American Journal of Respiratory and Critical Care Medicine

Airway inflammation is an important component of cystic fibrosis (CF) lung disease. To determine whether this begins early in the illness, before the onset of infection, we examined bronchoalveolar lavage (BAL) fluid from 46 newly diagnosed infants with CF under the age of 6 mo identified by a neonatal screening program. These infants were divided into three groups: 10 had not experienced respiratory symptoms or received antibiotics and pathogens were absent in their BAL fluid; 18 had clear evidence of lower respiratory viral or bacterial ( ⩾ 105 CFU/ml) infection; and the remaining 18 had either respiratory symptoms, taken antibiotics, or had < 105 CFU/ml of respiratory pathogens. Their BAL cytology, interleukin-8, and elastolytic activity were compared with those from 13 control subjects. In a longitudinal study to assess if inflammation develops or persists in the absence of infection, the results of 56 paired annual BAL specimens from 44 CF infants were grouped according to whether they showed absence, development, clearance, or persistence of infection. In newly diagnosed infants with CF, those without infection had BAL profiles comparable with control subjects while those with a lower respiratory infection had evidence of airway inflammation. In older children, the development and persistence of infection was accompanied by increased inflammatory markers, whereas these were decreased in the absence, or with the clearance, of infection. We conclude that airway inflammation follows respiratory infection and, in young children, improves when pathogens are eradicated from the airways.

Cystic fibrosis (CF) is caused by mutations to the CF transmembrane conductance regulator gene resulting in defective regulation of chloride transport by epithelial cells (1). Chronic lung disease is the most serious clinical expression of CF. The progressive lung injury associated with endobronchial infection, first from Staphylococcus aureus and then Pseudomonas aeruginosa eventually leads to death from respiratory failure.

Previous work has shown that even clinically stable patients with CF and mild lung disease have large numbers of bacteria within their lower airways, which are accompanied by a neutrophil influx and uninhibited elastase activity (2). Furthermore, following studies conducted in older children and adults with established lung disease, it is now believed that tissue damage resulting from airway inflammation has a major role in the pathogenesis of CF lung disease (3). For example, pulmonary secretions from CF patients have abundant neutrophils (4) and increased concentrations of interleukin-8 (IL-8), a pro-inflammatory α-chemokine (5). In addition, the proteolytic neutrophil enzymes, elastase, and cathepsin G, are present in sufficiently high concentrations to overwhelm the lung antiprotease defenses provided by α1-antitrypsin and secretory leukoprotease inhibitor (4, 6). Unopposed neutrophil enzyme activity interferes with local host immune mechanisms, impedes bacterial clearance and directly injures bronchial mucosal epithelium and its supporting tissues (3). Other inflammatory mediators, such as leukotrienes, C5a, tumor necrosis factor-α, and interleukin-1β, are also present in the lower respiratory secretions of patients with CF, but their precise role in the pathogenesis of CF lung disease is less certain (7, 8). However many questions remain, especially those relating to the onset of airway inflammation and whether in the young patient with CF such changes are reversible.

While gene replacement therapy may eventually become the treatment of choice for patients with CF (9), at present, reducing the lung damage associated with airway inflammation is likely to have the greatest impact upon disease progression. Of concern is that, although the lungs in infants with CF are histologically normal at birth (10), there is evidence that inflammatory changes begin early in life (11). Using bronchoalveolar lavage (BAL) to collect lower airway samples, we have shown that, by 3 mo of age, nearly 40% of infants identified by a neonatal CF screening program had a lower respiratory tract infection (12). Nearly one-third of the infected infants were asymptomatic and S. aureus was the predominant pathogen. As with the older patients (2), the presence of a large bacterial burden within the lower airways was accompanied by increased concentrations of Both IL-8 and inflammatory cells. Nevertheless, the role of infection in the establishment of CF lung disease has recently been challenged. Two limited studies in infants and children with CF have reported airway inflammation in the absence of infection (13, 14). It has since been suggested that the basic defect in CF itself may initiate or amplify the inflammatory response within the lung (15). If confirmed, such observations in young infants have implications for future gene therapy and the role of anti-inflammatory agents when managing infant CF lung disease.

To further investigate whether inflammation precedes infection, we examined BAL fluid collected from 10 newly diagnosed infants with CF identified prospectively from a neonatal screening program. All were younger than 6 mo of age, and none had experienced respiratory symptoms, received antibiotics, or had respiratory pathogens identified in their BAL fluid. Cytology, IL-8 concentrations, and free neutrophil elastase activity from their BAL specimens were compared with those from 13 disease control infants and, from the same newborn CF cohort, with 36 infants, 18 of whom had a lower airway infection.

To determine whether airway inflammation develops and persists in the absence of infection, we assessed cytology, IL-8 concentrations, and free neutrophil elastase activity in 56 paired annual BAL fluid specimens collected from 44 infants with CF who were part of a prospective statewide multiple birth cohort. Paired specimens were grouped according to whether they showed absence, development, eradication, or persistence of infection. BAL inflammatory data were compared for each of these categories.

Subjects and Controls

The state of Victoria, Australia (66,000 births per yr) uses a two-tiered newborn screening program for CF (16) based upon an estimation of the immunoreactive trypsin from a blood spot Guthrie card, followed by genetic analysis for the ΔF508 deletion and three exon 11 mutations: G542X, G551D, and R553X. A sweat chloride concentration ⩾ 60 Meq/L confirmed the diagnosis for heterozygotes. All patients are admitted to the Royal Children's Hospital, Melbourne, for parental education and baseline assessment. Subsequent management is by the Hospital's CF Clinic.

As previously reported (12, 17), infants born since 1990 were eligible for recruitment into a prospective, longitudinal study examining the early microbiology of CF lung disease. Subjects entered the study from February 1992 through November 1995, either within the first six months of life (newborn cohort) or at 12 or 24 mo of age if born in 1990 or 1991, respectively, giving a multiple cohort design. Each underwent an initial, and then an annual, BAL for 2 following yr to collect lower respiratory secretions. At bronchoscopy, and at their 3 monthly clinic appointments, subjects underwent a clinical evaluation, recording body weight expressed as z-scores (18), and the presence of any intervening respiratory symptoms or antibiotic therapy. Annual chest radiographs were given a Brasfield score by a blinded observer (19).

Otherwise well infants undergoing bronchoscopy for evaluation of congenital stridor formed a disease control group. These infants all came from the same population as the study subjects, and all had normal CF newborn screening results. Subjects were excluded from this group if there were symptoms or signs of respiratory infection or a history of antibiotic use within the previous 14 d.

Bronchoalveolar Lavage

Under halothane anesthesia and following topical administration of 4 mg/kg of 2% lignocaine hydrochloride to the vocal cords, a flexible fiberoptic bronchoscope (Olympus model BF 3C20; external diameter 3.6 mm, suction channel 1.2 mm; Olympus Corp. of America, Hyde Park, NY) was introduced into the lower airways through a laryngeal mask. The suction channel was not used until the bronchoscope tip was below the carina. The tip was wedged in the right middle lobe bronchus, and, to optimize sampling from endobronchial sites (20), a single, small volume lavage was performed by instilling 1 ml/kg (maximum 10 ml) of sterile nonbacteriostatic normal saline at room temperature through the bronchoscope over 3–5 s. The saline was immediately aspirated into a sterile suction set over 10–20 s using negative pressures of 100–150 mm Hg. The bronchoscope was next wedged into the lingula bronchus, and, using an identical technique, a further single aliquot lavage was performed. The BAL fluid from both lavages was then pooled and immediately transported on ice to the laboratory for processing and storage at −70° C.

Laboratory Procedures

Cytology. Total cell counts (21) were performed on 100 μl aliquots of pooled, uncentrifuged BAL fluid using a Neubauer hemocytometer counting chamber (Weber, Teddington, UK). The differential cell count was estimated after centrifugation of a 300 μl aliquot of BAL fluid at 72 g for 5 min in a Cytospin-2 cytocentrifuge (Shandon Southern Instruments, Sewickly, PA), Wright-Giemsa staining and counting 300 cells at magnification under oil ×1,000 (22).

Microbiology. The pooled BAL fluid was gently vortexed with sterile glass beads for 5 s and 500 μl was serially diluted from 10−1 and 10−5 in sterile phosphate buffered saline (pH 7.2). One hundred microliters of undiluted BAL fluid and 100 μl from each of the serial dilutions were then added by the spread plate method to six different selective media (horse blood, mannitol salt, MacConkey, chocolate bacitracin, cetrimide, and Pseudomonas cepacia agar) for quantitative bacterial colony counts yielding a lower detection limit of < 102 colony forming units (CFU) per ml of BAL fluid (17). Plates were incubated aerobically, and in 5% CO2, at 37° C and read at 24 and 48 h, before being discarded at 5 d. Standard microbiologic techniques identified respiratory bacterial pathogens (23). BAL and nasopharyngeal aspirate specimens were tested by immunofluorescence (24) and cultured for respiratory syncytial virus (RSV Direct IF; bioMerieux, Marcy l'Etoile, France), influenza virus (A and B), adenovirus and parainfluenza viruses 1-3 (Bartels Viral Respiratory Screening and Identification Kit; Baxter Diagnostic Inc., Deerfield, IL). The cells used for virus culture included primary monkey kidney (LLC-MK2), human epithelial (A549, HeLa), and fibroblast cell lines.

Biochemical assays. The BAL fluid was centrifuged at 500 g for 5 min at 4° C and the supernatant immediately stored in 100 μl aliquots, at −70° C, until the time of testing. IL-8 was measured by a double-sandwich, monoclonal antibody enzyme linked immunoabsorbent assay (Medgenix Diagnostics, Fleurus, Belgium), which measured free and receptor bound IL-8. The detection limit for the assay was 0.7 pg/ml.

After hydrolysis of the specific chromogenic substrate, MeO-Suc-Ala-Ala-Pro-Val-pNA (Sigma Chemical Co., St. Louis, MO), free neutrophil elastase activity was measured by a spectrophotometer at 405 nm (25). The absorbance was compared with a standard curve derived from purified human neutrophil elastase (Elastin Products Co., Inc., Owensville, MO) and the concentration calculated from a log-logit plot. A specific inhibitor of human neutrophil elastase, MeO-Suc-Ala-Ala-Pro-Val-CH2Cl (Sigma Chemical Co.), confirmed the specificity of the enzyme reaction. The sensitivity of the assay was 5.0 μg/ml.

The volume of recovered epithelial lining fluid was not estimated. Recent studies show that diffusion of urea and albumin into BAL fluid during bronchoscopy results in an overestimation of epithelial lining fluid volume (20). Consequently, quantitative culture results, cytologic values, IL-8 concentrations, and free neutrophil elastase activity are reported per ml of BAL fluid.

Definition of the CF Subject Groups

Newly diagnosed infants. Infants with newly diagnosed CF who underwent BAL during the first 6 mo of life were allocated into one of three groups: infants who had not experienced respiratory symptoms or received antibiotics and did not have any colonies of respiratory bacterial (Staphylococcus aureus, Haemophilus influenzae, Pseudomonas aeruginosa, Burkholderia cepacia, Stenotrophomonas maltophilia, Streptococcus pneumoniae, and Moraxella catarrhalis) or viral pathogens in their BAL were assigned to Group A; infants with CF with respiratory infection (defined by ⩾ 105 CFU of respiratory bacterial pathogens per ml of BAL fluid (17) or the detection of respiratory viruses) comprised Group B; and the remaining newly diagnosed CF infants became Group C (i.e., those with any combination of respiratory symptoms, antibiotic exposure, or low numbers of respiratory bacterial pathogens [< 105 CFU/ml] in their BAL fluid).

Longitudinal study. Paired, consecutive BAL samples collected on an annual (range 9–18 mo) basis were selected for longitudinal analysis. Specimens obtained outside this period were excluded. Paired BAL specimens were grouped according to the presence or absence of respiratory infection (17) in the following manner: paired BAL specimens in which there was no evidence for infection in either specimen were classified as NN; infection absent in the first BAL but present in the second became NP; infection in the first but absent from the second BAL were included in PN; and when both BAL specimens demonstrated infection these were assigned to the PP group.

Data Analysis

Cytologic data, IL-8, and free neutrophil elastase concentrations were logarithmically transformed and summarized as geometric means with 95% confidence intervals. When elastolytic activity fell below the assay detection threshold of 5.0 μg/ml, a value of 2.5 μg/ml was assigned to enable logarithmic transformations and pairwise comparisons. Two-group comparisons for continuous variables were conducted using t tests in the log scale. Proportions were compared by chi-square or Fisher's exact tests. Association between IL-8, neutrophils and free neutrophil elastase was assessed using Spearman correlation.

Longitudinal analysis was performed on all pairs of annual BAL samples where cytology, IL-8, and free neutrophil elastase data were available at both the beginning and end of the annual interval. Mean values of weight z-score and inflammatory markers, at the second BAL, and of their changes between the two time points, were obtained using analysis of covariance to adjust for value at the first BAL and the potentially confounding effect of age. For those outcomes analyzed in the log scale, this gave estimates of geometric mean values at the second BAL and geometric mean ratios, or fold changes, between the first and second BALs. A fold change < 1.0 indicated a reduction in value between the first and second BAL; whereas a fold change > 1.0 indicated an increase. Adjustment for level at the first BAL produces estimated means and fold changes that can be compared between groups as if each group had the same initial level, equal to the overall mean at the first BAL. Confidence intervals for these means were obtained using “robust” standard errors (Huber method) that allow for correlations potentially induced by the fact that some individuals contributed more than one pair of BAL specimens for analysis. All analyses were performed using the Stata statistical software package (26).

This study was approved by the Human Ethics Committee of the Royal Children's Hospital, and written informed consent was obtained from the parents of each child before bronchoscopy.

Newly Diagnosed Infants (Cross-sectional Study)

During the recruitment period, 59 of 75 (79%) newly diagnosed infants with CF underwent bronchoscopy within the first 6 mo of life. More than 5 ml of BAL fluid were obtained from 46 infants, which was sufficient to complete all tests on the BAL fluid, and these formed the study group. The numbers assigned to groups A, B, and C were 10, 18, and 18, respectively. Their clinical characteristics and those of 13 disease control subjects, at the time of BAL are shown in Table 1. The groups were comparable for gender and weight. However, by including only CF cases less than 6 mo of age, these infants were significantly younger than control infants whose ages ranged from 2–33 mo.


Group A (n = 10)Group B (n = 18)Group C (n = 18)Control Subjects (n = 13)
Mean age, mo (SD)2.4 (1.5)2.6 (1.1)2.3 (1.0)11.7 (10.2)
Sex (% male)40725069
Mean weight z-score−0.66−0.90−0.43−0.20
 (95% confidence interval)(−1.65, 0.34)(−1.49, −0.30)(−1.11, 0.24)(−0.85, 0.45)
CF Genotype
 Homozygous ΔF508 8 (80%)11 (61%)13 (72%)ND
 Heterozygous ΔF508 1 (10%) 6 (33%) 3 (17%)ND
 No copies ΔF508 1 (10%) 1 (6%) 2 (11%)ND
Brasfield chest X-ray score* 24.321.823.8ND
 (95% confidence interval)(21.5, 25.0)(19.6, 23.9)(22.9, 24.6)

A = CF infants without a history of respiratory symptoms, antibiotic use or respiratory pathogens (bacterial or viral) in their BAL; B = CF infants with respiratory infection (see Methods for definition of infection [17] ); C = CF infants without proven respiratory infection.

*Brasfield scores (18) were reported in 23 of 46 CF infants: groups A (n = 3), B (n = 12) and C (n = 8), but were not done (ND) in controls.

Forty-three of 46 (93%) cases were identified by neonatal screening. A further two were missed by screening, presenting as failure to thrive and another was diagnosed by sweat test because of an older sibling with CF. Almost 90% had one or more copies of the ΔF508 deletion, and these were evenly distributed among Groups A, B, and C. Meconium ileus occurred in 10 (17%) infants with CF where the diagnosis was made by sweat test and before the results of neonatal screening tests became available.

Almost 40% (18 of 46) of newly diagnosed infants with CF within the study population had a lower respiratory infection. These subjects formed Group B, and at the time of BAL 12 of 18 (67%) had respiratory symptoms severe enough to warrant hospitalization. Five (28%) were already receiving oral antibiotics (amoxycillin-clavulanate, cefaclor, or flucloxacillin). In Group C, five subjects had also received oral antibiotics (amoxycillin-clavulanate or cephalexin) for respiratory symptoms. Although three of these children were hospitalized and at BAL were being treated with parenteral flucloxacillin, noninfected subjects in Group C were less likely to be admitted to hospital for respiratory illness than those in Group B (RR = 0.25; ci 0.08, 0.74). None of the subjects had received oral or inhaled corticosteroids.

Seventeen bacterial and seven viral infections were identified in the 18 Group B subjects. S. aureus was present in 12, including three with mixed S. aureus and H. influenzae infections; H. influenzae and M. catarrhalis were detected in one subject each, respectively. RSV was present in two infants, parainfluenza virus type 3 in a third, a mixed adenovirus, and rhinovirus infection was discovered in another and two mixed rhinovirus and S. aureus infections were also found.

The BAL inflammatory data from the newly diagnosed infants with CF and control infants are shown in Figure 1, and the analysis is summarized in Table 2. BAL neutrophil and IL-8 concentrations (r = 0.61), neutrophil and free neutrophil elastase concentrations (r = 0.55), and IL-8, and free neutrophil elastase concentrations (r = 0.56) were significantly correlated with one another (p < 0.001). The percentage of BAL return from subjects in Group A was significantly greater than obtained from control infants (p < 0.01). In these younger infants, a single, small volume BAL returns a greater proportion of alveolar fluid. However, the BAL profiles between Group A and control infants were otherwise comparable. In contrast, subjects from Group B, the infants with CF and respiratory infection, had significantly greater total cell counts (p < 0.01), neutrophil counts and their percentages, IL-8 concentrations, and free neutrophil elastase activity (p < 0.001) than those found in the BAL fluid from subjects in Group A.


Group A*(n = 10)Group B (n = 18)Group C (n = 18)Control Subjects (n = 13)
% BAL fluid return (SD)a 51 (10)41 (14)46 (14)36 (14)
Cell counts × 103/ml
 Total cell countb134 454185143
(75, 240)(227, 909)(104, 329)(80, 253)
 Macrophages 73 114 94101
(32, 167)(66, 199)(44, 201)(61, 168)
 Neutrophilsc 15 153 25  7
(8, 29)(52, 449)(10, 63)(2, 22)
 Neutrophil %c 19  51 24  9
(2, 35)(34, 68)(13, 34)(3, 16)
 Lymphocytes  4  10  6  7
(2, 9)(5, 21)(3, 11)(3, 13)
 Ciliated cells 11   7 11 11
(4, 32)(3, 19)(5, 23)(5, 25)
 Squamous cells    1.1    1.5    1.5    1.5
(0.85, 1.5)(1.0, 2.3)(1.0, 2.2)(0.9, 2.5)
IL-8 (pg/ml)c 241,112 83 32
(9, 63)(582, 2,126)(26, 257)(12, 84)
% with neutrophil c  0  78 39 23
 elastase activity(0, 31)(52, 94)(17, 64)(5, 54)

*Groups as defined for Table 1.

a = a significant difference between group A and controls; p < 0.05. b = a significant difference between groups A and B; p < 0.01. c = a significant difference between groups A and B; p < 0.001.

The distribution of data points for the BAL outcome measures shown in Figure 1 for Groups A, C, and control infants were similar. Two of the three high outliers for IL-8 in Group C were from subjects receiving parenteral antibiotics for respiratory symptoms. These three BAL specimens also had elevated neutrophil counts and free neutrophil elastase activity and two grew 104 CFU/ml of S. aureus and H. influenzae, respectively. Another three symptomatic infants within Group C with elevated inflammatory indices had severe gastroesophogeal reflux, and aspiration lung disease was suspected. In contrast, the two control infants with the highest IL-8 readings did not have elevated total cell counts, neutrophil, or free neutrophil elastase concentrations, but one developed mild coryzal symptoms and the other diarrhea within 48 h of bronchoscopy.

Longitudinal Study

Forty-four of 92 (48%) subjects enrolled in an infant cohort study of CF lung disease had undergone at least two elective bronchoscopies, 9–18 mo apart, and had sufficient BAL fluid collected to complete all laboratory studies. These 44 children contributed 56 paired BAL specimens (12 children provided two sets of paired specimens) for analysis. None were receiving oral or inhaled corticosteroids. The characteristics of the subjects at the time of their second BAL are shown in Table 3.


NN (n = 26)NP (n = 9)PN (n = 16)PP (n = 5)All (n = 56)
Mean age (SD)24 (10)27 (12)23 (10)28 (12)25 (10)
Sex (% male)* 43445050
% BAL return (SD)40 (12)33 (9)38 (11)34 (9)

Definition of abbreviations: NN = no infection in either BAL specimens; NP = no infection in the first BAL, but present in the second; PN = infection present in the first BAL, but absent in the second; PP = infection present in both BAL specimens; see Methods for details of definition of infection (17).

*Sex status was assigned to 50 BAL pairs, the remaining six BAL pairs were from subjects with CF who contributed two BAL pairs but remained either within the NN (n = 5) or PP (n = 1) classification.

As a marker of airway inflammation, the data for free neutrophil elastase activity in the first and second BAL specimens for all four groups are shown in Figure 2 and changes in BAL inflammatory mediators between the two time points are presented in Table 4. The estimates incorporate adjustment by analysis of covariance to common mean values at the first BAL, but adjustment by age was found to have no effect and was omitted.


NN*(n = 26)NP (n = 9)PN (n = 16)PP (n = 5)
Fold change in neutrophil count 0.362.20.4128
(0.22, 0.58)(0.72, 6.7)(0.17, 0.97)(7, 114)
Change in neutrophil %−1717−1745
(−24, −10)(−2, 37)(−25, −9)(33, 57)
Fold change in IL-80.812.30.634.1
(0.34, 1.9)(1.06, 5.2)(0.26, 1.6)(1.7, 9.8)
Fold change in NE activity0.742.01.04.6
(0.58, 0.96)(1.2, 3.2)(0.72, 1.6)(2.9, 7.4)

*Groups as defined in Table 3.

Geometric means (arithmetic mean for neutrophil %) and confidence intervals adjusted to a common mean at first BAL (neutrophil count: 65 × 103/ml, neutrophil percentage: 36%, IL-8: 300 pg/ml and free neutrophil elastase activity: 6.2 μg/ml).

Twelve of the 26 (46%) paired BAL specimens in the NN group were from subjects with respiratory symptoms at their second bronchoscopy, and 10 were taking oral antibiotics. Nevertheless, this group had the lowest neutrophil and free neutrophil elastase concentrations which, with IL-8 concentrations, decreased over time. Three pairs of BAL specimens had both increasing neutrophil counts and free neutrophil elastase concentrations between the two time points. Two were from subjects receiving antibiotics for symptoms at the time of their second BAL, which grew 104 CFU/ml of S. aureus and 103 CFU/ml of P. aeruginosa, respectively.

Six of nine paired BAL specimens in the NP group were from symptomatic subjects, and five were taking oral antibiotics at the time of their BAL. Eight bacterial and two viral respiratory infections were present in the second BAL specimens. S. aureus was found in four including one with mixed S. aureus and H. influenzae, two had H. influenzae alone, one had P. aeruginosa, and RSV was detected in two BAL samples. Compared with their first BAL, this group had increased neutrophil counts and percentages, and had significant gains in IL-8 and free neutrophil elastase levels in their second BAL specimen. Two paired BAL samples from symptomatic subjects taking antibiotics at their second BAL grew ⩾ 105 CFU/ ml of S. aureus and had elevated IL-8 but did not show increased neutrophil or free neutrophil elastase concentrations.

Respiratory symptoms were reported at the time of the second BAL for nine of 16 (56%) paired BAL specimens from the PN group. Seven of these nine subjects were also receiving antibiotics. Previously, 21 bacterial and viral respiratory infections had been detected in the first of the 16 paired BAL specimens. There were nine with S. aureus, which included two with S. aureus and P. aeruginosa, two with S. aureus and H. influenzae and one with S. aureus and M. catarrhalis; H. influenzae was present in a further two BAL specimens, S. maltophilia was detected in one other, three had RSV, and one had a parainfluenza type 3 infection. The second BAL specimens showed significantly fewer neutrophils, decreased IL-8 and reduced free neutrophil elastase activity compared to values from the first BAL. Two BAL pairs had neutrophil and free neutrophil elastase concentrations which had not fallen over time. These were from subjects with respiratory symptoms at the time of the second BAL and both grew 104 CFU/ml of S. aureus from cultures of their BAL fluid.

Within the PP group, three of five paired specimens were from subjects taking antibiotics for respiratory symptoms at the time of their second BAL. Two of the first paired BAL specimens grew P. aeruginosa while S. aureus, S. maltophilia, and RSV were present in the remaining three, respectively. All of the second five BAL samples grew P. aeruginosa. These specimens yielded the greatest numbers of neutrophils and highest concentrations of IL-8 and free neutrophil elastase, which increased significantly between the two time points.

This study reaffirms the close relationship between large numbers of pathogenic organisms within the lower respiratory tract of CF infants and young children and an influx of inflammatory cells and mediators. For newly diagnosed CF infants identified by neonatal screening, the absence of infection in their lower airways is associated with a BAL profile of inflammatory cells, IL-8, and free neutrophil elastase activity similar to that of control subjects. Furthermore, if infection is eradicated, airway inflammation may reverse. This suggests that the inflammatory component of CF lung disease does not originate from a basic defect in airway epithelial cells. Instead, it is proposed that infection plays a role in initiating and maintaining airway inflammation.

Consistent with earlier reports, our data suggest that airway inflammation can occur in the first months of life in association with infection and is characterized by neutrophil infiltration, elevated IL-8 concentrations, and free neutrophil elastase activity (2, 12, 27). The predominant pathogens in infants are S. aureus and respiratory viruses. Lower airway inflammation in these young patients usually follows infection and, while these patients are often sicker than other CF subjects, some may be asymptomatic (2).

In contrast, two recent studies in young patients with CF have reported inflammation in the absence of infection. One study recruited 16 infants with CF from a screened neonatal population and found, during the first year of life, seven subjects without evidence for infection but who had increased neutrophils and IL-8 concentrations compared with eleven slightly older control subjects suffering from a variety of chronic respiratory disorders (13). The other study was in 14 older infants and children with stable lung disease who underwent 31 bronchoscopies over a 12-mo interval (14). Ten of these subjects had elevated neutrophil counts and increased cytokine and free neutrophil elastase activity, but pathogenic bacteria were only cultured from four of the children. Both reports concluded that airway inflammation occurs early in the course of CF lung disease and may antedate infection.

The current study suggests, however, that airway inflammation in young patients with CF is likely to be a secondary response to infection or some other airway insult such as aspiration. Infants with a lower respiratory tract infection had greater evidence of airway inflammation than either the noninfected subjects comprising Groups A and C or the control subjects. Furthermore, these latter three groups had similar measurements of these inflammatory markers. Those subjects in Group C with the highest neutrophil counts and IL-8 or free neutrophil elastase concentrations were either hospitalized and treated for suspected infection before BAL or were being treated for suspected aspiration.

It is possible that, in some cases, antibiotics treatment results in a rapid reduction of bacterial numbers within the CF lower airway, but with slower resolution of the accompanying inflammatory response. This appeared true for two subjects within Group C who had elevated neutrophil and IL-8 values in their BAL fluid while receiving parenteral antistaphylococcal antibiotics. Another possible cause of lower airway inflammation in the absence of infection is pulmonary aspiration from gastroesophageal reflux. We observed numerous fat-laden alveolar macrophages in BAL cytospin preparations from a further two Group C subjects with abnormal 24 pH tracings, increased BAL inflammatory indices and no growth of respiratory pathogens on culture. Increased inflammatory markers were also present in some controls. These children in the 2-wk period before their BAL, were reported to be asymptomatic other than for their persistent stridor, had not taken antibiotics and were judged fit to undergo anesthesia. The BAL findings in these older infants may therefore represent residual inflammation from an earlier respiratory infection.

Further evidence against the CF genetic defect being directly responsible for initial airway inflammation is from the subgroup of CF infants in Group A. These infants had no history of respiratory symptoms, exposure to antibiotics, and respiratory pathogens in their BAL specimens. This group is more tightly defined than either Group C or the controls, and did not show evidence of airway inflammation. The clustering of neutrophil counts in the region of 10 × 103 per ml of BAL fluid and a mean neutrophil percentage of almost 20% in this group is likely to be both a function of age and the technique of BAL adopted rather than an intrinsic defect predisposing to neutrophil infiltration into the lungs. The percentage of neutrophils, for example, in healthy non-CF children is approximately twofold greater in BAL fluid from infants than children older than 12 mo of age and a similar increase again in neutrophil percentage is observed if only the first aliquot is analyzed (20, 28, 29). Therefore, by using older children as control subjects, we could have positively biased differences in BAL fluid neutrophil counts between infants with CF and control subjects. That this was not observed in infants from both Groups A and C, and control subjects and that other inflammatory indices in these three groups were also comparable further strengthens the findings of this study. Finally, the longitudinal study showed increased inflammatory cells and mediators in the BAL fluid pairs from patients with new or persistent infection, and decreased inflammatory markers in patients without lower airway infection. This gives additional support to the close association between infection and inflammation in early CF lung disease.

There may be methodological reasons to explain the apparently conflicting results between the present study and the two earlier reports of airway inflammation in young children (13, 14). Unlike the present study which pooled BAL fluid samples from both lungs the previous studies only sampled BAL fluid from a single lobe. However, when BAL fluid from young children with CF is collected separately from each lung, we have found infection may be localized to one lung but accompanied by a generalized inflammatory response (30). A similar observation has been made in a rat lung model where local installation of P. aeruginosa resulted in an inflammatory response in the contralateral lung (31). BAL studies in infants with CF, which have sampled only from a single lobe, may have detected an inflammatory response but missed the inciting infection. Furthermore, only limited clinical information was provided in these studies of infant CF airway inflammation. It is possible that the BAL findings in these young children may represent a resolving inflammatory response to prior aspiration or infection, particularly as some were receiving antibiotics at the time of their BAL. The current study's demonstration of airway inflammation in asymptomatic control subjects and residual inflammation in subjects with CF, some of whom had lower respiratory symptoms, following eradication of earlier infection supports such an explanation.

The finding of significant correlations between levels of IL-8, neutrophils, free neutrophil elastase activity, and infection are consistent with observations in older patients with CF. This reflects the complex in-vivo interrelationships between bronchial epithelial cells, respiratory pathogens, cytokines, and the immune system (32). Respiratory viruses, such as RSV, the bacterial pathogens, S. aureus and P. aeruginosa, and even neutrophil elastase itself, can stimulate the synthesis of IL-8 by macrophages and bronchial epithelial cells. This cytokine is an important chemoattractant in the CF airways resulting in neutrophil infiltration and upregulation in their function. A vicious cycle of infection, inflammation, and airway damage is then established. This was seen in the longitudinal component of our study where young children with CF and persistent infection had marked airway inflammation. To avoid potential sampling bias, our study did not include data from children at times of acute exacerbations of their lung disease due to severe P. aeruginosa infection. However, it is worth noting that five of the 92 birth cohort children died from CF lung disease within the first 5 yr of life, including two before the age of 12 mo. Their deaths were directly attributable to persistent P. aeruginosa infection, and severe airway inflammation had been found in their previous BAL fluid analyses.

In contrast, inflammatory markers are decreased in BAL specimens which months earlier had shown evidence of infection. In some subjects, aggressive antimicrobial therapy eradicated P. aeruginosa from the airways for more than a year and this was accompanied by reduced airway inflammation. Reductions in some indices of inflammation have been observed in the BAL fluid and sputum from older patients with CF, but only immediately following acute treatment of their lung disease (5, 33). However, in adults pulmonary deterioration persists when, on rare occasions, P. aeruginosa is eradicated from the lower airways (34). Thus while infections in CF are required to initiate inflammation, with time this response is exaggerated and self-sustaining (35).

Many subjects in the longitudinal study did not fulfill the diagnostic criteria for infection during their BAL collections. Nevertheless, some of these children appeared to have elevated inflammatory markers. Many who exhibited respiratory symptoms were receiving antibiotics and had bacterial numbers just below the diagnostic threshold, suggesting at least some of these children had a partially treated or resolving infection. Others did not show any clinical features but had one or more elevated inflammatory markers. This underlines the difficulty in older infants and children of determining the cause of inflammation in the absence of current infection. While some subjects with CF may have been demonstrating low grade but persistent inflammation, as seen in asymptomatic non-CF control subjects signs of airway inflammation may persist for a variable period following presumed earlier infection.

This study shows that in infants with CF, lower respiratory infection occurs at an early age and is associated with the development of airway inflammation. In young patients with CF lower respiratory viral infections are self-limited and bacterial infections, including some caused by P. aeruginosa, are amenable to antibiotic therapy resulting in clearance of pathogens and reduction in inflammation. Understanding the pathogenesis of early CF lung disease, including the role of infection, will allow development of effective interventions before chronic infection is established and airway inflammation becomes irreversible (34).

1. Riordan J. R., Rommens J. M., Kerem B.-S., Alon M., Rozmahel R., Grzelczak Z., Zielenski J., Lok S., Plavsic N., Chou J.-L., Drumm M. L., Iannuzzi M. C., Collins F. S., Tsui L.-C.Identification of the cystic fibrosis gene: cloning and characterisation of complementary DNA. Science245198910661073
2. Konstan M. W., Hilliard K. A., Norvell T. M., Berger M.Bronchoalveolar lavage findings in cystic fibrosis patients with stable, clinically mild lung disease suggest ongoing infection and inflammation. Am. J. Respir. Crit. Care Med1501994448454
3. Stockley, R. A. 1995. Role of inflammation in respiratory tract infections. Am. J. Med. 99(Suppl. 6B):S8–S13.
4. Suter S., Schaad U. B., Tegner H., Ohlsson K., Desgrandchamps D., Waldvogel F. A.Levels of free granulocyte elastase in bronchial secretions from patients with cystic fibrosis: effect of antimicrobial therapy against Pseudomonas aeruginosa. J. Infect. Dis1531986902909
5. Dean T. P., Dai Y., Shute J. K., Church M. K., Warner J. O.Interleukin-8 concentrations are elevated in bronchoalveolar lavage, sputum and sera of children with cystic fibrosis. Pediatr. Res341993159161
6. O'Connor C. M., Gaffney K., Keane J., Southey A., Byrne N., O'Mahoney S., Fitzgerald M. X.α1-proteinase inhibitor, elastase activity, and lung disease severity in cystic fibrosis. Am. Rev. Respir. Dis148199316651670
7. Konstan M. W., Walenga R. W., Hilliard K. A., Hilliard J. B.Leukotriene B4 markedly elevated in the epithelial lining fluid of patients with cystic fibrosis. Am. Rev. Respir. Dis1481993896901
8. Wilmott R. W., Kassab J. T., Kilian P. L., Benjamin W. R., Douglas S. R., Wood R. E.Increased levels of interleukin-1 in bronchoalveolar washings from children with bacterial pulmonary infections. Am. Rev. Respir. Dis1421990365368
9. Johnson L. G.Gene therapy for cystic fibrosis. Chest107(Suppl.)199577S83S
10. Esterley J. R., Oppenheimer E. H.Cystic fibrosis of the pancreas: structural changes in peripheral airways. Thorax231968670675
11. Chow C. W., Landau L. I., Taussig L. M.Bronchial mucous glands in the newborn with cystic fibrosis. Eur. J. Paediatr1391982240243
12. Armstrong D. S., Grimwood K., Carzino R., Carlin J. B., Olinsky A., Phelan P. D.Lower respiratory infection and inflammation in infants with newly diagnosed cystic fibrosis. B.M.J310199515711572
13. Khan T. Z., Wagener J. S., Bost T., Martinez J., Accurso F. J., Riches D. W. H.Early pulmonary inflammation in infants with cystic fibrosis. Am. J. Respir. Crit. Care Med151199510751082
14. Balough K., McCubbin M., Weinberger M., Smits W., Ahrens R., Fick R.The relationship between infection and inflammation in the early stages of lung disease from cystic fibrosis. Pediatr. Pulmonol2019956370
15. Cantin A.Cystic fibrosis lung inflammation: early, sustained, and severe. Am. J. Respir. Crit. Care Med1511995939941
16. Balnaves M. E., Bonacquisto L., Francis I., Glazner J., Forrest S.The impact of newborn screening on cystic fibrosis testing in Victoria, Australia. J. Med. Genet321995537542
17. Armstrong D. S., Grimwood K., Carlin J. B., Carzino R., Olinsky A., Phelan P. D.Bronchoalveolar lavage or oropharyngeal cultures to identify lower respiratory pathogens in infants with cystic fibrosis. Pediatr. Pulmonol211996267275
18. Dean, A. G., J. A. Dean, D. Coulombier, K. A. Brendel, D. C. Smith, A. H. Burton, R. C. Dicker, K. Sullivan, R. F. Fagan, and D. G. Arner. 1994. Version 6: a word processing, database, and statistics program for epidemiology on microcomputers. Centers for Disease Control and Prevention, Atlanta, GA.
19. Brasfield D., Hicks G., Soong S.-J., Tiller P. E.The chest roentgenogram in cystic fibrosis: a new scoring system. Pediatrics6319792429
20. Walters E. H., Gardiner P. V.Bronchoalveolar lavage as a research tool. Thorax461991613618
21. Riedler J., Grigg J., Stone C., Tauro G., Robertson C. F.Bronchoalveolar lavage cellularity in healthy children. Am. J. Respir. Crit. Care Med1521995163168
22. Saltini C., Hance A. J., Ferrans V. J., Basset F., Bitterman P. B., Crystal R. G.Accurate quantitation of cells recovered by bronchoalveolar lavage. Am. Rev. Respir. Dis1301984650658
23. Balows, A., W. J. Hausler, Jr., K. L. Herrmann, H. D. Isenberg, and H. J. Shadowy, editors. 1991. In Manual of Clinical Microbiology, 5th ed. American Society for Microbiology, Washington DC. 209–572.
24. Uren E., Elsum R., Jack I.A comparative study of the diagnosis of respiratory virus infections by immunofluorescence and virus isolation in children. Aust. Paediatr. J131977282286
25. Bieth J., Spiess B., Wermuth C. G.The synthesis and analytical use of a highly sensitive and convenient substrate of elastase. Biochem. Med111974350357
26. StataCorp. 1995. Stata Statistical Software: Release 4.0. Stata Corporation, College Station, TX.
27. Birrer P., Mcelvaney N. G., Rudeberg A., Wirz C., Sommer, Liechti-Gallati S., Kraemer R., Hubbard R., Crystal R. G.Protease-antiprotease imbalance in the lungs of children with cystic fibrosis. Am. J. Respir. Crit. Care Med1501994207213
28. Ratjen F., Bredendiek M., Brendel M., Meltzert J., Costabel U.Differential cytology of bronchoalveolar lavage fluid in normal children. Eur. Respir. J.7199418651870
29. Midulla F., Villani A., Merollo R., Bjermer L., Sandstrom J., Ronchetti R.Bronchoalveolar lavage studies in children without parenchymal lung disease: cellular constituents and protein levels. Pediatr. Pulmonol201995112118
30. Gutierrez, J. P., K. Grimwood, R. Carzino, A. Olinsky, J. Carlin, and C. Robertson. 1996. Interlung differences on bronchoalveolar lavage fluid in cystic fibrosis infants and young children (abstract). Am. J. Respir. Crit. Care Med. 153(4, Part 2):A777.
31. Terashima T., Matsubara H., Nakamura M., Sakamaki F., Waki Y., Soejima K., Tasaka S., Nakamura H., Sayama K., Ishizaka A., Kanazawa M.Local Pseudomonas installation induces contralateral lung injury and plasma cytokines. Am. J. Respir. Crit. Care Med153199616001605
32. Levine S. J.Bronchial epithelial cell-cytokine interactions in airway inflammation. J. Invest. Med431995241249
33. Richman-Eisenstat J. B. Y., Jorens P. G., Hérbert C. A., Ueki I., Nadel J. A.Interleukin-8: an important chemoattractant in sputum of patients with chronic inflammatory airway diseases. Am. J. Physiol2641993L413L418
34. Sharma G. D., Tosi M. F., Stern R. C., Davis P. B.Progression of pulmonary disease after disappearance of pseudomonas in cystic fibrosis. Am. J. Respir. Crit. Care Med1521995169173
35. Bonfield T. L., Panuska J. R., Konstan M. W., Hilliard K. A., Hilliard J. B., Ghnaim H., Berger M.Inflammatory cytokines in cystic fibrosis lungs. Am. J. Respir. Crit. Care Med152199521112118
Correspondence and requests for reprints should be addressed to Dr. David Armstrong, Department of Thoracic Medicine, Royal Children's Hospital, Parkville, Victoria 3051, Australia. E-mail:

David Armstrong received the Smorgon Clinical Research fellowship and the study was supported by a grant from the Royal Children's Hospital Research Foundation.


No related items
American Journal of Respiratory and Critical Care Medicine

Click to see any corrections or updates and to confirm this is the authentic version of record