Eur J Pediatr (2006) 165: 585–589 DOI 10.1007/s00431-006-0139-7

ORIGINA L PA PER

Ryszard Lauterbach . Dorota Pawlik . Renata Radziszewska . Joanna Woźniak . Krzysztof Rytlewski

Plasma antithrombin III and protein C levels in early recognition of late-onset sepsis in newborns Received: 17 October 2005 / Revised: 28 February 2006 / Accepted: 9 March 2006 / Published online: 25 April 2006 # Springer-Verlag 2006

Abstract Objective: The aim of this prospective, observational study was to establish the prognostic values of plasma antithrombin III (ATIII) and protein C (PC) functional levels, determined in 150 neonates with suspected late-onset sepsis. Observations: Both the ATIII and PC concentrations were significantly lower in neonates with sepsis, either confirmed or not confirmed by blood culture, than data obtained in infants with excluded disease. The lowest values of plasma ATIII and PC were observed in neonates who had died in the course of sepsis. The differences between survivors and non-survivors were statistically significant (respectively, for ATIII P=0.003 and for PC P=0.00002). A highly statistically significant correlation (P=0.0016) between plasma PC functional level and risk of death was found in patients with sepsis. However, plasma ATIII concentration, evaluated in the same group of patients, did not correlate with the occurrence of death. Conclusion: We concluded that the measurement of plasma ATIII and, especially, PC levels may facilitate the recognition of sepsis and, with respect to PC, may have a prognostic value. Plasma PC functional concentration below 10% might be the indication for supplementation of PC concentrate. Keywords Antithrombin III . Protein C . Neonate . Sepsis

Introduction In sepsis, activation of the extrinsic pathway combined with depression of the inhibitory mechanisms of coagulation and fibrinolytic system result in a procoagulant state R. Lauterbach (*) . D. Pawlik . R. Radziszewska . J. Woźniak . K. Rytlewski Department of Neonatology, Jagiellonian University Medical College, Kopernika 23, 31-501 Cracow, Poland e-mail: [email protected] Tel.: +48-12-4248587 Fax: +48-12-4248587

that may lead to microvascular thrombosis and multi-organ dysfunction syndrome [1]. Antithrombin III (ATIII) and protein C (PC) play a major role in the regulation of coagulation, shifting thrombin from procoagulant to anticoagulant [10]. ATIII and PC plasma levels decrease in sepsis and, when low, predict high mortality in adults [5, 8]. At birth, PC and ATIII are present at approximately 20– 60% of adult levels [2]. It might cause newborns to be potentially susceptible to sepsis-induced disseminated intravascular coagulation [7]. Little is known about the prognostic value of plasma ATIII and PC levels determined in neonates with suspected sepsis. It prompted this prospective, observational study to assess the usefulness of measurement of plasma ATIII and PC functional concentrations in recognition of late-onset sepsis. The study was also designed to determine the possible relationship between the types of bacteria and the values of plasma ATIII and PC levels.

Materials and methods One hundred and fifty patients with suspected “late-onset sepsis” (diagnosed after their first week of life), who were admitted to the Neonatal Intensive Care Unit of the Neonatal Department of the Medical College of Jagiellonian University in Cracow, were enrolled in the study. Infants with major congenital malformations, intracranial ventricular haemorrhage or symptoms of a congenital infection were not included. The entry criteria of sepsis were physical signs of infection or rapid deterioration of respiratory and cardiovascular function. Physical symptoms of infection were defined by the presence of at least three of the following: feeding intolerance; abdominal distension; temperature instability; disordered peripheral circulation (defined as paleness or peripheral cyanosis and mottled skin with a delayed capillary refill >3 s); lethargy or irritability; hepatosplenomegaly. Respiratory dysfunction was established by the presence of tachypnea (>70 breaths/min) or episodes of apnea and a >15% increase in the fraction of inspired oxygen (FiO2). Tachy-

586

cardia (heart rate >190 beats/min) or bradycardia (heart rate <90 beats/min) and disordered peripheral circulation were regarded as symptoms of circulatory dysfunction. Immediately after the infants’ initial recruitment into study, the following laboratory measurements were determined, to confirm the diagnosis: leukocyte count, leukocyte differentiation, plasma C-reactive protein concentration and plasma procalcitonin level. The presence of at least three of the following: leukopenia <5,000/mm3, leukocytosis >20,000/mm3, C-reactive protein >50 mg/l and procalcitonin >10 ng/ml were considered indicative of an inflammatory phenomenon and induced the start of antibiotic therapy. Simultaneously with the laboratory diagnostic tests, a sample for blood culture was obtained before the first-line antibiotics were administered. A positive blood culture result was required for confirmation of sepsis. Positive results from two blood culture specimens from separate sites were required for confirmation of septicaemia due to Staphylococcus epidermidis, and only significant growth of bacteria was considered for diagnosis of sepsis. Blood specimens for culture were never obtained from the catheters. At the same time as the above-mentioned laboratory diagnostic tests were carried out, plasma ATIII and PC functional concentrations were evaluated as a single measurement, performed after the enrolment of the infant into the study. ATIII and PC were determined by colorimetric assay (Bio-Ksel System, Poland). Venous blood samples (1.0 ml) were drawn onto 0.1 ml of 0.1 M trisodium citrate anticoagulant and centrifuged at 2,000 g for 15 min. Quantitative measurements of plasma ATIII and PC functional levels were taken immediately, with a synthetic chromogenic substrate Bio-Ksel ATIII or PC kit and Chrom-7 (Bio-Ksel System) analyser. The routine medical management of sepsis in all infants was comparable. Sultamicillin (Unasyn Polfa) and tobramycin (Brulamycin, Biogal) were used as first-line antibiotics. After the results of blood culture tests were obtained, antibiotics were adjusted according to the sensitivity of the isolated bacteria. In case the clinical situation deteriorated during treatment with these antibiotics, therapy was changed to meropenem, and immunoglobulin was given. Essential parts of the management of septic shock were as follows: obligatory mechanical ventilation (patient sedated with sufentanil and midazolam), improvement of blood perfusion by pentoxifylline and volume resuscitation (crystalloid or colloid solutions) and/ or addition of vasopressors (protocols for dopamine and dobutamine or norepinephrine administration). Also, steroids were given in case of hypotension resistant to vasopressors. When disseminated intravascular coagulation developed, infants were additionally treated with infusion of fresh frozen plasma and/or concentrate of fibrinogen, administration of vitamin K, packed red blood cell transfusions or platelet transfusions. Neither ATIII nor PC was given in all patients. To check the hypothesis referring to distribution of ATIII and PC data, a non-parametric chi-square test (χ2) was used. In addition, analysis of variance (ANOVA) and the

two-tailed Fisher’s exact test or Scheffe test were applied for the purpose of re-examining possible dissimilarities between data anticipated in each group. Additionally, the hypothesis about the equality of the variance in each group was analysed with the Bartlett test. Logistic regression analysis was used to evaluate the influence of ATIII and PC on the dichotomous variable: survival or death. All statistical computations were performed with the use of a statistical package (Statistica for Windows, 6.0).

Results The mean birth weight of the infants investigated was 1,862.0 g (median 1,675.0 g; range 670–4,200 g). The infants were delivered between 25 and 41 weeks of pregnancy (mean 32.7 weeks; median 33 weeks). Those with suspected sepsis were recruited into the study at between 8 and 172 days of life (mean 21.1 days; median 14,5). In 57 of the 150 patients sepsis was excluded by laboratory measurements and blood cultures. In 19 infants, despite distinct physical and laboratory signs of infection, negative blood culture results were obtained. In only 74 infants sepsis was confirmed by positive results from both blood culture and laboratory measurements. Therefore, for further analysis, all patients were divided into three groups: group I, sepsis excluded, 57 infants; group II, sepsis confirmed, 74 infants; group III, clinical and laboratory signs of sepsis not confirmed by blood culture, 19 infants. The mean values of plasma ATIII and PC levels evaluated in neonates in group II and group III were statistically significantly lower than in the data obtained from infants without sepsis (respectively, for ATIII: group I vs group II, P=0.000001; group I vs group III, P=0.00002; for PC: group I vs group II, P=0.0000024; group I vs group III, P=0.00001) (Table 1). However, there were no significant differences between infants in groups II and III with respect to the values of plasma ATIII (P=0.7) and PC (P=0.6) concentrations. All patients in group I survived and were discharged, whereas seven of 74 newborns in group II and two of 19 Table 1 Comparative analysis of plasma antithrombin III and PC functional levels between the three groups of infants Variable

Group I n=57

Group II n=74

Group III n=19

70.6% 65.0% 43.0–121.0% 0.000001

50.4% 48.0% 6.0–105.0% 0.7

49.3% 47.0% 12.0–80.0% (Kruskal–Wallis test)

26.1% 24.0% 11.0–82.0% 0.0000024

17.9% 17.0% 2.0–57.0% 0.6

16.1% 17.0% 3.0–62.0% (Kruskal–Wallis test)

Antithrombin III Mean Median Range P Protein C Mean Median Range P

587

infants in group III died. In all cases the main symptoms leading to death were circulatory failure, generalized bleeding (venepuncture sites, alimentary tract, and lungs) and at least one organ failure (mostly renal or hepatic dysfunction). The diagnosis was confirmed by the findings of diffused foci inflammation and haemorrhagic diathesis during the autopsy. Moreover, the blood cultures obtained several hours before death in two patients in group III exhibited positive test results and Gram-negative bacteria were found. Also, in all cases of death in group II, sepsis was caused by Gram-negative bacteria. The mean values of plasma ATIII and PC functional concentrations evaluated in the neonates who died were statistically significantly lower than those of the survivors in both groups (groups II and III) analysed together (respectively: for ATIII P=0.003 and for PC P=0.00002) (Table 2). A highly statistically significant correlation (P=0.0016) between plasma PC level and risk of death was found in 93 of the patients investigated in both groups (group II and III) (Fig. 1). However, plasma ATIII concentration, evaluated in the same groups of patients, did not correlate significantly with the occurrence of death. On the other hand, the values of plasma ATIII levels, determined at the beginning of sepsis in both groups (group II and group III) statistically significantly correlated with respective data concerning PC (r=0.45; P=0.00006) (Fig. 2.). Plasma ATIII and PC concentrations evaluated in neonates with confirmed sepsis (group II) were significantly higher in cases of infection caused by Gramnegative bacteria than in those caused by Gram-positive strains (for ATIII: Gram-negative mean 35.2%, median 43%, range 6–64%; Gram-positive mean 54.0%, median 54.5%, range 10–98%; P=0.003; for PC: Gram-negative mean 10.5%, median 7.5%, range 3–26%; Gram-positive mean 20.1%, median 18.5%, range 7–48%, P=0.00002).

Discussion The neonatal PC system seems well prepared to match an acute thrombin formation. The duration of the procoaguTable 2 Plasma antithrombin III and PC functional levels in neonates with sepsis; a comparison between survivors and nonsurvivors Parameter

Antithrombin III Mean Median Range P Protein C Mean Median Range P

Survivors n=84

Non-survivors n=9

50.1% 48.0% 6.0–162.0% 0.003

35.9% 33.5% 5.0–70.0% (Mann–Whitney U test)

18.3% 17.0% 2.0–57.0% 0.00002

6.6% 7.0% 3.0–11.0% (Mann–Whitney U test)

lant challenge, however, may be a key factor. When thrombin formation stabilizes at an elevated level during sepsis, it is probable that the availability of PC is rate limiting for its activation. Thus, owing to the low levels of PC, the maintenance of the antithrombotic defences during sepsis seems to be more restricted in a neonate than in an older child. This speculation is supported by the observation that reduced PC levels in cord plasma predicted later occurrence of thrombosis in sick neonates [9]. The results of our study showed a marked decrease in the level of the physiological inhibition system of coagulation including ATIII and PC of septic neonates at the initial stage of disease. The comparable values of ATIII or PC found in patients with clinical signs and laboratory evidence of sepsis and either positive or negative results of blood culture tests, suggest that even small numbers of bacteria, not enough to be cultured, may significantly influence the regulation of thrombin generation. It is especially true with respect to cases of sepsis caused by Gram-negative bacteria. On the other hand, relatively high plasma ATIII and PC levels found in neonates with suspected sepsis may allow us to exclude the disease. One could suggest that evaluation of plasma ATIII and PC concentrations, simultaneously with measurement of leukocyte count, leukocyte differentiation, plasma C-reactive protein, and plasma procalcitonin concentration, might facilitate recognition of sepsis. It is known that low PC levels predict high mortality rates in children with meningococcal sepsis [12]. According to our knowledge, there are no similar observations described in neonates. Our data demonstrate the statistically significant correlation between plasma PC level, evaluated at the first stage of sepsis, and the risk of death in the course of the disease (Fig. 1). A sudden increase in the risk of mortality was observed when the plasma PC functional concentration dropped below 10%. Thus, we suggest this value might indicate the necessity for immediate supplementation of PC concentrate in neonates with sepsis. It is understood that results of conventional laboratory measurements (either plasma C reactive concentration or plasma procalcitonin level) may be of equal prognostic value in confirmed sepsis. However, they do not directly result in the modification of therapy. The attending physicians did know the values of plasma ATIII and PC functional concentrations. When the respective data were below the normal range (PC <20%; ATIII <40%) the immediate evaluation of fibrin split products, plasma fibrinogen level, partial thromboplastin time (PTT) and prothrombin time (PT) was performed. We speculate this “unblinding” might result in the earlier recognition of sepsis-induced coagulation disturbances. There has been one dose-finding, placebo-controlled study performed using PC concentrate in children [3]. This study showed a positive effect of PC administration on sepsis-induced coagulation disturbances. In neonates; the successful use of either ATIII or PC concentrate has been reported mostly in the treatment of hereditary deficiencies [4]. The anecdotal data, presenting the beneficial effect of activated protein C in neonates with sepsis, were published

588 1,1 1,0

Probability of the occurrence of death.

Fig. 1 Correlation between plasma PC functional level and the risk of death in the course of late-onset sepsis

0,9 0,8 0,7 0,6 0,5 0,4 0,3 0,2 0,1 0,0 -0,1 0

5

10

15

20

25

30

35

40

45

50

55

60

65

70

75

80

85

Protein C functional level (%)

[6, 11]. However, a randomized, double-blind, placebocontrolled, trial of drotrecogin alpha (Xigris) in paediatric patients with severe sepsis has been interrupted recently. The interim analysis showed a numerical increase in the rate of central nervous system bleeding in infants aged 2 months or less (Medscape alert—pediatric Xigris study stopped; April 2005). It is known that PC is an inactive substratum, which is transformed into the active form (A-PC) in an “ondemand” fashion. We speculate that administration of PC to newborn infants with sepsis might be a kind of supplementation of their reserves and exerts significantly minor adverse effects.

Fig. 2 Correlation between plasma PC and ATIII functional levels in the recognition of lateonset sepsis in neonates

The important consequences of our findings for daily practice seem to be following: a. Measurement of plasma ATIII and PC levels, together with C-reactive protein and procalcitonin as well as leukocyte count, may facilitate the recognition of sepsis. b. Evaluation of plasma PC concentration may have a prognostic value in the course of sepsis. c. Functional plasma level of PC below 10% measured in newborns with sepsis might be an indication for supplementation with PC concentrate.

120

AT III functional level (%)

100

80

60

40

20

0 0

10

20

30

40

Protein C functional level (%)

50

60

589

However, a randomized, placebo-controlled trial in neonates with severe sepsis is advisable to establish effectiveness of therapy with either ATIII or PC in reducing mortality in these patients.

References 1. Amaral A, Opal SM, Vincent JL (2004) Coagulation in sepsis. Intensive Care Med 30:1032–1040 2. Cvirn G, Gallistl S, Muntean W (1999) Effects of antithrombin and protein C on thrombin generation in newborn and adult plasma. Thromb Res 93:183–190 3. de Kleijn E, de Groot R, Hack C, Mulder P, Engl W, Moritz B, Joosten K, Hazelzet J (2003) Activation of protein C following infusion of protein C concentrate in children with severe meningococcal sepsis and purpura fulminans: a randomized, double-blind, placebo-controlled, dose-finding study. Crit Care Med 31:1839–1847 4. Dreyfus M, Magny J, Bridey F, Schwarz HP, Planche C, Dehan M, Tchernia G (1991) Treatment of homozygous protein C deficiency and neonatal purpura fulminans with purified protein C concentrate. N Engl J Med 325:1565–1568

5. Fisher CJ Jr, Yan SB (2000) Protein C levels as prognostic indicator of outcome in sepsis and related diseases. Crit Care Med 28 [Suppl 9]:S49–S56 6. Frommhold D, Birle A, Linderkamp O, Zilow E, Poschl J (2005) Drotrecogin alpha (activated) in neonatal septic shock. Scand J Infect Dis 37:306–308 7. Kreuz W, Veldmann A, Fisher D, Schlosser R, Volk WR, Ettingshausen CE (1999) Neonatal sepsis: a challenge in hemostaseology. Semin Thromb Hemost 25:531–535 8. Macias WL, Nelson DR (2004) Severe protein C deficiency predicts early death in severe sepsis. Crit Care Med 32 [Suppl 5]:S223–S228 9. Manco-Johnson M, Abshire T, Jacobson L, Marlar RA (1991) Severe neonatal protein C deficiency: prevalence and thrombotic risk. J Pediatr 119:793–798 10. Petaja J, Manco-Johnson MJ (2003) Protein C pathway in infants and children. Semin Thromb Hemost 29:349–361 11. Rawicz M, Sitkowska B, Rudziñska J, Kornacka MK, Bocheñski P (2002) Recombinant human activated protein C for severe sepsis in a neonate. Med Sci Monit 11:CS90–CS94 12. Rivard GE, David M, Farrel C (1995) Treatment of purpura fulminans in meningococcemia with protein C concentrate. J Pediatr 126:646–652