Main

Inflammatory mediator systems, such as complement and contact cascades, accompany surfactant inactivation by serum proteins in the pathogenesis of ARDS(13). It has also been demonstrated in ARDS that the degree of complement activation corresponds to the extent of pulmonary shunt and extravascular lung fluid(3). In contrast, the primary cause of IRDS in the immature lung is surfactant deficiency. Probably, mediator systems are secondarily activated, causing a vicious circle of permeable pulmonary capillaries, subsequent protein leakage into the alveolar space, and inactivation of endogenous surfactant(2,4). Previous studies have shown an activation of the contact system in preterm infants with RDS(47), but inconsistent results were reported regarding the complement system activation in these patients(4,810). It has been postulated that the type of response to surfactant treatment in IRDS may be determined by other pathophysiologic processes(11,12). In previous studies on complement and contact systems, the patients were not divided into surfactant responders and nonresponders.

The aim with this study was to evaluate whether the complement cascade and the contact cascade are activated in preterm infants with severe RDS and whether there is a difference regarding activation related to surfactant treatment.

METHODS

The study was approved by the local ethics committee, and written parental consent was obtained. Predefined inclusion criteria for the study group were: mechanical ventilation with initial FiO2 > 0.5, radiologic signs of RDS grade II-IV(13), and therapeutic surfactant treatment. Seventy-three preterm newborns, admitted to our neonatal intensive care unit between September 1995 and March 1997 were enrolled. Among them, the following were excluded: one infant who died within the first 24 h of life, two infants with lethal malformations, seven infants with fetal acidosis (umbilical arteria pH < 7.10), eight infants with prophylactic surfactant administration immediately after birth without blood gas analysis and x-ray examination, seven infants with an initial FiO2 < 0.5, seven infants with infection (perinatal infection was assumed if the following criteria were fulfilled: immature to total white blood neutrophil ratio ≥ 0.20 and C-reactive protein concentration > 10 mg/L during the first 24 h), and 11 infants without parental consent.

Therefore, a total of 30 preterm infants, treated with 100 mg/kg porcine natural surfactant (Curosurf, Serono Pharma GmbH, Germany), were investigated in the study group. After recruitment of a study infant, the next admitted preterm infant with the same gestational age but without clinical or radiologic signs of RDS, acidosis, or infection served as its matched control. Because 12 parents did not give their consent, the control group involved only 18 infants.

According to the decrease in FiO2 within 6 h after surfactant therapy, the study group was divided into two subgroups: surfactant responders with a reduction in FiO2 > 50% of the presurfactant level (n = 21) and poor responders (n = 9)(11). Median age at surfactant administration was 2 h in both groups. In study infants, the peak inspiratory pressure was recorded. In all infants, aAPO2 was calculated by dividing the arterial oxygen tension by the alveolar oxygen tension, using the following equation: FiO2(PATM-PH2O)-PaCO2. PATM is the atmospheric pressure (760 mm Hg), PH2O is the partial pressure of water vapor (47 mm Hg), and PaCO2 is the arterial carbon dioxide tension (Table 1).

Table 1 Comparison [median (quartiles)] of preterm infants with severe RDS (study group) and preterm infants without RDS (control group)

Two blood samples of 0.4 mL containing either disodium-EDTA (EDTA; Kabi-Labortechnik, Nuembrecht, Germany) or 0.07 mL sodium citrate (Fa. Saarstedt, Nuembrecht, Germany) were drawn 23-27 h after birth during a routine blood sampling to avoid an additional venous puncture. Within 20 min, the samples were centrifuged for 5 min at 10000 × g and plasma was separated and stored at -80°C for no longer than 1 mo.

Determination of complement and contact factors. The reliability of test results was monitored using controls with known concentrations or activities. The bioassays were blinded for group assignment and clinical data.

Kinetic determination of whole complement activity applies the lysis of sensitized sheep erythrocytes (Behring Diagnostica AG, Marburg, Germany) by activated complement factors in the plasma sample. The test measures the time needed for total lysis of a fixed number of erythrocytes, indicated photometrically at 578 nm(14). A within-run precision study (n = 20) showed a coefficient of variation of 5% using a plasma with 50% complement function.

Concentrations of C1 inhibitor, C1q, C4, and factor B were determined by single radial immunodiffusion(15) using the Nor-Partigen kit (Behring Diagnostica AG, Marburg, Germany). In within-run precision studies, coefficients of variation (n = 20) were C1q 4%, factor B 3%, C4 7%, and C1 inhibitor 6%.

C3a enzyme immunoassay (Fa. Progen Biotechnik GmbH, Heidelberg, Germany) selectively detects C3a-desArg using MAb(16). The coefficient of variation for this method (n = 20) was 8%. C5a was determined with a specific sandwich enzyme immunoassay (Fa. Behring, Marburg, Germany) and showed a coefficient of variation (n = 20) of 8%(17).

The concentration of activated factor XIIa was measured by a semiquantitative direct immunoassay(18) using specific MAb (WAK-Chemie Medical GmbH, Bad Homburg, Germany). The within-run coefficient of variation (n = 20) was 6%.

The functional activity of C1 inhibitor was determined in citrated plasma using the chromogenic substrate technique described by Heber et al.(19). The coefficient of variation in a within-run precision study was 5% (n = 20).

Statistical analysis. Because most data were not normally distributed, results were expressed as median with quartiles. Differences between the groups were assessed by Mann-Whitney U test or Fisher exact test for qualitative data. Statistical significance was assumed at p < 0.05. All calculations and tests were performed by means of the software package SPSS-PC (Chicago, Illinois).

RESULTS

Complement and contact activation in preterm infants with severe RDS. Clinical data from patient groups and results are shown in Fig. 1 and Table 1. No significant differences were found in birth weight, gestational age, kind of delivery, gender, use of prenatal steroids, and umbilical arterial pH. The mothers of one infant in the control group and six infants in the RDS group showed the syndrome of hemolysis, elevated liver enzymes, and low platelets and/or toxicosis (p = 0.17). In the control group, four mothers smoked, and in the study group six mothers (p = 0.34). A premature rupture of membrane occurred in four mothers of the control group and three mothers of the study group (p = 0.25). One infant of the controls and three study infants received erythrocyte transfusion before blood sampling (p = 0.59). No infants in the control group but two infants in the study group died during the further course (p = 0.52). Corresponding to the inclusion criteria, aAPO2 was different at admission to our intensive care unit and at the age of blood sampling. WBC and platelet counts were higher in controls at the age of blood sampling for complement analysis.

Figure 1
figure 1

Box-whisker plots illustrating median, quartiles, and ranges of complement and contact proteins in infants without (open bars; n = 18) and with (shaded bars; n = 30) RDS. All parameters show significant differences (C1q: p < 0.001; C3a: p = 0.002; C5a: p = 0.02; factor XIIa: p < 0.001) between the two groups (Mann-Whitney U test).

Complement precursor proteins C1q, C4, and factor B were lower in the study group than in the control group. The whole complement function was lower in the study group, whereas activated split products C3a and C5a were higher in this group. Contact system factor XIIa was also higher in the study group. Activity, as well as concentration of C1 inhibitor, which can inhibit both systems, did not vary between the groups (Table 2).

Table 2 Comparison [median (quartiles)] of preterm infants with a good response after surfactant treatment for RDS (FiO2 decrease > 50% within 6 h after surfactant) and preterm infants with a poor response (FiO2 decrease < 50% within 6 h after surfactant treatment)

Complement and contact activation in severe RDS related to surfactant response. The results of comparing the two subgroups are summarized in Fig. 2 and Table 2. Differences in gestational age, birth weight, mode of delivery, gender, use of prenatal steroids, umbilical arterial pH, and WBC and platelet counts at the age of blood sampling; FiO2; and aAPO2 before surfactant administration were not significant. Differences in FiO2 and aAPO2 after treatment, and peak inspiratory pressure are obviously caused by the different definitions of these groups. The aAPO2 was also different at the age of blood sampling.

Figure 2
figure 2

Box-whisker plots illustrating median, quartiles, and ranges of complement and contact proteins in RDS infants with good (open bars; n = 21) and poor (shaded bars; n = 9) response to surfactant treatment. All parameters show significant differences (C1q: p = 0.008; C3a: p = 0.01; C5a: p = 0.009; factor XIIa: p = 0.04) between the two groups (Mann-Whitney U test).

Looking at the RDS subgroups, we found that mothers of five infants in good responders and one infant in poor responders showed the syndrome of hemolysis, elevated liver enzymes, and low platelets and/or toxicosis (p = 0.43). In the group of good responders, four mothers smoked during pregnancy, compared with two mothers in the poor-responder group (p = 0.84). Premature rupture of membrane occurred in two mothers of the good-responder group and one mother of the poor responders (p = 0.89). Two good responders and one poor responder received erythrocyte transfusion before blood sampling (p = 0.89). One infant of each group died in the further course (p = 0.52). Complement precursor proteins of the classical pathway C1q and C4 were lower in the preterm infants with poor surfactant response, whereas no difference was found for factor B of the alternative pathway. In contrast, increased levels of activated split products C3a and C5a were found in preterm infants with poor surfactant response. Factor XIIa of the activated contact system was higher in these infants. Activity of C1 inhibitor was lower in patients with poor surfactant response, but concentration of C1 inhibitor as well as the complement function test did not show differences related to surfactant response.

DISCUSSION

Complement and contact activation in preterm infants with severe RDS. In this study, we have demonstrated a simultaneous activation of the complement and contact systems in preterm infants with severe RDS. The lower concentrations of precursor proteins and reduced functional activity of the terminal complement complex compared with control infants probably result from a consumption of these proteins during the activation of the cascade. The higher levels of split products, combined with the decrease of precursor proteins, reflect an activation of the cascade and not simply a loss into interstitial and alveolar spaces due to increased membrane permeability.

The activation of the contact or kallikrein-kinin system in preterm infants with RDS has been described by several authors(47). These findings are confirmed in our study by direct measurement of the concentration of activated Hageman factor (XII) as a central part of the contact system.

The study and control groups did not differ regarding prenatal and postnatal factors, which can influence the complement and contact activation. Differences in WBC and platelet counts were an expression of the described activation of leukocytes, clotting, fibrinolysis, and platelets(4,5). Because complement and contact activation were evident 22 to 28 h after birth, an interference of therapeutic interventions with the activation of these systems during this time period cannot be excluded. However, in the study group, we found no differences between the infants who were treated with transfusion or pasteurized plasma solutions and those who were not treated. None of the studied newborns received fresh frozen plasma, drugs, or interventions known to influence the immune system(20).

In contrast to our results, Cat et al.(9) were not able to show an activation of the complement system in preterm infants with RDS compared with healthy infants, but inclusion criteria were different: mean gestational age in their study group was 35.5 wk, and < 50% of these patients were mechanically ventilated, probably because of milder RDS. In infants of similar gestational age and weight, Brus et al.(4) reported elevated C3a levels compared with healthy preterm infants. Also Enskog et al.(10) and Schrod et al.(8) reported C3a values in RDS infants with immaturity similar to our study. Complications such as pneumothorax and intracerebral hemorrhages(10) or asphyxia(8) are associated with even higher anaphylatoxin concentrations.

In contrast to the studies mentioned above, we tried to exclude all patients with additional causes of complement activation; e.g. fetal acidosis and infection(8,21,22). Therefore, we assume that the detected complement activation is mainly due to RDS itself. Within the pathomechanism of RDS, tissue hypoperfusion and hypoxemia, local hyperoxia, and inflammatory response in the lung, but also interactions with the contact system and mechanical ventilation, can be discussed as triggers for complement activation(4,23).

The activation of complement and contact cascades in preterm infants with RDS may contribute to lung injury, and especially to the progression of pulmonary edema(24). Possible factors leading to increased endothelial permeability are the release of anaphylatoxins from the complement and of bradykinin from the contact system(1,25). Anaphylatoxins also contribute to leukocyte chemotaxis, aggregation, and local sequestration in the lung by releasing proteases, histamine, peroxidases, superoxides, and cytokines(4). Additionally, contact system activation leads to intra-alveolar fibrin deposition as a major compound of hyaline membranes(4,26).

Complement and contact activation in relation to surfactant response. Some infants do not respond to surfactant therapy(11,12). Segerer et al.(11) defined poor response as < 50% decrease of the initial oxygen supply within 6 h after surfactant instillation, and we used this definition to distinguish treatment groups. The groups did not differ regarding prenatal and postnatal factors that could influence the complement and contact activation. Activation of complement and contact systems was more pronounced in infants with poor surfactant response than in those with good response. This phenomenon is expressed in lower levels of the complement precursor proteins C1q and C4, and particularly in higher levels of anaphylatoxins and factor XIIa. In poor responders, the lower values of C1-inhibitor activity indicate a stronger activation of both systems. The consumption of C1 inhibitor by complement or contact activation primarily results in decreased C1 inhibitor activity and subsequently in decreased concentration(9).

If good responders are understood to be IRDS patients with immature surfactant and poor responders to be ARDS patients with surfactant inactivation(11,12), different pathophysiologic mechanisms could be conceivable, and the complement and contact activation could be responsible for the different surfactant response. Clinical and experimental observations suggest that aggregation of polymorphonuclear granulocytes in response to activated complement might contribute to the genesis of ARDS. These aggregated granulocytes cause pulmonary dysfunction by becoming lodged in the lung as leukoemboli, which leads to poor reaction to surfactant treatment(25). Because we determined the complement and contact parameters several hours after surfactant treatment, other reasons for our findings are possible. A similar activation for both groups may be started in RDS, but an already declining activity level in the responder group could be found 22 h after surfactant application. Therefore, the more pronounced complement and contact activation in the poor-responder group is probably a secondary effect after unsuccessful treatment of RDS. Another possible explanation is barotrauma, resulting from higher peak inspiratory pressures in the group of poor responders (Table 2), because activation of contact and complement systems can be triggered by tissue trauma(35). Whether the stronger activation in poor responders was a primary or secondary effect can be investigated only by blood sampling before surfactant treatment in a further study.

Conclusion. This study gives evidence that complement and contact systems are activated in premature infants with RDS. This activation is more pronounced in infants with a poor response to surfactant treatment. The reason for this occurrence is probably a persisting or increasing complement and contact activation, which leads to further lung injury. For those patients, approaches to inhibit inflammation and coagulation cascades should be discussed, in addition to surfactant treatment.