The usefulness of sputum culture is controversial. This method of diagnosis has evoked the same spectrum of emotions and suffers from most of the same potential drawbacks as Gram stain of sputum. Various studies have demonstrated that either sputum or bronchoscopic specimens are frequently contaminated by upper respiratory tract bacteria. Some of the contaminants, such as Staphylococcus aureus, Streptococcus pneumoniae, Haemophilus influenzae, enteric gramegative bacteria, and Candida organisms, are potential lower respiratory tract pathogens. In addition, bronchoscopy may introduce local anesthetic into the specimen. Transtracheal aspiration (insertion of a needle into the trachea) or direct needle aspiration of the lung has been shown to produce relatively uncontaminated specimens. However, these techniques have potential complications. Although there is general agreement on the possibility of contamination, there is difference of opinion in the literature on the possibility that sputum culture may sometimes fail to detect the bacteria responsible for pneumonia. The importance of a specimen from the lower respiratory tract rather than the mouth or nasopharynx must be reemphasized, especially in seriously ill, uncooperative, or mentally impaired patients. As mentioned previously, a “pure culture” or marked predominance of one organism enhances suspicion of pathogenicity.

Several studies have found that the number of squamous epithelial cells in sputum provides a useful index for degree of oropharyngeal contamination; the greater the number of squamous epithelial cells per low-power field (10 x microscope objective), the more likely to have significant degrees of contamination. There is some disagreement in the literature between 10 or 25 squamous epithelial cells per low-power field as the criterion cutoff number. Significant contamination strongly suggests that a new specimen should be collected with special attention given to obtaining true lower respiratory tract material with a minimum of oral contamination. Some investigators also quantitate the number of segmented neutrophils (WBCs) per low-power field; more than 25 suggests acute inflammation (however, this alone does not differentiate upper from lower respiratory tract origin).