J. Paediatr.

Child Health (1997) 33,471-475

Viewpoint

How does surfactant really work?

BA HILLS
Paediatric Respiratory Research Centre, Mater Children’s Hospital, South Brisbane, Queensland, Australia

Abstract: Although administration of exogenous surfactant to the neonate with RDS often relieves hypoxia rapidly, it is events
occurring some 18-48 h later that determine the ultimate clinical outcome, indicating a vital second stage to ‘rescue’. It is
proposed that, whereas a reduction in surface tension facilitates the initial penetration of the lungs by air, the second stage is
provided by surface-active phospholipid (SAPL) slowly adsorbing (binding) to alveolar epithelium to displace water in the same
way that surfactants are widely used industrially as ‘de-watering’ agents. In the normal newborn, this de-watering lining of SAPL
is already in place, explaining the extremely rapid expulsion of fluid from the lungs at birth. The selection or formulation of a
surfactant for rescue should thus take both stages into consideration.

Key words: alveolar models; lung fluid balance; respiratory distress syndrome; surfactant.

Surfactant replacement therapy, often termed ‘surfactant
rescue’ of premature infants afflicted with the respiratory
distress syndrome (RDS), is now a well accepted clinical
modality which has reduced infant morbidity and mortality
significantly despite improvements in neonatal care.lS2However,
many features are not explained by conventional theory,
especially those raised by Milnef‘ in which he emphasizes that
the establishment of normal lung physiology often lags behind
improvements in gas exchange by as much as 24 h.
CONVENTIONAL THEORY
The model adopted for the alveolus is as relevant as the
properties of surfactant itself in determining its role in the lung,
since it cannot reduce the surface tension of an interface which
does not exist. Conventional theory follows the Clements-
Hawgood in which surfactant locates at the air-
aqueous interface of a continuous liquid layer, or aqueous
hypophase, assumed to line the alveoli and adjacent terminal
airways at all times. This continuous fluid lining (Fig. l c ) can be
regarded as a one-sided bubble whose collapsing pressure
(AP) would be far too large to be physiologically compatible
unless surfactant greatly reduced the surface tension (y) from
the very high value for water, as related by the Laplace
equation:
AP=F?/r (1)
where r is the radius of curvature. Conventionally, the same
model is applied to both adult and neonatal lungs, this situation
arising largely because the whole field of pulmonary surfactant
tends to be dominated by perinatologists and scientists studying
RDS. They are undoubtedly starting with wet lungs, but the
Correspondence: Professor BA Hills, Paediatric Respiratory Research
Centre, Mater Children’s Hospital, South Brisbane, Old 4101, Australia.
BA Hills, PhD, DSc, ScD, Professor.
Accepted for publication 29 May 1997.
resulting ‘bubble’ model of the alveolus (Fig. l c ) has been
extrapolated unwittingly from pathological to physiological lungs
and from neonatal to adult lungs.
THE ADULT LUNG
The primary feature of the ‘bubble’ model is the continuous
liquid lining, but this does not appear to be present in the
mature or fully aerated infant lung. In the classical morphological
studies of GiI6 and Weibel? lungs were first frozen in order to
lock the fluid in position prior to fixation. In mature lungs they
demonstrated fluid confined to ‘pools’ at the septa1 corners and
to ‘pits’ elsewhere along the alveolar surface as though ‘pushed
aside’ to prevent a fluid barrier impeding gas exchange.
Elsewhere along this gas-exchange surface, the alveolar wall
appeared free of fluid as depicted in Fig. I d . The ‘pits’ and
‘pools’ display an edge by which they can ‘grip’ and, hence,
tension the alveolar wall to pull it into the same contour as the
liquid-air interface, although the contact angle is sometimes
obscured by the epithelium folding under the pool surface.’ In
tensioning the wall, excess epithelium is gathered into the
‘pools’ in the form which GiI6 and Weibe17 appropriately term
‘pleats’. Thus the ‘pits’ and ‘pools’ suffice to explain a major
mechanical feature of the excised mature lung; namely, the
roughly four-fold increase in compliance upon liquid fillingg and,
in this tensiolytic role, it is unnecessary for the liquid lining to
be continuous.
There are many other aspects of lung mechanics that are
more compatible with a discontinuous fluid lining as discussed
in detail elsewhere,” but two require special mention. The first
is that any series of interconnected bubbles is inherently
unstable, so this is no problem where there is no bubble in the
first place (Fig. 1d). It is interesting that, if one instills a wetting
agent such as Tween 20 into the lung, then atelectasis does
occur,” presumably because it has now wetted the intervening
epithelium, allowing the ‘pits’ and ‘pools’ to link up and form
472 01%Hills

\ alveolar
wall

\‘
aqueous
hypophase

surface

Fig. I Depicting (a) a central core of air being established in an alveolus at birth which then (b) enlarges as the pressure generated by inherent homeostatic
pumping mechanisms (F) exceeds the collapsing pressures (AP) arising from the air-aqueous (‘bubble’) interface-@) the conventional ‘bubble’ model of the
normal alve01us~~~ in which fluid would need to stop receding at a particular thickness in order to leave a continuous aqueous hypophase on which DPPC can
locate-(d) the alternative ‘adsorption’ model redrawn from Hills” based upon the distribution of fluid demonstrated morphologically in the normal adult alveolus.
Note how surfactant is directly adsorbed to epithelium and how any alveolar oedema reverses the curvature of the air-aqueous interface of a pool at a septal
corner (e), reversing the direction of the surface force so as to assist in resolving the additional fluid.

bubbles. However, this is now a pathological state and not the
normal physiological state of these lungs.
The other major requirement of the ‘bubble’ model is that
surface tension of the liquid-air interface is ‘near zero’;5”‘
otherwise the force (AP) of the concave fluid surface sucking
fluid into the air space at the septal corners ( r l l in eqn 1)
would be impossible to balance by normal homeostatic
mechanisms. Proponents of the bubble model have claimed
that surface tension (y) can reach ‘near-zero’ transiently during
the respiratory cycle, but these have been hotly disputed by
others as artefact10v13and described as ‘absurd’ by consider-
ations of basic surface physics.14
DE-WATERING
To this investigator, the fluid profile of ‘pits’ and ‘pools’ in the
adult alveolus is particularly reminiscent of surfaces to which
How does surfactant really work?
surfactants have been applied similar to what industrial chemists
term ‘de-watering’ agents in the United Kingdom and ‘de-
wetting’ agents in the United Sates of Amercia. It is very well
accepted in the physical sciences that these surfactants
function by directly binding to the surface via their positively
charged polar groups, orientating their non-polar (often fatty
acid) groups outwards to form a more hydrophobic (less
wettable) external surface.’0315This causes water to ‘bead up’
in extreme examples, as observed in everyday life when pouring
water on to a waxed surface. The same phenomenon has been
reported after exposing normal amnionic membrane to air when
a saline droplet displays a strong tendency to ‘bead up’.’6 This
observation could be particularly pertinent to alveolar epithelium
when considering that both surfaces have been in contact with
fetal surfactant aspirated in utero for ample time for adsorption
to occur. In this regard, it has been pointed out how the
extremely surface-active molecule in indigenous pulmonary
surfactant and its exogenous forms (dipalmitoyl phosphatidyl-
choline [DPPC]) possesses the same terminal quaternary
ammonium ion ideal for binding to negative sites on epithelium
as exploited in many commercial de-watering agents.” The
physical chemistry of surfactant adsorption to solid surfaces
has been described in more detail elsewhert?.”
There is a good example of de-watering in the eye where a
monolayer of surface active phospholipid (SAPL) adsorbed to
corneal epithelium is now known to cause ‘break-up’ of the
tear film unless we blink every 20 s or 50.’~
ADSORPTION OF SURFACTANT
The honeycomb structure of the distal airspaces makes the
surface physics of the lung much more difficult to study than
tissues such as the eye, but the morphological evidence for
adsorption consists firstly of electronmicroscopy in which Ueda
et a/. have produced a series of electron micrographs of superb
quality in which they demonstrate an additional monolayer
adjacent to the lipid bilayer of the alveolar membrane in
animals’* and multiple layers in human^.'^ We have confirmed
their findings in this laboratory. These findings are challenged
by advocates of the conventional (‘bubble’) model on the basis
that Ueda’s use of tannic acid as a fixative has allowed
surfactant to migrate during fixation. However, from his wealth
of experience using conventional fixatives, Weibel” claims that
the osmiophilic (surfactant) layer follows the tissue surface
rather than the air-aqueous interface (i.e. compatible with
Fig. 1d). Use of hydrophobic probes under epifluorescence
microscopy confirms Ueda’s finding^,'^ and there is no migration
of surfactant with this procedure.
WET-TO-DRY TRANSITION
Returning to the neonatal lung, there is no disputing that the
first (aeration) stage of ‘surfactant fescue’ involves the establish-
ment of ‘cores’ of air within the alveoli (Fig. l a ) in which the
reduction of its surface energy by DPPC enables that interface
to expand as rapidly as surfactant can be recruited. This stage
actually decreases compliance initially as the collapsing pressure
of the newly formed bubbles (AP in eqn 1) now augments
lung recoil.
The net force resolving the remaining fluid (Fig. 1b) remains
the difference between the collapsing pressure of the air core
473
(AP) sucking fluid into the air space and the approximately
constant force F pumping fluid out into the interstitium by
inherent homeostatic mechanisms manifest in the adult lung, at
least, by a negative interstitial pressure:”
Net force resolving fluid =F-2y/r (2)
As the residual fluid is resolved and the bubble expands (rt in
eqn 2), the net force should increase, causing the remaining
fluid to be resolved at an ever increasing rate: provided
surfactant recruitment to the air-fluid interface can be main-
tained to limit y. Thus it is very difficult to envisage the air-fluid
interface stopping at just the point needed to provide the
continuous liquid layer, or aqueous hypophase on which
surfactant ‘sits’ according to the conventional ‘bubble’ model
depicted in Fig. l c . In fact the proponents of the ‘bubble’ model
have never put forward any mechanism for controlling the
thickness of the continuous liquid layer which is otherwise
inherently unstable for reasons expounded in detail
elsewhere.10322
FLUID BALANCE
It would seem far more likely that, as fluid recedes at birth and
the bubble surface reaches epithelium, it will expose apparently
fluid-free areas, resembling an ebb tide exposing isolated pools
on a rocky foreshore, rather than stopping short at some
undefined fluid thickness. In the normal neonate the de-watering
of alveolar epithelium would be facilitated by prior adsorption
(i.e. direct binding of surfactant) to the tissue surface and we
have preliminary morphological evidence from a fetus stillborn
at term that this is the case. Fluid pushed aside from the gas
exchange surface into ‘pits’ and ‘pools’ can, if needed, assume
a convex profile with respect to air while still tensioning the
alveolar wall” (see Fig. 1d). The air-fluid interface will now push
fluid in the desired direction, (i.e. into the interstitium in support
of the inherent fluid pump) and will continue to do so until
excess fluid is resolved and the pool profile has become flat
(r= cc in eqn 2) or concave (r<O). Hence the surface mechanism
is self-regulating in Fig. I d . In adult lungs, it has been shown
that the ‘pools’ of alveolar fluid are normally slightly c~ncave,’~
but can become highly convex as oedema forrn~.’~,’~Thus,
once an ‘adult’ fluid profile is established, surface forces will
support the inherent homeostatic mechanisms maintaining fluid
balance (Fig. I d ) where, for a convex (oedematous) ‘pool’, A P
can now be added to the force ( F ) generated by inherent
homeostatic mechanisms.
In fact, in the oedematous state, when the pools are convex
(Fig. Id), an equilibrium surface tension of 26 dynes/cm would
be preferable to a ‘near-zero’ value of y by offering a larger
force (AP in eqn 1) returning oedema fluid to deeper structures.
Equilibrium values of surface tension are far more likely to
prevail whenever the liquid-air interface exists in the alveolus
for reasons discussed in detail elsewhere” and is the y-value
produced by exogenous surfactants administered for RDS,
including ExosurfZ6 as formulated by the proponents of the
‘bubble’ model. It is, perhaps, unfortunate that clinical success
in the use of Exosurf in reducing infant mortality and morbidity
has been taken as tacit confirmation of the ‘bubble’ model.
Only recently has debate been tolerated of the scientific issues
involved; while many features of the ‘bubble’ model are now
being ~hallenged,’~ as reviewed above and in more detail
el~ewhere.’~
474
EVENTS AT BIRTH
When air starts to replace fluid in the lung at birth, it forms
central cores within alveoli as demonstrated morphologicallyz8
and depicted in Fig. l a , the rapid location of surfactant at the
air-liquid interface enabling that surface to expand rapidly.
Hence, in premature lungs deficient in surfactant, it is easy to
envisage how the administration of exogenous SAPL will cause
the interface to advance, forming air ‘cores’ in more alveoli
whose recruitment to the gas exchange process largely
alleviates the hypoxia. Thus, the perinatologist can turn down
the ventilator and oxygen setting, often within a few minutes of
an initial bolus of Exosurf which is designed specifically to
spread rapidly over an air-water interface. However, in most
cases, the infant cannot be weaned from the ventilator because
lung compliance is too high and functional residual capacity
(FRC) is not established, while the chest radiograph has failed
to clear.3
SECOND STAGE TO RESCUE
Following successful and rapid aeration of the lung, there is
clearly a second stage to ‘surfactant rescue’ with events
occurring some 18-48 h later often determining the ultimate
clinical outcome,‘ there being no correlation between speed of
the initial response and overall mortality rate.’ This delay
applies to all formulations including extracts of natural surfactant
for which improvements in lung mechanics are often not
apparent until 24 h after initiation of treatment.”
THE VITAL QUESTION
The vital question is what causes the delay when, by the
conventional ‘bubble’ model, the two stages should coincide,
as they do so rapidly in the normal neonate.
If advance of the liquid-air interface in the normal neonate
does not stop at some undefined liquid thickness needed to
provide a ‘bubble’ but proceeds towards the epithelial surface
(Fig. 1d), then pre-adsorbed surfactant can rupture the remaining
film pushing fluid aside just as industrial de-watering agents
can dispose of water almost instantaneously. This can explain
the extremely rapid dissipation of lung fluid upon the first breath
in a normal birth where it is difficult to envisage the inherent
physiological pumping mechanisms providing such rapid expul-
sion of fluid without physical assistance.
Since adsorption is an equilibrium phenomenon between
deposited surfactant and that in the adjacent fluid,15 the same
deficiency which applied to enlargement of the air core (Fig. l b )
will also apply to the amount bound to the surface. However,
adsorption is a much slower process than spreading across a
liquid-air interface” and so, even when exogenous surfactant
has access to an alveolus, it will take a long time to diffuse
across the fluid adhering to the alveolar wall and be adsorbed
in sufficient quantity to act as a de-watering agent. The very
low solubility of DPPC in water exacerbates the problem. Thus,
establishment of an ‘adult‘ fluid profile can explain clearing of
the chest radiograph by de-watering and increase in lung
compliance (in establishing FRC) as the recoil contributed by
surface forces is reduced from that of a continuous bubble
surface (Fig. 1 b) to those contributed by individual fluid
‘pools’ (Fig. 1 d).
BA Hills
SELECTION OF SURFACTANT
This two stage model for surfactant replacement therapy has
certain implications for the choice of surfactant to use in the
’rescue’ of neonates with RDS. Exosurf is ideally formulated for
spreading over the air-aqueous interface, explaining rapid
alleviation of hypoxia, but the incorporation of Tyloxapol, used
commercially as a wetting agent, may not be compatible with
the de-watering action needed at the alveolar epithelial surface.
Hexadecanol (another component of Exosurf) is a powerful
solvent which might elutriate any adsorbed DPPC in the lung
or other organs where it has been claimed to render tight
junctions tight:’ possibly explaining some of the side-effects
such as intraventricular haemorrhage?’ Greater success in the
overall outcome of surfactant replacement therapy has been
reported’ in formulations which retain the surfactant proteins,
especially the hydrophobic apoproteins SP-B and SP-C whose
role is attributed to facilitating the location of DPPC at the air-
aqueous interfa~e.~’ These proteolipids could also be involved
in the transport and adsorption of DPPC to alveolar epithelium
necessary for de-watering, just as it has long been known that
a proteolipid is essential for the similar deposition of DPPC on
to axons as m ~ e l i n .Tubular
~~ myelin is a well accepted
intermediate step in the transformation of lamellar bodies after
secretion on to the alveolar surface.34
An interesting form of ‘surfactant rescue’ is provided by
ALEC, especially when dispersed as a micronised ‘fog’ of solid
particles into the ventilator air as it was originally a d m i n i ~ t e r e d . ~ ~
This is a mixture of DPPC:phosphatidylglycerol (PG) in the
proportions (7:3) which, if dispersed in air, can now reach the
‘leading edge’ of the air interface to deposit surfactant in its
more surface-active ‘dry’ state.36 Ongoing studies in this
laboratory are showing a propensity for PG to adsorb to solid
surfaces, and that this can occur across aqueous barriers,
although slowly. This could explain why the salutary effects of
ALEC are often not apparent until some 18 h after adminis-
tration: but the ultimate clinical outcome is good.
CONCLUSION
Despite rapid improvements in gas exchange imparted by
surfactant replacement therapy, the lag in establishing normal
lung physiology can be easily explained by extending the role
of surfactant from one of acting purely at the air-water interface
to its capability to be adsorbed to solid surfaces. Adsorbed to
alveolar epithelium, it can then facilitate ‘de-watering’, a concept
based upon principles very well established in the physical
sciences and the vast industrial use of ~ u r f a c t a n t s . ’ ~ ~ ’ ~
REFERENCES
1 Corbet A. Clinical trials of synthetic surfactant in the respiratory
distress syndrome of premature infants. Ctin. Perinatol. 1993;
20: 737-60.
2 Shapiro DL, Notter RF. Controversies regarding surfactant therapy.
C h . Perinatol. 1988; 15: 891 -02.
3 Milner AD. How does exogenous surfactant work? Arch. Dis.
Childhood 1993; 6 8 253-4.
4 Clements JA. Surface phenomena in relation to pulmonary function.
Physiologist 1962; 5: 11 -28.
5 Hawgood S, Clements JA. Pulmonary surfactant and its proteins.
J. Clin. Invest. 1990; 86: 1-6.
6 Gil J, Bachofen H, Gehr P, Weibel ER. Alveolar volume-surface
How does surfactant really work?
area relation in air- and saline-filled lungs fixed by vascular
perfusion. J. Appl. Physiol.: Respirat. Environ. Exercise Physiol.
1979; 47: 990-1 001.
7 Weibel ER, Limacher W, Bachofen H. Electron microscopy of rapidly
frozen lungs: Evaluation on the basis of standard criteria. J. Appl.
Physiol. 1982; 53:516-27.
8 Hills BA. Role of surfactant in the alveolus. J. Appl. Physiol. 1985;
59: 660-2.
9 von Neergaard K. Neue Auffasungen uber einen Grundbegriff der
Atemmechanik die Retraktionskraft der Lunge, abhangig von der
Oberflachenspannung in den Alveolen. Z. Ges. Exp. Med. 1929;
66: 373-94.
10 Hills BA. The Biology of Surfactant. Cambridge University Press,
Cambridge. 1988.
11 Nieman GF, Bredenberg CE, Clark WR, West NR. Alveolar function
following surfactant deactivation. J. Appl. Physiol. 1981; 51:
895-954.
12 Guyton AC, Moffatt DS, Adair TH. Role of alveolar surface tension
in transepthelial movement of fluid. In Robertson B, van Golde
LMG, Batenburg JJ, eds. Pulmonary Surfactant. Elsevier,
Amsterdam, 1984; 171-86.
13 Barrow RE, Hills BA. A critical assessment of the Wilhelmy method
in studying lung surfactants. J. Physiol. 1979; 295: 217-27.
14 Bangham AD. Lung surfactant: how it does and does not work.
Lung 1987; 165: 17-25.
15 Adamson AW. Physical Chemistry of Surfaces. 2nd edn. Wiley, New
York. 1967; 352-67.
16 Hills BA, Cotton DB. Release and lubricating properties of amniotic
surfactants and the very hydrophobic surfaces of the amnion,
chorion and their interface. Obstet. Gynecol. 1986; 68: 550-4.
17 Holly FJ. Tear film formation and rupture. In Holly FJ, Lamberts
DW, MacKeen DL, eds. The Preocular Tear Film. Dry Eye Institute,
Lubbock, Texas, 1986; 634-45.
18 Ueda S, lshii N, Matsumoto S, Hayashi K, Okayasu M. Ultrastructural
studies on surface lining layer (SLL) of the lungs. Part II. J. Jap.
Med. SOC.Biol. lnterface 1983; 14: 30-5.
19 Ueda S, Kawamura K, lshii N et a/. Ultrastructural studies on
surface lining layer of the lungs: Part IV. Resected human lungs.
J. Jap. Med. SOC.Biol. lnterface 1985; 16: 34-60.
20 Weibel ER. Lung cell biology. In Fishman AP, Fisher AB, eds.
Handbook of Physiology, The Respiratory System Section 3, Vol. 111.
Amercian Physiological Society, Washington. 1985; 47-91.
475
21 Guyton, AC. A concept of negative interstitial pressure based on
pressures in implanted perforated capsules. Circ. Res. 1963;
12: 399-414.
22 What is the true role of surfactant in the lung? Thorax 1981; 36: 1-4.
23 Gil J. Alveolar surface, intra-alveolar fluid pools and respiratory
volume changes. J. Appl. Physiol. 1983; 54: 321-4.
24 Derrner GB. The fixation of pulmonary surfactant for electron
microscopy. I. The alveolar surface lining layer. J. Ultrastructure
Res. 1969; 27: 88-104.
25 Kisch B. Electron microscopy of the lungs in acute pulmonary
edema. Exp. Med. Surg ,1958; 16: 17-8.
26 Perkins WR, Dause RB, Parente RA et a/. Role of lipid polymor-
phisrnin pulmonary surfactant. Science 1996 2 7 3 330-32.
27 Scarpelli EM, Mautone AJ. The surfactant monolayer theory does
not explain surfactant function in vivo. Pediatric Pulmonol. 1995;
19: 198-202 and 1995; 20: 324-33.
28 Scarpelli EM, Kumar A, Doyle C, Clutario BC. Functional anatomy
and volume-pressure characteristics of immature lungs. Respir.
PhySiOl. 1981; 45: 25-41.
29 Davis JM, Veness-Meechan K, Notter RH, Bhutani VK, Kendig JW,
Shapiro DL. Changes in pulmonary mechanics after the adminis-
tration of surfactant to infants with respiratory distress syndrome.
N. Engl. J. Med. 1988; 319: 476-9.
30 Hills BA. A hydrophobic oligolamellar lining to surfaces in various
tissues: A possible ubiquitous barrier. Med. Sci. Res ,1992;
20: 543-50.
31 Rollins M, Jenkins J, Tubman R, Corkey C, Wilson D. Comparison
of clinical responses to natural and synthetic surfactants. J. Perinat.
Med. 1993; 21: 341-7.
32 Whitsett JA, Weaver TE. Structure, function and regulation of
pulmonary surfactant proteins. In: Bourbon JR, ed. Pulmonary
Surfactant. CRC Press, Boca Raton, Florida, 1991; 77-96.
33 Benjamins JA, Morel1 P. Proteins of myelin and their metabolism.
Neurochem. Res. 1978; 3:137-74.
34 Morley CJ, Bangharn AD, Johnson P, Thorburn GD, Jenkins G.
Physical and physiological properties of dry lung surfactant. Nature
1978; 271: 162-3.
35 Stratton CJ. Morphology of surfactant producing cells and of the
alveolar lining. In Robertson B, van Golde LMG, Battenburg JJ,
eds. Pulmonary Surfactant. Elsevier, Amsterdam, 1984; 68-1 18.
36 Morley CJ, Bangharn AD, Miller N, Davis JA. Dry artificial lung
surfactant and its effect on very premture babies. Lancet 1981; 64-8.