Efficacy and safety of diltiazem versus metoprolol in the management of atrial fibrillation with rapid ventricular response

Introduction

Background

Acute atrial fibrillation (A-fib) is the most prevalent sustained cardiac arrhythmia encountered in emergency medicine (1). A-fib with rapid ventricular response (RVR) is characterized by an irregular heartbeat with a heart rate (HR) exceeding 120 beats per minute (bpm), resulting from the unsynchronized activation of atrial myocytes. This unsynchronized pumping causes blood to stagnate in the atria, which can lead to clot formation. These clots can travel through the bloodstream, giving rise to potential embolic complications such as myocardial infarctions or strokes (2). This can have debilitating effects, including permanent heart damage and neurologic dysfunctions. RVR is a subtype of A-fib where the irregular atrial contractions cause the ventricles to beat irregularly and much more rapidly than normal. The loss of atrial kick and rapid irregular ventricular rate interferes with diastolic filling and reduces ejection fraction in most cases thus reducing blood and oxygen delivery to tissues (3,4).

There has been a steady increase in the prevalence of A-fib over the years with the growth and aging of the U.S. population. In 1996, it was estimated that 2.3 million Americans had A-fib (5). Currently, approximately 5.3 million patients in the U.S. are diagnosed with A-fib, and this number is predicted to rise to 6.3 million by 2050 (5,6). In the Framingham heart study, both male and female patients were shown to have 1.5 to 1.9 times higher risk of death due to A-fib a decade after diagnosis (7). Furthermore, the economic burden associated with A-fib treatment is considerable, with annual costs ranging from $2,000 to $14,200 per patient, totaling roughly $6.65 billion in 2005 (2). It is expected that the economic costs will continue to rise over time.

Importance

Adequate rate control of A-fib with RVR is necessary to alleviate symptoms, as well as to prevent the development or reverse the subsequent development of tachycardia-mediated cardiomyopathy (8). There is also limited evidence for a mortality benefit in patients who received rate-control treatment with beta-blockers or calcium channel blockers, though this is limited by the nonrandomized nature of the dataset (9).

Current guidelines for the management of hemodynamically stable patients with A-fib with RVR include the use of beta-blockers, such as metoprolol, or calcium channel blockers like diltiazem (10). Metoprolol functions by diminishing the heart’s sympathetic response in the atria, atrioventricular (AV) node and ventricles, thus slowing the heart’s atrial and ventricular rate and slows AV conduction through the relatively selective blockade of β1 receptors. This medication is often the preferred choice for treating chronic A-fib, as well as for pregnant patients or those whose A-fib is triggered by hyperthyroidism (10-13). Diltiazem directly blocks L-type calcium channels to slow conduction through the AV node and by prolonging AV node refractoriness and is usually favored for acute A-fib and occasionally for chronic A-fib (10,11). Several smaller studies have shown intravenous (IV) diltiazem to be more effective in treating A-fib with RVR when compared to beta-blockers (14-16). However, these studies are limited by their small sample sizes, with the largest prospective randomized study consisting of only 100 patients. As such, additional, larger, and more rigorous studies are necessary to promote a change in recommendations and guidelines for the treatment of A-fib with RVR.

Goals of investigation

This retrospective study utilized the TriNetX database to compare the outcomes of mortality, HR, and intensive care unit (ICU) admissions in patients with A-fib with RVR who received IV diltiazem vs. IV metoprolol. We present this article in accordance with the STROBE reporting checklist (available at https://jeccm.amegroups.com/article/view/10.21037/jeccm-24-73/rc).

Methods

Design

The TriNetX database was selected for this study due to its robust international health research network, which grants access to retrospective electronic medical records encompassing comprehensive diagnostic, procedural, medication, laboratory, and genomic information. The database’s U.S. Collaborative Network includes nearly 92 million patients across 59 healthcare organizations (HCOs), primarily consisting of large academic tertiary care centers in America, along with their associated satellite facilities (17). This extensive and diverse patient population, coupled with the availability of detailed electronic medical records, makes the TriNetX database a valuable resource for conducting large-scale cohort studies with the potential to yield generalizable and clinically relevant findings.

Cohort definitions

Data for this retrospective cohort study was drawn from the U.S. Collaborative Network to identify two distinct cohorts (Table 1). Both cohorts consisted of adult patients aged 18 years and older at the time of the event. The patients were diagnosed with either Paroxysmal Atrial Fibrillation (ICD-10: CM: I48.0) or Unspecified Atrial Fibrillation (ICD-10: CM: I48.91) collectively referred to as A-fib with RVR, as per the International Classification of Diseases, Tenth Revision, Procedure Coding System (ICD-10). Additionally, these patients had a HR of at least 150 bpm (TNX: 9074). This HR of ≥150 bpm was chosen as the American Heart Association suggests that patients with A-fib at HR <150 bpm are unlikely to be unstable (18). Stability as an indication for electrical cardioversion relates to chest pain, altered mental status, severe heart failure, or hypotension in the setting of a HR >150 bpm. Cohort 1 includes patients who were given IV diltiazem (RXNORM: 3443), without metoprolol on the same day as their A-fib diagnosis. Cohort 2 consisted of patients administered IV metoprolol (RXNORM: 6918), without diltiazem on the day of the A-fib diagnosis. The study period was 20 years, from September 22, 2003, to September 22, 2023. Patients were excluded if the indexed event occurred more than 20 years ago.

Outcomes

Outcomes evaluated in both cohorts over the 7 days following diagnosis and treatment included mortality, acute myocardial infarction (ICD10: CM: I21), a HR equal to or exceeding 120 bpm (TNX: 9074), and ICU admissions (UMLS: CPT: 99291). There are varying definitions of HR control and American College of Cardiology (ACC)/American Heart Association (AHA) guidelines are to control ventricular response between 60–80 bpm at rest and between 90–115 bpm during moderate exertion (18). We chose to include <120 bpm to incorporate both groups. This analysis encompassed outcomes that transpired within a time frame starting from one day after the initial occurrence of the index event—A-fib with RVR, characterized by a HR of at least 150 bpm and treated with either IV diltiazem or IV metoprolol. This period concluded seven days post the index event’s first occurrence. A secondary exploratory analysis was performed analyzing the outcome of myocardial infarction (MI) in the above cohort on December 04, 2023, excluding patients who experienced the diagnosis of MI prior to day 0 of the analysis.

Mortality data within the TriNetX platform is obtained from electronic medical record (EMR) data and HCOs, in conjunction with the national death registries. There is potential for missed death events when a patient is treated at an HCO not affiliated with the TriNetX network and subsequently experiences a fatal outcome outside of this network. However, this represents only a minor issue, as currently, 94% of HCOs within the TriNetX network are also linked to the U.S. death registries. This percentage is steadily increasing as more HCOs continue to be linked with the registries.

Statistical analysis

The univariate analysis utilized both the Chi-square test and the t-test to examine the dataset. The measure of association tool in TriNetX was utilized to calculate risk differences (in percentages), relative risk (RR), 95% confidence intervals (CIs), and probability values (P values) for the outcomes.

A 1:1 propensity score matching was performed with linear and logistic regression on outcomes for demographics including age, sex, and race/ethnicity and comorbidities, such as hypertension, diabetes, acute and chronic renal failure, obesity, heart failure, prior cardiac arrest, ischemic heart disease, and malignant neoplasm of the bronchus/lung (Table 2). This helps control for potential confounders associated with mortality. All demographic data was self-reported by patients and recorded by HCOs to HL7 administrative standards. Statistical analyses were performed before and after propensity matching. A greedy nearest-neighbor matching algorithm was applied with a caliper width of 0.1 for the propensity scores. TriNetX randomizes the order of data rows to mitigate bias introduced by the nearest-neighbor algorithm. Propensity matching is generally performed without replacement. This study methodology has been previously validated (17). Statistical significance was set at a 2-sided alpha <0.05. The final analysis was performed on September 22, 2023.

Ethical considerations

The study protocol and data collection procedures were conducted in accordance with ethical guidelines and applicable regulations. Utilization of the data from TriNetX does not require UTMB IRB review as this is a secondary analysis of de-identified data. The UTMB IRB determined that this project is considered “not human subjects research”.

Results

In this study, data were analyzed from the TriNetX database, containing 91,785,977 patients from 59 HCOs within the U.S. Collaborative Network. Out of these, 19,366 patients met the inclusion criteria for our two cohorts prior to propensity score matching. Specifically, cohort 1 (diltiazem) comprised 8,762 patients, whereas cohort 2 (metoprolol) included 10,604 patients. Post propensity score matching, the total number of patients narrowed down to 14,598, with each cohort containing an equal number of 7,299 patients (Figure 1).

Figure 1 Cohort selection flow chart. RVR, rapid ventricular response.

The analysis revealed that patients with A-fib with RVR who were administered IV diltiazem had significantly better outcomes compared to those who received IV metoprolol within the first 7 days of initial presentation. The outcomes assessed were mortality, MI, the risk of HR exceeding 120 bpm, and ICU admissions. After propensity matching, the mortality rate was 3.3% in the diltiazem group compared to 5.1% in the metoprolol group (RR =0.643; 95% CI: 0.547–0.755; P<0.001) and MI was 2.4% compared to 3.4% (RR =0.703; 95% CI: 0.581–0.850; P<0.001). Additionally, the risk of having a HR greater than 120 bpm was lower in the diltiazem group (35.7% vs. 40.1%; RR =0.891; 95% CI: 0.855–0.929; P<0.001). ICU admissions were also less frequent among the diltiazem group (13.4% vs. 19.9%; RR =0.672; 95% CI: 0.624–0.724; P<0.001). In the secondary analysis evaluating patients with no outcomes of MI prior to day 0 of outcomes, those in the diltiazem group experienced lower rates of MI (0.6% vs. 1.0%; RR =0.562; 95% CI: 0.392–0.805; P=0.001). The results showed a similar trend before propensity matching (Tables 3,4).

Discussion

In this large, multicenter, propensity-matched retrospective cohort study, patients with A-fib with RVR treated with IV diltiazem experienced significantly more favorable outcomes compared to those treated with IV metoprolol. These outcomes include reduced 7-day mortality, reduced MI rate, an increased likelihood of achieving adequate rate control, and a lower probability of requiring ICU admission. Notably, this study to our knowledge is the largest to date, substantially surpassing previous studies in scale and takes into account several potential confounders, including CHF. In addition, when patients who were diagnosed with an MI on day 0 were excluded from the analysis, the relative risk of MI on day 1–7 was even lower in the diltiazem group after propensity matching.

The effectiveness of IV diltiazem compared to IV metoprolol in patients with A-fib with RVR has been subject to ongoing debate in recent studies. This study showed that favorable outcomes are more likely with the use of IV diltiazem. A recent 2022 meta-analysis published in the Indian Heart Journal analyzed the results of 14 studies [13 emergency department (ED), 1 ICU] with 1,732 patients (15). That analysis suggests IV diltiazem is more effective than IV metoprolol in achieving proper rate control for patients with A-fib with RVR. They highlight the significance of rate control, emphasizing that sub-optimal control is associated with an increased risk of adverse cardiovascular events (15). The meta-analysis found comparable rates of bradycardia and hypotension, after the administration of either drug. Their overall findings are consistent with similar meta-analyses of both retrospective and prospective data, such as Lan et al., with 1,214 patients which found IV diltiazem to be more efficient at rate control at 30 minutes than IV metoprolol. In addition, Jafri et al., in a review of prospective randomized trials, similarly concluded IV diltiazem to be more effective at 5, 10, and 15 minutes in patients with paroxysmal A-fib (14,19). The conclusions of these smaller studies that IV diltiazem is likely the preferable rate control treatment in patients with A-fib with RVR aligns with the findings of this study collected from the 19,366 patients.

In a retrospective cohort study conducted by Blackburn et al. in 2023, 100 patients were followed over a 22-month period following the initial administration of rate-control medication (20). It should be noted that this study differs from our own consideration, as it looks at long term outcomes. Their study measured several factors that were not considered in our analysis, such as total administered dose, time required to achieve HR reduction, adverse effects after administration, and pre-admission medication usage. Their conclusion was that the primary outcome of achieving a HR under 110 bpm was not statistically different between the metoprolol and diltiazem groups. The median time to achieve this outcome was 115.5 minutes for the metoprolol group and 99.0 minutes for the diltiazem group. This study found the frequency of bradycardia and hypotension not to be statistically different between the groups. The incidence of bradycardia was 7.1% in the metoprolol group and 4.9% in the diltiazem group. Similarly, the incidence of hypotension was not statistically different between groups with 20.2% in the metoprolol group and 18.4% in the diltiazem group. The relatively small sample size of 100 participants in their study limits the generalizability of the findings.

A-fib management typically involves either rhythm control, which aims to restore sinus rhythm, or rate control through pharmacotherapy. However, recent literature indicates no significant difference in outcomes between rhythm control and rate control in the short term (21,22). Although rate control alone is a viable long term management strategy for A-fib (23), many of these patients are referred for ablative therapy of the left atrium. Unstable patients with A-fib should be initially managed with electrical cardioversion for rhythm control regardless of their anticoagulation status. Many chronic A-fib patients will not respond to electrical cardioversion and may require primary management with rate control and chemical cardioversion (24).

The utilization of calcium channel blockers, specifically diltiazem, on patients experiencing RVR with prior history of heart failure with reduced ejection fraction (HFrEF), is sometimes discouraged due to the potential negative inotropic effect (25,26); however, the same can be said for metoprolol. The retrospective chart review conducted by Compagner et al. in 2022 concluded that although IV diltiazem and IV metoprolol had a nonsignificant difference between rate control at 30 minutes in HFrEF/heart failure with preserved ejection fraction (HFpEF) patients with RVR, diltiazem initiated a faster and greater reduction of HR (26). They concluded that there were no safety outcome differences between the two drugs being used in this context, despite what has been previously stated in prior literature. This review was limited by a small population size of 193 patients.

Diltiazem appears to provide more effective rate control in the acute setting of A-fib with RVR. Its action on the AV node can lead to a more rapid and pronounced decrease in ventricular rate compared to metoprolol which acts at the beta-adrenergic receptors (27). Identifying the optimal agent for rate control in the ED is important as the two agents identified in this study, metoprolol, and diltiazem (a non-dihydropyridine calcium channel blocker), can have additive effects on AV nodal blockade that can result in heart block or severe bradycardia (28). As a result, some authors suggest that inadequate rate control with one agent may potentially not allow a prompt switch to the other and instead necessitates the consideration of agents with less favorable side effect profiles such as digoxin or amiodarone. It is crucial to identify the agent that will have the greatest chance of initial success before escalating therapy. Our study provides additional evidence that this agent may be diltiazem.

Limitations

It is important to acknowledge the inherent limitations of a retrospective study, primarily in establishing causation. Propensity matching was performed for several risk factors in an attempt to ensure that the cohorts were evenly matched, however it is possible that an underlying risk factor or confounder, such as HFpEF vs. HFrEF, was not evaluated for due to limitations in covariate selection that could have contributed to the differences seen in this dataset. The duration of treatment and the dose appropriateness of diltiazem and metoprolol are not easily accessible in this database and were not included in the analysis.

The TriNetX database does not have sufficient granularity to distinguish between events that occur in close temporal proximity and determine the precise timeframe, within hours, of the response to treatment. Patients with myocardial infarction as an outcome attributed to day 1 could have possibly occurred on day 0, and A-fib with RVR due to myocardial injury and myocardial ischemia due to uncontrolled A-fib with RVR might not be distinguishable. Additionally, this limitation in granularity presents challenges in analyzing beta blocker administration and synchronized cardioversions, so these were not evaluated. Because this study focused on emergency treatment, patients that may have received diltiazem in the first 24 hours and then metoprolol or other betablocker after this period or the reverse (metoprolol IV then diltiazem IV or oral) were not analyzed or controlled for.

Finally, we could not assess whether there was a clustering of patients receiving IV diltiazem vs. IV metoprolol at specific centers. This is because the TriNetX privacy policy prevents us from accessing such information, owing to the deidentified nature of the dataset.

These limitations however, are mitigated as the number of patients queried and included in both cohorts help to substantially reduce the risk of random chance. Furthermore, the substantial number of academic centers that are included in the TriNetX database from which our dataset was drawn increases the generalizability and external validity of our study. These results align with previous studies supporting diltiazem and addresses a large gap in the existing literature, namely the small sample sizes involved in prior studies.

Conclusions

In conclusion, this comprehensive, multicenter, retrospective cohort study shows that in patients with A-fib with RVR, the administration of diltiazem for rate control is associated with a significantly lower risk of death, a higher likelihood of achieving an adequate reduction in HR, and decreased rate of ICU admission. Considering the large scale of this study, coupled with corroborating findings from previous randomized control studies, there is a convincing case favoring the use of diltiazem over metoprolol in the management of hemodynamically stable patients with A-fib with RVR. These findings may prompt a reevaluation and update of existing clinical guidelines.

Acknowledgments

The abstract was accepted for presentation at ACEP Research Forum 2023 on October 09, 2023.

Funding: This study was conducted with the support of the Institute for Translational Sciences at the University of Texas Medical Branch, supported in part by a Clinical and Translational Science Award from the National Center for Advancing Translational Sciences, National Institutes of Health (NIH) (No. UL1 TR001439). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://jeccm.amegroups.com/article/view/10.21037/jeccm-24-73/rc

Data Sharing Statement: Available at https://jeccm.amegroups.com/article/view/10.21037/jeccm-24-73/dss

Peer Review File: Available at https://jeccm.amegroups.com/article/view/10.21037/jeccm-24-73/prf

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jeccm.amegroups.com/article/view/10.21037/jeccm-24-73/coif). L.R.F. reports supports or endowments from University of Texas Medical Branch. Besides, she serves as the unpaid member on Texas Society of Anesthesiology Education Committee and the Committee on Resident and Medical Student Training, and the unpaid Advisor for American Society of Anesthesiology Program Directors Advisory Group. The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study protocol and data collection procedures were conducted in accordance with ethical guidelines and applicable regulations. Utilization of the data from TriNetX does not require UTMB IRB review as this is a secondary analysis of de-identified data. The UTMB IRB determined that this project is considered “not human subjects research”.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Wijesurendra RS, Casadei B. Mechanisms of atrial fibrillation. Heart 2019;105:1860-7. [Crossref] [PubMed]
2. Wolowacz SE, Samuel M, Brennan VK, et al. The cost of illness of atrial fibrillation: a systematic review of the recent literature. Europace 2011;13:1375-85. [Crossref] [PubMed]
3. Khoo CW, Lip GYH. Acute management of atrial fibrillation. Chest 2009;135:849-59. [Crossref] [PubMed]
4. Seet RC, Zhang Y, Rabinstein AA, et al. Risk factors and consequences of atrial fibrillation with rapid ventricular response in patients with ischemic stroke treated with intravenous thrombolysis. J Stroke Cerebrovasc Dis 2013;22:161-5. [Crossref] [PubMed]
5. Go AS, Hylek EM, Phillips KA, et al. Prevalence of diagnosed atrial fibrillation in adults: national implications for rhythm management and stroke prevention: the AnTicoagulation and Risk Factors in Atrial Fibrillation (ATRIA) Study. JAMA 2001;285:2370-5. [Crossref] [PubMed]
6. Benjamin EJ, Muntner P, Alonso A, et al. Heart Disease and Stroke Statistics-2019 Update: A Report From the American Heart Association. Circulation 2019;139:e56-528. Erratum in: Circulation 2020;141:e33. [Crossref] [PubMed]
7. Benjamin EJ, Wolf PA, D'Agostino RB, et al. Impact of atrial fibrillation on the risk of death: the Framingham Heart Study. Circulation 1998;98:946-52. [Crossref] [PubMed]
8. Umana E, Solares CA, Alpert MA. Tachycardia-induced cardiomyopathy. Am J Med 2003;114:51-5. [Crossref] [PubMed]
9. Chao TF, Liu CJ, Tuan TC, et al. Rate-control treatment and mortality in atrial fibrillation. Circulation 2015;132:1604-12. [Crossref] [PubMed]
10. Correction to. 2019 AHA/ACC/HRS Focused Update of the 2014 AHA/ACC/HRS Guideline for the Management of Patients With Atrial Fibrillation: A Report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the Heart Rhythm Society. Circulation 2019;140:e285. [PubMed]
11. Anter E, Jessup M, Callans DJ. Atrial fibrillation and heart failure: treatment considerations for a dual epidemic. Circulation 2009;119:2516-25. [Crossref] [PubMed]
12. Wilson JH. Arrhythmias during Pregnancy. In: Wilson JH, Schnettler WT, Lubert AM, et al. editors. Maternal Cardiac Care: A Guide to Managing Pregnant Women with Heart Disease. Elsevier; 2023:85-9.
13. Sandström B. Antihypertensive treatment with the adrenergic beta-receptor blocker metoprolol during pregnancy. Gynecol Invest 1978;9:195-204. [Crossref] [PubMed]
14. Lan Q, Wu F, Han B, et al. Intravenous diltiazem versus metoprolol for atrial fibrillation with rapid ventricular rate: A meta-analysis. Am J Emerg Med 2022;51:248-56. [Crossref] [PubMed]
15. Sharda SC, Bhatia MS. Comparison of diltiazem and metoprolol for atrial fibrillation with rapid ventricular rate: Systematic review and meta-analysis. Indian Heart J 2022;74:494-9. [Crossref] [PubMed]
16. Fromm C, Suau SJ, Cohen V, et al. Diltiazem vs. Metoprolol in the Management of Atrial Fibrillation or Flutter with Rapid Ventricular Rate in the Emergency Department. J Emerg Med 2015;49:175-82. [Crossref] [PubMed]
17. Murphy LR, Hill TP, Paul K, et al. Tenecteplase Versus Alteplase for Acute Stroke: Mortality and Bleeding Complications. Ann Emerg Med 2023;82:720-8. [Crossref] [PubMed]
18. Part 7.3: Management of symptomatic bradycardia and tachycardia. Circulation 2005;112:IV-67-77.
19. Jafri SH, Xu J, Warsi I, et al. Diltiazem versus metoprolol for the management of atrial fibrillation: A systematic review and meta-analysis. Am J Emerg Med 2021;48:323-7. [Crossref] [PubMed]
20. Blackburn M, Edwards L, Woolum J, et al. Metoprolol versus Diltiazem in the Emergency Department for atrial fibrillation with rapid ventricular response. JEM Reports 2023;2:100020. [Crossref]
21. Al-Khatib SM, Allen LaPointe NM, Chatterjee R, et al. Rate- and rhythm-control therapies in patients with atrial fibrillation: a systematic review. Ann Intern Med 2014;160:760-73. [Crossref] [PubMed]
22. de Denus S, Sanoski CA, Carlsson J, et al. Rate vs rhythm control in patients with atrial fibrillation: a meta-analysis. Arch Intern Med 2005;165:258-62. [Crossref] [PubMed]
23. Chiang CE, Naditch-Brûlé L, Murin J, et al. Distribution and risk profile of paroxysmal, persistent, and permanent atrial fibrillation in routine clinical practice: insight from the real-life global survey evaluating patients with atrial fibrillation international registry. Circ Arrhythm Electrophysiol 2012;5:632-9. [Crossref] [PubMed]
24. Long B, Robertson J, Koyfman A, et al. Emergency medicine considerations in atrial fibrillation. Am J Emerg Med 2018;36:1070-8. [Crossref] [PubMed]
25. Jandali MB. Safety of Intravenous Diltiazem in Reduced Ejection Fraction Heart Failure with Rapid Atrial Fibrillation. Clin Drug Investig 2018;38:503-8. [Crossref] [PubMed]
26. Compagner CT, Wysocki CR, Reich EK, et al. Intravenous metoprolol versus diltiazem for atrial fibrillation with concomitant heart failure. Am J Emerg Med 2022;62:49-54. [Crossref] [PubMed]
27. Alobaida M, Alrumayh A. Rate control strategies for atrial fibrillation. Ann Med 2021;53:682-92. [Crossref] [PubMed]
28. Gradman AH, Basile JN, Carter BL, et al. Combination therapy in hypertension. J Am Soc Hypertens 2010;4:42-50. [Crossref] [PubMed]