The sacrum is composed of five vertebral segments (S1-S5) that typically fuse during adulthood, making the sacrum a reinforced block of bone [1, 5, 13, 14]. This fusion of vertebrae not only increases the functional strength of the sacrum but also allows for the formation of a continuous osseous canal to contain the tapering spinal cord and dural sac, as well as four pairs of anterior and posterior foramina to allow the passage of sacral nerve roots [1, 14]. These neuroforamina serve a role anatomically as well because the sacrum is often divided into two regions based on proximity to the foramina. The body is composed of the sacrum medial to the neuroforamina and contains the spinal canal, and the ala is composed of bone lateral to the neuroforamina . The sacral roots exiting the foramina join the lumbosacral trunk to form the sacral plexus, which carries motor and sensory fibers to the pelvic organs and is vital to sphincter control and genital organ function . In addition to containing vital neurologic structures, the sacrum plays a major role in transferring forces from the upper body to the lower extremities during movement [1, 5].
In the sacrum, the S1 and S2 segments provide the most structural support because of their ability to directly articulate with the spine, and strong ligaments transfer some of the load to the pelvis and spine . Biomechanically, the sacrum counteracts anteriorly directed axial force from the spinal column and balances two opposite force vectors between the body’s weight (downward) and resistance to the ground (upward) . The sacrum and its surrounding ligaments form a foundation of structural integrity for both the lumbar spine and posterior pelvic ring, and allow a seamless transition of force from the upper body to the lower extremities . Although the sacrum is strong, the main weak points of the bony formation are the sacral foramina in the sagittal plane and anterior sacral crests in the transverse plane, and because of this, fractures in the sacrum most commonly occur in these locations . Due to the strength of the sacrum and its many attachment sites, a sacral fracture in a metabolically healthy individual without osteoporosis requires a high-energy injury to disrupt the structural integrity of this area .
The complexity of fracture patterns and anatomic regions has resulted in the development of several classification systems to better characterize sacral fractures. In 1939, Medelman  grouped sacral fractures into longitudinal, oblique, and horizontal types . Although this was one of the first classification systems, the lack of clinical relevance and treatment guidelines made it less useful than other systems in terms of its ability to guide clinical decision-making . In 1947, Bonnin  recognized the complexity of sacral fractures and subdivided them into six different types. These six types were based on fracture location in the sacrum and type (comminuted, fissure, or transverse). Although Bonnin’s classification recognizes commonly seen patterns of sacral fractures, it does not correlate the fracture with mechanism of injury or assist with clinical evaluation, treatment, or prognosis [3, 5]. Starting approximately three decades after Bonnin  published his work, several groups offered alternative approaches to classifying these injuries, although due to a lack of clinical correlation, none of these classifications were popularized [5,6,8].
In 1988, with the advancement in CT availability, Denis et al.  established a classification system based on the direction, location, and level of fracture with an emphasis on neurologic injury . The Denis classification has proven instrumental in describing these fractures and has resulted in the development of several subclassification systems that have arisen as an expansion of the original Denis description .
The aim of the classification system was to assist clinicians in recognizing and quantifying the likelihood of neurologic injuries in patients with sacral fractures. An ideal classification system will group similar diagnoses, guide treatment options, provide prognostic information, stratify research, and do all of this this with a high degree of reproducibility. Denis et al.  ultimately thought to provide an improved framework to explain the mechanism of injury, pathoanatomy, and overall severity of sacral fractures. Their analysis was based on thorough and detailed evaluations, radiographs, and surgical findings of patients with sacral fractures.
In their original description, Denis et al.  described three zones of injury: Zone I: injuries located lateral to the neuroforamina; Zone II: injuries that involve the neuroforamina, but not the spinal canal; and Zone III: injuries that extend into the spinal canal, with primary or associated fracture lines. The study found that neurologic injuries occurred in 5.9% of fractures lateral to the sacral foramina (Zone 1). In transforaminal fractures (Zone 2), 28.4% of patients had a neurologic deficit. Meanwhile, central fractures (Zone 3) had the highest likelihood of neurologic injury (56.7%)  (Fig. 1). Furthermore, Zone III fractures are often divided into four different types, including: (1) flexion fracture with anterior angulation, (2) flexion fracture with anterior angulation and posterior displacement, (3) extension fractures with anterior displacement, and (4) comminuted fracture of the upper segment of the sacrum without displaced alignment of sacrum .
One study with three observers (one spine surgeon and two residents) evaluated the CT images and MRIs of 53 patients with sacral fractures and repeated these evaluations 6 weeks later to test the reliability of the Denis classification. The results of that study revealed an interobserver kappa value of 0.60 for CT and a kappa value of 0.52 for MRI. Intraobserver reliability ranged from kappa = 0.66 to 0.93 based on the individual evaluator .
One element of the Denis classification that we think is useful is the way that it assesses fractures in relation to the neural foramina—though the kappa values described are indicative of fair/modest reproducibility. A sacral fracture’s location relative to the neural foramen can be used to predict neurologic deficits; the likelihood of identifying a sacral fracture is increased if an associated neurologic injury is present . Reports more commonly focus on a single Denis zonal injury, and these studies had findings similar to those of Denis et al. [6, 10]. One example is an anatomic study in which 35% of patients with transverse sacral fractures (most often Denis Zone III) had a sacral root transection on postmortem examination . Another study demonstrated a 65% incidence (22 of 34 patients) of neurologic deficits with vertically unstable patterns. Of these, 11 were classified as Denis Zone I injuries, 13 were Zone II injuries, two were Zone III injures, and eight were bilateral pelvic ring injuries . Expanding on this work, in another study, the same authors reported 18 patients with complete bowel and bladder dysfunction after presenting with a “U-type” sacral fracture .
Although it remains the standard classification system for sacral fractures because of its clarity and ability to predict neurologic deficits, the Denis classification has fundamental problems. Without question, its major limitation is its lack of validation regarding inter- and intraobserver reliability. Without robust validation of those parameters, the user cannot know whether two observers would make the same diagnosis of the same x-ray, CT, or MRI (or even whether a single observer can do so consistently). Insofar as a fracture classification is meant to guide treatment, prognosis, communication, and research, its ability to achieve those goals must be considered extremely limited until a classification is validated robustly. To our knowledge, only a single study  has sought to validate the Denis classification, and that study involved only a single spine specialist, no general orthopaedic surgeons, and two trainees doing the analyses of the images (and the number of images they evaluated was relatively small). Finally, the interobserver reliability of the Denis classification in that study was far from perfect. In light of all this, it is impossible to know whether the wide use of the Denis classification is justified or even safe.
Furthermore, the classification system is primarily descriptive and it generally lacks consideration of treatment options; this may lead to ambiguity, prompting surgeons to rely on treatment methods that are not necessarily evidence-based . If the Denis classification system had included the affected axial level and bilaterality of the fracture, it would have added more clinically relevant information because patients with transverse fractures involving the S1-S3 segments have higher rates of bladder dysfunction than do patients with fractures at more caudal levels . Furthermore, bilateral fractures in Zones 1 and 2 are uncommon, and a more careful evaluation of imaging frequently revealed a previously unrecognized Zone 3 injury with an occult transverse fracture line .
The development of the Denis classification coincided with and benefited from wider dissemination of transverse CT imaging technology around 1988. However, more recently developed technologies such as CT with multiplanar reformation and MRI reveal that these injuries may have a substantial degree of ligamentous involvement, especially in the structures of the posterior column, as noted by Oner et al. . Even with this knowledge, it is difficult to retroactively integrate this into the Denis classification scheme. As a result, these ligamentous injury patterns can only be categorized based on patterns of bony involvement, without consideration of other associated injuries. This makes it difficult for surgeons to reliably predict stability, especially that of the pelvic ring .
Even when focusing on the bony pathology that the classification is designed to characterize, surgeons can sometimes overlook horizontal fractures unless fine-cut CT scans and sagittal reconstructions are performed and carefully reviewed . Although these fractures rarely affect the stability of the pelvic ring, depending on the proximity of the fracture to the sacroiliac joints, spine stability might be compromised. For example, if a horizontal fracture is below the sacroiliac joint level, it is considered stable, although damage to the spinal nerve roots or cauda equina syndrome may result from occlusive fracture fragments. When the fracture is at the sacroiliac joint level, there is often an associated vertical component through the bilateral neuroforamina of the sacrum. These fractures can adopt a U- or H-shaped morphologic pattern and may reflect a loss of continuity between the pelvis and the spine, or “lumbopelvic dissociation.” These highly unstable fractures may require triangular osteosynthesis and spinopelvic fixation, and their omission from the original Denis classification overlooks a clinically important sacral fracture pattern.
An ideal classification system is reliable, reproducible, and able to accurately communicate information about prognosis to providers involved in patient care . Though commonly used, the authors of this paper recommend cautionary use of the Denis classification due to its overall lack of validation regarding inter- and intraobserver reliability. As stated, insofar as a fracture classification is meant to guide clinicians in treatment, prognosis, and communication, the Denis classification’s ability to achieve these goals must be considered limited until the classification is more thoroughly validated. In addition to a moderate kappa value of 0.6, it is difficult to assess the effectiveness, justification, and safety of the Denis classification. In light of this, we recommend caution if it is used to guide the treatment of sacral fractures, to discuss prognosis with patients, and to describe fracture patterns with colleagues. Further studies investigating the validity and reliability of this classification system are warranted to better elucidate the use of this classification system.
We thank Dr. N. K. Kanakaris and colleagues from the Department of Orthopaedic surgery, University of Leeds, Leeds, West Yorkshire, UK, for the figure used in this study. We also thank Dr. Ishaq Syed and Dr. Alan Jones for their supervision, guidance, and insight with this publication. Reprinted by permission from Springer Nature Sacral Fractures by Nikolaos K. Kanakaris and Peter V. Giannoudis.
1. Beckmann NM, Chinapuvvula NR, Sacral fractures: classification and management. Emerg Radiol.
2. Bellabarba C, Schildhauer TA, Vaccaro AR, Chapman JR. Complications associated with surgical stabilization of high-grade sacral fracture dislocations with spino-pelvic instability. Spine (Phila Pa 1976). 2006;31:S80-S88; discussion S104.
3. Bonnin JG. Sacral fractures and cauda equina lesions. Med World. 1947;67:140-144.
4. Burstein AH. Fracture classification systems: do they work and are they useful? J Bone Joint Surg Am.
5. Bydon M, Fredrickson V, Garza-Ramos R, Li Y, Lehman RA, Trost GR, Gokaslan ZL. Sacral fractures. Neurosurg Focus. 2014;37:E12.
6. Denis F, Davis S, Comfort T. Sacral fractures: an important problem. Retrospective analysis of 236 cases. Clin Orthop Relat Res.
7. Huittinen VM, Lumbosacral nerve injury in fracture of the pelvis. A postmortem radiographic and patho-anatomical study. Acta Chir Scand Suppl. 1972;429:3-43.
8. Kanakaris NK, Giannoudis PV. Sacral Fractures. In: Lasanianos NG, Kanakaris NK, Giannoudis PV, eds. Trauma and Orthopaedic Classifications. 2015, Springer: London, England.
9. Lehman RA, Kang DG, Bellabarba C. A new classification for complex lumbosacral injuries. Spine J. 2012;12:612-628.
10. Medelman JP. Fractures of the sacrum: their incidence in the fractures of the pelvis. Am J Roentgenol Radium Ther Nucl Med.
11. Mehta S, Auerbach JD, Born CT, Chin KR. Sacral fractures. J Am Acad Orthop Surg. 2006;14:656-665.
12. Oner FC, Ramos LM, Simmermacher RK, Kingma PT, Diekoerhof CH, Dhert WJ, Verbout AJ. Classification of thoracic and lumbar spine fractures: problems of reproducibility. A study of 53 patients using CT and MRI. Eur Spine J.
13. Pascal-Moussellard H, Hirsch C, Bonaccorsi R. Osteosynthesis in sacral fracture and lumbosacral dislocation. Orthop Traumatol Surg Res. 2016;102:S45-57.
14. Robles LA. Transverse sacral fractures. Spine J. 2009;9:60-69.
15. Roy-Camille R, Saillant G, Gagna G, Mazel C. Transverse fracture of the upper sacrum. Suicidal jumper's fracture. Spine (Phila Pa 1976). 1985;10:838-845.
16. Schildhauer TA, Bellabarba C, Nork SE, Barei DP, Routt ML Jr, Chapman JR. Decompression and lumbopelvic fixation for sacral fracture-dislocations with spino-pelvic dissociation. J Orthop Trauma. 2006;20:447-457.
17. Schildhauer TA, Josten C, Muhr G. Triangular osteosynthesis of vertically unstable sacrum fractures: a new concept allowing early weight-bearing. J Orthop Trauma. 1998;12:307-314.
18. Vaccaro AR, Kim DH, Brodke DS, Harris M, Chapman JR, Schildhauer T, Routt ML, Sasso RC. Diagnosis and management of sacral spine fractures. Instr Course Lect.