Abstract
BACKGROUND:
The comparative safety of sulfonylureas (SUs) in nursing home (NH) residents remains understudied despite widespread use. We compared the effects of three SU medications and initial SU doses on adverse glycemic and cardiovascular events among NH residents.
METHODS:
This national retrospective cohort study linked Medicare claims with Minimum Data Set 2.0 assessments for long-stay NH residents aged ≥65 years between January 2008 and September 2010. Exposures were the SU medication initiated (glimepiride, glipizide, or glyburide) and doses (standard or reduced). One-year outcomes were hospitalizations or emergency department visits for severe hypoglycemia, heart failure (HF), stroke, and acute myocardial infarction (AMI). After inverse probability of treatment and inverse probability of censoring by death weighting, we estimated hazard ratios (HR) using Cox regression models with robust 95% confidence intervals (CI).
RESULTS:
The cohort (N= 6,821) included 3,698 new glipizide, 1,754 glimepiride, and 1,369 glyburide users. Overall, the mean (standard deviation) age was 81.4 (8.2) years, 4,816 (70.6%) were female, and 5,164 (75.7%) were White non-Hispanic residents. The rates of severe hypoglycemia were 30.3 (95% confidence interval (CI) 22.3–40.1), 49.0 (95% CI 34.5–67.5), and 35.9 (95% CI 22.2–54.9) events per 1,000 person-years among new glipizide, glimepiride, and glyburide users, respectively (glimepiride versus glipizide HR 1.6, 95% CI 1.0–2.4, p=0.04; glyburide versus glipizide HR 1.2, 95% CI 0.7–1.9, p=0.59). The rates of severe hypoglycemia were 27.1 (95% CI 18.6–38.0) and 42.8 (95% CI 33.6–53.8) events per 1,000 person-years among new users of reduced and standard SU doses, respectively (HR 2.2, 95% CI 1.4–3.5, p<0.01). Rates of HF, stroke, and AMI were similar between medications and doses.
CONCLUSIONS:
Among long-stay NH residents, new use of glimepiride and standard SU doses resulted in higher rates of severe hypoglycemic events. Cardiovascular outcomes may not be affected by the choice of SU medication or dose.
Keywords: Sulfonylurea Compounds, Nursing Homes, Diabetes Mellitus, Frailty
INTRODUCTION
The prevalence of type 2 diabetes mellitus (T2DM) has increased over time and exceeds 30% among nursing home (NH) residents.1–3 Sulfonylureas (SUs) remain one of the most commonly prescribed glucose-lowering medication classes for NH residents despite warnings that these drugs, particularly glyburide and glimepiride, which have longer half-lives and potentially higher risks of hypoglycemia, may be unsafe in older adults.3–9 Concerns have persisted that sulfonylureas may also cause adverse cardiovascular events, and that the risk of these events may differ by individual sulfonylurea drug.10 Glyburide and glimepiride are less selective than glipizide for pancreatic beta cells, and thus are more likely to bind to receptors on cardiomyocytes and vascular smooth muscle, which could increase the risk of adverse cardiovascular events.10
NH residents represent a population with a high burden of multimorbidity that may be especially susceptible to SU adverse effects, such as hypoglycemia and related adverse outcomes (e.g., hospitalization).11–13 Unfortunately, current guidance on which SU and dose thereof to use in older NH residents is inconsistent.8,13 Studies comparing the safety of different SU medications and doses in NH residents and multimorbid older adults more broadly are scarce.14,15 Large randomized controlled trials examining medication use among NH residents are generally infeasible since it is challenging to randomize frail, cognitively impaired older adults to multiple treatments and doses, and then follow them for outcomes. High-quality observational studies that compare the safety of individual SU medications and doses are therefore needed to inform safe diabetes medication prescribing in multimorbid older adults.
To fill this knowledge gap, we applied the “target trial emulation” framework to emulate large randomized trials that would hypothetically be conducted if it were feasible to randomize heterogenous older NH residents to different SU medications and doses.16,17 We compared the effects of three SU medications and two initial SU dose categories on hospitalizations or emergency department visits for severe hypoglycemia and adverse cardiovascular events. We hypothesized that shorter-acting SUs (i.e., glipizide) and lower doses of SUs would result in decreased relative rates of adverse events.
METHODS
Study Design and Data Sources
This retrospective new-user cohort study was designed to emulate two hypothetical trials that could have been conducted (Table S1), one comparing different SU medications and one comparing different SU doses.16,17 We linked Medicare data (fee-for-service enrollment information, Part A inpatient claims, Part B carrier (outpatient) claims, and Part D prescription drug claims) to Minimum Data Set (MDS) version 2.0 assessments and Online Survey Certification and Reporting System (OSCAR) data. The MDS includes comprehensive clinical assessments of NH residents, including functional status, cognitive status, and other information, conducted at least quarterly for long-stay residents. OSCAR provided information on NH facility characteristics. This study was approved by the Brown University Institutional Review Board, which waived the requirement for informed consent.
Study Population
Our study population included long-stay NH residents (>100 days in the NH) aged ≥65 years residing in an NH between January 1, 2008 and September 30, 2010. All individuals were new users of SUs. Residents were included if they were continuously enrolled in Medicare fee-for-service (Parts A, B, and D) for 12 months prior to the date of the first eligible SU dispensing after becoming a long-stay resident. We excluded individuals who were in hospice, had cancer, were comatose or paralyzed, or had missing data on any covariate used in analyses at or within 12 months prior to the time of first eligible SU dispensing.
Exposures and Causal Contrast of Interest
The exposures of interest were 1) new use of a SU medication in the NH and 2) initial SU dose. We defined new use as the first dispensing of a SU (glimepiride, glipizide, or glyburide) after 6 months without a dispensing of a SU or another glucose-lowering treatment other than metformin.18 We classified SU dose as reduced or standard based on standardized daily dose of the first dispensing (the number of tablets multiplied by the strength in milligrams, then divided by the days supply). Reduced (i.e., “geriatric-friendly”) SU daily doses were glipizide 2.5 mg, glimepiride 1 mg, or glyburide 1.25 mg per day. Standardized daily doses above these thresholds were considered standard SU doses.
The causal contrasts of interest were the effects of initiating specific SU medications (glimepiride versus glipizide versus glyburide) and specific SU doses (standard versus reduced dose), regardless of subsequent treatment discontinuation or switching (i.e., the observational study analog of the intention-to-treat [ITT] estimand).16,17
Outcomes
Outcomes were hospitalizations or emergency department visits for severe hypoglycemia (International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes 251.0X, 251.1X, or 251.2X; algorithm positive predictive value [PPV], 89%),19 heart failure (HF) (ICD-9-CM codes 402.X1, 404.X1, 404.X3, or 428.X; PPV, 90%),20 stroke (including ischemic stroke [ICD-9-CM codes 433.X1, 434.X excluding 434.X0, or 436], intracerebral hemorrhage [ICD-9-CM code 431], and subarachnoid hemorrhage [ICD-9-CM code 430], but excluding traumatic brain injury [ICD-9-CM codes 800 to 804 and 850 to 854]; PPV, 97%),21 and acute myocardial infarction (AMI) (ICD-9-CM code 410.X; PPV, 67–97%).22–24 We defined outcomes based on Part A hospital claims and Part B emergency department claims in any coding position.
Follow-Up
Follow-up started on the day after the first SU dispensing and continued until disenrollment from Medicare fee-for-service (from Parts A, B, or D), death, an outcome (each evaluated separately), one year of follow-up, or study end (December 31, 2010), whichever occurred first. One year of follow-up was selected a priori to provide sufficient statistical power and because it is a clinically relevant time horizon for decision-making due to the high one-year mortality rate in this population.25
Baseline characteristics
We identified 203 resident and facility characteristics a priori that were expected to be variables (or proxies of variables) that would influence 1) both the probability of exposure to SUs and experiencing an outcome or 2) just the probability of experiencing an outcome.4,5,26 These characteristics were used in propensity score estimation models.27 All covariates were measured within the 12 months prior to time zero unless otherwise specified. Resident characteristics included demographics, comorbidities, cognitive and functional status, medications received in the facility, healthcare utilization, dietary factors, and social support. Facility characteristics included staffing, facility ownership, and quality indicators (e.g., percentage of residents restrained, percentage of residents with pressure sores).
Statistical Analyses
We estimated two sets of propensity scores for each patient using separate logistic regression models to predict each exposure of interest: 1) new use of a SU in the NH (glimepiride, glipizide, or glyburide; multinomial) and 2) receiving a standard versus reduced SU dose (binomial). We used the propensity scores to construct inverse probability of treatment weights (IPTW) to correct for potential confounding. We then used the same approach to predict the probabilities of death and construct inverse probability of censoring by death weights (IPCW) to correct for potential differential loss to follow up from death. We calculated final weights by multiplying the IPTW and IPCW (separately for drug and dose analyses) for each person.
To compare rates of severe hypoglycemia, HF, stroke, and AMI among new users of SUs by SU medication and dose, we used Cox regression models with robust standard errors to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) before and after weighting.
Software
We used SAS, version 9.4 (SAS Institute Inc., Cary, NC, USA) for data processing, and Stata version 14.0 (Stata Corp., College Station, TX, USA) for data analyses.
RESULTS
Study Cohort
The study cohort (N= 6,821) included 3,698 new glipizide, 1,754 glimepiride, and 1,369 glyburide users (Supplementary Figure S1). Of all SU users, 3,904 (57.2%) received a standard and 2,917 (42.8%) a reduced initial dose. Standard doses were used by 1,474 (39.9%) glipizide, 1,192 (68.0%) glimepiride, and 1,238 (90.4%) glyburide recipients. Of the 3,698 glipizide users, 888 (24%) received an extended-release formulation.
At baseline, overall, the mean (standard deviation (SD)) age was 81.4 (8.2) years, 4,816 (70.6%) were female residents, 5,164 (75.7%) were White non-Hispanic residents, 1,992 (29.2%) were prescribed metformin prior to SU initiation, and the median (interquartile range) length of NH stay prior to SU initiation was 589 (273–1,181) days (Table 1). Before weighting, few characteristics differed between sulfonylurea medication and initial dose groups based on standardized mean differences (SMDs) of 0.10 or greater. Notable differences across SU groups at baseline included race/ethnicity (non-Hispanic White 73.1% glipizide, 78.5% glimepiride, 79.3% glyburide), mean (SD) number of medications (12.6 (4.7) glipizide, 13.3 (4.9) glimepiride, 12.3 (4.8) glyburide), and proportion of residents taking angiotensin II receptor blockers (10.4% glipizide, 14.1% glimepiride, 9.1% glyburide) (Table 1, Supplementary Table S2). Differences across initial dose groups at baseline included age (mean (SD) 82.4 (8.0) years for reduced dose, 80.7 (8.2) years for standard dose) and proportion of residents taking beta blockers (47.7% reduced dose, 42.8% standard dose) (Table 2, Supplementary Table S3). However, characteristics were well-balanced across groups after weighting (Supplementary Tables S2–S4).
Table 1.
Demographic and Other Characteristics of Glipizide, Glimepiride, and Glyburide New Users at Baseline.
| Characteristic | n (%)a | |||
|---|---|---|---|---|
| All Sulfonylureas (N=6,821) | Glipizide (n=3,698) |
Glimepiride (n=1,754) |
Glyburide (n=1,369) |
|
| Age, mean (SD), years | 81.4 (8.2) | 81.4 (8.1) | 81.3 (8.2) | 81.4 (8.2) |
| Female sex | 4,816 (70.6) | 2,586 (69.9) | 1,278 (72.9) | 952 (69.5) |
| Race/ethnicity | ||||
| White | 5,164 (75.7) | 2,703 (73.1) | 1,376 (78.5) | 1,085 (79.3) |
| Black | 1,066 (15.6) | 644 (17.4) | 261 (14.9) | 161 (11.8) |
| Hispanic | 402 (5.9) | 243 (6.6) | 78 (4.5) | 81 (5.9) |
| Other | 189 (2.8) | 108 (2.9) | 39 (2.2) | 42 (3.1) |
| Number of MDS comorbidities, mean (SD) | 5.7 (2.7) | 5.7 (2.7) | 5.7 (2.8) | 5.5 (2.7) |
| Chronic conditions | ||||
| Ischemic heart disease | 1,039 (15.2) | 549 (14.9) | 279 (15.9) | 211 (15.4) |
| Heart failure | 1,934 (28.4) | 1,075 (29.1) | 480 (27.4) | 379 (27.7) |
| Hypertension | 5,261 (77.1) | 2,867 (77.5) | 1,352 (77.1) | 1,042 (76.1) |
| ADRD | 3,958 (58.0) | 2,132 (57.7) | 996 (56.8) | 830 (60.6) |
| Stroke | 1,668 (24.5) | 924 (25.0) | 437 (24.9) | 307 (22.4) |
| COPD | 1,365 (20.0) | 757 (20.5) | 341 (19.4) | 267 (19.5) |
| Renal failure | 639 (9.4) | 360 (9.7) | 167 (9.5) | 112 (8.2) |
| Diabetic retinopathy | 132 (1.9) | 75 (2.0) | 34 (1.9) | 23 (1.7) |
| ADL score, mean (SD)b | 15.8 (6.7) | 15.9 (7.6) | 15.7 (7.7) | 15.6 (7.9) |
| Cognitive performance score, mean (SD)c | 2.6 (1.6) | 2.6 (1.6) | 2.5 (1.6) | 2.6 (1.6) |
| Health instability score, mean (SD)d | 0.5 (0.8) | 0.5 (0.7) | 0.5 (0.8) | 0.4 (0.7) |
| Number of medications, mean (SD) | 12.7 (4.8) | 12.6 (4.7) | 13.3 (4.9) | 12.3 (4.8) |
| Medication use | ||||
| Metformin | 1,992 (29.2) | 1,048 (28.3) | 542 (30.9) | 402 (29.4) |
| ACE inhibitors | 2,578 (37.8) | 1,427 (38.6) | 657 (37.5) | 494 (36.1) |
| Alpha blocker | 216 (3.2) | 123 (3.3) | 48 (2.7) | 45 (3.3) |
| ARBs | 758 (11.1) | 386 (10.4) | 247 (14.1) | 125 (9.1) |
| Beta blocker | 3,063 (44.9) | 1,683 (45.5) | 808 (46.1) | 572 (41.8) |
| Calcium channel blockers | 2,027 (29.7) | 1,099 (29.7) | 549 (31.3) | 379 (27.7) |
| Clopidogrel | 959 (14.1) | 517 (14.0) | 262 (14.9) | 180 (13.2) |
| Steroids (oral) | 857 (12.6) | 475 (12.8) | 224 (12.8) | 158 (11.5) |
| Statins | 2,680 (39.3) | 1,456 (39.4) | 728 (41.5) | 496 (36.2) |
| Warfarin | 1,015 (14.9) | 449 (15.1) | 260 (14.8) | 196 (14.3) |
| Length of nursing home stay before treatment initiation, median (IQR) days | 589 (273–1181) | 580 (266–1152) | 618.5 (298–1232) | 588 (262–1174) |
| Number of overnight hospitalizations in prior 90 days, mean (SD) | 0.4 (0.6) | 0.4 (0.6) | 0.3 (0.6) | 0.4 (0.6) |
| Number of ED visits in prior 90 days, mean (SD) | 0.1 (0.4) | 0.1 (0.3) | 0.1 (0.4) | 0.1 (0.4) |
| Number of physician order changes in prior two weeks, mean (SD) | 1.9 (1.8) | 1.9 (1.8) | 1.9 (1.8) | 1.8 (1.8) |
| Nursing Home Facility Characteristics | ||||
| Staff hours per resident, mean (SD) | 2.9 (1.3) | 2.9 (1.3) | 2.8 (1.3) | 2.9 (1.3) |
| Ownership | ||||
| Profit | 4,835 (70.9) | 2,565 (69.4) | 1,260 (71.8) | 1,010 (73.8) |
| Non-profit | 1,514 (22.2) | 853 (23.1) | 388 (22.1) | 273 (19.9) |
| Government | 472 (6.9) | 280 (7.6) | 106 (6.0) | 86 (6.3) |
Abbreviations: SD, standard deviation; MDS, Minimum Data Set; ADRD, Alzheimer’s Disease and Related Dementias; COPD, chronic obstructive pulmonary disease; ADL, activities of daily living; ACE, angiotensin converting enzyme inhibitor; ARB, angiotensin receptor blocker; IQR, interquartile range; ED, emergency department.
Reported as number (%) unless otherwise stated.
ADL status was measured by the Minimum Data Set Morris ADL score, which ranged from 0 (independent) to 28 (extensive dependency).
Cognitive impairment measured using the Minimum Data Set Cognitive Performance Scale, which ranged from 0 (intact) to 6 (very severe impairment).
Health instability measured using the Minimum Data Set Changes in Health, End-stage disease and Symptoms and Signs score, which ranged from 0 (not at all unstable) to 5 (highly unstable).
Table 2.
Demographic and Other Characteristics of Standard versus Reduced Initial Sulfonylurea Dose Recipients at Baseline.
| n (%)a | ||
|---|---|---|
| Characteristic | Reduced Dose (n= 2,917) |
Standard Dose (n= 3,904) |
| Age, mean (SD) years | 82.4 (8.0) | 80.7 (8.2) |
| Female sex | 2,100 (72.0) | 2,716 (69.6) |
| Race/ethnicity | ||
| White | 2,191 (75.1) | 2,973 (76.2) |
| Black | 473 (16.2) | 593 (15.2) |
| Hispanic | 165 (5.7) | 237 (6.1) |
| Other | 88 (3.0) | 101 (2.6) |
| Number of comorbidities, mean (SD) | 5.8 (2.7) | 5.6 (2.8) |
| Chronic conditions | ||
| Ischemic heart disease | 422 (14.5) | 617 (15.8) |
| Heart failure | 843 (28.9) | 1,091 (27.9) |
| Hypertension | 2,243 (76.9) | 3,018 (77.3) |
| ADRD | 1,737 (59.5) | 2,221 (56.9) |
| Stroke | 698 (23.9) | 970 (24.8) |
| COPD | 582 (20.0) | 783 (20.1) |
| Renal failure | 293 (10.0) | 346 (8.9) |
| Diabetic retinopathy | 42 (1.4) | 90 (2.3) |
| ADL score, mean (SD)b | 16.2 (7.5) | 15.5 (7.8) |
| Cognitive performance score (CPS), mean (SD)c | 2.6 (1.6) | 2.5 (1.6) |
| Health instability score, mean (SD)d | 0.5 (0.8) | 0.5 (0.7) |
| Number of medications, mean (SD) | 12.6 (4.8) | 12.8 (4.8) |
| Medication use | ||
| Metformin | 797 (27.3) | 1,195 (30.6) |
| ACE inhibitors | 1,132 (38.8) | 1,446 (37.0) |
| Alpha blocker | 112 (3.8) | 104 (2.7) |
| ARBs | 319 (10.9) | 439 (11.2) |
| Beta blocker | 1,392 (47.7) | 1,671 (42.8) |
| Calcium channel blockers | 909 (31.2) | 1,118 (28.6) |
| Clopidogrel | 401 (13.7) | 558 (14.3) |
| Steroids (oral) | 393 (13.5) | 464 (11.9) |
| Statins | 1,126 (38.6) | 1,554 (39.8) |
| Warfarin | 454 (15.6) | 561 (14.4) |
| Length of nursing home stay before treatment initiation, median (IQR) days | 602 (282, 1188) | 583 (266, 1180) |
| Number of overnight hospitalizations in prior 90 days, mean (SD) | 0.2 (0.5) | 0.2 (0.6) |
| Number of ED visits in prior 90 days, mean (SD) | 0.1 (0.3) | 0.1 (0.3) |
| Number of physician order changes in prior two weeks, mean (SD) | 1.8 (1.8) | 1.8 (1.8) |
| Nursing Home Facility Characteristics | ||
| Staff hours per resident, mean (SD) | 3.5 (0.9) | 3.5 (0.9) |
| Ownership | ||
| Profit | 2,003 (68.7) | 2,832 (72.5) |
| Non-profit | 696 (23.9) | 818 (21.0) |
| Government | 218 (7.5) | 254 (6.5) |
Abbreviations: SD, standard deviation; ADRD, Alzheimer’s Disease and Related Dementias; COPD, chronic obstructive pulmonary disease; ADL, activities of daily living; ACE, angiotensin converting enzyme inhibitor; ARB, angiotensin receptor blocker; IQR, interquartile range; ED, emergency department. Note: Reduced doses included glipizide 2.5 mg, glimepiride 1 mg, or glyburide 1.25 mg. Standardized daily doses above these thresholds were considered standard doses.
Reported as number (%) unless otherwise stated.
ADL status was measured by the Minimum Data Set Morris ADL score, which ranged from 0 (independent) to 28 (extensive dependency).
Cognitive impairment measured using the Minimum Data Set Cognitive Performance Scale, which ranged from 0 (intact) to 6 (very severe impairment).
Health instability measured using the Minimum Data Set Changes in Health, End-stage disease and Symptoms and Signs score, which ranged from 0 (not at all unstable) to 5 (highly unstable).
Comparative Safety of SU Medications and Doses
Overall, there were 106 severe hypoglycemia events, 405 HF events, 64 stroke events, 76 AMI events, and 2,990 residents died during the one-year follow-up (Supplementary Table S5 reports mortality across SU medication and initial dose groups). The rates of severe hypoglycemia were 30.3 (95% confidence interval (CI) 22.3–40.1), 49.0 (95% CI 34.5–67.5), and 35.9 (95% CI 22.2–54.9) events per 1,000 person-years among new glipizide, glimepiride, and glyburide users, respectively (Supplementary Table S6). New glimepiride users had a 60% higher rate of severe hypoglycemic events versus new glipizide users (unweighted HR 1.6, 95% CI 1.1–2.5, p= 0.03; weighted HR 1.6, 95% CI 1.0–2.4, p= 0.04), but rates were similar between new glyburide versus glipizide users (unweighted HR 1.2, 95% CI 0.7–2.1, p= 0.44; weighed HR 1.2, 95% CI 0.7–1.9, p= 0.59) (Figure 1, Supplementary Table S6). Rates of HF, stroke, and AMI were statistically similar between new glimepiride versus glipizide and glyburide versus glipizide users.
Figure 1. Comparative Safety of New Glipizide, Glimepiride, and Glyburide Use among Nursing Home Residents.
Presents hazard ratios with robust 95% confidence intervals before and after inverse probably of treatment and censoring weighting based on 203 covariates (see Supplementary Table S4). New use of glipizide was the reference group. Abbreviations: AMI, acute myocardial infarction; HF, heart failure; HR, hazard ratio; IPW, inverse probably of treatment and censoring weighting.
The rates of severe hypoglycemia were 27.1 (95% CI 18.6–38.0) and 42.8 (95% CI 33.6–53.8) events per 1,000 person-years among new users of reduced and standard SU doses, respectively. Residents receiving standard doses of SUs thus had a higher rate of severe hypoglycemic events compared to residents receiving reduced doses (unweighted HR 1.6, 95% CI 1.0–2.4, p= 0.03; weighted HR 2.2, 95% CI 1.4–3.5, p<0.01) (Figure 2, Supplementary Table S7). The rates of cardiovascular outcomes were similar between new users of reduced and standard SU doses.
Figure 2. Comparative Safety of Initial Standard versus Reduced Sulfonylurea Dose among Nursing Home Residents.
Presents hazard ratios with robust 95% confidence intervals before and after inverse probably of treatment and censoring weighting based on 203 covariates (see Supplementary Table S5). Initial reduced dose was the reference group. Abbreviations: AMI, acute myocardial infarction; HF, heart failure; HR, hazard ratio; IPW, inverse probably of treatment and censoring weighting.
DISCUSSION
In this large national retrospective cohort study of 6,821 NH residents, we observed a higher rate of severe hypoglycemia events that required a hospitalization or emergency department visit among new users of glimepiride (versus glipizide) and among new users of standard (versus reduced) doses of SUs over one year of follow-up. We did not observe differences in the rates of HF, stroke, or AMI events across SU medication or dose groups. Since SUs are among the most commonly prescribed medications for T2DM in NHs,3–7 we believe that our findings should inform the selection of initial SU regimens to avoid adverse outcomes, particularly severe hypoglycemia, in older NH residents.
Leading geriatric and endocrine societies offer conflicting guidance on the choice of medication and dose when a SU is prescribed in older adults. The American Geriatrics Society Beers Criteria for Potentially Inappropriate Medication Use in Older Adults recommends avoiding the use of long-acting SUs such as glyburide and glimepiride among older adults,8 while guidance from the American Diabetes Association on treatment of T2DM in NHs recommends use of glimepiride or glipizide over glyburide.12,13 Since no randomized controlled trials have compared SUs among NH residents and future trials are unlikely to be conducted, our study used the target trial emulation framework to produce the first estimates of the effects of initiating different SUs and doses on safety outcomes to inform treatment decisions.
Avoiding hypoglycemia and other adverse effects may be one of the most important goals in treating T2DM in older NH residents.12,13 Our results suggest that glimepiride has a higher rate of severe hypoglycemia compared with glipizide. Although we found that the rate of severe hypoglycemia was statistically similar between glimepiride and glyburide, prior evidence has reported a higher risk of hypoglycemia with glyburide versus other SUs.8,28–30 Taken together, our results complement other guidance and suggest that glipizide is the safest first-line SU treatment to avoid severe hypoglycemia in NH residents.
Prior literature examining the association between different SU medications on cardiovascular events in older adults with multimorbidity or frailty is limited. A majority of studies compared SUs to other glucose-lowering medication classes.31–36 A systematic review and network meta-analysis including both new and prevalent users of SUs reported that SUs had a similar risk of cardiovascular-related mortality when compared to glibenclamide (glyburide), except for gliclazide (not available in U.S.), which had a significant lower risk.37 Several more recent observational studies found no intraclass differences in the risk of AMI, ischemic stroke, cardiovascular death, or hospitalization for HF.10,38,39 However, none of these studies were specific to NH residents or populations with a high burden of multimorbidity. We found no difference in the cardiovascular safety of newly initiated SUs, thus extending prior results to NH residents.
Literature comparing the safety of different initial SU doses is particularly scarcer. One study reported higher standardized rates (95% CIs) of hypoglycemia per 1,000 person-years for high dose glyburide (low dose 64.5 (60.8–68.2), high dose 74.3 (67.1–81.4)), glimepiride (low dose 50.3 (45.2–55.4), high dose 58.8 (50.3–67.2)), and glipizide (low dose 42.5 (39.6–45.4), high dose 59.4 (55.1–63.8)), though the CIs overlapped for glyburide and glimepiride.30 Another study found a higher risk of major adverse cardiovascular events with high versus low dose glyburide but not for high versus low dose gliclazide.40 A third study reported a greater risk of incident HF among individuals receiving high dose versus low dose SUs (96% received glyburide).41 Here, we found that the rate of severe hypoglycemia was higher among residents receiving standard versus reduced SU doses but the rates of cardiovascular events were similar across initial SU dose groups. Therefore, our findings support the use of reduced SU doses to avoid severe hypoglycemia in NH residents, but suggest that cardiovascular outcomes are not affected by the choice of SU medication or dose.
Limitations
Our results must be interpreted in light of several limitations. First, due to the nature of our claims-based outcome measures, we only ascertained the most severe hypoglycemia events that required a hospitalization or ED visit and did not capture less severe occurrences that could be managed in the NH. Second, although our data are relatively older, the biological effects of the drugs should not differ over time and SUs are still commonly used, so our results should remain highly generalizable to modern clinical practice. Third, some glipizide users (24%) received an extended release formulation, which may have attenuated estimates to the null if duration of action is the primary mechanism through which harms occur differentially by SU drug. Finally, the distribution of SU doses differed across treatments. Due to the rarity of the outcome events and insufficient sample size, we were unable to estimate the joint effects (i.e., interaction) of drug and dose, but future studies with a larger sample size should implement such an approach.42
Summary
Overall, new use of glimepiride and standard SU doses resulted in higher rates of severe hypoglycemic events among older NH residents. Adverse cardiovascular events were similar across SU medications and doses. Based on our results and available guidance on managing T2DM in older NH residents, when initiating an SU, reduced-dose glipizide (≤2.5 mg daily) may be the safest treatment to avoid severe hypoglycemia.
Supplementary Material
Supplementary Figure S1. Study Cohort Flow Diagram.
Supplementary Table S1. Protocol of the Hypothetical Target Trial Emulated to Estimate the Effects of Sulfonylurea Drugs and Starting Doses on Safety Outcomes among Older Nursing Home Residents.
Supplementary Table S2. Standardized Mean Differences Before and After Inverse Probability of Treatment Weighting by Sulfonylurea Type.
Supplementary Table S3. Standardized Mean Differences Before and After Inverse Probability of Treatment Weighting by Sulfonylurea Dose.
Supplementary Table S4. The Distribution of the Stabilized Inverse Probability of Treatment and Censoring by Death Weights.
Supplementary Table S5. Mortality by Sulfonylurea Medication and Initial Dose Groups.
Supplementary Table S6. Outcome Events, Person-Years, Rates, and Hazard Ratios for Glipizide, Glimepiride, and Glyburide New Users.
Supplementary Table S7. Outcome Events, Person-Years, Rates, and Hazard Ratios Stratified by Standard versus Reduced Initial Sulfonylurea Dose.
Key Points:
New use of glimepiride resulted in a 60% relative increase in the rate of severe hypoglycemia versus glipizide among older nursing home residents, yet the rate of severe hypoglycemia was similar between new glyburide versus glipizide users.
Standard doses versus reduced doses of sulfonylureas resulted in a more than 2-fold greater rate of severe hypoglycemia.
Rates of hospitalization or emergency department visits for heart failure, stroke, or acute myocardial infarction were similar between sulfonylurea medications, regardless of dose.
Why does this matter?
When prescribing a sulfonylurea, reduced-dose glipizide may be the safest option to avoid severe hypoglycemia for long-stay NH residents. Adverse cardiovascular outcomes may not be relevant to the prescribing decision.
ACKNOWLEDGEMENTS
Sponsors’ Role:
The sponsors had no role in the design, methods, subject recruitment, analysis, and preparation of the paper or any other aspect of the work.
Other Acknowledgements:
Dr. Zullo is a U.S. Government employee; the views expressed in this article are those of the author and do not necessarily reflect the position or policy of the Department of Veterans Affairs or the United States Government.
Funding Sources:
This study was supported by grants RF1AG061221, R01AG065722, R21AG061632, and R01AG077620 from the National Institute on Aging (NIA). The funders 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
Conflicts of Interest: Dr. Berry previously received grant money from Amgen unrelated to the current project. All other authors have no relevant conflicts of interest to report.
REFERENCES
- 1.Zhang X, Decker FH, Luo H, et al. Trends in the prevalence and comorbidities of diabetes mellitus in nursing home residents in the United States: 1995–2004. J Am Geriatr Soc. Apr 2010;58(4):724–30. doi: 10.1111/j.1532-5415.2010.02786.x [DOI] [PubMed] [Google Scholar]
- 2.Dybicz SB, Thompson S, Molotsky S, Stuart B. Prevalence of diabetes and the burden of comorbid conditions among elderly nursing home residents. Am J Geriatr Pharmacother. Aug 2011;9(4):212–23. doi: 10.1016/j.amjopharm.2011.05.001 [DOI] [PubMed] [Google Scholar]
- 3.Zarowitz B, Allen C, O’Shea T, Dalal MR, Haumschild M, DiGenio A. Type 2 diabetes mellitus treatment patterns in U.S. nursing home residents. Postgrad Med. Jun 2015;127(5):429–37. doi: 10.1080/00325481.2015.1035621 [DOI] [PubMed] [Google Scholar]
- 4.Zullo AR, Dore DD, Gutman R, Mor V, Smith RJ. National Glucose-Lowering Treatment Complexity Is Greater in Nursing Home Residents than Community-Dwelling Adults. J Am Geriatr Soc. Nov 2016;64(11):e233–e235. doi: 10.1111/jgs.14485 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Zullo AR, Dore DD, Daiello L, et al. National Trends in Treatment Initiation for Nursing Home Residents With Diabetes Mellitus, 2008 to 2010. J Am Med Dir Assoc. Jul 1 2016;17(7):602–8. doi: 10.1016/j.jamda.2016.02.023 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Lee SJ, Stijacic-Cenzer I, Barnhart C, McClymont K, Steinman MA. Changing Patterns of Glucose-Lowering Medication Use in VA Nursing Home Residents With Diabetes, 2005 to 2011. J Am Med Dir Assoc. Oct 1 2015;16(10):898 e9–14. doi: 10.1016/j.jamda.2015.06.020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.McNabney MK, Pandya N, Iwuagwu C, et al. Differences in diabetes management of nursing home patients based on functional and cognitive status. J Am Med Dir Assoc. Nov-Dec 2005;6(6):375–82. doi: 10.1016/j.jamda.2005.05.012 [DOI] [PubMed] [Google Scholar]
- 8.By the American Geriatrics Society Beers Criteria Update Expert P. American Geriatrics Society 2019 Updated AGS Beers Criteria(R) for Potentially Inappropriate Medication Use in Older Adults. J Am Geriatr Soc. Apr 2019;67(4):674–694. doi: 10.1111/jgs.15767 [DOI] [PubMed] [Google Scholar]
- 9.Riester MR, Goyal P, Steinman MA, et al. Prevalence of Potentially Inappropriate Medication Prescribing in US Nursing Homes, 2013–2017. Journal of general internal medicine. Sep 29 2022;doi: 10.1007/s11606-022-07825-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Douros A, Yin H, Yu OHY, Filion KB, Azoulay L, Suissa S. Pharmacologic Differences of Sulfonylureas and the Risk of Adverse Cardiovascular and Hypoglycemic Events. Diabetes care. Nov 2017;40(11):1506–1513. doi: 10.2337/dc17-0595 [DOI] [PubMed] [Google Scholar]
- 11.Lapane KL, Jesdale BM, Dube CE, Pimentel CB, Rajpathak SN. Sulfonylureas and risk of falls and fractures among nursing home residents with type 2 diabetes mellitus. Diabetes Res Clin Pract. Aug 2015;109(2):411–9. doi: 10.1016/j.diabres.2015.05.009 [DOI] [PubMed] [Google Scholar]
- 12.Munshi MN, Florez H, Huang ES, et al. Management of Diabetes in Long-term Care and Skilled Nursing Facilities: A Position Statement of the American Diabetes Association. Diabetes Care. Feb 2016;39(2):308–18. doi: 10.2337/dc15-2512 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Pandya N, Hames E, Sandhu S. Challenges and Strategies for Managing Diabetes in the Elderly in Long-Term Care Settings. Diabetes Spectr. Aug 2020;33(3):236–245. doi: 10.2337/ds20-0018 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Zullo AR, Duprey MS, Smith RJ, et al. Effects of dipeptidyl peptidase-4 inhibitors and sulphonylureas on cognitive and physical function in nursing home residents. Diabetes, obesity & metabolism. Feb 2022;24(2):247–256. doi: 10.1111/dom.14573 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Zullo AR, Smith RJ, Gutman R, et al. Comparative safety of dipeptidyl peptidase-4 inhibitors and sulfonylureas among frail older adults. Journal of the American Geriatrics Society. Jul 21 2021;doi: 10.1111/jgs.17371 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Huitfeldt A, Hernan MA, Kalager M, Robins JM. Comparative Effectiveness Research Using Observational Data: Active Comparators to Emulate Target Trials with Inactive Comparators. EGEMS (Wash DC). 2016;4(1):1234. doi: 10.13063/2327-9214.1234 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Hernan MA, Robins JM. Using Big Data to Emulate a Target Trial When a Randomized Trial Is Not Available. American journal of epidemiology. Apr 15 2016;183(8):758–64. doi: 10.1093/aje/kwv254 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Ray WA. Evaluating medication effects outside of clinical trials: new-user designs. American journal of epidemiology. Nov 1 2003;158(9):915–20. [DOI] [PubMed] [Google Scholar]
- 19.Ginde AA, Blanc PG, Lieberman RM, Camargo CA Jr. Validation of ICD-9-CM coding algorithm for improved identification of hypoglycemia visits. BMC Endocr Disord. Apr 1 2008;8:4. doi: 10.1186/1472-6823-8-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Saczynski JS, Andrade SE, Harrold LR, et al. A systematic review of validated methods for identifying heart failure using administrative data. Pharmacoepidemiol Drug Saf. Jan 2012;21 Suppl 1:129–40. doi: 10.1002/pds.2313 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Roumie CL, Mitchel E, Gideon PS, Varas-Lorenzo C, Castellsague J, Griffin MR. Validation of ICD-9 codes with a high positive predictive value for incident strokes resulting in hospitalization using Medicaid health data. Pharmacoepidemiol Drug Saf. Jan 2008;17(1):20–6. doi: 10.1002/pds.1518 [DOI] [PubMed] [Google Scholar]
- 22.Choma NN, Griffin MR, Huang RL, et al. An algorithm to identify incident myocardial infarction using Medicaid data. Pharmacoepidemiol Drug Saf. Nov 2009;18(11):1064–71. doi: 10.1002/pds.1821 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Rosamond WD, Chambless LE, Sorlie PD, et al. Trends in the sensitivity, positive predictive value, false-positive rate, and comparability ratio of hospital discharge diagnosis codes for acute myocardial infarction in four US communities, 1987–2000. American journal of epidemiology. Dec 15 2004;160(12):1137–46. doi: 10.1093/aje/ [DOI] [PubMed] [Google Scholar]
- 24.Petersen LA, Wright S, Normand SL, Daley J. Positive predictive value of the diagnosis of acute myocardial infarction in an administrative database. J Gen Intern Med. Sep 1999;14(9):555–8. doi: 10.1046/j.1525-1497.1999.10198.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Ogarek A. J, Mccreedy M. E, Thomas K S, Teno M. J, Gozalo L. P. Minimum Data Set Changes in Health, End-Stage Disease and Symptoms and Signs Scale: A Revised Measure to Predict Mortality in Nursing Home Residents. Journal of the American Geriatrics Society. 2018-05-01 2018;66(5):976–981. doi: 10.1111/jgs.15305 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Zullo AR, Dore DD, Gutman R, Mor V, Alvarez CA, Smith RJ. Metformin Safety Warnings and Diabetes Drug Prescribing Patterns for Older Nursing Home Residents. J Am Med Dir Assoc. Oct 1 2017;18(10):879–884 e7. doi: 10.1016/j.jamda.2017.05.020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Brookhart MA, Schneeweiss S, Rothman KJ, Glynn RJ, Avorn J, Sturmer T. Variable selection for propensity score models. American journal of epidemiology. Jun 15 2006;163(12):1149–56. doi:kwj149 [pii] 10.1093/aje/kwj149 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.van Staa T, Abenhaim L, Monette J. Rates of hypoglycemia in users of sulfonylureas. J Clin Epidemiol. Jun 1997;50(6):735–41. doi: 10.1016/s0895-4356(97)00024-3 [DOI] [PubMed] [Google Scholar]
- 29.Shorr RI, Ray WA, Daugherty JR, Griffin MR. Individual sulfonylureas and serious hypoglycemia in older people. J Am Geriatr Soc. Jul 1996;44(7):751–5. doi: 10.1111/j.1532-5415.1996.tb03729.x [DOI] [PubMed] [Google Scholar]
- 30.Leonard CE, Han X, Brensinger CM, et al. Comparative risk of serious hypoglycemia with oral antidiabetic monotherapy: A retrospective cohort study. Pharmacoepidemiol Drug Saf. Jan 2018;27(1):9–18. doi: 10.1002/pds.4337 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Zullo AR, Smith RJ, Gutman R, et al. Comparative safety of dipeptidyl peptidase-4 inhibitors and sulfonylureas among frail older adults. J Am Geriatr Soc. Oct 2021;69(10):2923–2930. doi: 10.1111/jgs.17371 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Varas-Lorenzo C, Margulis AV, Pladevall M, et al. The risk of heart failure associated with the use of noninsulin blood glucose-lowering drugs: systematic review and meta-analysis of published observational studies. BMC Cardiovasc Disord. Sep 26 2014;14:129. doi: 10.1186/1471-2261-14-129 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Pladevall M, Riera-Guardia N, Margulis AV, Varas-Lorenzo C, Calingaert B, Perez-Gutthann S. Cardiovascular risk associated with the use of glitazones, metformin and sufonylureas: meta-analysis of published observational studies. BMC Cardiovasc Disord. Jan 15 2016;16:14. doi: 10.1186/s12872-016-0187-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Monami M, Genovese S, Mannucci E. Cardiovascular safety of sulfonylureas: a meta-analysis of randomized clinical trials. Diabetes Obes Metab. Oct 2013;15(10):938–53. doi: 10.1111/dom.12116 [DOI] [PubMed] [Google Scholar]
- 35.Phung OJ, Schwartzman E, Allen RW, Engel SS, Rajpathak SN. Sulphonylureas and risk of cardiovascular disease: systematic review and meta-analysis. Diabet Med. Oct 2013;30(10):1160–71. doi: 10.1111/dme.12232 [DOI] [PubMed] [Google Scholar]
- 36.Du L, Chen Y, Shen J, Peng Y, Sun X. Cardiovascular safety of sulphonylurea: comment on the study by Monami et al et al. Diabetes Obes Metab. Jul 2014;16(7):667–9. doi: 10.1111/dom.12245 [DOI] [PubMed] [Google Scholar]
- 37.Simpson SH, Lee J, Choi S, Vandermeer B, Abdelmoneim AS, Featherstone TR. Mortality risk among sulfonylureas: a systematic review and network meta-analysis. Lancet Diabetes Endocrinol. Jan 2015;3(1):43–51. doi: 10.1016/S2213-8587(14)70213-X [DOI] [PubMed] [Google Scholar]
- 38.Fadini GP, Saragoni S, Russo P, et al. Intraclass differences in the risk of hospitalization for heart failure among patients with type 2 diabetes initiating a dipeptidyl peptidase-4 inhibitor or a sulphonylurea: Results from the OsMed Health-DB registry. Diabetes Obes Metab. Oct 2017;19(10):1416–1424. doi: 10.1111/dom.12979 [DOI] [PubMed] [Google Scholar]
- 39.van Dalem J, Brouwers M, Stehouwer CDA, et al. Risk of a first-ever acute myocardial infarction and all-cause mortality with sulphonylurea treatment: A population-based cohort study. Diabetes Obes Metab. Apr 2018;20(4):1056–1060. doi: 10.1111/dom.13168 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Abdelmoneim AS, Eurich DT, Senthilselvan A, Qiu W, Simpson SH. Dose-response relationship between sulfonylureas and major adverse cardiovascular events in elderly patients with type 2 diabetes. Pharmacoepidemiol Drug Saf. Oct 2016;25(10):1186–1195. doi: 10.1002/pds.4014 [DOI] [PubMed] [Google Scholar]
- 41.McAlister FA, Eurich DT, Majumdar SR, Johnson JA. The risk of heart failure in patients with type 2 diabetes treated with oral agent monotherapy. Eur J Heart Fail. Jul 2008;10(7):703–8. doi: 10.1016/j.ejheart.2008.05.013 [DOI] [PubMed] [Google Scholar]
- 42.VanderWeele TJ, Tchetgen Tchetgen EJ. Attributing effects to interactions. Epidemiology. Sep 2014;25(5):711–22. doi: 10.1097/EDE.0000000000000096 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Figure S1. Study Cohort Flow Diagram.
Supplementary Table S1. Protocol of the Hypothetical Target Trial Emulated to Estimate the Effects of Sulfonylurea Drugs and Starting Doses on Safety Outcomes among Older Nursing Home Residents.
Supplementary Table S2. Standardized Mean Differences Before and After Inverse Probability of Treatment Weighting by Sulfonylurea Type.
Supplementary Table S3. Standardized Mean Differences Before and After Inverse Probability of Treatment Weighting by Sulfonylurea Dose.
Supplementary Table S4. The Distribution of the Stabilized Inverse Probability of Treatment and Censoring by Death Weights.
Supplementary Table S5. Mortality by Sulfonylurea Medication and Initial Dose Groups.
Supplementary Table S6. Outcome Events, Person-Years, Rates, and Hazard Ratios for Glipizide, Glimepiride, and Glyburide New Users.
Supplementary Table S7. Outcome Events, Person-Years, Rates, and Hazard Ratios Stratified by Standard versus Reduced Initial Sulfonylurea Dose.