The main obstacle for using sputum as a diagnostic tool for LRTI is obtaining a good quality specimen. In our study, only 22.1 % of the samples were suitable when fewer than 10 SEC and >25 leukocyte per low power field was taken as acceptance criteria. This rate changes from 25% to 55% in different studies 9-13
. One of the major limitations of this paper is the lack of information of previous antibiotic use and final diagnosis of the patients. We wonder if there was an over-diagnosis of cases rather than obtaining low quality sputum. It has been suggested that the value of Gram stain and culture results are dependent upon the pretest probability that the patient has bacterial pneumonia and upon whether the patient has received antibiotics13
. Although more than 55% of the samples were recruited from what is presumed to be a specialty clinic for chest diseases, clinicians might be sending sputum samples to the laboratory without having supporting clinical and radiological data.
All accepted samples were examined microscopically by two microbiologists in our study and there was a good agreement between observers. Previous studies indicated a low concordance of Gram-stained specimens examined by different technicians, whereas others found the results to be reproducible 12,14-17. Overall sensitivity of Gram smear in our study was 78.6% with a specificity rate of 82%. Ewig et al 15 did not recommend sputum collection for diagnosis of community acquired pneumonia and suggested that Gram stain had a low diagnostic yield and a low number of positive samples had a corresponding growth in culture. Whereas Parry et al 16 suggested that sputum Gram smear can be a guide to the etiology of pneumonia, particularly pneumococcal pneumonia. Reed et al 17 revealed that, in good-quality sputum samples the sensitivity and specificity of Gram was 35.4% and 96.7% for S. pneumoniae and 42.8% and 99.4% for H. influenzae, respectively. The specificity of Gram smear was reached as 100% for H. influenzae and S.pneumoniae in our study. Previous studies showed that the washing procedure of sputum decreased the mean concentration of contaminants by 100 to 1000 fold and enhanced the value of the sputum samples 2,6. We theorized that washing procedure facilitates the detection of Gram negative bacteria, mainly thin H. influenzae bacilli in smears, probably related to the removal of mucus and contaminants from sputum related with saline washing and homogenization. Quantitative inoculation of sputum samples significantly increased overall culture positivity from 52% to 63.5% of our samples. In 32 samples, pathogenic bacteria were isolated only in quantitative cultures and in 24 of them the predominance of indicative bacteria was detected in the Gram smear. If bacterial count is greater than 106 cfu /ml in sputum with a predominance of related bacteria in Gram smear, with under 25 leukocytes per LPMF, this should improve the clinical management of the pneumonia.
The three most commonly isolated pathogens were Haemophilus influenzae , Pseudomonas aeruginosa and Streptococcus pneumoniae. These organisms were isolated at different rates in most outpatient studies of acquired LRTI in most series5,10,16. In 18 samples, multiple pathogens were isolated in significant numbers. However, it is not clear whether these agents act as co-pathogens or not.
The management of LRTI is remarkably simplified when the responsible pathogens are accurately identified 1,18-20. The main reason for detecting a microbiological cause of symptoms would be to select patients who could benefit from narrow-spectrum antibiotic treatment and to decrease bacterial resistance, side-effects and costs. Obtaining expectorated sputum samples is a non-invasive procedure and if done, it is best to use saline wash and then Gram stain and culture. Initial Gram staining of sputum samples, especially for for H. influenzae and S. pneumoniae is advisable when experienced microbiologists interpret the slides, since Gram stain is almost as effective as cultivation and results are available 48 hours sooner. Culture results are most convincing when the organism isolated in the culture is compatible with the morphology of organisms present in the Gram smear 21. One should keep in mind that discrepancy between direct examination and culture might also be associated to the fact that culture detects only viable bacteria whereas Gram staining may also detect non viable bacteria which might be related with previous antimicrobial consumption . On the other hand, it is mandatory to prepare guidelines for appropriate antimicrobials for empirical therapy and to reduce mortality with proper treatment cultivation and antimicrobial susceptibility data.
Acknowledgement. Major financial support for this study was provided by a research grant of MarmaraUniversity Research Fund (Project number SAG-TUS-200906-0168).