‘There are known knowns; these are things we know that we know. There are known unknowns; that is to say there are things that we now know we don’t know. But there are also unknown unknowns – there are things we do not know, we don’t know.’
Donald Rumsfeld, 2002
Just over a hundred years ago, rickets posed a frequent challenge in obstetric practice with reports of approximately three per cent of deliveries in Scotland requiring manoeuvers, including the revolutionary procedure of caesarean section, to deal with a contracted rachitic pelvis.1 The identification of a steroid compound able to prevent rickets led to the award of the 1928 Nobel Prize for Chemistry to Adolf Windaus. As the fourth ‘vital mineral’ to be identified, it was labelled ‘D’. Its biochemistry is now understood in great detail and the vital roles in calcium homeostasis and prevention of rickets and osteomalacia are well characterised. Encountering a rachitic pelvis is now rare, but a recent report confirms that childhood rickets has not disappeared from Australia, with over 850 cases notified annually.2
More recently, non-skeletal actions of vitamin D have been identified and over 900 genes are now recognised to be activated through the ubiquitous vitamin D receptor. It appears that vitamin D is vital to the adequate functioning of the immune system and deficiency is associated with increased risks of multiple sclerosis, rheumatoid arthritis, diabetes and an expanding list of cancers. A recent meta-analysis involving over 50 000 subjects found that all-cause mortality is reduced by seven per cent in vitamin D supplemented groups.3 Such data have led to a general perception that vitamin D is a ‘cure-all’. This thinking has also crept into obstetric practice, following the realisation that deficiency is linked to pregnancy complications such as preeclampsia. It is, however, important to remember the distinction between association and causation and to ensure there are no unexpected consequences on neonatal development before widespread supplementation can be endorsed.
Biochemistry of vitamin D and defining deficiency
Most vitamin D is derived from sunlight-stimulated conversion of provitamin D in the skin. The ability to synthesise vitamin D from sunlight is reduced in those with darker skin and at latitudes moving away from the equator. About ten per cent of vitamin D may be derived from the diet, in particular from certain fish (salmon, herring and mackerel) as well as the result of margarine supplementation. Subsequent liver metabolism results in the production of 25-OH vitamin D (25(OH)D), the index of vitamin D status in serum assays. Thereafter, activation occurs mainly in the kidney, although this process is now recognised to occur in every organ system in the body; with the placenta the most prominent site of extra-renal activation of 25(OH)D in pregnancy.4
Over the last decade, the 25(OH)D level that is thought to represent a healthy level of vitamin D has increased. Somewhat confusing terminology abounds, with various definitions for adequacy, insufficiency and deficiency. A recent Australian multidisciplinary position statement defines deficiency in adults as:
- Mild vitamin D deficiency: 30–49nmol/l
- Moderate vitamin D deficiency: 12.5–29nmol/l
- Severe vitamin D deficiency: <12.5nmol/l
It is worth noting the caveat that seasonal variation is best managed by targeting a level 10–20nmol/l higher at the end of summer to compensate for the natural decrease in most areas in winter. The statement also reiterates concerns over the reliability of some assays for 25(OH)D level, which are addressed by ensuring laboratories participate in external quality assurance programs.5
A pregnancy and childhood consensus group position statement is currently in preparation and seeks to recommend the same levels (personal communication). While some authorities, including the American Academy of Pediatrics, suggest pregnancy levels over 80nmol/l to be desirable, there is insufficient evidence of safety at the present time to support this target.
Vitamin D deficiency in pregnancy
There is a strong relationship between maternal and fetal/neonatal vitamin D levels such that deficient mothers are more likely to give birth to deficient neonates and extreme deficiency may even cause rickets in utero. Breastmilk is a poor source of vitamin D, with breastfed babies of deficient mothers being considerably more likely to become deficient than formula-fed babies. Therefore, ensuring adequate vitamin D levels in pregnant women is important, particularly among darker skinned and veiled women, who are often severely deficient and also over-represented in the groups whose children develop rickets. Studies from Melbourne, Sydney and New Zealand have identified between 50 and 94 per cent of women from these high-risk groups to be severely vitamin D deficient. Therefore recommendations to routinely screen and supplement these high-risk women and to supplement their children should be followed.6
Interestingly, recent reports have shown high levels of deficiency among groups traditionally thought to be at low risk. Thirty five per cent of pregnant women in Northern Victoria and in Canberra and 25 per cent of pregnant women in Campbelltown, New South Wales, have been found to be vitamin D deficient, many of these women being Caucasian.7,8 Given that obesity is a risk factor for low vitamin D, it is possible that the rapid rise in obesity levels explains some of the findings. Whatever the explanation, these are prevalence rates of a deficiency state with potentially serious neonatal consequences that are higher than any of the other conditions that are routinely screened for antenatally. Not all studies, however, have reported such widespread deficiency with a recent report from Far North Queensland, on 116 women, finding no women to have a vitamin D level below 50nmol/l.9 This finding makes a single Australia-wide recommendation on screening for vitamin D deficiency difficult.
Non-skeletal associations with vitamin D deficiency
Heightened interest in the detection and prevention of pregnancy vitamin D deficiency has resulted from the increasing list of associated diseases. Adding weight to the role of vitamin D in immunity has been the discovery that maternal hypovitaminosis D increases neonatal and childhood risk of eczema, asthma and wheeze as well as viral infections, including respiratory syncytial virus and bronchiolitis.
In obstetrics, a 2007 paper identifying an association of low vitamin D with preeclampsia was met with great excitement. Not only did the discovery potentially explain the long-recognised increased risk in African-Americans, whose darker skin is associated with lower vitamin D levels, but there was also a discernable ‘dose-response’ effect with a quantifiable reduction in preeclampsia risk with increasing 25(OH)D.10 To add to the tantalising possibility that an easily preventable vitamin deficiency might reduce the risk of a significant cause of perinatal mortality, came data linking vitamin D deficiency with increased risk of preterm labour, gestational diabetes and even caesarean delivery.4 As obstetricians, had we found the holy grail?
Routine testing and supplementation in pregnancy
Despite the excitement, the potential for significant reduction in pregnancy complications by vitamin D supplementation is, currently, unproven. A 2012 Cochrane review, utilising data from randomised controlled trials involving 1023 women, was not encouraging.11 While supplementation did increase maternal levels of 25(OH) D and there was no difference in overall incidence of adverse effects between groups, the preeclampsia risk was not significantly reduced (RR 0.67; 95 per cent CI 0.33 to 1.35) and there was only a borderline benefit in terms of reducing low birth weight (RR 0.48; 95 per cent CI 0.23 to 1.01).
So, where to now? It is difficult to dismiss the very strong associations of reduced pregnancy risks with increased levels of vitamin D. However, if there is the possibility of pregnancy benefit, but it is currently unproven, are there disadvantages and dangers to testing and supplementation that warrant a cautious approach?
The economic burden of the sudden rise in 25(OH)D assays in the context of unproven benefit has been questioned. If the aim is the prevention of childhood skeletal complications the potential benefit/risk ratio of routine testing will be influenced by prevalence; given the finding that deficiency is rare in Far North Queensland, routine screening is unlikely to be economically sensible. With much higher rates of deficiency in Victoria, Canberra, New South Wales and New Zealand, there may be more benefit. An alternative approach would be to consider routine supplementation rather than testing. This would avoid the laboratory costs, but would risk supplementing some women with already adequate vitamin D levels. It is possible with oral supplementation to raise serum 25(OH)D to potentially toxic levels and there have been reports of increased rates of eczema and asthma in offspring of women with 25(OH)D levels over 75nmol/l. This finding has recently been refuted, but the potential for a U-shaped benefit and risk curve remains.12 This warrants a cautious approach. Interestingly, sunlight exposure cannot lead to toxic levels and so skin exposure advice may be appropriate before conception and antenatally.5
On the basis of current evidence, screening is best undertaken in the context of risk factors such as dark skin colour, limited sun exposure, raised body mass index and a local assessment of prevalence. Supplementation should be undertaken to target a 25(OH)D level of 50nmol/l in the hope that this will prevent neonatal bone complications. Current recommendations support supplementing with 600–1000IU daily, but the utility of these doses has been questioned. Recent reports suggest that 4000IU/day is safe but, once again, the longer term consequences are unclear and so caution is necessary as the effects are clarified.13
Vitamin D and gynaecological cancers
There is potential for vitamin D to offer benefits in gynaecology.14 The wide range of cancers associated with vitamin D deficiency include breast, ovarian and endometrial. Teasing out the association is complicated by the recognised link between obesity and both hypovitaminosis D and cancer, but some studies suggest strong links even after accounting for body mass index. There is also evidence to suggest that survival after a diagnosis of breast or vulval cancer is higher in vitamin D replete individuals. Recent animal data have shown that vitamin D dramatically reduces the development of endometrial cancer in obese mice.15 Human trials are ongoing.
There is the potential that the correction of vitamin D deficiency could improve pregnancy outcomes, reduce gynaecological cancers and reduce a long list of autoimmune diseases. Unfortunately, these attractive possibilities must be counterbalanced by the, as yet, uncertain potential for unexpected adverse effects. There is good evidence that preventing maternal vitamin D deficiency is important for the prevention of neonatal rickets, but the hope that significant pregnancy complications can be prevented is unclear. This is a developing field, but at present a degree of caution is required before we can recommend routine supplementation without question. Primum non nocere.
I would like to thank Dr Claire McKie and Mrs Julie Jones for their help with proofreading the manuscript.