Pneumothorax

Background: Spontaneous pneumothorax is a complication that is commonly reported in patients with cystic fibrosis (CF). An understanding of the pathophysiology of this complication and its consequences is important for the management of patients with CF.

Objective: To identify risk factors associated with pneumothorax and to determine the prognosis of CF patients following an episode of pneumothorax.

Design: A retrospective observational cohort study of the National Cystic Fibrosis Patient Registry between the years 1990 and 1999.

Patients: The registry contained data on 28,858 patients with CF who had been followed up over Those 10 years at CF centers across the United States.

Results: Pneumothorax occurred with an average annual incidence of 0.64% and in 3.4% of patients overall. There was no increased occurrence by sex, but CF was more prevalent in older patients (mean [[+ or -] SD] age, 21.9 [+ or -] 9.1 years) with more severe pulmonary impairment (nearly 75% of patients with FE[V.sub.1] of < 40% predicted). The principal risks associated with an increased occurrence of pneumothorax included the presence of Pseudomonas aeruginosa (odds ratio [OR], 2.3), Burkholderia cepacia (OR, 1.8), or Aspergillus (OR, 1.3) in sputum cultures, FE[V.sub.1] < 30% predicted (OR, 1.5), enteral feeding (OR, 1.7), Medicaid insurance (OR, 1.1), pancreatic insufficiency (OR, 1.4), allergic bronchopulmonary aspergillosis (OR, 1.5), and massive hemoptysis (OR, 1.4). There is an increased morbidity (eg, increased number of hospitalizations and number of days spent in the hospital) and an increased 2-year mortality rate following pneumothorax.

Conclusion: Pneumothorax is a serious complication in CF patients, occurring more commonly in older patients with more advanced lung disease. Nearly 1 in 167 patients will experience this complication each year. There is an attributable mortality to the complication and considerable morbidity, resulting in increased health-care utilization and a measurable decline in lung function. (CHEST 2005; 128:720-728)

Key words: cystic fibrosis; lung disease; pneumothorax

Abbreviations: ABPA = allergic bronchopulmonary aspergillosis; CF = cystic fibrosis; CFF = Cystic Fibrosis Foundation; OR = odds ratio

**********

Pneumothorax is defined as air in the pleural cavity. Spontaneous pneumothorax is a complication that is commonly reported in patients with cystic fibrosis (CF). (1) The pathogenesis of spontaneous pneumothorax in the non-CF population has been attributed to an underlying disorder of the small airways, whereby inflammation causes airways obstruction resulting in alveolar hyperinflation. The lung ruptures into the pleural space in situations of increased transpulmonary pressure differences. (2) Intrapulmonary air pressure differences may increase in the case of a check valve phenomenon from mucus retention, (3) as occurs in CF patients. In addition, the occurrence of a spontaneous pneumothorax in CF is thought to be a bad prognostic indicator, (4) but it is not clear whether the severity of the lung disease or the pneumothorax is the independent predictor of shortened survival.

Using data from CF patients in the national Cystic Fibrosis Foundation (CFF) Patient Registry, (5) multiple clinical features were analyzed to try to determine the risk factors that may be associated with pneumothorax as well as to determine the prognosis of patients following an episode of pneumothorax.

MATERIALS AND METHODS

The CFF supports and accredits CF care centers nationwide. These centers provide specialized care for persons with CF, and offer comprehensive diagnosis and treatment, as well as participation in clinical trials of experimental therapies. The CFF Patient Registry contains data on patients with CF who have been seen at one of the accredited CF care centers in the United States. This represents about 90% of all CF patients in the United States. The CFF has sponsored the patient registry since 1966 and requires all CFF-accredited care centers to complete standardized questionnaires for all patients seen in their centers. (6) The questionnaires are completed by each center using the local resources available to them. The CFF serves as the coordinating center for data collection and quality, contacting the CF centers to retrieve or correct all missing and out-of-range values.

The CFF Patient Registry database was analyzed for the years from 1990 to 1999. For this period, the database contained 28,858 unique patients with a total of 197,156 separate entries (average number of entries per patient, 6.8). There were missing dates for 16,614 entries. This left a working database of 28,191 unique patients with 180,542 entries (average number of entries per patient, 6.4). This means that not all patients were seen and had reported data for all 10 years of the analysis. The CFF Patient Registry records only one entry per patient per year and so is a collection of year-end summary data. Complications are reported only if they are present during a specific year and may not reflect repeat occurrences. Therefore, the data are presented assuming no more than one episode of pneumothorax for each patient in a given year, although there could have been more. Patients who experienced a complication away from the CF center may not have had this included in the registry data.

Clinical variables that may influence pneumothorax were retrieved and analyzed. These included the following: (1) demographic data (eg, date of birth, sex, race, and insurance status); (2) complications related to CF (eg, pancreatic insufficiency, cirrhosis, diabetes, allergic bronchopulmonary aspergillosis [ABPA], and lung transplantation); (3) severity of disease (eg, FE[V.sub.1] percent predicted, FVC percent predicted, percentage of ideal body weight, body mass index, supplemental [tub] feeding, and supplemental oxygen); (4) organisms present in cultures of the airways (eg, Pseudollumas aeruginosa, Burkholderia cepacia, Staphylococcus aureus, methicillin-resistant S aureus, Stenotrophomonas maltophilia, Aspergillus, and atypical mycobacteria); (5) CF-related therapies (eg, dornase alfa, inhaled tobramycin, high-dose ibuprofen, and Corticosteroids); and (6) outcomes (eg, mortality, clinic visits, hospitalizations, hospital days, and IV antibiotic days).

Statistical analysis consisted of the [chi square] test for categoric comparisons, the Student t test for the comparison of normally distributed means, and logistic regression for the simultaneous control of multiple potential confounders. Logistic regression was used to obtain the independent effect of multiple risk factors for pneumothorax. The database did not collect all variables for all 10 years. For example, therapy with inhaled tobramycin, dornase alfa, and ibuprofen have been included in the database only since 1997. Thus, multiple models were constructed to include variables that were collected for specific time periods.

The change in lung function over time (eg, FE[V.sub.1]) was also assessed using a mixed-model analysis of repeated measures, which utilized restricted maximum likelihood methods for estimation. (7) Corey et al (8) demonstrated several advantages of mixed-model analysis in assessing pulmonary function decline in CF patients, including the ability to account for both between-subject and within-subject variability within the same model. The selection of a satisfactory covariance structure is important to make appropriate inferences from repeated-measures data. (7) Two models were fit with different covariance structures; an unstructured model and a first-order autoregressive model. The first-order autoregressive covariance model describes the correlation between observations in which adjacent observations are more highly correlated than observations that are further apart in time and was selected as the most appropriate for the data on the basis of information criteria assessing the fit of the model. A statistical software package used for these analyses (SAS Proc Mixed, SAS for Windows 9.0; SAS Institute; Cary, NC). The values reported are the mean [+ or -] SD, unless stated otherwise. A p value of < 0.05 was considered to be significant.

RESULTS

There were 965 patients (3.4% of total population) who had experienced at least one episode of pneumothorax. Two additional patients were reported to have had a pneumothorax, but there were no associated dates, and they were excluded from the study. There were a total of 1,180 entries with pneumothorax listed as a complication; most patients had only one recorded event (788 patients; 82%), although many (177 patients; 18%) had experienced more than one reported event, and one patient had 5 reported years with a pneumothorax (Table 1).

The average annual incidence of pneumothorax was 0.64% with no yearly increase over time (range, 0.57 to 0.74%; difference not significant). Episodes in different reporting years were counted as separate events. The mean age at the first pneumothorax episode was 21.9 [+ or -] 9.1 years (median age, 21.0 years; age range, 0 to 61 years). The occurrence of pneumothorax was more prevalent in adult patients, as can be seen in Figure 1. The proportion of pneumothorax patients who were < 18 years of age (27.6%) was far lower than those > 18 years of age (72.4%), which is different from the ratio in the entire population (57.1% vs 42.9%, respectively; p < 0.0001). The mean age of patients with pneumothorax for each year did not change during the 10 years analyzed (Fig 2).

There were 537 men (55.7%) and 428 women (44.3%) who ever had experienced a pneumothorax, which is similar to the proportion of the sexes in the entire database (men, 53.2%; women, 46.8%; difference not significant). The constitution of the population by race was 93.7% white, 3.0% African-American, and 1.8% Hispanic. This generally followed the racial distribution of all patients within the database (90.3%, 3.6%, and 5.5%, respectively), although there was a slight decrease in the occurrence of pneumothorax in the Hispanic population (p = 0.002).

Pneumothorax occurred in patients with a wide range of pulmonary function (including normal function) but occurred more frequently in patients with severe impairment of function(Table 2). Approximately 75% of patients had an FE[V.sub.1] of < 40% predicted. The association with the severity of pulmonary disease can be seen in Figure 3, in which there is a steep increase in numbers once FE[V.sub.1] falls below 50% predicted.

Risk factors

Specific variables that might be associated with the first episode of pneumothorax were evaluated, including infectious organisms, features of severity of disease, additional complications of CF, and specific therapies. All variables that met statistical significance in a univariate analysis are listed in Table 3. Logistic regression was used to examine the independent effects of risk factors simultaneously, and their adjusted odds ratios (ORs) are shown in Table 3. Patients with P aeruginosa (OR, 2.3), B cepacia (OR, 1.8), Aspergillus (OR, 1.3), FE[V.sub.1] < 30% predicted (OR, 1.5), enteral feeding (OR, 1.7), Medicaid insurance (OR, 1.1), pancreatic insufficiency (OR, 1.4), ABPA (OR, 1.5), and massive hemoptysis (OR, 1.4) had increased risk of pneumothorax. The only factor with a reduced risk was the presence of S aureus (OR, 0.9). There was also an association of pneumothorax with specific therapies, with an increased risk among patients using dornase alfa (OR, 2.1) and inhaled tobramycin (OR, 1.6), although these medications were included in the database for only the last 3 years.

Mortality

There were 469 deaths (48.6%) within the group of patients who had a pneumothorax. Patients who had experienced a pneumothorax had a higher occurrence of death than those who did not experience pneumothorax (48.6% vs 12.2%, respectively; p < 0.0001). The mean age at death in patients with pneumothorax was 23.5 [+ or -] 8.8 years (median age, 22 years; age range, 0 to 61 years), which was not different from the mean age of death (22.7 [+ or -] 10.8 years; median age, 22.7 years; age range, 0 to 72 years) of those who never had a reported episode of pneumothorax (difference not significant). Two hundred five patients (43.7%) who died did so within a year of experiencing a pneumothorax.

Patients with pneumothorax who died in the same year as the pneumothorax had worse pulmonary function than did those who survived at least 1 more year. The patients with pneumothorax who survived had a greater mean FVC (52.3 [+ or -] 19.9% vs 42.9 [+ or -] 17.3% predicted, respectively; p < 0.0001) and FE[V.sub.1] (37.7 [+ or -] 18.5% vs 28.0 [+ or -] 16.3% predicted, respectively; p < 0.0001) than did those who died in the same year. A Kaplan-Meier survival curve from the first pneumothorax episode is shown in Figure 4, demonstrating 50% mortality at about 4 years.

Since pneumothorax occurred in patients with more severe respiratory impairment, it was necessary to see whether the mortality was due solely to pulmonary impairment or could be attributed to the pneumothorax. The 2-year mortality rate was calculated for patients with pneumothorax, grouping them according to the severity of pulmonary impairment (ie, FE[V.sub.1] as a percentage of predicted values) at the time of the first episode of pneumothorax. These values were compared to the 2-year morality rates of patients who never had experienced an episode of pneumothorax, also grouping them according to their pulmonary impairment (Fig 5). The 2-year mortality rate was always greater for the patients with pneumothorax, with an attributable risk for mortality due to pneumothorax between 6.3% and 14.3%. The relative risk of dying for patients with pneumothorax and an FE[V.sub.1] of >50% predicted was 14.2 compared to those with similar pulmonary impairment but without pneumothorax.

Change in Lung Function

To evaluate the effects of pneumothorax on changes in lung function, data were analyzed from patients who had pulmonary function test results reported for all three time periods (ie, year before, year of, and year following pneumothorax; n = 326). There was a significant drop in both FVC and FE[V.sub.1] (in percentage of predicted values) from the year before the pneumothorax occurred to the year following (Table 2). The results of mixed-model analysis also showed a statistically significant change in lung function over time as measured by the change in FE[V.sub.1] (p < 0.001)

Utilization of Resources

The utilization of services by patients with pneumothorax was analyzed, including outpatient visits, hospitalizations, and IV antibiotics, comparing the year before, during, and following the occurrence of the pnenmothorax (Table 4). There was an increase in the median number of hospitalizations and hospital days in the year of and the year following the pneumothorax. The average number of office visits did not change, although the number of hospitalizations per patient increased in the year of the pneumothorax compared to the previous year, and there were greater overall numbers of each compared to the previous year. There was greater overall utilization of services in the year of pneumothorax, resulting in 2,275 more office visits, 1,573 more hospitalizations, and 23,748 more hospital days compared to the previous year.

DISCUSSION

Spontaneous pneumothorax is a well-known complication in patients with CF and is far more common than in the general population. In addition, this study shows that subsequent pneumothorax is common with more than one in five of the patients having experienced at least two events in separate years. It is impossible to know from this database whether the subsequent pneumothorax was more commonly ipsilateral or contralateral, but both have been described in the literature. (1)

The principle risk factors for pneumothorax are age and severity of airways obstruction. Prior reports of pneumothorax in CF patients from single centers have suggested that the mean age of occurrence was in the mid-teen years. (1) These reports were between the years 1968 and 1990. This analysis of patients in the 1990s demonstrates that pneumothorax generally occurs later in life with the mean age of occurrence in the early 20s. The median age of survival of CF patients has been increasing over time, (9) presumably due to more aggressive therapies. It is likely that these therapies have delayed the progression of pulmonary impairment, resulting in the older age of initial pneumothorax.

FE[V.sub.1] declines as CF patients get older, which is a consequence of chronic airways infection and inflammation, yet impairment of the FE[V.sub.1] remains an important risk factor for pneumothorax, even when controlling for age. This is similar to what has been reported as risk factors for spontaneous pneumothorax in patients with other diseases, in whom the most common associations are subpleural airspace enlargement, air-trapping, and smoking. (2,10) The use of high-resolution CT scanning of the chest at end-expiration has shown a lower lung density value in non-CF patients with spontaneous pneumothorax than in control subjects, which is consistent with the notion that air-trapping is part of the pathogenesis of this complication. (2) In addition, emphysema-like changes are commonly seen in non-CF patients with spontaneous pneumothorax either by CT scan or during video-assisted thoracoseopy. (11) There have been prior reports of air-trapping in CF patients with pneumothorax demonstrating a residual volume/total lung capacity ratio of 0.5612 and an average residual volume of 345% predicted. (13)

There is a known association between smoking and spontaneous pneumothorax. (10,14) Tobacco use is not reported in the registry database, so it is not possible to determine an association between smoking and pneumothorax in this population. However, studies (15-17) of smoking in CF patients have reported a prevalence of 8 to 21% of active CF smokers, so there is a potential for such an association.

Additional risk factors include the presence of P aeruginosa, B cepacia, and Aspergillus in the airways. The presence of specific pathogens may result in increased inflammation and airway secretions leading to obstruction of the distal airways with airtrapping. Aspergillus is commonly cultured from the sputum of CF patients, and is often thought to be a colonizer and not necessarily a pathogen. (18) The presence of Aspergillus in sputum cultures may indicate ABPA, an asthma-like complication that occurs commonly in patients with CF, (19) and this analysis found an increased association of ABPA with pneumothorax. The airways obstruction seen in ABPA may lead to air-trapping as a plausible pathogenesis for pneumothorax. However, most CF patients with Aspergillus present in their sputum cultures do not have ABPA. The association of Aspergillus with pneumothorax raises the possibility that the organism is a pathogen inducing airways inflammation leading to air trapping. Antibodies specific to Aspergillus fumigatus have been demonstrated in patients with CF, even in the absence of ABPA, and the authors suggested that Aspergillus may promote subclinical airways inflammation. (20) There are no data to suggest that the treatment of Aspergillus colonization in patients without ABPA reduces airways obstruction or the risk of pneumothorax.

Other risk factors that are associated with pneumothorax include Medicaid insurance, pancreatic insufficiency, massive hemoptysis, and the use of the inhaled medications dornase alfa and tobramyein. There is an inverse relationship between household income and outcomes in CF patients. Lower household income has been associated with an increased risk of death. (21) Medically indigent CF patients, as indicated by Medicaid insurance, have worse outcomes overall than do other CF patients. (22) This association has been attributed to multiple features of the medically indigent such as nutritional deficiencies, increased tobacco use, and a lower rate of adherence to therapies. (22) Trying to explain the association of pneumothorax and the other factors is challenging. Patients with pancreatic insufficiency generally have worse lung function, as do patients with massive hemoptysis.

Inhaled medications such as tobramycin and dornase alfa are used in patients with more severe lung disease. (230 Although they reduce lung infection (24) and promote clearance of airways secretions, (25) there is also the possibility that these inhaled medications may exacerbate airways disease. Some patients experience an acute drop in FE[V.sub.1] after inhaling nebulized tobramycin." (26,27) Such an effect could explain the association between inhaled medications and pneumothorax, although there have been no reported cases of pneumothorax in CF patients following nebulized therapy.

Pneumothorax is associated with a high rate of mortality and is a bad prognostic indicator for patients with CF. Thirty years ago, a scoring system developed as a prognostic tool for CF included pneumothorax because of a perceived association with severe disease and mortality. (28) A recent or recurrent pneumothorax warranted a score of 5 (ie, 5% of the maximum score) and a history of pneumothorax at any time received a score of 3. Some analyses of large databases, (29-31) including the CFF Patient Registry, have tried to identify clinical parameters that predict 2-year and 5-year survival rates, but none included pneumothorax as an important criterion. This may be because it was not found to be an independent predictor or because it was excluded from analysis as < 5% of patients were affected. (30) One model was found to be no better at predicting death within 2 years than was using an FE[V.sub.1] of < 30% predicted as the sole predictor. (31) It is evident by this analysis that a pneumothorax increases considerably the probability of dying within 2 years, especially for those with an FE[V.sub.1] of > 30% predicted.

Aside from mortality, there is considerable morbidity associated with pneumothorax. The treatment of a pneumothorax typically involves the placement of a thoracostomy tube that can cause considerable pain. Some patients will need more definitive therapy, such as pleurodesis. These therapies require hospitalization, and this analysis has demonstrated a considerable increase in the number of hospitalization days in the index year, which carries substantial morbidity and cost. The increase in outpatient visits, hospitalizations, and hospital days may be the direct result of the pneumothorax or an indirect result due to the worsening impairment of pulmonary function and infection. Indeed, there is a significant decrease in lung function measured in the year following the pneumothorax, more than would have been predicted based on the previous rate of decline.

There are some important limitations to this analysis. The most important limitation is that pneumothorax incidence is likely to be underestimated in this analysis. Other diseases that cause secondary spontaneous pneumothorax, such as COPD, have had an average incidence of recurrence of 85% reported, (32) with the majority of pneumothoraces occurring within the first few months after the first presentation. Although we would not be able to detect recurrent pneumothorax occurring in the same year, we might expect a higher rate of subsequent pneumothorax than we have reported here (approximately 20%). Also, patients could have experienced a pneumothorax prior to 1990 or entry into the database. There is no information in this database about the treatment of pneumothorax.

This study has several implications for therapy. First, CF patients with pneumothorax should be assessed for ABPA since specific therapy is available for this disorder. Second, it may be prudent to evaluate the effects of inhaled medications on airway reactivity in patients with more severe lung impairment. Finally, although pneumothorax increases the probability of death within 9. years, it may not play a role in the decision to refer for transplant those patients with an FE[V.sub.1] of > 30% predicted.

In conclusion, spontaneous pneumothorax is a common complication of patients with CF. It occurs in about 1 in 167 patients each year, and nearly 3.5% of all CF patients will experience a pneumothorax. It typically occurs in patients who are older and who have severe obstructive airways disease. Subsequent pneumothoraces are also common, and prior studies have reported treatment failures. (1) A specific study of early definitive therapy with recurrence prevention needs to be conducted in patients with this disease. There are multiple clinical factors associated with pneumothorax, including certain pathogens (especially P aeruginosa, B cepacia, and Aspergillus), pancreatic insufficiency, ABPA, and certain CF-specific inhaled therapies. Pneumothorax is a serious complication of CF, occurring in patients with advanced disease and resulting in a substantial increase in morbidity and mortality.

ACKNOWLEDGMENT: The authors are grateful to Dr. Preston Campbell for making the CFF Patient Registry data available, and to Dr. James Yankaskas for his editorial assistance.

REFERENCES

(1) Flume PA. Pneumothorax in cystic fibrosis. Chest 2003; 123:217-221

(2) Smit HJM. Spontaneous pneumothorax: pathogenesis and treatment related aspects. Amsterdam 1999

(3) Schramel FM, Postmus PE, Vanderschueren RG. Current aspects of spontaneous pneumothorax. Eur Respir J 1997; 10:1372-1379

(4) Spector ML, Stern RC. Pneumothorax in cystic fibrosis: a 26-year experience. Ann Thorac Surg 1989; 47:204-207

(5) Cystic Fibrosis Foundation. Patient registry 1999: annum data report. Bethesda, MD: Cystic Fibrosis Foundation, 2000

(6) FitzSimmons SC. The changing epidemiology of cystic fibrosis. J Pediatr 1993; 122:1-9

(7) Littell RC, Milliken GA, Stroup WW, et al. SAS system for mixed models. Cary, NC: SAS Institute 1996

(8) Corey M, Edwards L, Levison H, et al. Longitudinal analysis of pulmonary function decline in patients with cystic fibrosis. J Pediatr 1997; 131:809-814

(9) Cystic Fibrosis Foundation. Patient registry 2001: annual data report. Bethesda, MD: Cystic Fibrosis Foundation, 2002

(10) Bense L, Eklund G, Wiman LG. Smoking and the increased risk of contracting spontaneous pneumothorax. Chest 1987; 92:1009-1012

(11) Smit HJ, Deville WL, Schramel FM, et al. Atmospheric pressure changes and outdoor temperature changes in relation to spontaneous pneumothorax. Chest 1999; 116:676-681

(12) Stowe SM, Boat TF, Mendelsohn H, et al. Open thoracotomy for pneumothorax in cystic fibrosis. Am Rev Respir Dis 1975; 111:611-617

(13) McLaughlin FJ, Matthews WJ, Streider DJ, et al. Pneumothorax in cystic fibrosis: management and outcome. J Pediatr 1982; 100:863-869

(14) Jansveld CA, Dijkman JH. Primary spontaneous pneumothorax and smoking. BMJ 1975; 4:559-560

(15) Britto MT, Garrett JM, Dugliss MA, et al. Risky behavior in teens with cystic fibrosis or sickle cell disease: a multicenter study. Pediatrics 1998; 101:250-256

(16) Stern RC, Byard PJ, Tomashefski JF Jr, et al. Recreational use of psychoactive drugs by patients with cystic fibrosis. J Pediatr 1987; 111:293-299

(17) Verma A, Clough D, McKenna D, et al. Smoking and cystic fibrosis. J R Soc Med 2001; 40:29-34

(18) Nunley DR, Ohori P, Grgurich WF, et al. Pulmonary aspergillosis in cystic fibrosis lung transplant recipients. Chest 1998; 114:1321-1329

(19) Geller DE, Kaplowitz H, Light MJ, et al. Allergic bronchopulmonary aspergillosis in cystic fibrosis: reported prevalence, regional distribution, and patient characteristics; Scientific Advisory Group, Investigators, and Coordinators of the Epidemiologic Study of Cystic Fibrosis. Chest 1999; 116: 639-646

(20) Knutsen AP, Hutcheson PS, Mueller KR, et al. Serum immunoglobulins E and G anti-Aspergillus fumigatus antibody in patients with cystic fibrosis who have allergic bronchopulmonary aspergillosis. J Lab Clin Med 1990; 116:724-727

(21) O'Connor GT, Quinton HB, Kneeland T, et al. Median household income and mortality rate in cystic fibrosis. Pediatrics 2003; 111:e333-e339

(22) Schechter MS, Shelton BJ, Margolis PA, et al. The association of socioeconomic status with outcomes in cystic fibrosis patients in the United States. Am J Respir Crit Care Med 2001; 163:1331-1337

(23) Konstan MW, Butler SM, Schidlow DV, et al. Patterns of medical practice in cystic fibrosis: Part II. Use of therapies: Investigators and Coordinators of the Epidemiologic Study of Cystic Fibrosis. Pediatr Pulmonol 1999; 28:248-254

(24) Ramsey BW, Pepe MS, Quan JM, et al. Intermittent administration of inhaled tobramycin in patients with cystic fibrosis: Cystic Fibrosis Inhaled Tobramycin Study Group. N Engl J Med 1999; 340:23-30

(25) Ramsey BW, Astley SJ, Aitken ML, et al. Efficacy and safety of short-term administration of aerosolized recombinant human deoxyribonuclease in patients with cystic fibrosis. Am Rev Respir Dis 1993; 148:145-151

(26) Alothman GA, Alsaadi MM, Ho BL, et al. Evaluation of bronchial constriction in children with cystic fibrosis after inhaling two different preparations of tobramycin. Chest 2002; 122:930-934

(27) Nikolaizik WH, Trociewicz K, Ratjen F. Bronchial reactions to the inhalation of high-dose tobramycin in cystic fibrosis. Eur Respir J 2002; 20:122-126

(28) Taussig LM, Kattwinkel J, Friedewald WT, et al. A new prognostic score and clinical evaluation system for cystic fibrosis. J Pediatr 1973; 82:380-390

(29) Corey M, McLaughlin FJ, Williams M, et al. A comparison of survival, growth, and pulmonary function in patients with cystic fibrosis in Boston and Toronto. J Clin Epidemiol 1988; 41:583-591

(30) Liou TG, Adler FR, Fitzsimmons SC, et al. Predictive 5-year survivorship model of cystic fibrosis. Am J Epidemiol 2001; 153:345-352

(31) Mayer-Hamblett N, Rosenfeld M, Emerson J, et al. Developing cystic fibrosis lung transplant referral criteria using predictors of 2-year mortality. Am J Respir Crit Care Med 2002; 166:1550-1555

(32) Light RW, O'Hara VS, Moritz TE, et al. Intrapleural tetracycline for the prevention of recurrent spontaneous pneumothorax: results of a Department of Veterans Affairs cooperative study. JAMA 1990; 264:2224-2230

* From the Departments of Medicine (Drs. Flume and Strange) and Biometry and Epidemiology (Dr. Ye and Ms. Clark) and Pediatrics (Ms. Ebeling and Dr. Hulsey), Medical University of South Carolina, Charleston, SC.

This study was supported in part by a grant from the CFF.

Manuscript received July 28, 2004; revision accepted January 24, 2005.

Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (www.chestjournal. org/misc/reprints.shtml).

Correspondence to: Patrick A. Flume, MD, FCCP, Medical University of South Carolina, 96 Jonathan Lucas St, 812-CSB, Charleston, SC 2942,5; e-mail: flumepa@musc.edn

COPYRIGHT 2005 American College of Chest Physicians
COPYRIGHT 2005 Gale Group