com
Original article
Correspondence to
Dr Alison Buseld, Child and
Womens Health, Royal Devon
and Exeter NHS Foundation
Trust, Barrack Road,
Exeter EX2 5DW, UK;
alison.buseld@nhs.net
AB and RS are joint rst
authors of this paper.
Received 12 September 2011
Revised 16 October 2012
Accepted 17 October 2012
Published Online First
12 November 2012
ABSTRACT
Objective To survey vitamin K deciency bleeding
(VKDB) and document vitamin K (VK) prophylaxis
practice, and compare with ndings 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 (20062008)
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 ve (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
dened 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 efcacious, 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.
INTRODUCTION
To cite: Buseld A,
Samuel R, McNinch A, et al.
Arch Dis Child 2013, 98,
4147.
METHOD
Cases of suspected VKDB were recruited in the UK
and Irish Republic (IRE) from October 2006 to
September 2008 using the well-described BPSU
41
Original article
methodology. They were classied by the criteria used previously.1 Cases of late VKDB were also classied 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 nally to consultant paediatricians of nonresponding 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.69
Ethics and patient information advisory group approvals were
obtained before the start.
RESULTS
BPSU study of VKDB results
Monthly returns for the BPSU report cards averaged 93.9%
during the study.
Notications
After 46 notications, including two directly to the investigators
rather than via the BPSU, 43 questionnaires were returned,
yielding 10 conrmed and one probable case, which were considered as the 11 reported cases; 12 notications were duplicates, and 20 were no case or error. Three questionnaires
were not returned, two concerning the same notication,
leaving two notications unclassiable.
Incidence of VKDB
Overall incidence per 100 000 live births was 0.64, which is not
signicantly different from that in VKDB-02 (0.48)1; both
values are signicantly 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 condentiality.
VK prophylaxis
In six of the 11 cases, no VK was given, in ve because parents
withheld consent; in one, no reason was documented. In all six,
the recommended prophylaxis was IM VK. Of the ve 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.
42
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
1-antitrypsin deciency) were clinically jaundiced at presentation with bleeding, conjugated bilirubins 4493 mol/l; no additional VK had been given.
No other predisposing conditions were identied.
Outcome
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 longterm sequelae. No sequelae are expected in any others. Table 2
compares data for ICH between all BPSU studies.
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 g), 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.
Buseld A, et al. Arch Dis Child 2013;98:4147. doi:10.1136/archdischild-2011-301029
Clinical features of the 11 cases of VKDB notified during the 2 years of study, 20062008
Parental
consent
withheld?
Oral
alternative
offered
Age at
onset
(days)
Delay in
presentation
to hospital
(days)
Sites of
presenting
bleed
Other
bleeding
PT in
seconds
or (INR)
APTT in
seconds
or (APTR)
PIVKA-II
(arbitrary
units/ml)
Jaundiced at
presentation
Liver
disease
Adverse
outcome
expected
primarily
breast
Breast
Breast
<1
Heel prick
GIT
(>15)
(>5)
55
No
No
Unknown
34
5
2
3
Bruising
Bruising, heel
prick
GIT ICH
>180
30.6 (2.6)
>180
42.6 (1.6)
Yes
Yes
No
No
No
No
GIT ICH
Did not
clot
105
No
No
Yes
No
No
No
>180
Yes
1antitrypsin
deficiency
Yes
Case
number
Prophylaxis
Hospital
policy
None
IM
Yes
Unknown
2
3
IM
Oral 1mg
at birth
+weekly
for
8 weeks
Yes
No
No
None
Oral
Uncertain
when given,
perhaps day
3
Not taken
well
None
IM
Yes
Yes
Breast
20
Bruising
None
IM
No record
Unknown
Breast
Oral
1 mg at
birth,
50 g
daily
IM
No
Breast
23
Oral
1 mg at birth,
50 g daily,
but day 15
not given
None
Bruising, heel
prick
ICH
Did not
clot
Did not
clot
>180
Yes
Unknown
Unknown
No
No
No
None
IM 0.5 mg
IM
IM
Yes
No
Yes
Breast
Formula
5
86
0
2
Very
prolonged
(>10)
74.6
82
8
9
Bruising
GIT
Bruising
Bruising
>180
91.9
Yes
Yes
No
No
10
IM 1 mg
IM
No
Breast
66
146 (24)
101 (3.5)
11*
IM 0.3 mg
IM
0.4 mg/kg
No
Primarily
breast
91
42.3
59.5
No
Biliary
atresia
Biliary
atresia
No
Feed
Venepuncture
site
GIT
ICH
81
Yes
No
Table 1
Died from
liver failure
No
*Probable case.
APTR, activated partial thromboplastin time ratio; APTT, activated partial thromboplastin time; GIT, gastrointestinal bleeding; ICH, intracranial haemorrhage; IM, intramuscular; INR, international normalised ratio; VKDB, vitamin K deficiency bleeding;
PIVKA-II, protein induced by vitamin K absence/antagonism, factor II; PT, prothrombin time.
Original article
43
Original article
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
Study: cases of VKDB*
VKDB-90
(n=27)
VKDB-94
(n=32)
VKDB-02
(n=7)
VKDB-08
(n=11)
10 (1)
2 (1)
8
0
0
5
5
0
0
10
4
7
114 (6)
1
10 (2)
0
6 (2)
3
1
3
6
0
1 (1 mg)
8
3**
5
14 (2)
4
0
0
0
0
0
0
0
0
0
0
0
0
4
0
2
1
0
2
0
1
1 (1 mg)
4
2
2
12
0
*Definition of VKDB: infant under 6 months old with appropriate history of bleeding, documented prothrombin time or INR at least twice the control value, platelet count normal or
raised, no evidence of infection/disseminated intravascular coagulation.
Probable case: appropriate history of bleeding, diagnosis other than VKDB unlikely but lacking full laboratory confirmation.
One death due to liver failure.
Incomplete administration.
Soya feeds in two other cases.
**Pancreatic disease in one other.
Includes the case of IM prophylaxis and one probable case.
BPSU, British Paediatric Surveillance Unit; IM, intramuscular; VK, vitamin K; VKDB, vitamin K deficiency bleeding.
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
Study title (each last 2 years)
VKDB-90
VKDB-94
VKDB-02
VKDB-08
Birth population
(a) Idiopathic cases
(b) Secondary cases
(c) Predisposing illness
Total cases (probable included)
Total incidence a+b+c (95% CI)
True incidence a+b (95% CI)
Prophylaxis received
None
Oral1
Oral2
Oral3
Multiple oral
IM
Death
Handicap
Prophylaxis failures (95% CI)
Complete prophylaxis
Complete or incomplete
Unknown
1671000
11
5
0
16 (2)
0.96 (0.49 to 1.4)
0.96 (0.49 to 1.4)
1609785
10
13
1
24 (3)
1.5 (0.9 to 2.1)
1.4 (0.84 to 2.0)
1456200
0
2
0
2
0.14 (0.00 to 0.32)***
0.14 (0.00 to 0.32)***
1700000
2
2
0
4
0.24 (0.00 to 0.35)***
0.24 (0.00 to 0.35)***
10
6 (presumed in 1)
Regimen not used
Regimen not used
Regimen not used
0
1
8
6
13
2
2
Regimen not used
1
1
4 or 5
0
0
1
1 (probably given)
0
0
0
0
2
Regimen not used
Regimen not used
0
1 (incomplete)
1 (1 mg)
0
2
12 (5 to 19)
16 (8 to 24)
8
2 (0 to 5)*
2 (0 to 5)**
1 (0 to 5)**
4 (1 to 7)**
44
Original article
Table 4 Recommended route of administration for vitamin K
prophylaxis in the UK and Irish Republic (IRE), in 2003 and 2008
Year of survey
IM
Oral
Choice*
Total
responding
Total
units
2003 number of
units in UK (% of
responders)
2008 number of
units in UK (% of
responders)
2008 number of
units UK+IRE (% of
responders)
136 (59)
56 (24)
38 (17)
230 (95)
243
151 (69)
19 (9)
47 (21)
217 (100)
217
169 (72)
19 (8)
48 (20)
236 (100)
236
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).
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 ve did not
respond.
Jaundice
Of the 203 units that responded, 157 (77%) assessed infants
jaundiced at 23 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.
DISCUSSION
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
Table 5 Oral prophylaxis regimens for exclusively breastfed infants
Regimen
Konakion MM* 2 mg3 doses
Konakion MM 2 mg >3 doses
Konakion MM 1 mg3 doses
Orakay 1 mg at birth, then Neokay drops 50 mg daily for
3 months
Konakion MM 1 mg at birth, then Neokay drops 50 mg daily
for 3 months
Orakay 1 mg at birth, then Orakay 1 mg weekly for 8 weeks
Total
No of units
(%)
43
9
9
4
(64)
(13.5)
(13.5)
(6)
Table 6
Regimen
No of units (%)
31 (46)
20 (30)
1 (1.5)
10 (15)
5 (7.5)
67
*Licensed medicine by the MHRA in this dose regimen for bottle-fed infants.
MHRA, Medicines and Healthcare Product Regulatory Agency.
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 signicant 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 efcacy; 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 diseaseone
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 indenitely, 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.1618 Furthermore, with Konakion MM (ve times
more concentrated than Konakion Neonatal), the risk of
inaccuracy in administering the small volume required, especially for the smallest babies, is substantial.
1 (1.5)
1 (1.5)
67
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 prophylaxisalthough we
45
Original article
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 signicantly inuence
UK practice; the extent of their inuence 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 benet, NICE recommended IM administration as the most efcacious and costeffective 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 g IM not 1 mg. Oral Konakion MM was used off
licence for breastfed infants in 18 units (27%) and for bottlefed infants in 31 (46%). Another six (9%) used unlicensed products suitable for administration by parents.
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
46
CONCLUSION
Since the rst 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.
Correction notice This paper has been amended since it was published Online
First. It should now be noted that AB and RS are joint rst 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.
Buseld A, et al. Arch Dis Child 2013;98:4147. doi:10.1136/archdischild-2011-301029
Original article
REFERENCES
1
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
47
doi: 10.1136/archdischild-2011-301029
These include:
References
Email alerting
service
Receive free email alerts when new articles cite this article. Sign up in
the box at the top right corner of the online article.
Notes