Arch Dis Child doi:10.1136/archdischild-2011-301029
  • Original article

Vitamin K deficiency bleeding after NICE guidance and withdrawal of Konakion Neonatal: British Paediatric Surveillance Unit study, 2006–2008

  1. John H Tripp3
  1. 1Child and Women's Health, Royal Devon and Exeter NHS Foundation Trust, Exeter, UK
  2. 2Paediatric Department, Royal Devon and Exeter NHS Foundation Trust, Exeter, UK
  3. 3Health Services Research, Peninsula Medical School, Exeter, UK
  1. Correspondence to Dr Alison Busfield, Child and Women's Health, Royal Devon and Exeter NHS Foundation Trust, Barrack Road, Exeter EX2 5DW, UK; alison.busfield{at}
  • Accepted 17 October 2012
  • Published Online First 12 November 2012


Objective To survey vitamin K deficiency bleeding (VKDB) and document vitamin K (VK) prophylaxis practice, and compare with findings predating withdrawal of Konakion Neonatal and guidance from the National Institute of Health and Clinical Excellence (NICE), both occurring in 2006.

Design Two-year surveillance of VKDB (2006–2008) using British Paediatric Surveillance Unit methodology. Postal questionnaire to consultant-led maternity units.

Setting UK and Irish Republic.

Patients All newborns and infants under 6 months with suspected VKDB.

Main outcome measures VKDB incidence and predisposing factors, VK prophylaxis recommended/received.

Results Eleven cases of VKDB were found: six (55%) babies received no VK prophylaxis, in five (45.5%) because parents withheld consent; three (27.5%) babies with late VKDB received intramuscular (IM) Konakion MM (two had biliary atresia, and one was delivered preterm); two (18%) babies received incomplete oral prophylaxis. Nine babies (82%) were breast fed. Three (27%) babies had liver disease; four (36%), including all those with liver disease, were jaundiced at presentation after 21 days. Four (36%) babies had intracranial haemorrhage, two probably suffering long-term morbidity. VK prophylaxis practice was defined in 236 (100%) units. All units recommended prophylaxis for every newborn: 169 (72%) IM, 19 (8%) oral, and 48 (20%) offered parental choice. All units that recommended IM prophylaxis used Konakion MM. Oral prophylaxis always involved multidose regimens for breastfed babies; 61 (91%) units used Konakion MM, and six (9%) used unlicensed products suitable for administration by parents.

Conclusions IM Konakion MM is efficacious, but parents withholding consent for recommended IM prophylaxis reduces effectiveness. Reappraisal of NICE guidance would be appropriate. Prolonged jaundice demands investigation. Late VKDB occasionally occurs after IM prophylaxis.

What is already known on this topic

  • Vitamin K (VK) prophylaxis protects almost all babies if given intramuscularly or by current multidose oral regimens.

  • The incidence of vitamin K deficiency bleeding (VKDB) and the associated morbidity and mortality fell significantly between 1994 and 2002.

  • Early recognition of liver disease protects against VKDB.

What this study adds

  • The incidence of VKDB did not significantly increase after withdrawal of Konakion Neonatal.

  • Intramuscular prophylaxis does not prevent all cases of VKDB; investigating prolonged jaundice identifies liver disease and VK deficiency. Twenty-six units (13%) lacked a recognised pathway for this.

  • The incidence of VKDB might be halved if all parents consented to VK prophylaxis.

  • 31% of UK units deviated from the NICE guidance of 2006; the guidance should be reviewed to reflect new data and options.


This is the fourth British Paediatric Surveillance Unit (BPSU) 2-year study of vitamin K deficiency bleeding (VKDB); each is designated by the year of completion, making this VKDB-08. The third study, VKDB-02, found the lowest incidence of VKDB and no death or long-term morbidity, but showed parental refusal of prophylaxis to be a new important issue.1

Up to and including VKDB-02, Konakion Neonatal (Roche, Basel, Switzerland) was used by almost all units that recommended intramuscular (IM) prophylaxis2; with exceedingly rare exceptions, a single 1 mg IM dose at birth protected against VKDB. In 2006, there were two important developments: first, Konakion Neonatal was withdrawn, leaving Konakion MM (Roche Products Limited, Welwyn Garden City, UK) as the only licensed preparation for prophylaxis; second, the National Institute for Health and Clinical Excellence (NICE) recommended IM in preference to oral prophylaxis.3 Increased use of IM prophylaxis with Konakion MM was inevitable, despite the paucity of published efficacy data.4 The Medicines and Healthcare Product Regulatory Agency (MHRA) advised further surveillance, which this study represents.


Cases of suspected VKDB were recruited in the UK and Irish Republic (IRE) from October 2006 to September 2008 using the well-described BPSU methodology. They were classified by the criteria used previously.1 Cases of late VKDB were also classified by more exclusive criteria, agreed in 1994,5 to allow international comparisons.

Postal questionnaires to document policies of vitamin K (VK) prophylaxis were sent to labour ward senior matrons of all consultant-led maternity units from February 2008, repeatedly if necessary, and finally to consultant paediatricians of non-responding units in March 2009. Units failing to respond or providing inconsistent data were contacted by telephone. Information requested included unit birth rate, VK regimen, consent practice, prolonged jaundice follow-up, and a written policy if available.

Birth statistics were obtained from the appropriate national bodies.6–9

Ethics and patient information advisory group approvals were obtained before the start.


BPSU study of VKDB results

Monthly returns for the BPSU report cards averaged 93.9% during the study.


After 46 notifications, including two directly to the investigators rather than via the BPSU, 43 questionnaires were returned, yielding 10 confirmed and one probable case, which were considered as the 11 reported cases; 12 notifications were duplicates, and 20 were ‘no case’ or ‘error’. Three questionnaires were not returned, two concerning the same notification, leaving two notifications unclassifiable.

Incidence of VKDB

Overall incidence per 100 000 live births was 0.64, which is not significantly different from that in VKDB-02 (0.48)1; both values are significantly lower than those in VKDB-94 (1.98) and VKDB-90 (1.61).1 ,10

Age at bleeding

One baby, who had received no VK, bled within 24 h of birth (‘early VKDB’). Four bled between 24 h and 7 days (‘classical VKDB’); three had received no prophylaxis, and one probably had one oral dose and spat it out. Six babies (four with intracranial haemorrhage (ICH), three with underlying liver disease, and two with both) bled after 7 days (‘late VKDB’).

Case details

Case details can be found in table 1. One infant was delivered at 24 weeks, and the remainder at term. Gender and exact weight were not requested, to protect confidentiality.

Table 1

Clinical features of the 11 cases of VKDB notified during the 2 years of study, 2006–2008

VK prophylaxis

In six of the 11 cases, no VK was given, in five because parents withheld consent; in one, no reason was documented. In all six, the recommended prophylaxis was IM VK. Of the five cases for whom consent was withheld, oral prophylaxis was also refused in two, mistakenly not offered in one, and not documented in two.

Three babies bled after IM VK. The preterm infant received 0.4 mg/kg, bled on day 91, and had no liver disease. Patient nine received 0.5 mg and bled on day 86, and patient 10 received 1 mg and bled on day 66; both were found to have biliary atresia.

Two patients bled after oral VK. Patient six received 1 mg at birth but, through error, not the intended 50 μg daily doses for days 1–5. Prophylaxis was resumed on day 6, but ICH occurred on day 23; α1-antitrypsin deficiency was later diagnosed. Patient three, born at home, was intended to receive 1 mg orally at birth and repeated weekly for 8 weeks. The initial dose was reportedly spat out and not repeated. Bleeding occurred on day 5.

Liver disease

Liver function was tested in all cases. In three it was normal. One baby had raised alkaline phosphatase (411 IU/l on day 19) but no jaundice or liver disease. Four had unconjugated hyperbilirubinaemia and otherwise normal liver function. Three with liver disease (two with biliary atresia, and one with α1-antitrypsin deficiency) were clinically jaundiced at presentation with bleeding, conjugated bilirubins 44–93 μmol/l; no additional VK had been given.

No other predisposing conditions were identified.


No deaths were directly attributable to VKDB. Patient 10 had biliary atresia and died from liver failure 2 months after ICH. Three others with ICH were breast fed and received either no or incomplete oral prophylaxis; two are expected to have long-term sequelae. No sequelae are expected in any others. Table 2 compares data for ICH between all BPSU studies.

Table 2

Details of babies with intracranial haemorrhage (ICH) secondary to VKDB (classified by criteria of the BPSU studies) in the four BPSU incidence studies

Late VKDB—international comparison

Cases 2, 4, 6 and 10 met the international criteria for late VKDB.5 VKDB-08 and VKDB-021 show significantly lower incidences than VKDB-90 and VKDB-941 ,10 and compare favourably with other European data5 (table 3).

Table 3

Cases of late VKDB in UK and Ireland from four BPSU studies, reclassified as confirmed or probable according to internationally agreed criteria (which are more exclusive than those used in the BPSU reports) to allow comparison with published data from other countries

VK prophylaxis survey results

Of 236 consultant-led units identified (217 in UK; 19 in IRE), 222 provided data by postal questionnaire, and 14 by telephone. Seventy-five sent their written policy and 70 clarified data by telephone.

The national statistics offices reported 843 634 births for the study period,6–9 and the units reported 842 450.

Recommended VK prophylaxis after uncomplicated term delivery

All units recommended VK prophylaxis for every neonate. Table 4 shows the routes of administration recommended in 2008 (this survey) and in 2003.2 Parental consent was sought in 233 units, written in 49 (21%) of them; no data were available for three.

Table 4

Recommended route of administration for vitamin K prophylaxis in the UK and Irish Republic (IRE), in 2003 and 2008

Changing practice

Of 215 units that responded, 106 (49%) reported a policy change in the previous 3 years, usually because of withdrawal of Konakion Neonatal (78 units, 74%) or to comply with NICE guidance (37 units, 35%).

IM prophylaxis

Of 217 units (92% of all units) that used IM prophylaxis, 169 (78%) recommended this route, and 48 (22%) offered it as choice. All used Konakion MM; 214 (99%) gave the licensed 1 mg dose (three gave 500 μg), 194 (89%) gave one dose, and 23 (11%) stated no dose number.

Of 169 units that recommended IM VK, 162 (96%) offered oral prophylaxis if the injection was declined, and seven did not.

Oral prophylaxis

Oral VK prophylaxis was used in 67 (28%) units (recommended in 19 (28%), and offered for parental choice in 48 (72%)). All units gave three or more doses to breastfed babies (tables 5 and 6).

Seventeen (89%) of the 19 units that recommended oral prophylaxis and 20 (42%) of the 48 that offered choice had separate high-risk policies using Konakion MM (IM in 36, intravenous (IV) in one).

Table 5

Oral prophylaxis regimens for exclusively breastfed infants

Table 6

Oral prophylaxis for bottle-fed infants

IV prophylaxis

Of the 231 units that responded, 62 (27%) sometimes used IV prophylaxis (always Konakion MM); of these, only 19 (31%) gave follow-on doses, 38 (61%) did not, and five did not respond.


Of the 203 units that responded, 157 (77%) assessed infants jaundiced at 2–3 weeks in a hospital-based clinic, 20 (10%) measured conjugated bilirubin in the community, and 26 (13%), representing 90 000 deliveries/year, answered ‘No’ to both options, implying that they had no formal pathway for investigating prolonged neonatal jaundice.


The four 2-year BPSU studies of VKDB (VKDB-90, -94, -02 and -08) have documented the incidence and features of the condition over 20 years.1 ,10

The incidence was significantly lower in VKDB-02 and VKDB-08 than previously,1 ,10 probably for two reasons: (1) the decline in the proportion of neonates not offered VK prophylaxis from about 30% in 198211 to zero; (2) the changes in practices—increased use of single-dose IM prophylaxis, now recommended in 72% of units, and oral regimens becoming multidose (extending over several weeks) for breastfed infants.2

That all four studies were conducted using consistent methodology and criteria supports the reliability of the findings, but we acknowledge that occasional cases may have been missed; similar methodology was found to underestimate cases when compared with recruitment from a disease registry.12 Lack of PIVKA-II (protein induced by VK absence/antagonism, factor II) measurement in most cases may be considered another study limitation.

IM Konakion MM

The change from Konakion Neonatal as the most used IM preparation in VKDB-021 to Konakion MM used by all in VKDB-08 was not associated with a significant increase in VKDB incidence. The change of solubilising agent (from Cremophor in Konakion Neonatal to the glycocholic acid/lecithin mixture in Konakion MM) has not apparently reduced efficacy; the prolonged protection after IM rather than oral or IV administration may result from the lipophilic VK forming a depot in the muscle.13

Three babies had late VKDB after IM Konakion MM at birth, presenting on days 66, 86 and 91 (two had liver disease—one was bottle-fed and received 0.5 mg, the other was preterm and received the licensed dose, 0.4 mg/kg). Late VKDB after IM prophylaxis is extremely rare, but clearly a single dose cannot prevent VKDB indefinitely, even without liver disease; other cases have been reported.14 ,15

The dose of IM VK required to give adequate VK levels for a prolonged period in preterm infants is uncertain, as there are no data from a large population given IM Konakion MM in different doses.16–18 Furthermore, with Konakion MM (five times more concentrated than Konakion Neonatal), the risk of inaccuracy in administering the small volume required, especially for the smallest babies, is substantial.

Oral prophylaxis

The number of units recommending oral prophylaxis has decreased, but all doing so now use more effective multidose regimens in breastfed infants. Failure of compliance (possibly professional) occurred in both cases associated with oral prophylaxis, implying that an appropriate oral regimen correctly administered offers equivalent protection (indistinguishable in these data) to that achieved by IM prophylaxis—although we recognise that babies deemed at highest risk are usually selected for IM prophylaxis.

IV prophylaxis

This does not give the same enduring protection as that achieved with IM prophylaxis,19 yet 38 units (61% of those using IV) left infants vulnerable to late VKDB by not giving follow-up doses.

NICE recommendations

The UK Department of Health and NICE significantly influence UK practice; the extent of their influence in the IRE is unknown. NICE gives independent, national guidance on drugs and healthcare practices after assessing effectiveness, safety and economic data. When NICE published guidance for VK prophylaxis in 2006,3 Konakion MM was the only product licensed in the UK for oral use. Presented in glass ampoules, every dose requires administration by a healthcare professional, making three oral doses an expensive alternative to one IM injection. Without evidence of commensurate benefit, NICE recommended IM administration as the most efficacious and cost-effective practice, reserving oral prophylaxis for infants whose parents declined IM injection.3 This survey found that, while 37 units (17% of all units; 34 UK, three IRE) changed their practice because of NICE guidance, 67 UK units (31%) did not comply with it. The proportion of UK units recommending IM prophylaxis increased, but oral prophylaxis was still recommended in 19 (9%), parents had choice in 47 (21.5%), and one gave 500 μg IM not 1 mg. Oral Konakion MM was used ‘off licence’ for breastfed infants in 18 units (27%) and for bottle-fed infants in 31 (46%). Another six (9%) used unlicensed products suitable for administration by parents.

Newly licensed product—should NICE guidance be reviewed?

In 2010, the oral VK food supplement Neokay (Neoceuticals, Knaresborough, UK), presented as a 1 mg gelatine capsule and suitable for parents to give, was licensed as a medicine by the MHRA. Like Konakion MM, it contains phytomenadione, but in coconut oil rather than in mixed micelles with bovine bile salts. Containing no animal products may make Neokay more acceptable to some,20 the gelatine capsule (which is not ingested) notwithstanding; those who counsel parents should know of this option. The same preparation (Neokay) administered with a plastic dropper giving 50 μg daily (after 1 mg Konakion MM or Orakay orally at birth) was used in five units during the study, although unlicensed as a medicine. This preparation and presentation would be suitable for vegans and is acceptable to midwives and parents.21

Danish studies found oral prophylaxis using a Konakion cremophor preparation, 2 mg at birth and 1 mg weekly for 12 weeks, to be as effective as 1 mg IM at birth, even in babies with biliary atresia, and more effective than the three 2 mg doses of Konakion MM used by most units in our survey.22 Denmark abandoned this regimen for lack of a licensed product, but a suitable product now licensed in the UK could make it a more acceptable option. The current drug cost using weekly Neokay capsules equates to that of three doses of Konakion MM, but saves the cost of administration by healthcare professionals. Reappraisal by NICE would be appropriate.

Parental consent

Seeking parental consent is now standard practice. It is notable that, in nine (53%) of the VKDB cases in the last two BPSU studies, parents refused the recommended IM prophylaxis, whereas there were no refusals in units recommending oral prophylaxis or offering a choice (p=0.03). Similar data are reported from New Zealand: of 19 cases of VKDB (1998–2008), VK was not given in 17, in 11 because consent was withheld.23

The reasons for parents refusing IM prophylaxis is poorly documented, including whether oral prophylaxis was offered as an alternative. This needs addressing to improve understanding and to avoid possible litigation if bleeding occurs.

Liver disease and late VKDB

Late-onset VKDB is not prevented by VK prophylaxis alone. Liver disease is known to cause VK deficiency and VKDB.1 That the latest two studies have reported only four VKDB cases associated with liver disease among 3.15 million newborns probably reflects three factors: universal policies for prophylaxis, improved oral regimens, and increased vigilance for liver disease by professionals. All three cases of late VKDB associated with liver disease in this study were associated with clinical jaundice at presentation after 3 weeks of age. One other patient without liver disease was clinically jaundiced at presentation on day 34. It is imperative to investigate all cases of prolonged jaundice3 and provide additional VK prophylaxis, if required, but 26 units (13%) apparently had no formal pathway for this.

In the Netherlands, oral prophylaxis using VK 1 mg at birth and 25 μg daily for 3 months failed to prevent VKDB in babies with liver disease.24 Lack of community surveillance for prolonged jaundice and lack of the equivalent of the UK's ‘yellow alert’ publicity campaign probably contributed. The Netherlands now use 150 μg daily,25 equivalent to the successful 1 mg weekly Danish regimen described above.22


Since the first BPSU study, awareness of VKDB and the effectiveness of prophylaxis have improved considerably. Parental refusal of prophylaxis is now the most important issue. Those taking consent should understand local policy, and parents considering refusal should be offered a senior opinion. Documentation is essential, including reasons for withholding consent. The recent licensing of a vegetarian oral product, suitable for administration by parents, may increase the acceptability of prophylaxis for some parents.

VK prophylaxis alone will not prevent all cases of VKDB. Community surveillance for predisposing conditions, particularly liver disease, remains essential.


  • AB and RS are joint first authors of this paper.

  • Correction notice This paper has been amended since it was published Online First. It should now be noted that AB and RS are joint first authors of this paper.

  • Acknowledgements We thank the BPSU for supporting the studies, all reporting paediatricians for their collaboration and forbearance, and the reviewers for their helpful criticism. We thank Roche Pharmaceuticals for funding this study.

  • Contributors Ethical approval was sought by AB, the principal investigator, who collected and analysed data of VKDB cases. RS collected and analysed data from the survey of VK prophylaxis. RS and AB collaborated to write the paper. JHT and AWM helped with study design, interpretation of results and writing of the paper. JHT obtained funding from Roche.

  • Funding The funders, Roche Pharmaceuticals, had no involvement in the study design or data analysis and no editorial rights.

  • Competing interests JHT and AWM have previously received funding from Roche Pharmaceuticals. JHT acted as an expert witness to the MHRA in the appeal for approval of Neokay.

  • Ethics approval Cornwall Research Ethics Committee.

  • Provenance and peer review Not commissioned; externally peer reviewed.


Free sample

This recent issue is free to all users to allow everyone the opportunity to see the full scope and typical content of ADC.
View free sample issue >>

Don't forget to sign up for content alerts so you keep up to date with all the articles as they are published.