The cotton test redistributes nasal airflow in patients with empty nose syndrome

Abstract

OBJECTIVE:

Empty nose syndrome (ENS) remains a controversial disease primarily associated with inferior turbinate tissue loss. Cotton placement into the inferior meatus often alleviates ENS symptoms within minutes, but the physiologic explanation for this phenomenon is unknown. Computational fluid dynamics (CFD) was employed to evaluate the mechanisms of altered nasal airflow conferred by cotton testing.

METHODS:

6 ENS patients (12 sides) with pre-existing sinus CT imaging were enrolled following marked symptomatic improvement (ENS6Q decline >7 points) with office-based cotton testing. The fashioned cotton plug was labeled in-situ with iohexol contrast spray, and sinus CT was immediately obtained to detect cotton contouring in the inferior meatus. CT imaging from pre- and post-cotton placement was analyzed using comparative CFD techniques.

RESULTS:

Following cotton placement, significant symptomatic improvement and reduced ENS6Q scores (16.8±4.1 to 3.1±2.4; p<0.001) were recorded. Using CFD, cotton placement produced an expected 21% increase in upper airway resistance (p<0.05). However, a significant shift in the nasal airflow distribution was also detected, with a transition of airflow vectors away from a middle meatus jetstream (−41%; p<0.002).

CONCLUSION:

Objective CFD assessment confirms that the cotton test not only increases nasal resistance, but also restores airflow distribution to the inferior meatus in symptomatic ENS patients. These results provided potential mechanism for the efficacy of cotton test in ENS patients, and further bolster the utility of this tool in identifying appropriate candidates for the inferior meatus augmentation procedure.

Introduction

Empty Nose Syndrome (ENS) is one of the most controversial and rare complications following sinonasal surgeries, and is thought to be an iatrogenic phenomenon following possible over-resection of the inferior turbinate soft tissue^1,2. Typically ENS patients suffer from several symptoms that can become debilitating in nature, which include, but are not limited to: paradoxical perception of nasal obstruction, suffocation, nasal burning, nasal dryness, and nasal crusting³. Due to the severity of these symptoms, ENS patients often struggle with psychosocial issues and can present with comorbid psychiatric problems such as anxiety and depression⁴ that can significantly reduce the effectiveness of conventional treatment⁵. Treatments such as submucosal implants have been shown to restore the spatial contouring of the nasal cavity and improve both rhinological and psychological symptoms of ENS patients⁶. However, little is known about the underlying pathophysiology that may drive the successful procedural outcome and how this understanding can be translated into optimizing future procedures.

Clinically, ENS patients are probably best diagnosed using cotton test⁷. This test involves placement of a dry cotton plug into the region of the inferior lateral nasal wall/inferior meatus to restore turbinate-associated tissue bulk. Remarkably, many ENS patients report improvement in their nasal breathing within minutes in the office, which has been generally ascribed to restored nasal resistance and redirected airflow due to cotton placement. If patients have a >7-point reduction in their ENS6Q score, a validated disease specific questionnaire for ENS, these patients can be considered good candidates for inferior meatal implants. Given these observations and indications, our goal was to further understand the physiologic mechanisms that explain the changes in upper airway breathing following placement of a cotton plug. We hypothesized that the insertion of a cotton plug alters nasal airflow from the characteristic ENS flow patterns^8,9 to a more dispersed normal pattern.

Materials and Methods

Study Population

Following IRB approval, six consecutive patients with a diagnosis of ENS based on a ENS6Q score of >11 ³ and a cotton test that reduced the ENS6Q score by 7 points⁷ were enrolled candidates for the study. They required a pre-existing CT scan of their sinuses and were willing to receive a second research CT sinus scan following cotton placement bilaterally in the inferior meatus. Following the cotton test, the cotton plug in the inferior meatus was sprayed in-situ with iohexol contrast agent, in order to detect the outline of the cotton contour within the inferior meatus during the 2^nd research sinus CT imaging. Blue food coloring was added to the contrast agent to confirm spray coating of the cotton in the clinic, prior to imaging. It should be noted that no further ENS6Q testing or subjective questioning was performed following the use of the contrast spray.

The six patients were 3 males and 3 females, with averaged age of 51.6 ± 22.8 years ranging from 31 to 80 years. Three of the patients were white/Caucasian, two Hispanic/Latino and one Asian. All six patients had reported bilateral symptoms and had a history of multiple nasal surgeries that include a bilateral turbinate resection or reduction in all patients, septoplasty and FESS in 4 of the 6 patients. A detailed outline of the patient surgical history can be found in supplemental Table 1.

Computational Fluid Dynamics (CFD) Modeling

Both the pre- and post-cotton-placement CT scans were imported into AMIRA (Visualization Sciences Group, Hillsboro, OR, USA) software in order to create a 3-dimensional nasal cavity volume. After image smoothing, artifact correction, and visual assessment, the volume was imported into ICEM CFD (Ansys, Inc., Canonsburg, PA, USA) where it was meshed with 1.1 to 3.6 million tetrahedral volume elements as described by Li et al⁹.

ANSYS Fluent 16.2 (Ansys, Inc., Canonsburg, PA, USA) was used to solve the Navier-Stokes equations under the assumption of steady state conditions with incompressible, laminar flow. To simulate restful breathing, the state most relevant to patient symptoms and daily life, a 15 Pa pressure drop between the nostrils and nasopharynx was applied^10,11. The walls of the nasal cavity were assumed to be rigid with a no-slip boundary condition. Finite-volume method was used to evaluate the numerical solutions of the continuity and momentum equations and continuous pressure and velocity fields were discretized using a second-order upwind scheme for numerical simulations. Velocity and pressure were coupled using the SIMPLEC algorithm. A residual ceiling of 10⁻⁵ was set for each variable in order to identify simulation convergence. The numerical method applied in the current study has been experimentally validated in studies by Li et al.¹². Pre- and post-cotton models were run for the ENS patients and the resistance, regional flow rate, and pressure parameters were calculated and compared.

Statistical Analysis

A two-tailed paired t-test without assuming equal-variance was used for post-treatment comparison (pre- and post- cotton plug). For parameters such as resistance, flowrate, and flow distribution the analysis was performed unilaterally whereas parameters such as pressure and ENS6Q were analyzed bilaterally.

Results

First, we measured the location of cotton placement based on the contrast CT scan (Figure 1A, B and Supplementary Figure 1) at 1.74 ± 0.26 cm on average from the posterior edge of the nostril, and cotton plugs measured approximately 1.93 ± 0.35 cm in length (Table 1). Following the cotton placement, significant improvement in the ENS6Q scores among the ENS patients was reported with an average symptom score decrease of 13.6 ± 2.8 points (range 10–17 point improvement, Figure 1C). Prior to the cotton test, patients reported an average ENS6Q score of 16.8 ± 4.17 points, which significantly dropped to 3.1 ± 2.4 upon cotton placement (p < 0.001, paired two-tailed t-test).

Figure 1.

A, B) Cotton Placement in the Inferior Meatus of two ENS patients (patient 1 and 2). Cotton was sprayed in situ with radiocontrast prior to imaging, in order to be detectable by CT. C) Significant improvement of ENS symptoms (reduction of ENS6Q scoring) in all 6 patients following insertion of the cotton.

Table 1.

The location and length of cotton following placement within the inferior meatus of ENS patients.

Patient No.	Cotton Plug Distance from Posterior Edge of Nostril (cm; Left, Right)	Length of Cotton (cm; Left, Right)
1	1.78, 1.78	2.33, 2.33
2	1.39, 1.58	2.08, 2.38
3	1.62, 1.52	1.58, 1.42
4	1.78, 1.59	1.89, 2.18
5	1.79, 1.69	1.58, 1.48
6	2.23, 2.23	2.18, 1.78
AVG ± SD	1.75 ± 0.26	1.93 ± 0.35

Using CFD techniques, we analyzed various parameters that may contribute to this improvement. In Figure 2A, we reported that averaged unilateral nasal resistance significantly increased from 0.12 Pa*s/ml ± 0.07 to 0.15 ± 0.07 upon cotton placement (p < 0.05, paired two-tailed t-test). We also compared the ratio of flowrates between the left and right nostril both before and after cotton placement. On average there was a 16% change in airflow distribution between the left and right nostril once the cotton was inserted with highest shifts of 41% and 35% seen in two patients. Here, we found that there was a positive linear correlation between percentage of nostril airflow shift and change in ENS6Q score (Pearson r = 0.84, p = 0.07) suggesting patients with the highest lateral shift of airflow may have better ENS6q outcome post-cotton insertion. Although this left-right shift was not statistically significant, it should be further considered and evaluated in the future.

Figure 2.

A). CFD analysis showed a significant increase of unilateral nasal resistance (in Pa*s/mL) after the cotton placement. B). Regional airflow rate at 60% of the nasal cavity length is significantly reduced in the middle meatus after cotton placement. * = p <0.05

We further analyze airflow patterns at a coronal plane at ~60% of nasal cavity length from the nostril to be consistent with previously published ENS CFD studies (see Figure 2B and 4B). However, the plane may or may not cut across the cotton in different patients. In Figure 4B, the plane just cut immediately posterior to the cotton and in Figure 4C, the plane cut at the posterior region of the cotton. Prior to cotton insertion, we observe that the airflow predominately flows through the middle meatus region (as seen and labeled as the Middle 1/3 in Figures 4B and 4C), which matches our previously published observation on a different ENS cohort. After cotton insertion, we observe an average 41% decrease in middle meatus region among all patients with the overall airflow pattern becoming more evenly balanced between middle and inferior meatus (Figure 2B; p < 0.05, paired two-tailed t-test, flowrate decreases from 83.3 ± 38.8 ml/s to 48.6± 30.5 ml/s). It should be noted that while we do see a decrease in inferior flow volume (pre-cotton: 73.1 ± 36.1 mL/s; cotton: 57.6 ± 29.1 mL/s) the overall percentage of flow (or its distribution) does not change (pre-cotton: 45.7 ± 15.3%; cotton: 44.8 ± 19.8%). Instead we see a decrease in airflow percentage through the middle meatus (pre-cotton: 51.3 ± 24.4%; cotton: 38.9% ± 18.9%). An increase in airflow distribution through the superior meatus was also observed (pre-cotton: 11.1 ± 7.4%; cotton: 22.1 ± 18.1%)., We found that 5 of the 6 patients presented with said bilateral shift from the middle meatus to the inferior meatus, with the remaining patient showing a unilateral change in the left nasal passage. To further illustrate, the airflow velocity contour was plotted for two patients in coronal slices at baseline and after cotton placement, towards the posterior end of the cotton plug (as seen in Figure 4B, C). Both of these patterns resulted in a shift towards a more equal airflow distribution.

Figure 4.

A. Cotton placement results in a redirection of flow trajectories away from the middle meatus. B, C) Cross section slices taken at the posterior nasal region (at ~60% of nasal cavity length) among two patients show an increase in airflow in the inferior meatus and a redirection of airflow downwards with air velocity dispersed away from the middle meatus after cotton placement

We also see significant flow pathline changes between baseline ENS and cotton placement (Figure 4A). Again, anterior to the turbinates we observe a jet stream upwards towards and through the middle meatus. However, with cotton placement the flow trajectory angles more downward that impacts the mid to posterior nasal airway, which we tried to quantify manually using Tecplot 360 and ImageJ. First, the airflow velocity vectors were plotted in Tecplot 360. The images were then imported into ImageJ, overlaid and the angle between the two primary trajectories, pre and post cotton placement, was calculated using the angle calculation function in ImageJ. The primary trajectory on each image was manually determined based on the main velocity vector pattern. This trend was observed in all patients with degree changes ranging from 4 to 17 degrees.

Intranasal pressure gradient has also been suggested to be an important parameter related to breathing function. This gradient was quantified bilaterally for each patient before and after the cotton insertion (Figures 3A, B). Prior to cotton insertion, the pressure gradient plateaued immediately posterior to the nasal valve at values around −10 Pa before continuing to drop again posteriorly in the nasal cavity (as indicated by the red plot markers in Figure 3B). With the cotton insertion, the pressure gradient becomes more smooth and linear throughout the full length of the nasal cavity. This trend was seen in all patients.

Figure 3.

A). The regional bilateral pressure was calculated for 10 different locations moving from the nasal valve to posterior nasal airway (slice 1 – slice 10). B. Without cotton the pressure drop plateaus (red markers) verses with cotton placement where we see a more linear slope in pressure moving posteriorly.

Discussion

Prior to cotton placement, we have previously shown that ENS patients exhibit a jet-like air stream through the middle meatus with paradoxically restricted airflow to the inferior region of the nasal cavity despite no physical obstruction (Figure 4). This airflow pattern is consistent with previous studies of a different cohort of ENS patients^13,14. Implants that are either permanent or temporary have been suggested to improve ENS symptoms¹⁵. Performance of the cotton test, via directed placement of a cotton plug into the inferior meatus at the sites of turbinate tissue loss, was shown to improve patient reported symptoms of ENS⁷ and further supported in our study with a significant decline of nearly 14 points in ENS6Q scores (Figure 1B). In our study the contrast agent was sprayed only after ENS6Q or subjective test scores had been recorded. We therefore can confirm that the presence of increased moisture from the contrast agent did not contribute to patient reported subjective improvements. However, the physiologic mechanism that governed this rapid symptomatic improvement in nasal symptoms is not very well understood. With respect to overall nasal resistance, we observed a 31% increase in airflow resistance with the cotton test (Figure 2A). This is consistent with reduced nasal cavity airway volume due to placement of cotton plug into the inferior meatus.

The spatial profile of resistance was analyzed via intra-nasal pressure gradient (Figure 3A, B), which demonstrated few differences between baseline and cotton placement at locations anterior to the turbinates (slices 1 – 4). However in the turbinate region, the nasal airflow pressure gradient exhibits a plateau prior to cotton insertion until the nasopharynx (slice 8) where it begins to fall once more. Conversely, once the cotton plug is placed, the pressure gradient profile changed to a smoother and linear pressure gradient throughout the length of the nasal cavity. The existence of a pressure plateau among ENS patients reflects a region of extremely low nasal resistance, likely the result of extensive surgical tissue loss in the nasal cavity. This excess low resistance gradient has been implicated in disruption of pulmonary functions^16,17 and could help explain why ENS patients suffer from perceived suffocation and why there is improvement following cotton testing. Too low resistance (e.g. mouth breathing) has been shown to stimulate and increased presence of leukocytes in the blood, thus increasing lung hypersensitivity and decreasing its volumes and capacity which may result in symptoms of suffocation¹⁸. Furthermore, the low resistance gradient often corresponds with the characteristic jet stream of airflow seen in ENS patients. It is likely that this alteration in flow pattern also results in a compromised filtration, humidification, and heating of the inhaled air.

Also, the difference in the shape and slope of the pressure profile could also suggest that deeper, more forced breaths (which would increase the pressure drop) might be needed to mimic normal, healthy breathing. Mangin et al¹⁹ showed that one potential effect of continual deep breathing is hyperventilation as well as hypocapnia, which is one of the symptoms cited by long standing ENS patients. With the cotton insertion, it is possible that the need to compensate for these pressure discrepancies through deep breathing may no longer be necessary. Our analysis of the spatial discrepancies in pressure profiles is limited, but the preliminary findings may be extended in the future to various targeted flow rate conditions that may help to understand the range of upper airway breathing patterns affecting ENS symptoms.

But resistance is clearly not the only factor in the symptomatology of ENS patients; otherwise the insertion of cotton can be effective at multiple locations, including the middle meatus or septum. Several studies have reported that cotton placement and implants to the lateral nasal wall/inferior meatus yields better outcomes than the septum or middle meatus¹⁵. To better understand the regional airflow changes as a result of cotton placement, we quantified the flowrate parameters in the forms of regional flowrate as well as flow distribution.

As seen in Figure 4A, the insertion of a cotton plug results in an initial increased velocity jet anterior to the turbinate region of the nasal cavity that is then directed downwards or dispersed away from the middle meatus. These results are paradoxical as the cotton is inserted into the inferior turbinate region, yet we see the reduction of airflow through the middle meatus (Figure 2B). Figure 4B and 4C represent cross sectional slices taken at about 60% of the length of the nasal passage just posterior to the cotton plug pre and post cotton insertion. In Figure 2B, we observe that the effects of cotton placement and initial increase of airflow velocity actually bring the nasal airflow into a more balanced flow distribution similar to what is seen in healthy controls⁸. Coupled with the fact that the overall flowrate decreases with cotton insertion yet patients are reporting improved airflow suggests that flowrate quantity might be secondary to its spatial distribution when it comes to the perception of airflow. This decrease in flowrate through the middle meatus was observed among all patients. In Figure 4B for example, we see the decrease in middle meatus flow unilaterally (right panel/ left side of the nasal passage) with cotton placement whereas in Figure 4C a bilateral decrease with cotton placement is observed. For all patients, given a resting flowrate, the airflow through the middle meatus was significantly reduced by about 41% thus redistributing the flow through the middle and inferior meatus from a ratio of 1.13 to 0.84. This suggests that the nasal airflow profile is moving away from the characteristic ENS middle jet-stream and thus ENS symptoms appear to improve. Interestingly, the distribution of airflow through the inferior meatus does not change despite the insertion of cotton to the region. This suggests that the relationship between airflow distribution, regional airflow rate and regional resistance may be complex, but nevertheless important factors to consider when designing, implementing, or selecting implants for patients with ENS symptoms. This distribution of airflow in the presence of cotton placement is comparable with CFD results from healthy controls previously published^{8,9,11,13,20,21}, and cotton parameters such as volume and position should be further explored to characterize its impact on optimizing the distribution of nasal airflow.

Conclusion

Through CFD analysis, we have determined that the cotton test is not only successful in increasing overall nasal resistance of symptomatic ENS patients, but also redistributes nasal airflow away from the middle meatus. These results address the physiologic basis for the symptom relief of cotton insertion in ENS patients with significant turbinate tissue loss. Furthermore it bolsters the utility of this office-based tool in combination with CFD technique in identifying appropriate surgical candidates for the inferior meatus implant/augmentation procedure.