Use of the Protected Specimen Brush in Patients with Endotracheal or Tracheostomy Tubes

Use of the Protected Specimen Brush in Patients with Endotracheal or Tracheostomy Tubes· Robert P. Baughman, M.D., EC.C.P. ;joseph E. Thorpe, M.D., EC.C.P.; joseph Staneck, Ph.D .; MitcheU Rashkin, M.D., EC.C.P.; and Peter T. Frame, M.D .

Twenty-one patients on mechanical ventilators for greater than 48 hours who had new localized infiltrates were evaluated using a quantitative culture technique of the involved lung compared to the non-involved lung. Based on the clinical course, response to antibiotics, or subsequent analysis of pathologic specimens, eight patients were felt to have acute bacterial pneumonia, while the remaining 13 were felt to have an alternative cause of their infiltrate. Cultures of the protected brush specimen of the involved

lung in all eight cases of bacterial pneumonia had one or more organisms grown at a greater than 100colony forming units (cfu) per mI while only one of the 13 cases of nonpneumonia had a culture from the involved area having greater than 100 cfu per mI (p
The diagnosis of pneumonia of the mechanically ventilated patient is difficult. In these patients, the larynx has been bypassed and the trachea is often colonized with microbial flora.1,2 This is especially common in the intensive care unit, where colonization of patients with Gram-negative organisms is a wellrecognized phenomenon. Treatment in response to leukocytosis, fever,and new infiltrate on chest roentgenogram in this patient population is usually empiric. Antibiotics directed towards the organism recovered from secretions obtained by tracheal suctioning are often used. Studies with the protected brush have suggested that quantitative cultures from the lower respiratory tract can be useful in diagnosing the cause of pneumonia." Since these observations were made in an animal model within a research laboratory, we attempted to use the protected brush in the clinical setting of an intensive care unit in patients with new pulmonary infiltrates. Toassess the significance of the results of the brush specimens from the areas of infiltration noted on x-ray, we also obtained brush samples from uninvolved areas. We found the protective brush method does provide useful information on the intubated patient, provided quantitative culture procedures are carried out and that a comparison is made to non-involved portions of the lung.

MATERIAL AND METHODS

*From the Departments of Medicine and Pathology and Clinical Laboratory Medicine, University of Cincinnati Medical Center, Cincinnati. Supported in Part by Grant No. RROOO68-21 from the National Institutes of Health, General Clinical Research Center, University of Cincinnati. Manuscript received November 14, 1985; revision accepted August 12. Reprint requests: Dr. Baughl1l6n, Pulmonary Disease Division, University ofCincinnati Mei1lcal Center, Cincinnatl45267-o564

Patients were recruited from those seen in the medical and surgical intensive care units at the University of Cincinnati Medical Center and the Veterans Administration Medical Center, Cincinnati. Patients were eligible for study if they had been intubated with either an endotracheal tube or tracheostomy tube for at least 48 hours, if they had a new, localized infiltrate on chest roentgenogram either on a normal chest roentgenogram or in addition to diffuse infiltrate already present, as in patients with adult respiratory distress syndrome, and those who were not receiving antibiotics for at least 48 hours prior to the study. Patients were not candidates for bronchoscopy if their arterial POa could not be corrected to greater than 60 mm Hg, if they hadprolonged bleeding times or if they had significant thrombocytopenia (platelet count < 6O,OOOImm) or if the internal diameter of the endotracheal or tracheostomy tube was less than 7.5 mm , The patient or patient advocate signed an informed consent form approved by the Human Research Committee of the University of Cincinnati . The protected brush specimen was taken using a modification of the original procedure described by Wunberley et al." 1Wo bronchoscopes were used . The 6rst bronchoscope was used to suction the airways and was removed after the bronchi leading to the involved and uninvolved areas were cleared of secretions. A second bronchoscope, with no suction attached, was then advanced to the bronchus leading to the involved area, as determined by the most recent chest roentgenogram. A triple lumen protected brush was advanced and the gelatin plug pushed out into a dependent bronchus. The brush and inner catheter were then advanced to a separate bronchus subseg.nent thatwas leading to the involved area, the brush advanced and a sample taken. The brush was then retracted into the inner sheath and the entire bronchoscope removed with the brush still protruding distally through the bronchoscope. The outside of the catheter was then cleaned with 70 percent alcohol and a I em portion of the inner sheath cut off. The brush was then readvanced, cut off with sterile wire clippers, and placed into a sterile specimen container with 1 ml of balanced salt solution. Separate samples, similarly obtained, were taken from the noninvolved lung and from the trachea, in that order. The specimens were transported directly to the microbiology laboratory. There was never more thana 6G-minute interval between CHEST I 91 I 2 I FEBRUARY. 1987

233

Table I-Diagnosis Underlying Condition

Patient Group Pneumonia

Cerebrovascular accident Aspiration Esophageal rupture Chronic obstructive lung disease

Non-pneumonia

Adult respiratory distress syndrome Renal transplants with Pneumocyst/8 carinii Pulmonary embolism Cerebrovascular accident Chronic obstructive lung disease

infiltrate on chest roentgenogram temporally related to the use of antibiotics. In the other 13 cases (nonpneumonia), an alternate diagnosis was arrived at for the cause of the localized infiltrate. Six patients appeared to have the adult respiratory distress syndrome.' In these cases, the pulmonary infiltrate became more diffuse, the patient became severely hypoxemic, and did not improve with antibiotic therapy. One patient with a cerebrovascular accident had a mucous plug that responded to pulmonary suctioning only. The remaining three cases had evidence of other diseases. Two had pulmonary embolism documented by pulmonary angiogram, one had Pneumocqstis carinii infection, and one had pulmonary fibrosis without infection, the latter two documented by open lung biopsy. Tables 2 and 3 summarize the organisms found in the trachea, involved, and non-involved lobe of patients in the two groups. Seventeen of the 21patients had one or more organisms in their tracheal secretions, while 16 patients had no organisms grown from their uninvolved lung (p <0.02). The data also document differences in microbial Bora between the noninvolved and involved lobe in the two groups. In the nonpneumonia group, four of 13 patients had had organisms grown from the involved lobe ; however, in only one case was there more than 100 colony fonning units/ml. In the pneumonia group, however, all eight patients had at least one organism with 100 or more cfulml recovered from the area of the involved lung . When the bacterial count in the involved lobe was compared to that of the non-involved lobe (Table 4), all eight cases of pneumonia had at least a SO-fold increase in the number of organisms found in the involved versus noninvolved lobe. In only one on3 cases among the nonpneumonia group was such a difference noted (p <0.(01).

No of Patients

2 2 1 3 6 1 2 1 3

the first specimen collection and processing. Following 30 seconds of vigorous agitation using a vortex mixer, a1iquots of the neat, 100:1, and 10,000 :1 dilutions of the bronchial brush eluate were spread across the surface of six agar plates (Gibco Laboratories, Madison, WI), 0.1 ml per plate. Plates containing trypticase soy agar with 5 percent sheep blood, MacConkey agar, and chocolate blood agar, were incubated at 35"C under 7-10 percent COl' while plates containing Columbia agar with 5 percent sheep blood, kanamycinvancomycin agar, and neomycin agarwere incubated at 35"C under strictly anaerobic conditions. All cultures were reported as colony fonning units (cfu)per ml ofspecimen eluate. At 24 and 48 hours, the aerobic cultures were examined for bacterial isolates. Cultures plated for anaerobes were observed to 72 hours for growth. Identification of aerobic and anaerobic isolates was carried out using

standard techmques ." Statistics: Comparisons were made between groups using Fishers exact test. e

REsur.:rs We studied a total of 21 patients. The underlying condition in each case is shown in Table 1. All eight pneumonia patients had a clinical course and chest roentgenogram consistent with bacterial pneumonia. In all cases, there was resolution of the localized

Table 2-Quantitative Culture Results among Pneumonia Patient Group Patient No.

Involved Lung

Non-Involved Lung

8,OOO·Viridans group streptococci

70 Viridans group streptococci 140 mixed flora No growth 300 H inftuenwe 150 S pneumoniae

1,000 Haemophilus inftuenzae

2

. 22,000 Pseudomonas aeruginosa

3

20,000 Branhamella catarrhalia 500 S pneumoniae

4

30,000 100 100 1,000

5 6 7

8

P aeruginosa P aeruginosa

6,000 P aeruginosa No growth

Viridans group streptococci

No growth No growth

K pneumoniae

10,000 Serratia marcescens 1,000 Viridans group streptococci

No growth

Trachea 12,000 Viridans group streptococci 6,000 mixed Bora 5,000 P aeruginosa

50,000 Bacteroides oralis 4,000 Bacteroidesmelaninogenicus 1,000 H inftuenzae 900 S pneumoniae 40 P aeruginosa 200 P aeruginosa 20 K pneumoniae 900 50,000 10,000 1,000 900 1,000

Viridans group streptococci S marcescens P aeruginosa Viridans group streptococci Viridans group streptococci Coagulase neg . staphylococci

·Colony fonning units/ml.

234

Use at ProtectedSpecimen Brush (Baughman 8t aJ)

Table 3-Quantilative Culture &aulu among Non-pneumonia Patient Group Patient No. I 2 3 4

5 6

7 8 9 10 11 12 13

Involved Lung

Non-Involved Lung

Trachea

No growth 100 Corynebacterium sp. 20 Viridans group streptococci 20 Beta streptococci not group A 20 Escherichia coli No growth

No growth No growth 320 Vlridans group streptococci 500 Beta streptococci not group A No growth No growth

40 Peptostreptococcus No growth No growth No growth No growth No growth No growth No growth

40 Coagulase neg staphylococci No growth No growth No growth No growth No growth No growth No growth

I,OOO*mixed flora 600 Corynebacterium 140 Vlridans group streptococci 140 Beta streptococci not group A 120 E Coli 200 Viridans group streptococci 50 S aureus 500 Candida albicans No growth No growth 100 S pneumoniae 500 S aureus 500 Viridans group streptococci No growth No growth

*Colony forming unitslml.

In the pneumonia group, cultures of the trachea yielded 16 organisms while cultures of the involved lobe yielded 11 significant organisms. In two cases, organisms seen in the involved lobe were not recovered from the trachea culture and in seven cases, an organism wasgrown from the trachea, but not found in the involved lobe. In the non-pneumonia groups, 13 organisms were recovered from the trachea. In this group, however, only five organisms were cultured from the involved lobe and only one organism (a diphtheroid) was cultured at greater than 100 cfulml. 'Ireatment for the patients pneumonia was usually directed by the culture results. Patients were treated with broad spectrum antibiotics after the bronchoscopy but before the results of culture were available. Antibiotics were then directed towards the organisms found in the involved lung. For example, those patients with Streptococcus viridans were treated with ampicillin or penicillin. DISCUSSION

The degree of accuracy of diagnosis of pneumonia often depends on the particular patient. Several methods have been proposed for making the diagnosis of pneumonia, many trading safety for accuracy. For example, obtaining an expectorated sputum sample is Table 4-IJijference3 in Bacterial CountB/ml between Involved and Non-involved Lobe.· Greater Than 50nmes Difference

Less than 50 Times Difference

Pneumonia group

8

o

Nonpneumonia group

1

12

*p
an inexpensive, innocuous method of providing a specimen for culture of respiratory tract pathogens. However, the frequent contamination of the specimen with oral floracauses significant difficulty in interpreting sputum culture results. sThe usefulness ofthe more invasive transtracheal aspirate and protected brush specimen ofthe trachea are limited, since the results of such cultures are misleading in patients with tracheobronchitis or tracheal colonization." In our study, eight of13 patients without pneumonia had potentially clinically significant organisms growing from their tracheal secretions. Recent efforts to sample the lower respiratory tract have met with varying degrees of success. One difficulty has been in interpretation of results, since it is often not clear whether bacterial colonization is contaminating specimens of the lower respiratory systern." In our study, we found three-fourths of our patients had sterile specimens obtained from an area of the lung not felt to be infected as determined by chest roentgenogram. It was also shown that the involved area of the lung had at least a SO-fold increase in the number ofcfulml than the noninvolved area of the lung in patients with pneumonia. This study stresses that quantitation of the number of organisms should be performed whenever the protected brush specimen is used. This suggestion has been made by others.v' Our low number of cfu seen in patients with pneumonia could be due to dilution as well as the difficulties of transport to the microbiologylaboratory. Moser et al," studying an animal model for S pneumoniae pneumonia, found that plating the sample immediately had a significantly higher yield than sending the sample to a microbiology laboratory and plating there. This is a real limitation of the technique, since some organisms, especially anaerobes, may not grow in sufficient numbers to be clearly different from non-involved lungs. CHEST I 91 I 2 I FEBRUARY. 1987

235

This method also only investigated bacterial infections and did not examine possible viral or fungal infection. Although the tracheal specimens obtained from the patients in our study often contained the pathogens isolated from the involved lung, a significant organism was grown in the involved lung specimen but not from the tracheal specimen in two patients with pneumonia. More importantly, there were several incidences where a significant isolate, usually a Gram-negative organism, was grown from the tracheal specimen but not from the involved lung, suggesting the organism had no role in the disease etiology. Colonization of the trachea with Gram-negative bacteria is a very common occurrence among intubated patients in intensive care units fur over two days. U The finding of multiple organisms in several of our pneumonia patients is not surprising. It is becoming increasingly clear that pneumonias acquired while on mechanical ventilators are often polymicrobial. 12 This study proved useful in narrowing the number of potential pathogens from the number identified in the trachea. In summary, we have demonstrated that the protected brush specimen can be used to obtain a representative sample of the distal airways. This sample, especially when compared to a non-involved portion of the lung using a quantitative culture technique, is useful in diagnosing the causative agent of pneumonia in patients requiring mechanical ventilation who acquired new pulmonary infiltrates. ACKNOWLEDGMENT:The authors wish to thank Jean Walden fur

238

help with culturing specimens, Dr. Robert Loudon fur his review, and Naomi Sims fur secretarial help in preparing this manuscript.

REFERENCES 1 Johanson WG , Pierce AK, Sanford JP, Thomas GD. Nosocomial respiratory infections with gram-negative bacilli: the significance of colonization of the respiratory tract. Ann Intern Med 1972; 77:701-06 2 French GL , Faca JH. Insignificance of colonic bacteria in the sputum of patients in a new ICU . Crit Care Med 1979; 7:487-91 3 Higuchi JH , Coalson JJ, Johanson WG. Bacteriologic dlagnosis of nosocomial pneumonia in primates: usefulness of the protected specimen brush. Am Rev Respir Dis 1982; 125:52-7 4 Wimberley N, Faling LJ, Bartlett JG. A 6beroptic bronchoscopy technique to obtain uncontaminated lower airway secretions fur bacterial culture. Am Rev Respir Dis 1979; 119:337-43 5 Leonette EH, ed . Manual of clinical microbiology, 3rd ed. Washington , DC : American Society fur Microbiology, 1980 6 Fleiss JL . Statistical methods fur rates and proportions. New York: John Wiley and Sons, 1981 7 Hopewell PC, Murray JF. The adult respiratory distress syndrome. Ann Rev Med 1976; 27:343-56 8 Davidson M, Tempest B, Palmer DL. Bacteriological dlagnosis of acute pneumonia: Comparison of sputum, transtracheal aspirates, and lung aspirates. ]AMA 1976; 235:158-63 9 Bartlett JG. Diagnostic accuracy of transtracheal aspiration bacteriologic studies. Am Rev Respir Dis 1977; 115:777-82 10 Matthew EB , Holstrom FMG, Kasper RL. A simple method lOr dlagnosing pneumonia in intubated or tracheostomized patients. Crit Care Med 1977; 5:76-81 11 Moser KM, Maurer J, Jassy L, KremsdorfR, Konopka R, Shure D, et aI. Sensitivity, specificity, and risk of dlagnostic procedures in a canine model of Streptococcus pneumoniae pneumona. Am Rev Respir Dis 1982; 125:436-42 12 Chastre J, Viau F, Brun P, Pierre J, Dauge MC, Bouchama A, et al. Prospective evaluation of the protected specimen brush fur dlagnosis of pulmonary infections in ventilated patients. Am Rev Respir Dis 1984; 130:924-29

Use 01 Protected SpecImen Brush (Baughmllll st 81)