Introduction Abbreviations used in this article: BMP, bone morphogenic protein, CT, computed tomography, ENT, ear, nose and throat, FEV1, forced expiratory volume in 1 second, FVC, forced vital capacity, PFT, pulmonary function tests, TGF, transforming growth factor, TLC, total lung capacity, TO, tracheobronchopathia osteochondroplastica Tracheobronchopathia osteochondroplastica (TO) is a rare disorder of unknown cause affecting the tracheobronchial tree. It is characterized by the presence of multiple osseous and/or cartilaginous submucosal nodules protruding into the lumen of the trachea and large bronchi, typically sparing the posterior wall. Because of its usually benign course, TO was formerly an incidental finding at autopsy. It is presently increasingly discovered at fiberoptic bronchoscopy and/or chest computed tomography (CT) usually performed for other reasons. The first case of TO was described by Wilks (83) in 1857 under the title of "Ossific Deposits on the Larynx, Trachea and Bronchi," referring to the autopsy findings in a 38-year-old man who died from tuberculosis. In 1863, Virchow (82) considered TO to be an "ecchondrosis and exostosis" arising from the normal tracheal rings. The entity was named "tracheopathia osteoplastica" in 1910 by Aschoff (4), who considered it to be a disorder of the connective tissue, specifically affecting the internal elastic membrane of the trachea and major bronchi. Later, special interest was directed to the association of TO, atrophic rhinitis, and Klebsiella ozaenae(36,49,62,73), the latter being named after its isolation from patients with ozena, the other designation of atrophic rhinitis (1). Hence, most published clinical papers on TO refer to single cases or small series, the only previous large study being a report of 30 cases published in 1977 (32). The present study was undertaken to characterize better the clinical, functional, imaging, and endoscopic features of TO in a large series of cases, to estimate the rate of disease progression, and to determine its morbidity and mortality. Methods Case recruitment This retrospective multicentric study was undertaken by the Groupe d'Etudes et de Recherche sur les Maladies "Orphelines" Pulmonaires (GERM"O"P), a French group of pulmonary physicians interested in rare (so called "orphan") pulmonary diseases, and distributed nationwide, especially in university centers. Cases were recruited through the GERM"O"P's registry of rare pulmonary diseases. Starting from April 1995, participating physicians were requested to report to the registry all cases of selected rare pulmonary diseases, including TO. Records were nominative for patients who gave written consent, and anonymous otherwise. Data were collected by a coordination center based in Lyon, France. Data collection In March 1998, a detailed questionnaire was sent to all physicians who had reported cases of TO to the registry. Questionnaires were completed by reviewing the medical records. Retrieved items included past medical history, occupation, smoking status, symptoms and signs at presentation, laboratory data including markers of phosphate and calcium metabolism, bacteriologic analysis of bronchial washings, pulmonary function tests, treatment, and outcome. A copy of histopathologic reports was requested. Detailed charts were used to retrieve findings at fiberoptic bronchoscopy, based on reports of the procedure. Physicians were asked to evaluate the localization (anterolateral and/or posterior), elevation (mild, moderate, or important), and size of submucosal nodules, as well as the degree of airway obstruction if present. To estimate the concentration of nodules on the mucosa, lesions were classified as "scattered" (few nodules leaving areas of normal mucosa between them), "diffuse" (numerous nodules covering the entire mucosa without spared areas between them) or "confluent" (numerous nodules to the point that fusion occurs). Chest X-rays were evaluated for tracheobronchial stenosis and calcifications, atelectasis, parenchymal and mediastinal opacities, and possible tuberculous sequelae. Chest CT scans were analyzed for the presence and location of submucosal nodules, submucosal calcifications (distinct from calcifications of cartilaginous tracheal rings), tracheal deformity, mediastinal lymph nodes, bronchiectasis, atelectasis, pneumonia, and opacities consistent with possible tuberculous sequelae. Non-numerical variables were retrieved as binary propositions. Dates of all events were recorded. Data collection ended in May 2000. Numerical data were expressed as mean ± standard deviation, unless stated otherwise. Selection of cases The inclusion criterion used in the present series was the presence of typical endoscopic findings of TO consisting of characteristic multiple hard nodules, arising from the submucosa and projecting into the lumen of the tracheobronchial tree (46), with a beaded (60) to stalactitic cave appearance (81). Histopathologic diagnosis of TO was defined as the presence of either cartilage or bone in the submucosa at laryngotracheobronchial biopsies (2,21,37,59). However, histologic proof of TO was not required for inclusion. Data analysis All questionnaires were carefully reviewed. Ventilatory defects on pulmonary function tests were classified as obstructive ventilatory defects when the forced expiratory volume in 1 second/ forced vital capacity (FEV1/FVC) was <70%; as restrictive ventilatory defects when the total lung capacity (TLC) was <80% predicted or FVC was <80% predicted with FEV1/FVC >70%; or as combined ventilatory defects when the TLC was <80% predicted with FEV1/FVC <70% (See Table 3). Severity of obstructive ventilatory defect was further classified as mild (FEV1 >70% predicted), moderate (50% predicted < FEV1 < 70% predicted), or severe (FEV1 < 50% predicted). Evolution of pulmonary function over time was analyzed if 2 or more tests were available at least 1 year apart. Obstructive deterioration was considered when FEV1/ FVC decreased by more than 15% compared to baseline, and restrictive deterioration when TLC or FVC decreased by more than 15% compared to baseline. This 15% threshold was arbitrarily chosen to take into account both intraindividual and interinstitutional variations. Radiologic follow-up was assessed when 2 or more chest CT scans were available at least 1 year apart. New nodules, recent submucosal calcifications, or recent tracheal stenosis were considered as radiologic progression. In patients who underwent 2 or more fiberoptic bronchoscopies at least 1 year apart, disease progression was assessed by performing systematic site-by-site comparison of bronchoscopy reports, as recorded on data charts. For each patient and at each site of the entire laryngotracheobronchial tree, evolution was defined as stability, progression, or regression, with separate assessment of anterolateral and posterior tracheal walls. Progression was defined as the presence of new lesions, or modification in concentration or size of preexisting nodules. Progression was considered minimal when the changes were present at only 1 site of the tracheobronchial tree and lesions did not worsen previous obstruction. Progression was considered significant when new lesions or an increase in nodules size or concentration occurred in more than 1 site, or resulted in aggravation of preexisting airway stenosis.TABLE 3: Initial pulmonary function tests in 28 patients with TOResults Study population By May 1999, 51 cases of TO had been reported to the GERM"O"P registry. Forty-one detailed questionnaires (80%) were returned and completed for analysis by May 2000. After review, all 41 cases were considered eligible for inclusion in the present study. Eight of these cases have been previously published as case reports (7,17,24,63). Cases were diagnosed between 1951 and 1999, the majority (83%) after 1990. They included 22 women (54%) and 19 men (46%). The mean age at diagnosis was 63 ± 15 years (range, 25–85 yr; median, 63 yr). Only 10/41 patients (24%) were active or former smokers. The time interval between the first specialized medical visit for symptoms attributable to TO and diagnosis was 2 ± 5 years (range, 0–18 yr; median, 0.1 yr). The mean follow-up duration after diagnosis was 5 ± 8 years (range, 0–45 yr; median, 3 yr). Past medical history Twelve patients (29%) had a history of chronic or recurrent ear, nose, and throat (ENT) disease including atrophic rhinitis, sinusitis, or pharyngitis. Twenty-one (51%) had a history of chronic or recurrent lower respiratory tract disease (tracheitis, bronchitis, bronchiectasis, or chronic obstructive pulmonary disease). Two patients (5%) had a diagnosis of asthma, and 8/41 (20%) had a history of asthma-like symptoms. One patient had active dermatomyositis diagnosed before TO, and another had multiple myeloma. Nine patients (22%) had a past history of tuberculosis. One patient had idiopathic pulmonary fibrosis. One patient had been professionally exposed to silica, and another to glass fiber particles. Only 6/41 (15%) had no history of previous respiratory disease. Sarcoidosis, polychondritis, polyarthritis, and acromegaly were not reported. Family history did not disclose any case of TO in patients' relatives. In summary, most patients presented with a history of nonspecific chronic respiratory symptoms. Symptoms and signs at presentation The main symptoms at presentation are summarized in Table 1. Thirty-six of 41 patients (88%) had ENT or respiratory symptoms at the time of diagnosis, and 5/41 (12%) were asymptomatic. The onset of symptoms was acute in 15/36 patients (42%), and insidious in 21/36 (58%). The mean duration of symptoms before diagnosis was 4 ± 6 years. The most frequent symptom was chronic cough, present in 22/41 (54%). It was usually associated with other respiratory symptoms, and was isolated in only 2/41 (5%). Eight patients (20%) had hemoptysis as a main presenting symptom. Ten patients (24%) presented with acute pneumonia, 2 of whom had recurrent episodes. Dysphonia was present at diagnosis in 4/41 patients (10%). Two patients (5%) presented with unexpected difficulties in intubation, attributable to TO. In 1, intubation was required for acute respiratory failure in chronic obstructive pulmonary disease. Intubation attempts were unsuccessful due to severe tracheal stenosis, and emergency tracheostomy was performed. In the other patient, previously asymptomatic TO first manifested as difficult but successful intubation before anesthesia for elective breast cancer surgery. In the 4 other asymptomatic patients, TO was an incidental finding at fiberoptic bronchoscopy performed respectively for asymptomatic atelectasis, pleural effusion, pneumothorax, and isolated epistaxis. Chest auscultation was abnormal in 26/41 cases (63%), disclosing wheezing (32%), and/or ronchi (27%). Only 1 patient (2%) had stridor at diagnosis. Chest physical examination was unremarkable in 15/41 (37%) patients.TABLE 1: Symptoms at presentation in 41 patients with TOIn summary, chronic cough, sputum production, acute and chronic dyspnea, and hemoptysis were the most frequent features at diagnosis. Laboratory data At the time of diagnosis of TO, 2/28 (7%) patients had a white blood cell count > 10,000/mm 3 , 13/20 (65%) had a sedimentation rate >20 mm/h, 7/11 (64%) had a C-reactive protein level >5 mg/L, and 12/12 (100%) had serum phosphate and calcium levels within the normal range. In summary, apart from nonspecific markers of inflammation, laboratory data did not disclose any relevant abnormality. Microbiology Microbiologic findings at bronchial aspiration during fiberoptic bronchoscopy are summarized in Table 2. One or more infectious agents were identified in 25/41 (61%) patients. Cultures were positive for Gram-negative bacteria in 27 cases, Grampositive bacteria in 4, fungi in 5, and mycobacteria in 3. The most frequent strains were K. ozaenae, isolated in 8/41 (20%) of cases, and Pseudomonas aeruginosa, found in 7/41 (17%). Fourteen patients had symptoms of acute respiratory infection with 1 or more bacteria identified. Eleven others had acute symptoms, but no bacteria were found. Thus, among patients with acute respiratory symptoms, an infectious agent was identified in 14/25 (56%). Finally, bacteria were isolated in 5 asymptomatic patients. Insufficient data precluded any precise quantitative assessment of bacterial cultures.TABLE 2: Bronchial microbiology in 41 patients with TOPulmonary function tests Twenty-eight patients (68%) had pulmonary function tests (PFT) performed at the time of diagnosis. The frequency of PFT abnormalities is presented in Table 3. The mean values at initial PFT were FEV1 91% ± 22% predicted, FVC 94% ± 19% predicted and FEV1/FVC 74% ± 10%. Forty-three percent of patients had normal spirometry, and 39% had an obstructive ventilatory defect (2 mild, 6 moderate, and 3 severe). Five (12%) had no other identifiable cause than TO for their obstructive defect, whereas 5 others had a diagnosis of chronic obstructive pulmonary disease and 1 had a history of asthma. No patient had a combined defect. The peak flow was reduced in 9/14 (64%). Flow-volume loops were abnormal in 5/16 (31%), disclosing an expiratory plateau in 3 cases and an inspiratory plateau in 2. In all 5 patients, nodules were equally distributed along the trachea and main bronchi at fiberoptic bronchoscopy. An 80% obstruction of large airways and a 50% obstruction of main bronchi was present in 1 patient with an expiratory plateau, and a 40% obstruction was present in another patient with an inspiratory plateau. Expiratory collapse was seen in 1 patient with an inspiratory plateau and in 1 other with an expiratory plateau. Follow-up evaluation of lung function was possible in 11/41 patients who had 2 PFT at least 1 year apart. The mean time interval between first and second study was 5 ± 3 years (median, 5 yr). When comparing the mean values of FEV1, FVC, and FEV1/FVC in this subgroup, no significant deterioration was observed (data not shown). When evaluating each patient individually, 2/11 (18%) had an aggravation of a preexisting obstructive defect, defined as a decrease of FEV1/FVC >15% compared to baseline. Functional deterioration was explained by endoscopic progression of TO in both patients. Thoracic imaging Chest X-ray data were available in 38 patients. Tracheobronchial stenosis was suspected in 8/38 (21%), tracheobronchial calcifications in 4/38 (11%), atelectasis in 4/38 (11%), pneumonia in 10/38 (26%), and tuberculous sequelae in 7/38 (18%). Chest CT scan was available for 31 patients. The main findings are summarized in Table 4. Dense submucosal nodules, present in 74% of cases, were localized on the anterolateral tracheal walls in all patients, except for 1 case with both anterolateral and posterior nodules (Figure 1A and 1B). Submucosal calcifications were found in 61% of cases, and tracheal stenosis in 10%. The mean length of tracheal involvement, measured in 11 patients, was 7.1 ± 3.5 cm.TABLE 4: Chest CT scan findings at diagnosis in 31 patients with TOFig. 1: A. Chest CT scan showing diffuse anterolateral nodules protruding into the tracheal lumen. B. Chest CT scan of the upper trachea. Note the multiple anterolateral calcified nodules sparing the posterior walls.Ten patients had a second chest CT scan. In 2 (20%), progression was evidenced by the presence of new nodules, with submucosal calcifications in 1 case. Tracheal stenosis was absent on the initial scan in all 10 patients, but a mild tracheal stenosis was seen in 1 patient on the second examination. Tracheobronchial endoscopy All patients underwent fiberoptic bronchoscopy, disclosing a typical aspect of TO. Nodules were most frequently distributed along the proximal tracheobronchial tree, especially in the upper and middle trachea, affected in 76% and 78% of cases, respectively. The lower trachea was involved in 66% of cases, and the carina in 34%. Seventy-one percent had lesions along the whole trachea. The right and left main bronchi were equally affected (44% and 41% of cases, respectively). Lobar and segmental bronchi were less frequently involved (18% and 17% of cases, respectively). The infraglottic larynx was affected in only 10%. No patient had supraglottic lesions. Nodules were restricted to the anterolateral walls except in 6 patients (15%) who also had posterior involvement. In the trachea, the concentration of lesions was most frequently evaluated as diffuse (56%), and less commonly confluent (29%) or scattered (15%). In contrast, in the main bronchi, concentration was most frequently evaluated as scattered (54%), less commonly diffuse (39%), and rarely confluent (7%). Nodules elevation in the trachea was usually moderate or mild (82%), and less frequently important (18%) (Figure 2A and 2B). Airway obstruction was observed in the infraglottic larynx in 1 (2%) patient, in the trachea in 9 (21%), in the main bronchi in 4 (10%), and in the lobar/segmental bronchi in 2 (5%). A 50% reduction of airway caliber was found at the tracheal level in only 3 patients, with concomitant bronchial stenosis in 1 of them and segmental stenosis in another.Fig. 2: A. Tracheobronchial endoscopy showing diffuse anterolateral distribution of nodules resulting in the typical cobblestone appearance of tracheobronchopathia osteochondroplastica. The posterior wall is spared. B. Tracheobronchial endoscopy showing diffuse anterolateral nodules in the carina. Note the protruding sessile lesion at the left, giving the airway a stalactitic appearance.Hard consistency of the nodules was reported in 37/41 (90%), and grinding of the fibroscope on the lesions was noticed in 18/41 (44%). Saber-sheath trachea was observed in 6/41 (15%), and expiratory tracheal collapse in 13/41 (32%). The tracheal mucosa was ulcerated in 5/41 (12%), and hemorrhagic in 5/41 (12%). Tracheobronchial secretions were purulent in 21/41 (51%), mucous or serous in 18/41 (44%). Eighteen patients were evaluated twice, at least 1 year apart (mean interval, 7 ± 10 yr). No regression was seen. Ten of eighteen patients (55%) were stable, and disease progression occurred in 8/18 (45%). It was minimal in 5 patients (28%) and significant in 3 (17%). In 1 patient, nodules initially restricted to the trachea with a 30% lumen obstruction had extended along the entire tracheobronchial tree, with increase in lesion size and a consequent 80% obstruction of the tracheal lumen at the second bronchoscopy performed 9 years later. In summary, the disease largely predominated in the trachea and main bronchi, and was infrequent in more distal sites. The posterior wall was spared in most cases. The most frequent consequence was purulent infection of the tracheobronchial tree, whereas significant mechanical obstruction was uncommon. Disease progression over time was usually minimal or undetectable, but severe progression occurred in some cases. Histopathology Tracheobronchial biopsies were performed in 40/ 41 patients. Histopathology was considered diagnostic in 28/40 patients (70%). In 22/40 cases (55%), histologic diagnosis was achieved at first endoscopy, while in the other 18/40 cases (45%), the procedure did not allow histopathologic confirmation. Additional endoscopies were done in 8 of these 18 cases (up to 6 procedures in 1 patient), and allowed histopathologic diagnosis in 6 more patients. The main histopathologic findings are summarized in Table 5. The most frequent findings were the characteristic presence of bone in the bronchial submucosa (56% of cases) and squamous metaplasia of the tracheal epithelium (48%). Among 16 biopsies specifically examined for amyloidosis after Congo red staining, only 2/16 (13%) disclosed both TO and amyloidosis. Five of 6 patients with posterior tracheal involvement had histologically proven TO, with amyloidosis sought for in 3 of them and found in only 1. One patient had a small-cell carcinoma of the culmen and of the apical segment of the left lower bronchus, but lesions of TO were endoscopically restricted to the middle trachea, and the 2 disorders were considered unrelated. In another case with histologically confirmed TO, granulomatous giant-cell bronchitis was found in a lobar/segmental site remote from any TO lesion, but no specific etiology could be found (all cultures for mycobacteria were negative, and there was no further argument for sarcoidosis or other granulomatoses such as Wegener disease). Two of 3 patients with positive cultures for mycobacteria at bronchial washing had histologic confirmation of TO. The third patient did not have histologic confirmation, but, nevertheless, had typical disseminated, anterior tracheal and bronchial nodules characteristic of TO at endoscopy.TABLE 5: Histopathologic findings from tracheobronchial biopsies in 40 patients with TOTreatment Most patients received treatment directed to respiratory conditions associated with TO, such as infection, chronic obstructive pulmonary disease, bronchiectasis, or asthma. These consisted of inhaled corticosteroids in 8/41 (20%), systemic corticosteroids in 5/41 (12%), antibiotics in 7/41 (17%), bronchodilators in 9/41 (22%), and respiratory physical therapy in 8/41 (20%). Only 2 (5%) patients received specific treatment to relieve or bypass tracheal stenosis. One of them had emergency tracheostomy and was maintained on long-term home mechanical ventilation because of severe tracheal stenosis. Another patient received laser therapy, resulting in improvement. No surgical resection, curietherapy, cryotherapy, endoluminal prosthesis, or radiation therapy was done in our cases. Clinical outcome Ten of 41 patients (24%) were asymptomatic during follow-up, 5 of whom were already asymptomatic at diagnosis. The 31 other patients (76%) had chronic or recurrent symptoms, including cough (34%), sputum production (24%), hemoptysis (16%), and recurrent tracheobronchitis (22%). One patient died of septic shock of pulmonary origin, in the context of dermatomyositis with severe swallowing disturbances associated with severe tracheal stenosis secondary to TO (24). Four patients (10%) died of unrelated causes. Discussion This series of 41 patients with TO is, to our knowledge, the largest reported so far, and provides a thorough evaluation of imaging and endoscopic evaluation of TO, together with evaluation of the long-term clinical course. Study population In our study population the mean age at diagnosis was 63 years, in agreement with 2 previous series (41,81). Härma et al (32) reported a mean younger age at diagnosis of 47 years in a series of 30 patients with atrophic rhinitis who were systematically screened for TO. In any case, these figures probably poorly reflect the age of disease onset, since symptoms were usually present long before diagnosis: 4 ± 6 years in our series, and up to 18 ± 14 years in the series of Lundgren and Stjernberg (41). Disease may begin at various ages, even in the young adult. Our youngest patient was a 25-year-old woman, but TO has been reported even in an 11-year-old patient (32). This variable age at diagnosis may reflect different causal mechanisms of the disease. We found no difference in sex distribution, whereas a female predominance was reported in 2 previous series, with respectively 66% (n = 18) (81), and 73% (n = 30) of cases (32). In the latter series, the female predominance could partly result from the recruitment method, that is, patients with atrophic rhinitis, which is more frequently reported in women (14,68). A familial occurrence of TO has been previously reported in only 1 instance of a 83-year-old woman and her daughter diagnosed 14 years later at the age of 73 years (60). We found no family history of TO in our patients. Seventy-six percent of our patients were nonsmokers, in agreement with previous publications (41,78,81). Past medical history A history of chronic respiratory manifestations was found in 85% of our patients. Atrophic rhinitis was identified in 6 of our patients (15%). Atrophic rhinitis, also called ozena, is a chronic nasal disease characterized by progressive mucosal atrophy and resorption of the underlying bone, formation of thick crusts, and a distinct fetid odor (87). The simultaneous occurrence of atrophic rhinitis and TO was first described by Jepsen and Sorensen (36), and a causal relationship was suggested (32,36,62). They observed that the squamous metaplasia of the nasal and tracheal epithelium had similar characteristics, in support of Dalgaard's hypothesis of metaplasia of submucosal elastic fibers into elastic cartilage (22), eventually followed by calcification and ossification. Soulas et al (73) reported 4 cases of TO, of which 3 had a history of atrophic rhinitis. Rose (62) reported ozena in 9/10 patients with TO. Härma et al (32) reported that 18/30 of their patients with TO had a history of atrophic rhinitis and were thus specifically screened for TO in this context. The relatively low rate of atrophic rhinitis in our patients compared with the latter series probably results from different sources of recruitment, since an ENT clinic was the main source of recruitment for Härma et al (32). Symptoms and signs at presentation In this series, 88% of patients had 1 or more respiratory symptoms at presentation. Although these could retrospectively be attributed to TO in most cases, fiberoptic bronchoscopy was mainly performed in search of malignancy or infection, with TO being an incidental finding in all cases. Only 12% of our patients were asymptomatic at diagnosis. Hemoptysis was the presenting sign in 20% of our cases, and was attributed to TO in all but 1 patient who also had bronchiectasis. Clee et al (16) reported 2 cases with a history of recurrent hemoptysis before diagnosis of TO. Hemoptysis was also the presenting symptom in the unique familial case of TO (60). In a small series, 2 of 7 patients had important hemoptysis (12), and in another series 4/9 patients had hemoptysis associated with respiratory infections (41). TO was discovered unexpectedly following difficult intubation in 5% of our patients, a mode of presentation previously described in 2 reports (2,71). Chronic isolated cough as presenting symptom, found in 5% of our patients, has been previously reported (15,49,66). Histopathologic findings in 1 case disclosed absence of normal ciliated respiratory epithelium (15). The authors of this report speculated that cough could result from several factors, including air turbulence caused by irregularity of the trachea, increased sensitivity of the cough reflex, and sputum accumulation due to the lack of normal ciliated epithelium. TO was also diagnosed in 2/25 patients who underwent fibroscopy for chronic cough of undetermined origin which failed to resolve after conventional therapy (66). It is worth noting that 20% of our cases had a history of asthma-like symptoms, and 32% presented with wheezing at auscultation. TO simulating asthma has been described previously (56), and should be included in the differential diagnosis of large airways disorders mimicking asthma. The diagnosis of TO was thus frequently preceded by chronic or recurrent ENT or respiratory manifestations, in contrast with the usually silent course previously attributed to the disease at the time when it was commonly an incidental finding at autopsy (22). Laboratory data Except for a nonspecific inflammatory syndrome, all laboratory variables, including markers of phosphate and calcium metabolism, were normal in our patients. Phosphate and calcium metabolism was also reported to be normal in 5 patients with TO, 2 of whom had growth hormone hypersecretion (with overt acromegaly in 1 case) (74). Only 1 isolated case of TO with hyperphosphoremia was reported in the literature (26); probably an incidental association. Although it seems unlikely that TO could result from a single abnormality of phosphocalcic metabolism or growth hormone hypersecretion, this hypothesis has not been addressed systematically so far and may deserve more attention. Microbiology Although exposed to inhaled bacteria from the environment and pharyngeal flora, the normal lung is kept sterile by efficient defense mechanisms such as the mucociliary system (38,42,84). Presence of bacteria in the tracheobronchial tree reflects failure of these mechanisms (67). In TO, squamous metaplasia of the epithelium, chronic inflammation, and loss of normal airway architecture are likely to affect defense mechanisms, especially mucociliary clearance. Hence, bronchial cultures were positive in 61% of our cases. Similar to other chronic pulmonary diseases (38), TO leads to abnormal presence of bacteria in the lower respiratory tract. The most frequently identified strain in our TO population was K. ozaenae, found in 20% of patients, half of whom had a history of atrophic rhinitis. The second most frequent strain was P. aeruginosa, found in 7 patients (17%), of whom 3 had associated chronic obstructive pulmonary disease. Altogether, a bacterial infectious agent was identified in 56% of our cases presenting with acute respiratory symptoms, a rate similar to that observed in acute exacerbations of chronic bronchitis (67). In the latter, the most commonly identified bacteria are H. influenzae, S. pneumoniae, M. catarrhalis, and P. aeruginosa(67). This was similar to the findings of our study population, with the notable exception of K. ozaenae, usually not described in chronic bronchitis, which was the most common pathogen in our cases. Härma et al (32) found K. ozaenae in the tracheal cultures of 4/30 cases of their series. A case of TO associated with sinusitis due to K. ozaenae infection and positive bronchial cultures for both this bacteria and P. aeruginosa has been reported (40). Another patient with TO and a past history of atrophic rhinitis developed purulent tracheitis due to K. ozaenae, which