Abstract
Increasing evidence supports the use of embolisation to treat pulmonary arteriovenous malformations (AVMs). Most pulmonary AVM patients have hereditary haemorrhagic telangiectasia (HHT), a condition that may be associated with pulmonary hypertension.
The current authors tested whether pulmonary AVM embolisation increases pulmonary artery pressure (Ppa) in patients without baseline severe pulmonary hypertension. Ppa was measured at the time of pulmonary AVM embolisation in 143 individuals, 131 (92%) of whom had underlying HHT. Angiography/embolisation was not performed in four individuals with severe pulmonary hypertension, whose systemic arterial oxygen saturation exceeded levels usually associated with dyspnoea in pulmonary AVM patients.
In 143 patients undergoing pulmonary AVM embolisation, Ppa was significantly correlated with age, with the most significant increase occurring in the upper quartile (aged >58 yrs). In 43 patients with repeated measurements, there was no significant increase in Ppa as a result of embolisation. In half, embolisation led to a fall in Ppa. The maximum rise in mean Ppa was 8 mmHg: balloon test occlusion was performed in one of these individuals, and did not predict the subsequent rise in Ppa following definitive embolisation of the pulmonary AVMs.
In the present series of patients, which excluded those with severe pulmonary hypertension, pulmonary artery pressure was not increased significantly by pulmonary arteriovenous malformation embolisation.
Does embolisation of pulmonary arteriovenous malformations (AVMs) precipitate pulmonary hypertension (PH)? The reason this question is important is that for individuals with pulmonary AVMs, embolisation is an effective means of reducing lifetime risks of paradoxical embolic stroke and brain abscess 1, 2, improving oxygenation 3–17 and treating pulmonary AVM-related haemoptysis 18, 19. Conversely, embolisation may be expected to elevate pulmonary artery pressure (Ppa), since pulmonary AVMs are abnormal dilated vessels between pulmonary arteries and veins that provide low resistance pathways for pulmonary blood flow 20.
The question of whether pulmonary AVM embolisation increases Ppa is particularly pertinent, since most individuals with pulmonary AVMs have underlying hereditary haemorrhagic telangiectasia (HHT). Typically recognised by nosebleeds, mucocutaneous telangiectasia and visceral AVMs 21, HHT may be associated with PH 9, 22–30. The secondary causes of PH in HHT are diverse, as they are in the normal population 31, but PH particularly occurs either as a true pulmonary arterial hypertension (PAH) phenotype 9, 22, 28–30 or in the context of high output cardiac failure secondary to hepatic AVMs, when PH may be reversible after hepatic AVM treatment 32. The frequencies of PAH and hepatic AVMs differ with HHT genotype: HHT is caused by mutations in at least five genes, including endoglin (HHT type 1) and ALK-1 (HHT type 2), with pulmonary AVMs most common in HHT type 1 33. PAH phenotypes are more common in HHT type 2 22, 28, 29 than HHT type 1 30. Hepatic AVMs are also more frequent in HHT type 2 33.
Out of >700 reported pulmonary AVM embolisations 1–18, 34–42, data on Ppa measurements pre- and post-embolisation are scarce 9, 11, 17, 32. In three out of the four reported cases 9, 17, 32, each selected from larger series, Ppa increased post-embolisation, while in the fourth it was unchanged 11. There is also a report of worsening PH after surgical resection of a pulmonary AVM 43.
The current authors have previously reported that in a population of pulmonary AVM patients, 92% of whom had HHT, the overall prevalence of PH is low 44. It was hypothesised that, in contrast to the limited data in the literature, pulmonary AVM embolisation would not increase Ppa. Presented herein are results of a retrospective study in the series of pulmonary AVM patients reported recently 2, performed to determine whether pulmonary AVM embolisation affected Ppa, and whether it may be safe to extend this embolisation practice to individuals with severe PH.
METHODS
Study population
All studies were ethically approved by the Hammersmith, Queen Charlotte's, Chelsea and Acton Hospital Research Ethics Committee (LREC 00/5764; London, UK), and performed as part of routine clinical management of individuals with pulmonary AVMs.
Arterial oxygen saturation (Sa,O2) was measured as described previously 2, 4, 45. For Sa,O2 values reported in the present study, recordings were made every 60 s for 10 min standing, since Sa,O2 in the erect posture correlates better with right-to-left shunt 45. All patients with pulmonary AVMs of a size amenable to embolisation treatment underwent pulmonary angiography with a view to embolisation, unless there was a major medical contraindication. In view of theoretical concerns, angiography/embolisation was not considered for four females referred to the service with severe PH and well-preserved oxygen saturations. Pulmonary angiography was performed as described previously 4 in conscious patients who had not been pre-medicated or fluid-restricted before the procedure. Systolic and diastolic Ppa and mean Ppa (P̄pa) were recorded routinely prior to contrast injection via a multi-sidehole catheter (Grollman pigtail catheter; William Cook Europe, Bjaeverskov, Denmark). Measurements were repeated immediately after embolisation in the subgroup of individuals with higher Ppa. In the majority of patients, only a single angiography/embolisation session was required 2.
Statistics
For Sa,O2, the last 4 min readings for erect postures were entered as replicate data for each of the pre- and post-embolisation data. Quartile group data were incorporated from the full series of 219 pulmonary AVM patients, as reported previously 2.
Age was expected to influence Ppa, since catheterisation data at rest in two groups of healthy individuals (17 aged 16–28 yrs 46; 15 aged 61–83 yrs 47) demonstrated significantly higher systolic Ppa (mean 24.47 versus 19.94 mmHg; p = 0.003, Mann–Whitney) and P̄pa (mean 16.13 versus 13.65 mmHg; p = 0.026) in the older group. This has also been supported by more recent echocardiography data 48. In the present patient series, associations of Ppa measurements with age were performed using Spearman rank analyses and linear regression. The patients were also stratified into age quartile groups, and interquartile differences analysed by ANOVA with post-test correction for linear trend.
In order to test whether Ppa was elevated following embolisation of pulmonary AVMs, Ppa measurements recorded pre-embolisation were studied in all patients with at least two Ppa measurements. A total of 39 pairs of consecutive Ppa measurements recorded at the outset of at least two embolisation sessions were available from 35 patients. In addition, nine pairs of Ppa measurements recorded pre- and post-embolisation in the same session were available for eight patients with pre-existing mild to moderate PH, in whom the repeat measurements during the same procedure had been justified as part of their clinical management. To test the null hypothesis that embolisation of pulmonary AVMs does not increase Ppa, measurements pre- and post-embolisation, and pre-embolisation and age-adjusted post-embolisation measurements, were analysed by two-tailed paired t-test and significance assessed at a false discovery rate level of 0.05 49.
RESULTS
Patient populations
Patient ages ranged 8–78 yrs at the time of embolisation (fig. 1a⇓). In patients undergoing embolisation, baseline erect Sa,O2 ranged 73–98%, with a median value of 93% in the full population, as reported previously (fig. 1b⇓) 2. For the four individuals who did not undergo embolisation due to severe pre-existing PH, erect Sa,O2 was in the upper two quartiles of the full pulmonary AVM population, at levels for which the majority of individuals did not experience dyspnoea (fig. 1b⇓) 2.
Baseline characteristics of study population. Comparison of a) age, b) baseline erect arterial oxygen saturation (Sa,O2), and c) mean pulmonary artery pressure (P̄pa) for the 143 patients undergoing embolisation (▪) and the four patients not offered embolisation (○). For b), quartile groups are as previously reported for 219 pulmonary arteriovenous malformation (AVM) patients 2. d) Age-dependent increase in P̄pa in 143 patients with pulmonary AVMs; quartile boundaries are 34, 45 and 58 yrs.
Baseline Ppa measurements (systolic, diastolic and mean) varied widely, as illustrated in table 1⇓ and figure 1c⇑. The distribution of the embolised patients was skewed (fig. 1c⇑), but nevertheless, the four patients who were not offered embolisation represented significant outliers. Further details of their haemodynamic variables are presented in table 2⇓.
Pulmonary haemodynamic variables in 143 pulmonary arteriovenous malformation patients undergoing embolisation
Pulmonary haemodynamic variables in four pulmonary hypertension(PH) patients not undergoing pulmonary arteriovenous malformation (AVM) embolisation
Univariate analysis of first recorded Ppa measurements in the 143 embolised patients indicated that all three Ppa measurements were influenced by age. For age and P̄pa, the Spearman r correlation was 0.33 (95% confidence interval 0.17–0.47; p<0.0001), and P̄pa could be described by the expression 9.88+(0.107×age); r2 = 12.33%, p<0.0001. ANOVA indicated significant differences in Ppa between age quartile groups (fig. 1d⇑). Post-test analysis confirmed a linear trend (p<0.0001), with significant increases occurring between the third (45–58 yr) and upper (>58 yr) age quartiles (fig. 1d⇑).
Effect of embolisation
In order to test whether Ppa was elevated following embolisation of pulmonary AVMs, Ppa measurements recorded pre-embolisation were studied in all 43 patients for whom post-embolisation measurements were also available, either from consecutive sessions (35 patients), or from the same embolisation session (eight patients).
For the 35 patients in whom measurements were made prior to consecutive embolisation sessions, Sa,O2 increased in all except one of the patients following embolisation (p<0.0001; fig. 2a⇓). In contrast, there was no significant change in Ppa as a result of embolisation (fig. 2b⇓). There was a trend towards higher Ppa post-embolisation, but this was in part accounted for by increased patient age, as measurements were recorded at a mean interval of 19.9 (range 2–121) months.
Effects of pulmonary arteriovenous malformation embolisation. a and b) Erect arterial oxygen saturation (Sa,O2) and c) mean pulmonary arterial pressure (P̄pa) recorded at consecutive embolisation sessions (□: systolic Ppa; ░: diastolic Ppa; ▒: P̄pa) in 35 patients. In c), age-adjusted figures were calculated by P̄pa = 9.88+(0.107×age) as defined in the population (see text), and in contrast to Sa,O2, overall p-values did not reach significance at false discovery rate level of 0.05 49. #: p<0.0001; ¶: p = 0.25; +: p = 0.15; §: p = 0.14; ƒ: p = 0.12; ##: p = 0.08; ¶¶: p = 0.05.
Recognising that the pooled groups may have masked individual changes, individual Ppa responses were examined in a subgroup of 15 patients with mean Ppa in the upper quartile. In this small group, embolisation resulted in a highly significant improvement in Sa,O2 (p<0.0001; fig. 3a⇓). In contrast, Ppa measurements recorded at a mean interval of 26 (range 13–80) months demonstrated no consistent trend, and no significant difference between pre- and post-embolisation measurements (p = 0.76; fig. 3b⇓). Comparable findings were observed with systolic and diastolic Ppa (data not shown).
Individual patient details. a) Erect arterial oxygen saturation (Sa,O2) and b) mean pulmonary arterial pressure (P̄pa) recorded pre-embolisation at consecutive embolisation sessions in a subgroup of 15 patients with pre-existing elevated P̄pa (≥16 mmHg). c) Sa,O2 (erect) and d) P̄pa recorded pre- and post-embolisation at the same session in eight patients with pre-existing elevated P̄pa (≥16 mmHg). The two measurements for the patient with balloon test occlusion reported in table 2⇑ are illustrated by dotted lines. #: p<0.0001; ¶: p = 0.76; +: p = 0.0039; §: p = 0.93.
These measurements addressed whether embolisation led to a sustained change in Ppa. In order to explore whether there were any acute changes in Ppa, nine pairs of Ppa measurements recorded pre- and post-embolisation in the same session were examined. Embolisation resulted in a consistent and highly significant improvement in Sa,O2 (p = 0.0039; fig. 3c⇑). Again Ppa responses to pulmonary AVM embolisation varied between individuals. In half of all patients, post-embolisation P̄pa was lower than prior to embolisation and, overall, there was no significant difference as a result of embolisation (p = 0.93; fig. 3d⇑). Comparable findings were observed for systolic and diastolic Ppa (data not shown).
While overall there were no significant increase in Ppa as a result of embolisation, within both groups there were occasional individuals in whom Ppa did increase, by up to 8 mmHg between consecutive sessions, and up to 4 mmHg during the same session. It may be helpful to be able to identify these rare individuals prior to embolisation. Others have suggested test occlusion of the pulmonary AVM before definite embolisation 32. This technique was attempted for one patient early in the series. Balloon occlusion increased Sa,O2 to a similar degree as eventual complete embolisation (table 3⇓). This test occlusion did not significantly increase Ppa, whereas Ppa was significantly higher following maximal embolisation.
Comparison of balloon test occlusion and pulmonary arteriovenous malformation(AVM) embolisation
DISCUSSION
The key finding of the present study was that embolisation of pulmonary AVMs did not lead to a consistent increase in resting Ppa in a series which excluded individuals with severe PAH.
The strengths of the study included the relatively large patient group, correction for age, which could have been an important confounding variable in assessments over consecutive embolisation sessions, and strong evidence of embolisation efficacy. Weaknesses include the retrospective nature of the study, and the reliance on measurements performed for clinical purposes such that pulmonary vascular resistance measurements and same-session repeat measurements following embolisation were not available for the majority of patients. In addition, the study was only powered to address consistent changes pre- and post-embolisation.
Within the study limitations, in the present patient series embolisation of pulmonary AVMs did not generally increase Ppa, even in the setting of mild to moderate pre-existing PH. No patient in the present study developed clinical PAH (i.e. right heart failure) after embolisation.
The current authors were surprised by the significant fall in Ppa in one patient with pre-existing PH attributed to left ventricular disease, and also initially surprised to see that embolisation did not lead to a consistent increase in Ppa in other patients, since effective embolisation occludes vessels that provide a lower resistance to flow than the rest of the pulmonary vasculature 20. None of the patients illustrated in figure 3⇑ were known to have hepatic AVMs, and this may explain the differences between the results of the present study and those of others. Importantly, however, noting that embolisation leads to a reduction in cardiac output 11, 17, the present data suggest that the fall in cardiac output can have a greater effect on pulmonary vascular resistance than occlusion of individual pulmonary AVMs.
While the data indicate that pulmonary AVM embolisation does not necessarily lead to elevated Ppa, Ppa did rise in some individuals; rises which, in this study, were not predicted by balloon test occlusion pre-embolisation. It is important to recognise that in the setting of severe PAH and HHT-associated hepatic AVMs, there are reports that embolisation of pulmonary AVMs may precipitate a fatal increase in Ppa 32. Furthermore, in addition to previously reported cases, clinical PAH developed in another patient in the present series 2 in the years following pulmonary AVM resection and embolisation at another institution.
The current authors' interpretation of these considerations, and of observations from the series reported herein and elsewhere 2, is that for patients with pre-existing severe PAH, the risks of pulmonary AVM embolisation outweigh the potential benefits. The main indications for pulmonary AVM embolisation are to reduce the risk of paradoxical embolic stroke and, for individuals with hypoxaemia, to improve dyspnoea and exercise tolerance. The present authors have recently shown that the risk of paradoxical embolic stroke is substantially lower in individuals with higher Ppa 2. Furthermore, the present data serve as a reminder that pulmonary AVMs generally result in symptomatic dyspnoea only when resting Sa,O2 is <80%, perhaps suggesting that symptomatic relief should not be expected for patients with PH and Sa,O2 >90%, as was the case in all four excluded patients in the study. In the current authors' experience, the most difficult judgements relate to individuals with elevated Ppa and major haemoptysis, a consideration that was not required for the four individuals with PH in this series.
In summary, the data are a useful adjunct to case reports indicating increased pulmonary artery pressure post-embolisation, and indicate that embolisation may be undertaken with caution in the presence of pre-existing mild to moderate pulmonary hypertension in selected individuals.
Support statement
This work was supported by donations from families and friends of HHT patients. The authors are also grateful for support from the NIHR Biomedical Research Centre Funding Scheme. The funding sources played no part in the decision to submit the manuscript for publication.
Statement of interest
None declared.
Acknowledgments
The authors thank K. Sheares for provision of two sets of haemodynamic data for table 2.
- Received September 25, 2007.
- Accepted March 3, 2008.
- © ERS Journals Ltd
References
