PENETRATION OF AN AEROSOL, PRODUCED BY FILM ATOMIZATION, THROUGH THE CARINAL BIFURCATION

Transcription

1 British Journal of Anaesthesia 1993; 70: PENETRATION OF AN AEROSOL, PRODUCED BY FILM ATOMIZATION, THROUGH THE CARINAL BIFURCATION T. J. J. INGLIS, J. G. JONES AND S. PAXTON SUMMARY We have measured the size of spots produced by atomization of simulated biofilm particles from a trachea/ tube. Draughtsman's ink was atomized from 8.5-mm i.d. Portex trachea/ tubes, using an airflow of 1 litre s~', and trapped on vertical acetate sheets. The spots produced by these particles were compared with the spot size after entrainment and dissemination from an ink-lined trachea/ tube through a bifurcating tracheobronchial tube towards a glass plate. Spot dimensions were measured by computer-assisted image analysis of a video image. The smallest particles observed on the acetate sheets were 4~17 ftm in size. In the bifurcating tube experiment, large particles were deposited around the carinal bifurcation, but smaller particles traversed the bend. The smallest spot size observed (7 urn minimum; 17 /im median) corresponds to particles small enough to penetrate further into the lower respiratory tract. Travel via a tube similar to the carinal bifurcation suggests that fragments of biofilm entrained in the trachea/ tube could be propelled deeper into the respiratory tract during the inspiratory phase of mechanical ventilation. (Br. J. Anaesth. 1993; 70: ) KEY WORDS Airway: aerosol penetration Equipment: trachea/ tube. Critically ill patients are at increased risk of nosocomial pneumonia. Although pneumonia in patients undergoing artificial ventilation is often attributed to bacteria from endogenous flora, the mechanism by which these organisms are finally disseminated into the lungs has not been denned. It has been suggested that the material lining the inner surface of the tracheal tube ("tracheal tube biofilm") can be dislodged and carried further into the lung during suction catheter movement [1]. More recently, it was shown that the gas-liquid interaction between airflow along the tracheal tube and biofilm lining the tube could dislodge bacteria from the biofilm and scatter this material over a range of many centimetres [2]. That study did not explore the likelihood of dissemination through a branched tube system such as the trachea and bronchi. In the present study, biofilm dissemination from the tracheal tube was simulated to determine the size of the smallest particles produced and to examine if they would traverse a Y junction constructed to the dimensions of the trachea and main bronchi. MATERIALS AND METHODS Atomization Particle atomization produced by an interaction between gas flowing through a liquid lining the tracheal tube was studied in the following manner. A preweighed Portex tracheal tube of 8.5-mm i.d. was straightened and pointed horizontally towards a vertical acetate sheet held at 32 cm from thetipof the tube. A 1.0-ml bolus of draughtsman's ink (Rotring GmbH, Germany) was placed in the distal 3 cm of tube, and allowed to pool along the tube. A ventilator (Erica, Engstrom, Denmark) was connected to the tracheal tube connector using standard ventilator tubing and run for a single cycle to give aflow rate of 1 litre s" 1 (i.e. a tidal volume of 1 litre in 1 s) measured with a Fleisch II pneumotachograph.-the acetate sheet was removed immediately and air dried while held at room temperature in a horizontal position. This procedure was repeated nine times under identical conditions, using a fresh tracheal tube each time. After each experiment, the tube was sealed with paraffin film to contain the remaining liquid during transport and reweighed after removal of the film to calculate the mass of liquid entrained. A further series of experiments was carried out when particles entrained from the tracheal tube were directed via a horizontal bifurcating tube towards a vertical sheet of 2-mm picture glass. The bifurcating tube used in the procedure was constructed to the dimensions of the trachea and main bronchi as described by Weibel [3]. The bifurcating tube was designed and constructed in the workshops of the Royal Aircraft Establishment, Farnborough, and had a constant cross sectional area between "trachea!" and "bronchial" levels. The Portex tracheal tube was held horizontally and placed in the "tracheal" part of the bifurcating tube, after which the cuff was inflated to seal the tube in the model trachea. The glass sheet was placed perpendicular to the direction of flow, 2.5 cm from the distal end of each bifurcation (the total distance between tracheal TIMOTHY J. J. INGLIS*, D.M., M.R.C.PATH., D.T.M.&H. (Department of Microbiology); STEPHEN PAXTON (Department of Anatomy); University of Leeds, Leeds LS2 9JT. J. GARETH JONES, M.D., F.R.C.P., F.R.C.ANAES., Department of Anaesthesia, University of Cambridge, Addenbroke's Hospital, Cambridge CB2 2QQ. Accepted for Publication: December 3, *Present address, for correspondence: Department of Microbiology, National University of Singapore, Lower Kent Ridge Road, Singapore 0511.

2 528 BRITISH JOURNAL OF ANAESTHESIA -/\vi%- FIG. 1. Ink scatter produced at 32 cm from tips of adult tracheal tubes, replicated nine times. (For details of the particles, see table I.) tube tip and glass being 32 cm), and inspiratory flow was simulated by running the ventilator for a single cycle. The same ink volume, loading technique, ventilator and cycle were used in this experiment as described above. The experiment was repeated without the intervening Y-piece, and a fresh glass sheet held at 32 cm from the tracheal tube tip. Image analysis The nine acetate sheets were examined initially by light microscope (magnification x 400, with graticule eyepiece) to obtain an estimate of the range of spot sizes. To avoid losing spot images as a result of buckling or detachment from the acetate sheets, the sheets were photographed with 35-mm black-andwhite film. The negatives were then digitized using the CCTV camera and a dissecting microscope (Wild, Heerbrugg, Switzerland) and scanned in a single straight line from the area of most dense particle deposition at the centre of the scatter pattern to its furthest edge. The smallest spots were selected visually from the video screen and at least four possible spots were measured from each pattern. Total object number, total object area, mean object area and the D value (diameter of a circle of area identical to that of the scanned image of each spot) were obtained using the image analyser. Pixel dimensions were 0.6 x 1.1 urn. The software used was Vidi-pp version 1.0 (ULIS, Leeds) running under DOS 4.01 on an IBM-compatible microcomputer (Elonex PC 486B/25) with a framegrabber (ROMBO, Livingstone, Scotland), a CCTV monochrome camera (Hitachi, U.K.), and a photomicroscope (photomicroscope model II, Zeiss, Germany). Glass sheets were used in the Y-tube procedure to avoid problems with buckling and loss of ink spots during microscopy. Glass was also used to reduce any electrostatic repulsion of smaller particles by the surface on which impaction would occur. Spots produced by particles passing through the bifurcating tube were deposited in a broad crescentshaped area. A 2.5 x 2.5-mm square was chosen for image analysis in the wider part of the crescent, avoiding the larger spots, as these were large enough to obscure several smaller ones. This area was marked on the plate and subjected to image analysis, as above. RESULTS Extensive entrainment of liquid lining the tube resulted in extensive scatter of ink spots each time the experiment was repeated, although there was some variation in the quantity of material deposited on the acetate sheets (fig. 1, table I). In the bifurcating airway experiment, most spots large enough to be seen easily with the naked eye were deposited on the Y-tube wall at the bifurcation (fig. 2). Nevertheless, a collection of smaller spots was obtained on the glass plates positioned opposite the "bronchial" branches of the Y-tube. Most of these particles could not be seen easily without magnification. When their distribution was com-

3 529 DISSEMINATION OF TRACHEAL TUBE BIOFILM TABLE I. Characteristics of ink particles produced in each of nine replicate scatter experiments depicted in figure 1 Scatter pattern A B C D E F G H I Minimum Liquid entrsined (g) ciy\r diflltl i '* * * T i 11 Spot area (mm1) Gun) No. spots Total Mean pared with the scatter pattern obtained at the same distance without an intervening bifurcating tube, there was a readily visible difference between the spot size distribution on the respective plates (fig. 3). The minimum diameter of particles disseminated from the tracheal tube following flow through the Ypiece was 7 um (mode was 17 um) (fig. 4). DISCUSSION Particle deposition in the lungs of spontaneously ventilating subjects has been investigated previously [4] and two-phase gas liquid flow recognized as a mechanism for mucus transport in the airways of patients with increased thickness of the mucus layer [5]. There is some indirect evidence that two-phase gas-liquid flow occurs commonly in tracheal tubes containing a luminal biofllm [6], but the atomization and trajectory of biofilm particles in patients undergoing mechanical ventilation has not yet received detailed attention. An indwelling tracheal tube may increase substantially the likelihood of propulsion of atomized tracheal tube biofilm into the lungs, by FIG. 2. Deposition of ink particles at the bifurcation of a Y-shaped tube. bypassing the inertial impaction filter system of the upper respiratory tract in these patients. This simulation study suggests that low viscosity biofilm particles small enough to reach far into the lung can be disseminated with a velocity sufficient to -V. FIG. 3. Ink particle scaner obtained before (left) and after (right) placement of a bifurcating tube between the tracheal tube tip and a sheet of glass.

4 BRITISH JOURNAL OF ANAESTHESIA I5 CO CD Diameter ( FIG. 4. Size distribution (D value) histogram of ink spots obtained in a 2.5 x 2.5-mm area on glass scatter plate after passage through bifurcating tube. traverse the first order bifurcation in a model of an intubated trachea with symmetrically branched main "bronchi". The size of spots obtained in this series of experiments may overestimate particle size, because there may be some degree of particle spread at the moment of impact. A study of deposition of nebulized aerosols in spontaneously breathing volunteers found that 44 % nebulizer contents reached the lungs when the particle size was 10.3 urn mass median diameter [7]. In this study, percentage deposition increased with decreasing aerosol particle size. In another radiolabelled particle deposition study, particles of im (approximately the size of single bacterial cells) were found to reach the alveoli without significant loss, whereas particles > 12 urn were not able to penetrate the alveoli [8]. The various equations used to predict particle deposition in the Weibel morphological model of the human lung have been applied to inhalation of particles in the 5-um range [9], which would be expected to result in maximum deposition at the segmental and subsegmental bronchial level. If these results can be applied to particles entrained from tracheal tube biofilm, the majority would not be expected to penetrate further than this level; however, altered gas flow pathway and pattern expected during mechanical ventilation may alter substantially particle deposition. The process by which particles are produced in the tube can be described as gas-liquid interaction, in which the degree of interaction is sufficient to cause fracture of a gas-liquid interface (the luminal biofilm surface) with onward propulsion of liquid particles by the gas phase. Physical factors governing gas-liquid interaction in tubes with a liquid lining have been determined previously, and it was shown that the minimum thickness of liquid lining layer required for gas liquid interaction was 0.5 mm [10]. The critical flow required to generate gas liquid interaction in an 8.5-mm i.d. tube with a 0.5-mm liquid lining layer is equivalent to a Reynolds number of > A Reynolds number of 2000, synonymous with turbulent flow, is achieved readily in a tube of this size during mechanical ventilation, and at a flow of 1 litre s'1 the Reynolds number is more than This greater Reynolds number indicates a greater energy loss through gas-liquid interaction, with fluid expulsion from the tracheal tube and the production of mist or particleflow[11]. An important determinant of particle entrainment is the viscosity of the lining liquid. In clinical practice, tracheal tube biofilm is a heterogenous matrix and a wide range of liquid viscosity would be expected in a single biofilm. The biofilm layer adjacent to the luminal surface and therefore most subject to fragmentation is also likely to be the least viscous, because of factors such as condensation of humidifier water and the instillation of saline during tracheal suction procedures. The simulation experiments described in the present study were conducted therefore with a liquid of viscosity similar to that of water. More viscous biofilms are less likely to undergo fragmentation under the same conditions. Our results provide evidence to suggest that gas-liquid interaction in the tracheal tube provides an effective means of disseminating particles into the ventilated lung. Further work on particle entrainment and scatter from tracheal tube biofilm under different conditions of gas flow is required. The dissemination of bacteria and other components of tracheal tube biofilm as a result of physical processes occurring in the tracheal tube requires consideration as a possible step in the pathogenesis of nosocomial pneumonia. REFERENCES 1. Sottile FD, Marrie TJ, Prough DS, Hobgood CD, Gowr DJ, Webb LX, Costerton JW, Gristina AG. Nosocomial pulmonary infection: possible etiological significance of bacterial adhesion to endotracheal tubes. Critical Care Medicine 1986; 14: Inglis TJJ, Millar MR, Jones JG, Robinson DA. Tracheal

5 DISSEMINATION OF TRACHEAL TUBE BIOFILM 531 tube biofilm as a source of bacteria] colonization of the lung. Journal of Clinical Microbiology 1989; 27: Weibel ER. Morphometry of the Human Lung. Berlin: Springer-Verlag, Swift DL. Generation and respiratory deposition of therapeutic aerosols. American Review of Respiratory Disease] 1980; 122: Kim CS, Rodriguez CR, Eldridge MA, Sackner MA. Criteria for mucus transport in the airways by two-phase gas-liquid flow mechanism. Journal of Applied Physiology 1986; 60: Inglis TJJ. Evidence for dynamic phenomena in residual tracheal tube biofilm. British Journal of Anaesthesia 1993; 70: Clay MM, Clarke SW. Effect of nebulised aerosol size on lung deposition in patients with mild asthma. Thorax 1987; 42: Heyder J. Mechanisms of aerosol particle deposition. Chest 1981; 80: S820-S Agnew JE, Pavia D, Clarke SW. Aerosol particle impaction in the conducting airways. Physics in Medicine and Biology 1984; 29: Clarke SW, Jones JG, Oliver DR. Resistance to two-phase gas-liquid flow in airways. Journal of Applied Physiology 1970; 29: 464-^ Selsby D, Jones JG. Some physiological and clinical aspects of chest physiotherapy. British Journal of Anaesthesia 1990; 64: