INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

SURFACTANT REPLACEMENT THERAPY (SRT) is generally effective, yet 33% of those treated show only a transient response or no response (2). In an attempt to reduce this significant failure rate, a variety of factors related to treatment have been studied. Areas under investigation range from pretreatment ventilation strategies, where it has been hypothesized that large tidal volumes before treatment increase protein leaks into the alveoli, leading to surfactant inactivation (5), to variation of the volume of the instillate; larger volumes have been found to produce a more uniform distribution of surfactant, improving the response to treatment (4, 13). Recent investigations have focused on the method of instillation; they have addressed whether surfactant should be rapidly injected or slowly infused into the lungs (11, 12) and demonstrated that the method of instillation may determine the overall efficacy of treatment. Each of these studies has focused on a single aspect of SRT without an appreciation of the overall surfactant dispersal process. The functional relationships between administration rate, bolus properties, geometry, and lung orientation and the initial deposition or subsequent dispersal of the bolus have not been systematically studied.

The most common procedure for administration of surfactant starts with the neonate removed from the ventilator and placed in one of four positions: head-down, left or right lateral or head-up, or left or right lateral position. The first quarter-dose aliquot is then rapidly injected over 2-3 s into the trachea via a 5-Fr catheter threaded down the endotracheal tube. The neonate is then hand ventilated for 30 s at 30-60 breaths/min before being repositioned, and the next aliquot is injected. This procedure is repeated until all four aliquots have been administered. For example, a 1-kg neonate receives a total of 100 mg of phospholipid suspended in 4 ml of saline, with each quarter-dose being 1 ml (9). This procedure has been found to provide a homogenous distribution, with some transient effects of cyanosis, bradycardia, and increased PCO₂ (14). Alternative instillation strategies have recently been reported that explore the effects of instillation rate (11, 12), multiple instillations (8, 12), lung orientation with respect to gravity (1), and positive end-expiratory pressure (PEEP) (6) on surfactant dispersal and treatment efficacy.

In general, a wide range of instillation rates can be envisioned, ranging from a slow infusion to a rapid injection. On the basis of prior practice we classify delivery as 1) slow infusion, i.e., instillation of 1-10 ml of surfactant over ~1-45 min, or 2) rapid injection, i.e., delivery of the same volume over 2-15 s (1, 8, 11, 12). Segerer et al. (11) and Ueda et al. (12) reported that rapid injection was associated with a relatively uniform distribution and favorable response, in contrast to slow infusion, which resulted in a highly nonuniform distribution and poor response. Additionally, Ueda et al. found that with slow infusion a second dose entered into the same lung units as the first, with little increase in PO₂ after the second-dose. In a similar experiment, Plötz et al. (8) demonstrated that with slow infusion of surfactant the second-dose deposits in the same lung as the first dose, but additionally they found that some surfactant was delivered to surfactant-deficient alveoli. However, distribution remained grossly nonuniform. In contrast to the findings of Ueda et al., an increase in PO₂ was observed after the second dose, signifying that some additional recruitment had occurred. In both studies the second dose was delivered 2 h after the first. The difference might lie in anatomic differences in the animal models: Ueda et al. used preterm lambs, whereas Plötz et al. employed a rabbit model.

Broadbent et al. (1) demonstrated that surfactant accumulation occurring with the infusion mode of instillation was influenced by chest orientation, i.e., deposition favoring the dependent lung with respect to gravity. No redistribution was observed when the rabbit was repositioned after treatment. Recently, Merritt et al. (6) investigated whether keeping the airways inflated with PEEP during rapid bolus injection rather than removing the subject from ventilation during instillation would influence surfactant distribution or reduce adverse transient effects such as decreased oxygenation. The use of PEEP during instillation resulted in a more uniform distribution through the lungs with a smaller drop in O₂ saturation. However, no significant physiological differences were noted between the two instillation techniques after 12 h.

These studies suggest that the initial placement of the bolus may determine the ultimate distribution of surfactant within the lungs. The question as to why rapid injections result in more uniform distributions and, therefore, better therapeutic response (11, 12) has not been carefully studied. As will be shown, rapid injections help ensure meniscus formation in the trachea and main stem bronchi, from which a more uniform dispersal process can proceed. The first inspiration after instillation will drive the meniscus down all parallel pathways, coating the airways with surfactant, with immediate deposition of some fraction of the bolus in the periphery. Subsequently, delivery of surfactant to initially untreated air spaces can continue via surface tension gradients (3). In this manner, the surfactant is distributed more uniformly throughout both lungs rather than simply flowing into the dependent lung according to gravity, as observed with slow infusion (11, 12).

Our aim is to identify the critical parameters of surfactant instillation and their influence on the initial deposition of an instilled surfactant bolus. In particular, we seek to identify the conditions under which a meniscus will form in the trachea. In this in vitro experiment we considered the effects of bolus viscosity, injection rate, orientation of the trachea with respect to gravity, surface tension, and tracheal size. This study provides a framework for examining present instillation techniques and for exploring alternative strategies.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Assumptions. Some assumptions and simplifications have been made to focus on the fundamental physical factors influencing the injection process. Clinical practice involves surfactant administration by syringe through a 5-Fr catheter slipped through the endotracheal tube or via a side-port adapter atop the endotracheal tube. The former technique is shown schematically in Fig. 1. To capture both methods in a single experimental setup, we considered the case where the injected fluid from the catheter or the infused surfactant from the side-port adapter flows along the tracheal wall or down the endotracheal tube by placing the catheter near the model airway wall. The results are extrapolated to conditions where the catheter does not lie along the side of the trachea and are addressed in DISCUSSION.

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Schematic of endotracheal tube and catheter in relation to trachea during surfactant injection.

Bench-top experiments. The conditions under which a meniscus plug formed in the trachea was examined by using a simple bench-top model of the trachea-and-catheter combination (Fig. 2). The trachea was modeled by a 20-cm-long glass capillary tube. This length was used to avoid any end effects on meniscus formation. For accurate placement, the flexible catheter was replaced by an 18-gauge syringe needle (~1 mm ID) with a blunt tip and mounted on a positioning rail. The needle was connected by tubing to a 30-ml syringe mounted on a syringe pump (model 944, Harvard Apparatus, S. Natick, MA). The needle and capillary components were rigidly secured to a Plexiglas base that could be rotated in the vertical plane. We examined how fluid viscosity, capillary diameter, orientation, surface tension, and injection flow rate influenced meniscus formation.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Layout of experimental apparatus. Injection system consists of a syringe needle (d) mounted on a rail system for accurate placement within capillary (D). Needle is connected to a syringe pump by pressure tubing (not shown) to produce flow (Q). Entire setup is mounted on a rigid Plexiglas base that can rotate in vertical plane (). g, Gravity.

1

1

1

1

1

1

3

Analysis. To identify the parameters that determine whether a meniscus forms, a simple analysis is given before we present the experimental results. Figure 3 shows a bolus being injected into a capillary oriented at an angle  and accumulating at some downstream position. As a first approximation, the meniscus-forming process can be considered as one in which the liquid cannot drain away from the injection site as fast as it is introduced. If the liquid accumulates to some critical depth, surface tension forces eventually dominate, drawing the liquid around the circumference to form a meniscus that occludes the tube. The effects of gravity, viscosity, and jet momentum are considered in this analysis.

View larger version (11K): [in this window] [in a new window]   View larger version (8K): [in this window] [in a new window]   Fig. 3.   A: injected bolus collecting in capillary before meniscus formation. B: enlarged view of liquid pool showing control volume (dashed lines) used in analysis. L, length; h, pool depth; D, capillary diameter.

(1)

(2)

(3)

(4)

(5)

(6)

(7)

	RESULTS

Top Abstract Introduction Methods Results Discussion References

The purpose of these experiments was to obtain a criterion for the circumstances under which a meniscus would form. The approximate analysis in METHODS suggests that if the experimental conditions are mapped onto a plot of Stokes number vs. Reynolds number, it should be possible to distinguish a region of meniscus formation. This was done for a wide range of experimental conditions (Table 1) leading to 10³ < N_St < 1 and 10 < Re < 10³, with the results plotted in Fig. 4. A well-defined boundary is observed that separates conditions leading to meniscus formation from those that do not. Experiments with  = 0 are not plotted here, but all horizontal cases examined led to meniscus formation. All experiments were performed with and without surfactant; no difference was observed in terms of whether a meniscus formed.

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Plot of Stokes number (NSt) vs. Reynolds number (Re) maps experimental results into 2 separate regions of meniscus and no-meniscus formation. NSt = µQ/D4gsin and Re = QD/d2µ. Power law for dividing line is approximately NSt  0.004Re2/3. A-F, possible conditions during surfactant replacement therapy: µ = 0.2 and Q = 0.2 (A), µ = 0.2 and Q = 0.5 (B), µ = 0.5 and Q = 0.2 (C), µ = 0.5 and Q = 0.5 (D), µ = 0.2 and Q = 0.5 (E), and µ = 0.2 and Q = 0.01 (F); µ is expressed in g · s1 · cm1 and Q in ml/s;  = 20°, except for E, where  = 5°.

The separation boundary shows a weak Reynolds number dependence and is approximately described by a power law relation of

(8)

which is shown by the solid line in Fig. 4. Therefore, for N_St > 0.004Re^2/3, a meniscus will occlude the airways, provided a sufficient volume of liquid is injected.

The volume requirement was determined from examining recordings of the experiments. With knowledge of the time elapsed from when the bolus first emerged from the needle until a meniscus formed and the flow rate, the volume injected was determined. This gave rise to the additional constraint that a volume >1.6D³ is required for a meniscus to form. This relationship was confirmed with a smaller series of tests (n = 20) without surfactant.

Two examples are presented to illustrate how the physical variables interact to influence meniscus formation. Consider, for example, the effect of orientation with respect to gravity, as characterized by the term gsin. To see this effect, one starts at a point within the no-meniscus region (e.g., N_St = 0.005, Re = 40, Fig. 4A) and holds Re constant while varying gsin. Increasing the inclination angle increases the influence of gravity (decreases Stokes number) and moves one further from the meniscus criterion. The bolus drains away more rapidly because of an increased contribution of gravity, and no meniscus forms. In contrast, when the inclination angle is decreased, one moves toward the meniscus region (increasing Stokes number). Once the gravitational influence has been sufficiently reduced such that one crosses into the meniscus region, gravity carries the liquid away at a slower rate than it is supplied, increasing the thickness of the liquid layer until a meniscus is formed.

Now consider the dependence of meniscus formation on flow rate, a quantity easily controlled by a clinician. Substituting Eqs. 5 and 7 into Eq. 8 for Stokes and Reynolds numbers and collecting terms give a minimum flow rate (Q_crit), a criterion for meniscus formation as a function of the other variables

(9)

This result has implications for surfactant administration techniques and will be further discussed in connection with six possible clinical situations (Fig. 4, A-F), which are summarized in Table 2.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

We have identified the governing parameters and determined functional relationships that describe the initial fate of a bolus injected into the trachea. This process is the first phase of SRT and, according to previous animal studies, may determine the overall surfactant distribution and ultimate efficacy of treatment (8, 11, 12). The results from our in vitro experiments indicate that the instilled bolus accumulates locally to form a meniscus or drains away toward the dependent airways with respect to gravity. This behavior is characterized by two parameters: 1) Stokes number, a ratio of viscous to gravitational effects, and 2) the bolus Reynolds number, a ratio of inertial to viscous effects. These take into account the effects of orientation with respect to gravity, viscosity, flow rate, airway size, and momentum.

Animal experiments in the literature indicate that a more homogenous distribution and favorable response occur when the bolus is given as a rapid injection than when it is administered by slow infusion (11). From our results, we observe that a rapid injection promotes meniscus formation, filling the trachea and bronchi with the bolus. A meniscus effectively collects the injected liquid, preventing it from flowing down dependent airways. On the next inspiration the bolus could presumably be drawn into the lungs in proportion to regional lung expansion, producing a distribution more uniform than would be achieved by gravity. This study provides a basis for understanding how the bolus is initially situated after instillation, before it is dispersed through the lungs.

Current protocol calls for positioning the infant in different orientations for each quarter-dose administered (9). Table 2 lists some variations that might exist during treatment (Fig. 4). Cases A-D examine how changes in flow rate and viscosity could alter the deposition process at a fixed orientation of 20°. For example, for administration of Survanta (Ross Laboratories) to a 1-kg neonate, it is recommended that a quarter-dose of 1 ml be given over 2-3 s, corresponding to an injection flow rate of 0.5-0.3 ml/s. Even greater rates of 1 ml/s have been used in animal experiments (11). The injected volume of 1 ml is more than adequate to satisfy the constraint that volume is >1.6D³; therefore, Fig. 4 can be used to determine whether a meniscus forms. Beginning with case A, we start out at a location in Fig. 4 where no meniscus forms and the liquid flows freely to the dependent region of the lung. Increasing the flow rate (case B) or increasing the viscosity (case C) leads to more accumulation of surfactant in the trachea, producing situations near the boundary of the two regions, suggesting that the outcome would be inconsistent, sometimes leading to meniscus formation and other times not. The combination of rapid injection and high viscosity (case D) or a shallow angle of inclination (case E) results in a meniscus being formed. An alternative approach to rapid injection is administration of surfactant by slow infusion to reduce transient side effects (6). This condition (case F) clearly results in no meniscus being formed.

In light of the scenarios just presented, how might the instillation process be better controlled? Those parameters under clinician control are the flow rate, orientation, and to some degree the catheter diameter; the other variables are fixed by physiology or are physical properties of the liquid. Equation 9, derived from the criterion for meniscus formation (N_St > 0.004 Re^2/3), provides the functional relationship and sensitivity of each variable on the instillation flow rate. This criterion is strongly influenced by the size of the trachea or endotracheal tube (depending on the location of the catheter) and moderately affected by liquid density and viscosity, orientation, and catheter size. Equation 9 provides a lower limit for flow rate if a meniscus is to be achieved. Conversely, it can be viewed as an estimate of the maximum flow rate that can be used without producing airway obstruction. For example, inclination of 5°, airway diameter of 0.4 cm, catheter diameter of 0.1 cm, liquid density of 1 g/cm³, and gravity of 981 cm/s² reduce Eq. 9 to Q_crit = (5.2 × 10³)/µ² for viscosity expressed in grams per second per centimeter and Q_crit in milliliters per second. With viscosity of 0.2 g · s¹ · cm¹ (case E in Table 2), Q_crit = 0.18 ml/s (i.e., for a 1-ml aliquot, injection time is ~5.5 s). The bolus will occlude the airway for larger flow rates but would likely drain along the airway for smaller administration rates encountered during treatment.

To locate where current treatment lies, in Fig. 4 we obtained unused portions of Survanta (lot no. 16 916 Z7) and measured its viscosity using a Cannon-Fenske-Ostwald-type bulb viscometer (models 200 and 350, International Research Glassware). At 25 and 37°C the viscosity was ~100 and 75 g · s¹ · cm¹, respectively. For injection flow rates of 0.2 and 0.5 ml/s, this places one above and to the left of points C and D in Fig. 4, for inclination of 20°, airway diameter of 0.4 cm, catheter diameter of 0.1 cm, liquid density of 1 g/cm³, and gravity of 981 cm/s². Thus, in a clinical setting under these physical conditions and for this surfactant preparation, it is likely that a meniscus forms unless the injection time is >5 s (Q  0.2 ml/s).

It is appropriate to comment on the extent to which these results are valid and can be applied with confidence in different situations. The catheter position plays an important role in determining whether the bolus forms a meniscus in the trachea or drains into a main stem bronchus. The results presented here were obtained with the catheter positioned at the wall (Fig. 2). A meniscus will form if the criterion is met (N_St > 0.004 Re^2/3), and the liquid will drain away toward a main stem bronchus if the criterion is not met. In the instance when the catheter is located away from the wall and directed along the axis of the trachea, however, the liquid can leave the catheter as a stream, landing at a point downstream from the tip of the catheter. If the meniscus criterion is satisfied, a liquid plug will still form, but at the more distal location. As the bolus is more rapidly injected or the catheter is placed further within the trachea, the bolus may directly enter a main stem bronchus. With all other parameters fixed, the smaller diameter of the bronchus promotes meniscus formation because of the strong dependence of the Stokes number (N_St ~ D⁴) and volume requirement (V ~ D³) on diameter. With a daughter-to-parent diameter ratio of 2:3 for the bronchus and trachea bifurcation, the Stokes number for the bronchus is approximately five times greater than the tracheal value with a decrease in the Reynolds number by two-thirds of the tracheal value. The volume requirement also drops to 30% of the volume needed in the trachea. For example, a case where the Stokes number is 10² and the Reynolds number is 20 on the basis of tracheal diameter produces conditions near the border between the meniscus and no-meniscus regions (Fig. 4). However, if the bolus streams into the bronchus, the Stokes number increases to 0.05, moving well within the region where a meniscus would form.

Is it possible to form a meniscus in an adult lung? For a given set of bolus properties and fixed orientation, an adult tracheal diameter of ~2 cm reduces the Stokes number by a factor of 625 with respect to the value obtained for a neonate with a capillary diameter of 0.4 cm. From Fig. 4, clearly no meniscus will form in the trachea of the adult by this method, nor will a meniscus form in the adult main stem bronchus, where the Stokes number is 123 times smaller than in the neonatal trachea. Drastic measures such as increasing the viscosity 1,000-fold to offset the geometric size is neither realistic nor desirable, suggesting that meniscus formation may not be practical in the adult.

It may seem strange that surface tension was not found to play an important role in this process and was not included in our scaling analysis. The reason for omitting surface tension effects comes from several observations. First, as shown by Otis et al. (7), although surface tension affects the rate at which a meniscus forms from an initially uniform liquid layer on the wall, it does not alter the conditions (e.g., thickness of the liquid layer) necessary to produce meniscus formation. Second, the effects of surface tension gradients that might arise during injection to redistribute the liquid are likely to be small. This statement is based on the observation that in the present experiments in which the surface tension difference between the endogenous liquid (pure water) and exogenous liquid was maximized, there was little evidence that flows driven by surface tension gradients were affecting meniscus formation, at least on the time scales of injection. Finally, these heuristic arguments gain support from the observation that surface tension had no discernible influence on the experimental outcome.

Once the meniscus is formed, the bolus no longer preferentially drains to dependent regions of the lung with respect to gravity. The bolus volume can fill the major bronchi and part of the trachea. With the bolus in place, the surfactant can be delivered throughout the lungs as it is advanced distally on the next inspiration.