The effect of maxillary sinus antrostomy size on xenon ventilation in the sheep model

The effect of maxillary sinus antrostomy size on xenon ventilation in the sheep model KEVIN T. BRUMUND, MD, SCOTT M. GRAHAM, MD, KENNETH C. BECK, GEOFFREY MCLENNAN, MD, PHD, Iowa City, Iowa

OBJECTIVE: A major goal of maxillary antrostomy is to increase sinus ventilation. Limited data exist regarding the effect of maxillary antrostomy size on sinus ventilation. We sought to quantify the effect of uncinectomy, small antrostomy, and large antrostomy on maxillary sinus ventilation using xenonenhanced CT in the sheep model. MATERIALS, STUDY DESIGN, AND METHODS: A xenon-oxygen-air mixture was delivered to 8 fresh cadaveric sheep heads while repeated CT scans were performed through the maxillary sinuses. Baseline and postoperative studies were performed after an endoscopic uncinectomy, small antrostomy, or large antrostomy was created. Images were analyzed to measure the density of the xenon gas in the maxillary sinus as a function of time, generating a time constant. RESULTS: The time constants for both small antrostomy and large antrostomy were significantly different compared to baseline (P ⴝ 0.003 for both). The time constant comparison between small antrostomy and large antrostomy was not significant (P ⴝ 0.948). CONCLUSIONS: A small antrostomy produces a statistically significant increase in maxillary sinus ventilation over baseline. No significant further ventilation increase is obtained by creating a large antrostomy in the sheep model. This lends credence to the use of small antrostomies to improve maxillary sinus ventilation in human sinus surgery. (Otolaryngol Head Neck Surg 2004;131:528-33.)

From the Department of Otolaryngology-Head and Neck Surgery (Drs Brumund and Graham), the Division of Physiologic Imaging-Department of Radiology (Drs Beck and Hoffman), and the Division of Pulmonary Medicine-Department of Internal Medicine (Dr McLennan), University of Iowa, Iowa City, IA. Presented at the Annual Meeting of the American Academy of OtolaryngologyHead and Neck Surgery, Orlando, FL, September 21-24, 2003. Reprint requests: Scott M. Graham, MD, University of Iowa Hospitals and Clinics, Department of Otolaryngology-Head and Neck Surgery, 200 Hawkins Drive, 21201 PFP, Iowa City, IA 52242-1093; e-mail, [email protected]. 0194-5998/$30.00 Copyright © 2004 by the American Academy of Otolaryngology–Head and Neck Surgery Foundation, Inc. doi:10.1016/j.otohns.2004.04.010

528

T

PHD,

ERIC A. HOFFMAN,

PHD,

and

he air in the paranasal sinuses is continually exchanged during normal breathing. Local inflammatory processes may result in edema of the sinonasal mucosa affecting the patency of the ostiomeatal complex and subsequently obstructing airflow into the sinuses. A pathologic environment that serves as a medium for bacterial overgrowth can develop as a result of the obstruction. The classic theory of sinusitis involves a decrease in mucociliary transport resulting in pooling and stasis of secretions. Negative intrasinus pressure develops due to air resorption in the sinus cavity, anaerobic conditions evolve, the pH drops, and bacterial proliferation occurs, causing a sinus infection.1 An alteration in the ventilation of the paranasal sinuses is probably of great significance in the pathogenesis of sinus infections. In 1974, Aust and Drettner,2 stated that gas exchange depended on the following parameters: volume of the sinus, diameter of the ostium, nasal air flow, nasal respiratory pressure, size and shape of the nasal cavity, composition of the air, and gas absorption by the mucosa. Their later work led them to conclude that ostial patency was the most important factor in the pathophysiology of sinus disease.3 In his early studies of nasal physiology, Proetz4 estimated it would take several hours (more than 1,000 respirations) to completely exchange all the air in the sinuses. Later studies by Drettner and Aust3 estimated the average gas exchange time for the maxillary sinuses at 5 minutes. Subsequent work by Zippel and Streckenbach5 utilizing 133xenon produced comparable results. Xenon-enhanced sinus CT scans have been well studied6-14 and provide the opportunity to quantify sinus ventilation. A major goal of surgical manipulation of the maxillary sinus ostium in the treatment of sinus disease is to ensure ostial patency and to improve maxillary sinus ventilation. We were interested in the effect of a range of surgical manipulation of the maxillary sinus ostium on maxillary sinus ventilation. In studying this, we used the sheep model. The sheep model provides a useful representation of human sinuses and has been used previously in resident teaching15 and experimental studies of sinus surgery.16,17 We sought to quantify the effect of uncinectomy, small antrostomy, and large antrostomy on maxillary sinus ventilation using xenon (Xe)-enhanced computed tomography in the sheep model.

Otolaryngology– Head and Neck Surgery Volume 131 Number 4

MATERIALS AND METHODS Materials and Positioning Approval for the protocol was obtained from the University of Iowa Animal Care and Use Committee. Eight fresh cadaveric sheep heads were used in the experiment. All animals used in the experiment were adults of similar size, and each weighed 70 to 90 kg. Sheep are a useful animal model because both the general nasal anatomy and paranasal sinus anatomy have similarities in appearance and orientation to humans.15 The fresh cadaveric head was placed on the CT scanner table and secured with tape. The Xe/O2/room air gas delivery system was connected to an animal ventilator to maintain a constant flow of gas at 5 L/minute. A Y-connector was placed at the end of the gas delivery circuit and one end of tubing from the Y-connector was secured into each naris with putty. This was an open system with gas being delivered via one branch of the Y-connector into each nasal cavity and then escaping through the cut end of the trachea. CT Specifications Imaging was performed on a Philips MX8000 Quad CT Scanner (Phillips Medical Systems, Andover, MA). After positioning, a volume scan at 120 kV, 150 mA, and 1.3-mm slice thickness was obtained. This volume scan of the entire paranasal sinus region allowed for assessment of the anatomy and identification of the same area within the maxillary sinus to be evaluated in each subject. The subject was repositioned by moving the scanner table so the maxillary sinus region is scanned during the Xe protocol. The scanner settings for the images obtained during the xenon protocol were 90 kV, 190 mA, and 2.5-mm slice thickness. The quad scanner allows for a 1-cm coverage area (4 images 2.5 mm apart) with no table movement. Xenon Wash-in/Wash-out Protocol The protocol was initiated with a 25% O2/room air mixture delivered to the sheep head via the gas delivery circuit. Two minutes of baseline images were acquired, scanning every 30 seconds. The delivered gas is then switched to a 30% Xe/25% O2/room air mixture (washin). Sequential images were obtained every 30 seconds for the first 5 minutes, then every 60 seconds for the remaining 15 minutes. The delivered gas was then switched back to the original 25% O2/room air mixture (wash-out). Sequential images were then obtained every 30 seconds for the first 5 minutes, and every 60 seconds for the remaining 15 minutes. After completion of the baseline wash-in/wash-out, endoscopic sinus surgery was performed to generate either an uncinectomy, small antrostomy (2-4 mm in

BRUMUND et al

529

size or approximately 2⫻ to 3⫻ the size of the natural ostium), or a large antrostomy (6-9 mm in size or approximately 6⫻ the size of the natural ostium) on both maxillary sinuses. Careful endoscopic navigation around the large middle turbinate of the sheep revealed the uncinate process. Unlike the anteriorly based uncinate process in the human, the sheep uncinate is inferiorly based. A down-biting instrument was used to perform the uncinectomy, allowing visualization of the natural ostium, measured at 1.0 to 1.5 mm in size. Cutting instruments and the down-biter were used to generate the antrostomy. After surgery, another volume scan was obtained and the head was repositioned in the proper orientation on the scanner table so that the same region within the maxillary sinuses would be scanned. The entire wash-in/wash-out protocol is then repeated. Depending on the type of surgery performed initially, some subjects were reoperated on and the protocol was repeated a third time. The concentration of xenon delivered to the subject was recorded throughout the entire wash-in/wash-out protocol to confirm adequate Xe delivery. A total of 50 CT scans were obtained during the 40-minute wash-in/ wash-out protocol. Image acquisition was more frequent during the early portion of each phase because we expect a faster rate of change in xenon concentration within the sinuses when Xe is first being delivered or eliminated. Data Analysis The efficacy of sinus ventilation was quantified by analyzing regional xenon equilibration curves as follows. A custom computer program was used to extract radiodensity values [Hounsfield units (HU)] from regions of interest (ROI) indicated as small squares within the maxillary sinuses on the first of the timed series of images obtained in the scanning protocol. The program first extracted mean HU from all the images in a given Xe equilibrium run, resulting in an array of HU versus time values for the series of images in the ROI. This data was fitted into an exponential model:

冉

冋 册冊

HU(t) ⫽ HUb ⫹ MAG · 1 ⫺ exp ⫺

t

␶

where HU(t) represents the HU values at each time point, HUb is the baseline HU value before the start of Xe delivery, MAG is the magnitude of the total change in HU from baseline to complete equilibration of Xe within the sinus, and ␶ is the time constant for Xe equilibration in the sinus. We used a Levenberg-Marquardt-directed search routine to minimize the sum squared errors between estimated values and data

530 BRUMUND et al

Otolaryngology– Head and Neck Surgery October 2004

points while varying HUb, MAG, and ␶. The parameter ␶ is the time in seconds it takes for the Xe to reach 63% of the final equilibrium value, and therefore represents the efficacy of sinus ventilation. RESULTS A total of 8 cadaveric sheep heads were used. Data were collected from each side (right and left) and some heads had multiple sequential surgeries performed on them. Two of the subjects had unilateral maxillary sinus opacification on initial volume scan, precluding any data acquisition from the affected side. Repeat CT scans taken through the maxillary sinuses in each subject during the wash-in and wash-out of 30% Xe delivered into the nasal cavities yielded exponential local density curves for a chosen ROI within the maxillary sinuses. The ROI chosen was in the center of the maxillary sinus. Analyzing the same area in each subject in both maxillary sinuses allowed for uniform data collection and avoided the interference from surrounding mucus, soft tissue, and bone. All experiments were performed immediately after animal sacrifice to avoid the effect that tissue decay might have on gas exchange. The surrounding tissue can affect the noise level of the wash-in/wash-out curve and may have significant influence on the time constant as described by Marcucci et al.11 The mean Xe density in the ROI is measured in each image and plotted as a function of time (Fig 1). The time constant, ␶, in seconds, is calculated for both maxillary sinuses in each subject at baseline and after each surgical procedure. Analysis of the Xe wash-out density curves confirmed complete wash-out of all Xe and a return to baseline after each wash-in/wash-out protocol. Assuring no residual Xe within the maxillary sinuses after completion of the wash-in/wash-out protocol prevented skewing of Xe density curves between subsequent surgical procedures. Antrostomy size was measured radiographically by analysis of the CT images. The presence of accessory ostia in the sheep maxillary sinus could affect the wash-in and wash-out of Xe and the rate of gas exchange. The degree of effect would likely depend on their size. The presence or absence of accessory ostia in humans is certainly variable, with a reported incidence of 4% to 41%.18 Radiographic evaluation of the initial volume scan of the sheep head, and endoscopic evaluation during surgery revealed no evidence of accessory ostia. Others who use the sheep model extensively for studying endoscopic sinus surgery also have not observed any accessory ostia (P. J. Wormald, MD, personal communication, March 2004).

Fig 1. (A) Representative CT image of sheep head with ROI outlined in right maxillary sinus. (B) Xe wash-in/washout density curve of the ROI outlined in (A). HU, Hounsfield units.

The mean ␶ value at baseline (N⫽14) was 640.7 seconds with a range of 168.4-1464.6 seconds. The mean value of ␶ after uncinectomy (N⫽7) was 512.3 seconds with a range of 358.0-942.9 seconds. After small antrostomy (N⫽7) the mean ␶ was 79.8 seconds with a range of 57.9-129.7 seconds. After large antrostomy (N⫽6) the mean ␶ was 42.1 seconds with a range of 20.1-82.8 seconds (Table 1). The linear mixed model analysis was used to estimate the average change in ␶ from baseline for each surgery (Table 2). The change in ␶ after uncinectomy compared to baseline was not significant, P ⫽ 0.274, but the change after small antrostomy and large antrostomy compared to baseline was significant, P ⫽ 0.003 for both. Neither the side of operation (right vs. left, P ⫽ 0.086), or sequence of surgery (first vs. second, P ⫽ 0.759) proved to be of significance. Based on the linear mixed model analysis, the estimates of the differences in the average changes in ␶ from baseline between the uncinectomy and small antrostomy and between the small antrostomy and large

Otolaryngology– Head and Neck Surgery Volume 131 Number 4

BRUMUND et al

Table 1. Summary of ␶ values, in seconds

Baseline Uncinectomy Small antrostomy Large antrostomy

531

Table 3. Comparison of change in ␶ values between surgeries

N

Mean (␶)

Range (␶)

14 7 7 6

640.7 512.3 79.8 42.1

168.4-1464.6 358.0-942.9 57.9-129.7 20.1-82.8

Table 2. Change in ␶ values from baseline for each surgery

Surgery

Estimated mean change from baseline

95% Confidence interval for estimated mean

P value

Uncinectomy Small antrostomy Large antrostomy

⫺231.63 ⫺576.18 ⫺592.38

⫺661.40-198.14 ⫺935.49-216.87 ⫺951.40-233.36

0.274 0.003 0.003

antrostomy were calculated (Table 3). The change in ␶ comparing the uncinectomy with small antrostomy approached significance (P ⫽ 0.056), but there was not a significant change in ␶ between the small antrostomy and large antrostomy (P ⫽ 0.948). DISCUSSION The introduction of Xe-enhanced computed tomography has allowed for more accurate quantification of sinus ventilation. Stable Xe, is a radiodense gas, which is denser than air. When used as a CT contrast agent, the concentration of Xe can be measured by CT scan. This concentration is linearly related to its CT enhancement. In 1985, Kalender et al6,7 delivered xenon into the nasal cavities of human subjects by positive pressure insufflation. Serial CT scans after xenon delivery allowed for the generation of time constants from the Xe washout curves of the sinuses. The mean gas exchange time in the maxillary sinuses in this population of healthy subjects was 4.0 minutes. As described by Simon et al,8 serial CT scans taken through a fixed location during the wash-in or wash-out of a Xe/O2/air mixture yield exponential local density curves for a chosen ROI. The time constant tau, ␶, of the curve, determined from a nonlinear curve fitting procedure, is equal to the inverse of the local ventilation per unit volume (specific ventilation). A smaller tau means faster gas exchange and increased ventilation. This method was initially used for the noninvasive measurement of regional pulmonary ventilation in canine subjects.9 Leopold et al10,11 later applied this technique to quantify sinus ventilation in healthy human subjects. They were able to characterize maxillary sinus

Surgery Uncinectomy-Small antrostomy Small antrostomyLarge antrostomy

Estimated difference in estimated means ⫺344.55 ⫺16.20

95% Confidence interval for difference in means

P value

⫺698.18-9.08

0.056

⫺532.98-500.58

0.948

ventilation by delivering subanesthetic concentrations of Xe via a positive airway pressure mask. Their average baseline time constant, tau, for the maxillary sinuses was 5.59 minutes. Paulsson et al12 used dynamic single-photon emission CT to quantify xenon wash-out in healthy patients and those with sinus disease. On average, patients with sinus disease on CT scan tended to have longer xenon wash-out half-times, although this was not statistically significant. Of the patients with sinus disease, those with nasal polyposis had significantly longer half-times compared with healthy subjects. In subjects with chronic sinusitis only (mucosal swelling) without polyps, the difference in half-times was not significant. Surprisingly little is known about the effect of the medical and surgical interventions that we offer patients to treat sinus disease. In particular, the efficacy of these interventions in improving sinus ventilation is not known. Rettinger et al7 demonstrated a rapid wash-out of xenon in a patient who had a unilateral surgically created window between the maxillary sinus and inferior meatus as compared with the contralateral nonoperated side. They also found an increase in gas exchange time after the application of mucosal decongestion drops (xylometazoline), a phenomenon they could not explain. Leopold et al10 had similar results, demonstrating that the average time constant, ␶, for the xylometazoline decongested group was higher than when not decongested. They hypothesized that after decongestion, air moves through the nasal cavity against less resistance and less air enters the maxillary sinuses, thereby decreasing ventilation. There was also a patient in this study population that had a unilateral inferior meatal antrostomy who had substantially improved gas exchange on that side compared with the normal contralateral side. Paulsson et al13 demonstrated no significant effect on paranasal sinus ventilation in healthy subjects from application of either sodium chloride drops or oxymetazoline using xenon wash-out technique. They too, felt the change in intranasal aerodynamics after decongestion affected sinus ventilation.

Otolaryngology– Head and Neck Surgery October 2004

532 BRUMUND et al

Paulsson and colleagues14 also sought to determine the effect of ostial diameter on Xe wash-out using a synthetically constructed nose-sinus model. They found an increase in ostial diameter led to more rapid Xe washout in their model and that ostial diameter played the greatest role in determining xenon wash-out time. The modern rhinologic surgeon faces a variety of choices in dealing with the maxillary sinus ostium. These choices range from “surgical abstinence,” through uncinectomy alone, to the creation of a small or large antrostomy. Traditional endoscopic technique has favored enlargement of the natural ostium of the maxillary sinus.19 The wisdom of this has more recently been called into question by some authors who favor uncinectomy only20 or a small antrostomy. Although the literature is replete with discussion of technique and even patency rates,19 there is little that addresses indications in a systematic way, and even less that is helpful in choosing one technique over another. The last year has seen publications suggesting that nitric oxide levels in the maxillary sinus are decreased with antrostomy21 and that “minimally invasive” sinus surgery was effective in treating a population with mostly radiographic stage I and II chronic sinusitis.22 Indications other than maxillary sinus ventilation exist for antrostomy and there are certainly situations in which a large antrostomy is helpful. In circumstances such as fungal disease, extensive polyposis, antrochoanal polyps, endoscopic orbital decompression,23 or endoscopic transantral biopsy of the pterygopalatine space,24 large antrostomies are useful for a variety of reasons. While recognizing that in select circumstances other indications may take precedence, we have chosen to examine the common surgical indication for maxillary antrostomy, that of improving maxillary sinus ventilation. We were interested in the degree of surgery necessary to first demonstrate a benefit in ventilation and were also interested in the extent of surgery at which maximal benefit is achieved. Analysis of the data in the xenon sheep model has revealed no increase in maxillary sinus ventilation over baseline with uncinectomy alone (P ⫽ 0.274). A large or small antrostomy did produce a significant increase in maxillary sinus ventilation (P ⫽ 0.003). There was, however, no additional benefit in performing a large antrostomy over a small antrostomy (P ⫽ 0.948). This data has the limitations of dealing with a single indication for antrostomy, in an animal model without mucosal disease. Nonetheless, it lends credence to the concept that small antrostomies in human surgery may be all that is required to provide optimal ventilation of the maxillary sinus.

CONCLUSIONS Xenon-enhanced computed tomography is an effective method for quantifying ventilation in the paranasal sinuses. We were able to demonstrate an improvement in gas exchange in the maxillary sinuses after endoscopic sinus surgery and the generation of antrostomies in the sheep model. Establishment of improved ventilation in the paranasal sinuses is a major goal in the surgical treatment of sinus disease. The creation of small antrostomies may be all that is necessary to significantly improve gas exchange in human maxillary sinuses. The authors would like to thank Justine M. Ritchie, PhD, from the Department of Biostatistics for statistical analysis, Jered Sieren, RT, from the Department of Radiology for radiography assistance, and Jennifer Pierce, BA, for assisting in manuscript preparation. REFERENCES 1. Naclerio RM, Gungor A. Etiologic factors in inflammatory sinus disease. In: Kennedy DW, Bolger WE, Zinreich SJ, Diseases of the sinuses diagnosis and management, 1st ed. Hamilton, Ontario: B.C. Decker Inc; 2001. pp. 47-55 2. Aust R, Drettner B. The functional size of the human maxillary ostium in vivo. Acta Otolaryngol (Stockh) 1974;78:432-5. 3. Drettner B, Aust R. Pathophysiology of the paranasal sinuses. Acta Otolaryngol (Stockh) 1977;83:16-9. 4. Proetz AW. Applied physiology of the nose. St. Louis, MO: Annals Publishing Company; 1953. 5. Zippel R, Streckenbach B. 133Xenon washout in the paranasal sinuses—a diagnostic tool for assessing ostial function. Rhinology 1979;17:25-9. 6. Kalender WA, Rettinger G, Suess C. Measurement of paranasal sinus ventilation by xenon-enhanced dynamic computed tomography. J Comput Assist Tomogr 1985;9:524-9. 7. Rettinger G, Suss C, Kalender WA. Studies of paranasal sinus ventilation by xenon-enhanced dynamic CT. Rhinology 1986;24:103-12. 8. Simon BA, Marcucci C, Fung M, et al. Parameter estimation and confidence intervals for Xe-CT ventilation studies: a Monte Carlo approach. J Appl Physiol 1998;84:709-16. 9. Marcucci C, Nyhan D, Simon BA. Distribution of pulmonary ventilation using Xe-enhanced computed tomography in prone and supine dogs. J Appl Physiol 2001;90:421-30. 10. Leopold D, Zinreich SJ, Simon BA, et al. Xenon-enhanced computed tomography quantifies normal maxillary sinus ventilation. Otolaryngol Head Neck Surg 2000;122:422-4. 11. Marcucci C, Leopold DA, Cullen M, et al. Dynamic assessment of paranasal sinus ventilation using xenon-enhanced computed tomography. Ann Otol Rhinol Laryngol 2001;110:968-75. 12. Paulsson B, Dolata J, Larsson I, et al. Paranasal sinus ventilation in healthy subjects and in patients with sinus disease evaluated with the 133-xenon washout technique. Ann Otol Rhinol Laryngol 2001;110:667-74. 13. Paulsson B, Lindberg S, Ohlin P. Effects of oxymetazoline on the ventilation of paranasal sinuses in healthy subjects. Am J Rhinol 2002;16:125-9. 14. Paulsson B, Dolata J, Lindberg S, et al. Factors influencing 133-xenon washout in a nose-sinus model. Clin Physiol 2001;21:246-52. 15. Gardiner Q, Oluwole M, Tan L, et al. An animal model for training in endoscopic nasal and sinus surgery. J Laryngol Otol 1996;110:425-8. 16. Shaw CL, Cowin A, Wormald PJ. A Study of the normal temporal healing pattern and the mucociliary transport after endo-

Otolaryngology– Head and Neck Surgery Volume 131 Number 4

17. 18. 19. 20.

scopic partial and full thickness removal of nasal mucosa in the sheep. Immunol Cell Biol 2001;79:145-8. Shaw CL, Cowin A, Wormald PJ. Standardization of the sheep as a suitable animal model for studying endoscopic sinus surgery. Aust J Otolaryngol 2001;4:23-6. Hollinshead WH. The head and neck. In: Anatomy for surgeons, vol. 1, 3rd ed. Philadelphia: Harper and Row; 1982. pp. 223-67. Kennedy D, Zinreich SJ, Shaalan H, et al. Endoscopic middle meatal antrostomy: theory technique and patency. Laryngoscope 1987;98(suppl 43):1-9. Setliff RC. Minimally invasive sinus surgery: the rationale and the technique. Otolaryngol Clin North Am 1996;29:115-29.

BRUMUND et al 533

21. Kirihene RKDRA, Rees G, Wormald PJ. The Influence of the size of the maxillary sinus ostium on the nasal and sinus nitric oxide levels. Am J Rhinol 2002;16:261-4. 22. Catalano P, Rottman E. Outcome in patients with chronic sinusitis after the minimally invasive sinus technique. Am J Rhinol 2003;17:17-22. 23. Graham SM, Carter KD. Combined approach orbital decompression for thyroid-related orbitopathy. Clin Otolaryngol 1999;24:109-13. 24. Lane AP, Bolger WE. Endoscopic transmaxillary biopsy of pterygopalatine space masses: a preliminary report. Am J Rhinol 2002;16:109-12.