Prolonged antibiotic usage has been found to result in vitamin K-related coagulopathy in about one-third of children with higher incidence in malnourished children and infants [3]. Extrapolating the evidence from adult and limited pediatric studies, most of the NICUs in India prefer to administer vitamin K every 7 days to neonates on antibiotic therapy. However, evidence for this practice is lacking in the neonatal population. Studies on preterm neonates and neonates in surgical ICU on prolonged antibiotic therapy have demonstrated high levels of vitamin K after its single parenteral birth dose [4,5]. Therefore, this study was planned to revisit the practice of administering vitamin K during prolonged antibiotic therapy amongst neonates.

This open-label randomized clinical trial was conducted in a level-3 neonatal unit in Northern India, between July 2015 and August 2016. Neonates with first episode of symptomatic sepsis on antibiotics for 7 or more days were included. Neonates who received vitamin K prior to start of antibiotic therapy (except the birth dose), who had any clinical bleeding, cholestasis or prenatally diagnosed bleeding disorder were excluded. Eligible neonates were enrolled on the seventh day of antibiotic therapy after obtaining written informed consent from the parents. The study was approved by the Institute Ethics Committee, All India Institute of Medical Sciences, New Delhi, India.

Primary outcome was the proportion of neonates with PIVKA-II levels more than 2 ng/mL after 7±2 days of enrolment. Secondary outcome was the proportion of neonates with coagulation abnormalities (prothrombin time (PT): 2 seconds more than upper limit as per gestational age and post-natal age cut-off [6]) after 7±2 days of enrolment.

There was no published study that had used PIVKA-II levels as surrogate marker for vitamin K deficiency in neonates on prolonged antibiotics. So, sample size calculation was based on a study in patients aged 3 months to 23 years, suffering from cystic fibrosis and on prolonged antibiotic therapy where abnormal coagulation and raised PIVKA-II levels were found in 33% of cases [7]. Thus, a total of 94 neonates (47 in each group) were required to have 80% chance of detecting with significance at 5% level, a decrease in vitamin K deficiency from 33% in the control group to 10% in the intervention group.

Statistical analysis: Statistical analysis was performed using Stata 11.2 (Stata-Corp, College station, Texas, US). Categorical variables were compared by Chi square/Fisher Exact test. Continuous variables were compared by Student t-test (if normally distributed) or Wilcoxson rank sum Test (if skewed).

A total of 80 neonates were enrolled out of 120 neonates with sepsis during the study period. Among enrolled neonates, 41 were allocated to vitamin K group and 39 were allocated to ‘no vitamin K’ group (Fig. 1). Two neonates in the vitamin K group died before outcome assessment because of severe sepsis. In three neonates (2 in vitamin K group and 1 in no vitamin K group), second blood sample could not be collected after 7±2 days of enrolment. Thus, 37 neonates were analyzed in the vitamin K group and 38 in ‘no vitamin K’ group. Baseline demographic variables and postnatal morbidities were comparable between the two groups (Table I).

Fig. 1 Trial Flow.

TABLE I 	Baseline Demographic Variables and Postnatal Morbidities 

The baseline prevalence of vitamin K deficiency was 100% in both the groups. Quantitative PIVKA-II levels could be done in only 37 neonates (20 in the vitamin K group and 17 in the no vitamin K group). Mean (SD) PIVKA-II levels were very high in both the groups with no significant difference between the two groups (Table II).

TABLE II Prevalence of Vitamin K Deficiency and PIVKA II Levels at Enrolment 

The prevalence of vitamin K deficiency after 7±2 days of enrolment was 100% in both the groups. Quantitative PIVKA-II levels could be done in only 35 neonates (18 in the vitamin K group and 17 in the no vitamin K group). Mean (SD) PIVKA-II levels were comparable between the two groups (Table III). There was no difference in the prevalence of coagulation abnormalities as assessed by deranged PT or clinical bleeding between the two groups.

TABLE III Prevalence of Vitamin K Deficiency and PIVKA II Levels After 7 ± 2 Days of Enrolment

This open label randomized controlled trial, compared the effect of 1 mg parenteral vitamin K administration on seventh day of antibiotic therapy versus no vitamin K on the prevalence of vitamin K deficiency in neonates with sepsis. All neonates in both groups were deficient in vitamin K when assessed for outcome.

The strength of our study is that PIVKA-II levels were assessed by an independent clinician, blinded to the group allocation, thus minimizing bias. The major limitation of our study was that only 80 neonates out of the total planned 94 neonates could be enrolled because of limited budget and time factors. Blinding could not be done due to nature of the intervention. Both vitamin K1 and K3 were utilized in the intervention group, as per availability in the drug store of the institute. However, this is not likely to affect the outcome as both vitamin K1 and K3 have equal efficacy in preventing vitamin K deficiency [8]. Further, PIVKA-II level assessment, ELISA kits from two different manufacturers were used due to financial constraints and quantitative PIVKA-II levels could not be done in all the enrolled neonates due to non-availability of blood serum samples.

We conclude that neonates with sepsis receiving antibiotics for 7 or more days have universal biochemical vitamin K deficiency even after vitamin K administration, and question the utility of this practice in neonates on prolonged antibiotic therapy. However, it also raise questions about the usefulness of PIVKA-II measurement for assessing vitamin K deficiency in neonates. There is a need to establish normal values and standards of PIVKA-II measurements for assessing subclinical vitamin K deficiency in the newborn infant.

Contributors: AS: study conception and implementation, data management and writing the manuscript; MJS,AT,RS,SC,RA: supervised implementation of the study and contributed to writing of the manuscript. All authors approved the final version of manuscript, and are accountable for all aspects related to the study.

Funding: None; Competing interest: None stated.

1. Lipsky JJ. Mechanism of the inhibition of the gamma-carboxylation of glutamic acid by N-methylthiotetrazole-containing antibiotics. Proc Natl Acad Sci USA. 1984;81:2893-7.

2. Crowther MA, McDonald E, Johnston M, Cook D. Vitamin K deficiency and D-dimer levels in the intensive care unit: a prospective cohort study. Blood Coagul Fibrinolysis. 2002;13:49-52.

3. Bhat RV, Deshmukh CT. A study of Vitamin K status in children on prolonged antibiotic therapy. Indian Pediatr. 2003;40:36-40.

4. Najmaldin A, Francis J, Postle A, Griffiths DM, Burge DM, Karran S, et al. Vitamin K coagulation status in surgical newborns and the risk of bleeding. J Pediatr Surg. 1993;28:138-43.

5. Clarke P, Mitchell SJ, Wynn R, Sundaram S, Speed V, Gardener E, et al. Vitamin K prophylaxis for preterm infants: A randomized, controlled trial of 3 regimens. Pediatrics. 2006;118:e1657-66.

6. Pal S, Curley A, Stanworth SJ. Interpretation of clotting tests in the neonate. Arch Dis Child Fetal Neonatal Ed. 2015;100:F270-4.

7. De Montalembert M, Lenoir G, Saint-Raymond A, Rey J, Lefrère JJ. Increased PIVKA-II concentrations in patients with cystic fibrosis. J Clin Pathol. 1992;45:180-1.

8. Chawla D, Deorari AK, Saxena R, Paul VK, Agarwal R, Biswas A, et al. Vitamin K1 versus vitamin K3 for prevention of subclinical vitamin deficiency: a randomized controlled trial. Indian Pediatr. 2007;44:817-22.