MRIs are pictorial representations depicting the location of fat and water in tissues, sampled as thin sections through the human body. This technology relies on the nuclear magnetic resonance phenomenon and requires that the patient be placed in a strong magnetic field. Harmless radiofrequency pulses are applied, allowing detection of the amount and the location of hydrogen-containing material (fat and water) in a person's body [3,4]. A computer processes the faint electrical signal arising from the hydrogen nuclear magnetization into images that can be viewed on a computer monitor or printed onto film. The diagnostic value of MRI comes from its ability to discriminate different kinds of soft tissue. In part, this is possible because of the variations of fat and water content in the tissues but also because of the different rates at which the nuclear magnetization grows in tissues and the rates at which the detectable signals disappear with time. These growth and decay rates are termed T1 and T2, respectively. Radiologists have coined the terms T1-weighted image, T2W image, PD image, etc., to indicate the variables used to collect the image.
A T1-weighted image results when the radiofrequency pulses used to generate imaging signals are closely spaced in time. The parameter TR is used to report this time spacing, and if it is less than 800 milliseconds, the image is said to be T1-weighted. Tissues such as fat, which recover their nuclear magnetization quickly, will appear with bright intensity on these images.
If the TR parameter is set to be long (i.e., greater than 2200 milliseconds) during image collection and the signal is collected immediately after the radiofrequency pulse, a PD image is produced .
T2W images are produced by allowing the electrical signal to decay for a short period of time before collecting it in the computer. The time between the radiofrequency pulse generating the signal and its actual detection is called TE. If this parameter is set to be greater than 60 milliseconds, a T2W image results.
There are other variations of the variables in MRI that allow the detection of flowing material, permitting magnetic resonance angiography . Gradient echo sequences permit more rapid data collection, making possible three-dimensional imaging protocols with more 100 slices, each only 1 mm thick.
The brain and the spinal cord are surrounded by 150 mL of CSF. Approximately 50 mL of CSF is formed daily, most of which is secreted by choroid plexuses in the ventricles and absorbed by arachnoidal villi in the arachnoidal granulations of the venous sinuses. The CSF pressure is maintained at an average of 10 mm Hg . The etiology of PDPH is attributed to the effect of postdurally related decreases in CSF pressure caused by the persistent leak of CSF from the dural hole [1,8,9]. The loss of CSF causes the brain to descend when the patient assumes the upright position, thus stretching the pain-sensitive meninges and resulting in a headache. In animal studies, investigators have shown that blood clots epidurally and seals the dural hole, thereby stopping the CSF leak [10,11].
In our study, CSF leak was demonstrated on MRI as focal accumulation of clear or blood-stained fluid extrathecally in four of five patients. Interestingly, in two patients, we noted blood in the posterior margin of the thecal sac demonstrated as hemosiderosis on the preblood patch MRI. One of these patients received spinal anesthesia, and the other received epidural anesthesia with accidental dural puncture. Neither of these two patients showed any evidence of bleeding while the initial block was being performed.
In 1960, Gormley  reported the injection of autologous blood for controlling PDPH. He injected only 2-3 mL of autologous blood into the epidural space and reported relief of the headache. Since then, investigators have demonstrated that higher success rates can be achieved with larger volumes of blood [13-15]. Crawford [13,14] reported a 70% success rate with 6-15 mL of blood. The success rate increased to 98% when he increased the amount of blood to 20 mL. In our study, 20 mL of blood was injected in all patients with complete resolution of symptoms and no recurrence of the headache. This volume produced significant compression of the thecal sac as it spread over the area of the previous dural puncture. Its appearance on MRI supports the theory that the blood patch works by forming a dural tamponade that occludes the needle hole. The epidural space is expanded and the subarachnoid space relatively contracted, thus increasing the CSF pressure. This mechanism accounts for the immediate resolution of headache after placement of an epidural blood patch. Two reports in the British literature by Griffiths et al.  and Beards et al.  describe a similar appearance of the epidural blood patch on MRI. In the first case report published by them, the preblood patch MRI scan using T1-weighted images did not demonstrate extrathecal CSF or hemosiderosis. Also contrary to their observation, we did not notice subarachnoid extension of blood in any of our patients. We did, however, find extension of the blood anteriorly in one patient. Their second study examined the MRI appearance of the blood patch up to 18 hours after patching . An MRI scan was not performed prior to the placement of the blood patch; therefore, no CSF leak was demonstrated.
Postblood patch MRI demonstrated that the mean spread of the blood patch in the epidural space was 4.6 +/- 0.9 intervertebral levels. Most of the blood spread in the cephalad direction. This cephalad spread correlates with the findings in three prior studies, which described the spread of the blood patch using technetium-labeled red cells or MRI [15-17]. CSF flow studies, when positive, aid in better defining thecal compression, as shown in our study. Investigators have previously reported radicular pain and nerve root compression after the placement of the blood patch [18-20]. Even though there was MRI evidence of anterior displacement of descending nerves, none of our patients developed radicular symptoms during or after the injection of the blood patch.
Our study clearly demonstrates the tamponade effect of the 20-mL epidural blood patch, which we believe is responsible for the immediate resolution of PDPH. This was consistently shown by MRI studies. Due to cost and convenience, the use of MRI may not become a routine part of clinical management of PDPH. However, we believe that MRI and CSF flow studies can be used effectively as noninvasive tests to detect the site of CSF leak and to document the accurate placement of the blood patch.
The authors gratefully acknowledge Mary Corbett for her valuable contribution toward preparing this manuscript and Nicholas Szeverenyi, MD, Associate Professor of Radiology and Director of the MRI Laboratory, SUNY Health Science Center at Syracuse, NY, for the description of different MRI scans.
1. Olsen SK. Epidural blood patch in the treatment of post-lumbar puncture headache. Pain 1987;30:292-301.
2. Morewood GH. A rational approach to the cause, prevention and treatment of postdural puncture headache. Can Med Assoc J 1993;149(8):1087-93.
3. Bushong SC. Magnetic resonance imaging, physical and biological principles. St. Louis: Mosby, 1988.
4. Partain L, Price R, Patton J, et al. Magnetic resonance imaging. 2nd ed. Philadelphia: WB Saunders, 1988.
5. Kwong K. Functional magnetic resonance imaging with echo planar imaging. Magn Reson Q 1995;11:1-20.
7. Guyton AC, Hall AE. Textbook of medical physiology. 9th ed. Philadelphia: WB Saunders, 1996.
8. Kadrie H, Driedger AA, McInnis W. Persistent dural cerebrospinal fluid leak shown by retrograde radionuclide myelography: case report. J Nucl Med 1976;17(9):797-9.
9. Lieberman LM, Tourtellotte WW, Newkirk TA. Prolonged postlumbar puncture cerebrospinal fluid leakage from lumbar subarachnoid space demonstrated by radioisotope myelography. Neurology 1971;21(9):925-9.
10. Rosenberg PH, Heavner JE. In vitro study of the effect of epidural blood patch on leakage through a dural puncture. Anesth Analg 1985;64(5):501-4.
11. DiGiovanni AJ, Gilbert MW, Wahle WM. Epidural injection of autologous blood for post-lumbar puncture headache. II. Additional clinical experiences and laboratory investigation. Anesth Analg 1972;51(2):226-32.
12. Gormley JB. Treatment of post-spinal headache. Anesthesiology 1960;21:565-6.
13. Crawford JS. Experiences with epidural blood patch. Anaesthesia 1980;35:513-5.
14. Crawford JS. Epidural blood patch. Anaesthesia 1985;40:381.
15. Szeinfeld M, Ihmeidan IH, Moser MM, et al. Epidural blood patch: evaluation of the volume and spread of blood injected into the epidural space. Anesthesiology 1986;64(6):820-2.
16. Griffiths AG, Beards SC, Jackson A, Horsman EL. Visualization of extradural blood patch for post-lumbar puncture headache by magnetic resonance imaging. Br J Anaesth 1993;70(2):223-5.
17. Beards SC, Jackson P, Griffiths AG, Horsman EL. Magnetic resonance imaging of extradural blood patches: appearances from 30 minutes to 18 hours. Br J Anaesth 1993;71(2):182-8.
18. Heyman HJ, Salem MR, Klimov I. Persistent sixth cranial nerve paresis following blood patch for postdural puncture headache. Anesth Analg 1982;61(11):948-9.
19. Abouleish E, Vega S, Blendinger I, Tio TO. Long term follow-up of epidural blood patch. Anesth Analg 1975;54(4):459-63.
© 1997 International Anesthesia Research Society
20. Cornwall RD, Dolan WM. Radicular back pain after lumbar epidural blood patch. Anesthesiology 1975;43(6):692-3.