paprocki: conceptualization, writing–review and editing; and Vincent Willey: conceptualization, writing–original draft.

      Funding Sources

      This study was funded by Novo Nordisk.

      Appendix. Supplementary materials

      References

        • Centers for Disease Control and Prevention
        National Diabetes Statistics Report 2020: Estimates of Diabetes and Its Burden in the United States.
        2020 (Atlanta, GA;Accessed April 2020. URL: https://www.cdc.gov/diabetes/pdfs/data/statistics/national-diabetes-statistics-report.pdf)
        • Xu G
        • Liu B
        • Sun Y
        • et al.
        Prevalence of diagnosed type 1 and type 2 diabetes among US adults in 2016 and 2017: population based study.
        BMJ. 2018; 362: k1497
      1. National Health and Nutrition Examination Survey. Diabetes prevalence and glycemic control among adults aged 20 and over, by sex, age, and race and Hispanic origin. 2018. Accessed April 2020. URL: https://www.cdc.gov/nchs/data/hus/2018/014.pdf.

      2. Glycemic Targets: Standards of Medical Care in Diabetes—2019.
        Diabetes Care. 2019; 42: S61-S70
      3. Pharmacologic Approaches to Glycemic Treatment: Standards of Medical Care in Diabetes—2020.
        Diabetes Care. 2020; 43: S98-S110
        • Qaseem A
        • Wilt TJ
        • Kansagara D
        • Horwitch C
        • Barry MJ
        • Forciea MA.
        Hemoglobin A 1c Targets for Glycemic Control With Pharmacologic Therapy for Nonpregnant Adults With Type 2 Diabetes Mellitus: A Guidance Statement Update From the American College of Physicians.
        Ann Intern Med. 2018; 168: 569
        • Marso SP
        • Daniels GH
        • Brown-Frandsen K
        LEADER Steering Committee; LEADER Trial Investigators. Liraglutide and cardiovascular outcomes in type 2 diabetes.
        N Engl J Med. 2016; 375: 311-322
        • Gerstein HC
        • Colhoun HM
        • Gagenais GR
        • et al.
        Dulaglutide and Cardiovascular Outcomes in Type 2 Diabetes (REWIND): a Double-Blind, Randomised Placebo-Controlled Trial.
        Lancet. 2019; 394: 121-130
        • Zinman B
        • Wanner C
        • Lachin JM
        • et al.
        EMPAREG OUTCOME Investigators. Empagliflozin, cardiovascular outcomes, and mortality in type 2 diabetes.
        N Engl J Med. 2015; 373: 2117-2128
        • Neal B
        • Perkovic V
        • Mahaffey KW
        • et al.
        CANVAS Program Collaborative Group Canagliflozin and cardiovascular and renal events in type 2 diabetes.
        N Engl J Med. 2017; 377: 644-657
        • Marso SP
        • Bain SC
        • Consoli A
        • et al.
        Semaglutide and Cardiovascular Outcomes in Patients with Type 2 Diabetes.
        N Engl J Med. 2016; 375: 1834-1844
        • Lau J
        • Bloch P
        • Schäffer L
        • et al.
        Discovery of the Once-Weekly Glucagon-Like Peptide-1 (GLP-1) Analogue Semaglutide.
        J Med Chem. 2015; 58: 7370-7380
      4. Hai MTT. Center for Drug Evaluation and Research Approval Package for: Application Number: 209637Orig1s000. Food and Drug Administration. Silver Spring;2017. Accessed April 2020. URL: https://www.accessdata.fda.gov/drugsatfda_docs/nda/2017/209637Orig1s000Approv.pdf.

        • Pratley RE
        • Aroda VR
        • Lingvay I
        • et al.
        Semaglutide versus dulaglutide once weekly in patients with type 2 diabetes (SUSTAIN 7): a randomised, open-label, phase 3b trial.
        Lancet Diabetes Endocrinol. 2018; 6: 275-286
        • Ahrén B
        • Masmiquel L
        • Kumar H
        • et al.
        Efficacy and safety of once-weekly semaglutide versus once-daily sitagliptin as an add-on to metformin, thiazolidinediones, or both, in patients with type 2 diabetes (SUSTAIN 2): a 56-week, double-blind, phase 3a, randomised trial.
        Lancet Diabetes Endocrinol. 2017; 5: 341-354
        • Aroda VR
        • Bain SC
        • Cariou B
        • et al.
        Efficacy and safety of once-weekly semaglutide versus once-daily insulin glargine as add-on to metformin (with or without sulfonylureas) in insulin-naive patients with type 2 diabetes (SUSTAIN 4): a randomised, open-label, parallel-group, multicentre, mul.
        Lancet Diabetes Endocrinol. 2017; 5: 355-366
        • Lingvay I
        • Catarig A-M
        • Frias JP
        • et al.
        Efficacy and safety of once-weekly semaglutide versus daily canagliflozin as add-on to metformin in patients with type 2 diabetes (SUSTAIN 8): a double-blind, phase 3b, randomised controlled trial.
        Lancet Diabetes Endocrinol. 2019; 7: 834-844
        • Brown RE
        • Liu AR
        • Mahbubani R
        • Aronson R.
        Semaglutide in patients with Type 2 diabetes: Real-world analysis in the Canadian LMC Diabetes Registry: The SPARE study.
        Diabetes. 2019; 68 (995–P)
        • Visaria J
        • Dang-Tan T
        • Petraro PV
        • Nepal BK
        • Willey VJ.
        Real-World Effectiveness of Semaglutide in Early Users from a U.S. Commercially Insured (CI) and Medicare Advantage (MA) Population.
        Diabetes. 2019; 68 (1006–P)
      5. Ozempic® (semaglutide) [package insert]. Plainsboro, NJ: Novo Nordisk Inc.2017.

        • Nordisk Novo
        A Research Study to Look at How Semaglutide Compared to Placebo Affects Diabetic Eye Disease in People With Type 2 Diabetes (FOCUS).
        2021 (Available from) (Accessed March 2,)
        • Mody R
        • Huang Q
        • Yu M
        • et al.
        Clinical and economic outcomes among injection-naïve patients with type 2 diabetes initiating dulaglutide compared with basal insulin in a US real-world setting: the DISPEL Study.
        BMJ Open Diabetes Res Care. 2019; 7e000884