The bacteriological examination of sputum.

Pitfalls in the collection, transport, processing, and interpretation of the expectorated sputum culture are discussed. Data are presented which suggest that, under stricdy controlled conditions, the culture of sputum may be a valuable adjunct to the diagnosis of bacterial pneumonia. In 1971 there were 2,068,000 diagnosed cases of bacterial pneumonia resulting in 34,183,000 days of disabled activity and 10 percent of hospital admissions. As ex­ pected, since pneumonia is common, the demand for sputum cultures is great; in fact, over 37 million dollars is spent an­ nually on this laboratory test. Yet, for such a frequently performed test, sputum cul­ tures are still plagued by diagnostic uncer­ tainty with scientific obscurity. When an infectious agent is isolated from a site which is normally devoid of micro­ organisms such as the cerebral spinal fluid, the blood, or the pleural fluid, its patho­ genic role is easily established. Many parts of the body, such as the respiratory, gastro­ intestinal or genitourinary tracts have, how­ ever, a “normal flora.” Consequently, recov­ ery of a microorganism from these areas does not necessarily indicate infection. Spu­ tum cultures are one of the most notorious examples where the problem is not the de­ tection of a pathogen but rather the deter­ mination as to which one of the many bac­ teria found is responsible for the patient's infection. Of the many facets of the bacteriological examination of sputum, three are of par­ ticular importance. They include current methodology, difficulty encountered in the use of these methods and possible solutions for better laboratory diagnosis of respira­ tory infection. C u rren t M ethodology T h e G b a m S t a in The Gram stain is the most widely used and abused procedure in microbiology but it can be one of the most valuable, if cor­ rectly performed and intelligently inter­ preted. All respiratory secretions should be Gram stained and expeditiously examined. This simple procedure provides (1 ) an as­ sessment of the quality of the specimen, (2) a hint as to the microbial etiology of the disease and (3 ) information on the cellular response to infection. In preparing sputum for staining, the most purulent portion should be applied to BA CTERIO LO G ICA L EXA M IN A TIO N OF SPU TU M 6 1 the slide, spread so as to provide for read­ ing and air dried. The Gram stain involves four step wise procedures: (1 ) primary staining, (2 ) addition of a mordant, (3) decolorization and (4) counterstaining. The most critical step is decolorization. The time required to decolorize a smear will vary depending upon the ingredients of the decolorizing reagent ( ethanol being the slowest and acetone the most rapid) and the thickness of the smear. If this step is not well standardized, then quality control slides containing a mixture o f gram positive and gram negative organisms should be in­ cluded in a staining run at least once a day. Morphological evaluation of the stained secretions provides valuable information on the source of the specimen. Secretions, al­ legedly sputum but in fact contaminated by saliva, contain many epithelial cells but few, if any, polymorphonuclear (PM N) leukocytes. The bacterial nature of such a specimen reveals many different morpho­ logical types of bacteria suggestive of nor­ mal mouth flora. Frequently, gram positive, lancet shaped diplococci resembling Strep­ tococcus pneumoniae are superimposed on or attached to epithelial cells. Subsequent culture often reveals these organisms to be the alpha hemolytic streptococci of the oral cavity. In the sputa of patients with bac­ terial pneumonia, there is usually a pro­ liferation of PMN’s and mucous strands apparent on Gram stain. Streptococcus pneumoniae, if present, may be observed without resorting to a specific capsular stain. Other bacterial agents of pneumonia, such as the Staphylococci, Hemophilus in­ fluenzae and Klebsiella pneumoniae, may be presumptively identified by Gram stain. An attempt should be made to determine the predominant organism as all of the aforementioned bacteria can be seen in the respiratory secretions, especially those of hospitalized patients who may or may not have a pulmonary disease. C u ltu ra l T ech n iques The guide lines for the culture of respira­ tory secretions are very broad. Although the decision as to the choice of media is best left to the microbiologist or the pathol­ ogist, culture protocol should include an enrichment medium (e.g. blood agar), a medium which provides for the growth of fastidious organisms requiring hemin and NADH2, (e.g. chocolate agar) and a me­ dium which selectively suppresses flora and allows the growth of either gram negative or positive bacteria (e.g. MacConkey’s agar, Mannitol Salt agar). Incubation of non-selective media in a C 0 2-enriched en­ vironment is also beneficial. Sputum should be inspected grossly in the laboratory. The presence of excessive froth indicates a specimen which is pri­ marily saliva. The laboratory must be pre­ pared to reject saliva samples since they will provide meaningless and misleading data. Sputum should be inspected for areas of pus which provide the best material for smear and culture. Unfortunately, the large daily volume of sputum specimens, the frivolous nature of sputum collection and the failure to differentiate between sputum and saliva has led to a deterioration of specimen quality. The time a specimen is taken should be recorded on the requisition. Respiratory secretions are an effective growth medium for bacteria; if allowed to stand at room temperature for more than three to four hours, there will be bacterial overgrowth and subsequent difficulty in identifying the etiological agent. The results of culture must be evaluated in light of both the clinical presentation and known microbial ecological concepts. For example, more than a billion aerobic and anaerobic organisms are present as normal flora in a milliliter of oral secretions. These organisms can readily contaminate expectorated sputum. The normal anaero­ bic flora of the oral cavity will grow if 6 2 T ILT O N , M A DERAZA, IA N N IN I AND Q U IN TILIAN I sputum is cultured anaerobically. Thus, anaerobic incubation of sputum specimens is not indicated except for those secretions that are obtained by methods, such as transtracheal aspiration, which bypass the normal oral flora. E n coun tered Prob lem s Several studies have shown lack of cor­ relation between potential pathogens iso­ lated from sputum and the clinical diag­ nosis even when appropriate technique is employed.1’3 Bartlett and Melnick2 re­ ported that the incidence of identification in Gram stains or isolation in culture of Streptococcus pneumoniae was not signifi­ cantly different in infected and uninfected patients. A subsequent study by Hoeprich7 revealed that 34 percent of patients with no respiratory disease harbored Streptococcus pneumoniae in the sputum. Such results in­ crease the concern as to what relevance, if any, have cultures of expectorated sputum. The advisability of isolation, spéciation, and antimicrobial susceptibility on multiple potential pathogens in a sputum is ques­ tionable. Reports have documented the al­ teration in the normal upper respiratory flora of acutely ill patients as a function of length of hospital stay5 and use of inhala­ tion therapy equipment. The colonizing flora becomes predominately gram negative rods such as E. coli, Pseudomonas, and Enterobacter. Both laboratorian and clinician must recognize the distinctions between colonization and infection. In our labora­ tory, the presence of three or more indi­ vidual species of gram negative rods in a sputum specimen are suggestive of coloni­ zation, specimen deterioration or contam­ ination by normal flora. These organisms are not followed through unless subsequent consultation with the physician reveals an overriding reason to do so. The quintescence of the problem appears to be that sputum is contaminated with normal flora and the etiological agent is masked, that sputum is not always sputum but often saliva and that the visualization and/or isolation of a pathogen from sputum does not always indicate infection. L ab oratory D iagn osis o f R esp irato ry In fection Recognizing the problems that exist, there have been many attempts to over­ come them. Sputum has been washed to remove adhering saliva.4 This procedure was repeated in our laboratory on 50 spu­ tum specimens. There was no significant difference in the Gram stain or cultural re­ sults of washed and unwashed sputum. This washing procedure, if done in the open laboratory, produces potentially harm­ ful aerosols, especially from patients sus­ pected of mycobacterial disease. Other studies2’9’11 have proposed that sputum be liquified with a mucolytic agent, diluted in buffer and a quantitative bacterial culture performed. No increase in the numbers of positive cultures or in the numbers of path­ ogens detected was observed in our lab­ oratory with the use of the quantitative method. Furthermore, there was no reduc­ tion in the isolation of mixed flora. These data corroborate the conclusions of Bartlett and Melnick2 and Hahn and Beaty.6 An obvious approach to the problem of sputum specimen quality is to bypass the normal mouth flora by procuring respira­ tory secretions by means other than expec­ toration. Such procedures include endotra­ cheal aspiration, trans-tracheal aspiration (T T A ), bronchial washing, bronchial brushing, and percutaneous lung puncture and aspiration. Kalinske et al8 and Hahn and Beaty6 studied 102 and 61 patients, respectively, and concluded that TTA was a safe pro­ cedure with virtually no complications. Spencer and Beaty,10 however, reported cardiac arrest in three patients subjected to TTA. Distinct differences have been demonstrated in the microbial flora of spu­ tum versus secretions collected by these bypass procedures. BA CTERIO LO G ICA L EXA M IN A TIO N O F SPU TU M 6 3 Kalinske et als compared cultural results of sputum and transtracheal aspirates in 102 patients. In 30 percent of the patients, gram negative rods were isolated more fre­ quently from the sputum specimen than from the TTA. “False positive” sputums were seen less frequently with Staphylococ­ cus aureus (10 percent) and Streptococcus pneumoniae (8 percent). A similar but smaller study carried out by the authors (RQ, PI) revealed “false positive” sputum results attributable to gram negative rods (10 percent), Candida albicans (10 per­ cent), and Staphylococcus aureus (6 per­ cent). In yet another series by the authors, 22 patients with acute bacterial pneumonia, as indicated by fever, leukocytosis and pulmo­ nary infiltrate, underwent either TTA or percutaneous lung aspiration, expectorated sputum collection and blood cultures. Sev­ eral samples were obtained from each pa­ tient and the sample which appeared most purulent and free of saliva was selected for culture. TTA or percutaneous lung aspira­ tion was performed immediately following these collections, and blood cultures were obtained. All samples were immediately taken to the microbiology laboratory by the investigator (P I) and handled in a similar manner. A comparison was made in these patients between the cultural results of expecto­ rated sputum, TTA or lung biopsy, and blood culture. There was an 80 percent agreement between sputum culture and TTA or lung biopsy culture. Forty-one per­ cent of the patients (nine patients) had positive blood cultures, and in all nine cases the same organism was found in the sputum, the TTA, and the blood. In no case did TTA yield the offending organism when the expectorated sputum did not, that is, no “false negative” sputum cultures. The results of our studies suggest that expectorated sputum can be a valuable ad­ junct to diagnosis if collected, transported, gram-stained and processed under strictly controlled conditions. Although such ex­ perience is minimal, it is possible that con­ trolled sputum culture can provide data as reliable as the more invasive technique of trans-tracheal aspiration or percutaneous lung biopsy. If the laboratory and the clini­ cian are unwilling to adhere to these guide­ lines, then they should be prepared to ac­ cept the inevitable consequences, namely sputum culture and Gram stain reports that are unreliable and misleading. In fact, it could be argued that unless strict adher­ ence to an appropriate technique is main­ tained, it would be preferable to abandon