Management of Bacterial Pneumonia in Ventilated Patients

clinical investigations in critical care Management of Bacterial Pneumonia in Ventilated Patients· Protected Bronchoalveolar Lavage as a Diagnostic Tool G . Umberto Meduri , M .D ., F.C.C .P.; Richard G. Wunderink, M.D ., F.C.C .P.; Kenneth V. Leeper, M .D ., F.C.C.P.; and David H . Beals, M.D.

We conducted a prospective study to determine the effectiveness of protected bronchoalveolar lavage (PBAL) in diagnosing pneumonia in ventilated patients and the usefulness of bronchoscopic data in treating ventilated patients. Entrance criteria were (1) fever and a new or progressive in6Itrate on chest roentgenogram with either leukocytosis or a macroscopically purulent tracheal aspirate, and (2) no antibiotic therapy for at least 48 h before bronchoscopy. Twenty-6ve ventilated patients met entrance criteria for the study and completed the protocol. PBAL was effective in retrieving distal airway secretions with a minimal degree of contamination as indicated by a speci6city and a negative predictive value of 100 percent. Bacterial isolates grew in all patients with pneumonia at a concentration ~ 100,000 cfulml, with a median growth of 500,000 cfulmI. The presence of a two-log difference between the highest quantitative culture count in patients without pneumonia and the lowest quantitative culture count in patients with pneumonia allowed a clearer determination of a patient's

status, with regard to pneumonia, compared with the signmcant overlap in unprotected BAL. Gram and Giemsa stains of the PBAL were positive in all patients with pneumonia and negative in those without pneumonia. All but one patient with pneumonia received narrow-spectrum antibiotic therapy. All patients without infection had no antibiotic administered. Clinical and roentgenographic criteria could not discriminate between patients with and without pneumonia, con6rming the 6ndings of previous investigations. The results of microscopic and culture analyses of the PBAL effluent proved useful in directing antibiotic treatment in patients with pneumonia and in avoidingunnecessary antibiotic use in those patients without pneumonia. (Chest 1992; 101:500-08)

Ventilator-associated pneumonia (VAP) occurs in 9 to 24 percent of patients with different types of acute respiratory failure (ARF),1.2 and its incidence approaches 70 percent in patients who die of adult respiratory distress syndrome (ARDS). 3-5 Pneumonia is an important factor in the survival of patients with ARF. The mortality rate in ventilated patients who develop pneumonia compared with ventilated patients without lung infection is 67 percent vs 23 percent for those with ARDS3-5 and 48 percent vs 26 percent'< for those without ARDS . Two recent studies-" have shown, using multivariate analysis, that appropriate antibiotic treatment may

significantly increase survival of ventilated patients with pneumonia, particularly those without a terminal illness. In these two studies, combined mortality rates of patients with appropriate and inappropriate treatment were 28 percent (45/159) and 64 percent (25/39), respectively.v" Moreover, empiric broad-spectrum antibiotic use in patients without infection is potentially harmful, beause it facilitates colonization and superinfection with more virulent organisms," The mortality rate in ventilated patients who received antimicrobial therapy before the onset ofpneumonia was 83 percent, compared with a 48 percent mortality rate in those who had not received prior antibiotics (p
*From the Department of Medicine. Division of Pulmonary and Critical Care Med icine , The University of Tennessee Health Science Center and Veteran s Affairs Medical Cent er, Memphis. Presented at the annual meeting of the American Thoracic Societ y, Boston, May 20-24. 1990. Manuscript received July 1; revision accepted September 27 Reprint requests: Dr. Meduri, 956 Court Avenue . Memphis 38163 500

ARF = acute respiratory failure; ICO = intracellular organism; MV=mechanical ventilation; PBAL=protected BAL; PSB= protected specimen brush; TBT= transbronchoscopic balloon tipped; VAP= ventilator-associated pneumonia

Management of Bacterial Pneumonia in Ventilated Patients (Maduri af 8/)

(MV).8 The high rate of misdiagnosis can be attributed to the fact that ventilated patients commonly develop clinical conditions other than pneumonia that manifest with fever and pulmonary infiltrate(s) and freqQ,ently have purulent tracheal secretions of nonpneumonic origin." Analysis of tracheal aspirates has proved Inaccurates-? with a high rate of false-positive results l ()'15 as well as being misleading regarding the cause of the pneumonia when present." Bronchoscopy with quantitative cultures of lower respiratory tract secretions obtained by either a protected specimen brush (PSB) or bronchoalveolar lavage (BAL) is, at present, the most reliable means of determining the presence of pneumonia in ventilated patients and defining its microbiologic identity" Reliable bronchoscopic findings should facilitate medical intervention by allowing either formulation of an appropriate narrow-spectrum antibiotic regimen tailored to the organism(s) recovered in significant concentrations in the respiratory secretions or by indicating the need to search for an alternative cause when findings are negative. We have previously reported our experience with protected bronchoalveolar lavage (PBAL) performed though a transbronchoscopic balloon-tipped (TBlj catheter with a distal ejectable diaphragm for collecting distal respiratory tract secretions with a minimal degree of contamination.V We have conducted a prospective study to determine (1) the effectiveness of PBAL in diagnosing pneumonia in ventilated patients and (2) the usefulness of bronchoscopic data in managing patients on MV who have clinical pneumonia. METHODS AND MATERIALS

lbtient Pvpulation The study was conducted at The University of Tennessee Health Science Center and the Veterans Affairs Medical Center, Memphis. Twenty-five intubated and mechanically ventilated patients in the Intensive Care Unit (ICU) were prospectively entered into the study between February 1989 and February 1990. Their mean age was 57 years. The average duration of mechanical ventilation before bronchoscopy was 12 days (range, 1 to 90 days). Entrance criteria were (1) clinical suspicion of pneumonia with fever (temperature ~ 38.3°C) and a new or progressive infiltrate on chest roentgenogram with either leukocytosis (> 10,000 ceUslcu mm) or a macroscopically purulent tracheal aspirate and (2) no antibiotic therapy for at least 48 h before undergoing diagnostic bronchoscopy. Giemsa and Gram stains provided early recognition of patients with possible pneumonia (presence of intracellular organisms), and empiric antibiotics were directed to the organisms identified on Gram stains. In patients with Significant growth on quantitative cultures, narrow-spectrum antibiotics were instituted based on results of sensitivity testing. In patients with negative bronchoscopic stains and cultures, a postbronchoscopy decision algorithm was created and followed to identify alternative source(s) of fever and pulmonary infiltration ('Jable 1). In these patients, antibiotic therapy was not instituted unless a nonpulmonary source of infection was identified (ie, urine, central line, peritoneum, meninges) or the patient developed acute clinical deterioration (hypotension, metabolic acidosis, or worsened respiratory status).

Collecting and Processing Specimens

Bronchoscopic Technique: Premedication, monitoring, and bronchoscopic equipment were used as previously described." The fiberoptic bronchoscope (FOB) was inserted through the endotraclieai tuhe(~8 mm) via a sterile connector. Suction and injection of lidocaine through the bronchoscopic channel were avoided. The FOB tip was positioned next to the orifice of the sampling area to visualize the entrance to the desired bronchial subsegment. Sampllng Area: The sampling area was selected based on the location of the new or progressive infiltrate on chest roentgenogram or the segment visualized during bronchoscopy as having purulent secretions in patients with more than one new infiltrate. The sequence of sampling was always PSB followed by PBAL. Patients with ARDS had bilateral sampling . The technique of PSB sampling in ventilated patients was described previously. 8,17 Protected Bronchoaloeolar Ltwage: The protected TBT catheter was introduced into the suction channel of the FOB and advanced into the desired subsegment until the black ring marking the proximal portion of the balloon was visible at the subsegmental orifice. The balloon was then inHated with 1.5 ml of air to occlude proximally the subsegmental bronchial lumen. The catheter was gently retracted to verify a tight seal . The distal diaphragm was then expelled by Hushing 2 ml of sterile saline solution (with a 5-ml syringe) through the irrigation lumen. PBAL was performed with five 300ml a1iquots of sterile saline solution, gently aspirating back after each instillation. The PBAL efBuent was equally divided for bacteriologic and cytologic analysis, and 2 ml was placed in a transport medium (Port-A-Cul vial, Becton and Dickinson Co, Cockeysville , Md) for anaerobic cultures. The specimens were delivered immediately to the microbiology laboratory. Laboratory analysis of respiratory secretions was described previously. 17 Two O.5-ml samples of PBAL Huid were cytocentrifuged for Gram and Giemsa stains.

Table 1- The UnWenity of Thnneuee MedicallCU Diagnoatic Protocolfor Mechanically Ventilated lbtientB with Fever and a NeID Pulmonary Infiltrate on Cheat Roentgenogram Who Haoe Negative BronchoKopic (PSB and PBAL) Findings· Extrapulmonary Source of Sepsis Blood culture and sensitivity x 2 from two different sites. Urine for analysis and culture. Remove central lines and submit the subcutaneous portion and tip of catheter for semiquantitative culture. Computed tomographic (CI) scan of the sinuses to rule out sinusitis. RoentgenograrnlCT scan of the abdomen In patients at high risk for abdominal source of sepsis: Postabdominal surgery Pancreatitis GI bleeding or cancer High-dose corticosteroids Or with abnormal results of abdominal examination Atelectasis Nebulized p-agonist every 4 h Chest percussion (over area with new infiltrate) every 2 hr X 4 Repeated chest roentgenogram after 4 to 8 h Deep Venous ThrombosislPulmonary Emboli Ventilation/perfusion lung scan Noninvasive evaluation for deep venous thrombosis (impedance plethysmography or Doppler) Drug Fever Discontinue nonessential medications and observe. ·Modified with permission from reference 8. CHEST I 101 121 FEBRUARY, 1992

501

Table 2-Criteria to Define the Pruence ofPneumonia Histologic 1. Foci of consolidation with intense polymorphonuclear leukocyte accumulation in the bronchioles and alveolar spaces Clin ical 2. Pleural Huid: organisms on Gram stain or culture identical to those recovered in significant concentration in the PSB or PBAL culture 3. Blood culture: positive, with organism identical to that recovered in significant concentration in the PSB or PBAL culture 4. Roentgenographic: rapid cavitation in the absence of lung carcinoma 5-7 . Significant growth on (5) PSB (~IO' cfulml) or (6) PBAL (~IO' cfulml), or (7) both PSB and PBAL with appropriate response to spec ific narrow-spectrum antibiotic therapy (roentgenographic clearance and resolution of fever)

Data Analysis Cultures were classified according to positive and negative results. The diagnostic thresholds for positive quantitative bacterial culture (a culture exibiting significant growth) were a growth ~IO' colony forming unit/ml (cfulml) for PBAL specimens and ~IO' cfulml for PSB specimens.•1 Growth below these values was called insignificant or negative. The Gram stain was defined as negative if no organisms were seen. The Gram stain was considered positive only if the identified organism(s) correlated with the morphology of organisms grown in significant concentration in the PBAL. The Giemsa stain was used to define the differential of the alveolar cell population and the presence of intracellular organisms. The presence of squamous epithelial cells and neutrophils with intracellular organisms was reported as a percentage of total cells recovered. Patients were divided into two groups based on definitive presence or absence of pneumonia. Patients with an undetermined diagnosis were so reported but not considered for statistical analysis. Criteria used to define the presence or absence of pneumonia are shown in 'Iables 2 and 3. Pneumonia was excluded definitively in the patient without a significant growth on PBAL or PSB when the fever and infiltrate resolved without instituting antibiotic therapy or when an alternative diagnosis was established. A diagnostic evaluation for every ventilated patient with negative bronchoscopic findings (infection, malignancy, or others) was performed to identify the alternative etiology for the fever and infiltrate (Thble 1). Atelectasis was defined as the resolution of the new pulmonary infiltrate within 48 h of bronchoscopy in response to inhaled

Table 3-Criteria to Define the Absence ofPneumonia Histologic 1. Lack of consolidation with intense polymorphonuclear leukocyte accumulation in the bronchioles and alveolar spaces 2. Criterion 1 plus a definitive alternative etiology Cytologic 3. Identification of a process other than pneumonia (eg, malignant neoplasm) by PBAL without significant bacterial growth on both PSBand PBAL Clinical Lack of significant growth on quantitative cultures of the PSB «10' cfulml) and PBAL «10' cfulml), with one of the following: Resolution, without antibiotic therapy, of: 4. Fever 5. Roentgenographic infiltrate 6. Fever and roentgenographic infiltrate 7. Roentgenographic infiltrate and a definitive alternative diagnosis, or 8. Persistent fever or roentgenographic infiltrate with a definitive alternative diagnosis

502

Table 4-Clinical and Roentgenographic CharacteriBticB at Time ofBronchoscopy Clinical or Roentgenographic Variables

Pneumonia* (n=9)

No Pneumonia* (n=I4)

Fever, % WBC > 10,000 cells/cu mm, % Purulent secretions, % Radiologic Bndmgst ARDS(n=8) New localized Worse Unchanged Other abnormal (n = 7) New localized Worse Unchanged Previously normal (n=7) New localized

100 77 100

100 92 69

0 1 1

2 3 1

0 2 1

2 2 0

4

3

*None of the differences between clinical and roentgenographic variables reached statistical significance (p value range from 0.13 to 0.99). tThe chest roentgenograms (22, one missing) were retrospectively reviewed by a single blinded observer. bronchodilator therapy and chest percussion. The time without antibiotic therapy before bronchoscopy and the type of treatment received after the procedure were recorded for every patient. Postmortem data were analyzed only if they were obtained within one week of bronchoscopy. Because data for PBAL and PSB were obtained from the same patients, comparisons of the two methods for sensitivity and specificity were made with McNemar's test for paired data. Differences in BAL neutrophil count and binary clinical variables between patients with and without pneumonia were tested with the Wilcoxon rank sum test and Fishers exact two-tailed test, respectively. RESULTS

Leukocytosis (> 10,000 cells/cu mm) was found in 77 percent of patients with pneumonia and in 92 percent of those without (Table 4). Purulent tracheal secretions were found in all patients with pneumonia and in nine of 13 patients (69 percent) without pneumonia. The pattern of roentgenographic changes is shown in Table 4. Only four of 12 patients (33 percent) with previously abnormal chest roentgenograms who developed new localized or worsening infiltrates were found to have pneumonia. In addition, only one of seven patients with air bronchograms had pneumonia. Differences in results between clinical and roentgenographic variables were not statistically significant (p value range from 0.13 to 0.99). Bronchoscopy with PBAL was well tolerated. One patient with ARDS developed oxygen desaturation and transient bradycardia during bronchoscopy. Bronchial hemorrhage developed in three patients after PSB (two with >20 ml, one minimal) and in one patient after PBAL (minimal). The average return of PBAL efBuent was 22 percent (range, 2.5 to 37.5 percent). Management of Bacterial Pneumonia In Ventilated PaIienls(Maduri st eI)

Table 5-Patients with Pneumonia: Results ofMicrobiologic Studies, ClinicalFindings, Pre- and Post-Bronchoscopy Management

Cultures (cfulml) Patient No.

Organism

Diagnosis Mode·

PSB

Cultures PBAL

Stains

PSB PBAL Gram

Giemsat

Antibiotic Therapy Purulent Leukocyte MV Sputum Count Days Days Off Postbronchoscopy

1 2

Saureus 5aureus

6 6

10Saureus 80Saureus

1x 10'Saureus 1x10'5 aureus

FN FN

TP TP

TP TP (10%) TP TP (15%)

Yes Yes

18 7.9

2 9

>10 4

3 4 5 6

H inf/uen::ae

25x 10' 5t pneumo 5x10'P aerugil108tl 10 x10'5aureus 10 x10'Enterobacter 53x 10'P cepacia

FN FN TP TP

TP TP TP TP

TP TP (5%) TP TP FN (2%) TP TP (30%)

Yes Yes Yes Yes

15.7 24.3 11.1 10

2 1 90 6

>10 >10 2 >10

Polymicrobial

No growth 305t pneumoniIle 1x 10'P aerugiII08tI 10 x 10'5aureus 12 x 10'Enterobacter 3x10'Pcepacia

sx10'H inf/uen::ae

7

6 6 1 3 7 7

TP

TP

TP TP (53%)

Yes

8.3

8

5aureus

205aureus

15 x 10'5aureus 13 x 10'5aureus

FN

TP

TP TP (7.5%)

Yes

17.9

7

2

9 10

5aureus Aspiration

80 Saureus 150 E coli 100 Klebsiella po

OOKlebsiella po 10 E coli

FN N

TP N

N

N (0%)

Yes Yes

13.3 19.1

10 3

2 2

5tpneumonille Pa aerugil108tl Polymicrobial

6 Undet

1x 10'5aureus

>10

Nafcillin Vancomycin Rifampin Ceftriaxone Erythromycin Amikacin Clindamycin Gentamicin Trimethoprimsulfamathoxazole Vancomycin Piperacillin Gentamicin Vancomycin Cefazolin Gentamicin

·Forexplanation ofnumber, refer toTable 2. Undet =undetermined; MV =mechanical ventilation;Organisms: E=Escherichia, H=Hemophilus, Pneumo =Pneumoniae, P=Pseudomonas, S=Staphylococcus, St=Streptococcus; F=false, T=true, N=negative, P=positive. t Percentage oftotal alveolar cells with intracellular organisms.

Diagnosis

A definitive diagnosis was established in 23 patients. Of the nine patients who had pneumonia, two were confirmed by histologic postmortem examination, one by positive blood cultures, and six by clinical and roentgenographic response to narrow-spectrum antibiotic therapy directed toward the organism(s) recovered by culture in significant concentration. Pneumonia was excluded by lung histologic study in seven patients (four open-lung biopsies, three autopsies). An alternative diagnosis was established in four patients (positive cytologic study for malignant cells of the PBAL effiuent in one patient and alternative infections that resolved with specific antibiotic therapy in three) (Table 4). Pneumonia was excluded by clinical resolution without antibiotic therapy in three patients. Two patients with undetermined diagnoses (patients 9 and 25) received empiric antibiotic treatment because of worsening respiratory status (one patient) or hypotension (one patient). Results of Bacterial Cultures

Results of quantitative bacterial cultures of the PBAL and PSB are shown in Tables 5 and 6. In the nine patients with pneumonia (patients 1 through 9), all the PBAL samples and three of the PSB specimens had a significant growth (Table 5). In the 14 patients without pneumonia (patients 11 through 24), none of the specimens had a significant growth (Table6). Eight PBAL and ten PSB specimens had no growth, while contamination was seen in six PBAL specimens (mean, 1521 efulml) and in four PSB specimens (mean, 100 efulml). Overall, the PSB had three true-positive, six

false-negative, and 14 true-negative results, while the PBAL had only true-positive and true-negative results, with a sensitivity and specificity of 100 percent. Eight of nine patients grew an identical organism in both PBAL and PSB. Differences in results of quantitative bacterial cultures based on PBAL and PSB were statistically significant (p< .04). Results of Gram and Giemsa Stains

Gram stains of the PBAL cytocentrifuged specimens were available in 22 patients and were positive in all cases of pneumonia (Tables 5 and 6). Gram stain morphology matched the organism that grew in significant concentration in cultures. Two of the three patients with polymicrobial pneumonia grew both organisms identified by Gram stain. No patient without pneumonia had a positive Gram stain. Gram stains of the PSB specimens were positive in only three patients, two with a positive PSB culture. Giemsa stains of the PBAL effiuents were available from 16 of 22 patients. The mean percentage of neutrophils was 67 ± 9.7 in patients with pneumonia and 37 ± 23 in patients without lung infection (p = .009). In the pneumonia group, intracellular organisms (ICO) were seen in 5 percent or more (mean, 20 percent) of the retrieved alveolar cells in six patients and 2 percent in one patient. In the group without pneumonia, seven of eight patients had no ICO, and one patient had only 1 percent ICo. Subsequent Course

Two of the nine patients with pneumonia died, one (patient 8) within 48 h of bronchoscopy and the other CHEST I 101 121 FEBRUARY, 1992

503

Table 6-Patients without Pneumonia: Resulu afMicrobiologic Studies, CUnical Findings, Pre- and Poat-Bronchoacopy Management-

Antibiotic Cultures (efulml) Patient No.

Diagnosis

Modet

TN TN

TN TN

TN TN

TN (0%)

No

12 19.5

5 21

>7 4

No growth

TN

TN

TN

TN (0%)

No

19.7

4

3

No growth 60 Candida 450 Serratia

TN TN TN

TN TN TN

TN TN TN

TN (0%) TN (0%)

Yes Yes

12.5 15.3 33.4

14 7 18

4 >2 2

5 2 8

No growth No growth 280Alphast IOSerralia 150 Enterobaeter 10Saureus No growth

No growth No growth No growth

TN TN TN

TN TN TN

TN TN TN

Atelectasis ARDS Lung cancer Atelectasis

5 2 2 5

No growth No growth No growth No growth

TN TN TN TN

TN TN TN TN

TN TN TN TN

Neurogenic ARDS Upper Gi bleed ? Aspiration

8

No growth

TN

TN

TN

Undet

No growth

No growth No growth 70S epidennidi.l 2x I(}I Kleb cnytoc;J ISO I'ratew mirabilir 4x I(}I Sfviridall.l I x I(}I EnterrJcocx:us SO Sfviridall.l

N

N

N

Lymph mels and linesepsis Lungcancer ARDS Atelectasis

8

17 18 19

Atelectasis ARDS Line sepsis UTI

20 21 22 23 24

14 IS 16

25

Purulent Leulcocyte MY PSB PBAL Gram Giemsat Sputum Count Days Daysotr PIlsthronchoscopy

1.4 x I(}I Saureus No growth

7 2

13

PBAL

Therapy

Stains

No growth SO Betastreptococcus No growth

Atelectasis ARDS

II

12

PSB

Cultures

3 2 2

Yes

No NA

II

TN (0%) TN (1%)

Yes Yes

12.8 16.5

18 10 17

3 3 2

TN (0%)

Yes Yes Yes

TN (0%)

No

NR 14.9 22.5 IS

4 IS 20 4

2 >10 3 3

Yes

7.9

3

2

No

19.9

9

2

N

None None Vancomycin for Sepi line sepsis None None None None None v.mcomycin for S_1ine sepsis None None None None Amibcin for SepimeningiliI Empiric IicarciIIin and gentamicin

'Diagnosis: Lymph mels =Iympbangiticmetastasis, UTI =urinary tract infection. t For explanation ofnumber, refer toThhle 3. Undet =undetermined; MY =mechanical ventilation;Organisms: Kleb =Klebsiella, S=Staphylococcus, St=Streptococcus; F=false, T=true, N=negative, P=positive. tPercentage of total alveolar ceUs withintracellular organisms.

(patient 5) after three weeks. Subsequent evaluation of the 14 patients without conclusive evidence of pneumonia confirmed an alternative diagnosis in all (Table6). Three patients had an extrapulmonary source of infection and were treated with a narrow-spectrum antibiotic. One patient had Staphylococcus epidermidis meningitis with bacteremia, and two patients had line sepsis. The remaining 11 patients had no evidence of infection and did not receive antibiotic treatment subsequent to bronchoscopy. Four patients with ARpS (patients 12, 15, 18, and 21) had histologic documentation (three open-lung biopsies, one autopsy) of the proliferative phase of diffuse alveolar damage without evidence of pneumonia. Five patients with negative bronchoscopic findings had a diagnosis of atelectasis. One patient had the combination of line sepsis and congestive heart failure and responded clinically to diuresis and narrow-spectrum antibiotics directed to the bacteria identified on blood and catheter cultures. Two patients were diagnosed as having bronchogenic carcinoma, one by bronchial washing and PBAL cytologic study and the second by needle aspirate. An additional patient with breast cancer had fever and bilateral infiltrates without response to empiric antibiotic treatment. She had no evidence of pneumonia at bronchoscopy. Results of the diagnostic survey were 504

normal, and the infiltrate was attibuted to lymphangitic spread. She received no subsequent antibiotic treatment. Five of the 13 patients without pneumonia died during the leu stay, four of ARDS and the one with breast cancer. DISCUSSION

The role of bronchoscopy with PSB and BAL in diagnosing pneumonia and in directing therapy in ventilated patients has been reviewed recently.SolS To appreciate the potential value of PBAL in diagnosing bacterial pneumonia, a comparison between PSB and BAL is necessary to understand the advantages and disadvantages of each technique. Comparison between PSB and BAL Sampling Area: PSB samples a limited area of the peripheral airways,19 while BAL samples a larger portion of the lung. The alveolar surface area distal to the tip of the wedged bronchoscope is estimated to be 100 times greater than that of the peripheral airways.P' Since PSB samples such a relatively small area, proper placement of the PSB catheter is essential to avoid false-negative results, particularly in patients without visible secretions, or the one with small infiltrates or previously abnormal chest roentgenograms where exManagement of Bacterial PneumoniaIn Yentilated PatIents(Meduri ef aI)

act anatomic location may be difficult. Amount of Retrieved Respiratory Secretions: PSB retrieves an estimated 0.01 to 0.001 ml of respiratory secretions." The dilution of lung secretions in the BAL fluid of patients without pneumonia is estimated to be 10- to lOO-fold.21.22 Therefore, the recovered BAL should contain at least 1 ml of respiratory secretions. Diagnostic Threshold: Multiple studies have shown that bacterial infections of the lung, as well as infections at other anatomic sites, contain 1(Y5 or more bacteria per milliliter of exudate. 23 •24 The PSB specimen is diluted in 1 ml of transport media before processing. A growth of 1()3 cfuJml, proven by independent investigators to be the diagnostic threshold for pneumonia.P-" represents 1(Y5 to 1()6 bacteria per milliliter of the original lung secretions. A colony count of 1()4 cfuJml in the BAL fluid, considering a dilution factor of 10 to 100, represents 1(Y5 to 1()6 bacteria per milliliter of lung secretions. Contamination: The PSB catheter, when properly used , collects respiratory secretions with a minimal degree of contamination. 17. 25 •26 Contaminants, on the other hand, are found in 89 percent of BAL samples of patients without pneumonia27-29 and frequently (30 percent) at a high concentration (~1()4 cfuJml).13.28.29 TIming ofDiagnosis: Results of quantitative cultures of respiratory secretions are not available until 48 h after the procedure, a significant delay that may negatively affect the care of a critically ill patient. Microscopic analyses of the BAL effiuents have been shown to be useful for early identification of patients with bacterial pneumonia and in directing antimicrobial therapy. In one study,30 the presence of intracellular organisms in 7 percent or more of the retrieved alveolar cells (on Giemsa stain) was found in 80 percent of patients with pneumonia and in only 4 percent of patients without pneumonia. Gram stain of the cytocentrifuged BAL has also been shown to accurately identify the organism(s) that later grew in significant concentration on bacterial cultures.P'-" Results of Gram stains of the PSB specimens have been variable .13.15.32,33 In summary, both PSB and BAL have their own advantages and limitations and together appear to be complementary" PSB retrieves an uncontaminated specimen; however, the sampling area is limited and the quantity of secretions collected is small, possibly leading to false-negative results. On the other hand, BAL samples a more extensive area of the lung parenchyma but frequently becomes contaminated, compromising the specificity of culture results. These theoretical concerns have been confirmed by animal and human studies. In a baboon model of respiratory failure, 32 the results ofquantitative cultures of BAL and PSB were compared with the histologic

and bacteriologic findings of open-lung biopsy. Six ventilated animals not receiving antibiotics had moderate to severe pneumonia and polymicrobial growth on culture. BAL recovered 74 percent of all species present in the lung tissue compared with 41 percent by PSB and 56 percent by needle aspirate. Bacterial growth in the BAL, expressed as a bacterial index (sum of the logarithmic concentrations of individual organisms), was linearly related to tissue values. In animals receiving antibiotics, absence of growth was frequently found despite histologic evidence of pneumonia. The investigators concluded that BAL specimens, by sampling a larger area of lung parenchyma, provided the most accurate window on lung microbiology (in animals not receiving antibiotics), both quantitatively and qualitatively. Three investigators have evaluated the role of BAL in diagnosing nosocomial bacterial pneumonia in patients receiving mechanical ventilation. 13,29,31 Torres et al13 have compared the diagnostic yield of PSB and BAL in 25 patients with ventilator-associated pneumonia recently placed on antibiotic therapy «12 h). Microbiologicprocessing of the BAL fluid was different from other previously described methods.28,29 Quantitative cultures were performed on a concentrated BAL (ten-fold) sample , and a threshold of 103 cfuJml (similar to the one for PSB) was used . A significant growth was found in 72 percent of patients with each tech nique and in two of seven.BAL samples in a control group (71 percent specificity). Agreement between BAL and PSB regarding the type of organism recovered was 75 percent, but agreement dropped to 56 percent when concentration of organisms was also considered. Chastre et al29 compared the results of BAL and PSB in a group of 18 ventilated patients who developed clinical manifestations of pneumonia, all of whom had not received antibiotics for at least ten days. A growth ~1()4 cfuJml was seen in 80 percent of patients with pneumonia and in 31 percent of patients without lung infection. Guerra and Baughman" obtained BAL specimens for quantitative bacterial cultures in 54 patients receiving mechanical ventilation who underwent bron choscopy. Eighteen patients had bacterial pneumonia. Using a diagnostic threshold ofl()4 cfuJml, a significant growth was seen in 16 patients (89 percent) with pneumonia and in no patients without pneumonia. Of patients with bacterial pneumonia, 70 percent were receiving antimicrobial therapy, and none of the significant organisms (> 1()4 cfuJml) was sensitive to the administered antibiotics. Of patients in the control group, 75 percent were receiving antimicrobial therapy, and this treatment was discontinued after results of cultures were available. Combining the results of quantitative cultures of CHEST I 101 12 I FEBRUARY, 1992

505

these studies, BAL had a sensitivity of 77 percent and a specificity of 69 percent. The false-positive rate in patients without pneumonia was 31 percent.P-" It is also well recognized that previous antibiotic treatment affects the results of bacterial cultures and may preclude the recovery of organisms in respiratory secretions at a significant concentration. 32 ,34 ,35 Moreover, false-positive results have been reported previously in up to 50 percent of PSB25,36 and 30 percent of BAL specimens" in patients receiving antibiotics without clinical pneumonia.

Protected BAL in lbtients with Respiratory Failure Our study was designed to investigate the effectiveness of a novel technique , PBAL, in evaluating nosocomial pneumonia in ventilated patients. We believed that minimizing contamination of the BAL fluid should improve the specificity and negative predictive value of unprotected BAL while maintaining better sampling characteristics compared with PSB. The following discussion centers around the two goals established for this investigation: (1) the effectiveness of PBAL in diagnosing pneumonia in ventilated patients and (2) the usefulness of bronchoscopic data obtained in managing patients with clinical pneumonia receiving MY. Effectiveness of PBAL in Diagnosing Pneumonia in Ventilated lbtients: PBAL was effective in retrieving distal airway secretions with a minimal degree of contamination. In 14 patients without pneumonia, bacerial growth was absent in eight and nonsignificant (::54,000 cfulml) in six. Evidence of contamination of the BAL effluent by oropharyngeal flora (> 1 percent squamous epithelial cells) was found in 2 of 16 patients in this study, both with a insignificant bacterial growth on quantitative cultures. The low degree of contamination gave this technique a specificity and a negative predictive value of 100 percent. PBAL recovered bacterial organisms in all patients with pneumonia. Isolates grew in all patients at a concentration 2::100,000cfulml, with a median growth of 500,000 cfulml. Two patients had a polymicrobial growth (22 percent) similar to the findings of others. 1,13 ,26 .29,32 The sensitivity and positive predictive value of PBAL was 100 percent. The presence of a two-log difference between the highest quantitative culture count in patients without pneumonia and the lowest quantitative culture count in patients with pneumonia allows a clearer determination of a patient's status, with regard to pneumonia, compared with the significant overlap in unprotected BAL. The PBAL technique was superior to PSB in recovering organisms (p<.04). The PSB specimens grew organisms identical to the organisms of the PBAL in all but one sample. However, ·the concentration of bacteria was below the diagnostic 506

threshold of 1()3 cfulml in all but three patients. Previous antibiotic treatment may have affected the results of PSB specimen cultures," While all of our patients did not receive antimicrobial therapy for at least 48 h, this may not have been sufficient time to permit significant growth on PSB cultures in patients with pneumonia. This finding may have important clinical implications since three of nine patients with pneumonia were receiving antibiotic treatment for reasons other than a respiratory infection when they developed fever and a pulmonary infiltrate. Although the results of Guerra and Baughman" suggest that positive quantitative cultures can be obtained if the causative organism is resistant to the antibiotics, the effect of previous antibiotic use on quantitative growth of susceptible bacteria remains unknown. Usefulness of Bronchoscopic Data in lbtient Management: This study also addressed how data obtained by PBAL may affect management of ventilated patients with and without pneumonia. Gram and Giemsa stains of the PBAL were positive in all patients with pneumonia and negative in those patients without pneumonia. Rapid and accurate identification of pneumonia etiology facilitated decisions regarding the initiation of empiric antibiotic therapy while awaiting results of the quantitative cultures. Empiric therapy was avoided in the majority of patients without documented pneumonia (14 of 16), while antibiotic selections were more narrow spectrum and rational in those patients with pneumonia. Table 5 shows the organism(s) identified by PBAL and type of treatment received by the nine patients with pneumonia. All but one received narrow-spectrum antibiotics. Patient 8 had a rapid deterioration and failed to respond to empiric treatment initiated soon after the procedure. Thirteen patients had no bronchoscopic evidence of pneumonia. The diagnostic survey identified an infection outside of the lung in only three, for which they received narrow-spectrum treatment (Table 6). The other ten patients had conditions that did not require antimicrobial therapy: atelectasis (five), the proliferative phase of ARDS (four), and lung cancer (one). Withholding antibiotic treatment in these stable patients who had clinical manifestations suggestive of pneumonia but normal bronchoscopic findings (PSB, PBAL) had no adverse effects. Empiric broad-spectrum antimicrobial therapy cannot be considered innocuous, particularly in patients already at high risk for developing infections. Prior antibiotic use is associated with a significantly greater incidence of ventilator-associated pneumonia due to more serious organisms and the higher mortality rate attributable to those bacteria. In addition, falsely attributing fever and pulmonary infiltrates to pneumonia may lead to inappropriate treatment or misdiagnosis of other infectious or noninfectious problems in these patients Management of Bactarial Pneumonia in Ventilated PetIents (Maduri at aI)

with complex illnesses.37.38 Bronchoscopy with PBAL was well tolerated; one patient, however, developed oxygen desaturation and transient bradycardia. A drop in Pa02 during or after BAL is a common complication encountered in patients receiving MV undergoing diagnostic bronchoscopy and is linked to the severity of pulmonary dysfunction, size of the endotracheal tube, and reduction in alveolar ventilation.3 1•39 •40 The bronchoscope should be withdrawn immediately in patients who develop desaturation, arrhythmias, or hemodynamic instability.

10

11

12 13

CONCLUSION

Our findings support the use ofPBAL as a diagnostic tool in ventilated patients with clinical manifestations suggestive of pneumonia. Our results can be applied only to patients who have not received antibiotics for at least 48 h before the procedure. The value ofPBAL in patients receiving current antibiotic therapy needs to be established. Clinical and roentgenographic criteria could not discriminate between patients with and without pneumonia, confirming the findings of previous investigations}·26 The results of microscopic and culture analyses of the PBAL proved useful in directing antibiotic treatment in patients with pneumonia and in avoiding unnecessary antibiotic use in those patients without pneumonia. AKNOWLEDGMENTS: The writers thank Carol B. Jones, R.N .• B.S .N ., for data collection; Dr. Elizabeth A. Tolley for statistical analysis; Dr. Barbara J. Kuyper for helpful critique; and Ms. Vicky Franke for her secretarial expertise. REFERENCES

1 Fagon JY, Chastre J, Domart Y, Trouillet JL, Pierre J, Dame C, et al. Nosocomial pneumonia in patients receiving continuous mechanical ventilation: prospective analysis of 52 episodes with use of a protected specimen brush and quantitative culture techniques. Am Rev Respir Dis 1989; 139:877-84 2 Torres A, Aznar R, Gatell JM, Jimenez P, Gonzalez J, Ferrer A, et al. Incidence, risk, and prognosis factors of nosocomial pneumonia in ventilated patients. Am Rev Respir Dis 1990; 142: 523-28 3 Ashbaugh DG, Petty TL . Sepsis complicating the acute respiratory distress syndrome. Surgery 1972; 135:865-69 4 Bell RC , Coalson Smith JD, Johanson WG Jr. Multiple organ system failure and infection in adult respiratory distress syndrome. Ann Intern Med 1983; 99:293-98 5 Andrews CPo Coalson Smith JD, Johanson WG Jr. Diagnosis of nosocomial bacterial pneumonia in acute, diffuse lung injury. Chest 1981; 80:254-58 6 Celis R, Torres A, Gatell JM, A1mela M, Rodriquez-Roisin R, Agusti-Vidal A. Nosocomial pneumonia: a multivariate analysis of risk and prognosis. Chest 1988; 93:318-23 7 DiNubile MJ. Antib iotics: the antipyretics of choice? Am J Med 1990; 89:787-88 8 Meduri GU . Ventilator-associated pneumonia in patients with respiratory failure: a diagnostic approach. Chest 1990; 97:120819 9 Sottile FD, Marrie 1}, Prough DS, Hobgood CD, Gower DJ, Webb LX, et al. Nosocomial pulmonary infection: possible

n,

n,

14

15

16

17

etiologic significance of bacterial adhesions to endotracheal tubes. Crit Care Med 1986; 14:265-70 Hill JD, RatliffJL, Parrott JC\v, Lamy M, Fallat RJ, Koeniger E, et al. Pulmonary pathology in acute respiratory insufficiency: lung biopsy as a diagnostic tool. J Thome Cardiovasc Surg 1976: 71:64-71 Villers D, Derrienic M, Raffi F, Germaud P, Baron D, Nicolas F, et al. Reliability of the bronchoscopic protected catheter brush in intubated and ventilated patients. Chest 1985; 88:52730 Berger R, Arango L . Etiologic diagnosis of bacterial pneumonia in seriously ill patients. Crit Care Med 1985; 13:833-36 Torres A, De La BeI1acasaJP, Xaubet A, Gonzalez J, RodriquezRoisin R, Jimenez de Anta MT, et al. Diagnostic value of quantitative cultures of broncho-alveolar lavage and telescoping plugged catheters in mechanically ventilated patients with bacterial pneumonia. Am Rev Respir Dis 1989; 140:306-10 Torres A, Puig de la Bellacasa J, Ferrer M, Gene A, Gonzales J, Celis R, et al. Do quantitative cultures of endotracheal aspiration samples coincide with quantitative cultures of protected specimen brushes in mechanically ventilated patients with pneumonia. Am Rev Respir Dis 1990; 141:A277 Higuchi JH, Coalson Johanson WG Jr. Bacterial diagnosis of nosocomial pneumonia in primates: usefulness of the protected specimen brush. Am Rev Respir Dis 1982; 125:53-7 Eason AL, Wunderink RG , Meduri GU, Leeper Kv. Overrepresentation of Gram negative enterics in suspected ventilatorassociated pneumonia. Chest 1991; lOO:36S Meduri GU, Beals D , Maijub G , Baselski V. Protected bronchoalveolar lavage: a new bronchoscopic technique to retrieve uncontaminated distal airway secretions. Am Rev Respir Dis

n,

1991; 143:855-64 18 Chauncey JB , Lynch JP III, Hyzy RC, Toews GB. Invasive techniques in the diagnosis of bacterial pneumonia in the intensive care unit. Semin Respir Infect 1990; 5:215-25 19 Wimberley N, Faling LJ, Bartlett JG . A fiberoptic bronchoscopy technique to obtain uncontaminated lower airway secretions for bacterial culture. Am Rev Respir Dis 1979; 119:337-43 20 Reynolds HY. Bronchoalveolar lavage. Am Rev Respir Dis 1987; 135:250-63 21 Baughman RP, Bosken CH, Loudon RG, Hurtubise P, Wesseler T. Quantitation ofbronchoalveolar lavage with methylene blue. Am Rev Respir Dis 1983; 128:266-70 22 Rennard S, Basset G, Lecossier D, et al. Estimations of the absolute volume of epithelial lining 8uid recovered by bronchoalveolar lavage using urea as an endogenous marker of dilution. J Appl Physiol 1986; 60:532 23 Bartlett JG . Invasive diagnostic techniques in pulmonary infection s. In : Pennington JE, ed . Respiratory infections: diagnosis and management. 2nd ed . New York: Raven Press, 1989, 52-68 24 Bartlett JG, Finegold SM. Bacteriology of expectorated sputum with quantitative culture and wash technique compared to transtracheal aspirates. Am Rev Respir Dis 1978; 117:1019-27 25 Chastre J, Viau F, Brun P, Pierre J, Dauge MC, Couchama A, et al. Prospective evaluation of the protected specimen brush for the diagnosis of pulmonary infections in ventilated patients. Am Rev Respir Dis 1984; 130:924-29 26 Fagon JY, Chastre J, Hance AJ, Guiguet M, Trouillet JL, Domart Y, et al. Detection of nosocomial lung infection in ventilated patients: use of a protected specimen brush and quantitative culture techniques in 147 patients. Am Rev Respir Dis 1988; 138:110-16 'J:l Kirkpatrick MB, Bass JB Jr. Quantitative bacterial cultures of bronchoalveolar lavage 8uids and protected brush catheter specimens from normal subjects. Am Rev Respir Dis 1989; 139: 5MH8 28 Kahn F\v, Jones JM. Diagnosing bacterial respiratory infection CHEST I 101 121 FEBRUARY, 1992

507

by bronchoalveolar lavage. J Infect Dis 1987; 155:862-69 29 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 30 Chastre J. Fagon JY. Soler P, Domart Y. Pierre J. Dombret MC. et al. Quantification of BAL ceUs containing intraceUular bacteria rapidly identifies ventilated patients with nosocomial pneumonia. Chest 1989; 95:190S-1925 31 Guerra LF. Baughman RP. Use of bronchoalveolar lavage to diagnose bacterial pneumon ia in mechanically ventilated patients. Crit Care Med 1990; 18:169-73 32 Johanson WG Jr. Seidenfeld JJ. Gomez P, De Los Santos R. Coalson JJ. Bacteriologic diagnosis of nosocomial pneumonia foUowing prolonged mechanical ventilation. Am Rev Respir Dis 1988; 137:259-64 33 PoUock HM . Hawkins EL . Bonner JR. Sparkman T. Bass JB Jr. Diagnosis of bacterial pulmonary infections with quantitative protected catheter cultures obtained du ring bronchoscopy. J Clin Microbioll983; 17:255-59 34 Bartlett J. Diagnostic accuracy of transtracheal aspiration bacteriologic studies. Am Rev Respir Dis 1977; 115:777-82

508

35 Fox RC. Williams GJ. Wunderink RG. Leeper ICY, Jones CB. FoUowup bronchoscopy predicts therapeutic outcome in ventilated patients with nosocomial pneumonia. Am Rev Hespir Dis 1991; 143:Al00 36 Martos lA. Ferrer M. Torres A. Gonzales J. Puig De La Bellacasa J. Celis R. et al. Specificity of quantitative cultures of protected specimen brush and bronchoalveolar lavage in mechanically ventilated patients. Am Rev Respir Dis 1990; 141:A.276 37 Meduri GU. Belenchia JM. Estes RJ. Leeper ICY, Wunderink RG. Fibroproliferative phase of ARDS: clinical findings and effects of corticosteroids. Chest 1991; 100:943-52 38 Mauldin GL. Meduri GU. Wunderink RG. Leeper ICY, Jones CB. Estes, RJ. Causes of fever and pulmonary inBltrates in mechanically ventilated patients. Am Rev Respir Dis 1991; 143: Al00 39 Trouillet JL. Guiguet M. Gibert C. Fagon JY. Dreyfuss D. Blanchet F. Chastre J. Fiberoptic bronchoscopy in ventilated patients : evaluation of cardiopulmonary risk under midazolam sedation. Chest 1990; 97:927-33 40 Lindholm CE o Ollman B. Snyder ]v, Millen EG. Grenvik A. Cardiorespiratory effects of 8exible fiberoptic bronchoscopy in critically ill patients. Chest 1978;74:362-66

Management of Bacterial Pneumonia In VentilatlId PatientII (Medurl et aI)