Visual Abstract
Keywords: sitagliptin, chronic kidney disease, dosage, toxicity, heart failure
Abstract
Background and objectives
Sitagliptin, a dipeptidyl peptidase-4 inhibitor, is commonly prescribed to patients with type 2 diabetes. As this drug is primarily eliminated by the kidney, a reduced dose is recommended for patients with CKD. Some evidence suggests that sitagliptin is associated with a higher risk of congestive heart failure, particularly at higher doses. We compare the 1-year risk of death or hospitalization with congestive heart failure in patients with CKD newly prescribed sitagliptin at >50 versus ≤50 mg/d.
Design, setting, participants, & measurements
This population-based cohort study included older adults (>66 years) with type 2 diabetes and an eGFR<45 ml/min per 1.73 m2 (but not receiving dialysis) who were newly prescribed sitagliptin between 2010 and 2017 in Ontario, Canada. We used inverse probability of treatment weighting on the basis of propensity scores to balance baseline characteristics. The primary composite outcome was death or hospitalization with congestive heart failure. Secondary outcomes included hospitalization with pancreatitis or hypoglycemia, all-cause hospitalization, and glycemic control. Weighted hazard ratios were obtained using Cox proportional hazards regression, and 95% confidence intervals were obtained using bootstrap variance estimators.
Results
Of 9215 patients, 6518 started sitagliptin at >50 mg/d, and 2697 started sitagliptin at ≤50 mg/d. The 1-year risk of death or hospitalization with congestive heart failure did not differ significantly between groups (79 versus 126 events per 1000 person-years; weighted hazard ratio, 0.88; 95% confidence interval, 0.67 to 1.14); hospitalization with pancreatitis (weighted hazard ratio, 0.98; 95% confidence interval, 0.32 to 3.03) and hypoglycemia (weighted hazard ratio, 1.10; 95% confidence interval, 0.64 to 1.90) also did not differ significantly between groups. Patients starting sitagliptin at >50 mg/d had lower mean glycated hemoglobin concentrations (weighted between-group difference, −0.12%; 95% confidence interval, −0.19 to −0.06) and a lower risk of all-cause hospitalization (weighted hazard ratio, 0.81; 95% confidence interval, 0.66 to 0.98).
Conclusions
The risk of death or congestive heart failure was not higher in older adults with CKD starting sitagliptin at >50 versus ≤50 mg/d.
Podcast
This article contains a podcast at https://www.asn-online.org/media/podcast/CJASN/2020_11_25_CJN08310520_final.mp3
Introduction
Sitagliptin is an oral dipeptidyl peptidase-4 (DPP-4) inhibitor used to improve glycemic control in patients with type 2 diabetes mellitus (1–4). In patients with contraindications or intolerance to metformin, sitagliptin may be prescribed in monotherapy or as an add-on to insulin (1–4). In the United States, 8 million sitagliptin prescriptions were filled in 2016, and in Ontario, Canada, nearly 20% of all antihyperglycemic medication prescriptions are for sitagliptin (5).
Sitagliptin is eliminated primarily by the kidney (2,3). In a pharmacokinetic study, plasma concentrations of sitagliptin were two to four times higher in patients with moderate to severe CKD than in patients with normal kidney function after receipt of one oral dose of 50 mg sitagliptin (6). For this reason, the product monograph recommends starting sitagliptin at a lower dose in patients with an eGFR below 45 ml/min per 1.73 m2 (Supplemental Table 1) (2). However, no clinical studies have compared outcomes in patients with CKD who start sitagliptin at a higher versus lower dose (Supplemental Table 2). Although results are mixed, DPP-4 inhibitor use has been linked to a higher risk of congestive heart failure in meta-analyses of clinical trials and some observational studies (7–12). A higher risk of congestive heart failure was observed in sitagliptin users versus nonusers in three population-based studies (10,11,13). DPP-4 inhibitors may exert deleterious cardiovascular effects by increasing the production of endogenous peptides (Glucagon-like peptide-1 and stroma cell–derived factor-1), which can increase cardiac muscle contraction and promote cardiac fibrosis (14).
Approximately 40% of older adults with type 2 diabetes and CKD have an eGFR between 15 and 59 ml/min per 1.73 m2 (15). In practice, sitagliptin is rarely started at a lower dose in these patients (16). We conducted a population-based study of older adults with an eGFR below 45 ml/min per 1.73 m2 (excluding those receiving dialysis) to compare the 1-year risk of death or hospitalization with congestive heart failure in outpatients starting oral sitagliptin at >50 versus ≤50 mg/d. We hypothesized that starting sitagliptin at >50 mg/d would be associated with a higher risk of congestive heart failure.
Materials and Methods
Study Design and Setting
We conducted a population-based cohort study using linked administrative health care databases in the province of Ontario, Canada (2010–2018). All Ontario residents (approximately 14 million) have universal access to hospital care and physician services through a government-funded single-payer system (17). Those aged 65 years and older (approximately 2.2 million) also receive universal prescription drug coverage. The use of data in this study was authorized under section 45 of Ontario’s Personal Health Information Protection Act, which does not require review by a research ethics board. Study reporting follows recommended guidelines for observational studies that use routinely collected health data (Supplemental Table 3) (18,19).
Data Sources
Eight health care databases housed at ICES (ices.on.ca) were used to obtain information on patient characteristics, prescription drug use, covariates, and the outcomes (20). The datasets were linked using unique encoded identifiers and analyzed at ICES. We used the following databases: the Canadian Institute for Health Information Discharge Abstract Database, ICES-derived Physician Database, the National Ambulatory Care Reporting System, the Ontario Drug Benefit Database, the Ontario Health Insurance Plan database, the Ontario Laboratories Information System, the Ontario Mental Health Reporting System, and the Registered Persons Database. Hospital admissions and diagnoses are coded by trained personnel using the International Classification of Diseases 10th revision system; personnel only consider physician-recorded diagnoses in a patient’s medical chart when assigning codes and do not review or interpret symptoms or test results. These databases have been used previously to study adverse drug events and health outcomes (21–25). Except for prescriber data (7% missing; defined as a separate category) and neighborhood income quintile (0.3% missing; recorded as the middle quintile), the databases were complete for all variables used in this study. Emigration from the province, which occurs at a rate of 0.5%/yr, was the only reason for loss to follow-up (26). The codes used to ascertain comorbidities and outcomes are detailed in Supplemental Table 4.
The date sitagliptin was dispensed from the pharmacy served as each patient’s cohort entry date. Baseline comorbidities were assessed in the 5-year period before cohort entry and health care use in the 1-year period before cohort entry. A 120-day look-back period was used to ascertain prescription drug exposure because the Ontario Drug Benefits program allows a maximum prescription duration of 100 days.
Patients
We assembled a primary cohort of adults aged 66 years and older who had an eGFR<45 ml/min per 1.73 m2 and were newly dispensed oral sitagliptin from an outpatient pharmacy between June 2010 (when sitagliptin was first openly listed on the province’s formulary) and December 2017. We restricted the cohort to patients aged 66 years and older to ensure that all patients had at least 1 year of prior prescription drug coverage.
We calculated the eGFR for each patient using the Chronic Kidney Disease Epidemiology equation; the justification to use this equation for dose prescribing is provided in Supplemental Table 5 (27). We used the most recent outpatient serum creatinine measurement (using the isotope dilution mass spectroscopy–traceable enzymatic method) recorded before the cohort entry date (27). In Ontario, many older adults have at least one outpatient serum creatinine measured in routine care each year, and we have shown that single creatinine values are representative of chronic values (28). We excluded patients with no serum creatinine measurement in the year before the cohort entry date, kidney transplant recipients, and patients receiving dialysis at or before cohort entry.
To ensure that patients were new sitagliptin users, those with any evidence of sitagliptin use, including combination drug prescriptions (i.e., sitagliptin-metformin for example) and other DPP-4 inhibitors, in the 180-day period before the cohort entry date were excluded. We also excluded those who were discharged from the hospital or emergency department within 2 days before the cohort entry date (in Ontario, patients who start a sitagliptin prescription during a hospital admission would have their outpatient prescription dispensed on the same day or the day after hospital discharge). Patients could only enter the cohort once.
Sitagliptin Dose
The recommended dose of sitagliptin when the eGFR is between 30 and 45 ml/min per 1.73 m2 is 50 mg/d (Supplemental Table 1). We categorized patients as those who started oral sitagliptin at >50 mg/d and those who started at ≤50 mg/d. The sample was too small to include an additional category of 25 mg/d, which is the dose recommended in patients with eGFR below 30 ml/min per 1.73 m2 (Supplemental Table 1) (2).
Outcomes
The outcomes were prespecified. The diagnostic codes for all outcomes, their validation, and their interpretation are provided in Supplemental Table 6.
Primary Outcome.
The primary outcome was a composite of time to death or first hospitalization with congestive heart failure within 1 year of initiating sitagliptin. This time frame was defined on the basis of the The Saxagliptin Assessment of Vascular Outcomes Recorded in Patients With Diabetes Mellitus–Thrombolysis in Myocardial Infarction 53 trial (SAVOR-TIMI 53 trial), which showed a higher 1-year risk of congestive heart failure in patients with type 2 diabetes who received saxagliptin (a DPP-4 inhibitor) versus placebo (29). Components of the primary outcome were defined using codes proven to have good validity when compared with chart review (Supplemental Table 6).
Secondary Outcomes.
The first two secondary outcomes were the two components of the primary outcome analyzed separately. The remaining three secondary outcomes were time to first hospitalization or emergency department visit with pancreatitis (a potential DPP-4 inhibitor–related side effect) (12), time to first hospitalization or emergency department visit with hypoglycemia, and time to first hospitalization for any reason. (Supplemental Table 6).
Statistical Analyses
Analyses were conducted using SAS version 9.4 (SAS Institute, Cary, NC). We used inverse probability of treatment weighting on the propensity score to balance comparison groups on indicators of baseline health (30–32). We estimated the propensity score using multivariable logistic regression with 162 covariates chosen a priori (defined in Supplemental Table 7) because they are known confounders or risk factors for congestive heart failure (32–34). We weighted patients in the reference group (≤50 mg/d of sitagliptin) using average treatment effect for the treated weights defined as [propensity score/(1− propensity score)], with patients in the exposed group (>50 mg/d of sitagliptin) receiving weights of one (3–32). This method produces a weighted pseudosample of patients in the reference group with a similar distribution of measured covariates as the exposed group (>50 mg/d of sitagliptin) (30, 31). We compared between-group differences in baseline characteristics using standardized differences in both the unweighted and weighted samples (differences >10% were considered meaningful) (35). We obtained weighted hazard ratios using a Cox proportional hazards regression and 95% confidence intervals (95% CIs) using a bootstrap variance estimator (36). We assessed the proportional hazards assumption using a time-dependent covariate test, which was met for all outcomes. We conducted all primary analyses according to the intention-to-treat principle (i.e., patients were not censored if they discontinued sitagliptin or if their dose changed in follow-up). Death was treated as a censoring event when it was not part of an outcome. We interpreted two-tailed P values of 0.05 as statistically significant.
Additional Analyses
We conducted seven additional analyses. We recalculated the propensity scores (as done in the primary analysis) and used inverse probability of treatment weighting on the propensity score to balance comparison groups on indicators of baseline health.
In patients with blood glycated hemoglobin (HbA1c) concentrations available at baseline and within 1 year after cohort entry (the most recent measurement), we analyzed the between-group difference in the absolute change in HbA1c to determine if starting sitagliptin at higher dose was associated with better glycemic control. The between-group difference was analyzed using a binomial regression model with an identity link function (we assumed the data were normally distributed).
We compared the two groups of sitagliptin users (i.e., those prescribed >50 and ≤50 mg/d) with new linagliptin users on the 1-year risk of death or hospitalization with congestive heart failure. Unlike sitagliptin, which is eliminated primarily by the kidney, linagliptin is primarily eliminated by the enterohepatic system, and there is no recommendation to reduce the dose of linagliptin in patients with CKD.
We examined effect modification by baseline eGFR (≥30 versus <30 ml/min per 1.73 m2) and history of congestive heart failure (interaction terms were included in the models).
We performed an as-treated analysis censoring follow-up at sitagliptin discontinuation. A patient was considered to be continuously exposed during a series of prescriptions if the gap between prescriptions was within a period equivalent to 150% of the number of days of the previous prescription.
We extended the washout period before the cohort entry date from 180 to 365 days to reduce incident user misclassification.
We extended the follow-up period from 1 to 2 years after sitagliptin initiation in long-term users (i.e., patients with continuous sitagliptin use ≥180 days since the initiation). The median duration of continuous sitagliptin dispensing was 711 days (interquartile range [IQR], 400–1280.5) in the high-dose group and 738 days (IQR, 370–1252) in the low-dose group (Supplemental Table 8).
We excluded patients who received any antidiabetic prescriptions other than sitagliptin on the cohort entry date to isolate the effect of sitagliptin from the effect of other antidiabetic medications.
Results
Patients
The flow diagram for the cohort build is shown in Supplemental Figure 1. The primary cohort included 9215 older adults with an eGFR<45 ml/min per 1.73 m2 (median age 78 years; IQR, 73–83; 56% women) who were newly dispensed sitagliptin at an outpatient pharmacy. The outpatient serum creatinine to estimate baseline GFR was measured a median of 23 (IQR, 8–73) days prior to cohort entry. Overall, 77% of patients had an eGFR between 30 and 45 ml/min per 1.73 m2, and 23% had an eGFR<30 ml/min per 1.73 m2.
Patients received sitagliptin prescriptions primarily from primary care physicians (75%), endocrinologists (7%), and nephrologists (3%). Sitagliptin was prescribed by 4342 different physicians, and it was dispensed by 3139 different pharmacies. Most patients filled prescriptions for 100 mg/d, resulting in a median dose in each eGFR category of 100 mg/d (IQR, 50–100); 6518 (71%) started at >50 mg/d (median 100 mg/d; IQR, 100–100), and 2697 (29%) started at ≤50 mg/d (median 50 mg/d; IQR, 25.7–50).
Characteristics of patients who started sitagliptin at >50 versus ≤50 mg/d are shown in Table 1 (the full set of 173 characteristics is shown in Supplemental Table 9). After weighting, the standardized differences were <10% for 172 of 173 variables (99%), including prescriber specialty, comorbidities, baseline eGFR, diabetes characteristics, and diabetes medications (Supplemental Table 9). Although standardized differences were 11% for the lowest- and middle-income categories, this difference is unlikely to be clinically meaningful (Table 1, Supplemental Table 9).
Table 1.
Baseline characteristics of older adults with CKD who were newly prescribed sitagliptin in Ontario, Canada (2010–2017)
| Baseline Characteristics | Unweighted Data, n=9215a | Weighted Data, n=12,828b | ||||
|---|---|---|---|---|---|---|
| Sitagliptin Dose >50 mg/d, n=6518 | Sitagliptin Dose ≤50 mg/d, n=2697 | Standardized Difference,c % | Sitagliptin Dose >50 mg/d, n=6518 | Sitagliptin Dose ≤50 mg/d, n=6310 | Standardized Difference,c % | |
| Demographics | ||||||
| Women, no. (%) | 3653 (56) | 1475 (55) | 3 | 3653 (56) | 3470 (55) | 2 |
| Age, mean (SD), yr | 78 (7) | 80 (8) | 24 | 78 (7) | 78 (11) | 3 |
| Residence, no. (%) | ||||||
| Urban | 5833 (90) | 2398 (89) | 2 | 5833 (90) | 5575 (88) | 4 |
| Rural | 685 (11) | 299 (11) | 2 | 685 (11) | 735 (12) | 4 |
| Long-term care | 240 (4) | 257 (10) | 24 | 240 (4) | 250 (4) | 2 |
| Income quintile, no. (%)d | ||||||
| 1 (lowest) | 1519 (23) | 665 (25) | 3 | 1519 (23) | 1776 (28) | 11 |
| 2 | 1492 (23) | 613 (23) | 0 | 1492 (23) | 1259 (20) | 7 |
| 3 (middle) | 1363 (21) | 584 (22) | 2 | 1363 (21) | 1050 (17) | 11 |
| 4 | 1153 (18) | 440 (16) | 4 | 1153 (18) | 1293 (21) | 7 |
| 5 (highest) | 992 (15) | 395 (15) | 2 | 992 (15) | 932 (15) | 1 |
| Kidney function | ||||||
| eGFR, mean (SD),e ml/min per 1.73 m2 | 36 (7) | 33 (8) | 37 | 36 (7) | 36 (11) | 1 |
| eGFR category, no. (%), ml/min per 1.73 m2 | ||||||
| <30 | 1264 (19) | 859 (32) | 29 | 1264 (19) | 1234 (20) | 1 |
| 30 to <45 | 5254 (81) | 1838 (68) | 29 | 5254 (81) | 5076 (80) | 1 |
| Sitagliptin prescriber, no. (%) | ||||||
| General practitioner | 5125 (79) | 1784 (66) | 28 | 5125 (79) | 4900 (78) | 2 |
| Endocrinologist | 403 (6) | 246 (9) | 11 | 403 (6) | 458 (7) | 4 |
| Nephrologist | 62 (1) | 220 (8) | 35 | 62 (1) | 64 (1) | 0 |
| Internist | 267 (4) | 128 (5) | 3 | 267 (4) | 237 (4) | 2 |
| Cardiologist | 86 (1) | 21 (0.8) | 5 | 86 (1) | 59 (0.9) | 4 |
| Other | 152 (2) | 49 (2) | 4 | 152 (2) | 122 (2) | 3 |
| Missing | 423 (7) | 249 (9) | 10 | 423 (7) | 475 (7) | 4 |
| Comorbidities, no. (%)f | ||||||
| Alcohol-related disorder | 40 (0.6) | 20 (0.7) | 1 | 40 (0.6) | 86 (1) | 8 |
| Atrial fibrillation or flutter | 575 (9) | 298 (11) | 7 | 575 (9) | 500 (8) | 3 |
| Cancer | 2119 (33) | 941 (35) | 5 | 2119 (33) | 2055 (33) | 0 |
| Chronic obstructive pulmonary disease | 1592 (24) | 665 (25) | 1 | 1592 (24) | 1780 (28) | 9 |
| Coronary artery disease | 2396 (37) | 993 (37) | 0 | 2396 (37) | 2222 (35) | 3 |
| Dyslipidemia | 1726 (27) | 629 (23) | 7 | 1726 (27) | 1823 (29) | 5 |
| Hypertension | 6118 (94) | 2503 (93) | 4 | 6118 (94) | 5884 (93) | 2 |
| Peripheral vascular disease | 124 (2) | 45 (2) | 2 | 124 (2) | 126 (2) | 1 |
| Acute myocardial infarction | 364 (6) | 155 (6) | 0 | 364 (6) | 342 (5) | 1 |
| Ischemic stroke | 153 (2) | 64 (2) | 1 | 153 (2) | 244 (4) | 9 |
| Diabetic neuropathy | 100 (2) | 60 (2) | 5 | 100 (2) | 124 (2) | 4 |
| Diabetic retinopathy | 115 (2) | 39 (1) | 3 | 115 (2) | 105 (2) | 1 |
| Prior pancreatitis | 50 (0.8) | 32 (1) | 4 | 50 (0.8) | 62 (1) | 2 |
| Prior hypoglycemia | 184 (3) | 84 (3) | 2 | 184 (3) | 140 (2) | 4 |
| Prior congestive heart failure | 1439 (22) | 665 (25) | 6 | 1439 (22) | 1299 (21) | 4 |
| Charlson comorbidity index, mean (SD)g | 3 (1.7) | 3.2 (1.8) | 11 | 3 (1.7) | 3.1 (2.6) | 2 |
| Health care visits/testsh | ||||||
| Primary care visits, mean (SD) | 10.9 (9.7) | 11.3 (11.0) | 3 | 10.9 (9.7) | 10.7 (13.9) | 2 |
| Emergency department visits, mean (SD) | 0.7 (1.3) | 0.7 (1.4) | 5 | 0.7 (1.3) | 0.8 (2.8) | 8 |
| Serum creatinine tests, no. (%) | 3.4 (2.5) | 4.0 (2.8) | 22 | 3.4 (2.5) | 3.5 (3.5) | 1 |
| Medication use, no. (%)i | ||||||
| Insulin | 745 (11) | 499 (20) | 20 | 745 (11) | 730 (12) | 1 |
| Metformin | 4014 (62) | 1335 (50) | 25 | 4014 (62) | 3942 (63) | 2 |
| Glyburide | 1471 (23) | 286 (11) | 33 | 1471 (23) | 1403 (22) | 1 |
| Gliclazide | 2146 (33) | 875 (32) | 1 | 2146 (33) | 2230 (35) | 5 |
| Pioglitazone | 603 (9) | 79 (3) | 27 | 603 (9) | 495 (8) | 5 |
| Rosiglitazone | 159 (2.4) | 16 (0.6) | 15 | 159 (2.4) | 219 (3.5) | 7 |
| Repaglinide | 56 (0.9) | 29 (1) | 2 | 56 (0.9) | 38 (0.6) | 3 |
| Laboratory test valuesj | ||||||
| Urine ACR available | 4293 (66) | 1770 (66) | 1 | 4293 (66) | 4256 (67) | 3 |
| Baseline ACR categories, μg/mg | ||||||
| Missing | 2225 (34) | 927 (34) | 1 | 2225 (34) | 2054 (33) | 3 |
| <30 | 1998 (31) | 657 (24) | 14 | 1998 (31) | 1783 (28) | 5 |
| 30–300 | 1617 (25) | 695 (26) | 2 | 1617 (25) | 1638 (26) | 3 |
| >300 | 678 (10) | 418 (16) | 15 | 678 (10) | 835 (13) | 9 |
| Test for glycosylated hemoglobin levels available | 6211 (95) | 2577 (96) | 1 | 6211 (95) | 6023 (96) | 1 |
| Glycosylated hemoglobin level, mean (SD), % | 7.9 (1.4) | 7.9 (1.4) | 4 | 7.9 (1.4) | 7.8 (2.0) | 5 |
ACR, urine albumin-creatinine ratio.
Unless otherwise specified in the footnotes, baseline characteristics were assessed on the date that the patient filled the sitagliptin prescription—the cohort entry date.
Weighted using inverse probability of treatment weighting on the basis of propensity scores. The propensity score was estimated using multivariable logistic regression with 162 covariates chosen a priori (defined in Supplemental Table 7). Patients in the reference group were weighted as [propensity score/(1− propensity score)] (30–32). This method produces a weighted pseudosample of patients in the reference group with the same distribution of measured covariates as the exposure group (30,31).
Standardized difference is the difference between the groups divided by the pooled SD; a value >10% is interpreted as a meaningful difference (35).
Income was categorized into fifths of average neighborhood income on the cohort entry date; missing data on this variable (0.3%) were recorded as the middle quintile.
The most recent eGFR measurement in the 365-day period before the cohort entry date (including the cohort entry date); eGFR was calculated using the Chronic Kidney Disease Epidemiology Collaboration equation: 141×min ([serum creatinine concentration in micromoles per liter per 88.4]/ĸ, 1)α×max ([serum creatinine concentration in micromoles per liter per 88.4]/ĸ, 1)−1.209×0.993 age ×1.018 [if a woman] ×1.159 [if Black] (ĸ=0.7 if a woman and 0.9 if a man; α=−0.329 if a woman and −0.411 if a man; min = the minimum of serum creatinine concentration/ĸ or 1; max = the maximum of serum creatinine concentration/ĸ or 1. Information on race was not available in our data sources, and all patients were assumed not to be of African-Canadian race; African Canadians represented <5% of the population of Ontario in 2006. The lowest eGFR value was 15.1 ml/min per 1.73 m2.
Baseline comorbidities were assessed in the 5-year period before the cohort entry date.
Presence of kidney disease is a variable in the Charlson comorbidity index, which automatically results in all individuals receiving a minimum score of two.
Total number of health care visits/tests in the 12-month period before the cohort entry date.
Medication use was examined in the 120-day period before the cohort entry date (the Ontario Drug Benefit program dispenses a maximum 100-day supply). Some of these medications may have been discontinued after the initiation of sitagliptin.
Most recent laboratory test values in the 1- to 365-day period before the cohort entry date.
Additional details about the duration of sitagliptin use, rate of sitagliptin discontinuation, and the person-years of follow-up for sitagliptin users are presented in Supplemental Tables 10–12.
Primary Outcome
The primary outcome of death or hospitalization with congestive heart failure occurred in 495 of 6518 patients who started sitagliptin at >50 mg/d (7.6%; 79.2 events per 1000 person-years) and in 318 of 2697 patients who started sitagliptin ≤50 mg/d (11.8%; 126.1 events per 1000 person-years). Starting sitagliptin at >50 mg/d compared with ≤50 mg/d was not associated with a higher risk of death or a hospital admission with congestive heart failure within 1 year (weighted hazard ratio, 0.88; 95% CI, 0.67 to 1.14) (Figure 1, Table 2).
Figure 1.
The risk of death or congestive heart failure was not higher in older adults with CKD starting sitagliptin at >50 mg/d versus ≤50 mg/d. Kaplan–Meier estimates of survival probability without death or hospital admission with heart failure in older adults with moderate to severe CKD who started a new prescription for sitagliptin >50 versus ≤50 mg/d (weighted results). Inverse probability of treatment weighting on the propensity score was used to balance comparison groups on indicators of baseline health (30–32). The propensity score was estimated using multivariable logistic regression with 162 covariates chosen a priori (defined in Supplemental Table 7). Patients in the reference group were weighted as [propensity score/(1− propensity score)] (30–32). This method produces a weighted pseudosample of patients in the reference group with the same distribution of measured covariates as the exposed group (30,31). Weighted hazard ratios and 95% confidence intervals (95% CIs) were obtained using a Cox proportional hazards regression, and 95% CIs were obtained using a bootstrap variance estimator (36). The proportional hazards assumption was assessed using a time-dependent covariate test and was met for all outcomes.
Table 2.
Risk of death or heart failure in older adults with moderate to severe CKD who started a new prescription for sitagliptin >50 versus ≤50 mg/d
| Baseline Characteristics | Unweighteda | Weightedb | |||||||
|---|---|---|---|---|---|---|---|---|---|
| No. of Events (%) | No. of Events per 1000 person-yr | No. of Events (%) | No. of Events per 1000 person-yr | Hazard Ratio (95% Confidence Interval) | |||||
| Sitagliptin Dose >50 mg/d, n=6518 | Sitagliptin Dose ≤50 mg/d, n=2697 | Sitagliptin Dose >50 mg/d | Sitagliptin Dose ≤50 mg/d | Sitagliptin Dose >50 mg/d, n=6518 | Sitagliptin Dose ≤50 mg/d, n=6310 | Sitagliptin Dose >50 mg/d | Sitagliptin Dose ≤50 mg/d | ||
| Primary outcome | |||||||||
| Death or hospital admission with heart failure | 495 (7.6) | 318 (11.8) | 79.2 | 126.1 | 495 (7.6) | 543 (8.6) | 79.2 | 90.6 | 0.88 (0.67 to 1.14) |
| Secondary outcomes | |||||||||
| Death | 404 (6.2) | 271 (10.1) | 63.9 | 105.6 | 404 (6.2) | 432 (6.9) | 63.9 | 71.0 | 0.90 (0.68 to 1.20) |
| Hospital admission with heart failure | 126 (1.9) | 67 (2.5) | 20.1 | 26.4 | 126 (1.9) | 141 (2.2) | 20.1 | 23.4 | 0.86 (0.52 to 1.40) |
| Hospital encounter with pancreatitis | 25 (0.4) | 11 (0.4) | 4.0 | 4.3 | 25 (0.4) | 24 (0.4) | 4.0 | 4.0 | 0.98 (0.32 to 3.03) |
| Hospital encounter with hypoglycemia | 140 (2.2) | 55 (2.0) | 22.4 | 21.7 | 140 (2.2) | 122 (1.9) | 22.4 | 20.3 | 1.10 (0.64 to 1.90) |
| All-cause hospitalization | 1614 (24.8) | 754 (28.0) | 284.5 | 331.3 | 1614 (24.8) | 1879 (29.8) | 284.5 | 352.4 | 0.81 (0.66 to 0.98) |
Reference group: sitagliptin dose ≤50 mg/d.
Inverse probability of treatment weighting on the propensity score was used to balance comparison groups on indicators of baseline health (30–32). The propensity score was estimated using multivariable logistic regression with 162 covariates chosen a priori (defined in Supplemental Table 7). Patients in the reference group were weighted as [propensity score/(1− propensity score)] (30–32). This method produces a weighted pseudosample of patients in the reference group with the same distribution of measured covariates as the exposed group (30,31). Weighted hazard ratios and 95% confidence intervals were obtained using a Cox proportional hazards regression, and 95% confidence intervals were obtained using a bootstrap variance estimator (36). The proportional hazards assumption was assessed using a time-dependent covariate test and was met for all outcomes.
Secondary Outcomes
Starting sitagliptin at >50 versus ≤50 mg/d was not associated with a higher 1-year risk of a hospital encounter with pancreatitis (weighted hazard ratio, 0.98; 95% CI, 0.32 to 3.03), a hospital encounter with hypoglycemia (weighted hazard ratio, 1.10; 95% CI, 0.64 to 1.90), death (weighted hazard ratio, 0.90; 95% CI, 0.68 to 1.20), or hospitalization with congestive heart failure (weighted hazard ratio, 0.86; 95% CI, 0.52 to 1.40). The risk of hospitalization for any cause was lower in patients starting sitagliptin at >50 versus ≤50 mg/d (weighted hazard ratio, 0.81; 95% CI, 0.66 to 0.98) (Figure 2, Table 2).
Figure 2.
Neither baseline eGFR category or history of congestive heart significantly modified the association between starting sitagliptin at >50 versus ≤50 mg/day and the risk of death or hospitalization with congestive heart failure. Subgroup analysis for risk of death of heart failure by eGFR category and by history of heart failure (weighted results). Inverse probability of treatment weighting on the propensity score was used to balance comparison groups on indicators of baseline health (30–32). The propensity score was estimated using multivariable logistic regression with 162 covariates chosen a priori (defined in Supplemental Table 7). Patients in the reference group were weighted as [propensity score/(1− propensity score)] (30–32). This method produces a weighted pseudosample of patients in the reference group with the same distribution of measured covariates as the exposed group (30,31). Weighted hazard ratios and 95% CIs were obtained using a Cox proportional hazards regression, and 95% CIs were obtained using a bootstrap variance estimator (36). The proportional hazards assumption was assessed using a time-dependent covariate test and was met for all outcomes.
Additional Analyses
(1) A total of 7820 patients had both baseline and follow-up HbA1C measurements (5594 sitagliptin >50 mg/d, 2226 sitagliptin ≤50 mg/d). After weighting, the two groups were balanced on all baseline characteristics. Starting sitagliptin at >50 versus ≤50 mg/d was associated with a greater decrement in HbA1c (the weighted mean between-group difference was −0.12%; 95% CI, −0.19 to −0.06) (Table 3). (2) In comparison with linagliptin users, both groups of sitagliptin users (≤50 and >50 mg/d) did not have a higher 1-year risk of death or hospitalization with congestive heart failure: weighted hazard ratios were 0.93 (95% CI, 0.74 to 1.04) and 1.05 (95% CI, 0.92 to 1.19), respectively (Supplemental Tables 13 and 14). (3) Neither baseline eGFR category nor history of congestive heart significantly modified the association between starting sitagliptin at >50 versus ≤50 mg/d and the risk of death or hospitalization with congestive heart failure (Figure 2, test for interaction P=0.25 and P=0.14, respectively). (4–7) Results were consistent in the remaining additional analyses: when the follow-up time was censored at sitagliptin discontinuation (Supplemental Table 15), when the washout period was extended to 1 year (Supplemental Table 16), when the follow-up period was extended to 2 years (Supplemental Table 8), and after excluding patients who received antidiabetic prescriptions other than sitagliptin on the cohort entry date (Supplemental Table 17).
Table 3.
Changes in glycated hemoglobin within 365 days of new sitagliptin use
| Baseline Characteristics | Unweighteda | Weightedb | Weighted Mean Differences (95% Confidence Interval) | ||
|---|---|---|---|---|---|
| Sitagliptin Dose >50 mg/d, n=5594 | Sitagliptin Dose ≤50 mg/d, n=2226 | Sitagliptin Dose >50 mg/d, n=5594 | Sitagliptin Dose ≤50 mg/d, n=5413 | ||
| Change in Hba1c (%), mean (95% CI)c | −0.46 (−0.50 to−0.43) | −0.45 (−0.51 to−0.40) | −0.46 (−0.50 to−0.43) | −0.34 (−0.39 to−0.28) | −0.12 (−0.19 to −0.06) |
HbA1c, glycated hemoglobin; 95% CI, 95% confidence interval.
Reference group: sitagliptin dose ≤50 mg/d.
Inverse probability of treatment weighting on the propensity score was used to balance comparison groups on indicators of baseline health (30–32). The propensity score was estimated using multivariable logistic regression with 162 covariates chosen a priori (defined in Supplemental Table 7). Patients in the reference group were weighted as [propensity score/(1− propensity score)] (30–32). This method produces a weighted pseudosample of patients in the reference group with the same distribution of measured covariates as the exposed group (30,31).
Change in HbA1c was defined as the most recent serum HbA1c measurement in the 365 days after the index date minus the baseline serum HbA1c measurement. The most recent serum HbA1c level may have been measured in an outpatient laboratory, emergency department, or hospital setting.
Discussion
In this population-based cohort study of 9215 older adults with moderate to severe CKD, the 1-year risk of death or hospitalization with congestive heart failure was not significantly different in patients starting sitagliptin at ≤50 versus >50 mg/d. Furthermore, there was no association with the sitagliptin dose and the risk of other outcomes, including hospital encounters with pancreatitis or hypoglycemia. These findings were consistent across multiple additional analyses. The risk of congestive heart failure in sitagliptin users (i.e., those prescribed >50 mg/d and those prescribed ≤50 mg/d) was similar to that in linagliptin users. Our findings suggest that sitagliptin poses no higher risk of congestive heart failure at higher than recommended doses in patients with CKD.
To our knowledge, no previous studies have examined the safety of high- versus low-dose sitagliptin use in patients with type 2 diabetes and CKD; however, our results are consistent with findings from The Trial Evaluating Cardiovascular Outcomes with Sitagliptin (TECOS), which studied 3000 patients (mean age 69 years) with CKD (eGFR between 30 and 60 ml/min per 1.73 m2) (37,38). The TECOS trial examined the effect of sitagliptin versus placebo on major adverse cardiovascular outcomes (a composite of cardiovascular death, myocardial infarction, stroke, unstable angina, and congestive heart failure as a separate outcome) and adverse clinical events (hypoglycemia and acute pancreatitis). Sitagliptin was provided at a dose of 50 mg/d (or 100 mg if eGFR was >50 ml/min per 1.73 m2). Patients had a median follow-up time of 2.8 (2.2–3.6) years; no significant differences in outcomes were observed between patients receiving sitagliptin versus placebo (37,38). Other trials of DPP-4 inhibitors have shown neutral effects on cardiovascular outcomes overall, including congestive heart failure (39–41), although the SAVOR-TIMI 53 trial reported an increased risk of congestive heart failure in patients randomized to receive the DPP-4 inhibitor saxagliptin (29). In three prior observational studies, a higher risk of congestive heart failure was observed in patients receiving sitagliptin (including patients with kidney failure) compared with nonusers (10,11,13). The biologic plausibility of a deleterious cardiovascular effect of sitagliptin is unclear but may involve a hyperproduction of endogenous peptides, which can increase cardiac muscle contraction and promote cardiac fibrosis. Reconciling findings from these observational studies is challenging because the study population, exposures, patient characteristics, and outcomes differ across studies. For example, none of these studies used an active comparator. Furthermore, we based our hypothesis on the pharmacokinetics of sitagliptin, but these data are often derived from volunteers who receive a single dose of the drug, and results may not reflect the true pharmacokinetics of the drug in clinical practice nor in patients with CKD who often have multiple comorbidities. Our data do not support the hypothesis that starting a higher than recommended dose of sitagliptin in older adults with CKD is associated with significant harm. In fact, our findings suggest a possible benefit to patients with CKD who started a higher versus lower dose of sitagliptin. It is possible that patients who start at a higher dose achieve better glycemic control and have a lower risk of being admitted to the hospital; however, residual confounding cannot be ruled out—before weighting, sicker patients were less likely to be prescribed a higher dose of sitagliptin, and these results need to be replicated in other studies.
Our study has several strengths. It is the first population-based study to examine the risk of death or congestive heart failure associated with the use of sitagliptin in patients with moderate to severe CKD in routine practice. The findings of this study are likely generalizable to older adults because it was conducted in the setting of usual clinical care and included a representative sample of older adults with moderate to severe CKD in Ontario, Canada, where all residents aged 66 and older have universal prescription drug coverage. The results were consistent across multiple additional analyses, including when linagliptin (a DPP-4 inhibitor with a biliary excretion and a neutral effect on cardiovascular outcomes) was used as an active comparator group. By using inverse probability of treatment weighting, we were able to produce comparison groups that were balanced on a comprehensive set of baseline characteristics.
This study has some limitations. First, although we used an active comparator group and a robust statistical technique to control for confounding by indication, we are aware that residual confounding may be a factor in interpreting our results. For example, physicians who were aware of the association between sitagliptin and adverse cardiac events may have been more likely to prescribe sitagliptin at a lower dose to at-risk patients. Second, despite the use of highly accurate information on sitagliptin dispensing, the use of administrative data cannot provide information on the proportion of patients who took their pills as prescribed. This may introduce a nondifferential misclassification of the exposure, which may underestimate the true risk of death or congestive heart failure. Third, we studied only patients who were aged 66 and older, so our findings may not apply to younger patients. Fourth, because of the small sample size, we could not examine the 1-year risk of death or congestive heart failure in patients with an eGFR<30 ml/min per 1.73 m2 who started sitagliptin at a dose ≤25 mg/d (the recommended dose in these patients). Fifth, we did not censor patients when sitagliptin dose changed during follow-up. In health care databases, we have access to outpatient prescriptions only; therefore, any change in sitagliptin dose occurring during a hospital stay or emergency room visit would not be captured in the databases. However, the daily dose in each group did not change when we compared a pre- and a posthospitalization sitagliptin prescription (Supplemental Table 18), and thus, not censoring patients when their sitagliptin dose changed during follow-up is unlikely to affect our main conclusions.
In summary, we found that starting a higher than recommended dose of sitagliptin in older adults with CKD was not associated with significant harm. These findings suggest that sitagliptin dose reductions may not be needed in patients with CKD; however, our findings should be confirmed in future population-based studies using health care administrative databases in other jurisdictions.
Disclosures
K.K. Clemens received a Diabetes Canada Award sponsored in part by AstraZeneca, has attended Merck-sponsored conferences, and has received continuing medical education speaker fees from Sutherland Global Health and the Toronto Ontario Knowledge Translation Working Group Inc., outside the submitted work. V. Perkovic reports employment at The University of New South Wales, Sydney and The Royal North Shore Hospital; consultancy agreements with AbbVie, Astellas, AstraZeneca, Baxter, Bayer, Boehringer Ingelheim, Bristol-Myers Squibb, Chinook, Durect, Eli Lilly, Gilead, GlaxoSmithKline, Janssen, Merck, Mitsubishi Tanabe, Mundipharma, Novartis, Novo Nordisk, Pfizer, Pharmalink, Relypsa, Retrophin, Roche, Sanofi, Servier, and Vitae; research funding from GlaxoSmithKline and Pfizer (supplied drug and seed funding for the The Therapeutic Evaluation of Steroids in IgA Nephropathy Global [TESTING] trial); honoraria from AbbVie, Astellas, AstraZeneca, Baxter, Bayer, Boehringer Ingelheim, Bristol-Myers Squibb, Chinook, Durect, Eli Lilly, Gilead, GlaxoSmithKline, Janssen, Merck, Mitsubishi Tanabe, Mundipharma, Novartis, Novo Nordisk, Pfizer, Pharmalink, Relypsa, Retrophin, Roche, Sanofi, Servier, and Vitae; serving/served on steering committees for trials funded by AbbVie, AstraZeneca, Bayer, Boehringer Ingelheim, Chinook, Eli Lilly, Gilead, GlaxoSmithKline, Janssen, Novartis, Novo Nordisk, and Retrophin; and serving as a board director for Childrens Cancer Institute, Garvan Institute, George Clinical, George Institute, Mindgardens Network, and Victor Chang Cardiac Research Institute. M.M. Sood reports employment at The Ottawa Hospital; receiving honoraria from AstraZeneca; serving on the editorial boards for American Journal of Kidney Disease, Canadian Journal of Cardiology, and CJASN; serving as a deputy editor for Canadian Journal of Kidney Disease and Health; serving as a member of the American Society of Nephrology Highlights ESRD Team; and receiving CME speaker fees from AstraZeneca. All remaining authors have nothing to disclose.
Funding
This work was supported by Canadian Institutes of Health Research (CIHR) Award 415458 and MITACS Award IT13690.
Supplementary Material
Acknowledgments
This study was conducted by members of the ICES Kidney, Dialysis and Transplantation team at ICES Western Site. ICES is funded by an annual grant from the Ontario Ministry of Health and Long-Term Care. Core funding for ICES Western is provided by the Academic Medical Organization of Southwestern Ontario, the Schulich School of Medicine and Dentistry, Western University, and the Lawson Health Research Institute. A.X. Garg was supported by the Dr. Adam Linton Chair in Kidney Health Analytics and a Canadian Institutes of Health Research Clinician Investigator Award. M.M. Sood is supported by the Jindal Research Chair in the Prevention of Kidney Disease. Parts of this material are on the basis of data and information compiled and provided by the Canadian Institutes of Health Information (CIHI). However, the analyses, conclusions, opinions, and statements expressed herein are those of the author, and not necessarily those of CIHI. The opinions, results, and conclusions are those of the authors and are independent from the funding sources. No endorsement by ICES, the Academic Medical Organization of Southwestern Ontario, the Schulich School of Medicine and Dentistry, the Lawson Health Research Institute, the Canadian Institutes of Health Research, or the Ontario Ministry of Health and Long-Term Care is intended or should be inferred. The study sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Footnotes
Published online ahead of print. Publication date available at www.cjasn.org.
Supplemental Material
This article contains the following supplemental material online at http://cjasn.asnjournals.org/lookup/suppl/doi:10.2215/CJN.08310520/-/DCSupplemental.
Supplemental Figure 1. Flow diagram of cohort build.
Supplemental Table 1. Recommendations for sitagliptin dosing in patients with normal kidney function and in those with kidney impairment
Supplemental Table 2. Literature search.
Supplemental Table 3. Checklist of recommendations for reporting of observational studies using the Reporting of Studies Conducted Using Observational Routinely Collected Health Data guidelines.
Supplemental Table 4. Coding definitions for demographic and comorbid conditions.
Supplemental Table 5. Justification for the equation used to estimate the GFR for drug dosing adjustments.
Supplemental Table 6. Operating characteristics of hospital diagnosis codes used to define primary and secondary outcomes.
Supplemental Table 7. Variables included in the propensity score model.
Supplemental Table 8. Post hoc analysis comparing the 2-year risk of death or heart failure in older adults with moderate to severe CKD who started a new prescription for sitagliptin >50 versus ≤50 mg/d in patients with prolonged exposure.
Supplemental Table 9. Baseline characteristics of older adults with CKD newly prescribed sitagliptin in Ontario, Canada (2010–2018).
Supplemental Table 10. Dose and duration of continuous sitagliptin dispensing during the years 2010–2018.
Supplemental Table 11. Rate of sitagliptin discontinuation within 1 year after initiation of the medication.
Supplemental Table 12. Total follow-up (person-years) for sitagliptin users.
Supplemental Table 13. Risk of death or heart failure in older adults with moderate to severe CKD who started a new prescription for sitagliptin (>50 mg/d) versus linagliptin.
Supplemental Table 14. Risk of death or heart failure in older adults with moderate to severe CKD who started a new prescription for sitagliptin (≤50 mg/d) versus linagliptin.
Supplemental Table 15. Post hoc analysis comparing the risk of death or heart failure in older adults with moderate to severe CKD who started a new prescription for sitagliptin >50 versus ≤50 mg/d, censoring the follow-up time at sitagliptin discontinuation.
Supplemental Table 16. Post hoc analysis comparing the risk of death or heart failure in older adults with moderate to severe CKD who started a new prescription for sitagliptin >50 versus ≤50 mg/d when we extended the washout period to 1 year.
Supplemental Table 17. Post hoc analysis comparing the risk of death or heart failure in older adults with moderate to severe CKD who started a new prescription for sitagliptin >50 versus ≤50 mg/d after exclusion of patients who received any antidiabetic prescription on the cohort entry date.
Supplemental Table 18. Median dose of sitagliptin prescribed to patients with moderate to severe CKD who had a hospital admission for any reason within 1 year after sitagliptin initiation: A dose comparison between a pre- and a posthospitalization sitagliptin prescription.
References
- 1.Ahrén B: Use of DPP-4 inhibitors in type 2 diabetes: Focus on sitagliptin. Diabetes Metab Syndr Obes 3: 31–41, 2010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Merck Canada Inc.: Product Monograph: Januvia, 2018. Available at: https://www.merck.ca/static/pdf/JANUVIA-PM_E.pdf. Accessed December 13, 2019
- 3.UpToDate : Sitagliptin: Drug information. 2013. Available at: https://www.uptodate.com/contents/sitagliptin-drug-information?search=sitagliptin%20adult&source=panel_search_result&selectedTitle=1∼37&usage_type=panel&kp_tab=drug_general&display_rank=1#F3415869. Accessed December 13, 2019
- 4.American Diabetes Association : 9. Pharmacologic approaches to glycemic treatment: Standards of medical care in diabetes-2020. Diabetes Care 43[Suppl 1]: S98–S110, 2020. [DOI] [PubMed] [Google Scholar]
- 5.Clemens KK, Shariff S, Liu K, Hramiak I, Mahon JL, McArthur E, Garg AX: Trends in antihyperglycemic medication prescriptions and hypoglycemia in older adults: 2002-2013. PLoS One 10: e0137596, 2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Bergman AJ, Cote J, Yi B, Marbury T, Swan SK, Smith W, Gottesdiener K, Wagner J, Herman GA: Effect of renal insufficiency on the pharmacokinetics of sitagliptin, a dipeptidyl peptidase-4 inhibitor. Diabetes Care 30: 1862–1864, 2007. [DOI] [PubMed] [Google Scholar]
- 7.Verma S, Goldenberg RM, Bhatt DL, Farkouh ME, Quan A, Teoh H, Connelly KA, Leiter LA, Friedrich JO: Dipeptidyl peptidase-4 inhibitors and the risk of heart failure: A systematic review and meta-analysis. CMAJ Open 5: E152–E177, 2017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Monami M, Dicembrini I, Mannucci E: Dipeptidyl peptidase-4 inhibitors and heart failure: A meta-analysis of randomized clinical trials. Nutr Metab Cardiovasc Dis 24: 689–697, 2014. [DOI] [PubMed] [Google Scholar]
- 9.Wu S, Hopper I, Skiba M, Krum H: Dipeptidyl peptidase-4 inhibitors and cardiovascular outcomes: Meta-analysis of randomized clinical trials with 55,141 participants. Cardiovasc Ther 32: 147–158, 2014. [DOI] [PubMed] [Google Scholar]
- 10.Weir DL, McAlister FA, Senthilselvan A, Minhas-Sandhu JK, Eurich DT: Sitagliptin use in patients with diabetes and heart failure: A population-based retrospective cohort study. JACC Heart Fail 2: 573–582, 2014. [DOI] [PubMed] [Google Scholar]
- 11.Wang KL, Liu CJ, Chao TF, Huang CM, Wu CH, Chen SJ, Yeh CM, Chen TJ, Lin SJ, Chiang CE: Sitagliptin and the risk of hospitalization for heart failure: A population-based study. Int J Cardiol 177: 86–90, 2014. [DOI] [PubMed] [Google Scholar]
- 12.UpToDate : Dipeptidyl peptidase-4 (DPP-4) inhibitors for the treatment of type 2 diabetes mellitus. 2019. Available at: https://www.uptodate.com/contents/dipeptidyl-peptidase-4-dpp-4-inhibitors-for-the-treatment-of-type-2-diabetes-mellitus?search=dpp4%20inhibitors&source=search_result&selectedTitle=2∼61&usage_type=default&display_rank=1#H13200853. Accessed December 13, 2019
- 13.Hung YC, Lin CC, Huang WL, Chang MP, Chen CC: Sitagliptin and risk of heart failure hospitalization in patients with type 2 diabetes on dialysis: A population-based cohort study. Sci Rep 6: 30499, 2016. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Packer M: Worsening heart failure during the use of DPP-4 inhibitors: Pathophysiological mechanisms, clinical risks, and potential influence of concomitant antidiabetic medications. JACC Heart Fail 6: 445–451, 2018. [DOI] [PubMed] [Google Scholar]
- 15.Plantinga LC, Crews DC, Coresh J, Miller ER 3rd, Saran R, Yee J, Hedgeman E, Pavkov M, Eberhardt MS, Williams DE, Powe NR; CDC CKD Surveillance Team : Prevalence of chronic kidney disease in US adults with undiagnosed diabetes or prediabetes. Clin J Am Soc Nephrol 5: 673–682, 2010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Min T, Davies GI, Rice S, Chess J, Stephens JW: Treatment choices for the glycaemic management of patients with type 2 diabetes and chronic kidney disease: Analysis of the SAIL patient linked dataset. Diabetes Metab Syndr 12: 123–127, 2018. [DOI] [PubMed] [Google Scholar]
- 17.Statistics Canada: Population estimates on July 1st by age and sex, 2016. Available at: http://www.statcan.gc.ca/tables-tableaux/sum-som/l01/cst01/demo31a-eng.htm. Accessed March 22, 2019
- 18.Langan SM, Schmidt SA, Wing K, Ehrenstein V, Nicholls SG, Filion KB, Klungel O, Petersen I, Sorensen HT, Dixon WG, Guttmann A, Harron K, Hemkens LG, Moher D, Schneeweiss S, Smeeth L, Sturkenboom M, von Elm E, Wang SV, Benchimol EI: The reporting of studies conducted using observational routinely collected health data statement for pharmacoepidemiology (RECORD-PE). BMJ 363: k3532, 2018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Vandenbroucke JP, von Elm E, Altman DG, Gøtzsche PC, Mulrow CD, Pocock SJ, Poole C, Schlesselman JJ, Egger M; STROBE Initiative : Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): Explanation and elaboration. Int J Surg 12: 1500–1524, 2014. [DOI] [PubMed] [Google Scholar]
- 20.ICES : ICES Data. 1992. Available at: https://www.ices.on.ca/Data-and-Privacy/ICES-data. Accessed December 14, 2019
- 21.Gandhi S, Fleet JL, Bailey DG, McArthur E, Wald R, Rehman F, Garg AX: Calcium-channel blocker-clarithromycin drug interactions and acute kidney injury. JAMA 310: 2544–2553, 2013. [DOI] [PubMed] [Google Scholar]
- 22.Hwang YJ, Dixon SN, Reiss JP, Wald R, Parikh CR, Gandhi S, Shariff SZ, Pannu N, Nash DM, Rehman F, Garg AX: Atypical antipsychotic drugs and the risk for acute kidney injury and other adverse outcomes in older adults: A population-based cohort study. Ann Intern Med 161: 242–248, 2014. [DOI] [PubMed] [Google Scholar]
- 23.Patel AM, Shariff S, Bailey DG, Juurlink DN, Gandhi S, Mamdani M, Gomes T, Fleet J, Hwang YJ, Garg AX: Statin toxicity from macrolide antibiotic coprescription: A population-based cohort study. Ann Intern Med 158: 869–876, 2013. [DOI] [PubMed] [Google Scholar]
- 24.Fleet JL, Dixon SN, Kuwornu PJ, Dev VK, Montero-Odasso M, Burneo J, Garg AX: Gabapentin dose and the 30-day risk of altered mental status in older adults: A retrospective population-based study. PLoS One 13: e0193134, 2018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Muanda FT, Weir MA, Bathini L, Blake PG, Chauvin K, Dixon SN, McArthur E, Sontrop JM, Moist L, Garg AX: Association of Baclofen with encephalopathy in patients with chronic kidney disease. JAMA 322: 1987–1995, 2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Statistics Canada: Migration: Interprovincial, 2011/2012, 2015. Available at: http://www.statcan.gc.ca/pub/91-209-x/2014001/article/14012-eng.htm. Accessed December 14, 2019
- 27.Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF 3rd, Feldman HI, Kusek JW, Eggers P, Van Lente F, Greene T, Coresh J; CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration) : A new equation to estimate glomerular filtration rate [published correction appears in Ann Intern Med 155: 408, 2011]. Ann Intern Med 150: 604–612, 2009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Garg AX, Mamdani M, Juurlink DN, van Walraven C; Network of Eastern Ontario Medical Laboratories (NEO-MeL) : Identifying individuals with a reduced GFR using ambulatory laboratory database surveillance. J Am Soc Nephrol 16: 1433–1439, 2005. [DOI] [PubMed] [Google Scholar]
- 29.Scirica BM, Braunwald E, Raz I, Cavender MA, Morrow DA, Jarolim P, Udell JA, Mosenzon O, Im K, Umez-Eronini AA, Pollack PS, Hirshberg B, Frederich R, Lewis BS, McGuire DK, Davidson J, Steg PG, Bhatt DL; SAVOR-TIMI 53 Steering Committee and Investigators : Heart failure, saxagliptin, and diabetes mellitus: Observations from the SAVOR-TIMI 53 randomized trial. Circulation 130: 1579–1588, 2014. [DOI] [PubMed] [Google Scholar]
- 30.Sato T, Matsuyama Y: Marginal structural models as a tool for standardization. Epidemiology 14: 680–686, 2003. [DOI] [PubMed] [Google Scholar]
- 31.Brookhart MA, Wyss R, Layton JB, Stürmer T: Propensity score methods for confounding control in nonexperimental research. Circ Cardiovasc Qual Outcomes 6: 604–611, 2013. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Austin PC: An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivariate Behav Res 46: 399–424, 2011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Austin PC, Grootendorst P, Anderson GM: A comparison of the ability of different propensity score models to balance measured variables between treated and untreated subjects: A Monte Carlo study. Stat Med 26: 734–753, 2007. [DOI] [PubMed] [Google Scholar]
- 34.Hernán MA, Hernández-Díaz S, Werler MM, Mitchell AA: Causal knowledge as a prerequisite for confounding evaluation: An application to birth defects epidemiology. Am J Epidemiol 155: 176–184, 2002. [DOI] [PubMed] [Google Scholar]
- 35.Austin PC, Stuart EA: Moving towards best practice when using inverse probability of treatment weighting (IPTW) using the propensity score to estimate causal treatment effects in observational studies. Stat Med 34: 3661–3679, 2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Austin PC: Variance estimation when using inverse probability of treatment weighting (IPTW) with survival analysis. Stat Med 35: 5642–5655, 2016. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Green JB, Bethel MA, Armstrong PW, Buse JB, Engel SS, Garg J, Josse R, Kaufman KD, Koglin J, Korn S, Lachin JM, McGuire DK, Pencina MJ, Standl E, Stein PP, Suryawanshi S, Van de Werf F, Peterson ED, Holman RR; TECOS Study Group : Effect of sitagliptin on cardiovascular outcomes in type 2 diabetes. N Engl J Med 373: 232–242, 2015. [DOI] [PubMed] [Google Scholar]
- 38.Engel SS, Suryawanshi S, Stevens SR, Josse RG, Cornel JH, Jakuboniene N, Riefflin A, Tankova T, Wainstein J, Peterson ED, Holman RR; TECOS Study Group : Safety of sitagliptin in patients with type 2 diabetes and chronic kidney disease: Outcomes from TECOS. Diabetes Obes Metab 19: 1587–1593, 2017. [DOI] [PubMed] [Google Scholar]
- 39.Zannad F, Cannon CP, Cushman WC, Bakris GL, Menon V, Perez AT, Fleck PR, Mehta CR, Kupfer S, Wilson C, Lam H, White WB; EXAMINE Investigators : Heart failure and mortality outcomes in patients with type 2 diabetes taking alogliptin versus placebo in EXAMINE: A multicentre, randomised, double-blind trial. Lancet 385: 2067–2076, 2015. [DOI] [PubMed] [Google Scholar]
- 40.McGuire DK, Van de Werf F, Armstrong PW, Standl E, Koglin J, Green JB, Bethel MA, Cornel JH, Lopes RD, Halvorsen S, Ambrosio G, Buse JB, Josse RG, Lachin JM, Pencina MJ, Garg J, Lokhnygina Y, Holman RR, Peterson ED; Trial Evaluating Cardiovascular Outcomes With Sitagliptin (TECOS) Study Group : Association between sitagliptin use and heart failure hospitalization and related outcomes in type 2 diabetes mellitus: Secondary analysis of a randomized clinical trial. JAMA Cardiol 1: 126–135, 2016. [DOI] [PubMed] [Google Scholar]
- 41.McGuire DK, Alexander JH, Johansen OE, Perkovic V, Rosenstock J, Cooper ME, Wanner C, Kahn SE, Toto RD, Zinman B, Baanstra D, Pfarr E, Schnaidt S, Meinicke T, George JT, von Eynatten M, Marx N; CARMELINA Investigators : Linagliptin effects on heart failure and related outcomes in individuals with type 2 diabetes mellitus at high cardiovascular and renal risk in CARMELINA. Circulation 139: 351–361, 2019. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.