Main

Preterm newborns with respiratory distress syndrome and term newborns who develop cerebral palsy tend to have elevated blood levels of coagulation factors. In this essay, we suggest that these coagulation factors, when activated, produce their effects not so much by promoting coagulation as by promoting inflammation, a view supported by studies of adults with a systemic inflammatory response syndrome with multi-organ dysfunction (SIRS/MOD). We also suggest that activated systemic coagulation factors in the preterm newborn contribute to cerebral white matter damage (WMD) and to its clinical sequelae, including cerebral palsy (CP).

NEWBORNS

Respiratory distress syndrome (RDS) in preterm newborns.

Fibrin is a major part of hyaline membranes, which are viewed as locally produced clots in neonatal RDS (1, 2). Indeed, the severity of RDS has been linked to the systemic concentration of thrombin-antithrombin III complex (3). In addition, preterm babies with RDS are more likely than their peers to have systemic activation of inflammation (4, 5). We do not yet know if products of coagulation and inflammation co-occur in the blood of infants destined to develop bronchopulmonary dysplasia.

In newborn immature sheep with RDS, inflammation and clotting are activated shortly after birth (6). “Because activation of clotting and platelet activation are found later than activation of inflammation,” Jaarsma and colleagues recently wrote, “we consider activation of clotting is not causative in RDS, but is activated secondary to activation of inflammation”(6).

We postulate that this scenario of inflammation promoting coagulation also applies to neonatal WMD. Moreover, we postulate that inflammation promotes coagulation, which in turn further promotes inflammation (Fig. 1).

Figure 1
figure 1

This figure emphasizes how an inflammatory stimulus promotes thrombin formation, impairs each of the three anti-coagulant pathways, and interferes with fibrinolysis. The resulting increased concentration of thrombin and decreased concentrations of anti-thrombin III, activated protein C, tissue factor pathway inhibitor, and plasminogen activator inhibitor 1 all contribute to enhancing inflammation.

Full size image

CP in term newborns.

Support for the systemic activation of both inflammation and coagulation in newborn brain damage comes from a study of term newborns who later developed CP (7). On postnatal day 2, these term babies had elevated levels of a factor V Leiden mutation product. This indicator of impaired anti-coagulation, however, was accompanied by prominent elevations of proteins with anti-coagulant properties (e.g. protein C, protein S, and antithrombin III). The elevation of these coagulation factors was accompanied by elevated blood levels of many pro-inflammatory cytokines and chemokines, supporting the view that coagulation activation and inflammation coexist, and may even influence one another.

Cerebral WMD in preterm newborns.

Neurodevelopmental deficiencies are more common in preterm infants than in infants born at term (8). Many of these can be predicted by the echolucent ultrasound images indicative of focal WMD (9). However, a considerable proportion of the brain dysfunction in these infants appears to be a consequence of a diffuse intracranial process (10). Since the available evidence does not support claims that echolucencies are ischemic lesions (11), our discussion of activated coagulation factors and brain damage in no way implies that the damage is a consequence of intravascular coagulation leading to ischemia.

Chorioamnionitis predicts ultrasound-defined WMD and CP (12). The more this inflammatory response involves the fetus, the greater the risk of brain damage (13). Infection remote from the brain appears to increase the systemic availability of inflammatory products, which gain access to the fetal/newborn brain where they do their damage (14, 15). Indeed, newborns with elevated circulating levels of inflammatory cytokines are at increased risk of WMD and CP (16). The strongest support for a fetal contribution to this systemic inflammatory response is the presence of inflammatory products in umbilical cord blood (17).

The coexistence of inflammatory and thrombotic lesions in the preterm placenta is associated with heightened risk of neurologic impairment (18). These findings, along with the co-occurrence of markers of inflammation and coagulation in term newborns who develop CP (7), raise the possibility that increased circulating levels of activated coagulation factors enhance the influence of inflammation factors. In one highly selected sample, however, preterm newborns who developed CP had mean and median values of anti-thrombin that were similar to those of controls (19).

Preterm newborns with the Val34Leu polymorphism of the factor XIII gene appear to be at reduced risk of WMD (20). Adults with this polymorphism produce a less stable fibrin clot than people without (21), are at increased risk of intracerebral hemorrhage (22) and are at reduced risk of brain and myocardial infarction and venous thrombosis (23–25). These observations suggest that the reduced risk of WMD in preterm newborns is due to a reduced risk of fibrin clot obstruction of cerebral blood flow.

However, what the authors of this report (20) attribute to ischemia, might just as readily be due to inflammation. Fibrin has inflammatory properties (26). Circumstantial evidence for this comes from diverse sources. Depression of plasminogen activator-mediated fibrinolysis not only increases fibrin deposition in alveoli, but also contributes to inflammatory cell recruitment and migration (27). Fibrinolytic strategies to block pulmonary and pleural fibrin deposition and inflammation appear promising (28).

An imbalance between fibrin deposition and fibrin dissolution contributes to inflammation-induced peritoneal adhesions (29). In addition, mice deficient in tPA are more susceptible to inflammation-induced adhesion formation than wild-type mice (30).

Additional evidence for the inflammatory properties of fibrin comes from studies of bacteria and humans. The conversion of fibrinogen to fibrin at the surface of bacteria generates proinflammatory fibrinopeptides (31). People who have genetic polymorphisms that lead to excess expression of PAI-1 also have increased levels of TNF-alpha, and IL-1 (32). Some view the relationship between fibrin and inflammation so convincing that they advocate plasminogen activator inhibitors as therapy for sepsis (33).

Individuals with the factor XIII Val34Leu polymorphism have less fibrin than their peers (21) and should, therefore, be less able to mount a vigorous inflammatory response than individuals without this polymorphism. Since some WMD or CP appears to follow a vigorous fetal or neonatal inflammatory response (13), newborns with a factor XIII Val34Leu polymorphism should be at reduced risk of WMD and CP. This reasoning does not diminish the contribution of coagulation products to the occurrence of WMD. It just changes the focus from ischemia to inflammation.

Adult studies and animal models.

More research appears to be devoted to the study of systemic inflammation in the adult than in the newborn. Perhaps we can learn from the progress being made in studies of sepsis and SIRS/MOD (34).

Inflammation promotes coagulation.

Sepsis and SIRS/MOD are accompanied by thrombin generation, impaired anticoagulation, and impaired fibrinolysis (35). Each can be initiated by endotoxemia (36) and pro-inflammatory cytokines (37).

Coagulation promotes inflammation.

Coagulation products with pro-inflammatory effects include factor Xa (38), tissue factor (39), fibrinogen/fibrin (28), plasmin (40), and thrombin (41). Thrombin signaling is probably achieved by activating the PAR-1-type thrombin receptor (42). This in turn stimulates endothelial cell activation, resulting in the availability of adhesion molecules (43), which facilitate the transendothelial migration of leukocytes into the surrounding parenchyma (44). Endothelial cell activation also promotes the synthesis of pro-inflammatory cytokines (45).

Disseminated intravascular coagulation.

“Many investigators currently believe that it is not DIC, and particularly not fibrin formation itself that is harmful, but rather it is the generation of serine proteases and their potential interactions with pro-inflammatory mediators that contributes to organ failure and death”(46). Thus, markers of coagulation dysfunction in the blood (and perhaps the lung and brain) might provide more information about inflammation than they do about the obstruction of microvessels.

Inflammation and coagulation amplify each other.

Whether coagulation is a cascade or a sequence of overlapping stages (47), inflammation and coagulation amplify each other (43). A “vicious cycle” does not automatically follow because some proteins with anti-coagulant and anti-inflammatory properties are released in response to products of coagulation (43).

Activated protein C.

Impairments of the protein C system are a hallmark of adult sepsis (48). Activated protein C has anti-thrombotic, pro-fibrinolytic, and anti-inflammatory properties (49).

Activated protein C modulates coagulation by decreasing synthesis and expression of tissue factor (50). By forming a complex with protein S, activated protein C inactivates factors Va and VIIIa, thereby limiting production of thrombin (51), and limiting activation of factor X (52).

Activated protein C increases fibrinolysis by decreasing plasminogen activator inhibitor 1 and thereby preventing inhibition of tPA (53). Less directly, activated protein C promotes fibrinolysis by inhibiting thrombin formation (54), which in turn, limits the activation of thrombin activatable fibrinolysis inhibitor (55).

Reduced inflammation is achieved by limiting the production of thrombin (51, 54), inhibiting neutrophil binding to selectins (56), and limiting the production of monocyte chemoattractant protein-1 (57). Activated protein C also inhibits TNF-alpha production by monocytes and endothelial cells (58, 59), apparently by interfering with nuclear factor-κB nuclear translocation (60, 61) and the binding of STAT6 oligonucleotides to nuclear proteins (62). In addition, during translocation from the plasma membrane, the endothelial cell protein C receptor can carry activated protein C to the nucleus, where activated protein C is presumed to modulate inflammatory mediator responses in the endothelium (35).

Activated protein C in models of SIRS.

In a rat model of lipopolysaccharide-induced lung microvascular injury, a potent inhibitor of thrombin generation blocks neither lung injury nor the increase in plasma concentration of tumor necrosis factor-alpha (59). On the other hand, both are blocked by activated protein C.

In a baboon model of DIC induced with live Escherichia coli, an inhibitor of thrombin generation prevents DIC, but not shock and multiple organ failure (63). In contrast, activated protein C prevented DIC, shock and multiple organ failure (63).

Clinical trials of activated protein C to reduce mortality in SIRS/MOD.

Most attempts to reduce the very high mortality rate in people with SIRS/MOD have been disappointing. Some therapies were intended to reduce the inflammatory response, others to reduce DIC (e.g., by inhibiting thrombin generation) (64). DIC was reduced in some studies, but mortality was not. On the other hand, activated protein C, which suppresses thrombin generation and has indirect and direct anti-inflammatory properties, has reduced mortality in SIRS/MOD (48).

In the largest clinical trial that documented the effectiveness of activated protein C to prevent death in SIRS/MOD, drug recipients had lower plasma levels of IL-6 and thrombin-related biomarkers than placebo recipients (48). Apparently, both anti-inflammatory and anti-coagulation properties contribute to the effectiveness of activated protein C.

Inferences for newborns.

What we discussed above might be relevant to neonatal WMD. Inferences can be drawn in three areas: coagulation activation, endothelial activation, and the therapeutic benefits of endogenous activated protein C.

Both endotoxin and pro-inflammatory cytokines promote thrombin generation, impair anticoagulation, and diminish fibrinolysis (36, 37). Both endotoxin and pro-inflammatory cytokines have also been implicated in the pathogenesis of WMD in newborn animals (65) and humans (16). Perhaps thrombin generation, impaired anticoagulation, and diminished fibrinolysis contribute to WMD.

Funisitis (umbilical cord vessel inflammation) can be viewed as a histologic expression of endothelial activation. Infants with funisitis appear to be at increased risk of WMD, CP, and other neurodevelopmental dysfunctions (18, 66–68). In addition, products of inflammation synthesized by endothelial cells have been implicated in WMD pathogenesis (16). Since thrombin's activation of endothelial cells results in multiple inflammatory phenomena (35), thrombin activation of umbilical cord endothelial cells might contribute to the occurrence of WMD in preterm newborns.

Proinflammatory cytokines activate synthesis of inducible nitric oxide synthase, which is required for the production of nitric oxide and its breakdown products, free nitrogen and oxygen radicals (69, 70). Some of the brain damage initiated by inflammatory cytokines requires the presence of inducible nitric oxide synthase iNOS (71). Consequently, evidence that newborn white matter damage is due to fetal inflammatory phenomena (72) is entirely compatible with evidence that free oxygen radicals contribute to the demise of oligodendrocyte precursors in vitro(73), brain lesions in rodents (34), and lipid peroxidation in premyelinating oligodendrocytes in humans (74). Coagulation proteins can contribute to brain damage via the production of free radicals secondary to inflammatory phenomena, as well as by more direct means (75).

The success of activated protein C therapy in adults with SIRS/MOD suggests that therapies intended to reduce the organ damage accompanying fetal and neonatal inflammatory responses should have both anti-inflammatory and anti-coagulation properties. Activated protein C also has neuroprotective effects independent of its systemic anti-coagulant and anti-inflammatory functions (76). Nevertheless, in light of bleeding and other complications (77), we urge caution in considering activated protein C for preterm newborns with evidence of a systemic inflammatory response. A prudent course would be to delay consideration of clinical trials of activated protein C until laboratory models show that activated protein C is effective in reducing the risk of white matter damage. Although not needed to demonstrate effectiveness, additional epidemiologic studies would support the decision to test activated protein C in humans if they found that thrombin-related biomarkers are elevated in the blood of newborns who develop white matter damage and/or cerebral palsy.

Shortly after birth at term, protein C levels in the blood are almost 40% of those in adults (78). They are even lower in those born before term (79). In light of our suggestion that the fetal and neonatal inflammatory responses activate the clotting system, such low levels of protein C might place the newborn at especially high risk of the inflammatory consequences of clotting system activation. Because protein C levels rise following vitamin K administration (80), the recommendation that all babies receive vitamin K prophylaxis to prevent hemorrhagic disease of the newborn (81) might have the added benefit of reducing the heightened inflammatory response seen in preterm newborns.

CONCLUSION

Our hypothesis that activated coagulation products contribute to the occurrence of neonatal lung disease and WMD should be viewed as tenuous. We doubt that organ damage in preterm newborns who have elevated blood levels of coagulation factors can be attributed solely to vessel occlusion. Rather, we infer from the current literature that in preterm newborns with a systemic inflammatory response, activation of the coagulation system contributes to organ damage by promoting inflammation.