Basivertebral foramen (BF) is a bone defect in the posterior wall of vertebral body (VB), between the pedicles, containing basivertebral veins, arteries and nerves (Figure 1). BF existed in every vertebra of the spine (1–5). It has a bigger size than other vessel passage of bone (6). The intravertebral circulation links to vertebral venous system via BF, through so-called basivertebral vein, single or sometimes with two separate tributaries. All the venous system is valveless, and it eventually drains to Vena Cava (7–9). Studies implied that BFs had a relationship with lots of diseases and phenomenons, such as tumor metastasis, postero-superior fragment in burst fracture of thoracolumbar spine and cement leakage of percutaneous vertebroplasty or kyphoplasty (type B cement leakage) (10). Our former research also revealed the presence of a connection between the basivertebral foramen and the intravertebral cleft (11), and the possibility of direct cement leakage into the spinal canal through the basivertebral foramen when PKP is performed in senior patients of osteoporotic vertebral fracture with intravertebral cleft (12).

To this day there's no systemically study about the morphological features of BF, although which is the fundament of the recognition of the relationship between BF and vertebral diseases. Spinal metastasis was thought to be happen firstly in pedicle arch, because x-ray examination showed the malignant colony firstly settling in this region (13). But Algra et al. (14) found that, explicated by CT, spinal metastasis often happened firstly at posterior part of VB because of BV, since basivertebral vein served as an entrance site for hematogenous spread of metastases (9, 15–18). Algra et al. (14) also described that the lumbar region of the spine was the area most frequently affected by metastases; next were the thoracic and cervical portions. Lumbar spine bears the largest loading in the whole spine, while a big bony defect in the VB such as BF would impose important biomechanical influence on the VB, especially in the lumbar spine. Thus, the BFs of the lumbar spine became the first subjects we studied among the spine.

Spine degenerates as human being getting aged. Bone density decreases, and the shape of vertebrae also changes with age. In series of reports by Erichsen between 1976 and 1978, lumbar vertebrae showed relative lowering and broadening of their bodies with age (19–21). As to easterners, Kim et al. (22) found that VB height decreased with age in thoracic vertebrae (T5, T6) and lumbar vertebrae (L4, L5) among Koreans. As the VBs get less bone mass and flattened with age, we suppose that BFs would get shallower and higher compare to VBs. To prevent the influence of size change of BVs among different age groups, we calculated relative values of BFs to BVs and compared these relative values.

According to morphological characteristics of basivertebral foramen, we classified them as 3 types: triangular type (65%), trapezoid type (24%) and irregular type (11%). The bone interval exsist in 6% of the basivetebral foramen. The average basivertebral height is 9.2 ± 2.5 mm, which is approximately 1/3 of vertebral height at the same level, whilst the average basivertebral depth is 4.4 ± 1.6 mm. Compared with the lower endplates, the basivertebral foramen are statistically more close to the upper endplates at all five measured levels (P < 0.05, Figure 3). The BFHr at the lower level tends to be relatively smaller, while BFDr at L2 is the biggest among five vertebral levels (P < 0.05, Figure 4). People above 60 have significantly bigger BFHr than those in young adulthood and in the middle ages at all five levels, while they have shallowest BFs, especially at L3 (P < 0.05, Figures 5, 6). Females have higher BFHr and deeper BFDr than males in lumbar vertebrae, which only BFHr at L2 and L4 show statistically significant difference (P < 0.05, Figure 7). Multiple linear regression resulted in a formula that accounted for 30.1% of the variability in the height of basivertebral foramen. As detailed in Table 1, significant predictors included: gender, age, level, vertebral height and heavy work.

Although many studies have investigated the anatomy, circulation, innervation and biomechanical features of spine, the BV is still a blind zone of research. Our study and some previous studies showed that BV, such a big bony defect, lies in middle column according Denis and Ferguson's (23–25) three column theory of spine, anterior to spinal canal and posterior to the center of VB, a little closer to upper endplate. The size and location of BV could tell the truth of many clinical phenomenons such as burst fracture and type B cement leakage in PVP and PKP. The strain around the BV is the highest within the entire VB and along the surface of VB (26, 27), so posterior regions of the VB had greater bone volume, more connections, reduced trabecular separation and more plate-like isotropic structures than anterior regions (28). Although the superior part of posterior BV had better bone quality, the factors of the thinner upper endplate and closer BV seemed to share more in burst fracture, resulted in posterior superior fracture fragment (28–32). As the ratio of BF's height to VB's height largens according to aging, the posterior region and wall of VB would burden more strain, as a result, which would become fragile under violence. Thus, we can easily found that osteoporotic compression fracture of aging VB resulted in flattened VB while the younger non-osteoporotic VB were wedge-shaped after compression fracture. Thoracolumbar spine (T11–L1) (33, 34) is the most often fractured in spine, results from which previous studies argued that this span of spine is a transitional section and with lots of motion. In our study, the BFHr of L1 was the largest among the lumbar vertebrae, which could be a cause of the high risk of fracture of L1.

Cement leakage happened quite often in PKP/PVP (35–37), especially under CT examination post-operatively (38–40). Extravertebral cement leakage via the basivertebral veins (so-called type B) and fractured cortex (so-called type C) are more than other type. Type B is more dangerous than type C (37), and would be harmful (7, 39, 41). Because of BF, unilateral injection of cement had more cement leakage than bilateral injection (42). The needle head would get closer to BF when surgeon tried to put it into the center of VB. Our study showed that the depth of BF is up to 15% of anteroposterior diameter of VB, so the region anterior to the posterior 15% of VB is safe for cement injection. This is consistent with the founding of Yeom et al.'s study (37), who declared that cement existed in posterior 1/5 of VB might be a leakage.

The basivertebral nerves which might be the sole passage that enabled intravertebral innervations made BF more important (43, 44). Intravertebral nerves clustered at the center of VB, after descended from basivertebral nerves. That's why a little cement injected into the VB could also ameliorate symptom in osteoporotic-vertebral-fractured patients by PVP/PKP, just like larger volume (45). As mentioned before, the calibration of BF is larger than other nutritional foramen of bone. The existence of basivertebral nerves could make a reason. Although VB's volume changed under compressive loading via flow through the BF, and the basivertebral vein deformed in the course (46), it's still unimaginable that it needs a passage of over 1/3 of VB's height (except for L5 which is 29%). However, sufficient space for basivertebral nerves is essential not only to tolerate the enlarged BF but also to take precautions against deformation of BF that resulted from loading, fracture, hyperostegeny and so on.

The vascular system of VB is extremely complicated (10, 47). The central vascularity of the vertebral body which originated from vessels entering BF (48, 49) is established during fetal development, while peripheral vessels persist after adolescence (50). Tortuosity of the central intravertebral arteries and presence of vessels around the vertebral periphery increases with age (50) and is thought to be variable based on degeneration and nutrient availability. In our study, the depth of BF tended to be less as aging, especially in L3, which had significant tendency. It could be implied that as peripheral vessels increasing when aging, the nutrition supply via central vessels attenuated, so the depth of BF decreased. It is also possible that shallowed BF, because of microfractures or creep deformation, compromised blood supply of VB, causing VB degeneration under malnutrition, and then resulting in peripheral vessels infiltration. However, central vessels are still one of the main nutritional passages for the BFs in the aged VBs were just a little shallower than those in younger ones. Algra et al. reported that disappearing BF might have relation with bone marrow disease of VB (51).

Female had higher and deeper BF than male, although only L2 and L4's rates of BF's height to VB's height showed significance. Although female have smaller body size and VB size, the BFs were relatively bigger than male. Heavy work was a negative factor of BF height, which is intuitively understood, attributed to increased creep deformation and microfractures of VB.

This study has the following merits: the factors analyzed in our study are easily and reliably determined in the clinical practice. Besides factors that could be obtained by inquiry, the vertebral height could be acquired from cheap and time-saving x-ray examination, with acceptable radiation.

However, some limitations could not be ignored. First, as a bony defect, BF is most explicit in CT. In our region, the cost of MRI is just a little more expensive than CT, but with sparse radiation. Although MRI is time-consuming, patients still prefer MRI examination about spine. It costs us a lot of time to calibrate the parameters while carefully identified the low signal of cortex of VB, but it is not as accurate as calibration in CT image. Second, we couldn't get the width of BF. The images only contained transverse section of intervertebral discs. Although occasionally clinical CT examination of spine showed large width of BF (Figure 8), it needs further prospective study to reveal the truth. Third, some factors that might affect the BF such as blood pressure and race were not included in the study. Whether and how blood pressure, especially hypertension, influences BF is an interesting topic, we would carry out such a study to explicit the relationship.

Age is the highest weight in all factors on the height of BF. BF is closer to the upper endplate. The BF was relatively higher and deeper in the female lumbar spine. Heavywork results in lower BF. Last but not the least, as we supposed, BF gets shallower and higher compare to VB with age.