Should Fiberoptic Bronchoscopy Aspirates Be Cultured?·· 2
JOHN G. BARTLETT, JOHN ALEXANDER, JAMES MAYHEW, NADINE SULLIVAN-SIGLER, and SHERWOOD L. GORBACH
SUMMARY ________________________________________________________ The reliability of fiberoptic bronchoscopy as a method to study the bacteriology of the lower respiratory tract was tested. The procedure used was suction aspiration through the inner channel after topical anesthesia with lidocaine. To detect contamination by oropharyngeal bacteria, the aspirates were cultured in patients with no evidence of active infection, comparison was made with results of transtracheal aspiration cultures, and the aspirate was tested for the presence of an oral dye marker. Results with all 3 methods of analysis indicated contamination with oropharyn· geal bacteria that were presumably introduced during instrumentation through the upper airways. An additional factor studied was the effect of topical anesthetics. Analysis of aspirates showed that as much as 96 per cent of the specimen was anesthetic solution. Lidocaine also proved toxic to lower respiratory tract pathogens, although there were significant differences between bacterial species. It was concluded that fiberoptic bronchoscopy as performed in this study does not reliably reflect the bacteriology of the lower respiratory tract.
Introduction Fiberoptic bronchoscopy represents a major advance in the diagnostic armamentarium of pulmonary physicians. The procedure is often used to collect specimens for bacterial culture. For example, bronchoscopy may be performed in patients with bacterial infection of the lower respiratory tract to identify the responsible pathogen. Even when fiberoptic bronchoscopy is undertaken for other indications, the aspirated specimen is often submitted for microbiologic studies. The question posed in this study con(Received in original form january 20, 1976 and in revised form March 26, 1976) 1 From the Infectious Disease Section, Tufts-New England Medical Center, Boston, Mass.; Veterans Administration Hospital, Jamaica Plain, Boston, Mass.; Veterans Administration Hospital, Sepulveda, Calif.; and Department of Medicine, UCLA School of Medicine, Los Angeles, Calif. 2 Requests for reprints should be addressed to Dr. John G. Bartlett, Veterans Administration Hospital, 150 S. Huntington Ave., Boston, Mass. 021!!0.
cerns the reliability of inner channel aspirates for bacterial culture.
Materials and Methods Two potential limitations to the use of fiberoptic bronchoscopy specimens for bacteriologic culture were studied. (1) The possibility of contamination with oropharyngeal flora. This aspect was investigated by culturing inner channel aspirates, and also by analyzing the specimen for the presence of an upper airway dye marker. (2) The possible deleterious effects of topical anesthetics on the recovery of respiratory tract pathogens. To investigate this factor the concentration of lidocaine in aspirates was determined, and time-kill curves were constructed for li· docaine versus selected bacterial strains. Fiberoptic bronchoscopy procedure. The fiberop· tic bronchoscopy procedure was standardized for all studies. The upper airways were anesthetized initially with 10 per cent cocaine swabbed on the tonsillar _pillars and posterior pharynx. Two ml of 2 per cent lidocaine without preservative were then inoculated into the trachea. An uncuffed sterile orotracheal tube lubricated with xylocaine jelly was inserted. The bronchoscope was an Olympus model BS type 5 B2 sterilized by a 20·min soak in povidine-
AMERICAN REVIEW OF RESPIRATORY DISEASE, VOLUME 114, 1976
73
74
BARTLETT, ALEXANDER, MAYHEW, SULLIVAN-SIGLER, AND GORBACH
iodine (Betadine®) followed with a sterile saline wash. (Effectiveness of this sterilization procedure was examined in preliminary tests that showed that thioglycollate broth was consistently sterile after aspiration into the inner channel.) The bronchoscope was inserted through the oral airway and advanced to a mainstem bronchus. Specimens were collected by suction aspiration through the inner channel. Cultures of fiberoptic bronchoscopy aspirates. Cultural studies were performed in 2 groups of pa· tients. The first group had no evidence of active infection and the indication for bronchoscopy was usually a possible bronchogenic neoplasm. These patients were alert, clinically stable, and had not had recent intubation. The second group had suspected lower respiratory tract infections; these patients had previously undergone transtracheal aspiration, and this permitted a comparison of culture results by the 2 methods. The interval between the 2 procedures varied from 4 hours to 7 days; there was no antimicrobial therapy, intubation, or change in clinical status during this interval. In all instances the decision to perform transtracheal aspiration or fiberoptic bronchoscopy was based solely on principles of optimal patient management. Transtracheal aspirates and inner channel aspirates were handled in an identical fashion. The specimens were transferred in sterile containers to the Anaerobic Research Laboratory and placed in the anaerobic chamber for processing within 10 min after collection. Media used for aerobic and facultative strains were blood agar and peptic digest of blood agar for incubation in 10 per cent C0 2 and MacConkey agar for incubation in air. For the recovery of anaerobes the specimen was plated on pre-reduced brucella base agar with 6 per cent sheep's blood and lO ,ug per ml menadione, and on the latter with 75 ,ug per ml of kanamycin and 7.5 ,ug per ml of vancomycin (Clinical Standards, Los Angeles, Calif.). Aerobic plates were passed out of the chamber and read after 24 to 48 hourS' incubation; anaerobic media were retained in the anaerobic chamber at 37o C and held 7 days. Methylene blue studies. Methylene blue was used as a marker of oropharyngeal contamination in another group of patients. A 2 per cent concentration of the dye was sprayed onto the posterior pharynx by nebulizer during a Valsalva maneuver just before insertion of the oral airway. Aspirates were observed for gross visual evidence of the blue marker; they were also analyzed for maximal absorption at 609 nm on a Perkin-Elmer No. 124 double beam spectrophotometer. Lidocaine concentrations of aspirates. Aliquots of 0.1 or 0.05 ml of inner channel aspirates were diluted to a volume of 2.0 ml with distilled water. The sam· ples were then supplemented with 0.25 ml of 5N NaOH, 0.50 ml of aqueous internal standard (equivalent to 500 ,ug), and 1.0 ml of dichlormethane. The completed specimens were vigorously agitated and
then centrifuged (2,300 rpm for 1 min). Replicate 2.0-,ul aliquots were withdrawn directly from the organic phase for chromatographic analysis. Relative calibration curves were constructed by adding 500 ,ug of internal standard (w·N, N-methyltert-butylamino-2,6-dimethylanilide) in 0.5 ml of water to a graduated series of lidocaine concentrations in the range 0.05 per cent to 2.0 per cent. Standard mixtures and biologic specimens were processed in an identical manner. The relative standard devia· tion encountered was 1.4 per cent. Lidocaine concentrations were determined by gas chromatography using a Shimadzu GC4BMPF gas chromatograph equipped with flame ionization de· tectors. Chromatograms were traced by a Shimadzu R-201 recorder and peak areas were interpreted on a Shimadzu ITG-2A electronic digital integrator. The 2.0-,ul samples were injected with a 5-,ul Hamilton syringe onto coiled, silanized Pyrex® columns 2m long by 3 mm internal diameter. The analytic columns contained 3 per cent OV-101 on 80(100 mesh GasChrum Q (Applied Science, State College, Pa.). Peak identity was verified by chromatography on an alternate phase (3 per cent OV-17). Routine analysis conditions included: injection and detector blocks, 225° C; oven isothermal at 185° C; nitrogen, hydrogen, and air flows at 30, 30, and 300 ml per min, respectively. Input resistance of the electrometer was set at 109 fl, and the output range at 8 X I0-2V. The minimal reproducible limit under these operating conditions was 80 ng of lidocaine per 2 ,ul of CH 2Cl 2. Antibacterial activity of lidocaine and lidocaine with methylparaben preservative. The antibacterial activity of topical anesthetics was tested by constructing time-kill curves for 8 bacterial strains. Substances tested were I per cent lidocaine without preservative (Astra Pharmaceutical Products, Worcester, Mass), 1 per cent lidocaine with 0.05 per cent methylparaben preservative (Astra Pharmaceutical Products) and a
TABLE 1 CULTURE RESULTS OF FIBEROPTIC BRONCHOSCOPY ASPIRATES IN 16 PATIENTS WITHOUT EVIDENCE OF PULMONARY INFECTION Kind of Strain Aerobic and facultative Streptococci Neisseria sp.
Staphylococcus epidermidis K /ebsiel/a sp. Coliforms (other)
Pseudomonas sp.
Miscellaneous Anaerobic Peptostreptococcus Bacteroides sp. Veillonella Miscellaneous
No. of Isolates 17 7 5 5 13 4
6 7
5 4
10
SHOULD FIBEROPTIC BRONCHOSCOPY ASPIRATES BE CULTURED?
lactated Ringer's control. (The 1 per cent lidocaine concentration was used because our previous studies showed this was the approximate mean concentration of the topical anesthetic in fiberoptic bronchoscopy aspirates). Test organisms consisted of 8 recently isolated strains of bacteria obtained by transtracheal aspira. tion in patients with lower respiratory tract infections. These included 2 strains each of Streptococcus pneumoniae, Haemophilus influenzae, Bacteroides melaninogenicus, and Fusobacterium nucleatum. The test strains were grown overnight and diluted to a concentration of approximately 108 to 109 organisms per mi. A 0.8-ml aliquot of the broth growth of each organism was added to 7.2 ml of the test agents to achieve a final concentration of 1 per cent lidocaine. The final inoculum of bacteria, as verified by plate count, was 107 to 108 per ml, which approximates the number encountered in bronchial secretions of patients with pulmonary infections. The solutions were left at room temperature in air and cultured quantitatively in triplicate at 0, 10, 20, 30, 60, and 120 min. Quantitative cultures at each
75
sampling interval were made by adding 0.5 ml of the test solution to 4.5 ml of dilution salts of the Virginia Polytechnic Institute (I) followed by 5 additional 10-fold dilutions. Micropipettes were used to deliver 0.025 ml of each dilution onto appropriate media: peptic digest of blood agar incubated in 10 per cent C0 2 for H. influenzae, blood agar incubated in 10 per cent C0 2 for S. pneumoniae, and brucella base agar with sheep's blood and menadione incubated in the anaerobic chamber for anaerobes. To detect low concentrations the remaining portion of the 120-min specimen was expressed through a 0.45-l'm Millipore filter, washed with 20 ml of lactated Ringer's solution, and the filter was placed on appropriate media.
Results
Cultures of inner channel aspirates. Cultures of inner channel aspirates obtained from 16 patients without evidence of lower respiratory tract infection yielded 85 isolates, including 59 aerobic or facultative strains and 26 anaerobic
TABLE 2 COMPARISON OF CULTURE RESULTS OF TRANSTRACHEAL AND INNER CHANNEL ASPIRATES Patient
Final Diagnosis
Transtracheal Aspirate
93
Pulmonary fibrosis
No growth
2
Pulmonary fibrosis
a-hemolytic streptococcus 1 +
Inner Channel Aspirate
Klebsiella
sp. 2+ Enterococcus 2+
a-hemolytic streptococcus 1+
Staphylococcus epidermidis 2+ Peptostreptococcus 2+ Veillonella 1+
3
Pneumonia
Escherichia coli 1+ Streptococcus pneumoniae 3+
E. coli 2+ Neisseria sp 1 + a-hemolytic streptococci 1+
Proteus mirabilis 1 + 4
Bronchogenic carcinoma
S. epidermidis 1 + Diphtheroids 2+
S. epidermidis 1 + Pseudomonas aeruginosa 2+ Neisseria sp. 2+ a-hemolytic streptococci 1 +
5
Infected pulmonary cyst
Haemophilus influenzae 2+
H. influenzae 2+ . Staphylococcus aureus 2+ a-hemolytic streptococcus 2+ Neisseria sp. 1+ Lactobacillus 1 + Bifidobacteria 1 + Veillonella 2+
6
Primary lung adscess
Peptostreptococcus intermedius 4+
P. intermedius 3+ 0/:hemolytic streptococcus 1+ S. epidermidis 2+
7
Aspiration pneumonia
P. intermedius 3+
P. intermedius 1 +
HB-1 1+
HB-1 1+
Klebsiella sp. 3+ S. epidermidis 2+ Neisseria sp. 1+ Candida albicans 1 +
76
BARTLETT, ALEXANDER, MAYHEW, SULLIVAN.SIGLER, AND GORBACH
strains (table 1). This represents an average of more than 5 bacterial species per aspirate. The dominant isolates were facultative streptococci (17 isolates), coliforms (18), Neisseria sp. (7), peptostreptococcus (7), Bacteroides sp. (5), Staphylococcus epidermidis (5), and veillonella (4). Cultures were obtained in 7 additional patients who had companion transtracheal aspirates (table 2). Fiberoptic bronchoscopy as· pirates usually yielded the same bacterial spe· cies recovered in the transtracheal aspirates, but 2 or more additional species were also recovered from the bronchoscopy aspirate in every in· stance. For example, H. influenzae was found in pure culture by transtracheal aspiration in Patient 5, but it was present with 6 other bacteria in the bronchoscopy sample. In Patient 7, 2 potential pathogens were cultured from the transtracheal aspirate, whereas 6 microorgan· isms, including the original 2, were found in the inner channel specimen. Methylene blue marker studies. Ten patients had methylene blue marker studies. Our previous observations with the 2 per cent dye solu· tion showed that it could still be detected in water by visual inspection at a 1:10,000 dilution. Results with the inner channel aspirates showed the blue marker was readily apparent on visual inspection in 8 specimens. In each instance this observation was confirmed by spectrophotometric analysis. Dye could not be detected by either method in 2 specimens. Lidocaine concentrations. Concentrations of 13 fiberoptic bronchoscopy aspixylocaine rates were highly variable, ranging from 0.05 per cent to 1.91 per cent; the mean value for
m
all specimens was 1.02 per cent. Because the original xylocaine solution was a 2 per cent concentration, the amount of the final aspirate represented by topical anesthetic ranged from 3 per cent to 96 per cent (mean, 53 per cent) of the total specimen. Antibacterial properties of topical anesthetics. Sequential samplings of the 8 bacterial strains showed marked species variations in the antibacterial effect of the topical anesthetics (table 3). Lactated Ringer's control solutions showed that more than 90 per cent of the original inoculum could be· recovered throughout the 2-hour sampling; an exception was B. melaninogenicus, which showed no substantial loss until 60 min and 120 min, when recoveries were 20 per cent and 10 per cent, respectively. The topical anesthetics proved minimally inhibitory to S. pneumoniae. At 2 hours there was a loss of approximately 50 per cent with both 1 per cent lidocaine and l per cent lidocaine with methylparaben. The anesthetics proved more toxic to H. influenzae and F. nucleatum, which showed 28 to 50 per cent recovery compared with lactated Ringer's control at 20 min and 0.05 to 13 per cent recoveries at 2 hours. As anticipated, the preparation containing preservative showed a greater inhibitory effect. The most pronounced inhibition was noted with B. melaninogenicus. Using lidocaine, the recovery compared with lactated Ringer's control solution was 20 per cent at lO min, 1.6 per cent at 30 min, 0.4 per cent at 60 min, and 0.006 per cent at 120 min. With lidocaine and preservative there was no broth growth in the 60-min specimen; the Millipore filter of the 120-min specimen also failed to yield growth.
TABLE 3 ANTIBACTERIAL ACTIVITY OF TOPICAL ANESTHETICS Sampling Time
(min) Test Solution •
Bacteria
Streptococcus pneumoniae Haemophilus influenzae Fusobacterium nucleatum Bacteroides me/aninogenicus •L
= lidocaine;
LM
= lidocaine plus
L LM L LM L LM L LM
10 115t 100 50 100 66 80 20 6
20 100 107 50 50 33 28 12 6
30
60
120
107 60 50 40 33 28 1.6 3
100 60 50 5 13 8 0.4 0
54 53 10 0.05 13 1 0.006
0
methylparaben.
tPer cent recovered compared with lactated Ringer's control solution. Each value represents mean value of 2 strains tested in triplicate.
SHOULD FIBEROPTIC BRONCHOSCOPY ASPIRATES BE CULTURED?
Discussion
Several studies have shown that expectorated sputum is an unreliable indicator of the bacteriology of the lower respiratory tract (2-6). Fiberoptic bronchoscopy permits suction aspiration directly from the infected site and would appear to provide a superior culture source. Our results indicate that inner channel aspirates are contaminated by oropharyngeal bacteria that are presumably introduced during instrumentation through the upper airways. Cultures obtained from 16 patients without evidence of active infection yielded an average of more than 5 bacterial species per aspirate. The major isolates in the bronchoscopy aspirates were a-hemolytic streptococci, Neisseria sp., and peptostreptococcus, all normal inhabitants of the oropharynx. There were also 22 strains of coliforms and pseudomonads. These findings are consistent with the frequency with which such organisms colonize the upper airways in hospitalized patients (7) and represent an important source of possibly misleading results. Seven additional patients had transtracheal aspiration as well as the fiberoptic bronchoscopy. The former procedure bypasses the upper airways and has been established as a reliable method of documenting the bacteriology of lower respiratory tract infections (5, 6, 8-11). Comparison of the 2 culture sources generally showed a poor correlation because the fiberoptic bronchoscopy aspirates yielded multiple strains not present in the transtracheal aspirates. Another approach to determine oropharyngeal contamination was a methylene blue marker nebulized into the oropharynx before the bronchoscopy procedure. The advantage of this technique is that the contamination may be readily apparent by visual inspection. The presence of the indicator can be further documented by spectrophotometric analysis, but this is generally unnecessary for dyes in the visual range. We have previously used this technique with transtracheal aspirates and found that the specimen failed to reveal the dye in 25 consecutive patients. On the other hand, the methylene blue marker was readily apparent in 8 of 10 aspirates obtained by fiberoptic bronchoscopy. Topical anesthetics are an additional variable that can affect culture results. Analysis of fiberoptic bronchoscopy aspirates indicates that these solutions may constitute up to 96 per cent of the final specimen. It should be noted that
77
our standardized procedure used just 2 ml of lidocaine inoculated into the trachea. Many endoscopists use much larger volumes, and the routine use of as much as 20 to 40 ml has been reported (11). Thus, our results may be quite conservative on the basis of current practice elsewhere. In addition to simple dilution of bronchial secretions, topical anesthetics have a direct toxic effect on bacteria. This has been previously demonstrated for M. tuberculosis (12, 13), Staphylococcus aureus (13), Candida albicans (13), and most gram-negative bacilli other than Pseudomonas aeruginosa (14). Lidocaine was selected for in vitro antibacterial studies in the present report because of its frequent use and because this agent appears to be the least toxic to bacteria of topical anesthetics according to previous reports. Lidocaine with the preservative, methylparaben, was also tested because, despite obvious potential deleterious effects, this preparation is often used in clinical practice. Eight strains, including 4 relatively common lower respiratory tract pathogens, were tested: S. pneumoniae, H. influenaze, B. melaninogenicus, and F. nucleatum. Curiously, these organisms have seldom been examined with regard to inhibition by topical anesthetics despite their prevalence in lobar pneumonia, exacerbations of bronchitis, aspiration pneumonia, and lung abscess. Time-kill curves were constructed to determine the duration of bacteriaanesthetic contact that might alter culture results, thus simulating variable periods of delay in processing of specimens. Results showed major differences between bacteria in susceptibility to anesthetic with and without preservative. S. pneumoniae proved relatively resistant, with minimal loss during the 2-hour sampling interval. H. influenzae and F. nucleatum were more susceptible but this effect was not marked until after 30 min or more of contact. The major toxicity was with B. melaninogenicus, and 80 per cent to 94 per cent of the original inocula were no longer cultivable after just 10 min of exposure to lidocaine or lidocaine with preservative. The findings in this study indicate that cultures of fiberoptic bronchoscopy aspirates do not accurately reflect lower respiratory tract bacteriology. These conclusions are restricted to the procedural format used in this report, i.e., suction aspiration from the inner channel after intratracheal inoculation of lidocaine. It is con-
78
BARTLETT, ALEXANDER, MAYHEW, SULLIVAN-SIGLER, AND GORBACH
ceivable that oropharyngeal contamination could be avoided by utilizing polyethylene catheters containing retractable loops inserted through the inner channel and beyond the tip of the bronchoscope (15). An analogous device was previously found necessary for reliable bacteriologic studies with the Jackson brondlOscope (16). Others have used a tube within the inner channel of the fiberoptic bronchoscope in combination with the head-down position to prevent oropharyngeal contamination (17). The procedure used in the present study was select~d to reproduce the methods most frequently used in clinical practice. References 1. Anaerobe Laboratory Manual, L. V. Holdeman and W. E. C. Moore, ed., Virginia Polytechnic Institute and State University, Blacksburg, 1975, p.l26. 2. Potter, R. T., Rotman, F., Fernandez, F., McNeill, T. M., and Chamberlain, J. M.: Bacteriology of the lower respiratory tract, Am Rev Respir Dis, 1968, 97, 1051. 3. Pecora, D. V., and Yegian, D.: Bacteriology of the lower respiratory tract in health and chronic disease, N Eng! J Med, 1968, 268, 71. 4. Barrett-Conner, E.: The nonvalue of sputum culture in the diagnosis of pneumococcal pneumonia, Am Rev Respir Dis, 1970,103,845. 5. Hoeprich, P. D.: Etiologic diagnosis of lower respiratory tract infections, Calif Med, 1970, 112, 1. 6. Kalinske, R. W., Parker, R. H., Brandt, E., and Hoeprich, P. D.: Diagnostic usefulness and safety of transtracheal aspiration, N Engl J Med, 1967, 276,604. 7. Johanson, W. P., Pierce, A. K., and Sanford,
8.
9.
10.
II.
12.
13.
14.
15.
16.
17.
J. P.: Changing pharyngeal bacterial flora of hospitalized patients. Emergence of gram-negative bacilli, N Eng! J Med,l969, 281, 1137. Ries, K., Levison, M. E., and Kaye, D.: Transtracheal aspiration in pulmonary infection, Arch Intern Med, 1974, 133,453. Hahn, H. H., and Beaty, H. N.: Transtracheal aspiration in the evaluation of patients with pneumonia, Ann Intern Med, 1070, 72, 183. Bartlett, J. G., Rosenblatt, J. E., and Finegold, S. M.: Percutaneous transtracheal aspiration in the diagnosis of anaerobic pulmonary infection, Ann Intern Med, 1973, 79, 535. Patterson, J. R., Blaschke, T. F., Hunt, K. K., Jr., and Mellin, P. J.: Lidocaine blood concentrations during fiberoptic bronchoscopy, Am Rev Resp Dis, 1975,112,53. Conte, B. A., and LaForet, E. G.: The role of the topical anesthetic agent in modifying bacteriologic data obtained by bronchoscopy, N Eng! J Med, 1962,267,957. Erlich, H.: Bacteriologic studies and effects of anesthetic solutions on bronchial secretions during bronchoscopy, Am Rev Respir Dis, 1961, 84, 414. Schmidt, R. M., and Rosenkranz, H. S.: Antimicrobial activity of local anesthetics: Lidocaine and procaine, J Infect Dis, 1970,121,597. Wanner, A., Amikam, B., Robinson, M. J., and Sackner, M. A.: Comparison between the bacteriologic flora of different segments of the airways, Respiration, 1973, 30, 561. Lees, A. W., and McNaught, W: Bacteriology of lower respiratory tract secretions, sputum, and upper-respiratory tract secretions in "normals" and chronic bronchitics, Lancet, 1959,2, 1112. Jordan, G. W., Krajden, S. F., Hoeprich, P. D., Wong, G. A., Pierce, T. H., and Rausch, D. C.: Trimethoprim-sulfamethoxazole in chronic bronchitis, Can Med Assoc J, 1975,112,915.
JOHN G. BARTLETT, JOHN ALEXANDER, JAMES MAYHEW, NADINE SULLIVAN-SIGLER, and SHERWOOD L. GORBACH
SUMMARY ________________________________________________________ The reliability of fiberoptic bronchoscopy as a method to study the bacteriology of the lower respiratory tract was tested. The procedure used was suction aspiration through the inner channel after topical anesthesia with lidocaine. To detect contamination by oropharyngeal bacteria, the aspirates were cultured in patients with no evidence of active infection, comparison was made with results of transtracheal aspiration cultures, and the aspirate was tested for the presence of an oral dye marker. Results with all 3 methods of analysis indicated contamination with oropharyn· geal bacteria that were presumably introduced during instrumentation through the upper airways. An additional factor studied was the effect of topical anesthetics. Analysis of aspirates showed that as much as 96 per cent of the specimen was anesthetic solution. Lidocaine also proved toxic to lower respiratory tract pathogens, although there were significant differences between bacterial species. It was concluded that fiberoptic bronchoscopy as performed in this study does not reliably reflect the bacteriology of the lower respiratory tract.
Introduction Fiberoptic bronchoscopy represents a major advance in the diagnostic armamentarium of pulmonary physicians. The procedure is often used to collect specimens for bacterial culture. For example, bronchoscopy may be performed in patients with bacterial infection of the lower respiratory tract to identify the responsible pathogen. Even when fiberoptic bronchoscopy is undertaken for other indications, the aspirated specimen is often submitted for microbiologic studies. The question posed in this study con(Received in original form january 20, 1976 and in revised form March 26, 1976) 1 From the Infectious Disease Section, Tufts-New England Medical Center, Boston, Mass.; Veterans Administration Hospital, Jamaica Plain, Boston, Mass.; Veterans Administration Hospital, Sepulveda, Calif.; and Department of Medicine, UCLA School of Medicine, Los Angeles, Calif. 2 Requests for reprints should be addressed to Dr. John G. Bartlett, Veterans Administration Hospital, 150 S. Huntington Ave., Boston, Mass. 021!!0.
cerns the reliability of inner channel aspirates for bacterial culture.
Materials and Methods Two potential limitations to the use of fiberoptic bronchoscopy specimens for bacteriologic culture were studied. (1) The possibility of contamination with oropharyngeal flora. This aspect was investigated by culturing inner channel aspirates, and also by analyzing the specimen for the presence of an upper airway dye marker. (2) The possible deleterious effects of topical anesthetics on the recovery of respiratory tract pathogens. To investigate this factor the concentration of lidocaine in aspirates was determined, and time-kill curves were constructed for li· docaine versus selected bacterial strains. Fiberoptic bronchoscopy procedure. The fiberop· tic bronchoscopy procedure was standardized for all studies. The upper airways were anesthetized initially with 10 per cent cocaine swabbed on the tonsillar _pillars and posterior pharynx. Two ml of 2 per cent lidocaine without preservative were then inoculated into the trachea. An uncuffed sterile orotracheal tube lubricated with xylocaine jelly was inserted. The bronchoscope was an Olympus model BS type 5 B2 sterilized by a 20·min soak in povidine-
AMERICAN REVIEW OF RESPIRATORY DISEASE, VOLUME 114, 1976
73
74
BARTLETT, ALEXANDER, MAYHEW, SULLIVAN-SIGLER, AND GORBACH
iodine (Betadine®) followed with a sterile saline wash. (Effectiveness of this sterilization procedure was examined in preliminary tests that showed that thioglycollate broth was consistently sterile after aspiration into the inner channel.) The bronchoscope was inserted through the oral airway and advanced to a mainstem bronchus. Specimens were collected by suction aspiration through the inner channel. Cultures of fiberoptic bronchoscopy aspirates. Cultural studies were performed in 2 groups of pa· tients. The first group had no evidence of active infection and the indication for bronchoscopy was usually a possible bronchogenic neoplasm. These patients were alert, clinically stable, and had not had recent intubation. The second group had suspected lower respiratory tract infections; these patients had previously undergone transtracheal aspiration, and this permitted a comparison of culture results by the 2 methods. The interval between the 2 procedures varied from 4 hours to 7 days; there was no antimicrobial therapy, intubation, or change in clinical status during this interval. In all instances the decision to perform transtracheal aspiration or fiberoptic bronchoscopy was based solely on principles of optimal patient management. Transtracheal aspirates and inner channel aspirates were handled in an identical fashion. The specimens were transferred in sterile containers to the Anaerobic Research Laboratory and placed in the anaerobic chamber for processing within 10 min after collection. Media used for aerobic and facultative strains were blood agar and peptic digest of blood agar for incubation in 10 per cent C0 2 and MacConkey agar for incubation in air. For the recovery of anaerobes the specimen was plated on pre-reduced brucella base agar with 6 per cent sheep's blood and lO ,ug per ml menadione, and on the latter with 75 ,ug per ml of kanamycin and 7.5 ,ug per ml of vancomycin (Clinical Standards, Los Angeles, Calif.). Aerobic plates were passed out of the chamber and read after 24 to 48 hourS' incubation; anaerobic media were retained in the anaerobic chamber at 37o C and held 7 days. Methylene blue studies. Methylene blue was used as a marker of oropharyngeal contamination in another group of patients. A 2 per cent concentration of the dye was sprayed onto the posterior pharynx by nebulizer during a Valsalva maneuver just before insertion of the oral airway. Aspirates were observed for gross visual evidence of the blue marker; they were also analyzed for maximal absorption at 609 nm on a Perkin-Elmer No. 124 double beam spectrophotometer. Lidocaine concentrations of aspirates. Aliquots of 0.1 or 0.05 ml of inner channel aspirates were diluted to a volume of 2.0 ml with distilled water. The sam· ples were then supplemented with 0.25 ml of 5N NaOH, 0.50 ml of aqueous internal standard (equivalent to 500 ,ug), and 1.0 ml of dichlormethane. The completed specimens were vigorously agitated and
then centrifuged (2,300 rpm for 1 min). Replicate 2.0-,ul aliquots were withdrawn directly from the organic phase for chromatographic analysis. Relative calibration curves were constructed by adding 500 ,ug of internal standard (w·N, N-methyltert-butylamino-2,6-dimethylanilide) in 0.5 ml of water to a graduated series of lidocaine concentrations in the range 0.05 per cent to 2.0 per cent. Standard mixtures and biologic specimens were processed in an identical manner. The relative standard devia· tion encountered was 1.4 per cent. Lidocaine concentrations were determined by gas chromatography using a Shimadzu GC4BMPF gas chromatograph equipped with flame ionization de· tectors. Chromatograms were traced by a Shimadzu R-201 recorder and peak areas were interpreted on a Shimadzu ITG-2A electronic digital integrator. The 2.0-,ul samples were injected with a 5-,ul Hamilton syringe onto coiled, silanized Pyrex® columns 2m long by 3 mm internal diameter. The analytic columns contained 3 per cent OV-101 on 80(100 mesh GasChrum Q (Applied Science, State College, Pa.). Peak identity was verified by chromatography on an alternate phase (3 per cent OV-17). Routine analysis conditions included: injection and detector blocks, 225° C; oven isothermal at 185° C; nitrogen, hydrogen, and air flows at 30, 30, and 300 ml per min, respectively. Input resistance of the electrometer was set at 109 fl, and the output range at 8 X I0-2V. The minimal reproducible limit under these operating conditions was 80 ng of lidocaine per 2 ,ul of CH 2Cl 2. Antibacterial activity of lidocaine and lidocaine with methylparaben preservative. The antibacterial activity of topical anesthetics was tested by constructing time-kill curves for 8 bacterial strains. Substances tested were I per cent lidocaine without preservative (Astra Pharmaceutical Products, Worcester, Mass), 1 per cent lidocaine with 0.05 per cent methylparaben preservative (Astra Pharmaceutical Products) and a
TABLE 1 CULTURE RESULTS OF FIBEROPTIC BRONCHOSCOPY ASPIRATES IN 16 PATIENTS WITHOUT EVIDENCE OF PULMONARY INFECTION Kind of Strain Aerobic and facultative Streptococci Neisseria sp.
Staphylococcus epidermidis K /ebsiel/a sp. Coliforms (other)
Pseudomonas sp.
Miscellaneous Anaerobic Peptostreptococcus Bacteroides sp. Veillonella Miscellaneous
No. of Isolates 17 7 5 5 13 4
6 7
5 4
10
SHOULD FIBEROPTIC BRONCHOSCOPY ASPIRATES BE CULTURED?
lactated Ringer's control. (The 1 per cent lidocaine concentration was used because our previous studies showed this was the approximate mean concentration of the topical anesthetic in fiberoptic bronchoscopy aspirates). Test organisms consisted of 8 recently isolated strains of bacteria obtained by transtracheal aspira. tion in patients with lower respiratory tract infections. These included 2 strains each of Streptococcus pneumoniae, Haemophilus influenzae, Bacteroides melaninogenicus, and Fusobacterium nucleatum. The test strains were grown overnight and diluted to a concentration of approximately 108 to 109 organisms per mi. A 0.8-ml aliquot of the broth growth of each organism was added to 7.2 ml of the test agents to achieve a final concentration of 1 per cent lidocaine. The final inoculum of bacteria, as verified by plate count, was 107 to 108 per ml, which approximates the number encountered in bronchial secretions of patients with pulmonary infections. The solutions were left at room temperature in air and cultured quantitatively in triplicate at 0, 10, 20, 30, 60, and 120 min. Quantitative cultures at each
75
sampling interval were made by adding 0.5 ml of the test solution to 4.5 ml of dilution salts of the Virginia Polytechnic Institute (I) followed by 5 additional 10-fold dilutions. Micropipettes were used to deliver 0.025 ml of each dilution onto appropriate media: peptic digest of blood agar incubated in 10 per cent C0 2 for H. influenzae, blood agar incubated in 10 per cent C0 2 for S. pneumoniae, and brucella base agar with sheep's blood and menadione incubated in the anaerobic chamber for anaerobes. To detect low concentrations the remaining portion of the 120-min specimen was expressed through a 0.45-l'm Millipore filter, washed with 20 ml of lactated Ringer's solution, and the filter was placed on appropriate media.
Results
Cultures of inner channel aspirates. Cultures of inner channel aspirates obtained from 16 patients without evidence of lower respiratory tract infection yielded 85 isolates, including 59 aerobic or facultative strains and 26 anaerobic
TABLE 2 COMPARISON OF CULTURE RESULTS OF TRANSTRACHEAL AND INNER CHANNEL ASPIRATES Patient
Final Diagnosis
Transtracheal Aspirate
93
Pulmonary fibrosis
No growth
2
Pulmonary fibrosis
a-hemolytic streptococcus 1 +
Inner Channel Aspirate
Klebsiella
sp. 2+ Enterococcus 2+
a-hemolytic streptococcus 1+
Staphylococcus epidermidis 2+ Peptostreptococcus 2+ Veillonella 1+
3
Pneumonia
Escherichia coli 1+ Streptococcus pneumoniae 3+
E. coli 2+ Neisseria sp 1 + a-hemolytic streptococci 1+
Proteus mirabilis 1 + 4
Bronchogenic carcinoma
S. epidermidis 1 + Diphtheroids 2+
S. epidermidis 1 + Pseudomonas aeruginosa 2+ Neisseria sp. 2+ a-hemolytic streptococci 1 +
5
Infected pulmonary cyst
Haemophilus influenzae 2+
H. influenzae 2+ . Staphylococcus aureus 2+ a-hemolytic streptococcus 2+ Neisseria sp. 1+ Lactobacillus 1 + Bifidobacteria 1 + Veillonella 2+
6
Primary lung adscess
Peptostreptococcus intermedius 4+
P. intermedius 3+ 0/:hemolytic streptococcus 1+ S. epidermidis 2+
7
Aspiration pneumonia
P. intermedius 3+
P. intermedius 1 +
HB-1 1+
HB-1 1+
Klebsiella sp. 3+ S. epidermidis 2+ Neisseria sp. 1+ Candida albicans 1 +
76
BARTLETT, ALEXANDER, MAYHEW, SULLIVAN.SIGLER, AND GORBACH
strains (table 1). This represents an average of more than 5 bacterial species per aspirate. The dominant isolates were facultative streptococci (17 isolates), coliforms (18), Neisseria sp. (7), peptostreptococcus (7), Bacteroides sp. (5), Staphylococcus epidermidis (5), and veillonella (4). Cultures were obtained in 7 additional patients who had companion transtracheal aspirates (table 2). Fiberoptic bronchoscopy as· pirates usually yielded the same bacterial spe· cies recovered in the transtracheal aspirates, but 2 or more additional species were also recovered from the bronchoscopy aspirate in every in· stance. For example, H. influenzae was found in pure culture by transtracheal aspiration in Patient 5, but it was present with 6 other bacteria in the bronchoscopy sample. In Patient 7, 2 potential pathogens were cultured from the transtracheal aspirate, whereas 6 microorgan· isms, including the original 2, were found in the inner channel specimen. Methylene blue marker studies. Ten patients had methylene blue marker studies. Our previous observations with the 2 per cent dye solu· tion showed that it could still be detected in water by visual inspection at a 1:10,000 dilution. Results with the inner channel aspirates showed the blue marker was readily apparent on visual inspection in 8 specimens. In each instance this observation was confirmed by spectrophotometric analysis. Dye could not be detected by either method in 2 specimens. Lidocaine concentrations. Concentrations of 13 fiberoptic bronchoscopy aspixylocaine rates were highly variable, ranging from 0.05 per cent to 1.91 per cent; the mean value for
m
all specimens was 1.02 per cent. Because the original xylocaine solution was a 2 per cent concentration, the amount of the final aspirate represented by topical anesthetic ranged from 3 per cent to 96 per cent (mean, 53 per cent) of the total specimen. Antibacterial properties of topical anesthetics. Sequential samplings of the 8 bacterial strains showed marked species variations in the antibacterial effect of the topical anesthetics (table 3). Lactated Ringer's control solutions showed that more than 90 per cent of the original inoculum could be· recovered throughout the 2-hour sampling; an exception was B. melaninogenicus, which showed no substantial loss until 60 min and 120 min, when recoveries were 20 per cent and 10 per cent, respectively. The topical anesthetics proved minimally inhibitory to S. pneumoniae. At 2 hours there was a loss of approximately 50 per cent with both 1 per cent lidocaine and l per cent lidocaine with methylparaben. The anesthetics proved more toxic to H. influenzae and F. nucleatum, which showed 28 to 50 per cent recovery compared with lactated Ringer's control at 20 min and 0.05 to 13 per cent recoveries at 2 hours. As anticipated, the preparation containing preservative showed a greater inhibitory effect. The most pronounced inhibition was noted with B. melaninogenicus. Using lidocaine, the recovery compared with lactated Ringer's control solution was 20 per cent at lO min, 1.6 per cent at 30 min, 0.4 per cent at 60 min, and 0.006 per cent at 120 min. With lidocaine and preservative there was no broth growth in the 60-min specimen; the Millipore filter of the 120-min specimen also failed to yield growth.
TABLE 3 ANTIBACTERIAL ACTIVITY OF TOPICAL ANESTHETICS Sampling Time
(min) Test Solution •
Bacteria
Streptococcus pneumoniae Haemophilus influenzae Fusobacterium nucleatum Bacteroides me/aninogenicus •L
= lidocaine;
LM
= lidocaine plus
L LM L LM L LM L LM
10 115t 100 50 100 66 80 20 6
20 100 107 50 50 33 28 12 6
30
60
120
107 60 50 40 33 28 1.6 3
100 60 50 5 13 8 0.4 0
54 53 10 0.05 13 1 0.006
0
methylparaben.
tPer cent recovered compared with lactated Ringer's control solution. Each value represents mean value of 2 strains tested in triplicate.
SHOULD FIBEROPTIC BRONCHOSCOPY ASPIRATES BE CULTURED?
Discussion
Several studies have shown that expectorated sputum is an unreliable indicator of the bacteriology of the lower respiratory tract (2-6). Fiberoptic bronchoscopy permits suction aspiration directly from the infected site and would appear to provide a superior culture source. Our results indicate that inner channel aspirates are contaminated by oropharyngeal bacteria that are presumably introduced during instrumentation through the upper airways. Cultures obtained from 16 patients without evidence of active infection yielded an average of more than 5 bacterial species per aspirate. The major isolates in the bronchoscopy aspirates were a-hemolytic streptococci, Neisseria sp., and peptostreptococcus, all normal inhabitants of the oropharynx. There were also 22 strains of coliforms and pseudomonads. These findings are consistent with the frequency with which such organisms colonize the upper airways in hospitalized patients (7) and represent an important source of possibly misleading results. Seven additional patients had transtracheal aspiration as well as the fiberoptic bronchoscopy. The former procedure bypasses the upper airways and has been established as a reliable method of documenting the bacteriology of lower respiratory tract infections (5, 6, 8-11). Comparison of the 2 culture sources generally showed a poor correlation because the fiberoptic bronchoscopy aspirates yielded multiple strains not present in the transtracheal aspirates. Another approach to determine oropharyngeal contamination was a methylene blue marker nebulized into the oropharynx before the bronchoscopy procedure. The advantage of this technique is that the contamination may be readily apparent by visual inspection. The presence of the indicator can be further documented by spectrophotometric analysis, but this is generally unnecessary for dyes in the visual range. We have previously used this technique with transtracheal aspirates and found that the specimen failed to reveal the dye in 25 consecutive patients. On the other hand, the methylene blue marker was readily apparent in 8 of 10 aspirates obtained by fiberoptic bronchoscopy. Topical anesthetics are an additional variable that can affect culture results. Analysis of fiberoptic bronchoscopy aspirates indicates that these solutions may constitute up to 96 per cent of the final specimen. It should be noted that
77
our standardized procedure used just 2 ml of lidocaine inoculated into the trachea. Many endoscopists use much larger volumes, and the routine use of as much as 20 to 40 ml has been reported (11). Thus, our results may be quite conservative on the basis of current practice elsewhere. In addition to simple dilution of bronchial secretions, topical anesthetics have a direct toxic effect on bacteria. This has been previously demonstrated for M. tuberculosis (12, 13), Staphylococcus aureus (13), Candida albicans (13), and most gram-negative bacilli other than Pseudomonas aeruginosa (14). Lidocaine was selected for in vitro antibacterial studies in the present report because of its frequent use and because this agent appears to be the least toxic to bacteria of topical anesthetics according to previous reports. Lidocaine with the preservative, methylparaben, was also tested because, despite obvious potential deleterious effects, this preparation is often used in clinical practice. Eight strains, including 4 relatively common lower respiratory tract pathogens, were tested: S. pneumoniae, H. influenaze, B. melaninogenicus, and F. nucleatum. Curiously, these organisms have seldom been examined with regard to inhibition by topical anesthetics despite their prevalence in lobar pneumonia, exacerbations of bronchitis, aspiration pneumonia, and lung abscess. Time-kill curves were constructed to determine the duration of bacteriaanesthetic contact that might alter culture results, thus simulating variable periods of delay in processing of specimens. Results showed major differences between bacteria in susceptibility to anesthetic with and without preservative. S. pneumoniae proved relatively resistant, with minimal loss during the 2-hour sampling interval. H. influenzae and F. nucleatum were more susceptible but this effect was not marked until after 30 min or more of contact. The major toxicity was with B. melaninogenicus, and 80 per cent to 94 per cent of the original inocula were no longer cultivable after just 10 min of exposure to lidocaine or lidocaine with preservative. The findings in this study indicate that cultures of fiberoptic bronchoscopy aspirates do not accurately reflect lower respiratory tract bacteriology. These conclusions are restricted to the procedural format used in this report, i.e., suction aspiration from the inner channel after intratracheal inoculation of lidocaine. It is con-
78
BARTLETT, ALEXANDER, MAYHEW, SULLIVAN-SIGLER, AND GORBACH
ceivable that oropharyngeal contamination could be avoided by utilizing polyethylene catheters containing retractable loops inserted through the inner channel and beyond the tip of the bronchoscope (15). An analogous device was previously found necessary for reliable bacteriologic studies with the Jackson brondlOscope (16). Others have used a tube within the inner channel of the fiberoptic bronchoscope in combination with the head-down position to prevent oropharyngeal contamination (17). The procedure used in the present study was select~d to reproduce the methods most frequently used in clinical practice. References 1. Anaerobe Laboratory Manual, L. V. Holdeman and W. E. C. Moore, ed., Virginia Polytechnic Institute and State University, Blacksburg, 1975, p.l26. 2. Potter, R. T., Rotman, F., Fernandez, F., McNeill, T. M., and Chamberlain, J. M.: Bacteriology of the lower respiratory tract, Am Rev Respir Dis, 1968, 97, 1051. 3. Pecora, D. V., and Yegian, D.: Bacteriology of the lower respiratory tract in health and chronic disease, N Eng! J Med, 1968, 268, 71. 4. Barrett-Conner, E.: The nonvalue of sputum culture in the diagnosis of pneumococcal pneumonia, Am Rev Respir Dis, 1970,103,845. 5. Hoeprich, P. D.: Etiologic diagnosis of lower respiratory tract infections, Calif Med, 1970, 112, 1. 6. Kalinske, R. W., Parker, R. H., Brandt, E., and Hoeprich, P. D.: Diagnostic usefulness and safety of transtracheal aspiration, N Engl J Med, 1967, 276,604. 7. Johanson, W. P., Pierce, A. K., and Sanford,
8.
9.
10.
II.
12.
13.
14.
15.
16.
17.
J. P.: Changing pharyngeal bacterial flora of hospitalized patients. Emergence of gram-negative bacilli, N Eng! J Med,l969, 281, 1137. Ries, K., Levison, M. E., and Kaye, D.: Transtracheal aspiration in pulmonary infection, Arch Intern Med, 1974, 133,453. Hahn, H. H., and Beaty, H. N.: Transtracheal aspiration in the evaluation of patients with pneumonia, Ann Intern Med, 1070, 72, 183. Bartlett, J. G., Rosenblatt, J. E., and Finegold, S. M.: Percutaneous transtracheal aspiration in the diagnosis of anaerobic pulmonary infection, Ann Intern Med, 1973, 79, 535. Patterson, J. R., Blaschke, T. F., Hunt, K. K., Jr., and Mellin, P. J.: Lidocaine blood concentrations during fiberoptic bronchoscopy, Am Rev Resp Dis, 1975,112,53. Conte, B. A., and LaForet, E. G.: The role of the topical anesthetic agent in modifying bacteriologic data obtained by bronchoscopy, N Eng! J Med, 1962,267,957. Erlich, H.: Bacteriologic studies and effects of anesthetic solutions on bronchial secretions during bronchoscopy, Am Rev Respir Dis, 1961, 84, 414. Schmidt, R. M., and Rosenkranz, H. S.: Antimicrobial activity of local anesthetics: Lidocaine and procaine, J Infect Dis, 1970,121,597. Wanner, A., Amikam, B., Robinson, M. J., and Sackner, M. A.: Comparison between the bacteriologic flora of different segments of the airways, Respiration, 1973, 30, 561. Lees, A. W., and McNaught, W: Bacteriology of lower respiratory tract secretions, sputum, and upper-respiratory tract secretions in "normals" and chronic bronchitics, Lancet, 1959,2, 1112. Jordan, G. W., Krajden, S. F., Hoeprich, P. D., Wong, G. A., Pierce, T. H., and Rausch, D. C.: Trimethoprim-sulfamethoxazole in chronic bronchitis, Can Med Assoc J, 1975,112,915.
