[back] Vitamin K injection

For the last 40 or 50 years, it has become a generally accepted fact that vitamin K prevents haemorrhagic disease of the newborn, and routine administration of vitamin K to all newborns has been recommended.^{3, 6, 21, 34, 72} This recommendation has been questioned because results released in 1990 from a study by Golding and colleagues²⁶ in the UK showed a two to three times increased risk of childhood cancers, especially leukaemia, in children given prophylactic drugs (usually intramuscular vitamin K) in their first week. A further study in 1992 seemed to confirm this risk.²⁵

Vitamin K is a fat-soluble substance which triggers off the blood-clotting process. Blood clotting is a complex process and can be described as a sequence of three stages, requiring up to 12 different coagulation factors.⁷² The liver needs vitamin K to synthesise four of these factors. Vitamin K is also needed for the formation of other proteins d in plasma, bone and kidney.^{33, 58}

As with other fat-soluble vitamins, a normal flow of bile and pancreatic juice is necessary for digestion, and the presence of dietary fat, especially short-chain fatty acids, enhances absorption. Absorbed vitamin K is transported via the lymph into the systemic circulation.⁵⁸

Normally, a significant portion (up to 55%) of absorbed vitamin K is excreted so the amount in the body is small and its turnover is rapid (about 30 hours).⁵⁸ Vitamin K is stored and re-utilised in the body for 3-4 weeks.³³

Vitamin K is found in many foods. Leafy, dark green and deep yellow vegetables are the best sources.⁵⁸ Alfalfa¹⁸ is a good source; and milk and dairy products, eggs, cereals, fruits and other vegetables also provide small but significant amounts. As the liver of adults contains about equal amounts of plant and animal forms of Vitamin K, it is assumed that vitamin K is produced in the intestinal tract by bacterial flora. One of the reasons given for the low levels of vitamin K in newborn babies is because their gut has not yet been colonised by the required

bacteria.

Haemorrhagic Disease of the Newborn (HDN) is a bleeding disorder associated with low levels of vitamin K in newborn babies. It was first defined in 1894 by Townsend⁶⁹ as spontaneous external or internal bleeding occurring in newborn infants not due to trauma, accident or inherited bleeding disorders such as haemophilia. Previously, there were no generally agreed upon criteria to determine causes of haemorrhaging, so any diagnosis was based solely on the opinion of the attendant medical personnel.

Infants are born with low levels of vitamin K²³ compared to adults and this is termed ‘vitamin K deficiency’. Up to 50% of babies develop this ‘vitamin K deficiency’, but bleeding occurs in only a fraction of these cases.³⁷ In most it starts after birth, becomes progressively more severe over 48-60 hours, then corrects itself spontaneously by 72-120 hours.⁹

HDN has always been rare – in Britain where maternity units practised a selective policy of vitamin K administration, the incidence was no more than 1 in 20,000 in the years 1972-80. Estimates for late onset HDN are 4-8 per 100,000.⁴⁵ Incidence also seems to vary from country to country.

HDN is divided into three categories:

There has been some debate over the years as to whether or not HDN is actually caused by vitamin K deficiency. Certainly, giving vitamin K does arrest bleeding in the majority of cases, but this does not mean that vitamin K deficiency causes HDN. One may as well say that an antibiotic deficiency causes bacterial infection. There is also no consensus as to what level of vitamin K in plasma protects against HDN. Some researchers have found no evidence of vitamin K deficiency in babies in their studies^{43, 49} and other factors have also been suggested.^{52, 73, 74}

Most, if not all, of the reported cases of late onset HDN have presented with problems which affect the baby’s ability to absorb or utilise vitamin K.^{45, 56} These include: hepatitis, cystic fibrosis, chronic diarrhoea, bile duct atresia, alpha-1-antitrypsin deficiency, coeliac disease of insufficient plasma transport capacity. Subclinical cytomegalovirus has also been implicated. Vitamin K-responsive bleeding syndrome has been well documented after antibiotic therapy, especially with cyclosporins.³³

Several authors have noted the higher incidence of HDN in solely breastfed babies.^{9, 30} The incidence has been quoted as 1 in 1200.³⁰ Studies comparing breastmilk with formula and cow’s milk have shown that breastmilk is lower in vitamin K.^{22, 28, 32} Breastmilk substitutes are heavily supplemented with vitamin K. However, it is possible that, like iron, vitamin K is biologically more available to the baby from breastmilk, and so such high levels are not necessary.

Measured levels of vitamin K in breastmilk seemed to vary depending on the type of measurement used; however, they all come out lower than cow’s milk. Fournier²² and Greer²⁸ found levels of around 8-9m g/l, which would mean that if a baby was taking in about 500ml per day, it would be getting the recommended 3-5m g daily.

Vitamin K content and availability are greater in the hind milk because of its higher fat content and vitamin K levels are also higher in colostrum.³² As an extra plus, breastmilk contains thromboplastin, one of the factors in blood clotting.¹⁸

Vitamin K levels in the breastmilk rise markedly in response

to the mother eating vitamin K rich foods or taking vitamin K supplements.^{29, 54} Nishiguchi found no cases of low vitamin K levels in breastfed infants whose mothers had been given supplements, as opposed to infants who had only been given 1 or 2 doses of oral vitamin K.⁵⁴

Unrestricted access to the breast in the early days after birth is important, due to the higher levels of vitamin K in colostrum. The importance of early feeding has been recognised since the 1940’s. Babies who have been fed within their first 24 hours have significantly better coagulation times than babies not fed until after 24 hours.²⁴

The search for the cause of HDN began in 1913 when Whipple⁸² postulated that a lack of prothrombin activity could be a cause of HDN. In 1929, Henrik Dam¹⁴ noticed that chicks fed a fat-free diet suffered subcutaneous and intramuscular haemorrhages, which could be prevented if the chicks were fed seeds, cereals and green, leafy plants. Dam described the condition as a vitamin deficiency and named the deficient vitamin ‘vitamin K’, from the Danish word ‘koagulation’.

In 1939, vitamin K₁ was isolated from alfalfa by Dam, for which he later received the Nobel Prize, along with Edward Doisy, who isolated vitamin K₂.⁴⁵ Further research in 1939 by Waddell and Guerry⁸¹ found that low plasma prothrombin levels could be elevated by the administration of oral vitamin K.

Armed with this ‘proof’ that vitamin K deficiency caused HDN, vitamin K was synthesised and various trials were commenced to ascertain which was the most effective amount and route to use in prophylaxis.

It is difficult for us to assess these trials nowadays as they were mostly neither double blind nor well controlled. The dosage of vitamin K given, the route of administration and the time of administration all varied. In many cases, the conclusions did not seem to match the results.⁷²

Because of the variations in results from these early studies, further research focused on treating the baby after birth. One particular study done in 1942³¹ was intended to determine the minimal effective oral dose of Synkavite (K₃), a water-soluble synthetic form of vitamin K. The results showed that very small daily doses were effective and that a dose of 5m g daily would probably prevent the development of HDN, except in early onset cases. The study also found that 1.25mg was effective in lowering an excessively high prothrombin time to normal. However, the author admitted that several workers found prothrombin deficiencies in babies with no abnormal bleeding.

By 1950, most maternity units had a policy of giving infants oral vitamin K (usually Synkavite) immediately after birth.⁷⁰ This prevented the fall in prothrombin levels that occurred in the first few days and, presumably, the risk of excessive bleeding. This risk was higher in male babies because of routine circumcision, and, indeed, vitamin K proved to be of great clinical value in preventing post-circumcision bleeding.⁷⁵

Then, in the mid-1950’s, reports of increased jaundice and kernicterus (brain damage caused by high bilirubin levels) in infants given vitamin K prophylaxis began circulating. These reports caused a major scare. Reviews of maternity units found that some were giving Synkavite in doses exceeding 50mg.⁷⁰ It was established that high doses of Synkavite caused haemolysis (destruction of red blood cells) and high serum bilirubin levels.⁴⁸

Researchers and medical professionals queried the safety aspects of vitamin K, and there were many conflicting reports on the appropriate dosages. Some researchers queried the need for vitamin K at all, quoting results from studies that showed no difference in prothrombin times or vitamin K plasma levels between babies that bled and babies that didn’t.⁷²

Eventually, a newer preparation, intramuscular vitamin K1 (phytomenadione), was developed and approved for use, solely on the grounds that it appeared to cause less haemolysis. Phytomenadione (trade names Konakion (Roche) or Aquamephyton (Merck, Sharpe & Dohme)) is a synthetic petrochemical derived from 2-methyl 1,4-naptha-quinone in a polyethoxylated castor oil base.¹⁸ In the US, polysorbate-80 is used as a base instead of polyethoxylated castor oil.¹⁵

In spite of there being no long term trials of these prepara-

tions, the American Academy of Pediatrics recommended that phytomenadione be administered prophylactically to all newborn babies.⁷² The use of oral vitamin K preparations fell out of favour in the USA and the ‘safer’ intramuscular route became the route of choice.

In Britain, after the jaundice scare of the1950’s, many maternity units began to practice a selective policy, giving vitamin K only to babies at risk of haemorrhaging. McNinch reported in 1980 that less than half the maternity units in the UK gave vitamin K to all newborns.⁴⁷ Some of these babies were given oral prophylaxis and some were given intramuscular prophylaxis.

In Germany, almost all newborn infants who required medical care and instrumental deliveries were given intramuscular vitamin K, and some healthy newborns also received it.⁷⁶ Records have not always been kept in New Zealand hospitals, so it is impossible to say whether or not vitamin K was given routinely and by which route.¹⁷

Although vitamin K use seemed to prevent most cases of HDN, there was still controversy. Not everyone believed vitamin K deficiency was the cause of HDN. In 1977, van Doorm et al ^{52, 73, 74} suggested that HDN could be caused by a heparin-like inhibitor in the newborn and he concluded that babies given their first feed soon after birth do not have a vitamin K deficiency. Other researchers agreed with van Doorm.⁴⁹ In 1980, Malia et al⁴³ could find no evidence of vitamin K deficiency in babies in their study and concluded that low levels of vitamin K dependent clotting factors were due to the immature liver. The authors of these studies questioned whether vitamin K prophylaxis was really necessary for healthy newborns.

Late onset HDN was first reported in 1977.⁵ It mainly occurs in breastfed infants and ¯ to ¾ of cases have an underlying liver disorder or malabsorption syndrome,¹⁵ rather than insufficient dietary intake of vitamin K. This means the liver cannot adequately synthesise blood clotting factors or store adequate amounts of vitamin K. Liver function cannot be easily diagnosed at birth without a range of invasive tests and thus there exists an unknown risk of haemorrhaging.

Many factors contribute to poor liver function, including hepatitis, cystic fibrosis, antibiotic therapy, biliary atresia, alpha-1-antitrypsin deficiency, a-beta-lipoproteinaemia, coeliac disease, chronic diarrhoea and exposure to pharmacologic agents such as anticonvulsants, rifampin, isoniazid cephalosporins and coumarin compounds.³³ When tested, most of the reported cases of late onset HDN had hepatitis, liver malfunction or enzyme deficiencies.^{6, 35, 51, 80}

Birkbeck⁶ also reports that HDN is almost unknown in central Africa and he suggests an environmental mechanism as the cause. Associated with this, a discussion paper from the University of Amsterdam⁴² raises the idea that by-products of our industrial society such as PCBs, PCDDs and PCDFs are the cause of late onset HDN. These chemicals can induce enzymes in the liver which cause liver damage and prolong prothrombin time. Although overseas studies have reported contamination of breastmilk by these pollutants, a NZ Department of Health studyon breastmilk reported that levels of these contaminants were at the lower end of the scale.⁷ The Health Department is currently conducting another study to see if levels have changed over the past few years.

There seems to be a seasonal variance, with most cases of late onset HDN occurring in the warmer months.⁶ It has been suggested that the mother could have contracted a viral infection during pregnancy in the colder months and this has crossed the placenta. Since viruses have an affinity for the liver and mucous membranes, they can affect intestinal absorption and liver function.⁶⁷

Another suggested cause of late onset HDN includes use of the food antioxidant BHT (butylated hydroxytoluene), which has produced vitamin K deficiency.⁶⁸ BHT is present in many processed foods, including margarine. Our Western diets consist of a lot of processed food, and to reduce fat intakes, margarine is recommended rather than butter. The polyunsaturated fat in margarine is an inhibitor of vitamin K absorption.⁶⁸ Both of these factors could have an effect on the amount of vitamin K available to pass through to the baby. A high level of vitamin K in the mother’s blood is necessary to ensure adequate transplacental transfer of vitamin K.^{9, 33} It is important for the baby to have adequate stores of vitamin K in its liver at birth to prevent bleeding until its feeding and gut flora are established.

Of the six cases of HDN in Britain in 1980-1982, all were breastfed and none had received vitamin K at birth.⁴⁶ Two of the cases were in the high-risk group – one was born by caesarean section and had an epileptic mother treated with phenytoin, and the other had an alcoholic mother who had taken anti-depressants – and obviously should have received vitamin K at birth.

These cases prompted a call for the re-introduction of routine prophylaxis. Many opposed the idea of unnecessarily injecting otherwise healthy babies so studies^{40, 47, 55, 79} were therefore conducted to determine whether oral vitamin K was as effective as intramuscular. It was also proposed that oral vitamin K would be more cost-effective and thus better suited for use in Third World countries.⁵⁵ Results of these studies varied. Some showed that oral vitamin K was effective in preventing classic haemorrhagic disease but not as effective as intramuscular vitamin K in preventing late onset HDN.^{47, 55, 78} Others found oral as effective, especially a 10 year study conducted on 38,000 infants in Sweden where no cases of HDN were observed over that period.⁴⁰ Tripp and McNinch reported no cases in 25,000 babies in their maternity unit where only those at risk were given intramuscular prophylaxis and the rest oral prophylaxis.⁷⁰

In spite of these findings that oral vitamin K prophylaxis was not effective in preventing late onset HDN, it continued to be used in British maternity units, especially for low risk infants.

The authors of the study approached Roche, the manufacturers of Konakion, for funding for a further trial to examine the findings more closely. Roche was not interested until, a few months later, the media reported the results of the study and that vitamin K given to babies might cause childhood cancer. Roche then decided to fund a new study.²⁷

The new study²⁵ was a case-control study of 195 children with cancer born at either of two hospitals in Bristol, England, compared with 588 healthy children also born at these hospitals. One hospital predominantly gave vitamin K orally and the other intramuscularly. The authors found a nearly two-fold risk of leukaemia in children who had received intramuscular vitamin K.

These findings were extremely worrying. Golding calculated that the extra cases of leukaemia caused by vitamin K injection could be as many as 980 in the UK alone.²⁵ These results were supported by reports of the potential carcinogenicity of vitamin K from Israels et al, who suggested that low vitamin K levels in the newborn protect against the risk of mutations during a period of rapid cell growth and division.³⁹ Pizer et al did not find any association between the route of vitamin K administration and mutations in cells but concluded that his study was too small to show any real effect.⁶² Another study reported no increase in abnormalities in newborn infants, but, with only 12 infants, the study was too small to show any real effect.¹⁰ It is worth noting that after an intramuscular dose of vitamin K, the baby’s plasma levels are almost 9000 times the normal adult levels.⁴⁷ It has also been suggested that the cancer-causing agent could be a metabolite, N-epoxide, or some other component of the solution other than vitamin K itself.¹⁵

Golding’s study was criticised by many. One of the reasons was that the authors had to make assumptions for some cases, as the information on vitamin K administration was not clearly recorded. In spite of this, expert epidemiologists considered that the results were plausible and so could not be lightly dismissed.¹⁵ Further studies were proposed to answer the question of cancer and vitamin K.

In 1993, results from three retrospective studies on vitamin K and childhood cancer were published. The studies were done in the USA, Denmark and Sweden.^{41, 57, 19} These studies, although large, did not confirm the association between intramuscular vitamin K and childhood cancer. One of the studies not only showed no association between IM vitamin K and childhood cancer, it also showed no association between maternal smoking and childhood cancer, a finding totally at odds with the results from many other studies.¹⁹ The other two studies were also not comparable to the British study. One because of differences in type of vitamin K given⁴¹ and the other because of the use of birth cohorts with differing regimens of vitamin K usage.⁵⁷

Because of the design flaws in these studies, there was still a need for further case-control studies. Results from two were published in 1996.^{2, 77} They had carefully matched controls and more accurate information on whether vitamin K had been given or not, and by which route. One of the studies² reported no association between intramuscular vitamin K and childhood cancer and the other⁷⁷ found a risk of leukaemia, but only when cases were compared with local controls (i.e. from the same hospital) and not with controls randomly selected from the whole area under study. This, although suggestive, was not followed up but dismissed as a chance finding related to multiple testing.

The suggestion was then put forward that, as these studies had failed to show a definite association between intramuscular vitamin K and childhood cancers, worries about any potential cancer risk should be abandoned.⁸³

At that time, four more studies on vitamin K and cancer were in progress.^{44, 59, 60, 61} The results from these four studies were published in 1998. Two of them failed to confirm any increased risk of childhood cancers.^{44 61} One of the other studies showed a twofold risk of acute lymphoblastic leukaemia among 1-6 year olds,⁵⁹ the other showed a significant risk for all cancers.⁶⁰

After the publication of Golding’s studies, further trials were done on oral vitamin K prophylaxis and whether it gave longer term protection. In 1992, Cornelissen¹¹ found plasma vitamin K concentrations were higher in the group given IM vitamin K than the oral group, but blood coagulability, activities of factors VII, X and PIVKA-II concentrations showed no differences. By 3 mths follow-up, vitamin K levels had dropped in both groups but more in the oral group. He suggests that neither give long term protection. One would assume that babies should be producing their own vitamin K by 3 months and, if not, what other mechanism could be hindering this process.

Von Kries et al⁷⁸ studied repeated oral vitamin K prophylaxis in Germany, with 3x 1 mg doses and found that it was not as effective as a 1mg intramuscular dose at birth. Another study by Cornelissen et al¹² reported on the effectiveness of differing regimens of oral vitamin K in four different countries – the Netherlands, Germany, Switzerland and Australia (two differing regimes). In the Netherlands, babies are given 25 m g daily oral vitamin K for 3 months with I mg given at birth either orally for healthy newborns or intramuscularly for unwell babies. In Germany, the regime is 3 x 1 mg oral doses as was also the case in Australia from 1993 to 1994. In Switzerland 2 oral doses of a new ‘mixed-micellar’ oral vitamin K is given. The Netherlands had the lowest failure rate – 0 per 100,000. In Australia, where the regime was changed in 1994 from oral to IM, the failure rate was 1.5 per 100,000 for oral and 0.9 per 100,000 for IM, showing that 3 oral doses are less effective at preventing late onset HDN than one IM dose of vitamin K. The results from Switzerland (failure rate 1.2 per 100,000)¹² show that further study needs to be done on the most effective timing of the dose.

It would seem an anachronism that babies are born with a ‘deficiency’ of such an essential vitamin and require supplementation. It also begs the question of why we assess babies vitamin K levels against adult levels. We do not assess other factors (e.g. weight, hormone levels, etc) in this way. In fact, although there have been many studies on differing aspects of vitamin K prophylaxis, there seem to have been few³⁹ on the possible reasons for and the advantages (if any) of the physiological levels of vitamin K in newborns.

The risks of prophylaxis for the majority of babies who are at low risk of HDN are also not understood. As plasma vitamin K levels in newborns reach 300 times normal adult levels for oral and almost 9000 times for IM vitamin K⁴⁷, some research needs to be done on the effects this may have. Studies have shown that physiological levels of vitamin K maintain a careful balance between coagulation and anti-coagulation⁶⁵ and we have no idea what the effects of upsetting that delicate balance would be.

The number of children currently developing cancer during childhood is much higher than the number developing a life threatening or permanently disabling problem as a result of late onset HDN. The risk of childhood cancer is estimated to be 1.4 per 1000, from the 1970 British cohort. If IM vitamin K caused cancer, there would be 100 extra cases of cancer per case of HDN prevented.¹⁶ This could mean that giving IM vitamin K to every baby would be doing more harm than good.³⁶