Protected Specimen Brush in the Assessment of Ventilator-Associated Pneumonia

Protected Specimen Brush in the Assessment of Ventilator-Associated Pneumonia* ., . ,...... Selection of a Certain Lung Segment for Bronchoscoplc Sampling Is Unnecessary Charles H. Marquette, M.D.; Frederic Herengt, M.D.; mbienne Saulnier, M.D.; Remy Nevierre, M.D.; Daniel Mathieu, M.D.; Rene Courcol, M.D.; and Philippe Ramon, M.D.

The protected specimen brush (PSB) with quantitative cultures is one of the most reliable techniques for assessing pneumonia in mechanically ventilaled (MV) patients. The need to select a certain lung segment for broncboscopic sampling is still debated. We investigated whether tbe results of PSB specimens collected within an area radiographically involved with pneumonia (inv-PSB) differed from the results of PSB specimens collected within a lung area without radiographic abnormalities (non-inv-PSB) in 39 MV patients with suspected pneumonia. The comparison of bacterial titen of inv-PSB and non-inv-PSB cultures did not disclose significant differences. Agreement regarding the diagnosis of pneumonia according to recommended diagnostic threshold was observed in 34 of39 patients (87.1

a diagnosis of pneumonia in mechaniE stablishing cally ventilated (MV) patients is a troublesome

issue. Combination of clinical variables has proved to be inaccurate in predicting pneumonia in this setting.1-11 Simple diagnostic methods such as endotracheal or bronchoscopic aspiration are unreliable because of almost universal colonization of the proximal airways in MV patients. 1·6 The protected specimen brush (PSB) with quantitative cultures is one of the most reliable means of establishing the diagnosis of pneumonia and determining the causative microorganism7·16 and is now recommended for the diagnosis of pneumonia in MV patients (provided the threshold of 2::103 colony-forming units [cfu]/ml is used). 17·llll On the other hand the selection of the proper lung segment for bronchoscopic sampling is an unresolved problem.12 In patients with a new or progressive infiltrate evidenced on a chest radiograph it is theoretically advisable to place the catheter into the lobar or segmental airway leading to the radiographically dem*From Departement de Pneumologie (Ors. Marquette and Ramon), Service d'Urgence Respiratoire et de Reanimation (Ors. Herengt, Saulnier, Nevierre and Mathieu), and Service de Bact6riologie (Dr. Courcol), H6pital A. Calmette, Lille, France. This work was supported by the Comit6 National contre les Maladies Respiratoires et la Tuberculose. Manuscript received March 19; revision accepted May 28. &print requem: Dr. Marquette, Departement de Pneumologie,

HOpital Calmette, 59037 Ulle Ced.ex, France

percent). These results which are in accordance with tbe pathophysiology of ventilator-associated pneumonia and histologic studies do not support the need to select a certain lung segment for bronchoscopic sampling in most MV patients with suspected pneumonia. (Chat 1993; 103:243-47)

cfu=colony-formina unit; inv-PSB=protecled specimen brush specimens coUectecl within an area iowlvecl with • as shown radiographicall; MV=nieehanieally ven~ inv-PSB protAicleil speeimen brush speeimens collected within a lung area without ndiographieally shown ~ PSB =protected specimen LruSh; VAP ventilator-associ pneumonia

=

=

onstrated abnormality, whereas in patients with more

·than one new infiltrate the segment visualized as having purulent secretions should be selected. The problem actually is more complex because in MV patients the chest roentgenogram is unreliable both in predicting pneumonia13 and in locating the site of pneumonia when it is present. 24 Indeed in MV patients even localized infiltrates frequently are associated with noninfectious conditions such as pulmonary infarction, pulmonary hemorrhage, atelectasis, asymmetric adult respiratory distress syndrome and atypical pulmonary edema. 13 Determining which area of the lung is likely to be involved with suspected pneumonia by examining an anteroposterior roentgenogram may appear futile since histologic studies in ventilator-associated pneumonia (VAP) demonstrate that even localized infiltrates are associated with patchy and disseminated areas of pneumonia. 12•14 Bronchoscopic examination has never proved to be accurate in localizing infection and is generally difficult because all the airways appear erythematous and contain purulent secretions. 12 Studies partly addressing the issue of the sampling area selection note conflicting results. Baughman et al8 reported PSB cultures of the noninvolved lung to be sterile in three of four patients and positive in only one patient with pneumonia. Similarly, Belenchia et al111 have described CHEST/ 103 / 1 I JANUARY, 1993

243

significant bacterial growth in only one of two contralateral sampled bronchi in patients with ARDS and fever.. Conversely, Torres et al14 compared 6berscopeguided PSB with blindly advanced PSB (through a Metras catheter) and found agreement in 61 percent of the patients with proved pneumonia and complete disparity in only two patients (11 percent). Regarding decisions concerning antibiotic selection, Zucker et al• noted a high success rate using blind PSB in intubated patients. The purpose of this study was to compare the results of PSB directed into a radiographically involved lung area to the results of PSB directed into a lung area without radiographic abnormality in the contralateral lung.

Table I-Quantitative Cultura of the Involved Lung Protect«l Specimen B""h and the Noninvolved Lung Pro"cted Specimen B""h Case

Bacterial Species

lnv-PSB, log,0 cfu/ml

Non-inv-PSB, log,0 cfu/ml

2

Acinetobacter baurnannii Streptococcus species A baurnannii Xanthmnonas maltophilia Citrobacter dium1s

5 3 4 2 2

5 3 4 2

3 4* 5 6 7

P aeruginosa. type 9 Streptococcus species

.ftdfent Selection

S""1g Dengn The study regimen was perfurmed within 24 h after the clinical diagnosis of pneumonia was suspected. Patients were premedicated intravenously with midazolam. The fraction of inspired oxygen was adjusted to 100 percent and kept at this level throughout the procedure. In order to obtain clear airways. careful tracheobronchial aspiration was perfurmed through the endotracheal tube prior to Bbroscopy. The Bberoptic bronchoscope was then introduced through a special adapter (Bodai suction-safe Y; Sontek Medical, Lexington Mass) and advanced under direct vision (I) next to the oriftce of a segmental bronchus in the area of abnormal opacity thought to be pneumonia and (2) next to the oriftce of a segmental bronchus of the same order in an area without abnormal opacity in the contralateral lung. The order of passage was selected according to a random number design. No suction or topical anesthetics were used. Two plugged telescoping catheter brushes (Model 89.01; TAC medical, Bobigny, France) were successively passed through the working channel and advanced 2 cm beyond the tip of the fiberoptic bronchoscope. After ejecting the distal plug by protruding the inner cannula the brushes were advanced beyond the tip of the inner cannula 2 to 4 cm into the afurementioned selected bronchi, gently rotated several times and then retracted a few centimeters into the inner cannula. After removal from the fiberoptic bronchoscope, the whole catheters were immediately transported to the laboratory. The entire sampling procedure lasted appromnately 1 min. The distal portion of the catheters was wiped clean with 70 percent ethanol, cut with sterile scissors and discarded. The brushes were advanced beyond the tip of the transected inner cannula, aseptically

244

A bau1nannii Pseudmnonas aeruginosa.

... t ... t

2 2 2 ... t 3 4

3 2 2 ... t 3 2

5 5 ... t 3

5 5

type 1

METHODS Thirty-four consecutive MV patients (25 men and 9 women) were studied over a five-month period. Mean age was 62±3.5 (SEM) yean (range, 21 to 86 yean). Five patients were studied twice. In patients studied twice, more than eight days had elapsed since the lint specimens had been collected. The mean number of ventilator days was 11±3.5 (SEM); the range was 1to60 days. The primary indications fur ventilatory support included exacerbation of chronic obstructive pulmonary disease (n = 13), acute cardiac failure (n =3). status asthmaticus (n =2), community-acquired pneumonia (n = 1). tetanus (n =1), acute hepatic failure (n = I). drug overdose (n =2), postoperative respiratory failure (n =4), respiratory failure associated with multitrauma (n =3), multiorgan failure associated with sepsis (n =2) and neurologic emergencies (n =2). All patients met all four of the clinical criteria fur pneumonia described by Andrews and associates. 1 Seven patients had not received prior antibiotics. In the remaining 9:1, antibiotics were withheld 24 h prior to study protocol. Infurmed consent was obtained from all patients or from the nearest relative.

Candida albicans Staphylococcus aureus S pneurnoniae

... t

8 9 10 11

12* 13* 14 15 16 17 18 19

P aeruginosa C albicans Enterobacter cloacae A bau1nannii C albicans P aeruginosa, type 9 S pneu1noniae A bau1nannii S aureus Streptococcus species P aeruginosa. type 12 Saureus Escherichia coli X 1naltophilia A bau1nannii

20

23 24 27

P aeruginosa. type 12 S aureus Streptococcus species Haeinophilus inftuenzae

28

29 30* 31

A bau1nannii S aureus Ecali

32

33 34

35 36*

3 6 5 5 5 2 2

... t 6

A baumannii Klebsiella pneu1noniae A bau1nannii Abaumannii P aeruginosa, type 4 S aureus C albicans P aeruginosa. type 12

37 38*

Abaumannii

39

K pneu1noniae P aeruginosa, type 7 S pneumaniae

3 2 5 6

... t 5 2 4 4 4 6

... t

... t

6 5 6

5 4 5

... t 5

... t 6

... t ... t

... t . .. t

... t ... t

... t ... t ... t

... t

21 22

25 26

... t

... t

... t 2 2 4

... t ... t ... t 2 3 6 5 4 ... t 5 2 4 3 4 6 5 4

... t

... t

... t

2 2 5 3 ... t

... t 4

... t 3 5

... t ... t 5 2 2

... t ... t 6

... t 3 5 2

*Cases in which there was a lack of agreement between inv-PSB and non-inv-PSB. tQuantitative culture results below threshold of detection (10' cfu/ ml). PSS In AHHlmenl of VenlllaloMl88oc Pneumonia (Marquette et al)

severed with a wire cutter and dropped into a sterile glass vial containing 1 ml of lactated Ringer's solution. The vial was then mechanically vortexed for 60 s. Three tenfold dilutions were prepared and 0.1-ml aliquots of each dilution were plated on dill'erent media (Pasteur Production, Paris) for quantitative culture: purple lactose agar, 5 percent blood Columbia agar, 5 percent blood Columbia agar to which nalidixic acid (40 m!VL) was added for Streptococcus J1fl8U'llO'liae cultures), and "chocolate" agar (to which bacitracin (30 U/ml] was added for Haemophilus species cultures). H positive, counts of colony-forming units per milliliter and identi&cation were performed for each species. With this quantitative culture technique, the threshold of detection for a microorganism was 10" cfu/ml.

Statistical Methods Correlation between culture results &om PSB specimens collected within an area involved with pneumonia as shown radiographically (inv-PSB) and PSB specimens collected within a lung area without radiographically shown abnormalities (non-inv-PSB) was calculated using regression analysis. Bacterial titers of inv-PSB and non-inv-PSB cultures were compared using the paired Student t test. The PSB results were considered as indicative of pneumonia according to the recommended diagnostic threshold (PSB ;2:1()3 cfu/ ml). Accordingly the percentage of agreement between the inv-PSB and non-inv-PSB with respect to results indicative of pneumonia was calculated by adding the number of patients with both inv-PSB and non-inv-PSB culture results above the diagnostic threshold for at least one species and the number of patients with both inv-PSB and non-inv-PSB culture results below the diagnostic threshold for all the recovered species and by dividing this sum by the total number of patients. RESULTS

Thirty-nine paired specimens were collected in the 34 patients over a &ve-month period. Results of cultures are expressed in Table 1.

Qualitative Results Sixty bacterial pairs were cultured in the 39 patients CTuhle 1). With exception of nonpathogenic species, total agreement between microorganisms cultured from both inv-PSB and non-inv-PSB was observed in 31 of 39 patients (79.4 percent). There were partial disparities in 8 of 39 patients (20.6 percent). No total disparity was observed.

Quantitative Results Quantitative cultures of inv-PSB were positive (ie, above the 1()1 cfu/ml threshold of detection) in 61.5 percent of the patients and indicative of pneumonia (ie, above the 1()3 cfu/ml diagnostic threshold) in 43.5 percent of the patients. Similarly, quantitative cultures of non-inv-PSB were positive in 64.1 percent of the patients and indicative of pneumonia in 46.1 percent of the patients. The comparison of bacterial titers of inv-PSB and non-inv-PSB cultures did not disclose a signi&cant difference (p = 0.28). The inv-PSB and noninv-PSB quantitative cultures correlated well (r=0.62; p<0.0001). Except for patient 4, in whose case the lack of agreement concerned a nonpathogen, agreement regarding the diagnosis of pneumonia was oh-

served in 34 of 39 patients (87 .1 percent). Concerning the cases without agreement, it is noteworthy that the same pathogens were recovered in both inv-PSB and non-inv-PSB in three patients (patients 12, 13 and 30). In patients 36 and 38, the inv-PSB recovered an additional pathogen. In patient 36, the non-inv-PSB recovered an additional pathogen. In two of these &ve patients, the inv-PSB result was indicative of pneumonia, whereas the non-inv-PSB result was not (patients 12 and 36). Conversely, in two patients the noninv-PSB result was indicative of pneumonia whereas the inv-PSB result was not (patients 13 and 30). In the last patient (patient 38), polymicrobial Bora was recovered. The inv-PSB cultured Acinetobacter baumannii (105 cfu/ml) which was not present in non-inv-PSB, whereas non-inv-PSB cultured Pseudomonas aeruginosa (105 cfu/ml) which was not present in inv-PSB. Irrespective of the patients, lack of agreement regarding the diagnosis of pneumonia, with respect to the diagnostic threshold (~1()3 cfu/ml), occurred in 16 instances out of 60 bacterial pairs (26.6 percent); this lack of agreement concerned a total of 11 patients. Prior antibiotic therapy did not affect qualitative or quantitative comparisons. DISCUSSION

The present study demonstrates an acceptable level of agreement between inv-PSB specimens and noninv-PSB specimens in the contralateral lung. The major limitation of this study is the absence of histologic confirmation of pneumonia. In suspected VAP the best way to ascertain the diagnosis is to remove the entire lung for histologic examination since sampling small lung specimens in the immediate postmortem period may miss the diagnosis of pneumonia in one fourth of the patients. 24 For ethical and practical reasons we did not perform lung examination in those patients who died and thus could not ascertain the presence of pneumonia. We indirectly addressed this issue by considering the practical question facing the physician when there is reason to suspect pneumonia. That is, in an MV patient with fever, diffuse purulent tracheobronchial secretions, and abnormalities evidenced on an anteroposterior roentgenogram, is it warranted to precisely identify the lung segments potentially involved with pneumonia and to consequently try to wedge the protected specimen brush into the airway leading to this area? Since in this setting therapeutic decisions ensue from the qualitative and the quantitative results of PSB cultures we investigated whether the results of inv-PSB specimens differed from the results of non-inv-PSB specimens. With respect to the microorganisms recovered and to the results of quantitative cultures, we demonstrated a high level of agreement between inv-PSB and non-inv-PSB specimens. These results are consistent with the results of CHEST I 103 I 1 I JANUARY, 1993

245

previous clinical studies, pathophysiological considerations and autopsy studies. 1\vo studies disclosed good correlation between culture results of blindly advanced PSB and either results of bronchoscopicguided PSB 14 or clinical diagnosis of pneumonia.llll Studies using other techniques such as the plugged telescoping catheter27 or bronchoalveolar lavage22·28 evidenced good correlation between quantitative cultures of samples collected in an unselected area and samples collected in a selected area. When considering the pathophysiology of VAP, there is no reason for advocating bacteriologic sampling in a precisely selected lung area since in most instances VAP results from colonization of the lower respiratory tract with aspirated microorganisms. Multifocal damage in the terminal and respiratory bronchioles ensues. This is in contrast with the pathophysiology of lobar pneumonia where the damage initially occurs in well-delimited terminal airspaces. 29•30 Finally, autopsy studies in VAP indicate that multifocal areas of bronchopneumonia are disseminated throughout the lungs211.24 and predominantly involve the posterior segments of the lower lobes. 24 There is, however, controversy about these autopsy findings which may represent terminal stages of pneumonia in terminally ill patients. In certain subsets of patients VAP may indeed be well delimited, thus advocating sampling in precisely selected lung areas. This is supported by two clinical studies of ventilated patients with clinical pneumonia, which found significant differences between samples collected in the area of suspected pneumonia and samples collected from the noninvolved lung. 8 •25 The lack of agreement that we noted in five patients may as well be related to a nondisseminated stage of VAP in these patients. Further studies clearly are necessary to evaluate the time-dependent distribution of bacteria in the lungs of patients with VAP and to evaluate the need for precise selection of the sampling area in certain subsets of patients. Together with previous studies our findings, pathophysiologic considerations and reports of histologic findings lead us to conclude that in most MV patients with suspected pneumonia it is unnecessary to select a certain lung segment for bronchoscopic sampling. REFERENCES

1 Johanson WG, Pierce AK, Sanford JP, Thomas GD. Nosocomial respiratory infections with gram-negative bacilli: the significance of coloni7.ation of the respiratory tract. Ann Intern Med 1972; 77:701-06 2 Andrews CP, Coalson JJ, Smith JD, Johanson WG Jr. Diagnosis ofnosocomial bacterial pneumonia in acute, diffuse lung injury. Chest 1981; 80:254-58 3 Bell RC, Coalson JJ, Smith JD, Johanson WG Jr. Multiple organ system failure and infection in adult respiratory distress syndrome. Ann Intern Med 1983; 99:293-98 4 Chastre J, Viau F, Brun P, Pierre J, Dauge MC, Bouchama A, et al. Concise clinical studies: prospective evaluation of the pro-

tected specimen brush for the diagnosis of pulmonary infections in ventilated patients. Am Rev Respir Dis 1984; 130:924-29 5 Fagon JY, Chastre J, Hance AJ, Guiguet M, Trouillet JL, Domart Y, et al. Detection of nosocomial lung infection in ventilated patients. Am Rev Respir Dis 1988; 138:110-16 6 Cameron JL, Reynolds J, Zuidema GD. Aspiration in patients with tracheostomies. Surg Gynecol Obstet 1973; 136:68-70 7 Torzillo PJ, McWillian DB, Young IH, Woog RH, Benn R. Use of protected telescoping brush system in the management of bacterial pulmonary infection in intubated patients. Br J Dis Chest 1985; 79:125-31 8 Baughman RP, Thorpe JE, Staneck J, Rashkin M, Frame PT. Use of the protected specimen brush in patients with endotracheal or tracheostomy tubes. Chest 1987; 91:233-36 9 Chastre J, Fagon JY, Soler P, Bornet M, Domart Y, Trouillet JL, et al. Diagnosis of nosocomial bacterial pneumonia in intubated patients undergoing ventilation: comparison of the usefulness of bronchoalveolar lavage and the protected specimen brush. Am J Med 1988; 85:499-506 10 Faling LJ. New advances in diagnosing nosocomial pneumonia in intubated patients. Am Rev Respir Dis 1988; 137:253-55 11 Villers D, Derriennic M, Raffi F, Gennaud P, Baron D, Nicolas F, et al. Reliability of the bronchosoopic protected catheter brush in intubated and ventilated patients. Chest 1988; 4:52730 12 Winterbauer RH, Bass JB. Fiberoptic bronchoscopy with protected brush catheterization should be used for the speci&c diagnosis of nosocomial pneumonia. Am Rev Respir Dis 1988; 138:1072-74 13 Baigelman W. Bellin S, Cupples LA, Berenberg MJ. Bacteriologic assessment of the lower respiratory tract in intubated patients. Crit Care Med 1989; 14:864-68 14 Torres A, Puig de la Bellacasa J, Rodriguez-Raisin R, Jimenez de Anta MT, Agusti-Vidal A. Diagnostic value of telescoping plugged catheters in mechanically ventilated patients with bacterial pneumonia using the Metras catheter. Am Rev Respir Dis 1988; 138:117-20 15 Torres A, Puig de la Bellacasa J, Xaubet A, Gonzalez J, RodriguezRoisin R, Jimenez de Anta MT, et al. Diagnostic value of quantitative cultures of bronchoalveolar lavage and telescoping plugged catheters in mechanically ventilated patients with bacterial pneumonia. Am Rev Respir Dis 1980; 140:306-10 16 De Castro FR, Violan SJ, Capuz BL, Luna JC, Rodriguez BG, Alonso JL. Reliability of the bronchoscopic protected catheter brush in the diagnosis of pneumonia in mechanically ventilated patients. Crit Care Med 1991; 19:171-75 17 Johanson WG. Ventilator-associated pneumonia: light at the end of the tunnel? Chest 1990; 97:1026-27 18 Kalm RJ, Arich CH, Baron D, Gutmann L, Hemmer M, Nitenberg G, et al. Diagnostic des pneumopathies nosocomiales en reanimation. Bean Soins Intens Med Urg 1990; 6:91-9 19 Meduri GU. Ventilator-associated pneumonia in patients with respiratory failure: a diagnostic approach. Chest 1990; 97:120819 20 Torres A. Accuracy of diagnostic tools for the management of nosocomial respiratory infections in mechanically ventilated patients. Eur Respir J 1991; 4:1010-19 21 Torres A, Gonzalez J, Ferrer M. Evaluation of the available invasive and non-invasive techniques for diagnosing nosocomial pneumonias in mechanically ventilated patients. Intensive Care Med 1991; 17:439-48 22 Johanson WG, Seidenfeld JJ, Gomez P, De Los Santos R, Coalson JJ. Bacteriologic diagnosis of nosocomial pneumonia following prolonged mechanical ventilation. Am Rev Respir Dis 1988; 137:259-64 23 Wunderink RG, Woldenberg LS, Zeiss J, Day MD, Ciemins J, Lacher DA. The radiologic diagnosis of autopsy-proved ventilaPSS in Asaessment ol Venlilafor-associafad Pneumonia (Marquette et al)

to~associated pneumonia. Chest 1992; 101:458-63 24 Rouby JJ, Rl>SSignon MD, Nicolas MH, Martin de Lassale E, Cristin S, Grosset J, et al. A prospective study ofbronchoalveolar lavage in the diagnosis of nosocomial pneumonia. Anesthesiology 1989; 71:679-8.5 2.5 Belenchia JM, Wunderink RG, Meduri GU, Leeper Kv. Alternate causes of fever in ARDS patients suspected of having pneumonia. Am Rev Respir Dis 1991; 143:A683 26 Zucker A, Pollack M, Katz R. Blind use of the double-lumen plugged catheter for diagnosis of respiratory tract infections in critically ill children. Crit Care Med 1984; 12:867-70 27 Pham LH, Brun-Buisson C, Legrand P, Rauss A, Verra F, Brochard L, et al. Diagnosis of nosocomial pneumonia in

XVII

x. x ... Ar-..~'TIH

.'\~.·

.,f

""f

mechanically ventilated patients: comparison of a plugged telescoping catheter with the protected specimen brush. Am Rev Respir Dis 1991; 143:1055-61 28 Pugin J, Auckenthaler R, Mili N, Janssens JP, Lew PD, Suter PM . Diagnosis of ventilato~associated pneumonia by bacteriologic analysis of bronchoscopic and nonbronchoscopic blind bronchoalveolar lavage fluid. Am Rev Respir Dis 1991; 143:112129 29 Palmer LB. Bacterial colonization: pathogenesis and clinical significance. Clio Chest Med 1987; 8:45.5-66 30 Heitzman ER. Pneumonia and lung abscesses. In: The lung, 2nd ed. St Louis: CV Mosby, 1984; 194-234

Plan to Attend ACCP's XVII World Congress on Diseases of the Chest

H f f ! ; . •. 1 ,

CHEST I 103 I 1 I JANUARY, 1993

247