Implants in Spinal Surgery Part III: Intervertebral Spacers

In our last post we discussed pedicle screws, the most commonly used form of posterior instrumentation.  Even in the recent past, pedicle screws were the only form of instrumentation used in a spinal fusion.  Unfortunately, these posterior-only fusion constructs didn’t have excellent rates of fusion: nearly 30% of these cases ended up in non-unions.  Today, most spinal fusion constructs utilize pedicle screws at the back of the spine and intervertebral spacers (or cages) in the front of the spine.  These so-called 360-degree constructs (also referred to as interbody fusions) have dramatically decreased the rates of non-union and thus are now the standard way spinal fusions are done.  

Of note, interbody fusion procedures are abbreviated with the suffix –IF. Common examples include the anterior lumbar interbody fusion (ALIF), posterior lumbar interbody fusion (PLIF) and extreme lateral interbody fusion (XLIF).  We’ll highlight the differences between the various –IFs in a future post.  The most commonly used interbody fusion procedure is actually the ACDF which we’ve already discussed here.

In the 1950s anterior interbody fusions were used in the treatment of Pott’s disease, or tuberculosis of the spine.  These anterior fusions proved to be more effective than posterior fusions because they allowed for reconstruction of the front of the spine, which is where most of the destruction of Pott’s disease occurs.  Once the infected vertebral bodies were resected (via a corpectomy), large defects in the front of the spine were reconstructed with pieces of bone fashioned into struts.  These pieces of bone were either harvested from cadavers (allograft), or may have been harvested from somewhere within the patient’s body (autograft) like the fibula or iliac crest.  These bone struts provided immediate support to the front of the spine and also promoted bony fusion.  They were the earliest intervertebral spacers.

While long structural bone struts are important in the reconstruction of the front of the spine after corpectomy for infection, tumor or trauma, the more commonly used intervertebral spacers are smaller spacers inserted into the disc space for fusion for degenerative conditions.  I will be referring to these spacers primarily for the remainder of the post.  The first such anterior interbody fusion utilizing bone inserted into the disc space was reported by Dr. Burns in 1933 for the treatment of lumbar spondylolisthesis (“slipped bones”, to discussed in a future post.)  Unfortunately, in this case, as with early anterior cervical fusions, often only loose shards of bone were packed into the disc space.  While this may have promoted bony fusion across the disc space it certainly didn’t provide any structural support and may have allowed for further collapse of the spinal segment with resultant kyphosis of the spine (see figure 1).  In the subsequent decades since Dr. Burns’ case, intervertebral spacers of various materials and shapes have been developed that are better at providing structural support and are also more readily available than spacers made of bone.  Today most intervertebral spacers are made of a fancy plastic called PEEK (polyetheretherketone) and this is what I use almost exclusively when I perform interbody fusions.  Titanium is also making a comeback (either comprising the entire spacer or as a coating on the surface of a PEEK spacer) as some feel it may be better at incorporating new bone in the early stages of bony fusion (see figure 2).  I could spend multiple posts discussing the merits of different intervertebral spacer materials but that you would likely bore you (and me) to death.  Regardless of what it’s made of, every intervertebral spacer should serve two purposes: a) correct spinal deformity, and b) promote a bony fusion.   

Postoperative Deformity of the Cervical Spine Clinical Gate 

Figure 1: Anterior cervical fusion done with only bone graft and no spacer for structural support.  Note that the fused bones have collapsed somewhat and are now allowing the spine the fall forward  The red line indicates the forward angle of the collapse and should, in fact, be straight (source: Riew et al, 2015).

Harms cage Google Search

Figure 2: Various intervertebral spacers used for lumbar interbody fusion (source: Williams et al, 2005). 

First, let’s talk about the correction of spinal deformity.  In my opinion spinal reconstruction should take place where the degenerative process starts: the disc space.  As we’ve talked about previously, as the intervertebral disc (IVD) degenerates it dries out and collapses.  Without the structural integrity of the IVD the adjacent vertebral bodies fall out of alignment and the spine then becomes deformed.  Spinal deformity is a complicated topic that I won’t delve too far into at this point.  All you need to know now is that a deformed spine causes both back pain and radicular (nerve) pain.  The back pain is caused by the mechanical stress of the deformity.  The radicular pain is caused by foraminal stenosis and the resulting nerve root compression.  Recall that the neural foramen, the hole on the side of the spine where the nerve exits, is an aperture comprised of the top of the pedicle of the vertebral body below and the bottom of the pedicle of the vertebral body above.  As the IVD collapses the diameter of the foramen narrows significantly and the nerve is guillotined within.  An intervertebral spacer inserted into a collapsed disc space corrects deformity by acting like a wedge to a) restore alignment to relieve mechanical stress, and b) restore foraminal height to allow for indirect decompression of compressed nerves.  The concept of indirect decompression is an immensely important one that I’ll get to in an upcoming post.  (See figures 3 and 4.)


Figure 3: Image on left shows collapsed disc with resultant foraminal stenosis and nerve root compression.  Image on right shows the result after a spacer is inserted with restored foraminal height and indirect nerve decompression.   

Coronal deformityNewImage

Figure 4: Preoperative MRI showing a focal coronal deformity with unilateral foraminal stenosis and nerve root compression resulting in severe R leg pain (the red lines, indicating the vertebral endplates, should be parallel and instead are collapsed to one side.)  Image on right is a postoperative anterior view Xray after a spacer (with plate) is inserted.  Notice that the vertebral endplates (red lines) are now parallel after the deformity has been corrected.  Of note, because it is PEEK the spacer isn’t seen well on Xray and instead the borders of the spacer are indicated by markers within.  

The other important function of the intervertebral spacer is to promote bony fusion.  Recall from previous posts that it’s the bone graft material that is applied between parts of the spine that promotes the new bone growth in a spinal fusion.  There are many types of graft material used in spine surgery that I’ll discuss in a later post.  Most modern intervertebral spacers have large chambers within their structure where bone graft is packed prior to insertion into the disc space (see figure 5).  These chambers then hold the graft material in place, nicely apposed against the adjacent vertebral bodies, in order to promote bony ingrowth (see figure 6).  

 ALIF spacer bone graft Google Search

Figure 5: standard ALIF spacer used in lumbar fusions.  Note the large graft chambers (red arrows) that are packed with bone graft material prior to insertion into the disc space. Source: Globus Medical. 


AH L4 5 spondy XLIF pptx

Figure 6: Lateral postoperative CAT scan demonstrating robust bone growth (red arrow) across intervertebral spacer (spacer material isn’t seen in this cut of the scan.)  

Lastly, I believe that bigger is better when selecting an intervertebral spacer in an interbody fusion.   Not only do larger spacers provide more structural support for deformity correction but they can also carry more graft material to better promote bony fusion.  THIS is why I prefer anterior procedures such as XLIF or ALIF with their large spacers rather than TLIF or PLIF with their quite puny spacers.  More on that later.

Thanks for reading!

J. Alex Thomas, M.D.


1) Riew et al, 2015, Postoperative Deformity of the Cervical Spine, online access at iKnowledge:

2) Williams et al, 2005, CT Evaluation of Lumbar Interbody Fusion: Current Concepts, ANJR, 20: 2057-2066, September 2005.









Implants in Spinal Surgery Part II: pedicle screws and other (very rarely used) forms of posterior spinal instrumentation

In our last post we discussed how various types of spinal implants help stabilize the spine to promote a more robust bony fusion.  Recall that the main goal of a spinal fusion procedure is to promote bone growth which in turn will stabilize a painful unstable or deformed segment of the spine.  Historically, non-instrumented in-situ fusions had very high rates of non-union in which the spine didn’t fuse despite weeks of bedrest and bracing.  This may have required subsequent revision surgeries and often left the patient severely disabled.

In order to mitigate the problem of non-union, surgeons developed forms of posterior instrumentation (instrumentation applied to the back of the spine) to help stabilize the spine.  These implants also provided surgeons a more powerful way to restore alignment to the spine to correct spinal deformity.  The most primitive form of posterior spinal instrumentation, first described in the late 19th century, was posterior spinal wiring.  Here, various parts of the spine (usually the spinous process or lamina) were wired together for immobilization during spinal fusion.  While useful in the cervical spine (see figure 1), this technique never proved to be effective in the thoracic and lumbar spine, where larger forces prevail, and thus has largely been abandoned today. 


Figure 1: Cahill technique of posterior spinous wiring in the cervical spine.  Source: Omeis et al

In the 1950s Paul Harrington began performing spinal fusions using long rods anchored to the spine with hooks under the lamina (see figure 2).  The rod construct was periodically lengthened to slowly straighten the deformed spines of scoliosis patients. Harrington rod systems were state of the art for decades and I’ll still see patients in clinic today with these long rods in their spine.  Unfortunately, with only a few points of fixation anchoring the long, straight rod in place, these constructs were prone to failure and also predisposed patients to a painful “flat back” deformity.  (When a surgeon fuses a spinal segment without being mindful of its normal degree of curvature he may inadvertently cause more harm by creating spinal deformity.  More on this very important topic in later posts.)  Moving away from Harrington rods, surgeons began to develop new forms of segmental instrumentation in which each segment of the fused section of the spine was anchored with instrumentation (versus the long Harrington rod which spanned multiple non-instrumented segments.)  This segmental instrumentation vastly increased the strength and stability of the fusion construct and therefore further decreased the rate of non-union.  Early forms of segmental instrumentation include variations of trans-facet screws described by King in 1944 and Boucher in 1959 (the latter incorporated part of the pedicle in the screw trajectory and thus is considered by some to be the first pedicle screw.)  In 1970 Roy Camille described the precursor to today’s pedicle screw systems.  In his system he attached screws inserted via the pedicle to a multi-holed plate which would span across two spinal segments from screw to screw.  The screw-plate concept was expanded upon in the U.S. by Dr. Arthur Steffee in the early 1980s with his Steffee Plate/VSP stainless steel pedicle screw system (see figure 3).  Steffee’s company AcroMed was the subject of a 1993 ABC 20/20 expose which profiled several patients who were left disabled after Steffee’s VSP pedicle screws broke (all pedicle screws will eventually break, by the way, if the spinal segment doesn’t fuse properly.)  This prompted hundreds of pedicle screw-related lawsuits around the country in the mid 1990s (part of the issue was that the FDA never cleared the VSP screw for use specifically in the spine).  Ultimately most of the pedicle screw litigation was thrown out.  Today pedicle screws are a mainstay of treatment of a variety of thoracic and lumbar spinal pathology. 


Figure 2: AP (A) and lateral (B) postoperative Xrays demonstrating a single Harrington rod used in correction of a thoracic scoliosis.  Note in image B how flat the fused segment is. 

 Mendeley Desktop

Figure 3: Steffee plate and VSP pedicle screw system in L4-S1 fusion.  Source: Kabins et al.

Today’s pedicle screws are generally made of titanium and have polyaxial heads in which a rod is seated and locked in place using a locking cap.  The rod is used in favor of Camille’s and Steffee’s plates as it allows for easier insertion as well as contouring to correct spinal deformity (see figure 4). Screws vary in size depending on the size of the patients and the part of the spine being instrumented (i.e. smaller screws in the thoracic spine and larger screws for the lumbar spine.)  Typically in lumbar spine fusions I will use screws that are 6.5mm in diameter and 45mm long with a 4.5mm diameter rod.   Once the screws are inserted various attachments can be used to manipulate the screws to rotate, compress or distract across spinal segments in order to correct spinal deformity prior to fusion. 


Figure 4: Modern pedicle screw with polyaxial head and rod locked in place. Source: Zimmer Biomet.

As you can probably guess, a “pedicle” screw traverses part of the vertebral body called the pedicle.  This bridge of bone connects the anterior elements of the spine (i.e. the vertebral body) with the posterior elements (i.e. lamina, facets, spinous process, see figure 5).  In order to properly insert a screw into the pedicle the surgeon must access a starting point at the junction of the transvers process and the facet complex, several centimeters off of midline.  In traditional open spinal surgery via midline incisions the surgeon has to do quite a bit of destructive, bloody dissection to get a wide enough exposure to access this starting point of the pedicle.  Often, in order to gain enough laxity in the tissue to get out to the starting point, the surgeon must also expose the level above and below the level that is being fused.  This “collateral damage” of healthy levels in open spinal surgery is what sets patients up for adjacent segement degeneration and other problems later in life. 

Lumbar vertebra jpg 648×778 pixelsPasted Image 3 30 17 1 22 PM

Figure 5: Axial (left) and lateral (right) views of lumbar spine demonstrating pedicle (outlined in red) bridging the anterior and posterior elements.

Pedicle Screw Incisions

Figure 6: Small stab incisions used to placed 4 percutaneous pedicle screws for an L4/5 spinal fusion.  The red line indicates the length of incision that would have been needed to place the same number of screws using traditional open techniques. 

In order to avoid the increased blood loss and tissue destruction of open spinal surgery, I insert pedicle screws percutaneously with tiny, minimally-invasive incisions off midline (see figure 6.) I’ll first identify the starting point of the pedicle using a fluoroscope (like an xray machine) and then will hammer in a large needle into the pedicle via the starting point.  I can identify the correct starting point not only using imaging but also with the tactile feedback of the hard bone of the pedicle starting point.  Once I’m happy with my position in the pedicle I’ll insert a long, rigid wire called a K-wire into the pedicle and will remove the needle.  The pedicle is tapped to make a pilot hole and the screw is then inserted over the wire (see figure 7.)  The nerve that exits the spine passes just below the pedicle so one potential complication of pedicle screw insertion is nerve injury resulting from improper positioning of the screw within the pedicle. In order to avoid this complication I always use electromyographic (EMG) monitoring in which the insertion needle and tap are stimulated with low-voltage electrical current.  If the screw trajectory is too close to a nerve the current will stimulate the nerve, the corresponding muscles in the leg will twitch and I’ll be alerted to the problem so that I can plan a new trajectory.  After the screws are inserted a rod is passed through the heads of the screw and locked in place.  While some surgeons will try to get away with placing only unilateral screws, I always place bilateral pedicle screws (i.e. on both sides of the spine), as this is the gold standard for maximum spinal stabilization (see figure 8.)  Finally, recall that the screws just serve as an internal brace to allow the bony fusion to occur.  Theoretically we could remove the screws in a year after the fusion has healed but we almost never do this.

 IMG 6277

Figure 7: 4 K-wires placed prior to placement of percutaneous pedicle screws for lumbar fusion (red wire).  Note that wires and thus screws can be placed with the patient in the lateral position (in this case patient’s right side is up.) 

Session Working Studies AccessionNumber CT1607469 17 WindowNewImage

Figure 8: Axial CT image (left) and schematic image (right) demonstrating bilateral pedicle screws traversing the lumbar pedicles. Source for schematic: DePuy Synthes

There are other types of posterior instrumentation that are used in the lumbar spine.  These include interspinous clamps as well as facet screws.  I don’t believe that these implants provide the same amount of stability as a bilateral pedicle screw construct and therefore I never use them.  Not discussed here are specialized screws used in posterior cervical spine fusions called lateral mass screws (named for the part of the cervical vertebral body that they enter.)  These are very similar to thoracic and lumbar pedicle screws (titanium, polyaxial heads, connected by rods passed through the heads) but are much smaller with a typical diameter of 3.5mm and a length of 12-14mm.  

In summary, pedicle screws act as an internal brace to immobilize the spine so that a more robust bony fusion may occur. These screws can be safely inserted into the spine using minimally-invasive percutaneous techniques.

Thanks for reading!

J. Alex Thomas, M.D.


  1. Hasler CC: A brief overview of 100 years of history of surgical treatment for adolescent idiopathic scoliosis. J Child Orthop 7:57–62, 2013.
  2. Omeis I, DeMattia J a, Hillard VH, Murali R, Das K: History of instrumentation for stabilization of the subaxial cervical spine. Neurosurg Focus 16:E10, 2004.
  3. Kabins MB, Weinstein JN: The History of Vertebral Screw and Pedicle Screw Fixation. Iowa Orthop J 11:127–136, 1991.

Implants in Spinal Fusion, Part I: In-situ Fusions Rarely Fused

As I was writing the previous few posts I realized that I was relying on terms such as pedicle screw and intervertebral spacer to begin the explain techniques used to achieve spinal fusion.  Before we get further into our discussion on these techniques I think it would benefit you, devoted Spinal (con)Fusion reader, if I spent a few posts discussing the various implants used during spinal fusion procedures.   It seems like in my clinic everyone knows someone who didn’t do well after getting “a whole bunch of screws and rods in their back”.  Granted, on their own these sound like medieval torture devices that no sane person would want implanted into their spine.  Hopefully by shining some light on the screws, rods and various other spinal implants used during fusion procedures, I can put prospective patients’ minds at ease if they’re considering a spinal fusion.

In order to appreciate the benefits of today’s spinal instrumentation, you must first understand how terribly inadequate early non-instrumented spinal fusions were.  Recall that we discussed that the main goal of a spinal fusion procedure is to promote bone growth across one or more spinal motion segments.  This bone growth immobilizes what is felt to be an unstable, and thus painful, part of the spine.   While spinal fusions have been done since the early 20th century, the only strategy that early spine surgeons could employ to achieve a bony fusion was to harvest autograft (bone harvested from a site within the patient such as the spinal lamina or the iliac crest) and lay it down over an exposed part of the spine that they wished to fuse.  This primitive in-situ fusion technique, first described by Albee and Hibbs in the early 1900s, was problematic for two reasons.  First, there was no good way to correct spinal deformity while promoting a bony fusion.  Thus, after a long, morbid surgery, patients were often fused with painfully deformed spines and no better than they were prior to surgery.  Second, in order for any bone to fuse together the adjacent pieces of bone have to be immobilized (think of a cast on a broken arm.)  In an in-situ fusion, with the bone graft simply laying on top of a segment of spine, the only way to immobilize a patient’s spine to promote bone growth was to keep them on bedrest, often in full body braces or casts, FOR MONTHS.  UGH!  What’s worse is that because of these inadequate forms of immobilization, in half the cases the new bone wouldn’t grow, the spinal segment wouldn’t fuse and the patient would be left with a painful condition called a non-union, or failed fusion, often requiring subsequent revision surgeries.  It’s no surprise, then, why many at the time considered early spinal fusion procedures to be painfully ineffective.     

Beginning in the mid-20th century, various forms of spinal instrumentation were developed in order to help mitigate the above limitations of early in-situ fusions.  First, spinal implants provide the necessary internal bracing that immobilizes the diseased motion segment so that robust bone growth can occur.  No more full-body casts!   Also, spinal implants, particularly the intervertebral spacers inserted into the disc space at the front of the spine, allow for correction of spinal deformity.  This deformity correction, equally important as correction of instability, restores the spine to its normal form and alignment prior to it being permanently immobilized by the new bone growth of a spinal fusion.  In short, spinal implants create the optimal conditions for new bone to grow to achieve a spinal fusion and thus correct painful spinal instability and deformity. 

In our next post I’ll dive right into the world of spinal implants with a discussion on pedicle screws and other forms of posterior instrumentation. 

Thanks for reading!

J. Alex Thomas, M.D.


True spinal instability is a clear indication for spinal fusion

As we illustrated in our last post there is a wide spectrum of indications for lumbar spinal fusion.   As you move along this spectrum from unstable to more stable pathology the odds of a successful outcome decrease.  At the far end of the spectrum of diagnoses, the end at which there is a lesser chance of a favorable outcome after fusion, is degenerative disc disease (DDD) and spondylosis (without instability) causing back pain.  In my opinion this is softest indication for spinal fusion.  I’m not saying that you should never perform a spinal fusion on a patient with only DDD, the patient just has to be properly vetted and they must understand that a good outcome isn’t guaranteed in these cases.  On the opposite end of the spectrum is acute spinal instability caused by trauma or some other acutely destructive process such as tumor or infection.  This is the clearest indication for a spinal fusion.  NOTE: we’ve already discussed cervical spinal fusion (ACDF) here and here so this discussion will pertain primarily to the lumbar spine.

Classically, spinal stability is defined as the spine’s ability, under normal physiological loads (“normal” obviously varies widely depending on whether you’re a bank clerk or a mixed martial arts fighter), to a) protect the neural elements (i.e. nerve roots and spinal cord), and b) avoid painful deformity.  Sounds complicated right?  It may be easier to think about what happens when the spine becomes unstable: a) it may not be able to maintain proper alignment and thus may become deformed which causes severe pain; and b) it may not be able to properly protect the spinal cord within which could cause paralysis.  So in a nutshell: a stable spine is one that is protecting you against pain and/or paralysis. 

The concept of traumatic spinal fractures is a vast one that I won’t get into too much here.  Generally, though, fractures are classified as stable or unstable (hopefully you’re starting to pick up on a theme here.)  There are many complicated grading schemes that allow spine surgeons to look at a fracture on imaging and determine if it’s unstable or not.  One classic scheme is that of Denis which divides the spine into three columns.  Stable fractures typically only involve one column of the spine. Examples of stable fractures include fractures of the spinous process (a so-called clay shoveler’s fracture, see Figure 2), compression fractures and transverse process fractures.  Stable fractures may be painful from the local trauma of the injury but they do not cause painful deformity nor do they threaten the spinal cord or nerve roots.  Thus these types of fractures may be treated conservatively such as with bracing. 

Mendeley Desktop

Figure 1: Illustration from Denis’ 1983 paper discussing his three spinal columns and their involvement in traumatic injuries.

Clay shoveler fracture Image Radiopaedia org

Figure 2: Fracture of the C6 spinous process (clay shoveler’s fracture).  Source:

Generally speaking if two or more of Denis’ columns are involved in a fracture it is considered unstable (again let me reiterate that Denis’ model is quite simplistic and analyzing a fracture isn’t always as easy as looking at the spine in only 3 columns.)  When a fracture is determined to be unstable a spinal fusion may be indicated to restore stability.  If an unstable fracture is left to heal without surgery it may heal poorly resulting in a painful deformity. Worse, if a patient with an unstable fracture is allowed to get up out of bed and loads their spine the fracture may shift resulting in injury to the spinal cord and paralysis. 

Trauma isn’t the only cause of acute spinal instability.  Indeed, aggressive tumors or infections can destroy the integrity of the spine thereby causing painful spinal deformity and perhaps paralysis.  These lesions are treated in a similar manner as acute fractures depending on which part of the spinal column has been damaged. The case presentation below describes a case I had a few years ago of an elderly gentleman with severe damage to his spine caused by a staph infection. 

Generally speaking when deciding which type of spinal fusion to perform for acute spinal instability, I’ll go to where the problem is:  if the pathology primarily involves the vertebral body in front of the spine, for example, I’ll do a corpectomy to remove the fractured vertebral body.  Once the body is removed I’ll reconstruct the spine with a spacer inserted where the damaged vertebral body was, and a combination of plating or screws to provide extra stability (we’ll talk about these devices in more detail in future posts.)  The main goal of all of that surgery is to promote new bone growth across the damaged segment of the spine.  It’s this new bone growth that restores spinal stability.  


The patient is a 75yo male with methicillin-resistant staph aureus (MRSA) bacteremia (in his bloodstream) who presents with worsening mid-back pain.  Imaging reveals T11-12 discitis.  (Discitis is an infection of the intervertebral disc space that is probably the most painful condition that I see.  You can usually make the diagnosis by very gently bumping the patient’s bed when you approach the bedside; if the patient screams out in pain it’s probably discitis.  That’s how bad it is.)  The medicine doctors tried a long course of antibiotics but unfortunately his pain didn’t improve.  Repeat imaging revealed that the infection hadn’t been cleared and in fact had caused further destruction of the T11 and T12 vertebral bodies (see Figure 3.) This destruction resulted in spinal instability and kyphosis (a painful deformity in which the spine falls forward.)  

T11 12 discitis

Figure 3: CT scan illustrating T11-12 discitis resulting in severe bony destruction (red arrow) and resultant kyphotic deformity (blue arrow indicates top of spine falling forward). 

When I met this patient he looked like he had given up and wanted to die.  He’d been bedbound from his infection for weeks and now was quite debilitated.  He agreed to undergo surgery and underwent a T11 and T12 corpectomy (via a lateral approach through the chest and behind the lung) followed by reconstruction of the spine with an expandable cage and percutaneous pedicle screws (see Figure 4.)  By one month post-op the patient reported no pain and was walking without assistance.  The last time I saw him about a year after his surgery he was living a normal life at home with his family.  He looked like he’d been given a new chance at life. 

Post op T11 12 corpectomy

Figure 4: Postoperative AP (left) and lateral (right) X-rays with expandable corpectomy spacer at T11-12 (red arrow) and percutaneous pedicle screws from T9-L2 (blue arrows).

I think I’ll spend the next post or two talking about the various forms of spinal implants that we use to achieve a spinal fusion. I had planned to do this later but I think that by presenting it first it will help you better understand the various spinal fusion procedures discussed in later posts. 

 Thanks for reading!

 J. Alex Thomas, M.D.


Denis F: The three column spine and its significance in the classification of acute thoracolumbar spinal injuries. Spine (Phila Pa 1976) 8:817–31, 1983.

What is a spinal fusion?

Of all the procedures that I perform, the spinal fusion is one of the most misunderstood and maligned.  Indeed, the name of this blog is a nod to the confusion surrounding the procedure.  Before we get to the specific conditions that necessitate spinal fusions and the techniques used to achieve a fusion let’s talk about what a spinal fusion is in general terms.  Throughout the article I may refer to spinal fusion by its preferred medical name, spinal arthrodesis.  The discussion will generally pertain to fusions of the lumbar spine (if you want to read about cervical spine fusions you can do so here and here.) 

Whenever the spine becomes deformed or unstable it becomes painful.  It’s not as easy as you may think to pinpoint the cause of spinal pain as often it’s multifactorial.  Generally, though, the pain is caused by either compression of a nerve by a deformed spine (causing a radiculopathy) or by the structural stress of instability or deformity (causing neck or back pain.)  The goal of a spinal arthrodesis is to promote bone growth between one or more vertebral bodies in order to correct spinal deformity and instability and thus relieve pain.  Classically this bone growth is achieved between the posterior elements or transverse processes (posterolateral arthrodesis), between the vertebral bodies within the disc space (anterior/interbody arthrodesis) or a combination of both (see figure 1).  

Posterolateral fusionInterbody fusion

Figure 1: Left, robust posterolateral fusion (arrows) with pedicle screw fixation.  Right, interbody fusion with robust bone growth between two vertebral bodies (arrow). (Source:

Today, advanced forms of spinal stabilization are used to stabilize the spine to allow the arthrodesis to occur more robustly and more rapidly (you can think of spinal stabilization as internal bracing.)   In the early 1900’s, when spinal fusion was first described to treat the destructive Pott’s Disease, or spinal tuberculosis, these technologies weren’t available. In these early procedures surgeons attempted to achieve a spinal arthrodesis by simply laying down harvested bone graft over the posterior aspect of the spine (either from the lamina, the iliac crest or from the fibula in the leg) and hoping that bone would eventually grow into a robust posterolateral arthrodesis.  This typically required a long, arduous surgery followed by 6 months of bedrest in a body cast.  Just imagine that for a moment.  Patients who were lucky enough to achieve a solid bony fusion may have had a permanently deformed (and painful) spine, as methods to adequately correct spinal deformity hadn’t been developed yet.  Over time surgeons realized that the simply fusing a patient in-situ, or in its original place, wasn’t adequate and that correction of spinal deformity at the time of arthrodesis was paramount.   In the 100 years since the early spinal arthrodeses for Pott’s Disease, various metallic implants, intervertebral spacers and grafting materials have been developed to dramatically improve the outcomes of spinal arthrodesis.  Today, minimally-invasive spinal arthrodesis can be performed on an outpatient basis with a near 100% success rate (see figure 2). We’ll get into the specifics of implant and graft technologies in future posts.

Post op XLIF CT

Figure 2: Postoperative Xray showing successful minimally-invasive spinal fusion (XLIF).  A large intervertebral spacer is used to restore foraminal height (blue arrow) while percutaneous pedicle screws (red arrow) are used to stabilize the posterior aspect of the spine. 

Despite recent technological advances of spinal arthrodesis, some patients are still downright terrified when I recommend the procedure.  Everyone knows someone who “was never the same” after having a spinal fusion.  Unfortunately these fears aren’t entirely unfounded as some spinal fusions are still being done on the wrong patients for the wrong reasons.  As illustrated in table 1 the odds of success (generally defined as relief of pain and disability) of spinal arthrodesis vary based on the indication for surgery.  The clearest indication for spinal arthrodesis is acute instability or deformity caused by trauma, infection or tumor.   In these instances spinal arthrodesis is almost always associated with an excellent outcome.  In contrast, the murky, poorly defined indications of degenerative disc disease (DDD) or spondylitic back pain (back pain caused by arthritis) often are NOT good indications for spinal surgery.  The problem in these cases is that back pain is notoriously cryptic and thus it can be difficult to correlate a patient’s pain with a certain structural abnormality on an MRI or X-ray.  The surgeon then has to make an educated guess as to what is generating the pain and then target it with a spinal fusion.   It’s no surprise when this effort is unsuccessful.  I typically will NOT perform spinal fusion on patients with only DDD or spondylosis and back pain (i.e. in the absence of gross instability or deformity.)  There are exceptions to this, of course.  For example, in cases of severe DDD with disc space collapse and resultant foraminal stenosis and radiculopathy (as opposed to just back pain) an interbody fusion may be needed to restore foraminal height and indirect decompression of a compressed nerve (see figure 3).  While this isn’t a case of frank instability I consider the disc space collapse with foraminal stenosis a deformity that requires correction. We’ll talk about this and other indications for spinal arthrodesis in future posts.    

Screenshot 11 27 16 11 11 AM

Table 1: Odds of success (defined as reduction of pain and disability) of spinal fusion is dependent on the indication for surgery. 

XLIF jpg

Figure 3a: severely collapsed disc space results in foraminal stenosis and compression of exiting nerve root.  3b, interbody fusion with large intervertebral spacer results in restore foraminal height and indirect decompression of nerve root.  

I realize that this post may have generated more questions than answers.  Stay tuned for future posts about indications for spinal arthrodesis as well as advanced techniques used to achieve successful arthrodesis.  I hope to prove to you that in the properly selected patient the properly executed spinal fusion can provide life-altering improvement in a patient’s quality of life. 

Thanks for reading!

J. Alex Thomas, M.D.

I specialize in “useless” surgery.

On August 3, 2016 the New York Times published an essay called “Why ‘Useless’ Surgery Is Still Popular”.   In the essay the author decries the continued use of medical procedures “despite clinical trials that cast doubt on their effectiveness.”  One of the procedures discussed in the essay, the spinal fusion, is a procedure that I routinely perform on my patients, almost uniformly with great success.  Unfortunately, this essay irresponsibly cites only one review article about spinal fusions and thus unfairly describes the procedure as ineffective and “useless”.  On a professional level I am disappointed in this essay because I think it is misleading to the public and may prevent delivery of a potentially very effective therapy.  Not only may a patient be scared away from getting a spinal fusion after reading the essay, insurers are starting to take notice and have pounced on the opportunity to not have to pay for their patients to have fusions, deeming them “medically unnecessary”.   On a deeper, more personal level, articles like this really burn me up (and I really have to bite my tongue to remain professional there.)  I work tirelessly to provide the best possible care for my patients and spinal fusions comprise a large part of my practice.  You can imagine how I feel when articles like this in the mass media attempt to discredit what I do to help so many people.

Like any surgical procedure, the key to a desirable outcome is to only perform spinal fusions on patients with the proper indications for the procedure.  The “useless” author cites a review article by Mirza et al in 2007 that compiled data from 4 randomized trials of lumbar spinal fusion for discogenic back pain.  These trials found that spinal fusions were no better than physical and cognitive therapy for treating chronic low back pain.  The issue here, as it usually is whenever a surgery fails, is the poor indication for surgery.  First of all, any reputable spine surgeon knows that you should never offer surgery to a patient with only back pain.  There must be a corresponding structural cause of the patient’s pain that is amenable to surgery and the patient’s physical exam findings must correlate with these structural problems.  A degenerating disc causing so-called discogenic pain is NOT a structural cause of back pain!!  We’re not even certain that a degenerating intervertebral disc (IVD) can cause pain.  The thought is that by removing the degenerated and thus painful disc and fusing the adjacent vertebral bodies you will relieve the patient’s pain.  Unfortunately, so called “black disc surgery” (because the discs get darker as they degenerate) usually doesn’t work.  In my opinion this is the softest indication for spinal fusion and in fact most insurers won’t even approve the procedure for this indication. The vast majority of patients who present to my clinic with chronic discogenic back pain are sent right back out for pain management, physical therapy or other forms of conservative management.  

Spinal fusions are clearly effective in correcting structural problems of the spine such as spondylolisthesis and degenerative scoliosis.  That’s not just my anecdotal belief; multiple clinical studies have proven so.  For example, in the landmark randomized, controlled SPORT study published in the New England Journal of Medicine in 2007, Weinstein et al looked at spinal fusion versus nonsurgical treatments (i.e. physical therapy, epidural steroid injections, etc.) for the treatement of spondylolisthesis (a painful condition where one vertebral body slips over the one below it.)  The study demonstrated clear superiority of spinal fusion over nonsurgical treatments (see figure 1.)  The benefits of spinal fusion have been found to persist out to at least 8 years in subsequent analyses.  Patients who underwent nonsurgical treatment also got better, just not as rapidly or to the same extent as patients who underwent spinal fusion.  Finally, it’s important to note that the benefits of spinal fusion in the SPORT study were seen for fusion techniques that in my opinion are a bit archaic in the age of advanced minimally-invasive techniques.   Of course, the “useless” author didn’t discuss seminal studies such as SPORT in her essay.   


Figure 1: A successful minimally-invasive spinal fusion done at L4/5 for spondylolisthesis. 

Over the next several posts we’ll discuss the indications for spinal fusion as well as the various techniques used to achieve a spinal fusion.  Hopefully you’ll learn what I already know: that spinal fusions, when done for proper indications, can dramatically improve a patient’s function and quality of life.   

Thanks for reading!

J. Alex Thomas, M.D.


Weinstein, J. N., Lurie, J. D., Tosteson, T. D., Hanscom, B., Tosteson, A. N. a, Blood, E. a, … Hu, S. S. (2007). Surgical versus nonsurgical treatment for lumbar degenerative spondylolisthesis. The SPORT authors. The New England Journal of Medicine, 356(22), 2257–70. 

What makes a far lateral disc herniation unique (and PAINFUL)?

We made an important distinction in the last post: that of the far lateral disc herniation.  We’ve just discussed the more common central herniated nucleus pulposus (HNP) in which the disc herniates into the center of the spinal canal (see figure 3 of this post.)  This centrally herniated fragment hits the traversing nerve that is still within the spinal canal (e.g. a central L4/5 disc herniation causes a radiculopathy of the L5 nerve.)  In a far lateral HNP, occurring only in about 10% of cases, the piece of disc herniates on the side of the spine and compresses the nerve along its course within the neural foramen as it exits the spine (e.g. a far lateral L4/5 disc causes a radiculopathy of the L4 nerve.)   See figure 1 for an MRI showing a far lateral HNP. 

Far lateral HNP 2

Figure 1: Axial MRI showing large right far lateral HNP (outlined in pink.)  Note the displaced nerve root (pink arrow) as compared to the normal nerve root free in its neural foramen (green arrow). 

Often I can identify a patient with a far lateral HNP right when I enter the room because they’re MISERABLE.  The pain associated with far lateral HNPs is typically much worse than that seen with central HNPs.  Just as it exits from within the neural foramen, the nerve dilates into an important junction point called the dorsal root ganglion (DRG).  It’s this exquisitely sensitive part of the nerve that is compressed by a far lateral HNP.  Couple that with the fact there’s a very limited amount of space within the bony neural foramen for both the herniated disc fragment and the DRG and one understands why this type of disc herniation is so debilitating (see figure 2).

DRG in foramen 

Figure 2: Image of the right side of the lumbar spine showing the nerve roots exiting via the bony neural foramina.  Note the dilation of the DRG within the confines of the foramen (red arrow.)  

A far lateral HNP requires a different approach than the standard discectomy we discussed in the last post.  In the far lateral discectomy, I’ll typically employ an “outside-in” approach to find the fragment under the nerve as it exits the foramen.  First, the incision for a far lateral discectomy is made a few more centimeters off of midline compared to that of a standard discectomy.  Next, I’ll dock a tubular retractor in between the transverse processes at the level in question (see figure 3).  I’ll then work my way into the foramen and look for exiting nerve within the soft tissue of the intertransverse space.  One benefit of using an outside-in approach is that usually I don’t have to drill away any of the facet joint and avoid potentially destabilizing the spine.  Once I’ve found the nerve (the DRG is usually what I see first) I move it out of the way and the piece of herniated typically found right underneath.  In order to mitigate some of the pain caused by my manipulation of the DRG I will apply some steroids to the nerve when I’m done removing the disc.

Discectomy docking points

Figure 3: Image depicting the docking points for discectomy in relation to bony anatomy at the right L4/5 level. The blue circle illustrates the docking point for the tubular retractor in a standard central discectomy.  The green circle illustrates the docking point for a far lateral discectomy. 

Recovery from far lateral discectomies is typically rougher than after standard discectomies.  The DRG is already inflamed and manipulating it to get to the herniated fragment can often make the patient’s pain and numbness worse before it gets better.  Thus, I always have to get my patients with far lateral HNP mentally prepared for a tough couple of weeks after their discectomy.  In the end, though, patients do very well after a far lateral discectomy. 

Thanks for reading!

J. Alex Thomas, M.D.