Diabetes Ther (2018) 9:1169–1184 https://doi.org/10.1007/s13300-018-0429-x ORIGINAL RESEARCH Treatment Intensiﬁcation in Type 2 Diabetes: A Real- World Study of 2-OAD Regimens, GLP-1 RAs, or Basal Insulin . . . . Lawrence Blonde Denis Raccah Elisheva Lew Juliana Meyers . . . . Elena Nikonova Mayank Ajmera Keith L. Davis Monica Bertolini Bruno Guerci Received: January 23, 2018 / Published online: April 19, 2018 The Author(s) 2018 receptor agonists (GLP-1 RAs), or basal insulin ABSTRACT in a real-world setting. Methods: Adults with T2D on OAD monother- Introduction: Treatment guidelines recom- apy were identiﬁed in the MarketScan claims mend a stepwise approach to glycemia man- database (2007–2014). Those initiating two agement in patients with type 2 diabetes (T2D), OADs (simultaneously or sequentially), GLP-1 but this may result in uncontrolled glycated RAs, or basal insulin were selected (date of ini- hemoglobin A1c (HbA1c) between steps. This tiation was termed the ‘index date’); patients retrospective analysis compared clinical and were required to have HbA1c [ 7.0% in the economic outcomes among patients with 6 months pre-index date. HbA1c was compared uncontrolled T2D initiating two oral antidia- from 6 months pre- to 1-year post-index. betes drugs (OADs), glucagon-like peptide-1 Annual all-cause healthcare utilization and costs were reported over the 1-year follow-up period. Enhanced Digital Features To view enhanced digital Results: Data for 6054 patients were analyzed features for this article, go to https://doi.org/10.6084/ (2-OAD, n = 4442; GLP-1 RA, n = 361; basal m9.ﬁgshare.6115208. insulin, n = 1251). Baseline HbA1c was high in Electronic supplementary material The online all cohorts, but highest in the basal-insulin version of this article (https://doi.org/10.1007/s13300- cohort. Treatment initiation resulted in reduc- 018-0429-x) contains supplementary material, which is available to authorized users. tions in HbA1c in all cohorts, which was E. Nikonova L. Blonde (&) Eisai Inc, Woodcliff Lake, NJ, USA Department of Endocrinology, Ochsner Medical Center, New Orleans, LA, USA M. Bertolini e-mail: firstname.lastname@example.org Sanoﬁ, Paris, France D. Raccah B. Guerci University Hospital Sainte Marguerite, Marseille, Diabetology Department, University of Lorraine, France Vandoeuvre-Le`s-Nancy, France E. Lew Sanoﬁ, Chilly-Mazarin, France J. Meyers M. Ajmera K. L. Davis RTI Health Solutions, Research Triangle Park, Durham, NC, USA 1170 Diabetes Ther (2018) 9:1169–1184 generally maintained throughout the follow-up metformin. The ADA recommends the addition period. Average HbA1c reductions from the of one of the following six available treatment 6 months pre- to 1 year post-index date were options: sulfonylurea, thiazolidinedione, -1.2% for GLP-1 RA, -1.6% for OADs, and dipeptidyl peptidase-4 (DPP-4) inhibitor, -1.8% for basal insulin. HbA1c \ 7.0% at 1 year sodium-glucose co-transporter 2 (SGLT-2) inhi- occurred in 32.6%, 47.5%, and 41.1% of bitor, glucagon-like peptide-1 receptor agonist patients, respectively. Annual healthcare costs (GLP-1 RA), or basal insulin . If the HbA1c (mean [SD]) were lowest for OAD (US$10,074 goal is not achieved or the patient’s glycemic [$22,276]) followed by GLP-1 RA (US$14,052 control is not maintained with dual therapy, [$23,829]) and basal insulin (US$18,813 triple therapy can be initiated . For patients [$37,332]). with marked hyperglycemia (HbA1c C 10.0%) Conclusion: Despite robust HbA1c lowering at diagnosis, initial treatment with insulin may following treatment initiation, many patients be considered . Combination injectable ther- did not achieve HbA1c \ 7.0%. Basal insulin, apy comprising basal insulin plus preferably a generally prescribed for patients with high GLP-1 RA or rapid-acting insulin (or a change to baseline HbA1c, was associated with a large premixed insulin twice daily) is usually the next reduction in HbA1c and with higher costs. recommended step in the treatment course if Therapy intensiﬁcation at an appropriate time HbA1c remains above target . The latest could lead to clinical and economic beneﬁts algorithm from the American Association of and should be investigated further. Clinical Endocrinologists (AACE)/American Funding: Sanoﬁ U.S., Inc. College of Endocrinology (ACE) recommends initiation with combination therapy rather than monotherapy in patients presenting with Keywords: Basal insulin; Clinical inertia; HbA1c C 7.5% [if the HbA1c goal is B 6.5%] GLP-1 RA; Oral antidiabetes drugs; and initiating treatment with insulin, with or Treatment intensiﬁcation; Type 2 diabetes without other agents, in symptomatic patients presenting with HbA1c [ 9.0% . AACE/ACE recommends antihyperglycemic agents in INTRODUCTION descending order of preference. However, the often-used stepwise approach The American Diabetes Association (ADA) Standards of Medical Care in Diabetes recom- to diabetes treatment may not be ideal, as it can leave patients exposed to periods of uncon- mend individualized treatment to achieve con- trol of glycated hemoglobin A1c (HbA1c) levels trolled HbA1c between treatment steps [4, 5]. In a study by Khunti et al., for example, among soon after diagnosis . The ADA-recom- mended HbA1c goal for most people with dia- type 2 diabetes (T2D) patients with HbA1c C 7.0% who were receiving one oral antidia- betes is \ 7.0%, while that from the American betes drug (OAD), patients had 2.9 years Association of Clinical Endocrinologists and the between initiation of the ﬁrst OAD and the American College of Endocrinology is B 6.5% second and [ 7 years before initiation of basal . Treatment strategies should be designed to attain and maintain goals to prevent or delay insulin . Furthermore, a stepwise approach may not be effective for many patients—a pre- hyperglycemia-related complications . HbA1c targets may be achieved and maintained vious study observed that only 53% of patients in the USA receiving stepwise therapy achieved with a stepwise approach to treatment intensi- ﬁcation, based on regular HbA1c testing to and maintained a target HbA1c level of \ 7.0% . identify the need for additional therapy if treatment goals are not met . Often, the rec- While having a choice of therapies can be beneﬁcial, the variety of treatment options ommended ﬁrst step in pharmaceutical treat- available to clinicians and patients for initiating ment is metformin. Guidelines recommend that if a patient’s HbA1c level is above goal after intensiﬁed regimens can also complicate indi- vidualized patient management. Despite 3 months, a second agent should be added to Diabetes Ther (2018) 9:1169–1184 1171 elevated HbA1c levels, a number of physician-, individuals covered by [ 100 large employers patient-, and healthcare delivery system-related and 12 unique health plans as well as retirees resistance factors may contribute to delays in with employee-sponsored Medicare supple- treatment intensiﬁcation [7–10]. These include mental insurance. Information is available on senses of loss of control or of personal failure medical and pharmacy claims (including dates and injection-related anxiety among some and place of service, diagnosis and procedure patients who have had recommendations for codes, and costs) as well as patient demo- initiating insulin or GLP-1 receptor agonist graphics and periods of continuous health-plan therapy and concerns about hypoglycemia and/ enrollment. Linked laboratory information is or weight gain and perceived low efﬁcacy of also available for approximately 1 million indi- insulin among some patients and providers viduals, including laboratory tests (types, [7, 10]. This failure of therapy intensiﬁcation results, and dates) ordered by ofﬁce-based despite an inadequate achievement of target practices. This article does not contain any glycemia with the current treatment regimen is studies with human participants or animals known as clinical inertia  or clinical myopia performed by any of the authors. Research Tri-  and is an established problem contributing angle Institute (RTI) International’s institu- to prolonged elevated glycemic levels despite tional review-board committee determined that treatment with OADs [4, 5, 12]. Results from this study met all criteria for exemption from published studies continue to demonstrate that full review as the data were retrospective, de- clinical inertia, as a delay or a lack of treatment identiﬁed, and anonymous . The decision intensiﬁcation in people with T2D, is common follows Health and Human Services policy in clinical practice [5, 13, 14]. detailing which studies can be considered Alongside the medical consequences for exempt from meeting the requirements of the patients with diabetes, clinical inertia also has federal human subjects regulations, including an impact on healthcare expenditures; over- the requirement for initial and annual Institu- coming clinical inertia to provide a sustained tional Review Board review . reduction in HbA1c levels has been associated with signiﬁcant cost savings and reduced Patient Selection Criteria healthcare utilization [15–17]. Thus, overcom- ing clinical inertia has the potential to improve Data were selected from patients both clinical and economic outcomes in T2D aged C 18 years with a diagnosis of T2D (ICD-9- patients. The Assessment of Clinical Treatment CM codes 250.x0 or 250.x2) between 1 January Inertia and Outcomes aNalysis (ACTION) was 2007 and 31 December 2014 who were treated conducted to compare clinical and economic with a single OAD and also had linked labora- outcomes in patients with uncontrolled T2D tory information. The date of initiation of (HbA1c [ 7.0%) escalating current OAD ther- 2-OAD therapy, GLP-1 RA, or basal insulin was apy to 2-OAD therapy (either simultaneously or designated as the index date. Patients were sequentially) vs. initiation of GLP-1 RA or basal required to have C 6 months of continuous insulin in a real-world clinical setting. health-plan enrollment prior to the index date (the baseline period) and C 12 months after the index date (the follow-up period). All patients METHODS were required to have C 1 HbA1c test result of [ 7.0% in the baseline period (inclusive of Data Source the index date). Patients were selected for inclusion in the This was a retrospective analysis of data from 2-OAD cohort if they either simultaneously or the MarketScan Commercial Claims and sequentially initiated treatment with two Encounters and Medicare Supplemental and OADs; these OADs included the biguanide Coordination of Beneﬁts database. This data- metformin, sulfonylureas, DPP-4 inhibitors, base includes information on [ 60 million 1172 Diabetes Ther (2018) 9:1169–1184 thiazolidinediones, a-glucosidase inhibitors, Study Measures and Data Analyses amylin analogs, meglitinides, and SGLT-2 inhibitors. Patients could be treatment naive or Eligible patients were categorized into one of experienced with prior OAD monotherapy, three cohorts (2-OAD, GLP-1 RA, or basal insu- andwererequiredtoremain on bothnewly lin) based on the ﬁrst observed qualifying initiated OAD medications for a minimum of treatment or regimen. The following parameters 30 days. Patients were included in the GLP-1 were assessed for each patient cohort and ana- RA cohort if they received the GLP-1 RA exe- lyzed on an intent-to-treat basis: patient natide (short-acting oronce-weekly formula- demographics, clinical characteristics, longitu- tion) or liraglutide (other GLP-1 RAs were not dinal HbA1c, HbA1c change from baseline, and available during the study period), regardless of total all-cause healthcare resource utilization receipt of prior or simultaneous OAD medica- and costs in the follow-up period. Clinical tions. Patients were included in the basal-in- characteristics included the Charlson Comor- sulin cohort if they received insulin glargine, bidity Index (CCI) score measured during the insulin detemir, or NPH insulin, regardless of baseline period, OADs received during the receipt of prior or simultaneous OAD medica- baseline period, and duration of OAD, GLP-1 tions; patients receiving premix insulin were RA, or basal-insulin treatment during the fol- excluded. Patients were required to have low-up period. Longitudinal HbA1c was recor- received their 2-OAD regimen, GLP-1, or basal ded at baseline (the closest value to the index insulin for a minimum of 30 days. date, inclusive of the index date, was reported) HbA1c changes were evaluated during each for the four quarters of the follow-up period (if quarter of the follow-up period and for the last more than one measurement was available in available measure of HbA1c among patients each quarter, the measurement closest to the with follow-up HbA1c measurements available end of the quarter was used) and the last avail- (note: the last available measure of HbA1c was able follow-up (deﬁned as the last available reported for up to 15 months rather than up to HbA1c test value observed during the 15-month 12 months post-index date). HbA1c was repor- post-index date period, using the last observa- ted among patients with a measurement in the tion carried forward). HbA1c change from quarter, and it was observed that 38, 38, 34, and baseline included the change from baseline to 29% of patients had a measurement during the last available follow-up. The percentage of quarters 1, 2, 3, and 4, respectively, with 73% of patients in the following categories of HbA1c patients having at least one follow-up mea- was also reported for the baseline and follow-up surement. The database timeline for patient measurements: \ 7.0%, 7.0 to \ 8.0%, 8.0 selection is shown in Fig. 1. Fig. 1 Database timeline for patient selection. MarketScan database not designed to capture clinical outcomes Diabetes Ther (2018) 9:1169–1184 1173 to \ 9.0%, 9.0 to \ 10.0%, and C 10.0%. Total indicating higher discontinuation rates among all-cause cost data were obtained from medical patients initiating injectable therapy. and pharmacy claims over the 1-year period During follow-up, patients may have inten- following the index date, comprising inpatient, siﬁed treatment with additional therapy or emergency-outpatient, physician-ofﬁce, other switched therapy, as this study used an intent- ancillary (including laboratory claims and to-treat approach; patients were not censored at claims in other healthcare settings, such as the time of medication discontinuation, outpatient hospital), and pharmacy costs. All switching, or augmentation. Biguanides and analyses were descriptive. Cost values represent sulfonylureas were used more commonly by the total amounts paid to the provider by both patients who had initiated injectable therapy— health plans and patients (i.e., cost values rep- especially those initiating GLP-1 RA therapy— resent the sum of health plan payments, patient than by those in the 2-OAD cohort. A minority out-of-pocket expenses, and payments by sup- of patients in the basal-insulin cohort and the plemental insurance plans). Cost values were GLP-1 RA cohort added the other updated to 2014 US dollars using the medical injectable therapy to their index therapy. Rapid- care component of the consumer price index. acting insulin use during the follow-up period was more common and used for a longer period of time among patients in the basal-insulin RESULTS cohort compared with patients in the other two cohorts (Table 2). A total of 591,550 T2D patients aged C 18 years Longitudinal HbA1c data and change in were identiﬁed as receiving two OADs, GLP-1 HbA1c from baseline are shown in Fig. 3. HbA1c RA, or basal insulin in the database. Of these at baseline was high, indicating the presence of patients, 6054 met all inclusion criteria: 4442 clinical inertia. Following treatment initiation (73.4%) in the 2-OAD cohort, 361 (6.0%) in the with 2-OAD therapy, GLP-1 RA, or basal insulin, GLP-1 RA cohort, and 1251 (20.6%) in the basal- HbA1c levels generally decreased and the insulin cohort (Fig. 2). decrease was maintained for the duration of the Patients’ demographic and clinical charac- follow-up period (Fig. 3a). The greatest overall teristics at baseline are shown in Table 1. Mean HbA1c reductions from baseline were seen in HbA1c at baseline was lowest in the GLP-1 RA quarter 2 for the basal-insulin (- 2.1%) and cohort (8.8%) and highest in the basal-insulin 2-OAD (- 1.8%) cohorts (Fig. 3b). cohort (10.1%). The basal-insulin cohort also The percentage of patients achieving had the highest percentage of patients with HbA1c \ 7.0% was lower in the basal-insulin HbA1c [ 10.0% (45.7%). Biguanides were the cohort than in the other two cohorts (Table 3). most common OADs received at baseline; Despite a greater reduction in HbA1c compared 50.0% of the 2-OAD cohort, 47.4% of the GLP-1 with the other two cohorts, patients in the RA cohort, and 25.3% of the basal-insulin basal-insulin cohort, who had a higher mean cohort were receiving this OAD class. Sulfony- HbA1c at baseline, were less likely to reach the lureas were the second most common OAD, target HbA1c. with 23.4% of the 2-OAD cohort, 15.8% of the Patients in the basal-insulin cohort had more GLP-1 RA cohort, and 15.7% of the basal-insulin inpatient days and emergency-department (ED) cohort receiving sulfonylureas. visits than patients in the other two cohorts. Duration of treatment, measured as the Patients in the 2-OAD cohort had fewer claims number of days between the index date and last for physician-ofﬁce visits than patients in the available evidence of a prescription for the other two cohorts (Supplementary Table 1). index OAD, basal insulin, or GLP-1 RA, was Patients in the 2-OAD cohort had the lowest longest for the 2-OAD cohort (338.9 days) fol- mean (SD) all-cause total healthcare costs dur- lowed by 285.0 days for the basal-insulin cohort ing the 12-month follow-up period, followed by and 281.1 days for the GLP-1 RA cohort, those in the GLP-1 RA cohort, while total healthcare costs were highest in the basal- 1174 Diabetes Ther (2018) 9:1169–1184 Diabetes Ther (2018) 9:1169–1184 1175 bFig. 2 Flow chart showing selection of patients for this levels that are so high that it is difﬁcult to analysis. Percentages do not add up to 100% because of effectively achieve the guideline-recommended rounding. A total of 3897 (74.8%) patients had HbA1c HbA1c target, despite robust glycemic lowering values recorded in the 12-month post-index date period. with a variety of agents—as seen in this study. A total of 274 (72.9%) patients had HbA1c values The greatest overall HbA1c reductions in this recorded in the 12-month post-index date period. A total study were seen in quarter 2 for both the basal of 878 (68.8%) patients had HbA1c values recorded in the insulin and 2-OAD cohorts. These ﬁndings 12-month post-index date period suggest that if patients do not achieve HbA1c \ 7% with 2-OAD, basal insulin, or GLP- 1 RA therapy within the ﬁrst 6 months follow- insulin cohort. Inpatient and ED visit costs were ing initiation, they have a low probability of achieving control thereafter. Altering treatment the main drivers of total cost in the basal-in- sulin cohort, while pharmacy costs (i.e., any regimens by adding or changing antihyper- glycemic agents should be considered when costs directly related to drugs, excluding moni- toring costs) were the main drivers of total cost HbA1c goals are not achieved within for patients in the other two cohorts (Supple- 3–6 months, as recommended by the ADA . In this analysis, most patients in the basal-in- mentary Table 2). sulin cohort did not achieve target HbA1c levels, despite having a greater reduction in DISCUSSION HbA1c from baseline compared with the other two cohorts. Potential reasons may be that, on This retrospective analysis of data from a large average, they had higher HbA1c at baseline and US claims database was conducted to compare a greater percentage of patients in the basal-in- outcomes in patients with uncontrolled T2D sulin cohort had an HbA1c [ 10.0% at follow- intensifying anti-hyperglycemic therapy by up compared with the other cohorts. Patients initiating treatment with two OADS, GLP-1 RA, initiating basal insulin also had a higher CCI or basal insulin. The study results show that score, suggesting a cohort of more difﬁcult-to- many patients are not at optimal glycemic treat patients. control: HbA1c at baseline was high in all In this study we speculate that the high cohorts but highest in patients initiating basal HbA1c levels observed at baseline prior to insulin. Average HbA1c levels decreased in all treatment intensiﬁcation could be the result of cohorts following treatment intensiﬁcation; the clinical inertia. It is plausible that either clinical greatest reduction was seen in the basal-insulin inertia or the physician’s perception regarding cohort, which also had the highest HbA1c at when treatment goals have been achieved is baseline. Despite robust HbA1c lowering fol- partly responsible for the observation that a lowing treatment intensiﬁcation, many patients large proportion of patients in the current study still failed to achieve target HbA1c levels. More had poorly controlled diabetes in the baseline timely treatment intensiﬁcation, including ear- period. This observation is in line with other lier use of combination therapy, could help recently published evidence; based on an anal- improve the achievement of glycemic control ysis of the prevalence of clinical inertia among and impact of timely treatment intensiﬁcation patients with T2D with personalized HbA1c on attaining and maintaining glycemic goals goals in the US managed-care setting, Lin et al. and should be investigated further. concluded that clinical inertia—deﬁned as fail- While the recommended stepwise approach ure to have therapy initiated or intensiﬁed for diabetes management [1, 3] is appropriate despite an inadequate treatment response—is for many patients, it may not always be effec- high and has increased in recent years . tive. For a variety of reasons, the next step in Results from the current study are in line with anti-hyperglycemic treatment may often be those from another real-world retrospective initiated very late: by the time intensiﬁcation analysis of T2D patients from the MarketScan occurs, HbA1c values may have already reached database, which reported low rates of treatment 1176 Diabetes Ther (2018) 9:1169–1184 Table 1 Baseline demographic and clinical characteristics of the patients included in this analysis (N = 6054) Characteristic Cohort 2-OAD GLP-1 RA Basal insulin Patients, n (%) 4442 (73.4) 361 (6.0) 1251 (20.6) Demographic characteristics Mean (SD) age at index, years 56.1 (9.8) 52.5 (10.9) 53.3 (13.1) Female, n (%) 1782 (40.1) 210 (58.3) 590 (47.2) Region, n (%) Northeast 427 (9.6) 32 (8.9) 114(9.1) North central 591 (13.3) 59 (16.3) 202 (16.1) South 1966 (44.3) 228 (63.2) 637 (50.9) West 1454 (32.7) 41 (11.4) 296 (23.7) Missing 4 (0.1) 1 (0.3) 2 (0.2) Commercial payer, n (%) 3852 (88.1) 332 (92.7) 1053 (87.5) Plan type, n (%) Preferred provider organization 2214 (49.8) 248 (68.7) 747 (59.7) Health maintenance organization 1562 (35.2) 63 (17.5) 343 (27.4) Other 636 (14.3) 50 (13.8) 151 (12.1) Missing 30 (0.7) 0 (0.0) 10 (0.8) Clinical characteristics HbA1c, % Mean (SD) 9.1 (1.8) 8.8 (1.6) 10.1 (2.38) Median (min, max) 8.5 (7.1, 17.9) 8.2 (7.1, 16.1) 9.7 (7.1, 18.9) HbA1c categories, % patients [ 7.0–8.0% 37.1 44.3 21.7 [ 8.0–9.0% 22.4 23.3 19.1 [ 9.0–10.0% 14.1 12.5 13.4 [ 10.0% 26.4 19.9 45.7 CCI score Mean (SD) 1.04 (1.30) 1.04 (1.21) 1.47 (1.91) Median (min, max) 1.00 (0.00, 14.00) 1.00 (0.00, 8.00) 1.00 (0.00, 15.00) OADs received during the baseline period, n (%) Biguanides 2220 (50.0) 171 (47.4) 316 (25.3) SUs 1037 (23.4) 57 (15.8) 196 (15.7) DPP-4 inhibitors 193 (4.3) 19 (5.3) 34 (2.7) Diabetes Ther (2018) 9:1169–1184 1177 Table 1 continued Characteristic Cohort 2-OAD GLP-1 RA Basal insulin TZDs 164 (3.7) 18 (5.0) 32 (2.6) Other OAD medication classes 8 (0.2) 3 (0.8) 14 (1.1) Patients with any OAD treatment during the baseline 3320 (74.7) 239 (66.2) 537 (42.9) period, n (%) Duration of any OAD treatment during the baseline period, days Mean (SD) 112.1 (58.4) 102.8 (58.1) 87.4 (59.4) Median (min, max) 127 (1.0, 180.0) 106 (1.0, 180.0) 90.0 (1.0, 180.0) OADs received on the index date, n (%) a-Glucosidase inhibitor 9 (0.20) 0 (0) 0 (0) Amylin analog 3 (0.07) 1 (0.3) 2 (0.16) Biguanide 3179 (71.6) 163 (45.2) 327 (26.1) DPP-4 inhibitor 686 (15.4) 11 (3.0) 39 (3.1) Meglitinide 30 (0.68) 0 (0) 6 (0.48) SGLT-2 inhibitor 1 (0.02) 0 (0) 0 (0) SU 2823 (63.6) 56 (15.5) 172 (13.7) TZD 571 (12.9) 21 (5.8) 36 (2.9) Biguanide/DPP-4 inhibitor ﬁxed-dose combination 445 (10.0) 5 (1.4) 17 (1.4) Biguanide/meglitinide ﬁxed-dose combination 1 (0.02) 0 (0) 0 (0) Biguanide/SGLT-2 inhibitor ﬁxed-dose combination 0 (0) 0 (0) 0 (0) Biguanide/SU ﬁxed-dose combination 126 (2.8) 1 (0.3) 1 (0.08) Biguanide/TZD ﬁxed-dose combination 199 (4.5) 2 (0.55) 6 (0.48) DPP-4 inhibitor/TZD ﬁxed-dose combination 0 (0) 0 (0) 0 (0) SU/TZD ﬁxed-dose combination 20 (0.45) 1 (0.3) 0 (0) SD standard deviation, SU sulfonylurea, TZD thiazolidinedione Information on the duration of diabetes was unavailable because of a lack of information on enrollees’ medical history in the database Includes alphaglucosidase inhibitors, amylin analogs, and meglitinides. No patients received SGLT-2 inhibitors during the baseline period Days between index date and last available OAD, basal insulin, and GLP-1 RA intensiﬁcation among patients with elevated patients with HbA1c [ 7.0% at 4 years after HbA1c levels at baseline or during a 4-year fol- initiation with a new class of anti-hyper- low-up period (8.0–9.2%). The percentage of glycemic medication ranged from 48% among 1178 Diabetes Ther (2018) 9:1169–1184 Table 2 Medications used during follow-up among patients initiating 2-OAD therapy, GLP-1 RA, or basal insulin (among C 2.0% of patients) Medication class Cohort 2-OAD (n = 4442) GLP-1 RA (n = 361) Basal insulin (n = 1251) Biguanides Patients, n (%) 175 (3.9) 69 (19.1) 208 (16.6) Duration, mean (SD), days 141.12 (91.24) 170.35 (99.30) 163.38 (99.88) Sulfonylureas Patients, n (%) 230 (5.2) 59 (16.3) 122 (9.8) Duration, mean (SD), days 156.83 (99.11) 170.68 (100.99) 170.56 (97.64) DPP-4 inhibitors Patients, n (%) 187 (4.2) 20 (5.5) 30 (2.4) Duration, mean (SD), days 131.70 (91.04) 139.15 (85.34) 137.20 (105.57) Thiazolidinediones Patients, n (%) 155 (3.5) 19 (5.3) 46 (3.7) Duration, mean (SD), days 133.30 (97.09) 205.05 (109.80) 152.30 (99.18) GLP-1 RA Patients, n (%) 133 (3.0) – 78 (6.2) Duration, mean (SD), days 128.89 (90.13) – 116.09 (105.74) Basal insulin Patients, n (%) 296 (6.7) 49 (13.6) – Duration, mean (SD), days 120.68 (87.33) 118.41 (80.43) – Rapid-acting insulin Patients, n (%) 88 (2.0) 20 (5.5) 264 (21.1) Duration, mean (SD), days 2.13 (18.93) 8.02 (40.43) 27.78 (60.50) patients initiating OADs to 74% of patients considered eligible for treatment intensiﬁcation initiating basal insulin . A retrospective . Likewise, less than half (49%) of older study involving databases from the US and adult T2D patients with a baseline HbA1c of multiple European countries found that after 2 C 8.0% despite treatment with two OADs years of basal-insulin therapy, only 27.8% of received intensiﬁed treatment during follow-up, patients achieved the HbA1c goal of \ 7.0%, with median time to intensiﬁcation of down from a mean baseline HbA1c level of 9.0% 18.5 months [14, 23], although the reasons for . Fu and Sheehan reported that 52% of adult clinical inertia can be multifactorial . T2D patients who were uncontrolled on OADs The results from the current study argue for with a baseline HbA1c of C 8.0% had not more intensive treatment and the implementa- received treatment intensiﬁcation after 1 year of tion of strategies to overcome clinical inertia. A follow-up even though these patients were detailed discussion of the causes of and Diabetes Ther (2018) 9:1169–1184 1179 Fig. 3 Mean HbA1c by cohort and time period among deﬁned as the last available HbA1c test value observed patients with uncontrolled T2D initiating 2-OAD ther- during the 15-month post-index period apy, GLP-1 RA, or basal insulin. a Longitudinal HbA1c, b change from baseline. Last available follow-up measure potential solutions to clinical inertia is outside treatment regimen intensiﬁcation in patients the scope of this article. Even so, new thera- with very high HbA1c at diagnosis (e.g., peutic options (with fewer side effects including patients who may have been undiagnosed for fewer hypoglycemic events and/or less weight an extended period of time) may reduce the gain), less complex treatment regimens (for burden of the treatment and provide ways for- example, fewer injections), and earlier ward to overcome clinical inertia in T2D. 1180 Diabetes Ther (2018) 9:1169–1184 Table 3 HbA1c category at last available follow-up mea- in comparable US studies describing retrospec- surement (up to 15 months) among patients with tive observational database analyses [16, 25–29]. uncontrolled T2D initiating 2-OAD therapy, GLP-1 RA, Several limitations, however, should be or basal insulin taken into account when considering the ﬁnd- ings of this study, some of which are due to the HbA1c Cohort limitations and biases inherent to retrospective categories (% 2-OAD GLP-1 Basal analyses of claims data . Further, the Mar- patients) (n = 4442) RA insulin ketScan database is not designed to capture (n = 361) (n = 1251) clinical information, and this is reﬂected in the low number of patients with HbA1c values \ 7.0% 47.5 41.1 32.6 available at all time points in the follow-up [ 7.0–8.0% 27.2 29.6 21.4 period; HbA1c data for quarters 1, 2, 3, and 4 [ 8.0–9.0% 13.1 14.1 20.5 were available for an average of 38%, 38%, 34%, and 29% of patients, respectively, although this [ 9.0–10.0% 6.2 9.6 9.7 could have been due to a failure of clinicians to [ 10.0% 6.0 5.6 15.8 obtain HbA1C levels at guideline-recommended intervals. Findings of this analysis may also be confounded by the fact that some patients may have received further treatment intensiﬁcation with additional therapy, may have switched In our analysis, patients in the basal-insulin therapy during follow-up, or may have cohort had higher total all-cause healthcare increased medication doses. As it was not pos- costs and more inpatient and ED visits com- sible to randomize patients to treatment groups, pared with patients in the other two cohorts. the data may be subject to selection bias in These patients had higher HbA1c levels and CCI terms of representativeness of the study popu- scores at baseline, and it is possible that they lation and their treatment regimens to the had a longer duration of diabetes and more overall population of patients with T2D. medical problems overall. Clinicians may have Patients receiving treatment with GLP-1 RAs, reserved high-efﬁcacy therapy (i.e., basal insu- for example, may not be the same as those lin) for these patients. However, despite basal- being treated with basal insulin; those treated insulin therapy, HbA1c remained uncontrolled with basal insulin will likely have more and patients incurred higher costs. It is possible advanced disease, more comorbidities, and that overcoming clinical inertia and treating higher HbA1c, while patients treated with a patients earlier would help reduce HbA1c more GLP-1 RA may have received these agents rather effectively and could therefore have the poten- than basal insulin because of concerns about tial to contain healthcare costs. The ﬁnding of a body weight (exempliﬁed by the higher pro- higher baseline HbA1c and poorer health status portion of female patients in the GLP-1 RA of patients initiating basal insulin compared cohort) or a need to minimize weight gain with those initiating a GLP-1 RA was also during treatment, or to reduce the risk of reported in a real-world clinical and economic hypoglycemia, or as a result of health-plan outcomes study in T2D patients in the USA . coverage and co-pay policies. Likewise, the Further studies are needed to better understand magnitude of the change in HbA1c from base- the association among earlier treatment inten- line is not only inﬂuenced by the HbA1c level at siﬁcation, HbA1C goal attainment, and health- baseline, but also by the selection of included care costs. patients (e.g., duration of diabetes). Further- A strength of this analysis is that the large more, the three cohorts compared in this study administrative claims database included infor- had different sample sizes: the sample size of the mation for US patients across all care settings GLP-1 RA cohort was one-third that of the BI (e.g., inpatient, ED, physician’s ofﬁce) and for cohort and 12 times smaller than the 2-OAD prescriptions and laboratory tests. Patient base- cohort, which hampers between-group line characteristics were similar to those found Diabetes Ther (2018) 9:1169–1184 1181 comparison and increases the risk of selection (59.6% had a co-prescription of basal insulin bias. An unexpected ﬁnding in this study was only and metformin) . The low use of met- that the HbA1c reduction from baseline formin in the current study may be the result of observed in this study was numerically smaller the fact that baseline treatment utilization for for patients in the GLP-1 RA cohort compared all patients was measured during a 6-month pre- with the 2-OAD cohort, while in clinical trials, index period (potentially missing earlier pre- GLP-1 RAs have shown an HbA1c-lowering scriptions with long lengths of script). There- potential comparable to that of basal insulin. fore, the low metformin use could be due to a The gap between clinical efﬁcacy observed in failure to observe baseline medication history randomized clinical trials and effectiveness for a longer period or the result of patients not observed in real-world studies in terms of being able to tolerate metformin in the long HbA1c reduction with newer antihyperglycemic term. therapies has been described recently, even when baseline characteristics are comparable. CONCLUSIONS Poor medication adherence in the real world is thought to be a key reason for real-world The high percentage of T2D patients not at patients having poorer outcomes than the optimal glycemic control in this retrospective counterparts participating in randomized clini- analysis of data may be the result of clinical cal trials . In addition to the potential effect inertia in the period prior to treatment inten- of cohort size, this observation could be siﬁcation, although the study did not look at explained by the relatively better HbA1c control clinical inertia itself. Despite robust lowering in at baseline in the GLP-1 RA cohort. Also, the HbA1c following treatment intensiﬁcation, cohort might have been exposed to slightly less many patients still did not achieve guideline- intensive treatment approaches compared with recommended HbA1c levels, suggesting a need the other cohorts. HbA1c levels at last available for earlier and/or more intensive treatment, follow-up were similar for the 2-OAD and GLP-1 including combination pharmacotherapy. Such RA cohorts. Further, patients were evaluated strategies may be helpful in getting more based on an intent-to-treat approach and were patients to goal earlier and provide better not censored at the point of medication dis- durability of glycemic control, thus reducing continuation, switching, or augmentation. It their exposure to hyperglycemia, which may be was not within the scope of this analysis to associated with both clinical and economic measure the effect of a treatment change on beneﬁts. HbA1c. Next steps for further research could include an evaluation of HbA1c following treatment changes such as further intensiﬁca- tion or treatment switching or an analysis of ACKNOWLEDGEMENTS matched patient cohorts. Additionally, this study did not examine clinical inertia itself or the reasons why glycemic targets were not met. Funding. Sponsorship for this study and Finally, the current study reported a fairly low article processing charges were funded by Sanoﬁ use of metformin (25.3%) in combination with US, Inc. All authors had full access to all of the basal insulin. Treatment guidelines recommend data in this study and take complete responsi- continuing metformin (unless there are con- bility for the integrity of the data and accuracy of the data analysis. traindications or intolerance to its use) as background therapy and in combination with Medical Writing and Editorial Assis- insulin and other agents in patients who do not tance. Writing and editorial assistance in the reach their glycemic target on monotherapy preparation of this article was provided by Tessa [1, 3]. Only 50.1% of patients from the French Hartog, PhD, of Excerpta Medica. Support for national health insurance database had co-pre- this assistance was funded by Sanoﬁ US, Inc. scription claims for insulin and metformin 1182 Diabetes Ther (2018) 9:1169–1184 Authorship. All named authors meet the committee determined that this study met all International Committee of Medical Journal criteria for exemption from full review as the Editors (ICMJE) criteria for authorship for this data were retrospective, de-identiﬁed, and article, take responsibility for the integrity of anonymous . The decision follows Health the work as a whole, and have given their and Human Services policy detailing which approval for this version to be published. studies can be considered exempt from meeting the requirements of the federal human subjects Disclosures. Lawrence Blonde has received regulations, including the requirement for ini- grant/research support from AstraZeneca, Jans- tial and annual Institutional Review Board sen Pharmaceuticals, Inc., Lexicon Pharmaceu- review . ticals, Inc., Merck & Co., Novo Nordisk, and Sanoﬁ, speaker honoraria from AstraZeneca, Data Availability. The data sets generated Janssen Pharmaceuticals, Inc., Novo Nordisk, and/or analyzed during the current study are Sanoﬁ, and consultant honoraria from Intarcia not publicly available because of the proprietary Therapeutics, Inc., Janssen Pharmaceuticals, nature of the database. Questions regarding the Inc., Merck & Co., Inc., Novo Nordisk, and study database should be directed to the corre- Sanoﬁ. Denis Raccah has acted as a member of sponding author. advisory boards and as a speaker during sym- Open Access. This article is distributed posia for AstraZeneca, Janssen, Lilly, Novartis, under the terms of the Creative Commons NovoNordisk, and Sanoﬁ. Elisheva Lew is an Attribution-NonCommercial 4.0 International employee of Sanoﬁ. Juliana Meyers is an License (http://creativecommons.org/licenses/ employee of RTI Health Solutions, which by-nc/4.0/), which permits any non- received funding from Sanoﬁ to conduct this commercial use, distribution, and reproduction analysis. Elena Nikonova was an employee of in any medium, provided you give appropriate Artech Information Systems LLC, under con- credit to the original author(s) and the source, tract with Sanoﬁ, at the time the study was provide a link to the Creative Commons license, performed. Mayank Ajmera is an employee of and indicate if changes were made. RTI Health Solutions, which received funding from Sanoﬁ to conduct this analysis. Keith L. Davis is an employee of RTI Health Solutions, which received funding from Sanoﬁ to conduct REFERENCES this analysis. Monica Bertolini is an employee of Sanoﬁ. Bruno Guerci has acted as an advisory 1. American Diabetes Association. Standards of medi- panel/board member for Sanoﬁ, Eli Lilly, cal care in diabetes—2017. Diabetes Care. NovoNordisk, Novartis, GSK, MSD, Boehringer 2017;40(Suppl. 1):S1–135. Ingelheim, AstraZeneca, Abbott, Medtronic, Roche Diagnostics, and as Clinical investigator 2. Garber AJ, et al. Consensus statement by the American Association of Clinical Endocrinologists for Sanoﬁ, Eli Lilly, NovoNordisk, GSK, BMS, and American College of Endocrinology on the AstraZeneca, Medtronic, Abbott, Roche Diag- comprehensive type 2 diabetes management algo- nostics, MSD, Novartis, Janssen, Boehringer rithm—2018 executive summary. Endocr Pract. Ingelheim; he has received research support 2018;24(1):91–120. from Medtronic, Vitalaire, Sanoﬁ, Eli Lilly, 3. Garber AJ, Abrahamson MJ, Barzilay JI, et al. Con- Novo Nordisk; the GetGoal-Duo2 trial sensus statement by the American Association of (NCT01768559) was sponsored by Sanoﬁ. Clinical Endocrinologists and American College of Endocrinology on the comprehensive type 2 dia- Compliance with Ethics Guidelines. This betes management algorithm—2017 executive summary. Endocr Pract. 2017;23:207–38. article does not contain any studies with human participants or animals performed by 4. Brown JB, Nichols GA. Slow response to loss of any of the authors. Research Triangle Institute glycemic control in type 2 diabetes mellitus. Am J Manag Care. 2003;9:213–7. (RTI) International’s institutional review-board Diabetes Ther (2018) 9:1169–1184 1183 5. Khunti K, Wolden ML, Thorsted BL, Andersen M, 18. Research Triangle Institute. Ethics and Human Davies MJ. Clinical inertia in people with type 2 Research Protection Web site. www.rti.org/ethics- diabetes: a retrospective cohort study of more than and-human-research-protection. Accessed 19 Oct 80,000 people. Diabetes Care. 2013;36:3411–7. 2017. 6. Stark Casagrande S, Fradkin JE, Saydah SH, Rust KF, 19. Health and Human Services. Code of Federal Reg- Cowie CC. The prevalence of meeting A1C, blood ulations. Title 45—Public Welfare Department of pressure, and LDL goals among people with dia- Health and Human Services. Part 46—Protection of betes, 1988–2010. Diabetes Care. 2013;36:2271–9. Human Subjects. www.hhs.gov/ohrp/regulations- and-policy/regulations/45-cfr-46/index.html.Acces- 7. Peyrot M, Rubin RR, Khunti K. Addressing barriers sed 19 Oct 2017. to initiation of insulin in patients with type 2 dia- betes. Prim Care Diabetes. 2010;4(Suppl. 1):S11–8. 20. Bonafede M, Chandran A, DiMario S, Saltiel-Berzin R, Saliu D. Medication usage, treatment intensiﬁ- 8. Lin J, Zhou S, Wei W, Pan C, Lingohr-Smith M, cation, and medical cost in patients with type 2 Levin P. Does clinical inertia vary by personalized diabetes: a retrospective database study. BMJ Open Diabetes Res Care. 2016;4:e000189. A1C goal? A study of predictors and prevalence of clinical inertia in a U.S. managed-care setting. Endocr Pract. 2016;22:151–61. 21. Mauricio D, Meneghini L, Seufert J, et al. Glycaemic control and hypoglycaemia burden in patients with 9. Reach G. Patient non-adherence and healthcare- type 2 diabetes initiating basal insulin in Europe provider inertia are clinical myopia. Diabetes Metab and the USA. Diabetes Obes Metab. J. 2008;34(4 pt 1):382–5. 2017;19:1155–64. 10. Reach G, Pechtner V, Gentilella R, Corcos A, Cer- 22. Fu AZ, Sheehan JJ. Treatment intensiﬁcation for iello A. Clinical inertia and its impact on treatment patients with type 2 diabetes and poor glycaemic intensiﬁcation in people with type 2 diabetes mel- control. Diabetes Obes Metab. 2016;18:892–8. litus. Diabetes Metab. 2017 (Epub ahead of print). 23. Ajmera M, Raval A, Zhou S, et al. A real-world 11. Phillips L, Branch W, Cook C, et al. Clinical inertia. observational study of time to treatment intensiﬁ- Ann Intern Med. 2001;135(9):825–34. cation among elderly patients with inadequately controlled Type 2 Diabetes Mellitus. J Manag Care 12. Grant RW, Cagliero E, Dubey AK, et al. Clinical Spec Pharm. 2015;21:1184–93. inertia in the management of type 2 diabetes metabolic risk factors. Diabet Med. 2004;21:150–5. 24. Thayer S, Wei W, Buysman E, et al. The INITIATOR study: pilot data on real-world clinical and eco- 13. Nichols GA, Koo YH, Shah SN. Delay of insulin nomic outcomes in US patients with type 2 diabetes addition to oral combination therapy despite initiating injectable therapy. Adv Ther. inadequate glycemic control: delay of insulin ther- 2013;30:1128–40. apy. J Gen Intern Med. 2007;22:453–8. 25. Levin P, Lin J, Wei W, et al. Initiating 14. Roussel R, Gourdy P, Gautier JF, et al. Low rate of injectable therapy in US managed care patients with intensiﬁcation in type 2 diabetic patients inade- diabetes. Am J Pharm Beneﬁts. 2014;6:e93–102. quately controlled with basal insulin: the INTER- DIA study. Diabetes Manag. 2016;6:119–27. 26. Davis SN, Wei W, Garg S. Clinical impact of initi- ating insulin glargine therapy with disposable pen 15. Wagner EH, Sandhu N, Newton KM, McCulloch versus vial in patients with type 2 diabetes mellitus DK, Ramsey SD, Grothaus LC. Effect of improved in a managed care setting. Endocr Pract. glycemic control on health care costs and utiliza- 2011;17:845–52. tion. JAMA. 2001;285:182–9. 27. Dalal MR, Xie L, Baser O, DiGenio A. Adding rapid- 16. Levin PA, Zhou S, Gill J, Wei W. Health outcomes acting insulin or GLP-1 receptor agonist to basal associated with initiation of basal insulin after 1, 2, insulin: outcomes in a community setting. Endocr or C 3 oral antidiabetes drug(s) among managed Pract. 2015;21:68–76. care patients with type 2 diabetes. J Manag Care Spec Pharm. 2015;21:1172–81. 28. Bieszk N, Reynolds SL, Wei W, Davis C, Kamble P, Uribe C. ‘‘Act on threes’’ paradigm for treatment 17. Smolen HJ, Murphy DR, Gahn JC, Yu X, Curtis BH. intensiﬁcation of type 2 diabetes in managed care: results of a randomized controlled study with an The evaluation of clinical and cost outcomes asso- educational intervention targeting improved gly- ciated with earlier initiation of insulin in patients cemic control. J Manag Care Spec Pharm. with type 2 diabetes mellitus. J Manag Care Spec 2016;22:1028–38. Pharm. 2014;20:968–84. 1184 Diabetes Ther (2018) 9:1169–1184 29. Ayyagari R, Wei W, Cheng D, Pan C, Signorovitch J, efﬁcacy in randomized controlled trials and effec- Wu EQ. Effect of adherence and insulin delivery tiveness in real-world use of GLP-1 RA and DPP-4 system on clinical and economic outcomes among therapies in patients with type 2 diabetes. Diabetes patients with type 2 diabetes initiating insulin Care. 2017;40(11):1469–78. treatment. Value Health. 2015;18:198–205. 32. Roussel R, Charbonnel B, Behar M, Gourmelen J, 30. Riley GF. Administrative and claims records as Emery C, Detournay B. Persistence with Insulin sources of health care cost data. Med Care. therapy in patients with type 2 diabetes in France: 2009;47(7 Suppl. 1):S51–5. an insurance claims study. Diabetes Ther. 2016;7:537–49. 31. Carls GS, Tuttle E, Tan RD, Huynh J, Yee J, Edelman SV, Polonsky WH. Understanding the gap between
Diabetes Therapy – Springer Journals
Published: Apr 19, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera