When massive hemoptysis is untreated, it has a mortality rate of over 50 percents. It is considered as one of most dreaded of all respiratory emergencies and can have a variety of underlying causes.

Bronchial artery embolization (BAE) was first reported in 1973 by Remy1 and it has become an established procedure in the management of massive and recurrent hemoptysis. Its efficacy is widely documented thereafter by number of articles1-11. An immediate control of hemoptysis is achieved in 73 to 98%, with a mean follow of less than one month 1-4. Immediate success rates have increased recently because of the introduction of superselective embolization and the refinement of embolic agents and techniques4. However, the long-term success rate of BAE is known to be unfavorable. Longterm recurrence rates are reported to be 10 to 52%, with a mean follow up period ranging from one to 46 months1-7.

The variety of factors influencing that control failure has been described.

Bronchiectatic change on high resolution CT scan (HRCT)4, broncho-pulmonary shunt4, pleural thickening5,7, underlying lung diseases6, the amount of bleeding8, multiple feeding vessels9, incomplete embolization10, and previous hemoptysis history11 are possible risk factors of recurrent bleeding events. But these findings vary from article to article and there is not yet a proven condition to predict the recurrence.

This study is designed to survey previously documented possible risk factors of recurrence in those who underwent BAE in our hospital, during a long period. Furthermore, since all patients had taken HRCT, we focused on radiological findings such as pleural thickening to be possible risk factors of the recurrence.

Seventy-five patients underwent bronchial artery embolization due to massive hemoptysis more than 100 cc amount of bleeding in Severance Hospital from Jan. 2000 to Jan. 2005. Among them, nine patients' data were not available and could not be contacted with. Finally 66 (48 males, 18 females) patients' medical records were analyzed retrospectively with a mean follow up period of 20.4 months (ranging from 1 month to 56 months).

The recurrence of massive hemoptysis after BAE was defined that a gross hemoptysis more than 100 cc amount of blood occurred again which needed to undergo treatments such as BAE or lung resection surgery.

Demographic characteristics such as gender, age, the duration of symptom, the amount of hemoptysis, previous history of treatment for hemoptysis, the main medical condition causing hemoptysis, bilaterality of pulmonary lesion, the number of feeding vessels, and the presence of pleural thickening were our parameters analyzed.

For comparison of various risk factors between recurrent massive hemoptysis group and controlled hemoptysis group, Pearson's Chi-square test (for n > 5) and Fisher's exact test (for n ≤ 5) were used. For continuous variables, Student's T-test was used. Statistical validity is defined if p-value is less than 0.05.

To verify the compounding factors between the risk factors whose characters were diverse confounding variables, we did multivariate analysis using logistic regression analysis with 95% confidence interval (Table 1).

Among 66 patients whose data were available, 23 (34.9%) patients had recurred massive hemoptysis during a mean follow up period of 20.4 months (ranging from 1 month to 54 months). Eight out of 23 recurred patients had pneumonectomy or lobectomy of lung (3 pneumonectomies and 5 lobectomies) in following event, and remaining 12 patients had to undergo another BAE. Four patients died due to uncontrolled hemoptysis (one died after pneumonectomy).

Concerning the factors influencing relapse, age and sex did not play any role and duration of symptom had not any significance, but the amount of hemoptysis had a statistical significance (p =0.008 by T-test; Table 2).

As to the underlying diseases, we had 20 active tuberculosis, 19 benign diseases including 19 bronchiectasis, 14 aspergillomas, and 6 others (including 1 bronchial artery aneurysm, 2 lung abscesses, and 2 bronchitis), and 7 malignancies (4 primary lung cancer and 3 metastatic malignancies), we did not find any statistical significance between them (Table 3). Nine patients had a history of previous BAE treatments for hemoptysis, but they showed no increased risk of relapse of major hemoptysis (Table 3). Bilateral lesion on radiographic finding and pleural thickening on HRCT had increased risk of recurred major hemoptysis (p = 0.018 and 0.001, respectively, by Chi-square test and Fisher's exacttest; Table 4). Number of feeding vessels had not any statistical value concerning the relapse.

Univariate analysis showed that reliable risk factors for recurrent hemoptysis after BAE were amount of hemoptysis, bilaterality of lesion, and pleural thickening on HRCT. To verify the compounding factors between the risk factors whose characters were diverse confounding variables, we did multivariate analysis using logistic regression analysis with 95% confidence interval. Multivariate analysis showed same results as univariate analysis (Table 5).

We analyzed the result of BAE in our institution during a long period. In previous literatures, longterm recurrence rate have been reported to be 10 to 52%, with a mean follow up period ranging from one to 46 months1-11.

Remy et al.1 reported that of 49 patients treated for hemoptysis, an immediate arrest was achieved in 41, but 34 patients had experienced re-bleeding in follow up period beyond the 18 months. Ulfacker et al.3 reported that an immediate control of hemoptysis was achieved in 33 out of 41 patients (80.5%) while hemoptysis recurred in 9 of 33 patients (27.3%) in the long-term follow up (mean 24.8 months). In our analysis, BAE effectively controlled 65.1% of life threatening massive hemoptysis (23 re-bleedings in 66 cases) in a mean follow up period of 20.4 months.

Regarding the factor of recurrence after BAE, Osaki et al.4 concluded that bronchiectatic change on CT scan and pulmonary-bronchial shunt had some statistical significance. Kim et al.6 described the underlying lung disease and amount of bleeding as reliable risk factors for the recurrence, in a study involving 75 patients with a result estimating 54.5% of re-bleeding rate after 3 years. But in the study of Kim et al14 published 5 years earlier, those factors had no significant impact on the recurrence of hemoptysis. The former used Caplan-Mayer survival analysis with each variable and the latter used Chi-square univariate analysis. The diversity of previously proposed risk factors may be explained by variability of their criteria on recurrence, sample size, underlying diseases, follow-up time, and statistic tool used.

In this study, the amount of hemoptysis had some statistical relation with the recurrent event. Though the analysis of underlying disease had no statistical significance, active tuberculosis tended to have more control rate compared to aspergilloma and cancer. The effective anti-tuberculosis drug therapy must have reduced the recurrent hemoptysis but its relatively modest prevalence in our series (30% compared to 43-52% in other domestic studies)5,6,12,14 lead to overall no statistical significance.

The fact that bilaterlity of lesion on initial chest X-ray was higher in relapse group can be explained by the extent of the lung disease accounts for more serious pathology as it was commented by Kim et al.6

In 1993, Tamura et al.8 described pleural thickening as a risk factor for recurrent bleeding after BAE. According to them, in the presence of pleural thickening, non-bronchial systemic feeder vessels that originate from various arteries (e.g., intercostals artery, branches of the subclavian and axillary arteries, internal mammary artery and inferior phrenic artery) may develop along the pleural surface and become enlarged as a result of the inflammatory process. In our study 16 cases showed pleural thickening on chest radiography and 12 (75%) of them experienced recurrent massive bleeding, which was significantly higher than 22% of no pleural thickening group.

This study is a retrospective review of medical records, which often should underestimate strength of variable. But the presence of pleural thickening which had the highest odds ratio can be a reliable risk factor for the recurrence of hemoptysis after BAE, and this should be verified in a prospective study with larger number of patients because this study has the limitation caused by small size population involved with various underlying diseases.