Severe Community-acquired Pneumonia Epidemiology and Prognostic Factors 1 - 3

ANTONI TORRES, JUAN SERRA-BATLLES, ANTONI FERRER, PATRICIO JIMENEZ, ROSA CELIS, ERIK COBO, and ROBERT RODRIGUEZ-ROISIN

Introduction

Although a large number of patients with community-acquired pneumonia can safely be treated on an outpatient basis, this infection is still a common cause of hospital admission (1, 2). In Spain the annual rate of community-acquired pneumonia hospital admissions is 77 of 100,000inhabitants (3). Despite the improvement in both antibiotic chemotherapy and a better knowledge of the etiologies that can cause community-acquired pneumonia (i.e., Legionellapneumophila and Branhamella catarrhalis), the mortality rate of community-acquired pneumonia still averages between 10and 20% (2,4). Different studies have shown that mortality is higher in those patients who develop severe acute respiratory failure and require mechanical ventilatory support because of their critical condition (2, 5-7). Although recent reports have partially addressed this issue (5-7), the incidence, etiology, prognostic factors, and outcome of patients with severe community-acquired pneumonia requiring intensive care are still insufficiently known. For example, prognostic factors of severe community-acquired pneumonia have only been studied using univariate statistical approaches. In the present study, weinvestigated the etiology,the clinical features, and the prognostic factors (using both univariate and recursive partitioning analyses), and the outcome, of 92 patients with severe communityacquired pneumonia admitted to a Respiratory Intensive Care Unit (RICU) over a period of 4 yr. Methods We prospectively studied all the patients admitted to the RICU at Hospital Clinic of Barcelona with the diagnosis of communityacquired pneumonia from January 1984 through December 1987. Patients wereadmitted to the RICU becausethey needed mechanical ventilation or were potential candidates for ventilatory support. Patients severelyim312

SUMMARY Over a period of 4 consecutive yr, 92 nonlmmun08uppressed patients (21 women and 71 men aged 53 ± 16 yr, i ± SO) with critical acute re8plratory failure (PaOlFlo2, 209 ± 9 mm Hg) caused by severe community-acquired pneumonia Were admitted to the respiratory Intensive care unit (RICU)of a general hospital. The most frequent underlying clinical condition was chronic obstructive pUlmonary disease (44 patients, 48%). A total of 56 patients (61%) required mechanical ventilation for a mean period of 10.7 ± 12.5 days, 29 of them (52%) needing PEEP (9.9 ± 3.8 cm H2 0 ). A group of 23 (25%) patients had criteria of adult respiratory distress syndrome (ARDS). A causal microorganism was identified in 48 patients (52%), the two m08t frequent etiologies being Streptococcus pneumonlae (14, 15%) and Leglonella pneumophlla (13, 14%). Pseudomonas aerug/nosa (5, 5%) wasalwaysassociated with bronchiectasis. Mortality due to severecommuntty-acqulred pneumonia was 22% (20 patients). According to univariate analysis, mortality was associated with anticipated death within 4 to 5 yr, Inadequate antibiotic treatment before RICUadmission, mechanical ventilation requirements, use of PEEp, FI02 > 0.6, coexistence of ARDS, radiographic 8pread of the pneumonia during RICUadmission, septic 8hock, bacteremia, and P. aeruglnosa a8the cause of the pneumonia. Further, recursive partitioning analysis selected two factors significantly related to the prognosis: the radiographic spread of the pneumonia during RICU admission and the presence of septic shock. Our statistical analysis could accurately predict prognosis In 75 patients (82%). Recommendations for antibiotic chemotherapy In these patients with both community-acquired pneumonia and severe acute respiratory failure are suggested. AM REV RESPIR DIS 1991; 144:312-318

munosuppressed (namely, hematologic malignancies, organ transplantation [renal and bone marrow], and the acquired immunodeficiency syndrome [AIDS)) were excluded from the study. Diagnostic criteria for community-acquired pneumonia were cough with or without sputum production, fever, leukocytosis or leukopenia, and pulmonary infiltrates on chest radiograph present at the time of hospitalization. The diagnosis of pneumonia was confirmed by a positive response to antibiotic treatment (38 cases), by isolation of a microorganism capable of producing a pulmonary infection (48 cases), and/or by postmortem pathologic findings (10cases). Furthermore, an alternative diagnosis to pneumonia was ruled out (using clinical, bronchoscopic, and/or pulmonary hemodynamic criteria). Patients who had been hospitalized within 10 days before developing pneumonia were excluded. In all cases the following variables were recorded: sex, age, hospital of origin, year of hospital admission, month of RICU admission, smoking and alcohol habits, prior illnesses and underlying clinical characteristics of the patients, initial signs and symptoms of pneumonia, antibiotic regimen given before and during RICU admission, chest X-ray features at RICU admission (unilateral or

bilateral involvement, abscess, and pleural effusion), arterial blood gas measurements at RICU admission (at maintenance fraction of inspired oxygen [F'Ioz] and also while breathing 100010 Os), hematocrit, hemoglobin, leukocytecount, percentage of band neutrophils, platelet count, serum glucose level, serum creatinine and electrolyte levels, aspartate aminotransferase (ASAT), alanine aminotransferase (ALAT), alkaline phosphatase, and gamma glutamyltransferase(GOT) levels, prothrombin time activity and plasma level of fibrinogen, mechanical ventilation requirements, the application of positive end-expiratory pressure (PEEP), coexistence of adult respiratory distress syndrome (ARDS), and maximal Fro, required. Regarding laborato-

(Received in original form January 30, 1990 and in revised form July 18, 1990) 1 From the Serveide Pneumologia, Hospital Clinic, Facultat de Medicina, Universitat de Barcelona, and the Escola de Alta Administracio i Direccio (EADA), Barcelona, Spain. 2 Supported in part by Grant Fisss 90/0758. 3 Correspondence and requests for reprints should be addressed to A. Torres, M.D., Servei de Pneumologia, Hospital Clinic, Villaroel170, Barcelona 08036, Spain.

CRITICAL

COMMUNIT~CQUIRED

PNEUMONIA

ry parameters, only initial values were used. Complications during RI CU admission, length of RICU stay, radiographic evolution of pneumonia, hemodialysis requirements, and patient outcome were also recorded. All radiographs were read by two of the authors (AT and JS-B) and also by the attending physician. The final assessment of each film represented the median interpretation of these three observers. A patient was considered a smoker if he or she had smoked more than 10 pack-years. Only three subjects were former smokers, being classified as smokers for purposes of these analyses. Alcohol habit was considered present when more than 80 g/day was consumed. Chronic obstructive pulmonary disease (COPD) was diagnosed combining clinical criteria and the coexistence of an obstructive ventilatory impairment assessed either before (in the majority, 38 of 44) or after RICU stay (6 of 44) (8). The underlying clinical condition of the patients was classified according to the criteria proposed by McCabe and Jackson (9) in three categories: (1) nonfatal; (2) ultimately fatal (death anticipated within 4 to 5 yr) (28 of 36 patients included in this group had severe COPD); and (3) fatal (death anticipated during hospitalization). Initial signs and symptoms of pneumonia were also classified in three categories: (1) typical (suggesting a common bacterial etiology) when two or more of the following criteria were present: fever, pleuritic pain, initial chills, purulent bronchial secretions, and leukocytosis (> 10.0 x 109 /L) or leukopenia « 4.0 x 109/L); (2) atypical when only dry cough, fever, and pulmonary infiltrates were present; and (3) aspiration pneumonia, when an episode of aspiration of gastric content was confirmed or suspected before the appearance of fever and pulmonary infiltrates. Antibiotic chemotherapy was considered adequate if the empiric drugs chosen accomplished the recommendations of the medical literature and were changed according to the antibiotic susceptibility reports, and their timing, route, dosage, and duration of the administration were all considered correct (10, 11). The diagnosis of respiratory failure was made when the Pao, was less than 60 mm Hg and/or Pac02 was equal to or greater than 50 mm Hg while breathing room air. The severity of respiratory failure was assessed by means of the alveolar-arterial Po2 difference (AaPo 2) and the Pao/PAo2 and Pao2/FI02 ratios, calculated according to standard formulas (assuming a respiratory gas exchange ratio of 0.8). ARDS was defined according to recent criteria (12). A poor radiographic evolution was defined as a spread of pulmonary infiltrates despite both antibiotic and intensive care treatment. Atelectasis, pleural effusion, or other noninfectious causes of the spread of pulmonary infiltrates were not included in this variable. Renal failure was defined when the serum creatinine level was greater than 1.2 mg/dl. The presence of shock was determined using the same criteria reported elsewhere (13, 14). Oth-

313

er causes of shock, such as hypovolemia or cardiogenic shock, were not considered. A microbiologic diagnosis was always attempted. Sputum specimens were not used as criteria for microbiologic diagnosis because of their low specificity. In all patients two blood cultures and serologies against L. pneumophi-

la, Mycoplasma pneumoniae, Coxiella burnetii, and Chlamydia psittaci were routinely performed. In 18 patients a pleural effusion culture was performed. In 16 patients direct immunofluorescence against L. pneumophila was carried out using endotracheal bronchoaspirate samples. In 14 patients pulmonary samples were obtained using nonfluoroscopic percutaneous ultrathin needle aspiration. Telescoping plugged catheter combined with bronchoalveolarlavage samples obtained via fiberoptic bronchoscope were retrieved from 10 patients requiring mechanical ventilation. All the diagnostic procedures were carried out shortly after admission to RI CD. A microorganism was considered the etiologic agent only when it could be isolated from blood or pleural fluid or from a lung needle aspiration sample, from a telescoping plugged catheter (cutoff point ~ 1

Severe community-acquired pneumonia. Epidemiology and prognostic factors.

Over a period of 4 consecutive yr, 92 nonimmunosuppressed patients (21 women and 71 men aged 53 +/- 16 yr, means = SD) with critical acute respiratory...
909KB Sizes 0 Downloads 0 Views