Otology & Neurotology 36:481Y485 Ó 2014, Otology & Neurotology, Inc.

Calvarium Thinning in Patients with Spontaneous Cerebrospinal Fluid Leak *†Rick F. Nelson, *Kameron R. Hansen, *Bruce J. Gantz, and *Marlan R. Hansen *Department of Otolaryngology, Head and Neck Surgery, University of Iowa Hospitals and Clinics, Iowa City, Iowa, U.S.A.; and ÞDepartment of Otolaryngology, Head and Neck Surgery, Indiana University, Indianapolis, Indiana, U.S.A.

Objective: To determine the thickness of the calvarium in patients with spontaneous cerebrospinal fluid (CSF) leaks. Study Design: Case control study. Setting: University of Iowa Hospitals and Clinics. Patients: Those with a confirmed spontaneous CSF leak compared to non-obese (body mass index, BMI G 30) and obese (BMI Q 30) cochlear implant (CI) control groups. All patients had to have temporal bone CT scans that fit specified criteria. Intervention: Bilateral volumetric analysis of the squamous temporal bone and the zygoma in all patients. Assessment of patient age, sex, BMI, and medical comorbidities. Main Outcome Measure: Assessment of the average thickness of the squamous temporal bone and zygoma compared to control groups. Results: The average BMI of patients with spontaneous CSF leaks was significantly higher than non-obese CI controls (43.73 T 9.19 vs. 24.60 T 3.10; P G 0.0001). The calvarium in patients with spontaneous CSF leaks was 23% thinner than both non-

obese CI controls (3.29 T 0.68 vs. 4.25 T 0.58; P G 0.0001) and obese CI controls (3.29 T 0.68 vs. 4.27 T 0.68; P G 0.0001). In addition, the skull thickness of obese CI patients (body mass index, BMI = 37.34 T 6.1) was not significantly different from non-obese CI controls (4.27 T 0.68 vs. 4.25 T 0.58; P = 0.92). The extracranial zygoma was not significantly different among the three groups (ANOVA = 0.9). In our study groups, 5.8% of both CI control groups had the diagnosis of obstructive sleep apnea (OSA), whereas 46.2% of the spontaneous CSF leak patients presented with the diagnosis of OSA. Conclusion: Patients with spontaneous CSF leak are more likely to be obese, have the diagnosis of OSA, and show thinning of their entire calvarium that is independent of BMI. These data suggest an additional obesity-associated intracranial process contributes to skull thinning. Key Words: CalvariumVCerebrospinal fluidV Intracranial hypertensionVObesityVObstructive sleep apneaV Spontaneous CSF leakVSquamous temporal boneVThinning. Otol Neurotol 36:481Y485, 2015.

Spontaneous cerebrospinal fluid (CSF) leakage occurs when there is both a defect in the bony skull base (either anterior or lateral) and a disruption of the meninges, without preceding trauma, surgery, or inciting cause. The anterior skull base overlies the pneumatized nasal sinuses whereas the lateral skull base (also known as the tegmen) overlies the pneumatized middle ear and mastoid. These areas of the skull base are very thin, often a millimeter or thinner. Spontaneous CSF leaks have a very low rate of spontaneous closure. Lateral spontaneous CSF leaks are repaired via a temporal craniotomy with elevation of the temporal lobe and repair of both the meninges and the skull base or via a transmastoid approach. Anterior

spontaneous leaks can be repaired with anterior craniotomy or endoscopically through the nose (1). Currently, it is unknown whether patients with spontaneous CSF leaks inherently have a thin skull base, or if the skull base erodes as a result of a pathologic process over time. It is possible that chronically elevated intracranial pressure (ICP) over many years could lead to erosion of the skull base, but this has not been proven. Nevertheless, spontaneous CSF leaks are associated with idiopathic intracranial hypertension (IIH), also known as benign intracranial hypertension or pseudotumor cerebri (2). In addition, measurements of CSF pressures during lumbar puncture in patients undergoing repair of spontaneous CSF leak show that some (36%), but not all, patients have elevated ICP (3). Furthermore, there is a strong association between obesity and spontaneous CSF leaks (4,5). However, it is currently unknown if obesity is a direct cause of spontaneous CSF leaks. It is known that obstructive sleep apnea (OSA) is highly associated with obesity (6,7) and studies

Address correspondence and reprint requests to Marlan R. Hansen, M.D., Division of Otolaryngology, Head and Neck Surgery, University of Iowa Hospitals and Clinics, 200 Hawkins Road, Iowa City, IA 52246, U.S.A.; E-mail: [email protected] The authors disclose no conflicts of interest.

481

Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.

482

R. F. NELSON ET AL.

have shown that OSA patients have transient large spikes in ICP during apnea spells (8,9). Despite these disease associations, it is still unknown what leads to skull base defects in patients with spontaneous CSF leaks. We hypothesize that spontaneous CSF leak patients have erosion of the skull base and that the same pathologic process also leads to calvarial thinning.

MATERIALS AND METHODS IRB approval was obtained from the University of Iowa (#201307714). We developed a consistent and accurate method of measuring cortical bone thickness to quantify the degree of bone loss in patients with spontaneous CSF leaks. Our control groups consisted of non-obese and obese cochlear implant patients who had a preoperative temporal bone CT at our institution. Inclusion criteria for all groups included (1) temporal bone CT performed at the University of Iowa, (2) axial images (1-mm slices) that span the optic nerve canal and zygoma, and (3) patient age above 18 years and below 80 years. Patients under 18 and over 80 years old were eliminated because there were no spontaneous CSF leak patients in this age range and age-related changes in bone thickness may artificially change cortical thickness measurements in controls groups. Many CT scans not obtained at the University of Iowa had a slice thickness of 0.6 mm and/or did not include the optic canal, which were required for consistent measurements. The following patient characteristics were documented from the medical record: 1) body mass index (BMI) calculated as kilograms per square meter (kg/m2), 2) age, 3) sex, and 4) obstructive sleep apnea. No patients were independently tested for obstructive sleep apnea. For adults aged 20 years and older, obesity was defined as a BMI of 30.0 or higher (10). Obstructive sleep apnea was confirmed if the patient had a positive polysomnogram.

Volumetric analysis was performed by persons blinded to the patient’s diagnosis. Calvarial bone thickness was determined using axial temporal bone CT scans (1-mm slices). The extracranial zygoma bone was measured as a control. Axial images were analyzed using VitreaCore software (version 6.5.3) from Vital Images, Inc. (Minnetonka, MN, USA). CT VScore is a clinical application for calcium scoring that provides the visualization and measurement of calcification or bone volume, which has a signal density greater than or equal to 130 Hounsfield units (11). This application was used to allow for more accurate volumetric bone measurement. Squamous temporal bone volume was measured starting at the level of the optic nerve canal as follows: 1) All volumetric measurements began 10 mm from the lateral orbital wall to eliminate the thickened insertion of the squamous temporal bone. 2) A measurement of 30 mm of squamous temporal bone was taken. 3) The bone was circled and the bone volume was determined using the CT VScore application. 4) Five consecutive inferior horizontal measurements were taken to obtain the total volume (i.e., 5 mm deep, 30 mm long). 5) Bilateral squamous temporal bone measurements were obtained and each patient’s average was determined. 6) Knowing that volume of the temporal bone is = depth (5 mm)  length (30 mm)  thickness, we could calculate the thickness. The thickness (in millimeters) of the squamous temporal bone was determined by dividing the volume obtained by CT VScore by 150 (30 mm long  5 mm deep) (Fig. 1, A). Zygoma bone volume was determined as follows: 1) A measurement of 30 mm of the thickest, middle portion of the zygoma was taken. 2) The bone was encircled and the bone volume was determined using the CT VScore application. 3) Two consecutive measurements were used to obtain the total volume (i.e., 2 mm deep, 30 mm long). 4) Bilateral measurements were obtained and the patient’s average was determined. 5) The thickness (in millimeters) of the zygoma bone was determined by dividing the volume obtained by CT VScore by 60 (30 mm long  2 mm thick) (Fig. 1, B).

FIG. 1. Axial temporal bone CT volumetric measurement method. A, The image is centered at the level of the optic canal. The squamous temporal bone volume is measured starting 10 mm lateral to the lateral orbital wall. The next 30 mm is circled and volume is calculated using VitreaCore software. Measurements from a total of five consecutive 1-mm-thick slices are taken to calculate the total volume. B, The image is centered on the horizontal portion of the zygoma. The central 30 mm is circled and the volume is calculated using VitreaCore software. Measurements from a total of two consecutive 1-mm-thick slices were taken to calculate the total volume. Otology & Neurotology, Vol. 36, No. 3, 2015

Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.

CALVARIAL THINNING IN CSF LEAKS

FIG. 2. Axial temporal bone CT scan images of a normal cochlear implant patient and a spontaneous CSF leak patient centered at the level of the optic canal. Note the difference in the thickness of the calvarium between the patients.

The ratio of the squamous temporal bone to zygoma bone was determined by dividing the thickness (in millimeters) of the squamous temporal bone by the thickness (in millimeters) of the zygoma bone. The average of each group was calculated.

RESULTS Interestingly, we have noted decreased calvarial thickness during middle fossa craniotomy for repair of spontaneous CSF leaks compared to patients undergoing the same approach for treatment of vestibular schwannomas. Calvarial thinning on preoperative temporal bone CT scans corroborated this clinical observation (Fig. 2). We examined the medical records of 31 patients with spontaneous CSF leaks from 2008 to 2012. CT images from patients before 2008 were not digitized. Thirteen patients met criteria for analysis. Seventeen patients were eliminated because the CT was obtained at an outside hospital or the images were not adequate (did not include

TABLE 1.

BMI (kg/m2) Subjects (no.) Male Female Age (y) OSA (%)

483

the optic nerve or not parallel with the zygoma). One patient had only an MRI scan and was not included. We examined medical records of 124 patients who underwent cochlear implantation in 2012. Thirty-one were eliminated because they were below 18 years and 13 were eliminated because they were over 80 years. Thirty were eliminated because they had external CT images or the images were inadequate (did not include the optic nerve or were not parallel with the zygoma). Twelve only had MRI scans and were not included. One patient had external beam radiation and was not included. Twenty patients with a BMI less than 30 met criteria and 17 patients with a BMI more than 30 met criteria. The squamous temporal bone was measured in patients with spontaneous CSF leaks and the thickness was compared to age-matched cochlear implant patients with a BMI less than 30 (Table 1). All CSF leak patients were obese (BMI 930) and the group had an average BMI of 43.73 T 9.19 (Table 1). The calvarium was 0.96 mm thinner in patients with spontaneous CSF leaks (3.29 T 0.68 vs. 4.25 T 0.58; P G0.0001) (Fig. 3). Next, we sought to determine if obesity was independently associated with thinning of the calvarium. Thus, we measured the squamous temporal bone in obese cochlear implant patients. The calvarial thickness was not significantly different compared to non-obese cochlear implant patients (4.27 T 0.68 vs. 4.25 T 0.58; P = 0.89). Similar to non-obese CI patients, the calvarial thickness of spontaneous CSF leak patients was significantly thinner than obese CI patients (4.27 T 0.68 vs. 3.29 T 0.68; P G0.0001; Fig. 3). The extracranial zygoma, which would not be subject to intracranial pressure or pathology, was not significantly different among the three groups (ANOVA, P = 0.92). These data suggest that there is not a systemic

Patient characteristic of the three study groups; kg = kilograms; m = meter; y = year Cochlear Implant (Non-Obese)

Cochlear Implant (Obese)

Spontaneous CSF Leak

24.60 T 3.10 20 8 12 60.4 T 16.5 5.9

37.34 T 6.10 17 8 9 59.9 T 13.1 5.9

43.73 T 9.19 13 4 9 56.0 T 8.3 46.2

FIG. 3. Squamous temporal bone and zygoma thickness in the three study groups. mm = millimeters. *** = P G 0.001. Otology & Neurotology, Vol. 36, No. 3, 2015

Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.

484

R. F. NELSON ET AL. thickness-to-zygoma thickness ratio between the six CSF leak patients with OSA (0.73 T 0.12) with the seven CSF leak patients without OSA (0.71 T 0.13) patients.

DISCUSSION

FIG. 4. Scatter plot of the ratio of the temporal bone to zygoma thickness for individual patients in all three study groups. The horizontal bar is the average of the group. The arrows designate patients with documented obstructive sleep apnea.

process, such as osteopenia, that is causing the calvarial thinning in spontaneous CSF leak patients. To control for inter-patient variability in bone density, we plotted a ratio of the squamous temporal bone thickness to the zygoma thickness (Fig. 4). The average ratio for spontaneous CSF leak patients (0.71 T 0.12) was significantly lower than the non-obese CI (0.91 T 0.12; P G 0.0001) and obese CI (0.93 T 0.21; P = 0.007). There was no statistical difference between the two CI groups (P = 0.72). In addition, the ratio of the squamous temporal bone to zygoma bone in every spontaneous CSF leak patient was at or below the mean of the two CI control groups. We did not directly test patients for intracranial hypertension with lumbar puncture as this is not our practice. In addition, we did not perform polysomnograms to evaluate for OSA in all patients. Rather, medical records were examined and showed that 46.2% of patients (6 of 13 patients had a positive polysomnogram) with spontaneous CSF leaks included in this study presented with the diagnosis of OSA (Table 1). One patient with a spontaneous CSF leak had witnessed apneas but no formal polysomnogram and thus was not included in this study. One patient in the non-obese CI group (5.8%) and one patient in the obese CI group (5.8%) presented with the diagnosis of OSA (positive polysomnogram; Table 1). Interestingly, both of these patients had thin skulls similar to the spontaneous CSF leak patients demonstrated by the squamous temporal bone-to-zygoma bone ratio (Fig. 4). Next, we compared skull thickness of CSF leak patients with and without a presenting diagnosis of OSA. There was no statistical difference (P = 0.58) between the squamous temporal bone thickness of the six CSF leak patients with OSA (3.42 T 0.76 mm) compared to the seven CSF leak patients without a formal diagnosis of OSA (3.19 T 0.70 mm). There was no statistical difference (P = 0.78) between the squamous temporal bone

Here we show that patients with spontaneous CSF leaks have thinning of the squamous temporal bone that appears to be independent of body mass. It was not feasible to measure the entire skull thickness on the CT scans that we analyzed. Therefore, we chose the squamous temporal bone as an area of the skull that would be represented in the scans and infer that thinning of the squamous temporal bone reflects global thinning of the calvarium. These patients have bony skull base defects in addition to thinning of the squamous temporal bone further supporting the notion that there is thinning of the entire calvarium in spontaneous CSF leak patients. In addition, many patients with spontaneous CSF leaks have a squamous temporal bone thickness-to-zygoma ratio that is below the range of control groups implicating an ongoing erosive process as opposed to an intrinsically thin skull. The extracranial zygoma bone thickness was the same in all groups, indicating the thinning of the squamous temporal bone is not the result of a systemic bone disease such as osteopenia. Interestingly, another study demonstrated two patients with superior semicircular canal dehiscence to also have thinning of the squamous temporal bone (12). We also show that patients with spontaneous CSF leaks are likely to be obese. The obesity epidemic developed in the United States starting in the 1990s (13). The percentage of American that are obese (BMI Q 30) has risen from 12% in 1991 to 19.8% in 2000 to 29% in 2010 (14). Our data along with others (4,5) show a strong correlation between obesity and spontaneous CSF leaks. One limitation of these data is that they only demonstrate an association between CSF leaks, skull thinning, and BMI. Demonstration of a causal relationship requires experimental manipulation of these variables in some patients but not others creating feasibility and ethical barriers. Cortical thinning in patients with a spontaneous CSF leak suggests that there is an intracranial process that leads to erosion of the skull and skull base. Areas of the skull base overlying pneumatized cavities (e.g., tegmen mastoideum) are very thin and the loss of 0.96 mm in this group of patients physically correlates with the bone loss necessary to have a CSF leak. In addition, there is likely a time requirement on the order of months to years for erosion of the skull base. One could hypothesize that elevated ICP contributes to skull and skull base erosion. Previous studies associated spontaneous CSF leaks with IIH (2), and some patients with spontaneous CSF leaks (~36%) have elevated ICP during lumbar puncture (3). We do not routinely perform lumbar puncture to measure ICP and we do not routinely place lumbar drains for management. Thus, we cannot yet make a direct causal link between elevated ICP and skull thinning.

Otology & Neurotology, Vol. 36, No. 3, 2015

Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.

CALVARIAL THINNING IN CSF LEAKS Here we demonstrate a high prevalence of a previous diagnosis of OSA and skull thinning in patients with spontaneous CSF leaks. Others have shown that patients with spontaneous CSF leaks have a higher rate of OSA (5) and patients with OSA have been found to have elevations in ICP and arterial blood pressure during apnea events (8,9). It is postulated that apneas lead to hypercarbia, which in turn leads to cerebral vasodilation and elevated ICP. OSA is defined as an apnea/hypopnea index above 5 on polysomnogram and the prevalence of OSA is approximately 4% in men and 2% in women in the United States (7). Our data show that patients with spontaneous CSF leaks have a much higher prevalence of OSA (~50%) than the national average. These associations do not necessarily imply a casual role for OSA in the development of spontaneous CSF leak or skull thinning. In this study, patients with a spontaneous CSF leak and OSA did not have thinner skulls than those patients without the formal diagnosis of OSA. The power of this comparison is limited because of the small sample size. In addition, polysomnograms were not performed on all patients in this study, so the true incidence of OSA in the populations in this study is not known. However, our data, taken together with other studies, highlight the need to directly determine the contribution of OSA to skull thinning and CSF leak. Spontaneous CSF leaks are highly associated with obesity. However, obesity alone likely does not cause calvarial thinning and skull base erosion. Mechanistically, we propose that another obesity-related intracranial process likely contributes to erosion of the skull and skull base. Further studies are needed to establish this intracranial process. Clinically, we propose that cortical thickness and OSA diagnosis should be evaluated in patients with a spontaneous CSF leak. Acknowledgment: We thank Mary Burr for assistance with VitreaCore.

485 REFERENCES

1. Banks CA, Palmer JN, Chiu AG, et al. Endoscopic closure of CSF rhinorrhea: 193 cases over 21 years. Otolaryngol Head Neck Surg 2009;140:826Y33. 2. Schlosser RJ, Woodworth BA, Wilensky EM, et al. Spontaneous cerebrospinal fluid leaks: a variant of benign intracranial hypertension. Ann Otol Rhinol Laryngol 2006;115:495Y500. 3. Allen KP, Perez CL, Kutz JW, et al. Elevated intracranial pressure in patients with spontaneous cerebrospinal fluid otorrhea. Laryngoscope 2013;124:251Y4. 4. Stucken EZ, Selesnick SH, Brown KD. The role of obesity in spontaneous temporal bone encephaloceles and CSF leak. Otol Neurotol 2012;33:1412Y7. 5. LeVay AJ, Kveton JF. Relationship between obesity, obstructive sleep apnea, and spontaneous cerebrospinal fluid otorrhea. Laryngoscope 2008;118:275Y8. 6. Li C, Ford ES, Zhao G, et al. Prevalence of self-reported clinically diagnosed sleep apnea according to obesity status in men and women: National Health and Nutrition Examination Survey, 2005Y2006. Prev Med 2010;51:18Y23. 7. Lee W, Nagubadi S, Kryger MH, et al. Epidemiology of obstructive sleep apnea: a population-based perspective. Expert Rev Respir Med 2008;2:349Y64. 8. Jennum P, Borgesen SE. Intracranial pressure and obstructive sleep apnea. Chest 1989;95:279Y83. 9. Sugita Y, Iijima S, Teshima Y, et al. Marked episodic elevation of cerebrospinal fluid pressure during nocturnal sleep in patients with sleep apnea hypersomnia syndrome. Electroencephalogr Clin Neurophysiol 1985;60:214Y9. 10. Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults: executive summary. Expert Panel on the Identification, Evaluation, and Treatment of Overweight in Adults. Am J Clin Nutr 1998;68:899Y917. 11. Hoff JA, Chomka EV, Krainik AJ, et al. Age and gender distributions of coronary artery calcium detected by electron beam tomography in 35,246 adults. Am J Cardiol 2001;87:1335Y9. 12. Friedland DR, Michel MA. Cranial thickness in superior canal dehiscence syndrome: implications for canal resurfacing surgery. Otol Neurotol 2006;27:346Y54. 13. Mokdad AH, Bowman BA, Ford ES, et al. The continuing epidemics of obesity and diabetes in the United States. JAMA 2001; 286:1195Y200. 14. Mokdad AH, Serdula MK, Dietz WH, et al. The spread of the obesity epidemic in the United States, 1991Y1998. JAMA 1999;282:1519Y22.

Otology & Neurotology, Vol. 36, No. 3, 2015

Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.

Calvarium thinning in patients with spontaneous cerebrospinal fluid leak.

To determine the thickness of the calvarium in patients with spontaneous cerebrospinal fluid (CSF) leaks...
1MB Sizes 4 Downloads 11 Views