* Neuroradiology *
* *


Key Points 1
General 1
Basic Anatomy and Physiology 1
Indications for Myelography 2
Alternative Examinations 2
Steroid Premedication Regimen 2
Contraindications and Acceptable Lab Values 2
Drugs that Lower Seizure Threshold 3
Possible Complications 4
Post Procedure Instructions 5
Post Myelogram Headache 5
Equipment 6
Table Weight Limits 6
Iodinated Contrast 6
Admonition 6
Vasovagal Reaction 10
Lumbar Myelogram with Congenital Lumbar Anomaly 13
Cervical Approach for Lumbar Myelogram 13
Cervical Myelography--Lumbar Approach 14
Cervical Myelography--Cervical Approach 15
Total Spine (not for AVM) 17
Total Spine Myelogram to Assess for AVM 17
Lumbar Puncture under Fluoroscopy 19
CSF Opening Pressure 19
CT Cisternogram 20
Radionuclide Cisternogram 20
Chemotherapy Injection 21

Instructions for Sending Patients to the Medical
Procedures Area (MRA)



Post Dural Headaches and Epidural Blood Patches


Mallinckrodt Institute of Radiology

Key Points

1.  Only Non-ionic contrast can be injected into the thecal sac.  We use Omnipaque.  Injection of ionic contrast into the thecal sac can result in death.

2.  While performing the procedure remain aware of the patient's condition.  Vasovagal reactions are fairly common with myelograms, especially in young muscular males (football player types).  By intervening early one can avoid a potentially life‑threatening reaction.  See the discussion later in this handout.

3.  Proper labeling of the vertebral bodies is a critical aspect of the myelographic examination.  Surgeons rely on correct labeling and communication of the lesion location.

4.  When performing Cervical Myelograms, take great care when extending the patient's neck.  Prolonged extension or over-extension of the neck in a patient with a high-grade cervical canal stenosis can result in permanent cord damage, even quadriplegia.  Obtain and review any previous imaging studies prior to performing the procedure.

5.  Avoid passing a needle through an epidural abscess into the thecal sac.  This could result in meningitis.  This also applies to any abscess or infected decubitus ulcer, etc.

6.  During a cervical myelogram via a cervical needle placement, avoid injecting contrast directly into the cord by never injecting contrast unless there is good flow of CSF out through the needle.


Myelography refers to the examination of the contents of the thecal sac after administration of intrathecal radiographic contrast.  The examination is done with both plain films and CT.

Keep in mind that you will be doing multiple examinations during your Neuroradiology rotation; therefore, it is important to limit your exposure to radiation during each examination.  Stand away from the tube and limit your fluoroscopy time.

Basic Anatomy and Physiology

The total adult CSF volume is about 150 ml (50% intracranial, 50% spinal).  About 500-750 ml of CSF is produced each day (0.4 ml/min, 20-30 ml/hr).  Adult opening pressure is normally 7-15 cm fluid, >18 abnormal (although young adult can be slightly higher with normal <18-20).

According to Dr. Hodges:  The AP diameter of the cord is 7 mm down to C7, 6 mm from C7 to the conus, then 7 mm at the conus.  The cord size can be considered abnormal if it is over 8 mm or under 6 mm.

There are normally 7 cervical vertebral bodies, 12 thoracic vertebral bodies, 5 Lumbar vertebral bodies, the sacrum and the coccyx.  There are corresponding nerve roots: 8 Cervical, 12 Thoracic, 5 Lumbar, 5 Sacral, and the coccygeal nerve.

The position of the tip of the conus at birth is debated.  By 3 months it is usually at the normal adult level of mid L1 to mid L2.  It is considered abnormal if it is below the L2-L3 disk space level (ref. Barkovitch).

Indications for Myelography

This examination is usually performed to assess for HNP or spinal stenosis.  Less often it is used to determine the level of spinal cord compression from metastatic disease or trauma.

Myelography often yields better information about the bony structures than does MRI. 

Alternative Examinations

MRI of the spine or a noncontrast CT can also yield useful information.

Steroid Premedication Regimen

Patients with a history of reaction to iodinated contrast should receive premedication with steroids and Benadryl (diphenhydramine) prior to contrast administration.  However, contrast reaction is extremely rare following myelography.  The usual premedication regimen (Manual on Contrast Media, 4th Ed. ACR, 1998) is: prednisone 50 mg PO 13, 7, and 1 hour before contrast administration; Benadryl 50 mg PO (or IM or IV) one hour before contrast administration.  If Benadryl is given make sure someone else is available to drive the patient home.  For patients who can not take oral medication, substitute hydrocortisone 200 mg IV for each prednisone dose and give diphenhydramine IV.  A period of at least 6 hours between the onset of corticosteroid administration and the injection of contrast medium is recommended, regardless of the route of steroid administration.

H2 blockers can be used for patients with a history of severe reaction; cimetidine (Tagamet) 300 mg PO (or IV; diluted in 20 ml or more normal saline and infused slowly over a minimum of 5 minutes) 1 hour prior to contrast administration.  Cimetidine should be given only to patients also receiving diphenhydramine.

In emergency situations, intravenous medications can be administered: hydrocortisone (Solu-Cortef) 200 mg IV stat and q 4 hour until the examination is complete.  Benadryl 50 mg IV 1 hour before contrast administration.  Premedication with Epinephrine can also be considered, but caution is advised in patients with unstable angina, arrhythmia or hypertension (Manual on Contrast Media, ACR).  A period of at least 6 hours between the onset of corticosteroid administration and the injection of contrast medium is recommended, regardless of the route of steroid administration.

ref:  Greenberger PA, et al.  (Oral) J Allergy Clin Immunol 1991; 87:867-872 and (IV) J Allergy Clin Immunol 1986; 77:630-634.

Manual on Contrast Media, 4th Edition, ACR monograph, 1998.

Contraindications and Acceptable Lab Values

Lab values

PT (nl 10-12) acceptable < 15.0 seconds

Platelets (nl 150,000-450,000) (bleeding time nl for platelets >100,000,  Transfuse < 50,000)

Coagulation parameters should be within normal limits.  Check PT and PTT (and platelets) if there is a clinical condition that predisposes the patient to bleeding.  It is preferable to cancel or delay the examination if PT > 15.0.  If possible, heparin should be held for 4 hours prior to the procedure.  It can be restarted 2 hours after the procedure.  Coumadin should be held for  3-4 days, until PT is normal.  If platelets are below 50,000 (some prefer 70,000), a platelet transfusion can be given prior to LP (this situation occurs especially in patients in whom an LP is being performed for administration of intrathecal chemotherapy).  Transfusion of one unit of platelets raises the platelet count by 10‑20,000.  Check the platelet count before continuing with the LP.

Creatinine is usually not checked prior to the myelogram.  If the patient has very poor renal function you could time the myelogram to precede dialysis, or perform an MRI.

If a normal healthy outpatient arrives without recent laboratory results it may be possible to continue the examination at the discretion of the staff.  Ask the patient about history of renal disease or bleeding abnormalities before proceeding.

Medications that lower the seizure threshold should not be administered for 48 hours prior to or following the procedure.  The patient should be informed during the consent procedure that their risk of seizure is higher than if they were not taking the medication.  This mostly applies to tricyclic antidepressants.  Check the PDR if uncertain whether the medication lowers the seizure threshold.  (Many of these medications have half-lives well over 48 hours; however, usually we still hold the medication for 48 hours.)

Drugs that lower seizure threshold (not an exhaustive list)

(source: Neuroradiology, The Requisites, R.I. Grossman, D.M. Yousem):

Phenothiazines (chlorpromazine [Thorazine], prochlorperazine [Compazine], perphenazine [Etrafon, Trilafon], thioridazine [Mellaril])

Antipsychotics (thiothixene [Navane], haloperidol [Haldol], droperidol [Fentanyl])

Tricyclic antidepressants (amitriptyline [Elavil], desipramine [Norparmin], imipramine [Tofranil], nortryptyline [Pamelor], doxepin [Sinequan])

CNS stimulants (methylphenidate [Ritalin], ephedrine, pseudoephedrine)

Monoamine oxidase inhibitors (tranylcypromine [Parnate], procarbazine [Matulane])

Others (lithium, reserpine, isoniazid)

If the patient has a history of seizures or is taking seizure medications, there is no premedication for seizures.  Discuss with the referring physician and the patient that there is a higher risk of the contrast inducing a seizure than in a patient without a seizure history.  The patient should continue to take their usual seizure medications prior to the myelogram.

Caution must be exercised in patients taking the oral antihyperglycemic agent Glucophage (metformin) because of the risk of renal failure or lactic acidosis after receiving iodinated contrast.  The FDA package insert states that Glucophage should be withheld temporarily in patients undergoing radiological studies using intravenous iodinated contrast media.  We apply this to intrathecal iodinated contrast media also.

The Manual on Contrast Media, 4th Ed. by the ACR (1998) states that "metformin should be discontinued before or at the time of the procedure, withheld for 48 hours after the procedure, and be reinstated only after renal function has been re-evaluated and found to be normal."  Be sure to call the patient's physician so that he/she can manage the patient's diabetes while off Glucophage.  Call the patient's physician before the procedure to verify that the proposed plan to manage the diabetes is acceptable and discuss with the physician instructing the patient as to when to resume taking Glucophage.

Low Molecular Weight Heparin (Fragmin, Lovenox, Normiflow, Orgaran) are a contraindication to LP or Myelogram.

Of course, never pass a needle through an epidural abscess into the thecal sac.  If a patient has a suspected lumbar epidural abscess, an MRI is usually the best means of evaluation.  If a myelogram must be done, introduce the contrast into the thecal sac using the cervical approach.

A complete CSF Block is a contraindication to collecting CSF below the block.  Reduction in CSF volume and pressure below the block can cause downward herniation of the cord.  The risk versus benefit of the procedure must be considered in each case individually.

A Complete CSF Block is a also a contraindication to injecting contrast below the block (of course, it is acceptable to inject contrast above the block).  This is because the block prohibits resorption of the intrathecal contrast by the arachnoid villi in the head and leaves neurotoxic contrast in contact with the spinal cord and nerve roots.  Resorption through the ependyma is only minimal and will not suffice in this situation.

If a complete block is suspected, one can inject two cc of contrast below the suspected level of the block as a test.  Run the contrast up into the head to assess whether a block is present at any level.  If no block is present, inject the remainder of the contrast and continue with the study.  If a block is present, do not inject any more contrast.  The two cc of contrast already injected should be visible on the CT examination and can help define the lower extent of the block.

If the patient has immediate pain during the contrast injection it might be due to distention of the thecal sac below an unsuspected block.  Stop injecting and investigate whether a block is present.

Other general contraindications include medical conditions that might lead to complications.  For instance, a patient with bacteremia from a tooth abscess should not undergo myelography because of the risk of meningitis.

Possible Complications

The most common complications are due to meningeal reactions, spinal headache, vomiting, vertigo, and neck pain.  This is partly the result of CSF loss due to dural injury from the puncture.  This complication is minimized by using a small needle.  It also helps to orient the bevel of the needle parallel to the longitudinal fibers of the thecal sac during puncture (to separate the fibers rather than cut them).  The limiting factor for needle caliber is the viscosity of the injected contrast material.

The typical headache after puncture can be distinguished from migraine or other types of headache by the increased severity in the upright position and the spontaneous improvement in recumbency.  It has its onset immediately after puncture or within a few hours.

Other complications include nerve root damage, meningitis, epidural abscess, contrast reaction, CSF leak, or hemorrhage,

Unlikely complications include damage to the spinal cord, such as due to a low conus or tethered cord with a lumbar approach or direct cord damage in a cervical approach.  Other complications include death or paralysis from cord damage due to injection of contrast into the cord or hemorrhage in the cord from needle damage.

If it is discovered during the injection that a large fraction of the contrast has gone into the subdural space it is usually best to discontinue the study and reschedule it for two weeks later.  This is because the enlarged (contrast containing) subdural space fills the region of the canal that previously contained subarachnoid space.  It is thus difficult to reposition the needle tip into the subarachnoid space.

If a small subdural injection is discovered early, it may be possible to reposition the needle and continue with the study.

Post Procedure Instructions

The routine post procedure orders are:

Strict bedrest for 4-8h with HOB elevated 30 degrees, at discretion of the physician, then bathroom privileges the rest of the day.

Light activity for 24-48 h after discharge.

Force Fluids.

No phenothiazines, tricyclic antidepressants or Tigan for 48 hours after study.

Also, the patient should be told that if there are any signs of meningitis (severe headache with stiff neck, fever) within 48 hours he or she  should contact their doctor or go to the emergency room.

Post Myelogram Headache

Tylenol, horizontal position, forced fluids, and caffeine all help relieve the headache.  The headache can persist up to a week after the procedure.  If headache persists over 24 hours after the myelogram or if there is a fever or signs of meningitis the patient should contact his referring clinician or go to the emergency room. 

It may be necessary for the patient to receive a blood patch to alleviate the headache.  Blood patches are effective at stopping post myelogram headaches anytime between 24 hours and several weeks after the procedure.  The first blood patch is effective in 70% of patients, a second blood patch increases the effectiveness to 95%.  If a headache persists at 48 hours after the procedure a blood patch should be considered.  In the case of a severe headache, a blood patch should be considered at 24 hours.  This involves an injection of 10-15 cc of autologous blood into the epidural space.   The Neuroradiology Service performs the blood patch procedure on our post‑myelogram patients.  See Dr. Wippold's monograph about the blood patch procedure, included here as an Appendix.  The headache usually stops immediately after the injection of blood.  This suggests that the mechanism of headache is not simply CSF leak.  The procedure is effective even if the epidural blood injection is not at the same level as the original myelogram needle puncture.  (The Pain Service also performs the blood patch procedure, 362-8840.)


Spinal needles come in 3.5-, 4.5- and 6‑inch lengths.  The standard needle is 22- or 25-gauge 3.5 inch.  The longer needles are larger-gauge (20-gauge 4.5 cm; and 18-gauge 6 cm).  The use of smaller gauge needles has been shown to reduce the incidence of spinal headaches.  The flow of CSF through a 25-gauge needle is insufficient to allow for CSF collection.  If a CSF sample is to be obtained, use a 22 gauge or larger needle.  

Before starting the procedure, estimate the necessary length of needle.  It is far preferable to start with a 4.5-cm needle than to find out that you are slightly short after the attempted placement of a 3.5-cm needle.

Table Weight Limits

According to the Special Procedures Supervisor, Roberta McQueen-Eads, the published table weight limits are:

Rooms 321 and 326 (Siemens Angio Rooms): 330 pounds.

Room 309 (Philips Myelo Room): 297 pounds (135 kg).

CT on 3rd floor:  400 pounds (according to Roberta).

CT Somatom Plus 4 on 2nd floor:  500 pounds (according to Roberta).

Iodinated Contrast

Nonionic contrast must be used for the examination.  Intrathecal administration of Ionic iodinated contrast can cause death.

We use Omnipaque (Iohexol).  It comes in 20-ml vials with concentrations of 180 mg I/ml and 300 mg I/ml.  The adult dose limit for myelography is 3 g total of iodine (i.e., 17 ml of 180 mg I /ml; or 10 ml of 300 mg I/ml).  Use 180 for lumbar myelograms and 300 for cervical, thoracic, or combined myelograms.  (Some prefer to use 180 for cervical myelograms with a cervical approach).  Consult the contrast material package insert for pediatric dose limits.

Before drawing the contrast into the syringe, the technologist should show you the bottle.  Verify that (1) non -ionic contrast is being used (Omnipaque),  (2) the expiration date has not passed (the date will appear, for example, as 07 02 indicating July 2002), and (3) the desired concentration of contrast (180 or 300 mg I/ml) is being used.


Procedures in neuroradiology are based on finesse, not brute force.  Plan what you want to do, set it up, recheck it, then do it.  Avoid multiple passes with the myelography needle by knowing ahead of time what you want to do and how to do it.


Preliminary steps

Before beginning, talk to the patient.  Obtain a history and learn what information is desired from the test.  Explain the risks, benefits, and alternatives to the patient.  Obtain written informed consent.  The post myelogram orders should be explained to the patient.

Write a short note in the chart before the procedure including: indication for examination, labs, allergies, documentation of the consent, planned procedure.

Check the lab values.  Check whether the referring clinician has any special requests concerning the myelographic examination or the CSF analysis.

Obtain old MR, CT, plain films and myelograms and reports.  It is important to look at the old films prior to starting the procedure.  Determine the level of the suspected pathology; usually you should avoid placing the needle at this level.  Exclude tethered cord or low-lying conus.  Count ribs from chest x-rays and old myelograms.

Place the patient prone on the table.  It is useful to have a pillow under the abdomen producing a slight flexion of the L-spine.  Using fluoroscopy, carefully verify the number of ribs and the number of vertebrae.  Document with plain films.

Have everything needed for the procedure set up before beginning.  Have the contrast drawn up and flushed through the long connection tube.  The short connection tube is used for CSF collection.  The long connection tube holds 1.9 cc and the short connection tube holds just over 1.0 cc.

Locate the desired entry point with fluoroscopy and place an ink mark on the skin.  Prep and drape the patient.  Position the patient in true prone, such that the transverse processes project in the midline between the pedicles on fluoroscopy.  Locate the desired level for needle placement.  (See Figures 1 and 2 at the end of this section.)  The usual level is L2-3.  Above this, at L1-2, there is a risk of damaging the conus (which normally lies at L1-2 or above).  Degenerative disease is usually below this level, and the presence of a large disk herniation or bulge can make it difficult to place the needle tip in the thecal sac.  The thecal sac is also smaller at the lower levels, especially if there is epidural lipomatosis.  The normal lumbar lordosis makes the L2-3 spinous processes and inter laminar space relatively perpendicular to the table for easier needle placement than at other levels.

Estimate the approximate depth of the thecal sac (i.e., is it superficial in a thin patient or deep in a large patient?).   Anesthetize the skin with lidocaine.  Deep lidocaine may not be necessary, a single pass of a 22- or 25-gauge spinal needle is less painful than multiple passes with a 22-gauge lidocaine injection needle and lidocaine injection.

Lidocaine mixed with bicarbonate is less painful than straight lidocaine.  Use a mixture of 10 cc of 1-% lidocaine and 1 cc of 8.4% sodium bicarbonate.  (ref: Hospital Pharmacy, 31(10), p. 1275-1282, (1996).

The attending spine neuroradiologist must be present for the surgical procedure portion of the examination, i.e. needle placement and contrast injection.  This is because a procedure surgical code is used in billing in addition to the film interpretation code.

The two most common approaches for lumbar needle placement are midsagittal and parasagittal.  It is useful to be proficient at both approaches because in some difficult cases one may work better than the other.

In the midsagittal approach the needle is usually placed between the L2 and L3 spinous processes, through the interspinous ligament.  (See Figures 1 and 2 at the end of this section.)  The needle should be positioned during placement such that the fluoroscopy beam looks "down the barrel."  The needle should appear to be a dense dot projected between the transverse processes on AP fluoroscopy.  Hold the needle with two hands; one index finger should hold the stylet in place, the other hand should be at the skin.

AP fluoroscopy should be used periodically to verify the needle trajectory.  It is often not necessary to move the fluoroscopy unit back and forth, completely out of the way; just reach under the unit to advance the needle (for large patients this may not be possible).  Insert the needle to the estimated depth of the thecal sac.  Check the depth on lateral fluoroscopy.  Check for CSF return by tilting the table slightly head up and removing the stylet.  Move the image intensifier out of the way to remove the stylet.  Check for CSF return.  (Note:  It is very important that the head of the table is tilted up so that the column of CSF rises above the hub of the needle before withdrawing the stylet, otherwise there will be no CSF return.  This might lead you to erroneously believe that the needle tip is not in the thecal sac, leading to unnecessary repositioning.)  With a 25‑gauge needle the CSF return might be minimal and/or slow.

The needle can be "steered" slightly by using the bevel.  The bevel side is indicated by a raised mark on the hub.  By rotating the needle, the bevel can be used to deflect the needle away from obstructions.

The advantage of the mid sagittal approach is that it is less painful than the parasagittal approach because the ligament is not innervated to the same degree as the paraspinal muscles.

The disadvantage of the midsagittal approach is that in a severely degenerated back it may be difficult to place the needle between the spinal processes.  The space between the processes may be diminished, and the ligament may be severely calcified.  This, along with the normal caudally oriented spinous process and interspinous space, makes it difficult to access the thecal sac with a needle that is oriented perpendicular to the table.  Do not negate the advantages of fluoroscopy by trying to place the needle at an angle other than parallel to the fluoroscopic x-ray beam.

The oblique parasagittal technique is performed by placing the needle on either side of the spinous process.  The side with the greatest interlaminar space is chosen.  The patient is positioned with the knee slightly turned outward on the side of desired needle placement.  This rotates the spine a minimal amount (about 10 degrees) to demonstrate the interlaminar space just to the side of the spinous process.  Proceed as in the midsagittal needle placement, being sure to use the fluoroscopy beam correctly by looking "down the barrel" of the needle and seeing that it is directly superimposed on the desired target space.

The advantage of the oblique parasagittal technique is that the needle is not forced through the interspinous ligament.  It need not negotiate a narrowed obliquely oriented interspinous space.

The disadvantage of the oblique parasagittal approach is that it may be more painful than the midsagittal approach because the needle passes through muscle instead of through ligaments.  However, if the needle can be placed more quickly and with fewer passes, the overall patient discomfort is less than with the other technique.

Figure 1
Figure 1.   An AP view of the lumbar spine showing the osseus structures.  The needle is oriented parallel to the fluoroscopy beam in a midsagittal approach.  The fluoroscopic beam also has caudal angulation.

Figure 2

Figure 2.  A lateral view of the lumbar spine showing the relationship between the osseus structures, the needle and the image intensifier.  The x-ray beam is parallel to the path of the needle.  The image intensifier is angled to optimally demonstrate the space between the spinous processes and laminae of L2 and L3.

Vasovagal Reaction

Occasionally a patient may have a vasovagal reaction during placement of the needle.  This occurs most often in large young physically fit men (football player types).  Be aware of the patient fidgeting and sweating.  Stop the procedure, place the patient in Trendelenburg position and place a cool wet towel on the back of his or her neck and head before the reaction becomes severe.  Assess their pulse rate and blood pressure.  Let the patient recover for an extra few minutes after they seem to have recovered.  Trendelenburg position and smelling salts are useful in moderate reactions (vials of smelling salts are taped to the image intensifier).  A severe reaction (hypotension, bradycardia) may require IV Atropine and rapidly infused IV normal saline.  The Atropine comes in premeasured syringes containing 1 mg.  The usual Atropine dose is 0.8 - 1.0 mg IV slow push.  It can be repeated in 3-5 minutes, up to 2 mg total.  But the key point is to notice the early signs, attend to the patient and avoid a severe reaction.  After the patient recovers from a vasovagal reaction they usually will not have another during the remainder of the procedure.

Once the needle is placed and there is CSF return, gently spin the needle around 180 degrees (with the stylet in place).  This clears the needle from the dura and helps avoid subdural or epidural injections.

Starting October 20, 1997 we will collect CSF for laboratory analysis only when it is requested by the referring clinician.  If CSF is to be collected, collect at least 1 cc in each of the 3 test tubes in the order of the test-tube label (1, 2, then 3).  The rate the CSF collects can be increased by having the patient periodically cough or bear down.  Slightly more than 1 ml of fluid fills the short connection tube.  Collect the CSF in the collection tubing when filling test-tube 3 by holding it at a downward slope and gently disconnecting the tube from the needle and allowing it to drain into the test-tube by gravity and siphon action.  Tightly close the tubes and give them to the technologists for labeling.  The technologists will transport them to the laboratory.  The samples are analyzed for:  cell count and differential, VDRL, glucose, and protein.  Patients tend not to like to have CSF drip onto their backs; try to avoid it.

Sometimes there is a request to collect a large volume of CSF.  Neurosurgeons suggest that we do not collect more than 30 cc.  Removal of larger volumes tends to cause subdural hematomas.

Connect the contrast syringe.  Note that if 20 ml are drawn up initially, and 2 ml are flushed through the tube, there remains 18 ml contrast in the syringe.  Thus the maximum allowable dose of 3 gm contrast can be administered without flushing through the contrast in the connection tube.  (Note that the maximal dose of 17-cc contrast is seldom necessary, 13 to 14 cc are usually sufficient.)  The goal is to administer sufficient contrast so that the thecal sac is completely filled below the level of the midbody of L3 when the patient is upright.

The contrast is injected under fluoroscopic observation to avoid subdural or epidural injections.  The table should be tilted with the patient's head slightly upward.  Lateral fluoroscopy should be used initially to clearly see the first puff of contrast freely fall away from the needle tip to form a dense horizontal line of contrast along the ventral side of the canal.  A subdural or epidural injection can be identified by accumulation of contrast at the needle tip or in the dorsal (non-dependent) part of the spinal canal.

After one or two tests puffs are injected, then a steady slow injection can be performed.  During this time, periodic lateral (or AP) fluoroscopy shows filling of the caudal portion of the sac, with no accumulation of contrast near the needle tip.  Fluoro periodically throughout the injection to verify that the needle tip does not migrate and cause a subdural or epidural injection.  If contrast starts to accumulate near the needle tip during the injection it usually indicates that the injection has become subdural.  Stop the injection and consult the staff or fellow. 

Usually 13 to 14 cc of intrathecal contrast are sufficient.  This dose reduces the rate of complications relative to the highest allowable dose of 17 cc.  If the patient has a patulous thecal sac the highest allowable dose of 17 cc may be necessary.  The tubing is then disconnected and the stylet is replaced.  The needle and stylet are left in place (or withdrawn slightly so that the tip is in the soft tissues of the back) during filming to document its location and to help localize lumbar vertebral levels.

Digital plain film images are obtained.  Make sure that the entire extent of the very distal caudal thecal sac is filled with contrast before imaging.  This may require standing the patient almost completely upright (at least 45 degrees of table tilt should be done in all patients).  If there is a tight stenosis that impedes passage of the contrast into the lower lumbar/sacral canal, stand the patient up and have him or her flex and extend.  This usually opens the canal sufficiently to allow for some passage of contrast.  It is very rare to have such a tight stenosis that insufficient contrast passes for adequate CT evaluation, even if the plain films do not show adequate contrast.

AP, lateral, shallow oblique and steep oblique images are obtained for the lower L Spine.  (The shallow and steep obliques are obtained at about 15 degrees and 25 degrees off AP, respectively.)  The Tavaras View profiles the L5 and S1 nerve roots.  Head of the bed is elevated 30 degrees so that the contrast accumulates in the lower thecal sac.  An image is obtained with the tube angled 30 degrees relative to the patient (so that it is perpendicular to the floor).

The patient is then positioned to move the contrast column higher and the same images are obtained for the upper L spine.  The lower images should include the sacrum and the upper images should include the lowest rib to provide unequivocal landmarks for localization of levels (this is a good habit to adopt even though the needle is also in place for localization.)

Supine and weight-bearing flexion and extension lateral views are obtained. 

Finally, a conus view is obtained with the patient supine so that the contrast accumulates near the thoraco-lumbar junction.

During the plain film examination note levels of pathology so that these can be included in the CT examination.  This is especially true for pathology that is outside of the usual limits of the CT examination (i.e., low thoracic or sacral).

A short note is left in the chart documenting the procedure, and describing any preliminary fluoroscopic findings.

The patient is transferred off of the myelogram table onto a gurney and transported to the CT scanner.  While waiting for the CT scan, the patient should remain basically in the prone position so that the contrast stays dependent in the lumbar lordosis (unless he or she is unable to tolerate this position).  The patient is then turned supine for the CT scan. 

If there is a long delay before the CT scan (10 minutes or more), the contrast might settle too much in the spinal canal causing layering.  If the delay between the myelogram and the CT is longer than 10 minutes, the CT technologist should tell the patient to alternately roll gently and slowly toward one side, then toward the other side, etc.  This mixes the contrast and helps avoid layering.  (You can have the patient wait for the CT in the prone position; this ensures that he or she will turn over at least once into the supine position for the CT.  Also, in the prone position, the contrast collects in the lumbar lordosis--where you want it--rather than pooling in the sacrum.)

The technologists should also check the scanogram at the beginning of the CT scan.  If the contrast is seen layering in the lumbar or sacral regions, the patient should be rolled from side to side or completely around to mix the contrast before proceeding with the CT scan.

CT is performed.  For the "disk protocol" the images cover L1-S1 with 3-mm slices.  The disk protocol slices are grouped in sets (L12, L23, L34, L45, L5S1), each parallel to the disk space.  For the "stenosis protocol" the images also cover L1-S1 but 5-mm slices are used and the images are all parallel and contiguous over the imaged region.  For suspected pathology outside of the usual levels, instruct the technologists to obtain images at additional levels.

CT images are photographed at a "myelographic window" between usual CT bone and soft tissue windows (W 2000/C 0).  This window is optimal for viewing the contrast in the thecal sac. 

The films are hung on the myelogram board in the spine reading room along with comparison old examinations.  The examination must be read and dictated on the day of the procedure since many patients will have surgery the following morning.  The films are left up on the myelogram board after dictation until the following morning.  The procedure and findings should be documented in the myelogram book in the reading room.

The dictation must include as the last line of the procedure section, "Dr. Smith, the attending neuroradiologist, was present during the procedure."

The findings portion of the report is best organized level by level, i.e. all of the findings at L12 are described, then L23, etc.

At North Campus, Dr. Cho does his own puncture and dictates his own procedure note.  The neuroradiologist performs the fluoroscopy during needle placement and then interprets and dictates the findings of the examination.

Lumbar Myelogram with Congenital Lumbar Anomaly

These patients should usually be evaluated by MRI before the myelogram in order to delineate the exact nature of their anomaly.  Often a lumbar myelogram is obtained by injecting the contrast in the cervical region and running it down to the lumbar region.  This approach avoids placing the needle through a region of anomaly.  However, many of these patients have associated cervical anomalies such as low lying cerebellar tonsils.  Before placing a needle into the cervical region you must be sure that there is no accompanying cervical lesion.  A cervical MRI is the best and most positive proof.  If you are sure that the patient's particular lumbar anomaly is not associated with a cervical anomaly you might proceed without an MRI.

Cervical Approach for Lumbar Myelogram

Occasionally it is necessary to do a cervical approach for lumbar myelogram.  Perform the puncture as described below.  It is usually best to use 300 mg I/ml contrast so that adequate opacification will be obtained after the contrast dilutes on its way down to the lumbar region.  Inject the contrast with the head of the bed up and allow the contrast to flow into the lumbar region (and not into the head). 

Note that only Neuroradiology Fellows and Staff are allowed to perform C1-2 punctures and injections.


Cervical Myelography--Lumbar Approach

The history, consent, and old films are as described above.

Using the lumbar approach for cervical myelography, the contrast is administered in the lumbar region and manipulated under gravity to the cervical region.  The patient can be positioned either prone or in the lateral decubitus position while the contrast flows to the cervical region.

Beware, spinal cord damage can result if a patient with severe cervical canal stenosis is over extended.  Look at previous imaging studies before positioning the patient to assess for cervical spinal canal stenosis.  If there is  severe stenosis consider either a cervical approach for introduction of contrast or use the lateral decubitus position for running the contrast up from a lumbar introduction.  Note that a severe stenosis will impede passage of the contrast so there is no need to extend the neck to create a lordotic depression to trap the contrast.

Patient in Prone Position:  Pillow supports are placed around the head to gently extend the neck and produce a lordotic cervical curvature.  The contrast will collect in this lordotic depression rather than running into the head. 

Access to the thecal sac is as was described above for the lumbar myelogram.  10 cc of 300-mg I/ml contrast are administered.  During administration the table is tipped with the patient's head slightly downward.  Care is taken to avoid too steep of an angle, which would send the contrast beyond the cervical lordosis and into the head.  However, the angle must be steep enough to allow the contrast to flow over the normal thoracic kyphosis.  The contrast can be observed either with lateral or AP fluoro.  First verify the subarachnoid injection in the lumber region, then move to the cervical region to monitor the contrast accumulation.  Proceed with imaging as below.

Patient in Lateral Decubitus Position:  By having the spine horizontal, this position avoids the problem of getting the contrast over the thoracic kyphosis without having it flow into the head.  This method is most useful in patients with an exaggerated thoracic kyphosis and least useful in patients with scoliosis.  Access the lumbar thecal sac.  With the table level, administer the contrast.  Adjust the angle of the table to avoid having the contrast flow into either the sacral or thoracic region.

Position the patient in the lateral decubitus position facing away from you.  Have the patient turn his/her face toward the ceiling and tip their upper ear toward the upper shoulder to create a cervical depression.  Using LATERAL fluoro observe the contrast as it flows cephalad while tilting the table.  Once the contrast is in the cervico-thoracic region have the patient turn prone with his/her head extended.  Adjust the table tilt to maximize the contrast in the cervical region.

Images are obtained: AP, lateral, steep and shallow obliques.  A swimmer's view of the cervical-thoracic junction is obtained.  The plain film images should be acquired expeditiously while the contrast is well concentrated in the cervical region.

CT of the cervical spine is obtained with 4-mm slices from the foramen magnum to C4, and 2-mm thick images from C4 to T1.  Additional levels can be obtained if necessary.

Cervical Myelography--Cervical Approach

The cervical approach should be performed only by Neuroradiology Staff or Fellows.  Residents are encouraged to observe. 

The cervical approach shares the same risks and complications as the lumbar approach.  In addition, there is also the risk of direct trauma to the cord by the needle.  Contraindications to the procedure include chiari malformation with low-lying cerebellar tonsils or a low course of the posterior inferior cerebellar artery (PICA).  If the CSF cistern posterior to the cord at C1-2 is too small for safe puncture, a lumbar approach should be used.  If previous MR or myelograms are available they should be studied to exclude the presence of these entities.  When filling out the consent form, remember to include the risks that are unique to cervical myelography.

Ankle braces or a body harness should be used to avoid the patient slipping off the table.

There are two approaches for cervical puncture, the horizontal needle approach and the vertical needle approach.  In the horizontal needle approach the patient is positioned prone on the table, whereas for the vertical needle approach the patient is positioned in the lateral decubitus position.  Both involve placement of the needle at the C1-C2 level. 

In the horizontal needle approach, the patient is placed in the prone position with the neck positioned such that the contrast introduced into the C1-C2 area will pool in the cervical region (i.e. so that it will not run into the head or down into the thoracic spine).  This does NOT usually require extending the neck, the normal lordosis of the neck in neutral position usually suffices.  The important aspect is to tilt the table so that the cervical spine is overall horizontal (between C1 and C7) with the lowest point in the mid‑cervical region.

Use of a head-extending device is usually NOT necessary (and not desirable).  When the neck is extended the cord is moved into the dorsal spinal canal where the needle will be placed.  The extended neck position also closes the C1-C2 interlamina space where the needle is to be placed.  Finally, extending the neck accentuates stenosis; this could potentially injure the cord in a patient with severe stenosis or a large disk herniation.

Care is taken to position the patient such that horizontal fluoroscopy views the spine in a true lateral projection.  Line up the right and left external auditory canals in the head and the right and left sides of the arch of C1 in the cervical spine.

The horizontal needle is placed between the pedicles of C1 and C2 at the junction of the ventral 2/3 and the dorsal 1/3 of the spinal canal.  Lateral fluoroscopy is used.  The needle is positioned so that a "down the barrel" view is projected directly over the target point.  Estimate the depth from skin to thecal sac (the needle can be placed over the patient's neck for direct estimation).  Insert the needle.  Check the depth of the needle tip by AP fluoroscopy.  A characteristic "pop" is felt when entering the dura.  Do not rotate the needle to clear it from the dura, as is commonly done in the lumbar region.  If the needle happens to be in the cord this maneuver would increase the damage.

Common mistakes include:  Placing the needle too far posteriorly; this can cause one to pass posterior to the thecal sac.  Another common mistake is to not advance the needle slightly after the dura is entered; this may lead to a mixed epidural/subdural/subarachnoid injection because the needle does not completely clear the dura.

Either 10 cc of 300 (or 17 cc of 180 contrast) can be used.  Verify the location of the needle in the thecal sac by fluoroscopy and by ample CSF return before injecting contrast.  Do not inject contrast if there is any chance of the needle being in the cord.  The first injection should be a small puff viewed under lateral fluoroscopy.  The contrast should be seen to cascade down away from the needle tip, over and around the cord, and into the ventral side of the thecal sac.  Epidural and subdural injections are identified by contrast pooling around the tip of the needle.  During the injection, care is taken to avoid contrast running into the head (which may cause a seizure).  Periodically fluoroscopic observation is used throughout the injection to verify that the needle tip does not migrate and cause a subdural or epidural injection.

As contrast begins to accumulate in the cervical region the patients usually experience a crampy pain in the neck, across the back and down the arms. This is probably related to direct irritation from the contrast.  The patient will want to move his/her neck to relieve the cramp.  If you are satisfied that the pain is due to this and not some other cause, tell the patient that this is a common but transient effect.   It usually lasts about 10-15 minutes.  Advise that changing positions will not help and that soon the pain will lessen on its own.  Ask the patient to remain still for the filming.

Filming is as described in the lumbar approach/cervical myelogram section.  The plain film images should be acquired expeditiously while the contrast is concentrated in the cervical region.

The vertical needle approach for cervical myelography involves placement of the needle at C1-C2 under vertical fluoroscopy.  This approach is most useful when horizontal beam fluoroscopy is not available, such as when using a "GI" x-ray table.  The patient is positioned in the lateral position.  The needle is again parallel to the fluoroscopy beam, but both are vertical, rather than horizontal as above.  The contrast can be injected with the patient in the lateral position, then the patient is rolled prone with head extended after the needle has been removed.  Alternately, the patient can be rolled prone with the needle in place, then the contrast can be injected (care must be exercised to avoid cord damage from the needle during patient repositioning).

The dictation must include as the last line of the procedure section, "Dr. Smith, the attending neuroradiologist, was present throughout the procedure".


Thoracic myelography is relatively rare compared to lumbar and cervical myelography. 

A critical aspect of the thoracic examination is documentation of location.  Before beginning the examination, document the number of ribs so that later the location of any lesions can be unambiguously determined by counting ribs either from the top or the bottom.

Contrast is placed in the thecal sac, usually with a lumbar approach.  In the rare instance of a complete spinal block in the thoracic region, it may be necessary to also place contrast above the block by the cervical approach to delineate the superior extent of the block.

The patient is then placed supine with the head elevated on a pillow to allow the contrast to pool in the thoracic region.

Images are obtained.  It is critical that each image allows unambiguous determination of the rib levels.  If it is not possible to include either the top or bottom rib on the image, it may be necessary to use a radiodense pointer to point to a particular rib in the image.

The CT portion of the examination is usually done with contiguous 5-mm-thick images from T1 through T12.  If pathology is seen on the plane films or on another examination, obtain 2-mm images over the area of pathology.

Total Spine (not for AVM)

The CT is usually done as:
Routine Cervical
5 mm contiguous slices through the T spine
Stenosis L spine (i.e., also 5-mm contiguous slices).

Total Spine Myelogram to Assess for AVM

If possible, determine the anatomic level of clinical concern, or at least an uppermost level of concern.  Statistically, the most common location of a spinal AVM is in the lower thoracic region.  However, they can occur anywhere in the thoracic region.  They are distinctly rare in the cervical region.  If no clinical level can be determined, concentrate the examination on the lower thoracic region.

Theory:  Speed is of the essence in the myelogram in order to keep the contrast bolus together.  Obtain wide field of view AP and Lateral films only. Do not obtain obliques as one would for a routine myelogram.  Obtain Lateral views of the lumbar/conus region with the patient both supine and prone so that the contrast layers are dependent in both the ventral and dorsal canal.  In order to obtain a lateral of the thoracic spine with the patient supine it may be necessary to build a spacer cushion between the table and the patient (Important: do this before the contrast is administered).  By the time the contrast is in the thoracic region it should be mixed enough that both supine and prone lateral views are not necessary.

Procedure:  Use a lumbar approach to administer the contrast (10 cc of 300).  Immediately remove the needle completely so that the patient is free to roll supine.  Do not let the contrast pool in the sacrum, try to keep it in the mid and upper lumbar region.  Obtain AP and Lateral plain films of the lumbar/conus region.  Turn the patient supine and obtain another Lateral of the lumbar/conus region.  Keep the patient supine and manipulate the contrast into the thoracic region.  Obtain AP and Lateral plain films of the thoracic region.  Tip the patient slightly head down for views of the upper thoracic region, including a swimmer's view.  Then turn the patient prone with neck extended while manipulating the contrast into the cervical region.  The contrast bolus might be too diffuse by this time for good plain films of the cervical spine.

Send the patient to CT.  Obtain CT scans from the foramen magnum to the sacrum (5 mm thick through, 8 mm in cervical regions, contiguous).

Summary: Non-magnified images for AVM of Thoracic spine:


1. AP/Lat upper lumbar/conus


1.  Lat upper lumbar/conus
2.  AP/Lat mid lumbar
3.  AP/Lat/Swimmer upper lumbar


AP/Lat cervical

Have the patient lie supine while waiting for the CT so that the contrast pools in the thoracic region.

CT (Supine)

8 mm cervical, contiguous
5 mm thoracic/lumbar, contiguous
2 mm over areas of concern, contiguous


Lumbar Puncture under Fluoroscopy

If an LP can not be performed by the clinicians, the patient can be sent for LP under fluoroscopy.  Before beginning, check whether the referring clinician has special instructions for the CSF analysis so that it is collected and handled in the proper manner.  CSF is obtained.  If access is not possible in the lumbar spine, a cervical approach may be necessary.

In order to ensure that the proper laboratory tests are ordered on the CSF sample, the collected CSF is given to the referring clinician for disposition.

LP for Meningitis:  It is a good idea to start with a 20-gauge needle.  In frank meningitis the purulent CSF may be too viscous to flow through a 22-gauge needle.

LP for Multiple Sclerosis:  When collecting CSF to examine for the diagnosis of MS (usually cell count and differential, total protein, gamma globulin, and oligoclonal bands), a minimum of 8 cc of CSF must be collected.

LP for cells:  CSF is often collected to diagnose cells in a malignancy.  If there is a large intrathecal or intramedullary mass, the removal of CSF below the mass might cause it to herniate downward.  Previous images are helpful to assess for risk of herniation.  One might consider a C1-2 or cisterna magna tap if there is a great risk of herniation of a spinal lesion from lumbar puncture.

The dictation needs to explicitly state the staff member's name and that he or she was present for the procedure.

CSF Opening Pressure

   A true CSF opening pressure must be measured with the patient in the lateral position so that the entire CSF column is at the same horizontal level.  This horizontal position of the CSF in the spine is not possible in the prone position due to the normal cervical and lumbar lordosis and thoracic kyphosis.

Set up the manometer before starting the procedure.  Have the stopcock valve set to be open from the needle attachment port to the manometer.

Place the patient in the lateral position and place the spinal needle at L23 with fluoro using a midsagittal approach (to keep the needle level with the CSF).  When in the thecal sac, attach the stopcock to the needle.  Wait for the CSF to stop rising in the manometer.  This takes a few breaths of the patient.  Measure the pressure in cm of water (i.e. not mm Hg).  The pressure varies with the cardiac cycle; use the maximum pressure over the cycle.  Abnormal is more than 18 cm water. 

If CSF is to be collected, open the stopcock to allow the CSF in the manometer to flow into the test-tube.

(An alternate, as done by Dr. Hodges.  From practical experience he has found that using the level of the third ventricle as the zero point of the manometer, a reasonable estimate of the opening pressure can be obtained regardless of the patient's position.  This is useful in patients who cannot be placed in the lateral decubitus position.  Attach the manometer to the needle using the long connecting tube and place the bottom of the manometer at the level of the third ventricle.  The position of the third ventricle can be approximated by the superior most point where the ear connects to the head [where eyeglasses rest].)

Include the opening pressure in your dictation.  The dictation needs to explicitly state the staff member's name and that he or she was present for the procedure.

CT Cisternogram

A CT cisternogram involves injection of intrathecal contrast followed by head CT.  This is used for evaluation of arachnoid cysts and for investigation of CSF leaks, particularly into the nasal cavity after trauma or surgery.

In the case of CSF leak, the CT cisternogram is used to locate the source of a known CSF leak prior to surgery.  A radionuclide study (below) is usually used to determine whether a leak is present.  Slow leaks might not be identified on the CT study.

Contrast is administered into the lumbar spine or less often the cervical spine as described above.  Because a large amount of contrast in the head may produce seizures, only 5-7 cc of 180 nonionic contrast (Omnipaque [Iohexol]) are administered.

To look for arachnoid cysts one wants the contrast mixed well with CSF.  The head is tipped down to allow contrast to flow into the head.  Immediate and 1 hour delayed CT allows for assessment of the amount of communication between the cyst and the CSF.

To investigate CSF leaks into the sinuses, the contrast is best not mixed with CSF.  Obtain pre contrast CT of the sinuses (2 mm axial and coronal).  Have the patient face down and head down for contrast to pool near the sinuses.  Obtain an immediate post contrast CT (axial and coronal).  If this is negative, obtain a 30 minute delayed CT.  If you obtain a delayed image much later than 30 minutes the contrast will be too dilute to see well.

The dictation needs to explicitly state the staff member's name and that he or she was present for the procedure.

Radionuclide Cisternogram

This study is used to look for CSF leaks into the nasal cavity and to assess the CSF flow in normal pressure hydrocephalus. 

In the case of CSF leak, the radionuclide cisternogram is used to determine whether a CSF leak is present.  It is quite sensitive to slow leaks.  However, it can not determine the exact location of the leak.  A CT cisternogram (above) is used for preoperative planning to identify the source of a known leak.

The technologists usually order and pick up the radionuclide from Nuclear Medicine (Hot Lab 362-2799).  Before starting the procedure, verify that the radionuclide is present and ready for injection.

Obtain the patient's consent for the LP.  It is also a good idea to obtain consent for contrast administration.  (Occasionally it is useful to inject a little contrast prior to administering the radionuclide to verify that you are indeed in the thecal sac.)

A needle is placed into the lumbar thecal sac. 

In the case of R/O CSF leak, CSF is usually collected because of the risk of meningitis. 

In the case of normal pressure hydrocephalus, NPH, ask the clinician ahead of time if they would like an opening pressure, or just obtain the measurement.

The radionuclide for injection (0.5 mCi of 111-indium DTPA) is sent from Nuclear Medicine in a lead container.  Keep it in the lead container until you are ready to inject it.  Nuclear Medicine says that you need not wear special leaded gloves while holding the syringe during the injection, just wear the usual sterile gloves and then discard them.  Note that the syringe containing the radiopharmaceutical is not sterile.

All items that touched the radiopharmaceutical need to be disposed of in a special radioactive disposal container (syringe, needle, etc).

The dictation needs to explicitly state the staff member's name and that he or she was present for the procedure.

On the rare occasion that a CT cisternogram and a radionuclide cisternogram are done simultaneously, nuclear medicine prefers that the radiopharmaceutical is injected before the iodinated contrast so that it has a little longer to circulate before scanning.  After injection, obtain the post injection CT, then send the patient to Nuclear Medicine for the first scan.  Then have the patient return to Radiology for a delayed CT, if necessary.

Chemotherapy Injection

Intrathecal chemotherapy is often administered, especially for intracranial lymphoma and leukemia.  These patients are often thrombocytopenic, so check the coagulation factors and the platelet count before proceeding.  A platelet count below 50,000 requires a platelet transfusion before LP.  Check before the procedure that the technologists ordered (362-1518) and picked up the chemotherapeutic agent.

Discuss the risks of LP in addition to the risks of chemotherapy injection to the patient before the procedure.  As with administration of any drug, the risk of chemotherapy reaction needs to be explained in the consent process.  It is also a good idea to obtain consent for contrast administration.  (Occasionally it is useful to inject a little contrast prior to administering the chemotherapeutic agent to verify that you have indeed accessed the thecal sac.)

The lumbar approach is most common.  Sometimes, after many doses, there are adhesions and sclerosis in the lumbar thecal sac, making access difficult.  Delivery of chemotherapy should probably not be done via a cervical approach.  An Ommaya reservoir may need to be placed by Neurosurgery if there is no access to the thecal sac via LP.

Occasionally the clinician requests CSF collection for laboratory analysis.  This is usually cell count (at least 1 cc in one tube) and cyto-spin (at least 4 cc in a second tube).

Wear safety goggles while handling and administering the chemotherapeutic agent.  A protective gown may also be appropriate.

The outside (and inside) of the syringe containing the chemotherapeutic agent is sterile.  It is delivered within a special container to maintain sterility.  Inject the agent at a moderate rate; there are no special guidelines for the rate of injection.  Insert the stylet before removing the needle so that no chemotherapeutic agent is deposited in the soft tissues during needle removal. 

All items that touched the chemotherapeutic agent need to be disposed of in a special chemotherapy disposal container (syringe, needle, gloves, etc).

If the patient is an outpatient, he or she can leave immediately after the injection.  No observation period is necessary.

The dictation needs to explicitly state the staff member's name and that he or she was present for the procedure.


Residents should be maintaining a procedure book in which all procedures are logged.  The entry should include: date, patient name, procedure, and staff.  It is necessary to document procedures performed for future board certification and hospital accreditation.  The myelogram entries should include the site of puncture (cervical or lumbar) and the region(s) imaged.  Other procedures to be put in the book include angiograms, drainage procedures, biopsies, arthrograms, etc. 

Fellows should maintain a procedure book until after passing the CAQ examination.

Note:  Medicine is an ever-changing science.  Although the author has made every effort to ensure the accuracy of the information in this article, readers are encouraged to confirm the information herein with other sources.  Drug and procedural information are designed as guidelines.  Nuances in performing safe procedures depend on the operator's experience and the patient's condition.  Certain applications described herein may not be appropriate for a given patient situation.  Moreover, readers are advised to check product information available in the package insert for specific drugs and to tailor usage for each individual patient cognizant that dosages, indications and contraindications may change.


Appendix:  Instructions for Sending Patients to the Medical Procedures Area (MPA)

Dear colleagues:                                                           March 1997

We have completed the administrative preparation for the use of the MPA observation unit for myelogram patients.  This arrangement is provided in order to facilitate the workup and performance of outpatient myelograms.  The MPA unit provides nursing care from 8 AM to 5 PM.  We will begin scheduling the first two myelogram patients per day for this service.  This will enable us to discharge the patients by 5 PM.  Patients scheduled for myelography later in the day will use the traditional 23-hr. admission option that we have been currently using.

Therefore, the staff and fellows will be responsible for the following:

1.  Performing a brief history and physical note on progress note paper prior to the myelogram.  This is to be a very brief statement of chief complaint and a review of systems.  A detailed examination is not necessary.

2.  The post-procedural orders and post-procedural note should be completed prior to discharging the patient to the MPA unit. 

3.  A discharge order must be written in person at the end of the afternoon provided the patient is well enough to leave the hospital.  A post-procedural instruction sheet should also be completed and given to the patient.  Patients who are too ill to be discharged must be converted to a 23-hr. admission by discussing the situation with the referring doctor. 

4.  Staff and fellows must be within beeper range to deal with any situations that arise while the patient is in the unit.

5. Remember, you are the patient's primary physician during the myelogram outpatient encounter.  This will hopefully enable us to expedite the examinations. 

Please contact me if you have any questions.

Jay Wippold


From: KRUZICH, STEPHEN                                   05/29/97    9:22 AM

     To: BH Neuro Technologists________      cc: @ZBNRMD_______________________

Subject: MPA PATIENT_____________________________




MPA 362-0813




Headaches occur in 1% to 30% of patients following dural punctures (Choi, 1996).  Symptoms vary from mild to incapacitating and may persist for days or even weeks.  Although this usually self-limiting complication may be considered merely an annoying nuisance to the busy physician, the post-dural-puncture headache (PDPHA) comprises a significant aspect of patient care for those neuroradiologists performing myelograms and fluoroscopically guided diagnostic lumbar punctures.

RECOGNIZING THE PDPHA – Prior to treating PDPHA, the diagnosis must be secured and differentiated from mimics such as meningitis (usually heralded by fever, stiff neck, and a nonpostural headache).  Key aspects of the PDPHA include (Hardman, 1996; Leibold, 1993; Weakland, 1994);

1.      Severe postural headache aggravated by Valsalva maneuver.

-         worse when patient sitting up; better when lying down.

2.      Headache affecting any part of the cranium, but typically located in the frontal or occipital region; rarely unilateral.

-         pressure, pounding, worse on movement of head.

3.   Onset usually within 24 hours of dural tap, but may be delayed.

-         90% of patients develop PDPHA within 3 days.

-         25% resolve within 7 days.

4.   Associated findings:

Visual or auditory symptoms.
Nausea, vomiting.
Neck stiffness.
Visual symptoms such as photophobia, blurred vision, diplopia (10%).
Hearing loss.
Occasionally cranial nerve palsies due to traction.

MIMICKERS OF PDPHA – The positional nature of PDPHA provides the key to diagnosis.  Nevertheless, other headache syndromes can mimic PDPHA.  Clinical differential diagnoses include:

1.      Meningitis – fever, patient usually quite ill.  Assuming sterile technique for the original spinal puncture, this complication is quite rare.  Nevertheless, any patient with a question of infection should be referred to the Emergency Department immediately.

2.      Pneumocephalus – not positional.

3.      Cortical vein thrombosis – not positional; patient usually quite ill.

4.      Migraine – not positional, may have prodrome.

5.      Hypoglycemia – not positional.

6.      Dehydration – not positional.

7.      Fatigue – not positional.

8.      Subarachnoid hemorrhage – not positional; patient usually quite ill.


MECHANISMS OF THE PDPHA – The currently held theory for the PDPHA states that the tear in the dura left by the spinal needle produces a leak of cerebrospinal fluid (CSF).  The leak causes intracranial hypotension, reflex arterial dilatation, and resultant headache.  Intracranial hypotension also causes loss of intracranial buoyancy and traction on pain-sensitive structures such as the basal meninges (Hardman, 1996; Weakland, 1994).  In its most severe form, intracranial hypotension has been linked with development of subdural hematoma (Tekkok, 1996).

            Research has revealed that younger patients are at higher risk, possibly due to differences in meningeal thickness or compliance compared to older adults (Leibold, 1993).  Women are affected twice as commonly as men.  Development of PDPHA is inversely related to needle size.  Therefore, smaller needles, such as 25G, theoretically should be associated with fewer headaches compared to 18G or 20G needles.  Special needles, such as the pencil point needle, may reduce the incidence of PDPHA; however, benefits must be weighed against added cost.  Needles puncturing the dura at an oblique angle to the long axis of the spine cause almost 10% fewer headaches than punctures perpendicular to the long axis of the spine.  Apparently, this approach tunnels into the dura and allows for the edges of the dural hole to close more effectively than if the dura is cleanly cut.  Orienting the bevel of the needle parallel to the longitudinally arranged dural fibers may also diminish likelihood of PDPHA.  Hypovolemia and dehydration prior to puncture increase likelihood of PDPHA.  Interestingly, operator skill has not been shown to correlate with PDPHA (Leibold, 1993).  Finally, traumatic taps seem to produce fewer PDPHAs than nontraumatic taps, possibly due to “self-administered” blood patches.


TREATMENT OF THE PDPHA – Conservative measures include:

1.      Bedrest – symptoms improve with the patient in the horizontal position.  This position may also relieve hydrostatic pressure around the leak.

2.      Hydration.

3.      Analgesics – Over-the-counter analgesics usually adequately control pain.

Analgesics should not substitute bedrest and horizontal positioning.

a.       Aspirin, acetaminophen (Tylenol), or ibuprofen (Advil, Motrin) following dosage recommendations on the label.

NOTE: Aspirin 650 mg po qid, acetaminophen (Tylenol) 650 mg po qid (not to exceed 3900 mg/day) , or ibuprofen (Advil, Motrin) 200-300 mg po q4-6 hrs. (not to exceed 3200 mg/day) is usually adequate.   Some preparations of ibuprofen offer 800 mg tablets.  Cautions related to anticoagulation, gastrointestinal upset, prior allergy, and aspirin sensitivity apply.

b.      Narcotics should be avoided.

4.      Caffeine 300 mg po – This often-overlooked drug found in coffee and cola drinks offers excellent short-term relief by causing mild vasoconstriction.  Caution with rebound phenomena pertains.  Heavy coffee drinkers should be encouraged to maintain their usual intake to avoid withdrawal headache, which can exacerbate PDPHA.  Caffeine can be administered as caffeine sodium benzoate 500 mg IV q8h in severe cases.  Caution regarding coronary status and seizures applies.

Caffeine contents of common foods
Coffee, 5-8 oz

Regular brewed                 40 – 180 mg
Instant                               30 – 120

Soft drinks                           20 - 110

NOTE: Some over-the-counter preparations such as Excedrin Extra strength (Acetaminophen 250 mg, aspirin 250 mg and caffeine 65 mg per tables; average dose 2 tablets po q6h not to exceed 8 tablets in 24 hrs.) offer analgesia and caffeine

5.      Epidural blood patch – See discussion below.

EPIDURAL BLOOD PATCH – Epidural blood patch (EBP) is the preferred therapy for sealing the dural leak and attacking the cause of PDPHA.

What is EBP? – EBP consists of delivering autologous blood into the epidural space.

How successful is EBP? – EBP is typically 85%-90% effective after the first patch and 95%-98% successful after a second patch, if a second EBP is necessary (Choi, 1996; Hardman, 1996).  Relief may occur instantly.

When is the best time to administer EBP? – EBP is 96% successful when administered after 24 hours following lumbar puncture (Weakland, 1994).  Interestingly, EBP is only 24% successful if performed within 24 hours of puncture (Hardman, 1996).  Prophylactic EBP has not been shown effective.

How does EBP work? – The short-term effect may be due to tamponade of the epidural space, which temporarily elevates intracranial pressure.  This reinstates brain buoyancy and often provides instant relief (Weakland, 1994; Carrie, 1991).  Magnetic resonance imaging has demonstrated the initial mass effect of the clot.  Long-term relief probably occurs because the clot occludes and fibroses the hole in the dura.

How does one perform EBP? – Following the diagnosis of PDPHA, counseling and advised consent, begin an intravenous line for hydration with normal saline (20-30 ml/hr).  Attach a three-way stopcock to the line.

            An 18G spinal needle is introduced into the ligamentum flavum of the chosen lumbar level.  Blood used for the EBP tends to track cranially, as demonstrated in tagged red cell studies; therefore, the EBP needle should be placed at the site or one interspace inferior to the original tap site (Hardman, 1996; Taivainen, 1993).  If the original tap was paramedian, the needle should be placed on the same side of the midline as the tap.  Fluoroscopy may be helpful for visually positioning the needle.

            A 5-ml syringe is filled with approximately 2 ml of normal saline.  This 5-ml syringe is then attached to the spinal needle hub.  The spinal needle is gently advanced while applying gentle pressure on the syringe piston.  When the epidural space is entered, resistance to the piston markedly diminishes.  Withdraw the piston slightly to exclude the possibility of entering the subarachnoid space.  If CSF freely flows (indicating an intrathecal needle), attempt a second tap at a lower level (Miller, 1994).

            Once the epidural space has been identified, 20 ml of blood is withdrawn from the intravenous line using the three-way stopcock; care must be taken to avoid sample dilution by the normal saline.  It may be necessary to discard the first few saline diluted mls.  (Because withdrawing venous blood prior to the epidural tap risks having a clot clog the syringe, we recommend that the venous blood be obtained after the epidural space is identified.)  Inject the blood slowly into the epidural needle.  The optimal volume of blood for injection is controversial; however, based on the current literature, the recommended volume should be approximately 15 ml (Taivainen, 1993; Hardman, 1996).  Volumes less than 10 ml are often ineffective and volumes greater than 15-20 ml have a higher incidence of complications such as backache (Weakland, 1994).  Document the procedure in the patient's outpatient record.

Is blood the best medium? – Substitutes of normal saline, crystalloid, and dextran have not been shown to be as effective as blood and even carry a higher rate of adverse effect.

When is EBP contraindicated? – EBP is contraindicated in the following situations:
1.      Coagulopathy – Risk of epidural bleeding with needle placement; lack of clot formation with instilled blood.

2.      HIV positive status – Risk of epidural abscess.

3.      Meningitis – Risk of epidural abscess.

4.      Septicemia – Risk of epidural abscess

5.      Local skin infection – Risk of epidural abscess.

6.      CNS or marked nerve root dysfunction – The temporary rise in intracranial and/or epidural pressure with epidural tamponade could exacerbate intracranial or peripheral nerve root conditions.  Most of the conditions for which patients are referred for myelography should be eligible for EBP; however, patients with serious motor impairment or bowel or bladder dysfunction should only be referred for EBP after consultation with the clinician.

What are the complications of EBP?  Blood extravasating into the subcutaneous tissues of the back may contribute to the backache (Carrie, 1993).  This mild transient complication occurs in 35% of patients and usually occurs with higher volumes of injected blood.  Leg pain or paresthesia is rare (1%) and should be investigated with MR imaging.  Bradycardia and mild hyperpyrexia are usually transient.  Facial nerve palsy has been reported and may be due to increased intracranial pressure and compromise of blood flow to the nerve.  Arachnoiditis is rare due to the small amount of blood injected (Choi, 1996; Weakland, 1994).

What if the EBP is ineffective? -  A second patch may be considered.  At this point, consultation with the referring clinician is advised.  CT or MR imaging of the brain or spine or other evaluation may be necessary for excluding:
1.      Meningitis – fever, child, stiff neck

2.      Intracranial hypertension – papilledema

3.      Intracranial mass lesion (hematoma, tumor) – focal signs.

4.      Cortical vein thrombosis.

NOTE:  Medicine is an ever-changing science.  Although the author has made every effort to ensure the accuracy of the information in this article, readers are encouraged to confirm the information herein with other sources.  Drug and procedural information is designed as guidelines.  Nuances in performing safe procedures depend on the operator's experience and the patient's condition.  Certain applications described herein may not be appropriate for a given patient situation.  Moreover, readers are advised to check product information available in the package insert for specific drugs and to tailor drug usage for each individual patient cognizant that dosages, indications, and contraindications may change.


Braune H-J, Huffman G.  A prospective double-blind clinical trial, comparing the sharp Quincke needle (22G) with an “atraumatic” needle (22G) in the induction of post-lumbar puncture headache.  Acta Neurol Scand 1992; 86:50-54.

Camann WR, Murray RS, Mushlin PS, Lambert DH.  Effects of oral caffeine on postdural puncture headache.  Anesth Analog 1990: 70;181-184.

Carrie LES.  Epidural blood patch: Why the rapid response?  Anesth Analg 1991; 72:129-30.

Carrie LES.  Postdural puncture headache and extradural blood patch.   Br J Anaesthesia 1993; 71:179-181.

Choi A, Laurito CE, Cunningham FE.  Pharmacologic management of postdural puncture headache.  Ann Pharmacotherapy 1996; 30:831-839.

Englehardt A, Ohiem S, Neundörfer B.  Post-lumbar puncture headache: experiences with Sprotte's atraumatic needle.  Cephalalgia 1992; 12:259.

Gerancher JC, A R, Carpenter R.  Caudal epidural blood patch for the treatment of postdural puncture headache.  Anest Analg 1998; 87:394-395.

Gurmarnik S, Kandror KV.  Postdural puncture headache: The betadine factor.  Regional Anesthesia 1996; 21:375-376.

Hardman JG.  Epidural blood patch.  Br J Hosp Med 1996; 56:268-269.

Harrison PB.  The contribution of needle size and other factors to headache following myelography.  Neuroradiology 1993; 35:487-489.

Heide W, Diener H-C.  Epidural blood patch reduces the incidence of post lumbar puncture headache.  Headache 1990; 30:280-281.

Leibold RA, Yealy DM, Coppola M, Cantees KK.  Post-dural-puncture headache:  Characteristics, management, and prevention.  Ann Emer Med 1993; 22:1863-1870.

Luzza F, Imeneo M, Maletta M, Pallone F.  Smoking, alcohol and coffee consumption, and H. pylori infection.  Br Med J 1998; 316:1019.

Parkinson D.  Reducing headaches after a myelogram.  AJR 1998; 171:266-267.

Peterman SB.  Postmyelography headache rates with Whitacre versus Quincke 22-gauge spinal needles. Radiology 1996; 200:771-778.

Peterman SB.  Postmyelography headache: A review.  Radiology 1996; 200:765-770.

Prager JM, Roychowdhury S, Gorey MT, Lowe GM, Diamond CW, Ragin A.  AJR 1996; 167:1289-1292.

Quaynor H, Tronstad A, Heldaas O.  Frequency and severity of headache after lumbar myelography using a 25-gauge pencil-point (Whitacre) spinal needle.  Neuroradiology 1995; 37:553-556.

Raskin NH.  Lumbar puncture headache: A review.  Headache 1990; 30:197-200.

Suojanen JN, Punzak ST.  Needle characteristics related to headaches after myelograms.  AJR 1997; 169:1749-1750.

Taivainen T, Pitkänen M, Tuominen M, Rosenberg PH.  Efficacy of epidural blood patch for postdural puncture headache.  Acta Anaesthesiol Scand 1992; 37:702-705.

Tekkök IH, Carter DA, Brinker R. Spinal subdural haematoma as a complication of immediate epidural blood patch.   Can J Anaesth 1996; 43:306-309.

Tourtellotte WW, Henderson WG, Tucker RP, Gilland O, Waler JE, Kokman E.  A randomized, double-blind clinical trial comparing the 22 versus 26 gauge needle in the production of the post-lumbar puncture syndrome in normal individuals.  Headache 1972; 22: 73-78.

Vilming ST, Kloster R.  Post-lumbar puncture headache: clinical features and suggestions for diagnostic criteria.  Cephalalgia 1997; 17:778-784.

Vilming ST, Kloster R.  The time course of post-lumbar puncture headache.  Cephalalgia 1998; 18:97-100.

Weakland HJ.  The epidural blood patch—current practices and concerns. CRNA: The Clinical Forum for Nurse Anesthetists 1994; 5:156-163.


1.      Determine if the headache is truly a post dural puncture headache (PDPHA).  Measure the patient's temperature and obtain the white blood cell count.  Review contraindications.

2.      Explain the procedure and potential complications.  Obtain informed consent.

3.      Begin an intravenous line for hydration with normal saline.  Attach a three-way stopcock to line.

4.      Position the patient and determine the desired vertebral level for puncture.  Aim for the site of the original puncture or one interspace inferior to the original tap site.

5.      Place an 18G needle into the ligamentum flavum.  Using a 5-ml syringe containing approximately 2 ml of normal saline, advance the needle into the epidural space while applying gentle pressure on the syringe piston.  Upon entering the epidural space, resistance markedly diminishes.  Withdraw the piston slightly to exclude the possibility of entering the subarachnoid space.  If CSF freely flows, attempt a second tap at a lower level.

6.      Withdraw 20 ml of blood from the intravenous line using the three-way stopcock.  Be sure to avoid dilution with the saline.

7.      Inject 15 ml of blood slowly into the epidural needle.

8.      Keep the patient horizontal (no bathroom privileges) for 1-2 hours while infusing normal saline (20-30 ml/hr).

9.      Instruct the patient to avoid straining or lifting for 4 to 5 days and to seek medical advice if headache symptoms recur or fever or chills develop.  A stool softener [i.e., Docusate sodium (COLASE) 50 mg po bid up to 200 mg/day for 5 days] is recommended.

10.  Document the procedure in the patient's records

* *