Keywords
pulmonary arteriovenous malformation - embolization - hereditary hemorrhagic telangiectasia
Introduction
Pulmonary arteriovenous malformations (PAVMs) are abnormal connections of the pulmonary
artery to the pulmonary vein, which bypass the normal pulmonary capillary bed.[1] This leads to a right-to-left shunt the diameter of the feeding vessel. The normal
pulmonary capillary bed measures only 8 to 10 mcm in diameter and serves not only
to oxygenate blood but also to filter small thrombi and bacteria from passing through
the pulmonary vascular bed.[2] Approximately 70% of patients with PAVMs are symptomatic from their right-to-left
shunt. Symptoms may be mild due to poor oxygen exchange resulting in dyspnea or hypoxemia,
or more severe as a result of paradoxical emboli causing transient ischemic attacks
(TIAs), stroke, brain abscess, systemic emboli, or septic emboli. PAVMs can also bleed
spontaneously, which can cause massive life-threatening hemoptysis.[3]
[4]
[5]
Most PAVMs are congenital with the vast majority of patients having hereditary hemorrhagic
telangiectasia (HHT), otherwise known as Osler-Weber-Renu syndrome. The literature reports that the incidence of HHT in patients with PAVMs may be as
high as 90%,[2]
[3]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13] with up to 50% of patients with HHT having PAVMs. The remaining PAVMs can be congenital
idiopathic lesions, or acquired such as in the setting of cirrhosis with hepatopulmonary
syndrome or after treatment of complex cyanotic cardiac disease.[14]
PAVM embolization, which was first described by Porstmann in 1977,[15] is now the current standard of care for a PAVM with a feeding vessel ≥ 2 to 3 mm.[3]
[5]
[6]
[10]
[16]
[17]
[18]
[19] Technical success rate has been reported to be as high as 88 to 100%;[6] however, up to 25% of technically successful embolizations may require retreatment.[17] Retreatment has been indicated in the patients who on follow-up imaging have signs
of an enlarged PAVM or persistence whether it is from recanalization, growth of a
missed accessory vessel, or collateral reperfusion from pulmonary or systemic branches.
Reperfusion rates have been reported ranging from 5 to 57%.[17]
[20] Currently, reperfused pulmonary arteriovenous malformations (rPAVMs) are treated
regardless of whether or not the patient is symptomatic for fear that he/she is at
continued risk for complications from paradoxical embolization. In this study, the
authors sought to assess the clinical significance of rPAVMs.
Methods
Patient Population
All patients with PAVMs who underwent angiography and embolization between January
1, 2006 and December 30, 2007 from a single HHT center were included in this retrospective
review. Approval from our institutional review board (Human Investigation Committee
Protocol 131001294) was obtained. Patients with diffuse PAVMs, defined by a PAVM involving
each subsegmental artery of at least one lobe, were excluded. However, patients with
multiple focal PAVMs were included.
Patient charts and imaging studies were reviewed at the time of the embolization and
during a follow-up period through December 30, 2013. The following information was
recorded: patient demographics, HHT diagnosis and genetic subtype (when available),
clinical symptoms both at the time of embolization and during the follow-up period,
and imaging findings at the time of embolization and during the follow-up period.
The clinical symptoms recorded included hypoxic-related symptoms, anemia, TIA, stroke,
septic emboli, hemoptysis, hemothorax, and migraines. Note was taken of any factors
that may have contributed to or caused the symptom not related to the PAVM. The clinical
symptoms were subsequently graded as mild (hypoxic-related events, anemia, and migraine)
or severe (TIA, stroke, septic emboli, hemoptysis, and hemothorax).
Procedure
Procedures were performed via a right common femoral vein approach. Initial pulmonary
artery pressures and bilateral pulmonary artery angiograms were performed. Once the
PAVM was identified, the White Lumax Guide Catheter (Cook Medical, Bloomington, Indiana,
United States) was introduced. Often, the feeding vessel was selected with the 5F
inner catheter, and if needed, microcatheters were used. Embolization was performed
with appropriately oversized coils.
Imaging Findings
All angiographic and computed tomographic images were reviewed by one of the authors
for imaging performed at the time of initial embolization as well as any further studies
performed during the follow-up period. Note was made of the size of the feeding vessel
and the location for all PAVMs, whether or not recanalization was present on previously
embolized PAVMs, as well as the presence, number, and location of any new significant
PAVMs. Additionally, note was made of the pulmonary artery pressure prior to treatment.
Statistical Analysis
Descriptive statistics were performed to examine patient demographics. A Fischer exact
test was performed to compare symptoms in patients with rPAVMs alone versus those
with rPAVM and additional new significant PAVMs.
Results
During the 2-year treatment period from January 1, 2006 and December 30, 2007, there
were 101 patients at our institution who underwent PAVM embolization ([Table 1]) with a total of 216 treated for PAVMs. Twenty-seven of these patients were treated
for rPAVM for lesions treated either before this time period at our own institution
or lesions previously treated at another institution. Twenty-six of these patients
were previously embolized with coils, and one had been embolized with a detachable
balloon. None of these patients developed a recurrent rPAVM during follow-up period.
Three patients who underwent initial coil embolization of a PAVM at our institution
developed rPAVM during the follow-up period through December 30, 2013. This resulted
in a total of 30 patients with rPAVMs with a total of 36 rPAVMs.
Table 1
Demographics of 101 patients undergoing PAVM embolization from January 1, 2006 to
December 30, 2007
|
Age
|
|
Mean
|
42 y old
|
|
Range
|
11–85 y old
|
|
Sex
|
|
Male
|
41
|
|
Female
|
60
|
|
HHT
|
|
Yes
|
97
|
|
No/Negative genetic Testing
|
4
|
|
HHT type
|
|
Type 1
|
37
|
|
Type 2
|
7
|
|
VUS ENG mutation
|
2
|
|
SMAD4 mutation
|
1
|
|
Negative genetic testing
|
3
|
|
No genetic testing
|
51
|
Abbreviation: HHT, hereditary hemorrhagic telangiectasia.
Out of the 30 patients with rPAVM, 18 (60%) patients were found to be symptomatic
at the time of presentation of the rPAVM ([Table 2]). Two of the symptomatic rPAVM patients were excluded on the basis of respiratory
manifestations that were felt to be related to medical comorbidities, not their PAVM
([Tables 3], [4]).
Table 2
Patients with reperfused pulmonary arteriovenous malformations
|
Patients
|
Comment
|
|
Total patients treated
|
101
|
|
|
rPAVM
|
30
|
|
|
Reperfused previously embolized PAVM at the time of entry into the study period
|
27
|
26 patients with coil embolized, 1 patient embolized with detachable balloon. None
were again reperfused during follow-up period.
|
|
rPAVM during follow-up
|
3
|
All patients were newly coil-embolized
|
|
Symptomatic with rPAVM
|
18 (60%)
|
|
|
Symptomatic attributed to PAVM or rPAVM
|
16 (53%)
|
|
Abbreviations: PAVM, pulmonary arteriovenous malformation; rPAVM, reperfused pulmonary
arteriovenous malformation.
Table 3
Patients excluded for symptoms felt to be due to comorbidities not rPAVM
|
rPAVM
|
Symptomatic
|
Hypoxic-related symptoms
|
Other symptoms (migraine, anemia, hemothorax, hemoptysis, migraine, TIA, stroke, septic
emboli)
|
Cause for symptom not PAVM related
|
|
Patient 1
|
Yes
|
Yes
|
Yes
|
No
|
OSA, smoker, weight gain
|
|
Patient 2
|
Yes
|
Yes
|
Yes
|
No
|
OSA
|
Abbreviations: OSA, obstructive sleep apnea; PAVM, pulmonary arteriovenous malformation;
rPAVM, reperfused pulmonary arteriovenous malformation; TIA, transient ischemic attack.
Table 4
Symptoms in patients with rPAVMs
|
No. of patients
|
No. of patients with just rPAVM
|
No. of patients with rPAVM and nonembolized significant PAVM
|
|
No symptoms
|
12
|
10[*]
|
2[*]
|
|
Any symptoms
|
16
|
2[*]
|
14[*]
|
|
Mild symptoms only—total
|
13
|
2
|
11
|
|
Respiratory
|
11
|
1
|
10
|
|
Migraines
|
7
|
2
|
5
|
|
Severe symptoms—total
|
2
|
0
|
3
|
|
Hemoptysis
|
1
|
0
|
1
|
|
Stroke or TIA
|
2
|
0
|
2
|
|
Brain or systemic abscess
|
0
|
0
|
0
|
Abbreviations: PAVM, pulmonary arteriovenous malformation; rPAVM, reperfused pulmonary
arteriovenous malformation; TIA, transient ischemic attack.
*
p = 0.0004 on Fischer exact test.
Fourteen of the remaining 16 symptomatic patients with rPAVMs had at least one other
significant PAVM at the time of diagnosis of reperfusion. Only two symptomatic patients
had only an rPAVM. Both these patients presented with mild symptoms: one patient had
respiratory symptoms and migraines whereas the other only had respiratory symptoms.
There were 12 asymptomatic patients with rPAVM. Of these, 10 presented with only rPAVM,
whereas 2 had additionally a new significant nonembolized PAVM. Only two patients
with rPAVM presented with severe symptoms: one presenting with stroke and the other
with TIA. Both these patients had another significant PAVM in addition to rPAVM ([Table 5]).
Table 5
Etiology of PAVM reperfusion on 27 patients included in this study
|
Cause of reperfusion
|
No. of patients
|
|
Recanalization
|
16
|
|
Missed accessory vessel
|
3
|
|
Recanalization and missed accessory vessel
|
3
|
|
Collateral perfusion from a pulmonary artery branch
|
3
|
|
Collateral perfusion from a pulmonary artery branch and recanalization
|
1
|
|
Deflated gold valve balloon and coil recanalization
|
1
|
|
Collateral perfusion from a systemic or bronchial branch
|
1[*]
|
Abbreviation: PAVM, pulmonary arteriovenous malformation.
* This patient presented with recurrent hemoptysis of uncertain etiology; either collateral
reperfusion from a hypertrophied bronchial artery or coil erosion into the right upper
lobe bronchus.
Angiographic examination and possible embolization were performed on all rPAVMs regardless
of symptomatology. Two patients without symptoms and five patients with mild symptoms
underwent angiographic evaluation but not embolization. These rPAVMs were felt to
be too small to cause patients symptoms or were technically challenging. All these
patients underwent subsequent imaging follow-up performed at or reviewed by our institution
with the exception of one patient whose follow-up imaging was performed elsewhere.
Fig. 1 Angiogram with selective catheterization demonstrates recanalization of a right lower
lobe PAVM previously embolized with coils. The patient initially presented 4 years
prior with a brain abscess thought to be due to septic emboli from the PAVM. At the
time of this follow-up angiogram demonstrating recanalization, the patient was asymptomatic
and found to have no additional newly enlarged PAVMs.
The etiologies of the rPAVM included in our study included isolated recanalization
of a previously coil embolized PAVM ([Fig. 1]) in 16 patients, isolated missed accessory vessel in 3 patients, 3 patients with
both coil recanalization and a missed accessory vessel, 3 patients with isolated collateral
perfusion from a pulmonary artery branch, 1 patient with both coil recanalization
and collateral reperfusion from a pulmonary artery branch, and 1 patient reperfusion
due to a deflated gold valve balloon that was placed in 1994 in addition to coil recanalization
([Table 5]). One patient presented with recurrent hemoptysis and previous bronchial artery
embolization. Imaging showed a hypertrophic bronchial artery measuring 2 mm suggesting
this to be the possible etiology; however, upon bronchoscopy, a coil was seen projecting
into the bronchial lumen. Though this patient clearly has a complication related to
PAVM embolization, it is unclear whether the hemoptysis was due to bronchial/systemic
collateralization versus coil erosion.
Symptomatic patients with rPAVM were more likely to have additional significant nonembolized
PAVM than asymptomatic patients with rPAVM, with a significant p value of 0.0003.
Discussion
PAVMs are known to put patients at risk of serious and even life-threatening complications
due to paradoxical embolization and hemorrhage, especially if left untreated.[12] This has led to a recommendation that PAVMs with feeding vessels ≥ 3 mm should be
embolized to prevent migration of larger clots, although more recent literature has
suggested that paradoxical embolization can occur with even smaller feeding vessels.
It may not be possible to determine the patient's risk directly related to feeding
vessel size.[21]
[22]
[23]
[24]
Reperfusion of a previously embolized PAVM has been shown to occur from 5 to 57% of
patients, with a recent study by Woodward et al reporting persistence in 25%.[12]
[17]
[20] Our study showed reperfusion in 16.7% of the PAVMs treated. Reperfusion can occur
through recanalization of the embolized vessel (most commonly described with coils),
a missed accessory vessel, collateral reperfusion from adjacent pulmonary arteries
distal to the site of embolization, or collateral reperfusion from bronchial or other
systemic vessels distal to the site of embolization.[6]
[10]
[17]
[18]
[25] Patients with reperfusion typically undergo retreatment of the rPAVM as it is thought
they have increased risk of paradoxical embolization due to incomplete closure of
the shunt and a continued risk of hemorrhage.
A previous study by Woodward et al looked at 23 patients with PAVM who underwent 71
procedures treating 162 PAVMs, 57 of which had been previously treated. They found
that none of the patients with persistence of a PAVM had brain abscess or stroke during
the time of persistence.[17]
Our study specifically examined the clinical symptoms of patients presenting with
rPAVMs, in addition to documenting any additional new significant PAVMs. To our knowledge,
this is the first study correlating patient symptoms and rPAVMs with additional new
significant PAVMs.
The authors found that patients with clinical symptoms were significantly more likely
to have a nonembolized now significant-sized PAVM (believed to be enlargement of previously
tiny lesions), in addition to the rPAVM. A prior study has shown that growth or enlargement
of PAVMs previously too small to embolize occurs in up to 18% of patients.[6] Of note, none of the patients with severe symptoms in this study had an rPAVM alone.
This leads the authors to believe that it may not be the reperfusion of the PAVM,
which caused patient symptoms but rather the newly significant-sized other PAVMs related
to the underlying pathology and HHT.
It should be noted that all the coil-embolized rPAVMs during the time of this study
were densely packed. This may also account for a reduced risk of macroscopic paradoxical
embolization through such lesions given the small channels available through the coil
matrix. Indeed, one patient embolized with stainless steel coils in the early 1990s
has since developed a small stroke after the end of this study's period from a solitary
significant-sized rPAVM, with only very tiny other PAVMs present. The coils in this
lesion were indeed found to be loosely packed.
There were several limitations to our study, which are in large part due to the retrospective
review design and limited patient sample size. Additionally, patient charts were reviewed
for clinical symptoms. This is prone to recall bias as the patient had to report the
subjective symptoms such a migraines and dyspnea. Finally, the annual occurrence of
a symptomatic paradoxical embolus appears to be fairly low. Though our follow-up spanned
7 to 8 years, this may not be enough to assess the lifelong risk to patients with
rPAVM and so could have underestimated their significance.
Conclusion
The majority of symptomatic patients with rPAVM had additional newly enlarged PAVMs
that required embolization. None of the symptomatic patients with rPAVM alone had
severe manifestations, suggesting that rPAVM plays a small role in patients with recurrent
severe symptoms and may not need to be reembolized. However, given the nature of HHT
and low annual risk from PAVM, close follow-up remains important for to evaluate for
newly enlarging PAVMs as well as rPAVM.
