L2 and L3 are the second and third vertebrae of the lumbar spine, sitting in the upper-to-middle portion of the lower back. They occupy the region roughly behind the navel and just above the waistline, bridging the transition from the relatively stiff thoracolumbar junction above to the more mobile lower lumbar segments below. Though L4 and L5 get most of the attention in back-pain discussions, L2 and L3 play a quietly critical role: the spinal cord typically ends just above them, major nerve roots exit through their openings, and the forward curve of the lower back often reaches its deepest point near their level.
Counting Down from the Top
The lumbar spine consists of five vertebrae stacked between the thoracic spine (mid-back) and the sacrum (the fused bone at the base of the spine). L1 is the highest, sitting just below the last rib-bearing thoracic vertebra, and L5 is the lowest, resting on the sacrum. L2 and L3 fall in between, forming the upper-middle segment of the lumbar column. They are slightly smaller than L4 and L5 but larger than L1, following a general pattern in which lumbar vertebral bodies grow progressively wider and thicker as they descend to bear increasing body weight.
In terms of the overall spinal curve, L2 and L3 sit near the peak of the lumbar lordosis, the inward curve of the lower back that helps keep the torso balanced over the pelvis. In a large study of nearly 470 people without back pain, the apex of this curve fell at L3 in about 40% of subjects and at L4 in roughly 57%, with a small percentage peaking as high as L2.1Spine. Variation in Lumbar Shape and Lordosis in a Large Asymptomatic Population That places L2 and L3 squarely in the zone where the lower back bends most prominently forward. The lower lumbar segments, especially L4 through S1, contribute the lion’s share of total lordosis angle, while L2 and L3 contribute a smaller portion of the curve but serve as the transitional region between the relatively straight upper lumbar spine and the sharply curved lower lumbar spine.2PubMed Central. Current strategies for the restoration of adequate lordosis during lumbar fusion
Finding L2 and L3 from the Outside
If you place your hands on your hips and press your thumbs into the top of the bony ridges you feel on each side, you are touching the iliac crests, the upper rim of the pelvis. Textbooks have long taught that an imaginary line drawn between the two iliac crests crosses the spine at approximately the L4 vertebral body or the L4–L5 disc space. This “intercristal line” is the standard surface landmark used when positioning spinal needles or choosing injection sites.
The reality is less tidy. An imaging study of 75 patients found that while the intercristal line does correspond to L4 or L4–L5 on X-ray or MRI in about 87% of people, palpating that same line with your fingers tends to identify a higher level. In that study, the palpated line landed at L3 or L3–L4 in roughly 77% of cases, and the mismatch was more pronounced in women and in people with higher body mass.3PubMed Central. Which spinal levels are identified by palpation of the iliac crests and the posterior superior iliac spines? The palpated level was never lower than the imaged level; it was always the same or higher, sometimes by as many as four segments. In practical terms, if someone presses your hips and says “this is L4,” they may well be touching L3 or even L2. The upshot is that L2 and L3 are closer to the waistline landmark than most people assume, and clinicians are increasingly aware that palpation alone can be unreliable.
What Lies Inside the Spinal Canal at L2 and L3
One of the most clinically significant facts about L2 is its relationship to the end of the spinal cord. The spinal cord does not extend all the way down to the sacrum. Instead, it tapers into a cone-shaped tip called the conus medullaris, which in most adults terminates around the lower third of L1.4PubMed. Magnetic Resonance Imaging-Based Anatomy of the Conus Medullaris: Variations of Location and Morphology Below the conus, the spinal canal is filled with a bundle of individual nerve roots called the cauda equina, which looks something like a horse’s tail fanning out downward.
This means that at the L2 and L3 levels, you are typically below the spinal cord proper and within the territory of the cauda equina. But “typically” does important work in that sentence. A pediatric MRI study found the conus termination ranging from the upper edge of T12 all the way down to the L2–L3 disc, with the average sitting at lower L1.5PubMed Central. Assessment of the levels of termination of the conus medullaris and thecal sac in the pediatric population In a small percentage of adults, the cord still extends to L2 or slightly below, which is why spinal anesthesia is typically performed at L3–L4 or lower. A study examining spinal cord position in an East African population found that if a needle were placed at the L2–L3 interspace, about 2% of spinal cords could be at risk of being contacted, a number that jumped to over 33% if the needle drifted up one space to L1–L2.6PubMed Central. Safety of spinal anaesthesia in the East African population: assessment of spinal cord termination level at and above L2 vertebra in adults That steep increase in risk is exactly why L2 marks a meaningful boundary for safe needle insertion.
The Nerve Roots That Exit at L2 and L3
Each lumbar vertebra has openings on its sides, called intervertebral foramina, through which nerve roots pass on their way out of the spinal canal. The L2 nerve root exits below the L2 vertebra, and the L3 nerve root exits below L3. These nerves are part of the lumbar plexus, a network that ultimately forms larger nerves supplying the front and inner portions of the thigh.
Anatomical dissections show that the L1 and L2 nerve roots tend to cross the transverse processes (the bony wings that stick out sideways from each vertebra) near their inner half, while the L3 nerve root crosses more toward the middle or outer portion.7PubMed. The relation between the lumbar vertebrae and the spinal nerves for far lateral lumbar spinal approaches This shift in position matters when surgeons are planning approaches that go around the side of the spine rather than through the back.
A cadaver and CT imaging study measured the space between the exiting nerve root and the bony joint surface in the triangular corridor that surgeons use for minimally invasive procedures. Even at its widest, that space was quite narrow, and the clearance at the proximal end of the triangle measured only about one to two millimeters.8PubMed. Relationship of the Exiting Nerve Root and Superior Articular Process in Kambin’s Triangle There is very little room for error when working near L2 and L3 nerve roots.
How L2 and L3 Symptoms Differ from Lower Back Problems
Most disc herniations occur at L4–L5 or L5–S1 and cause the classic pattern of pain shooting down the back of the leg into the foot. Problems at L2 and L3 look different. When the L2 or L3 nerve root is compressed, the pain and sensory changes tend to show up in the front of the thigh and the groin rather than running down the back of the leg.
A study examining L3 nerve root radiculopathy specifically found that the dominant complaint was thigh pain, reported by the majority of patients, with some experiencing hip or knee pain instead. Sensory disturbance was common, but significant muscle weakness was less frequent.9PubMed. Characteristics of L3 nerve root radiculopathy Research on patients who had the L2 nerve root deliberately cut during tumor surgery found that most developed pain or numbness along the front of the thigh and in the groin. Initially, over 60% experienced weakness in the muscles that flex the hip and extend the knee, but nearly all recovered walking ability over time.10Spine. Motor and Sensory Impairments of the Lower Extremities After L2 Nerve Root Transection During Total en Bloc Spondylectomy
A broader review of lumbar plexus lesions, which involve the nerve network formed partly by L2 and L3 roots, describes weakness of hip bending, knee straightening, and thigh adduction, along with sensory loss across the front and inner surfaces of the thigh and inner lower leg.11PubMed Central. Clinical Presentations of Lumbar Disc Degeneration and Lumbosacral Nerve Lesions The pattern is distinct enough that a clinician who hears “my front thigh goes numb and my knee buckles” should be thinking about upper lumbar levels rather than the usual L4–L5 suspects.
Disc Herniations at L2 and L3 Are Uncommon but Tricky
Disc herniations at L1–L2 and L2–L3 are relatively rare compared to the lower lumbar levels, but they carry some extra risk. The spinal canal is narrower at these upper levels, which means a herniated disc does not have to be very large before it starts compressing multiple nerve roots or, in some individuals, the bottom of the spinal cord itself.12PubMed Central. Clinical features and treatments of upper lumbar disc herniations In one surgical series of upper lumbar herniations, L2–L3 was affected in about 30% of cases and L3–L4 in about 42%, with L1–L2 making up the remainder.13Coluna/Columna. Prevalence and Characteristics of Upper Lumbar Disc Herniations in Our Practice
A systematic review of surgical outcomes for upper lumbar disc herniations found a trend toward slightly better satisfaction rates when the herniation was at L3–L4 compared to L1–L3, though the difference did not reach statistical significance.14PubMed Central. Surgical Outcomes for Upper Lumbar Disc Herniations: A Systematic Review and Meta-analysis The takeaway is that disc problems at L2–L3 are uncommon enough that they can be missed on initial evaluation if the clinician is not thinking about upper lumbar pathology, especially since the symptom pattern does not match the stereotypical sciatica presentation that most people associate with a “slipped disc.”
Range of Motion and Biomechanics
L2–L3 is part of the spine’s biomechanical sweet spot for side-to-side bending. A cadaveric study of 42 lumbar spines found that lateral bending range of motion was greatest at the middle lumbar segments, with L2–L3, L3–L4, and L4–L5 all showing significantly more movement than L1–L2 or L5–S1.15PubMed Central. Range of Motion of the Intact Lumbar Segment: A Multivariate Study of 42 Lumbar Spines A similar pattern held for twisting, though L1–L2 was the only level that clearly lagged behind the others.
Another study looking at how repeated loading affects the spine’s neutral zone found that the motion segments between L2 and L5 all loosened slightly with repetitive bending, gaining a small but measurable increase in flexibility after just ten loading cycles.16PubMed. Effects of repetitive movement on range of motion and stiffness around the neutral orientation of the human lumbar spine This is relevant to everyday life: when you warm up before exercise or feel stiffer first thing in the morning, part of what is happening is that these mid-lumbar segments are going through that initial loosening cycle. Stiffness in the surrounding structures was not significantly affected, suggesting the change is more about the disc and ligaments settling into their loaded state than about the joints themselves becoming lax.
Muscles, Blood Vessels, and the Surgical Corridor
The psoas major muscle runs along the front and sides of the lumbar vertebrae, and its anatomy at L2–L3 is distinctly different from its anatomy lower down. An MRI study found that at L1–L2 and L2–L3, the psoas is relatively thin and positioned more toward the back of the vertebral body, while at L3–L4 and L4–L5 it becomes considerably thicker and shifts forward.17PubMed Central. An MRI study of psoas major and abdominal large vessels with respect to the X/DLIF approach The aorta runs along the left side of the vertebral bodies at these levels, typically sitting to the left of the anterior zone from L1–L2 through L3–L4 before splitting into the two iliac arteries around L4–L5.
The lumbar plexus, the nerve network inside the psoas that receives contributions from L2 and L3, sits in the back third of the muscle at the L2–L3 level.18PubMed Central. Localization of the Lumbar Plexus in the Psoas Muscle At L4–L5, the plexus shifts toward the outside edge of the muscle, which makes lateral surgical approaches at L4–L5 riskier for nerve injury. A cadaveric study quantified this: the percentage of patients with neurovascular structures in the danger zone during a lateral approach was about 7% at L2–L3 compared to roughly 21–44% at L4–L5, depending on the side of approach.19PubMed Central. Anatomy of the psoas muscle and lumbar plexus with respect to the surgical approach for lateral transpsoas interbody fusion This makes L2–L3 one of the safer levels for lateral surgery from a nerve-avoidance standpoint.
The artery of Adamkiewicz, the major blood vessel feeding the lower portion of the spinal cord, also has a notable relationship to L2 and L3. It typically branches off from the aorta somewhere between T9 and L2, entering the spinal canal through a nerve root opening on the left side. In the literature it has been described as arising from as low as L3 but generally not below that level.20Springer Link / Neuroradiology. Unusual origin of the artery of Adamkiewicz from the fourth lumbar artery Damage to this artery during aortic surgery or spinal procedures can cause devastating spinal cord injury, so knowing that its territory overlaps with L2 and L3 is a concern for vascular and spine surgeons alike.
Aging and Degeneration at the Upper Lumbar Levels
There is a common assumption that the most degenerated lumbar segments are always L4–L5 and L5–S1, since those are the levels that bear the most load and tend to develop the most symptomatic disc herniations. The picture is more nuanced than that. A large MRI study of 709 patients spanning ages 16 to 82 found that age-related degenerative changes were actually more pronounced in the upper lumbar segments, L1 through L4, than in the lower segments.21Journal of Neurosurgery: Spine. Redefining lumbar spinal stenosis as a developmental syndrome The study challenged the standard thinking that spinal stenosis at L4–L5 and L5–S1 is primarily caused by wear and tear, suggesting instead that narrowing at those lower levels has a significant developmental component, while the upper levels degenerate more with age.
This finding has practical implications. Spinal stenosis at L2–L3 in an older adult is more likely to be a product of age-related disc bulging, ligament thickening, and joint overgrowth than stenosis at the same level in a younger person. Surgical options exist for single-level stenosis at these upper levels, and comparative studies have examined both minimally invasive and traditional open decompression approaches at L1–L2 and L2–L3.22PubMed. Comparative Analysis of Microendoscopic and Open Laminectomy for Single-Level Lumbar Spinal Stenosis at L1-L2 or L2-L3
When Counting Vertebrae Goes Wrong
One underappreciated challenge with L2 and L3 is knowing that you have the right L2 and L3. Not everyone has exactly five lumbar vertebrae. Transitional vertebrae at the lumbosacral junction, where the lowest lumbar vertebra partially fuses with the sacrum or the lowest thoracic vertebra takes on lumbar characteristics, occur in a substantial minority of people. One imaging study of 147 subjects found lumbosacral transitional vertebrae in 15% and thoracolumbar transitional vertebrae in about 4%.23PubMed. Effect of spinal segment variants on numbering vertebral levels at lumbar MR imaging When radiologists count vertebrae from the bottom up, which is common on lumbar MRI, a transitional vertebra can shift every label by one level. What gets called “L2” might really be L1 or L3, depending on the direction of the error.
Several anatomical landmarks have been tested for their ability to correctly number vertebrae in people with transitional anatomy. The iliac crest tangent sign, a line drawn across the top of the pelvis on imaging, performed reasonably well in one study, with a sensitivity of 81% and good interreader agreement.24PubMed. Merits of different anatomical landmarks for correct numbering of the lumbar vertebrae in lumbosacral transitional anomalies Other landmarks, such as the aortic bifurcation and the origin of the renal arteries, proved unreliable. For patients about to undergo surgery at a specific lumbar level, many centers now require whole-spine imaging or at least full-length films that allow counting from the top down, starting from the easily identifiable C2 vertebra in the neck. Getting the level wrong is one of the oldest and most preventable errors in spine surgery.
The Developing Spine in Babies and Young Children
In premature infants, the spinal cord has not yet finished its relative “ascent” within the canal, and the conus medullaris sits lower than it does in adults. Ultrasound studies of premature babies born at 27 to 29 weeks of gestational age have found the conus at the L2–L3 disc space, roughly two vertebral segments lower than the adult position of L1–L2. The cord gradually moves upward in relative terms as the vertebral column grows faster than the cord, reaching the adult level around the 40th week of gestation.25PubMed. Sonographic determination of the position of the conus medullaris in premature and term infants
A study of infants in the first months of life confirmed this upward trend continues after birth. The proportion of babies whose conus was still below L2 dropped from 13% in the youngest group to under 5% in infants several months old.26PubMed. Sonographic determination of normal Conus Medullaris level and ascent in early infancy This information matters for pediatricians performing lumbar punctures or for radiologists evaluating whether an infant’s cord position is abnormally low, which can be a sign of tethered cord syndrome. Knowing that a low-lying conus is normal for a premature baby prevents unnecessary alarm and intervention.
Posture, Pressure, and the Mid-Lumbar Discs
Much of the classic research on intradiscal pressure, the force pushing outward inside a spinal disc, was measured at the L3–L4 or L4–L5 levels. A comprehensive literature review of pressure studies found that the relationship between posture and disc pressure is not as simple as “sitting is always worse than standing.” For back flexion angles under about 20 degrees, sitting does produce higher intradiscal pressure than standing. But once you bend forward past that angle, standing with a flexed back actually produces more pressure than sitting at the same angle.27PubMed Central. Differences in lumbar spine intradiscal pressure between standing and sitting postures Adding weight to your hands while seated and slightly bent forward increases disc pressure by roughly half.
For L2 and L3 specifically, these pressure findings are relevant because the mid-lumbar discs sit near the top of the lordotic curve, where the load distribution shifts from primarily compressive (straight down) to a combination of compression and shear (angled forces). The ergonomic advice people commonly receive, like “sit up straight” or “don’t slouch,” is a blunt approximation of this biomechanics. The real story is that varying your position frequently matters more than finding one perfect posture, because sustained pressure in any single configuration is what the discs tolerate least well.
Why the Lumbar Curve Exists at All
The lordotic curve that L2 and L3 help form is not just a quirk of human anatomy. It is an adaptation to walking upright. The forward curve of the lumbar spine positions the body’s center of mass over the hips, knees, and ankles, reducing the muscle effort needed to stand and walk.28PubMed Central. Inferring lumbar lordosis in Neandertals and other hominins Research comparing modern human spines to those of Neanderthals and earlier hominins has focused on how much lordosis different species had and how it was distributed across the lumbar vertebrae. The degree of lordosis at the L2–L3 level appears to be part of what distinguishes the fully upright modern human spine from earlier forms that were adapted to a more flexed trunk posture. L2 and L3 contribute to this curve not so much through dramatic wedging of their vertebral bodies (most of the wedge shape is at L4 and L5) but through the orientation of their facet joints and the geometry of their discs, which together permit and guide the curve’s shape.
Pedicle Anatomy and Hardware Placement
When spine surgeons place screws in the lumbar vertebrae for fusion or stabilization, those screws typically pass through the pedicles, the short bony bridges connecting the back part of each vertebra to the main body in front. A morphometric study of lumbar pedicles found that at all levels, including L2 and L3, the pedicle has an oval cross-section rather than a round one, with the taller dimension running top to bottom and the narrower dimension running side to side.29Acta Ortopédica Brasileira. Morphometric study of lumbar vertebrae’s pedicle The cortical wall of the pedicle is thicker at the upper lumbar levels (L1–L3) than at L4 and L5, while the overall pedicle area increases as you move downward. At L2 and L3, surgeons are working with pedicles that have relatively sturdy walls but smaller overall diameters than those lower down, which can constrain screw size. The medial wall is thicker than the lateral wall, an asymmetry that matters because a screw breaching the medial wall enters the spinal canal where nerve roots live.