Abstract
Background
Beta-adrenergic antagonists (“beta-blockers”) have metabolic effects that influence risk of hypoglycemia in persons with diabetes. This risk may be elevated upon initial exposure and may vary among agents in the class.
Objective
To explore differences in the risk of hypoglycemia among new and prevalent users of individual beta-blockers in people with diabetes.
Design
Population-based nested case–control study. Cases of hypoglycemia, defined using emergency department or hospital admitting diagnosis, were matched with up to ten controls on age, sex, diabetes duration, and at least one prescription record for a beta-blocker in the 365 days preceding the hypoglycemic event or the equivalent date for controls (index date). Beta-blocker exposure was categorized as new, prevalent, recent, and remote use based on a pre-index 90-day (primary) and 30-day (secondary) exposure ascertainment period.
Participants
Persons aged 66 years or older with diabetes in Ontario, Canada.
Main Measures
Conditional logistic regression models were used to calculate adjusted odds ratios with 95% confidence intervals for the association between initiation of a beta-blocker and the risk of hypoglycemia.
Key Results
We identified 26,545 cases and 265,450 matched controls. The median age of patients was 78 years, 50% were female, and their median time since diabetes diagnosis of 11 years. Compared with prevalent use of beta-blockers, new use was associated with an increased odds of hypoglycemia (adjusted odds ratio [aOR] 1.87, 95% confidence interval [CI] 1.75–1.99). In contrast, neither recent use (aOR 1.03, 95% CI 0.98–1.08) nor remote use (aOR 0.82, 95% CI 0.77–0.87) was associated with an increased odds of hypoglycemia. Among individual agents, nonselective beta-blockers had the highest risk of hypoglycemia associated with new use.
Conclusions
In older patients with diabetes, the initiation of beta-blockers is associated with a near-doubling of the odds of a hospital encounter for hypoglycemia.
KEY WORDS: nested case–control study, type 2 diabetes, beta-blockers, hypoglycemia, drug safety
INTRODUCTION
Beta-blockers are prescribed for a variety of indications, including hypertension, cardioprotection following acute myocardial infarction, migraine prophylaxis, and essential tremor.1 Despite their potential benefits, beta-blockers carry potential harms including bradycardia, syncope, fatigue, increased airway resistance, and adverse metabolic effects. One such effect in patients with diabetes is hypoglycemia.2,3 By antagonizing the effects of epinephrine, beta-blockade may slow the correction of hypoglycemia due to suppression of glycogenolysis and gluconeogenesis, and dull symptoms of early neuroglycopenia, the latter leading to hypoglycemic unawareness.
Beta-blockers are pharmacologically diverse which may impact hypoglycemic risk.4–6 Beta-1-specific blockers such as bisoprolol, acebutolol, atenolol, and metoprolol are “cardioselective” and thought to have fewer metabolic effects. Likewise, beta-blockers with concurrent alpha-1 blockade and nitric oxide release such as carvedilol and labetalol may have positive effects on metabolic parameters including a diminished reduction in insulin sensitivity. Following initiation of a beta-blocker, upregulation or downregulation of beta-adrenoreceptors may also impact hypoglycemic risk with the early period prior to upregulation being one of increased risk of hypoglycemia.7
Previous studies evaluating the risk of hypoglycemia among patients receiving beta-blockers have had important methodologic limitations including unexposed reference groups, 8 small sample sizes,9,10 and no comparison of agents within the class.8,10,11 We aimed to expand upon this literature and evaluate the comparative risk of hospital presentation with hypoglycemia among users of beta-blockers with diabetes.
METHODS
Design and Data Sources
We conducted a nested case–control study using population-based administrative healthcare data housed and analyzed at ICES, a non-profit research institute in Ontario, Canada’s most populous province. The data included information on demographics, health insurance status, and vital status from the provincial health insurance registry and the Ontario Health Insurance Plan (OHIP) Registered Persons Database; outpatient and hospital encounter data from the OHIP Claims History Database and the Canadian Institute for Health Information Discharge Abstract Database (DAD) and National Ambulatory Care Reporting System (NACRS), respectively; and prescription drug claim data from the Ontario Drug Benefit Plan (ODB) Database. Because comprehensive drug claim data are not available for all residents of Ontario, ascertainment of drug exposures and outcomes was confined to Ontarians eligible for the ODB program aged ≥ 65 years. We used a validated case definition to identify persons with diabetes. This algorithm requires at least two physician visits with an International Classification of Diseases, 9th Revision (ICD-9) diagnosis code for diabetes (250) over 1-year period, at least one hospitalization with an ICD-10, Canadian modification (ICD-10-CA) diagnosis code for diabetes (E10, E11, E13, E14), or at least one prescription drug claim for an antihyperglycemic agent in a 1-year period; and has a sensitivity of 90%, specificity of 98%, and a positivity predictive value (PPV) of 83%.12 The ODB Database was used to characterize treatment episodes (drug identification numbers [DINs] used to identify medications) in persons aged ≥ 65 years; the DAD and the NACRS to identify the occurrence and timing of hypoglycemia hospitalizations and emergency department visits via ICD-10 codes, respectively; and the OHIP Registered Persons Database to capture demographics and vital statistics. Other databases including the Ontario Mental Health Reporting System, which captures mental health hospital admissions; the ICES Physician Database, which describes the characteristics of all physicians licensed to practice in Ontario; and validated chronic disease cohorts for asthma, heart failure, chronic obstructive pulmonary disease, human immunodeficiency virus, Crohn’s and colitis, myocardial infarction, rheumatoid arthritis, and dementia,13–20 which were used as potential confounding variables.
Study Patients
Entry into the nested cohort occurred on the date of diagnosis of diabetes identified in the Ontario Diabetes Database from April 1, 2003, to March 31, 2019, among persons aged ≥ 66 years. Persons remained in the cohort until the earliest date of a hypoglycemic event (case), loss of Ontario Health Insurance Plan coverage, death, or March 31, 2020 (the end of the study period). Cases and controls were then identified after applying a series of exclusion criteria including missing sex or birth date, age < 66, invalid date of death, not an Ontario resident, ineligible for the Ontario Medical Insurance Plan on the index date, less than 365 days of continuous Ontario Medical Insurance Plan and Ontario Drug Benefit coverage prior to index date, no oral beta-blocker treatment episode during the 365 days before index date, any prescription claims for nadolol, pindolol, or timolol due to very low prescription counts, and > 1 type of beta-blocker in the 365 days prior to index.
We defined cases as those within the cohort who attended an emergency department or were hospitalized with an admitting diagnosis code of hypoglycemia, using ICD-10-CA diagnosis codes E15, E160, E161, E162, E1063, E1163, E1363, or E1463. This algorithm has been reported to have high positive predictive value (94%), but limited sensitivity (12.7%).21
For each case, we defined the index date as the date of the first episode of hypoglycemia during the accrual period. Up to 10 controls were randomly selected for each case and matched on sex, age (± 1 year), and diabetes duration (< 3 years or 3 + years), with the index date for each control identical to that of the matched case. Cases were eligible to serve as a control in risk sets prior to each case’s index date. Cases and controls were required to have a prescription drug claim for a single oral beta-blocker within 365 days prior to their index date.
Exposure
The primary exposure definition used a 90-day assessment window prior to the index date for each risk set. We categorized cases and controls into mutually exclusive groups as new users, prevalent users, recent users, or remote users of a beta-blocker (Fig. 1). New users were defined as persons who initiated their first prescription for a single oral beta-blocker in the 90 days preceding the index date, with the treatment duration (based on the days supplied) overlapping index date. If there were multiple claims for the same beta-blocker on the same date, the days supplied were summed. Prevalent users were defined as persons who had one or more prescriptions for a single oral beta-blocker more than 90 days before the index date with the treatment duration overlapping the index date. Recent users were defined as persons who had one or more prescriptions for a single oral beta-blocker with an end of treatment duration within 90 days before the index date. Remote users were defined as persons who had one or more prescriptions for a single oral beta-blocker with an end of treatment duration greater than 90 days before the index date. Eligible beta-blockers were those commonly used and available on the Ontario Drug Benefit Formulary: acebutolol, atenolol, bisoprolol, carvedilol, labetalol, metoprolol, and propranolol. Using the same risk sets of cases and controls, we also ascertained the individual beta-blocker used in the year prior to the index date for each person.
Figure 1.
Exposure ascertainment flow diagram for defining new, prevalent, recent, and remote use of beta-blockers.
We also created a secondary exposure definition in which cases and controls were categorized as above, but with new user and prevalent user defined using a 30-day exposure interval prior to the index date. For all exposure definitions, the first prescription was defined as no prior claim for any beta-blocker in the previous 365 days.
Statistical Analysis
Baseline characteristics of cases and controls were compared using standardized differences. Differences greater than 10% were considered meaningful. The first exposure contrast of interest was to compare new users to prevalent users of beta-blockers. Recent and remote users were also compared to prevalent beta-blocker users. The second exposure contrast of interest was to compare new users to all other users (mix of prevalent, recent, and remote) for individual beta-blockers. Conditional logistic regression was used to estimate the independent association of the new, past, and remote beta-blocker use compared to prevalent use on the occurrence of serious hypoglycemia. To control for potential confounders, we included the following covariates in a multivariable conditional logistic regression model using a 365-day assessment window prior to cohort entry: number of hospitalizations (0, 1, 2, 3 +), the presence of chronic kidney disease (indicator variable based on presence of physician visit, emergency department visit, or hospitalization diagnostic or service codes), insulin use (indicator variable based on any insulin prescription record), insulin secretagogues (indicator variable based on any prescription record for a sulfonylurea or meglitinide), and the number of distinct drugs used (tertiles). We conducted an additional sensitivity analysis to explore the impact of including additional covariates in the conditional logistic regression model. Specifically, we adjusted for the number of hospitalizations, presence of chronic kidney disease, drug count (tertiles) in the 365 days prior to cohort entry, diabetes duration, plus the following additional variables assessed during the 365 days prior to the index date: alcohol use (yes/no), insulin duration, sulfonylurea duration, other antidiabetic agents (yes/no), and long-term care resident (yes/no). Insulin and sulfonylurea duration were assessed in the same manner as beta-blocker exposure (prevalent use, new use based on a 90-day pre-index ascertainment window, recent use, remote use). Lastly, we evaluated subgroup effects by including a multiplicative interaction term in our conditional logistic regression model for the following subgroups: selective vs. non-selective beta-blocker users, insulin users vs. non-users, and secretagogue users vs. non-users. Odds ratios and 95% confidence intervals were estimated for all the above exposure contrasts for both the primary and secondary exposure definitions. Data analysis was conducted using SAS version 9.4 (Cary, NC, USA).
RESULTS
Within a cohort of 1,744,287 persons with diabetes, we identified 26,545 cases of hypoglycemia (Fig. 2). Eligible matches were found for all cases with a total of 265,450 matched controls. The median age of study patients at index was 78 years, 50% were female, and their median time since diabetes diagnosis was 11 years. On average, cases had longer duration of diabetes, were prescribed more drugs and glucose test strips, used more health system resources, and had more comorbidities compared to controls (Table 1). Of the 291,995 patients selected from the cohort for the nested case–control study, the majority were metoprolol users (n = 117,644 [40%]), followed by atenolol (n = 81,147 [28%]), bisoprolol (n = 63,899 [22%]), acebutolol (n = 10,984 [4%]), carvedilol (n = 9215 [3%]), propranolol (n = 6686 [2%]), and labetalol (n = 2420 [1%]) users.
Figure 2.
Study flow diagram for identification of hypoglycemia cases.
Table 1.
Characteristics of Study Patients
| Controls | Cases | STD DIFF | |
|---|---|---|---|
| Number of individuals | N = 265,450 | N = 26,545 | |
| Age at index (years), median (IQR) | 78 (72–83) | 78 (72–83) | 0.00 |
| Female, n (%) | 131,481 (49.5%) | 13,138 (49.5%) | 0.00 |
| Time since diabetes diagnosis (years), median (IQR) | 10 (6–14) | 14 (10–18) | 0.51 |
| Living in LTC home at index, n (%) | 15,006 (5.7%) | 2184 (8.2%) | 0.10 |
| Number of distinct drugs used in 365 days before index date, mean (SD) | 13.25 ± 5.66 | 17.69 ± 6.54 | 0.71 |
| Tertile 1 | 93,302 (35.1%) | 3110 (11.7%) | 0.57 |
| Tertile 2 | 93,994 (35.4%) | 7993 (30.1%) | 0.11 |
| Tertile 3 | 78,154 (29.4%) | 15,442 (58.2%) | 0.59 |
| Number of glucose strips dispensed in 365 days before index date, mean (SD) | 81.75 ± 130.77 | 113.67 ± 113.99 | 0.25 |
| Number of outpatient physician visits in 365 days before index date, mean (SD) | 12.67 ± 9.20 | 15.81 ± 10.74 | 0.31 |
| Number of hospitalizations in 365 days before index date, mean (SD) | 0.32 ± 0.74 | 0.85 ± 1.26 | 0.51 |
| Calendar year of index date | |||
| 2003–2008 | 102,300 (38.5%) | 10,216 (38.5%) | 0 |
| 2009–2014 | 89,580 (33.7%) | 8956 (33.7%) | 0 |
| 2015–2020 | 73,570 (27.7%) | 7373 (27.8%) | 0 |
| Income quintile | |||
| 1st (lowest) | 61,940 (23.3%) | 7006 (26.4%) | 0.07 |
| 2nd | 59,600 (22.5%) | 5917 (22.3%) | 0.00 |
| 3rd | 51,860 (19.5%) | 5205 (19.6%) | 0.00 |
| 4th | 48,507 (18.3%) | 4528 (17.1%) | 0.03 |
| 5th (highest) | 42,425 (16.0%) | 3757 (14.2%) | 0.05 |
| Comorbidities, n (%) | |||
| Alcohol-use disorder | 4273 (1.6%) | 883 (3.3%) | 0.11 |
| Acute myocardial infarction | 27,684 (10.4%) | 4619 (17.4%) | 0.20 |
| Angina | 79,931 (30.1%) | 9607 (36.2%) | 0.13 |
| Atrial fibrillation | 74,497 (28.1%) | 9624 (36.3%) | 0.18 |
| Dyslipidemia | 50,529 (19.0%) | 5199 (19.6%) | 0.01 |
| Heart failure | 75,330 (28.4%) | 12,927 (48.7%) | 0.43 |
| Hypertension | 248,648 (93.7%) | 25,034 (94.3%) | 0.03 |
| Stroke or transient ischemic attack | 33,059 (12.5%) | 5113 (19.3%) | 0.19 |
| Peripheral vascular disease | 18,442 (6.9%) | 3386 (12.8%) | 0.20 |
| Asthma | 18,041 (6.8%) | 2166 (8.2%) | 0.05 |
| Chronic obstructive pulmonary disease | 65,488 (24.7%) | 8939 (33.7%) | 0.20 |
| Inflammatory bowel disease | 691 (0.3%) | 79 (0.3%) | 0.01 |
| Chronic kidney disease | 46,294 (17.4%) | 10,343 (39.0%) | 0.49 |
| Dementia | 25,880 (9.7%) | 4173 (15.7%) | 0.18 |
| Rheumatoid arthritis | 6335 (2.4%) | 829 (3.1%) | 0.05 |
| Major depressive disorder | 12,604 (4.7%) | 1975 (7.4%) | 0.11 |
| Generalized anxiety disorder | 53,882 (20.3%) | 6084 (22.9%) | 0.06 |
| Schizophrenia | 1544 (0.6%) | 264 (1.0%) | 0.05 |
| Hyperthyroidism | 2568 (1.0%) | 268 (1.0%) | 0.00 |
| Hypothyroidism | 12,748 (4.8%) | 1705 (6.4%) | 0.07 |
| Use of antihyperglycemic medications, n (%) | |||
| Insulin | 38,549 (14.5%) | 11,171 (42.1%) | 0.64 |
| Agents associated with hypoglycemia | 93,225 (35.1%) | 12,656 (47.7%) | 0.26 |
| Agents not associated with hypoglycemia | 164,102 (61.8%) | 16,192 (61.0%) | 0.02 |
All data are expressed as n (%) unless otherwise noted
Figure 3 shows the number of cases and controls in each exposure category based on the primary exposure ascertainment window of 90 days prior to the index date. After controlling for potential confounders, risk of serious hypoglycemia was 1.9 times greater among new users of beta-blockers compared to prevalent users. Recent users had similar chances of serious hypoglycemia compared to prevalent users and remote users had a slightly lower risk. As expected, these associations strengthened in a secondary analysis employing a 30-day exposure ascertainment window (Fig. 4). The sensitivity analysis with additional covariates showed results consistent with the main analysis for both the 90-day (new user: OR 1.83, 95% CI 1.71–1.95; past user OR 1.05, 95% CI 1.00–1.11; remote user OR 0.86, 95% CI 0.81–0.91) and 30-day (new user: OR 2.31, 95% CI 2.11–2.53; past user OR 1.04, 95% CI 0.99–1.09; remote user OR 0.84, 95% CI 0.79–0.90) exposure ascertainment window. Subgroup analysis showed a larger magnitude of association for new-users compared with prevalent users of beta-blockers who were also using insulin during the 90-day exposure window (OR 2.18, 95% CI 1.95–2.43) versus those not using insulin (OR 1.73, 95% CI 1.60–1.87) with an interaction p-value < 0.01. No other subgroup analyses showed statistically significant effects (with all p-values > 0.1).
Figure 3.
New users, recent users, remote users versus prevalent users of beta-blockers and the occurrence of hypoglycemia using the primary exposure definition.
Figure 4.
New users, recent users, remote users versus prevalent users of beta-blockers and the occurrence of hypoglycemia using the secondary exposure definition.
The association between new use of individual beta-blockers and the occurrence of hypoglycemia using a 90-day and 30-day exposure ascertainment period is shown in Fig. 5. There was variation in the magnitude of association among new users of individual beta-blockers and hypoglycemia. Notably, acebutolol was not associated with an increased odds of hypoglycemia. New use of carvedilol, labetalol, and propranolol were all associated with a more than doubling of the odds of hypoglycemia compared to prevalent use.
Figure 5.
New users versus all other users (mix of prevalent, recent, remote) for individual beta-blockers using the primary (left) and secondary (right) exposure definitions.
DISCUSSION
In this population-based, nested case–control study of older adults with diabetes from Ontario, Canada, we observed that recent initiation of a beta-blocker was associated with an increased risk of an emergency department visit or hospitalization for hypoglycemia relative to prevalent use. The highest risk was evident with non-selective beta-blockers, for which we observed at least a twofold higher chance of serious hypoglycemia among new vs prevalent users. A statistically significant positive association was observed for cardioselective betablockers, atenolol, bisoprolol, and metoprolol but not for acebutolol.
Plausible Biologic Mechanisms
The observed increased odds of hypoglycemia could be the result of several mechanisms related to beta-blocker exposure. One such mechanism is a change in density of beta-receptors following beta-blocker exposure.7 Specifically, exposure to nonselective beta-blockers leads to upregulation of multiple beta-adrenergic receptor isoforms, whereas exposure to selective beta-1 blockers leads to upregulation of beta-1 receptors. In contrast, beta-blockers with intrinsic sympathomimetic activity downregulate beta-adrenoceptors.22 The initial period of exposure to a beta-blocker may be one of the greater risks because of the accentuation of the blunting of warning symptoms of hypoglycemia mediated by the sympathetic nervous system until upregulation occurs. The clinical relevance of such cellular level changes is unknown. A related contributing factor is the time required to recognize symptoms of hypoglycemia. A person’s lack of awareness of the fact that early adrenergic symptoms including tachycardia, tremor, and anxiety may be blunted could increase the occurrence of serious hypoglycemia. Moreover, evidence suggests that sweating is enhanced and that other hypoglycemic related symptoms may continue to be present, such as nausea and feelings of hunger and hyperthermia.23 It is unlikely that nonselective beta-blockers cause absolute hypoglycemia unawareness, but rather shift the glycemic threshold for symptoms to a lower concentration of plasma glucose.24
The observed variation in the magnitude of association among individual beta-blocker medications may be due to selective binding affinity for beta-adrenergic isoforms, the presence of intrinsic sympathomimetic activity, alpha-1 blockade, and nitric oxide release. We included seven different beta-blockers in our study, four that are relatively beta-1 or cardioselective (acebutolol, atenolol, bisoprolol, metoprolol), and three that are nonselective (carvedilol, labetalol, and propranolol). Acebutolol also has intrinsic sympathomimetic properties and thus may have fewer metabolic adverse effects. Carvedilol and labetalol also block alpha-1 adrenoceptors resulting in vasodilation and may not impair insulin sensitivity to the same extent as beta-blockers without alpha-blocking properties.25 In addition, carvedilol may further enhance vasodilation by promoting nitric oxide accumulation.4
Comparison with Previous Studies
Empirical evidence demonstrating differences for the occurrence of hypoglycemia among individual beta-blockers is limited. There are few studies evaluating differences in the glycemic effects of beta-blockers based on the aforementioned pharmacologic properties. Shorr et al. reported a significant twofold association between nonselective beta-blocker use vs. non-use of antihypertensive drugs and serious hypoglycemia among insulin users. This study also reported a non-statistically significant decreased risk of serious hypoglycemia associated with beta-1 selective vs. nonselective beta-blockers (RR = 0.58, 95% CI 0.22–1.54).10 This study had limited statistical power to demonstrate differences, with only 12 and 6 events in the nonselective and beta-1 selective groups. Likewise, Thamer et al. did not report a significant association between beta-blocker use and serious hypoglycemia (OR = 1.2, 95% CI 0.8–1.7) in a case–control study using Medicaid data from Pennsylvania.11 Tsujimoto et al. found that beta-blocker use compared to non-use was associated with an increased risk of severe hypoglycemia among patients with diabetes (HR = 1.36, 95% CI 1.074–1.74).8 Dungan et al. conducted a hospital-based cohort study with 1020 carvedilol users, 886 selective beta blocker users, and 10,216 non-users upon admission.9 They observed that the odds of hypoglycemia were higher for carvedilol and selective beta-blocker users compared to non-users but only in a subpopulation of basal insulin non-users. Overall, the strength of the evidence assessing the association between beta-blocker use and hypoglycemia is low given the methodological heterogeneity and imprecision of results from previous studies.
Our study expands upon this evidence by analyzing a sample of over 25,000 persons with a hospital-encounter for hypoglycemia, focusing on a pharmacological relevant period of interest (i.e., 30 to 90 days after initiating beta-blocker therapy), and providing real world evidence for the association between individual beta-blockers and the occurrence of hypoglycemia.
Limitations
Our study has several limitations that warrant emphasis. We used prescription claims as a proxy for exposure to beta-blockers. We expect non-adherence to be roughly similar among individual beta-blockers. However, it is possible that non-adherence may be greater among new users compared to other exposure categories. Second, there was limited precision for point estimates for the analysis of new users of individual beta-blockers due to the low numbers of cases and controls for some drugs. Third, we were not able to measure all potential confounders known to affect decisions around the reason for use for beta-blockers. For example, changes in nutrition such as skipping a meal, changes in activity levels, low body mass index, changes in alcohol consumption, or acute changes in carbohydrate intake due to illness are not well captured in administrative data. Fourth, prior validation work tells us that most cases of hypoglycemia do not present to hospital.21 Finally, we do not know the indication for beta-blocker therapy. There are multiple indications for beta-blocker therapy, some of which are specific to individual agents (e.g., use of nadolol for esophageal varices and sotalol for ventricular arrhythmias).
In conclusion, we found a near doubling of the odds of hypoglycemia among initiators of beta-blockers compared to prevalent users with the highest risk among persons exposed to nonselective beta-blockers. These findings suggest that the weeks immediately after starting beta-blocker therapy deserve increased vigilance in monitoring for signs and symptoms of hypoglycemia.
Acknowledgements:
This study was supported by ICES, which is funded by an annual grant from the Ontario Ministries of Health and Long-Term Care. Parts of this material are based on data and/or information compiled and provided by the Ontario Ministry of Health and the Canadian Institute for Health Information. This document also used data adapted from the Statistics Canada Postal CodeOM Conversion File, which is based on data licensed from Canada Post Corporation, and/or data adapted from the Ontario Ministry of Health Postal Code Conversion File, which contains data copied under license from ©Canada Post Corporation and Statistics Canada. We thank IQVIA Solutions Canada Inc. for use of their Drug Information File.
Author Contribution:
JMG conceived the idea for this study. All authors were involved in the study design and interpretation of study findings. JMG was responsible for drafting the first version of the manuscript. All authors provided critical revisions to the manuscript.
Funding
This study was supported by JMG’s research operating funds from the University of Waterloo.
Data Availability
The data set from this study is held securely in coded form at ICES. Although data sharing agreements prohibit ICES from making the data set publicly available, access may be granted to those who meet prespecified criteria for confidential access at https://www.ices.on.ca/DAS. Please contact the authors for any supplemental information related to the study such as the study protocol, analysis plan, or analytic code. The full data set creation plan and underlying analytic code are available from the authors upon request, understanding that the computer programs may rely on coding templates or macros that are unique to ICES and are therefore either inaccessible or may require modification.
Declarations:
Ethics Approval:
The use of data in this project is authorized under Sect. 45 of Ontario’s Personal Health Information Protection Act and does not require review by a research ethics board. ICES is an independent, non-profit research institute with legal status under Ontario’s health information privacy law that allows it to collect and analyze healthcare and personal data without consent for health system evaluation and improvement.
Disclaimer:
The opinions, results, and conclusions reported in this article are those of the authors and do not reflect those of the funding or data sources; no endorsement is intended or should be inferred.
Conflict of Interest:
BRS is funded by the University of Toronto as the Novo Nordisk Research Chair in Equitable Care of Diabetes and Related Conditions. All other authors declare no competing interests.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Oliver E, Mayor Jr F, D’Ocon P. Beta-blockers: Historical Perspective and Mechanisms of Action. Rev Esp Cardiol Engl Ed. 2019;72(10):853-862. 10.1016/j.rec.2019.04.006. [DOI] [PubMed] [Google Scholar]
- 2.Abramson EugeneA, Arky RonaldA, Woeber KennethA. Effects of Propranolol on the Hormonal and Metabolic Responses to Insulin-Induced Hypoglycemia. Lancet. 1966;288(7478):1386-1389. 10.1016/S0140-6736(66)90422-3. [DOI] [PubMed] [Google Scholar]
- 3.Mills GA, Horn JR. Beta-blockers and glucose control. Drug Intell Clin Pharm. 1985;19(4):246-251. 10.1177/106002808501900401. [DOI] [PubMed] [Google Scholar]
- 4.Poirier L, Tobe SW. Contemporary use of β-blockers: clinical relevance of subclassification. Can J Cardiol. 2014;30(5 Suppl):S9-S15. 10.1016/j.cjca.2013.12.001. [DOI] [PubMed] [Google Scholar]
- 5.Cruickshank JM. The clinical importance of cardioselectivity and lipophilicity in beta blockers. Am Heart J. 1980;100(2):160-178. 10.1016/0002-8703(80)90112-x. [DOI] [PubMed] [Google Scholar]
- 6.Majumdar SR. Beta-blockers for the Treatment of Hypertension in Patients with Diabetes: Exploring the Contraindication Myth. Cardiovasc Drugs Ther. 1999;13(5):435-439. 10.1023/A:1007860108024 [DOI] [PubMed] [Google Scholar]
- 7.Haeusler G. Pharmacology of beta-blockers: classical aspects and recent developments. J Cardiovasc Pharmacol. 1990;16 Suppl 5:S1-9. [PubMed] [Google Scholar]
- 8.Tsujimoto T, Sugiyama T, Shapiro MF, Noda M, Kajio H. Risk of Cardiovascular Events in Patients With Diabetes Mellitus on β-Blockers. Hypertens Dallas Tex 1979. 2017;70(1):103–110. 10.1161/HYPERTENSIONAHA.117.09259. [DOI] [PMC free article] [PubMed]
- 9.Dungan K, Merrill J, Long C, Binkley P. Effect of beta blocker use and type on hypoglycemia risk among hospitalized insulin requiring patients. Cardiovasc Diabetol. 2019;18(1):163. 10.1186/s12933-019-0967-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Shorr RI, Ray WA, Daugherty JR, Griffin MR. Antihypertensives and the risk of serious hypoglycemia in older persons using insulin or sulfonylureas. JAMA. 1997;278(1):40-43. [PubMed] [Google Scholar]
- 11.Thamer M, Ray NF, Taylor T. Association between antihypertensive drug use and hypoglycemia: a case-control study of diabetic users of insulin or sulfonylureas. Clin Ther. 1999;21(8):1387-1400. 10.1016/s0149-2918(99)80039-3. [DOI] [PubMed] [Google Scholar]
- 12.Lipscombe LL, Hwee J, Webster L, Shah BR, Booth GL, Tu K. Identifying diabetes cases from administrative data: a population-based validation study. BMC Health Serv Res. 2018;18(1):316. 10.1186/s12913-018-3148-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Gershon AS, Wang C, Guan J, Vasilevska-Ristovska J, Cicutto L, To T. Identifying patients with physician-diagnosed asthma in health administrative databases. Can Respir J J Can Thorac Soc. 2009;16(6):183-188. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Schultz SE, Rothwell DM, Chen Z, Tu K. Identifying cases of congestive heart failure from administrative data: a validation study using primary care patient records. Chronic Dis Inj Can. 2013;33(3):160-166. [PubMed] [Google Scholar]
- 15.Gershon AS, Wang C, Guan J, Vasilevska-Ristovska J, Cicutto L, To T. Identifying Individuals with Physcian Diagnosed COPD in Health Administrative Databases. COPD J Chronic Obstr Pulm Dis. 2009;6(5):388-394. 10.1080/15412550903140865. [DOI] [PubMed] [Google Scholar]
- 16.Antoniou T, Zagorski B, Loutfy MR, Strike C, Glazier RH. Validation of Case-Finding Algorithms Derived from Administrative Data for Identifying Adults Living with Human Immunodeficiency Virus Infection. PLoS One. 2011;6(6):e21748. 10.1371/journal.pone.0021748. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Benchimol EI, Guttmann A, Mack DR, et al. Validation of international algorithms to identify adults with inflammatory bowel disease in health administrative data from Ontario, Canada. J Clin Epidemiol. 2014;67(8):887-896. 10.1016/j.jclinepi.2014.02.019. [DOI] [PubMed] [Google Scholar]
- 18.Austin PC, Daly PA, Tu JV. A multicenter study of the coding accuracy of hospital discharge administrative data for patients admitted to cardiac care units in Ontario. Am Heart J. 2002;144(2):290-296. 10.1067/mhj.2002.123839. [DOI] [PubMed] [Google Scholar]
- 19.Widdifield J, Bernatsky S, Paterson J, et al. Accuracy of Canadian Health Administrative Databases in Identifying Patients With Rheumatoid Arthritis: A Validation Study Using the Medical Records of Rheumatologists - Widdifield - 2013 - Arthritis Care & Research - Wiley Online Library. Arthritis Care Res. 2013;65(10):1582-1591. [DOI] [PubMed] [Google Scholar]
- 20.Jaakkimainen RL, Bronskill SE, Tierney MC, et al. Identification of Physician-Diagnosed Alzheimer’s Disease and Related Dementias in Population-Based Administrative Data: A Validation Study Using Family Physicians’ Electronic Medical Records. J Alzheimers Dis. 2016;54(1):337-349. 10.3233/JAD-160105. [DOI] [PubMed] [Google Scholar]
- 21.Hodge MC, Dixon S, Garg AX, Clemens KK. Validation of an International Statistical Classification of Diseases and Related Health Problems 10th Revision Coding Algorithm for Hospital Encounters with Hypoglycemia. Can J Diabetes. 2017;41(3):322-328. 10.1016/j.jcjd.2016.11.003. [DOI] [PubMed] [Google Scholar]
- 22.Michel MC, Pingsmann A, Beckeringh JJ. Selective regulation offl- and fl2-adrenoceptors in the human heart by chronic f-adrenoceptor antagonist. Br J Pharmacol. 1988;94:685-692. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Viberti GC, Keen H, Bloom SR. Beta blockade and diabetes mellitus: Effect of oxprenolol and metoprolol on the metabolic, cardiovascular, and hormonal response to insulin-induced hypoglycemia in normal subjects. Metabolism. 1980;29(9):866-872. 10.1016/0026-0495(80)90126-2. [DOI] [PubMed] [Google Scholar]
- 24.Hirsh I, Boyle P, Craft S, Cryer P. Higher Glycemic Thresholds for Symptoms During β-Adrenergic Blockade in IDDM | Diabetes | American Diabetes Association. Diabetes. 1991;40(9):1177-1186. 10.2337/diab.40.9.1177. [DOI] [PubMed] [Google Scholar]
- 25.Poirier L, Cléroux J, Nadeau A, Lacourcière Y. Effects of nebivolol and atenolol on insulin sensitivity and haemodynamics in hypertensive patients. J Hypertens. 2001;19(8):1429. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data set from this study is held securely in coded form at ICES. Although data sharing agreements prohibit ICES from making the data set publicly available, access may be granted to those who meet prespecified criteria for confidential access at https://www.ices.on.ca/DAS. Please contact the authors for any supplemental information related to the study such as the study protocol, analysis plan, or analytic code. The full data set creation plan and underlying analytic code are available from the authors upon request, understanding that the computer programs may rely on coding templates or macros that are unique to ICES and are therefore either inaccessible or may require modification.