Does Total Inferior and Middle Turbinectomy Always Lead to Empty Nose Syndrome?

Abstract

Objective

Empty Nose Syndrome (ENS) is a rare and paradoxical complication often associated with aggressive turbinate reductions. Previous computational fluid dynamic (CFD) modeling has indicated that distorted nasal airflow patterns could contribute to ENS. However, no data have shown that aggressive turbinate reductions consistently lead to ENS.

Study Design

Retrospective case series.

Setting

Tertiary.

Methods

We retrospectively recruited 6 nasal obstruction patients who underwent turbinate reduction surgery, and performed total inferior and middle turbinectomies on all patients using a published virtual surgical simulator. We compared the CFD modeling results to that of patients' presurgery and postsurgery and published 27 ENS patients.

Results

As expected, the virtual turbinectomy significantly expanded nasal airway cross‐sectional area, more than that of actual surgery (inferior: 0.9 ± 0.4 to1.8 ± 0.3 cm², middle: 0.6 ± 0.3 to 1.0 ± 0.3 cm², all P < .01); however, it does not create the same distorted nasal airflow patterns as seen in ENS patients, with no significant difference in nasal resistance after actual surgery (0.12 ± 0.04 Pa/mL*s) versus virtual surgery (0.10 ± 0.03 Pa/mL*s) nor ENS (0.11 ± 0.04 Pa/mL*s, all P > .05). However, all had significantly higher inferior wall shear force (WSF) distribution, an important indicator of air/mucosa stimulation, than that of ENS patients (baseline 54.0 ± 11.9%, surgery: 51.5 ± 15.1%; virtual: 46.5 ± 11.5%, P > .05; ENS: 32.2% ± 12.5%, P < .001). The opposite was seen for the middle turbinate region—WSF among all actual or virtual surgeries conditions was significantly lower than that of the ENS group (baseline: 36.1 ± 11.0%; virtual: 30.2 ± 13.1%; ENS: 43.8 ± 10.1%; all P ≤ .05).

Conclusions

ENS cannot be solely attributed to aggressive turbinate reduction surgeries with various factors potentially playing a role.

Empty Nose Syndrome (ENS) is a rare, debilitating condition. ¹ Patients who suffer from this condition experience a variety of symptoms, including nasal congestion or obstruction, paradoxical dryness, burning, suffocation, crusting, and a sensation of their nose being too open. ² , ³ , ⁴ , ⁵ Patients may have difficulty describing their symptoms, but the paradoxical sense of their nose being too open yet feeling suffocation or obstruction is the most common experience. ⁶ This coupled with psychological symptoms of depression or anxiety, as well as computerized tomography (CT) scans indicating no obstructive tissue can be difficult to manage for both patients and providers.

While the exact pathogenesis of ENS is still unknown, it was thought that ENS could be attributed to aggressive turbinate surgery ¹ that result in: (1) changes in anatomy that alters the local nasal environment (2) disruption of mucosal cooling and drying, and (3) changes in the underlying mechanism of neurosensation. ⁷ Our lab previously reported that both nasal aerodynamics and trigeminal sensation is paradoxically distorted in ENS patients, and correlate to patients' symptoms. ⁸ ENS can be categorized into ENS‐IT (absence of inferior turbinates, ENS‐MT [absence of the middle turbinates], and ENS‐both, ⁹ however the characterization of these subgroups are not widely accepted.

Endoscopic sinus surgery (ESS) is a common procedure used to treat sinus diseases such as chronic rhinosinusitis, nasal polyps, benign nasal tumors, and structural issues such as a deviated septum or enlarged turbinates. ¹⁰ Since ESS is an invasive procedure, it is usually a last resort after other treatments (steroid sprays, irrigation, etc.) have failed to relieve symptoms. ¹⁰ Although the surgeon relies on visual and haptic feedback from the instruments to perform the surgery, no technique currently exists to accurately predict sinus surgery treatment outcomes. Due to the hypothesis that ENS could be caused by aggressive turbinate surgery, some studies have recommended preserving at least 50% of the turbinates to minimize the risk of developing ENS, ⁹ however, this is not based on objective findings. This coupled with its ability to predict distorted airflow within the nasal airway suggests that virtual surgery planning (VSP) could be an effective method at preventing ENS. In this study, we attempt to use VSP to investigate whether aggressive turbinate surgeries can consistently produce airflow distortions seen among ENS.

Methods

Virtual Surgery Planning System

Our research team have previously developed a VSP prototype that can load patient CT scans and create a highly detailed 3D visual display of the sinonasal airway (Figure 1). ¹¹ Surgeons currently plan sinus surgery based on visual assessment of CT scans and endoscopic findings. However, predicting changes in airflow dynamics solely based on CT and endoscopy findings is challenging. This simulator incorporates force‐feedback (haptic) device that provides interaction between a virtual endoscopic tool and 3D CT‐based models. ¹¹ This system was validated by a rhinology/skull base attending, who conducted virtual nasal valve surgery, septal body volume reduction, and bilateral inferior turbinate reduction. After each virtual surgery, the updated scan was exported, and computational fluid dynamics modeling and analysis was validated. ¹¹

Figure 1.

Screenshot of performing virtual turbinectomy surgeries using a virtual surgery planning system with built‐in image guidance (right panel). A microdebrider with real‐time force‐feedback allows virtual tissue removal. Post virtual‐surgery CT scans are shown below, which is then exported for CFD modeling and analysis. CFD, computational fluid dynamic; CT, computerized tomography.

Recruitment

The Ohio State University Institutional Review Board approved our human subjects protocol for this study (2015H0262). We retrospectively recruited 6 patients that had undergone turbinate reduction surgery and did not develop symptoms of ENS. We excluded patients with severe inflammation, prior head trauma, significant atopy, viral‐related smell loss, cystic fibrosis, Wegener's, or other connective tissue disorders. Presurgical CT scans were uploaded into the VSP system and a bilateral total inferior (ITR) and middle turbinectomy (MTR) was performed for each patient. These virtual surgeries were reviewed by a rhinologist to confirm surgical techniques and accuracy. Once this was complete, presurgery, postsurgery, and postvirtual surgery CT scans were used to create 3D computational fluid dynamics models. These models were compared to a collection of 27 ENS patients data from our previous studies to determine if similar airflow patterns were found after aggressive inferior and middle turbinectomy as ENS ¹² , ¹³ , ¹⁴ and differ to the actual successful surgery.

Computational Fluid Dynamics

Patient CT scans were imported into AMIRA (Visualization Sciences Group) software. ¹⁴ Each 3D model of the nasal cavity was created from coronal, lateral, and axial images. The results were imported into the ICEM CFD software (Ansys, Inc.) to generate the interior airway portion of the model and a 4‐layer boundary mesh. Once this was complete, computational analysis was conducted in ANSYS Fluent version 16.2.

Statistical Analysis

Data analysis was performed via repeat t‐tests to compare all variables between previrtual and postvirtual surgery independent variables (nasal resistance, cross‐sectional area). Changes in airflow rate between virtual surgeries and ENS patients' data were compared as well.

Results

Among 6 patients with nasal obstruction, all underwent bilateral submucous inferior turbinate resection; four also had septoplasty, 1 had right maxillary antrostomy with partial ethmoidectomy, and 1 had an out‐fracture (Table 1). Unilateral Visual analog scale of nasal obstruction (VAS: 0 being no obstruction at all, and 10 being completely obstructed) and NOSE scores significantly decreased from baseline to postsurgery (Table 1: VAS: 6 ± 2.5 to 1.2 ± 1; NOSE: 72.5 ± 13.1 to 10.8 ± 9.8; both P < .001), indicating good surgical outcomes.

Table 1.

Demographics, Symptom Scores, and Surgeries of the Study Cohort

Demographic		Number (n = 6)
Mean age (range)		40 (23‐54)
Ethnicity
Latino (%)		1 (17%)
Not Hispanic or Latino (%)		5 (83%)
Sex
Male (%)		4 (67%)
Female (%)		2 (33%)

Symptom Score (n = 6)	Baseline	Surgery	P
VAS (Unilateral)	6 ± 2.5	1.2 ± 1	<.001
NOSE	72.5 ± 13.1	10.8 ± 9.8	<.001

Patient	Surgery performed
Patient	Surgery performed
1	Septoplasty, bilateral submucous resection of inferior turbinates (IT)
2	Septoplasty, bilateral submucous resection of IT
3	Septoplasty, bilateral submucous resection of IT, right maxillary antrostomy, partial ethmoidectomy
4	Nasal septum fracture repair, bilateral submucous resection of IT and out‐fracture
5	bilateral submucous resection of IT
6	Septoplasty, bilateral submucous resection of IT

Bold values indicate statistical significance.

Table 2 summarizes changes in cross‐sectional area, resistance, airflow, and wall shear force (WSF). As expected, the virtual ITR significantly increased cross‐sectional area in the inferior regions (inferior: 0.8 ± 0.3 to 1.9 ± 0.4; P < .05), more than that of the actual surgery (inferior: 0.9 ± 0.4, P < .001). The ITR + MTR further significantly enlarged the middle turbinate regions (middle: 0.47 ± 0.18 to 1.0 ± 0.3 cm²; P < .001), more than that of the actual surgery (middle: 0.6 ± 0.3 to 1.0 ± 0.3 cm²; P < .01). However, no significant differences were found between virtual ITR + MTR and ENS group, which indicated that these ENS patients indeed underwent aggressive reduction to both the inferior and middle turbinate regions and that our virtual ITR and MTR adequately captured the aggressiveness in surgeries.

Table 2.

Comparison of Variables Between Baseline, Postsurgery, Postvirtual Inferior Turbinectomy, and ENS Patient Cohort

			Cross‐section 4.5/10 (45% slice location)
Variables	Baseline (B) (n = 6)	Surgery (S) (n = 6)	Virtual ITR (ITR) (n = 6)	Virtual ITR + MTR (ITR + MTR) (n = 6)	ENS (n = 27)	p (B vs S)	p (B vs ITR)	p (B vs ITR + MTR)	p (B vs ENS)	p (S vs ITR)	p (S vs ITR + MTR)	p (S vs ENS)	p (ITR vs ENS)	p ( ITR vs ITR + MTR)	p ( ITR + MTR vs ENS)
Resistance (pa/mL/s)	0.21 ± 0.13	0.12 ± 0.04	0.10 ± 0.03	0.10 ± 0.03	0.11 ± 0.04	0.06	0.01	0.01	<0.001	0.16	0.26	0.49	0.46	0.25	0.47
			Nasal cross‐sectional area (cm2)
Inferior	0.8 ± 0.3	0.9 ± 0.4	1.9 ± 0.4	1.8 ± 0.3	1.2 ± 1.0	0.47	<0.001	<0.001	0.18	<0.001	<0.001	0.28	0.03	0.32	0.09
Middle	0.47 ± 0.18	0.6 ± 0.3	0.5 ± 0.2	1.0 ± 0.3	1.8 ± 1.7	0.34	0.05	<0.001	0.01	0.42	0.02	0.02	0.01	<0.001	0.17
Superior	0.21 ± 0.18	0.2 ± 0.3	0.2 ± 0.2	0.2 ± 0.2	0.7 ± 0.7	0.79	0.72	0.29	0.02	0.79	0.90	0.03	0.02	0.30	0.05
			Cross‐sectional airflow rate (mL/s)
Inferior	42.0 ± 26.5	82.0 ± 27.6	130.0 ± 41.7	98.6 ± 41.2	36.6 ± 25.2	0.003	<0.001	0.002	0.51	0.002	0.12	<0.001	<0.001	0.01	<0.001
Middle	48.3 ± 32.8	57.0 ± 29.7	44.3 ± 35.1	83.1 ± 63.4	103.8 ± 49.7	0.43	0.21	0.02	<0.001	0.27	0.37	0.002	<0.001	0.003	0.26
Superior	2.5 ± 1.9	1.9 ± 2.8	1.3 ± 1.0	0.8 ± 1.2	11.9 ± 15.1	0.59	0.01	0.01	0.04	0.51	0.30	0.03	0.02	0.35	0.03
			Wall shear force distribution (%)
Anterior	10.0 ± 4.7	16.0 ± 5.4	25.8 ± 7.0	23.3 ± 4.0	22.7 ± 7.6	0.006	<0.001	<0.001	<0.001	0.002	0.009	0.006	0.21	0.54	0.82
Inferior	54.0 ± 11.9	51.5 ± 15.1	47.3 ± 11.3	46.5 ± 11.5	32.2 ± 12.5	0.54	0.06	0.05	<0.001	0.33	0.53	<0.001	<0.001	0.57	0.002
Middle	36.1 ± 11.0	32.4 ± 15.0	26.9 ± 11.4	30.2 ± 13.1	43.8 ± 10.1	0.30	0.004	0.04	0.02	0.13	0.31	0.002	<0.001	0.04	<0.001

Variables that are significantly different between the groups are bolded.

Abbreviations: ENS, Empty Nose Syndrome; ITR, inferior turbinectomy; MTR, middle turbinectomy.

Nasal resistance were significantly lower post different surgeries (actual and virtual) than that of presurgey (0.21 ± 0.13 Pa/mL*s; P < .01), but with no significant difference between the different surgical conditions and the ENS groups (Figure 2: actual surgery: 0.12 ± 0.04 Pa/mL*s; virtual ITR: 0.10 ± 0.03 Pa/mL*s; virtual ITR + MTR: 0.10 ± 0.03 Pa/mL*s; ENS: 0.11 ± 0.04 Pa/mL*s; P > .05). This is consistent with increased airway caliber and may reflect a ceiling effect that more aggressive turbinate reduction even with the added middle turbinectomy didn't decrease the resistance any further.

Figure 2.

Nasal resistance, flow rate, cross‐sectional area and wall shear force distribution within different nasal airway regions among baseline, postactual, and virtual surgeries, and ENS patients (n = 27).

In Figure 2, airflow rates showed significant group differences: inferior region airflow increased significantly from baseline (42.0 ± 26.5 mL/s) to actual surgery (82.0 ± 27.6 mL/s) and further significantly to virtual ITR (130.0 ± 41.7; all P < .01), but with decreased superior airflow (2.5 ± 1.9 to 1.3 ± 1.0 mL/s; P < .01). Similarly, middle region airflow increased significantly from baseline (48.3 ± 32.8 mL/s) to actual surgery (57 ± 29.7 mL/s) and further significantly to after ITR + MTR (83.1 ± 63.4 mL/s). However, ENS patients differed significantly from all virtual and actual surgical groups across most regions. They surprisingly have significantly lower inferior region airflow rate (ENS inferior: 36.6 ± 25.2 mL/s), which is even slightly lower than presurgery, but with significantly higher middle airflow rate (103.8 ± 49.7 mL/s), except when compared with virtual ITR + MTR (83.1 ± 63.4 mL/s, P = .26), still higher but not reaching significance. An example patient with coronal CT cross‐sectional scans and CFD model cross‐sections is shown in Figure 3 (Left to right: baseline, post‐actual surgery, post virtual ITR, post virtual ITR + MTR, and an ENS patient).

Figure 3.

An example patient with CT cross‐sections and corresponding CFD for baseline, postactual surgery, virtual ITR, virtual ITR and MTR, and an ENS patient. CFD, computational fluid dynamic; CT, computerized tomography; ENS, Empty Nose Syndrome; ITR, inferior turbinate; MTR, middle turbinectomy.

WSF distribution, an important indicator of air/mucosa stimulation, also demonstrated significant regional changes (Figure 2). In the anterior region, WSF increased significantly from baseline to all other groups (10.0 ± 4.7 to 16.0‐25.8; all P < .01) and significantly differed between the actual surgery and both virtual and ENS groups (P < .01). No significant differences were observed between the 2 virtual models (P > .5). However, Inferior WSF distribution does not significantly differ between baseline, postsurgery and post virtual ITR + MTR (baseline 54.0 ± 11.9%, surgery: 51.5 ± 15.1%; virtual: 46.5 ± 11.5%, P > .05), but all are significantly higher than that of ENS patients (WSF: 32.2% ± 12.5%, P < .001). In contrast, WSF in the middle turbinate region is reversed—all actual or virtual surgeries conditions was significantly lower than that of the ENS group (baseline: 36.1 ± 11.0%; virtual: 30.2 ± 13.1%; ENS: 43.8 ± 10.1%; all P ≤ .05).

Collectively, these findings suggest that ENS patients exhibit marked regional differences in nasal aerodynamic and mucosal shear forces, even when comparing with most aggressive ITR + MTR, reflecting a significantly distortion in air‐mucosa interaction that may lead to ENS symptomology.

Discussion

This study expands on our previous work that examined how aggressive turbinate reductions, including inferior turbinectomy, reshape nasal airflow. ¹⁵ Using computational fluid dynamics, we further compared baseline, postsurgical, virtual inferior turbinectomy [ITR] and combined virtual inferior + middle turbinectomy [ITR + MTR] against a cohort of ENS patients to better define how anatomy and airflow interact after turbinate losses.

Structurally, actual surgery did not markedly change cross‐sectional areas, yet it produced clear symptomatic improvement: unilateral VAS scores fell from 6 ± 2.5 to 1.2 ± 1 and NOSE scores from 72.5 ± 13.1 to 10.8 ± 9.8 (both P < .001), confirming substantial relief of obstruction. Whereas virtual turbinectomy significantly enlarged the inferior and middle regions (all P < .01), more than the actual surgery. Nasal resistance dropped compared with baseline (P < .01) but plateaued across surgical and virtual conditions, suggesting a physiological ceiling beyond which further enlargement offers limited additional resistance relief.

Airflow patterns told a more nuanced story. Virtual ITR + MTR increased flow through the inferior and middle regions significantly compared to actual surgery, yet none of the virtual models reproduced the distinctive, unbalanced flow distribution seen in ENS patients, who exhibited exaggerated middle region airflow while paradoxical significant reduction of inferior region airflow. Thus, while tissue removal predictably alters flow magnitude, it does not consistently recreate the pathological airflow signature of ENS, implying that anatomic loss alone cannot explain the disorder. Figure 4 shows the comparison of airflow jetstream of a virtual ITR + MTR patient and an ENS patient that clearly displayed different patterns.

Figure 4.

Airflow streamline patterns in one ENS and one virtual ITR + MTR patient. There was a jetstream towards the middle meatus region in the ENS patient, but not seen in the virtual ITR + MTR patient.

WSF analysis underscored this divergence. In the anterior region, WSF rose sharply across all surgical and virtual models (10.0 ± 4.7 to 16.0‐25.8; all P < .01), while the inferior region showed progressive declines from baseline to ENS (54.0 ± 11.9 to 32.2 ± 12.5; P < .001). The middle region displayed the highest WSF in ENS compared with all other groups (P ≤ .05), highlighting a regionally distorted air‐mucosa interaction unique to ENS. Collectively, these data indicate that ENS is defined not only by simply airflow velocity distortions, but by an uneven redistribution of shear forces, especially within the middle region, potentially driving the abnormal nasal sensations patients describe.

Overall, aggressive turbinate reduction significantly reshapes nasal aerodynamics but fails to replicate the distorted airflow and shear force profiles of ENS. This suggests that ENS pathophysiology extends beyond the aggressive surgery only, likely involving individual unique anatomical feature that may be prone to airflow distortion after aggressive surgery. Altered mucosal feedback, disrupted sensory signaling, or changes in mucosal hydration and temperature regulation are other potential confound factors to the disease.

The study's limitations include a small sample size (n = 6) and the time‐intensive nature of virtual modeling (1‐2 days per case). Moreover, computational models cannot directly capture neural or sensory dysfunction. Future studies combining CFD with mucosal thermosensory, humidity, and receptor analyses are needed to clarify how structural and neurophysiologic factors converge to produce ENS symptoms. Finally, total turbinectomies were included solely as a modeling scenario to represent the most extreme surgical approaches. We certainly do not advocate for total turbinectomy and surgical decisions should be individualized and guided by patient‐specific clinical needs.

Conclusion

The virtual turbinectomy significantly altered nasal cross‐sectional area of the inferior and middle turbinate regions, more than that of actual surgery, and produce alterations in airflow patterns within the nasal airway. However, there was significant differences in airflow and WSF distributions when comparing the virtual turbinectomy to ENS patients. Future directions should investigate the impact of aggressive turbinate surgery on nasal airflow distribution on a larger cohort of patients to understand how it impacts the individual nasal cavity differentially. This could provide direction to potentially prevent the paradoxical ENS from occurring in the future. As of now, ENS cannot be solely attributed to aggressive turbinate reduction surgeries with multiple factors potentially playing a role in its disease process.

Author Contributions

Nidhi Jha, wrote the manuscript, presented the findings at AAO‐HNSF, created figures; Jed Speers, did the MT virtual surgeries on the virtual surgery simulation; Ahmad Odeh, did the IT virtual surgeries and recruited patients; Zhenxing Wu, analyzed computational fluid dynamics (CFD) models and compiled data; James Mihalich, created CFD models; Lauren E. Gastineau, recruited patients; Bradley M. Hittle, created the virtual surgery simulator program (VSP); Bradley Otto, consulted about current sinus surgery practices; Kathleen M. Kelly, consulted about current sinus surgery practices; Gregory J. Wiet, consulted about current sinus surgery practices and helped create VSP; Kai Zhao, principle investigator.

Disclosures

Competing interests

None.

Funding source

NIH‐NIDCD R21 DC017530 and R01 DC020302 to KZ.