History of Management of Spasticity with Selective Dorsal Rhizotomy in Children

Understanding of Disease

• 1880s – Bennett and Abbe:  In the late 1880s both Bennett, a London surgeon, and Abbe, a New York surgeon, published the observation that cutting the sensory fibers in a spastic limb not only treated pain in the limb but also decreased the limb’s spasticity (21, 22).

Robert Abbe, MD (left) and William Bennett, KCVO, FRCS (right).

• 1890s – Sherrington and the decerebrate cat model: Sherrington validated Bennett’s and Abbe’s observation that deafferentation of a spastic limb allieviated its spasticity using his decerebrate cat model (48).

Charles Scott Sherrington, OM, PRM, FRCP, FRCS

• 1910s – Foerster: The procedure was applied to humans who had painful spasticity by Foerster in the First World War era but rapidly fell into disfavor because of the inability to maintain limb function and because of the creation of phantom limb pain as a result of too-aggressive deafferentation (24).

Otfried Foerster, MD

• 1960s and 1970s – Montpellier experience: The procedure was reintroduced to neurosurgery for the management of spasticity by Gros with his preservation of tone at the knee and Fasano (33,34) with the introduction of nerve stimulation in the procedure. Sindou then furthered this technique and added detailed knowledge of the DREZ.
• 1980s – Peacock: Peacock introduced evaluation criteria for candidates for the procedure and documented the occurrence of several important complications of the procedure (25). He moved the approach to a cauda equina exposure to allow confirmation of nerve root levels at the exit foramina to reduce urinary complications.

Technological Development

• Ultrasound: The use of percutaneous ultrasound has enabled surgeons to minimize the length of incision for SDR (20). In children ≤8 years of age, ultrasound is used to pinpoint the position of the conus medullaris, a critical operative landmark for the conus approach, prior to incision.
• Operative microscope: The operative microscope greatly enhances lighting and visualization for SDR and other spinal and brain operations. For SDR, the microscope is brought into the field after the dura has been opened. The microscope can be  used for the entire intradural dissection, including the separation of the nerve roots and the rhizotomy.
• Intraoperative neurophysiological monitoring: Although its use is controversial (28, 30), we believe that IOM is an important element of SDR. Electromyography (EMG) permits delineation of the innervation of a nerve root prior to division into rootlets and assessment of response strength of individual rootlets prior to sectioning them (27, 29).

Surgical Technique

Two categories of surgical techniques for SDR:

In essence, the SDR techniques currently in use are based on one of two historic techniques:

• Cauda equina approach (“Peacock technique”): The cauda equina approach, described by Peacock in 1982, uses a multisegment laminotomy from L2 to L5 to allow visualization of the corresponding nerve roots as they exits the foraminae from L2 to S1 (25).
• Conus approach (“Park technique”): The conus approach, described by Park and Johnston in 2006, uses a limited laminectomy (usually a single lamina) to exposes the nerve roots as they exit the conus (20). This is like the techniques used by Foerster and later Gros.

Common features for both categories:

• Selective lesioning using neurophysiologic monitoring: Fasano was the first to use some form of intraoperative neurophysiological response to grade the abnormal responses to electrical stimulation of sensory nerves coming from spastic legs (33,34).

Fasano’s illustration of normal and abnormal responses (74).  On the left are EMG recordings of leg muscle contraction patterns seen in response to electrical stimulation of sensory nerves innervating the muscle.  ‘A’ is a so-called normal response with degradation in the amplitude of response with repeated stimulus pulse of a stimulation train.  The degree of degradation increases as the stimulus train’s frequency increases as shown in ‘A’.  ‘B’ and ‘C’ are abnormal recordings with ‘B’ showing persisting response to stimuli of a train and ‘C’ showing fluctuating amplitude of  response to the stimulus train.  On the right are EMG recordings of two muscles contracting in response to a stimulus train of 50 Hz with ‘B’ being the EMG recording of a muscle not innervated by the sensory nerve being stimulated, the so-call “spreading” response that is considered “more” abnormal by most surgeons performing SDR.

• 1990s – Park grading responses: The object of SDR is to identify the rootlets that are participating in the abnormal segmental circuitry responsible for spasticity. Although several papers have looked at different responses to stimulation, the most widely used criterion remains the spread of the stimulus to contralateral musculature or to the upper-extremity musculature, a so-called grade 4 compound motor action potential response as described by Phillips & Park (4). It has been shown that the response to sensory root stimulation is relatively consistent, rarely varying by more than one grade on this scale when restimulated (19).
• 2000s – Park and thoracolumbar conus exposure for surgery: Though the  procedure is still done at the cauda equina level, modifications have recently been introduced by Park for the procedure to be done at the conus level or thoracolumbar junction (20). Advantages of the cauda equina exposure include a secure anatomical identification of each of the nerve roots. Advantages of a thoracolumbar junction approach include a slightly shorter duration of the procedure, a shorter incision, and avoidance of the need to separate the dorsal roots into their component rootlets.
• Anesthetic technique: After the introduction of neurophysiological monitoring for selection of nerves to be lesioned, the importance of anesthetic technique for SDR became apparent. As with other surgeries requiring neurophysiologic monitoring, the combination of propofol and remifentanil was found to be the best regimen to obtain good responses to stimulation without having adverse hemodynamic effects related to “too-light” an anesthesia. The use of a “BIS” monitor to determine the degree of sedation has been used by many anesthetists. Patient awareness is not a common problem with this regimen. Additionally, too-light anesthesia prior to the advent of propofol led patients with a history of bronchopulmonary dysplasia of prematurity to develop reactive airway disease intraoperatively (35). With the propofol-remifentanyl regimen, with or without perioperative bronchodilators, this is almost never seen