Safety and Efficacy of Apixaban Following Bioprosthetic Valve Replacements: A Retrospective Evaluation

Abstract

Background: Data on the use of direct oral anticoagulants in the setting of bioprosthetic valve replacements is limited. Objective: The purpose of this study was to describe outcomes among patients who underwent a bioprosthetic valve replacement and were subsequently prescribed apixaban. Methods: A retrospective cohort study was completed for inpatients at a community hospital who were prescribed apixaban following a bioprosthetic valve replacement from 2015 to 2020. Endpoints assessed included incidence of all-cause readmission and emergency visits within 3 months following valvular surgery, incidence of mortality, and all-cause major and minor bleeding. A post hoc analysis was conducted comparing outcomes among patients who underwent mitral versus aortic valve replacement, as well as patients with and without atrial fibrillation. The study was deemed exempt by hospital and university institutional review boards. Results: A total of 54 patients were included for analysis. All-cause readmission or emergency visit occurred in 33% (n = 18) of patients and the mortality rate was 6% (n = 3). A minor bleeding event was reported in 6% (n = 3) of patients and a major bleeding event was reported in 2% (n = 1) of patients. A thrombotic event was reported in 2% (n = 1) of patients. Conclusion: Within this cohort of patients requiring anticoagulation following bioprosthetic valve replacement, apixaban was safe and well-tolerated. However, more prospective data are needed to further correlate the safety and efficacy of apixaban, particularly in the setting of mitral valve replacement.

Background

Direct oral anticoagulants (DOACs) are preferred alternatives to warfarin for the management and prevention of many thromboembolic diseases. However, patients who have had valve replacements have typically comprised a small percentage of the patient populations in clinical trials, leaving practitioners questioning the safety and efficacy of DOACs in this patient population.^1,2

Patients who receive a bioprosthetic valve replacement are recommended to receive 3 months of anticoagulation followed by lifelong aspirin therapy.^3,4 These recommendations are consistent for both mitral and aortic valve replacements. ⁴ While a bioprosthetic valve may not require lifelong anticoagulation, many patients have other conditions that do require a lifelong anticoagulant including atrial fibrillation, a previous unprovoked thromboembolic event, or a hypercoagulable condition such as Factor V Leiden. Without the option of DOACs, prescribers may rely on warfarin, which requires a substantial amount of international normalized ratio monitoring and has numerous drug interactions. Additionally, due to age, medical history, and other patient specific factors, some patients are poor candidates for warfarin treatment. ⁵

Studies using DOACs in patients with prosthetic valves have had variable results, depending on the specific DOAC analyzed. A randomized controlled trial comparing warfarin with dabigatran in patients with mechanical heart valves was terminated prematurely due to a higher risk of both thromboembolic and bleeding events in the dabigatran arm. ⁶ The GALILEO trial compared rivaroxaban with antiplatelet therapy after a transcatheter aortic valve replacement (TAVR) in patients without atrial fibrillation. This trial was stopped early as a result of increased rates of all-cause mortality, thromboembolic events, and bleeding in the rivaroxaban group. ⁷

Favorable results have been seen, however, in other studies. The recently completed RIVER trial showed that in bioprosthetic mitral valve replacement patients with atrial fibrillation, rivaroxaban was noninferior to warfarin. ⁸ A case series study completed by the Mayo Clinic considered the outcomes of patients who underwent TAVR who were also prescribed a DOAC, and found DOACs were generally safe and well-tolerated. ⁹ In a subset of the ARISTOTLE trial, patients who had a bioprosthetic valve replacement were randomized to apixaban. ¹⁰ Overall clinical event rates in this population were low, with no significant differences between apixaban and warfarin for any outcomes. ¹⁰

Due to the mixed and limited data, DOACs are not frequently used by providers to prevent thromboembolism in patients who have had valvular replacements, making research in this setting imperative to determine their potential effectiveness. At this point, it remains uncertain whether apixaban may be a reasonable alternative for the prevention of thromboembolism in patients with a bioprosthetic valve replacement. Our study sought to describe outcomes among patients who underwent a bioprosthetic valve replacement procedure and were subsequently prescribed apixaban for anticoagulation, with or without atrial fibrillation.

Methods

A single-center, retrospective observational cohort study was conducted on patients who were admitted to a 344-bed community hospital. Patients were included if they underwent a valve replacement surgery with a bioprosthetic valve and were prescribed apixaban for anticoagulation at discharge. This study focused solely on apixaban due to possible improved safety profile compared with other DOACs in this setting.^1,2,9,10 Patients were identified via electronic report of all valve surgeries performed at the hospital from January 1, 2015, to January 1, 2020. Exclusion criteria included those who were <19 years of age, pregnancy, prisoners, those who had only valvular repairs, and those with mechanical heart valves, as documented in either the history and physical note and/or cardiovascular surgeon’s clinical note. If a patient was admitted multiple times during the report time period, only the first admission was included as a data point; subsequent admissions were assessed for readmission reasons.

The data were collected from hospital medical records by the researchers of the study. Age, gender, and race were recorded for demographic purposes. The type of valve replacement received by the patient was also recorded along with any alternate indications for anticoagulation. Alternate indications for anticoagulation included previous thromboembolic event such as a pulmonary embolism or deep vein thrombosis, atrial fibrillation with a CHA₂DS₂-VASc score of 2 or more, or other hypercoagulable condition such as Factor V Leiden. To calculate a CHA₂DS₂-VASc score, additional data including history of congestive heart failure (CHF), hypertension, stroke/transient ischemic attack (TIA)/thromboembolism, vascular disease (previous myocardial infarction [MI], peripheral artery disease, or aortic plaque), and diabetes history were recorded. Additional data were collected to calculate an estimated HAS-BLED score to determine a patient’s bleeding risk in patients with atrial fibrillation. These data included indicators of renal disease (dialysis, transplant, or serum creatinine >2.26), history of major bleeding, or liver dysfunction (cirrhosis, bilirubin 2 times the upper limit of normal, or aspartate aminotransferase/alanine aminotransferase 2 times the upper limit of normal.) Serum creatinine was assessed at the closest point in time to apixaban prescribing, which was usually the day of discharge.

The dosing for apixaban in patients with atrial fibrillation varies according to the patient’s age, weight, indication, and renal function. Renal function was assessed by recording serum creatinine and creatinine clearance. Doses of apixaban for atrial fibrillation are typically 5 mg twice daily unless a patient has any 2 of the following criteria: age ≥80 years, body weight ≤60 kg, or serum creatinine ≥1.5 mg/dL then the dose would be 2.5 mg twice daily. ¹¹ Prescribed doses of apixaban were assessed for appropriateness in patients with atrial fibrillation due to these criteria. Additionally, medication lists were reviewed for any potential drug interactions with apixaban using the tertiary drug database, LexiComp. Use of antiplatelet medications on discharge summary was recorded. Additionally, prescriber rationale and planned duration for apixaban use was recorded, when noted.

All-cause readmission and emergency visits to the hospital in the 3 months following valve surgery were recorded, and the medical records were used to ascertain the reason for readmission. Mortality was also noted. The following efficacy endpoints were recorded: stroke or TIA, readmission due to cardiovascular cause (MI or thromboembolic event), or need for follow-up surgery. Safety readmissions due to all-cause bleeding, divided into major and minor, were also recorded. Major bleeding was defined as fatal bleeding, intracranial bleeding, intraspinal bleeding, intraocular bleeding, retroperitoneal bleeding, intraarticular or pericardial bleeds, or intramuscular with compartment syndrome, or bleeding that required the transfusion of 2 or more units of whole blood or red blood cells, or use of a reversal agent. ¹² Minor bleeding was defined as non-major bleeding that the patient sought treatment for at the study hospital.

The analysis was primarily descriptive in nature, focusing on researching readmissions and safety endpoints, as well as the appropriateness of apixaban dose, of any patients prescribed apixaban following bioprosthetic valve replacement. The electronic health records of patients readmitted were thoroughly explored to determine a cause for the follow-up event. Categorical variables were reported as percentages. Continuous variables were reported as a mean and standard deviation (SD) or median and interquartile range (IQR). A post hoc analysis was conducted comparing outcomes among patients who had a mitral versus aortic valve replacement to determine any potential differences between groups with proportions compared via χ² analysis (MedCalc Statistical Calculator, 2021 MedCalc Software Ltd).

The study was deemed exempt by hospital and university institutional review boards. All data were recorded in a de-identified manner.

Results

A total of 226 charts were reviewed. After reviewing for inclusion/exclusion criteria, 54 patients were identified as having been prescribed apixaban at discharged following a bioprosthetic valve replacement (see Figure 1 for inclusion pathway). Patient characteristics can be found in Table 1. The average age was 70.5 years (SD = 9.0), with 63% males. Most valve replacements were aortic (61%), though a substantial portion received mitral valve replacement (39%). Of note, 1.9% (n = 1) had both a mitral and aortic valve replacement concurrently. Within this cohort, 25.6% (n = 14) had a diagnosis of atrial fibrillation.

Figure 1.

Inclusion pathway.

Table 1.

Patient Characteristics.

Characteristics	Entire cohort (N = 54)
Age (years), mean (SD)	70.5 (9)
Male gender, n (%)	33 (61)
Race, n (%)
African American	17 (32)
Caucasian	35 (65)
Other	2 (4)
Type of valve replacement, n (%)
Surgical aortic valve replacement	31 (57)
Mitral valve replacement	21 (39)
TAVR	2 (4)
Tricuspid valve replacement	1 (2)
Valvular defect, n (%)
Valvular insufficiency	18 (33)
Mild	2 (4)
Moderate	5 (9)
Severe	11 (20)
Stenosis	21 (39)
Mild	0
Moderate	3 (6)
Severe	18 (33)
Regurgitation	42 (78)
Mild	5 (9)
Moderate	13 (24)
Severe	24 (44)
Comorbidities, n (%)
Hypertension	48 (89)
Diabetes mellitus	22 (39)
CHF	41 (76)
Previous stroke	9 (17)
Previous VTE	4 (7)
Previous MI	12 (22)
Peripheral artery disease/vascular diseases	32 (59)
Atrial fibrillation/atrial flutter	14 (26)
Active malignancy	2 (4)
Hypercoagulable condition	0
Prior major bleeding	1 (2)
Liver disease	5 (9)
Renal replacement therapy	3 (6)
Medications prior to admission for valve replacement
Anticoagulation prior to admission	9 (17)
Apixaban	7 (13)
Warfarin	2 (4)
Antiplatelet prior to admission	33 (61)
Aspirin	28 (52)
Clopidogrel	3 (6)
Aspirin/dipyridamole	1 (2)
Aspirin + clopidogrel	1 (2)
Serum creatinine (mg/dL), mean (SD)	1.15 (1)
CHA2DS2-VASc of patient’s with Afib a , median [IQR]	4.4 [4, 5]
HAS-BLED of patients with Afib a , median [IQR]	2 [1, 2]
Left ventricular ejection fraction percentage, mean (SD)	46 (16)

Abbreviations: CHF, congestive heart failure; IQR, interquartile range; MI, myocardial infarction; SD, standard deviation; TAVR, transcatheter aortic valve replacement; VTE, venous thromboembolism.

^aN = 14.

Clinical endpoints are documented in Table 2. Readmission or emergency visits primarily occurred due to heart failure exacerbation (n = 7) or infection (n = 6). In the post hoc analysis of patients undergoing mitral versus aortic valve replacements, all-cause readmission or emergency visits occurred in 48% (n = 10) of patients who underwent a mitral valve replacement compared with 24% (n = 8) of patients who underwent an aortic valve replacement (95% confidence interval = −1.6% to 57.0%, P = .0708). When comparing the subgroup of patients receiving mitral versus aortic valve replacements, there were no notable differences in terms of age, gender, race, comorbidities, or concomitant antiplatelet therapies that would explain this trend. Fewer patients in the mitral valve subgroup received the higher dose of apixaban (5 mg twice a day) as compared with the aortic subgroup (19% vs 30%). Nonetheless, none of the readmissions in the mitral valve subgroup were secondary to thrombosis or cardiovascular cause. There was no significant difference in mortality, incidence of major or minor bleeding, or incidence of stroke/TIA in patients undergoing mitral versus aortic valve replacements.

Table 2.

Clinical Endpoints.

Outcomes	Entire cohort (N = 54)
All cause readmission, n (%)	18 (33)
Readmission due to heart failure exacerbation	7 (13)
Readmission due to infection	6 (11)
Readmission due to bleeding	1 (2)
Readmission due to cardiac arrest	1 (2)
Readmission due to stroke/TIA	1 (2)
Readmission due to other/unknown cause	2 (4)
Mortality, n (%)	3 (6)
Secondary to infection	1 (2)
Secondary to cardiac arrest	1 (2)
Secondary to unknown cause	1 (2)
Stroke or TIA, n (%)	1 (2)
Major bleeding, n (%)	1 (2)
Minor bleeding, n (%)	3 (6)

Abbreviation: TIA, transient ischemic attack.

A post hoc analysis was also performed, which assessed the subgroups of patients with and without atrial fibrillation. There was a trend toward increased mortality in patients with atrial fibrillation versus without atrial fibrillation (14% vs 3%) but this did not reach statistical significance (95% confidence interval = −3.7% to 36.7%, P = .135). There was no difference in incidence of all-cause readmission, incidence of major or minor bleeding, or incidence of stroke/TIA in patients with and without atrial fibrillation. Clinical outcomes among subgroups (mitral versus aortic, atrial fibrillation versus no atrial fibrillation) can be found in Figure 2.

Figure 2.

Subgroup analysis of clinical outcomes. TIA, transient ischemic attack.

*One patient underwent both mitral and aortic valve replacement and is counted in both mitral and aortic subgroups.

Mortality occurred in 3 patients (6%). One death was secondary to cardiac arrest; one death was due to mediastinitis and prosthetic valve endocarditis; and one patient had an unknown cause of death. Of note, the patient who expired secondary to cardiac arrest had a history of a previous MI, CHF with LVEF (left ventricular ejection fraction) of 20%, and vascular disease that may have contributed to mortality. The patient who expired secondary to unknown cause had a past medical history including hypertension, previous stroke or TIA, previous venous thromboembolism, previous MI, and atrial fibrillation. The average age of patients who met the mortality end point was 74.8 years (SD = 7.5 years).

One patient experienced a thrombotic adverse event with a stroke/TIA while taking apixaban 2.5 mg twice daily and aspirin 81 mg daily; this patient had undergone mitral valve replacement. In total, 7.7% (n = 4) of patients had a bleeding event occur (Table 2). All of the patients who experienced a bleeding event were taking aspirin concomitantly. Three of the 4 patients experiencing a bleeding event were taking apixaban 2.5 mg twice daily, and one patient was taking apixaban 5 mg twice daily. Clinical characteristics of patients experiencing major bleed, minor bleed, and TIA can be found in Table 3.

Table 3.

Summary of Clinical Events, With Patient-Specific Characteristics.

Event	Medication	Time from apixaban initiation	Type of valve replacement	Age, years	Sex	Hemoglobin	INR	Units of PRBC administered	HAS-BLED*	Medication interactions with apixaban	Outcome
Major bleeding (n = 1)
Gastrointestinal bleed	Apixaban 2.5 mg BID	3 days	SAVR	71	Male	7.9	1.54	2	N/A	Aspirin	Apixaban discontinued; patient survived
Minor bleeding (n = 3)
Hematochezia	Apixaban 2.5 mg BID	7 days	SAVR	76	Male	8.5	2.16	0	N/A	Aspirin	Patient experienced asystole and expired
Hematoma with bloody discharge at surgical site	Apixaban 5 mg BID	28 days	SAVR	79	Male	11.3	1.33	0	3	Aspirin, diltiazem	Apixaban discontinued; ultimately expired from unknown cause after discharge from emergency department
Nosebleed	Apixaban 2.5 mg BID	1 days	SAVR	79	Female	10.3	1.74	0	N/A	Aspirin, escitalopram, and megestrol	Apixaban discontinued; patient survived
Event	Medication	Time from apixaban initiation/time from surgery	Type of valve replacement	Age, years	Sex	Atrial fibrillation diagnosis	Drug interactions	Outcome
TIA (n = 1)
TIA, NIH stroke scale 2	Apixaban 2.5 mg BID	29 days/36 days	MV	77	Male	No	Aspirin	Discharged from emergency department; instructed to continue apixaban and follow-up with cardiovascular surgeon

Abbreviations: BID, twice a day; CrCl, creatinine clearance; INR, international normalized ratio; N/A, not applicable; NIH, National Institute of Health; PRBC, packed red blood cell; SAVR, surgical aortic valve replacement; TIA, transient ischemic attack.

^*HAS-BLED calculated only in patients with atrial fibrillation.

Identified drug-drug interactions for all included study patients can be found in Table 4. Of note, 5.8% (n = 3) patients were also prescribed phenytoin, which represents a Category X: Avoid Combination interaction with apixaban. ¹³ The majority of patients (83%) were receiving concomitant antiplatelet therapy in addition to apixaban (Category D: Modify Regimen ¹³ ), including 2 patients receiving dual antiplatelet therapy. The other interactions identified included concomitant SSRI/SNRI (selective serotonin reuptake inhibitor/serotonin and norepinephrine reuptake inhibitor) (Category C: Monitor Therapy) or megestrol (Category D: Modify Regimen). ¹³ Of note, one patient receiving apixaban and concomitant escitalopram and megestrol experienced a minor bleed (Table 3).

Table 4.

Apixaban Dosing Information.

Characteristic	Entire cohort (N = 54)
Drug interactions with apixaban, n (%)
Aspirin	45 (83)
SSRI/SNRI	7 (13)
Megestrol	7 (13)
Phenytoin	3 (6)
Aspirin and clopidogrel (dual antiplatelet)	2 (4)
Diltiazem	2 (4)
Meloxicam	1 (2)
Apixaban dosing, n (%)
2.5 mg twice daily	40 (74)
5 mg twice daily	14 (26)
Prescribed duration of apixaban, n (%)
Unspecified	19 (35)
≤30 days	6 (11)
31 to 60 days	16 (30)
61 to 90 days	8 (15)
Other finite duration	1 (2)
Indefinite	4 (7)

Abbreviations: SSRI, selective serotonin reuptake inhibitor; SNRI, serotonin and norepinephrine reuptake inhibitor.

Table 4 details the dosage and duration of apixaban that was prescribed, with 2.5 mg twice daily the most commonly prescribed regimen. Among the 14 patients with atrial fibrillation, 9 were prescribed the recommended dose based on age, weight, and serum creatinine, while the other 5 patients received 2.5 mg twice daily without meeting package insert criteria for dose reduction. The rationale for dose reduction was not documented. The majority of patients were prescribed a duration of <90 days, though 37% of patients had an unspecified duration of apixaban. Rationale for using apixaban over warfarin was available for only 2 patients; one patient refused warfarin therapy, and the other experienced a thrombus while taking warfarin therapy.

Discussion

This retrospective cohort study assessed clinical outcomes in patients prescribed apixaban following bioprosthetic valve replacement. All-cause mortality occurred in 3 patients, and only one patient expired secondary to cardiovascular cause (cardiac arrest). Four patients experienced a bleeding event, of which only one was considered major. One patient experienced a stroke/TIA during the study period.

Our results are similar to results available in the current literature. Within a case series of 21 patients conducted by the Mayo Clinic Hospital, 15 patients received apixaban after TAVR. ⁹ In this case series, only 3 patients experienced a bleeding event, and no thromboembolic events were reported. Researchers concluded similarly that apixaban particularly may represent a safe and effective alternative to warfarin in TAVR patients. ⁹ Within this case series, 95% of patients had an alternative indication for anticoagulation (atrial fibrillation or atrial flutter), which is in contrast with our study in which only 26% of patients carried an atrial fibrillation/flutter diagnosis. Furthermore, our study included TAVR patients in addition to mitral, tricuspid, and surgical aortic valve replacement patients, providing evidence on apixaban’s safety in the setting of additional valve surgeries.

Investigators from the ARISTOTLE trial also assessed efficacy and safety of apixaban versus warfarin in the setting of bioprosthetic valve, including both aortic and mitral valves as our study did. ¹⁰ However, they only included patients with atrial fibrillation, and only 27% to 36% of patients were receiving concomitant aspirin. Similar to our findings, clinical event rates were low and no difference was found between apixaban and warfarin with respect to bleeding, stroke, or mortality. ¹⁰

Izumi et al assessed anticoagulant therapy in patients with bioprosthetic valves, of whom 16 patients received a DOAC. As in our study, clinical event rates were low. Among those patients receiving a DOAC (specific DOAC not specified), none experienced a stroke, systemic embolism, major bleeding, or cardiovascular event. Two patients receiving a DOAC did experience clinically relevant minor bleeding. Of note, all patients had concomitant atrial fibrillation, unlike our study, and it is unclear how many patients receiving a DOAC had a mitral versus aortic prosthesis. ¹⁴

Pasciolla et al evaluated the use of DOACs compared with warfarin following bioprosthetic aortic and mitral valve replacement. ¹⁵ A total of 86 patients received apixaban, 40 received rivaroxaban, and 1 received dabigatran. Clinical event rates were similar between warfarin and DOACs during the 6-month follow-up period, which was longer than the 3 months assessed in our study. Three thromboembolic events occurred, all of which were in the DOAC group. There was no difference in major bleeding in the warfarin versus DOAC arms (2.9% vs 7.1%, P = .22), as well as no difference in readmissions within 6 months (37% vs 27%, P = 0.16). Within the patient cohort described by Pasciolla et al, 49% to 64% of patients were receiving anticoagulation prior to valve surgery. ¹⁵ This is in contrast with the 17% of patients that were receiving anticoagulation prior to surgery within our cohort.

After beginning this study, a trend toward increased readmissions was noticed within the mitral valve patient subgroup, which prompted a post-hoc analysis. While this was a post hoc analysis that was not anticipated prior to beginning, other studies similarly contribute to this hypothesis. For example, the GALILEO trial found harm when using rivaroxaban in bioprosthetic TAVR patients, but the RIVER trial found rivaroxaban to be noninferior after bioprosthetic mitral valve replacements.^7,8 Additional research is necessary to determine if the valve being replaced affects patient outcomes.

Despite the retrospective design, our study does possess a few strengths. We had a relatively large sample size when compared with prior literature on this topic. Additionally, a high percentage of patients received concomitant aspirin therapy, which provides reassuring evidence of apixaban’s safety even when used in conjunction with antiplatelet therapy. A low percentage of patients were receiving apixaban prior to valve replacement (7%), so our study sheds light on the safety and efficacy of apixaban initiation in this patient population. Last, our study included a variety of patients, including those with and without atrial fibrillation, and both mitral and aortic valve replacements. This furthers the generalizability of our findings.

Our study also possesses several limitations. Readmission, emergency visit, bleeding, and mortality data were only able to be collected from the single institution and does not account for visits that may have occurred elsewhere. The study was retrospective in nature and there was not a comparator group. We were unable to assess apixaban compliance and were unable to assess all potential comorbidities that could have served as confounding variables, such as active malignancies. We elected not to include patients receiving warfarin, owing to our inability to monitor time within therapeutic international normalized ratio range and dose adjustments, which could have significantly confounded any clinical findings. At the time of study design, the RIVER trial results were unknown, and we elected to exclude patients receiving rivaroxaban due to the potential harm documented in the GALILEO trial.^7,8 Future studies should compare clinical outcomes between different DOACs, particularly rivaroxaban, and assess outcomes based on type of valve replaced.

Conclusion and Relevance

Data from this study suggest apixaban may be a safe option in patients with bioprosthetic valve replacements who require anticoagulation, with or without atrial fibrillation. In the patient cohort analyzed, apixaban was well-tolerated with low incidences of major/minor bleeding and thrombotic events. Further research, including prospective studies, is necessary to determine fully the safety and effectiveness of apixaban in this setting.