Quality measure and weight loss assessment in patients with type 2 diabetes mellitus treated with canagliflozin or dipeptidyl peptidase-4 inhibitors
BMC Endocrine Disorders volume 17, Article number: 32 (2017)
Achieving control of glycated hemoglobin (HbA1c), blood pressure (BP), and body weight (BW) remains a challenge for most patients with type 2 diabetes mellitus (T2DM). In clinical trials, canagliflozin (CANA), an inhibitor of sodium-glucose co-transporter 2, has shown significant improvement compared to some dipeptidyl peptidase-4 (DPP-4) inhibitors in the achievement of such quality measures. This study used recent electronic medical records (EMR) data to assess quality measure achievement of HbA1C, BP, and BW loss in patients treated with CANA versus DPP-4 inhibitors.
Adult patients with ≥1 T2DM diagnosis and ≥12 months of clinical activity (baseline) before first CANA or DPP-4 prescription (index) were identified in the QuintilesIMS Health Real-World Data EMRs–US database (03/29/2012–10/30/2015). Patients were observed from the index to last encounter. Inverse probability of treatment weighting (IPTW) was used to adjust for observed baseline confounders between groups. Kaplan-Meier (KM) rates and Cox proportional hazard models were used to compare achievement of HbA1c < 7% (among patients <65 years old), HbA1c < 8%, systolic BP < 140 mmHg, diastolic BP < 90 mmHg, and BW loss ≥ 5% among patients not meeting these respective targets at baseline.
A total of 10,702 CANA and 17,679 DPP-4 patients were selected. IPTW resulted in balanced baseline demographic, comorbidity, and disease characteristics (CANA: N = 13,793, mean age: 59.0 years; DPP-4: N = 14,588, mean age: 58.9 years). Up until 24 months post-index, CANA patients were more likely to reach an HbA1c < 7% (hazard ratio [HR] = 1.10, P = 0.007, KM rates: 42.8% vs. 40.3%), an HbA1c < 8% (HR = 1.16, P < 0.001, KM rates: 63.7% vs. 60.0%), and a BW loss ≥ 5% (HR = 1.46, P < 0.001, KM rates: 55.2% vs. 46.2%), compared to DPP-4 patients. Up until 12 months post-index, CANA patients were more likely to reach a systolic BP < 140 mmHg (HR = 1.07, P = 0.04, KM rates: 87.8% vs. 83.9%). but not a diastolic BP < 90 mmHg (HR = 0.95, P = 0.361), compared to DPP-4 patients.
This retrospective study of EMR data covering up to 30 months after CANA approval (March 2013) suggests that patients initiated on CANA were more likely to reach HbA1c, systolic BP, and weight loss objectives specified by general diabetes care guidelines than patients initiated on DPP-4 inhibitors.
Diabetes is a chronic condition characterized by high glucose levels in the blood. The American Diabetes Association (ADA) estimates that 9.3% of the US population have diabetes in 2012, among which 90–95% had type 2 diabetes mellitus (T2DM) . Diabetes, including T2DM, can lead to several serious and sometimes life-threatening complications including microvascular conditions such as retinopathy and nephropathy, as well as cardiovascular disease .
Patients with better glycemic (e.g., glycylated hemoglobin [HbA1c]) and blood pressure (BP) control have better outcomes, including lower risks of complications and better survival, which in turn help to maintain quality of life and reduce complication-related treatment costs [3,4,5]. Thus, monitoring for and managing high levels of HbA1c and BP as well as body weight (BW) are key components of the standard of care for patients with T2DM .
Although personalization of goals is recommended by the Healthcare Effectiveness Data and Information Set (HEDIS) and the ADA, both organizations suggest the following general goals for HbA1c management: HbA1c <7% for nonpregnant adults, and less stringent goals such as HbA1c <8% for those older than 65 or with complications (specifically, “history of severe hypoglycemia, limited life expectancy, advanced microvascular or macrovascular complications, extensive comorbid conditions, or long-standing diabetes”) [6, 7]. The ADA also recommends an objective of BP <140/90 mmHg for all patients . As reported by Bailey et al., various healthcare organizations and community quality improvement programs such as the Centers for Medicare & Medicaid Services: Accountable Care Organization, the Minnesota Community Measurement, the Wisconsin Collaborative for Healthcare Quality, and the Health Collaborative of Greater Cincinnati have established HbA1c goals below 8% and systolic BP <140 mmHg in diabetes patients . Overweight and obesity are highly prevalent among patients with T2DM;  furthermore, managing obesity has been shown to have beneficial impacts on T2DM management . Therefore, both the American Association of Clinical Endocrinologists (AACE) and the American College of Endocrinology (ACE) recommend to target a BW loss goal of ≥5% in this population .
Various antihyperglycemic agents have been developed to improve glycemic control as well as other vital signs in T2DM patients, as a complement to diet and lifestyle changes . Sodium-glucose co-transporter 2 (SGLT2) inhibitors such as canagliflozin (Invokana® and Invokamet® [CANA]) are one class of agents that work to lower glucose levels by increasing the amount of glucose excreted by the kidneys. Due to their mechanism of action, SGLT2 inhibitors also contribute to reductions in BW and BP . This insulin-independent mechanism of action makes CANA a candidate for add-on therapy with all other antihyperglycemic agents . In recent Phase 3 clinical trials and real-world studies, CANA has been shown to improve HbA1c, BP, and BW in patients with T2DM [13,14,15,16,17,18,19,20,21].
Another class of antihyperglycemic agents is dipeptidyl peptidase-4 (DPP-4) inhibitors. DPP-4 inhibitors prevent the degradation of incretin hormones by inhibiting the DPP-4 enzyme. Incretin hormones exert their action by stimulation of insulin secretion and inhibition of glucagon release . While DPP-4 inhibitors have been shown to improve HbA1c, only modest improvement in BP and no improvement in BW have been found [23, 24].
In clinical trials comparing CANA 100 mg and CANA 300 mg to sitagliptin 100 mg (a DPP-4 agent) [8, 15, 20], CANA 100 mg was shown to be non-inferior to sitagliptin in achieving diabetes-related quality measure goals for HbA1c, BP, and weight loss at week 52, while CANA 300 mg was shown to be statistically superior. Patients enrolled in these two trials and treated with CANA 100 mg or CANA 300 mg also showed significantly decreased BP and BW as compared to patients treated with sitagliptin.
Few studies have compared glycemic control, BP, and BW loss outcomes between patients initiated on CANA or a DPP-4 agent in a real-world setting . Furthermore, no real-world study has assessed these outcomes over more than one year of follow-up. The aim of the present study was to expand on the clinical benefits mentioned in the clinical trials above and assess achievement of HbA1c, BP, and BW loss goals between a large sample of inadequately controlled patients initiated on CANA versus a DPP-4 agent in a real-world setting where patients can be observed for more than two years following the initiation. In addition, this study assessed the pre- and post-initiation use of antihyperglycemics, antihypertensives, and lipid-lowering agents for patients initiated on CANA versus a DPP-4 agent.
Recent electronic medical records (EMRs) from March 29, 2012 to October 30, 2015 were retrieved from the QuintilesIMS Health Real-World Data Electronic Medical Records – US database (formerly known as Cegedim Strategic Data [CSD] US patient database). This study period was chosen to allow patients initiated on CANA, which was approved by the Federal Drug Administration (FDA) on March 29, 2013 for the treatment of T2DM , to have up to one year of medical records prior to CANA initiation while maximizing the length of the observation period following initiation of CANA or DPP-4 treatment. The QuintilesIMS database, in which 40% of contributors are primary care providers and 60% are specialists, includes clinical information and demographic characteristics, such as gender, age, or ethnicity, on more than 35 million patients from all payer types, including Medicare, Medicaid, commercial, and cash. The data are de-identified in compliance with the Health Insurance Portability and Accountability Act (HIPAA) of 1996 to preserve patient confidentiality.
Study Design and Patient Selection
A retrospective observational study design was used to compare characteristics, attainment of diabetes-related quality measures, and body weight loss between patients treated with CANA or a DPP-4 agent. The index date was defined as the first day with a prescription for CANA or a DPP-4 agent (alogliptin, linagliptin, saxagliptin, or sitagliptin) after March 29, 2013 (the date of CANA approval by the FDA). Patients were classified in the CANA or DPP-4 cohort based on the agent initiated on the index date. Patients were included in the study if they had ≥1 diagnosis for T2DM (ICD-9-CM: 250.x0, 250.x2), ≥12 months of clinical activity prior to the index date (baseline period), ≥1 HbA1c measurement during the baseline period, and if they were 18 years or older at the index date. Moreover, patients initiated on CANA or a DPP-4 agent were excluded if they had a diagnosis for type 1 diabetes mellitus (T1DM, ICD-9-CM: 250.x1, 250.x3) or at least one prescription for a DPP-4 agent during the baseline period. Additionally, patients initiated on CANA and on a DPP-4 agent on the same day were excluded.
The observation period spanned from the index date until the last encounter (i.e., a prescription, a diagnosis, a laboratory test, or a clinical measurement [e.g., vital signs]).
Demographic and clinical characteristics were analyzed during the 12-month baseline period. Clinical characteristics included the Quan-Charlson Comorbidity Index (Quan-CCI), which was designed to predict a patient’s ten-year mortality risk based on a range of comorbid conditions , and the Diabetes Complications Severity Index (DCSI), which was designed to quantify the severity of diabetes complications and help predict the risk of adverse outcomes (for both indexes, a higher score is associated with a higher risk) . Using the Agency for Healthcare Research and Quality (AHRQ) comorbidity definitions, the most common comorbidities were identified .
The duration of the index treatment during the observation period (i.e., the number of days between the index date and last day of final prescription for the index treatment) and the daily dose prescribed at the index date were reported. Utilization of antihyperglycemic, lipid-lowering, and antihypertensive agents was also assessed at baseline and during the observation period.
Several goals for diabetes-related quality measures and BW reduction were evaluated during the entire observation period and included HbA1C <7% among patients with a baseline HbA1C ≥7% and aged <65 years old (this goal was also evaluated for all patients with a HbA1C ≥7%), HbA1C <8% among patients with a baseline HbA1C ≥8% and aged ≥65 years old (this goal was also evaluated for all patients with a HbA1C ≥8%), BW loss ≥5%, systolic BP <140 mmHg among patients with a baseline systolic BP ≥140 mmHg, and diastolic BP <90 mmHg among patients with baseline diastolic BP ≥90 mmHg.
For baseline demographic and clinical characteristics as well as treatment patterns, means, standard deviations, and medians were reported for continuous variables; frequencies and percentages were reported for categorical variables. Standardized differences were used to compare baseline characteristics between the two groups and to assess whether cohorts were well-balanced after weighting according to an accepted threshold of ≤10% [30, 31]. Treatment patterns were compared between weighted cohorts using Chi-squared tests for categorical variables and two-sided Student’s t-tests for continuous variables.
To minimize potential confounding and selection bias while retaining a sufficiently large study population, an inverse probability of treatment weighting (IPTW) approach was used [32,33,34]. First, the propensity score (PS), defined as the probability of being initiated on CANA treatment, was estimated using a multivariate logistic regression model conditional on the baseline covariates (i.e., gender; age categories; ethnicity; region; quarter and year of the index date; use of antihyperglycemic agents at baseline; use of a fixed-dose combination at index date; Quan-Charlson comorbidity index at baseline; number of visits at baseline; closest baseline HbA1c measurement to the index date, closest baseline body mass index [BMI] measurement to the index date, and diagnosis of obesity at baseline). The IPTW approach then used weights derived from PS to create a weighted population that retains all patients and makes both study cohorts similar to the overall study population, such that the distribution of covariates is independent of treatment. Therefore, results obtained with IPTW should be interpreted as the average treatment effect among the overall population. Weights were calculated as the inverse of patients’ estimated probabilities of having their observed initiation treatment (i.e., 1/PS for the CANA group and 1/(1-PS) for the DPP-4 group). Finally, the normalized weights were calculated by dividing each weight by the overall mean weight. After weighting, the sum of weights attributed to each patient in a given cohort may not be equal to the sample size of this cohort; consequently, the effective sample size after weighting may be different than the original sample size (i.e., before weighting). For the analysis of HbA1c among patients aged <65 years, PS and weights were recalculated within this group.
Weighted Kaplan-Meier (KM) survival curves and Cox proportional hazards models (i.e., time-to-event analyses) were used to compare the KM rates and the hazards of reaching diabetes-related goals between the two cohorts. The KM curves of both groups were compared using log-rank tests up to 24 months for all the outcomes and up to 12 months for BP outcomes. The hazards of both cohorts were compared using hazard ratios (HRs), 95% confidence intervals (CIs), and p-values which were estimated using weighted Cox models containing a single indicator for treatment cohort.
Baseline demographic and clinical characteristics
Among 292,248 patients diagnosed with T2DM, 28,381 met all inclusion and exclusion criteria. Of those, 10,702 were initiated on CANA and 17,679 were initiated on a DPP-4 agent. Most patients in the CANA group were prescribed CANA 100 mg (62.8%) and most patients in the DPP-4 group were prescribed sitagliptin (73.4%), followed by linagliptin (14.3%), saxagliptin (11.2%), and alogliptin (1.8%).
Table 1 presents the baseline demographic and clinical characteristics for the study cohorts. Prior to weighting and relative to patients initiated on DPP-4, patients initiated on CANA were younger (mean age: 57.4 vs. 59.8 years, standardized difference 21.4%), and more likely white (76.9% vs. 74.0%, standardized difference 6.6%). In addition, CANA patients initiated treatment later in the study period, which means that they were also observed and treated for a shorter period of time on average, post-index, compared to patients initiated on a DPP-4 agent. Most patients in both cohorts used other antihyperglycemic (94.2% of CANA patients and 88.7% of DPP-4 patients), lipid-lowering (71.6% of CANA patients and 68.0% of DPP-4 patients), and antihypertensive agents (80.7% of CANA patients and 78.3% of DPP-4 patients) prior to the index date.
Quan-CCI and DCSI scores were equivalent for patients initiated on CANA compared to those of patients initiated on a DPP-4 (mean Quan-CCI: 1.4; mean DCSI: 0.6). Among the most common DCSI and AHRQ complications, neuropathy, depression, and obesity were more likely for patients initiated on CANA, while cardiovascular complications, nephropathy, and anemia were slightly more likely for patients initiated on a DPP-4 agent. Patients initiated on CANA had higher HbA1c values at index (8.4 vs. 8.3%, standardized difference 9.5%) as well as a higher baseline BMI and weight (mean BMI: 35.7 vs. 34.1 kg/m , standardized difference 26.5%; mean weight: 233.9 vs. 219.3 lbs, standardized difference 27.1%), compared to patients initiated on a DPP-4 agent. Baseline characteristics of the IPTW-weighted CANA and DPP-4 cohorts (CANA: N = 13,793; DPP-4: N = 14,588) were, overall, well balanced (Table 1).
Index treatment and medication use during the observation period
After weighting, most CANA patients were prescribed 100 mg of CANA (59.6%) at the index date, and the average treatment duration was 267 days. For DPP-4 patients, the majority of patients were prescribed sitagliptin (73.6%) among which 64.9% were initiated on 100 mg. The average treatment duration for DPP-4 patients was 282 days, longer than that of the CANA cohort (P < 0.001). Similarly, the average observation period was shorter for CANA patients compared to DPP-4 patients (343 vs. 372 days, P < 0.001).
Fewer patients initiated on CANA added or switched to a new antihyperglicemic agent compared to patients initiated on a DPP-4 agent (30.3 vs. 45.4%, P < 0.001; Table 2). Notably, fewer CANA patients added or switched to a new glucagon-like peptide 1 agonist, insulin, biguanide, sulfonylurea, or thiazolidinedione agent during the observation period, compared to patients initiated on a DPP-4 agent (all P < 0.001; Table 2). CANA patients were also less likely to add or switch to a new SGLT2 agent. This result was expected, since patients would rather switch to an agent with a different mechanism of action. However, CANA patients were as likely as DPP-4 patients to add or switch to another DPP-4 agent (P = 0.100).
Goal achievement of diabetes-related quality measures and weight loss
HbA1c quality measures
Among patients <65 years old with a baseline HbA1c ≥7% (CANA: N = 7,771; DPP-4: N = 8,332; mean baseline HbA1c = 8.9% in both cohorts), a higher proportion achieved an HbA1c <7% in the CANA group than in the DPP-4 group at 6, 9, 12, 18 and 24 months post-index (6-month KM rates: CANA vs. DPP-4: 19.6% vs. 16.0%, log rank test P < 0.001; 24-month KM rates: CANA vs. DPP-4: 42.2% vs. 38.7%, log rank test P < 0.001; Table 3) and CANA patients were 19% more likely to reach an HbA1c <7% compared to DPP-4 patients (HR = 1.19, P < 0.001; Fig. 1). Similarly, among all patients with a baseline HbA1c ≥7%, the proportion of patients achieving an HbA1c <7% (CANA: N = 11,427; DPP-4: N = 12,024; mean baseline HbA1c = 8.8% in both cohorts) was higher in the CANA group than in the DPP-4 group (HR: 1.10, P = 0.007; 24-month KM rates: CANA vs. DPP-4: 42.8% vs. 40.3%, log rank test P < 0.001, Fig. 1 and Table 3).
Among patients ≥65 years old with a baseline HbA1c ≥8% (CANA: N = 1,950; DPP-4: N = 1,980; mean baseline HbA1c = 9.2% in both cohorts), a similar proportion achieved an HbA1c <8% in the CANA group and in the DPP-4 group at 6, 9, 12, 18 and 24 months post-index (6-month KM rates: CANA vs. DPP-4: 37.5% vs. 35.4%, log rank test P = 0.511; 24-month KM rates: CANA vs. DPP-4: 63.5% vs. 65.3%, log rank test P = 0.591; Table 3) and CANA patients were 2% more likely to reach an HbA1c <8% compared to DPP-4 patients, although the result was not found to be statistically significant (HR = 1.02, P = 0.710; Fig. 1). Among all patients with a baseline HbA1c ≥8%, the proportion of patients achieving an HbA1c <8% (CANA: N = 7,124; DPP-4: N = 7,523; mean baseline HbA1c = 9.5% in both cohorts) was higher in the CANA group than in the DPP-4 group (HR: 1.16, P < 0.001; 24-month KM rates: CANA vs. DPP-4: 63.7% vs. 60.0%, log rank test P < 0.001, Fig. 1 and Table 3)
The proportion of patients achieving a BW loss ≥5% among those with ≥1 BW measurement at baseline (CANA: N = 13,661, mean baseline BW = 226.1 lb; DPP-4: N = 14,410, mean baseline BW = 224.4 lb) was higher in the CANA group than in the DPP-4 group starting at 3 months post-index (3-month KM rates: CANA vs. DPP-4: 6.2% vs. 4.5%, log rank test P < 0.001, Table 3); over the entire post-index period, CANA patients were 46% more likely to have a weight loss of 5% than DPP-4 patients (P < 0.001; Fig. 1).
BP quality measures
Starting at 6 months after the index date, the proportion of patients achieving a systolic BP <140 mmHg among those with a baseline systolic BP ≥140 mmHg (CANA: N = 3,670, mean baseline systolic BP = 150.0 mmHg; DPP-4: N = 3,738, mean baseline systolic BP = 150.1 mmHg) was higher in the CANA group than in the DPP-4 group (6-month KM rates: CANA vs. DPP-4: 70.7% vs. 67.1%, log rank test P = 0.056, Table 3). Moreover, over 12 months of observation, CANA patients were more likely to reach a systolic BP <140 mmHg relative to DPP-4 patients (HR = 1.07, P = 0.040; Fig. 1). However, among those with a diastolic BP ≥90 mmHg (CANA: N = 1,501, mean baseline diastolic BP = 94.3 mmHg; DPP-4: N = 1,600, mean baseline diastolic BP = 94.2 mmHg), no differences were found between the two cohorts for the proportion of patients achieving a diastolic BP <90 mmHg (12-month KM rates: CANA vs. DPP-4: 89.6% vs. 89.7%, log rank test P = 0.401; HR: 0.95, P = 0.361).
In this study of EMRs collected between March 2012 to October 2015 including 28,381 T2DM patients initiated on CANA or DPP-4 agents, patients initiated on CANA were more likely to attain HbA1c measurements of <7% or 8% during the observation period compared to patients initiated on a DPP-4 agent. Patients initiated on CANA were also significantly more likely to achieve BW reduction of at least 5% and reduction of systolic BP below 140 mmHg compared with patients initiated on a DPP-4 agent. Patients initiated on CANA were as likely as patients initiated on DPP-4 agents to reach a diastolic BP lower than 90 mmHg, this result was expected because greater reductions in systolic BP than diastolic BP were observed in clinical trials for CANA . Overall, attainment of diabetes-related quality measure goals was more likely in CANA patients than in DPP-4 patients despite the fact that a higher proportion of DPP-4 patients added or switched to a new antihyperglycemic agent during the follow-up period, which may have improved quality measure attainment among DPP-4 patients and reduced observed differences between cohorts.
These results were obtained using IPTW to account for observable differences in patients’ characteristics at treatment initiation or over the baseline period. Before weighting, many patient demographic and clinical characteristics were found to be significantly different between the CANA and DPP-4 patient cohorts. Compared to patients initiated on DPP-4: CANA patients were younger; had used a higher number of antihyperglycemic agents at baseline; suffered more from neuropathy, but less from nephropathy; were more often obese, with on average almost 15 more pounds in terms of BW; and had higher baseline HbA1c values, but similar systolic BP. Such differences were also observed in other retrospective studies comparing patients initiated on CANA and patients initiated on sitagliptin, the most commonly prescribed DPP-4 agent [36, 37]. In particular, Grabner et al. found that patients initiated on CANA were on average younger compared to patients initiated on sitagliptin, suffered more from neuropathy and obesity, and had a higher initial HbA1c . However, after applying IPTW, the weighted populations of the current study were well balanced.
This study adds to the growing body of literature reporting on comparisons of CANA and DPP-4s in terms of quality measures [8,38,, 15, 20, 25, 37–39]. Of particular interest, two clinical trials assessed the efficacy and safety of CANA versus sitagliptin over 52 weeks, respectively on background antihyperglycemic treatment with metformin  and metformin plus sulfonylurea . Significant differences in favor of CANA 100 mg were identified compared to sitagliptin 100 mg at 52 weeks (12 months) in terms of HbA1c < 8%, BP, BMI and BW loss measure achievement, while no difference was found in terms of HbA1c < 7% . When pooling results from both trials, significant differences in favor of CANA 300 mg were found for the same endpoints, including HbA1c < 7%, when compared to sitagliptin 100 mg .
In the present study and following weighting to balance characteristics between patient populations, modest but statistically significant differences between all CANA and DPP-4 patients reaching an HbA1c < 7% appeared at 6 months post-index date, and were observed for those with up to 24 months of follow-up (weighted KM rates respectively 42.8% and 40.3% for CANA and DPP-4, P < 0.001). Among patients younger than 65, significant differences in the proportion of patients reaching an HbA1c < 7% were also found as early as 6 months after treatment initiation and consistently for up to 24 months (42.2% vs. 38.7% respectively for CANA and DPP-4 patients, P < 0.001). Consistent and numerically larger differences were also found as early as 6 months after treatment initiation and up to 24 months in the proportion of patients reaching HbA1c <8% between CANA and DPP-4 cohorts, although no difference was found in the subgroup of patients aged 65 years or older. However, it is important to consider the clinical context of these older patients when interpreting this finding. Specifically, older age is associated with lower eGFR , which may impact dosing considerations. Improving renal and other outcomes (as opposed to targeting HbA1c) may also be of higher priority in older patients, particularly given the promising findings reported for SGLT2s and renal outcomes that have been recently reported [41, 42]. Therefore, future investigations may benefit from exploring other clinical endpoints in older sub-populations of patients. Also of note, although SGLT2 inhibition lowers glucose starting on the first day of initiation, differences with DPP-4 agents are not likely to be found prior to 6 months post-index because HbA1c is likely measured at 6-month intervals among patients with stable glycemic control, as per treatment guidelines (3-month intervals for those not meeting treatment goals) .
To the best of our knowledge, only a single study has compared changes in HbA1c and proportions of patients achieving HbA1c <8% and <7% among patients with T2DM treated with CANA versus DPP-4 inhibitors, using data from the real-world setting . In particular, Thayer et al identified that among matched CANA and DPP-4 inhibitor cohorts of 2,776 patients each, change in HbA1c was greater among patients in the CANA cohort than for those in the DPP-4 inhibitor cohort (−0.92% vs. −0.63%, P < 0.001), and greater percentages of the CANA cohort relative to the DPP-4 inhibitor cohort achieved HbA1c < 7% (35.4% vs. 29.9%, P = 0.022) over a 9 month follow-up . The shorter time to reach the endpoint compared to our study can potentially be explained by the relatively younger population (mean age 56 years vs. 58 years in our study). The aim of the present study was to additionally expand on the clinical benefits other than HbA1c and compare achievement BP, and BW loss goals between a large sample of inadequately controlled patients initiated on CANA or a DPP-4 agent in a real-word setting where patients can be observed for more than two years post-index.
This study was subject to some limitations. First, the data used came from EMRs with a physician centric perspective without any link to healthcare claims. As a consequence, there was no direct link to secondary care data which means that information on hospital visits was not collected. Second, this study relied on prescription data without any knowledge of whether the medication was filled and taken as prescribed. Third, although the CANA prescribing information recommends that patients be initiated on a daily dose of 100 mg , a minority of patients in the present study had their first CANA prescription for 300 mg. Two possible explanations for this observation could be that patients were actually initiated on 300 mg, which is consistent with some clinical trials, or that they received samples of CANA 100 mg prior to the index date, which would not have been observed in this database. Fourth, although the study took appropriate measures to reduce confounding, there remains a risk of residual confounding due to unmeasured confounders, which is inherent to observational studies. For instance, sociodemographic characteristics such as economic status and type of insurance coverage were not available while waist circumference and serum creatinine were available for less than 1% of patients. Fifth, the current study assessed goal attainment using recommended thresholds from various healthcare organizations. However, not every patient has the same goals based on his/her demographic and clinical profile; therefore personalization of goals remains important. Sixth, 75% of the study population was White; consequently, results from the current study may not be generalizable to other races/ethnicities. Finally, adverse events were not assessed in this study given that our data source does not allow tracking of patients across different providers and does not include direct link to secondary care data such as hospital visits. Authors believe that the evaluation of adverse events would be impacted by this limitation. As a consequence, it was not possible to evaluate the impact that such adverse events may have had on the outcomes observed in this study. Although the findings should be interpreted in the context of these limitations, large observational studies such as the one currently presented provide a rich insight into real-world clinical populations and practices.
This real-world retrospective study showed that inadequately controlled patients initiated on CANA were more likely to achieve diabetes-related quality measure goals (HbA1c <7%, HbA1c <8%, BW loss ≥5%, and systolic BP <140 mmHg) compared to patients initiated on a DPP-4 agent. These findings suggest that compared to DPP-4 inhibitors, CANA may be a more effective therapeutic option for improving quality measure goals, which is relevant both for patient outcomes as well as for providers and payers, especially given the importance of quality measure achievement on provider evaluation and reimbursement in the US.
American Association of Clinical Endocrinologists
American College of Endocrinology
American Diabetes Association
Agency for Healthcare Research and Quality
Body mass index
Cegedim Strategic Data
Diabetes Complications Severity Index
Electronic medical records
Federal Drug Administration
Healthcare Effectiveness Data and Information Set
Health Insurance Portability and Accountability Act
Inverse probability of treatment weighting
Quan-Charlson Comorbidity Index
Sodium-glucose co-transporter 2
Type 1 diabetes mellitus
Type 2 diabetes mellitus
Centers for Disease Control and Prevention. National Diabetes Statistics Report: Estimates of Diabetes and Its Burden in the United States, 2014. Atlanta, GA: U.S. Department of Health and Human Services; 2014.
Stolar M. Glycemic control and complications in type 2 diabetes mellitus. Am J Med. 2010;123(3):S3–S11.
Palmer AJ, Roze S, Valentine WJ, Minshall ME, et al. Impact of changes in HbA1c, lipids and blood pressure on long-term outcomes in type 2 diabetes patients: an analysis using the CORE diabetes model. Curr Med Res Opin. 2004;20 Suppl 1:S53–8.
Lafeuille M-H, Grittner AM, Gravel J, Bailey RA, et al. Quality measure attainment in patients with type 2 diabetes mellitus. Am J Manag Care. 2014;20(1 Suppl):s5–15.
Lafeuille M-H, Grittner AM, Gravel J, Bailey RA, et al. Economic simulation of canagliflozin and sitagliptin treatment outcomes in patients with type 2 diabetes mellitus with inadequate glycemic control. J Med Econ. 2015;18(2):113–25.
American Diabetes Association. Standards of medical care in diabetes. Diabetes Care. 2017;40(Supplement 1):S25–32.
National Committee for Quality Assurance (NCQA). HEDIS 2016 - Volume 1: Narrative.; 2016. http://www.ncqa.org/hedis-quality-measurement/hedis-measures/hedis-2016.
Bailey RA, Vijapurkar U, Meininger G, Rupnow MFT, et al. Diabetes-related composite quality End point attainment: canagliflozin versus sitagliptin based on a pooled analysis of 2 clinical trials. Clin Ther. 2015;5:1–10.
Handelsman Y, Bloomgarden ZT, Grunberger G, Umpierrez G, et al. American association of clinical endocrinologists (AACE) and American college of endocrinology (ACE) – clinical practice guidelines for developing a diabetes mellitus comprehensive care plan – 2015. Endocr Pract. 2015;21 Suppl 1:1–87.
Inzucchi SE, Bergenstal RM, Buse JB, Diamant M, et al. Management of hyperglycemia in type 2 diabetes, 2015: a patient-centered approach: update to a position statement of the American diabetes association and the European association for the study of diabetes. Diabetes Care. 2015;38(1):140–9.
Battise DM. Efficacy and safety of canagliflozin as Add-on therapy to metformin in type 2 diabetes. Clin Diabetes. 2014;32(2):81–6.
Seufert J. SGLT2 inhibitors - an insulin-independent therapeutic approach for treatment of type 2 diabetes: focus on canagliflozin. Diabetes, metab syndr obes: targets ther. 2015;8:543–54.
Buysman EK, Chow W, Henk HJ, Rupnow MFT. Characteristics and short-term outcomes of patients with type 2 diabetes mellitus treated with canagliflozin in a real-world setting. Curr Med Res Opin. 2015;31(1):137–43.
Forst T, Guthrie R, Goldenberg R, Yee J, et al. Efficacy and safety of canagliflozin over 52 weeks in patients with type 2 diabetes on background metformin and pioglitazone. Diabetes Obes Metab. 2014;16:467–77.
Lavalle-González FJ, Januszewicz A, Davidson J, Tong C, et al. Efficacy and safety of canagliflozin compared with placebo and sitagliptin in patients with type 2 diabetes on background metformin monotherapy: a randomised trial. Diabetologia. 2013;56:2582–92.
Nardolillo A, Kane M, Busch R, Watsky J, et al. A clinical perspective of canagliflozin in the management of type 2 diabetes mellitus. Clin Med Insights: Endocrinol Diabetes. 2014;7:25–30.
Stenlöf K, Cefalu WT, Kim K, Alba M, et al. Efficacy and safety of canagliflozin monotherapy in subjects with type 2 diabetes mellitus inadequately controlled with diet and exercise. Diabetes Obes Metab. 2013;15:372–82.
Cefalu WT, Leiter LA, Yoon KH, Arias P, et al. Efficacy and safety of canagliflozin versus glimepiride in patients with type 2 diabetes inadequately controlled with metformin (CANTATA-SU): 52 week results from a randomised, double-blind, phase 3 non-inferiority trial. Lancet. 2013;382(9896):941–50.
Meckley L, Miyasato G, Kokkotos F, Bumbaugh J, et al. An observational study of glycemic control in canagliflozin treated patients. Curr Med Res Opin. 2015;Online:1–23.
Schernthaner G, Gross JL, Rosenstock J, Guarisco M, et al. Canagliflozin compared with sitagliptin for patients with type 2 diabetes who do not have adequate glycemic control with metformin plus sulfonylurea: a 52-week randomized trial. Diabetes Care. 2013;36(9):2508–15.
Fulcher G, Matthews DR, Perkovic V, de Zeeuw D, et al. Efficacy and safety of canagliflozin when used in conjunction with incretin-mimetic therapy in patients with type 2 diabetes. Diabetes Obes Metab. 2015;d:82–91.
Thornberry NA, Gallwitz B. Mechanism of action of inhibitors of dipeptidyl-peptidase-4 (DPP-4). best practice & research. Clin endocrinol metab. 2009;23(4):479–86.
Zhang X, Zhao Q. Effects of dipeptidyl peptidase-4 inhibitors on blood pressure in patients with type 2 diabetes: a systematic review and meta-analysis. J Hypertens. 2016;34(2):167–75.
Phung OJ, Scholle JM, Talwar M, Coleman CI. Effect of noninsulin antidiabetic drugs added to metformin therapy on glycemic control, weight gain, and hypoglycemia in type 2 diabetes. JAMA. 2010;303(14):1410–8.
Thayer S, Chow W, Korrer S, Aguilar R. Real-world evaluation of glycemic control among patients with type 2 diabetes mellitus treated with canagliflozin versus dipeptidyl peptidase-4 inhibitors. Current medical research and opinion. 32(6):1087–96.
United States Food and Drug Administration. FDA approves Invokana to treat type 2 diabetes. https://www.jnj.com/media-center/press-releases/us-fda-approves-invokana-canagliflozin-for-the-treatment-of-adults-with-type-2-diabetes. Accessed 1 June 2017.
Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40(5):373–83.
Young BA, Lin E, Von Korff M, Simon G, et al. Diabetes complications severity index and risk of mortality, hospitalization, and healthcare utilization. Am J Manag Care. 2008;14(1):15–23.
Elixhauser A, Steiner C, Kruzikas D. Comorbidity Software Documentation. HCUP Methods Series Report # 2004-1. U.S. Agency for Healthcare Research and Quality. 2004. https://www.hcup-us.ahrq.gov/reports/methods/2004_1.jsp. Accessed 1 June 2017.
Cohen J. Statistical power analysis for the behavioral sciences. Toronto: Academic; 1977.
Austin PC. Using the standardized difference to compare the prevalence of a binary variable between two groups in observational research. Commun Stat-Simul Comput. 2009;38(6):1228–34.
Austin PC, Stuart EA. The performance of inverse probability of treatment weighting and full matching on the propensity score in the presence of model misspecification when estimating the effect of treatment on survival outcomes. Statistical methods in medical research. 2015;1–21.
Curtis LH, Hammill BG, Eisenstein EL, Kramer JM, et al. Using inverse probability-weighted estimators in comparative effectiveness analyses with observational databases. Med Care. 2007;45(10 Supl 2):S103–7.
Hernán MA, Hernández-Díaz S, Robins JM. A structural approach to selection bias. Epidemiology. 2004;15(5):615–25.
Weir MR, Januszewicz A, Gilbert RE, Vijapurkar U, et al. Effect of canagliflozin on blood pressure and adverse events related to osmotic diuresis and reduced intravascular volume in patients with type 2 diabetes mellitus. J clin hypertens (Greenwich, Conn. 2014;16(12):875–82.
Grabner M. Demographic and clinical profiles of type 2 diabetes mellitus patients initiating canagliflozin versus DPP-4 inhibitors in a large U.S. Managed care population. J manag care spec pharm. 2015;21(12):1204–12.
Bailey RA, Damaraju CV, Martin SC, Meininger GE, et al. Attainment of diabetes-related quality measures with canagliflozin versus sitagliptin. Am J Manag Care. 2014;20(1 Suppl):s16–24.
Ektare VU, Lopez JMS, Martin SC, Patel DA, et al. Cost efficiency of canagliflozin versus sitagliptin for type 2 diabetes mellitus. Am J Manag Care. 2014;20(10 Suppl):S204–15.
Bailey RA, Vijapurkar U, Meininger GE, Rupnow MFT, et al. Diabetes-related quality measure attainment: canagliflozin versus sitagliptin based on a pooled analysis of 2 clinical trials. Am J Manag Care. 2014;20(13 Suppl):s296–305.
Solini A, Penno G, Bonora E, Fondelli C, et al. Age, renal dysfunction, cardiovascular disease, and antihyperglycemic treatment in type 2 diabetes mellitus: findings from the renal insufficiency and cardiovascular events Italian multicenter study. J Am Geriatr Soc. 2013;61(8):1253–61.
Heerspink HJL, Desai M, Jardine M, Balis D, et al. Canagliflozin slows progression of renal function decline independently of glycemic effects. J Am Soc Nephrol. 2017;28(1):368–75.
Wanner C, Inzucchi SE, Lachin JM, Fitchett D, et al. Empagliflozin and progression of kidney disease in type 2 diabetes. N Engl J Med. 2016;375(4):323–34.
American Diabetes Association. Standards of medical care in diabetes. Diabetes Care. 2015;38(Supplement 1):S1–S94.
Janssen Pharmaceuticals Inc. Prescribing information for InvokanaTM (canagliflozin). 2015. https://www.invokana.com/prescribing-information.pdf. Accessed 1 June 2017.
This research was funded by Janssen Scientific Affairs, LLC.
Availability of data and materials
The data that support the findings of this study are available from QuintilesIMS but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available.
CHW, PL, DP, MI, MHL, BE, RK, WC, MP, and MSD contributed to the conception and design of the current study. Analyses were carried out by PL, DP, MHL, BE, RK, and MSD. CHW, PL, DP, MI, MHL, BE, RK, WC, MP, and MSD were involved in interpreting the findings and drafting the manuscript. CHW, PL, DP, MI, MHL, BE, RK, WC, MP, and MSD approve the final version.
PL, DP, MHL, BE, RK, and MSD are employees of Analysis Group, Inc., a consulting company that has received research grants from Janssen Scientific Affairs, LLC, to conduct this study. MI, WC, and MP are employees and stockholders of Johnson & Johnson. CHW is a paid consultant for Astra Zeneca, Boehringer Ingelheim, Janssen Scientific Affairs, LLC, Novo Nordisk, and Sanofi, and is on the speaker’s bureau for Astra Zeneca, Boehringer Ingelheim, Eli Lilly, Insulet, Janssen Scientific Affairs, LLC, Novo Nordisk, Sanofi.
Consent for publication
Ethics approval and consent to participate
This research activity was submitted for review to the New England Independent Review Board (NEIRB) and was exempt from NEIRB review, granted that information was recorded in such a manner that subjects could not be identified. The data used in this study was de-identified in compliance with the patient confidentiality requirements of the Health Insurance Portability and Accountability Act (HIPAA). Permission to use this data was granted to Analysis Group by QuintilesIMS.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Wysham, C.H., Lefebvre, P., Pilon, D. et al. Quality measure and weight loss assessment in patients with type 2 diabetes mellitus treated with canagliflozin or dipeptidyl peptidase-4 inhibitors. BMC Endocr Disord 17, 32 (2017). https://doi.org/10.1186/s12902-017-0180-8