ORIGINAL RESEARCH published: 14 July 2021 doi: 10.3389/fendo.2021.708372 Analysis of the Adherence and Safety of Second Oral Glucose-Lowering Therapy in Routine Practice From the Mediterranean Area: A Retrospective Cohort Study Bogdan Vlacho 1,2,3, Manel Mata-Cases 1,4, Xavier Mundet-Tudurı́ 1,5, Joan-Antoni Vallès-Callol 1, Jordi Real 1,4, Magi Farre 3,6, Xavier Cos 1,7, Kamlesh Khunti 8, Dı́dac Mauricio 1,4,9,10* and Josep Franch-Nadal 1,4* 1 Grup de Recerca Epidemiològica en Diabetis des de l'Atenció Primària (DAP-CAT) Group, Unitat de Suport a la Recerca Barcelona, Fundació Institut Universitari per a la recerca a l’Atenció Primària de Salut Jordi Gol i Gurina (IDIAPJGol), Edited by: Barcelona, Spain, 2 Clinical Trials Unit, Germans Trias i Pujol Health Science Research Institute (IGTP), Barcelona, Spain, Alexandre Gabarra Oliveira, 3 Pharmacology Department, Universitat Autònoma de Barcelona (UAB), Cerdanyola del Vallès, Spain, 4 CIBER of Diabetes São Paulo State University, Brazil and Associated Metabolic Diseases (CIBERDEM), Instituto de Salud Carlos III (ISCIII), Madrid, Spain, 5 Departament of Reviewed by: Medicine, Universitat Autònoma de Barcelona, Bellaterra, Spain, 6 Clinical Pharmacology Unit, Hospital Universitari Germans Karel Kostev, Trias i Pujol, Badalona, Spain, 7 Primary and Hospital Innovation Department, Innovation Office at Institut Català de la Salut, IQVIA, Germany Barcelona, Spain, 8 Diabetes Research Centre, College of Life Sciences, University of Leicester, Leicester, United Kingdom, Mario Saad, 9 Department of Endocrinology and Nutrition, Hospital Universitari de la Santa Creu i Sant Pau, Universitat Autònoma de State University of Campinas, Barcelona, Bellaterra, Spain, 10 Departament of Medicine, University of Vic - Central University of Catalonia, Vic, Spain Brazil *Correspondence: The aims of our study was compare adherence measured by the medical possession ratio Dı´dac Mauricio didacmauricio@gmail.com (MPR), time until discontinuation and describe adverse events after adding a DPP-4i, Josep Franch-Nadal SGLT-2i, or sulfonylureas (SU) to metformin in a primary care population with insufficient josep.franch@gmail.com glycemic control. We used routinely-collected health data from the SIDIAP database. The Specialty section: included subjects were matched by propensity score. The follow-up period was up to 24 This article was submitted to months or premature discontinuation. The primary outcomes were the percentage of Obesity, a section of the journal subjects with good adherence, treatment discontinuation and adverse events among Frontiers in Endocrinology treatment groups. The proportion of patients with good adherence (MPR> 0.8) after the Received: 11 May 2021 addition of DPP-4i, SGLT-2i or SU was 53.6%, 68.7%, and 43.0%, respectively. SGLT-2i Accepted: 29 June 2021 users were 1.7 times more likely to achieve good adherence compared with DPP-4i users Published: 14 July 2021 (odds ratio [OR]:1.72, 98% confidence interval [CI]:1.51, 1.96), and 2.8 times more likely Citation: Vlacho B, Mata-Cases M, compared with SU users (OR: 0.35, 98% CI: 0.07, 0.29). The discontinuation hazard Mundet-Tudurı´ X, Vallès-Callol J-A, ratios were 1.43 (98%CI: 1.26; 1.62) and 1.60 (98%CI: 1.42; 1.81) times higher among Real J, Farre M, Cos X, Khunti K, Mauricio D and Franch-Nadal J (2021) SGLT-2i and SU users than DPP-4i users during the follow-up period. No differences were Analysis of the Adherence and Safety observed for adverse events among the treatment groups. In conclusion, in our real-world of Second Oral Glucose-Lowering setting, the combination of SGLT-2i with metformin was associated with better Therapy in Routine Practice From the Mediterranean Area: A adherence. The mean time until discontinuation was longer in the SGLT-2i group in Retrospective Cohort Study. comparison with the DPP-4i or SU groups. Front. Endocrinol. 12:708372. doi: 10.3389/fendo.2021.708372 Keywords: adherence - compliance – persistence, glycemia control, type 2 diabetes, primary care, observational study Frontiers in Endocrinology | www.frontiersin.org 1 July 2021 | Volume 12 | Article 708372 Vlacho et al. Adherence of Second Oral Glucose-Lowering Therapy INTRODUCTION suboptimal (67.9%), while the persistence to initial oral antihyperglycemic agents ranged from 41.0% to 81.1% (22). Good quality management of type 2 diabetes mellitus (T2DM) Adverse events can directly influence adherence and involves a combination of changes in lifestyle and persistence to antidiabetic treatment. Hypoglycemias associated pharmacological interventions to achieve target glycated with sulphonylureas (SU) and genital tract infections associated hemoglobin (HbA1c) and, thus, reduced risk of macrovascular with sodium-glucose Cotransporter 2 Inhibitors (SGLT-2i) and microvascular complications (1). However, over time, combined with metformin were the most frequently reported insulin secretory capacity declines, and most people with adverse events in recently published meta-analyses (23–25). T2DM will require escalation of pharmacotherapy to achieve We previously published efficacy results regarding the good metabolic control (2). This is common in real clinical addition of dipeptidyl peptidase-4 inhibitor (DPP-4i), SGLT-2i, practice where first-line treatment with metformin will, in time, or SU as second-line therapies to metformin, showing that users require intensification with a second antidiabetic drug to achieve initiating SGLT-2i in combination with metformin achieved good glycemic control (3–5). According to the current greater reduction in weight and combined target HbA1c therapeutic guidelines, the selection of a second antidiabetic (≥0.5%) and weight (≥3%) reduction among the cohorts (26). drug should be based on patient-specific treatment goals, In the present study, we assessed adherence using the MPR and presence of comorbidities, and drug characteristics (6–10). time till discontinuation of DPP-4i, SGLT-2i, or SU added to Unfortunately, intensification with additional antidiabetic metformin in subjects with T2DM with insufficient glycemic drugs is often delayed, leaving patients with prolonged periods control in a primary care setting. Additionally, we described the of poor glycemic control with worse long term outcomes (11). adverse events associated with these drug combinations. We recently reported a lack of treatment intensification in 1 in 5 patients with HbA1c values >8% (12) and only 20% of the persons with T2DM were treated with dual antidiabetic therapy (13). MATERIAL AND METHODS There are a number of complex barriers to the proper implementation of antidiabetic treatment both on the Study Design and Data Source healthcare professionals’ and patients’ side. Adherence to This was a retrospective cohort study to compare subjects pharmacological treatment plays an important role in initiating add-on treatment with DPP-4i, SGLT-2i, or SU to achieving treatment goals (2). Moreover, data from a meta- metformin. Exposure to these drugs was defined if the user had analysis suggests that good adherence to antidiabetic treatment more than one drug dispensation/prescription register for the was associated with a lower hospitalization rate and all-cause first time between January 1st, 2010, and December 31st, 2017. mortality among the persons with T2DM (14). Treatment Subjects were followed up for a period of 24 months or until adherence and persistence are similar, yet distinct, premature discontinuation. measurements of the degree to which a patient continues Data were obtained from the primary care SIDIAP database treatment after initiation (15). Adherence is defined as “the (The Information System for the development of Primary Care extent to which a patient acts following the prescribed interval Research) (27). This database contains anonymized data from and dosing regimen” (16). Treatment persistence is defined as electronic medical records of the people attended in the 279 the length of time from initiation until discontinuation of Primary Care Teams that belong to the Catalan Health Institute, therapy (16), measured by the drug’s availability, expressed as Catalonia, Spain. The Institute’s assigned population is about the continuous filling of prescriptions (17). 5,835,000 individuals (75% of the total Catalan population). Evidence suggests that adherence to medication in T2DM is Furthermore, the SIDIAP database incorporates laboratory less than optimal, and many patient factors could influence it, data, prescriptions, and data on drug dispensations extracted such as comprehension of the treatment regimen and its benefits, from pharmacy-billing records provided by the Catalan Health emotional well-being, regimen complexity, medication cost and Service (CatSalut). The SIDIAP database has been extensively adverse events (18). RWE (real-world evidence) studies have used for other epidemiologic and pharmacoepidemiologic shown that non-adherence to oral antidiabetic drugs is frequent: national and international research studies, and it is established over 50% in the first year and even higher at the two-year follow- as a well-validated primary care Spanish database for the study of up (19). Low adherence may explain, at least in part, the efficacy diabetes (28, 29). gap in the reduction of HbA1c between RWE studies and randomized clinical trials (RCT) (20). Inclusion and Exclusion Criteria In RWE studies, where prescription or pharmacy claims data Patients were included if they were 18 years or older, diagnosed are available, adherence is usually measured through the with T2DM (ICD-10: E11), and had poor glycemic control medication possession ratio (MPR), where a value of 0.80 (HbA1c ≥7%). We defined the inclusion date when the second (80%) is the cut-off point that stratifies adherent and non- add-on treatment (DPP-4i, SGLT-2i, or SU) was introduced to adherent patients (21). Results from a recently published meta- metformin for the first time. For each treatment group, we analysis confirm the high variability in adherence (38.5 to 93.1%) identified drug exposure (index medication) using ATC codes among different observational studies (2). In another meta- (Anatomical Therapeutic Chemical classification system) from analysis, the proportion of adherent patients was found to be the World Health Organization (WHO) (30), the date of Frontiers in Endocrinology | www.frontiersin.org 2 July 2021 | Volume 12 | Article 708372 Vlacho et al. Adherence of Second Oral Glucose-Lowering Therapy prescription and dispensation. Patients registered with other Main Analysis types of diabetes such as diabetes mellitus type 1, gestational or The MPR and persistence were described by mean, standard secondary (ICD-10: E8, E9, E10, O24, E13), and those subjects deviation, median and interquartile range, while good and poor with missing baseline values for HbA1c and weight were adherence and adverse events were reported by frequency and excluded. Subjects could enter the study groups only once. percentage. We used linear regression models to analyze the differences in MPR as an interval variable among the three Study Variables treatment groups. The associations between good/poor At inclusion, we collected routine information on the social- adherence among the treatment groups were analyzed by logistic demographic characteristics of subjects (age, gender and toxic regression models, summarized as odds ratios (OR), with 98% habits) and clinical characteristics such as laboratory and clinical confidence intervals (CI). All pairwise comparisons (2X2) was parameters related to diabetes control and comorbidities. We conducted between the three groups, where the family significance collected information about drug prescriptions in each treatment level (alpha=0.05) was corrected for multiple paired groups group, dispensations, and both adverse events and (Bonferroni correction), so the individual test was prefixed at discontinuation events during the follow-up period. 0.017, and the confidence level at 98%. To analyze the time to a discontinuation event, we used Cox proportional hazards analysis, Outcomes and hazard ratio (HR), CI, and p-value were summarized. We Adherence was estimated using the medication possession ratio used Kaplan-Meier curves to graphicly visualize treatment (MPR), calculated as the number of days covered by dispensation persistence up to 24 months of the observation period in each divided by the number of days covered by prescription, which is treatment group. As a sensitivity analysis, adjusted estimates were defined as days between the date of initiation of index medication calculated with multivariable models. The variables used for and discontinuation event or up to 24 months. MPR is a adjustment were age, sex, number of comorbidities, weight, validated and standard method to evaluate adherence in HbA1c, year of inclusion, duration of diabetes, and glomerular studies with routinely-collected health data; good adherence filtration rate. The statistical analyses were performed using R3.6.1 was defined as an MPR value >0.8 (>80%), whereas poor software (https://www.r-project.org/). adherence was defined as an MPR value ≤0.8 (≤80%) (21). Persistence was defined as the time between index treatment initiation and the first discontinuation event. For this study, we Ethical Review considered treatment discontinuation events if there was any gap The study was approved by the Ethics Committee of the Primary of at least 90 days (15) without index medication dispensation, Health Care University Research Institute (IDIAP) Jordi Gol, any changes in antidiabetic treatment, death, or moving to Barcelona (approval code: P17/205). another healthcare provider. We calculated the proportion of subjects who discontinued treatment for each treatment group at 6, 12, and up to 24 months of follow-up period. Adverse events were classified into eight categories based on RESULTS the affected system orga n class (SOC) (metabolic, gastrointestinal, hepatic, renal, musculoskeletal, dermatological, Patient Characteristics A total of 75,808 poorly controlled T2DM subjects initiating a hematological, and genitourinary events); these SOCs were second antidiabetic drug in addition to metformin were chosen as they are the most frequently reported adverse included: 27,878 (36.7%) initiated a DPP-4i, 2,198 (2.89%) a reactions in the summary of product characteristics for each SGLT-2i and 45,732 (60.3%) an SU. The study flow chart is drug group. We described the mortality events (any cause) for shown in Supplementary Figure 1. After matching, 6,310 the three groups during the follow-up period. subjects were compared: 2,124 for DPP-4i, 2,124 for SGLT-2i and 2,062 for SU (Supplementary Figure 2). The baseline Statistical Methods characteristics of subjects in each study group are shown in Propensity Score Matching Supplementary Table 1. Overall, the mean age was 60.8 years ( ± The matching criteria were the same as for the previously 11.7), with a mean diabetes duration of 7.61 years ( ± 6.59), and published effectiveness analysis related to changes in glycated an HbA1c of 8.8% ( ± 1.45) (72.3 mmol/mol ( ± 15.9). Subjects in hemoglobin (HbA1c) and the effect on body weight following the the DPP-4i group were older with a mean age of 61.2 ( ± 12.1), addition of DPP-4i, SGLT-2i, or SU as second-line therapies to while those in the SGLT-2i treatment group had a longer diabetes metformin (26). The three treatments groups were matched for duration of 7.89 ( ± 6.67) and had a higher BMI 33.9 ( ± 5.80) the following baseline characteristics: weight, HbA1c, sex, age, compared to the other groups. SGLT-2i users also had slightly diabetes duration, year of inclusion, and kidney function. higher triglycerides and a worse comorbidity profile, especially Matching was done by the “Nearest Neighbor algorithm” for cardiovascular complications. The baseline characteristics (caliper=0.01), using the “MatchIt” library of the R (v3.6.1) and analysis of effectiveness among the three treatment groups statistical package (31). have been recently published (26). Frontiers in Endocrinology | www.frontiersin.org 3 July 2021 | Volume 12 | Article 708372 Vlacho et al. Adherence of Second Oral Glucose-Lowering Therapy Adherence to Treatment we only observed statistical differences between SU and DPP-4i Table 1 summarises the data related to adherence and drug users (1.02 fewer packages in the former group; adjusted DR: dispensations. Comparison between study groups showed that -1.02, 98% CI: -1.59, -0.46). The odds ratios for good adherence, good adherence (MPR>0.8) was achieved for most of the SGLT- MPR differences and the number of dispensed packages among 2i and DPP-4i treated subjects (68.7% and 53.6%, respectively), the treatment groups are shown in Table 2. while the majority of SU users had poor adherence (43.0%). Supplementary Table 2 shows adherence for different drug Treatment Persistence within the drug groups. Alogliptin in combination with Table 1 summarizes the results of discontinuation and metformin had the highest mean MPR in the DPP-4i group persistence in the 3 study groups. The mean time until (0.81 ± 0.28), canagliflozin in combination with metformin in the discontinuation was longer in the SGLT-2i group in SGLT-2i group (0.82 ± 0.30), and glimepiride in combination comparison with the DPP-4i or SU groups: 385 ( ± 289), 372 with metformin in the SU group (0.91 ± 0.21). Multiple logistic ( ± 330) and 343 ( ± 306) days, respectively. During the initial six regression analysis showed that SGLT-2i users were 1.7 and 2.8 month period, 21.3% of SU users discontinued treatment, times more likely to be associated with good adherence than compared with 18.6% of SGLT-2i users and only 12.9% of DPP-4i users (adjusted OR: 1.72, 98% CI: 1.51, 1.96), or SU users DPP-4i users. At the end of the 24-month follow-up period, (adjusted OR: 0.35, 98% CI: 0.07, 0.29), respectively. 43.0% of SU users, 39.7% of SGLT-2i users, and 28.8% of DPP-4i The DPP-4i users were 1.6 times more likely to be associated users had ceased treatment. with good adherence than SU users (OR: 0.59, 98% CI: 0.52, The Kaplan-Meier curves of persistence are shown in 0.67). A mean difference in MPR of 6% was observed between Figure 1 and summarized in Supplementary Table 3. We SGLT-2i users and DPP-4i users (adjusted DR: 0.06, 98% CI: performed a Cox proportional hazards analysis to compare the 0.04, 0.08) and 14% compared with SU users (adjusted DR: -0.14, hazard risk ratios for discontinuation events. The risk of 98% CI: -0.16, -0.11); the difference was 8% between DPP-4i and discontinuation was 1.4 times higher for SGLT-2i (HR: 1.43, SU users (adjusted DR: -0.08, 98% CI: -0.10, -0.06). Comparing 98% CI: 1.26, 1.62) and 1.6 times higher for SU (HR: 1.60, 98% the difference in number of packages dispensed between groups, CI: 1.42, 1.81) compared to DPP-4i. Furthermore, the risk of TABLE 1 | Medical possession ratio, adherence, persistence (time until discontinuation and discontinuations) among the three treatment groups. MET+ DPP-4i (n = 2113) MET+SGLT-2i (n = 2117) MET+ SU (n = 2056) Medical possession ratio (MPR) Medication possession ratio, Mean (SD) 0.71 (0.34)** 0.78 (0.34)** 0.63 (0.35)** Medication possession ratio, Median [IQR: 25th;75th] 0.86 [0.43;1.00]** 1.00 [0.62;1.00]** 0.64[0.33;1.00]** Number medicine packages dispensed, Mean (SD) 10.6 (8.62)* 10.7 (8.52)* 9.88 (12.9)* Poor adherence (≤ 0.8) 981 (46.4%)** 662 (31.3%)** 1172 (57.0%)** Good adherence (>0.8) 1132 (53.6%)** 1455 (68.7%)** 884 (43.0%)** Persistence Persistence time on treatment, Mean (SD) 372 (330)** 385 (289)** 343 (306)** Persistence time on treatment, Median, [IQR: 25th;75th] 274 [121;548]** 333 [150;600]** 272 [91.2;486]** Discontinuation events Discontinuation of treatment 6 m: % (98% CI Linf, Lsup) 12.9 (11.4, 14.5) 18.6 (16.8, 20.4) 21.3 (19.4, 23.2) Discontinuation of treatment 12 m: % (98% CI Linf, Lsup) 20.1(18.0, 22.0) 28.6 (26.4, 30.7) 32.1 (29.7, 34.3) Discontinuation of treatment 24 m: % (98% CI Linf, Lsup) 28.8 (26.0, 31.4) 39.7(36.8, 42.5) 43.0 (39.8, 48.9) CI, confidence interval; DPP-4i, dipeptidyl peptidase-4 inhibitors; IQR, inter-quartile range; Linf, inferior limit; Lsup, superior limit; m, months; MET, metformin; SD, standard deviation; SGLT-2i, sodium/glucose cotransporter 2 inhibitors; SU, sulphonylureas; *p-value =0.018; **p-value <0.001. TABLE 2 | Odds ratios for good adherence, MPR differences and number of dispensed packages among the cohorts. Good adherence (MPR>0.8) MPR differences Number of dispensed packages differences Unadjusted OR Adjusted OR Unadjusted DR Adjusted DR Unadjusted DR Adjusted DR (98% CI) (98% CI) (98% CI) (98% CI) (98% CI) (98% CI) SGLT-2i+MET, ref: DPP-4i+MET 1.90 (1.67, 2.16)* 1.72 (1.51, 1.96)* 0.08 (0.06, 0.1)* 0.06 (0.04, 0.08)* 0.13 (-0.49, 0.74) -0.63 (-1.19,-0.06) SU+ MET, ref: DPP-4i+MET 0.65 (0.57,0.73)* 0.59 (0.52, 0.67)* -0.07 (-0.09,-0.05)* -0.08 (-0.10, -0.06)* -0.71 (-1.33,-0.08) -1.02 (-1.59,-0.46)* SU +MET, ref: SGLT-2i+MET 0.34 (0.29, 0.40)* 0.35 (0.07, 0.29)* -0.15(-0.17, -0.13)* -0.14 (-0.16, -0.11)* -0.83 (-1.6,-0.08)* -0.34 (-1.09,0.29) *Statistically significant p-value (p-value <0.017). CI, confidence interval; IDPP-4i, dipeptidyl peptidase-4 inhibitors; OR, odds ratio; DR: differences; SGLT-2i, sodium/glucose cotransporter 2 inhibitors; SU, sulphonylureas; MET, metformin. Frontiers in Endocrinology | www.frontiersin.org 4 July 2021 | Volume 12 | Article 708372 Vlacho et al. Adherence of Second Oral Glucose-Lowering Therapy FIGURE 1 | Kaplan-Meier discontinuation probability curves for the three treatment groups. discontinuation among SU users was 1.1 times higher than that Comparing the adherence among the study groups, 68.7% of of SGLT-2i users (HR: 1.12, 98% CI: 1.00, 1.26). users in the SGLT-2i treatment group had good adherence (MPR>0.8), while this percentage was lower in both DPP-4i Adverse Events and SU users (53.6% and 43%, respectively). In an observational The results for adverse events are reported in Table 3. We observed study with 11,961 subjects in the US, the percentage of good that gastrointestinal, musculoskeletal, dermatological, and urogenital adherence (MPR≥0.8) for subjects initiating an SGLT-2i was were the most frequent adverse events during the follow-up period in 56.2-58.8% for canagliflozin, 36.4-36.7% for dapagliflozin and all three groups. In the SGLT-2i group, urogenital, metabolic and 45.7% for sitagliptin after 12 months (15). Our study showed a dermatological adverse reactions were more frequent than in the similar tendency, although with a higher level of good adherence: other treatment groups (10.5%, 0.19% and 3.01%, respectively) but canagliflozin 62.5-73.7%, dapagliflozin 71.2-71.7%, and without statistically significant differences between groups. There sitagliptin 49.9-51.8%. In an observational study with 171,220 were no significant differences in the frequency of gastrointestinal T2DM subjects from Sweden during 2005 and 2006, the refill adverse events between the groups. adherence for dual therapy with SUs (glibenclamide, glipizide, glimepiride) was high (91.3%, 91.0%, and 91.7%, respectively) (32), however the proportion of subjects with good adherence in DISCUSSION our study was lower for the same drugs (50%, 85.3% and 44.4%, respectively). In the current study, among 6,310 propensity score-matched Our results show that SGLT-2i users were more likely to have users who initiated a second line add-on therapy to metformin good adherence to treatment than DPP-4i and SU users (1.7 and with DPP-4i, SGLT-2i or SU in Catalonia, the highest adherence 2.8 times higher, respectively). Results from an administrative- and persistence was observed in SGLT-2i users. claims study in the US reported that patients who initiated an TABLE 3 | Adverse events among the treatment groups. Adverse event, n (%) MET+ DPP-4i (n = 2113) MET+SGLT-2i (n = 2117) MET+ SU (n = 2056) Metabolic adverse event 0 (0.00%) 4 (0.19%) 0 (0.00%) Gastrointestinal adverse events 156 (7.34%) 149 (7.02%) 145 (7.03%) Hepatic adverse events 26 (1.22%) 20 (0.94%) 23 (1.12%) Kidney adverse events 25 (1.18%) 28 (1.32%) 12 (0.58%) Musculoskeletal system adverse events 57 (2.68%) 57 (2.68%) 48 (2.37%) Dermatological adverse events 54 (2.54%) 64 (3.01%) 49 (2.38%) Hematological adverse events 1 (0.05%) 2 (0.09%) 0 (0.00%) Urogenital adverse events 161 (7.58%) 223 (10.5%) 160 (7.76%) Death by any cause 33 (1.55%) 29 (1.37%) 39 (1.89%) *No statistically significant difference were observed among the groups; DPP-4i, dipeptidyl peptidase-4 inhibitors; SGLT-2i, sodium/glucose cotransporter 2 inhibitors; SU, sulphonylureas; MET, metformin. Frontiers in Endocrinology | www.frontiersin.org 5 July 2021 | Volume 12 | Article 708372 Vlacho et al. Adherence of Second Oral Glucose-Lowering Therapy SGLT-2 inhibitor were 1.36 times more likely to be adherent to adherence and may be influenced by both efficacy and their medication and 1.35 times less likely to discontinue their tolerability; a subject who experiences undesired side effects of medication than patients who initiated an SU (33). In another medication is less likely to take the prescribed drug (15). One UK RWE study from the US, comparing the DPP-4i sitagliptin with study reported that gastrointestinal side effects, hypoglycemia, SUs as an add-on therapy to metformin, subjects in the SU group weight change, and efficacy were the most important factors had lower adherence and persistence (34). Similar findings were determining patient preferences for oral antidiabetic drugs (42). observed in a retrospective RWE study with 238,372 subjects, In our study, efficacy in combination with weight reduction where DPP-4i users had a significantly greater OR of being could be the reason for good adherence and persistence among adherent than users initiating an SU; the authors pointed to a the SGLT-2i and DPP-4i users. Our previous study showed that better tolerability profile of DPP-4i as an explanation of their the addition of SGLT-2i or DPP-4i to metformin was associated findings (19). with a greater weight reduction (3.47 kg and 1.21 kg, About 79.9% of DPP-4i users, 71.4% of SGLT-2i users, and respectively) (26). Additionally, the proportion of subjects who 67.9% of SU users persisted with their initial therapy during the achieved a combined target of HbA1c (≥0.5%) and weight (≥3%) first year of treatment in our study. In an RWE study by Farr reduction was greater in these two treatment groups (26). et al. (19), the authors reported that over 40% of SU initiators Previous studies have shown that the presence of certain stopped refiling in the first year. In another RWE study from conditions, such as depression and mental disorders, before the Hungary, the persistence rate after the first 12 months was 69.6% first antidiabetic drug prescription is associated with non- for DPP-4i users and 67.8% for SGLT-2i users (35). A meta- persistence to antidiabetic treatment (43). In our study, the analysis of previous studies of treatment persistence to oral lowest persistence and adherence were observed among SU antidiabetic drugs reported persistence rates ranging from 33 users; however, at baseline, these users had fewer mental to 61%, with an overall mean percentage of persistence of 49.2% disorder comorbidities than the users included in the other (95% CI: 40.1%– 58.3%) among the studies that investigated only treatment groups. persistence to the index medication (22). We found that the risk There are some limitations to our study. As per the study of treatment discontinuation among SGLT-2i and SU users design, we only included subjects with complete data for baseline compared with DPP-4i users during the follow-up period was HbA1c and weight; the study population is a highly selected 43% and 60% higher, respectively. These results are in line with sample which potentially diminishes the external validity. other RWE studies, where the risk of discontinuation was 40% Moreover, we cannot rule out that patients having both higher among SU initiators (adjusted HR: 1.390, 95% CI: 1.363, variables at baseline were treated more proactively to favor a 1.418) (19) and 6% higher among SGLT-2i users (HR: 1.066, 95% better T2DM control; however, due to the matching process, we CI: 1.036–1.096) (35) compared with DPP-4i users. High would expect this limitation to be the same for the three study discontinuation rates and poor adherence are important factors groups. Another limitation is the relatively small sample size to that may induce possible issues with the initially prescribed observe the number of adverse events, mainly due to the treatment. In the current study, among the users who initiated propensity score matching, which drastically reduces the SU in combination with metformin, despite the relatively large population size and, thus, the total number of adverse events. gap period (90 days) without dispensation, two of ten subjects However, our goal was to describe the number of events among stopped the initial treatment during the initial six months. On the treatment groups as opposed to analyzing statistically the other hand, in the SGLT-2i group, the percentage of users significant differences. Finally, the retrospective nature of the with good adherence was higher, but discontinuation rates were study precludes explaining the reason for treatment higher than in the DPP-4i users. A possible explanation for this discontinuation, so we are not able to discern whether the could be an improved tolerability in the DPP-4i group (36). differences in adherence could be due to the drug itself, to the With regards to safety, the most frequently reported adverse risk of adverse events, to the number of pills per day or to the use events were gastrointestinal and urogenital disturbances. We of available fixed-dose combinations. Strengths of our study found higher percentages of urogenital, metabolic and include a population-based cohort, long follow-up of two dermatological adverse events in the SGLT-2i group and more years, propensity matching and outcomes for adherence, frequent gastrointestinal events in the DPP-4i group but without persistence and adverse events. significant differences. Indeed, it is well reported that SGLT-2i In conclusion, the results of the present study show better drugs are often associated with a higher incidence of urogenital drug adherence and longer persistence among subjects on SGLT- infections (mycotic genital infections such as vaginitis in women 2i as an add-on to metformin compared with DPP-4i or SU and balanitis in man) (37–39). However, despite the occurrence users. Subjects being treated with DPP-4i combined with of these adverse effects, these episodes are often regarded as mild metformin had the fewest discontinuation events during the by patients (40); additionally, their incidence tends to decline follow-up period. These results may help clinicians better over time without the need for halting SGLT-2i therapy (41). understand the treatment trajectory following the addition of Studies have previously shown that achieving better DPP-4i, SGLT-2i, or SU to metformin. However, further studies adherence is associated with improved glycemic control (17), in real-world conditions are needed to identify factors related to but many factors could influence adherence and persistence to good adherence, persistence and safety amongst these three treatment. The patient is the primary driver of treatment commonly prescribed drug combinations. Frontiers in Endocrinology | www.frontiersin.org 6 July 2021 | Volume 12 | Article 708372 Vlacho et al. Adherence of Second Oral Glucose-Lowering Therapy DATA AVAILABILITY STATEMENT FUNDING The data analyzed in this study is subject to the following licenses/ This study received funding from AstraZeneca, Spain [grant restrictions: The data controller for SIDIAP does not allow the number ESR-16-12628]. The funder was not involved in the sharing of raw data. The source code is available at https://github. study design, collection, analysis, interpretation of data, the com/jrealgatius/METPLUS. Requests to access these datasets writing of this article or the decision to submit it for publication. should be directed to JF-N, dap.cat.info@gmail.com. ETHICS STATEMENT ACKNOWLEDGMENTS The studies involving human participants were reviewed and We thank Amanda Prowse for the editorial support of this approved by Ethics Committee of the Primary Health Care manuscript. KK is supported by the National Institute for University Research Institute (IDIAP) Jordi Gol, Barcelona Health Research (NIHR) Applied Research Collaboration East (approval code: P17/205). Written informed consent for Midlands (ARC EM) and the NIHR Leicester Biomedical participation was not required for this study in accordance Research Centre (BRC). BV is supported by the Instituto de with the national legislation and the institutional requirements. Salud Carlos III (ISCIII, Fondo Investigació n sanitaria-FIS- Fondo Europeo de Desarrollo Regional-FEDER: Plataforma de Unidades de Investigació n Clı́nica y Ensayos Clı́nicos-SCREN. AUTHOR CONTRIBUTIONS JF-N, MM-C, JR, DM, XM-T, JAV-C, and BV conceived the research and participated in its design. JR performed the SUPPLEMENTARY MATERIAL statistical analysis. BV wrote the initial draft of the manuscript, which JF-N, MM-C, JR, DM, XM-T, JAV-C, XC MF and KK The Supplementary Material for this article can be found online edited. All authors contributed to the article and approved the at: https://www.frontiersin.org/articles/10.3389/fendo.2021. submitted version. 708372/full#supplementary-material REFERENCES Diabetes (SED, 2015). Atenció n Primaria (2016) 48:325–36. doi: 10.1016/ j.aprim.2015.05.002 1. Ritz E, Rychlı́k I, Locatelli F, Halimi S. End-Stage Renal Failure in Type 2 10. Seidu S, Cos X, Brunton S, Harris SB, Jansson SPO, Mata-Cases M, et al. A Diabetes: A Medical Catastrophe of Worldwide Dimensions. Am J Kidney Dis Disease State Approach to the Pharmacological Management of Type 2 (1999) 34:795–808. doi: 10.1016/S0272-6386(99)70035-1 Diabetes in Primary Care: A Position Statement by Primary Care Diabetes 2. Krass I, Schieback P, Dhippayom T. Adherence to Diabetes Medication: A Europe. Prim Care Diabetes (2021) 15:31–51. doi: 10.1016/j.pcd.2020.05.004 Systematic Review. Diabetes Med (2015) 32:725–37. doi: 10.1111/dme.12651 11. Khunti K, Wolden ML, Thorsted BL, Andersen M, Davies MJ. Clinical Inertia 3. Turner RC, Cull CA, Frighi V, Holman RRGroup for the UKPDS (UKPDS). in People With Type 2 Diabetes: A Retrospective Cohort Study of More Than Glycemic Control With Diet, Sulfonylurea, Metformin, or Insulin in Patients 80,000 People. Diabetes Care (2013) 36:3411–7. doi: 10.2337/dc13-0331 With Type 2 Diabetes Mellitusprogressive Requirement for Multiple 12. Mata-Cases M, Franch-Nadal J, Real J, Gratacòs M, Ló pez-Simarro F, Khunti Therapies (UKPDS 49). JAMA (1999) 281:2005–12. doi: 10.1001/ K, et al. Therapeutic Inertia in Patients Treated With Two or More jama.281.21.2005 Antidiabetics in Primary Care: Factors Predicting Intensification of 4. Mata-Cases M, Franch-Nadal J, Real J, Mauricio D. Glycaemic Control and Treatment. Diabetes Obes Metab (2018) 20:103–12. doi: 10.1111/dom.13045 Antidiabetic Treatment Trends in Primary Care Centres in Patients With 13. Mata-Cases M, Franch-Nadal J, Real J, Cedenilla M, Mauricio D. Prevalence Type 2 Diabetes Mellitus During 2007–2013 in Catalonia: A Population- and Coprevalence of Chronic Comorbid Conditions in Patients With Type 2 Based Study. BMJ Open (2016) 6:e012463. doi: 10.1136/bmjopen-2016- Diabetes in Catalonia: A Population-Based Cross-Sectional Study. BMJ Open 012463 (2019) 9:e031281. doi: 10.1136/bmjopen-2019-031281 5. Turner LW, Nartey D, Stafford RS, Singh S, Alexander GC. Ambulatory 14. Khunti K, Seidu S, Kunutsor S, Davies M. Association Between Adherence to Treatment of Type 2 Diabetes in the U.S., 1997–2012. Diabetes Care (2014) Pharmacotherapy and Outcomes in Type 2 Diabetes: A Meta-Analysis. 37:985–92. doi: 10.2337/dc13-2097 Diabetes Care (2017) 40:1588–96. doi: 10.2337/dc16-1925 6. American Diabetes Association. 9.Pharmacologic Approaches to Glycemic 15. Cai J, Wang Y, Baser O, Xie L, Chow W. Comparative Persistence and Treatment: Standards of Medical Care in Diabetes—2020. Diabetes Care Adherence With Newer Anti-Hyperglycemic Agents to Treat Patients With (2020) 43:S98–S110. doi: 10.2337/dc20-S009 Type 2 Diabetes in the United States. J Med Econ (2016) 19:1175–86. 7. Davies MJ, D’Alessio DA, Fradkin J, Kernan WN, Mathieu C, Mingrone G, doi: 10.1080/13696998.2016.1208208 et al. Management of Hyperglycaemia in Type 2 Diabetes, 2018. A Consensus 16. Cramer JA, Roy A, Burrell A, Fairchild CJ, Fuldeore MJ, Ollendorf DA, et al. Report by the American Diabetes Association (ADA) and the European Medication Compliance and Persistence: Terminology and Definitions. Value Association for the Study of Diabetes (EASD). Diabetologia (2018) 61:2461– Health (2008) 11:44–7. doi: 10.1111/j.1524-4733.2007.00213.x 98. doi: 10.1007/s00125-018-4729-5 17. Guerci B, Chanan N, Kaur S, Jasso-Mosqueda JG, Lew E. Lack of Treatment 8. Mediavilla Bravo JJ. Guı́as En El Manejo De La Diabetes Mellitus Tipo 2. Persistence and Treatment Nonadherence as Barriers to Glycaemic Control in Semer - Med Fam (2014) 40:11–8. doi: 10.1016/S1138-3593(14)74392-8 Patients With Type 2 Diabetes. Diabetes Ther (2019) 10:437–49. doi: 10.1007/ 9. Arrieta F, Iglesias P, Pedro-Botet J, Té bar FJ, Ortega E, Nubiola A, et al. s13300-019-0590-x Diabetes Mellitus Y Riesgo Cardiovascular: Recomendaciones Del Grupo De 18. Rubin RR. Adherence to Pharmacologic Therapy in Patients With Type 2 Diabetes Trabajo Diabetes Y Enfermedad Cardiovascular De La Sociedad Española De Mellitus. Am J Med (2005) 118:27–34. doi: 10.1016/j.amjmed.2005.04.012 Frontiers in Endocrinology | www.frontiersin.org 7 July 2021 | Volume 12 | Article 708372 Vlacho et al. Adherence of Second Oral Glucose-Lowering Therapy 19. Farr AM, Sheehan JJ, Curkendall SM, Smith DM, Johnston SS, Kalsekar I. Peptidase-4 Inhibitors, Sodium-Glucose Co-Transporter-2 Inhibitors, and Retrospective Analysis of Long-Term Adherence to and Persistence With Glucagon-Like Peptide-1 Receptor Agonists) in People With Type 2 DPP-4 Inhibitors in US Adults With Type 2 Diabetes Mellitus. Adv Ther Diabetes: A Nationwide Cohort Study. Diabetes Ther (2018) 9:2133–41. (2014) 31:1287–305. doi: 10.1007/s12325-014-0171-3 doi: 10.1007/s13300-018-0483-4 20. Carls GS, Tuttle E, Tan R-D, Huynh J, Yee J, Edelman SV, et al. Understanding the 36. Avogaro A, Delgado E, Lingvay I. When Metformin Is Not Enough: Pros and Gap Between Efficacy in Randomized Controlled Trials and Effectiveness in Real- Cons of SGLT2 and DPP-4 Inhibitors as a Second Line Therapy. Diabetes World Use of GLP-1 RA and DPP-4 Therapies in Patients With Type 2 Diabetes. Metab Res Rev (2018) 34:e2981. doi: 10.1002/dmrr.2981 Diabetes Care (2017) 40:1469–78. doi: 10.2337/dc16-2725 37. Chen Z, Li G. Sodium-Glucose Co-Transporter 2 Inhibitors Compared With 21. Karve S, Cleves MA, Helm M, Hudson TJ, West DS, Martin BC. Good and Sulfonylureas in Patients With Type 2 Diabetes Inadequately Controlled on Poor Adherence: Optimal Cut-Point for Adherence Measures Using Metformin: A Meta-Analysis of Randomized Controlled Trials. Clin Drug Administrative Claims Data. Curr Med Res Opin (2009) 25:2303–10. Investig (2019) 39:521–31. doi: 10.1007/s40261-019-00781-w doi: 10.1185/03007990903126833 38. Li D, Wang T, Shen S, Fang Z, Dong Y, Tang H. Urinary Tract and Genital 22. Iglay K, Cartier SE, Rosen VM, Zarotsky V, Rajpathak SN, Radican L, et al. Infections in Patients With Type 2 Diabetes Treated With Sodium- Meta-Analysis of Studies Examining Medication Adherence, Persistence, and Glucose Co-Transporter 2 Inhibitors: A Meta-Analysis of Randomized Discontinuation of Oral Antihyperglycemic Agents in Type 2 Diabetes. Curr Controlled Trials. Diabetes Obes Metab (2017) 19:348–55. doi: 10.1111/ Med Res Opin (2015) 31:1283–96. doi: 10.1185/03007995.2015.1053048 dom.12825 23. Wang Z, Sun J, Han R, Fan D, Dong X, Luan Z, et al. Efficacy and Safety of 39. Zaccardi F, Webb DR, Htike ZZ, Youssef D, Khunti K, Davies MJ. Efficacy and Sodium-Glucose Cotransporter-2 Inhibitors Versus Dipeptidyl Peptidase-4 Safety of Sodium-Glucose Co-Transporter-2 Inhibitors in Type 2 Diabetes Inhibitors as Monotherapy or Add-on to Metformin in Patients With Type 2 Mellitus: Systematic Review and Network Meta-Analysis. Diabetes Obes Diabetes Mellitus: Systematic Review and Meta-Analysis. Diabetes Obes Metab (2016) 18:783–94. doi: 10.1111/dom.12670 Metab (2018) 20:113–20. doi: 10.1111/dom.13047 40. Leiter LA, Yoon K-H, Arias P, Langslet G, Xie J, Balis DA, et al. Canagliflozin 24. Maruthur NM, Tseng E, Hutfless S, Wilson LM, Suarez-Cuervo C, Berger Z, Provides Durable Glycemic Improvements and Body Weight Reduction Over et al. Diabetes Medications as Monotherapy or Metformin-Based 104 Weeks Versus Glimepiride in Patients With Type 2 Diabetes on Combination Therapy for Type 2 Diabetes. Ann Intern Med (2016) 164:740. Metformin: A Randomized, Double-Blind, Phase 3 Study. Diabetes Care doi: 10.7326/M15-2650 (2015) 38:355–64. doi: 10.2337/dc13-2762 25. Li J, Gong Y, Li C, Lu Y, Liu Y, Shao Y. Long-Term Efficacy and Safety of 41. Del Prato S, Nauck M, Durá n-Garcia S, Maffei L, Rohwedder K, Theuerkauf Sodium-Glucose Cotransporter-2 Inhibitors as Add-On to Metformin A, et al. Long-Term Glycemic Response and Tolerability of Dapagliflozin Treatment in the Management of Type 2 Diabetes Mellitus. Med Versus a Sulphonylurea as Add-On Therapy to Metformin in Patients With (Baltimore) (2017) 96:e7201. doi: 10.1097/MD.0000000000007201 Type 2 Diabetes: 4-Year Data. Diabetes Obes Metab (2015) 17:581–90. 26. Vlacho B, Mundet-Tudurı́ X, Mata-Cases M, Vallès-Callol JA, Real J, Farré M, et al. doi: 10.1111/dom.12459 Analysis of the Effectiveness of Second Oral Glucose-Lowering Therapy in Routine 42. Gelhorn HL, Stringer SM, Brooks A, Thompson C, Monz BU, Boye KS, et al. Clinical Practice From the Mediterranean Area: A Retrospective Cohort Study. Preferences for Medication Attributes Among Patients With Type 2 Diabetes Diabetes Res Clin Pract (2021) 171:108616. doi: 10.1016/j.diabres.2020.108616 Mellitus in the UK. Diabetes Obes Metab (2013) 15:802–9. doi: 10.1111/ 27. Bolı́bar B, Avilés FF, Morros R, Gil M del MG, Hermosilla E, Blanes RR, et al. Base dom.12091 De Datos SIDIAP: La Historia Clı́nica Informatixada De Atenció n Primaria Como 43. Guénette L, Moisan J, Breton M-C, Sirois C, Grégoire J-P. Difficulty Adhering to Fuente De Informació n Para La Investigació n Epidemioló gica. Med Clin (Barc) Antidiabetic Treatment: Factors Associated With Persistence and Compliance. (2012) 138:617–21. doi: 10.1016/j.medcli.2012.01.020 Diabetes Metab (2013) 39:250–7. doi: 10.1016/j.diabet.2012.12.005 28. Mata-Cases M, Franch-Nadal J, Mauricio D, Bolı́bar B. Investigar En Diabetes Desde Una Base De Datos De Atenció n Primaria: La Experiencia Del Sistema Conflict of Interest: MM-C has received an advisory and or speaking fees from De Informació n Para El Desarrollo De La Investigació n En Atenció n Primaria Astra-Zeneca, Bayer, Boehringer Ingelheim, GSK, Lilly, MSD, Novartis, Novo (SIDIAP). Av en Diabetol (2013) 29:169–74. doi: 10.1016/j.avdiab.2013.09.002 Nordisk, and Sanofi; he has received research grants to the institution from Astra- 29. Mata-Cases M, Mauricio D, Real J, Bolı́bar B, Franch-Nadal J. Is Diabetes Zeneca, GSK, Lilly, MSD, Novartis, Novo Nordisk, and Sanofi. JF-N has received Mellitus Correctly Registered and Classified in Primary Care? A Population- advisory and or speaking fees from Astra-Zeneca, Ascensia, Boehringer Ingelheim, Based Study in Catalonia, Spain. Endocrinol Nutr (2016) 63:440–8. GSK, Lilly, MSD, Novartis, Novo Nordisk, and Sanofi; he has received research doi: 10.1016/j.endonu.2016.07.004 grants to the institution from Astra-Zeneca, GSK, Lilly, MSD, Novartis, Novo 30. WHO Collaborating Centre for Drug Statistics Methodology. Purpose of the Nordisk, Sanofi, and Boehringer. DM has received advisory and/or speaking fees ATC/DDD System. Norwegian Institute for Public Health (2018). Available at: from Astra-Zeneca, Ascensia, Boehringer Ingelheim, GSK, Lilly, MSD, Novartis, https://www.whocc.no/atc_ddd_methodology/purpose_of_the_atc_ddd_ Novo Nordisk, and Sanofi; he has received research grants to the institution from system/. Astra-Zeneca, GSK, Lilly, MSD, Novartis, Novo Nordisk, Sanofi, and Boehringer. 31. Ho DE, Imai K, King G, Stuart EA. Matching as Nonparametric Preprocessing KK has acted as a consultant, speaker or received grants for investigator-initiated for Reducing Model Dependence in Parametric Causal Inference. Polit Anal studies for Astra Zeneca, Novartis, Novo Nordisk, Sanofi-Aventis, Lilly and Merck (2007) 15(3):199–236. doi: 10.1093/pan/mpl013 Sharp & Dohme, Boehringer Ingelheim, Bayer, Berlin-Chemie AG/Menarini 32. Haupt D, Weitoft GR, Nilsson JLG. Refill Adherence to Oral Group, Janssen, and Napp. Antihyperglycaemic Drugs in Sweden. Acta Diabetol (2009) 46:203–8. doi: 10.1007/s00592-008-0076-1 The remaining authors declare that the research was conducted in the absence of 33. Bell KF, Cappell K, Liang M, Kong AM. Comparing Medication Adherence any commercial or financial relationships that could be construed as a potential and Persistence Among Patients With Type 2 Diabetes Using Sodium- conflict of interest. Glucose Cotransporter 2 Inhibitors or Sulfonylureas. Am Health Drug Benefits (2017) 10:165–74. Copyright © 2021 Vlacho, Mata-Cases, Mundet-Tudurı,́ Vallès-Callol, Real, Farre, 34. Bloomgarden ZT, Tunceli K, Liu J, Brodovicz KG, Mavros P, Engel SS, et al. Cos, Khunti, Mauricio and Franch-Nadal. This is an open-access article distributed Adherence, Persistence, and Treatment Discontinuation With Sitagliptin under the terms of the Creative Commons Attribution License (CC BY). The use, Compared With Sulfonylureas as Add-Ons to Metformin: A Retrospective distribution or reproduction in other forums is permitted, provided the original Cohort Database Study. J Diabetes (2017) 9:677–88. doi: 10.1111/1753-0407.12461 author(s) and the copyright owner(s) are credited and that the original publication in 35. Jermendy G, Kiss Z, Rokszin G, Abonyi-Tó th Z, Wittmann I, Kempler P. this journal is cited, in accordance with accepted academic practice. No use, Persistence to Treatment With Novel Antidiabetic Drugs (Dipeptidyl distribution or reproduction is permitted which does not comply with these terms. Frontiers in Endocrinology | www.frontiersin.org 8 July 2021 | Volume 12 | Article 708372