GLUCAGON-LIKE PEPTIDE 1 RECEPTOR AGONISTS AND OBSTRUCTIVE SLEEP APNEA RISK USING MEDICARE DATA 2007 - 2019

Study type

• Study topic:
  Disease /health condition

  Human medicinal product
• Study type:
  Non-interventional study
• Scope of the study:
  Effectiveness study (incl. comparative)

  Safety study (incl. comparative)
• Data collection methods:
  Secondary use of data

Non-interventional study

• Non-interventional study design:
  Cohort

Study drug and medical condition

• Medicinal product name:
  OZEMPIC

  RYBELSUS

  WEGOVY

  VICTOZA

  SAXENDA

  TRULICITY

  MOUNJARO

  JARDIANCE

  INVOKANA

  STEGLATRO
• Medicinal product name, other:
  FARXIGA
• Study drug International non-proprietary name (INN) or common name:
  EXENATIDE

  LIRAGLUTIDE

  LIXISENATIDE

  ALBIGLUTIDE

  DULAGLUTIDE

  SEMAGLUTIDE

  DAPAGLIFLOZIN

  CANAGLIFLOZIN

  EMPAGLIFLOZIN

  ERTUGLIFLOZIN
• Anatomical Therapeutic Chemical (ATC) code:
  (A10BJ01) exenatide

  exenatide

  (A10BJ02) liraglutide

  liraglutide

  (A10BJ03) lixisenatide

  lixisenatide

  (A10BJ04) albiglutide

  albiglutide

  (A10BJ05) dulaglutide

  dulaglutide

  (A10BJ06) semaglutide

  semaglutide

  (A10BK01) dapagliflozin

  dapagliflozin

  (A10BK02) canagliflozin

  canagliflozin

  (A10BK03) empagliflozin

  empagliflozin

  (A10BK04) ertugliflozin

  ertugliflozin
• Medical condition to be studied:
  Type 2 diabetes mellitus

  Sleep apnoea syndrome

Population studied

• Short description of the study population:
  Medicare Fee-for-Service (FFS) Database (Parts A, B, and D) 2007-2019 (or additional years of data if available). This US federal database contains deidentified individual-level, longitudinal information on demographics, diagnoses, and procedures, and outpatient prescription dispensations recorded during billing of all health care encounters.
• Age groups:
  • Elderly (≥ 65 years)

    • Adults (65 to < 75 years)
    • Adults (75 to < 85 years)
    • Adults (85 years and over)

Study design details

• Study design:
  Active comparator, new user cohort study
• Main study objective:
  Aim 1: To estimate the comparative effect of GLP-1RA versus sodium-glucose co-transporter-2 (SGLT-2) inhibitors on the incidence of OSA in older adults with type 2 diabetes. ; Aim 2: To investigate whether comorbid OSA at baseline modifies the comparative effect of GLP-1RA (vs. SGLT-2 inhibitors) on the incidence of adapted major adverse cardiovascular events (MACE) defined as stroke, myocardial infarction, coronary artery bypass grafting, percutaneous transluminal coronary angioplasty, or all-cause mortality in older adults with type 2 diabetes.
• Setting:
  We will include an active comparator new user cohort where GLP-1 agonists are compared with SGLT-2 inhibitors. New users are defined as individuals who initiate the drugs of interest or their active comparator after a preceding washout period of at least 12 months without a prescription for the drug classes compared. Participants are allowed to have other anti-hyperglycemic drugs during the washout period except the drugs being compared. Study subjects are required to have 2 diagnosis codes for T2DM and at least 12 months of continuous part A, B, and D coverage before the first prescription date. Since SGLT2 inhibitors were approved in March 201348, the earliest possible first prescription date for this study will be January 1, 2014, allowing for adoption and uptake in clinical practice and mitigating channeling bias immediately after approval.; We will exclude all patients who do not refill the same drug class within the days supply and a grace period of 30 days after the first prescription. Requiring two prescriptions increases the probability that patients actually started therapy. We will describe the patients who do not meet the refill criterion to assess the potential selection imposed by requiring a refill. Patients with evidence of pre-existing OSA (diagnosis of OSA, prescriptions for modafinil and armodafinil, prior use of continuous positive airway pressure (CPAP) therapy, oral device) and adapted Major Adverse Cardiovascular Events (MACE) (see outcome section) within all the available lookback period prior to the first prescription or between the first and 2nd prescription will be excluded from the analysis for Aim 1 and Aim 2 respectively. For both Aims 1 and 2, we will also exclude patients with a history of bariatric surgery, stage 4 chronic kidney disease, end-stage renal disease or dialysis during the lookback period prior to the first prescription. Additionally, for Aim 1, we will exclude patients with any sleep disorder, tracheostomy and current hom
• Comparators:
  Exposure will be defined by at least two same drug class prescription dispensing claims of either GLP-1 agonists or SGLT-2 inhibitors (active comparator) for Aim 1 and at least one prescription dispensing claim of either GLP-1 agonists or SGLT-2 inhibitors for Aim 2 between January 1, 2014, and December 31, 2018, identified using Anatomical Therapeutic Chemical (ATC) classification codes and National Drug Codes (NDCs). We will identify an active comparator cohort where GLP-1 agonists are compared with SGLT-2 inhibitors.
• Outcomes:
  Aim 1: incident obstructive sleep apnea (OSA). The American Academy of Sleep Medicine (AASM) guidelines recommend ordering sleep studies (in-lab polysomnography or technically adequate home sleep apnea tests) only when there is clinical suspicion of moderate-to-severe OSA and note that repeat testing after a negative result is uncommon unless symptoms persist or the initial test was inconclusive. OSA will therefore be defined if there are both a polysomnography and least one diagnosis code for OSA (ICD-9 code 327.23 or ICD-10 code G47.33) within 60 days after the polysomnography (Healthcare Common Procedure Coding System 95808, 95810, 95811, G0398-G0400). Date of diagnosis will be assigned at the first OSA claim associated with polysomnography. However, because this algorithm was used for cohort definition and has not been validated in Medicare claims data, we will implement four other algorithms for defining OSA: 1) A polysomnography and at least one procedure code for OSA treatment (positive airway pressure (PAP) therapy or oral device) or diagnosis code for OSA within 60 days after the polysomnography. 2) A polysomnography and at least one diagnosis code for unspecified sleep apnea plus one procedure code for OSA treatment (positive airway pressure (PAP) therapy or oral device) within 60 days after the polysomnography. 3) At least two diagnosis codes of OSA within 60 days. 4) At least one procedure code for OSA treatment (positive airway pressure (PAP) therapy or oral device). The performance of all five algorithms will be assessed in a validation study using linked Medicare and EHR data.; Aim 2: adapted MACE, a composite of myocardial infarction (MI), stroke, coronary artery bypass grafting (CABG), percutaneous transluminal coronary angioplasty (PTCA), and all-cause mortality. We use an adapted definition that substitutes all-cause mortality for cardiovascular death because Medicare claims data do not reliably distinguish cause-specific mortality.
• Data analysis plan:
  Propensity Score (PS) methods will be used to control for measured confounders. Specifically, logistic regression will be utilized to estimate propensity scores - the probability of initiating GLP-1RA compared to SGLT-2 inhibitors, conditional on baseline covariates. Our primary aim is to estimate the counterfactual scenario of what would have happened to the initiators of GLP-1RA if they had initiated SGLT-2 inhibitors instead. To achieve this goal, we will estimate the average treatment effect in the treated (ATT) by reweighting the comparator drug initiators by the propensity score odds (PS/(1-PS)) (Standardized Morbidity Ratio Weighting).56 The adequacy of covariate balance will be evaluated based on standardized absolute mean differences (SAMD), with a threshold of less than 0.1 indicating satisfactory balance.