Canadian Medical Association Journal 1996; 154: 347-349
The suggestion that the risk of childhood cancer is increased after intramuscular administration of vitamin K shortly after birth(3,4) was dismissed in the 1993 AAP recommendations.(2) Brousson and Klein also effectively dispel any concern that intramuscular administration is associated with an increased incidence of childhood cancer. It is seldom advisable to rely on reports from a single group to determine safety and efficacy -- especially reports of "inadvertent" discoveries, which have a greater risk of bias than other reports. Investigators often analyse multiple studies to determine efficacy; a similar analysis should be required to determine safety.
Brousson and Klein cite a review that reported no significant complications after 420 000 intramuscular injections of vitamin K to newborns.(5) They did not discuss the psychologic effects of intramuscular injections on babies and parents. At least one article has shown that pain experienced during the neonatal period may have long-term effects.(6) However, these effects may be less important than the benefits of routine vitamin K administration, which have been clearly shown.(7) The CPS recommendations aimed to obtain the benefit of vitamin K for newborns without incurring the pain.(1) They supported the oral route of administration of vitamin K with a formulation designed for parenteral use, a regimen reported to be practical and economical.(8)
To prevent early HDNB (which occurs during the first 24 hours of life), the CPS also recommended administering vitamin K to expectant mothers who are taking drugs that impair vitamin K metabolism.(1) Classic HDNB (occurring in the first week of life) is rarely seen when vitamin K is given to babies after birth.(7) Late HDNB (at 3 to 8 weeks of age), which occurs almost exclusively among babies who are breast-fed, is a concern in Germany and Britain as well as in Sweden and Australia.(9,10) In these countries the incidence of this uncommon, but serious, problem increased at the same time as an increase in the use of oral, rather than intramuscular, administration of vitamin K. On the basis of biochemical studies, this has led authorities in Australia to recommend repeated oral doses of vitamin K.(10)
By 5 days after oral administration of 1.0 mg of vitamin K at birth, protein induced by vitamin K absence-II (PIVKA-II) disappears.(11) However, by 4 to 6 weeks, biochemical signs of vitamin K deficiency are observed in up to 19% of babies given 2.0 mg of vitamin K orally at birth; by comparison, only 5.5% of those given 1.0 mg intramuscularly have biochemical signs of vitamin K deficiency.(12) Brousson and Klein mention the development of a mixed-micelle form of vitamin K, which would be better absorbed; however, a study showed that, even with this formulation, there is a greater incidence of vitamin K deficiency when the formulation is given orally than when it is administered intramuscularly.(13) However, the problem common to all of these studies is the poor clinical correlation of these biochemical indicators to abnormal bleeding in babies.
In analysing the effectiveness of vitamin K in preventing late HDNB, Brousson and Klein rely on an epidemiologic review by Loughnan and McDougall.(14) That review cited a similar German epidemiologic study by von Kries,(5) who showed that the failure rate (occurrence of late HDNB) after intramuscular administration was 0.25 per 100 000 infants, compared with a rate of 1.4 per 100 000 infants after oral administration. In three other countries in which oral administration is the primary form of vitamin K prophylaxis, the incidence rate of late HDNB was up to 6.4 per 100 000 infants.(5,14) (It should be kept in mind that some of these babies may have had underlying disorders that affected vitamin K metabolism.(15)) These incidence rates of late HDNB are similar to the rate of 1 per 100 000 infants in two Canadian hospitals, reported in the CPS statement.(1) Unfortunately, the specific incidence rate of late HDNB in Canada after oral or intramuscular administration of vitamin K cannot be determined because information is unavailable.
The relative risk of HDNB among infants after a single dose of vitamin K given orally versus intramuscularly has been estimated at 5.47 (95% confidence interval 0.5 to 65.8).(16) Although not statistically significant (the confidence interval includes 1.0), reduction of the rate of HDNB from 1 incident per 70 000 newborns to 1 incident per 419 000 newborns(16) may be of sufficient clinical significance to recommend intramuscular rather than oral administration of vitamin K. Repeated oral doses of vitamin K, as Brousson and Klein suggest, may be as safe and effective as a single dose administered intramuscularly after birth, but such a regimen may not be practical because of poor patient compliance.(17)
Brousson and Klein correctly point out that even intramuscular administration of vitamin K does not provide complete protection from HDNB. The incidence rate of HDNB among babies given vitamin K intramuscularly after birth is approximately 1 per 400 000 infants.(16) It is reasonable to consider administering further doses of vitamin K to infants at a high risk of HDNB: those who fail to thrive or have liver disease or long-term diarrhea. Physicians should also consider the possibility of vitamin K deficiency at an early stage in the evaluation of any bleeding that occurs during the first 6 months of life. Appropriate therapy with vitamin K should be instituted when required.
The large number of patients required to conduct a prospective study comparing intramuscular and oral administration of vitamin K (with or without repeated doses) makes it unlikely that such a study will be carried out. In the absence of such information, the review by Brousson and Klein provides an important, clinically significant analysis on which to base clinical decisions. The review should lay to rest concerns that there is a higher incidence of childhood cancer when vitamin K is given intramuscularly than when it is given orally or not at all. Furthermore, given the higher risk of late HDNB after a single dose taken orally than after a dose administered intramuscularly, and the 50% chance that babies with late HDNB will have an intracranial hemorrhage,(16) administration of vitamin K by the intramuscular route, as recommended by the AAP, seems most prudent. As the CPS reviews its recommendations in this area, mechanisms to determine outcome in relation to clinical practice should also be considered. This would help to better determine the effects of guidelines and analyses like the review by Brousson and Klein as we strive to improve outcomes for all children.