Abstract
In 2 distinct entities, left ventricular noncompaction (LVNC) and peripartum cardiomyopathy (PPCM), routine anticoagulation therapy is often used in current practices. However, our systematic review showed that LVNC itself was not associated with the increase in thromboembolism event rates and therapeutic anticoagulation therapy should not be considered only for LVNC, unless there is risk factor for thromboembolism. Current literature justifies prophylactic therapeutic anticoagulation in LVNC with low left ventricular ejection fraction (EF < 40%) and/or atrial fibrillation. Although not specifically studied, the presence of intracardiac thrombi by echocardiography or other imaging studies should also prompt anticoagulation therapy. There is limited evidence available for the use of anticoagulation in patients with PPCM, but our systematic review showed that anticoagulation should be recommended only for patients with PPCM especially with an EF < 35% until EF is recovered, as well as for patients with PPCM treated with bromocriptine.
There is a consensus that heart failure (HF) increases the risk of thromboembolic complications. Cardiac dysfunction with impaired contractility results in reduced blood flow, increased viscosity, stasis of blood, activation of platelets, and the coagulation cascade and clot formation in dilated cardiac chambers and the venous system, creating the potential for both pulmonary and systemic embolism. 1,2 Several studies demonstrated the trend for increasing thromboembolism with lower ejection fraction (EF). 3,4 However, routine anticoagulation in patients with HF is not currently recommended because a randomized trial and meta-analysis showed that reduction in embolic complications was offset by an increase in major bleeding. 5,6
However, in 2 distinct entities, left ventricular noncompaction (LVNC) and peripartum cardiomyopathy (PPCM), anticoagulation therapy is often recommended. We systemically review the evidence of the need for anticoagulation in these 2 conditions.
Literature Search
Medline and Cochrane Library database searches for relevant articles through November 12, 2017, were performed using the following keywords: (LVNC OR left ventricular hypertrabeculation) AND (warfarin OR anticoagulation OR aspirin OR apixaban OR rivaroxaban OR dabigatran OR edoxaban). The results were limited to human studies published in the English language. Pertinent evaluations identified in the references of articles found via the PubMed search were also included. A total of 13 studies evaluating anticoagulation management in patients with LVNC were identified. (Table 1) 4,7 –18 Three studies were excluded because they did not describe the effect of anticoagulation on Thromboembolism (TE) event rates. 19 –21 No study evaluating direct oral anticoagulants (apixaban, rivaroxaban, dabigatran, or edoxaban) was identified. One case report describing the use of rivaroxaban was excluded due to lower evidence level of case report. 22
Summary of 13 Clinical Studies Regarding Anticoagulation Use in Patients With LVNC.
Abbreviations: EF, ejection fraction; LVEF, left ventricular ejection fraction; LVNC, left ventricular noncompaction; N/A, not available.
Medline and Cochrane Library database searches were also performed using the following keywords: (peripartum cardiomyopathy) AND (anticoagulation OR anticoagulant) through November 12, 2017. Only 1 retrospective article by Mandal et al 23 and 4 case reports were included. 24 –27 No study for direct oral anticoagulant was identified.
Left Ventricular Noncompaction
Challenges in Diagnosis of LVNC
Isolated LVNC is a rare cardiomyopathy that was first described in 1926 and was classified as a genetic cardiomyopathy according to the American Heart Association. 28 The distinct feature of this cardiomyopathy is extensive trabeculation of the left ventricular (LV) cavity, with multiple pockets and recesses along the walls with a noncompacted myocardium. These trabeculations and pockets create a perfect environment for thrombi formation. The treatment approaches for patients with LVNC are similar to those for other cardiomyopathies, but the prevention of cardioembolic complications resulting from mural thrombosis remains controversial. Thus, some recommend systemic prophylactic anticoagulants in all patients with LVNC. 14 However, most believe that warfarin should be given or considered in higher risk patients with noncompaction, such as with atrial fibrillation, left ventricular ejection fraction (LVEF) <30%, or a history of thromboembolic complication. Currently, there is no guideline recommendation on the use of anticoagulation in patients with LVNC, and there is no consistent approach in practice.
Additionally, the diagnosis of LVNC is based on morphologic criteria, mostly on the appearance and ratio of compacted and noncompacted myocardium. The Jenni criteria assess for a maximal end-systolic ratio of noncompacted to compacted layers >2 (among other criteria). 29 The Chin criteria focus on the trabeculae at the left ventricular apex on the parasternal short-axis and apical views and on the LV free-wall thickness at end-diastole. 8 The distance from the epicardial surface to the trough of the trabecular recess and the distance from the epicardial surface to peak of trabeculation are compared; a ratio ≤0.5 is considered diagnostic. Unfortunately, these criteria are nonspecific, and sometimes healthy people with a benign prognosis can meet these criteria. Gati et al alluded to the overreporting of noncompaction in healthy athletes, with 8.1% of African/Afro-Caribbean males fulfilling diagnostic criteria by echocardiogram for noncompaction cardiomyopathy, but having no real adverse events when followed longitudinally. 30 One hypothesis is that chronic increases in preload may be associated with the development of increased LV trabeculation. This theory was previously tested in patients with chronic sickle-cell anemia in which 20.8% of patients fulfilled the Chin criteria and 10% fulfilled the Jenni criteria for LVNC. 31
Hypertrabeculation can also be a transient phenomenon, as 102 primigravida pregnant women underwent serial echocardiographic evaluation with 26 developing increased trabeculations during gestation, 73% of which showed complete resolution 2 years postpartum. 32 The inability to distinguish noncompaction cardiomyopathy from hypertrabeculation has significant clinical implications because patients with true noncompaction cardiomyopathy are at higher risk for thromboembolism and ventricular arrhythmias. Giving anticoagulation to all patients who fit the diagnostic criteria of LVNC may result in iatrogenic bleeding in people without a risk for TE events.
Left Ventricular Noncompaction as an Indication for Anticoagulation
Case series were first reported to describe the clinical courses of 8 cases with LVNC and evaluate the incidence of thromboembolic complications. 8 It showed that 3 (38%) patients developed systemic embolization comprising 2 cerebral emboli and 1 episode of a saddle embolus in the abdominal aorta. Two out of 3 patients were also found to have LV mural thrombi. Of note, the percentage of patients taking anticoagulation therapy was not reported. Another case series by Ritter et al included 17 adult patients with high thromboembolic event rates (24%) during the mean follow-up of 30 months. 15 The authors recommended anticoagulation for all patients with LVNC regardless of the thromboembolic history, atrial fibrillation, or LVEF. A prospective cohort study of 34 patients with LVNC including 17 patients from the previous study 15 was performed, and it supported the findings of the above study 14 . Although it did not report the percentage of patients on anticoagulation, it reported long-term clinical outcomes with high thromboembolic event rates (24%), independent of LV size or function. The authors recommended oral anticoagulants for every patient with LVNC regardless of ventricular function. However, this study was not designed to evaluate the use of anticoagulation to prevent TE, and because of the lack of a control group, no percentage of anticoagulation therapy was reported in the small sample size.
Left Ventricular Noncompaction by Itself Is not an Indication for Anticoagulation
Italian multicenter registry by Fazio et al investigated whether LVNC by itself is a risk factor for thromboembolic complications. 9 It included 229 patients without atrial fibrillation who were diagnosed with LVNC based on echocardiography and observed over the average follow-up of 7.3 years. None of the patients was on anticoagulation therapy. It showed that the thromboembolic event rate was low at 2.1% without prophylactic anticoagulation. There were 4 cases of ischemic stroke and 1 episode of LV thrombus that occurred in a 1-year-old patient with Behcet disease; this disease is a risk factor for thromboembolism by itself. The authors concluded that prophylactic anticoagulation is not necessary for patients with LVNC without atrial fibrillation, but they also acknowledged the need for further investigation regarding the use of prophylactic anticoagulation in patients with LV dysfunction and LVNC.
Since the previous study by Fazio et al did not have a control group and included all patients with LVNC, it could not evaluate how LVNC affected TE event rates directly. Thus, another cohort study investigated TE event rates in 62 patients with LVNC and 62 control patients matched by age, sex, and LVEF. 17 In the LVNC cohort, 5 patients had ischemic stroke and 1 patient had right iliac artery embolism. Of note, 5 of 6 patients with TE had systolic dysfunction. In the control group, 9 patients developed stroke; 3 of the 9 patients had atrial fibrillation. Two of 3 patients with atrial fibrillation were not on anticoagulation. Thus, LVNC itself was not associated with the increase in TE event rates, and the authors concluded that it should not be an indication for prophylactic anticoagulation; prophylaxis may be needed for patients with LVNC and systolic dysfunction.
When to add prophylactic anticoagulation: LVNC with systolic dysfunction or atrial fibrillation?
Another retrospective study aimed to assess the risk factors for TE in patients with LVNC by comparing cohorts with or without TE (22 vs 122 patients). 18 Among 14 episodes of cardioembolic stroke, 6 (42.9%) had atrial fibrillation, 11 (68.8%) had systolic dysfunction, and 4 (28.6%) had a combination of systolic dysfunction and atrial fibrillation, while only 1 (7.1%) patient who had a cardioembolic stroke did not have either. Although there were not enough patients to statistically assess the risk factors for atrial fibrillation and systolic dysfunction, the authors concluded that prophylactic anticoagulation is needed for patients with LVNC with atrial fibrillation or systolic dysfunction. Another study also supported systolic dysfunction as an important risk factor for prophylactic anticoagulation in patients with LVNC. 12 Although this study did not evaluate anticoagulation as a main objective, it reported that no patients with EF >30% developed TE, although none of the patients was on therapeutic anticoagulation.
Another study by Murphy et al also investigated the TE event rate retrospectively in 45 patients with LVNC. 13 Of note, only 6.6% had atrial fibrillation, and the mean EF was 21%. Although the study did not specify the criteria for anticoagulation, 40% of the patients were on anticoagulation therapy, and the TE event rate was only 4% in patients with a sinus rhythm. Thus, the authors supported anticoagulation therapy in patients with LVNC with an EF < 25% or a history of TE.
One study did not support the use of anticoagulation in patients with LVNC and systolic dysfunction. The prospective cohort study in the sub-Saharan African population included 55 patients with LVNC and HF with an EF <50% (mean EF [SD] = 29.6% [11.8%] ). 4 Of note, only 2 (3.6%) patients had atrial tachycardia. Sixteen patients were on warfarin, and indications for anticoagulation were pulmonary embolism (n = 1), LV thrombus (n = 8), and LVEF <30% (n = 7). None of the patients had TE during the follow-up period (mean 16.7 months), while 1 patient had TE in the nonwarfarin group. Thus, the authors concluded that anticoagulation is not recommended in patients with LVNC and HF with a reduced EF unless patients have atrial tachyarrhythmia. However, because there was no control group, the study was nonconclusive.
In summary, current literature justifies prophylactic anticoagulation in LVNC with low LVEF (EF < 40%) and/or atrial fibrillation. Although not specifically studied, the presence of intracardiac thrombi by echocardiography or other imaging studies should also prompt anticoagulation therapy.
Peripartum Cardiomyopathy
Peripartum cardiomyopathy is an uncommon heart disorder with nonischemic cardiomyopathy and marked systolic dysfunction that occurs in the last month of pregnancy or within 5 months after delivery. 33 Although PPCM shared common features with idiopathic dilated cardiomyopathy and both the cardiomyopathies may lead to LV systolic dysfunction, a key difference is that pregnant women have increased levels of coagulation factors VII, VIII, and X and plasma fibrinogen during late pregnancy and up to 4 to 6 weeks postpartum. 27,34 A depressed LV function and hypercoagulable state causes a higher incidence (17%) of LV thrombus than in patients with dilated cardiomyopathy (DCM). 35
Currently, there are guideline statements regarding the use of anticoagulation in patients with PPCM. However, those statements are mainly expert opinion based, and there is limited evidence available for the use of anticoagulation in patients with PPCM. The American Heart Association/American College of Cardiology guidelines recommend anticoagulation, especially in persistent LV dysfunction. 36 Additionally, the European Society of Cardiology supports the use of anticoagulation resulting from a hypercoagulable state during pregnancy in patients with PPCM and severe LV dysfunction (LVEF < 30%). 37 An expert opinion by Bhattacharyya et al recommended low-molecular-weight heparin antepartum and warfarin postpartum when the EF is reduced to less than 30%. 38 There is no consensus on the duration of anticoagulation, but the experts recommend serial evaluation of LVEF over 6 to 12 months and discontinuation of anticoagulation upon clinical and echocardiographic recovery. 38 Another group also recommends anticoagulation in patients with PPCM on bromocriptine as a result of an increased risk of stroke. 39
Mandal et al followed the cohort of 36 patients with PPCM (100% had EF <45%, 75% had EF <35%) who were not on anticoagulation for primary prophylaxis. 23 During the 2 year follow-up period, 5 (14%) thromboembolic events occurred (3 cerebral events and 2 peripheral thromboses), and 2 patients with stroke died. Thus, the authors recommended anticoagulation for primary prevention until at least 6 months postpartum.
To our knowledge, there are 4 case reports published in patients with PPCM who developed thromboembolism and were treated with anticoagulation (Table 2). 24 –27 Amos et al 35 reported LV thrombus in 17% of their PPCM population, compared with 3% of patients with nonischemic cardiomyopathy of other etiologies. He suggested that patients with PPCM should either be empirically anticoagulated or followed up closely. 35 Although the evidence level is limited, all 4 case reports also recommend using anticoagulation as a primary prevention measure in patients with PPCM until EF is normalized. 24 –27
Summary of 4 Cases Regarding Anticoagulant Use in Patients With PPCM.
Abbreviations: LVEF, left ventricular ejection fraction; POD, postoperative day Abbreviations: EF, ejection fraction; LV, left ventricular; LVEF, left ventricular ejection fraction; POD, postoperative day; PPCM, peripartum cardiomyopathy; aPTT, activated partial thromboplastin time; INR, international normalized ratio.
For many years, there was a discussion about the utility of bromocriptine in PPCM. Recently, the results of a multicenter randomized clinical trial were published. 40 Although there was no placebo arm, and the authors compared the outcomes in their trial with outcomes in another prospective cohort that were selected in another country by different criteria and with different racial representation, 41 superior outcomes in the randomized trial led to the conclusion that bromocriptine is beneficial. Currently, a phase 3 randomized trial with major adverse cardiac events as the primary outcome comparing bromocriptine plus guideline-driven medical therapy with only guideline-driven therapy as the control group is currently recruiting participants. 42 Until results are available, bromocriptine will continue to be used in PPCM. The question is whether anticoagulation should be given in patients with PPCM treated with bromocriptine. Bromocriptine is a hydrogenated semisynthetic ergot alkaloid derivative. Hydrogenation decreases vasopressor activity and increases vasodilatory effects. Nevertheless, ergot alkaloids are known to have the potential of causing vasoconstriction and a propensity to TE events. The drug package insert for bromocriptine indicates that in postmarketing experience, 30 cases of stroke and 9 cases of myocardial infarction were reported among postpartum women. 43 Given the relatively high rate of thromboembolic complications on bromocriptine, in combination with the hypercoagulable state of late pregnancy and early postpartum period, anticoagulation in patients with PPCM treated with bromocriptine is reasonable.
Conclusion
The reviewed evidence demonstrates that therapeutic anticoagulation therapy should not be considered only for LVNC, unless there is risk factor for thromboembolism. Anticoagulation should be recommended for patients with LVNC with reduced EF (EF < 40%), atrial fibrillation, presence of thrombi in the heart, or thromboembolism history including intracardiac thrombi or venous thromboembolism.
Although there is limited evidence available for the use of anticoagulation in patients with PPCM, anticoagulation should be recommended for patients with PPCM especially with an EF <35% until EF is recovered, as well as for patients with PPCM treated with bromocriptine.
Footnotes
Author Contributions
Kazuhiko Kido contributed to conception and design, contributed to acquisition, analysis, and interpretation, drafted the manuscript, critically revised the manuscript, gave final approval, and agrees to be accountable for all aspects of work ensuring integrity and accuracy. Maya Guglin contributed to design, contributed to acquisition and analysis, drafted the manuscript, critically revised the manuscript, gave final approval, and agrees to be accountable for all aspects of work ensuring integrity and accuracy
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
