|
|
||||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From the Institut Clínic de Pneumologia i Cirurgia Toràcica (Mr. Vallès, Ms. Pinart, Ms. Piñer, and Dr. Torres), Servei de Microbiologia (Drs. Marcos and Marco), and Servei de Malalties Infeccioses (Dr. Mensa), Hospital Clínic de Barcelona, Barcelona, Spain.
Correspondence to: Antoni Torres, PhD, FCCP, Institut Clínic de Pneumologia i Cirurgia Toràcica, Hospital Clínic de Barcelona, Villarroel 170, 08036, Barcelona, Spain; e-mail: atorres{at}ub.edu
Abstract
Study objectives: To determine the incidence and trends of pneumococcal community-acquired pneumonia (CAP) resistant to antibiotics, to describe clinical and microbiological features of pneumococcal CAP, and to ascertain prognostic risk factors in a third-level hospital.
Design and setting: We performed a prospective study of all well-defined pneumococcal CAP hospitalizations in the Hospital Clínic de Barcelona (Spain) over 2 years of follow-up, and results were compared with a previous study.
Measurements and results: One hundred twenty-five patients were included (mean age, 59.6 years; 71.2% male and 28.8% female). Mortality was 7% (n = 9). Twenty-four percent were HIV-1 seropositive (n = 30), and 53% had at least one comorbidity (n = 65). Nonsusceptibility to penicillin, ceftriaxone, and erythromycin accounted for 34%, 9%, and 33%, respectively. A decrease in penicillin (p = 0.01) and cephalosporin (p < 0.001) resistance was observed on comparison with a previous study, while macrolide resistance remained unchanged. Serotype 1 infection was overrepresented (8%, n = 10). A bad outcome was related to female gender (relative risk [RR], 9.1; confidence interval [CI], 1.3 to 61.3), pleural effusion (RR, 13.35; CI, 1.9 to 93.1), and prior oral corticoid intake (RR, 10.59; CI, 1.2 to 91.2), whereas drug-resistant strains were not.
Conclusions: We found a decrease in drug resistance compared with a previous report and a relatively high incidence of serotype 1 pneumococcal CAP. We also observed a high prevalence of HIV-1 infection among individuals with pneumococcal pneumonia. We confirm the lack of association of drug resistance with mortality and length of hospitalization. Mortality was associated with female gender, pleural effusion, and previous oral corticoid treatment. These results should be better ascertained in further studies.
Key Words: antimicrobial resistance community-acquired pneumonia mortality serotype Streptococcus pneumoniae
Community-acquired pneumonia (CAP) is one of the most important causes of morbidity and mortality in developed countries, with a specific mortality ranging from 7 to 36% in different studies.1 The leading cause of CAP is Streptococcus pneumoniae, accounting for up to 50% of the cases.2 In developed countries, CAP is also the major cause of death by infectious disease.3 Furthermore, the incidence of pneumococcal pneumonia is related to age, with a peak in elderly people.4 The aging of the population in developed countries makes this prospect worse.5 The incidence of drug and multidrug resistance in pneumococcal pneumonia is increasing worldwide,6 being particularly high in Spain.7 S pneumoniae resistance to penicillin, defined as a minimum inhibitory concentration (MIC)
2 µg/mL, has not shown to influence the mortality by pneumococcal pneumonia, but the influence of highly resistant strains (MIC > 4 µg/mL) on morbidity and mortality remains controversial.8 One of the tools to fight against pneumococcal infections is the nonconjugate 23-valent vaccine and the newly conjugated 7-valent vaccine. The 7-valent vaccine has a restricted spectrum of coverage and has been designed to prevent pneumococcal disease in children, while its indication in adults is restricted to HIV-infected individuals or patients with immunosuppression or comorbidity in Spain.9 However, it is necessary to monitor patterns and trends of S pneumoniae antimicrobial susceptibility and serotype distribution. Besides drug resistance and serotype distribution, it is fundamental for clinicians to ascertain risk factors related to prognosis and morbidity, in order to establish appropriate treatment and management. The aim of this study was to describe microbiological and clinical features and trends on mortality and morbidity of a 2-year consecutive cohort of cases of well-defined pneumococcal pneumonia in a third-level hospital. Furthermore, we determined the risk factors for mortality and long-term hospitalization in our hospital.
Materials and Methods
Design and Study Population
A prospective study was carried out in the Hospital Clinic of Barcelona, a 750-bed third-level hospital covering an urban population of 500,000 inhabitants in Barcelona, Spain. Included were all consecutive cases of pneumonia admitted to the hospital between January 1, 1999, and December 31, 2002, in patients > 16 years old in whom S pneumoniae isolate was isolated. Results were compared with a previous study10 in our hospital that used the same methodology. The study received the approval from the ethics board of our institution.
Diagnostic Criteria
Pneumonia was defined as the presentation of acute onset of signs and symptoms of lower respiratory tract infection at hospital admission and radiographic evidence of a pulmonary infiltrate previously nonexistent or of any other known cause. All patients fulfilled the definite diagnosis of pneumococcal pneumonia, defined as clinical and radiologic pneumonia with the following: (1) validated sputum (< 10 squamous epithelial cells and > 25 polymorphonuclear cells per low-power field11), or tracheobronchial aspirate (BAS) cultures with
105 cfu/mL of S pneumoniae plus positive pneumococcal urinary antigen test result; or (2) the isolation of a S pneumoniae strain from blood or pleural effusion. Pneumonia was considered as CAP in patients not hospitalized for at least the previous 3 weeks.
Data Collection
The following data were collected at the time of hospital admission: age, gender, current smoking, alcohol habits, comorbid illness, antimicrobial treatment for pneumonia prior to hospital admission, and clinical symptoms: body temperature, presence of chills, chest pain, cough, expectoration, dyspnea, confusion, physical examination (presence of rales, heart rate, arterial systolic and diastolic pressure), chest radiograph pattern (alveolar, interstitial, or mixed infiltrate, multilobar involvement, bilateral involvement, pleural effusion), and blood analysis (leukocyte count, serum creatinine, Na, Ka, bilirubin, glutamic, oxaloacetic transaminase and glutamic pyruvic transaminase concentration, PCO2, PO2, and pH), and antimicrobial treatment at hospital admission.
During hospital admission, the following information was recorded: ICU admission, pleural effusion, septic shock as defined below, renal failure (serum creatinine level > 2 mg/dL from known baseline values in patients with previous impairment of renal function, or acute renal failure requiring dialysis), requirement for mechanical ventilation, in-hospital antimicrobial treatment, and in-hospital outcome (days of hospital stay and survival or death within 30 days of in-hospital treatment). All laboratory tests were performed using conventional equipment.
Antimicrobial Treatment Regimen
The general guideline for initial treatment consisted of the combination of a third-generation nonantipseudomonal cephalosporin with a macrolide or a quinolone in monotherapy (levofloxacin). The design of the initial antimicrobial treatment regimen was the responsibility of the physician in charge. Adjustments in antimicrobial treatment were made as soon as the results of susceptibility testing were available. Generally, cephalosporin treatment in the presence of intermediate resistance to cephalosporins was modified only in case of clinical nonresponse to this treatment.
Microbiologic and Laboratory Methods
Samples were collected in the emergency department in all patients with pneumonia following a standard protocol: (1) blood for culture and urine for antigen detection were collected in all patients, and sputum in all patients when possible; (2) BAS was performed in all intubated patients; and (3) pleural fluid was obtained in all patients with pleural effusion through thoracocentesis. Sample cultures were performed following standard procedures established in our laboratory. Antigen urine determination was done using a commercial kit (Binax; Binax; Portland, ME), following standardized procedures in our center.12 The following antibiotics were tested for antimicrobial susceptibility: penicillin, erythromycin, cefotaxime, vancomycin, tetracycline, and quinolones (ciprofloxacin, levofloxacin, ofloxacin, clinafloxacin, gatifloxacin, grepafloxacin, and sparfloxacin). Susceptibility or resistance of pneumococci to ß-lactams, macrolides, vancomycin, and tetracycline was determined by a microdilution test in accordance with the 2000 National Committee for Clinical Laboratory Standards guidelines.13 Susceptibility to quinolones was determined by the Epsilometer test (AB Biodisk; Solna, Sweden) as described elsewhere.1415 Strains were classified as susceptible, intermediate, or resistant following MIC cutoff interpretative standards.16 In summary, a pneumococcal isolate was considered susceptible to penicillin if the MIC was
0.06 µg/mL, to be of intermediate resistance if the MIC was 0.12 to 1 µg/mL, and to be fully resistant with a MIC of
2 µg/mL. Intermediate resistance to cephalosporin was defined as a MIC of cefotaxime of 1 µg/mL and fully resistance as a MIC
2 µg/mL. Resistance to macrolide antibiotic was defined by a MIC
1 µg/mL to erythromycin. All isolates were subcultured and stored frozen at 70°C and were submitted to the national Centre for Microbiology (Majadahonda, Madrid) for serotyping. Serotyping was performed by Qellung reaction, following standard procedures.
Definitions
Immunosuppression was defined as the presence of HIV-1 infection, solid-organ transplantation, or hematologic malignancy with known associated immune defects. Oral steroid treatment was defined as current regular use of immunosuppressive treatment (> 20 mg/d of prednisone for > 2 weeks). Underlying illness (comorbidity) included diabetes mellitus, heart disease, neurologic illness, cirrhosis, chronic renal insufficiency, and neoplastic disease. HIV-1 infection and pulmonary disease were considered separate from other comorbidities. Heart disease was defined as treatment for coronary artery disease or congestive heart failure or the presence of cardiovascular disease. Pulmonary disease was defined as treatment for asthma, COPD, the presence of interstitial lung disorders, antecedents of pulmonary tuberculosis, chronic bronchitis, and bronchiectasis. Renal disease was defined as preexisting renal disease with documented abnormal serum creatinine levels before the pneumonia episode. Cirrhosis was defined as previous diagnosis of cirrhosis of alcoholic, idiopathic, or viral origin. Neoplastic disease was defined as any solid tumor active at the time of presentation. Current smoking was defined as tobacco consumption
10 pack-years. Alcohol abuse was considered as the current intake of at least 80 g/d of alcohol. Shock was defined as BP
90 mm Hg not corrected by the administration of IV fluids or the need for vasopressor drugs. Acute respiratory failure at hospital admission was present if PO2 was < 60 mm Hg or PCO2 was > 45 mm Hg. Nonsusceptible strain includes intermediate or full resistance to a given antibiotic. Multidrug resistance was defined as penicillin nonsusceptibility (intermediate or full resistance) associated with nonsusceptibility to another non-ß-lactam drug. Prolonged length of stay (LOS) was defined as LOS > 75th percentile in our study samples, as defined by previous investigators.17
Statistical Analysis
Data were analyzed using statistical software (STATA 8.1; StataCorp; College Station, TX). Means, proportions, 95% confidence intervals (CI), and SDs were calculated. When comparing two proportions, the z test was used. For multivariate analysis, relative risks (RRs) were calculated using stepwise logistic regression. All p values are two tailed, and the level of significance was set at 5%.
Results
Patient Description
A total of 125 patients were included in our study, accounting for 13% (125 of 972 patients) of the total CAP attended (admitted and nonadmitted) and 14% of patients admitted (125 of 876 patients) to our hospital during the study period. The mean age was 59.6 years (SD, 19.8 years; range, 17 to 101 years); 89 patients (71.2%) were male, and 36 patients (28.8%) were female. The mean hospital stay was 9 days (SD, 8.0 days; range, 1 to 46 days). The description of comorbidities, clinical symptoms and resolution, and radiologic features compared with a previous study10 in our hospital is shown in Table 1
. Briefly, the mortality rate was 7% (n = 9), and 65 patients (53%) had some comorbidity: 39 patients had one concurrent disease, 14 patients had two concurrent diseases, and 12 patients had three concurrent diseases. Of all the patients, 61 patients (52%) had pulmonary comorbidity: 33 patients fulfilled the criteria of COPD, 8 patients had chronic bronchitis, 8 patients had bronchiectasia, 6 patients had ex-tuberculosis infection, 5 patients had asthma, and 1 patient had interstitial disease. Nine percent (n = 11) had immunosuppression other than HIV-1 infection. Patients with HIV-1 infection accounted for 24% (n = 30), and among whom the CD4 count was determined at hospital admission in 14 patients (mean, 405/µL; SD, 410/µL; range, 30 to 1,500/µL). Twelve percent (n = 15) were alcohol abusers. The main finding on chest radiography was an alveolar infiltrate (81%, n = 94). Pleural effusion developed in 19 patients (15%); among them, 8 patients had empyema. Differences between male and female patients were found. Female patients were less prone to smoke (17% vs 42%, p < 0.001), to have lung disease (32% vs 53%, p = 0.06), had a higher mortality (17% vs 3%, p = 0.06), to receive mechanical ventilation (25% vs 11%, p = 0.02), and to have shock (17% vs 5%, p = 0.01).
|
|
Serotype Distribution and Antimicrobial Susceptibility Patterns
Among the 122 isolates recovered for serotyping, 28 different serotypes were identified. Briefly, the most frequent serotypes were 3, 1, 14, and 6B. Four strains were nontypeable. Multidrug resistance was significantly associated with serotype 14, 9V, and 19 (p < 0.001). Invasive isolates showed a certain trend to be inversely associated with penicillin resistance (p = 0.06). We did not find significant differences among the serotype distribution of isolates from invasive samples (blood and pleural effusion) compared with BAS and sputum. Overall, 34% of strains in which antibiotic testing was performed (n = 123) were nonsusceptible to penicillin (n = 41, 30 intermediate, 11 fully resistant), 33% were nonsusceptible to erythromycin (n = 37), 9% were nonsusceptible to cefotaxime (n = 11, 10 intermediate, 1 fully resistant), and 1.5% were resistant to ciprofloxacin (n = 2). All strains were fully susceptible to imipenem, vancomycin, and the quinolones tested (levofloxacin, ofloxacin, clinafloxacin, gatifloxacin, grepafloxacin, and sparfloxacin). Among the isolates nonsusceptible to penicillin, 27 were nonsusceptible to erythromycin with statistical significance (p < 0.001). All strains nonsusceptible to cephalosporins were associated with nonsusceptibility to penicillin (six with full resistance and five with intermediate resistance to penicillin). The two ciprofloxacin resistant isolates were associated with penicillin, cephalosporin, and erythromycin resistance, and were serotypes 14 and 19, respectively. Antimicrobial susceptibility and serotype distribution are summarized in Table 2
. Compared with a previous study13 of pneumococcal CAP, we observed a decrease in the incidence of penicillin resistance strains (p = 0.01) and third-generation cephalosporin (p < 0.001). However, when stratifying for invasive isolates, the difference in third-generation cephalosporin resistance remained significant (p = 0.001), and penicillin resistance showed a nonsignificant trend (p = 0.1). These results are shown in Table 3
.
|
|
|
This study represents a series of cases of well-defined community-acquired pneumococcal pneumonia in a third-level hospital; thus, the results have a high degree of reliability. The current prevalence of nonsusceptible pneumococcal pneumonia in Spain is 35.7%,18 similar to the findings of our study but significantly lower than a previous study10 in our hospital. However, our data agree with the decrease observed in a national surveillance survey19 and could be partially attributed to the vaccination that covers serotypes typically associated with drug resistance, by a direct effect or a herd-immunity effect through 7-valent vaccination of children, the introduction of which has been stressed throughout the private sector in Spain.20 The success of campaigns to rationalize antibiotic administration in primary care centers may have contributed to this observed decrease in the incidence of drug-resistant pneumonia. Regarding macrolides, we did not observe any variation in nonsusceptibility compared with the referred studies. Furthermore, the low resistance of quinolones observed was of note.
Five of the six most frequent serotypes were the same than as those found in a survey carried out in Spain,21 except for the serotype 1, which is overrepresented in our study population (n = 10). Serotype 1 is not included in the 7-valent vaccine. The overrepresentation of serotype 1 could reflect the swing in the incidence of nonvaccine serotype-related strains in our study population. Geographic, temporal (seasonality), and patients (sample type, origin, age group) are factors that can affect the variations in susceptibility and serotype distribution observed among different pneumococcal clinical isolates.22
Other interesting findings were the high proportion of HIV-1infected people (24%) and their relationship to short LOS in the crude analysis. The HIV-1infected individuals probably did not share the risk factors typically associated with LOS (most were younger than the seronegative patients with pneumonia). The proportion of HIV-1infected individuals was higher (nonsignificant) than in a previous study10 in our hospital (30 of 125 patients vs 18 of 101 patients, p = 0.3). The strategies to prevent pneumococcal infection in HIV-1infected individuals are not conclusive. The effectiveness of vaccination with either the 23-valent or 7-valent conjugate vaccine remains controversial, and the major risk factor for pneumococcal pneumonia continues to be CD4 count.23 Nevertheless, despite successful highly active antiretroviral therapy, pneumococcal pneumonia remains a threat among HIV-1infected individuals.24 The CD4 count at the time of admission helps to better ascertain the reasons for the high burden of pneumococcal pneumonia among HIV-1seropositive people.
We must to point out that compared with a previous study10 in our hospital, the current series have a higher prevalence of comorbidities (pulmonary, renal, hepatic, and cardiac), which could reflect a trend of increase of underlying disease in patients with pneumococcal CAP. Interestingly, female gender showed a significantly higher risk for death by pneumococcal CAP. We observed a trend among women to have more serious disease, with a higher percentage of ICU admission (p = 0.02) and mechanical ventilation (p = 0.01) but fewer pulmonary comorbidities (p = 0.06). It seems that hospital-admitted women are far lesser in number than men but tend to present with more severe pneumonia. The reason for this finding remains unclear and should be better ascertained in further studies. Pulmonary infiltrates in patients receiving oral corticoids have been associated with attenuated local and systemic inflammatory response and high mortality, supporting our results in which oral corticoids intake was associated with a higher mortality.25
According to previous studies,262728 nonsusceptibility to penicillin is not related to a higher mortality. It is difficult to determine whether the results from more recent studies are due to the recent changes introduced in international guidelines29 that otherwise have increased treatment costs. However, the effect of highly resistant strains remains largely unknown8 and could not be determined in this study due to the lack of strains with MIC
4 µg/mL.
The main limitation of our study is the relatively low number of cases recorded regarding prognostic analysis (exitus, n = 9), which has not enough statistical power to generalize results (wide CI). Further studies should confirm these findings. Another limitation of this study is that we used a stricter pneumococcal pneumonia definition than the previous series10 including urinary antigen, and noninvasive isolates (typically associated to resistance) may be underrepresented. We have overridden this limitation by comparing only invasive isolates in both series.
In conclusion, we found a decrease in the incidence of pneumococcal CAP nonsusceptible to ß-lactams, while macrolide resistance remains unchanged. The serotype distribution and population findings are of great interest for prevention strategies against pneumococcal CAP.
Footnotes
Abbreviations: BAS = tracheobronchial aspirate; CAP = community-acquired pneumonia; CI = confidence interval; LOS = length of stay; MIC = minimum inhibitory concentration; RR = relative risk
This work was performed at the Hospital Clínic de Barcelona.
This study was supported by Red Gira FIS ISCIII 03/063, Respira FIS ISCIII 03/11, and FIS 00/0505.
Received for publication December 7, 2005. Accepted for publication February 21, 2006.
References
This article has been cited by other articles:
![]() |
I. Obando, L. Arroyo, D. Sanchez-Tatay, X. Valles, and A. Torres Decrease in Drug Resistance in Pneumococcal Community-Acquired Pneumonia Chest, July 1, 2007; 132(1): 359 - 360. [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |