PIVKA-II correlates with INR but not protein C or protein S concentrations in cord blood among newborns

Abstract

BACKGROUND: Protein induced by vitamin K absence (PIVKA)-II, inactive precursor of prothrombin, is elevated in vitamin K (VK) deficiency. Our aims were to find the prevalence of VK deficiency in neonates, assess the utility of international normalized ratio (INR) as a screening tool, and explore the relationship between PIVKA-II, activated partial thromboplastin time (aPTT) and VK dependent anticoagulants.

METHODS: INR, aPTT, PIVKA-II, and proteins C and S activities were measured in neonatal cord blood prior to VK administration.

RESULTS: We found 45% of neonates had subclinical VK deficiency based on PIVKA-II levels and 7% based on INR. Receiver operating characteristic (ROC) analysis assessed the utility of INR in detecting >4 ng/mL of PIVKA-II and ROC of the area under the curve was 0.70 (95% CI 0.46–0.92, p = 0.07). Proteins C and S activities were normal for age and did not correlate with PIVKA-II [(r = 0.40, p = 0.14) and (r = 0.29, p = 0.29), respectively]. There was no association between aPTT and PIVKA-II (p = 0.83).

CONCLUSION: PIVKA-II seems to be a sensitive indicator of mild VK deficiency. Further studies are needed to investigate the lack of relationship between PIVKA-II and functional protein C or S levels.

Keywords

PIVKA-II vitamin K deficiency newborns INR protein C protein S

1 Background

Protein induced by vitamin K absence (PIVKA)-II is a non-carboxylated, inactive precursor of coagulation factor II (prothrombin). Factor II precursor and other vitamin K dependent coagulation factor precursors normally undergo post-translational carboxylation by gamma-glutamyl carboxylase in the liver prior to being secreted into plasma. Patients with vitamin K deficiency, liver disease or warfarin therapy have underproduction of active vitamin K-dependent coagulation factors and accumulation of PIVKA-II.

In neonates, the placenta acts as a relative barrier for fluctuations in maternal vitamin K concentrations and therefore only very small amount of vitamin K can cross the placenta [1]. All vitamin K-dependent coagulation factors are decreased at birth possibly due to decreased synthesis of precursor, non-availability of enzymes needed to produce vitamin K, poor stores of vitamin K itself, and premature overall liver synthetic function, thus increasing the risk for hemorrhage [1 –3]. At birth, the levels of the vitamin K-dependent factors (factors II, VII, IX, and X) are about 25–70% of adult plasma levels and they remain at this level the next six months of post-natal life [4].

Low vitamin K-dependent procoagulant factor levels in newborns may reflect as a prolonged prothrombin time (PT) and consequently a higher international normalized ratio (INR). Traditionally, an elevated INR is used as an indicator of vitamin K deficiency and has been used to monitor resolution after vitamin K supplementation. However, the INR is relatively insensitive to vitamin K deficiency and a change in INR usually occurs only after factor II levels fall to less than 50% of its baseline value [5]. PIVKA-II, on the other hand, correlates more closely to the degree of vitamin K deficiency, rendering it a more sensitive marker [6]. Based on these facts, we sought to determine if PIVKA-II could be a better surrogate measure of vitamin K deficiency in neonates rather than the INR which is currently used as a laboratory test for this purpose. The second aim was to determine if the activated partial thromboplastin time (aPTT) due to its responsiveness to factor IX, provides additional useful information in neonatal subclinical vitamin K deficiency. The third aim was to determine if other vitamin K dependent proteins, specifically protein C and protein S activities correlate with PIVKA-II levels.

2 Methods

This study was approved by Institutional Review Board at Baylor College of Medicine. After delivery of the placenta, 5–10 mL of cord blood was collected within 30 minutes of delivery using a 21-gauge needle. Cord blood samples were collected from 29 healthy newborns born after 38 weeks gestation prior to the routine administration of vitamin K. Per protocol reported previously by Teruya et al., specimens were processed into platelet-free plasma [7]. Hemoglobin fractionation was performed by high performance liquid chromatography to confirm that the cord blood samples originated from the neonates. Hemoglobin F in the specimens was 82% (range 70–94%), indicating that the specimens were from neonates. Plasma aliquots were stored at –80°C until PT/INR, aPTT, protein C activity, protein S activity, and PIVKA-II were measured. PIVKA-II was measured by enzyme immunoassay (Asserachrom PIVKA-II™, Stago, France). PT, aPTT, and protein C and protein S activity assays were performed on the STA-R analyzer according to manufacturer’s instructions. PT was measured using Neoplastine CI™, aPTT was measured using STA-PTT™ and functional protein C (STA-Staclot Protein C™) and functional protein S (STA-Staclot Protein S™) were measured by an aPTT based assay. All reagents used in this study were purchased from Diagnostica Stago, Asnieres, France. Non-normal data were analyzed using non-parametric tests such as Mann-Whitney to compare two groups and Spearman’s correlation method was used to assess the correlation between continuous variables. The patients were dichotomized into two groups based on the median value of PIVKA-II (PIVKA-II ≤4 ng/mL and PIVKA-II >4 ng/mL). Chi-square analysis was done to assess the association between both the categorical PIVKA-II groups and the two patient groups with normal aPTT and prolonged aPTT. Receiver operating characteristic (ROC) analysis was done to evaluate the ability of INR to discriminate between the two groups of PIVKA-II.

3 Results

The values of routine coagulation parameters measured in the cord blood samples of 29 healthy newborn infants along with PIVKA-II, protein C and protein S activities are listed in Table 1. A correlation coefficient was computed to assess the relationship between the dependent variables. Overall, there was significant positive correlation between INR and PIVKA-II (Fig. 1). On the other hand, there was no correlation between PIVKA-II and aPTT, protein C and protein S activities (Figs. 2 and 3).

The median (IQR) PIVKA-II level in neonates with an elevated INR of >1.30 was 56.9 ng/mL (13.6–153.8), whereas the level was 2.4 ng/mL (1.05–8.15) in neonates with normal INR ranges (Table 2). The difference was statistically significant (p < 0.05). Forty five percent of neonates had PIVKA-II levels above the normal reference interval (>4 ng/mL). Only two subjects had INR elevated above >1.30. The distribution of protein C and protein S activities were the same across both the PIVKA-II groups (p = 0.39, p = 0.78).

Sensitivity and specificity for the different cut-off points were calculated by receiver operating characteristic curve methodology to further explore the diagnostic accuracy of PIVKA-II. Our results indicate that at a cut-off level of 13.4 ng/mL of PIVKA-II, there was a concomitant elevation of INR; the sensitivity and specificity were 100% and 85%, respectively and accuracy of 87%. If PIVKA-II levels exceed 13.4 ng/mL, then the sensitivity drops to 50% indicating 50% of neonates do not show elevation of INR at >13.4 ng/mL of PIVKA-II.

4 Discussion

PIVKA-II has been detected in 10–30% of neonatal cord blood samples in the US according to several studies [8 –10]. Our study showed about 45% of the cord blood samples with a PIVKA-II levels of >4 ng/mL, which may be partially due to a delay in maturation of fetal liver enzymes accompanied by insufficient placental transfer of vitamin K [11]. Even studies that have used 100 ng/mL as cut-off values for normal levels report a high prevalence of PIVKA-II in about 21.5% of cord blood samples [12]. Furthermore there was a significant linear correlation between INR and PIVKA-II (r = 0.790, p < 0.001). This clearly indicates that PIVKA-II levels can be used as an alternative to PT/INR for assessing the vitamin K deficient status in neonates. A careful examination of the correlation reveals that newborns with PIVKA-II of >0.4 ng/mL had INR well within the normal range of <1.5. Only one neonate had INR of 1.79 and PIVKA-II level of 153.8 ng/mL, however protein C and protein S levels were normal for age at 37% and 48%, respectively. There is a significant difference between the median values of PIVKA-II in newborns with and without an elevation of the INR. In addition, of the 14 neonates with PIVKA-II levels of >4 ng/mL, only two had an elevated INR, indicating that INR is affected by higher concentration of PIVKA-II, making it useful to assess moderate deficiency states and not subclinical vitamin K deficiency. This finding has been confirmed by several other studies done in the neonatal population [11].

ROC analysis reveals that when the cut-off level of PIVKA-II is reduced from 13.4 ng/mL, the sensitivity increases (>85%) and specificity decreases from 80% to approximately 10%. For PIVKA-II above 13.4 ng/mL, the sensitivity falls to 50% indicating that INR does not show comparable increase. This indicates the poor utility of INR as a screening tool for subclinical vitamin K deficiency in neonatal population. Some studies have considered 2 ng/mL as abnormal PIVKA-II levels in cord blood as indicated by the PIVKA-II assay kit manufacturer (Asserachrom PIVKA-II, Stago, France). PIVKA-II level of >100 ng/mL has been reported in cord blood of neonates enrolled in a randomized controlled trial to test the efficacy of vitamin K1 and K3 [8].

In our study, aPTT shows poor correlation with PIVKA-II. Reagent sensitivity is such that a reduction of <30% of one or more intrinsic factor levels is necessary before the aPTT will be prolonged. However, in subclinical vitamin K deficiency, there is only a mild reduction in all the vitamin K dependent factors which is likely the reason for the lack of association [13].

Functional protein C and protein S levels did not correlate with PIVKA-II. There could be many reasons behind this finding. PIVKA proteins include decarboxylated forms of factors II, VII, IX, and X, protein C, and protein S. It has been shown that PIVKA-II and PIVKA-protein C exhibit significant levels of positive correlation indicating that there may be possible links in the mechanisms involved in the production of these two PIVKA proteins [14]. However, relationship between PIVKA-II and protein C still remains elusive. Protein C activity, free protein S, and PIVKA-protein C have been shown by Matzusaka et al. to be positively correlated with each other, and protein C activity has been reported to be inversely correlated with PIVKA-II concentration in neonates [6]. However, in Matzusaka’s study full term neonates had high PIVKA-II levels indicating vitamin K deficiency as evidenced by <20% of clotting activity and subsequent increased risk for hemorrhagic disease during the first six days of life. On the contrary, the PIVKA-II levels in our study indicate subclinical deficiency. Chromogenic protein C assay measures some portion of PIVKA-protein C whereas the aPTT based protein C assay that was employed in our study does not measure PIVKA-protein C. In addition, protein C exhibits shorter half-life of 4–6 hours compared to 24 hours for PIVKA-II. The inverse relationship between PIVKA-II and PIVKA-protein C that was previously reported may indicate possible links in the mechanisms involved in the production of these two PIVKA proteins. This may explain the fact that our patients with detectable PIVKA-II levels have low normal protein C activity in our neonatal population.

Protein S is a cofactor of activated protein C. In adults, approximately 40% of protein S circulates as free protein S and the rest is bound to C4b binding protein which regulates the complement system [15]. In neonates, a premature liver may result in low normal activity of protein S ranging from 33–46% as seen in our study. Moreover, studies have shown that PIVKA-II correlates well with free antigenic protein S but not functional protein S. In vitamin K deficiency, there is evidence that a portion of free protein S is converted into PIVKA-protein S thereby leading to a decrease in the protein S activity [6].

The limitation of our study is the non-availability of vitamin K levels in our cohort of neonates. Vitamin K was not measured due to insufficient specimen volumes. However, it is known that PIVKA-II correlates very well with vitamin K levels and therefore could be used in its place [6]. Another limitation of our study is the relatively small sample size and its homogeneity. It would limit the extrapolation of the results to the general population. Future larger studies are warranted to confirm these findings.

In conclusion, PT/INR and aPTT are not reliable coagulation tests to screen for subclinical vitamin K deficiency. PIVKA-II seems to be a sensitive indicator of mild vitamin K deficiency. The effect of PIVKA-II on protein C and protein S is elusive and more studies are necessary to define the role of all PIVKA proteins in neonatal hemostasis.

References

Hathaway

. New insights on vitamin K. Hematol Oncol Clin North Am 1987;1(3):367–79.

Lippi

, Franchini

. Vitamin K in neonates: Facts and myths. Blood Transfus 2011;9(1):4–9.

Pichler

, Pichler

. The neonatal coagulation system and the vitamin K deficiency bleeding - a mini review. Wien Med Wochenschr 2008;158(13-14):385–95.

Andrew

, Paes

, Milner

, et al . Development of the human coagulation system in the healthy premature infant. Blood 1988;72(5):1651–7.

O’Shaughnessy

, Allen

, Woodcock

, Pearce

, Harvey

, Shearer

. Echis time, under-carboxylated prothrombin and vitamin K status in intensive care patients. Clin Lab Haematol 2003;25(6):397–404.

Matsuzaka

, Tanaka

, Fukuda

, Aoki

, Tsuji

, Kondoh

. Relationship between vitamin K dependent coagulation factors and anticoagulants (protein C and protein S) in neonatal vitamin K deficiency. Arch Dis Child 1993;68(3 Spec No):297–302.

Edwards

, Naik-Mathuria

, Gay

, Olutoye

, Teruya

. Parameters of thromboelastography in healthy newborns. Am J Clin Pathol 2008;130(1):99–102.

Chawla

, Deorari

, Saxena

, et al . Vitamin K1 versus vitamin K3 for prevention of subclinical vitamin deficiency: A randomized controlled trial. Indian Pediatr 2007;44(11):817–22.

Kumar

, Greer

, Super

, Suttie

, Moore

. Vitamin K status of premature infants: Implications for current recommendations. Pediatrics 2001;108(5):1117–22.

10.

Laosombat

, Kenpitak

, Wongchanchailert

, Wiriya-sateinkul

. Acarboxy prothrombin (PIVKA-II) in cord plasma in the south of Thailand. J Med Assoc Thai 1993;76(5):278–84.

11.

Dituri

, Buonocore

, Pietravalle

, et al . PIVKA-II plasma levels as markers of subclinical vitamin K deficiency in term infants. J Matern Fetal Neonatal Med 2012;25(9):1660–3.

12.

Motohara

, Takagi

, Endo

, Kiyota

, Matsuda

. Oral supplementation of vitamin K for pregnant women and effects on levels of plasma vitamin K and PIVKA-II in the neonate. J Pediatr Gastroenterol Nutr 1990;11(1):32–6.

13.

Burns

, Goldberg

, Wenz

. Paradoxic effect of multiple mild coagulation factor deficiencies on the prothrombin time and activated partial thromboplastin time. Am J Clin Pathol 1993;100(2):94–8.

14.

Bovill

, Soll

, Lynch

, et al . Vitamin K1 metabolism and the production of des-carboxy prothrombin and protein C in the term and premature neonate. Blood 1993;81(1):77–83.

15.

Dahlback

, Stenflo

. High molecular weight complex in human plasma between vitamin K-dependent protein S and complement component C4b-binding protein. Proc Natl Acad Sci U S A 1981;78(4):2512–6.