Antibiotic Treatment of Severe Exacerbations of Chronic Obstructive Pulmonary Disease with Procalcitonin: A Randomized Noninferiority Trial

Background The duration of antibiotic treatment of exacerbations of COPD (ECOPD) is controversial. Serum procalcitonin (PCT) is a biomarker of bacterial infection used to identify the cause of ECOPD. Methods and Findings We investigated whether a PCT-guided plan would allow a shorter duration of antibiotic treatment in patients with severe ECOPD. For this multicenter, randomized, non-inferiority trial, we enrolled 184 patients hospitalized with ECOPD from 18 hospitals in Italy. Patients were assigned to receive antibiotics for 10 days (standard group) or for either 3 or 10 days (PCT group). The primary outcome was the rate of ECOPD at 6 months. Having planned to recruit 400 patients, we randomized only 183: 93 in the PCT group and 90 in the standard group. Thus, the completed study was underpowered. The ECOPD rate at 6 months between PCT-guided and standard antibiotic treatment was not significant (% difference, 4.04; 90% confidence interval [CI], −7.23 to 15.31), but the CI included the non-inferiority margin of 15. In the PCT-guided group, about 50% of patients were treated for 3 days, and there was no difference in primary or secondary outcomes compared to patients treated for 10 days. Conclusions Although the primary and secondary clinical outcomes were no different for patients treated for 3 or 10 days in the PCT group, the conclusion that antibiotics can be safely stopped after 3 days in patients with low serum PCT cannot be substantiated statistically. Thus, the results of this study are inconclusive regarding the noninferiority of the PCT-guided plan compared to the standard antibiotic treatment. The study was funded by Agenzia Italiana del Farmaco (AIFA-FARM58J2XH). Clinical trial registered with www.clinicaltrials.gov (NCT01125098). Trial Registration ClinicalTrials.gov NCT01125098

List of study centers 1 Exacerbations of chronic obstructive pulmonary disease (COPD) are the first cause of admission to pulmonary departments in Italy and worldwide. Guidelines recommend treating most patients hospitalized for exacerbations of COPD with antibiotics, even if the role of bacterial infections is often uncertain and the effect of antibiotics small. The recommendation to use antibiotics is guided by clinical signs and symptoms that have an insufficient diagnostic accuracy, whereas serum biomarkers, eg procalcitonin (pro-ct), may guide the selection of COPD patients who need antibiotic treatment. The main aim of our study is to investigate whether antibiotics can be safely stopped after 3 days or continued for 10 days according to a pro-ct-guided algorithm in patients hospitalized for exacerbations of COPD for whom guidelines recommend 3-10 days antibiotic treatment based on the presence of increased dyspnoea, sputum and purulence. The study is designed to assess the non inferiority of the pro-ct guided plan as compared to the standard guideline recommended plan. The pro-ct guided withholding of antibiotics is viewed as an experimental intervention associated with less antibiotic-associated complications, eg antibiotic resistance and drug-related side-effect and lower costs. The proposed study is a prospective, randomised controlled, single-blinded intervention trial comparing the standard with a pro-ct guided antibiotic treatment plan. Patients will be recruited from the Pulmonary Departments of 18 University or City Hospitals in Italy, starting October 2006 and continuing until April 1, 2008. Results expected by end of 2008. The 200 patients randomised to the standard non pro-ct guided antibiotic treatment plan will continue antibiotics for 10 days, whereas the 200 patients randomised to the pro-ct-guided antibiotic treatment plan, will continue for 10 days or stop antibiotics on day 3 depending on pro-ct levels measured at admission, day 1 and day 2. Serum pro-ct will be measured in a central laboratory. Patients will be examined at admission, discharge, 10 days, 1, 3 and 6 months. A telephone interview will be obtained at 2, 4 and 5 months. The primary outcome of the study will be the rate of exacerbations. Secondary outcomes will be hospital readmission, admission to ICU, change in FEV1, length of the hospital stay, and death. The sample size was estimated according to the primary outcome of the study. This study will be guided by a Steering Committee and monitored by an independent Safety Committee. Informed Consent x Medical history x Physical examination

Background information
Chronic obstructive pulmonary disease (COPD) is a chronic disease caused mainly by cigarette smoking that is characterized by progressive and poorly reversible airflow limitation that is usually associated with chronic dyspnoea, cough and sputum (1,2). COPD is a leading cause of morbidity and mortality worldwide and results in an economic and social burden that is both substantial and increasing (3).
COPD often exacerbates. COPD exacerbations are defined as a change in the patient's chronic respiratory symptoms sufficient to warrant a change in management (1,2). COPD exacerbations are a frequent cause of hospital admission and severely affect quality of life and prognosis. Mortality rates during hospitalizations are around 10% and during the year following hospitalization may be as high as 40% (4).
In Italy the 2001 mortality for COPD was 17.403, a number stable in the last 10 years (5).
The 2002 hospital admission caused by COPD (DRG 88) were 120.188, the majority (97.473) in patients > 65 years. If added to patients admitted to hospital for acute respiratory failure and pulmonary oedema (DRG 87) and of respiratory failure requiring mechanical ventilation (DRG 475), there would be >150.000 hospital admissions caused COPD exacerbation (6). The total estimated cost of DRG 88 and of the portion of DRG 87 and 475 that can be attributed to COPD exacerbation is > than € 500.000.000/year (6,7).
The main causes of COPD exacerbation are believed to the bacterial and viral infections and/or air pollution. More than 50% of COPD exacerbation are associated with viral infection, the majority of which are due to rhinovirus (4,8). The lower airways of 25 to 50% of COPD patients are colonized by bacteria, especially non capsulated Haemophilus influenzae, Streptococcus pneumoniae, and Moraxella catarrahalis (9). The predominant bacteria recovered in the lower airways of patients with mild exacerbation are H. influenzae, S. pneumoniae and M. catarrahalis, whereas in severe COPD requiring mechanical ventilation the predominant bacteria are gram negative bacilli, and P. aeruginosa is more frequent (4). Although bacteria can be isolated from the sputum of the majority of the patients with exacerbations of COPD, whether this finding represents colonization or infection remains controversial. The uncertain role of bacterial infections is also suggested by the limited evidence that antibiotics are effective in the treatment of COPD exacerbations. The strongest evidence available remains one randomized controlled trial published in 1987 that showed that the beneficial effect of antibiotics is limited to COPD patients who presented with an increase in all three of cardinal symptoms, i.e. dyspnoea, sputum volume, sputum purulence afterward named the "Anthonisen's criteria" (9). There was also some benefit in those patients with an increase in only two of these cardinal symptoms. A study on non-hospitalized patients with exacerbation of COPD showed a relationship between the purulence of the sputum and the presence of bacteria, suggesting that these patients should be treated with antibiotic if they also have at least one of the other two cardinal symptoms (dyspnoea or sputum volume) (10). However, these criteria for exacerbation of COPD have not been validated in other interventional trails and in fact the overall sensitivity and specificity of signs and symptoms for exacerbations of COPD requiring antibiotic therapy is poor (10). Thus exacerbations of COPD are often treated with antibiotics without evidence of clinically relevant bacterial disease. Based on the current available albeit limited evidence, guidelines (1,2,11) recommended antibiotics for: • patients with exacerbation of COPD with all 3 cardinal symptoms (increase of dyspnoea, and sputum volume and sputum purulence) • patients with exacerbation of COPD with 2 of the 3 cardinal symptoms, if increased purulence of sputum is one of the 2 symptoms • patients with a severe exacerbation of COPD that requires invasive mechanical ventilation (invasive and non-invasive).
Regarding hospitalized patients, in addition to the treatment with bronchodilators and systemic steroids, guidelines strongly recommend antibiotics treatment for most exacerbations of COPD even if their effect has not been properly investigated.
In fact, of 9 placebo-controlled trials included in one meta-analysis, 5 did not show a significant benefit in the group treated with antibiotics, and 4 were conducted in hospitalized patients. The overall results favored antibiotics, but the benefit was clinically marginal in all studies. Although inpatients seemed to benefit more from antibiotics than outpatients, extrapolation of the marginal effect seen in this meta-analysis (12), to properly diagnosed and characterized patients is not possible. Nonetheless, use of antibiotics in patients hospitalized with COPD exacerbation remains usual practice, although this treatment is not at all based on evidence.
In particular whether antibiotic treatment is effective in patients with COPD exacerbations characterized according to above described clinical criteria (dyspnoea, sputum and purulence) has never been examined in a properly designed and powered prospective study. The only properly designed and conducted randomized placebo controlled study was performed in patient admitted to the intensive care unit because of an exacerbation of COPD associated with acute respiratory failure requiring mechanical ventilation (13), and showed that treatment with antibiotics (oral moxafloxacin for 10 days) reduced the number of deaths, the duration of mechanical ventilation and hospital stay. However, the positive results obtained in this study can not be extrapolated to all patients hospitalized for COPD exacerbations, firstly because only one part of these patients are managed in the Intensive Care Unit and require mechanical ventilation, secondly because symptoms and particularly purulent sputum were not considered in the inclusion criteria.
In conclusion, the few data available and current concern about use of antibiotics and their contribution to selection of resistant bacteria, show that a properly designed and powered study is needed to investigate the role of antibiotics in the management of these patients. To limit antibiotic use, rapid and accurate differentiation of clinically relevant bacterial lower respiratory tract infections from other, mostly viral or non-infections, causes is pivotal. After obtaining a patient's medical history, physical examination, laboratory test, and chest radiograph, the clinician is often left with diagnostic uncertainty, because signs and symptoms of bacterial and viral infections widely overlap (14). Also, bacteria can be isolated from sputum in up 50% of patients with acute exacerbation of COPD, but whether this finding represents colonization or infection in controversial (14), also because viruses can also be isolated in a similar proportion, and in up to 25% of patients with COPD exacerbation there is neither evidence of bacteria nor of viruses (14,15).
In conclusion the prescription of antibiotics in COPD exacerbation is guided by clinical signs and symptoms with very poor scientific evidence.
The majority of routinely used clinical parameters have insufficient diagnostic predictive value to identify patients with bacterial infection requiring antibiotics, whereas some biomarkers, e.g. Creactive protein (CRP) and procalcitonin (pro-ct), might provide a more adequate discrimination between patients with and without bacterial respiratory infections requiring antibiotic therapy (16).

Rationale
Pro-ct is a precursor peptide from the hormone calcitonin. After translation from calcitoninmessenger RNA (mRNA), pro-ct is cleaved enzymatically into smaller peptides (17). Circulating amounts of calcitonin precursor, particularly pro-ct, are increased in severe bacterial infections, but remains fairly low in viral infection and non-specific inflammatory diseases. Previous studies have established the superior diagnostic accuracy of pro-ct in severe infection, and pro-ct guidance substantially reduces antibiotic use in lower respiratory tract infection without compromising clinical outcomes (16). According to a recently completed study: • the large majority of patients with COPD exacerbations are treated with antibiotics • the Anthonisen's criteria marginally reduce the use of antibiotics • a pro-CT guided antibiotic treatment almost halves the use of antibiotics.
C reactive protein also increases during bacterial infection and COPD exacerbations but, at variance with pro-ct: • its increase is delayed (17) • is blunted in steroid treated patients with bacterial infections (18) • is less sensitive in the early phase and less specific in the late phase of a bacterial infection, thus less adequate to guide the initiation and duration of antibiotics (19) which limits its clinical usefulness.
Other biomarkers (interleukin (IL)-6, IL-8, neopterin, TNFα, or soluble triggering receptor expressed on myeloid cells have not been adequately examined in COPD patients. The validation of a specific biomarker of bacterial respiratory infection such as pro-ct would provide a valuable information to the Italian Agency of Drug to reduce the inappropriate use of antibiotics for exacerbation of COPD, and would help the Italian Ministry of Health to reduce both the direct cost and the indirect costs of antibiotics due to side effects (resistance, diarrhea, allergy, etc).

Primary Objective
The main objective of study is to investigate whether antibiotics can be safely stopped at 3 days or continued for 10 days according to a pro-ct-guided plan in patients hospitalized for exacerbation of COPD requiring antibiotic treatment and associated with increased dyspnoea, increased sputum volume and sputum purulence.
The study is designed to assess the non inferiority of the pro-ct guided plan as compared to the standard 10 days antibiotic treatment plan recommended by guidelines. All patients will receive 3 days of antibiotic treatment regardless of the randomization arm, as recommended by guidelines.

Secondary Objectives
Secondary objectives are: 1. to estimate the direct and indirect cost of hospitalized exacerbation of COPD and the cost/benefits of experimental pro-ct guided antibiotic treatment plan; 2. to determine whether the pro-ct guided continuation of antibiotic treatment for 10 days is superior to the standard non pro-ct guided antibiotic treatment of 10 days because pro-ct identifies patients whose exacerbation are caused by bacterial infection; 3. to determine the rate of antibiotic-associated side effects (i.e. frequency of gastrointestinal symptoms like diarrhea or nausea, allergy, other drug-specific adverse reaction); 4. to estimate, through pro-ct, the proportion of exacerbation of COPD defined with the "Anthonisen's" clinical criteria that need to start antibiotics in the first place 5. to investigate the heterogeneity of COPD exacerbation, as the symptoms of COPD exacerbation (increasing dyspnoea and cough and sputum) are non specific, and they may be often due to other diseases (acute heart failure, pneumonia, pulmonary thromboembolism) 6. to compare the predictive value of recommended clinical criteria with traditional (eg. CRP, leukocyte, fever, etc.) and novel (acute phase proteins, prohormones) biomarkers (Ancillary project 2) 7. to determine the relationship between serum pro-ct and sputum bacterial/viral load/etiology (Ancillary project 3).

STUDY DESIGN
This is a prospective, randomized, controlled, single-blinded interventional trail comparing the standard antibiotic therapy recommended by current guidelines (1,2), with a pro-ct guided antibiotic treatment for patient with COPD exacerbation admitted to the respiratory units of 18 sites (University or City Hospital). Only COPD patients hospitalized because of an exacerbation characterized by increased dyspnoea, sputum and purulence, who according to guideline require treatment with antibiotic, will be recruited starting October 2006 and continuing until April 2008.
The eligible patients will be randomly assigned to the standard or to the pro-ct guided antibiotic treatment plan according to a 1:1 computer-generated scheme. Randomization will be stratified according to hospital.
Initially all patients will start antibiotics at admission to hospital and pro-ct will be measured at admission, day 1 and 2. All pro-ct samples will be sent by courier on day 2 to the central laboratory (Central laboratory, University of Padova), where serum pro-ct will be measured.
Patients randomized to the standard non pro-ct guided antibiotic treatment plan, will continue antibiotics for 10 days. Patients randomized to the pro-ct guided antibiotic treatment plan, will continue or stop antibiotics on day 3 depending on pro-ct levels, using cut-offs previously validated (16). The final decision on the antibiotic treatment will be left to the investigator according to guidelines (1,2). Likewise, other diagnostic procedures, concomitant therapeutic regimen or change of regimen and final decision to initiate antibiotic treatment will be, in all cases, left to the discretion of investigator. In the pro-ct group all investigator will be advised to follow the antibiotic treatment algorithm based on the pro-ct levels. Thereafter, patients will be prospectively followed up during hospitalization and after hospital discharge.

Number of Subjects
All patients admitted to the participating hospital respiratory units for exacerbation of COPD will be potential subjects. The Investigators will define and classify the severity of COPD according to the Global Initiative for Chronic Obstructive Lung Disease guidelines (http://www.goldcopd.com).
A total of 400 patients (200 for arm) hospitalized for COPD exacerbation will be enrolled in the study. For the sample size see chapter 8.1.

Inclusion criteria
• Male and female patients who give written informed consent • Age: adults >18 years old • COPD diagnosis according to GOLD guidelines: FEV1/FVC <70% with FEV1<80% of predicted. If a COPD diagnosis is not evident from previous spirometric exams of the patient, a new spirometry will be obtained within 3 days of admission and a diagnosis of COPD will be made if all of the following conditions are true: 1) FEV1/FVC <70% with FEV1 <80% of predicted; 2) history of cigarette smoking; 3) exclusion of an asthma diagnosis. At the end of the study (i.e. at 6 months) a new spirometric exam will be obtained from these patients to confirm the new diagnosis of COPD made at admission.
• Diagnosis of COPD exacerbation: -defined as acute-onset dyspnoea and/or cough associated with increased purulent sputum production (

Exclusion criteria
• Female subjects: pregnant, lactating mother or lack of efficient contraception in a subject with child-bearing potential (e.g. contraceptive methods other than oral contraceptives, IUD, tubal ligature) • Diagnosis of asthma

Withdrawal criteria
Patients have the right to withdraw from the study at any time for any reason, including personal reasons. The Investigators have also the right to withdraw patients from the study in the event of intercurrent illness that necessitates pharmacological treatment with a disallowed drug, in case of an adverse event, treatment failure, protocol violations, poor compliance, pregnancy or any other reason.

Treatment Assignment and blinding
On day 2 each eligible patient will be randomly assigned to one of the two treatment plans.
The randomization will be performed via website. The allocation schedule will be computer generated with random numbers and prepared before the study begins by an independent randomization center, which will not reveal the allocation sequence to the Investigators. After baseline assessments, the Investigators will connect to the web site and, in case of eligibility, patients will be allocated to one of the arms. The randomization will be stratified by centre.
The Investigators will not be aware of randomization sequence before allocation of the study group. Every patient's assignment will occur at the time of admission to the clinical department by the Investigators and antibiotics will be given according to the protocol.

Treatment Compliance
Compliance will be assessed on the basis of capsule counts/medical record.

Concomitant Medication
Starting at admission each patient will also receive inhaled bronchodilators and either oral or intravenous glucocorticosteroids for 14 days according to guidelines (1). These 14 days of gluococorticosteroids will include any days of glucocorticoids taken by the patient at home before admission. Additional use, in case of need, of short acting bronchodilators will also be allowed and recorded throughout all the study.
Patients can be enrolled even if antibiotic treatment was started before hospital admission.
The selection of antibiotic (amoxycilline clavulanate or quinolones), its way of administration and dose will be decided by the Investigators according to guidelines (1).
The patients will remain hospitalized for at least 3 days afterwards they will be discharged according to guidelines (1). At discharge, all patients will be prescribed regular treatment according to guidelines (1). Treatment will be monitored and recorded regularly at each visit.

CLINICAL PROCEDURES VISIT 1 -DAY 0 -HOSPITALIZATION
When the patient arrives in the hospital with a diagnosis of COPD exacerbation, he/she will be examined by the investigator. The inclusion/exclusion criteria will be examined and if the patient satisfies these criteria the investigator will obtain written informed consent to participate in the study. The procedures for Visit 1 will follow. • Pharmacoeconomy questionnaire • spirometry performed according to guidelines (22) • sputum collection: 1.5 ml of sputum produced by the patient will be obtained. ½ of this sample will be sent to the local laboratory for bacteriology and intracellular pathogens examination. Each center will explicitly ask its local laboratory to recognize 10 10 CFU/ml. ¼ of the sputum sample will be stored at -80°C and sent to the University of Ferrara for respiratory viruses RT-PCR analysis (ANCILLARY PROJECT 3). The final ¼ of the sample will be used for inflammatory cell counts, slide preparation and the supernatants will be stored at -80°C for immunoistochemistry. Each center will be responsible for preparing a sputum slide whose reading will be centralized at the University of Ferrara. The presence of blood in the sputum does not invalidate the sample. • respiratory symptoms using previously validated questionnaires (24)(25)(26) • Pharmacoeconomy questionnaire • spirometry performed according to guidelines (22) • sputum collection: 1.5 ml of spontaneous or induced sputum will be obtained. ½ of this sample will be sent to the local laboratory for bacteriology and intracellular pathogens examination. Each center will explicitly ask its local laboratory to recognize 10 10 CFU/ml. ¼ of the sputum sample will be stored at -80°C and sent to the University of Ferrara for respiratory viruses RT-PCR analysis (ANCILLARY PROJECT 3). The final ¼ of the sample will be used for inflammatory cell counts, slide preparation and the supernatants will be stored at -80°C for immunoistochemistry. Each center will be responsible for preparing a sputum slide whose reading will be centralized at the University of Ferrara. The presence of blood in the sputum does not invalidate the sample. • arterial blood gas analysis • Treatment compliance monitoring • Adverse events monitoring (including any episodes of pulmonary tromboembolism, cardiovascular events including acute onset pulmonary edema, and pneumonia)

VISIT 5 -DAY 30
The Visit 5 procedures will include: • physical examination • Pharmacoeconomy questionnaire • spirometry performed according to guidelines (22) • sputum collection: 1.5 ml of spontaneous or induced sputum will be obtained. ½ of this sample will be sent to the local laboratory for bacteriology and intracellular pathogens examination. Each center will explicitly ask its local laboratory to recognize 10 10 CFU/ml. ¼ of the sputum sample will be stored at -80°C and sent to the University of Ferrara for respiratory viruses RT-PCR analysis (ANCILLARY PROJECT 3). The final ¼ of the sample will be used for inflammatory cell counts, slide preparation and the supernatants will be stored at -80°C for immunoistochemistry. Each center will be responsible for preparing a sputum slide whose reading will be centralized at the University of Ferrara. The presence of blood in the sputum does not invalidate the sample. • Pharmacoeconomy questionnaire • spirometry performed according to guidelines (22) • arterial blood gas analysis • Adverse events (including any exacerbation of COPD since the last visit, any hospitalization and/or admission to an Intensive Care Unit, any episodes of pulmonary tromboembolism, cardiovascular events including acute onset pulmonary edema, pneumonia occurred since the last visit)

VISIT 7 -AT DAY 180
The Visit 7 procedures will include: • physical examination • spirometry performed according to guidelines (22) • sputum collection: 1.5 ml of spontaneous or induced sputum will be obtained. ½ of this sample will be sent to the local laboratory for bacteriology and intracellular pathogens examination. Each center will explicitly ask its local laboratory to recognize 10 10 CFU/ml. ¼ of the sputum sample will be stored at -80°C and sent to the University of Ferrara for respiratory viruses RT-PCR analysis (ANCILLARY PROJECT 3). The final ¼ of the sample will be used for inflammatory cell counts, slide preparation and the supernatants will be stored at -80°C for immunoistochemistry. Each center will be responsible for preparing a sputum slide whose reading will be centralized at the University of Ferrara. The presence of blood in the sputum does not invalidate the sample. • arterial blood gas analysis • Adverse events (including any exacerbation of COPD since the last visit, any hospitalization and/or admission to an Intensive Care Unit, any episodes of pulmonary tromboembolism, cardiovascular events including acute onset pulmonary edema, pneumonia occurred since the last visit) In addition to the above procedures, all patients will be contacted by telephone at months 2, 4 and 5 to obtain information on clinical status, treatment monitoring, adverse events monitoring and pharmacoeconomy.

Central Laboratory Procedures
All procedures will be detailed in the Study Procedures Handbook, which will be distributed to all Investigators before the beginning of the study.

Study endpoints
The primary endpoint of the study will be the rate of exacerbation at 6 months.
Secondary endpoints will be: • Hospital readmission • Admission to ICU (Intensive Care Unit) • Change in FEV1 • Length of the Hospital stay

Safety Evaluation / Adverse Event
The Investigator will evaluate the patients for any possible adverse event at each visit, using a standard clinical criteria in particular for pneumonia, thromboembolism, acute pulmonary edema and other cardiovascular events. The Investigator will also question the patients about other possible adverse event.

Sample size
The sample size has been estimated according to the primary outcome of the study, i.e. the exacerbation rate at 6 months calculated by dividing the number of patients experiencing at least one exacerbation over a period of 6 months by the total number of patient initially randomized.
The rate of COPD exacerbation/treatment failure has been chosen as primary endpoint because it is clinically relevant for the patients and because this is where differences between the 2 groups, if any, may occur.
To estimate the frequency of this primary outcome data from a previous study have been used. These data suggest that exacerbation occur in about 50% of patient within 6 months (10).
To define non-inferiority with regard to the primary outcome, we settled on a 15% difference in the percentage of patients with an exacerbation within 6 months of follow up as clinically tolerable upper limit. We hypothesized that the exacerbation rates for both treatment plans are equal (at 50%), and that a difference of 15% or less would be unimportant, setting β=0.20 and α=0.05 (one tail) the estimated simple size is 140 patients per arm.
Formally, the null hypothesis is that the event rate for the standard treatment plan is 15% lower than the pro-ct guided plan, and the study has a power of 80% to reject this null hypothesis.
Similarly, the likelihood is 80% than the 95% confidence interval for the difference in event rates will exclude a 15% difference in favor of the standard treatment.
Taking into account an estimated 30% drop-out of patients at 6 months follow up, we increased the sample size up to 200 patient per arm.
Sample size has been estimated by using power and precision, which is a computer program for statistical power analysis and confidence intervals.

Data Analysis
For each single patient, all study data obtained by interview, clinical and laboratory tests and reviewing of the medical records will be reported in a Case Report Form (CRF) that will be collected at the end of study by the external Clinical Research Organization (CROM Srl, Verona) The study will be terminated prematurely if a disadvantage become evident in one treatment arm assessed. Because in non-inferiority trials intention to treat and per protocol analyses may both falsely favour the conclusion (6), we will analyze our data in both ways. Means, standard deviation, median, proportion, hazard rate and statistical test for significance will be calculated, along with 95% confidence interval. The t test will be used for continuous variables, whereas the chi square test will be used for discrete variables. The association between treatment and outcomes will be estimated by calculating odd ratios. Logistic regression models will be fitted to the data to adjust for potential unbalances of prognostic factors. In addition, COX's regression models will be fitted at 6 months to account for the precise event times and patients lost to follow up. In order to avoid potential bias due to differences in subgroup characteristics, we will randomize patients within the pre-specified strata at the time of study entry.
This study will be guided by a Steering Committee consisting of clinical experts in COPD, infectious diseases, and clinical pharmacology (L. Fabbri, Chair, A. Papi, L. Richeldi, R. D'Amico, B. Muller) and monitored by an independent Safety Committee (P. Geppetti, Chair, C. Giuntini, S. Johnston) that will review the results of one interim analysis of the primary efficacy endpoint and cumulative data on serious adverse events.
An external professional Clinical Research Organization (CROM Srl, Verona) will be responsible for the training of the personnel of the individual centers, regular monitoring (7 visits throughout the study) and quality control of the procedures and data collection.

Safety analysis
The assessment of safety will be based mainly on the frequency of adverse events and on the number of laboratory values that fall outside of pre-determined ranges. Other safety data (e.g. electrocardiogram, vital signs, special tests) will be considered as appropriate.
Adverse events will be listed summarizing the number and percentage of patients experiencing adverse event. Any other information collected (e.g. severity or relationship to the study medication) will be listed as appropriate.
Laboratory data will be presented by summary statistics of raw data and change from baseline values (means, medians, standard deviations, ranges). Inflammatory markers and exacerbation aetiology and severity.
The aim of the project is to identify novel markers of immunity and/or of inflammatory response able to discriminate whether a patients is at risk of infection disease or complication related to infection disease. We will focus our attention on recently identified molecules called long pentraxin (PTX3). An original assay to quantify this molecules has been validated in recent studies (1). PTX3 is structurally related to C-reactive protein that play a non-redudant role in innate immune response against lung pathogens, fungi and bacteria (i.e. Klebsiella pneumoniae, etc). Furthermore, PTX3 plays a regulatory role in innate immune response and in inflammatory cascade. Increased level of PTX3 have been described in patients with infections including pulmonary infections. We will also measure levels of interlukin-1 (IL-1) decoy receptor. IL-1 decoy receptor is essential for the negative regulation of IL-1 system that can be used as marker of endogenous anti-inflammatory pathways (2). PTX3 and IL-1 decoy receptor will be measured first in a subgroup of subjects enrolled in the study (50 subjects) and than the analysis will be extended to all the study population.
Serum amyloid A, a protein of the acute-phase inflammatory response, and additional biomarkers (such as C reactive protein and RANTES) will be also measured in plasma samples and relationship with exacerbation aetiology and/or severity analysed. C -reactive protein (CRP) and serum amyloid A (SAA) are members of a family of proteins that are major components of the acute-phase  Relationship between COPD exacerbation's etiology and procalcitonin serum levels.
The aim of this project is to validate the rationale for the use of serum procalcitonin to guide the requirement for antibiotic treatment in COPD exacerbations. Given the fact that a great proportion of COPD patients are colonized by bacteria in stable conditions, and since an increased sputum bacterial load is expected in COPD exacerbation associated to bacterial infection, we will analyze the relationship between bacterial load (quantitative bacteriology) and serum procalcitonin levels to evaluate whether a direct relationship is present between the severity of airway infection and serum levels of procalcitonin. We will also analyze serum procalcitonin levels (and changes from stable conditions) in COPD exacerbation not associated to bacterial infection (viral infection and non infective exacerbations). Bacterial isolation and identification will be performed with the use of standard techniques.