Upper Respiratory Tract

This region originates at the nostrils and mouth and extends through the larynx; a diagram of the human upper respiratory tract is shown in figure 1. Air entering the nostrils passes first through the vestibule, the narrowest cross-section in the entire nasal region, before entering the main nasal passages. These consist of two airways separated almost symmetrically by the nasal septum. They are convoluted (due to the folds of the nasal turbinates), downward-curving shelf-like structures, resulting in a large surface area and a relatively narrow distance between opposing airway walls. Here, exchange of heat and moisture modify the temperature and humidity of the inhaled air. The nasopharynx begins at the posterior end of the turbinates, where the septum ends and the nasal passage narrows and turns downward. Although the basic structure of the nasal airways is similar in humans and most other mammals, there are considerable interspecies differences in the relative position, shape, and size of individual components, as shown in figure 2. For example, the nasopharynx in the rat encompasses a greater percentage of the total length of the nasal passages than in the human, whereas that in rabbits and dogs is intermediate between rats and humans.

Figure 1.

Diagram of the human upper respiratory tract.

Figure 2.

Silicone rubber replica casts of the nasopharyngeal region of different species: (A) human; (B) rabbit; (C) rat; (D) guinea pig; (E) hamster; (F) baboon; scale in cm. (Adapted with permission from Patra 1986, and from Hemisphere Publishing Corporation.) (more...)

The oral passages begin at the mouth and are characterized by much greater inter-and intraindividual variation in shape and cross-section than the nasal passages. At the posterior of the mouth, inhaled air enters the oropharynx. The oro- and nasopharynx join to form the hypopharynx, an airway that extends to the entrance of the larynx. The latter extends to the trachea and has a variable cross-section depending upon the rate of airflow through it.

Tracheobronchial Tree

The tracheobronchial tree consists of airways from the trachea through the terminal bronchioles. The trachea divides into two main bronchi which then enter the lungs at the hilar region. These main bronchi further subdivide into smaller airways. Support for the trachea and bronchi are derived from cartilagenous rings or plates. As the bronchial tree proceeds distally, the cartilage eventually disappears, and these airways—the bronchioles—are supported by smooth muscle. In humans, the transition from bronchi to bronchioles occurs in airways of ~1-mm diameter.

Simplistically, the tracheobronchial tree can be considered to be a system of tubes connected at specific division points. In most cases, division is by dichotomy, whereby a single branch (the parent) gives rise to two branches (the daughters). To describe this structure, the position of an individual airway is usually assigned a code number. There are two basic coding systems: the numbering of divisions up from distal end branches or, alternatively, down from the trachea. For example, in the of-ten-used Weibel ordering system (Weibel 1963), each branching division is known as a generation; the trachea is generation 0, and each distal division increases by one number.

In a dichotomous branching system, the pattern can vary in terms of the degree of symmetry (figure 3). If both daughters have the same diameter and length, and branch from the parent at the same angle, the mode of division is known as regular or symmetrical. If the two daughters differ from each other in one or more dimensions, the mode of branching is termed irregular or asymmetrical, the extreme case of which is monopody. In a monopodial branching system, the larger-diameter daughter (major daughter) may not be easily distinguishable from the parent since the change in diameter and direction from the parent may be negligible.

Figure 3.

Schematic diagram of tracheobronchial tree branching patterns: (A) human lung; (B) monopodial system common in experimental animals.

A major difference in respiratory tract anatomy between humans and most other mammals commonly used in inhalation studies is in the pattern of bronchial airway branching. Figure 4 shows casts of the upper bronchial tree in humans and in a number of other species, and figure 5 presents a quantitative analysis that allows characterization of branching patterns. In a regular dichotomous branching system, the ratio of daughter diameters is 1, whereas in a perfect monopodial branching system, the ratio of major daughter diameter to parent diameter is 1. The human bronchial tree, at least for the first six generations, exhibits the most symmetrical branching of all of the species shown, whereas the dog's bronchial tree is almost ideally monopodal. The other species exhibit various degrees of irregularity. Recent qualitative observations on the tracheobronchial trees of two nonhuman primates—the rhesus monkey and the baboon—suggest a branching pattern that is more irregular than that of humans, but not to the extent of the experimental animal species shown in figure 5 (Patra 1986). But although there may be striking interspecies differences in the upper bronchial tree, the branching patterns in most mammals tend to approach more regular symmetry in distal conducting airways.

Figure 4.

Silicone rubber replica casts of the tracheobronchial tree of different species. (A) human; (B) baboon; (C) dog; (D) rhesus monkey; (E) rabbit; (F) guinea pig; (G) rat; (H) hamster; (I) mouse. Both photos are reproduced here at the same scale, given in (more...)

Figure 5.

Morphometric relationships for the bronchial trees of different species. Each panel is derived from measurements of a single silicone rubber cast. (A) Ratios of airway diameters as a function of branching generation; (B) ratios of branching angles as (more...)

An important difference between regular and irregular dichotomous branching modes concerns the number of airways between the trachea and the terminal bronchioles. In a regular dichotomous branching system, the number of divisions and, therefore, the path length, between the trachea and the most distal conducting airways is the same along any pathway. In addition, all airways at any branch level have exactly the same dimensions. In an irregular dichotomous system, the number of branch divisions from the trachea to each distal bronchiole is not the same along every pathway, and not all airways at a given branch level have the same dimensions. Table 1 presents the average number of branching generations from the trachea through the terminal bronchioles for various species. Humans have the narrowest range of branching generations, a reflection of the greater symmetry of their lungs.

Table 1.

Airway Path Lengths.

Because of the complexity of airway branching structure, the geometry of the tracheobronchial tree has been represented by models; these are idealizations derived from experimental data, usually from measurements performed on castings prepared from actual lungs. One of the most widely used human structural models is the symmetrically dichotomous Weibel Model A (Weibel 1963). This is a 23-generation system, with generations 0–16 representing conducting airways. Although the assumption of regular dichotomy simplifies the treatment of morphometric data, the actual bronchial tree is asymmetrical, and a number of models of human airways that account for asymmetry have been described (Horsfield and Cumming 1968; Olson et al. 1970; Horsfield et al. 1971; Parker et al. 1971). In addition, Phalen and coworkers (1978) and Yeh and Schum (1980) developed structural models of the human lung which consist of “typical pathways,” based on mean dimensions, for each lobe within the lung. Although the models were developed with symmetrical branching within each lobe, they do account for the asymmetry, and resultant variable path length, between different lobes. Most of the tracheobronchial models have been based upon measurements made in only one lung. The very limited data base suggests that there is significant variability in airway dimensions between individuals (Nikiforov and Schlesinger 1985), but the only model that accounts for this is a statistical description of the tracheobronchial tree based upon the Weibel geometry (Soong et al. 1979).

Structural models of the bronchial tree have also been developed for experimental animals. These include symmetrical dichotomous models for the rabbit (Kliment 1974), the rat (Kliment 1973), and the guinea pig (Kliment et al. 1972) and typical pathway models for the dog, the rat, and the hamster (Yeh 1980).

Pulmonary Region

The pulmonary region extends from the respiratory bronchioles through the alveoli and contains airways involved in gas exchange between the air and blood (figure 6a). In the human lung, the final generation of airways that merely conduct air—the terminal bronchioles—branch into several generations of respiratory bronchioles, which are characterized by the presence of alveoli. The degree of alveolarization increases toward the lung periphery; when the airway becomes totally alveolarized, it is termed an alveolar duct. This may branch into other ducts, or into blind-ended alveolar sacs. The adult human lung contains~375 million alveoli, the number varying with body size, and the average alveolar diameter is 250–300 µm. This results in a total alveolar surface area on the order of 150–180 m² (Weibel 1980).

Figure 6.

(A) Diagram of the human airways in the pulmonary region; (B) diagram of the cellular makeup and surrounding structures of the alveolus.

There are large interspecies differences in the gross structure of the pulmonary region (Gehr et al. 1981; Tyler 1983). The number of branching generations of respiratory bronchioles and alveolar ducts varies, and some species appear to have no respiratory bronchioles. The degree of alveolarization of the respiratory bronchioles also differs, as does alveolar size and total alveolar surface area, the latter increasing in direct proportion to body mass.

The alveolar surface is lined with a continuous layer of two distinct cell types (figure 6b). About 93–95 percent is covered by type I cells, which are characterized by a central nucleus surrounded by cytoplasm stretching out in thin winglike processes to form part of the alveolar wall. The remaining surface is covered by cuboidal-shaped type II cells, which are actually more numerous than the type I cells. The relative numbers of these cell types, as well as the percentage of the alveolar surface covered by each, are similar in humans and most other mammals (Crapo et al. 1983; Gehr 1984).

The alveoli are supported by a framework of connective tissue termed the interstitium. Capillary endothelial cells are joined through the interstitium to alveolar epithelial cells, to form the “alveolo-capillary membrane.” This membrane is about 2 µm thick in humans, but appears to be thinner in most experimental animals (Meessen 1960; Crapo et al. 1983; Gehr 1984). The interstitium and associated structures form the part of the lung known as the parenchyma. This region also includes the pulmonary lymphatic vessels.

The lungs contain two lymphatic networks. One set (superficial or pleural network) is located within the connective tissue layer of the visceral pleura, whereas the other (deep or peribronchovascular network) consists of interconnecting vessels within the connective tissue surrounding both the airways (to the level of the respiratory bronchioles) and the pulmonary vascular system. A plexus of vessels connects the two sets. In both systems, the network begins as blind-ended capillaries and fluid flows toward the hilar region of the lung. Many larger lymphatic vessels are interspersed with nodes (encapsulated aggregates of lymphoid tissue); the most prominent of these are located along the trachea and main bronchi, and at branching sites between these airways. More diffuse lymphoid aggregates occur near the branching regions of smaller bronchi and bronchioles. Eventually, the entire pulmonary lymphatic system drains into the general venous circulation.

Research Recommendations

Quantitative anatomy—or morphometry—of the respiratory tract is essential for understanding the dosimetry of inhaled particles. The structure of the various components of the respiratory tract influences the airflow and, thus, the resultant pattern of particle deposition and the distribution of sites of potential damage. Morphometry must be assessed in humans as well as experimental animals, the latter so as to assist in the extrapolation of toxicologic data to humans. Data are available for normal adult humans and some other species, but critical gaps remain.

■ Recommendation 1. Variability in morphometry of the tracheobronchial and pulmonary regions in normal humans as well as experimental animals (including different strains) should be studied. Better statistical descriptions of interindividual variation at all levels of the respiratory tract are needed to validate conclusions drawn from current theoretical or empirical deposition models, which are generally based upon a single morphometric model.

■ Recommendation 2. Lung morphometry should be assessed in potentially “susceptible” subsegments of the human population: children, the elderly, and people with respiratory disease. Although data are becoming available on the morphometry of children's lungs at different ages, these are not yet sufficient to develop a comprehensive morphometric model describing growth of the tracheobronchial tree. No information exists at all for assessment of morphometric changes due to aging or disease.

■ Recommendation 3. Comparative morphometry of human and animal upper respiratory tracts should be assessed. Because of large interspecies differences in the nasopharyngeal region, more quantitative information is needed to allow better comparison with that in humans. For example, rodents have essentially a straight pathway from the nostrils to the trachea, a situation radically different from that in humans and nonhuman primates. In humans, more detailed information on dimensions of the oral passages under different ventilatory conditions is also needed to assess particle removal by the upper respiratory tract.

■ Recommendation 4. Comparative structure and physiology of human and animal pulmonary lymphatic systems should be studied. This knowledge is needed for better comparisons of particle clearance by this route in humans and experimental animals.

Ventilatory Parameters

Ventilation is the movement of air in and out of the respiratory tract and is a factor in determining the amount of an exposure atmosphere that is actually inhaled. Ventilatory parameters also affect the deposition of particles once inhaled.

The amount of air inspired (or expired) during a normal breath is the tidal volume (V _T ); it averages 450–600 ml in resting healthy males and slightly less for females. The fraction of the V _T that does not reach the alveolated airways—about 150–200 ml in resting males and 120–160 ml in females—is termed the anatomic dead space volume

Not all of the inspired air reaching the pulmonary region is equally effective in oxygenating the blood, since air may enter alveoli that are ventilated but poorly perfused. The portion of V _T that does not equilibrate with gas pressure in the pulmonary capillary blood is the alveolar dead space volume The total volume of inhaled air that does not participate in gas exchange, is termed the total or physiological dead space

During expiration, air within the tracheobronchial tree—largely from the previous inspiration—is expelled along with some alveolar air which is a mixture from a number of inspirations. Particles inhaled into the pulmonary region can therefore be exhaled over a number of breaths. Thus, the time available to deposit inhaled particles in the conducting airways is fairly short (a few seconds), whereas the residence time in pulmonary air may be longer (about a minute).

Total ventilation (V̇ _E ), or minute volume (MV), is defined as the volume of air expired each minute and is equal to V _T times the breathing frequency (f). The average f during normal quiet breathing in adults is 11–17 breaths/min, and the resting V̇ _E averages 5–10 liters/min. The V̇ _E consists of anatomic dead space ventilation and total alveolar ventilation (V̇ _A ), the latter being the amount of air entering the pulmonary region each minute. The effective portion of V̇ _A that participates in gas exchange is equal to

Ventilation is affected by numerous exogenous factors such as altitude, ambient temperature, and smoking, as well as endogenous factors such as body size. Two of the major modifiers in any particular individual are physical activity and age.

Physical Activity. Healthy humans at rest normally breath through the nose, but when respiratory demand increases above a certain level there is a shift to oronasal (combined nose and mouth) breathing. Maximum inspiratory nasal airflow occurs at a V̇ _E of 30–40 liters/min (Swift and Proctor 1977; Niinimaa et al. 1980), at which point ~40–60 percent of total airflow occurs through the nose. As respiratory demand increases further, the proportion of air entering the mouth increases, but even at high demand the oral pathway accounts for no more than 60 percent of the inhaled air (Swift and Proctor 1987).

With mounting respiratory demand, V _T and f increase, and the maximum volume of air that can be inhaled per minute, or the maximum voluntary ventilation, may rise to more than 10 times the resting ventilatory level. As breathing frequency increases, expiratory time diminishes, but inspiratory time remains relatively constant. Furthermore, respiratory pauses, the gaps between expiration and inspiration which can occupy 25 percent of the breathing cycle in resting individuals, become shorter with increasing level of activity.

Growth and Aging. The volume of air in the lungs and the ventilatory capacity depend on body and lung size and, thus, increase with growth from childhood. In addition, the contribution of V _T and f to total ventilation also changes; V _T increases while f decreases until maturity is reached (Mauderly 1979).

Ventilatory function reaches a peak between the ages of 20 and 35 and then begins to decline. Although various models have been proposed to describe these changes, they differ in their assumptions about the rate of functional decline (Buist 1982). Furthermore, most of the reported data for age-related changes in lung function are derived from cross-sectional population studies and may not reflect the true aging process, especially since these studies may be measuring the heartiest survivors. The best way to avoid possible bias is to examine true aging patterns in longitudinal studies in which the same people are tested over a number of years. Such analyses are scarce, and those that do exist have measured only a few parameters (Fowler 1985).

Changes in lung function with aging are the result of deterioration of the lung tissue itself, a decrease in the strength of the respiratory muscles, and an increase in the stiffness of the thoracic cage. The time course varies from individual to individual and may be aggravated by chronic pollutant exposure. Some ventilatory indices are affected by age, whereas others are not. Figure 7 shows a diagram of the various divisions into which the volume of air in the lungs may be separated. With age, functional residual capacity (FRC) and residual volume (RV) increase, whereas vital capacity (VC), inspiratory capacity (IC), and expiratory reserve volume (ERV) decrease. Anatomic dead space increases with age because of a decrease in lung elasticity and a resultant increase in lung volume at the same pressure differentials.

Figure 7.

Diagram of subdivisions of lung volumes as measured with a spirometer. A typical spirometer tracing is shown on the right. TLC=total lung capacity, VC=vital capacity, RV=residual volume, FRC=functional residual capacity, IRV=inspiratory reserve volume, (more...)

Aging is associated with regional inequalities in the distribution of ventilation and a decrease in the uniformity of perfusion (Holland et al. 1968). Nonuniform mixing of inspired air may result when sections of the lungs communicate poorly with others and, because of this, some alveolar regions may not be continuously ventilated during normal tidal breathing. Nonuniform perfusion results in an increase in which, together with the increase in results in an aging-related rise in Although this does not affect resting levels of V̇ _E , which show no major change with aging, the ability of the lungs to respond to increased activity is altered, and maximum voluntary ventilation declines by about 30 percent between ages 30 and 70.

Comparative Aspects of Ventilation

Since much of the toxicologic work with inhaled particles involves experimental animals, it is essential that their respiratory mechanics be quantitated. Various animal data exist (see, for example, Guyton 1947; Spell 1969), but the methods used to obtain them were not standardized, so there is much variability, even for similarly sized animals of the same species. “Representative” ventilatory values for a particular species are therefore difficult to specify, so generalized values based on scaling procedures are used. Scaling is based on the principle that respiratory mechanical properties may be related to body size or mass in some consistent fashion, even though there may be interspecies differences in the mechanisms that determine these properties. This allows quantitative comparisons of function between animals of different sizes, within or between species. Scaling makes use of dimensional or dimensionless parameters that either remain constant with body size or can be related to body size by some proportionality factor (Leith 1983). For example, V̇ _E is proportional to body mass (M) raised to the ¾ power, whereas lung volumes, such as V _T , tend to vary with M to the first power. Similarly, breathing frequency is proportional to M ^−¼, whereas the ratio of to V _T is independent of body size.

Stahl (1967), after an extensive literature search, developed predictive equations relating respiratory variables in mammals to body weight. These equations can be used to scale values between animals of different species as well as between individuals of different body weights within one species, as long as the animals are in comparable physiological states. Scaling is not a precise technique, however, and is only as good as the values upon which the exponents and proportionality factors are based. For example, many of these values have been obtained in anesthetized animals, in which actual lung volumes and ventilation may be less than normal (Sweeney et al. 1983).

Airflow Patterns

Patterns of airflow in the conducting airways are a major determinant of particle deposition sites. Basic principles of airflow are presented by Ultman (this volume). Aspects of airflow critical to particle deposition are addressed below.

Within straight tubes, two main types of flow may occur: laminar and turbulent. In laminar flow, gas molecules move in parallel as a smooth stream, with the highest velocity occurring at the center of this stream. The flow can be imagined as concentric layers of air sliding or telescoping lengthwise along each other, with no transverse mixing between layers. In turbulent flow, gas molecules are in an agitated state, and there is erratic mixing of concentric layers. Random secondary flows (eddies) are superimposed on the average longitudinal motion of flow velocity. Flow that is partially laminar and partially turbulent is termed transitional.

The type of flow that occurs depends upon the strength of the inertial forces in the moving air in relation to the frictional and viscous forces acting on it. For example, turbulence occurs when the former exceed the latter. Airflow may thus be described in terms of the ratio of inertial forces to viscous and frictional forces, which is expressed as the dimensionless Reynolds number (Re). The Reynolds number depends on the geometry of the conduit through which the air passes and the velocity of airflow, and flow characteristics change as Re passes certain critical values. Thus, for steady flow in a straight, smooth-walled, circular tube, flow will be laminar when Re is less than 2100, transitional when Re is between 2100 and 4000, and fully turbulent when it exceeds 4000 (Hinds 1982).

Within the respiratory tract, bends, bifurcations, constrictions, surface roughness and convolutions, and other features of airway shape that add inertial forces may lead to turbulent flow at a velocity lower than that at which turbulence would be initiated in a smooth, straight, obstacle-free tube having the same cross-section. Thus, flow instability and turbulence may occur in the upper respiratory tract and upper tracheobronchial tree at Reynolds numbers well below 2100 (West and Hugh-Jones 1959; Dekker 1961; Sekihara et al. 1968; Olson et al. 1973; Swift and Proctor 1977). Turbulence is also produced by the continuous acceleration and deceleration of air during the breathing cycle (Lakin and Fox 1974). But although turbulent flow generated in the upper airways upon inspiration may be propagated into a few generations of downstream bronchi, air velocity decreases with depth into the lung, and in the smaller conducting airways, flow is always laminar.

Because of structural differences between the tracheobronchial trees of humans and most other mammals, one would expect differences in resultant flow patterns. For example, the trachea of most mammals is much longer relative to its diameter than is the human trachea. Thus, any turbulence introduced by flow through the larynx is much less likely to persist into the downstream bronchi of nonhuman mammals. Unfortunately, there are few data on airflow patterns in the airways of most commonly used experimental animals (see, for example, Snyder and Jaeger 1983).

Research Recommendations

Ventilatory patterns and airflow dynamics are critical determinants of dose to the respiratory tract from inhaled particles. The following important gaps in our knowledge of ventilation in humans and in experimental animals should be filled.

■ Recommendation 5. Patterns and distribution of airflow in the tracheobronchial tree of healthy adult experimental animals and humans should be determined. This information is important for the development of deposition models and for the extrapolation of results of toxicologic studies to humans.

■ Recommendation 6. Effects of aging on ventilation in humans and experimental animals should be determined by use of longitudinal studies of humans and experimental animals involving numerous ventilatory parameters. In animals, a cross-correlation of age equivalencies between species should be performed, so that parameters of toxicologic studies may be better related to lung function in humans.

■ Recommendation 7. Ventilatory mechanics and airflow in children should be analyzed. Although data are available for some stages of growth, there is a gap between birth and~9 years of age.

■ Recommendation 8. Flow patterns in the upper respiratory tracts of experimental animals and humans should be studied. Most experimental animals are obligate nasal breathers, so only their nasal passages need be studied. But in humans, analyses of the nature of flow in the oral passages through the oropharynx, including the effects of speech and increased physical activity, are also needed.

Deposition Mechanisms and Controlling Factors

Specific Deposition Mechanisms. The size of inhaled particles is a critical factor affecting their deposition; thus, resultant biological effects are, to some extent, particle-size dependent. Size may, however, be expressed in various ways. For spherical particles, actual measured diameter is unambiguous, but for nonspherical or irregularly shaped particles some “effective” diameter is more appropriate. Such particles are often described in terms of equivalent spheres, on the basis of equal volume, mass, or aerodynamic drag.

In order to compare deposition data obtained using particles of different materials, a diameter that accounts for the factors affecting deposition should be used; the most common of these is aerodynamic equivalent diameter (D _ae). This term accounts for shape and density and is defined as the diameter of a spherical particle with unit density that has the same terminal settling velocity (see below) as the particle in question. Particles that have higher than unit density will have actual diameters smaller than their D _ae.

The significant mechanisms by which particles are deposited in the respiratory tract are impaction, sedimentation, Brownian diffusion, interception, and electrostatic precipitation (figure 8). The relative contribution of each depends on characteristics of the particles as well as on ventilatory patterns and respiratory tract anatomy.

Figure 8.

Mechanisms for particle deposition in the respiratory tract. (Adapted with permission from Lippmann and Schlesinger 1984.)

Impaction onto an airway surface occurs when a particle's momentum prevents it from changing course in an area where there is a rapid change in the direction of bulk airflow. It is the main deposition mechanism for particles having D _ae≳0.5 µm in the upper respiratory tract and at or near tracheobronchial tree branching points. The probability of impaction increases with increasing air velocity, rate of breathing, particle density, and size.

Sedimentation is deposition due to gravity. When the gravitational force on a particle is balanced by the total forces due to air buoyancy and air resistance, the inspired particle will fall out of the airstream at a constant rate, known as its terminal settling velocity. The probability of sedimentation increases with increasing residence time in the airway, particle size, and density, but decreases with increasing breathing rate. Sedimentation is important for particles with D _ae≳0.5 µm in medium to small airways where air velocity is relatively low.

Submicrometer-size airborne particles, especially those with diameters ≲0.2 µm, have imparted to them a random motion due to bombardment by surrounding air molecules; this motion may then cause such particles to come into contact with the airway wall. The displacement sustained by the particle is a function of the diffusion coefficient, which is inversely related to particle cross-sectional area. Brownian diffusion is a major deposition mechanism in airways where bulk flow is low or no longer occurring, that is, in the bronchioli and alveoli. However, molecular-size particles may be deposited by diffusion in the upper respiratory tract, trachea, and larger bronchi.

As mentioned, particles with D _ae≳0.5 µm are subject to impaction and sedimentation, whereas the deposition of particles ≲0.2 µm is diffusion dominated. Particles with diameters between these values are only minimally influenced by these mechanisms and tend to have prolonged suspension times in air. They may, thus, undergo little deposition, being carried out of the respiratory tract in the exhaled air.

Interception is a significant deposition mechanism for elongated particles, such as fibers, and occurs when the edge of the particle contacts the airway wall. The aerodynamic diameter of a fiber is related to its transverse diameter. Thus, fibers that are long (for example, 50–100 µm) but thin (for example, 0.5 µm) behave aerodynamically like small particles, penetrating into distal airways. Fiber shape is also important, since straight fibers penetrate more distally than do curly ones.

Some freshly generated particles can be electrically charged and may exhibit deposition greater than that expected on the basis of size alone. Electrostatic deposition results from image charges induced on the surface of the airways by charged particles and/or from space/charge effects, whereby repulsion of similarly charged particles causes increased migration toward the airway wall. The effect of charge on deposition increases with decreasing particle size and airflow rate. Since most ambient particles become neutralized naturally by air ions, electrostatic deposition is a minor contributor to particle collection by the respiratory tract.

Factors Controlling Deposition. An understanding of the extent and loci of particle deposition in the respiratory tract requires an appreciation of various controlling factors: characteristics of the inhaled particles, anatomy of the respiratory tract, and ventilation pattern.

Characteristics of Inhaled Particles. The major particle characteristic that influences deposition is size. However, particles are inhaled not singly, but as constituents of aerosols, which are suspensions of liquid or solid particles in a gas. The components of the particulate phase may differ, but even if this consists of a single material, a spectrum of particle sizes is often present. In general, the size distribution of particles in commonly encountered aerosols fits reasonably well with a lognormal distribution; that is, the logarithm of particle diameter is normally distributed. Such a distribution can be described by a geometric mean size (which is also the median diameter) and by an index of dispersion—the geometric standard deviation (σg). This latter is the ratio of the diameter at 84.1 percent (or 15.9 percent) cumulative probability, that is ±1 standard deviation (SD) of the normal curve, to the diameter corresponding to 50 percent cumulative probability (figure 9). Depending upon the specific size parameter used to develop the distribution, the resultant median diameter may be count median (CMD, using the physical diameter of the particles), mass median (MMD, using the particle mass distribution relative to diameter), or aerodynamic mass median (MMAD, using aerodynamic equivalent diameter). If not directly measured, the MMD and MMAD may be calculated from the measured CMD for spherical particles. Radioactive or toxic aerosol size distributions are often expressed as activity median aerodynamic diameter (AMAD).

Figure 9.

Cumulative frequency distribution plots of particle number for two polydisperse aerosols. Both aerosols have the same count median diameter (CMD) (50 percent probability), but they have different geometric standard deviations (σ _g ). Because (more...)

The deposition probability of particles with physical diameters ≳0.5 µm is governed largely by particle aerodynamic diameter, whereas deposition probability of smaller ones is governed by actual physical diameter. Thus, use of the MMAD parameter is appropriate only in describing aerosols in which most particles are physically ≳0.5 µm; the median size of aerosols containing particles with actual diameters less than this is usually expressed in terms of a diffusion diameter or actual physical size. Aerodynamic diameter is, therefore, the most appropriate unit for describing deposition by sedimentation and impaction, but not by diffusion.

The size distribution of an aerosol, which largely depends on its method of production, is characterized as monodisperse or polydisperse (heterodisperse). A monodisperse aerosol consists of particles of uniform size. Since, in reality, perfect monodispersity does not exist, an aerosol is considered monodisperse if the σ _g is<1.2 (Fuchs and Sutugin 1966). But use of this term in deposition analyses means that all of the particles are assumed to behave as if they were exactly the same size, that is, the median size. In polydisperse aerosols, particles of widely differing sizes may be present, and the σ _g is≥1.2.

If the σ _g of a polydisperse aerosol is<2, its total respiratory tract deposition will probably not differ substantially from that for a monodisperse aerosol having the same median size (Morrow 1981; Diu and Yu 1983). However, size distribution is important in determining the spatial pattern of initial dose. This is because the effect of size dispersion on regional deposition depends upon the sequential “filtering” action of each component of the respiratory tract, which in turn depends upon particle size. For example, as σ _g increases for aerosols with median sizes between 0.01 and about 0.07 µm, tracheobronchial deposition will likely increase, but pulmonary deposition will decrease because of less penetration into this region. On the other hand, as σ _g increases for aerosols with median sizes of 0.07 to about 1 µm, bronchial as well as pulmonary deposition will increase (Diu and Yu 1983).

A particle characteristic that may dynamically alter its size is hygroscopicity. Hygroscopic particles may grow substantially while they are still in transit in the respiratory tract and will be deposited according to their hydrated, rather than their initial dry size. The deposition pattern of specific hygroscopic aerosols can generally be related to their particle growth characteristics, if known.

Respiratory Tract Anatomy. Respiratory tract geometry affects particle deposition in various ways. For example, airway diameter sets the displacement required for a particle to contact a surface, whereas cross-section determines the air velocity for a given flow rate. Differences in pathway lengths in different lung lobes affect regional deposition. Lobes with the shortest average path length between the trachea and terminal bronchioles may have the highest concentration of deposited particles ≥1 µm in the alveoli. Regional differences become less obvious for submicrometer particles, which tend to deposit evenly in all lobes regardless of path length but in proportion to relative ventilation (Raabe et al. 1977).

Although humans differ from most other mammals in various aspects of respiratory tract anatomy, the implications of this for particle deposition have not been adequately appreciated. For example, alveolar size differs among species; since particles with D _ae ≳0.5 µm that reach the alveoli will be deposited primarily by sedimentation, and different-size alveoli have different characteristics as sedimentation chambers, the net result will be that the pulmonary region of various species will have different deposition efficiencies. Differences in deposition patterns affect the dosimetry of inhaled particles and the ability to use the results of toxicity tests in experimental animals for human risk assessments. In addition to interspecies differences, the data available indicate that size of tracheobronchial airways and alveoli vary considerably within species. Such variation is probably a major factor responsible for the observed differences in deposition efficiency among individuals of one species (Heyder et al. 1982).

Ventilatory Parameters and Mode of Inhalation. The pattern of respiration during particle exposure influences regional deposition sites and efficiencies. For example, high inhalation velocities enhance deposition by impaction but decrease that due to sedimentation and diffusion. Thus, a rise in flow rate, such as during increased physical activity, may shift regional deposition, increasing collection in the upper respiratory tract and central bronchi and reducing it in more distal conducting airways and the pulmonary region (Valberg et al. 1982; Morgan et al. 1984; Bennett et al. 1985). Increased flow velocities may also result in the development of turbulence, which tends to enhance particle deposition, the degree of potentiation depending on particle size (Schlesinger et al. 1982).

Tidal volume affects deposition by determining how deep into the lungs the inspired air penetrates. At a constant breathing frequency, increasing tidal volume deepens penetration of inhaled particles, thus increasing deposition in the smaller conducting airways and pulmonary region. Alterations in tidal volume can also dramatically affect total respiratory tract deposition. For example, Schum and Yeh (1980) suggested that, in the rat, doubling tidal volume from 1.4 ml to 2.8 ml increases the deposition of a 1-µm (median D _ae) aerosol by seven times. Finally, the duration of respiratory pauses influences sedimentation or diffusion deposition by affecting particle residence time in relatively still air.

A major ventilatory change that occurs in humans when activity level increases is a switch from nasal to oronasal breathing. Since the nasal passages remove inhaled particles more efficiently than the oral passages, bypassing the nose increases the penetration of particles into the lungs. The actual magnitude of this increase is influenced by particle size, since larger particles are more effectively filtered in the nose than are smaller ones.

Measurement of Deposition

Measurement Techniques. Various techniques have been used to measure particle deposition in the respiratory tract of humans and experimental animals (Valberg 1985). Unfortunately, the use of different experimental methods and assumptions, especially in assessment of regional deposition, has resulted in large variations in reported values, even within the same species.

Total respiratory tract deposition has often been determined by a procedure that compares the concentration of test particles administered in inhaled air with that in collected exhaled air, the difference representing the total amount deposited. If assumptions are made about mixing and dead space, estimates of regional deposition can be derived from measurements of particle concentration in different volume fractions of the expired air, but such assumptions cannot be validated.

Specific particle characteristics may be used to measure deposition. Most commonly, radioactively tagged tracer particles are used with various types of detector systems. Total deposition is estimated by monitoring the thoracic and head regions immediately after exposure, whereas regional deposition is usually defined functionally on the basis of subsequent clearance. For example, it is often assumed that any particles remaining in the thorax 20 to 24 hr after exposure are in the pulmonary region, and particles that deposited in the tracheobronchial region were cleared from the lungs prior to this time. This is a reasonable assumption for healthy subjects but may not be for subjects with disease states where clearance is slower, and its use could result in an overestimation of pulmonary deposition and an underestimation of tracheobronchial deposition.

Deposition in the upper respiratory tract is inferred from measurements on the head immediately after exposure. Since this region clears rapidly to the stomach, even the first measurement may not accurately reflect actual deposition; accordingly, some investigators include an initial measurement of material in the gastrointestinal tract in their reported value for upper respiratory deposition. However, the upper respiratory tract, as defined in various studies, may include any or all of the following anatomic regions: nasopharynx, oropharynx, larynx, or upper trachea.

Another technique for deposition analysis in experimental animals is chemical and/ or radiological assay of tissues or whole organs removed by dissection after exposure. Obtaining accurate deposition values requires immediate sacrifice, and the assumption of no particle translocation (except to the gastrointestinal tract) prior to or during dissection.

Experimental Deposition Assessment. The species of choice for deposition analyses is the human. However, experimentation with human subjects is not always possible, and various animals are therefore used instead, with the ultimate goal of extrapolating the results to humans. If the results are to be valid, the extrapolation must take into account interspecies differences in total and regional deposition.

It is difficult to systematically compare deposition patterns obtained from reported studies in one species, and it is even harder to do this between species, because of variations in experimental protocols, measurement techniques, definitions of specific respiratory tract regions, and so on. For example, tests with humans are generally conducted under protocols that standardize tidal volume and breathing frequency (although the standardization parameters often differ in different laboratories), whereas those using experimental animals involve a wider variation in respiratory exposure conditions (for example, spontaneous breathing versus controlled breathing as well as various degrees of sedation). Much of the variability in the reported data for individual species is due to the lack of normalization for specific respiratory parameters during exposure.

In addition, experimental inhalation studies use different exposure techniques, such as nasal mask, oral mask, oral tube, or tracheal intubation. Regional deposition fractions are affected by the exposure route and delivery technique used (Wolfsdorf et al. 1969; Swift et al. 1977a). Even the specific size of the delivery device can affect inspired airflow rates, which influence the extent of deposition in the upper respiratory tract and the degree of particle penetration into the lungs (Heyder et al. 1980b).

Compilations of experimentally determined deposition values in humans and those experimental animals commonly used in inhalation toxicology studies are shown in figure 10. Not all deposition studies reported in the literature were included in this survey, since the objective was to make the intercomparisons as valid as possible. Thus, only studies where regional deposition values as a fraction of the amount of particles inhaled were provided, or could be derived, were included. Most studies describe regional fractions as a percentage of total deposition rather than in terms of amount of material inhaled and were, therefore, excluded. In addition, only studies using nonhygroscopic, nonviable, nonfibrous aerosols and reporting an aerodynamic or diffusion-related diameter were included. Most studies with humans used monodisperse aerosols, whereas many of those with experimental animals used polydisperse aerosols. Since some of these latter may have consisted of particles of widely different sizes, it is often difficult to evaluate deposition based upon the median size alone. However, it is necessary to include some of these studies, since a substantial amount of the existing data base is derived using such aerosols. Finally, although the tracer aerosols in some studies were not charge neutralized, data using these tracer aerosols were included. The presence of electrical charges could account for some of the variability between different studies using the same species and similar size particles.

Figure 10.

Deposition efficiency (that is, percentage deposition of amount inhaled) as a function of particle size for (A) total respiratory tract, (B) upper respiratory tract, (C) tracheobronchial tree, and (D) pulmonary region. All values are means (with standard (more...)

Total Respiratory Tract. Figure 10a shows total respiratory tract deposition. In humans, nasal inhalation results in somewhat greater total deposition than oral exposures for particles with diameters >0.5 µm because the nasal passages collect larger particles more efficiently than the oral passages. There is little difference in total deposition between nasal or oral breathing for particles from 0.02 to 0.5 µm. With even smaller particles, total deposition should be greater with nose breathing than with mouth breathing, although the difference would be small, amounting to, for example, only about 5 percent for particles with diameters of 0.005 µm (Schiller et al. 1987).

Dogs and guinea pigs exhibit greater total deposition of 0.1–1–µm particles than do nasal-breathing humans. However, for particles >1 µm, deposition is less in dogs than in humans, but deposition in guinea pigs is similar to that in humans. On the other hand, both rats and hamsters generally show less total deposition than nasal-breathing humans.

In some cases, the data indicate that total deposition for the same size particle can be quite similar in experimental animals and humans. It therefore follows that deposition efficiency is independent of body (or lung) size (McMahon et al. 1977; Brain and Mensah 1983). However, different species exposed to the same size particles at the same exposure concentration will not receive the same initial mass deposition. If the total amount of deposition is divided by body (or lung) weight, smaller animals would receive greater initial particle burdens per unit weight per unit exposure time than would larger ones. For example, for 1–µm (D_ae) particles, it is predicted that the rat would receive an initial deposit 5 to 10 times that of humans, and the dog would receive 3 times that of humans, if deposition was calculated on a per unit lung (or body) weight basis (Phalen et al. 1977).

Not all atmospheric particles to which an individual is exposed will be inhaled. The inspirable fraction is the portion of the ambient concentration that actually enters the upper respiratory tract. In humans, for example, the fraction for particles with D _ae <10 µm is greater than 80 percent, whereas that for particles ranging from 30 to 80 µm is about 50 percent (Vincent and Armbruster 1981). The probability of particles being inhaled into the respiratory tract depends upon particle size as well as the orientation of the individual to external air currents and the size of the entrance to the respiratory tract. Thus, inspirable fractions likely differ among species.

An additional point concerns hygroscopicity. If figure 10a is examined, it is evident that total deposition of hygroscopic particles<0.5 µm inhaled by humans would tend to decrease if particles grow no larger than 0.5 µm, and deposition will only begin to increase if particle final diameter is >1 µm. Furthermore, since particles>5 µm may grow minimally in one respiratory cycle, their deposition may not increase at all compared to nonhygroscopic particles (Ferron et al. 1987). On the other hand, deposition probability for 0.3 to 0.5–µm hygroscopic particles may change substantially.

Upper Respiratory Tract. Figure 10b shows upper respiratory tract deposition. There is substantial variability between species as well as large differences between individuals of the same species. Most experimental animals are obligate nasal breathers, and a large part of the intraspecies variability may be due to nasal geometry variation (Brain and Valberg 1979) as well as to different breathing patterns during exposure. Note the large intraspecies variability in deposition for particle sizes subject to impaction; this is probably responsible for a large portion of the intraspecies variation in total respiratory tract deposition (Stahlhofen et al. 1981 a; Heyder et al. 1982).

In humans, nasal inhalation results in enhanced deposition compared to oral inhalation. In all species shown, there is a rapid increase in deposition with increasing particle size about 1 µm, although the apparent “rate” of deposition increase with size is not the same. Thus, in humans, deposition appears to plateau somewhat for sizes>2 µm, whereas in rodents, deposition increases more rapidly.

Upper respiratory deposition of particles >1 µm is sometimes greater in nasal-breathing humans than in the experimental animals. This is not necessarily expected, since the nasal passages of animals are more intricate than are those in humans and should therefore be more efficient particle collectors. However, the actual observations may be a reflection of exposure conditions. Many of the experimental animals were sedated or anesthetized and therefore breathed slower than fully awake animals. Since the dominant mechanism for deposition of particles>1 µm in the upper respiratory tract is impaction, low flow rates should reduce deposition efficiency. Inasmuch as smaller particles can penetrate the upper respiratory tract at all flow rates, deposition for these is similar in all species. If deposition were plotted in a manner that would normalize for flow, which is not possible for most of the experimental animal studies because of the lack of such data, the experimental animals would probably show greater deposition efficiency for larger particles than would humans at equivalent size/flow normalization parameters.

Hounam et al. 1969; Johnson and Zeimer 1971; Kanapilly et al. 1982; Landahl et al. 1951, 1952; Lippmann 1970, 1977; Lippmann and Albert 1969; Lippmann and Altshuler 1976; Martens and Jacobi 1974; McMahon et al. 1977; Moores et al. 1980; Muir and Davies 1967; Palm et al. 1956; Pattle 1961; Raabe et al. 1977, 1987; Schiller et al. 1987; Stahlhofen et al. 1981a,b; Swift et al. 1977b; Tu and Knutson 1984; Wilson et al. 1985; Wolff et al. 1981, 1982; Yeh et al. 1980.

The extent of particle removal by the upper respiratory tract may vary depending upon whether an aerosol is mono- or polydisperse. For example, Thomas and Raabe (1978) compared the deposition in hamsters of a monodisperse and a poly disperse aerosol having similar median diameters (AMAD, 1.53 µm vs. 1.87 µm). The major difference was that the polydisperse aerosol deposited to a greater extent in the upper respiratory tract because of the presence of a certain percentage of larger particles that were effectively removed by impaction. Total respiratory tract deposition of the two aerosols, expressed as a percentage of inhaled amount, was the same. The less the deposition in the head, the greater is the amount available for removal in the lungs. Thus, the extent of removal in the upper respiratory tract may affect deposition patterns in distal regions.

Tracheobronchial Tree. Figure 10c shows tracheobronchial deposition; the amount of data available is less than for other regions. The figure indicates that the percentage of inhaled aerosol that is removed is greater in the oral-breathing human than in the nasal-breathing dog, hamster, or rat, at least in the limited region where particle sizes overlap. As mentioned above, a lower tracheobronchial deposition in experimental animals may be a reflection of greater upper respiratory tract deposition. On the other hand, the differences in regional deposition may be due to differences in flow in the upper bronchial tree and/or in airway branching patterns. In all cases, especially in the experimental animals, there is no well-defined trend relating deposition to particle size, unlike the situation in the other respiratory tract regions; on the contrary, fractional tracheobronchial deposition is relatively constant over a wide particle size range.

Pulmonary Region. Deposition in the pulmonary region is shown in figure 10d. In general, deposition in humans breathing orally increases as particle size decreases, after a minimum deposition is reached at about 0.5 µm. In nasal-breathing humans and experimental animals, deposition tends to decrease with increasing particle size.

The removal of particles in more proximal airways determines the shape of the pulmonary deposition curves. Increased upper respiratory and tracheobronchial deposition of particles≥1 µm results in a reduction of pulmonary deposition that occurs more sharply in smaller animals than in humans. This is due not to reduced efficiency of pulmonary deposition of larger particles, but to the fact that only a small fraction of these large particles reach the lower respiratory tract. Similarly, nasal breathing in humans results in less pulmonary penetration of larger particles; thus, there is a lesser fraction of deposition for entering aerosol than for oral inhalation. In oral-breathing humans, the peak for pulmonary deposition shifts upward to a larger-size particle compared to nasal breathing humans, and is more pronounced. On the other hand, with nasal breathing, there is a relatively constant pulmonary deposition over a wider size range, that is, 0.7–3

Pulmonary deposition is much less in hamsters and rats, which are similar to each other, than in dogs, guinea pigs, or humans. However, deposition in nasal-breathing humans is less than in these other species when available data for comparable size ranges are compared. Patterns are similar for oral inhalation, although the particle size for peak deposition is greater in humans than in guinea pigs or dogs. This is probably due to the more efficient removal; of larger particles in the upper respiratory tract and tracheobronchial airways of these experimental animals.

Factors Modifying Deposition

Various factors cause deposition patterns to differ from those in normal healthy adults, the greatest contributors to the human data base. These include exposure to airborne irritants, lung disease, growth, and aging, all of which can affect particle deposition by changing ventilation patterns and/or airway geometry.

Bronchoconstriction induced by inhaling irritants increases impaction deposition in the upper airways. Thus, for example, cigarette smokers with no clinical disease exhibit somewhat greater tracheobronchial deposition of tracer particles ≳1 µm than do nonsmokers. As a consequence, smokers also exhibit a reduction in pulmonary region deposition compared to nonsmokers (Lippmann et al. 1972).

There are some deposition data in disease states. In humans with chronic bronchitis, an obstructive airway disease, tracheobronchial and upper respiratory tract deposition of particles≳1 µm is quite variable, but greater than in healthy individuals (Thomson and Short 1969; Lippmann et al. 1972). Airway obstruction associated with lung disease in humans reduces the peripheral deposition of particles and may even entirely eliminate deposition in some parts of the lungs (Lourenco et al. 1972; Thomson and Pavia 1974). Total deposition has been found to be lower in rodents with enzyme-induced emphysema than in normal controls (Hahn and Hobbs 1979; Damon et al. 1983). This is probably due to an increase in alveolar size, resulting in greater distances to deposit on a surface and a concomitant reduction in pulmonary region deposition efficiency (Brain and Valberg 1979).

There are some data on the effects of fibrotic disease on deposition. Heppleston (1963) found that inhaled hematite particles deposited more distally in rats with coal- or silica-derived pneumoconiosis than in normal animals. On the other hand, Love and coworkers (1971) found no difference in the total respiratory tract deposition of 1-µm particles in coal workers with simple pneumoconiosis compared to normal people.

Most particle deposition studies in humans are performed with young to middle-aged adults, and few data are available on the growing lung. The available information is based on estimates of the influence of anatomic and ventilatory changes upon deposition during postnatal growth, obtained by using assumed respiratory parameters and mathematical particle deposition modeling techniques in conjuction with actual child lung morphometric values or scaled versions of available adult morphometric models (see, for example, Crawford 1982; Hofmann 1982; Martonen 1985; Phalen and Oldham 1985; Phalen et al. 1985; Phalen 1987). These studies suggest that the relative effectiveness of the major deposition mechanisms may differ at various times in the growth of the individual and that this, in turn, may alter the pattern of regional particle deposition. Thus, all age groups may not have the same distribution of deposition after exposure to the same particles. Although the results of the modeling studies are not consistent in terms of which regions or specific age groups differ, they all suggest that deposition efficiency, in at least some regions, is greater in children than adults. Taking into account the greater ventilation per unit body weight in children, the deposition fractions in certain regions could be well above those measured in adults. Since there are also regional differences in clearance rates, it follows that the dose to specific lung compartments will vary with age from newborn to adult.

There are no systematic data that would allow an analysis of deposition in the aging lung, that is, between adulthood and senescence. There are also few data on deposition differences according to gender. Available evidence indicates that under equivalent inspiratory conditions, total respiratory tract and tracheobronchial deposition of particles with D _ae from 2.5 to 5 µm, inhaled orally at rest, is similar in men and women (Pritchard et al. 1987). However, women's smaller-diameter airways result in higher flow rates and, hence, more impaction in the upper respiratory tract, resulting in less deposition in the pulmonary region. This suggests that as particle size increases, women (and perhaps children) may be at less risk from material in pulmonary airways but at a greater risk from deposition in the upper respiratory tract and tracheobronchial tree.

Localized Patterns of Deposition

Specific patterns of enhanced local deposition within various regions of the respiratory tract are important to consider, since tissue dose depends on the surface density of deposited particles. The occurrence of nonuniform deposition suggests that the initial dose delivered to specific sites may be greater than that occurring if a uniform density of surface deposit is assumed. This is especially important for inhaled particles, such as irritants, that affect tissues on contact.

In the human upper respiratory tract, enhanced deposition may occur in the larynx, oropharyngeal bend, and soft palate (Swift 1981). Deposition is also nonuniform in the nasal passages; varying relative amounts occur in the anterior and posterior regions, depending largely on particle size (Itoh et al. 1985; Swift and Proctor 1987). The change in airflow direction at the vestibule in the nasal passages, together with the fact that it is an area of high velocity, produces locally enhanced deposition posterior to this region.

Studies in models and hollow casts of the human upper tracheobronchial tree have shown that deposition of aerosols >1 µm is not homogeneous. Entrance conditions produced by the larynx result in enhanced deposition in the upper trachea. At bronchial bifurcations, deposition is greatly enhanced relative to the rest of the airway length (Schlesinger et al. 1982; Cohen et al. 1987). This occurs by impaction during inspiration, although deposition is also enhanced downstream of bifurcations during exhalation (Schlesinger et al. 1983a). Enhanced deposition of submicrometric particles at bifurcations also occurs (Cohen et al. 1987); since this size aerosol is not subject to impaction, the effect is probably from turbulent diffusion.

The experimental conditions used in the numerous microdistribution studies varied widely, yet the relative distribution of enhancement among the airways was quite similar. This suggests that within the larger bronchi, local patterns of deposition may be fairly insensitive to particle size and airflow rate. Measurements in hollow human airway casts have also shown that the proportional distribution of deposition in specific airways is relatively constant over a wide range of particle sizes and overall deposition efficiencies (Schlesinger and Lippmann 1978). In addition, inhalation studies with rodents indicate that the distribution of deposition in the various lobes of the lungs is also relatively constant over a range of particle sizes and different total lung deposition efficiencies (Raabe et al. 1977).

There are few data on local deposition patterns for distal airways. Available information is based on examination by microscopy of tissues after in vivo exposures of experimental animals (Holma 1969; Brody and Roe 1983). These studies indicate that bronchiolar and alveolar duct bifurcations are preferential sites for deposition of a wide range of particles small enough to reach these regions.

Differences in the geometry of airways in humans and other species may result in differences in the microdistribution patterns of particle deposition, a factor that should be accounted for in extrapolation modeling. For example, nasal turbinates of rodents are more complex than those of primates and, as a result, the bulk of impaction-dominated deposition occurs more anteriorly in the nasal passages of rodents (Gooya and Patra 1986; Schreider 1986). Unfortunately, there are no other data to relate geometry to microdeposition. Speculated differences in the site and extent of localized deposition in the tracheobronchial tree are depicted in figure 11.

Figure 11.

Location and relative intensity of enhanced tracheobronchial particle deposition for (A) inspiratory flow; and (B) expiratory flow, in humans (a) and nonprimate laboratory animals (b). (Adapted with permission from Lippmann and Schlesinger 1984.)

Mathematical Modeling

Mathematical models are needed to predict deposition sites and efficiencies since it is not possible to study all conditions of exposure experimentally. A mathematical model relates the main factors that control deposition to various geometric parameters and is used to predict the mean probability of particle deposition in the respiratory tract. Although most models have been designed for assessing deposition of spherical particles in humans, some have been developed for experimental animals (Kliment et al. 1972; Kliment 1973, 1974; Schreider and Hutchens 1979; Schum and Yeh 1980).

The first mathematical treatment of regional particle deposition in humans was performed by Findeisen (1935). This was later refined by others (Landahl et al. 1951, 1952; Beeckmans 1965; Task Group on Lung Dynamics 1966). Because of their very nature, these analytical models adopted assumptions and idealizations of almost all aspects of the respiratory tract and of particle dynamics. This simplification resulted in the loss of some important characteristics of the real system and often limited their ultimate usefulness and reliability. When compared to results from human experiments, these early models tended to overestimate pulmonary and underestimate tracheobronchial deposition (Mercer 1975). They were, however, very useful in quantitating the influence of various controlling parameters on deposition.

One of the major components of any deposition model, and one subject to the greatest oversimplification, is the representation of airway geometry. Most of the early predictive models made use of a very simple stylized lung structure. Recently, however, more realistic anatomic descriptions have become available, some developed specifically for use in deposition analyses and others easily adapted for such applications. As discussed earlier, some of these are symmetrical, others asymmetrical. Since the actual human bronchial tree is asymmetrical, and because the amount of deposition depends upon the path length over which the inhaled aerosol passes, realistic computations require consideration of asymmetry.

In order to assess the effect of anatomic structure, Yu and Diu (1982) calculated deposition in humans, using various symmetrical and asymmetrical geometries, and compared results to experimental data. They found that predicted total deposition did not differ greatly from geometry to geometry, and all compared reasonably well with experimental values. However, predicted regional deposition was quite sensitive to the particular geometry used. Since the anatomic models are generally derived from examination of single casts, they reflect, to some extent, actual interindividual structural variability. These results may provide the reason for variations from individual to individual in deposition experiments, even under identical breathing conditions. The sensitivity of predicted deposition to the specific anatomic model has also been noted by others (Martonen and Gibby 1982).

Another drawback of analytical models is the oversimplification of airflow pattern, a necessity since there are no exact expressions for flow dynamics in noncylindrical, tapering tubes that undergo repeated branchings and often have asymmetrical, nonlaminar flow profiles. Analytical models generally assume laminar flow, no disturbances produced by bifurcations, and uniform ventilation of airways. In addition, many assume a constant velocity of air during inspiration and expiration.

Any predictive model must also contain expressions for deposition probability. However, equations for deposition within a realistic geometry and flow pattern are not available. Thus, semiempirical expressions based upon analyses of simplified analogues of sections of the bronchial tree are often used. For example, impaction expressions are often obtained from the analytical solution to the equation of motion of a particle in ideal flow in a bent tube; diffusion expressions are obtained from the analytical solution for flow in an infinitely long, horizontal tube with ideal flow; and sedimentation expressions are obtained from the solution for deposition along a long, horizontal tube.

Recent mathematical deposition models have increased in sophistication and flexibility. Some allow for variations in air velocity, mode and pattern of breathing, polydispersity, hygroscopic growth, and even for changes in linear airway dimensions over the breathing cycle (see, for example, Taulbee and Yu 1975; Ferron 1977; Diu and Yu 1983; Egan and Nixon 1985). Some also include a more realistic treatment of the upper respiratory tract (Scott et al. 1978; Yu et al. 1981; Yu and Diu 1982). When predictions from these recent models are compared to experimental data, there is more often agreement for total than for regional deposition. However, given the complexity of the respiratory tract and intersubject variability, this is not surprising.

A relatively recent approach to modeling in humans is dimensional analysis, in which deposition is related to some dimensionless parameter. Heyder and coworkers (1980a) formulated a parameter that is a function solely of particle size and flow rate and upon which total deposition was dependent; single parameters for regional deposition have been reported by Rudolf et al. (1987).

Because of the importance of drawing conclusions about humans from experiments with animals, special attempts have been made to develop methods for direct interspecies extrapolation. Stauffer (1975) used dimensional analysis to develop scaling factors based on particle physics and the assumption of a geometric similarity among all mammalian lungs. He predicted that interspecies particle deposition probabilities would be similar for sedimentation but a function of body weight for diffusion. McMahon and co workers (1977) attempted to scale the collection efficiency of the respiratory tract in different species, based on inhalation studies in mice, hamsters, rats, rabbits, and dogs. They concluded that the overall collection efficiency of the lung would be independent of body size (this is essentially what is observed in figure 10a).

Research Recommendations

There is a considerable body of data on the deposition of inhaled aerosols in humans and experimental animals. We also have a fairly good understanding of some of the factors that control deposition. But our knowledge in certain critical areas is still not adequate.

■ Recommendation 9. Particle removal in the human upper respiratory tract should be assessed experimentally for oral and nasal breathing as should the influence of breathing mode on deposition in other regions of the respiratory tract. Information on removal of particles in the upper tract will allow the prediction of appropriate starting concentrations for modeling particle transport into the lungs, since particle removal in the upper respiratory tract determines the concentration penetrating to distal regions. It is still not well defined how the specific route of entry affects regional deposition in humans, especially those with respiratory disease. No systematic studies have been done in which oral geometry, flow, and deposition of particles have been measured during natural oral breathing, that is, without an inhalation tube or mouthpiece placed in the anterior oral passages. These types of studies will enable better interpretation of the large available data base on deposition in humans, which was obtained with mouthpiece breathing. In addition, the effect of various degrees of oronasal breathing upon deposition should be assessed.

■ Recommendation 10. Microdistribution patterns of deposition should be studied under a wide range of exposure conditions. The nonuniformity of deposition in both the tracheobronchial tree and the pulmonary region may be important factors in ultimate dose. Further assessment of microdistribution is needed for incorporation of “enhancement functions” into deposition models.

■ Recommendation 11. Regional deposition efficiencies should be determined for ultrafine (<0.1 µm) particles in the human respiratory tract. Much of the lack of such data is due to the difficulty in generating monodisperse aerosols in this size range, as well as in accurately detecting the generated particles. More studies are needed that evaluate deposition in the upper respiratory tract where such small particles, if soluble, may be rapidly absorbed into the blood, as well as in the tracheobronchial tree, where increased deposition compared to larger submicrometer sizes may occur. The available theoretical models appear to be inconsistent in that predicted deposition for particles in this size range depends very much on the particular model being used.

■ Recommendation 12. The effects of anatomic variability on deposition should be analyzed systematically, and appropriate statistical descriptions developed for incorporation into deposition models. Theoretical predictions and in vivo studies show a dependence of regional deposition upon morphometry, and there is interindividual variability in structural characteristics of the human lungs. Yet the significance of this variability in affecting deposition is not currently known.

■ Recommendation 13. Effects of specific aspects of ventilation upon deposition should be determined. More data are needed with exercise breathing patterns, which may result in greater risk because of increased ventilation. More information is also needed on the relation of changes in tidal volume and breathing rate to the uniformity of deposition in the lungs. Studies should also be performed in experimental animals relating ventilation to deposition for use in extrapolation models.

■ Recommendation 14. Deposition in sensitive subsegments of the human population, such as children, the aged, and people with chronic lung disease (for example, emphysema and bronchitis), should be examined. Since children cannot be used in experimental studies, the development of deposition models based upon accurate ventilatory and morphometric information is critical. Although it is difficult to study deposition in individuals with lung disease, because of ventilatory and anatomic dysfunction which result in a large variability in deposition, more studies performed using well-controlled in vivo testing procedures and/or hollow airway cast systems would provide a better basis for assessing deposition in the compromised lung.

■ Recommendation 15. The deposition of hygroscopic particles in the human respiratory tract should be evaluated. Many important pollutant aerosols are hygroscopic, and there may be substantially greater deposition during inhalation as well as exhalation compared to dry particles of the same initial size, making mathematical predictions of deposition based on nonhygroscopic particles difficult. Most calculations of the growth of hygroscopic particles are based upon growth curves developed for sodium chloride or sulfate particles, and few data for other dynamic material exist. Studies of such deposition in the head, especially during nasal breathing, are needed.

■ Recommendation 16. Intercomparisons of regional deposition patterns among experimental animals (unsedated) and humans should be made using comparable monodisperse particles over a wide size range and comparable experimental techniques. Most of the regional deposition data that allow any cross-species comparison are for particles>1 µm. In addition, there are no consistently applied methods for assessment of regional deposition in experimental animals and humans. “Calibration factors” need to be developed that may be used to relate results of toxicologic studies in experimental animals to human exposure assessment and health effects.

■ Recommendation 17. Models that allow calculation of deposition by airway generation should be expanded to other species. Coupled with data on ventilation and morphometry, this will allow better estimations of delivered dose in experimental animals.

Clearance Mechanisms: Basic Structure and Function

Particles are cleared from the respiratory tract by several different processes, some of which are regionally distinct as shown in table 2.

Table 2.

Respiratory Tract Clearance Mechanisms.

Upper Respiratory Tract. The nasal passages beyond the vestibular region are lined with a ciliated epithelium overlaid by mucus (see Overton and Miller, this volume, figure 1). The mucus is produced by specialized epithelial cells and submucosal glands and consists of two layers: a low-viscosity hypophase that surrounds the cilia and within which they move, and a high-viscosity epiphase lying on top of the cilia (Lucas and Douglas 1934). The composition and characteristics of mucus are described in detail by Overton and Miller (this volume). Material depositing on the mucus is cleared by movement of the epiphase due to coordinated beating of the cilia.

The general flow of mucus in the ciliated nasal passages is toward the nasopharynx. In the region just distal to the nonciliated vestibule, however, mucous flow is forward, clearing deposited material to an area where sneezing, wiping, or blowing can occur. Soluble material deposited on the ciliated nasal epithelium will be accessible to underlying cells if diffusion through the mucus occurs at a rate faster than mucous flow. Since there is a rich vasculature in the nose, uptake into the blood can occur rapidly.

Insoluble particles deposited in the oral passages are cleared by swallowing into the gastrointestinal tract. Although there are no data on the clearance of soluble particles, oral tissue has the capacity for rapid systemic absorption (Swift and Proctor 1987). In the larynx, mucus moving toward the head from the trachea passes into the hypopharynx and is swallowed.

Tracheobronchial Tree. Most of the surface of the tracheobronchial tree through the terminal bronchioles is lined with ciliated epithelium overlaid by mucus, and insoluble particles are cleared primarily by the net movement of fluid toward the oropharynx. Some insoluble particles may traverse the tracheobronchial epithelium, entering the peribronchial region (Masse et al. 1974; Sorokin and Brain 1975), while soluble particles can be absorbed through the mucus into the circulation.

The bronchial surfaces are not homogeneous. For example, there are openings of daughter bronchi and islands of nonciliated squamous cells at bifurcations. Within these regions, the usual progress of mucous movement is interrupted (Hilding 1957), and clearance can be retarded. The efficiency with which nonciliated obstacles are passed depends on the traction of the mucous layer.

Pulmonary Region. A number of mechanisms and pathways contribute to clearance from the pulmonary region, but their relative importance is uncertain and depends to some extent on the physicochemical properties and amount of material deposited. The mechanisms involve either absorptive (dissolution) or nonabsorptive processes, which can occur simultaneously or at different times.

Nonabsorptive clearance processes, which are outlined in figure 12, are mediated primarily by alveolar macrophages. These large, mononuclear cells (figure 13) originate from precursors in bone marrow, reach the lung as circulating monocytes, and mature in the pulmonary interstitium, from which they traverse the epithelium to reach the alveolar surface. As macrophages move freely on alveolar surfaces, they phagocytize, transport, and detoxify deposited material, which they contact by chance or by directed motion due to chemotactic factors.

Figure 12.

Flowchart of clearance pathways for particles depositing in the pulmonary region (dissolution is not included).

Figure 13.

(Top) Electron micrograph (12,000x) of an alveolar macrophage. (Bottom) Light micrograph of an alveolar macrophage that has phagocytosed polystyrene latex particles. The dark area is the nucleus.

Macrophages normally comprise about 3 percent of total alveolar cells in healthy, nonsmoking humans and in other mammalian species (Gehr 1984). However, the actual count can be influenced by particle deposition. Few particles may not result in an increase in cell number. Above a certain loading, however, macrophage numbers increase in proportion to particle number until a saturation point is reached (Brain 1971; Adamson and Bowden 1981). This increase is due to monocytic egress, proliferation of interstitial mononuclear cells, and/or to actual division of alveolar macrophages (Bowden and Adamson 1980; Blusse van Oud Alblas et al. 1983), and appears to be a generalized response that follows exposure to many types of particles, although the extent differs for particles of different composition (Adamson and Bowden 1981). Furthermore, the magnitude of the increase is more closely related to the number of deposited particles than to the total dose by weight, so equal masses of an identical material will stimulate more macrophages if the material is delivered as many small particles rather than as fewer large ones.

Particle-laden macrophages are cleared from the pulmonary region along a number of pathways. The primary route is by the mucociliary system, but the mechanism(s) by which cells reach it is not certain. One possibility is movement along the alveolar epithelium due to surface tension gradients between the alveoli and conducting airways; alternatively, locomotion could be directed along a density gradient, such as that produced by chemotactic factors released by macrophages actively ingesting deposited material (Kilburn 1968; Sorokin and Brain 1975; Ferin 1976). Another possible route to the mucociliary system involves passage through the alveolar epithelial wall into the interstitium (Brundelet 1965; Green 1973; Corry et al. 1984; Harmsen et al. 1985). Macrophages could then reach the surface of ciliated airways, perhaps through small collections of lymphatic tissue that exist at the alveolar/bronchiolar junction (Macklin 1955). Some of the cells that follow interstitial clearance pathways are probably resident interstitial macrophages that have ingested particles transported through the alveolar epithelium by endocytosis by type I pneumocytes (Brody et al. 1981; Bowden and Adamson 1984).

Particle-laden macrophages that do not clear by way of the bronchial tree may actively migrate within the interstitium to a nearby lymphatic channel or, along with uningested particles, be carried in the flow of interstitial fluid toward the lymphatic system. Passive entry into lymphatic vessels is fairly easy since endothelial cells are loosely connected with wide intercellular junctions (Lauweryns and Baert 1974); lymphatic endothelium has also been observed to actively engulf particles from the surrounding interstitium (Leak 1980). Deposited particles can then be translocated to the tracheobronchial lymph nodes, which often become reservoirs of retained material. Some particles subsequently appear in postnodal lymph, from which they enter the blood and may then translocate to extrapulmonary sites. Alternatively, uningested particles or macrophages in the interstitium may cross the alveolar capillary endothelium, entering the blood directly (Robertson 1980; Holt 1981; Raabe 1982). Whatever the route to the systemic circulation, particle-laden macrophages as well as free particles have been found in various extrapulmonary organs (Hourihane 1965; Pooley 1974; Lee et al. 1981; LeFevre et al. 1982).

Free particles in macrophages within the interstitium can end up in perivenous or subpleural sites, where they then become trapped. The migration and grouping of particles and macrophages can lead to the redistribution of deposits into focal aggregates in the lungs (Heppleston 1953).

The specific clearance route for particles depositing in the pulmonary region may depend upon loading. Earlier reports have suggested that at low-exposure concentrations, most particles are removed within macrophages via the bronchial mucociliary system, whereas at higher exposure levels, a larger proportion of free particles are translocated by the lymphatic system (Ferin 1977). More recently, however, researchers have found that the percentage of initial lung burden cleared by the lymphatic system after exposure to a high particle level is the same as that after exposure to a lesser burden (Snipes and Clem 1981; Snipes et al. 1983; Lehnert et al. 1986). Thus, some free particles are likely cleared by the lymphatics under most conditions of exposure.

The most important mechanism for absorptive clearance is dissolution. Free particles that dissolve in the alveolar fluid can diffuse through the epithelium and interstitium into the lymph or blood, whereas particles initially translocated to and trapped in interstitial sites may undergo dissolution there. Dissolution is a major clearance route even for particles usually considered to be relatively insoluble (Morrow et al. 1964; Mercer 1967). The factors affecting the solubility of deposited particles are poorly understood, although they are influenced by the particle's surface-to-volume ratio and other surface properties (Morrow 1973).

Some deposited material can dissolve after phagocytic uptake by macrophages. For example, certain metals dissolve within the acidic milieu of phagosomes. It is, however, not certain whether the dissolved material then leaves the cell. This is discussed further in a subsequent section.

Clearance Kinetics

Kinetic data are essential for determining the dosimetry of inhaled particles. Although the lungs may clear deposited material completely, the time frame over which clearance occurs determines the dose delivered to the lungs as well as to other organs. Tissue doses to the upper respiratory tract and tracheobronchial tree are often limited by the rapid clearance from these regions and are thus proportional to concentration and exposure duration. On the other hand, doses from material deposited in the pulmonary region depend much more on the characteristics of both the particle matrix and any associated materials.

Both the pulmonary region and the tracheobronchial clearance rates (that is, the fraction cleared per unit time) are well-defined functional characteristics of an individual human or experimental animal when repeated tests are performed under the same conditions (Gibb and Morrow 1962; Schlesinger et al. 1978; Lippmann et al. 1980; Bohning et al. 1982); but there is substantial interindividual variability. In addition, because of differences in experimental techniques and the fact that measured rates are strongly influenced by the specific methodology, comparisons between studies by different investigators are difficult to make.

Measurement Techniques. Methods for measuring clearance have been reviewed recently by Schlesinger (1985). Some of the techniques are identical to those used in assessing deposition, since the first measurement after aerosol inhalation is assumed to represent initial deposition.

The velocity of mucous transport in the nasal passages, trachea, and main bronchi can be measured directly by monitoring inert marker particles placed on the epithelium, or by measuring the movement of a bolus of radioactive particles selectively deposited within these airways or, in the case of the trachea and main bronchi, moving through them from more distal airways. The advantage of local velocity techniques is that they allow measurement in anatomically defined, albeit limited, airways. Because of this specificity, there is no doubt about whether clearance rates altered by toxicant exposure resulted from actual alterations in the mucociliary system or from a change in deposition pattern, a doubt not easily resolved when using whole-lung clearance assays, as discussed below. However, local velocity techniques have a number of disadvantages. Some marker techniques are invasive, since particles may have to be selectively introduced into the airway of interest. Anesthesia, necessary in many cases, may affect the observed transport rates. Finally, the actual method of marker introduction can result in trauma to the airway.

In a number of studies aimed at assessing the effects of inhaled toxicants upon mucociliary clearance, alteration in tracheal transport rate has been used as the sole end point because it is easier to measure than is whole-lung clearance. However, an overall effect cannot necessarily be inferred from a change in this index, since toxicant-induced changes in bronchial clearance are not always associated with an alteration in tracheal transport rate. This could occur if the toxic particles are of a size too small to provide significant deposition within the trachea.

The most common technique to measure whole-lung clearance uses inhalation of radiolabeled (nonleaching) tracer particles. The total amount of radioactivity remaining in the lungs at selected intervals is then measured by external detector systems. The decline in emission rate, corrected for radioactive decay, represents clearance. Various types and configurations of mobile and/or fixed-scintillation detector systems have been used, and each has its own advantages and disadvantages in terms of spatial resolution and sensitivity.

One of the problems in using external monitors to assess tracheobronchial mucociliary clearance is that the observed clearance pattern depends on the initial deposition of the tracer particles. This is because the techniques are indirect, and clearance rate is proportional to transit pathways. For example, an apparent increase in clearance rate after toxicant exposure could be due to a proximal shift in deposition of the tracer rather than to an effect on the clearance system itself. This presents a special problem when different groups are being compared; for example, persons with chronic obstructive pulmonary disease tend to have greater central bronchial deposition than do healthy subjects for the same size tracer particles. There is, however, no basis for any kinetic distinction between mucous transport rates measured using different particles, as long as the deposition sites are the same. The rate of mucous transport in the trachea has been found to be independent of the shape, size, or composition of the insoluble marker used to measure it (van Antweiler 1958; Man et al. 1980). There is no a priori reason to assume that this does not hold true for transport in more distal airways as well, provided there is no toxic effect from the deposited material.

Pulmonary region clearance is relatively slow, and therefore measurements should be performed over, perhaps, several months. When radioactively tagged tracer particles are used, a nuclide having a long half-life is required. In addition, since the total dose to the subject should be minimized, especially if humans are used, long counting times may be required to obtain statistically reliable data. Health risks may therefore rule out such long-term radioactive tracer clearance studies.

A technique that avoids this problem is magnetopneumography (Cohen 1973). In this procedure, the subject is exposed to magnetizable particles and, at various times, a magnetic field is applied externally to the thorax in order to magnetize the bulk of the deposited particles. After the external field is removed, a remanent field remains, which is measured with an appropriate sensor (Freedman et al. 1982). Magnetopneumographic techniques have some advantages over radioaerosol techniques in terms of temporal and spatial resolution. Furthermore, certain information can be obtained using them that is not obtainable by other whole-animal in vivo techniques, for example, the assessment of in situ phagocytosis of tracer particles by macrophages (Brain et al. 1987). However, magnetopneumography has some significant drawbacks: all sources of external magnetic contamination on the subject or on the measurement apparatus must be removed; critical positioning is required since the measurements are highly sensitive to distance from the source; and there are difficulties in deducing actual particle distribution in the lungs from the data.

Fecal analysis is a technique for indirect monitoring of clearance using experimental animals that involves radioaerosols but not external monitoring. The radioactivity in feces collected at fixed intervals after exposure to tracer particles is measured. The fecal excretion activity curve presumably represents material cleared by the mucociliary system into the gastrointestinal tract and can thus be used to provide an index of tracheobronchial clearance. This technique assumes that all material cleared from the lungs is transported to the gastrointestinal tract and subsequently excreted in the feces. It is also very sensitive to feeding behavior of the animals; those that do not eat or do not excrete for a particular fraction of the sampling interval cannot be included in the analysis.

Clearance of deposited particles can also be assessed in experimental animals by serial sacrifice at various intervals after exposure, followed by microscopic, chemical, or radiological analysis. Various parts of the respiratory tract or the lungs as a whole can be sampled. The measured burden plotted as a function of time provides an index of clearance. Although microscopy can provide only a qualitative assessment of particle distribution and clearance from various regions, other techniques alone or combined with microscopy allow quantitative determination of the material retained regionally, and without interference from material in adjacent areas. Sacrifice techniques have the advantage of being very sensitive, but major disadvantages are that a large number of animals is needed for statistical reliability, the intraindividual variability in clearance cannot be examined, and the effects of toxicants on the course of clearance in the same individual on different occasions or under different conditions cannot be assessed.

Clearance Rates and Times. Upper Respiratory Tract. Nasal mucous flow rates are nonuniform. Regional velocities in the healthy adult human range from <2 to >20 mm/min (Proctor 1980), with the fastest flow in the midportion of the nasal passages; average velocities for the nasal passages as a whole range from 4 to 12 mm/ min (Bang et al. 1967; Phipps 1981). The resultant mean transport time for insoluble particles over the nasal passages is about 10 to 13 min (Rutland and Cole 1981; Majima et al. 1983; Stanley et al. 1985). If soluble particles diffuse through the mucus within this time period, they become accessible to underlying epithelial cells.

Particles deposited in the anterior, nonciliated portion of the nasal passages can be cleared slowly by mucous movement; a flow rate of 1 to 2 mm/hr has been suggested for fluid moved by traction from more distal cilia (Hilding 1963). Particles may take over 12 hr to be cleared by this mechanism and are usually more effectively removed by sneezing or wiping, in which case clearance may occur in under 30 min (Fry and Black 1973; Morrow 1977).

The velocity of mucous transport in the larynx has not been measured. However, it is probably about the same as that in the trachea (Swift and Proctor 1987).

Tracheobronchial Tree. Clearance of particles deposited on tracheobronchial airways occurs by the parallel processes of mucous transport and absorption. The fraction of deposited material cleared by either of these pathways is a function of its physicochemical properties, but because of the short time frame for mucociliary clearance, relatively insoluble material will be cleared solely by this route.

Mucous clearance occurs at different rates in different local regions; mucus moves fastest in the trachea, and progressively slower in more distal airways. Measured rates in the human trachea range from 4 to 20 mm/min, depending upon the experimental technique used (Yeates et al. 1981a). Anesthesia and invasive procedures affect transport, resulting in rates apparently slower than normal. Using noninvasive measurement procedures on unanesthetized, healthy nonsmokers, researchers have observed average tracheal mucous transport rates of 4.3 to 5.7 mm/min (Yeates et al. 1975, 1981b; Foster et al. 1980; Leikauf et al. 1981, 1984).

The mean mucous velocity in the human main bronchi has been found experimentally to be about 2.4 mm/min (Foster et al. 1980). Although rates of mucous movement in smaller airways cannot be measured directly, transport rates in human medium bronchi have been estimated at 0.2 to 1.3 mm/min, and those in the most distal ciliated airways as low as 0.001 mm/ min (Morrow et al. 1967b; Yeates and Aspin 1978).

The total duration of bronchial clearance or some other time parameter is often used as an index of mucociliary function. In healthy, nonsmoking adult humans, 90 percent of insoluble particles depositing in the tracheobronchial tree will be cleared within 2.5 to 20 hr after deposition. The actual time depends on the individual subject and the size of the tracer aerosol used, which affects the depth of deposition and subsequent pathway length for removal (Albert et al. 1973). Clearance will be 99 percent completed by 48 hr after deposition (Bailey et al. 1985a).

In humans, normal tracheobronchial mucociliary clearance exhibits a two-phase pattern (figure 14): a short initial phase characterized by rapid clearance lasting a few hours, followed by a slower, second phase extending until 24 to 48 hr after exposure. These probably represent clearance of the tracer particles deposited in the “upper” and “lower” tracheobronchial tree, respectively. As the size of the tracer particles is reduced, resulting in more distal deposition, there is an increase in the fraction of total tracheobronchial clearance which is accounted for by the slower phase. Although some portion of the above clearance pattern may include rapid early clearance of material deposited in the pulmonary region, the contribution of this to the apparent bronchial clearance rate appears minimal.

Figure 14.

Schematic representation of tracheobronchial clearance after exposure to tracer particles. Particles remaining beyond 24 to 48 hr (shaded area) are assumed to have deposited in the pulmonary region.

Studies in rodents have shown that a small fraction of insoluble material is retained for prolonged periods within the upper respiratory tract or tracheobronchial tree (Patrick and Stirling 1977; Watson and Brain 1979; Gore and Patrick 1982). In humans, it has been estimated that the average residence time in bronchial tissue of insoluble particles derived from cigarette smoke is 3 to 5 months (Radford and Martell 1977). Soluble material may also be retained in ciliated airways for long periods because of binding to cells or specific macromolecules (Boecker et al. 1983).

The mechanism(s) underlying the long-term retention of insoluble particles is unknown. It may involve endocytosis by epithelial cells with subsequent translocation into deeper tissue or merely passive movement into the tissue (Sorokin and Brain 1975; Watson and Brain 1979; Gore and Patrick 1982). In addition, long-term tracheobronchial retention patterns for insoluble particles are not uniform. Enhanced retention occurs at bifurcation regions (Radford and Martell 1977; Henshaw and Fews 1984; Cohen et al. 1987), which may be the result of greater deposition as well as ineffective mucous clearance. Because of this nonuniformity, doses calculated using uniform surface retention density may be misleading.

Pulmonary Region. Particles are cleared from the pulmonary region by a number of pathways and mechanisms. Their effectiveness depends on the nature of the particles, but just what this dependence is has not been completely resolved. Consequently, the kinetics of clearance from the pulmonary region are not definitively understood, although particles deposited there generally remain longer than do those deposited in ciliated airways. Data on clearance rates in humans are limited, and those for experimental animals (and humans) vary widely because of different properties of the particles in the various studies. Many of these studies used high concentrations of particles, which may of itself have interfered with normal mechanisms, producing rates different from those that would occur at lower exposure levels.

Pulmonary region clearance data appear to fit an exponential model, and each component is believed to represent removal by a different mechanism or pathway (Casarett 1972). For example, an initial fast phase having a clearance half-time of about 2 to 6 weeks presumably represents rapid clearance by macrophages; an intermediate phase, with a half-time of months, may represent macrophage clearance by interstitial pathways; and a phase of prolonged clearance with a half-time of months to years represents removal by dissolution. Rates of removal by dissolution are extremely variable but likely dominate the long-term retention of relatively insoluble particles.

Rates that correspond to the various clearance phases can only be obtained if clearance is measured until all the deposited particles are removed from the lungs. This is usually not possible, and many studies are terminated when the radioactivity levels of retained particles fall below detectable limits. Clearance of inert insoluble particles in healthy, nonsmoking humans has been observed experimentally to consist of two phases: the first has a half-time measured in days, and the second in hundreds of days (Bailey et al. 1982; Bohning et al. 1982; Philipson et al. 1985). Table 3 summarizes data from numerous studies for the half-times of the longer, second phase of clearance. Wide variations in clearance times indicate a dependence upon the nature of the material being cleared. For example, when polydisperse aerosols are used, various size fractions clear by different routes and, thus, with varying rates (Snipes et al. 1984a,b). Different clearance rates have also been observed when using different-size particles (Morgan and Holmes 1980; Bailey et al. 1982); but if dissolution is accounted for, mechanical removal to the gastrointestinal tract and lymph nodes is independent of particle size (Snipes et al. 1983). There is also considerable intersubject variation in the clearance rates of similar particles, which increases with time postexposure (Bailey et al. 1985a; Philipson et al. 1985). This large difference in pulmonary region clearance kinetics among different individuals suggests that equivalent exposures to insoluble particles will result in differences in respiratory tract burdens.

Table 3.

Long-Term Particle Clearance from the Pulmonary Region in Human Nonsmokers.

Even less is known about relative rates along specific pathways than about overall pulmonary region clearance kinetics. After deposition, the uptake of particles by alveolar macrophages is very rapid, unless the particles are cytotoxic (Lehnert and Morrow 1985; Naumann and Schlesinger 1986). The actual rate of subsequent clearance of these cells is not certain; perhaps 5 percent or less of their total number is translocated from the lungs each day (Masse et al. 1974; Lehnert and Morrow 1985).

Uningested particles may penetrate into the interstitium within a few hours after deposition (Sorokin and Brain 1975; Ferin and Feldstein 1978; Brody et al. 1981). The amount transported via transepithelial passage seems to increase as particle loading increases, especially when loading surpasses the level at which the number of macrophages saturate (Ferin 1977; Adamson and Bowden 1981). Similarly, a depression of phagocytosis by toxic particles may increase the number of free particles in the alveoli, enhancing removal by other routes. Free particles or those within alveolar macrophages reach the lymph nodes within a few days after deposition (Harmsen et al. 1985; Lehnert et al. 1987). However, most clearance by the lymphatic system is very slow (Sorokin and Brain 1975; Ferin 1976).

Soluble particles deposited in the pulmonary region are cleared rapidly by absorption through the epithelial surface into the blood, but there are few data on dissolution and transfer rates in humans. The rate does depend upon the size of the particle, with smaller ones clearing faster than larger ones. Some dissolved material may be retained in lung tissue because of binding with cellular components, preventing it from passing into the circulation (Cuddihy 1984).

Comparative Clearance Kinetics and Modeling. The retention of certain materials cannot be studied experimentally in humans, so experimental animals must be used. Since dosimetry depends upon clearance rates and routes, adequate toxicologic assessment necessitates relating clearance kinetics in animals to that in humans. Although the basic mechanisms of respiratory tract clearance are similar in humans and most other mammals, regional clearance rates vary substantially among species, even for similar particles deposited under comparable exposure conditions. It is likely that dissolution rates and rates of transfer of dissolved substances into blood are related solely to the properties of the material being cleared and are essentially independent of species (Cuddihy et al. 1979; Griffith et al. 1983; Bailey et al. 1985b). On the other hand, different rates of mechanical transport, such as macrophage clearance from the pulmonary region (Bailey et al. 1985b) or mucociliary transport in conducting airways (Felicetti et al. 1981), are found, resulting in species-dependent rate constants for these clearance pathways. Differences in regional (and perhaps total) clearance rates among species are probably due to these latter processes. Accordingly, respiratory tract clearance in humans can be predicted by using dissolution rates in experimental animals and mechanical clearance rates in humans, as long as lung damage or binding to lung molecules has not occurred (Bailey et al. 1985b).

Another approach used to predict clearance in humans is mathematical modeling. Various theoretical and empirical models have been developed to predict regional as well as total respiratory tract clearance of particles. Most of these models have been used for dosimetry of inhaled radionuclides (see, for example, International Commission on Radiological Protection 1959, 1972; Bailey and James 1979). In these models, fractional allocations are made between mechanical clearance processes and dissolution on the basis of properties of each specific material being assessed.

Mathematical models have also been developed that describe overall tracheobronchial clearance patterns by calculating mucous transport rates in each generation (Yeates and Aspin 1978; Lee at al. 1979; Yu et al. 1983). These models make various assumptions: for example, all mucus is produced in the terminal airways; no fluid reabsorption occurs; or the thickness of the mucous layer is constant in all airways. In addition, the overall clearance rates generated by some of these models are very sensitive to the rates assumed in the smallest airways; this is because this region has the slowest rate and a large surface area, and the models assume transport rates to be inversely proportional to surface area or circumference. Only limited testing of the accuracy of these models is possible because actual transport rates are not known for distal airways. Thus, predicted results are often compared to actual observations for total time of tracheobronchial clearance and to values for mucous transport rates in the trachea. Internal adjustments are made so that the predicted time is the same as that observed experimentally.

Factors Modifying Clearance

A number of host and environmental factors modify normal clearance patterns, affecting the dose delivered by exposure to inhaled particles. These factors include aging, gender, work load, disease, and irritant inhalation exposure. In many cases, however, their exact role is not resolved.

The evidence for aging-related effects on mucociliary function in healthy individuals is contradictory, with studies showing either no change or a slowing in clearance with age after maturity (Goodman et al. 1978; Yeates et al. 1981a). One problem is that it is difficult to determine whether an apparent decrement in function is due to aging alone, or to long-term, low-level ambient pollutant exposure (Wanner 1977). There are no data for changes in overall pulmonary region clearance related to aging. Functional differences have been found between alveolar macrophages from mature and senescent mice (Esposito and Pennington 1983), although no age-related decline in human macrophage function has been seen (Gardner et al. 1981).

There are not sufficient data to assess changes in clearance in the growing lung. Nasal mucociliary clearance time in a group of children (average age 7 yr) has been found to be about 10 min (Passali and Ciampoli 1985), which is within the range for adults. There is one report of bronchial clearance in 12 yr olds, but this study was performed in hospitalized patients (Huhnerbein et al. 1984).

In terms of gender, no difference in nasal mucociliary clearance has been observed between male and female children (Passali and Ciampoli 1985), nor in tracheal transport rates in adults (Yeates et al. 1975). Slower bronchial clearance has been noted in male compared to female adults, but this was attributed to differences in lung size rather than inherent gender differences in transport velocities (Gerrard et al. 1986).

The effect of increased physical activity on mucociliary clearance is also unresolved, with the available data indicating either no change or an increase with exercise (Wolff et al. 1977; Pavia 1984). There are no data relating changes in pulmonary region clearance to increased activity levels, but Valberg and coworkers (1985) found that CO₂-stimulated hyperpnea had no effect on early pulmonary clearance and redistribution of particles.

Various diseases are associated with altered clearance. Nasal mucociliary clearance is prolonged in humans with chronic sinusitis, bronchiectasis, or rhinitis (Majima et al. 1983; Stanley et al. 1985), and with cystic fibrosis (Rutland and Cole 1981). Bronchial mucous transport may be impaired in people with bronchial carcinoma (Matthys et al. 1983), chronic bronchitis (Vastag et al. 1986), asthma (Pavia et al. 1985), and various acute respiratory infections (Lourenco et al. 1971b; Camner et al. 1979; Puchelle et al. 1980). In some of these conditions, coughing may enhance mucous clearance but is generally effective only if excess secretions are present.

Rates of pulmonary region particle clearance appear to be reduced in humans with chronic obstructive lung disease (Bohning et al. 1982), and the viability and functional activity of macrophages has been found to be impaired in human asthmatics (Godard et al. 1982). Reduced clearance from the pulmonary region of experimental animals with viral infections has also been observed (Cresia et al. 1973). On the other hand, Tryka and coworkers (1985) found increased pulmonary clearance in hamsters with interstitial fibrosis. Damon and coworkers (1983) observed no clearance difference in rats with emphysema. Hahn and Hobbs (1979), however, found that the copresence of inflammation resulted in prolonged retention. Inflammation may enhance the penetration of free particles and macrophages through the alveolar epithelium into the interstitium by increasing the permeability of the epithelium and the lymphatic endothelium (Corry et al. 1984).

Cigarette smoking in humans is associated with persistently slowed mucociliary clearance in both the nasal passages and the tracheobronchial tree (Lourenco et al. 1971 a; Camner and Philipson 1972; Goodman et al. 1978; Stanley et al. 1984), and the extent of decline appears related to the amount of smoking (Vastag et al. 1986). Smokers can also exhibit specific clearance abnormalities, including intermittent retrograde mucous flow in the trachea and intermittent periods of stasis that alternate with abrupt drops in particle retention in the bronchi (Albert et al. 1971, 1973). The rate of particle clearance from the pulmonary region also appears to be reduced in heavy cigarette smokers (Cohen et al. 1979; Bohning et al. 1982).

In addition to cigarette smoke, other inhaled irritants have an effect on mucociliary clearance function in humans as well as experimental animals (Wolff 1986). Single exposures to a particular material may increase or decrease the overall rate of tracheobronchial clearance, depending upon the exposure concentration. Although alterations in clearance rate following single exposures to moderate concentrations of irritants are transient—lasting <24 hr— repeated exposures may persistently retard clearance. The effects of irritant exposure may be enhanced by exercise, or by coexposure to other materials.

Acute and chronic exposures to inhaled irritants can also alter clearance from the pulmonary region. For example, nitrogen dioxide (NO₂), ozone (O₃), sulfuric acid (H₂SO₄), and some metals (for example, cadmium) have been shown to change the rate of tracer particle clearance (Ferin and Leach 1977; Oberdörster and Hochrainer 1980; Driscoll et al. 1986; Schlesinger and Gearhart 1986). Clearance may be accelerated or depressed, depending upon the specific material and/or length of exposure. Alterations in alveolar macrophage function may underly some of the observed changes, since numerous irritants have been shown to impair the functional properties of these cells (Gardner 1984).

Specific macrophage properties, which include phagocytosis and mobility, allow them to adequately perform their role in clearance. However, the relation between these characteristics and overall clearance is not certain. For example, in comparisons among a number of species, no positive correlation was found between macrophage mobility and clearance rate since slower movement was often associated with an acceleration in clearance (Metivier 1984; Naumann and Schlesinger 1986).

Research Recommendations

■ Recommendation 18. Interspecies comparisons of short-term (mucociliary) and long-term (pulmonary) clearance kinetics should be made, with an emphasis on mechanical processes, using equivalent insoluble particles and experimental techniques. These studies should assess the effects of exposure characteristics—for example, particle size and mass concentration—on retention patterns, and should examine the effects of differences in lung anatomy. The use of equivalent experimental conditions is essential to avoid differences in the results due to lack of standardization in measurement techniques and particle characteristics. Newer, nonradioactive experimental techniques should be used to expand the data base on long-term pulmonary region clearance kinetics in humans.

■ Recommendation 19. The long-term fate of particles in conducting airways should be examined. To determine whether specific cell types preferentially handle certain particles or to assess whether cells critical in clearance change during their normal activity, these studies should include morphological techniques to precisely locate sites of particle deposition and retention.

■ Recommendation 20. Studies should be undertaken on the pathways and circumstances by which macrophages and free particles are removed from alveoli to regions of potential long-term retention in the lung, such as peribronchial and peri-vascular sites, as well as lymph nodes. Effects of particle loading should be considered, since it is not clear how this affects transport by various routes.

■ Recommendation 21. Alveolar macrophages should be characterized. There is a need to study how alterations in cell functional properties are manifested in changes in overall clearance patterns, so as to determine whether or not specific functional changes induced by inhaled materials critically affect the cell's ability to participate in lung defense. The effects of various characteristics of exposure (for example, mass loading), on phagocytosis, mobility, release of chemotactic factors for neutrophils, or production of other mediators, requires further examination. In addition, since macrophages are not homogeneous, and different subsets exhibit functional differences, more work is needed to characterize the heterogeneity of macrophages recovered after particle exposures.

■ Recommendation 22. In vitro effects on macrophages should be related to those produced in vivo. Dose to cells is better defined in in vitro studies, but calibration factors are needed to relate exposure concentration to actual target tissue dose in extrapolating in vitro to in vivo results.

■ Recommendation 23. The effects of exogenous factors on retention should be determined. For example, exercise may alter deposition pattern, but how clearance and ultimate retention are affected is not known.

■ Recommendation 24. Mucus should be characterized as to average and regional variations in thickness, physicochemical properties, and synthesis rates. Data are needed on humans as well as most experimental animals.

■ Recommendation 25. Fate of soluble particles in the tracheobronchial tree should be determined. There are uptake data for the nasal passages but not for other conducting airways.

■ Recommendation 26. Animal models that mimic human respiratory disease should be developed further. These models should be used to examine clearance and retention of potentially sensitive subsegments of the human population.

■ Recommendation 27. Clearance kinetics of the growing lung and the aging lung should be studied. Studies of mucociliary clearance and pulmonary region clearance are needed; when combined with deposition studies, a comprehensive picture of defense capabilities in the young and elderly segments of the population can be developed. Animal models can be used to some extent, since it is difficult in human populations to separate true aging effects from certain environmental influences, such as air pollution.

■ Recommendation 28. Gender-related differences in the disposition of inhaled particles in humans should be investigated. A comparison of the efficiency of defense mechanisms is needed to determine whether, as has been suggested, doses in males and females may indeed differ.

Diesel Exhaust Particles

Diesel exhaust particles have a diameter (MMD) of 0.2 to 0.3 µm and a dense carbonaceous core. They usually contain adsorbed organic matter, but this section discusses the general fate of inhaled diesel particles without regard to specific adsorbed components. Available data are based largely on inhalation studies using rodents exposed to diluted diesel exhaust containing particles, usually at predetermined concentrations, as well as various gases.

About 15 to 20 percent of diesel particles inhaled by rodents are initially deposited (Chan et al. 1981; Dziedzic 1981; Lee et al. 1983); this value is quite close to that for total respiratory tract deposition of similar-sized particles in humans (figure 10a). Although there are no actual experimental studies in humans, diesel particle deposition in different age groups under various breathing conditions has been estimated using mathematical modeling (Xu and Yu 1985). Results suggest that total and regional deposition vary with age. For nasal breathing at rest, total and pulmonary deposition in infants and children are predicted to be greater than that in adults, with maximum deposition occurring at about two years of age. Because of its particle size, diesel particle deposition is predicted to be unaffected by the mode of inhalation or by the frequency of respiration but should increase with increasing tidal volume.

The clearance routes for diesel particles depend on their regional deposition. The fraction deposited in the tracheobronchial tree, about 40 to 50 percent of the initial deposit, is rapidly cleared by mucociliary transport in about one to two days (Chan et al. 1981; Lee et al. 1983), although there is some evidence that diesel particles depress mucociliary clearance rates (Battigelli et al. 1966). Most of the particles that reach the pulmonary region are phagocytized by macrophages (Barnhart et al. 1981; White and Garg 1981), and the numbers of these cells seem to increase in relation to the rate of particle entry into the lung rather than to the total cumulative exposure (Strom 1984; Mauderly et al. 1987). Although diesel particle exposures to levels up to about 2 mg/m³ do not reduce viability of macrophages, phagocytic activity has been variously reported to be either depressed or unaltered (Chen et al. 1980a,b; Weller et al. 1980; Barnhart et al. 1981; Castranova et al. 1985).

Diesel particles can also be taken up by type I alveolar epithelial cells (Barnhart et al. 1981). Enhanced uptake probably occurs if the macrophages are overloaded, since the number of type I cells containing particles increases as particle concentration and exposure duration increase.

Following deposition, diesel particles are fairly evenly distributed throughout the pulmonary region. Gradually, within macrophages, the particles are moved from peripheral lung regions toward the terminus of the mucociliary transport system, from where they may be cleared via the tracheobronchial tree (White and Garg 1981). Clusters of particle-laden macrophages are often found at the distal ends of the terminal bronchioles after high- or low-level exposures (Barnhart et al. 1979; Puro 1980; Garg 1985). Aggregates of free particles have been observed in focal areas of the tracheobronchial tree, perhaps because of a general depression of mucociliary clearance, local accumulation in areas of inadequate transport, or longer-term retention after uptake by or through the epithelium.

Another clearance pathway from the pulmonary region is by the lymphatic system. Free diesel particles, as well as particle-laden macrophages, have been found in parenchymal lymphoid aggregates, lymphatic vessels, and mediastinal lymph nodes (Vostal et al. 1979). The amount cleared along this pathway increases with increasing duration and level of exposure (Chan et al. 1981; White and Garg 1981).

The kinetics of diesel particle clearance have been examined in rodents. However, exposure concentrations were generally high, and it is not known whether the kinetics are the same when exposure levels are lower. In rats, two phases of pulmonary clearance were observed, with half-times of 6 and 80 days, respectively (Lee et al. 1983). The faster clearance was ascribed to the mucociliary transport of particles deposited in proximal respiratory bronchioles, whereas the slower phase was ascribed to other alveolar removal processes. The kinetics were the same with exposure to either 7 mg diesel particulate/m³ for 45 min, or 2 mg/m³ for 140 min. Guinea pigs exposed to 7 mg/m³ for 45 min showed little clearance from days 10 to 432 after exposure, even though initial deposition percentages and mucociliary clearance times were the same as in the rat. On the other hand, the clearance was found to occur at the same rate in rats and mice chronically exposed to 0.35 to 7.0 mg diesel particulate/m³ (Henderson et al. 1982).

It has been suggested that the observed increase in macrophage numbers after exposure to diesel particles should increase the pulmonary clearance rate of this material relative to clearance in nonexposed controls (Lee 1981). Experiments do not always support this hypothesis; diesel particle exposure has been associated with depressed and accelerated clearance from the pulmonary region (table 4). Lung burden appears to be a critical factor affecting the overall efficiency of pulmonary clearance, perhaps by altering relative amounts cleared by different pathways. Diesel particles may be retained in the lungs for long periods of time after exposure, with the amount increasing with increasing deposition (Chan et al. 1981; Rudd and Strom 1981; Henderson et al. 1982). A long residence time provides an extended period for elution of adsorbed material.

Table 4.

Effects of Diesel Particles on Respiratory Tract Clearance.

The development of dust clusters and their residence times probably depend on exposure concentration and duration (Moore et al. 1978). Perhaps accumulation and persistence begin, or increase, when normal clearance processes are overloaded during chronic exposures; this could occur at exposure levels lower even than those used in most studies, but there are few data at realistic concentrations. Most of the data on diesel particle disposition were obtained from chronic studies in rodents using diluted exhaust with diesel particulate levels ≳0.25 mg/m³. Furthermore, the observed rates and routes of clearance could have been affected by the various combustion products, many of which are irritants, found either in the gas phase of diesel exhaust or adsorbed on particle surfaces.

Metals

Many metals present in motor vehicle fuel are emitted in the exhaust (Lee and van Lehmden 1973). They are distributed in the atmosphere as individual particles or adsorbed onto the surfaces of other particles. Particle sizes range from submicrometer to about 2 to 3 µm, depending on the metal.

Once deposited in the respiratory tract, the disposition of a metal varies with its valence state and the compound containing it. The solubility of metals and their compounds in biological fluids strongly influences their biological availability, utilization, and toxicity. Insoluble forms of some metals can accumulate in the lungs over time because of continuous exposure and slow systemic absorption, whereas more soluble forms are rapidly absorbed into the blood and translocated to other organs. But some soluble forms can actually undergo greater retention than insoluble ones because of binding to protein in the lungs. Unfortunately, most ambient measurements determine the total concentration of the metal and do not discriminate among different compounds or valence states. Thus, for dosimetric purposes, it is usually assumed that all compounds, whatever their source, will dissociate to some degree after deposition in the respiratory tract, releasing metal ions that will be absorbed and redistributed in the body in a similar manner.

The absorption efficiency for most metals is about 50 to 80 percent from the pulmonary region, and 5 to 15 percent from the upper respiratory tract and tracheobronchial tree (Natusch et al. 1974). This may reflect the more efficient extraction of metals from the smaller particles that would preferentially deposit in the pulmonary region, and/or the shorter residence time of particles depositing on the tracheobronchial tree. After absorption from the respiratory tract, these metals will be distributed rapidly to blood-rich organs and more slowly to other organs and to fat, and will very slowly equilibrate with poorly perfused tissues (Luckey and Venugopal 1977).

Metals that deposit in the pulmonary region have the greatest toxic potential because of the likelihood of extended residence times. The more insoluble the metal, the more likely it is to be cleared from this area by movement to the tracheobronchial tree, followed by swallowing. Systemic absorption from the respiratory tract is minimal in the course of this movement. In general, metals are absorbed less effectively from the gastrointestinal tract than from the lungs, in part because of differences in residence times (Natusch et al. 1975; Luckey and Venugopal 1977). Thus, material processed by the gastrointestinal tract often has less toxic impact, and may also be influenced by dietary factors.

Dissolution is a major clearance mechanism for metal particles deposited in the pulmonary region, and it may be enhanced if the particles are first phagocytized by macrophages. When deposited particles are ingested and subsequently exposed to the acidic environment of the phagosome, metal ions can be released. Studies of rabbit and human alveolar macrophages exposed in vitro to submicrometer manganese dioxide (MnO₂) particles have shown that the cells of both species were able to dissolve two to three times more of the material than was dissolved in the culture media within the same time (Lundborg et al. 1984, 1985). Dissolved metals can then leave the macrophage, and the lungs, at rates faster than their normal dissolution rate in lung fluid. The early clearance rate of a metal can therefore vary with the form in which it is inhaled. For example, MnO₂, which is insoluble in lung fluid, dissolves in the macrophage, but soluble manganese chloride (MnCl₂) probably dissolves extracellularly and is not ingested, so deposited Mn may clear at different initial rates depending upon its original state (Camner et al. 1985). Furthermore, intracellular dissolution, by enhancing the release of metals into the cellular milieux, can be a mechanism for the local cytotoxic action of some phagocytosed metal particles, for example, lead (Pb) (DeVries et al. 1983). Intracellular dissolution must therefore be considered in models of the pulmonary clearance of particles and in assessment of mechanisms of differential toxicity of metal compounds.

Various characteristics of macrophages have been examined by performing bronchopulmonary lavage after in vivo exposure to metal particles or after direct in vitro exposures. Subchronic particle inhalations may or may not produce a nonspecific increase in macrophage number, depending upon the metal particle. For example, macrophages in rats increased after exposure to nickel oxide (NiO) but not nickel chloride (NiCl₂) or lead chloride (PbCl₂), whereas lead oxide (Pb₂O₃) exposure reduced the numbers of recovered cells (Bingham et al. 1972). Exposures of rabbit alveolar macrophages to soluble chlorides of cadmium (Cd²⁺), Ni²⁺, Mn²⁺, and chromium (Cr³⁺) resulted in significantly reduced viability, with Cd²⁺ being the most toxic (Waters et al. 1975). Except for Cd, these metals produced cell lysis in a roughly concentration-dependent manner and with a relative potency similar to that which characterized the change in viability. It is conceivable that the mode of cell death affects subsequent clearance pathways for a metal. If the cell dies after phagocytosis but does not lyse, the particle-laden cell may be cleared by the mucociliary system. Alternatively, phagocytosis-induced cell lysis would liberate particles for systemic absorption or reengulfment.

Examination of the viability of rabbit alveolar macrophages exposed in vitro to fly ash with adsorbed oxides of Pb, Ni, or Mn showed Pb to be the most cytotoxic, with Mn and Ni exhibiting somewhat lower toxicity (Aranyi et al. 1977). At any specific particle concentration, the effect on viability decreased with increasing particle size over a 2- to 8-µm range, even though the percentage of metal in the ash was the same for all sizes. This finding was ascribed both to the fact that bigger particles were ingested in fewer numbers than were smaller ones and that, once ingested, the larger particles presented less surface area to the phagosomal contents, resulting in less leaching of metal ions. Thus, the cytotoxicity of a metal at a specific concentration depends upon the size of its associated carrier particle and, conversely, the nature of the carrier particle influences the fate and effects of the compound it transports into the cell.

The rate of phagocytosis for equivalent-sized particles depends upon the nature of the adsorbed surface coating. Rabbit alveolar macrophages exposed to 5 µm Teflon particles coated with various metals phagocytosed those with carbon or Cr to a greater degree than those coated with Pb, Mn, or silver (Ag) (Camner et al. 1973, 1974). Direct in vitro exposure to Ni²⁺, Cd²⁺, vanadate (VO₃−), Mn²⁺, or Cr³⁺ impaired the phagocytic ability of macrophages (Graham et al. 1975; Waters et al. 1975; Castranova et al. 1980); in vivo exposure to MnO₂ also depressed phagocytosis (Bergstrom 1977). The effect may depend upon the concentration of the metal. In rats exposed to cadmium chloride (CdCl₂) at 1.5 or 5 mg/m³, the lower level stimulated phagocytosis whereas the higher level depressed it (Greenspan and Morrow 1984). Reduced phagocytosis could result in increased lung burdens of the offending metal or of other deposited substances.

The effects of metals on various cytologic end points may not be equal. For example, Ni²⁺ impaired phagocytosis at a concentration much lower than that required to decrease viability, whereas Cd²⁺ and Cr³⁺ depressed phagocytosis as well as viability at comparable concentrations (Waters et al. 1975). Concentrations of VO₃− that caused macrophage lysis did not reduce phagocytosis in the surviving cells (Graham et al. 1975). Thus, examination of viability as the only toxic end point may not be appropriate. Other functions may be more sensitive and may also play a role in determining the ultimate disposition of deposited particles.

Examples of the disposition of selected inhaled metals found in exhaust—namely Pb, Cd, Cr, Mn, Ni, and vanadium (V) — are discussed below.

Lead. Until recently, Pb was the metal of greatest concern, but as the amount of Pb in gasoline has been gradually reduced, so has the concern. About 20 to 60 percent of inhaled Pb particles deposit in the adult human respiratory tract (Nozaki 1966; Moore et al. 1980; Morrow et al. 1980). Most is rapidly cleared by absorption; lung retention half-times of 13 to 14 hr have been measured (Morrow et al. 1980). Of the Pb cleared to the gastrointestinal tract, only about 5 to 15 percent is subsequently absorbed in adults (Kehoe 1961; Goyer and Chisolm 1972; Baksi 1982), with the rest excreted in the feces. Gastrointestinal absorption is, however, greater in infants and children (Rabinowitz et al. 1976; Ziegler et al. 1978).

The total contribution of airborne Pb to total blood Pb levels is hard to determine, but the percentage of airborne Pb ultimately found in the blood is in the range of 7 to 40 percent (Patterson 1965; Rabinowitz et al. 1973, 1974; Manton 1977), and the biological half-time in blood is about 25 days (Baksi 1982). Absorbed Pb is excreted primarily in urine but also in bile and by exfoliation of epithelial tissue. About 25 to 40 percent of inhaled Pb is retained in the body, almost all in bone, where it accumulates slowly with age and continued exposure (Goyer and Chisolm 1972; Barry 1975; Gross et al. 1975).

Cadmium. The solubility of Cd salts varies widely, and the relative rates of clearance by specific pathways probably depend on specific form. For example, whereas soluble CdCl₂ and insoluble CdO₂ have similar long-term clearance rates, with a retention half-time of ~67 days (Oberdörster et al. 1979), a larger proportion of the oxide is cleared by an earlier fast phase, perhaps mediated by macrophages or mucociliary transport.

Although Cd is absorbed from the respiratory tract, the relation between exposure and uptake in humans is not known. On the basis of experimental animal studies, up to 30 percent appears to be absorbed, depending on the specific form of Cd (Friberg 1950). Absorption from the gastrointestinal tract is quite poor—less than 10 percent in experimental animals and adult humans— but there is increased absorption in young individuals (Rahola et al. 1972; Fleischer et al. 1974).

Once absorbed, Cd accumulates primarily in the liver and kidneys, which together account for about 50 percent of the total body burden (Fleischer et al. 1974). The primary route of excretion of absorbed Cd is urine, and its biological half-time in the human body is estimated at 19 to 38 years (Friberg et al. 1974).

Chromium. After deposition, the water-soluble salts of Cr are rapidly absorbed from the respiratory tract into the circulation, but the less-soluble, and more-toxic, forms remain primarily in the respiratory tract, where their concentration increases with age (Baselt 1982). Lesser amounts of Cr accumulate in skin, muscle, fat, and liver. There is low gastrointestinal absorption, only up to about 25 percent of the initial dose (Baselt 1982). Of the Cr that is absorbed, at least 80 percent is excreted in urine.

Manganese. The use of Mn additives as alternatives to Pb as antiknock ingredients in gasoline will probably result in an increase of Mn in exhaust emissions. A number of studies have examined the clearance of deposited particulate Mn from the lungs of humans and experimental animals (Morrow et al. 1964, 1967a; Maigetter et al. 1976; Bergstrom 1977; Drown et al. 1986). Unfortunately, the data are not directly comparable; residence times vary widely because of differences in particle size and resultant deposition, exposure duration, and concentration. However, like Cd, insoluble and soluble forms may clear at similar long-term rates, but dissimilar short-term rates.

Once it is absorbed, Mn is stored primarily in liver, kidneys, intestines, and pancreas, but does not accumulate with age; it is excreted primarily in bile. Injected radiolabeled Mn has been found to disappear from the human body at two rates: about 70 percent is removed with a half-time of 39 days, and 30 percent with a half-time of 4 days (Mahoney and Small 1968).

Nickel. Respiratory tract clearance mechanisms for Ni are not very effective, and lung levels remain high for years after exposures have ended (Torjussen and Anderson 1979; Williams et al. 1980). The passage of Ni across lung epithelium is slow, and studies in rodents show no significant removal by the lymphatic system (Williams et al. 1980). Because of this slow removal, the concentration of Ni within the respiratory tract increases with age, even during chronic exposure to low levels. In addition, the lung actually sequesters significant amounts of Ni because of binding to a variety of macromolecules; this may be an additional cause of Ni accumulation and toxicity. Of the Ni that reaches the gastrointestinal tract, more than 90 percent is excreted unabsorbed in the feces (Sunderman 1977). Aside from lungs, absorbed Ni tends to localize in connective tissue and kidney. Excretion of absorbed Ni is in the urine.

Vanadium. Vanadium concentrates primarily in fat, which can account for 90 percent of the total body burden, but also in bones and teeth (Schroeder et al. 1963; Myron et al. 1978). Of the other organs, the lungs contain the greatest concentration, but lung kinetics for V are not known. It is possible that insoluble forms may accumulate in the lungs with age, but this is also not known. Any V that reaches the gastrointestinal tract is excreted unabsorbed.

Sulfates

Sulfates in exhaust, primarily sulfuric acid (H₂SO₄) and its neutralization products with atmospheric ammonia, occur in ambient air as submicrometric aerosols. These particles are hygroscopic and their deposition depends upon their effective diameter within the respiratory tract which, in turn, depends upon the rate of particle growth. In guinea pigs and rats, total respiratory tract deposition of H₂SO₄ aerosols ranging in size from 0.4–1.2 µm (MMAD) increased with increasing initial droplet size (Dahl and Griffith 1983). Desposition models developed for hygroscopic sulfate particles (Martonen and Patel 1981) predict that total respiratory tract deposition efficiencies should be greater than those for nonhygroscopic particles only if the sulfate originated from dry particles with diameters greater than about 0.3 µm. Although the regional deposition of H₂SO₄ has not been studied experimentally in humans, predictive deposition models indicate that patterns of deposition for 0.5–1 µm (final size) H₂SO₄ particles are similar, with deposition concentrated in the distal conducting airways (Leikauf et al. 1984).

Sulfate is cleared from the lungs by diffusion (Charles et al. 1977), but the exact rate depends upon the inhaled concentration and associated cations. Using radioactive ³⁵S label, Dahl and coworkers (1983) studied the clearance of inhaled submicrometer H₂SO₄ from the lungs of experimental animals. For dogs, rats, and guinea pigs, they found that the half-time of sulfate clearance from all sites in the lungs ranged from 2 to 9 min, with smaller airways clearing faster than larger ones. The lungs cleared H₂SO₄ much faster than the nasal region, suggesting that clearance from the nose was not primarily by the blood. There were some interspecies differences; the dog cleared slower than the guinea pig, which cleared slower than the rat.

In humans as well as experimental animals, either acute or chronic H₂SO₄ exposure alters the bronchial mucociliary clearance rates of tracer particles. Acceleration or depression of clearance may occur, depending on the concentration and exposure regime. These effects, recently reviewed by Schlesinger (1986), are likely due to the deposition of hydrogen ion (H⁺), rather than sulfate (SO₄ ²⁻), on the airway surfaces. Sulfuric acid exposures of experimental animals have also been associated with alterations in the rate of clearance of tracer particles from the pulmonary region (Phalen et al. 1980; Naumann and Schlesinger 1986; Schlesinger and Gearhart 1986) and with changes in macrophage function (Naumann and Schlesinger 1986; Schlesinger 1987). Effects of H₂SO₄ on respiratory tract clearance are summarized in table 5.

Table 5.

Effects of Sulfuric Acid on Respiratory Tract Clearance.

Research Recommendations

■ Recommendation 29. Factors that control the bioavailability of material adsorbed onto particles should be examined. A major effort should be made to evaluate the effects of carrier particle characteristics, such as size, composition, surface area, and surface characteristics, on translocation and redistribution, intra- as well as extrapulmonary, of adsorbed nonorganic pollutants. Another aspect of this effort involves examination of modifiers of toxic action. For example, macrophages dissolve metals, but there are no comprehensive data to determine if this occurs for all metals of interest, or whether the extent varies significantly among animal species.

■ Recommendation 30. Studies should be undertaken at low concentrations of diesel exhaust which simulate actual human exposure. There is a need to determine whether all aspects of diesel particle disposition are the same for chronic exposures at low dose levels. In addition, studies of diesel particle retention beyond a 100-day postexposure observation time, to more completely assess long-term clearance, translocation, and body retention, should be performed.

■ Recommendation 31. Effects of particulate as well as gas-phase components of diesel exhaust should be studied. Because diesel exhaust is a complex mixture of particles and gases, it is essential that the effects of these components be separated to determine underlying toxicologic mechanisms.

■ Recommendation 32. In experimental animal systems, effects should be determined of concurrent exposures to more than one specific material on clearance pathways, retention patterns, and extrapulmonary disposition. This could involve coexposures to diesel exhaust with other components of ambient air, including cigarette smoke.

■ Recommendation 33. Effects of components of exhaust emissions on susceptible populations, such as those with respiratory disease, should be examined. This can be performed with animal models of specific human diseases. There is a need to use these models, and to develop others, so as to be able to study effects of exhaust products on specific sensitive individuals. Studies are also needed to assess the disposition of pollutants in exercising adults, the young, and the elderly. In addition, the role of concomitant stresses should be assessed in these individuals, as should the question of dose distribution in various extrapulmonary tissues as a function of age.

■ Recommendation 34. The effects of particulate emissions on clearance of other deposited particles should be examined. In addition to sulfates, other materials need investigation in this regard. Studies should assess underlying mechanisms of alteration; for example, changes in bronchial and alveolar epithelial permeability may affect ultimate clearance rates.