Secondary Logo

Journal Logo

Original Research Articles: Original Clinical Research Report

Adverse Outcomes Associated With Delaying or Withholding β-Blockers After Cardiac Surgery: A Retrospective Single-Center Cohort Study

Chanan, Emily L. MD*; Kendale, Samir M. MD*; Cuff, Germaine PhD*; Galloway, Aubrey C. MD; Nunnally, Mark E. MD*

Author Information
doi: 10.1213/ANE.0000000000005051

Abstract

KEY POINTS

  • Question: How do clinical outcomes differ in patients who receive β-blockers during the early postoperative period after cardiac surgery with patients who either do not receive β-blockers during the early period or at any time in-hospital?
  • Findings: In this retrospective study of 2526 patients who underwent open cardiac surgery, we used propensity score matching and logistic regression analyses on weighted sets and found that patients who do not receive postoperative β-blockers within the first 5 days after cardiac surgery or at any time in-hospital did not have significantly increased adjusted odds ratios of in-hospital mortality but did have higher adjusted odds of postoperative atrial fibrillation and pulmonary complications after cardiac surgery.
  • Meaning: Among patients undergoing cardiac surgery, not receiving postoperative β-blockers within the first 5 days after cardiac surgery or at any time in-hospital is not associated with in-hospital mortality and is associated with postoperative atrial fibrillation and pulmonary complications.

β-Blockers play an important role in postoperative patient care. Administration of β-blockers after cardiac surgery decreases arrhythmias, length of hospital stay, and mortality.1–4 Patients who receive β-blockers after noncardiac surgery experience lower rates of adverse cardiac events. Some studies have demonstrated higher mortality and cerebrovascular events in this population.5–9 However, β-blocker dosing varied, and adverse events may be attributed to excessive β-blocker dosing.5–9 Many organizations support the use of perioperative β-blockers for patients undergoing cardiac surgery.10–12

Our study investigates the use of β-blockers for postoperative atrial fibrillation prophylaxis after open cardiac surgery. Postoperative atrial fibrillation occurs after cardiac surgery, regardless of whether a coronary artery bypass grafting (CABG), valve, or combined procedure was performed. Approximately one-third of patients who undergo open aortic valve replacement develop postoperative atrial fibrillation.13 Postoperative atrial fibrillation after aortic valve replacement is associated with increased mortality.14 Perioperative administration of β-blockers has been shown to decrease the incidence of postoperative atrial fibrillation after aortic valve replacement.15 Some institutions implement β-blockers for postoperative atrial fibrillation prophylaxis by the first postoperative day (POD), in the absence of contraindications.15

The ideal time to administer β-blockers after cardiac surgery remains unclear. In chronic β-blocker users who undergo noncardiac surgery, resumption of β-blockers by the end of the first POD was associated with lower rates of new-onset in-hospital atrial fibrillation.15 At our institution, we administer β-blockers for atrial fibrillation prophylaxis once patients are hemodynamically stable after cardiac surgery.16–18 We predicted that patients who did not receive β-blockers by POD 5 would have increased in-hospital mortality and higher rates of postoperative atrial fibrillation, renal failure, cerebrovascular events, and pulmonary complications. We used POD 5 to define the early postoperative period because postoperative atrial fibrillation typically occurs between POD 2 through 5.19 In chronic β-blocker users undergoing cardiac surgery, we anticipated poor outcomes if they did not receive β-blockers by POD 5. For the subset of patients taking β-blockers before cardiac surgery, we derived our hypothesis from literature describing the increased rates of adverse cardiac events and mortality in patients whose chronic β-blockers are stopped.20–24

METHODS

We performed a retrospective cohort study after obtaining institutional review board (IRB) approval. The requirement for written informed consent was waived by our IRB. Our study and article adhere to the Enhancing the Quality and Transparency of health Research (EQUATOR) guidelines. We identified patients who underwent CABG, valve, or combined surgeries from January 1, 2013 to September 30, 2017. January 1, 2013 was the earliest date for available electronic medical records. We acquired procedural data and medication administration records from our institutional Epic electronic medical record. Preoperative and postoperative medication administration information was obtained from Epic and the Society of Thoracic Surgery (STS) database. From our electronic medical record, we received a list of medical record numbers with surgery dates and all β-blockers taken after surgery until discharge. β-Blockers were defined according to the STS database (Supplemental Digital Content, Table 1, http://links.lww.com/AA/D133). Efforts were made to obtain missing data from the STS database from the electronic medical record. Patients missing outcome data were omitted from the study. Patients scheduled for “other” types of cardiac surgery which includes surgery for congenital heart disease, ventricular assist devices, and transplants were excluded because our institutional postoperative β-blockade protocol does not apply to our study patient population.

For the primary analysis, we divided all patients presenting for CABG, valve, or combined surgery into groups according to the timing of β-blocker administration (Figure). In a subset analysis, we similarly divided the patients presenting for CABG, valve, or combined surgery who were taking preoperative β-blockers (Supplemental Digital Content, Figure 1, http://links.lww.com/AA/D133). The subset analysis was modeled after the β-blocker withdrawal analysis performed by Wallace et al.25 We used the STS definition of preoperative β-blocker status indicating that the medication had been given within 24 hours of surgical incision. Our institutional practice is to initiate β-blockers for all patients presenting for CABG or valve surgery, in the absence of contraindications, and to continue β-blockade in chronic users. The early group included patients who received β-blockers before or on POD 5. The absent or delayed group included patients who never received postoperative β-blockers while they were inpatients or received them after POD 5. The exposure variable will compare the absent or delayed group with the early group. Our primary outcome was in-hospital mortality. Secondary outcomes were postoperative cerebrovascular events, postoperative atrial fibrillation, pulmonary complications, and renal failure or new dialysis initiation.

Figure.
Figure.:
Flowchart of patient analysis (all patients). CABG indicates coronary artery bypass grafting; NYU, New York University; POD, postoperative day.

Confounding and outcome variables were defined per the STS database versions 2.9, 2.81, and 2.73. Confounding variables included preoperative risk factors (age, sex, preoperative ejection fraction, previous myocardial infarction, chronic lung disease, cerebrovascular disease, diabetes mellitus, dialysis), type of procedure (reoperation or not, CABG or not), preoperative inotrope use, postoperative tamponade, postoperative cardiac arrest, and presence of an intra-aortic balloon pump at any time. The data set included patient demographics, risk factors, preoperative, intraoperative, postoperative, and discharge information (Supplemental Digital Content, Table 2, http://links.lww.com/AA/D133). Some modifications were made to the original database variables. Postoperative cerebrovascular accident and postoperative transient ischemic attack were combined into a single variable denoting a postoperative cerebrovascular event. Prolonged mechanical ventilation and postoperative endotracheal reintubation were combined into the single variable postoperative pulmonary complications. Postoperative renal failure and postoperative renal dialysis were combined into a single variable denoting a postoperative renal complication. Reoperation for bleeding, valve replacement, graft dysfunction, and other reoperations were combined into a single variable denoting a reoperation. Records without a mortality status or without a specific procedure identified were excluded from the analysis.

Propensity scores for absent or delayed β-blocker administration were estimated using a generalized boosted model (GBM) based on preoperative characteristics and comorbidities and postoperative events.26 The GBM is a logistic model used here to model the probability of absent or delayed β-blocker administration. It generates the propensity score model iteratively by starting with a globally constant model and incrementally increases model complexity. In propensity score applications, the GBM complexity is calibrated by optimizing covariate balance between the inverse probability–weighted treatment and control samples. We used GBM because it does not assume linearity in unknown covariates.27 The late treatment was modeled here (Y = 1 = late).

Variables included in the propensity scores were age, sex, type of surgery (CABG or not), left ventricular ejection fraction on preoperative echocardiogram, history of prior myocardial infarction, chronic lung disease, history of cerebrovascular disease, history of diabetes mellitus, last creatinine level, history of dialysis, history of peripheral vascular disease, preoperative β-blocker use, reoperation status, cardiogenic shock, presence of preoperative inotropic agents, occurrence of cardiac tamponade, occurrence of cardiac arrest, and timing of intra-aortic balloon pump (none, intraoperative, postoperative). The number of iterations to achieve balance was visualized for each estimated propensity score. Balance was considered adequate for standardized mean difference <0.1. Covariate balance was assessed using standardized differences on the weighted samples.

Inverse probability of treatment weighting (IPTW) was used to show the achieved balance between exposure groups by applying the weights to patients when comparing at baseline. IPTW facilitates estimation of more types of treatment effects than propensity score matching, including the average treatment effect (ATE). Two common estimates are the ATE or average treatment effect on the treated (ATT). ATE looks at the estimated overall effect of the treatment, while ATT looks at the estimated effect on only those treated. Here, it is helpful that we can estimate the ATE using IPTW. We estimated ATE to get a better estimate of the relationship between β-blocker timing and outcomes across the whole sample. The twang package for the R statistical software (The R Foundation for Statistical Computing, Vienna, Austria) was utilized to estimate the propensity score.

Univariate logistic regression analyses were performed on the weighted sets using absent or delayed β-blocker as the independent variable. The primary outcome (in-hospital mortality) and each of the 4 secondary outcomes (postoperative cerebrovascular accident, postoperative atrial fibrillation, postoperative pulmonary complications, and renal failure or new dialysis initiation) were dependent variables in separate analyses. Patients who died after cardiac surgery were excluded from the analyses for secondary outcomes. Our data set included in-hospital mortality status but not the date of death, and we were unable to ascertain whether patients died before or after the POD of interest.

E-values were used to assess the magnitude of unmeasured confounders needed to explain away significant outcomes.28 The e-value estimates the magnitude of an unmeasured confounding variable (or set of variables) that would be needed to reduce an observed relative risk estimate to 1.0, and to reduce the confidence limit to 1.0 (either the upper or lower limit). For example, an E-value of 2.0 for reducing an observed relative risk to 1.0 means that there would need to be an unmeasured confounding variable which was associated with the exposure with a relative risk of 2.0 and also associated with the outcome with relative risk of 2.0, both after adjusting for the confounding variables already included in the analyses. A P value of .005 was set for significance due to multiple comparisons. Confidence intervals (CIs; 99.5%) were reported. All statistical operations were performed using the R statistical software (v. 3.1.1; The R Foundation for Statistical Computing).

To explore the relationship between timing of β-blocker administration after surgery and outcome, a set of additional analyses were performed in which IPTW was done using absent or delayed β-blocker administration within each of 1, 2, 3, 4, 5, and 6 days after surgery as the treatment variable. A separate propensity score model was fit for each of these different exposures. Like the primary analysis, each of these weighted sets was subsequently analyzed with univariate logistic regression using absent or delayed β-blocker as the independent variable and postoperative atrial fibrillation and postoperative pulmonary complications as dependent variables in 2 separate analyses. We did not have the exact dates for the secondary outcomes, only whether or not they occurred during each patient’s in-hospital course. For the sensitivity analyses for secondary outcomes, the exposure and outcome periods overlap.

The number of iterations to achieve balance was visualized for each estimated propensity score and was noted to be appropriate and that further iterations would likely not improve balance. Among 2526 patients, unstabilized and stabilized weights were calculated with the goal of estimating the ATE. We used ATE instead of ATT because ATE provides an estimate of the population-level effect, whereas ATT would be an estimate of effect of treatment only on those receiving the treatment, in this case absent or delayed β-blocker administration. After trimming the top and bottom 1% of propensity scores, the range of unstabilized weights was 1.029–8.109 and range of stabilized weights was 0.3386–2.95. Balance was achieved with absolute standardized differences <0.1. We used stabilized weights for the analyses. For treated cases, weight is P (T = 1)/P (T = 1|X). For untreated cases, the weight is 1 − P (T = 1)/(1 − P (T = 1|X)), where T is treatment and X is the vector of observed covariates.

RESULTS

Patients (3617) underwent cardiac surgery of any type at our institution from January 1, 2013 to August 30, 2017. Table 1 describes the characteristics of the patients included in the study. Our study included the 2526 patients who underwent CABG, valvular, or combined CABG and valvular surgeries. Preoperative β-blockade was given to 1671 (66%) patients within 24 hours before surgery. In decreasing order of frequency, 887 (35%) patients underwent CABG, 650 (26%) patients underwent aortic valve replacement, 590 (23%) patients underwent mitral valve repair, and 148 (5.8%) patients underwent mitral valve replacement. The average preoperative ejection fraction was 60% (55%–65%). Sixty-four (2.5%) were on inotropic agents within 48 hours before surgery. Eighty-three patients were supported by an intra-aortic balloon pump at any time during the perioperative period.

Table 1. - Patient Characteristics
Characteristics Total Patients = 2526
No. of Patients (% or Interquartile Range)
Patients on preoperative β-blockers 1671 (66)
Age, y 66 (58–74)
Sex (male) 1714 (68)
Type of surgery
 Isolated CABG 887 (35)
 Isolated aortic valve replacement 650 (26)
 Aortic valve replacement + CABG 137 (5.4)
 Isolated mitral valve replacement 148 (5.8)
 Mitral valve replacement + CABG 19 (0.7)
 Isolated mitral valve repair 590 (23)
 Mitral valve repair + CABG 42 (1.7)
 Aortic valve + mitral valve replacement 53 (2.1)
Preoperative ejection fraction (%) 60 (55–65)
Chronic lung disease 100 (4.0)
Diabetes mellitus 776 (31)
Cerebrovascular disease 235 (9.3)
Last serum creatine before surgery 0.9 (0.8–1.1)
History of dialysis 64 (2.5)
Reoperation 66 (2.6)
Preoperative cardiogenic shock 55 (2.2)
Preoperative inotrope 64 (2.5)
Intra-aortic balloon pump 83 (3.3)
Postoperative tamponade 3 (0.12)
Postoperative cardiac arrest 36 (1.4)
Abbreviation: CABG, coronary artery bypass grafting.

Of all patients undergoing cardiac surgery, 1695 (67%) received β-blockers during POD 0 to 5. Eight hundred thirty-one (33%) received no β-blockers at any time or received them after POD 5. For patients who received preoperative β-blockers within the 24 hours before surgery, 1078 (65%) received β-blockers during POD 0 to 5. Five hundred seventy-eight (35%) did not receive postoperative β-blockers or received them after POD 5. The median day to restart β-blockers was POD 4 [2–5].

Primary and Secondary Outcomes

Table 2 shows the results for the primary outcome of in-hospital in all patients. The patients who did not receive postoperative β-blockers by POD 5 did not have significantly increased adjusted odds ratios (ORs) of in-hospital mortality (OR = 1.6; 99.5% CI, 0.49–5.1; P = .28). Table 3 shows results for the secondary outcomes in all patients. In patients who did not receive postoperative β-blockers by POD 5, there was an increased adjusted OR for postoperative atrial fibrillation (OR = 1.5; 99.5% CI, 1.1–2.1; P < .001) and the calculated e-value was 2.4. In patients whose postoperative β-blockers were not given by POD 5, there was an increased association with postoperative pulmonary complications (OR = 3.0; 99.5% CI, 1.8–5.2; P < .001) and the calculated e-value was 5.6. Rates of postoperative cerebrovascular events (OR = 2.3; 99.5% CI, 0.82–6.4; P = .023), renal failure, or new need for dialysis (OR = 4.7; 99.5% CI, 0.53–41; P = .16) were not significantly different between the groups.

Table 2. - Odds Ratio for the Primary Outcome in All Patients
β-Blockers Given POD 0–5 (n = 1695) Absent or Delayed β-Blockers After POD 5 (n = 831) Adjusted OR (99.5% CI) Adjusted P Value
In-hospital mortality 10 (0.59%) 19 (2.3%) 1.6 (0.49–5.1) .28
Univariate logistic regression analyses were performed on the weighted sets using absent or delayed β-blocker administration as the independent variable. Propensity scores were estimated using a generalized boosted model based on preoperative characteristics and comorbidities and postoperative events. The number of iterations to achieve balance was visualized for each estimated propensity score. Balance was considered adequate for standardized mean difference <0.1. Covariate balance was assessed using standardized differences on the weighted samples. Inverse probability of treatment weighting was used to show the achieved balance between exposure groups by applying the weights to patients when comparing at baseline. A P value <.005 was set for significance due to multiple comparisons. CIs (99.5%) were reported. All statistical operations were performed using the R statistical software (v. 3.1.1; The R Foundation for Statistical Computing, Vienna, Austria).
Abbreviations: CI, confidence interval; OR, odds ratio; POD, postoperative day.

Table 3. - Odds Ratios for Secondary Outcomes in All Patients
β-Blockers Given POD 0–5 (n = 1682) Absent or Delayed β-Blockers After POD 5 (n = 815) Adjusted OR
(99.5% CI)
Adjusted P Value E-Value E-Value Lower CI
Postoperative cerebrovascular event 13 (0.78%) 22 (2.7%) 2.3 (0.82–6.4) .023
Postoperative atrial fibrillation 275 (16%) 196 (24%) 1.5 (1.1–2.1) <.001 2.4 1.5
Pulmonary complications 51 (3.0%) 102 (13%) 3.0 (1.8–5.2) <.001 5.6 3.0
Renal failure or new need for dialysis 1 (0.06%) 5 (0.61%) 4.7 (0.53–41) .16
Univariate logistic regression analyses were performed on the weighted sets using absent or delayed β-blocker as the independent variable. Propensity scores were estimated using a generalized boosted model based on preoperative characteristics and comorbidities and postoperative events. The number of iterations to achieve balance was visualized for each estimated propensity score. Balance was considered adequate for standardized mean difference <0.1. Covariate balance was assessed using standardized differences on the weighted samples. Inverse probability of treatment weighting was used to show the achieved balance between exposure groups by applying the weights to patients when comparing at baseline. A P value <.005 was set for significance due to multiple comparisons. CIs (99.5%) were reported. All statistical operations were performed using the R statistical software (v. 3.1.1; The R Foundation for Statistical Computing, Vienna, Austria). The e-value assesses the magnitude of an unmeasured confounding variable (or set of variables) that would be needed to reduce an observed relative risk estimate to 1.0 and to reduce the confidence limit to 1.0 (either the upper or lower limit). For example, an e-value of 2.0 for reducing an observed relative risk to 1.0 means that there would need to be an unmeasured confounding variable which was associated with the exposure with a relative risk of 2.0 and also associated with the outcome with relative risk of 2.0, both after adjusting for the confounding variables already included in the analyses.
Abbreviations: CI, confidence interval; OR, odds ratio; POD, postoperative day.

We performed the same analyses in the subset of patients who received β-blockers within the 24 hours before surgery, and the results are shown in Supplemental Digital Content, Tables 4–5, http://links.lww.com/AA/D133. Among patients on preoperative β-blockers who did not receive β-blockers by POD 5, we did not detect an increased adjusted ORs for the primary outcome of in-hospital mortality (OR = 1.3; 99.5% CI, 0.43–4.1; P = .63). Among patients on preoperative β-blockers who did not receive β-blockers by POD 5, there was an increased adjusted OR for postoperative atrial fibrillation (OR = 1.6; 99.5% CI, 1.1–2.4; P < .001) with an e-value of 2.7. Among patients on preoperative β-blockers who did not receive β-blockers by POD 5, there was an increased adjusted OR of postoperative pulmonary complications (OR = 2.8; 99.5% CI, 1.6–5.2; P < .001) and an e-value of 5.1. Significantly increased adjusted ORs of postoperative cerebrovascular events (OR = 2.7; 99.5% CI, 0.82–9.0; P = .01), renal failure, or new need for dialysis (OR = 3.1; 99.5% CI, 0.32–30; P = .32) were not detected in this patient subset.

Sensitivity Analysis

For the POD 1 group, we were unable to find a good match balance between the treated group and the untreated group and did not perform the analysis. Either the differences between the groups were too large or there were too few patients in the POD 1 group. When we used sensitivity analysis to investigate outcomes when β-blockers were given within POD 2 or 3, we did not detect a statistically significant difference in the adjusted ORs of postoperative atrial fibrillation or pulmonary complications (Supplemental Digital Content, Tables 3–6, http://links.lww.com/AA/D133). We found that patients who did not receive β-blockers by POD 4 had higher adjusted ORs of postoperative atrial fibrillation (OR = 1.4; 99.5% CI, 1.1–1.9; P < .001). Patients who did not receive β-blockers by POD 6 also had increased adjusted ORs of postoperative atrial fibrillation (OR = 1.5; 99.5% CI, 1.1–2.0; P < .001). Among patients who did not receive β-blockers by POD 2 through 6, increased ORs for pulmonary complications were detected (Supplemental Digital Content, Table 6, http://links.lww.com/AA/D133). The observed significant ORs could be explained away by an unmeasured confounder that was associated with both the treatment and the outcome by an OR of the reported e-values and lower CI e-values, but weaker confounding could not do so.

DISCUSSION

Our study illustrates how delaying or withholding β-blocker administration after cardiac surgery corresponds with higher adjusted odds of postoperative atrial fibrillation and pulmonary complications. Our sensitivity analysis suggests that each day in delaying β-blockers beyond POD 4 or thereafter corresponds with increasing adjusted ORs of postoperative atrial fibrillation. A recent study of chronic β-blocker users who had noncardiac surgery had similar findings.15 Noncardiac surgical patients who resumed taking their chronic β-blockers by the first POD had decreased rates of postoperative atrial fibrillation.15 Our results may reflect the residual confounding effect of hemodynamic instability in patients who do not receive β-blockers during the early postoperative period, defined in our study as POD 0 to 5. We were not able to adjust for hemodynamic instability in patients who required perioperative vasoactive medications, pacing, or mechanical circulatory support other than intra-aortic balloon pumps. Whether there is a direct effect of β-blockers on rates of postoperative atrial fibrillation or pulmonary complications or whether the timing of β-blocker administration after cardiac surgery is a marker of hemodynamic instability warrants further study.

Our study suggests an association between delaying or discontinuing β-blockers with postoperative respiratory complications including prolonged mechanical ventilation or postoperative respiratory distress requiring intubation. Risk factors for extended postoperative mechanical ventilation include depressed left ventricular ejection fraction, increased cardiopulmonary bypass times, and lactic acidosis.29–33 Clinicians at our institution postpone administering β-blockers to patients with these characteristics. Our practice is to avoid premature extubation in hemodynamically unstable patients, the same group of patients who would also have contraindications for β-blockers.34

How early β-blocker administration prevents respiratory complications after cardiac surgery is unclear, but many studies demonstrate this association. Cardioselective β-blockers have been shown to benefit patients with chronic obstructive pulmonary disease and are not associated with disease exacerbations.35 β-Blocker administration is associated with improved survival in patients with chronic obstructive pulmonary disease.36 β-Blockers may confer a mortality benefit in critically ill patients with acute respiratory failure.37 β-Blockers’ respiratory benefits may stem from their ability to blunt the postoperative stress response mediated by the sympathetic nervous system.36 More studies are needed to explain why.

While we did not compare outcomes for patients on preoperative β-blockers directly with patients without preoperative β-blockers, our results for patients on preoperative β-blockers were similar to results for all patients combined. In the group of patients who received preoperative β-blockers and did not receive β-blockers by POD 5, the adjusted OR for postoperative atrial fibrillation was 1.6 (OR = 1.6; 99.5% CI, 1.1–2.4; P < .001) and for pulmonary complications, 2.8 (OR = 2.8; 99.5% CI, 1.6–5.2; P < .001). These results resemble the adjusted ORs for postoperative atrial fibrillation and pulmonary complications obtained from the entire group of patients undergoing cardiac surgery regardless of preoperative β-blocker status. Our results suggest that administration of preoperative β-blockers may not confer additional benefit for cardiac surgical patients.38

Our study strengths include the large cohort of patients at our institution and novel clinical questions investigating the timing of postoperative β-blockade. Limitations of the study include its retrospective nature, the exclusion of patients who died after cardiac surgery from the secondary outcome analyses, the small number of patients with in-hospital mortality, postoperative cerebrovascular events, and renal failure or new need for dialysis. The estimated CIs should be given more importance than the P values. For our primary outcome of in-hospital mortality, the 99.5% CI for the adjusted OR was 0.49 to 5.1. Our best estimate of the true effect of absent or delayed postoperative β-blockers falls within these limits. We were unable to control for the presence of postoperative inotropic and vasopressor agents and pacing due to inability to obtain these data from the electronic medical record. For the sensitivity analyses for secondary outcomes, exposure and outcome periods overlap and make inference difficult. The outcome events could have occurred before or after postoperative β-blocker administration. Delayed or absent postoperative β-blocker administration may not necessarily cause atrial fibrillation or pulmonary complications. Future studies could investigate reasons why β-blockers are delayed after cardiac surgery and how they affect pulmonary function.

In summary, patients who do not receive postoperative β-blockers within the first 5 days after cardiac surgery or at any time in-hospital did not have significantly increased adjusted ORs of in-hospital mortality. Patients who do not receive postoperative β-blockers by POD 4 through 6 have higher adjusted odds of postoperative atrial fibrillation and pulmonary complications after cardiac surgery. These associations persist whether or not patients received β-blockers before cardiac surgery. Each day in delaying β-blockers after cardiac surgery from POD 4 through 6 corresponded with increasing adjusted ORs of postoperative atrial fibrillation.

DISCLOSURES

Name: Emily L. Chanan, MD.

Contribution: This author helped conceive and design the study, acquire and interpret the data, and draft and finalize the manuscript.

Name: Samir M. Kendale, MD.

Contribution: This author helped analyze and interpret the data and draft the manuscript.

Name: Germaine Cuff, PhD.

Contribution: This author helped conceive and design the study and acquire the data.

Name: Aubrey C. Galloway, MD.

Contribution: This author helped conceive and design the study.

Name: Mark E. Nunnally, MD.

Contribution: This author helped conceive and design the study and draft the manuscript.

This manuscript was handled by: Stefan G. De Hert, MD.

FOOTNOTES

    REFERENCES

    1. Blessberger H, Kammler J, Domanovits H, et al. Perioperative beta-blockers for preventing surgery-related mortality and morbidity. Cochrane Database Syst Rev. 2018;3:CD004476.
    2. Sun J, Ding Z, Qian Y. Effect of short-acting beta blocker on the cardiac recovery after cardiopulmonary bypass. J Cardiothorac Surg. 2011;6:99.
    3. Crystal E, Connolly SJ, Sleik K, Ginger TJ, Yusuf S. Interventions on prevention of postoperative atrial fibrillation in patients undergoing heart surgery: a meta-analysis. Circulation. 2002;106:75–80.
    4. Chan AY, McAlister FA, Norris CM, Johnstone D, Bakal JA, Ross DB; Alberta Provincial Program for Outcome Assessment in Coronary Heart Disease (APPROACH) Investigators. Effect of beta-blocker use on outcomes after discharge in patients who underwent cardiac surgery. J Thorac Cardiovasc Surg. 2010;140:182–187, 187.e1.
    5. Bouri S, Shun-Shin MJ, Cole GD, Mayet J, Francis DP. Meta-analysis of secure randomised controlled trials of β-blockade to prevent perioperative death in non-cardiac surgery. Heart. 2014;100:456–464.
    6. Stevens RD, Burri H, Tramèr MR. Pharmacologic myocardial protection in patients undergoing noncardiac surgery: a quantitative systematic review. Anesth Analg. 2003;97:623–633.
    7. Devereaux PJ, Yang H, Yusuf S, et al.; POISE Study Group. Effects of extended-release metoprolol succinate in patients undergoing non-cardiac surgery (POISE trial): a randomised controlled trial. Lancet. 2008;371:1839–1847.
    8. Bangalore S, Wetterslev J, Pranesh S, Sawhney S, Gluud C, Messerli FH. Perioperative beta blockers in patients having non-cardiac surgery: a meta-analysis. Lancet. 2008;372:1962–1976.
    9. Lindenauer PK, Pekow P, Wang K, Mamidi DK, Gutierrez B, Benjamin EM. Perioperative beta-blocker therapy and mortality after major noncardiac surgery. N Engl J Med. 2005;353:349–361.
    10. Kulik A, Ruel M, Jneid H, et al.; American Heart Association Council on Cardiovascular Surgery and Anesthesia. Secondary prevention after coronary artery bypass graft surgery: a scientific statement from the American Heart Association. Circulation. 2015;131:927–964.
    11. Montalescot G, Sechtem U, Achenbach S, et al.; Task Force Members; ESC Committee for Practice Guidelines; Document Reviewers. 2013 ESC guidelines on the management of stable coronary artery disease: the Task Force on the management of stable coronary artery disease of the European Society of Cardiology. Eur Heart J. 2013;34:2949–3003.
    12. Shahian DM, O’Brien SM, Sheng S, et al. Predictors of long-term survival after coronary artery bypass grafting surgery: results from the Society of Thoracic Surgeons Adult Cardiac Surgery Database (the ASCERT study). Circulation. 2012;125:1491–1500.
    13. Yokota J, Nishi H, Sekiya N, Yamada M, Takahashi T. Atrial fibrillation following aortic valve replacement: impact of perioperative use of intravenous β-blocker. Gen Thorac Cardiovasc Surg. 2017;65:194–199.
    14. Filardo G, Hamilton C, Hamman B, Hebeler RF Jr, Adams J, Grayburn P. New-onset postoperative atrial fibrillation and long-term survival after aortic valve replacement surgery. Ann Thorac Surg. 2010;90:474–479.
    15. Khanna AK, Naylor DF Jr, Naylor AJ, et al. Early resumption of β blockers is associated with decreased atrial fibrillation after noncardiothoracic and nonvascular surgery: a cohort analysis. Anesthesiology. 2018;129:1101–1110.
    16. January CT, Wann LS, Alpert JS, et al.; ACC/AHA Task Force Members. 2014 AHA/ACC/HRS guideline for the management of patients with atrial fibrillation: executive summary: a report of the American College of Cardiology/American Heart Association Task Force on practice guidelines and the Heart Rhythm Society. Circulation. 2014;130:2071–2104.
    17. Arsenault KA, Yusuf AM, Crystal E, et al. Interventions for preventing post-operative atrial fibrillation in patients undergoing heart surgery. Cochrane Database Syst Rev. 2013:CD003611.
    18. Ha AC, Mazer CD, Verma S, Yanagawa B, Verma A. Management of postoperative atrial fibrillation after cardiac surgery. Curr Opin Cardiol. 2016;31:183–190.
    19. Banach M, Goch A, Misztal M, Rysz J, Jaszewski R, Goch JH. Predictors of paroxysmal atrial fibrillation in patients undergoing aortic valve replacement. J Thorac Cardiovasc Surg. 2007;134:1569–1576.
    20. Kwon S, Thompson R, Florence M, et al.; Surgical Care and Outcomes Assessment Program (SCOAP) Collaborative. β-blocker continuation after noncardiac surgery: a report from the surgical care and outcomes assessment program. Arch Surg. 2012;147:467–473.
    21. Wallace A, Layug B, Tateo I, et al. Prophylactic atenolol reduces postoperative myocardial ischemia. McSPI Research Group. Anesthesiology. 1998;88:7–17.
    22. Psaty BM, Koepsell TD, Wagner EH, LoGerfo JP, Inui TS. The relative risk of incident coronary heart disease associated with recently stopping the use of beta-blockers. JAMA. 1990;263:1653–1657.
    23. Neumann A, Maura G, Weill A, Alla F, Danchin N. Clinical events after discontinuation of β-blockers in patients without heart failure optimally treated after acute myocardial infarction. Circ Cardiovasc Qual Outcomes. 2018;11:e004356.
    24. Alharbi FF, Souverein PC, de Groot MC, Maitland-van der Zee AH, de Boer A, Klungel OH. Risk of acute myocardial infarction after discontinuation of antihypertensive agents: a case-control study. J Hum Hypertens. 2017;31:537–544.
    25. Wallace AW, Au S, Cason BA. Association of the pattern of use of perioperative β-blockade and postoperative mortality. Anesthesiology. 2010;113:794–805.
    26. Schulte PJ, Mascha EJ. Propensity score methods: theory and practice for anesthesia research. Anesth Analg. 2018;127:1074–1084.
    27. Setodji CM, McCaffrey DF, Burgette LF, Almirall D, Griffin BA. The right tool for the job: choosing between covariate-balancing and generalized boosted model propensity scores. Epidemiology. 2017;28:802–811.
    28. VanderWeele TJ, Ding P. Sensitivity analysis in observational research: introducing the E-value. Ann Intern Med. 2017;167:268–274.
    29. Wise ES, Stonko DP, Glaser ZA, et al. Prediction of prolonged ventilation after coronary artery bypass grafting: data from an artificial neural network. Heart Surg Forum. 2017;20:E007–E014.
    30. Sharma V, Rao V, Manlhiot C, Boruvka A, Fremes S, Wąsowicz M. A derived and validated score to predict prolonged mechanical ventilation in patients undergoing cardiac surgery. J Thorac Cardiovasc Surg. 2017;153:108–115.
    31. Reddy SL, Grayson AD, Griffiths EM, Pullan DM, Rashid A. Logistic risk model for prolonged ventilation after adult cardiac surgery. Ann Thorac Surg. 2007;84:528–536.
    32. Hessels L, Coulson TG, Seevanayagam S, et al. Development and validation of a score to identify cardiac surgery patients at high risk of prolonged mechanical ventilation. J Cardiothorac Vasc Anesth. 2019;33:2709–2716.
    33. Hsu H, Lai HC, Liu TJ. Factors causing prolonged mechanical ventilation and peri-operative morbidity after robot-assisted coronary artery bypass graft surgery. Heart Vessels. 2019;34:44–51.
    34. Grant MC, Isada T, Ruzankin P, et al.; Johns Hopkins Enhanced Recovery Program for the Cardiac Surgery Working Group. Results from an enhanced recovery program for cardiac surgery. J Thorac Cardiovasc Surg. 2020;159:1393–1402.e7.
    35. Duffy S, Marron R, Voelker H, et al.; NIH COPD Clinical Research Network and the Canadian Institutes of Health Research. Effect of beta-blockers on exacerbation rate and lung function in chronic obstructive pulmonary disease (COPD). Respir Res. 2017;18:124.
    36. Etminan M, Jafari S, Carleton B, FitzGerald JM. Beta-blocker use and COPD mortality: a systematic review and meta-analysis. BMC Pulm Med. 2012;12:48.
    37. Noveanu M, Breidthardt T, Reichlin T, et al. Effect of oral β-blocker on short and long-term mortality in patients with acute respiratory failure: results from the BASEL-II-ICU study. Crit Care. 2010;14:R198.
    38. Kohsaka S, Miyata H, Motomura N, et al. Effects of preoperative β-blocker use on clinical outcomes after coronary artery bypass grafting: a report from the Japanese cardiovascular surgery database. Anesthesiology. 2016;124:45–55.

    Supplemental Digital Content

    Copyright © 2020 International Anesthesia Research Society