Exudative phase

To facilitate gas exchange the alveolar wall is highly specialised in structure. Unfortunately this specialisation renders it susceptible to injury by a wide variety of agents. The principal cells of the air–blood barrier, the type I alveolar epithelial cell and the capillary endothelial cell are exceptionally thin (see Fig. 1.27, p. 16) and this makes them particularly vulnerable to non-specific damage. Injury to these two cells underlies the development of DAD. At an early stage of alveolar injury the type I epithelial cells show cytoplasmic blebbing, which is soon followed by necrosis resulting in denudation of the basement membrane (see Fig. 7.2.5, p. 376).^16,17 Similar blebbing is seen in the alveolar capillary endothelium but denudation of the endothelial basement membrane is seldom observed, probably because of differences in the ways epithelial and endothelial cells regenerate (see below). The consequences of this damage include the escape of fibrin-rich exudates into the interstitial and air spaces, loss of the surface-active alveolar lining film and pulmonary collapse.

The exudative phase lasts about 1 week, during which the lungs are heavy, often weighing over 1 kg each, dark and airless. The changes are often patchy, with the dorsal and basal regions being most severely affected (Fig. 4.3).^18–20 Slicing shows that they are wet, the cut surface exuding blood or heavily blood-stained watery fluid. Microscopically there is widespread collapse, intense congestion of the capillaries, interstitial oedema and distension of the lymphatics, a pattern that is sometimes known as congestive atelectasis (Fig. 4.4). Alternatively, there may be haemorrhagic oedema (Fig. 4.5). At the air–tissue interface, which in these collapsed lungs is at the respiratory bronchiole or alveolar duct level, respiratory movements compact a fibrin-rich exudate mixed with necrotic epithelial debris into a thin layer that covers an otherwise denuded epithelial basement membrane (Fig. 4.6), leading to the formation of hyaline membranes (Fig. 4.7)^{11–13,21,22}: these are identical to those that paediatric pathologists recognise as the hallmark of the infantile respiratory distress syndrome (compare Fig. 4.3 with Fig. 2.8, p. 44). Similar changes are also seen in acute eosinophilic pneumonia but here the hyaline membranes contain eosinophils, possibly in small focal collections that are easily overlooked (see p. 462).

Figure 4.3 Shock lung. The lower lobe shows large irregular areas of collapse and congestion.

(Reproduced from Corrin (1980)¹⁸ by permission of the editor of the Journal of Clinical Pathology.)

Figure 4.4 Congestive atelectasis in septic shock. There is severe capillary congestion and alveolar collapse.

Figure 4.5 Haemorrhagic pulmonary oedema in septic shock. The alveolar capillaries are congested and the alveolar spaces are filled by oedema fluid in which there are free erythrocytes.

Figure 4.6 Diffuse alveolar damage. The alveolar epithelium terminates just above centre and from this point (arrow) upwards a mixture of electron-dense fibrin and cell debris (seen at the light microscopic level as a hyaline membrane) is closely applied to the denuded basement membrane. Part of a necrotic cell is seen within the alveolus. Transmission electron micrograph.

(Courtesy of Miss A Dewar, Brompton, UK.)

Figure 4.7 Diffuse alveolar damage, exudative phase. Hyaline membranes are a conspicuous feature. They often outline alveolar ducts. From a pregnant woman who suffered respiratory arrest due to angioneurotic oedema. She needed to be ventilated for 2 weeks before death and the oxygen concentration in the inspired air constantly increased to prevent hypoxaemia, reaching 70% for the last 8 days of her life.

(Courtesy of Dr D Melcher, Brighton, UK.)

The congested alveolar capillaries sometimes contain increased numbers of platelets or neutrophil leukocytes. This selective sequestration of formed blood elements in the microvasculature of the lungs is particularly noticeable in shock, and is considered in more detail under that heading below.

Regenerative phase

As with any exudative process, healing may be by resolution, which involves fibrinolysis and permits the lungs to return to normal, or repair, which involves fibrosis and leaves the lungs permanently scarred. Resolution and repair are both accompanied by epithelial and endothelial regeneration.

The regenerative (or proliferative) phase becomes prominent 1–2 weeks after the initial injury. It involves proliferation of both epithelial and connective tissue cells. The stem cell concerned in epithelial regeneration is the type II alveolar epithelial cell.^23,24 These cells first proliferate and then differentiate into type I cells, thereby re-epithelialising the denuded basement membranes. The dividing type II cells form a simple cuboidal epithelium (Fig. 4.8), or the alveoli may be lined by plump pleomorphic spindle cells that represent cell forms intermediate between types II and I epithelial cells (Figs 4.9 and 4.10). Eosinophilic cytoplasmic inclusions indicative of cell damage (see Fig. 7.1.25, p. 348) are often present,^25,26 and sometimes there is squamous metaplasia instead of orderly differentiation into type I cells, a change that the unwary pathologist may mistake for neoplasia.²⁷

Figure 4.8 Diffuse alveolar damage, regenerative phase. (A) The alveoli have a simple cuboidal epithelial lining.

(Courtesy of Dr D Melcher, Brighton, UK.)

(B) Electron microscopy shows that such cells represent hyperplastic type II pneumocytes, which form a continuous row rather than being scattered singly in the corners of the alveoli. They are readily recognisable by their surface microvilli and the osmiophilic lamellar secretory vacuoles of alveolar surfactant.

(Courtesy of Mrs D Bowes, Midhurst, UK.)

Figure 4.9 Diffuse alveolar damage, regenerative phase. Regenerating alveolar epithelial cells are atypical, consisting of plump, elongated forms that represent type II pneumocytes differentiating into type I.

Figure 4.10 Diffuse alveolar damage, regenerative phase. An alveolar epithelial cell having the squamous form of a type I cell but the microvilli and lamellar vacuoles of a type II cell. Such intermediate cells are indicative of epithelial regeneration.

The regenerating epithelium usually grows beneath the exudates lining the denuded basement membrane, casting the hyaline membranes off into the air space (Fig. 4.11), but it may grow over them so that they are incorporated into the interstitium (Figs 4.12 and 4.13),^11–13 where their subsequent organisation contributes to the fibrosis.²² The regenerating epithelial cells may also bridge the mouths of collapsed alveoli so that these air spaces never re-expand and there is permanent shrinkage of the lung, a process termed atelectatic induration (Figs 4.14 and 4.15).^28–30

Figure 4.11 Diffuse alveolar damage, regenerative phase. The regenerating alveolar epithelium grows underneath the hyaline membranes, casting them off into the air space.

Figure 4.12 Diffuse alveolar damage, regenerative phase. The regenerating alveolar epithelium may also grow over the hyaline membranes to incorporate them into the alveolar wall. In this electron micrograph the hyaline membranes are largely represented by electron-dense fibrin which is closely applied to previously denuded alveolar epithelial basement membrane and are in turn covered by regenerating epithelium laying down a new basement membrane.

(Courtesy of Miss A Dewar, Brompton, UK.)

Figure 4.13 Diffuse alveolar damage, repair phase. Hyaline membranes, which are stained red, are in the process of being incorporated into the alveolar interstitium. Martius scarlet blue stain.

Figure 4.14 Atelectatic induration. Denuded alveolar epithelial basement membranes are closely apposed (centre). The regenerating alveolar epithelium (top) bridges the mouth of the alveolus, which consequently has little chance of ever re-expanding.

(Courtesy of Miss A Dewar, Brompton, UK.)

Figure 4.15 Atelectatic induration. Silver staining of the basement membranes shows that what appear to be thickened single alveolar walls represent the closely apposed walls of several collapsed alveoli.

In contrast to the type I epithelial cells, which have no regenerative powers and are replaced by differentiation of proliferating type II cells, endothelial cells are replaced by lateral spread of their own kind. An effete endothelial cell is first undermined by its healthy neighbours and only cast off when these have completely covered the basement membrane.³¹ Therefore, although segments of bare basement membrane have been described on the vascular side of the air–blood barrier,^32,33 they are not seen to the same extent as on the epithelial side. Nevertheless, thrombosis may complicate such endothelial damage^21,32 and subsequent organisation of such thrombi is probably responsible for some of the vascular remodelling that is seen in the repair phase of DAD. This remodelling consists of fibrocellular intimal thickening that narrows the lumen of small vessels throughout the lung and can be visualised as decreased background filling on postmortem arteriograms.³⁴

Repair phase

If healing is by repair, interstitial connective tissue cells proliferate and, as in any scarring, myofibroblasts are involved at an early stage.³⁵ Whilst myofibroblast contracture is beneficial when it promotes early closure of an open wound, in the lungs it largely results in harmful distortion of the bronchioloalveolar architecture and shrinkage of the lungs. Possible sources of the proliferating interstitial connective tissues include resting interstitial connective tissue cells, the bone marrow and the alveolar epithelium. Epithelial–mesenchymal transition is being increasingly recognised throughout the body and in the lung in conditions such as idiopathic pulmonary fibrosis, asthma and obliterative bronchiolitis.^36–38

Whatever the source of the proliferating connective tissue cells, those with a fibroblastic phenotype lay down collagen, leading to the development of interstitial fibrosis.^39,40 Interactions between fibroblasts and the alveolar epithelium through gaps in the basement membrane have been described,^41–45 suggesting that the regenerating epithelial cells play a further role in the underlying process of fibrosis. It is known that these cells synthesise fibrogenic cytokines such as tumour necrosis factor-α; the secretion of tumour necrosis factor-α into the underlying connective tissue would further promote interstitial fibrosis.^46,47

Fibroblasts also migrate into the alveolar exudates through defects in the epithelial basement membrane to lay down collagen within the hyaline membranes.^48,49 As epithelial cells grow over the newly formed connective tissue, a new basement membrane is formed, thereby incorporating the collagen into the interstitium,⁴⁹ a process known as fibrosis by accretion. Involvement of the alveolar ducts results in these structures being lined by a ring of granulation tissue (Fig. 4.16). Less frequently, the organised exudates retain a predominantly intra-alveolar position, resulting in loose buds of granulation tissue similar to those seen in organising pneumonia due to other causes (see Box 6.2.1, p. 309). However, here they are more widespread and there is more prominent type II cell hyperplasia, interstitial fibroblastic proliferation and interstitial inflammation (Fig. 4.17). Nevertheless, such an organising pneumonia pattern is worth recognising as it carries a better prognosis than interstitial fibrosis.⁵⁰ It is claimed that this intra-alveolar pattern of repair is particularly found when generalised sepsis is the cause of the initial damage whereas interstitial fibrosis is more characteristic of injury caused by cytotoxic drugs and the idiopathic cases.⁵¹

Figure 4.16 ‘Ring’ fibrosis, representing the organisation and incorporation of exudates lining the alveolar ducts into the walls of these airways.

Figure 4.17 Diffuse alveolar damage, repair phase. (A) A lower lobe seen at autopsy shows diffuse consolidation, with occasional lobules showing more exudative appearances. (B) Microscopy shows diffuse intra-alveolar organisation with abundant interstitial inflammation.

An increase in lung collagen can be detected in patients with ARDS who survive longer than 14 days and this progressively increases with the duration of the disease.⁵² The identification of pulmonary fibrosis on transbronchial biopsy is closely related to mortality.⁵³ Fibrosis can be well established by 2 weeks, at which time the lungs may be contracted and firm with a fine sponge-like pattern on their cut surfaces, representing bronchiolectasis and irregular microcystic distortion of the alveolar architecture (Figs 4.18 and 4.19).^54–57 The changes are similar to those seen in any fibrotic process but are reached remarkably quickly (see Fig. 4.19).²² However, at this early stage the fibrosis differs from that seen in chronic fibrosis in being more cellular and less collagenous: extensive fibroblast proliferation is evident. With small air cysts alternating with solid areas of fibrosis and foci of squamous metaplasia there is a resemblance to the bronchopulmonary dysplasia seen in the late stages of the infantile respiratory distress syndrome.^55,56 In survivors the granulation tissue undergoes progressive collagenisation and such patients may suffer from debilitating fibrotic lung disease. The pattern is then that of fibrotic non-specific interstitial pneumonia,⁵⁸ making it important to determine whether there has been an episode of acute lung injury in such patients.

Figure 4.18 Diffuse alveolar damage, repair phase. The repair process has resulted in interstitial fibrosis.

Figure 4.19 Suicidal ingestion of kerosene resulting in advanced pulmonary fibrosis and honeycombing within 2 weeks of the initial injury.

Shock

Shock is a state of prolonged hypotension, generally attributable to trauma, hypovolaemia, cardiac failure, sepsis or anaphylaxis.^63,64 The hypotension leads to inadequate tissue perfusion and, if this is not corrected, multiorgan failure is inevitable. At necropsy, the lungs are the organs most commonly affected.^65–67

Severe pulmonary injury was well described in patients suffering from shock in the Second World War⁶⁸ but it was not until the war in Vietnam that the importance of respiratory failure as a complication of shock was fully appreciated.^69,70 By this time there had been major improvements in medical care. Casualties could be transported rapidly by helicopter to well-equipped field hospitals where intensive care with mechanical ventilatory support was available. Despite this, injured patients often developed fatal respiratory insufficiency, typically after an interval of between 48 and 72 hours. A number of terms graphically described this syndrome – ‘shock lung’, ‘posttraumatic respiratory insufficiency’, ‘traumatic wet lung’ and ‘Da Nang lung’. Pathological examination showed congestive atelectasis or haemorrhagic oedema proceeding to fully developed DAD, as described above.

The pathogenesis of ‘shock lung’ is complex and requires special consideration. In hypovolaemic and cardiogenic shock, compensatory mechanisms such as peripheral vasoconstriction initially maintain cerebral oxygenation, but if the underlying cause is untreated, there follows a state of decompensation characterised by vascular unresponsiveness: vasodilatation develops, the blood pressure plummets and there is widespread hypoxic cell death. Anaphylactic and septic shock are characterised from the outset by such vasodilatation, which is caused by a variety of mediators that are released from inflammatory and other cells.⁷¹ The identification of the same mediators in experimental hypovolaemia⁷² suggests that the pathogenesis of shock may be similar regardless of the cause.

In numerical terms, septic shock is the most important cause of diffuse damage.^59,60 It represents the culmination of a clinicopathological continuum of infection-driven sepsis syndromes:

These terms have replaced ‘septicaemia’, which proved difficult to define and has now been abandoned. The multiorgan dysfunction is due to inflammatory cytokines released into the circulation from the site of infection, the common sites of which are, in decreasing frequency, lung, blood, intestine and peritoneum, urinary tract and surgical wounds. Gram-positive infections are slightly commoner than Gram-negative, which are slightly commoner than polymicrobial. The role of lipopolysaccharide derived from the cell walls of Gram-negative bacteria has been particularly well studied in the pathogenesis of septic shock.^73–77 The many effects of this endotoxin include the release of tumour necrosis factor from monocytes, macrophages and polymorphonuclear leukocytes,^{74–76,78,79} the production of cytokines such as interleukin-1 and interferon-γ that act synergistically with tumour necrosis factor,⁸⁰ the widespread induction of nitric oxide synthesis^81–83 and the activation of both the coagulation and complement cascades.⁸⁴ Exotoxins released by Gram-positive bacteria appear to act similarly. Some of the most fulminant forms of shock are seen with group A streptococci causing puerperal sepsis and necrotising fasciitis, meningococci, Staphylococcus aureus related to tampon retention and a meticillin-resistant strain that secretes an exotoxin known as Panton–Valentine leukocidin.⁸⁵

Tumour necrosis factor has been demonstrated on the luminal surface of pulmonary endothelium in endotoxin-induced shock.⁸⁶ It causes vascular smooth muscle to relax but this action is reduced if the endothelium is removed,⁸⁷ indicating that tumour necrosis factor-induced vasodilatation is partially dependent upon the integrity of the endothelium. A factor that causes vascular dilatation has been detected as coming from the vascular endothelium and acting on the medial muscle coat of the vessel. At first termed endothelium-derived relaxing factor, this factor is now known to be nitric oxide, a remarkably simple chemical that has long been recognised to be poisonous.⁸⁸ Fortunately, its half-life in the vessel wall is very short, timed in seconds rather than minutes. The enzyme responsible for its production (from L-arginine) is nitric oxide synthase, which is found in endothelium and can be induced in the vascular medial smooth muscle. Both the constitutive and inducible forms of the enzyme are activated by bacterial lipopolysaccharide and it therefore seems likely that in septic shock bacterial products act directly on the vessel wall resulting in the production of excess amounts of nitric oxide. Even momentarily increased levels of nitric oxide might be expected to cause arterial dilatation and hence capillary congestion. It would appear that in septic shock circulating bacterial products such as lipopolysaccharide cause vascular dilatation and possibly increased permeability by direct action on the blood vessels and indirectly through the induction of nitric oxide synthase and the release of the cytokines mentioned in the preceding paragraph. Pulmonary endothelial upregulation is indicated by widespread expression of intercellular adhesion molecule-1 (ICAM-1: CD54) in septic shock. Indeed, positive immunostaining for CD54 is proving to be a useful marker of shock.⁸⁹

Nitric oxide is a powerful vasodilator but it also mediates many other processes throughout the body. In host defence nitric oxide plays a very different and more aggressive role. It enables macrophages to generate free oxygen radicals, the principal means by which these cells eliminate both bacteria and cancer cells, but which, if not inactivated, also damage healthy host cells.⁹⁰ Macrophages release much more nitric oxide than endothelial cells but, as in the blood vessels, the amounts released are generally inactivated within seconds. However, overwhelming bacterial infections result in the release of very large amounts of nitric oxide and the overproduction of toxic oxygen radicals. Although the oxygen radicals are countered by the protective action of enzymes such as superoxide dismutase,⁹¹ their excessive release results in oxidation of lipids and protein sulphydryl groups, and DNA damage.⁹² Damaged cell membrane phospholipids release free arachidonic acid which in turn is degraded to produce leukotrienes, such as prostaglandins and thromboxane, that are capable of altering vessel calibre and permeability.

The direct vasodilatory action of nitric oxide and the toxic action of the free oxygen radicals that nitric oxide generates account for most of the pathological features of shock but there are other processes in the microvasculature of the lung that contribute to the pulmonary damage. The vascular engorgement characteristic of shock lung is occasionally accompanied by sequestration of neutrophil polymorphonuclear leukocytes in the pulmonary microvasculature (Figs 4.20 and 4.21).^18,18,93,94 Acting through the complement cascade,^84,95,96 endotoxin activates these cells within the systemic circulation so that they lose their normal deformability and aggregate into microemboli, with the result that they cannot traverse the alveolar capillaries.^73,97,98 Their arrest there is promoted by activated endothelial intercellular adhesion molecules.^99–103 The unique position of the pulmonary capillaries in the circulation is probably responsible for the lungs being the organs most severely affected in shock.^65–67 Trapped in the alveolar capillaries, activated neutrophils damage the alveolar wall by producing reactive oxygen radicals and releasing enzymes such as elastase, collagenase and cathepsins that are able to degrade protein constituents of the wall.^104–110 Impairment of neutrophil migration into the air spaces has been referred to above.⁶²

Figure 4.20 Shock lung. In shock, the alveolar walls are sometimes hypercellular due to the accumulation of neutrophil polymorphonuclear leukocytes.