Myeloperoxidase and eosinophil cationic protein in serum and sputum during antibiotic treatment in cystic fibrosis patients with Pseudomonas aeruginosa infection

I order to study the time-course of myeloperoxidase (MPO) and eosinophil cationic protein (ECP) as parameters for monitoring inflammation in cystic fibrosis (CF), we investigated ten patients during both a 14-day intravenous antibiotic treatment and a corresponding self control. Modified Shwachman-Kulczycki score improved significantly (p < 0.008), C-reactive protein (CRP) levels decreased significantly (p < 0.05) during antibiotic treatment, while in the control phase there were no significant differences. Lung function parameters did not change significantly during antibiotic treatment or control phase. Serum MPO concentration (p < 0.006) and peripheral blood neutrophil granulocyte counts (p < 0.04) decreased significantly during antibiotic treatment, but not during the control phase. Sentm ECP concentration showed a tendency to decrease during antibiotic treatment, but this failed to reach significance. In general, sputum concentrations of MPO and ECP Were 500- to 1000-fold higher than in serum. However, neither MPO nor ECP in sputum showed a significan variation over time during antibiotic treatment or control phase. From our data we conclude that: (1) measurements of MPO, neutrophils and CRP in peripheral blood do correlate with clinical parameters such as the modified Shwachman-Kulczycki score; (2) neutrophils and MPO seem to reflect inflammatory changes induced by antibiotic treatment; (3) eosinophils may play a role in CF by an enhanced ‘releasability’ and (4) Sputum measurements of mediators of inflammation cannot be recommended.

I order to study the time-course of myeloperoxidase (MPO) and eosinophil cationic protein (ECP) as parameters for monitoring inflammation in cystic fibrosis (CF), we investigated ten patients during both a 14-day intravenous antibiotic treatment and a corresponding self control. Modified Shwachman-Kulczycki score improved significantly (p < 0.008), C-reactive protein (CRP) levels decreased significantly (p < 0.05) during antibiotic treatment, while in the control phase there were no significant differences. Lung function parameters did not change significantly during antibiotic treatment or control phase.
Serum MPO concentration (p < 0.006) and peripheral blood neutrophil granulocyte counts (p < 0.04) decreased significantly during antibiotic treatment, but not during the control phase. Sentm ECP concentration showed a tendency to decrease during antibiotic treatment, but this failed to reach significance. In general, sputum concentrations of MPO and ECP Were 500-to 1000-fold higher than in serum. However, neither MPO nor ECP in sputum showed a significan variation over time during antibiotic treatment or control phase. From our data we conclude that: (1) measurements of MPO, neutrophils and CRP in peripheral blood do correlate with clinical parameters such as the modified Shwachman-Kulczycki score; (2) neutrophils and MPO seem to reflect inflammatory changes induced by antibiotic treatment; (3) eosinophils may play a role in CF by an enhanced 'releasability' and (4) Sputum measurements of mediators of inflammation cannot be recommended.

Introduction
Patients with cystic fibrosis (CF) suffer from chronic Pseudomonas aeruginosa infection and a dramatic neutrophil recruitment into the lungs. This results in a proteinase imbalance and a subsequent progressive proteolytic destruction of lung connective tissue. -Deterioration of lung function and parameters of inflammation (e.g., Creactive protein CRP) parallels the time-course of progressive host inflammatory responses to chronic Pseudomonas aeruginosa infection. 4 Myeloperoxidase (MPO), as a neutrophil granulocyte-specifi inflammatory mediator, could be shown to have cytotoxic properties to the lung epithelium in vitro. 5 Other neutrophil granule products correlate with lung function parameters and CRP in acute exacerbations in CF.
Eosinophils and their granule proteins play a major role for chronic inflammation in bronchial asthma, but there is controversy about the role 282 Mediators of Inflammation Vol 4 1995 of the eosinophils in CF. While an Austdan working group saw evidence for an eosinophilic activation in CF, 7 other authors did not find elevated numbers of eosinophils in post-mortem CF lungs.
Because there is little information concerning the role of the inflammatory mediators MPO and eosinophil cationic protein (ECP) as parameters for monitoring inflammation in CF patients, we studied these mediators in serum and sputum during a 14-day intravenous antibiotic treatment and a 14-day in-patient control phase and compared them with clinical parameters.

Patients and Methods
Patients: Ten patients (three males, seven females) aged 14 to 44 years (mean age 24 years) with cystic fibrosis were enrolled in the study. Patient data are shown in Table 1 Study design: Patients were admitted to hospital in a cross-over design both for 14 days of intravenous antibiotic treatment against Pseudomonas aeruginosa 9 and (4 to 6 weeks before or after) for the same period without antibiotic treatment.
Therefore, each patient seeeed as their own control. Time of admittance was chosen when patients were free of acute exacerbations. During both hospitalizations, usual conservative treatment, such as intensive physiotherapy, inhalation therapy etc., was performed identically. Blood and sputum samples were collected in the morning of days 1, 3, 5, 8, 10, 12 and 15 between 08:00 and 09:00. Lung function tests were obtained on the same days.
Methods: The clinical severity of CF was assessed by the Shwachman-Kulczycki score, m modified by omitting the chest X-ray evaluation. The highest possible score was therefore 75 points. Chest X-rays were assessed by Chrispin-Norman score.
Pulmonary function tests were performed in a whole-body plethysmograph (E. Jaeger, Wrzburg, Germany). A sweat test was performed by quantitative pilocarpine iontophoresis. Serum was prepared according to guidelines published recently, and was stored at -20C until analysed. First morning sputum samples were diluted 1:1 with phosphate-buffered saline (PBS). After mixing, I mg of DNase (Sigma, Deisenhofen, Germany) and 100ml acetylcysteine was added. After vortexing, incubation was performed at 37C for 30 to 60 min until complete liquifaction. Subsequently, samples were centrifuged for 10 min at 1 000 x g. The supernatants were stored at -80C until analysis.
Leucocyte counts were determined from peripheral EDTA-blood using a Coulter counter (Coulter Electronics, Krefeld, Germany). Eosinophil and neutrophil counts were calculated by differential blood cell count. C-reactive protein was analysed by nephelometry (Behringwerke, Marburg, Germany) using routine methods. Total and specific IgE was determined by solid phase immunoassay, using the 'Pharmacia CAP System' (Kabi Pharmacia, Uppsala, Sweden). 17 Mediators of Inflammation Vol 4 1995 283 Statistical analysis: For statistical analysis, the (6.2-13.0mg/1)on day 1 to 11.4mg/1 (5.0-Wilcoxon matched-pairs signed rank and the 17.3 mg/1) on day 12. Spearman correlation tests were used. A value of Lung function (assessed by forced expiratory p < 0.05 was considered significant. Original data volume in 1 s (FEV1) showed a tendency to are given as medians and quartiles, increase during antibiotic treatment (from 37% of the predicted value on day 1 to 39% on day 12), while in the control phase there was a tend-Reu ency to decrease (from 38% of the predicted value on day 1 to 32% on day 12) ( Table 3).
Clinical parameters: Modified Shwachman-These results failed to reach significance. How-Kulczycki scores improved significantly during ever, comparison of the antibiotic treatment and antibiotic treatment from 47.5 on day 1 to 54.5 the control phase revealed a statistically signifion day 12 (p < 0.008, Fig. 1A). In the control cant difference for FEV on day 12 (p < 0.03).
phase there were no statistically significant differences: 49.5 on day 1 to 50.5 on day 12. Comparison of antibiotic treatment and control phases revealed a statistically significant difference on day 12 (p < 0.05, Fig. 1a).
Serum ECP concentration showed a tendency to decrease during antibiotic treatment from median 12.5 lg/1 on day 1 to 9.5 I.tg/1 on day 5 and 8.9btg/1 on day 12, but this failed to reach significance (Table 3, Fig. 3A). In the control phase, no changes in serum ECP could be noted: 13.8g/1 on day 1, 12.21.tg/1 on day 5, and 13.6 I.tg/1 on day 12 (Table 3, Fig. 3A). Peripheral blood eosinophil counts increased during antibiotic treatment (p < 0.03) but not during the control phase (Table 3, Fig. 3B).
Sputum data: Microbial growth of Pseudomonas aeruginosa in sputum did not show a significant decrease both during the antibiotic treatment and the control phase (Table 3). In general, sputum concentrations of MPO and ECP were 500to 1 000-fold higher than in serum (Table  4). There was no statistically significant correlation between serum and sputum MPO or between serum and sputum ECP concentration.

Discussion
Originally, a placebo-controlled design was intended and the present study was started in this manner. However, after two patients suffered from pulmonary exacerbations during a placebo 'treatment', the study was continued as an intraindividual cross-over design (each patient serving as its own control). Conditions during the antibiotic treatment were completely comparable (e.g., in terms of physiotherapy, inhalation therapy, dietetic intervention) with those during the control phase, with the exception of the intravenous treatment.
The reason for choosing a control-phase in addition to antibiotic treatment was the possible bias of the admission to hospital, which may improve a patient's well-being by intensified basic therapy. In order to minimize observer bias, the investigator of the study was not the physician responsible for the care of the patients on the ward.
In order to study the monitoring character of the parameters investigated, blood was sampled on 7 days during the antibiotic treatment and impact of antibiotic treatment on serum MPO control phases. All data points of each individual concentration. were analysed; however, three time-points (day 1, Concerning the role of the eosinophil and its 5 and 12) are given in the figures and tables to granule proteins in CF, there is little data in the avoid unclear presentation and because complete literature: Koller et al. found increased serum data did not add any information to those pre-levels of ECP in most of their 42 CF patients. 7 In sented, contrast, only three of our ten CF patients had Lung function parameters, such as FEV1, did an increased serum ECP concentration (> 20tg/ not differ significantly on day 1 either before 1), although CF in our patients was more severe starting antibiotic treatment (median 37% of pre-than in the Austrian group. The four atopic dicted) or control phase (median 38% of prepatients did not show a different pattern from -dicted). The same was true for modified the non-atopic patients.
Shwachman-Kulczycki scores: median 47.5 Conflicting results continue, e.g., Azzawi et al. points for day 1 of antibiotic treatment and 49.5 observed an increased number of EG2 positive points for day 1 of the control phase. Therefore, cells in post-mortem lungs of CF ,patients, but starting points for both phases were comparable normal numbers of eosinophils. This dis-( Table 2). crepancy may be interpreted as an increased FEV1 values showed an improvement at the propensity of the eosinophils to release their end of antibiotic treatment compared with the granule proteins as suggested by Koller. 7 Serum control phase. This effect of antibiotic treatment ECP levels showed a tendency to decrease under has been observed by other authors. 6'8'9 The antibiotic treatment, but this failed to reach sigimprovement of lung function was paralleled by nificance in either the Austrian study (elevated other parameters such as the Shwachman-Kulclevels of ECP in 90% of their patients), or in our zycki score, serum MPO concentration and CRP. investigation (30% of patients had an increased It seems that this positive effect is even seen at a ECP concentration).
time-point where patients did not suffer from Peripheral blood eosinophils increased during clinical exacerbations, antibiotic treatment in our study. Because of the CRP reflects the pulmonary inflammatory state low absolute numbers of eosinophils, our obserin chronically infected CF patients, is increased vation is probably not of clinical relevance, during pulmonary exacerbations, and decreases despite the statistical significance. Further studies during antibiotic treatment, 6 but it is not a reliare needed to elucidate the role of the eosinoable indicator of intermittent bacterial coloniza-phil for airway inflammation in CF. tion in early lung disease, i Very high In the local compartment, the sputum, MPO as concentrations of CRP are seen days and weeks well as ECP concentrations, were 500to 1 000before death. 2 However, in bronchoalveolar fold higher than in serum. This reflects their role lavage (BAL), the degree of airway obstruction, in local inflammatory lung injury. However, the neutrophil elastase activity, myeloperoxidase failure to reduce sputum concentrations of these activity, or total neutrophils did not correlate highly toxic proteins by intravenous antibiotic with the density of P. aeruginosa (CFU/ml)or therapy makes additional forms of treatment total pathogen burden in BAL fluid. 22 necessary. The Austrian group found a significant MPO, as a neutrophil inflammatory mediator, correlation between serum and sputum MPO and is not only increased in chronically infected ECP concentration. 7 Our data does not confirm cystic fibrosis patients, but also in other dis-this. The inherent problem lies in the noneases. It was found increased in BAL of idio-homogenous composition of sputum, depending pathic pulmonary fibrosis, 23 transiently on local factors of the lung region from which it increased in BAL during recovery from airway was sampled. In our view, sputum measurements hyperresponsiveness 24 and increased in BAL do not seem to be a reliable parameter for from some patients with bronchial asthma. 25 monitoring inflammatory changes in CF. Furthermore, the release of MPO from neut-From our data we conclude that: (1) measurerophils of patients with asthma was somewhat ments of MPO, neutrophils and CRP in periphhigher than in controls. 26 Serum MPO con-eral blood correlate with clinical parameters such centration was even suggested to be a good as the modified Shwachman-Kulczycki score; (2) indicator of exposure to noxious agents neutrophils and MPO seem to reflect inflammacausing respiratory disorders, such as in 'sick tory changes induced by antibiotic treatment; (3) building syndrome', iv From this study, there is eosinophils may. play a role in CF by an evidence for a role of MPO in chronic lung enhanced 'releasability'; and (4) sputum meainflammation in CF, as shown by elevated surements of mediators of inflammation cannot blood levels in some patients, as well as the be recommended.