This retrospective study was approved by the institutional review board with a waiver of written informed consent.
Study population
A retrospective review of a prospectively maintained institutional database was performed to identify consecutive patients undergoing PAVM embolisation under general anaesthesia (mainly due to patients’ refusal of local anaesthesia) between December 2015 and May 2018. Following the introduction of HFJV at our institution in April 2017, all cases undergoing PAVM embolisation under general anaesthesia were systematically treated using HFJV assistance (group A); prior to this, patients undergoing PAVM embolisation under general anaesthesia were treated using conventional IPPV without HFJV (group B).
All patients were affected by HHT and referred for treatment following multidisciplinary discussion between pulmonologists, interventional radiologists, and anaesthesiologists, to prevent future cerebrovascular accident or ameliorate respiratory symptoms. All PAVMs were confirmed on echocardiography/bubble test, demonstrating right-to-left shunts, and contrast-enhanced computed tomography (CECT). No patients in group A had contraindications to HFJV (chronic airways disease with forced expiratory volume in the first second < 1.5, severe obesity, or recent pneumothorax/thoracic surgery).
Procedures
All procedures were performed on an in-patient basis by the same interventional radiologist (> 7 years of experience in embolisation) under sterile surgical conditions. Procedures were performed in an angiography suite equipped with a flat-panel C-arm cone-beam computed tomography (CBCT) system and XperCT/Embo-Guide tools (Philips Healthcare, Best, the Netherlands). Anticoagulant/anti-platelet therapy and blood clotting parameters were managed according to Society of Interventional Radiology guidelines [26]. Antibiotic prophylaxis (cefazolin, 1 g) and 5000 IU heparin were administered intravenously prior to and during the procedure, respectively.
Anaesthesia and HFJV assistance
Following induction of anaesthesia, muscle relaxation, and orotracheal intubation, conventional IPPV was commenced in all patients (6–8 mL/kg predicted body weight tidal volume; respiratory rate adjusted according to end-tidal carbon dioxide [ETCO2] level; target 30–40 mmHg). For group A patients, IPPV was discontinued and a HFJV dual-lumen cannula (length 40 cm, diameter 2 mm; Acutronic Medical Systems AG, Hirzel, Switzerland) was introduced into the endotracheal tube. HFJV was initiated using a Monsoon II jet ventilator (Acutronic Medical Systems AG, Hirzel, Switzerland) with frequency 150–250 breaths/min, driving pressure 1–2 bar, inspiratory time 30%, and fraction of inspired O2 (FiO2) adjusted to maintain peripheral oxygen saturation (SpO2) > 95%. Blood pressure, ECG, and SpO2 were continuously monitored. ETCO2 monitoring was performed intermittently after 5 min and then every 30 min using the jet ventilator capnography device, by injecting five long insufflations with IPPV followed by measurement over 10 s (with the jet switched off). Once embolisation was completed, the HFJV cannula was removed and conventional ventilation was resumed until patient’s awakening/extubation.
PAVM embolisation
Following ultrasound-guided placement of a 6–8-Fr sheath (Pinnacle, Terumo Corporation, Tokyo, Japan) in the femoral vein, a 4-Fr Imager TM II pig-tail catheter (Boston Scientific, Marlborough, MA, USA) was advanced into the target main pulmonary artery, and diagnostic angiography was performed to localise PAVMs and identify number/size of feeding arteries. Feeding arteries were selectively catheterised using a 6-Fr catheter (Neuron catheter, Penumbra Inc., Alameda, CA, USA), and when required, a co-axially 4-Fr MPA 125-cm catheter (Cordis, Johnson & Johnson Company, USA) was used. The 6-Fr catheter was continuously flushed with heparin-saline solution (2500 IU/L) via a Y connector to avoid thrombus formation and minimise neuro-embolic risk. For group A, navigation was assisted by a simple 2D roadmap (Fig. 1) in four patients, and a continuous 3D roadmap automatically delineating feeding arteries (generated by Embo-Guide software following initial 3D rotational angiography with 40 mL of contrast agent at 8 mL/s) was used in one patient. No 2D/3D roadmaps could be generated for group B. Embolisation was performed using conventional or microvascular plugs (MVP, Medtronic, Minneapolis, MN, USA) and/or platinum detachable microcoils with/without hydrogel coating (Ruby Coil, Penumbra Inc., Alameda, CA, USA; Azur Coils, Terumo Corporation, Tokyo, Japan), depending on feeding artery anatomy and available technologies during the study period.
Devices were deployed as distally as possible within feeding arteries, via 4–6-Fr catheters for conventional vascular plugs, or via a co-axial 2.8-Fr microcatheter (Progreat, Terumo Corporation, Tokyo, Japan; inner diameter 0.027″) for microvascular plugs/microcoils. Vascular occlusion was confirmed via selective feeding artery angiogram. A final angiographic check was performed after 3–4 min to confirm satisfactory device position and occlusion. For group B, angiograms throughout the procedure were variably performed with induction of patient’s apnoea, depending on operator requirements. For group A, all angiograms were performed with HFJV, other than a final acquisition under conventional IPPV to compare differences in diaphragmatic excursion.
Following the procedure, patients were admitted to a recovery ward for 24-h observation and discharged when medically fit.
Follow-up
All patients were clinically reviewed by the treating interventional radiologist and underwent CECT to confirm PAVM occlusion 1 month post-procedure. Serial echocardiography/bubble tests were performed at 1, 6, and 12 months to monitor for right-to-left shunts.
Data collection and statistical analysis
Patients’ demographics; number, location, and type (simple or complex) of PAVMs treated per session; and type of embolic device were tabulated. Primary outcomes were patient radiation dose and procedural time. Secondary outcomes included mean intraprocedural diaphragmatic excursion, considered as a surrogate for whole-lung movement in group A versus B; mean diaphragmatic excursion in group A on final angiogram with HFJV versus IPPV; technical success; complications (according to the classification System of the Cardiovascular and Interventional Radiological Society of Europe [27]); and clinical success.
The patient radiation dose was calculated as dose-area product by the proprietary angiographic software (Philips Healthcare, Best, the Netherlands). Procedural time was calculated between the first and last angiographic acquisition.
Diaphragmatic excursion was calculated for each case on digital subtracted angiographic series (duration at least 15 s), measuring the cranio-caudal distance between the upper- and lower-most positions of the superimposed hemidiaphragmatic cupolas and recording the lowest measurement for each procedure (Fig. 2).
Technical success was defined as distal occlusion of the feeding artery < 1 cm from the PAVM aneurysmal sac, with complete devascularisation on final angiogram. Clinical success was defined as the absence of neurologic/respiratory symptoms and echocardiographic right-to-left shunt at the last available clinical/sonographic follow-up.
Statistical analysis was performed using R software (R v3.4.5, R Foundation for Statistical Computing, Vienna, Austria). Non-parametric Wilcoxon test was used to compare numerical and discrete variables; p values < 0.05 were considered statistically significant.