Date Published: March 5, 2019
Publisher: Public Library of Science
Author(s): Liehua Liu, Shiming Cheng, Qian Wang, Qiang Liang, Yong Liang, Weidong Jin, Qiang Zhou, Zili Wang, Dennis Maiman.
To observe the regional anatomy of the lumbar artery (LA) associated with the extrapedicular approach applied during percutaneous vertebroplasty (PVP) and percutaneous kyphoplasty (PKP), we collected 78 samples of abdominal computed tomography angiography imaging data. We measured the nearest distance from the center of the vertebral body puncture point to the LA (distance VBPP-LA, DVBPP-LA). According to the DVBPP-LA, four zones, Zone I, Zone II, Zone III and Zone IV, were identified. LAs that passed through these zones were called Type I, Type II, Type III and Type IV LAs, respectively. A portion of the lumbar vertebrae had an intersegmental branch that originated from the upper segmental LA and extended longitudinally across the lateral wall of the pedicle; it was called Type V LA. Compared with the DVBPP-LA in L1, L2, L3 and L4, the overall difference and between-group differences were significant (P < 0.05). In L1, L2, L3, L4 and L5, there were 8, 4, 4, 0 and 1 Type I LAs, respectively. There were no Type V LAs in L1 and L2, but there were 2, 16 and 26 Type V LAs in L3, L4 and L5, respectively. In L1-L5, the numbers of Type I LA plus Type V LA were 8, 4, 6, 16 and 27, and the presence ratios were 5.1%, 2.6%, 5.6%, 10.3% and 17.3%, respectively. In L4 and L5, the male presence ratios of Type I LA plus Type V LA were 7.1% and 10.7%, respectively, and the female presence ratios were 13.9% and 25.0%, respectively. Thus, extrapedicular PVP (PKP) in lumbar vertebrae had a risk of LA injury and was not suggested for use in L4 and L5, especially in female patients.
In 2005, Han et al.  reported extrapedicular percutaneous vertebroplasty (PVP) in the treatment of thoracic vertebral compression fracture. In 2007, Ryu et al.  reported the surgical technique and clinical effects of extrapedicular percutaneous kyphoplasty (PKP) with a single balloon in 13 lumbar vertebrae. In 2011, Cho et al.  reported extrapedicular PVP and PKP in 74 lumbar vertebrae, confirming the efficacy and feasibility of the extrapedicular approach for lumbar PVP (PKP). The main advantage of the extrapedicular approach is that the puncture needle can easily reach the midline of the vertebral body to facilitate bone cement diffusion in the central part of the vertebral body. Moreover, bilateral puncture is not needed, reducing trauma and saving operative time. In addition, an extrapedicular approach puncture can maintain the integrity of the pedicle cortex, preserving the axial or lateral biomechanical stability of the spine. 
This study included 42 males and 36 females, aged 44.0 ± 10.5 years old (20–70 years). The average male age was 45.7 ± 11.7 years old (22–70 years), and the average female age was 42.0 ± 8.7 years old (20–61 years). Age was not significantly different between males and females (P = 0.112). The LA number and absence, intersegmental branches and DVBPP-LA in each lumbar vertebra are shown in S1 Table and Table 1 (Fig 2A and 2B). The 1st and 2nd LA origin were mostly parallel to the upper edge of the inferior vertebral body, and the 3rd LA origin was mostly parallel to the L3/4 intervertebral space. LAs in L1-L3 inclined on the outward and upward sides of the vertebral body. The 4th LA origin was in front of the middle part of the L4 vertebral body, which passed backward and downward through both sides of the L4 vertebral body (Fig 3A). In L1 and L2, LAs entered intervertebral foramens in the anterior and superior positions of the intervertebral foramen, while in L3 and L4, LAs entered intervertebral foramens in front of the intervertebral foramen. The presence ratios of LAs in L1-L3 were greater than 98%, and the ratio in L4 was 84.6%, indicating that 24 LAs were missed. Only 16 (10.3%) LAs were observed in L5, and 8 of these LAs originated from the iliolumbar artery, 4 from the abdominal aorta, 3 from the internal iliac artery and 1 from the common iliac artery (Fig 3B, 3C, 3D and 3E). In the absence of normal LAs, an intersegmental branch appeared in some lumbar vertebrae. The numbers of intersegmental branches in L3, L4 and L5 were 2, 16 and 26, respectively. The DVBPP-LA in L2 was significantly different between males and females (P = 0.001), but the DVBPP-LA in L1, L3 and L4 was not significantly different between males and females (P > 0.05) (Fig 2B). Compared with the DVBPP-LA in L1, L2, L3 and L4, the overall difference and between-group differences were significant (P < 0.05). In L5, a smaller number of LAs were observed, and only 2 branches were observed in females. Thus, the DVBPP-LA in L5 was not included in comparative statistical analysis. LA injury is more common in abdominal trauma and knife stab wounds, [13–14] but there has been no lack of reports on iatrogenic LA injury, which is more common in spine surgery [6–9,15–18] and urinary surgery. [19–20] In a previous meta-analysis, the surgical technical complication incidence associated with PVP (PKP) ranged from 1.8% to 3.8%. [21–22] However, vascular complications associated with PVP (PKP), such as aortic adventitial injury,  infection-induced aortic aneurysm  and LA injury, [6–9] have not been systematically documented. Giordano et al.  reported a ruptured LA pseudoaneurysm after PVP and cryoablation for a L2 metastatic tumor, resulting in a sudden and large number of retroperitoneal hemorrhages and hypovolemic shock. The authors noted that it was necessary to preoperatively confirm the anatomy of LAs through enhanced CT and puncture path design. Ajit et al.  reported two cases of 3rd LA pseudoaneurysms after PVP and suggested that it was very important to master the bony anatomical landmarks and the vascular region of the vertebral body through clear fluoroscopy, especially being alert to LA injury during the extrapedicular approach puncture. Source: http://doi.org/10.1371/journal.pone.0213164