If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Address correspondence to Amy Valkovec, PharmD, Duke University Hospital, Department of Pharmacy, 2301 Erwin Road, Duke University Medical Center 3089, Durham, NC 27710
The altered pharmacokinetics of milrinone in renal impairment could result in an increased risk of cardiac arrhythmias. This study aimed to determine if there is an association between new-onset arrhythmias and renal impairment after cardiac surgery following milrinone administration.
Design
A retrospective cohort study.
Setting
A single-center tertiary care hospital.
Participants
Adult patients who received a milrinone infusion in the intensive care unit (ICU) setting after coronary artery bypass graft, valvuloplasty, annuloplasty, or a combination of these surgeries from July 1, 2014 to July 1, 2021. Renal impairment was defined using a creatinine clearance <60 mL/min, calculated using the Cockcroft-Gault equation.
Interventions
Patients received a weight-based continuous intravenous infusion of milrinone.
Measurements and Main Results
The primary outcome was the presence of new arrhythmias after the initial administration of a weight-based continuous intravenous infusion of milrinone postcardiac surgery. Of the 197 patients who met inclusion, there was no difference in the presence of new arrhythmias (42.9% v 40.3%, p = 0.76) or in the time to first new arrhythmia from milrinone initiation in those with renal impairment compared to those without renal impairment (29.1 hours v 33.3 hours, p = 0.54). Patients with renal impairment had a longer hospital stay than patients without renal impairment (17.5 days v 13.9 days, p = 0.016). Arrhythmia type, length of ICU stay, ICU mortality, and hospital mortality were not different between the cohorts.
Conclusions
There was no association between new arrhythmias, milrinone, and renal impairment in patients postcardiac surgery.
PATIENTS WHO REQUIRE cardiac surgery are at an increased risk of developing low-cardiac-output syndrome that can continue for days after surgery, potentially leading to critical organ hypoperfusion and cardiogenic shock.
Treatment options to reverse this acute low-output state include volume expansion, vasopressors, inodilators, and mechanical circulatory support devices.
Milrinone is a type III phosphodiesterase inhibitor that enhances cardiac output and improves hemodynamics by increasing cyclic adenosine monophosphate, increasing calcium in cardiac myocytes, augmenting contractility, and producing vasodilation in pulmonary and systemic vascular beds.
Milrinone has a prolonged effect in patients with renal impairment, with concern for potential arrhythmias because milrinone is predominantly renally eliminated.
Some practitioners avoid using milrinone in patients with renal dysfunction because of the increased elimination half-life and the potential for elevated plasma concentrations of milrinone.
Studies of milrinone have not evaluated the potential association of prolonged half-life and plasma concentration in patients with renal impairment and an increase in new cardiac arrhythmias.
This study aimed to determine if there was a difference in new cardiac arrhythmias between patients with renal impairment and patients without renal impairment who received a weight-based intravenous (IV) infusion of milrinone after cardiac surgery.
Methods
A retrospective cohort study of adult patients, who received an infusion of milrinone after a coronary artery bypass graft (CABG), valvuloplasty, annuloplasty, or a combination of these cardiac surgeries, was conducted at a single-center tertiary-care hospital from July 1, 2014 to July 1, 2021. Patients were included if they received a weight-based continuous infusion of IV milrinone for at least 12 hours during the index encounter.7 Data were collected on the first infusion of milrinone for the encounter, deemed the first if it was started intraoperatively or after the index cardiac surgery. Patients were excluded if they had a history of cardiac arrhythmia prior to surgery, experienced a cardiac arrhythmia after surgery but prior to milrinone initiation, received an antiarrhythmic within 5 half-lives (Supplementary Appendix Table 1), had an out-of-hospital cardiac arrest, received milrinone within 12 hours prior to cardiac surgery, received a loading dose of milrinone, used dobutamine within 1 hour prior to administration of milrinone or concurrently with milrinone, received a heart or lung transplant, received a permanent ventricular assist device, or had congenital heart disease. Patients were identified using an institutional data repository, and data were collected through chart review by A.M.V. and validation by S.J.K. The Institutional Review Board approved the study protocol with a waiver of informed consent.
The primary endpoint was new-onset cardiac arrhythmia, defined as atrial fibrillation, atrial flutter, atrial tachycardia, ventricular tachycardia, or ventricular fibrillation lasting >30 seconds or requiring an intervention.
The presence of an arrhythmia was determined from chart review of vitals flow sheets. This was cross-referenced with a 12-lead electrocardiogram, if available, and a second author reviewed the chart if there was any uncertainty regarding the arrhythmia. New-onset arrhythmia was identified as occurring after the first infusion of milrinone and within 12 hours after the end of the infusion.
The secondary endpoints were the time (in hours) to onset of the first new cardiac arrhythmia, type of arrhythmia that occurred, length of hospital stay, length of intensive care unit (ICU) stay, ICU mortality, hospital mortality, and cause of death. The interventions were defined as the administration of an antiarrhythmic, cardioversion or defibrillation, initiation of mechanical circulatory support, or titration of inotropes or vasoactive medications. The infusion rate of milrinone immediately prior to the arrhythmia was collected as the rate most closely charted to the onset of the arrhythmia. The infusion rates for milrinone and vasopressors were captured at the initiation of the milrinone infusion and the onset of arrhythmia in order to calculate the vasoactive-inotropic score, a weighted sum of all administered inotropes and vasoactive medications.
Serum creatinine was collected at the start and end of the milrinone infusion, prior to the onset of the arrhythmia, and at discharge from the ICU or death. Patients originally were identified as having renal impairment if their creatinine clearance was <30 mL/min, calculated using the Cockcroft-Gault equation. However, the original protocol was amended to identify patients as having renal impairment if their creatinine clearance was <60 mL/min.
Descriptive statistics were used to characterize the study population's baseline characteristics. Comparison between groups with continuous outcomes was performed with Student's t test or Wilcoxon rank sum test for parametric or nonparametric data, respectively. Comparison between groups with binary outcomes was performed with the Fisher's Exact test. Comparison between groups with ordinal or nominal outcomes was performed with the Pearson chi-square test. A multivariate binomial logistic regression with adaptive Elastic Net estimation and corrected Akaike information criterion was performed to determine the parameter impact on the presence of new cardiac arrhythmias. Parameters were chosen that maximized number of model parameters with the lowest Akaike information criterion value. The time to first new cardiac arrhythmia was analyzed using a Kaplan-Meier estimate with the Wilcoxon method, censoring noncardiac death prior to the end of milrinone infusion. A Cox proportional hazards estimate was performed using variables identified with the regression model described above. No adjustment for repetitive analyses was made as the analysis is hypothesis-generating. An a priori sensitivity analysis of the impact of renal dysfunction classifications was performed. The power for this study was calculated using a 58% estimated risk of postoperative atrial fibrillation in patients receiving milrinone.
An a priori power analysis determined a total of 190 patients would be needed to have 80% power to detect a 20% difference in arrhythmias with the use of a 2-sided alpha level of 0.05. Statistical analysis was performed using JMP Pro, version 16.2.0 (SAS Institute Inc, Cary, NC).
Results
A total of 684 patients were treated with milrinone postcardiac surgery, and 197 were included in the study (Fig 1). The most common reasons for exclusion were previous arrhythmia (362 patients) and the administration of an antiarrhythmic medication prior to receiving milrinone (61 patients). The 2 cohorts were well-matched for the type of cardiac surgery performed, ethnicity, median left ventricular ejection fraction (LVEF), and median milrinone infusion duration (Table 1). The cohort of patients with renal impairment was older, had a higher percentage of female patients, and had a lower body mass index than the cohort of patients without renal impairment (Table 1). More patients in the group with renal impairment had hypertension and diabetes mellitus, and this cohort also had a higher median European System for Cardiac Operative Risk Evaluation II score (EuroSCORE II) (Table 1). The baseline median creatinine clearance for the cohort with renal impairment was 43.7 mL/min (IQR 36.6-51.8), compared to 87.4 mL/min (IQR 73.6-110) for the cohort without renal impairment (p < 0.001), and these differences remained at both the time the milrinone infusion started, and the time the infusion ended (46.2 mL/min v 91.4 mL/min and 47.7 mL/min v 85.4 mL/min, respectively, p < 0.001). The median initial milrinone infusion was the same between groups, but the IQR varied with a statistical significance (Table 2). In addition, the median duration of the milrinone infusion and the maximum milrinone infusion rate were similar between groups (Table 2).
A new arrhythmia occurred in 27 patients (42.9%) with renal impairment and 54 patients (40.3%) without renal impairment (p = 0.76; Table 3). The a priori sensitivity analysis of the impact of renal dysfunction classifications is shown in the Supplementary Appendix Table 2 After controlling for prespecified covariates, renal impairment was not associated with the development of new-onset postoperative arrhythmias, but the milrinone infusion duration was (Tables 4 and 5). At the time the first arrhythmia occurred, the milrinone infusion rate was higher in patients without renal impairment compared to those with renal impairment (Table 3).
Model parameters: age >70 years, male sex, left ventricular ejection fraction <40%, mitral valve disease, obesity, chronic renal failure, withdrawal of beta blocker in immediate postoperative period, obstructive sleep apnea, diabetes, chronic obstructive pulmonary disease, valve surgery, EuroSCORE II, Vasopressor-inotrope score, renal impairment prior to milrinone infusion, maximum milrinone infusion rate, and milrinone infusion duration.
Variable
Estimate (Beta)
SE
F-ratio
95% CI
p
Milrinone infusion duration, h
0.007
0.003
5.44
0.002 to 0.01
0.006
Milrinone dosing weight, kg
−0.002
0.009
0.05
−0.02 to 0.02
0.84
EuroSCORE II
0.04
0.03
1.8
−0.02 to 0.1
0.24
NOTE. Model Akaike information criterion = 163.
Abbreviations: EuroSCORE, European System for Cardiac Operative Risk Evaluation.
Model parameters: age >70 years, male sex, left ventricular ejection fraction <40%, mitral valve disease, obesity, chronic renal failure, withdrawal of beta blocker in immediate postoperative period, obstructive sleep apnea, diabetes, chronic obstructive pulmonary disease, valve surgery, EuroSCORE II, Vasopressor-inotrope score, renal impairment prior to milrinone infusion, maximum milrinone infusion rate, and milrinone infusion duration.
There was no difference in the time to first new arrhythmia from milrinone initiation in patients with renal impairment compared to those without renal impairment (Table 3; Fig 2). Patients with renal impairment had a longer hospital stay than patients without renal impairment (Table 3). Arrhythmia type, length of ICU stay, ICU mortality, and hospital mortality were not different between patients with and without renal impairment (Table 3). All patients who died were in the ICU at the time of death (Table 3). There were 4 deaths in the renal impairment group and 2 deaths in the group without renal impairment. Only 1 death in the renal impairment group was related to arrhythmia, with the cause of death identified as a lack of neurologic recovery postventricular fibrillation arrest (Table 3).
Fig 2Time to Arrhythmia after Milrinone Infusion Initiation. Kaplan-Meier Time-to-Event Analysis. New cardiac arrhythmia defined as lasting > 30 seconds or requiring an intervention censored for noncardiac death prior to the end of the milrinone infusion.
No significant differences were found in the incidence of new arrhythmias postcardiac surgery after the administration of a weight-based continuous infusion of milrinone in patients with renal impairment compared to patients without renal impairment. Only 1 parameter, milrinone infusion duration, was associated with an increased risk of arrhythmias. Hospital length of stay and arrhythmia duration were the only clinical outcomes associated with renal impairment.
The lack of difference in arrhythmias between patients with renal impairment and those without renal impairment suggested that both groups were equally likely to experience an arrhythmia due to myocardial dysfunction regardless of exposure to milrinone. A recent prospective trial found that up to 50% of patients with cardiogenic shock treated with dobutamine or milrinone experienced an arrhythmia leading to medical team intervention, suggesting these patients have an increased risk of arrhythmias regardless of inodilator selection.
In a retrospective analysis of the OPTIME-CHF study using blood urea nitrogen to determine renal function in patients receiving milrinone, there was no statistically significant difference in the incidences of atrial or ventricular arrhythmias.
Admission or changes in renal function during hospitalization for worsening heart failure predict postdischarge survival: Results from the Outcomes of a Prospective Trial of Intravenous Milrinone for Exacerbation of Chronic Heart Failure (OPTIME-CHF).
Admission or changes in renal function during hospitalization for worsening heart failure predict postdischarge survival: Results from the Outcomes of a Prospective Trial of Intravenous Milrinone for Exacerbation of Chronic Heart Failure (OPTIME-CHF).
In the authors' study, the baseline LVEF was low between both cohorts of patients, with a median of 35%, and at least 79% of patients in each cohort had heart failure at baseline (Table 1). In 1 study, patients with an ejection fraction <35% were associated with a 5-fold increased risk of sustained postoperative ventricular tachycardia-fibrillation, and LVEF was identified as the strongest risk factor for new-onset sustained postoperative ventricular tachycardia-fibrillation.
Atrial fibrillation has been reported in up to 40% of patients in the postoperative period after CABG, up to 37%-to-50% of patients after valve surgery, and up to 60% of patients with a combination of CABG and valve replacement.
In this study, all patients underwent cardiac surgery, which would place them at a similar risk for arrhythmia based on the procedure. Another possible explanation for lack of difference in arrhythmias between the cohorts in this study was that regardless of renal dysfunction and exposure to milrinone, patients in a low-cardiac-output state postcardiac surgery could be at an increased risk of cardiac arrhythmia due to myocardial dysfunction.
The median milrinone infusion duration was 75.6 hours in the study, which was consistent with the 2 large randomized controlled trials that had a median milrinone infusion duration of 36-to-72 hours.
Admission or changes in renal function during hospitalization for worsening heart failure predict postdischarge survival: Results from the Outcomes of a Prospective Trial of Intravenous Milrinone for Exacerbation of Chronic Heart Failure (OPTIME-CHF).
Both the multiple logistic regression model and Cox proportional hazards model identified milrinone duration as the only parameter associated with cardiac arrhythmias. Milrinone infusion duration may be associated with arrhythmias due to an accumulation of the drug over time. Increased exposure to milrinone postcardiac surgery also may place patients at an increased risk of arrhythmias regardless of renal function. This could be investigated in follow-up studies.
The initial milrinone infusion rate was statistically different between the groups, but this was not clinically significant. The recommended dose adjustments for patients with a creatinine clearance of 50 mL/min or less ranges from 0.2-to-0.43 µg/kg/min, which is above the median initial infusion rate for this study.
Patients without renal impairment weighed more than patients with renal impairment and, therefore, received higher overall doses in µg/min (Table 2). This disparity could have increased the incidence of arrhythmias, but both the multiple logistic regression model and Cox proportional hazards model did not identify milrinone dosing weight as a factor predicting arrhythmias (Tables 4 and 5). Dosing weight refers to the patient's weight that was entered into the infusion pump and used to calculate the continuous infusion of milrinone. Due to the increased elimination half-life and plasma concentration of milrinone in patients with renal impairment, the difference in milrinone dose due to weight could still mean both cohorts had equal milrinone exposure. Plasma milrinone concentrations were not measured to identify variations between the cohorts. However, milrinone concentrations are not routinely used in clinical practice for monitoring patients receiving milrinone. Patients enrolled in OPTIME-CHF received 0.5 µg/kg/min of milrinone, whereas patients in DOREMI were initiated on 0.125-to->0.5 µg/kg/min, and titrated according to a scoring tool.
Admission or changes in renal function during hospitalization for worsening heart failure predict postdischarge survival: Results from the Outcomes of a Prospective Trial of Intravenous Milrinone for Exacerbation of Chronic Heart Failure (OPTIME-CHF).
The multivariate binomial logistic regression model adjusting for these comorbidities found that none of the parameters was associated significantly with the development of new-onset postoperative cardiac arrhythmias (Table 4). Previous literature has identified patients >70 years as having the highest risk for postoperative atrial fibrillation.
Patients with renal impairment in this study exceeded this age, with a median of 71 years, which placed this cohort at an increased risk of arrhythmias. This could have served as a confounder for the study, but no single risk factor was identified as significant, using the multivariate binomial logistic regression model. The lack of signal from the multivariate binomial logistic regression model, in combination with the complexity of postoperative arrhythmias, demonstrated that differences in the baseline characteristics between the groups did not likely serve as confounders for the incidence of cardiac arrhythmias in this study. Overall, the study had a lower incidence of atrial fibrillation than expected, with only 61 (31%) patients experiencing atrial fibrillation postcardiac surgery.
The power for this study was calculated a priori using the risk of postoperative atrial fibrillation in patients receiving milrinone.
However, this study only excluded patients with chronic atrial fibrillation, whereas the authors' study excluded patients who experienced any arrhythmia or were treated with an antiarrhythmic within 5 half-lives prior to milrinone infusion.
In DOREMI, which reported up to 50% of patients with cardiogenic shock treated with dobutamine or milrinone experienced an arrhythmia, at least 48% of patients in each treatment group had atrial fibrillation at baseline.
Excluding patients with arrhythmias at baseline may have decreased the overall incidence of arrhythmia postcardiac surgery observed in the authors' study.
This study had several limitations. First, this was a single-center retrospective study. The institution is a large, academic medical center that performs high-risk cardiac surgeries. The results could have been different if this study had been performed at other institutions with a low-risk patient population. Second, patients were identified as having renal dysfunction if their creatinine clearance was <30 mL/min, calculated using the Cockcroft-Gault equation in the original protocol. Because only 9 patients included in this study fell into this category, the protocol was amended to identify patients as having renal impairment if their creatinine clearance was <60 mL/min. A pharmacokinetic study of milrinone concentrations in patients with stage-D heart failure with and without renal dysfunction found that patients with a creatinine clearance between 30 mL/min and 60 mL/min had a mean milrinone concentration of 591 ng/mL, whereas those with a creatinine clearance <30 mL/min had a significant increase to 1,575 ng/mL.
For patients receiving thrice weekly hemodialysis, the predialysis milrinone level was also 20 times higher than the upper limit of the therapeutic index.
Although these milrinone plasma concentrations were above the upper limit of the therapeutic index of 500 ng/mL, the increased accumulation of milrinone in patients with creatinine clearance <30 mL/min could have impacted the primary outcome of this study if more patients in this category were eligible for inclusion. Additionally, at least 29% of patients in each cohort experienced atrial fibrillation, which could have resulted from left atrial dysfunction, and this was not measured in the study. Next, although many risk factors for arrhythmia after cardiac surgery were included in the multivariate logistic regression, it was not feasible to include every risk factor. Finally, fewer patients than anticipated experienced a new-onset arrhythmia, and the study did not meet power. A post hoc power analysis determined that 9,762 patients would be required to have 80% power to detect a difference in arrhythmias with the use of a 2-sided alpha level of 0.05. Conducting a study of this magnitude was beyond the scope of a single-center study.
In a population of adults who received a continuous IV infusion of milrinone postcardiac surgery, there was no association between renal impairment and new cardiac arrhythmias. However, large, multicenter studies are needed to further evaluate if the altered pharmacokinetics of milrinone in renal impairment affect clinical outcomes in this population.
Admission or changes in renal function during hospitalization for worsening heart failure predict postdischarge survival: Results from the Outcomes of a Prospective Trial of Intravenous Milrinone for Exacerbation of Chronic Heart Failure (OPTIME-CHF).
The use of milrinone in patients with renal insufficiency is controversial because the drug is excreted exclusively by the kidney.1 As such, anesthesiologists may avoid milrinone in patients experiencing post-cardiotomy shock with renal impairment because of fears about drug accumulation and subsequent tachyarrhythmias, myocardial ischemia, and death—a notion perpetuated in the 2022 American College of Cardiology/American Heart Association guidelines.2,3 However, is this practice truly evidence-based?
Multiple arrhythmia types possible.
00850-3&id=ga1.jpg){kind=link}
00850-3&id=gr1.jpg){kind=link}
00850-3&id=gr2.jpg){kind=link}