Article Identification

Article Title: Clinical Efficacy of Beta-1 Selective Beta-Blockers Versus Propranolol in Patients With Thyroid Storm: A Retrospective Cohort Study

Authors: Yuichiro Matsuo, MD, MPH; Taisuke Jo, MD, PhD; Hideaki Watanabe, MD, MPH; Hiroki Matsui, MPH; Kiyohide Fushimi, MD, PhD; Hideo Yasunaga, MD, PhD

Journal Name: Critical Care Medicine

Year: 2024

Volume: 52

Issue: 00

Type of Study: Retrospective Cohort Study

DOI/PMID: 10.1097/CCM.0000000000006285

Quick Reference Summary

Key Findings: The study found that the use of beta-1 selective beta-blockers was not associated with a lower in-hospital mortality rate compared to propranolol in patients with thyroid storm (p=0.412). After adjustment, in-hospital mortality was 6.3% for the beta-1 selective group versus 7.4% for the propranolol group (OR, 0.85; 95% CI, 0.57–1.26).

Main Results: No significant difference in mortality was observed across various subgroups, including patients with acute heart failure or atrial fibrillation/flutter.

Core Clinical Question

Primary Research Question: Do beta-1 selective beta-blockers exhibit superiority over propranolol in reducing in-hospital mortality among patients with thyroid storm?

Background

Disease Overview

Thyroid storm is the most severe manifestation of thyrotoxicosis, characterized by extreme hyperthyroidism with a short-term mortality rate of 10%–20%.

Prior Data

Combination therapy with antithyroid drugs, glucocorticoids, and beta-blockers is the recommended treatment approach.

Historically, propranolol has been the preferred beta-blocker due to its nonselective properties affecting both beta-1 and beta-2 adrenergic receptors.

Current Standard of Care

2016 guidelines by the Japan Thyroid Association and the Japan Endocrine Society recommend beta-1 selective beta-blockers over nonselective beta-blockers like propranolol.

Knowledge Gaps Addressed

Limited evidence supporting the superiority of beta-1 selective beta-blockers over propranolol in reducing mortality in thyroid storm patients.

Study Rationale

To investigate in-hospital mortality outcomes based on the choice between beta-1 selective beta-blockers and propranolol, thereby providing evidence to support or refine current treatment guidelines.

Methods Summary

Study Design

Retrospective cohort study.

Setting and Time Period

Utilized the Diagnosis Procedure Combination database, a national inpatient database in Japan, encompassing data from April 2010 to March 2022.

Population Characteristics

Participants: 2,462 patients hospitalized with thyroid storm (ICD-10 code: E055).

Division: 1,452 received beta-1 selective beta-blockers; 1,010 received propranolol.

Inclusion/Exclusion Criteria

Included: Patients aged ≥15 years hospitalized with a diagnosis of thyroid storm.

Excluded:

1. Patients younger than 15 years at admission.
2. Multiple hospitalizations during the study period, retaining only the final hospitalization.
3. Patients who did not receive antithyroid drugs within the first 2 days of admission.
4. Patients who did not receive beta-blockers within the first 2 days of admission.
5. Patients who received both beta-1 selective beta-blockers and propranolol on the day of treatment initiation.

Intervention Details

Beta-1 Selective Group: Included IV landiolol, esmolol, oral bisoprolol, metoprolol, atenolol, and betaxolol.

Propranolol Group: Included both IV and oral propranolol.

Control/Comparison Group Details

Propranolol users compared against users of beta-1 selective beta-blockers.

Primary and Secondary Outcomes

Primary Outcome: In-hospital mortality rate.

Secondary Outcomes: Various subgroup analyses based on acute heart failure, atrial fibrillation/flutter, and mode of beta-blocker initiation.

Statistical Analysis Approach

Propensity-score overlap weighting to adjust for baseline confounders.

Generalized linear regression with a clustered sandwich estimator for standard errors.

Sample Size Calculations

Not explicitly detailed.

Ethics and Funding Information

Approved by the institutional review board of the University of Tokyo (Approval Number: 3501-(5)).

Funding received from the Ministry of Health, Labor, and Welfare of Japan (Grants 23AA2003 and 22AA2003).

No conflicts of interest reported.

Detailed Results

Participant Flow and Demographics

Total Eligible Patients: 2,462 (1,452 beta-1 selective; 1,010 propranolol).

Age: Mean age 48.6 years (SD=16.9) for beta-1 selective vs. 44.2 years (SD=17.2) for propranolol.

Sex: 30.6% male in the beta-1 selective group vs. 25.6% male in the propranolol group.

Comorbidities: Higher prevalence of heart failure (41.3% vs. 20.2%) and atrial fibrillation/flutter (32.4% vs. 15.7%) in the beta-1 selective group.

Primary Outcome Results

Crude In-hospital Mortality:

• Beta-1 selective: 9.3%
• Propranolol: 6.2%

Adjusted In-hospital Mortality:

• Beta-1 selective: 6.3%
• Propranolol: 7.4%
• Odds Ratio (OR): 0.85
• 95% Confidence Interval (CI): 0.57–1.26
• P-value: 0.412 (not statistically significant)

Secondary Outcome Results

Subgroup Analyses

• Acute Heart Failure:
  • Mortality: 15.3% (beta-1 selective) vs. 14.9% (propranolol)
  • OR: 1.03 (95% CI, 0.54–1.94)
  • P-value: 0.932
• Without Acute Heart Failure:
  • Mortality: 4.1% vs. 5.1%
  • OR: 0.80 (95% CI, 0.47–1.36)
  • P-value: 0.411
• Atrial Fibrillation or Atrial Flutter:
  • With: 8.1% vs. 10.8% (OR: 0.73; 95% CI, 0.34–1.57; P=0.417)
  • Without: 5.9% vs. 6.4% (OR: 0.92; 95% CI, 0.58–1.44; P=0.703)
• Mode of Beta-blocker Initiation:
  • IV Initiation: 11.4% vs. 13.4% (OR: 0.83; 95% CI, 0.45–1.54; P=0.554)
  • Oral Initiation: 3.9% vs. 4.7% (OR: 0.81; 95% CI, 0.45–1.47; P=0.490)

Adverse Events/Safety Data

Not explicitly reported in the provided text.

Results Tables

Table 2. Results of the Main Analysis

Outcome	Intervention Group (Beta-1 Selective)	Control Group (Propranolol)	Difference (95% CI)	P-value
In-hospital mortality	6.3%	7.4%	OR: 0.85 (0.57–1.26)	0.412

Table 3. Results of Subgroup Analyses

Subgroup	Intervention Group (%)	Control Group (%)	Difference (OR [95% CI])	P-value
Patients with acute heart failure	15.3	14.9	OR: 1.03 (0.54–1.94)	0.932
Patients without acute heart failure	4.1	5.1	OR: 0.80 (0.47–1.36)	0.411
Patients with atrial fibrillation/flutter	8.1	10.8	OR: 0.73 (0.34–1.57)	0.417
Patients without atrial fibrillation/flutter	5.9	6.4	OR: 0.92 (0.58–1.44)	0.703
Patients initiated with IV beta-blocker	11.4	13.4	OR: 0.83 (0.45–1.54)	0.554
Patients initiated with oral beta-blocker	3.9	4.7	OR: 0.81 (0.45–1.47)	0.490

Authors' Conclusions

Primary Conclusions

The choice between beta-1 selective beta-blockers and propranolol did not significantly affect in-hospital mortality in patients with thyroid storm.

Authors' Interpretation of Results

Both beta-1 selective beta-blockers and propranolol are viable options for beta-blocker therapy in thyroid storm, depending on the clinical context.

Clinical Implications Stated by Authors

Clinicians can consider either beta-1 selective beta-blockers or propranolol without expecting differences in in-hospital mortality outcomes.

Future Research Recommendations

Further studies are needed to explore specific patient populations that might benefit more from one type of beta-blocker over the other, especially considering cardiac function and other clinical variables not fully captured in this study.

Literature Review

A. Previous Studies and Meta-Analyses:

1. Sterling et al. (2016)
   • Findings: Favored the use of beta-1 selective beta-blockers over propranolol in thyroid storm management.
   • Citation: Sterling, et al. Medicine (Baltim) 2016; 95:e2848
2. Isozaki et al. (2016)
   • Findings: Retrospective cohort study indicating higher mortality with nonselective beta-blockers in thyroid storm patients with acute heart failure.
   • Citation: Isozaki, et al. Medicine (Baltim) 2016; 95:e2848

B. Contrasting Methodological Quality:

The current study utilized a larger sample size (2,462 patients) compared to Isozaki et al.’s study (286 patients), enhancing the statistical power and generalizability of the findings. Additionally, this study employed propensity-score overlap weighting to adjust for baseline disparities, whereas Isozaki et al. had limited adjustments due to smaller sample sizes.

C. Comparisons with Guidelines:

2016 Japan Thyroid Association and Japan Endocrine Society Guidelines

• Recommendations: Preferential use of beta-1 selective beta-blockers over nonselective beta-blockers like propranolol in thyroid storm management.
• Citation: Ross, et al. Thyroid 2016; 26:1343–1421

D. This Trial's Contribution:

Addition to Existing Evidence:

• Provides robust evidence from a large national database indicating no significant mortality difference between beta-1 selective beta-blockers and propranolol in thyroid storm patients.
• Contradiction: Challenges previous smaller studies suggesting superiority of beta-1 selective agents in certain subgroups.
• Citation: Yasunaga, et al. Thyroid Storm: Analysis Using National Inpatient Data in Japan. Medicine (Baltim) 2016; 95:e2848

Critical Analysis

A. Strengths:

• Large Sample Size: Included 2,462 patients, enhancing the study’s statistical power.
• Robust Data Source: Utilized the Diagnosis Procedure Combination database, capturing approximately 50% of all acute care hospitalizations in Japan.
• Advanced Statistical Methods: Employed propensity-score overlap weighting to adjust for confounders and balance covariates between treatment groups.
• Comprehensive Subgroup Analyses: Evaluated outcomes based on acute heart failure, atrial fibrillation/flutter, and mode of beta-blocker initiation.
• Real-World Applicability: Findings are based on a nationwide inpatient database, reflecting diverse clinical settings.

B. Limitations:

• Observational Design: Potential for unmeasured confounders influencing treatment choice and outcomes.
• Lack of Clinical Detail: Absence of baseline cardiac function data, dosage information, and specific clinical parameters like left ventricular ejection fraction.
• Diagnosis Validity: Reliance on ICD-10 codes for thyroid storm diagnosis, which might not capture all clinical nuances despite reported high validity.
• Incomplete Safety Data: Adverse events and safety profiles were not extensively reported.
• Potential Selection Bias: Exclusion of patients who switched beta-blockers or had multiple hospitalizations may impact the generalizability of results.

C. Literature Context:

• Comparison to Previous Studies: Contrasts with Isozaki et al. (2016) who favored beta-1 selective agents in acute heart failure subsets, likely due to smaller sample sizes and different adjustment methodologies.
• Positioning Within Existing Evidence: Aligns with studies in hyperthyroidism management showing no significant difference between beta-1 selective and nonselective beta-blockers regarding symptom improvement.
• Guideline Alignment: Supports current guidelines recommending either beta-1 selective or propranolol without a clear mortality advantage.
• Identified Knowledge Gaps: Highlights the need for studies incorporating detailed cardiac function and dosing information to further refine beta-blocker selection in thyroid storm.

Clinical Application

Changing Current Practice:

The findings suggest that clinicians can confidently use either beta-1 selective beta-blockers or propranolol in managing thyroid storm without expecting differences in in-hospital mortality. This flexibility allows for individualized patient care based on other clinical factors.

Applicable Patient Populations:

Particularly relevant for patients without severe acute heart failure, as no mortality benefit was observed even in those with acute heart failure. Clinicians should remain cautious in patients with compromised cardiac output.

Implementation Considerations:

Choice of beta-blocker can be guided by patient-specific factors such as existing comorbidities, potential side effects, and clinician familiarity with the medication.

How To Use This Info In Practice:

Practitioners can consider both beta-1 selective beta-blockers and propranolol as effective options for treating thyroid storm, selecting based on individual patient contexts and clinical judgment.

References

1. Akamizu T, Satoh T, Isozaki O, et al. Thyroid 2012; 22:661–679.
2. Ono Y, Ono S, Yasunaga H, et al. Medicine (Baltim) 2016; 95:e2848.
3. Bourcier S, Coutrot M, Kimmoun A, et al. Crit Care Med 2020; 48:83–90.
4. Ross DS, Burch HB, Cooper DS, et al. Thyroid 2016; 26:1343–1421.
5. Satoh T, Isozaki O, Suzuki A, et al. Thyroid 2016; 22:263–277.
6. Ramsay I. Br J Clin Pharmacol 1975; 2:385–388.
7. Silva JE, Bianco SD. Thyroid 2008; 18:157–165.
8. Wiersinga WM. Thyroid 1991; 1:273–277.
9. Isozaki O, Satoh T, Wakino S, et al. Medicine (Baltim) 2016; 95:e2848.
10. Yasunaga H. J Epidemiol 2017; 27:476–482.
11. Ono Y, et al. Medicine (Baltim) 2016; 95:e2848.
12. Li F, Morgan KL, Zaslavsky AM. J Am Stat Assoc 2018; 113:390–400.
13. Kanaoka K, Iwanaga Y, Okada K, et al. Circ J 2023; 87:536–542.
14. Isozaki O, et al. Clin Endocrinol (Oxf) 2016; 84:912–918.
15. Sterling et al. Chest 2016; 145:779–786.
16. Morales DR, Jackson C, Lipworth BJ, et al. Chest 2016; 145:779–786.
17. Metra M, Nodari S, Dei Cas L. Am J Cardiovasc Drugs 1977; 1:1505–1507.
18. Morales DR, Jackson C, Lipworth BJ, et al. Chest 2016; 145:779–786.
19. Teerlink JR, Alburikan K, Metra M, et al. Curr Cardiol Rev 2015; 11:53–62.
20. Polikar R, Burger AG, Scherrer U, et al. Circulation 1993; 87:1435–1441.
21. Klein I, Ojamaa K. N Engl J Med 2001; 344:501–509.
22. Siu CW, Yeung CY, Lau CP, et al. Heart 2007; 93:483–487.
23. Ikram H. Br Med J 1977; 1:1505–1507.
24. Abubakar H, Singh V, Arora A, et al. Am J Cardiovasc Drugs 1977; 1:1505–1507.
25. Burch HB, Wartofsky L. Endocrinol Metab Clin North Am 1993; 22:263–277.
26. Yamana H, Moriwaki M, Horiguchi H, et al. J Epidemiol 2017; 27:476–482.

Conflicts of Interest

The remaining authors have disclosed that they do not have any potential conflicts of interest.

Funding Sources

Dr. Matsuo’s institution received funding from the Ministry of Health, Labor, and Welfare of Japan (grants 23AA2003 and 22AA2003).

Dr. Matsui’s institution received funding from Japanese Kakenhi (grants 20H03907 and 21H03159).

Areas of Uncertainty

Potential benefits of beta-1 selective beta-blockers in specific subpopulations not fully explored.

Lack of dosage information and detailed cardiac function data limits nuanced interpretation.