Advertisement

Cost-Effectiveness of Canagliflozin versus Sitagliptin as Add-on to Metformin in Patients with Type 2 Diabetes Mellitus in Mexico

Open ArchivePublished:June 03, 2015DOI:https://doi.org/10.1016/j.vhri.2015.01.002

      Abstract

      Objective

      To assess the cost-effectiveness of canagliflozin versus sitagliptin for the treatment of type 2 diabetes mellitus (T2DM) as an add-on to metformin in Mexico.

      Methods

      A validated model (Economic and Health Outcomes [ECHO]-T2DM) was used to estimate the cost-effectiveness of canagliflozin 300 or 100 mg versus sitagliptin 100 mg in patients with T2DM inadequately controlled on metformin monotherapy. Data from a head-to-head, phase III clinical trial, including patients’ baseline demographic characteristics, biomarker values, and treatment effects, were used to simulate outcomes and resource use over 20 years from the perspective of the Mexican health care system. Costs of complications and adverse events were tailored to the Mexican setting and discounted at 5%. Cost-effectiveness was assessed using willingness-to-pay thresholds equivalent to 1 times the gross domestic product per capita (locally perceived to be “very cost-effective”) and 3 times the gross domestic product per capita (locally perceived to be “cost-effective”) on the basis of recommendations of the Mexican government and the World Health Organization.

      Results

      Owing primarily to better glycated hemoglobin (HbA1c), body weight, and systolic blood pressure values, canagliflozin 300 and 100 mg were associated with an incremental benefit of 0.16 and 0.06 quality-adjusted life-years (QALYs) versus sitagliptin 100 mg, respectively, over 20 years. The mean differences in cost for canagliflozin 300 and 100 mg versus sitagliptin 100 mg were Mexican pesos (MXP) 1797 (US $134) and MXP 7262 (US $540), respectively, resulting in a cost per QALY gained of MXP 11,210 (US $834) and MXP 128,883 (US $9590), respectively. Both of these cost-effectiveness ratios are below the very cost-effective willingness-to-pay threshold in Mexico. The general finding that canagliflozin is cost-effective versus sitagliptin in Mexico was supported by sensitivity analyses.

      Conclusion

      In Mexico, both doses of canagliflozin are likely to be cost-effective versus sitagliptin in patients with T2DM who have inadequate glucose control on metformin, primarily because of better biomarker control and higher QALYs.

      Keywords

      Introduction

      The number of people with diabetes in Latin American countries is growing, likely because of widespread increases in obesity in the region, and it is expected to increase by approximately 60%, from 24.1 million today to 38.5 million by 2035 [
      The Lancet Diabetes Endocrinology
      Obesity prevention in Latin America: now is the time.
      ]. About 90% of patients have type 2 diabetes mellitus (T2DM) [
      International Diabetes Federation
      IDF Diabetes Atlas.
      ]. In Mexico, the prevalence of diagnosed patients increased from 7.3% of the population in 2006 to 9.2% in 2012 [
      • Corona A.J.
      • Martinez R.R.
      • Villalpando S.
      • SBCA Salinas
      Diabetes mellitus: La urgente necesidad de reforzar la respuesta en políticas públicas para su prevención y control.
      ], and it is believed that many cases remain undiagnosed [
      International Diabetes Federation
      IDF Diabetes Atlas.
      ]. Diabetes has been the leading cause of death since 2000, and was estimated to account for nearly 14% of deaths in 2009 [
      • Barquera S.
      • Campos-Nonato I.
      • Aguilar-Salinas C.
      • et al.
      Diabetes in Mexico: cost and management of diabetes and its complications and challenges for health policy.
      ].
      T2DM imposes a significant economic burden on health care in Latin America due to the increasing prevalence and chronic nature of T2DM and associated comorbidities. There are direct costs incurred in managing the hyperglycemia associated with T2DM. It is notable, however, that most of the costs are attributable to T2DM-related complications (e.g., myocardial infarction, stroke, nephropathy, neuropathy, and retinopathy), the rates of which are inversely related to disease control [
      • Guzman J.R.
      • Lyra R.
      • Aguilar-Salinas C.A.
      • et al.
      Treatment of type 2 diabetes in Latin America: a consensus statement by the medical associations of 17 Latin American countries. Latin American Diabetes Association.
      ]. In Mexico, the direct costs of diabetes were estimated at approximately US $1.16 billion (~15 billion Mexican pesos [MXP]) in 2006, and these figures have steadily increased [
      • Barquera S.
      • Campos-Nonato I.
      • Aguilar-Salinas C.
      • et al.
      Diabetes in Mexico: cost and management of diabetes and its complications and challenges for health policy.
      ]. Estimates suggest that diabetes-related complications can substantially increase patient costs in Mexico [
      • Barquera S.
      • Campos-Nonato I.
      • Aguilar-Salinas C.
      • et al.
      Diabetes in Mexico: cost and management of diabetes and its complications and challenges for health policy.
      ]. As noted in a recent consensus statement from the Latin American Diabetes Association, the Latin American health care system has historically focused on the treatment of acute health conditions, primarily due to limited resources [
      • Guzman J.R.
      • Lyra R.
      • Aguilar-Salinas C.A.
      • et al.
      Treatment of type 2 diabetes in Latin America: a consensus statement by the medical associations of 17 Latin American countries. Latin American Diabetes Association.
      ]. As chronic diseases such as T2DM become more common in Latin America, payers and clinicians will face challenges unique to this region in helping patients achieve health-related goals.
      Maintaining near-normal blood glucose levels has been shown to improve key T2DM-related outcomes [
      UK Prospective Diabetes Study (UKPDS) Group
      Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33).
      ]. The Latin American Diabetes Association and the Institute of Mexican Social Security (IMSS) recommend maintaining a glycated hemoglobin (HbA1c) level of less than 7.0%, blood pressure of less than or equal to 130/80 mmHg, and low-density lipoprotein cholesterol (LDL-C) level of less than 100 mg/dL in most patients [
      • Guzman J.R.
      • Lyra R.
      • Aguilar-Salinas C.A.
      • et al.
      Treatment of type 2 diabetes in Latin America: a consensus statement by the medical associations of 17 Latin American countries. Latin American Diabetes Association.
      ,

      Instituto Mexicano del Seguro Social. Diagnóstico y Tratamiento de la Diabetes Mellitus Tipo 2 México: Guía de Práctica Clinica. Instituto Mexicano del Seguro Social, 2009. Actualización 2012. Instituto Mexicano del Seguro Social 2012. Available from: http://www.imss.gob.mx/sites/all/statics/guiasclinicas/000GER_DiabetesMellitus.pdf. [Accessed October 22, 2014].

      ], consistent with the recommendations of the American Diabetes Association [
      • Inzucchi S.E.
      • Bergenstal R.M.
      • Buse J.B.
      • et al.
      Management of hyperglycemia in type 2 diabetes: a patient-centered approach. Position statement of the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD).
      ]. In addition, guidelines emphasize the importance of weight loss/control in T2DM management, reflecting acknowledgement of the detrimental effects of excess weight on health outcomes [
      • Guzman J.R.
      • Lyra R.
      • Aguilar-Salinas C.A.
      • et al.
      Treatment of type 2 diabetes in Latin America: a consensus statement by the medical associations of 17 Latin American countries. Latin American Diabetes Association.
      ,

      Instituto Mexicano del Seguro Social. Diagnóstico y Tratamiento de la Diabetes Mellitus Tipo 2 México: Guía de Práctica Clinica. Instituto Mexicano del Seguro Social, 2009. Actualización 2012. Instituto Mexicano del Seguro Social 2012. Available from: http://www.imss.gob.mx/sites/all/statics/guiasclinicas/000GER_DiabetesMellitus.pdf. [Accessed October 22, 2014].

      ,
      • Inzucchi S.E.
      • Bergenstal R.M.
      • Buse J.B.
      • et al.
      Management of hyperglycemia in type 2 diabetes: a patient-centered approach. Position statement of the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD).
      ].
      Many patients do not meet or maintain glycemic goals with available treatments [
      • Ali M.K.
      • Bullard K.M.
      • Saaddine J.B.
      • et al.
      Achievement of goals in U.S. diabetes care, 1999-2010.
      ,
      • Stark Casagrande S.
      • Fradkin J.E.
      • Saydah S.H.
      • et al.
      The prevalence of meeting A1C, blood pressure, and LDL goals among people with diabetes, 1988-2010.
      ]. According to data from the 2006 Mexican National Nutrition Survey, only 5.3% of the patients with T2DM were found to have an HbA1c level of 7.0% or less despite treatment [
      • Villalpando S.
      • de la Cruz V.
      • Rojas R.
      • et al.
      Prevalence and distribution of type 2 diabetes mellitus in Mexican adult population: a probabilistic survey.
      ]. Notably, more than half had an HbA1c ≥11.0%. Similarly, data indicate widespread failure to meet blood pressure and lipid goals in Mexico. In 2006, for example, approximately one-third of the Mexican population had a systolic blood pressure (SBP) of ≥140 mmHg and approximately 75% had LDL-C ≥100 mg/dL [
      • Barquera S.
      • Campos-Nonato I.
      • Hernandez-Barrera L.
      • et al.
      Hypertension in Mexican adults: results from the National Health and Nutrition Survey 2006.
      ,
      • Aguilar-Salinas C.A.
      • Gomez-Perez F.J.
      • Rull J.
      • et al.
      Prevalence of dyslipidemias in the Mexican National Health and Nutrition Survey 2006.
      ]. Moreover, two-thirds were classified as being overweight or obese [
      • Gomez L.M.
      • Hernandez-Prado B.
      • Morales M.C.
      • Shamah-Levy T.
      Physical activity and overweight/obesity in adult Mexican population: the Mexican National Health and Nutrition Survey 2006.
      ].
      Canagliflozin is an agent that inhibits sodium glucose co-transporter 2 (SGLT2), which is approved in numerous countries [

      Johnson & Johnson, Inc. New long-term data show INVOKANA® (canagliflozin) lowers blood glucose in older patients with type 2 diabetes inadequately controlled on other therapies. 2014. Available from: http://www.jnj.com/news/all/New-Long-Term-Data-Show-INVOKANA-canagliflozin-Lowers-Blood-Glucose-in-Older-Patients-with-Type-2-Diabetes-Inadequately-Controlled-on-Other-Therapies. [Accessed October 20, 2014].

      ], including Mexico, for the treatment of adults with T2DM [
      INVOKANA® (canagliflozin) tablets, for oral use [package insert].
      ,

      INVOKANA. Informacion Para Prescribir Amplia INVOKANA® F.F Tableta V1.0 CCDS Junio. 2012.

      ]; the efficacy and safety of canagliflozin have been demonstrated in phase III clinical trials of up to 2 years in a broad range of patients with T2DM [
      • Bode B.
      • Stenlöf K.
      • Sullivan D.
      • et al.
      Efficacy and safety of canagliflozin treatment in older subjects with type 2 diabetes mellitus: a randomized trial.
      ,
      • Cefalu W.T.
      • Leiter L.A.
      • Yoon K.-H.
      • et al.
      Efficacy and safety of canagliflozin versus glimepiride in patients with type 2 diabetes inadequately controlled with metformin (CANTATA-SU): 52 week results from a randomised, double-blind, phase 3 non-inferiority trial.
      ,
      • Lavalle-González F.J.
      • Januszewicz A.
      • Davidson J.
      • et al.
      Efficacy and safety of canagliflozin compared with placebo and sitagliptin in patients with type 2 diabetes on background metformin monotherapy: a randomised trial.
      ,
      • Schernthaner G.
      • Gross J.L.
      • Rosenstock J.
      • et al.
      Canagliflozin compared with sitagliptin for patients with type 2 diabetes who do not have adequate glycemic control with metformin plus sulfonylurea: a 52-week, randomized trial.
      ,
      • Stenlöf K.
      • Cefalu W.T.
      • Kim K.A.
      • et al.
      Long-term efficacy and safety of canagliflozin monotherapy in patients with type 2 diabetes inadequately controlled with diet and exercise: findings from the 52-week CANTATA-M study.
      ,
      • Stenlöf K.
      • Cefalu W.T.
      • Kim K.-A.
      • et al.
      Efficacy and safety of canagliflozin monotherapy in subjects with type 2 diabetes mellitus inadequately controlled with diet and exercise.
      ,
      • Wilding J.P.
      • Charpentier G.
      • Hollander P.
      • et al.
      Efficacy and safety of canagliflozin in patients with type 2 diabetes mellitus inadequately controlled with metformin and sulphonylurea: a randomised trial.
      ,
      • Yale J.F.
      • Bakris G.
      • Cariou B.
      • et al.
      Efficacy and safety of canagliflozin in subjects with type 2 diabetes and chronic kidney disease.
      ,
      • Forst T.
      • Guthrie R.
      • Goldenberg R.
      • et al.
      Efficacy and safety of canagliflozin over 52 weeks in patients with type 2 diabetes on background metformin and pioglitazone.
      ,
      • Cefalu W.T.
      • Leiter L.A.
      • Yoon K.-H.
      • et al.
      Canagliflozin demonstrates durable glycemic improvements over 104 weeks versus glimepiride in subjects with type 2 diabetes mellitus on metformin.
      ]. Canagliflozin leads to inhibition of glucose reabsorption and increased urinary glucose excretion, thereby reducing blood glucose, body weight (predominantly due to fat loss), and SBP (from weight loss and mild osmotic diuresis), with a low risk of hypoglycemia, which can be a limiting factor for achieving treatment goals [
      American Diabetes Association
      Standards of medical care in diabetes–2014.
      ]. This insulin-independent mechanism differentiates canagliflozin from other classes of antihyperglycemic agents (AHAs), such as dipeptidyl peptidase-4 inhibitors, including sitagliptin, which act directly on β cells to lower blood glucose.
      The present analysis is based on results from a clinical study that directly compared canagliflozin 300 and 100 mg versus sitagliptin 100 mg in dual therapy with metformin in patients with T2DM [
      • Lavalle-González F.J.
      • Januszewicz A.
      • Davidson J.
      • et al.
      Efficacy and safety of canagliflozin compared with placebo and sitagliptin in patients with type 2 diabetes on background metformin monotherapy: a randomised trial.
      ]. This was a randomized, double-blind, four-arm, parallel-group, placebo- and active-controlled, phase III study. Change in HbA1c from baseline to week 52 was a key end point, with a hypothesis that canagliflozin 300 mg or both doses would demonstrate noninferiority in lowering HbA1c versus sitagliptin 100 mg. In the clinical study, canagliflozin 300 mg demonstrated superiority and canagliflozin 100 mg demonstrated noninferiority compared with sitagliptin 100 mg in lowering HbA1c at 52 weeks (–0.88%, –0.73%, and –0.73%, respectively). Canagliflozin 300 and 100 mg also provided reductions compared with sitagliptin 100 mg in body weight (–4.2%, –3.8%, and –1.3%, respectively) and SBP (–4.7, –3.5, and –0.7 mmHg, respectively). Both doses of canagliflozin were generally well tolerated. Although the incidences of adverse events (AEs) potentially related to the mechanism of SGLT2 inhibition, such as male and female genital mycotic infections (e.g., yeast infections), osmotic diuresis–related AEs (e.g., pollakiuria, polyuria, and nocturia), and volume depletion–related AEs (e.g., orthostatic hypotension and postural dizziness), were higher with both canagliflozin doses than with sitagliptin in the study, AE-related discontinuation rates were similar across treatment groups.
      Because T2DM is chronic and progressive, the costs and health benefits of interventions are fully realized only over long time horizons. Ideally, therefore, cost-effectiveness analyses of T2DM interventions would be informed by long-term, naturalistic, randomized clinical trials [
      • Palmer A.J.
      The Mount Hood 5 Modeling Group. Computer modeling of diabetes and its complications: a report on the fifth Mount Hood Challenge meeting.
      ,
      American Diabetes Association
      Guidelines for computer modeling of diabetes and its complications.
      ]. Clinical trials of sufficient duration, however, are rarely (if ever) available at the time that initial coverage decisions are made. As such, economic computer modeling that extrapolates the available clinical trial data to long-term health economic outcomes has been widely accepted as a way to assess the cost-effectiveness of alternative T2DM treatment strategies [
      • Palmer A.J.
      The Mount Hood 5 Modeling Group. Computer modeling of diabetes and its complications: a report on the fifth Mount Hood Challenge meeting.
      ,
      American Diabetes Association
      Guidelines for computer modeling of diabetes and its complications.
      ].
      Given the growing economic burden of T2DM in Latin America and specifically in Mexico, cost-effectiveness evaluations can inform decisions about the efficient allocation of limited health care resources. Mexico’s independent health technology assessment body, Centro Nacional de Excelencia Tecnológica en Salud, encourages the use of cost-effectiveness analysis and states a willingness-to-pay (WTP) threshold of 1 times the gross domestic product (GDP) per capita as “very cost-effective” (MXP 141,120 or US $10,500; exchange rate as of September 26, 2014 of US $1 = MXP 13.44) [

      Mexican Peso exchange rate - US dollar - Historical exchange rates. 2014. Available from: http://www.exchange-rates.org/history/MXN/USD/T. [Accessed October 22, 2014].

      ]. Centro Nacional de Excelencia Tecnológica en Salud further states that for treatments with costs per QALY gained of ≥1 and ≤3 times the GDP per capita, a detailed analysis should be performed; those with costs per QALY gained of >3 times the GDP per capita should not be considered “cost-effective” [

      Quintiles Consulting, van Engen A, Heemstra L. The challenges of introducing HTA in Latin America. HTA Uncovered 2014;(5):1–9.

      ,
      • Rizo Rios P.
      • Gonzalez Rivera A.
      • Rivas Oropeza I.
      • Rivas Bocanegra R.E.
      Análisis del Proceso de Actualización de Medicamentos en el Cuadro Básico y Catálogo de Insumos del Sector Salud.
      ]. These WTP thresholds are in line with those recommended by the World Health Organization [

      World Health Organization. Cost-effectiveness thresholds. Available from: http://www.who.int/choice/costs/CER_thresholds/en/. [Accessed October 21, 2014].

      ].
      Comparing diabetes treatment alternatives over the long term from the perspective of the Mexican health care system is necessary to direct resources in the most efficient manner, enabling better patient outcomes from available resources. In this study, the cost-effectiveness of adding canagliflozin 300 or 100 mg versus sitagliptin 100 mg in patients with T2DM inadequately controlled on metformin monotherapy was determined using the 1 and 3 times the GDP per-capita thresholds. The incremental cost-effectiveness ratios (ICERs) were estimated using a validated microsimulation model, Economic and Health Outcomes (ECHO)-T2DM, with local cost data [
      • Willis M.
      • Asseburg C.
      • He J.
      Validation of economic and health outcomes simulation model of type 2 diabetes mellitus (ECHO-T2DM).
      ].

      Methods

       Model Overview and Simulation Description

      ECHO-T2DM is a stochastic microsimulation (patient-level) cost-effectiveness model of the treatment of T2DM (see Fig. 1 for a diagrammatic overview) [
      • Willis M.
      • Asseburg C.
      • He J.
      Validation of economic and health outcomes simulation model of type 2 diabetes mellitus (ECHO-T2DM).
      ]. The physiology of T2DM is captured using Markov health states for microvascular and macrovascular complications and death. The cycle length is 1 year, and the time horizon is defined by the user. ECHO-T2DM accounts explicitly for both first-order uncertainty (associated with interpatient variability) and second-order uncertainty (uncertainty regarding the true value of the underlying parameters) and is programmed in R using user-friendly front- and back-end Excel interfaces. Because of space limitations, technical details including a conceptual walk through, parameterization of macrovascular and microvascular complications (i.e., chronic kidney disease, neuropathy, retinopathy), and parameters related to uncertainty and heterogeneity can be found in the Appendix in Supplemental Materials found at doi:10.1016/j.jval.2014.12.017.
      Figure thumbnail gr1
      Fig. 1 ECHO-T2DM model. BDR, background diabetic retinopathy; CHF, congestive heart failure; CVD, cardiovascular disease; ECHO-T2DM, Economic and Health Outcomes model of type 2 diabetes mellitus; eGFR, estimated glomerular filtration rate; HDL-C, high-density lipoprotein cholesterol; IHD, ischemic heart disease; LDL-C, low-density lipoprotein cholesterol; LEA, lower extremity amputation; ME, macular edema; MI, myocardial infarction; PDR, proliferative diabetic retinopathy; PVD, peripheral vascular disease; QALY, quality-adjusted life-year; SBP, systolic blood pressure; TGR, triglycerides; UKPDS, UK Prospective Diabetes Study.
      Cohorts of hypothetical patients are generated at the start of the simulation. Each patient is defined by age, sex, disease duration, HbA1c, biomarker values, smoking status, and preexisting health conditions (micro- and macrovascular disease). Biomarker values at the individual level tend to be correlated; for example, the clustering of poor glycemic control, hypertension, dyslipidemia, and overweight affected 41.6% of the Mexican population in 2006 [
      • Rojas R.
      • Aguilar-Salinas C.
      • Jimenez-Corona A.
      • et al.
      Metabolic syndrome in Mexican adults: results from the National Health and Nutrition Survey 2006.
      ]. As such, ECHO-T2DM samples an individual patient’s biomarkers, accounting for key correlations between the biomarkers themselves (and age) explicitly [
      • Brown J.B.
      • Russell A.
      • Chan W.
      • et al.
      The global diabetes model: user friendly version 3.0.
      ].
      ECHO-T2DM includes a comprehensive set of health complications. Microvascular health states reflect increasing severity of chronic kidney disease (tracked by both the albumin-to-creatinine ratio and the estimated glomerular filtration rate [eGFR]), retinopathy, and neuropathy (see Appendix Sections 2.1, 2.2, and 2.3 in Supplemental Materials found at doi:10.1016/j.jval.2014.12.017). HbA1c- and T2DM-duration–specific progression rates steer transition between the different health states, and are largely taken from previous modeling work [
      • Bagust A.
      • Beale S.
      Modelling EuroQol health-related utility values for diabetic complications from CODE-2 data.
      ,
      • Eastman R.C.
      • Javitt J.C.
      • Herman W.H.
      • et al.
      Model of complications of NIDDM, II: analysis of the health benefits and cost-effectiveness of treating NIDDM with the goal of normoglycemia.
      ,
      • Hoerger T.J.
      • Wittenborn J.S.
      • Segel J.E.
      • et al.
      A health policy model of CKD, 1: model construction, assumptions, and validation of health consequences.
      ]. Macrovascular health states consist of ischemic heart disease, myocardial infarction, stroke, and congestive heart failure; myocardial infarction and stroke can occur multiple times. ECHO-T2DM includes four different sets of macrovascular risk equations, including both the new and the original UK Prospective Diabetes Study risk functions [
      • Clarke P.M.
      • Gray A.M.
      • Briggs A.
      • et al.
      A model to estimate the lifetime health outcomes of patients with type 2 diabetes: the United Kingdom Prospective Diabetes Study (UKPDS) Outcomes Model (UKPDS no. 68).
      ,
      • Hayes A.J.
      • Leal J.
      • Gray A.M.
      • et al.
      UKPDS outcomes model 2: a new version of a model to simulate lifetime health outcomes of patients with type 2 diabetes mellitus using data from the 30 year United Kingdom Prospective Diabetes Study: UKPDS 82.
      ], the Action in Diabetes and Vascular Disease: Preterax and Diamicron Modified-release Controlled Evaluation (ADVANCE) risk function, and the Swedish National Diabetes Registry risk function [
      • Zethelius B.
      • Eliasson B.
      • Eeg-Olofsson K.
      • et al.
      A new model for 5-year risk of cardiovascular disease in type 2 diabetes, from the Swedish National Diabetes Register (NDR).
      ]. The new UK Prospective Diabetes Study risk functions (UK Prospective Diabetes Study 82 [
      • Hayes A.J.
      • Leal J.
      • Gray A.M.
      • et al.
      UKPDS outcomes model 2: a new version of a model to simulate lifetime health outcomes of patients with type 2 diabetes mellitus using data from the 30 year United Kingdom Prospective Diabetes Study: UKPDS 82.
      ]) were used in this application to model macrovascular (and mortality) risks because they include separate risks for second events for some complications and because the developers suggest they do not have the tendency to overpredict risks, as was seen with the original set (likely due, in part, to the doubling of follow-up to >20 years). Mortality consists of four equations covering general mortality; case fatality in the year following a primary event (e.g., stroke); excess mortality risk related to a history of diabetes complications such as myocardial infarction, stroke, and end-stage renal disease; and excess mortality risk in the years following an event [
      • Hayes A.J.
      • Leal J.
      • Gray A.M.
      • et al.
      UKPDS outcomes model 2: a new version of a model to simulate lifetime health outcomes of patients with type 2 diabetes mellitus using data from the 30 year United Kingdom Prospective Diabetes Study: UKPDS 82.
      ].
      The algorithm for drug intensification in ECHO-T2DM is designed to maintain HbA1c target values, including an option for contraindication and discontinuation due to health conditions (including, e.g., end-stage renal disease and macrovascular disease). Failure to achieve the glycemic target can lead to dose increases (if available) or the addition of new agents (previous agents can be continued or discontinued at failure).
      Patient biomarker values (i.e., HbA1c, SBP, body mass index [BMI], and lipids) are updated each cycle and incorporate annualized drug-specific “drifts,” which capture deterioration in these biomarkers over time, net of the treatment effects [
      UK Prospective Diabetes Study (UKPDS) Group
      Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33).
      ,
      • Chaudhry Z.W.
      • Gannon M.C.
      • Nuttall F.Q.
      Stability of body weight in type 2 diabetes.
      ]. To account for the observed changes in the efficacy of agents with SGLT2 activity (such as canagliflozin) with declining eGFR, HbA1c-, SBP-, and BMI-lowering can be adjusted. Specifically in these analyses, when eGFR was simulated to drop below 60 mL/min/1.73 m2, the treatment effects of canagliflozin on these biomarkers were adjusted to correspond to estimates obtained from a pooled analysis of subjects in the clinical trial program with moderate renal impairment (eGFR ≥45 and <60 mL/min/1.73 m2) [

      Janssen Research & Development. Integrated summary of efficacy: canagliflozin as an adjunctive treatment to diet and exercise alone or co-administered with other antihyperglycemic agents to improve glycemic control in adults with type 2 diabetes mellitus. 2012.

      ]. Treatment with canagliflozin (though not with sitagliptin) was discontinued if eGFR fell below 30 mL/min/1.73 m2 per the Mexican label at the time of the analysis [

      INVOKANA. Informacion Para Prescribir Amplia INVOKANA® F.F Tableta V1.0 CCDS Junio. 2012.

      ].
      Although neither canagliflozin nor sitagliptin is associated with an increased risk of hypoglycemia on the basis of their mechanism of action [
      INVOKANA® (canagliflozin) tablets, for oral use [package insert].
      ,
      JANUVIA® (sitagliptin) tablets [package insert].
      ], nonsevere symptomatic hypoglycemic episodes and severe hypoglycemic episodes (i.e., those requiring third-party assistance) were modeled using data from the clinical study [
      • Lavalle-González F.J.
      • Januszewicz A.
      • Davidson J.
      • et al.
      Efficacy and safety of canagliflozin compared with placebo and sitagliptin in patients with type 2 diabetes on background metformin monotherapy: a randomised trial.
      ]. Data from the clinical study [
      • Lavalle-González F.J.
      • Januszewicz A.
      • Davidson J.
      • et al.
      Efficacy and safety of canagliflozin compared with placebo and sitagliptin in patients with type 2 diabetes on background metformin monotherapy: a randomised trial.
      ] were also used as model inputs for AEs potentially related to the mechanism of action of SGLT2 inhibition, including male and female genital mycotic infections (i.e., yeast infections) [
      • Nyirjesy P.
      • Sobel J.D.
      • Fung A.
      • et al.
      Genital mycotic infections with canagliflozin, a sodium glucose co-transporter 2 inhibitor, in patients with type 2 diabetes mellitus: a pooled analysis of clinical studies.
      ], lower and upper urinary tract infections [
      • Nicolle L.E.
      • Capuano G.
      • Fung A.
      • Usiskin K.
      Urinary tract infection in randomized phase III studies of canagliflozin, a sodium glucose co-transporter 2 inhibitor.
      ], and osmotic diuresis–related AEs and volume depletion–related AEs [
      • Weir M.
      • Januszewicz A.
      • Gilbert R.
      • et al.
      Effect of canagliflozin on blood pressure and adverse events related to osmotic diuresis and reduced intravascular volume in patients with type 2 diabetes mellitus.
      ]. The risk of volume depletion–related AEs was adjusted using the pooled renal impairment data for those with an eGFR of less than 60 mL/min/1.73 m2, for those aged 75 years or older, and for those on a loop diuretic [

      Janssen Research & Development. Integrated summary of safety. Canagliflozin as an adjunctive treatment to diet and exercise alone or co-administered with other antihyperglycemic agents to improve glycemic control in adults with type 2 diabetes mellitus. Raritan, NJ: Janssen Research & Development, 2012.

      ]. Note that AEs potentially associated with the mechanism of action of dipeptidyl peptidase-4 inhibitors (e.g., pancreatitis [
      • Meier J.J.
      • Nauck M.A.
      Risk of pancreatitis in patients treated with incretin-based therapies.
      ]) were not modeled.
      Antihypertension and antidyslipidemia treatment algorithms based on meeting blood pressure and lipid targets can also be applied. The hypothetical patients in the present analysis were treated to an SBP target of 140 mmHg and an LDL-C target of 100 mg/dL using a treatment algorithm consistent with Mexican guidelines [

      Instituto Mexicano del Seguro Social. Diagnostico y Tratamiento de la Diabetes Mellitus Tipo 2. Guía de Referencia Rápida. México: Instituto Mexicano del Seguro Social, 2009. Actualización 2012. Instituto Mexicano del Seguro Social 2012. Available from: http://www.imss.gob.mx/sites/all/statics/guiasclinicas/000GRR_DiabetesMellitus.pdf. [Accessed May 27, 2014].

      ], with treatment effect data obtained from the literature [
      • Law M.R.
      • Morris J.K.
      • Wald N.J.
      Use of blood pressure lowering drugs in the prevention of cardiovascular disease: meta-analysis of 147 randomised trials in the context of expectations from prospective epidemiological studies.
      ,
      • Brown W.V.
      • Bays H.E.
      • Hassman D.R.
      • et al.
      Efficacy and safety of rosuvastatin compared with pravastatin and simvastatin in patients with hypercholesterolemia: a randomized, double-blind, 52-week trial.
      ,
      • Wolffenbuttel B.H.
      • Franken A.A.
      • Vincent H.H.
      Cholesterol-lowering effects of rosuvastatin compared with atorvastatin in patients with type 2 diabetes -- CORALL study.
      ].
      Costs can be assigned for treatment interventions and AEs, initial-year medical events, and resources related to the disease history and follow-up. Patient preferences for health are captured using quality-adjusted life-year (QALY) disutility weights (i.e., decrements in quality of life associated with a negative impact of particular health states) for patients’ demographic characteristics (including age and sex), clinical characteristics (including disease duration and BMI), the presence of individual microvascular and macrovascular complications, hypoglycemia, and AEs.
      Outcomes include the cumulative incidence of each type of health outcome and relative risk reductions in microvascular and macrovascular events, life-years and QALYs, biomarker evolution curves, and cost and cost-effectiveness metrics.
      The external (predictive) validity of ECHO-T2DM has been assessed by replicating the design of a number of clinical trials and simulating various outcomes over the respective time horizons, and then comparing model predictions with the observed trial results (see Willis et al. [
      • Willis M.
      • Asseburg C.
      • He J.
      Validation of economic and health outcomes simulation model of type 2 diabetes mellitus (ECHO-T2DM).
      ] for a full description). Results demonstrated that ECHO-T2DM simulated patient outcomes for various microvascular and macrovascular outcomes observed in important clinical trials with an accuracy that is consistent with that of other well-accepted models.

       Simulation Parameters

      In the base case, a total of 1000 cohorts of 2000 hypothetical patients with T2DM (i.e., 2 million unique hypothetical patients) were randomly generated and simulated over 20 years to ensure that long-term costs and benefits of treatment intervention in chronic and progressive T2DM were captured. The large number of patients per cohort ensures that the first-order uncertainty is small, and the large number of cohorts ensures that the second-order uncertainty is captured. All future costs and health benefits (i.e., life-years, QALYs) were discounted at 5% per Centro Nacional de Excelencia Tecnológica en Salud guidelines [

      Centro Nacional de Excelencia Tecnológica en Salud. Evaluación de Tecnologías de la Salud, Documento metodológico. Available from: http://www.cenetec.salud.gob.mx/descargas/detes/Documento_metodologico_ETES.pdf. [Accessed October 22, 2014].

      ]. In this analysis, cost-effectiveness was assessed using the WTP thresholds of 1 times GDP per capita (i.e., MXP 141,200/QALY) and 3 times GDP per capita (i.e., MXP 423,600/QALY) suggested by the Mexican government and the World Health Organization [
      • Ríos P.R.
      • Rivera A.G.
      • Oropeza I.R.
      Bocanegra RER. Análisis del Proceso de Actualización de Medicamentos en el Cuadro Básico y Catálogo de Insumos del Sector Salud.
      ]. Note that an ICER below the more conservative threshold of 1 times GDP per QALY is considered to be very cost-effective per the World Health Organization.

       Patient Profiles

      Baseline demographic and disease characteristics reflecting the distribution of those from the head-to-head study of canagliflozin versus sitagliptin described above [
      • Lavalle-González F.J.
      • Januszewicz A.
      • Davidson J.
      • et al.
      Efficacy and safety of canagliflozin compared with placebo and sitagliptin in patients with type 2 diabetes on background metformin monotherapy: a randomised trial.
      ] were randomly generated for each hypothetical patient (Table 1).
      Table 1Key baseline patient characteristics and clinical inputs for canagliflozin 300 and 100 mg and sitagliptin 100 mg.
      All data are from the clinical study unless otherwise noted [20].
      Characteristic
      Data are mean ± SD from baseline to week 52 (modified intent-to-treat) unless otherwise indicated.
      Overall population total (N = 1284)Latin American sample total (N = 240)
      Age (y)55.4 ± 9.454.8 ± 9.2
      Sex (%)
       Male47.136.1
       Female52.963.9
      Duration of T2DM (y)6.9 ± 5.37.7 ± 5.6
      HbA1c (%)7.9 ± 0.97.9 ± 0.9
      Total cholesterol (mg/dL)189.8 ± 39.9196.2 ± 35.5
      LDL-C (mg/dL)107.9 ± 33.6112.3 ± 30.5
      HDL-C (mg/dL)45.3 ± 11.245.2 ± 11.4
      SBP (mmHg)128.2 ± 13.0124.4 ± 13.4
      BMI (kg/m2)31.8 ± 6.230.7 ± 5.5
      Current smoker (%)12.57.5
      Atrial fibrillation (%)0.80
      Background diabetic retinopathy (%)6.62.9
      Macular edema (%)0.10
      Proliferative diabetic retinopathy (%)0.10
      Microalbuminuria (%)0.51.1
      Macroalbuminuria (%)0.40
      Symptomatic neuropathy (%)2.60.4
      Treatment effects
      Data are least squares mean changes from baseline to week 52 (modified intent-to-treat) unless otherwise indicated.
      Overall populationLatin American sample
      CANA 300 mg (n = 367)CANA 100 mg (n = 368)SITA 100 mg (n = 366)CANA 300 mg (n = 79)CANA 100 mg (n = 80)SITA 100 mg (n = 81)
      HbA1c (%)–0.88–0.73–0.73–0.94–0.95–0.73
      Total cholesterol (mg/dL)7.086.113.378.423.6311.23
      LDL-C (mg/dL)4.404.333.125.634.259.29
      HDL-C (mg/dL)5.514.502.365.565.132.32
      Triglycerides (mg/dL)–16.61–11.05–13.51–9.22–27.50–6.36
      SBP (mmHg)–4.65–3.53–0.66–4.36–3.61–0.40
      BMI (kg/m2)
      Some weight loss was seen in the sitagliptin arm because all patients followed a regimen of diet and exercise in addition to pharmacologic treatment (a requirement in registration trials in diabetes).
      –1.35–1.22–0.45–1.31–1.11–0.22
      Annual “drift” in HbA1c (%)
      Sourced from the literature [6,44].
      0.140.140.140.140.140.14
      Annual “drift” in SBP (mmHg)
      Sourced from the literature [6,44].
      0.30.30.30.30.30.3
      Annual “drift” in lipid parameters (mg/dL)
      Sourced from the literature [6,44].
      0.030.030.030.030.030.03
      AEs, per patient-year of exposure (%)
       Severe hypoglycemia00.30.301.41.4
       Nonsevere symptomatic hypoglycemia8.58.84.415.827.19.7
       Volume depletion–related AEs1.31.32.71.41.41.4
       Osmotic diuresis–related AEs4.18.51.04.36.61.4
       Male genital mycotic infection2.75.91.310.616.02.8
       Female genital mycotic infection11.112.93.019.015.82.4
       Lower UTI7.49.37.812.617.211.8
       Upper UTI0.30.60.3001.3
      AE, adverse event; BMI, body mass index; CANA, canagliflozin; HbA1c, glycated hemoglobin; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; SBP, systolic blood pressure; SD, standard deviation; SITA, sitagliptin; T2DM, type 2 diabetes mellitus; UTI, urinary tract infection.
      low asterisk All data are from the clinical study unless otherwise noted
      • Lavalle-González F.J.
      • Januszewicz A.
      • Davidson J.
      • et al.
      Efficacy and safety of canagliflozin compared with placebo and sitagliptin in patients with type 2 diabetes on background metformin monotherapy: a randomised trial.
      .
      Data are mean ± SD from baseline to week 52 (modified intent-to-treat) unless otherwise indicated.
      Data are least squares mean changes from baseline to week 52 (modified intent-to-treat) unless otherwise indicated.
      § Some weight loss was seen in the sitagliptin arm because all patients followed a regimen of diet and exercise in addition to pharmacologic treatment (a requirement in registration trials in diabetes).
      || Sourced from the literature [
      UK Prospective Diabetes Study (UKPDS) Group
      Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33).
      ,
      • Chaudhry Z.W.
      • Gannon M.C.
      • Nuttall F.Q.
      Stability of body weight in type 2 diabetes.
      ].

       Treatment Comparisons and Algorithm

      Two treatment comparisons were simulated: (1) canagliflozin 300 mg versus sitagliptin 100 mg and (2) canagliflozin 100 mg versus sitagliptin 100 mg added to maximally tolerated metformin. Treatment effects for canagliflozin 300 and 100 mg and sitagliptin 100 mg were obtained from the clinical trial data (Table 1; see Appendix Table 5 in Supplemental Materials found at doi:10.1016/j.jval.2014.12.017). Treatment effects (i.e., HbA1c, weight gain) and hypoglycemia incidences for basal and prandial insulin were sourced from the literature [
      • Rosenstock J.
      • Davies M.
      • Home P.D.
      • et al.
      A randomised, 52-week, treat-to-target trial comparing insulin detemir with insulin glargine when administered as add-on to glucose-lowering drugs in insulin-naive people with type 2 diabetes.
      ,
      • Riddle M.C.
      • Vlajnic A.
      • Zhou R.
      • Rosenstock J.
      Baseline HbA1c predicts attainment of 7.0% HbA1c target with structured titration of insulin glargine in type 2 diabetes: a patient-level analysis of 12 studies.
      ,
      • Fonseca V.
      • Gill J.
      • Zhou R.
      • Leahy J.
      An analysis of early insulin glargine added to metformin with or without sulfonylurea: impact on glycaemic control and hypoglycaemia.
      ]. Details of the AHA treatment intensification algorithm applied in these simulations are given in Table 2.
      Table 2AHA treatment intensification algorithm
      • Hypothetical patients entered the model uncontrolled on background metformin and were assigned to treatment with canagliflozin 300 or 100 mg or sitagliptin 100 mg.
      • Patients meeting the HbA1c target level of <7.0% continued on their assigned treatment.
      • Patients were discontinued from treatment with canagliflozin or sitagliptin because of AEs in the first cycle at the rates observed in the clinical trial. Otherwise, patients remained on their assigned treatment unless the HbA1c level was ≥7.0% or eGFR was <30 mL/min/1.73 m2 (eGFR threshold applied only to canagliflozin).
      • When the HbA1c level was ≥7.0%, AHA treatment was intensified by adding basal insulin (i.e., neutral protamine Hagedorn [NPH]) insulin), initially at 10 IU/d; the dose could be titrated up to 60 IU/d as needed to maintain HbA1c <7.0%.
      • When the HbA1c level was ≥7.0% at the maximum dose of basal insulin, prandial insulin was added (and titrated as required to maintain control up to 200 IU/d).
      AE, adverse event; AHA, antihyperglycemic agent; eGFR, estimated glomerular filtration rate; HbA1c, glycated hemoglobin; NPH, neutral protamine Hagedorn.

       Costs

      The costs of myocardial infarction, stroke, and lower extremity amputation were sourced from the diagnostic-related groups issued by the IMSS in 2012 (Table 3). Costs for macular edema, proliferative retinopathy, and blindness were obtained from the Mexican health care administration agency, the Instituto de Seguridad y Servicios Sociales de los Trabajadores del Estado, the costing table, and the IMSS. Costs for end-stage renal disease were obtained from literature searches of studies and articles related to diabetes [
      • Lopez-Cervantes M.
      • Rojas-Russell M.E.
      • Tirado-Gomez L.L.
      • et al.
      Enfermedad renal crónica y su atención mediante tratamiento sustitutivo en México.
      ]. The remaining cost inputs were obtained from interviews that were conducted with clinicians in public Mexican hospitals to assess how most patients would be treated and mapping to the Instituto de Seguridad y Servicios Sociales de los Trabajadores del Estado costing table [

      Instituto de Seguridad y Servivios Sociales de los Trabajadores del Estado. Tabulador de no derechohabientes 2012, Dirección Médica, Subdirección de Regulación y Atención Hospitalaria, Jefatura de Servicios, Instituto de Seguridad y Servicios Sociales de los Trabajadores del Estado. ISSSTE, 2012.

      ,

      Costos directos de complicaciones relacionadas con la diabetes en México. Internal report by IMS. 2013.

      ]. Resource utilization was matched with cost data using several sources as follows. Costs for laboratory studies, physician visits, hospitalizations, intensive care, and emergency visits were matched to those from established sources [

      Instituto Nacional de Cancerología INCAN. Tabulador de Cuotas de Recuperación, Subsecretaría de Ingresos, Unidad de política de ingresos. 2012. Available from: http://www.policiabancaria.df.gob.mx/portal_pbi/articulo15/fraccion8/2012/24%20Instituto%20Nacional%20de%20Cancerolog%C3%ADa%20POLICIA%20BANCARIA%20INCAN%202012.pdf. [Accessed May 30, 2014].

      ,

      Hospital General de Mexico. Tabulador de Cuotas de Recuperación, 2012. 2012. Available from: http://www.policiabancaria.df.gob.mx/portal_pbi/articulo15/fraccion8/2012/19%20Hospital%20General%20de%20M%C3%A9xico%202012.pdf. [Accessed May 30, 2014].

      ]. Unit costs of medications used to treat comorbidities were sourced from the IMSS and the Diario Oficial de la Federación [

      Comisión Interinstitucional del Cuadro Básico de Insumos del Sector Salud. Cuadro básico y catálogo de medicamentos. 2012. Available from: http://www.dof.gob.mx/nota_detalle.php?codigo=5299596&fecha=21/05/2013. [Accessed May 30, 2014].

      ,

      Instituto Mexicano del Seguro Social. Portal de compras del IMSS. Instituto Mexicano del Seguro Social. 2014. Available from: http://compras.imss.gob.mx. [Accessed May 30, 2014].

      ]. The unit prices for canagliflozin 300 and 100 mg and sitagliptin 100 mg were specified as MXP 27.11, MXP 27.11, and MXP 26.07 per day, respectively. These prices were sourced from a local retail pharmacy (i.e., Walmart) because they are not yet widely available in the institutional formularies of the public sector, and represent the cost that would be paid out of pocket. All costs were converted to 2012 MXP using the National Consumer Price Index released by the Mexican government [

      Indice Nacional de Precios al Consumidor, Servicio de Administración Tributaria. 2014. Available from: http://www.sat.gob.mx/informacion_fiscal/tablas_indicadores/Paginas/inpc_2012.aspx. [Accessed May 30, 2014].

      ].
      Table 3Event-year and state costs and QALY utility weight inputs
      Clinical eventEvent-year cost
      Event costs are associated with management of the acute episode and any subsequent care in the first year.
      (MXP)
      State, annualized cost (MXP)QALY utility weights
      Macrovascular
       Ischemic heart disease39,81616,898–0.028 [
      • Bagust A.
      • Beale S.
      Modelling EuroQol health-related utility values for diabetic complications from CODE-2 data.
      ]
       Myocardial infarction172,87216,898–0.028 [
      • Bagust A.
      • Beale S.
      Modelling EuroQol health-related utility values for diabetic complications from CODE-2 data.
      ]
       Congestive heart failure27,46911,991–0.028 [
      • Bagust A.
      • Beale S.
      Modelling EuroQol health-related utility values for diabetic complications from CODE-2 data.
      ]
       Stroke54,83516,313–0.115 [
      • Bagust A.
      • Beale S.
      Modelling EuroQol health-related utility values for diabetic complications from CODE-2 data.
      ]
       Peripheral vascular disease26,29023,669–0.061 [
      • Bagust A.
      • Beale S.
      Modelling EuroQol health-related utility values for diabetic complications from CODE-2 data.
      ]
      Microvascular
       Microalbuminuria723410 [
      • Bagust A.
      • Beale S.
      Modelling EuroQol health-related utility values for diabetic complications from CODE-2 data.
      ]
       Macroalbuminuria72341–0.048 [
      • Bagust A.
      • Beale S.
      Modelling EuroQol health-related utility values for diabetic complications from CODE-2 data.
      ]
       eGFR <15 mL/min/1.73 m2 (1st year)0NA–0.175 [
      • Bagust A.
      • Beale S.
      Modelling EuroQol health-related utility values for diabetic complications from CODE-2 data.
      ]
       eGFR <15 mL/min/1.73 m2 (end-stage renal disease)179,003170,923–0.175 [
      • Bagust A.
      • Beale S.
      Modelling EuroQol health-related utility values for diabetic complications from CODE-2 data.
      ]
       Macular edema1,307652NA
       Proliferative retinopathy5,5466520 [
      • Bagust A.
      • Beale S.
      Modelling EuroQol health-related utility values for diabetic complications from CODE-2 data.
      ]
       Blindness in both eyes14,231652–0.057 [
      • Bagust A.
      • Beale S.
      Modelling EuroQol health-related utility values for diabetic complications from CODE-2 data.
      ]
       Symptomatic neuropathy15,9607,980–0.084 [
      • Bagust A.
      • Beale S.
      Modelling EuroQol health-related utility values for diabetic complications from CODE-2 data.
      ]
       Diabetic foot ulcer7,745NA–0.17 [
      • Bagust A.
      • Beale S.
      Modelling EuroQol health-related utility values for diabetic complications from CODE-2 data.
      ]
       Lower extremity amputation74,373NA–0.272 [
      • Bagust A.
      • Beale S.
      Modelling EuroQol health-related utility values for diabetic complications from CODE-2 data.
      ]
      Hypoglycemia
       Nonsevere symptomatic hypoglycemic event0NA–0.005 [
      • Evans M.
      • Khunti K.
      • Mamdani M.
      • et al.
      Health-related quality of life associated with daytime and nocturnal hypoglycaemic events: a time trade-off survey in five countries.
      ]
       Severe hypoglycemic event2,651NA–0.06 [
      • Evans M.
      • Khunti K.
      • Mamdani M.
      • et al.
      Health-related quality of life associated with daytime and nocturnal hypoglycaemic events: a time trade-off survey in five countries.
      ]
      Overweight/obesity
       For each 1 kg/m2 over 25 kg/m2NANA–0.0171 [
      • Lane S.
      • Levy A.R.
      • Mukherjee J.
      • et al.
      The impact on utilities of differences in body weight among Canadian patients with type 2 diabetes.
      ]
      AEs
       Lower UTI (male)925NA–0.00123 [
      • Shingler S.
      • Fordham B.
      • Evans M.
      • et al.
      Utilities for treatment-related adverse events in type 2 diabetes.
      ]
       Lower UTI (female)881NA–0.00123 [
      • Shingler S.
      • Fordham B.
      • Evans M.
      • et al.
      Utilities for treatment-related adverse events in type 2 diabetes.
      ]
       Upper UTI (male)6,018NA–0.00729 [
      • Shingler S.
      • Fordham B.
      • Evans M.
      • et al.
      Utilities for treatment-related adverse events in type 2 diabetes.
      ]
       Upper UTI (female)5,883NA–0.00729 [
      • Shingler S.
      • Fordham B.
      • Evans M.
      • et al.
      Utilities for treatment-related adverse events in type 2 diabetes.
      ]
       GMI (male)6,348NA–0.0046 [
      • Shingler S.
      • Fordham B.
      • Evans M.
      • et al.
      Utilities for treatment-related adverse events in type 2 diabetes.
      ]
       GMI (female)1,518NA–0.0046 [
      • Shingler S.
      • Fordham B.
      • Evans M.
      • et al.
      Utilities for treatment-related adverse events in type 2 diabetes.
      ]
       Volume depletion–related AEs2,080NA–0.005 [
      • Evans M.
      • Khunti K.
      • Mamdani M.
      • et al.
      Health-related quality of life associated with daytime and nocturnal hypoglycaemic events: a time trade-off survey in five countries.
      ]
      Assumed equal to the value of a nonsevere symptomatic hypoglycemic event.
       Osmotic diuresis–related AEs1,418NA–0.005 [
      • Coyne K.S.
      • Sexton C.C.
      • Irwin D.E.
      • et al.
      The impact of overactive bladder, incontinence and other lower urinary tract symptoms on quality of life, work productivity, sexuality and emotional well-being in men and women: results from the EPIC study.
      ]
      AE, adverse event; GMI, genital mycotic infection; MXP, Mexican peso; NA, not applicable; QALY, quality-adjusted life-year; UTI, urinary tract infection.
      low asterisk Event costs are associated with management of the acute episode and any subsequent care in the first year.
      Assumed equal to the value of a nonsevere symptomatic hypoglycemic event.

       QALY Disutility

      QALY disutility weights are given in Table 3. Because detailed utility studies do not exist for diabetes in the Mexican population, utility values were obtained from established sources based on other T2DM populations. Values assigned to microvascular and macrovascular complications were obtained from a widely used source in the economic modeling of T2DM [
      • Bagust A.
      • Beale S.
      Modelling EuroQol health-related utility values for diabetic complications from CODE-2 data.
      ], which used multivariate regression techniques to isolate the unique contribution of each health outcome. The study did not estimate disutility estimates for hypoglycemia and AEs. Disutilities for nonsevere symptomatic and severe hypoglycemic events were sourced from a recent time trade-off study in 1603 individuals with T2DM [
      • Evans M.
      • Khunti K.
      • Mamdani M.
      • et al.
      Health-related quality of life associated with daytime and nocturnal hypoglycaemic events: a time trade-off survey in five countries.
      ]. Disutility values for urinary tract infections and genital mycotic infections were taken from a new time trade-off study in T2DM that was, in large part, performed to provide estimates for these selected AEs potentially related to SGLT2 inhibition [
      • Shingler S.
      • Fordham B.
      • Evans M.
      • et al.
      Utilities for treatment-related adverse events in type 2 diabetes.
      ]. The disutility for the group of osmotic diuresis–related AEs was derived from a study of symptoms associated with overactive bladder in the general population [
      • Coyne K.S.
      • Sexton C.C.
      • Irwin D.E.
      • et al.
      The impact of overactive bladder, incontinence and other lower urinary tract symptoms on quality of life, work productivity, sexuality and emotional well-being in men and women: results from the EPIC study.
      ]. The disutility associated with volume depletion–related AEs was assumed to be equal to that of nonsevere symptomatic hypoglycemia because symptoms of these events are somewhat similar. The estimate of the disutility-associated weight gain of 1 BMI unit (kg/m2; applied only when the BMI was >25 kg/m2) was obtained from the most recent article investigating this issue [
      • Lane S.
      • Levy A.R.
      • Mukherjee J.
      • et al.
      The impact on utilities of differences in body weight among Canadian patients with type 2 diabetes.
      ].

       Sensitivity Analyses

      One-way sensitivity analyses were conducted to explore the robustness of the base-case results to scenarios potentially more reflective of Mexico and to key assumptions related to the profile of canagliflozin. These sensitivity analyses are described in logical groups:
      • 1.
        Clinical characteristics, efficacy, and AE rates: set to those obtained from a post hoc subgroup analysis of subjects from Latin America in the clinical trial (SA1; see Table 1, Latin American sample).
      • 2.
        Treatment rules: AHA treatment intensification threshold set to 8.0% (vs. the base-case assumption of 7.0%; SA2), a level that may be more reflective of Mexico.
      • 3.
        Time horizon: shorter time horizon of 10 years (SA3) and longer time horizons of 30 years (SA4) and 40 years (SA5).
      • 4.
        Disutility associated with weight change: more conservative disutility estimate associated with weight gain in individuals with a BMI of more than 25 kg/m2 (0.0061 per unit gained above 25 kg/m2) [
        • Bagust A.
        • Beale S.
        Modelling EuroQol health-related utility values for diabetic complications from CODE-2 data.
        ] (SA6).
      • 5.
        AEs potentially related to SGLT2 inhibition: double associated costs (SA7); double associated QALYs (SA8).
      Note that the number of hypothetical patients simulated in each of the 1000 cohorts was 1000 for the sensitivity analyses (i.e., 1 million hypothetical patients). This has been shown to be of sufficient size to generate stable results (data available on request).

       Outcomes Reporting and Statistical Analyses

      The cumulative incidence rates for each type of microvascular and macrovascular outcome, event rates for AEs, their associated costs (both total and disaggregated), and mean life-years and QALYs experienced over the simulation period were calculated for each treatment arm. Relative risk reductions for canagliflozin versus sitagliptin were generated for microvascular and macrovascular complications; hazard ratios for the AEs and differences for the costs, life-years, and QALYs were also generated. The ICERs were computed, as were scatterplots of cost-effectiveness planes and cost-effectiveness acceptability curves.

      Results

       Base-Case Analyses

      Canagliflozin 300 and 100 mg were associated with improved life expectancy and QALY gains of 0.16 and 0.06, respectively, relative to sitagliptin 100 mg (Table 4). The incremental lifetime costs per patient associated with adding canagliflozin 300 or 100 mg for patients with uncontrolled HbA1c on a background of metformin were estimated at MXP 1797 and MXP 7262, respectively. The resulting ICERs of MXP 11,210 and MXP 128,883 for canagliflozin 300 and 100 mg versus sitagliptin 100 mg, respectively, are below the WTP threshold of 1 times the GDP per capita for a QALY gained, suggesting that both doses are very cost-effective (Table 4).
      Table 4Base case, costs of complications and AEs, average costs, and benefits per person, over 20 y (modeled).
      Outcome parameterCANA 300 mgSITA 100 mgDifferenceCANA 100 mgSITA 100 mgDifference
      Costs (MXP), discounted
       Macrovascular
        Myocardial infarction24,95126,906–95524,42824,787–540
        Ischemic heart disease13,11113,611–50012,92413,306–382
        Congestive heart failure3,7203,915–1953,7383,884–146
        Stroke4,2204,436–2154,1214,285–163
       Microvascular
        Retinopathy9,4559,461–59,4619,44913
        Nephropathy1,5281,539–121,5281,535–7
        Neuropathy35,05435,111–5835,11635,02194
       AHA
        Oral agents155,433150,7814,652152,916150,5732,343
        Insulin56,24258,927–2,68660,49958,8371,662
       Prescription treatment
        Hypoglycemia357428–7146444519
        AEs3,6752,4391,2365,6572,4393,218
        Hypertension3,4954,649–1,1543,6994,645–947
        Dyslipidemia18,84617,0861,76019,21517,1172,099
      Total costs (MXP)330,087328,2901,797333,587326,3247,262
      Health indicators, discounted
       Life-years10.6610.640.0210.6410.630.02
       QALYs6.356.190.166.236.180.06
       Survival (%)
      Percentage alive at simulation end.
      56.856.40.456.556.20.3
      Cost per QALY (MXP)11,210128,883
      AE, adverse event; AHA, antihyperglycemic agent; CANA, canagliflozin; MXP, Mexican peso; QALY, quality-adjusted life-year; SITA, sitagliptin.
      low asterisk Percentage alive at simulation end.
      Because the hypothetical patients were treated with rescue medications to meet HbA1c, SBP, and lipid goals, differences in key biomarkers driving event risks for macrovascular and especially microvascular events were effectively minimized, with the consequence that cost offsets and avoided QALY losses related to these complications were smaller than would have been the case without strict treatment-to-goal rescue medication (see base case results for cumulative incidences and relative risk reductions in Supplemental Table 1 in Supplemental Material found at doi:10.1016/j.jval.2014.12.017 and associated costs in Table 4).
      The greater initial HbA1c lowering for canagliflozin 300 mg translated into a reduced need for insulin rescue therapy, which resulted in cost offsets of MXP –2686 in overall insulin therapy use associated with canagliflozin 300 mg relative to sitagliptin 100 mg. In the simulation of canagliflozin 100 mg versus sitagliptin 100 mg, estimated insulin costs were higher in the canagliflozin arm, owing largely to the discontinuation of canagliflozin (but not sitagliptin) if an eGFR of less than 30 mL/min/1.73 m2 was eventually reached. In addition, decreases in SBP contributed to decreased costs for the treatment of hypertension with canagliflozin 300 and 100 mg (MXP –1154 and –947, respectively).
      As previously mentioned, both doses of canagliflozin generated QALY gains versus sitagliptin. The largest drivers of QALY differences were lower disutility associated with overweight/obesity (–0.129 and –0.097, respectively) and with mortality (–0.017 and –0.012, respectively; see Supplemental Table 2 in Supplemental Materials found at doi:10.1016/j.jval.2014.12.017). For canagliflozin 300 mg, less need for insulin rescue therapy also contributed to reduced disutility associated with hypoglycemic events. For canagliflozin 100 mg, the difference in the estimated disutilities associated with hypoglycemia favored sitagliptin, in part owing to the discontinuation of canagliflozin and start of insulin for patients who reached an eGFR of less than 30 mL/min/1.73 m2. This result may be compounded, moreover, by the likely spurious differences observed in rates of hypoglycemia between canagliflozin 100 mg and sitagliptin 100 mg in the clinical trial; both canagliflozin and sitagliptin have mechanisms of action that have been shown to not independently increase the risk of hypoglycemia [
      INVOKANA® (canagliflozin) tablets, for oral use [package insert].
      ,
      JANUVIA® (sitagliptin) tablets [package insert].
      ]. The disutilities associated with AEs potentially related to the SGLT2 mechanism of action favored sitagliptin, but their impact was small in magnitude (0.003 and 0.006 for canagliflozin 300 and 100 mg vs. sitagliptin 100 mg, respectively).
      The effect of parameter uncertainty on outcomes is illustrated in the scatterplot of the incremental costs and QALYs from each of the 1000 cohorts. Canagliflozin 300 mg was associated with greater QALYs in all 1000 simulations (northeast and southeast quadrants; Fig. 2A), and nearly half of the cohort replications had lower costs as well (southeast quadrant; Fig. 2A). Canagliflozin 100 mg was associated with more QALYs than sitagliptin 100 mg in all but 1 of the 1000 cohort replications (Fig. 2B). Canagliflozin 100 mg was also associated with greater costs in most of the replications, though in a minority of cases it was associated with greater QALYs and lower costs (Fig. 2B). These scatterplots indicate a high degree of confidence that canagliflozin 300 mg is very cost-effective versus sitagliptin 100 mg, which is depicted graphically in Fig. 2C. The likelihood that canagliflozin is cost-effective versus sitagliptin 100 mg naturally increases as the WTP increases (i.e., gains in QALYs become increasingly rewarded), reaching approximately 95% for canagliflozin 300 mg at both specified WTP thresholds. Canagliflozin 100 mg has an approximately 65% probability of being cost-effective versus sitagliptin 100 mg at the 3 times the GDP per-capita threshold versus approximately 50% probability of being very cost-effective at the more conservative WTP threshold (Fig. 2C). The probability that canagliflozin is cost saving is indicated by the probability at a WTP of 0 (the Y axis), and is approximately 37% for the 300-mg dose and approximately 10% for the 100-mg dose.
      Figure thumbnail gr2
      Fig. 2 Cost-effectiveness planes for (A) CANA 300 mg and (B) CANA 100 mg* and (C) cost-effectiveness acceptability curves for CANA 300 and 100 mg over time. CANA, canagliflozin; MXP, Mexican peso; QALY, quality-adjusted life-year; WTP, willingness to pay. *Open circles represent the mean value across cohorts.

       Sensitivity Analyses

      Findings from the sensitivity analyses generally support the findings of the base case that canagliflozin is cost-effective versus sitagliptin. In fact, QALY gains for canagliflozin 300 and 100 mg versus sitagliptin 100 mg increased to 0.19 and 0.27, respectively, from 0.16 and 0.06 in the base case, respectively, in the case in which the Latin American subsample data from the clinical study were used to define clinical characteristics, efficacy, and AE rates. Total cost offsets were MXP –2256 and –8088 with canagliflozin 300 and 100 mg, respectively. Both canagliflozin doses dominated sitagliptin 100 mg (i.e., lower costs and higher QALYs) in this scenario (Table 5).
      Table 5Sensitivity analyses: ICERs.
      ScenarioICERs (LY), MXPICERs (QALY), MXP
      CANA 300 mgCANA 100 mgCANA 300 mgCANA 100 mg
      Base case82,647468,06911,210128,883
      Baseline characteristics
       SA1: Latin American subsampleDominatingDominatingDominatingDominating
      Treatment rules
       SA2: 8.0% HbA1c intensification trigger148,801670,14424,85980,450
      Time horizon
       SA3: 10-y time horizon166,0141,192,91614,83981,149
       SA4: 30-y time horizon63,896462,86512,158205,500
       SA5: 40-y time horizon64,579392,63513,931186,313
      Disutility associated with weight change
       SA6: Alternative disutility estimate for weight loss per BMI unit >25 kg/m282,647468,06923,301Dominated
      AEs potentially related to SGLT2 inhibition
       SA7: Double costs137,6271,166,37018,361195,928
       SA8: Double QALYs80,267804,18210,921154,759
      BMI, body mass index; CANA, canagliflozin; HbA1c, glycated hemoglobin; ICER, incremental cost-effectiveness ratio; LY, life-year; MXP, Mexican peso; QALY, quality-adjusted life-year; SA, sensitivity analysis; SGLT2, sodium glucose co-transporter 2.
      Using the less stringent HbA1c treatment target of less than 8.0%, QALY gains for canagliflozin 300 and 100 mg versus sitagliptin 100 mg were 0.14 and 0.08, respectively, with total cost increases of MXP 3477 and 6521, respectively. In this scenario, both canagliflozin 300 and 100 mg were very cost-effective (Table 5).
      When the time horizon was shortened to 10 years, QALY gains were 0.11 and 0.06 for canagliflozin 300 and 100 mg versus sitagliptin 100 mg, respectively, with total cost increases of MXP 1592 and 4582, respectively. QALY gains were 0.17 and 0.04 for canagliflozin 300 and 100 mg versus sitagliptin 100 mg, respectively, over 30 years, and 0.18 and 0.05, respectively, over 40 years. Cost increases were seen with canagliflozin 300 and 100 mg versus sitagliptin 100 mg of MXP 2076 and 9084, respectively, over 30 years, and MXP 2483 and 9508, respectively, over 40 years. In each of these analyses, canagliflozin 300 mg was very cost-effective and canagliflozin 100 mg was cost-effective at the WTP threshold of 3 times the GDP per capita (Table 5).
      Using a more conservative disutility for overweight/obesity, treatment with canagliflozin 300 mg versus sitagliptin 100 mg resulted in a QALY gain of 0.08 (smaller compared with the base case) and a cost increase of MXP 1797 for canagliflozin 300 mg, compared with a small decrease in QALYs of –0.01 and a cost increase of MXP 7262 with canagliflozin 100 mg versus sitagliptin 100 mg. Thus, in this conservative scenario, canagliflozin 100 mg was dominated by sitagliptin 100 mg, but canagliflozin 300 mg remained very cost-effective (Table 5).
      In the sensitivity analysis that doubled costs of AEs potentially related to the SGLT2 mechanism of action, canagliflozin 300 and 100 mg showed QALY gains of 0.16 and 0.05, respectively, relative to sitagliptin 100 mg; costs were higher by MXP 2964 and 10,353, respectively. When the QALYs associated with AEs potentially related to the SGLT2 mechanism of action were doubled, canagliflozin 300 and 100 mg versus sitagliptin 100 mg resulted in QALY gains of 0.16 and 0.05, respectively, and cost increases of MXP 1729 and 7138, respectively. Canagliflozin 300 mg was very cost-effective when either costs or quality of life associated with SGLT2-related AEs was doubled; canagliflozin 100 mg was cost-effective at the WTP threshold of 3 times the GDP per capita in these analyses (Table 5).

      Discussion and Conclusions

      Economic simulations in the Mexican setting based on the 52-week clinical study data of patients with T2DM on background metformin suggest that the use of canagliflozin 300 and 100 mg is associated with improved health outcomes, associated QALY gains, and a relatively marginal increase in total costs relative to sitagliptin 100 mg. Both canagliflozin doses were found to be cost-effective for patients with T2DM in Mexico, even when considering the most conservative WTP threshold suggested by the Mexican government. The base-case cost-effectiveness estimates of MXP 11,210/QALY gained (US $834) and MXP 128,883/QALY gained (US $9590) for canagliflozin 300 and 100 mg, respectively, fell below this conservative threshold, suggesting that both doses were very cost-effective.
      Note that the ICER for canagliflozin 100 mg versus sitagliptin 100 mg is influenced by the difference observed in the incidence of nonsevere symptomatic hypoglycemia with canagliflozin 100 mg versus sitagliptin 100 mg in the clinical trial. As noted above, however, both agents have a low inherent risk of hypoglycemia [
      INVOKANA® (canagliflozin) tablets, for oral use [package insert].
      ,
      JANUVIA® (sitagliptin) tablets [package insert].
      ] and the observed difference is likely spurious. Sensitivity analyses modeling equal rates of nonsevere symptomatic hypoglycemia with canagliflozin 100 mg and sitagliptin 100 mg were performed in various scenarios, and, indeed, lower ICERs for canagliflozin 100 mg were estimated (data on file).
      All ICERs generated for canagliflozin 300 mg in the sensitivity analyses fell below the very cost-effective WTP threshold. ICERs generated for canagliflozin 100 mg were very cost-effective, except in the sensitivity analyses related to the unlikely scenarios of a doubling of costs or QALYs for AEs potentially associated with the SGLT2 mechanism of action or analyses over longer time horizons (30 and 40 years); however, these were cost-effective at the WTP threshold of 3 times the GDP per capita. The increase in ICERs seen over longer time horizons with canagliflozin 100 mg is likely related to the fact that the simulations accounted for the observed changes in the efficacy of agents with SGLT2 inhibition activity (such as canagliflozin) in patients with lower eGFR values [

      Janssen Research & Development. Integrated summary of efficacy: canagliflozin as an adjunctive treatment to diet and exercise alone or co-administered with other antihyperglycemic agents to improve glycemic control in adults with type 2 diabetes mellitus. 2012.

      ]. As a result, additional costs for insulin were incurred and more hypoglycemic events occurred over time. In a sensitivity analysis using the Latin American subsample data, both canagliflozin doses dominated sitagliptin, likely because of the greater efficacy associated with canagliflozin 300 and 100 mg versus sitagliptin 100 mg in Latin American patients enrolled in the clinical study, suggesting that canagliflozin may be especially favorable in the Latin American setting. Future analyses based on a larger Latin American or Mexican population would be of value given the relatively smaller sample size of this subsample compared with the total trial population. In the scenario more likely mimicking the real-world setting in Mexico (therapy intensification threshold of HbA1c ≥8.0%), where only a minority of patients achieve the generally recommended HbA1c target of less than 7.0% [
      • Villalpando S.
      • de la Cruz V.
      • Rojas R.
      • et al.
      Prevalence and distribution of type 2 diabetes mellitus in Mexican adult population: a probabilistic survey.
      ], canagliflozin was again very cost-effective.
      There were several challenges in performing this analysis in the Mexican health care setting. The evidence base on T2DM outcomes in Mexico is limited (e.g., the most recent data on goal attainment were from 2006). Enhanced data collection in Mexico and Latin America, including detailed studies on disutility weights for the complications associated with T2DM in the Mexican population, would be beneficial in terms of understanding the true burden of T2DM in this region and assessing the effectiveness of alternative treatment strategies. Obtaining accurate costs in Mexico is complicated given the fragmented health care system. Better unit cost data at the micro-level for the Mexican context would improve estimation of the true economic consequences of treatment intervention in T2DM.
      This article described economic modeling of T2DM using results from a clinical trial of canagliflozin versus sitagliptin with local costs sourced from Mexican sources and a well-validated model, ECHO-T2DM [
      • Nyirjesy P.
      • Sobel J.D.
      • Fung A.
      • et al.
      Genital mycotic infections with canagliflozin, a sodium glucose co-transporter 2 inhibitor, in patients with type 2 diabetes mellitus: a pooled analysis of clinical studies.
      ,
      • Nicolle L.E.
      • Capuano G.
      • Fung A.
      • Usiskin K.
      Urinary tract infection in randomized phase III studies of canagliflozin, a sodium glucose co-transporter 2 inhibitor.
      ]. Findings suggest that canagliflozin is likely to be cost-effective versus sitagliptin as an add-on therapy to metformin in patients with T2DM from the perspective of the Mexican health care system, owing to the differential benefits of improved control of hyperglycemia, hypertension, and weight with canagliflozin versus sitagliptin.
      Source of financial support: This analysis was sponsored by Janssen Global Services, LLC, and was based on data from a study supported by Janssen Research & Development, LLC. Editorial support was provided by Cherie Koch, PhD, of MedErgy, and was funded by Janssen Global Services, LLC. Canagliflozin has been developed by Janssen Research & Development, LLC, in collaboration with Mitsubishi Tanabe Pharma Corporation.

      Supplemental Materials

      References

        • The Lancet Diabetes Endocrinology
        Obesity prevention in Latin America: now is the time.
        Lancet Diabetes Endocrinol. 2014; 2: 263
        • International Diabetes Federation
        IDF Diabetes Atlas.
        6th ed. International Diabetes Federation, Brussels, Belgium2013
        • Corona A.J.
        • Martinez R.R.
        • Villalpando S.
        • SBCA Salinas
        Diabetes mellitus: La urgente necesidad de reforzar la respuesta en políticas públicas para su prevención y control.
        Encuesta Nacional de Salud Nutrición. 2012; (Available from: http://ensanut.insp.mx/doctos/seminario/M0302.pdf. [Accessed May 30, 2014])
        • Barquera S.
        • Campos-Nonato I.
        • Aguilar-Salinas C.
        • et al.
        Diabetes in Mexico: cost and management of diabetes and its complications and challenges for health policy.
        Global Health. 2013; 9: 3
        • Guzman J.R.
        • Lyra R.
        • Aguilar-Salinas C.A.
        • et al.
        Treatment of type 2 diabetes in Latin America: a consensus statement by the medical associations of 17 Latin American countries. Latin American Diabetes Association.
        Rev Panam Salud Publica. 2010; 28: 463-471
        • UK Prospective Diabetes Study (UKPDS) Group
        Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33).
        Lancet. 1998; 352: 837-853
      1. Instituto Mexicano del Seguro Social. Diagnóstico y Tratamiento de la Diabetes Mellitus Tipo 2 México: Guía de Práctica Clinica. Instituto Mexicano del Seguro Social, 2009. Actualización 2012. Instituto Mexicano del Seguro Social 2012. Available from: http://www.imss.gob.mx/sites/all/statics/guiasclinicas/000GER_DiabetesMellitus.pdf. [Accessed October 22, 2014].

        • Inzucchi S.E.
        • Bergenstal R.M.
        • Buse J.B.
        • et al.
        Management of hyperglycemia in type 2 diabetes: a patient-centered approach. Position statement of the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD).
        Diabetes Care. 2012; 35: 1364-1379
        • Ali M.K.
        • Bullard K.M.
        • Saaddine J.B.
        • et al.
        Achievement of goals in U.S. diabetes care, 1999-2010.
        N Engl J Med. 2013; 368: 1613-1624
        • Stark Casagrande S.
        • Fradkin J.E.
        • Saydah S.H.
        • et al.
        The prevalence of meeting A1C, blood pressure, and LDL goals among people with diabetes, 1988-2010.
        Diabetes Care. 2013; 36: 2271-2279
        • Villalpando S.
        • de la Cruz V.
        • Rojas R.
        • et al.
        Prevalence and distribution of type 2 diabetes mellitus in Mexican adult population: a probabilistic survey.
        Salud Publica Mex. 2010; 52: S19-S26
        • Barquera S.
        • Campos-Nonato I.
        • Hernandez-Barrera L.
        • et al.
        Hypertension in Mexican adults: results from the National Health and Nutrition Survey 2006.
        Salud Publica Mex. 2010; 52: S63-S71
        • Aguilar-Salinas C.A.
        • Gomez-Perez F.J.
        • Rull J.
        • et al.
        Prevalence of dyslipidemias in the Mexican National Health and Nutrition Survey 2006.
        Salud Publica Mex. 2010; 52: S44-S53
        • Gomez L.M.
        • Hernandez-Prado B.
        • Morales M.C.
        • Shamah-Levy T.
        Physical activity and overweight/obesity in adult Mexican population: the Mexican National Health and Nutrition Survey 2006.
        Salud Publica Mex. 2009; 51: S621-S629
      2. Johnson & Johnson, Inc. New long-term data show INVOKANA® (canagliflozin) lowers blood glucose in older patients with type 2 diabetes inadequately controlled on other therapies. 2014. Available from: http://www.jnj.com/news/all/New-Long-Term-Data-Show-INVOKANA-canagliflozin-Lowers-Blood-Glucose-in-Older-Patients-with-Type-2-Diabetes-Inadequately-Controlled-on-Other-Therapies. [Accessed October 20, 2014].

      3. INVOKANA® (canagliflozin) tablets, for oral use [package insert].
        Janssen Pharmaceuticals, Titusville, NJ2013
      4. INVOKANA. Informacion Para Prescribir Amplia INVOKANA® F.F Tableta V1.0 CCDS Junio. 2012.

        • Bode B.
        • Stenlöf K.
        • Sullivan D.
        • et al.
        Efficacy and safety of canagliflozin treatment in older subjects with type 2 diabetes mellitus: a randomized trial.
        Hosp Pract. 2013; 41: 72-84
        • Cefalu W.T.
        • Leiter L.A.
        • Yoon K.-H.
        • et al.
        Efficacy and safety of canagliflozin versus glimepiride in patients with type 2 diabetes inadequately controlled with metformin (CANTATA-SU): 52 week results from a randomised, double-blind, phase 3 non-inferiority trial.
        Lancet. 2013; 382: 941-950
        • Lavalle-González F.J.
        • Januszewicz A.
        • Davidson J.
        • et al.
        Efficacy and safety of canagliflozin compared with placebo and sitagliptin in patients with type 2 diabetes on background metformin monotherapy: a randomised trial.
        Diabetologia. 2013; 56: 2582-2592
        • Schernthaner G.
        • Gross J.L.
        • Rosenstock J.
        • et al.
        Canagliflozin compared with sitagliptin for patients with type 2 diabetes who do not have adequate glycemic control with metformin plus sulfonylurea: a 52-week, randomized trial.
        Diabetes Care. 2013; 36: 2508-2515
        • Stenlöf K.
        • Cefalu W.T.
        • Kim K.A.
        • et al.
        Long-term efficacy and safety of canagliflozin monotherapy in patients with type 2 diabetes inadequately controlled with diet and exercise: findings from the 52-week CANTATA-M study.
        Curr Med Res Opin. 2014; 30: 163-175
        • Stenlöf K.
        • Cefalu W.T.
        • Kim K.-A.
        • et al.
        Efficacy and safety of canagliflozin monotherapy in subjects with type 2 diabetes mellitus inadequately controlled with diet and exercise.
        Diabetes Obes Metab. 2013; 15: 372-382
        • Wilding J.P.
        • Charpentier G.
        • Hollander P.
        • et al.
        Efficacy and safety of canagliflozin in patients with type 2 diabetes mellitus inadequately controlled with metformin and sulphonylurea: a randomised trial.
        Int J Clin Pract. 2013; 67: 1267-1282
        • Yale J.F.
        • Bakris G.
        • Cariou B.
        • et al.
        Efficacy and safety of canagliflozin in subjects with type 2 diabetes and chronic kidney disease.
        Diabetes Obes Metab. 2013; 15: 463-473
        • Forst T.
        • Guthrie R.
        • Goldenberg R.
        • et al.
        Efficacy and safety of canagliflozin over 52 weeks in patients with type 2 diabetes on background metformin and pioglitazone.
        Diabetes Obes Metab. 2014; 16: 467-477
        • Cefalu W.T.
        • Leiter L.A.
        • Yoon K.-H.
        • et al.
        Canagliflozin demonstrates durable glycemic improvements over 104 weeks versus glimepiride in subjects with type 2 diabetes mellitus on metformin.
        Poster presented at: the 73rd Scientific Sessions of the American Diabetes Association (ADA). June 21–25, 2013; (Chicago, IL)
        • American Diabetes Association
        Standards of medical care in diabetes–2014.
        Diabetes Care. 2014; 37: S14-S80
        • Palmer A.J.
        The Mount Hood 5 Modeling Group. Computer modeling of diabetes and its complications: a report on the fifth Mount Hood Challenge meeting.
        Value Health. 2013; 16: 670-685
        • American Diabetes Association
        Guidelines for computer modeling of diabetes and its complications.
        Diabetes Care. 2004; 27: 2262-2265
      5. Mexican Peso exchange rate - US dollar - Historical exchange rates. 2014. Available from: http://www.exchange-rates.org/history/MXN/USD/T. [Accessed October 22, 2014].

      6. Quintiles Consulting, van Engen A, Heemstra L. The challenges of introducing HTA in Latin America. HTA Uncovered 2014;(5):1–9.

        • Rizo Rios P.
        • Gonzalez Rivera A.
        • Rivas Oropeza I.
        • Rivas Bocanegra R.E.
        Análisis del Proceso de Actualización de Medicamentos en el Cuadro Básico y Catálogo de Insumos del Sector Salud.
        Value Health Regional. 2013; 3 (387–291)
      7. World Health Organization. Cost-effectiveness thresholds. Available from: http://www.who.int/choice/costs/CER_thresholds/en/. [Accessed October 21, 2014].

        • Willis M.
        • Asseburg C.
        • He J.
        Validation of economic and health outcomes simulation model of type 2 diabetes mellitus (ECHO-T2DM).
        J Med Econ. 2013; 16: 1007-1021
        • Rojas R.
        • Aguilar-Salinas C.
        • Jimenez-Corona A.
        • et al.
        Metabolic syndrome in Mexican adults: results from the National Health and Nutrition Survey 2006.
        Salud Pública Méx. 2010; 52: 11-18
        • Brown J.B.
        • Russell A.
        • Chan W.
        • et al.
        The global diabetes model: user friendly version 3.0.
        Diabetes Res Clin Pract. 2000; 50: S15-S46
        • Bagust A.
        • Beale S.
        Modelling EuroQol health-related utility values for diabetic complications from CODE-2 data.
        Health Econ. 2005; 14: 217-230
        • Eastman R.C.
        • Javitt J.C.
        • Herman W.H.
        • et al.
        Model of complications of NIDDM, II: analysis of the health benefits and cost-effectiveness of treating NIDDM with the goal of normoglycemia.
        Diabetes Care. 20. 1997: 735-744
        • Hoerger T.J.
        • Wittenborn J.S.
        • Segel J.E.
        • et al.
        A health policy model of CKD, 1: model construction, assumptions, and validation of health consequences.
        Am J Kidney Dis. 2010; 55: 452-462
        • Clarke P.M.
        • Gray A.M.
        • Briggs A.
        • et al.
        A model to estimate the lifetime health outcomes of patients with type 2 diabetes: the United Kingdom Prospective Diabetes Study (UKPDS) Outcomes Model (UKPDS no. 68).
        Diabetologia. 2004; 47: 1747-1759
        • Hayes A.J.
        • Leal J.
        • Gray A.M.
        • et al.
        UKPDS outcomes model 2: a new version of a model to simulate lifetime health outcomes of patients with type 2 diabetes mellitus using data from the 30 year United Kingdom Prospective Diabetes Study: UKPDS 82.
        Diabetologia. 2013; 56: 1925-1933
        • Zethelius B.
        • Eliasson B.
        • Eeg-Olofsson K.
        • et al.
        A new model for 5-year risk of cardiovascular disease in type 2 diabetes, from the Swedish National Diabetes Register (NDR).
        Diabetes Res Clin Pract. 2011; 93: 276-284
        • Chaudhry Z.W.
        • Gannon M.C.
        • Nuttall F.Q.
        Stability of body weight in type 2 diabetes.
        Diabetes Care. 2006; 29: 493-497
      8. Janssen Research & Development. Integrated summary of efficacy: canagliflozin as an adjunctive treatment to diet and exercise alone or co-administered with other antihyperglycemic agents to improve glycemic control in adults with type 2 diabetes mellitus. 2012.

      9. JANUVIA® (sitagliptin) tablets [package insert].
        Merck & Co., Inc., Whitehouse Station, NJ2012
        • Nyirjesy P.
        • Sobel J.D.
        • Fung A.
        • et al.
        Genital mycotic infections with canagliflozin, a sodium glucose co-transporter 2 inhibitor, in patients with type 2 diabetes mellitus: a pooled analysis of clinical studies.
        Curr Med Res Opin. 2014; 30: 1109-1119
        • Nicolle L.E.
        • Capuano G.
        • Fung A.
        • Usiskin K.
        Urinary tract infection in randomized phase III studies of canagliflozin, a sodium glucose co-transporter 2 inhibitor.
        Postgrad Med. 2014; 126: 7-17
        • Weir M.
        • Januszewicz A.
        • Gilbert R.
        • et al.
        Effect of canagliflozin on blood pressure and adverse events related to osmotic diuresis and reduced intravascular volume in patients with type 2 diabetes mellitus.
        J Clin Hypertens (Greenwich). 2014; 16: 875-882
      10. Janssen Research & Development. Integrated summary of safety. Canagliflozin as an adjunctive treatment to diet and exercise alone or co-administered with other antihyperglycemic agents to improve glycemic control in adults with type 2 diabetes mellitus. Raritan, NJ: Janssen Research & Development, 2012.

        • Meier J.J.
        • Nauck M.A.
        Risk of pancreatitis in patients treated with incretin-based therapies.
        Diabetologia. 2014; 57: 1320-1324
      11. Instituto Mexicano del Seguro Social. Diagnostico y Tratamiento de la Diabetes Mellitus Tipo 2. Guía de Referencia Rápida. México: Instituto Mexicano del Seguro Social, 2009. Actualización 2012. Instituto Mexicano del Seguro Social 2012. Available from: http://www.imss.gob.mx/sites/all/statics/guiasclinicas/000GRR_DiabetesMellitus.pdf. [Accessed May 27, 2014].

        • Law M.R.
        • Morris J.K.
        • Wald N.J.
        Use of blood pressure lowering drugs in the prevention of cardiovascular disease: meta-analysis of 147 randomised trials in the context of expectations from prospective epidemiological studies.
        BMJ. 2009; 338: b1665
        • Brown W.V.
        • Bays H.E.
        • Hassman D.R.
        • et al.
        Efficacy and safety of rosuvastatin compared with pravastatin and simvastatin in patients with hypercholesterolemia: a randomized, double-blind, 52-week trial.
        Am Heart J. 2002; 144: 1036-1043
        • Wolffenbuttel B.H.
        • Franken A.A.
        • Vincent H.H.
        Cholesterol-lowering effects of rosuvastatin compared with atorvastatin in patients with type 2 diabetes -- CORALL study.
        J Intern Med. 2005; 257: 531-539
      12. Centro Nacional de Excelencia Tecnológica en Salud. Evaluación de Tecnologías de la Salud, Documento metodológico. Available from: http://www.cenetec.salud.gob.mx/descargas/detes/Documento_metodologico_ETES.pdf. [Accessed October 22, 2014].

        • Ríos P.R.
        • Rivera A.G.
        • Oropeza I.R.
        Bocanegra RER. Análisis del Proceso de Actualización de Medicamentos en el Cuadro Básico y Catálogo de Insumos del Sector Salud.
        Value Health. 2013; 2: 387-391
        • Rosenstock J.
        • Davies M.
        • Home P.D.
        • et al.
        A randomised, 52-week, treat-to-target trial comparing insulin detemir with insulin glargine when administered as add-on to glucose-lowering drugs in insulin-naive people with type 2 diabetes.
        Diabetologia. 2008; 51: 408-416
        • Riddle M.C.
        • Vlajnic A.
        • Zhou R.
        • Rosenstock J.
        Baseline HbA1c predicts attainment of 7.0% HbA1c target with structured titration of insulin glargine in type 2 diabetes: a patient-level analysis of 12 studies.
        Diabetes Obes Metab. 2013; 15: 819-825
        • Fonseca V.
        • Gill J.
        • Zhou R.
        • Leahy J.
        An analysis of early insulin glargine added to metformin with or without sulfonylurea: impact on glycaemic control and hypoglycaemia.
        Diabetes Obes Metab. 2011; 13: 814-822
        • Lopez-Cervantes M.
        • Rojas-Russell M.E.
        • Tirado-Gomez L.L.
        • et al.
        Enfermedad renal crónica y su atención mediante tratamiento sustitutivo en México.
        Facultad de Medicina, Universidad Nacional Autónoma de México, Mexico2009
      13. Instituto de Seguridad y Servivios Sociales de los Trabajadores del Estado. Tabulador de no derechohabientes 2012, Dirección Médica, Subdirección de Regulación y Atención Hospitalaria, Jefatura de Servicios, Instituto de Seguridad y Servicios Sociales de los Trabajadores del Estado. ISSSTE, 2012.

      14. Costos directos de complicaciones relacionadas con la diabetes en México. Internal report by IMS. 2013.

      15. Instituto Nacional de Cancerología INCAN. Tabulador de Cuotas de Recuperación, Subsecretaría de Ingresos, Unidad de política de ingresos. 2012. Available from: http://www.policiabancaria.df.gob.mx/portal_pbi/articulo15/fraccion8/2012/24%20Instituto%20Nacional%20de%20Cancerolog%C3%ADa%20POLICIA%20BANCARIA%20INCAN%202012.pdf. [Accessed May 30, 2014].

      16. Hospital General de Mexico. Tabulador de Cuotas de Recuperación, 2012. 2012. Available from: http://www.policiabancaria.df.gob.mx/portal_pbi/articulo15/fraccion8/2012/19%20Hospital%20General%20de%20M%C3%A9xico%202012.pdf. [Accessed May 30, 2014].

      17. Comisión Interinstitucional del Cuadro Básico de Insumos del Sector Salud. Cuadro básico y catálogo de medicamentos. 2012. Available from: http://www.dof.gob.mx/nota_detalle.php?codigo=5299596&fecha=21/05/2013. [Accessed May 30, 2014].

      18. Instituto Mexicano del Seguro Social. Portal de compras del IMSS. Instituto Mexicano del Seguro Social. 2014. Available from: http://compras.imss.gob.mx. [Accessed May 30, 2014].

      19. Indice Nacional de Precios al Consumidor, Servicio de Administración Tributaria. 2014. Available from: http://www.sat.gob.mx/informacion_fiscal/tablas_indicadores/Paginas/inpc_2012.aspx. [Accessed May 30, 2014].

        • Evans M.
        • Khunti K.
        • Mamdani M.
        • et al.
        Health-related quality of life associated with daytime and nocturnal hypoglycaemic events: a time trade-off survey in five countries.
        Health Qual Life Outcomes. 2013; 11: 90
        • Shingler S.
        • Fordham B.
        • Evans M.
        • et al.
        Utilities for treatment-related adverse events in type 2 diabetes.
        J Med Econ. 2015; 18: 45-55
        • Coyne K.S.
        • Sexton C.C.
        • Irwin D.E.
        • et al.
        The impact of overactive bladder, incontinence and other lower urinary tract symptoms on quality of life, work productivity, sexuality and emotional well-being in men and women: results from the EPIC study.
        BJU Int. 2008; 101: 1388-1395
        • Lane S.
        • Levy A.R.
        • Mukherjee J.
        • et al.
        The impact on utilities of differences in body weight among Canadian patients with type 2 diabetes.
        Curr Med Res Opin. 2014; 30: 1267-1273