A review of the English language literature since 1966 (MEDLINE) revealed 10 previous cases of spontaneous CSF leak emanating from the posterior aspect of the petrous pyramid ().3,7,11,17,18,19,20,21,22
Some of these cases were published before the era of high-resolution imaging and in others the defect was poorly characterized. We report a series of three patients with high-resolution imaging corroborating their clinical history and precise surgical observation.
Summary of Published PF CSF Leak Cases to Date Including Our Case Series
On evaluation of all cases, including our own, PF leak patients do appear to share the demographics of MF leaks. Of the 12 PF cases, 58% were female compared with the 60 to 90% incidence reported in published MF leak cases.1,3,5,6,7,8,10
Interestingly, only 5 of the 12 cases fell in the characteristic age range of 40 to 60 years, however the average age was still 59. Not enough information was available with regards to obesity and IIH. It should be noted that all three of our patients were overweight, but none had a BMI greater than 30. Clinical presentation was similar though meningitis was a more common presentation in PF leak patients.
Prior cases confirm that punctate lesions, solitary or multiple, predominate. In our series, two PF CSF leaks were of narrow diameter whereas one had an extensive defect in the posterior petrous face. In the earlier literature, 2 of the 10 cases had bilateral defects in the posterior aspect of the petrous bone. We recognized a bilateral deformity in our case 1, but actual leakage on one side.
A difference between the cases we provide compared with previously published PF leak cases was that two of the three patients presented with coexisting MF and PF leaks. For one case, not recognizing the PF dehiscence resulted in continued leak following a MF repair necessitating a second surgery. The concomitant MF and PF leak as well as shared demographics point to a possible similar etiology for tegmen and posterior plate defects.
It has been proposed that raised intracranial pressure may be the central etiologic factor for transtemporal CSF leaks. Long-term pressure may lead to erosion of often thin dural plates on the faces of the petrous pyramid, the MF being more susceptible than the PF.14
The bilateral scalloping of the posterior petrous face in case 1 suggests underlying high pressure and perhaps erosion by prominent AGs.12,13
Alternatively, preformed pathways could provide an avenue for dural herniation into the pneumatic spaces of the temporal bone. Schuknecht suggests that “embryogenic faults” in the dura and bone can slowly break down with normal CSF pressure over time.23
One possible site is at the endolymphatic sac and operculum where the cartilage of the embryonic otic capsule fuses with the temporal bone.24
The narrow opening in the vicinity of the endolymphatic sac in case 3 suggests such a mechanism. The large focused posterior defect with cerebellar herniation in case 2 may also suggest such a mechanism; however, the contralateral cerebellum abutting the posterior petrous face also implicates a role for increased intracranial pressure.
It is imperative that when evaluating a patient for a transtemporal CSF leak the PF plate be carefully evaluated on preoperative imaging. If irregularity or defect is noted, exploration via mastoidectomy is warranted. It should be cautioned that antrotomy alone is insufficient and could easily overlook a subtle posterior surface deficiency. The entire petrous face from sigmoid sinus laterally to posterior semicircular canal medially should be inspected. Special scrutiny should be given to the area of the endolymphatic sac and its operculum.