Osborn AJ, Leung R, Ratjen F, James AL. Effect of Endoscopic Sinus Surgery on Pulmonary Function and Microbial Pathogens in a Pediatric Population With Cystic Fibrosis. Arch Otolaryngol Head Neck Surg. 2011;137(6):542-547. doi:10.1001/archoto.2011.68
To assess whether improvements in pulmonary function and microbial pathogenic findings can be achieved by endoscopic sinus surgery in a pediatric population with cystic fibrosis.
Retrospective medical record review.
Forty-one patients with cystic fibrosis who had undergone endoscopic sinus surgery at a single tertiary academic pediatric otolaryngology practice.
Main Outcome Measures
Changes in pulmonary function test values or respiratory tract microbial pathogens after endoscopic sinus surgery were examined.
Endoscopic sinus surgery did not improve pulmonary function test results in this population. Examination of respiratory tract microbial colonization showed that endoscopic sinus surgery did not affect microbial pathogens. The most common organisms isolated were Staphylococcus aureus and Pseudomonas aeruginosa.
The lack of effect of endoscopic sinus surgery on pulmonary function test results and respiratory tract microbial pathogens in our study highlights the need for prospective assessments of postoperative quality-of-life improvement and of adjunct medical therapy efficacy.
Pathologic conditions of the upper airway have been shown to influence pathologic conditions and function of the lower airway, leading to development of the unified airway model as reviewed by Krouse et al.1 Demonstrations of the unified airway model have been most valid in allergic rhinitis and chronic rhinosinusitis and their respective effects on asthma. Topical corticosteroids2,3 and subcutaneous immunotherapy4 for allergic rhinitis have been shown to decrease asthma symptoms, reduce bronchiolar reactivity, and improve performance on pulmonary function tests (PFTs). Similarly, medical5 and surgical6,7 treatment in the case of chronic rhinosinusitis improves asthma symptoms or pulmonary function.
It is tempting to apply the unified airway model to cystic fibrosis (CF), a single-gene recessive disorder affecting approximately 1 in 3000 births of white race/ethnicity. These patients have mutations inactivating the CF transmembrane conductance regulator, an adenosine triphosphate–dependent chloride channel.8 Dysregulation of chloride transport leads to thickened static mucus along the respiratory tract, subsequently causing recurrent infection, bronchiectasis, and chronic sinusitis. The incidence of rhinosinusitis approaches 100% in patients with CF,9 and nasal polyposis is seen in up to 86% of children with CF.10 Furthermore, a direct correlation has been established between severe sinonasal symptoms and poor pulmonary function, suggesting that treatment of sinonasal pathologic conditions might improve pulmonary function in accord with the unified airway model.11
The effect of sinus surgery on CF was initially reported in a small study12 among adults and showed that pulmonary symptoms, but not function, improved after surgery. The largest study13 to date consisted of 66 young adult and pediatric patients and found that surgical intervention did not improve pulmonary function as measured by forced vital capacity (FVC) and forced expiratory volume in the first second of expiration (FEV1). However, a reduction in hospital days was observed in the 6 months following surgery. Although smaller studies3,14,15 have been published, a large study limited to a young pre–lung transplantation pediatric population is missing from the literature. Herein, we examine the effects of endoscopic sinus surgery on PFT results and bacterial colonization in such patients with CF.
Approval for this study was granted by the Hospital for Sick Children Research Ethics Board, Toronto, Ontario, Canada. A retrospective medical record review of pediatric patients with CF undergoing endoscopic sinus surgery for nasal polyps was performed. The principal indication for surgery was symptomatic nasal obstruction. All patients underwent bilateral maxillary antrostomy, anterior ethmoidectomy, and as much debridement as possible of the nasal and sinus polyps during the initial surgical procedure. Revision surgery consisted of debridement of recurrent polyps and of any scar tissue that was causing occlusion of the ethmoid or maxillary outflow tracts. Therefore, the postoperative status of patients after both initial and revision surgical procedures should be equivalent.
The Hospital for Sick Children surgical database was searched for the terms nasal polyp, polypectomy, endoscopic sinus surgery, ESS, and FESS over the last 10 years. These patients were then cross-referenced with patients in the CF pulmonary database. Diagnosis was confirmed in all patients by sweat chloride testing.
Pulmonary function test values were extracted from the respirology department laboratories for a period of 1 year before surgery and 1 year after surgery. Values obtained were FEV1, FVC, and the forced expiratory flow of the midexpiratory phase (FEF25%-75%). These values were subjected to standardized conversions to yield the percentage of expected performance based on sex and height.16
Nonparametric statistical methods were used because PFT values did not follow a normal distribution. Accordingly, Wilcoxon matched-pair signed-rank test was used to compare preoperative and postoperative values, with each patient serving as his or her control. McNemar χ2 test was performed for analysis of the effect of surgery on microbial pathogens because the samples were paired.
Our study included 41 patients, 24 male and 17 female. The mean age at surgery was 11.9 years (age range, 5-18 years). These 41 patients underwent a total of 91 procedures, with a median of 2 (range, 1-11) procedures per patient. Eighteen patients underwent a single procedure, while 23 required revision surgery. Preoperative and postoperative data were available for 85 of these procedures. The mean populationwide values for FVC, FEV1, and FEF25%-75% were 87.6%, 83.0%, and 68.1%, respectively. The median values for FVC, FEV1, and FEF25%-75% were 89.4%, 86.8%, and 66.6%, respectively. Preoperative and postoperative means and medians are given in Table 1. We also compared PFT values of patients who required only 1 surgical procedure with those of patients who required at least 1 revision. By Mann-Whitney test, preoperative and postoperative FVC, FEV1, and FEF25%-75% values for patients who had multiple procedures were significantly better than those for patients who had only 1 procedure (P < .05 for all). There was no difference in age distribution between the 2 groups (P = .37).
We examined PFT data in several ways, bearing in mind the limitations of a retrospective study. Initially, we evaluated each procedure and PFT values that were temporally closest to each procedure before and after surgery (Table 2). In most of these cases, PFTs were performed within 3 months preceding or following surgery. Preoperative and postoperative PFT results for each of 85 procedures were compared using the Wilcoxon matched-pair signed-rank test. This initial analysis failed to reveal any improvement or worsening of pulmonary function after surgery. We elected to stratify our analysis under the assumption that those individuals with PFT scores above 80% of predicted would not be likely to show improvement. Accordingly, we performed the Wilcoxon matched-pair signed-rank test for all procedures and then only for those in which the patient's pulmonary function was less than 90%, less than 80%, and less than 70% of the predicted value before surgery (Table 2).
We considered the possibility that using the mean PFT values over 12 months preceding and following surgery might give a better picture of the patient's overall pulmonary function and the effect of surgery. However, some patients had multiple procedures over a period of months. This complicates the analysis because some PFT values would be considered preoperative values for multiple procedures or considered preoperative values for some and postoperative values for others. Accordingly, we eliminated all procedures that were performed less than 2 years before or after another sinus surgery. This ensured that preoperative PFT values truly reflected a baseline state and that the influences of any prior surgery were not affecting the pulmonary function of the patient. This also eliminated multiple surgical procedures contributing to the data set by a handful of patients and reduced any skewing of the results due to those patients' being overrepresented. Fifty-six procedures were analyzed using Wilcoxon matched-pair signed-rank test, and no significant differences were found between preoperative and postoperative values.
Patients often undergo weekly PFTs during an acute CF exacerbation. Therefore, this might bias the 12-month mean PFT values to be lower than the patient's true baseline. To remove this and other elements of variability from preoperative and postoperative PFT values, we compared the best PFT values in the 12 months preceding surgery with the best values in the 12 months following surgery. Again, only those 56 procedures that were separated from other sinus surgical procedures by 2 years were considered. A statistically significant decrease in FEF25%-75% was seen among the unstratified patients after surgery (P = .04) (Table 2). This decrease disappeared once patients with poorer preoperative PFT results were considered separately.
Finally, we considered the possibility that the effects of surgery might be short lived and that an effect might be seen only when comparing the patient's preoperative baseline with the first PFT performed after surgery. Accordingly, we compared 6-month and 12-month preoperative PFT means with the values of the first PFTs performed after surgery. The FVC trended toward a significant improvement after surgery when the 6-month preoperative means were compared with the first postoperative value. We further restricted the procedures analyzed by mandating that this postoperative PFT had to be performed within 3 months of surgery, and when we did so, there was a statistically significant improvement in FVC after surgery (P = .02). The mean change in FVC was an improvement of 9.2% after surgery in this subset of the patient population. A similar change was not seen when considering the 12-month preoperative PFT mean.
Our patient population had a total of 552 respiratory tract cultures obtained during the study period from sputum or throat swabs. Staphylococcus aureus and Pseudomonas aeruginosa were the most commonly isolated organisms and were seen 33% and 31% of the time, respectively (Table 3). Other organisms included Aspergillus species, Haemophilus influenzae, Streptococcus, yeast and Candida species, Burkholderia cepacia, Stenotrophomonas maltophilia, and Alcaligenes species. Escherichia coli and Klebsiella species were not isolated. Cultures revealing no organisms were obtained approximately 24% of the time. Preoperative and postoperative cultures were available for 62 procedures, and McNemar χ2 test for paired samples did not show a significant effect of surgery on the number of cultures that revealed no organisms (Table 4). Of 39 procedures for which preoperative and postoperative cultures revealed organisms, the cultured flora was identical in 19 cases (49%). Among 20 cases in which there was an alteration in flora, there were 8 episodes in which a new pathogen appeared after surgery, 6 episodes in which one of the preoperative pathogens disappeared, and 6 episodes in which preoperative and postoperative pathogens were entirely different. No single pathogen was more common before or after surgery to any statistically significant degree; however, it should be noted that this sample size is small.
This retrospective review was conducted to determine the effects of endoscopic sinus surgery on PFT values, microbial pathogens, and hospital days in a pediatric population with CF. Our results overall indicate that sinus surgery does not significantly affect objective measures of pulmonary function in this population. This is in accord with previous studies12- 14,17 among adult or mixed populations. Why does one fail to see an effect of surgery on pulmonary function as is seen with asthma? One explanation stems from the infection-free status of the lungs in patients with asthma. The only source of airway inflammation in these patients may be from the sinonasal region; therefore, addressing this source of inflammation can dramatically reduce the reaction in the lower airway. In contrast, CF leads to chronic infection of both the sinuses and the lungs. Eliminating the inflammation or bacterial load in the sinuses still leaves a significant inflammatory source in the lungs.
The importance of end-organ infection and inflammation is emphasized when one examines the treatments that improve PFT values in CF. Inhaled tobramycin, which reduces pulmonary bacterial load, has been shown to improve FEV1 in a randomized placebo-controlled crossover study.18 Furthermore, chest percussion therapy and other mucus clearance techniques have been shown to lead to improvement in PFT values.19 Typically, the improvements in PFT values from inhaled antibiotics or mucous clearance techniques are modest outside the setting of a CF exacerbation, highlighting the fact that even directed therapy does not drastically change these values.
When we compared patients who underwent 1 procedure with those who had at least 1 revision, we found that PFT values were worse in the group who underwent 1 procedure. This implies an inverse relationship between pulmonary function and sinonasal disease, which is surprising given previous results that seem to correlate better sinonasal symptom scores with better PFT values in children with CF.11 A possible explanation for this finding may be selection bias in determining which patients will undergo surgery. Patients with higher PFT values might be healthier overall, represent better surgical candidates, and be referred for surgery more often. Counterparts with lower PFT values might be managed medically to avoid the potential morbidity associated with surgery and general anesthesia.
Our analysis shows that FVC increases after surgery under certain circumstances. We stratified our procedures based on preoperative PFT values and found that FVC, but not FEV1 or FEF25%-75%, trended toward improvement after surgery as the preoperative value decreased. This is somewhat surprising because lung disease caused by CF is primarily obstructive and is best monitored by FEV1 or FEF25%-75%.20 Restrictive lung disease is seen in few patients with CF, and our population showed a similar incidence of this PFT pattern as demonstrated previously.21 Little has been reported concerning restrictive lung disease in CF, and it is possible that surgery has an effect on the worst cases, as shown in our study. Another explanation, and perhaps more likely, is that other perioperative treatments or inherent variability of PFTs will naturally lead to an “improvement” after surgery when considering only the worst presurgical cases (ie, a regression-to-the-mean bias). We believe that a similar explanation is applicable to the “decrease” in FEF25%-75% seen after surgery in one of our analyses. The fact that consistent results are not obtained when analyzing the data in multiple ways also supports this interpretation.
We analyzed the data in different ways for 2 reasons. First, after initial analysis failed to demonstrate any effect of surgery on objective measures of pulmonary function, we elected to see if any more subtle or less straightforward effect might be seen. Second, demonstrating that an effect or phenomenon exists from a logical and practical standpoint is often more straightforward than demonstrating that it does not. Proving that surgery has an effect would require 1 positive result, while demonstrating that it does not requires multiple negative results.
The most common pathogens isolated in our study were S aureus and P aeruginosa. This is similar to previous work, although our rates of colonization with both pathogens seemed lower than those described elsewhere.22,23 This was also the case for the more recently emerging bacterial species, such as B cepacia and S maltophilia. Surgery did not seem to have an effect on culture positivity for organisms. This is not surprising for at least 2 reasons. First, surgery did not primarily address the presence of any infection in the lungs. Second, previous work studying patients undergoing lung transplantation has demonstrated that pulmonary recolonization occurs within 2 to 3 weeks after transplantation.24,25 This does not seem to be significantly altered by pre–lung transplantation sinus surgery.24 Accordingly, any effect that sinus surgery might have on pulmonary microbial pathogens would be expected to disappear within 2 or 3 weeks after the procedure. This interval is too small to be captured in our retrospective analysis because most postoperative cultures were obtained outside of this critical window.
Previous studies12,13 have demonstrated a decrease in the need for hospital admission among patients with CF in the months following sinus surgery. However, the frequency of hospital admission for exacerbation of pulmonary disease among our population was too infrequent to detect a significant change in hospital stays during the period following surgery.
The primary limitation of our study is its retrospective nature. Because data were not prospectively collected, there was little coordination between the surgical and medical teams concerning the timing of PFTs or cultures with respect to surgery. Accordingly, some preoperative and postoperative tests analyzed were performed 3 or more months before and after the surgery date, respectively. Therefore, our study was biased to detect only long-lasting effects of surgery on pulmonary function and microbial pathogens. A prospective study would be able to uniformly assess patients at short predetermined preoperative and postoperative intervals. Although this would allow short-lived improvements in pulmonary function or bacterial colonization to be detected, the clinical significance of any such findings may be limited in the absence of longer-term benefit.
In addition, the retrospective nature of our study necessarily limits its focus to objective measures of pulmonary function. The FEV1 is accepted by the US Food and Drug Administration as a clinical outcome measure for trials involving CF. It correlates with prognosis and responds to treatment in the setting of exacerbation. However, it is worth noting that PFT values are by no means the sole standard of pulmonary health for patients with CF, and subjective improvements can occur in the absence of PFT value improvements. Future studies are needed to focus specifically on the quality-of-life improvements and other subjective or symptomatic improvements from which this population might benefit after surgery.
We believe that a lack of apparent benefit from endoscopic sinus surgery with respect to objective measures of pulmonary function and microbial colonization does not obviate the need for surgery in these patients. Patients derive symptomatic benefit from surgery, and further characterization is feasible with established validated tools for measuring sinonasal symptoms.26 Nasal polypectomy with or without more extensive sinus surgery remains indicated for symptomatic relief of nasal obstruction, rhinorrhea, cosmetic deformity, and potentially anosmia. Chronic sinusitis in patients with CF is a tenacious entity that responds to surgery; however, adjunct medical therapy, such as saline lavage, topical antibiotic washes, and topical corticosteroids, is likely to prove essential for truly effective management of sinus disease in these patients. Routine antibiotic lavage has been shown to drastically reduce the need for revision sinus surgery in patients with CF.27
In conclusion, patients with asthma can demonstrate an improvement in objective and subjective measures of pulmonary function with effective management of upper airway sources of inflammation, such as infection. We observed no improvement in objective measures, such as PFT values or bacterial colonization, among our pediatric population with CF after endoscopic sinus surgery. Prospective studies are needed to evaluate potentially short-lived effects of surgery on these variables, as well as to assess subjective or symptomatic improvements from a quality-of-life standpoint.
Correspondence: Adrian L. James, MA, DM, Department of Otolaryngology–Head and Neck Surgery, University of Toronto, Hospital for Sick Children, 555 University Ave, Toronto, ON M5G 1X8, Canada (firstname.lastname@example.org).
Submitted for Publication: January 5, 2011; final revision received February 23, 2011; accepted March 22, 2011.
Author Contributions: Drs Osborn and Leung had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Osborn, Leung, Ratjen, and James. Acquisition of data: Leung. Analysis and interpretation of data: Osborn, Leung, Ratjen, and James. Drafting of the manuscript: Osborn. Critical revision of the manuscript for important intellectual content: Osborn, Leung, Ratjen, and James. Statistical analysis: Osborn. Study supervision: Ratjen and James.
Financial Disclosure: None reported.
Previous Presentation: This study was presented as a poster at the 2011 Annual Meeting of American Society of Pediatric Otolaryngology in conjunction with the Combined Otolaryngological Spring Meetings; April 29 to May 1, 2011; Chicago, Illinois.