Computational fluid dynamics after endoscopic endonasal skull base surgery- Possible ENS in the context of middle turbinate resection

Abstract

Introduction

Empty nose syndrome (ENS) is a rare and debilitating disease with controversial definition, etiology and treatment. One puzzling fact is that patients who undergo an endoscopic endonasal approach (EEA) often receive resection of multiple anatomical structures, yet seldom develop ENS. In this pilot study, we analyzed and compared the computational fluid dynamics (CFD) and symptoms among patients following EEA, ENS patients and healthy cohorts.

Methods

CT scans of four post-EEA patients were collected and analyzed utilizing CFD techniques. Two had significant ENS symptoms based on an ENS6Q questionnaire (score >11), while the other two were asymptomatic. As a reference, their results were compared to previously published CFD results of 27 non-EEA ENS patients and 42 healthy controls.

Results

Post-EEA patients with ENS symptoms had a similar nasal airflow pattern as non-EEA ENS patients. And this pattern, significantly differed from that of EEA patients without ENS symptoms and healthy controls. Overall, groups with ENS symptoms exhibited airflow dominant in the middle meatus region and significantly less percentage of airflow in the inferior turbinate region (EEAwENS:17.74%±4.00% vs EEAw/oENS:51.25%±3.33%, t-test p<0.02; non-EEA ENS:25.8±17.6%; healthy:36.5±15.9%) as well as lower peak wall-shear-stress (EEAwENS:0.30±0.13 vs EEAw/oENS:0.61±0.03, p=0.003; non-EEA ENS:0.58±0.24; healthy:1.18±0.81 Pa).

Conclusion

These results suggest that turbinectomy and/or posterior septectomy may have a varying functional impact and that ENS symptoms go beyond anatomy, but rather correlate with aerodynamic changes. The findings open the door for CFD as potential objective diagnosis of ENS.

INTRODUCTION

Empty Nose Syndrome (ENS) is an enigmatic disease that remains controversial in regards to its definition, etiology and treatment. Currently, it is thought to be caused by altered patterns of nasal airflow and impaired neurosensory functions.^{1, 2} Although ENS is most commonly linked to procedures involving an aggressive resection of the inferior turbinate (IT), it has also been described after resection of the middle turbinate (MT) in much rarer incidents³ with its physiopathology even more poorly understood.

Surgical techniques employed to address lesions of the skull base and adjacent structures, have evolved significantly during the last two decades. Endoscopic endonasal approaches (EEA) have gradually replaced some traditional open approaches that required a craniotomy and brain retraction. The sinonasal corridor offers a direct route to the ventral skull base; however, despite the frequent resection of the middle and superior turbinates and posterior septum, these patients seldom develop ENS symptoms.

Mechanisms, by which this at-risk population rarely develops ENS symptoms, are not yet understood; however, their elucidation could offer critical insights about the disorder. To study this phenomenon, the computational fluid dynamic (CFD) models of EEA patients with and without clinical symptoms of ENS were analyzed, and their airflow patterns were compared with those of healthy controls as well as patients with ENS who did not undergo EEA surgery.

MATERIALS AND METHODS

Subjects

The subjects were patients treated during the past 5 years at The Ohio State University - Wexner Medical Center with a history of EEA that volunteered to fill out an Empty Nose Syndrome 6-items questionnaire ⁴ (ENS6Q). Patients with a history of radiotherapy, chronic rhinosinusitis or nasal obstructive symptoms before surgery were excluded. Due to that the postoperative imaging of these patients typically centers on the skull base, we were only able to find 4 patients with a high resolution CT/MRI scan comprising the entire sinonasal cavity (see Figure 1), among which two had an ENS6Q score above 11 (EEAwENS) and two below this thresholdENS development, ²⁵ particularly (EEAw/oENS). The imaging was analyzed using CFD techniques and compared between groups. The Ohio State University Institutional Review Boards approved the current study.

Figure 1.

Coronal sections of patient’s imaging, including the sinonasal cavity at similar antero-posterior depth. Patient 1: Coronal T1 MRI in the anterior (A), middle (B) and posterior (C) thirds. Patient 2: Coronal T1 MRI in the anterior (D), middle (E) and posterior (F) thirds. Patient 3: Coronal T1 MRI in the anterior (G), middle (H) and posterior (I) thirds. Patient 4: Coronal Head CT in the anterior (J), middle (K) and posterior (L) thirds. Pt: Patient.

A separate group of confirmed ENS patients (n=27) that did not undergo EEA (Non-EEA ENS) was obtained from previously published data² (Table 1), constituting likely the largest published series of CFD studies of ENS patients². This group, whose ages ranged from 25 to 67 years (median 47 years), had a mean ENS6Q score of 19.78, and underwent multiple rhinologic procedures (I.e. septoplasty, FESS, turbinate reduction), but not EEA. Their ENS condition was diagnosed with a combination of the ENS6Q, SNOT-22, and NOSE score, along with evaluation of their clinical history, and their imaging.

Table 1.

Summary of demographics, clinical scores and results.

Patient	EEAw/oENS	EEAwENS	Non-EEA ENS	Controls
N	2	2	27	42
Gender, n
Male	-	2	19	15
Female	2	-	8	27
Age (mean, years)	54	48	43.59	31.69
FU (mean, months)	21	48	93	-
ENS6Q score (0–30), mean	4	12.5	19.78	-
Dryness (0–5)	1.5	3	3.70	-
Lack of air sensation (0–5)	1	1.5	3.96	-
Suffocation (0–5)	-	0	3.52	-
Nose feels too open (0–5)	-	2.5	3.41	-
Nasal crusting (0–5)	1.5	5	2.78	-
Nasal burning (0–5)	-	0.5	2.41	-
CSA (mean, cm2)	3.79	3.47	3.62	1.34
TITV (mm3)	8175.83	8389.355	-	-
RITV (mm3)	5068.225	4488.82	-	-
LITV (mm3)	3107.65	3900.535	-	-
LADi (%)	51.25%±3.3***	17.74%±4***	25.8±17.6	36.5±15.9
Peak WSSi (Pa)	0.61±0.03**	0.30±0.13**	0.58±0.24	1.18±0.81

N: Total number. M: Male, F: Female, EEAwENS: Patients with endoscopic endonasal approaches and ENS symptoms, EEAw/oENS: Patients with endoscopic endonasal approaches, without ENS symptoms. FU: Follow-up time, ENS6Q: Empty Nose Syndrome 6-items questionnaire. CSA: Cross-sectional Area. TITV: Total mean volumetric measurement of both inferior turbinates. RITV: Mean volumetric measurement of the right inferior turbinates. LITV: Mean volumetric measurement of the left inferior turbinates. LADi: Local airflow distribution in the inferior nasal airway. WSSi: Wall shear stress along the inferior nasal airway.

^**P< 0.01

^***P< 0.0001

A healthy control group (n=42) was also obtained from the previously published study^{2, 5, 6} (Table 1). These subjects ranging from 21 to 60 years (median 27 years) received a research CT-scan, filled SNOT-22 and NOSE questionnaires. This group underwent evaluation and medical history to rule out those with preexisting nasal conditions or previous nasal surgery.

CT scan and CFD model

The CFD nasal airway models were created for each subject, using techniques already described in detail by Zhao et al⁷ and Li et al.² To summarize, each CT or MRI scan was handled with commercial software AMIRA (Visualization Sciences Group, Hillsboro, OR) to create a 3-D model of the nasal cavity, with delineation of the air-mucosa interface. The resulting product was processed with the commercial software package ICEM CFD (Ansys, Inc., Canonsburg, PA), in order to generate tetrahedral elements inside the model. Then a 4-layer prism mesh was included in the geometry, and ANSYS Fluent 16.2 (Ansys, Inc., Canonsburg, PA) was employed to simulate restful breathing, calculating a pressure drop between the nares and the rhinopharynx of 15 Pa.^5,8–15 The final mesh used to simulate the nasal cavity ranged from 1.55 million to 3.6 million hybrids finite elements. A finite-volume method was employed to solve the equations of continuity and momentum, while a second-order upwind scheme for numerical simulation approximated the fields of pressure and velocity. Pressure and velocity were coupled through the SIMPLEC algorithm. A comparison to experimental measurements validated these numerical methods.¹⁶ Airflow rate (ml/s) is calculated at the surface integration of the perpendicular velocity component, and subsequently, the percentage of flow in each region was calculated. Each of the EEA patients had cross-sectional areas; local airflow distribution and peak wall-shear stress (WSS) quantified and statistical comparisons were made between groups with independent two-tailed T-Tests.

RESULTS

The population of this study comprised two men and 2 women (Table 1). Their ages ranged from 40 to 56 years, with a median of 54. All of these patients had a history of EEA including middle and superior turbinectomy, posterior septectomy, and a wide sphenoidotomy along with other procedures as required per case (Table 2). The time from surgery-to-study ranged between 15 to 60 months, with a median of 31 months. Their baseline pathologies included a skull base meningioma invading the optic canal, a middle cranial fossa encephalocele, and two recurrent pituitary adenomas, one of the latter infiltrated the cavernous sinus. Two of these patients had significant ENS symptoms, with a mean ENS6Q score of 13 (EEAwENS). Conversely, the other two had a mean ENS6Q score of 4 (EEAw/oENS).

Table 2.

Summary of performed EEA procedures, along with resected endonasal structures.

Patient	Performed EEA	MT	IT	ST	PS	A	Eth	Sph	Recon	ENS6Q score (0–30)
1	TPt	L- T	-	L - P	E	L- E	L - P	W	NSF	2
2	TS, TT/TP	R- T	-	R - P	S	R- E	RT - LP	W	CoMa, NSF	6
3	TS	R- T	LR- Pa	LR - T	S	R- S	RT - LT	W	CoMa, LPRF	13
4	TS	R- T	LR- Pa	R - T	S	R- S	R - T	W	CoMa, Mgraft	12

EEA: Endoscopic endonasal approaches. TPt: Transpterygoid approach. TS: Transsellar approach. TT/TP: Transtuberculum transplanum approach. FU: Follow-up time. MT: Middle Turbinectomy. IT: Inferior Turbinectomy. ST: Superior Turbinectomy. PS: Posterior Septectomy. A: Antrostomy. Eth: Ethmoidectomy. Sph: Sphenoidotomy. Recon: Reconstruction method of skull base defect. L: Left. R: Right. LR: Bilateral. T: Total. P: Partial. E: Extended. S: Standard. W: Widened. NSF: Nasoseptal flap. CoMa: Collagen Matrix. Mgraft: Mucosal free graft from resected middle turbinate. LPRF: Leucocyte-and-platelet-rich fibrin.

^aminor surgery as endosed by clinical, imaging and statistical evaluation.

A coronal cross-section was selected at equal distance in all models to quantify local airflow distribution (inferior, middle and superior). All EEA patients underwent expansive procedures, and as expected, there was not any significant difference in the cross-sectional areas between EEAwENS (3.47cm² ± 0.23) and EEAw/oENS (3.79cm² ± 1.20). A significantly lower percentage of airflow was observed through the inferior region of EEA patients with significant ENS symptoms (EEAwENS: 17.74%±4.00% vs EEAw/oENS: 51.25%±3.33%, p<0.002) (Figure 2). For the EEAw/oENS group, the airflow was more uniformly spread along the IT area, while the symptomatic EEAwENS group presented a preponderance of airflow peaks upwards, above IT (Figure 1). These findings follow the trend previously observed when comparing non-EEA ENS patients to healthy controls (non-EEA ENS: 25.8±17.6%; healthy: 36.5±15.9%)² (Figure 2).

Figure 2.

Comparative analysis of airflow distribution and peak wall shear stress in all groups. (A) Coronal CT of patient 4 included in the group of EEA with ENS symptoms. (B) Distribution of airflow peaks in patient 4 from the group of EEA with ENS symptoms, with predominant air movement in the middle and superior nasal passageways. (C) Airflow peaks as presented by patient 2 included in the group of EEA patients without ENS symptoms, with inferior and superior predominance of airflow distribution. (D) Representation of the division of airway passage into superior, middle and inferior segments to compare regional airflow distribution (E) Distribution of regional airflow percentage in EEA patients, with significantly reduced inferior air movement in patients with ENS symptoms. Airflow rate is calculated as the surface integration of perpendicular velocity component. (F) Regional airflow distribution of Non-EEA ENS patients, with predominant air movement in middle and superior areas, and significantly diminished airflow and peak WSS in the inferior nasal passageway (G) Healthy controls present a uniform distribution of the airflow along the nasal cavity, including the inferior areas, in similarity to EEA patients without ENS symptoms (H) Regional airflow distribution of Non-EEA ENS patients, compared to healthy controls with predominant air movement in middle and superior areas, and significantly diminished airflow in the inferior nasal passageway

In order to visualize and compare the nasal airflow pattern, the streamlines, from 100 uniformly distributed buoyant seeds, were evaluated (Figure 3). The obtained streamline patterns across the EEAwENS subjects were very similar to the pattern observed in non-EEA ENS patients. They showed an airflow jet turning narrowly towards the middle meatus, leaving the IT area with minimal airflow. This pattern was significantly different in EEAw/oENS subjects and healthy controls, which presented distributions of the airflow along both the inferior and middle turbinate regions.

Figure 3.

Spatial distribution of airflow streamlines in a patient of each group, colored according to the velocity magnitude. A similar pattern of streamlines is observed in patients with ENS symptoms with a narrow airflow (see thick arrows) directed to the middle meatus, and located in the middle corridor of the nasal cavity (A, C). Contrastingly, the streamlines in patients without ENS symptoms and healthy controls are more uniformly found in the middle, superior and inferior regions. Special emphasis is made to the inferior airflow current present only in asymptomatic patients (thin arrows) (B, D). Pt: Patient.

WSS corresponds to the shear force exerted onto a unit area of nasal mucosa by airflow. The peak value was calculated in three areas: Anteriorly, in the nasal vestibule, and posteriorly in the inferior and middle air passages, at the level of the IT and MT respectively (Figure 4). As with the nasal airflow, a significantly lower peak WSS was observed in the inferior region, in the EEAwENS group (EEAwENS: 0.30±0.13 vs EEAw/oENS: 0.61±0.03, p=0.003). These findings follow the same tendency observed while comparing WSS in non-EEA ENS patients to controls (non-EEA ENS: 0.58±0.24; healthy controls: 1.18±0.81 Pa)² (Figure 4).

Figure 4.

Distribution of wall shear stress along the nasal passageway of EEA patients 1 (A) and patient 2 (B) without ENS symptoms and EEA patients 3 (C) and patients 4 (D) with ENS Symptoms. (E) The peak WSS presented a significant reduction in the inferior compartment of EEA patients with symptoms, as compared to those without symptoms. (F) The Peak WSS was compared between the anterior, middle and inferior areas of the nasal cavities, as shown in the model of patient 2. (G) Previously published data (Li, 2017) comparing Peak Wall Shear stress in Non-EEA ENS patients and controls, showing also a significant reduction of the peak WSS in in the inferior compartment of patients with ENS. White line arrow: Presence of WSS in the inferior compartment of the nasal cavity. White dashed arrow: Absence of WSS in the inferior nasal cavity. WSS: Wall Shear Stress

Both patients from the EEAwENS group had a history of minor sub-mucosal reduction of the ITs, which was hardly evident at imaging evaluations. Still, volumetric measurements of the ITs (head, body, and tail) were compared to their counterparts in EEA w/o ENS patients (Figure 5), and they are very similar (8175.83 mm³ Vs. 8393.35mm³). Due to the small sample size, we were not able to declare this lack of difference being statistical significant..

Figure 5.

Example of volumetric measurement of the inferior turbinates. Shown in the image is Patient 2, with a history of minor inferior turbinate reduction.

DISCUSSION

ENS is a disease surrounded by uncertainty. Its clinical appearance as reported by patients, remains as the basis for diagnosis, as other objective criteria are lacking. On the other hand, a frequent association to psychiatric comorbidities such as severe depression and anxiety¹⁷ has led to skepticism over the disease´s origin by some health professionals.

Since Kern coined the term “Empty Nose Syndrome,” this disease has always been linked to diminished turbinate volume, mostly after procedures involving resection of the inferior turbinate.^{18, 19} Albeit on a lesser, not yet defined incidence, middle turbinate resection has also been associated to the disease.^{2, 3} On healthy patients the partition of the nasal airflow against the turbinates allow uniform distribution across the nasal cavity, in an adequate length of time for air-mucosal stimulation. This phenomenon not only permits nasal conditioning (humidification, warming and filtering of the inspired airflow, nasal resistance), but also proper activation of the cold-receptors TRPM8 (responsible for perception of nasal patency²⁰).

In contemporary literature, ENS is considered to occur by a multifactorial interaction of an aberrant nasal airflow pattern and loss of sensorineural function. Both processes are altered in ENS patients after surgery, ^{1, 2} with a significant decline of the cold-receptor TRPM8 function, as observed in menthol lateralization detection threshold testing.^{1, 2, 21} Likewise, previous studies on patients with ENS have confirmed a sui-generis pattern of airflow at CFD analysis (as observed in the data of the nonEEA-ENS group), where the air current brusquely deviates to the middle meatus, decreasing air transport across the inferior nasal compartment.^{1, 2} This phenomena is counterintuitive, since these patients typically have a history of IT resection, hence also having increased inferior nasal airspace for air to circulate. Additionally, Li et al² also determined that the lower peaks of airflow and WSS along the inferior nasal cavity were correlated to higher scores on the ENS6Q. But, it hasn’t been defined if patients experiencing ENS after a middle turbinate resection share the same mechanism.

Currently, the majority of skull base surgeons favor EEA techniques over the traditional open approaches requiring craniotomy, postoperative scarring and the perils of brain retraction. But in order to facilitate and take complete advantage of the 2-surgeons, 4-handed approach, an expanded endonasal resection is usually required.^{22, 23} This frequently involves resection of the middle turbinate, superior turbinates, and the posterior septum. This approach often employs a total ethmoidectomy, wide sphenoidotomy and diverse reconstructive techniques, including the harvesting of nasal mucosal flaps. By itself, a total middle turbinectomy shouldn´t translate into important changes to the overall total airflow, nor in a high reduction of nasal conditioning capacity compared to total inferior turbinectomies. ²⁴ But still, it could create a significant localized redistribution of the airflow that, combined with a reduction of the number of cold-receptors, could potentially explain ENS development, ²⁵ particularly, if the receptors are equally distributed on the surface of both the ITs and MTs.²⁶ Notwithstanding, significant nasal morbidity is rarely observed in these patients after six months, ²⁷ while ENS is generally not reported at all.²⁸

An overall rise in disease awareness has led to an increased search for knowledge of its origins and for a more accurate diagnostic methodology. The ENS6Q questionnaire has been validated for ENS diagnosis with high sensitivity (86.7%) and specificity (96.6%)⁴, with an advantage of simplicity. This short questionnaire evaluates some of the most common symptoms of the disease (paradoxical nasal obstruction, nasal crusting, nasal dryness, suffocation, burning and a sensation of nasal openness) but relies on patient’s subjectivity. Additional, and potentially more objective diagnostic testing could account for a recently improved understanding of the disease’s physiopathology.

This study suggests that some patients that undergo EEA do report significant ENS symptoms. Most importantly, the EEAwENS group had a similar nasal airflow pattern as the one observed in nonEEA-ENS patients. On the other hand, patients that undergo EEA without presenting ENS symptoms had very different airflow patterns, more closely resembling those observed in controls. It is very significant that similar airflow findings were found at very different postoperative scenarios (EEA procedures without inferior turbinate changes vs. less aggressive procedures with inferior turbinate changes) and conversely, very different airflow patterns were found among patients with similar procedures (EEAs). As it appears, the variables capturing this aerodynamic variance correlated with the presence or absence of ENS clinical symptoms.

CFD studies in skull base patients are scarce, which is not surprising as endoscopic skull base surgery is a fairly young subspecialty that developed almost at the same time that CFD studies grew in the rhinology field. One of the few available works comes from Frank-Ito et al, who examined nasal airflow changes after artificially created endoscopic skull base approaches on three patients with and without septal deviation. ²⁹

In the present study, CFD was performed on patients that underwent actual EEA procedures, with documented symptoms evaluation. A limitation of this study is the small sample of the population, which can be addressed with future larger scale prospective studies. Another factor to take into account is that both patients of the EEAwENS group had a history of minor inferior turbinate surgery. But, although minor surgical procedures can devolve into ENS, the imaging and statistical evaluation of these inferior turbinates volume suggests that they were not very different than their unblemished counterparts (Table 1). Interestingly, the two patients with ENS symptoms also received slightly less aggressive approaches (see Figure 1 and Table 2). It is possible that postoperative anatomy does not completely dictate the aberrant pattern of airflow that abandons the inferior nasal area, giving further traction to natural predisposition as a possible explanation of origin. Furthermore, the submucosal nature of the reported procedure decreases the possible contribution of damaged mucosa to the observed symptoms

Conclusions

The pattern of airflow in patients with ENS symptoms appears to be similar, regardless of the expansiveness of the procedure (FESS vs. EEA), and different from controls or non-ENS patients. These results suggest that the aberrant pattern of airflow observed in ENS patients after inferior turbinate surgery, may also be seen in the context of EEA procedures. Further studies are required to confirm that airflow changes are inexorably correlated to ENS symptoms, regardless of the type of surgery that originated them. And, a potential confirmation of these findings could support CFD as an objective and dynamic test to diagnose ENS, within instances of diagnostic doubt, pushing us beyond the anatomical or dimensional evaluation.