Magnetic Resonance Imaging Reveals the Complementary Effects of Decongestant and Breathe Right Nasal Strips on Internal Nasal Anatomy

Transcription

1 The Laryngoscope VC 2016 The American Laryngological, Rhinological and Otological Society, Inc. Magnetic Resonance Imaging Reveals the Complementary Effects of Decongestant and Breathe Right Nasal Strips on Internal Nasal Anatomy Courtney A. Bishop, PhD; Steven M. Johnson, PharmD; Matthew B. Wall, PhD; Robert L. Janiczek, PhD; Gilbert Shanga, PhD; Richard G. Wise, PhD; Rexford D. Newbould, PhD; Philip S. Murphy, PhD Objectives/Hypothesis: This magnetic resonance imaging (MRI) study of 26 subjects with nasal congestion was performed to assess in the complete nasal passage both the anatomical effect of the marketed Breathe Right Nasal Strip (BRNS) relative to placebo and the potential adjunctive effect of using a decongestant in combination with the BRNS. Study Design: Randomized, crossover study. Methods: The study consisted of two parts, the first involving application of either the BRNS or the placebo strip in a randomized, crossover design with evaluator blinding, and repeated MRI scanning; and the second a sequential process of decongestant administration, MRI scanning, application of the BRNS, and repeated MRI. The same anatomical MRI protocol was used throughout. Nasal patency was assessed in the whole nasal passage and eight subregions (by inferior superior, anterior posterior division). Numerical response scores representing subjective nasal congestion were also obtained. Results: Results demonstrate significant anatomical enlargement with the BRNS relative to placebo (P <.001), as well as an additive effect of using a decongestant in combination with the BRNS; both supported by a strong and significant negative correlation with the subjective nasal response measures of nasal congestion (r , P 5.002). Furthermore, analysis of the nasal subregions indicates that this adjunctive effect arises from a partially localized action of the complementary products: the BRNS acting primarily anteriorly in the nose and the decongestant mainly posteriorly. Conclusions: The BRNS alone significantly increases nasal patency and alleviates perceived nasal congestion, and additional relief of symptoms can be obtained with simultaneous use of a decongestant. Key Words: Nose and paranasal sinuses, clinical, airway and voice modeling, evidence-based medicine. Level of Evidence: 1b. Laryngoscope, 126: , 2016 From Imanova, London, United Kingdom (C.A.B., M.B.W., R.D.N.); Imperial College London, London, United Kingdom (C.A.B., R.D.N.); GlaxoSmithKline Consumer Healthcare, Parsippany, New Jersey, U.S.A. (S.M.J., G.S.); GlaxoSmithKline, Uxbridge, United Kingdom (R.L.J., P.S.M.); and Cardiff University Brain Research Imaging Centre, School of Psychology, Cardiff University, Cardiff, United Kingdom (R.G.W.) Editor s Note: This Manuscript was accepted for publication January 6, C.A.B. designed and performed the study, processed data, analyzed results, and wrote the manuscript; S.M.J. designed the study and wrote the manuscript; M.B.W. designed and performed the study and wrote the manuscript; R.L.J. designed the study and performed data and manuscript review; G.S. designed the study and statistically analyzed and interpreted the data; R.G.W. designed the study and wrote the manuscript; R.D.N. designed and performed the study and wrote the manuscript; P.S.M. designed the study, interpreted data, and critically reviewed the manuscript. The study was funded by GlaxoSmithKline. S.M.J., G.S., R.L.J., P.S.M. are employees of GlaxoSmithKline, and the study was funded by GlaxoSmithKline. R.G.W. was a consultant for this work, paid by GlaxoSmithKline; there was no involvement of his employer, Cardiff University, in this work. The authors have no other funding, financial relationships, or conflicts of interest to disclose. Send correspondence to Dr. Courtney Bishop, Imanova Limited, Burlington Danes Building, Hammersmith Hospital, Du Cane Road, London, W12 0NN, United Kingdom. courtney.bishop@imanova. co.uk DOI: /lary INTRODUCTION Nasal congestion affects around 25% of the European population, 1 usually caused by inflammation of the membranes lining the nasal passage, resulting in nasal blockage. Frequent triggers are allergic reactions and common colds. Subjects with nasal congestion generally have nasal breathing difficulties and a reduced ability to smell, which may impact upon general feelings of wellbeing and mood state. 2 Nasal obstruction has also been associated with sleep apnea- 3 increasing the likelihood of personal injury, burdening health services, and causing lost work productivity- 4,5 so there is both high social and economic demand for alleviation of the deleterious effects of nasal congestion and its associated conditions. Magnetic resonance imaging (MRI) measures of nasal cross-sectional area and volume (patency) have previously been compared with the more common acoustic rhinometry (AR) measures, 6 as well as with measures derived from primate nasal casts. 7 Potential advantages of MRI over these methods include increased reliability, because MRI measures do not rely so heavily on patient cooperation; the ability to visualize both the nasal passage and the surrounding structures in three dimensions, facilitating the identification of various tissue types; and the opportunity to align/coregister longitudinal data from a given subject under different treatment 2205

2 conditions, to clearly identify anatomical changes on a slice-by-slice or regional basis. Corey et al. 6 found that the correlation between the AR and MRI measures of cross-sectional areas and volumes was high after administration of a decongestant (0.969), but low beforehand (0.345), suggesting that one of the techniques is more affected by relatively high congestion within the nasal passage; hypothesized to be the result of interference of the nasal cycle during the long MRI scan (1 hour), but equally likely to be the result of the relatively higher interobserver variability in the AR data and possible disturbance of the nasal cavity and congestion by the AR apparatus. Enlargement of the nasal airspace and alleviation of nasal congestion can be achieved with a nasal strip or a decongestant. In a subset of congested subjects, the use of the Breathe Right Nasal Strip (BRNS) was shown to be associated with a feeling of relief, 8 and increased maximum nasal inspiratory flow rates. 9 A number of studies have examined the nasal valve with the BRNS or similar dilator using AR. 10 However, no study has yet used MRI to examine the complete nasal passages to determine the magnitude and locations of the anatomical enlargement effects of the BRNS compared to placebo, and whether there is added benefit from using the BRNS in combination with short-term decongestant use. This study aims to compare both structural MRI measures of nasal patency and subjective numerical response (NR) reported feelings of congestion in the BRNS and placebo states, as well as to assess whether there is an adjunctive effect of the BRNS after administration of a decongestant. The anatomical locations of these effects will be determined via three-dimensional (3D) deformation of the coregistered MRI data. An exploratory analysis will examine covariation in nasal patency and NR scores. MATERIALS AND METHODS Ethics Statement This study was reviewed by the London Brent Research Ethics Committee and was performed in accordance with the ethical standards of the 2000 Declaration of Helsinki. All subjects gave written informed consent prior to inclusion in the study. Clinical Trial This research was registered on ClinicalTrials.gov (identifier: NCT ). Subjects At offsite screening, on a day prior to the study proper, subjects were recruited based on a self-rating of 3 to 8 (inclusive) for nasal congestion on a 0 to 10 numerical rating scale (from Breathe Freely 5 0 to Totally Blocked 5 10), and a nasal congestion frequency score of 1 to 4 on a 1 to 5 scale (Always 5 1, Often 5 2, Sometimes 5 3, Rarely 5 4, Never 5 5). A total of 28 subjects who suffer from nasal congestion were recruited. Two subjects dropped out before the first MRI scan. The scanned cohort consisted of nine females and 17 males (mean age years). Exclusion criteria at screening included a history of psychiatric illness that may affect assessments, any condition that may have an effect on nasal breathing (e.g., obstructive nasal polyps, or nasal tract structural malformations such as a deviated septum), use of medications that may alter nasal breathing capacity (e.g., corticosteroids within the past month, decongestant within the past 12 hours), and any contraindications to MRI scanning; all of which were assessed via interview and questionnaires prior to scanning. After MRI scanning, a subset of subjects (n 5 5) were excluded from image processing for the following reasons: incidental finding (n 5 1), scan images poor quality (n 5 1), incomplete scanning (n 5 1), contraindicated dental implant (n 5 1), imaged on a different scanner (n 5 1). Study Design An offsite screening visit for informed consent and determination of subject eligibility was conducted before subjects attended for all study assessments on a separate, single day. No scanning was performed at the offsite screening visit. The postscreening study assessment timeline is depicted in Figure 1. The schedule consisted of two parts: the first involved application of either the BRNS or the placebo strip in a randomized, crossover design with evaluator blinding, and repeated MRI scanning; the second involved a sequential process of decongestant administration, MRI scanning, application of the BRNS, and repeated MRI. Each subject received all scans during a single day, with a break (90 minutes) between scan sessions. The same anatomical MRI protocol was used throughout. Nasal Strips and Administration Procedure The BRNS is an external, mechanical nasal dilator used to provide temporary relief from nasal congestion and stuffiness. The placebo strip used was a matching sham device without the nasal dilatory mechanism. The nasal strip (either BRNS or placebo), was applied to the outside of the subject s nose, across the bridge from alar crease to alar crease. This study was evaluator-blinded, and attempted to keep the intervention as blinded as practically possible. Subjects were asked to close their eyes during application of the strip, clinical staff applying the strip were blinded to the strip used for each application, and the analyst processing the imaging data remained blinded to the active/placebo condition. Decongestant Administration Procedure At a time 20 minutes (65 minutes) before the start of the third scanning session (part 2 of the study), subjects were observed while being instructed to administer one spray per nostril of the decongestant Vicks Sinex (Procter & Gamble, Cincinnati, OH), containing 0.5 mg/ml oxymetazoline hydrochloride. MRI Data Acquisition High-resolution anatomical MRI data were acquired duringeachscansession(1,2,3a,and3binfig.1)onasiemens (Erlangen, Germany) Verio 3-T scanner equipped with a 32- channel phased-array receive-only head coil; providing four anatomical MRI scans for each subject. A dual-contrast turbo spin echo sequence acquired two effective echo times: 10 milliseconds for a proton density (PD)-weighted volume and 84 milliseconds for a T2-weighted volume. Each volume consisted of mm-thick slices in a 256 by 256-mm field of view 2206

3 Fig. 1. Postscreening study assessment timeline. Depiction of the magnetic resonance imaging scan timeline for each subject, showing the scheduling of the numerical response (NR) recordings at the predefined time points throughout a single day. 0AAp 5 immediately (0 minutes) after application of strip (BRNS/placebo); 0AR 5 immediately (0 minutes) after removal of strip (BRNS/placebo); 0BAp 5 immediately (0 minutes) before application of BRNS strip (scan 3 only); 20AAd 5 20 minutes after administration of decongestant; 30AAp 5 30 minutes after application of strip (BRNS/placebo); BRNS 5 Breathe Right Nasal Strip; BSL 5 baseline; DECON 5 decongestant. [Color figure can be viewed in the online issue, which is available at with a 320 by 320 matrix size resulting in 0.8-mm isotropic resolution. A repetition time of 8,480 milliseconds with a parallel imaging factor of 2 gave a scan time of 9 minutes 11 seconds. Data were checked for motion artifacts or incidental findings and excluded from further processing if necessary. A total of 21 subjects passed initial image quality control. MRI Data Processing The three subsequent anatomical MRI scans for each evaluable subject (from scan sessions 2, 3a, and 3b) were initially coregistered to the first MRI scan using a rigid-body transform. 11 A multistep segmentation process defined both the whole nasal passage and eight subregions on the coregistered images (Analyze 11.0, Specifically, following delineation of the whole nasal passage as a single region of interest (ROI), eight subregion bounding boxes acted to subdivide the whole nasal passage into eight subregions, by inferior superior and anterior posterior location as well as laterality (Fig. 2). The spatial limits of the bounding boxes demarcating the spatial extent of the bilateral nasal passage ROIs were defined by the following anatomical landmarks for consistency: the lower limit by the horizontal axis of the inferior turbinate, the midpoint by the horizontal axis of the middle turbinate, the upper limit by the cribriform plate, the anterior posterior separation by a vertical line perpendicular to the head of the inferior turbinate, and the posterior extent by a vertical line perpendicular to the tail of the inferior turbinate. The left and right side of the whole nasal passage was thus divided into inferior, superior, anterior, and posterior subregions (inferior anterior [IA], superior anterior [SA], inferior posterior [IP], superior posterior [SP]), providing a total of eight subregional ROIs. The images in Figure 2 depict the above-described segmentation processing steps for a representative subject s PD-weighted MRI scan. 3D models of the segmented regions (airways/bounding boxes, shown on the top right of each panel) aided visualization of the resultant segmentations in the workspace. Measures of cross-sectional area (CSA) were extracted from each coronal imaging slice of the whole/single-region nasal ROI, and measures of nasal volume from both the single/whole-region ROI and the eight subregions. Bilateral averages were computed for the subregion volumes to account for lateral shift in nasal congestion (and therefore nasal patency) throughout the scanning period, and are reported herein. Supplementary Measures An NR score representing subjective nasal congestion was obtained from each subject at defined time points, depicted in Figure 1. For the first part of the study (scans 1 and 2), the time points were baseline (before any intervention), immediately after application of each nasal strip (BRNS/placebo), 30 minutes (65 minutes) after the start of MRI scanning, and when the nasal strip was removed at the end of each scan. For the second part of the study (scan session 3), the NR was recorded before administration of the decongestant (baseline), 20 minutes (65 minutes) after administration of the decongestant (corresponding to a time immediately before commencement of scanning), then immediately before and after application of the BRNS. Baseline (and change from baseline) measurements at each scan session allowed us to account for possible nasal cycle contributions to the NR scores. Statistical Analysis Analysis of variance was used to analyze the anatomical measures of CSA and volume (with treatment and period as fixed effects, subject as random effect), whereas analysis of covariance was used to assess the change from baseline for NR (with the above factors plus baseline NR as a covariate). All tests were done at the 5% significance level. Pearson correlation coefficients were used to explore (post hoc) the linear relationship between change from baseline for NR and the anatomical measures. Comparisons were made between BRNS and placebo within part 1, and between decongestant and BRNS with decongestant in part 2. RESULTS Anatomical Measures (CSA and Volume) CSA and volume measurements for the whole/single nasal ROIs, as well as volumes of the smaller subregions (IA, SA, IP, SP), are provided in Table I. The subregional percentage volume changes from placebo are plotted in Figure 3. Results show that BRNS provided statistically higher nasal CSA (P <.0001), higher total volume (P ), and higher volume of both the IA and SA subregions (P and.020, respectively) compared to the placebo strip. There was no treatment order (period) effect. There was no statistical difference between the 2207

4 Fig. 2. Definition of the bounded nasal passage regions of interest (ROIs). For a representative subject, the screenshots depict (A) segmentation of the whole nasal airway, (B) definition of the eight subregion bounding boxes, and (C) the eight subregion nasal ROIs (generated from subdivision of the whole nasal airway by the eight subregion bounding boxes). In A, labels give the image orientations: A 5 anterior; I 5 inferior; L 5 left; P 5 posterior; R 5 right; S 5 superior. The individual bounding boxes are labeled according to their relative positioning and by laterality (L/R) inferior anterior (IA [L/R]), inferior posterior (IP [L/R]), superior anterior (SA [L/R]), and superior posterior (SP [L/R]) in B, and likewise for the eight masked subregions in C. BRNS and placebo strips on measurements of volume for the posterior subregions (IP and SP). Mean CSA and single volume measurements for decongestant and decongestant 1 BRNS were higher than both BRNS and placebo, and the combination of decongestant 1 BRNS was numerically higher than decongestant only. The magnitude of the effect in the decongestant 1 BRNS treatment arm was comparable to the sum of the effects in the individual treatment arms (i.e., BRNS alone plus decongestant alone). Comparing anatomical measures under the decongestant and decongestant 1 BRNS conditions, the decongestant 1 BRNS had statistically higher total volumes (P <.0001) compared to decongestant alone, as well as statistically higher volume of both anterior subregions (IA, P <.0001; SA, P <.0001) and the IP region (P <.0001). There was no statistical difference between the decongestant and decongestant 1 BRNS conditions for the SP subregion. Observation of the subregional percentage volume change from placebo, for each of the treatment 2208 conditions, shows clearly the regional differences in the primary mechanism of action of the BRNS and the decongestant; the BRNS primarily enlarging the anterior subregions, whereas the effect of the decongestant is primarily in the posterior subregions (Fig. 3). NR Scores Table II shows results for the NR scores in part 1 of the study (scan sessions 1 and 2). There was a statistically significant improvement in NR scores from baseline at both the immediate time point (P ) and after 30 minutes of application (P ) with the BRNS, but not with the placebo strip (P and P , respectively). However, there was no statistically significant head-to-head difference between the BRNS and placebo strip scores at either time point (P and P , respectively). In part 2 of the study, the NR score at baseline was (mean 6 standard deviation). Twenty

5 TABLE I. Anatomical Measures (CSA and Volume). Measure BRNS, n 5 21 Placebo, n 5 21 Decongestant, n 5 21 Decongestant 1 BRNS, n 5 21 CSA, cm (0.29) 1.17 (0.31) 1.57 (0.29) 1.77 (0.29) P <.0001* n.s VOL T,cm (3.12) 11.7 (3.25) 15.8 (3.60) 17.7 (3.69) P.0004* <.0001 VOL IA,cm (1.20) 2.96 (1.32) 3.16 (1.41) 3.86 (1.38) P.011* <.0001 VOL SA,cm (0.657) (0.493) (0.557) (0.684) P.020* <.0001 VOL IP,cm (1.41) 5.36 (1.32) 8.19 (1.95) 9.02 (1.84) P n.s* <.0001 VOL SP,cm (1.27) 2.79 (1.13) 3.66 (1.25) 3.86 (1.49) P n.s.* n.s. *Between BRNS and placebo strip. Between decongestant and decongestant 1 BRNS. All values are mean (standard deviation). BRNS 5 Breathe Right Nasal Strip; CSA 5 cross-sectional area; n.s. 5 not significant; VOL IA/SA/IP/SP 5 bilateral average volumes of the multiple subregions (inferior anterior, superior anterior, inferior posterior, superior posterior, respectively); VOL T 5 total (single airway) volume. minutes after the administration of the decongestant, the NR score dropped by 1.35 points to As expected, the decongestant spray reduced the selfreported feeling of congestion (P 5.011). There was an even greater change from baseline after the application of the BRNS, with a mean NR score of (P 5.001). was an adjunctive effect of using a decongestant in combination with the BRNS, as determined by anatomical MRI. Findings demonstrated both the marked anatomical enlargement of the nasal airway with the BRNS relative Correlation of Anatomical Measures With NR Scores Plotting the mean NR immediate change from baseline against the mean percentage volume change for each treatment arm normalized to the placebo arm (Fig. 4A), we observe a significant negative correlation between these two measures (r , P 5.002). Figure 4B D shows scatterplots of the individual subject data points for each non-placebo treatment arm. R 2 values for the relationship between the NR immediate change from baseline (change from placebo) scores and the mean percentage volume change from placebo are 0.90, 0.71, and 0.77 for the BRNS, decongestant, and decongestant 1 BRNS conditions, respectively. Post Hoc Power Calculations Our primary finding of increased nasal patency with the BRNS versus placebo was assessed with a post hoc power analysis of the total nasal volume using G*power software ( which showed an effect size (Cohen d) of 0.903, and power of 0.99/99%, with our analyzed cohort of 21 congested subjects. DISCUSSION This single-center, evaluator-blind study in subjects with nasal congestion was performed primarily to assess the effect of the marketed BRNS relative to placebo on nasal patency, and secondly, to determine whether there Fig. 3. Percentage volume change from placebo for the nasal passage subregions. Plots represent the percentage volume change from placebo for the nasal subregions located in the inferior anterior (IA), superior anterior (SA), inferior posterior (IP), and superior posterior (SP) under each treatment condition (from left to right: the Breathe Right Nasal Strip [BRNS], decongestant [DECON], and DECON 1 BRNS). Statistically significant differences between the BRNS and placebo conditions, as well as between the DECON 1 BRNS and DECON conditions, are indicated as *P <.05 and ***P <

6 TABLE II. Temporal Change in NR Scores During Part 1 of the Study (Scan Sessions 1 and 2). BRNS, n 5 26, Mean (SD) Placebo, n 5 26, Mean (SD) Baseline NR 5.1 (1.50) 4.8 (1.49) Immediate change from baseline (1.07), P P minute change from baseline 21.0 (1.30), P P (0.90), P (1.63), P BRNS 5 Breathe Right Nasal Strip; NR 5 numerical response; SD 5 standard deviation. to placebo, and the adjunctive anatomical effect of a decongestant with the BRNS. The lack of a treatment order (period) effect in the crossover study (part 1) suggests no nasal cycle contributions during the break (90 minutes) between scan sessions. There was also a statistically significant improvement in the subjective measures of nasal congestion (NR scores) immediately after application of the BRNS and after administration of the decongestant. The addition of the BRNS after the decongestant further improved the NR scores. Regional ROI analysis revealed that the two treatments, BRNS and decongestant, were primarily operating in different subregions of the nasal passage; the BRNS selectively enlarged the anterior subregions (IA and SA regions in Table II), whereas the decongestant had greatest effect at the posterior of the nasal passage (IP and SP regions). Supportive findings have previously been described in a cohort of 89 subjects with complaints of nasal obstruction, 12 where mean CSA and volumetric measurements were obtained with AR in the anterior and posterior nasal regions, before and after application of the BRNS or administration of the decongestant xylometazoline. However, the use of MRI to measure such effects is unique to this study, and in this analyzed cohort of 21 congested subjects a power of 99% was achieved on the volumetric changes between BRNS and placebo, suggesting that relatively small subject numbers are sufficient with an MRI study design such as employed here. The strong and significant correlation between 1) the NR immediate change from baseline change from placebo and 2) the percentage total volume change from placebo (r ) suggests that the anatomical enlargement effects observed with each of the treatment Fig. 4. Correlation of total/whole region nasal volumes with subjective numerical response (NR) scores. (A) The relationship between (x) the cohort mean NR immediate change from baseline scores and (y) the mean percentage change in single volume readings, both expressed as change from placebo (n 5 21). From left to right, the points correspond to the decongestant (DECON) 1 Breathe Right Nasal Strip (BRNS), DECON, BRNS, and placebo conditions. Error bars are standard error. (B D) Scatterplots of the individual subject values (gray diamond symbols) for (x) NR immediate change from baseline scores and (y) the percentage volume change, both expressed as change from placebo (n 5 21), in each non-placebo treatment arm (BRNS, DECON, and DECON 1 BRNS, respectively); NR scores are grouped into intervals of fixed width across the observed range of NR score changes (25 to24, 23 to22, 21 to 0, and 1 to 2) and plotted at their midinterval values; linear fits are generated for the mean percentage volume change from placebo in each NR interval (black square symbols). 2210

7 conditions is directly proportional to the subjective improvement in nasal breathing capacity of the subjects. This finding indicates that anatomical MRI is a valuable tool to assess the substrate for perceived changes in breathing sensation. It could therefore be used in future experimental studies to improve nasal dilator devices and decongestants, and with such strong correlations to NR scores, may prove to be a more reproducible and sensitive measure than other techniques such as AR. One limitation of this study is the lack of a postdecongestant placebo strip arm, which would allow the actual contribution of the BRNS to the decongestant to be quantified. This study instead made the assumption based on known pharmacodynamics of oxymetazoline nasal spray that the decongestant effect would not change significantly in the short time period between measurements approximately 5 minutes after the decongestant NR score was taken but before application of the BRNS in part 2 of the study but it cannot strictly be differentiated from a possible increased effect of the decongestant at the later time point. However, the magnitude of the effect seen in the decongestant 1 BRNS condition was similar to the effect of decongestant only plus the effect of BRNS alone, adding confidence that this is truly an adjunctive effect. CONCLUSION In conclusion, it appears that there is a complementary mechanism of action of the BRNS and decongestant products, so that when the products are used in combination, there is added relief from nasal congestion. We do, however, appreciate that the long-term use of a decongestant is actively discouraged, because of the development of rhinitis medicamentosa. We are therefore not recommending decongestant as a long-term treatment for nasal congestion, but merely as an occasional complementary product for the BRNS. The BRNS alone can provide significantly increased nasal patency, and perceived alleviation of nasal congestion, so it may be a suitable drug-free, short-term treatment option for nasal congestion. BIBLIOGRAPHY 1. Bauchau V, Durham SR. Prevalence and rate of diagnosis of allergic rhinitis in Europe. Eur Respir J 2004;24: Litvack JR, Mace J, Smith TL. Role of depression in outcomes of endoscopic sinus surgery. Otolaryngol Head Neck Surg 2011;144: Lofaso F, Coste A, d Ortho MP, et al. Nasal obstruction as a risk factor for sleep apnoea syndrome. Eur Respir J 2000;16: Colten HR, Altevogt BM. Functional and economic impact of sleep loss and sleep-related disorders. In: Sleep Disorders and Sleep Deprivation: An Unmet Public Health Problem. Washington, DC: National Academies Press; Rosekind MR, Gregory KB, Mallis MM, Brandt SL, Seal B, Lerner D. The cost of poor sleep: workplace productivity loss and associated costs. J Occup Environ Med 2010;52: Corey JP, Gungor A, Nelson R, Fredberg J, Lai V. A comparison of the nasal cross-sectional areas and volumes obtained with acoustic rhinometry and magnetic resonance imaging. Otolaryngol Head Neck Surg 1997; 117: Yeh HC, Brinker RM, Harkema JR, Muggenburg BA. A comparative analysis of primate nasal airways using magnetic resonance imaging and nasal casts. J Aerosol Med 1997;10: Latte J, Taverner D. Opening the nasal valve with external dilators reduces congestive symptoms in normal subjects. Am J Rhinol 2005;19: Di Somma EM, West SN, Wheatley JR, Amis TC. Nasal dilator strips increase maximum inspiratory flow via nasal wall stabilization. Laryngoscope 1999;109: Ellegard E. Mechanical nasal alar dilators. Rhinology 2006;44: Jenkinson M, Smith S. A global optimisation method for robust affine registration of brain images. Med Image Anal 2001;5: Hoyvoll LR, Lunde K, Li HS, Dahle S, Wentzel-Larsen T, Steinsvag SK. Effects of an external nasal dilator strip (ENDS) compared to xylometazolin nasal spray. Eur Arch Otorhinolaryngol 2007;264: