| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Clinical Observations |
1 Instituto Neurológico de Antioquia and Universidad de Antioquia,
Neuroradiology Department, Calle 55 No. 46–36 Medellín,
Antioquia, Colombia.
2 Division of Neuroradiology, Department of Radiology, University of North
Carolina, School of Medicine, Chapel Hill, NC.
Received July 24, 2005;
accepted after revision July 31, 2006.
Address correspondence to E. Medina
(elcy.medina{at}gmail.com).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CONCLUSION. MR cisternography was useful for evaluating patients with rhinorrhea and suspected CSF fistula. It depicted the fistula site in most patients. No adverse effects were found in any patient.
Keywords: CSF fistula • intrathecal Gd-DTPA • MR cisternography
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Accurate diagnosis and localization of a dural defect often involve multiple imaging studies [1, 3]. Noninvasive techniques include conventional CT and unenhanced MR cisternography in which fat-suppressed heavily T2-weighted images are obtained [1–7]. Invasive diagnostic techniques performed during active leaking are performed with contrast media or radiotracers for labeling the CSF [1, 2, 8–13]. The purpose of this study was to evaluate the utility of MR cisternography with intrathecal gadolinium enhancement in patients with suspected nasoethmoidal CSF fistula.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We used a 1.5-T MRI unit and positioned the patients prone in a Waters-like position. Unenhanced sequences that included the brain and sinonasal cavities were obtained before intrathecal injection of gadopentetate dimeglumine. The sequences were as follows: coronal fat-suppressed T1-weighted images (TR/TE, 1,050/12; number of excitations, 2; section thickness, 3 mm; matrix size, 168 x 256), axial fat-suppressed T1-weighted images (550/12; number of excitations, 3; section thickness, 3 mm), and axial fat-suppressed T2-weighted images (4,500/105; number of excitations, 1; section thickness, 5 mm). We then performed lumbar puncture at the L4–5 or L5–S1 space with a 23-gauge needle and introduced 1 mL (0.05 mmol, 469.01 mg) of gadopentetate dimeglumine (Magnevist, Schering) into the subarachnoid space at a rate of 0.03 mL/s.
Patients were placed in Trendelenburg position for 40 minutes to achieve satisfactory concentration of the small volume of contrast medium in the subarachnoid intracranial space. After this period, with the patient prone in a Waters-like position, image acquisition was repeated with the protocol used before contrast administration. If contrast medium was visualized in the ethmoidal or sphenoidal air cells, the study was finished. If not, the patient was asked to wait for 30 minutes, and imaging was repeated. Exact site was defined as the place where the dural defect was found and thus the site through which the subarachnoid space communicated with the sinonasal cavities.
|
|
|
|
|
MR images were evaluated by two experienced neuroradiologists who compared the images obtained before with those obtained after intrathecal administration of gadopentetate dimeglumine. The neuroradiologists were blinded to the surgical findings. When their opinions differed, agreement was achieved by consensus.
Because our department is a referral center, patients were initially evaluated at other hospitals according to protocols particular to those institutions. The techniques included one or more of the following: CT, including high-resolution studies; CT cisternography; radionuclide cisternography; and MRI without and with contrast enhancement. Because of this variability, we did not attempt to compare those individual studies. Because the patients had already undergone multiple studies at the referring institutions, additional studies were not performed at our hospital. The only study performed on all patients was contrast-enhanced MR cisternography.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
|
Within 24 hours of lumbar puncture, eight (36%) of the 22 patients had headaches attributed to low pressure. The symptoms were self-limited in all patients, stopping after the patient lay flat and took minor analgesics for 48 hours. No neurologic, cognitive, behavioral, or hemodynamic changes were detected during or after the procedure. No allergic reactions to gadopentetate dimeglumine occurred. Neurologic examinations and telephone interviews did not reveal adverse symptoms.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Traumatic CSF rhinorrhea occurs in 2–3% of all head injuries and is most frequent among men in the third to fifth decades of life. The onset of rhinorrhea is usually abrupt and occurs within the first 48 hours after head trauma. In most patients, the leakage resolves spontaneously without intervention within 1 week of onset [1, 3, 8, 9]. Prolonged rhinorrhea can be complicated by meningitis. The incidence of meningitis among these patients is 9–50%, and the frequency of recurrent meningitis is less than 10% [1, 8, 17–19]. Persistent leaks usually must be closed surgically. Precise localization of the defect greatly facilitates treatment.
Various combinations of planar tomography, high-resolution CT, contrast-enhanced CT cisternography, radionuclide cisternography, MR cisternography, and, more recently, MR cisternography enhanced with intrathecal administration of gadolinium have been used in the diagnosis of CSF leak. Preoperative evaluation of patients with suspected CSF leakage requires precise localization of an anatomic defect and the site of the dural tear [1, 3, 8, 15].
Radionuclide cisternography is a reliable, safe, and accepted procedure for the study of CSF flow dynamics [2, 4, 14]. The incidence of minor side effects is low, with less than 25% of patients reporting headaches, which are often related to the lumbar puncture [2]. If leakage is not occurring at the time of imaging, the ability to identify a fistula is limited. Radionuclide cisternography combined with the use of nasal pledgets helps to localize a leak in 53% of cases [20]. In an attempt to improve detection of CSF fistula, overpressure cisternography with 99mTc diethylenetriamine pentaacetic acid (DTPA) has revealed rhinorrhea in 65% of patients [2]. Indium-111-labeled DTPA radionuclide cisternography has been used to increase the sensitivity of radionucleotide cisternography by allowing imaging over 48–72 hours [2, 21]. A major disadvantage of radionuclide cisternography is its poor spatial resolution; it shows the leak but not the precise location in many cases.
High-resolution CT of the maxillofacial region performed in the axial and coronal planes in 1- to 3-mm contiguous sections with a field of view of 150–180 mm, a matrix size of 512 x 512, and a bone algorithm shows bone defects in 71% patients with CSF leaks [3]. This technique has been found approximately 71% accurate in establishing the presence or absence of CSF fistula [15, 22]. It has been suggested that this technique is all that is needed for effective visualization of the site of CSF leak and that it is a noninvasive alternative to CT cisternography in the diagnosis of CSF fistula [3, 12, 22]. When clinical and imaging findings (bone defect) coincide, further evaluation with CT cisternography and radionuclide cisternography often is unnecessary [3]. This approach eliminates the need for lumbar puncture and provides anatomic localization of the defect. Unfortunately, the specific site of a dural tear and therefore the active CSF leak cannot be confirmed with the high-resolution CT approach alone [1, 12]. This approach relies on the presence of indirect signs, such as fractures, bone defects, pneumocephalus, meningocele, cephalocele, mucous swelling, and air–fluid levels in the paranasal sinuses, to establish the presence of CSF leak without confirmation that the defect depicted is the actual cause of dural disruption [15].
The traditional method for evaluating a patient with suspected CSF rhinorrhea is a combination of thin-section high-resolution CT followed by contrast-enhanced CT cisternography. The latter requires injection of iodinated contrast medium into the intrathecal space by lumbar puncture. The sensitivity of CT cisternography for showing the exact site of dural tears has been reported to be 72–81% [9, 15]. A controversial overpressure technique has been recommended for increasing sensitivity, but it is not widely used [2, 15]. Cisternographic techniques depend on the timing of the examination and should be performed during active leakage to improve the likelihood of establishing a diagnosis. The use of CT cisternography is limited in the detection of inactive or low-flow fistulas and tiny communications [15]. In addition, a small amount of diluted contrast material leaking through a tiny defect can be difficult to differentiate from adjacent bone. CT cisternography is not without side effects, which include headache, nausea and vomiting due to the lumbar puncture, and more severe reactions, such as seizures, allergic reactions and, rarely, intracerebral hemorrhage caused by the contrast medium [23–25].
MR cisternography with fat-suppressed heavily T2-weighted 2D or 3D images without intrathecal administration of contrast medium has been used in an attempt to identify the presence of CSF leaks and has a sensitivity of 80–90% [1, 5–8, 15, 22, 26, 27]. This indirect technique eliminates the need for lumbar puncture and does not depend on active CSF leakage. It is based on the high signal intensity of CSF on T2-weighted images, which depict the area of the leak surrounded by the low-signal-intensity background of bone and air. This high signal intensity of CSF can outline bone defects in the cribriform plates and even outline brain herniation [1, 5–8, 15, 26, 27]. Diagnosis can be difficult in the presence of inflammatory changes, which also have high signal intensity on T2-weighted images. Conversely, findings suggesting CSF leak sometimes are seen in the absence of a fistula [1, 15, 27].
The safety and patient tolerance of gadolinium products administered IV have been documented [28]. Data collected in experiments on animals show that intrathecal injection of gadolinium at low doses is a safe procedure causing no clinically significant neurologic abnormalities, CSF changes, or electroencephalographic alterations [29–33]. Studies with rodents have shown that intraventricular introduction of gadopentetate dimeglumine at doses of 5–15 µmol/g brain causes severe motor disturbances associated with lesions in the spinal cord, brainstem, and thalamus. At that dose, transient coordination disturbances have been seen [32], but at doses less than 3.3 µmol/g brain, these symptoms do not occur. Although several studies [1, 29, 30, 33] have shown that it is safe and leads to excellent visualization of the site of CSF fistulas, intrathecal administration of gadopentetate dimeglumine is currently not approved in many countries [1, 15, 34].
Results of studies with human subjects have shown no clinical evidence of acute or chronic neurologic or physical abnormalities after low-dose intrathecal injection of gadopentetate dimeglumine [1, 15, 35]. The first studies [35] were conducted with a dose 30–50 times smaller than doses causing toxic changes in laboratory animals. The findings from these studies were used in calculations that showed the median lethal dose for a 70-kg person is 52.5 mmol of intrathecal gadopentetate dimeglumine [15]. In several studies [1, 15, 36], use of a single dose of 1 mL (0.5 mmol) intrathecal gadopentetate dimeglumine showed no neurologic, behavioral, or electrophysiologic alterations. Long-term (9–12 months) follow-up after the procedure revealed no delayed neurologic or behavioral abnormalities. These observations are in accordance with our findings.
In our patients, a single intrathecal dose of 0.5 mmol gadopentetate dimeglumine was distributed freely and widely in the subarachnoid space, and contrast enhancement of the CSF-containing spaces was excellent. After intrathecal injection of gadopentetate dimeglumine into the subarachnoid space, CSF spaces should show high signal intensity on T1-weighted images owing to shortening of the T1 relaxation time of the CSF hydrogen protons [15, 31]. Because of its excellent definition of the CSF spaces, MR cisternography with intrathecal injection of gadopentetate dimeglumine has high sensitivity in the detection of CSF fistula [1, 15, 36]. The degree of high signal intensity in the subarachnoid space was judged excellent in all our patients.
In our experience, owing to excellent differentiation of CSF-containing spaces and adjacent bone, brain parenchyma, and paranasal sinuses, fat-suppressed T1-weighted images obtained after intrathecal injection of gadopentetate dimeglumine have contrast resolution sufficient for detection of even subtle points of extradural contrast accumulation. Small amounts of leaking contrast medium were easily detected as areas of high signal intensity in small defects in dura and bone. No artifacts related to bone were found. Larger amounts of leaking contrast medium were detected as accumulation of contrast medium in the paranasal sinuses with the presence of fluid levels. Pachymeningeal disruption was detected as leakage of contrast medium beyond or as irregularities in the dural surfaces. In most instances, dural disruption was adjacent to the bone defect, and leakage of contrast medium was seen as a direct communication between the subarachnoid space, the bone defect, and the paranasal sinus. If dural disruption was not directly related to a bone defect and in cases of multiple fractures, contrast medium was seen in the epidural space reaching a bone defect or as an irregularity in the dural surface. In two patients with multiple dural defects, we did not find the exact site of the disruption, although clear leakage of contrast medium into the sinuses was present.
Our results represent the consensus opinion of two experienced neuroradiologists who reached a definite diagnosis in all cases. We did not have difficulty visualizing the passage of gadolinium through defects. We attribute this success to the inherent high signal intensity of gadolinium on T1-weighted images compared with the low signal intensity of adjacent fat on T1-weighted fat-suppressed images and with adjacent inflammatory changes, which tend to be isointense on T1-weighted images. We had no difficulty differentiating the low signal intensity of bone and the high signal intensity of the contrast material.
In our study, we identified 24 CSF fistulas by showing leakage of contrast medium from the anterior cranial fossa into the ethmoidal, frontal, or sphenoidal air cells. The site of leakage was confirmed surgically in 14 of these patients and was repaired with dural grafting. No leak was found in two patients, who had findings of sinonasal infection and negative results of a ß2 transferrin test. In these patients, MR cisternography was performed first, and because the findings were normal, the ß2 transferrin test was performed for further characterization of the fluid.
Mild-to-moderate (in most instances mild) and self-limited postprocedural headache was found in eight patients. Because intrathecal injection of gadolinium is not a standard and commonly used procedure, any neurologic symptoms, including minor headache, were carefully monitored and recorded. Overall, the incidence of postprocedural headache in our patients was similar to that reported for lumbar puncture.
We had no false-positive findings. Eight patients who had evidence of intermittent CSF leakage were observed and monitored for infection. In all cases the CSF leak resolved within 1 year. In our patients, MR cisternography after intrathecal injection of gadopentetate dimeglumine was an effective, safe, and minimally invasive technique for the evaluation of suspected CSF fistula arising from the ethmoidal and sphenoidal areas. This technique allowed visualization of the site of leakage in 22 of 24 patients, and surgery confirmed the site in 14 of these patients. We believe that because of the low rate of complications and the benefits of MR cisternography, the informed consent procedure can be similar to that for lumbar puncture. None of the patients had allergic or neurologic complications.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  K. M. Lloyd, J. M. DelGaudio, and P. A. Hudgins Imaging of Skull Base Cerebrospinal Fluid Leaks in Adults Radiology, September 1, 2008; 248(3): 725 - 736. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
