Keywords
anticoagulation - intracerebral hemorrhage - mechanical heart valve - atrial fibrillation
Introduction
Patients with mechanical heart valves and coexisting atrial fibrillation (AFib-MHV)
who suffer an intraparenchymal hemorrhage (IPH, defined as bleeding solely within
the brain parenchyma and/or ventricle) are at a considerably high risk of thromboembolism
when off their anticoagulation.[1]
[2]
[3] There are no studies determining when it is safe to restart anticoagulation after
IPH for this select population.[1]
[4]
[5] Neurologists, cardiologists, and intensivists must weigh the risk of re-bleeding
against the high risk of thromboembolism, and due to the lack of data, decisions must
be made with little evidence, only that extrapolated from studies of patients with
atrial fibrillation or mechanical heart valves (MHV) alone.
The European Society of Cardiology (ESC) and American Heart Association/American Stroke
Association (AHA/ASA) emphasize the need for earlier anticoagulation for patients
with MHV due to their higher risk of embolism than patients with atrial fibrillation
or other indications for anticoagulation, but there are no clear guidelines as to
when it is most safe.[1]
[2]
[3]
[6] At best, the ESC states it “may be safe” to start systemic heparin 3 days after
IPH and switch to a vitamin K antagonist (VKA) at 7 days.[1]
[2]
[3]
[6]
Patients with AFib-MHV have a higher risk of thromboembolic events compared with patients
with high-risk atrial fibrillation (CHA2DS2-VASc >2)[7] and those with MHV with sinus rhythm.[8] The emphasis of early anticoagulation for patients with MHV is extrapolated to those
with AFib-MHV but without better evidence of risk factors associated with good outcomes,
the worry of IPH expansion after reinitiating anticoagulation will strongly persist.
To address the gap in the literature, we performed a retrospective analysis of patients
admitted to our institution with non-traumatic supratentorial IPH in patients with
AFib-MHV and compared those who were started on anticoagulation while in-patient with
those who were not. We characterized the patients and their IPH and evaluated outcomes
in terms of thromboembolic events, bleeding rates, and disposition. We hypothesized
that anticoagulation within a few days post-IPH would be tolerated in select patients
of this high-risk group.
Patients and Methods
We performed a single-institution retrospective analysis of patients with non-traumatic,
supratentorial IPH. This study was approved by the Duke University Institutional Review
Board. Adult patients ≥ 18 years old who were admitted to any Duke University Hospital
for non-traumatic, supratentorial IPH between July 1, 2013, and June 20, 2017, were
retrospectively identified from the DENDRITe database (DukE Neurocritical Care Patient
Data ReposITory) and the Get With The Guidelines Stroke Database. Based on past medical
history, we identified patients who had the following indications for anticoagulation
prior to their hospitalization: atrial fibrillation or MHV. For this type of study,
formal consent was not required.
The following data were extracted from the Duke electronic health record: patient’s
medical history; admission laboratory values; systolic blood pressure at the time
of presentation to the hospital; IPH characteristics (including etiology, location—deep,
hemispheric, lobar, multiple); intracerebral hemorrhage (ICH) score[9]; if applicable, type of neurosurgical intervention performed; code status; and if
applicable, type, timing, and dose of anticoagulation or antiplatelet agent reinitiated.
By chart review, we determined the occurrence of thromboembolic complications, defined
as ischemic stroke, transient ischemic attack, myocardial infarction, valvular thrombus,
or peripheral arterial embolism, and the occurrence of bleeding complications, defined
as hematoma expansion >33% or >6 mL,[10] any major extracranial hemorrhage, or any overt, actionable sign of hemorrhage (i.e.,
more bleeding than would be expected for a clinical circumstance, including bleeding
found by imaging alone) that met at least one of the following criteria: (1) requiring
nonsurgical, medical intervention by a healthcare professional, (2) leading to the
increased level of care, or (3) prompting evaluation, according to Bleeding Academic
Research Consortium Type 2 or greater.[11] We also reviewed the patient’s chart for the modified Rankin Scale (mRS) at discharge[12]; discharge disposition; and date of death, if applicable. Hematoma volume was calculated
based on the ABC/2 method.[13] The following were calculated based on extracted data whenever applicable: for patients
with atrial fibrillation, and CHA2DS2-VASc and HAS-BLED scores[7]
[14] and for patients with a thromboembolic or hemorrhagic complication, time from IPH
to complication and time from anticoagulation (if started) until complication.
Results
Non-traumatic supratentorial IPH was identified in 395 patients admitted to the Duke
Health System during the study period, of whom 6 patients were identified to have
Afib-MHV and were analyzed.
All patients suffered an IPH in the setting of anticoagulation and received reversal
agents to correct their coagulopathy. All patients were managed medically according
to the ASA/AHA Guidelines for the Management of Spontaneous Intracerebral Hemorrhage,[2] which included treatment of coagulopathy and thrombocytopenia and acute lowering
of blood pressure to < 140 mm Hg until stable intracranial imaging. One patient underwent
hematoma evacuation out of concern for abscess as the etiology of the hemorrhage ([Table 1], Patient 1). The final pathology report revealed only blood products. One patient
underwent placement of an extra-ventricular drain ([Table 2], Patient 2).
Table 1
Patients with mechanical heart valve and concomitant atrial fibrillation initiated
on anticoagulation prior to discharge
|
Patient
|
Age
|
Sex
|
GCS on admission
|
Location of hemorrhage
|
INR on admission
|
ICH volume (mL)
|
ICH score
|
Atrial fibrillation rate control
|
Mechanical valve type
|
CHA2DS2
|
HASBLED
|
Anticoagulation start date (days post-bleeding)
|
Anticoagulation restarted
|
LOS (days)
|
DC mRS
|
DC Dispo
|
|
Abbreviations: Afib, atrial fibrillation; DC, discharge; Dispo, disposition; F, female;
GCS, Glasgow coma scale; ICH, intracerebral hemorrhage; INR, international normalized
ratio; LOS, length of stay; M, male; mRS, modified Rankin Scale; SNF, skilled nursing
facility; VKA, vitamin K antagonist.
aVKA started with a heparin.
|
|
1
|
45
|
F
|
15
|
Frontal lobe
|
4
|
6.6
|
0
|
Yes
|
Mitral
|
1
|
1
|
1
|
VKAa
|
18
|
1
|
Home
|
|
2
|
44
|
M
|
15
|
Frontal lobe
|
3.4
|
3
|
0
|
No
|
Aortic
|
1
|
3
|
2
|
VKA
|
2
|
1
|
Home
|
|
3
|
80
|
F
|
14
|
Basal ganglia
|
2.4
|
10
|
1
|
Yes
|
Mitral
|
1
|
4
|
13
|
VKA
|
12
|
2
|
SNF
|
|
4
|
72
|
M
|
15
|
Parietal lobe
|
4.8
|
6.6
|
1
|
No
|
Aortic and Mitral
|
6
|
5
|
18
|
VKAa
|
26
|
1
|
Home
|
Table 2
Patients with a mechanical heart valve and concomitant atrial fibrillation not initiated
on anticoagulation
|
Patient
|
Age
|
Sex
|
GCS on admission
|
Location of hemorrhage
|
INR on admission
|
ICH volume (mL)
|
ICH score
|
Atrial fibrillation rate control agent
|
Mechanical valve type
|
CHA2DS2
|
HASBLED
|
Why no anticoagulation
|
Time from IPH until death (days)
|
|
Abbreviations: GCS, Glasgow coma scale; ICH, intracerebral hemorrhage; INR, international
normalized ratio; IPH, intraparenchymal hemorrhage; mL, milliliters.
|
|
1
|
84
|
M
|
10
|
Temporal
|
4.2
|
32
|
3
|
No
|
Mitral
|
7
|
2
|
Critical illness
|
10
|
|
2
|
85
|
M
|
5T
|
Frontal
|
5
|
60
|
5
|
Yes
|
Aortic
|
7
|
5
|
Comfort care within 48 h
|
2
|
Four patients were initiated on anticoagulation prior to discharge (median 8.5 days
post-IPH, range 1–18), and two of these patients were initiated on anticoagulation
within 3 days post-IPH ([Table 1]). Patients started on early anticoagulation had small hematoma volumes (<10 mL,
mL), were relatively young (<50 years old), and tolerated anticoagulation without
any hemorrhagic complications. There were no thromboembolic complications in these
patients.
Two patients were not initiated on anticoagulation and died during their admission
([Table 2]). These patients were older (>60 years) with large hematoma volumes (>30 mL). One
died of multiorgan failure; the other was made comfort care.
Discussion
This is a descriptive study evaluating solely patients with AFib-MHV who suffered
non-traumatic, supratentorial IPH and were reinitiated on anticoagulation while in-patient.
As far as we know, this is the first study specific to this population to date. Four
patients were started on therapeutic anticoagulation prior to hospital discharge,
and two within 3 days after IPH. The etiology of IPH for all was predominantly due
to therapeutic anticoagulation and all patients’ coagulopathies were adequately reversed.
Patients initiated on early anticoagulation were younger (<50 years old) and had smaller
hematoma volumes (<10 mL). No patients suffered hemorrhagic complications of initiating
anticoagulation. There was no institutional protocol to guide decision-making for
whom and when to restart anticoagulation.
This study is unique from prior studies regarding atrial fibrillation and MHV because
of its focus on patients with only concomitant Afib-MHV, who should be uniquely studied
because of their different risk factors of both thromboembolic and bleeding complications
off anticoagulation compared with patients with sole atrial fibrillation or MHV.[8] Kuramatsu et al’s study,[8] the largest analysis of patients with MHVs who suffered vitamin-K antagonist associated
IPH, revealed a statistically significant higher risk of bleeding complications and
a non-statistically significant but higher risk of thromboembolic complications off
anticoagulation for those with MHV-afib (n = 21) compared with those with sinus rhythm (n = 50).[8] There was no difference in bleeding complications when both groups were therapeutically
anticoagulated.[8] The authors recommended waiting until day 14 to re-initiate anticoagulation for
patients with MHV, except for patients with MHV-afib, who had a higher risk of thromboembolic
complications. For these patients, anticoagulation 6 days post-bleeding seemed reasonable.[8]
Our patient population was also unique from prior studies because of its focus on
patients who suffered only spontaneous, supratentorial IPH. A systematic review by
Chandra et al. that has influenced guidelines determined that initiating heparin infusion
within 3 days and oral anticoagulation within 7 days after intracerebral hemorrhage
can be safe for patients with prosthetic heart valves[2]
[6]; however, their data were extrapolated from studies of patients with various indications
for anticoagulation and various types of intracerebral hemorrhage, not just IPH.[2]
[6] Given its incomparable rate of re-bleeding and different pathophysiology than traumatic
hemorrhages or subarachnoid hemorrhages, patients with IPH should not be concomitantly
studied with other types of intracerebral hemorrhage when determining the safe timing
of re-initiating anticoagulation post-bleeding.
For the time being, patients with Afib-MHV can only be anticoagulated with VKAs. There
is currently no approval for the newer, direct oral anticoagulants for these indications.
VKA-associated IPH can be reversed by several agents. Uncorrected coagulopathy in
vitamin K antagonist-associated IPH is associated with worse outcomes, namely increased
frequency of hematoma enlargement and rate of in-hospital mortality.[15] In light of this, the impact of the Kuramatsu et al’s study is dampened by the fact
that international normalized ratio (INR) reversal was achieved in only 25% of all
patients.[8]
[15] Kuramatsu et al’s conclusion to wait 6 days based on the hemorrhagic risk profile
was likely skewed, given the known association between inadequate correction of coagulopathy
and increased bleeding complication rate.
We acknowledge that a sample of two patients initiated on early anticoagulation is
the main limitation of this study, in addition to its retrospective design. Given
the small sample size and a sparse number of adverse events, we could not evaluate
significant risk factors for those who tolerated early anticoagulation or risk of
thromboembolic events if anticoagulation was delayed. Moreover, as a retrospective
study, there was no systematic method of identifying thromboembolic and bleeding complications.
A complication would have only been identified if the primary provider was concerned
enough to order a diagnostic study.
This study highlights the rarity of this population. Prior studies emphasize the fact
that the population is understudied, and there is a need for more robust data for
this population of patients whose thromboembolic risk off anticoagulation is greater
than those with sole MHV or Afib. A prospective or larger retrospective study is needed
to inform the management and guidelines of this select population who face no option
to stay off anticoagulation. Our study did show tolerance of early anticoagulation
(within 3 days) in young patients with small hematoma volumes and coagulopathy-associated
etiology of their IPH, which was corrected. Age, hematoma volume, and correction of
coagulopathy are suggested risks factors to explore in future studies. Uncontrolled
hypertension, poor neurological status, and the presence of cerebral amyloid angiopathy
are also other known risk factors for recurrent IPH and worse outcomes. Although these
were not a concern in our patient group, these could also be explored as additional
risk factors.[2]
[16]
[17]
[18]
[19] Long-term risk of re-hemorrhage should also be studied.
Conclusion
Patients with MHV and coexistent atrial fibrillation who suffer spontaneous IPH are
a rare population to study, and further research regarding risk factors for early
anticoagulation in these patients is warranted to best inform the management of these
patients with a high risk of thromboembolism off anticoagulation. We report two patients
who tolerated early anticoagulation (within 3 days) after suffering a spontaneous
IPH attributed to anticoagulation. We note that their small hematoma size, young age,
adequate reversal of coagulopathy, and no comorbid hypertension or amyloid angiopathy
may have been favorable risk factors for early anticoagulation.
