SGLT2 Inhibitor Adherence and Diabetes Hospitalization Risk in Type 2 Diabetes

ABSTRACT

Objectives: Emerging evidence shows that sodium-glucose cotransporter 2 (SGLT2) inhibitors improve glycemic control and reduce the risk of cardiovascular-related mortality. This is contingent on adherence, yet evidence of SGLT2 inhibitor adherence and associated diabetes-related clinical outcomes in real-world settings is lacking. Therefore, we examined SGLT2 inhibitor longitudinal adherence trajectories and their association with the risk of diabetes-related hospitalizations.

Study Design: SGLT2 inhibitor new-user retrospective cohort study using a 25% random sample of IQVIA PharMetrics Plus for Academics US health plan claims, 2014 to 2022.

Methods: We included new users of SGLT2 inhibitors who had type 2 diabetes with prior cardiovascular disease or risk factors. Group-based trajectory modeling identified SGLT2 inhibitor adherence subgroups using monthly proportions of days covered over the 12 months post SGLT2I initiation. Cox proportional hazards models, adjusted for baseline covariates, estimated the HRs for diabetes-related hospitalizations/emergency department (ED) visits in the year following 12-month treatment with SGLT2 inhibitors.

Results: We classified 5051 new SGLT2 inhibitor users into 5 SGLT2 inhibitor adherence trajectory groups: nonadherent (14.7%), early decline (13.5%), late decline (10.5%), dynamic adherence (13.0%), and adherent (48.3%). A diabetes-related hospitalization/ED visit occurred in 2.3% of the overall cohort, with a median (IQR) time to event of 151.5 (76-263) days. In the adjusted Cox proportional hazards model, the late decline group (HR, 2.52; 95% CI, 1.38-4.60) and the dynamic adherence group (HR, 2.42; 95% CI, 1.52-3.86) had a statistically significantly greater risk of diabetes-related hospitalizations/ED visits than the adherent group.

Conclusions: These findings suggest that variable adherence to SGLT2 inhibitors can negatively impact diabetes-related outcomes.

Am J Manag Care. 2026;32(9):In Press

Takeaway Points

This study examined medication use patterns for sodium-glucose cotransporter 2 (SGLT2) inhibitors among individuals with type 2 diabetes and identified 5 distinct use patterns: nonadherent, early decline, late decline, dynamic adherence, and adherent.

• We found that individuals in the adherent trajectory had a lower risk of diabetes- and cardiovascular-related hospitalizations compared with those in other adherence trajectories.
• These findings underscore the importance of sustained medication use in reducing the likelihood of complications and increasing the likelihood of better outcomes.
• Managed care organizations can leverage these insights to develop adherence-focused interventions and targeted support programs aimed at improving health outcomes in this population.

Type 2 diabetes (T2D) is a chronic disease that, if poorly managed, can lead to severe complications, such as neuropathy, retinopathy, nephropathy, cardiovascular disease (CVD), and mortality.¹ Effective management of T2D is crucial to mitigate long-term micro- and macrovascular complications.² However, the progressive nature of T2D often mandates complex medication regimens, potentially leading to issues with adherence.^2,3 Consistent adherence to antidiabetic medications has been associated with reduced health care resource utilization and costs.⁴

Sodium-glucose cotransporter 2 (SGLT2) inhibitors are a class of oral antidiabetic medications used by approximately 4% of adults with T2D.^2,5 SGLT2 inhibitors are recommended as a first-line or add-on therapy for individuals at high risk for CVD, particularly those with established CVD or additional cardiovascular risk factors.^3,4 Clinical trials have consistently demonstrated that SGLT2 inhibitors not only improve glycemic control but also significantly reduce the risk of major adverse cardiovascular events, including hospitalization for heart failure, myocardial infarction, and stroke.^6-9 In addition, substantial evidence supports SGLT2 inhibitors’ renal protective effects, with trial data demonstrating slowed chronic kidney disease (CKD) progression and reduced risk of kidney failure, leading to recommendations for SGLT2 inhibitors as first-line therapy in individuals with T2D and CKD.¹⁰ SGLT2 inhibitors provide significant glycemic control benefits within 6 months, which may be maintained for up to a year with continuous use.^11,12 However, these findings stem largely from clinical trial settings where medication adherence is closely monitored.^13,14 In real-world clinical practice, adherence to SGLT2 inhibitors is often suboptimal,¹⁵ potentially due to factors such as adverse events, cost, and access barriers, which may diminish their full therapeutic benefit.^16-18 It is unknown how variable adherence to SGLT2 inhibitors in real-world settings affects clinical outcomes such as hospitalizations and emergency department (ED) visits.

Distinct longitudinal adherence trajectories can be identified using group-based trajectory modeling (GBTM).¹⁹ Research that incorporates this method to study the impact of SGLT2 inhibitor adherence trajectories on clinical outcomes remains limited,^20,21 representing a gap given the need for long-term adherence to improve diabetes outcomes. Therefore, this study aimed to identify longitudinal SGLT2 inhibitor adherence trajectories and to evaluate their association with the risk of diabetes- and cardiovascular-related hospitalizations and ED visits. Understanding longitudinal adherence patterns and their impact on clinical outcomes is crucial for optimizing T2D management.

METHODS

Study Design/Data Source

This retrospective cohort study utilized a 25% random sample of the IQVIA PharMetrics Plus for Academics database from 2014 to 2022. The data comprise fully adjudicated health plan claims and enrollment data for more than 117 million enrollees in the US. Data contributors are largely commercial health plans. The data provide a longitudinal view of inpatient and outpatient services, outpatient prescription dispensing, costs, and detailed enrollment information. Diagnoses are recorded using International Classification of Diseases, Ninth Revision (ICD-9) or International Statistical Classification of Diseases, Tenth Revision (ICD-10) codes, and procedures are recorded using Current Procedural Terminology (CPT) or Healthcare Common Procedure Coding System (HCPCS) codes. Because ICD-10 coding was implemented in our data set starting in 2016, both coding systems were necessary to account for diagnostic information before and after the transition. IQVIA’s Uniform System of Classification (USC) for pharmaceutical products facilitates the complete capture of prescription medications filled at outpatient pharmacies.²²

Study Cohort

We identified individuals aged 18 to 64 years who were newly initiated on oral SGLT2 inhibitors (ie, canagliflozin, dapagliflozin, empagliflozin, ertugliflozin) at any time from January 1, 2014, through December 31, 2020. The date of the first SGLT2 inhibitor pharmacy claim was the index date (eAppendix Figure 1 [eAppendix available at ajmc.com]). Individuals were required to have at least 2 prescription claims for SGLT2 inhibitors during the identification period to enable measurement of adherence, at least 1 medical ICD-9 or ICD-10 diagnosis code for T2D (250.02 or E11.xx) in the 6 months prior to the index SGLT2 inhibitor, and at least 18 months of continuous enrollment (ie, 6 months prior to the index date [baseline period] and 12 months post index date [adherence assessment period]) to obtain information on baseline covariates and to allow for the 12-month adherence measurement. Given our interest in individuals with T2D and high risk for CVD, we further required individuals to have at least 2 ICD-9 or ICD-10 diagnosis codes for CVD or an additional cardiovascular risk factor in the baseline period (eAppendix Table 1). Individuals were excluded if they had at least 1 medical claim with a diagnosis of type 1 diabetes, neoplasms, HIV, or pregnancy at any time during the study baseline period (eAppendix Table 2). Excluding these populations ensured a more homogenous study cohort, as medication use and adherence in these patients are likely to differ from those of the overall population with T2D.

Study Variables

SGLT2 inhibitor adherence trajectories. Based on the dispensing date and days supplied, we calculated an interval-based monthly proportion of days covered (PDC) for 12 months starting from the index date. We implemented a 12-month adherence measurement period because this has been shown to be sufficient time to observe the impact of SGLT2 inhibitors on clinical outcomes, such as blood sugar control and cardiovascular risk reduction.⁷ The PDC was calculated as the number of days that an SGLT2 inhibitor was refilled in a given month divided by 30 to generate a proportion from 0 (ie, nonadherent) to 1 (ie, full adherence). We selected 30-day monthly intervals because 95% of oral hypoglycemic prescriptions are dispensed as 30-day supplies.^23,24 We shifted days forward with overlapping coverage, assuming an individual would finish the existing medication supply before starting the new supply. We used GBTM to identify subgroups with similar medication refill patterns over the 12 months following the index date.^19,25,26 GBTM was modeled with the PDC as a continuous outcome and months following the index date as the time variable.

We adopted a censored-normal GBTM, given our continuous SGLT2 inhibitor adherence outcome.^19,25,26 We tested 2- to 6-group models, allowing for second- and third-order polynomials. Individuals were assigned to the trajectory group for which their longitudinal adherence showed the highest probability of membership. The final number of latent trajectory groups was determined by a combination of (1) the Bayesian information criterion, where smaller values indicate a better model fit; (2) a 5% minimum membership requirement for each group; (3) the average posterior probability of at least 0.7 in all groups; and (4) the odds of correct classification (> 5.0 for all groups).^19,27 These models were estimated using PROC TRAJ in SAS Studio 9.4 (SAS Institute Inc).

Outcome measures. All outcomes were assessed in the 1 year following the end of the 12-month SGLT2 inhibitor adherence assessment. The primary outcome was the time to the first diabetes-related hospitalization/ED visit in the 1-year follow-up period (eAppendix Figure 1). Diabetes-related hospitalizations/ED visits were defined as an inpatient admission or an ED encounter related to diabetes, including hyperglycemia, hypoglycemia, electrolyte imbalances, diabetic retinopathy, nephropathy, or neuropathy. The outcome was operationalized using the primary discharge ICD-9/ICD-10 diagnosis codes (eAppendix Table 3).^28-30 These conditions were selected because they are established complications of T2D medication nonadherence that can result in hospitalization.^31,32 Given clinical trial reports of reduced CVD hospitalizations with SGLT2 inhibitor exposure,³³ we examined the secondary outcome as the time to first CVD-related hospitalization/ED visit during the 1-year follow-up period. A CVD-related hospitalization/ED visit was defined as an inpatient admission or ED visit with a primary diagnosis of cardiac arrhythmia, congestive heart failure, stroke, myocardial infarction, or angina (eAppendix Table 3).

Baseline covariates. The baseline covariates, including sociodemographic characteristics, clinical diagnoses, medication utilization, and health service use, were selected based on prior SGLT2 inhibitor adherence studies examining diabetes- and CVD-related hospitalization/ED visit outcomes.^34-36 Covariates were measured during the 6-month baseline period to avoid including factors that might change due to the treatment. Sociodemographic characteristics were age, sex (female or male), and clinical diagnoses, including preexisting cardiovascular risk factors and diseases. Using the USC classification system, we categorized baseline cardiovascular medications as statins, antidiabetics, and antihypertensives, which included angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, diuretics, β-blockers, and calcium channel blockers (eAppendix Table 4). Health service use measures in the 6 months before the index date included prior hospitalizations/ED visits.³⁰

Statistical Analysis

We compared baseline characteristics across adherence trajectory groups using χ² tests for categorical variables and analyses of variance for continuous variables. We used multivariable Cox proportional hazards models to estimate the HR of diabetes-related hospitalizations/ED visits. We selected the group with the highest adherence trajectory as the reference group because we hypothesized this group would have the lowest risk of the study outcomes. We included the baseline covariates in the adjusted models. The time to first diabetes-related hospitalization/ED visit was the number of days from the day 1 start of follow-up until the date of the earliest of either the study outcome or a censoring (ie, the loss of insurance coverage or the end of the 1-year follow-up period). The violation of the proportional hazard assumption was tested using Schoenfeld residuals. Statistical significance was determined using 95% CIs and 2-tailed P values less than .05.

Statistical analyses were performed using SAS Studio 9.4 (SAS Institute Inc). The University of Maryland, Baltimore Institutional Review Board deemed this study exempt.

Sensitivity Analyses

Sensitivity analysis was conducted excluding individuals with a prior history of established CVD, evidence of hospitalizations/ED visits at baseline, and insulin use at baseline. These exclusions were made to minimize potential confounding due to underlying disease severity, which may influence both adherence to SGLT2 inhibitors and risk of subsequent diabetes-related hospitalizations/ED visits.

RESULTS

Characteristics of the SGLT2 Inhibitor Initiator Study Cohort

Our study cohort included 5051 individuals who initiated SGLT2 inhibitors. Figure 1 shows the flow diagram for cohort selection. More than half the cohort was male (59.7%), and the cohort had a mean (SD) age of 54.2 (7.4) years. Over the 12-month SGLT2 inhibitor adherence measurement period, the mean (SD) PDC was 0.66 (0.28), and 43.8% of the cohort had a PDC of at least 0.80.

SGLT2 Inhibitor Adherence Trajectory Groups

Guided by the Nagin criteria, the 5-group model best fit the data (eAppendix Tables 5 and 6 and eAppendix Figure 2).^25,37,38 The nonadherent group comprised 14.7% of the cohort. Two trajectory groups discontinued treatment, one by 9 months (13.5%; ie, early decline) and one gradually declining in adherence after 6 months (10.5%; ie, late decline). One group displayed adherence fluctuating between a PDC of 0.84 and 0.50 (13%; ie, dynamic). One group had high adherence, with a PDC greater than 0.85, throughout the 12-month adherence measurement period (48.3%; ie, adherent) (Figure 2).

Characteristics of the SGLT2 Inhibitor Adherence Trajectory Groups

Individuals in the adherent trajectory group were older (mean age, 54.6 years) than those in the other trajectory groups. The proportion of women was highest in the nonadherent group (44.6%) compared with the late decline (40.3%) and adherent (37.0%) groups. The dynamic adherence group had higher proportions of individuals with prior glucagon-like peptide-1 receptor agonist use (4.3%) than other trajectory groups. The late decline group had higher proportions of individuals with congestive heart failure (3.6%) than the other trajectory groups (Table 1).

Risk of Diabetes-Related Hospitalizations/ED Visits

Overall, 2.3% of the study cohort had a diabetes-related hospitalization/ED visit, with a median (IQR) time to event of 151.5 (76-263) days. The crude incidence rates of diabetes-related hospitalizations/ED visits per 10,000 person-years were 0.68 (adherent), 1.23 (nonadherent), 0.83 (early decline), 1.79 (late decline), and 1.75 (dynamic adherence) (Table 2). The cumulative incidence of diabetes-related hospitalizations/ED visits across the SGLT2 inhibitor adherence trajectory groups is presented in Figure 3. After adjusting for baseline covariates (eAppendix Figure 3), the late decline (HR, 2.52; 95% CI, 1.38-4.60) and the dynamic adherence (HR, 2.42; 95% CI, 1.52-3.86) groups had a significantly greater risk of diabetes-related hospitalizations/ED visits than the adherent group (Table 2).

Risk of Cardiovascular-Related Hospitalizations/ED Visits

For the secondary outcome, 2.6% of the study cohort had a CVD-related hospitalization/ED visit, with a median (IQR) time to event of 155.0 (77-259) days. In the adjusted model (eAppendix Figure 4), the risk of CVD-related hospitalizations/ED visits was higher in the nonadherent (HR, 1.71; 95% CI, 1.15-2.54) and the dynamic adherence (HR, 1.67; 95% CI, 1.16-2.41) groups than in the adherent group (Table 2).

Results from the sensitivity analysis were consistent with those of the primary analysis (eAppendix Table 7).

DISCUSSION

In this study of 5051 individuals with T2D who initiated SGLT2 inhibitors, we identified 5 distinct adherence trajectories over the first 12 months of use. The adherent group exhibited a lower risk of diabetes- and CVD-related hospitalizations/ED visits. Those in the dynamic adherence (ie, fluctuating) group had a higher risk of diabetes- and CVD-related outcomes relative to those in the adherent group. Our findings show a higher risk of diabetes-related hospitalizations/ED visits among the late decline group relative to the adherent group. Although the nonadherent and early decline groups had a higher risk of diabetes- and CV-related hospitalizations/ED visits, the small number of events resulted in imprecise estimates, leading to wide CIs and a lack of statistical significance. These findings carry important implications for understanding the relationship between SGLT2 inhibitor adherence and clinical outcomes among adults with T2D.

For one, both the dynamic adherence group and declining adherence groups were primarily composed of women, and these were the SGLT2 inhibitor adherence trajectories with the highest likelihood of diabetes- and CVD-related hospitalizations/ED visits. Lower adherence to SGLT2 inhibitors in women relative to men has been reported in other studies.^15,20,39 SGLT2 inhibitor use has been associated with urinary tract and vaginal yeast infections,⁴⁰ which may explain the lower adherence among women.

Three adherence trajectory groups discontinued SGLT2 inhibitors at different points over the course of treatment. Discontinuation of second-line treatments for T2D, which includes SGLT2 inhibitors, is reported to be 21%.^41,42 This study could not determine the reason for discontinuation, but other studies have reported poor tolerability of the medication, ineffectiveness, or cost as potential reasons for discontinuation.^16,41,43 Less than half (43.8%) of our cohort maintained a PDC of at least 0.80 consistently over the 12-month assessment period, which is consistent with data from other studies.¹⁵

Poor adherence to SGLT2 inhibitors early in the course of treatment carries important consequences. In the dynamic adherence group, suboptimal PDC (ie, < 0.8), starting as early as month 2, likely limited the glycemic benefits. Data from clinical trials have shown that maximal reductions in hemoglobin A_1c for SGLT2 inhibitors are typically achieved approximately 6 months after initiation and sustained up to 1 year.¹² Therefore, failure to maintain consistent medication use during this period may compromise glycemic control, increasing the risk of diabetes-related complications requiring hospitalizations or ED visits. The proportion of individuals with microvascular complications was highest in the late decline group, which may further explain their increased risk of diabetes-related hospitalizations. In the absence of prior research specifically examining SGLT2 inhibitor adherence trajectories and diabetes-related outcomes, our findings are consistent with those of studies on oral hypoglycemic agents, which have shown that individuals with low and moderate adherence trajectories are at increased risk of diabetes-related hospitalizations/ED visits compared with those with adherent adherence.³⁰

Discontinuation and inconsistent adherence can result in CVD-related complications. The cardiovascular benefits of SGLT2 inhibitor therapy can begin as early as 1 month after initiation and can be sustained for approximately 4 months or more.⁴⁴ The elevated risk observed in the nonadherent and dynamic adherence groups may be attributable to discontinuation of treatment by 6 months or earlier. Although it is possible that some individuals switched to alternative antidiabetic therapies, these treatment changes were not captured in our study.

Strengths and Limitations

This study has several strengths. Defining longitudinal SGLT2 inhibitor adherence trajectories captured medication use behaviors. Second, the use of a commercial administrative data set reflects real-world settings that can generalize to a broader population. Unlike randomized clinical trials, which are usually restricted to highly selective trial populations, our study included a more heterogeneous population with comorbid CVD conditions.

Several limitations should be considered when interpreting these results. Adherence using PDC is assessed through pharmacy claims data, which indicate only that a medication was dispensed but not whether it was consumed. In addition, PDC does not measure primary nonadherence, which happens when newly prescribed medications are never filled. Furthermore, administrative claims data lack clinical data (eg, hemoglobin A_1c level) and the reason for treatment discontinuation (eg, physician decision, switching to another medication). We excluded individuals who were not continuously enrolled for 12 months to ensure complete data to assess adherence over the follow-up period. If those excluded were more likely to have poor adherence, this study may underestimate the diabetes and CVD complications associated with nonadherence. The possibility of unmeasured confounders (eg, healthy user bias) cannot be ruled out; individuals with better adherence to SGLT2 inhibitors may be more likely to engage in healthy behaviors, which could influence the outcome.

CONCLUSIONS

We identified distinct SGLT2 inhibitor adherence trajectories that were associated with diabetes- and CVD-related hospitalizations/ED visits. Describing longitudinal changes in adherence over time enables a better understanding of adherence patterns, offering valuable insights for targeted interventions to improve adherence. Research and intervention development efforts should prioritize groups at high risk of nonadherence.

Author Affiliations: Department of Practice, Sciences, and Health Outcomes Research, University of Maryland School of Pharmacy (UD, WCC, SdR), Baltimore, MD; Health Outcomes Division, College of Pharmacy, The University of Texas at Austin (GO), Austin, TX.

Source of Funding: None.

Author Disclosures: Dr Damachi is supported by a 2024 PhRMA Foundation Predoctoral Fellowship in Value Assessment and Health Outcomes Research. Dr Camelo Castillo was supported in part by a research grant unrelated to this work from the Investigator Studies Program of Merck Sharp & Dohme Corp. Dr dosReis has received funding from the National Institutes of Health, the Patient-Centered Outcomes Research Institute, and GSK for work unrelated to the work described in this manuscript. Mr Okoye reports no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Authorship Information: Concept and design (UD, GO, WCC, SdR); acquisition of data (SdR); analysis and interpretation of data (UD, GO, WCC, SdR); drafting of the manuscript (UD, GO, WCC); critical revision of the manuscript for important intellectual content (UD, GO, WCC, SdR); statistical analysis (UD, GO, SdR); administrative, technical, or logistic support (UD); and supervision (WCC, SdR).

Address Correspondence to: Udim Damachi, PhD, MS, Department of Practice, Sciences, and Health Outcomes Research, School of Pharmacy, University of Maryland, Baltimore, 220 N Arch St, 12th Floor, Baltimore, MD 21201. Email: udamachi@umaryland.edu.

REFERENCES

1. Diabetes: coexisting conditions and complications. CDC. Updated September 30, 2022. Accessed February 13, 2023. https://web.archive.org/web/20230209072427/https://www.cdc.gov/diabetes/data/statistics-report/coexisting-conditions-complications.html

2. Tiwari K, Bisht M, Kant R, Handu SS. Prescribing pattern of anti-diabetic drugs and adherence to the American Diabetes Association’s (ADA) 2021 treatment guidelines among patients of type 2 diabetes mellitus: a cross-sectional study. J Family Med Prim Care. 2022;11(10):6159-6164. doi:10.4103/jfmpc.jfmpc_458_22

3. American Diabetes Association Professional Practice Committee. 9. pharmacologic approaches to glycemic treatment: Standards of Care in Diabetes—2024. Diabetes Care. 2023;47(suppl 1):S158-S178. doi:10.2337/dc24-S009

4. Heidenreich PA, Bozkurt B, Aguilar D, et al; ACC/AHA Joint Committee Members. 2022 AHA/ACC/HFSA Guideline for the Management of Heart Failure: a report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation. 2022;145(18):e895-e1032. doi:10.1161/CIR.0000000000001063

5. Sodium-glucose cotransporter-2 (SGLT2) inhibitors. FDA. Updated August 20, 2018. Accessed March 6, 2023. https://www.fda.gov/drugs/postmarket-drug-safety-information-patients-and-providers/sodium-glucose-cotransporter-2-sglt2-inhibitors

6. Pinto LC, Rados DV, Remonti LR, Kramer CK, Leitao CB, Gross JL. Efficacy of SGLT2 inhibitors in glycemic control, weight loss and blood pressure reduction: a systematic review and meta-analysis. Diabetol Metab Syndr. 2015;7(suppl 1):A58. doi:10.1186/1758-5996-7-S1-A58

7. Zaccardi F, Webb DR, Htike ZZ, Youssef D, Khunti K, Davies MJ. Efficacy and safety of sodium-glucose co-transporter-2 inhibitors in type 2 diabetes mellitus: systematic review and network meta-analysis. Diabetes Obes Metab. 2016;18(8):783-794. doi:10.1111/dom.12670

8. Yen FS, Wei JCC, Yu TS, Hung YT, Hsu CC, Hwu CM. Sodium-glucose cotransporter 2 inhibitors and risk of retinopathy in patients with type 2 diabetes. JAMA Netw Open. 2023;6(12):e2348431. doi:10.1001/jamanetworkopen.2023.48431

9. Rådholm K, Wu JH, Wong MG, et al. Effects of sodium-glucose cotransporter-2 inhibitors on cardiovascular disease, death and safety outcomes in type 2 diabetes – a systematic review. Diabetes Res Clin Pract. 2018;140:118-128. doi:10.1016/j.diabres.2018.03.027

10. Elenjickal EJ, Travlos CK, Luu J, Lemay S, Suri RS, Mavrakanas TA. Efficacy and safety of sodium-glucose cotransporter-2 inhibitors in patients with and without advanced CKD: systematic review and meta-analysis. Clin J Am Soc Nephrol. 2025;20(6):788-809. doi:10.2215/CJN.0000000693

11. Thewjitcharoen Y, Yenseung N, Malidaeng A, et al. Effectiveness of long-term treatment with SGLT2 inhibitors: real-world evidence from a specialized diabetes center. Diabetol Metab Syndr. 2017;9:96. doi:10.1186/s13098-017-0297-y

12. Monami M, Nardini C, Mannucci E. Efficacy and safety of sodium glucose co-transport-2 inhibitors in type 2 diabetes: a meta-analysis of randomized clinical trials. Diabetes Obes Metab. 2014;16(5):457-466. doi:10.1111/dom.12244

13. Zinman B, Wanner C, Lachin JM, et al; EMPA-REG OUTCOME Investigators. Empagliflozin, cardiovascular outcomes, and mortality in type 2 diabetes. N Engl J Med. 2015;373(22):2117-2128. doi:10.1056/NEJMoa1504720

14. Wiviott SD, Raz I, Bonaca MP, et al; DECLARE-TIMI 58 Investigators. Dapagliflozin and cardiovascular outcomes in type 2 diabetes. N Engl J Med. 2019;380(4):347-357. doi:10.1056/NEJMoa1812389

15. Ofori-Asenso R, Sahle BW, Chin KL, et al. Poor adherence and persistence to sodium glucose co-transporter 2 inhibitors in real-world settings: evidence from a systematic review and meta-analysis. Diabetes Metab Res Rev. 2021;37(1):e3350. doi:10.1002/dmrr.3350

16. Lorenzo MG, Romero LY, Lorenzo VT, Lara Mjm. 4CPS-013 discontinuation of sodium-glucose co-transporter 2 inhibitors due to recurrent genitourinary infections. Eur J Hosp Pharm. 2018;25(suppl 1):A47. doi:10.1136/ejhpharm-2018-eahpconf.104

17. McGovern AP, Hogg M, Shields BM, et al; MASTERMIND consortium. Risk factors for genital infections in people initiating SGLT2 inhibitors and their impact on discontinuation. BMJ Open Diabetes Res Care. 2020;8(1):e001238. doi:10.1136/bmjdrc-2020-001238

18. Bak M, Chi SA, Jeon K, et al. Discontinuation rates, clinical effects and provocation factors of SGLT-2 inhibitor in the real world. Sci Rep. 2024;14(1):30653. doi:10.1038/s41598-024-71231-7

19. Nagin DS, Odgers CL. Group-based trajectory modeling in clinical research. Annu Rev Clin Psychol. 2010;6:109-138. doi:10.1146/annurev.clinpsy.121208.131413

20. Hawley CE, Lauffenburger JC, Paik JM, Wexler DJ, Kim SC, Patorno E. Three sides to the story: adherence trajectories during the first year of SGLT2 inhibitor therapy among Medicare beneficiaries. Diabetes Care. 2022;45(3):604-613. doi:10.2337/dc21-1676

21. Huang HC, Touchette DR, Tadrous M, Schumock GT, Awadalla S, Lee TA. Adherence patterns 1 year after initiation of SGLT2 inhibitors: results of a national cohort study. Am J Manag Care. 2024;30(8):e226-e232. doi:10.37765/ajmc.2024.89591

22. Fact sheet: IQVIA PharMetrics Plus. IQVIA. Accessed May 16, 2023. https://www.iqvia.com/locations/united-states/library/fact-sheets/iqvia-pharmetrics-plus

23. Peterson AM, Nau DP, Cramer JA, Benner J, Gwadry-Sridhar F, Nichol M. A checklist for medication compliance and persistence studies using retrospective databases. Value Health. 2007;10(1):3-12. doi:10.1111/j.1524-4733.2006.00139.x

24. Adherence: PQA adherence measures. Pharmacy Quality Alliance. Accessed January 28, 2025. https://www.pqaalliance.org/adherence-measures

25. Nagin DS. Group-based trajectory modeling: an overview. Ann Nutr Metab. 2014;65(2-3):205-210. doi:10.1159/000360229

26. Franklin JM, Shrank WH, Pakes J, et al. Group-based trajectory models: a new approach to classifying and predicting long-term medication adherence. Med Care. 2013;51(9):789-796. doi:10.1097/MLR.0b013e3182984c1f

27. Twisk J, Hoekstra T. Classifying developmental trajectories over time should be done with great caution: a comparison between methods. J Clin Epidemiol. 2012;65(10):1078-1087. doi:10.1016/j.jclinepi.2012.04.010

28. Ginde AA, Blanc PG, Lieberman RM, Camargo CA Jr. Validation of ICD-9-CM coding algorithm for improved identification of hypoglycemia visits. BMC Endocr Disord. 2008;8:4. doi:10.1186/1472-6823-8-4

29. McCoy RG, Lipska KJ, Van Houten HK, Shah ND. Association of cumulative multimorbidity, glycemic control, and medication use with hypoglycemia-related emergency department visits and hospitalizations among adults with diabetes. JAMA Netw Open. 2020;3(1):e1919099. doi:10.1001/jamanetworkopen.2019.19099

30. Lo-Ciganic WH, Donohue JM, Jones BL, et al. Trajectories of diabetes medication adherence and hospitalization risk: a retrospective cohort study in a large state Medicaid program. J Gen Intern Med. 2016;31(9):1052-1060. doi:10.1007/s11606-016-3747-6

31. Pednekar P, Heller DA, Peterson AM. Association of medication adherence with hospital utilization and costs among elderly with diabetes enrolled in a state pharmaceutical assistance program. J Manag Care Spec Pharm. 2020;26(9):1099-1108. doi:10.18553/jmcp.2020.26.9.1099

32. Farmaki P, Damaskos C, Garmpis N, Garmpi A, Savvanis S, Diamantis E. Complications of the type 2 diabetes mellitus. Curr Cardiol Rev. 2020;16(4):249-251. doi:10.2174/1573403X1604201229115531

33. Patel SM, Kang YM, Im K, et al. Sodium-glucose cotransporter-2 inhibitors and major adverse cardiovascular outcomes: a SMART-C collaborative meta-analysis. Circulation. 2024;149(23):1789-1801. doi:10.1161/CIRCULATIONAHA.124.069568

34. Lee DSU, Lee H. Adherence and persistence rates of major antidiabetic medications: a review. Diabetol Metab Syndr. 2022;14(1):12. doi:10.1186/s13098-022-00785-1

35. Murray MD, Morrow DG, Weiner M, et al. A conceptual framework to study medication adherence in older adults. Am J Geriatr Pharmacother. 2004;2(1):36-43. doi:10.1016/s1543-5946(04)90005-0

36. Capoccia K, Odegard PS, Letassy N. Medication adherence with diabetes medication: a systematic review of the literature. Diabetes Educ. 2016;42(1):34-71. doi:10.1177/0145721715619038

37. Hickson RP, Annis IE, Killeya-Jones LA, Fang G. Opening the black box of the group-based trajectory modeling process to analyze medication adherence patterns: an example using real-world statin adherence data. Pharmacoepidemiol Drug Saf. 2020;29(3):357-362. doi:10.1002/pds.4917

38. Niyonkuru C, Wagner AK, Ozawa H, Amin K, Goyal A, Fabio A. Group-based trajectory analysis applications for prognostic biomarker model development in severe TBI: a practical example. J Neurotrauma. 2013;30(11):938-945. doi:10.1089/neu.2012.2578

39. Stolfo D, Lund LH, Benson L, et al. Real-world use of sodium-glucose cotransporter 2 inhibitors in patients with heart failure and reduced ejection fraction: data from the Swedish Heart Failure Registry. Eur J Heart Fail. 2023;25(9):1648-1658. doi:10.1002/ejhf.2971

40. Li D, Wang T, Shen S, Fang Z, Dong Y, Tang H. Urinary tract and genital infections in patients with type 2 diabetes treated with sodium-glucose co-transporter 2 inhibitors: a meta-analysis of randomized controlled trials. Diabetes Obes Metab. 2017;19(3):348-355. doi:10.1111/dom.12825

41. Liss DT, Cherupally M, O’Brien MJ, et al. Treatment modification after initiating second-line medication for type 2 diabetes. Am J Manag Care. 2023;29(12):661-668. doi:10.37765/ajmc.2023.89466

42. Montvida O, Shaw J, Atherton JJ, Stringer F, Paul SK. Long-term trends in antidiabetes drug usage in the U.S.: real-world evidence in patients newly diagnosed with type 2 diabetes. Diabetes Care. 2018;41(1):69-78. doi:10.2337/dc17-1414

43. García-Pérez LE, Alvarez M, Dilla T, Gil-Guillén V, Orozco-Beltrán D. Adherence to therapies in patients with type 2 diabetes. Diabetes Ther. 2013;4(2):175-194. doi:10.1007/s13300-013-0034-y

44. Chen K, Nie Z, Shi R, et al. Time to benefit of sodium-glucose cotransporter-2 inhibitors among patients with heart failure. JAMA Netw Open. 2023;6(8):e2330754. doi:10.1001/jamanetworkopen.2023.30754