CSF Leakage in the Setting of Pituitary Adenomas
A comprehensive search of the published literature revealed 29 articles with 52 cases of spontaneous and medically induced CSF leaks associated with pituitary adenomas. Although this phenomenon has been described anecdotally or as single case reports in the past, the purpose of this review was to provide an improved characterization of the clinical and pathological settings in which this phenomenon occurs. The majority of cases of nonsurgical CSF leaks, the leaks occur in the setting of invasive pituitary tumors (typically prolactinomas) following initiation of standard medical management (Table 1). In addition, some patients may present with spontaneous CSF leaks as the presenting symptom of a newly diagnosed pituitary tumor (Table 2). Aside from prolactinomas, other (typically invasive) tumors that may result in nonsurgical CSF rhinorrhea include nonfunctioning macroadenomas, GH-secreting adenomas, and ACTHsecreting adenomas.
Based on the observations made and an understanding of the anatomy and pathophysiology of these lesions and the sellar space, the following mechanism has been proposed to explain the development of CSF rhinorrhea in the setting of pituitary adenomas. In larger and more invasive pituitary adenomas, tumor expansion into the surrounding dural and bony structures is commonly observed. In addition, various functional subtypes of pituitary adenomas are known to preferentially invade various regions of the parasellar spaces. For instance, GH-secreting adenomas and prolactinomas are known to frequently invade the sellar floor and infrasellar space (sphenoid sinus and clivus), whereas nonfunctioning macroadenomas have a tendency to invade the suprasellar space via bowing or invasion of the diaphragma sellae. In cases where the arachnoid and/or brain parenchyma have also been violated, the potential for developing a CSF fistula has been established. These communications have little clinical significance as long as the tumor occludes the opening and serves as a "plug," thus preventing the escape of CSF. Any significant reduction in tumor size, however, can provide a conduit for the escape of CSF, typically resulting in CSF rhinorrhea. Rapid tumor shrinkage is frequently associated with initiation of dopamine agonist therapy for prolactin-secreting tumors. In cases of spontaneous (noniatrogenic) rhinorrhea, the cause of tumor shrinkage is not well understood, but it is likely related to intratumoral infarction and/or hemorrhage with subsequent reduction in tumor volume, ongoing invasion through the arachnoid or bony skull base, and/or increases in intracranial pressure resulting in a CSF fistula. The development of meningitis should be anticipated and addressed in advance in cases of nonsurgical CSF leak.
Surgical repair is the recommended initial treatment for definitive management of DA-induced rhinorrhea, and was ultimately required in nearly 90% of patients reviewed in the literature. Our preferred operation in the majority of cases is the endoscopic endonasal approach, although open surgical repair via a craniotomy may be warranted in a minority of cases. Depending on the tumor subtype and growth patterns, tumor resection may be concurrently achieved prior to reconstruction of the skull base in these instances, although complete tumor resection may be limited in many of these cases as these tumors are by definition invasive into bone and dura. Following tumor debulking, standard techniques for repairing the skull base can be used. For smaller "weeping" CSF leaks, dural substitutes and/or fibrin glue may be used to achieve a successful repair. For larger CSF leaks, the use of autologous fat or fascia is recommended, with or without a sellar floor buttress. For extremely large or refractory CSF leaks, rotation of a pedicled nasoseptal flap may be required to definitively address the CSF fistula. If the location of CSF leakage is not obvious, the use of intrathecal fluorescein may also be attempted to achieve a more targeted repair.
Although reduction or discontinuation of medical therapy was attempted as the primary treatment in 24% of cases, we do not recommend this strategy in patients who are candidates for surgical repair. The half-life of cabergoline is 63–69 hours and that of bromocriptine is 12–14 hours, suggesting that no immediate effects from medication discontinuation will be achieved via this strategy and that patients may be at risk for the development of meningitis. Furthermore, the effects on tumor volume and the CSF fistula may be irreversible in the short term, often mandating eventual surgical repair of the CSF leak. We recommend surgical repair of spontaneous or medically induced CSF leaks in the setting of pituitary adenomas, along with safe maximal tumor resection, in patients who are surgical candidates. In patients who are not surgical candidates, a more conservative approach would include temporary cessation of medical therapy and temporary insertion of a lumbar drain, but this strategy is not recommended as a first-line therapy unless deemed medically necessary.
Although the current review provides improved characterization of factors associated with CSF fistulas associated with nonsurgical treatment of pituitary tumors, there are inherent limitations of this particular study design. There is a certain degree of publication bias associated with amassing data from numerous case reports and small clinical series to describe a relatively uncommon phenomenon. In an attempt to further understand the clinical scenarios in which this type of CSF leak most commonly occurs, however, we aimed to identify which tumor subtypes and particular treatments posed the greatest risk for the development of CSF rhinorrhea.
Discussion
A comprehensive search of the published literature revealed 29 articles with 52 cases of spontaneous and medically induced CSF leaks associated with pituitary adenomas. Although this phenomenon has been described anecdotally or as single case reports in the past, the purpose of this review was to provide an improved characterization of the clinical and pathological settings in which this phenomenon occurs. The majority of cases of nonsurgical CSF leaks, the leaks occur in the setting of invasive pituitary tumors (typically prolactinomas) following initiation of standard medical management (Table 1). In addition, some patients may present with spontaneous CSF leaks as the presenting symptom of a newly diagnosed pituitary tumor (Table 2). Aside from prolactinomas, other (typically invasive) tumors that may result in nonsurgical CSF rhinorrhea include nonfunctioning macroadenomas, GH-secreting adenomas, and ACTHsecreting adenomas.
Based on the observations made and an understanding of the anatomy and pathophysiology of these lesions and the sellar space, the following mechanism has been proposed to explain the development of CSF rhinorrhea in the setting of pituitary adenomas. In larger and more invasive pituitary adenomas, tumor expansion into the surrounding dural and bony structures is commonly observed. In addition, various functional subtypes of pituitary adenomas are known to preferentially invade various regions of the parasellar spaces. For instance, GH-secreting adenomas and prolactinomas are known to frequently invade the sellar floor and infrasellar space (sphenoid sinus and clivus), whereas nonfunctioning macroadenomas have a tendency to invade the suprasellar space via bowing or invasion of the diaphragma sellae. In cases where the arachnoid and/or brain parenchyma have also been violated, the potential for developing a CSF fistula has been established. These communications have little clinical significance as long as the tumor occludes the opening and serves as a "plug," thus preventing the escape of CSF. Any significant reduction in tumor size, however, can provide a conduit for the escape of CSF, typically resulting in CSF rhinorrhea. Rapid tumor shrinkage is frequently associated with initiation of dopamine agonist therapy for prolactin-secreting tumors. In cases of spontaneous (noniatrogenic) rhinorrhea, the cause of tumor shrinkage is not well understood, but it is likely related to intratumoral infarction and/or hemorrhage with subsequent reduction in tumor volume, ongoing invasion through the arachnoid or bony skull base, and/or increases in intracranial pressure resulting in a CSF fistula. The development of meningitis should be anticipated and addressed in advance in cases of nonsurgical CSF leak.
Surgical repair is the recommended initial treatment for definitive management of DA-induced rhinorrhea, and was ultimately required in nearly 90% of patients reviewed in the literature. Our preferred operation in the majority of cases is the endoscopic endonasal approach, although open surgical repair via a craniotomy may be warranted in a minority of cases. Depending on the tumor subtype and growth patterns, tumor resection may be concurrently achieved prior to reconstruction of the skull base in these instances, although complete tumor resection may be limited in many of these cases as these tumors are by definition invasive into bone and dura. Following tumor debulking, standard techniques for repairing the skull base can be used. For smaller "weeping" CSF leaks, dural substitutes and/or fibrin glue may be used to achieve a successful repair. For larger CSF leaks, the use of autologous fat or fascia is recommended, with or without a sellar floor buttress. For extremely large or refractory CSF leaks, rotation of a pedicled nasoseptal flap may be required to definitively address the CSF fistula. If the location of CSF leakage is not obvious, the use of intrathecal fluorescein may also be attempted to achieve a more targeted repair.
Although reduction or discontinuation of medical therapy was attempted as the primary treatment in 24% of cases, we do not recommend this strategy in patients who are candidates for surgical repair. The half-life of cabergoline is 63–69 hours and that of bromocriptine is 12–14 hours, suggesting that no immediate effects from medication discontinuation will be achieved via this strategy and that patients may be at risk for the development of meningitis. Furthermore, the effects on tumor volume and the CSF fistula may be irreversible in the short term, often mandating eventual surgical repair of the CSF leak. We recommend surgical repair of spontaneous or medically induced CSF leaks in the setting of pituitary adenomas, along with safe maximal tumor resection, in patients who are surgical candidates. In patients who are not surgical candidates, a more conservative approach would include temporary cessation of medical therapy and temporary insertion of a lumbar drain, but this strategy is not recommended as a first-line therapy unless deemed medically necessary.
Although the current review provides improved characterization of factors associated with CSF fistulas associated with nonsurgical treatment of pituitary tumors, there are inherent limitations of this particular study design. There is a certain degree of publication bias associated with amassing data from numerous case reports and small clinical series to describe a relatively uncommon phenomenon. In an attempt to further understand the clinical scenarios in which this type of CSF leak most commonly occurs, however, we aimed to identify which tumor subtypes and particular treatments posed the greatest risk for the development of CSF rhinorrhea.
Source...