CSF Leakage in the Setting of Pituitary Adenomas
CSF Leakage in the Setting of Pituitary Adenomas
Twenty-nine articles were identified from between the years 1980 and 2011, describing 52 patients with spontaneous or medically induced CSF leaks in the setting of a pituitary adenoma. A majority of these cases (38 patients, 73%) occurred following initiation of medical therapy, whereas the CSF leak developed as the presenting symptom of a pituitary adenoma in the remaining 14 patients (27%). The mean age was 42.8 years (range 23– 68 years). There were 35 men and 17 women. Forty-two patients (81%) had a prolactinoma. The remaining 10 patients had the following tumor subtypes: nonfunctioning pituitary adenoma (6 patients, 11%), mammosomatotroph cell adenoma (1 patient, 2%), GH-secreting adenoma (2 patients, 4%), and ACTH-secreting adenoma (1 patient, 2%). Thirty-eight patients (73%) had a previously diagnosed pituitary tumor and developed CSF rhinorrhea following initiation of medical treatment. Thirty-six of the 38 patients who developed CSF leaks following initiation of medical therapy had prolactinomas, whereas patients who developed spontaneous (noniatrogenic) CSF leaks had a variety of pituitary adenomas (6 prolactinomas, 6 nonfunctioning pituitary adenomas, 1 GH-secreting adenoma, and 1 ACTH-secreting adenoma). The medical agents most closely associated with CSF leakage were DA medications in 37 patients (97%) and a somatostatin analog (lanreotide) in 1 patient (3%). Of the patients who developed CSF rhinorrhea while taking DA medications, 24 were taking bromocriptine and 13 were taking cabergoline. The average time from initiation of medical therapy to onset of rhinorrhea was 3.3 months (range 3 days–17 months). Seven patients (14%) presented with meningitis in conjunction with CSF rhinorrhea. Tumor size was reported in 9 of 52 cases (17%), with a mean maximum tumor diameter of 3.6 cm. Forty-nine of 52 patients (94%) were reported to have tumors with neuroimaging evidence of extrasellar tumor invasion. Specific invasion into the sphenoid sinus, ethmoid sinus, or clivus was documented in 29 patients (56%). In 47 patients, a discrete skull base defect could be identified on imaging studies, most frequently with thin-slice CT images in the coronal plane. The average initial prolactin level in patients with spontaneous CSF leakage was 9169 ng/ml, compared with 4917 ng/ml in those with medically induced leakage. Due to a small number of values from the spontaneous leakage group, however, this difference was not statistically significant. Interestingly, no cases of spontaneous CSF leakage following radiation therapy or radiosurgery for pituitary adenomas were identified. One case of a spontaneous CSF leak developing 4 months following Gamma Knife surgery for a metastatic renal cell carcinoma to the skull base was identified, but this case was excluded from the present study. The patient underwent successful CSF leak repair via an endoscopic sinus approach.
In 4 patients, nonoperative management was successfully employed. Of these, 1 patient was successfully treated with temporary lumbar drainage, and the other 3 with bed rest or withdrawal of medications. Forty-six patients (88%) ultimately underwent surgical intervention as the definitive treatment for the CSF leak. Definitive procedures for the CSF leaks included: transsphenoidal surgery (32 patients), craniotomy (5 patients), lumboperitoneal shunt (2 patients), and unknown approach (7 patients). In 2 patients, surgery was not performed, due to medical contraindications or patient preference, resulting in ongoing CSF rhinorrhea.
In nearly 90% of cases reported, operative repair of the CSF fistula was eventually required. In 21 of 38 cases of DA-induced leak, there was no documented recurrence after the initial treatment. In these cases, the initial treatment strategies were as follows: transsphenoidal surgery (in 16 cases), transsphenoidal surgery combined with lumbar puncture (1 case), frontal craniotomy (1 case), unspecified surgery (1 case), and treatment withdrawal or bed rest (2 cases).
Recurrence was documented in the remaining 17 cases, and a combination of treatment approaches was employed. In 8 cases, temporary cessation of rhinorrhea occurred with treatment reduction or withdrawal, but the rhinorrhea recurred within days or weeks of restarting medical treatment. Of these cases, definite resolution of rhinorrhea was achieved via transsphenoidal surgery in 5 cases and a combination of craniotomy and transbasal surgery in the sixth. In the seventh and eighth cases, surgical repair was not performed because of patient refusal or contraindications to surgery, and rhinorrhea continued to occur.
In 7 cases, rhinorrhea recurred despite initial surgical treatment. Of these cases, rhinorrhea was ultimately resolved with a transsphenoidal approach in 3 patients, a transfrontal approach in 1 patient, lumboperitoneal shunt placement in 1 patient, and craniotomy in 1 patient, and it subsided without treatment in 1 patient.
In 2 cases, the rhinorrhea initially subsided with bed rest. Of these cases, the first recurred 3 three years later, even though treatment was uneventfully reinstated a month after the initial rhinorrhea episode ceased, and the recurrence subsided again with bed rest. The second was promptly resolved with transsphenoidal surgery.
Finally, CSF rhinorrhea occurred spontaneously as the presenting symptom of pituitary adenoma in 14 cases, of which 7 resolved with surgical repair via a transsphenoidal approach, and 2 resolved with unknown operations. In 1 patient, a temporary lumbar drain was successfully used as the only definitive treatment. In the remaining 4 cases, initial management with frontal craniotomies or transsphenoidal surgery failed to resolve the leak. One of these patients was treated with bed rest, and the others were treated definitively with first-time or repeat transsphenoidal surgery.
Results
Patient and Tumor Characteristics
Twenty-nine articles were identified from between the years 1980 and 2011, describing 52 patients with spontaneous or medically induced CSF leaks in the setting of a pituitary adenoma. A majority of these cases (38 patients, 73%) occurred following initiation of medical therapy, whereas the CSF leak developed as the presenting symptom of a pituitary adenoma in the remaining 14 patients (27%). The mean age was 42.8 years (range 23– 68 years). There were 35 men and 17 women. Forty-two patients (81%) had a prolactinoma. The remaining 10 patients had the following tumor subtypes: nonfunctioning pituitary adenoma (6 patients, 11%), mammosomatotroph cell adenoma (1 patient, 2%), GH-secreting adenoma (2 patients, 4%), and ACTH-secreting adenoma (1 patient, 2%). Thirty-eight patients (73%) had a previously diagnosed pituitary tumor and developed CSF rhinorrhea following initiation of medical treatment. Thirty-six of the 38 patients who developed CSF leaks following initiation of medical therapy had prolactinomas, whereas patients who developed spontaneous (noniatrogenic) CSF leaks had a variety of pituitary adenomas (6 prolactinomas, 6 nonfunctioning pituitary adenomas, 1 GH-secreting adenoma, and 1 ACTH-secreting adenoma). The medical agents most closely associated with CSF leakage were DA medications in 37 patients (97%) and a somatostatin analog (lanreotide) in 1 patient (3%). Of the patients who developed CSF rhinorrhea while taking DA medications, 24 were taking bromocriptine and 13 were taking cabergoline. The average time from initiation of medical therapy to onset of rhinorrhea was 3.3 months (range 3 days–17 months). Seven patients (14%) presented with meningitis in conjunction with CSF rhinorrhea. Tumor size was reported in 9 of 52 cases (17%), with a mean maximum tumor diameter of 3.6 cm. Forty-nine of 52 patients (94%) were reported to have tumors with neuroimaging evidence of extrasellar tumor invasion. Specific invasion into the sphenoid sinus, ethmoid sinus, or clivus was documented in 29 patients (56%). In 47 patients, a discrete skull base defect could be identified on imaging studies, most frequently with thin-slice CT images in the coronal plane. The average initial prolactin level in patients with spontaneous CSF leakage was 9169 ng/ml, compared with 4917 ng/ml in those with medically induced leakage. Due to a small number of values from the spontaneous leakage group, however, this difference was not statistically significant. Interestingly, no cases of spontaneous CSF leakage following radiation therapy or radiosurgery for pituitary adenomas were identified. One case of a spontaneous CSF leak developing 4 months following Gamma Knife surgery for a metastatic renal cell carcinoma to the skull base was identified, but this case was excluded from the present study. The patient underwent successful CSF leak repair via an endoscopic sinus approach.
Treatment Characteristics in Patients With CSF Leaks and Underlying Pituitary Adenomas
In 4 patients, nonoperative management was successfully employed. Of these, 1 patient was successfully treated with temporary lumbar drainage, and the other 3 with bed rest or withdrawal of medications. Forty-six patients (88%) ultimately underwent surgical intervention as the definitive treatment for the CSF leak. Definitive procedures for the CSF leaks included: transsphenoidal surgery (32 patients), craniotomy (5 patients), lumboperitoneal shunt (2 patients), and unknown approach (7 patients). In 2 patients, surgery was not performed, due to medical contraindications or patient preference, resulting in ongoing CSF rhinorrhea.
In nearly 90% of cases reported, operative repair of the CSF fistula was eventually required. In 21 of 38 cases of DA-induced leak, there was no documented recurrence after the initial treatment. In these cases, the initial treatment strategies were as follows: transsphenoidal surgery (in 16 cases), transsphenoidal surgery combined with lumbar puncture (1 case), frontal craniotomy (1 case), unspecified surgery (1 case), and treatment withdrawal or bed rest (2 cases).
Recurrence was documented in the remaining 17 cases, and a combination of treatment approaches was employed. In 8 cases, temporary cessation of rhinorrhea occurred with treatment reduction or withdrawal, but the rhinorrhea recurred within days or weeks of restarting medical treatment. Of these cases, definite resolution of rhinorrhea was achieved via transsphenoidal surgery in 5 cases and a combination of craniotomy and transbasal surgery in the sixth. In the seventh and eighth cases, surgical repair was not performed because of patient refusal or contraindications to surgery, and rhinorrhea continued to occur.
In 7 cases, rhinorrhea recurred despite initial surgical treatment. Of these cases, rhinorrhea was ultimately resolved with a transsphenoidal approach in 3 patients, a transfrontal approach in 1 patient, lumboperitoneal shunt placement in 1 patient, and craniotomy in 1 patient, and it subsided without treatment in 1 patient.
In 2 cases, the rhinorrhea initially subsided with bed rest. Of these cases, the first recurred 3 three years later, even though treatment was uneventfully reinstated a month after the initial rhinorrhea episode ceased, and the recurrence subsided again with bed rest. The second was promptly resolved with transsphenoidal surgery.
Finally, CSF rhinorrhea occurred spontaneously as the presenting symptom of pituitary adenoma in 14 cases, of which 7 resolved with surgical repair via a transsphenoidal approach, and 2 resolved with unknown operations. In 1 patient, a temporary lumbar drain was successfully used as the only definitive treatment. In the remaining 4 cases, initial management with frontal craniotomies or transsphenoidal surgery failed to resolve the leak. One of these patients was treated with bed rest, and the others were treated definitively with first-time or repeat transsphenoidal surgery.
Source...