Overview of vitamin D
Role of vitamin D
The role of vitamin D in skeletal health is robust; the major biological function of the active form of vitamin D (1,25(OH)2D) is to maintain serum calcium and phosphorus homeostasis, essential for bone mineralisation( Reference DeLuca 1 ) and neuromuscular function( Reference Holick, Binkley and Bischoff-Ferrari 2 ). The discovery that the vitamin D receptor is expressed in virtually all cells in the human body( Reference Norman 3 ) has led to the recognition that vitamin D may have a role to play in many more processes than previously thought. This concurs with the growing body of observational data showing associations between 25-hydroxyvitamin D (25(OH)D) concentrations and chronic diseases and conditions such as CVD, diabetes and colorectal cancer( Reference Autier, Boniol and Pizot 4 ). However, whether low 25(OH)D concentrations are a cause or effect of such conditions has not been established( Reference Autier, Boniol and Pizot 4 ).
Sources of vitamin D
Vitamin D is the generic term used for both vitamin D2 and vitamin D3, also known as ergocalciferol and cholecalciferol, respectively. Vitamin D2 is synthesised by the exposure of ergosterol in plants to UVB radiation, whereas vitamin D3 is synthesised in the skin of human subjects and animals by the action of UVB radiation.
There are therefore two sources of vitamin D; diet and exposure to sunlight. In human subjects, the action of direct sunlight, containing UVB radiation of wavelengths 290–315 nm, on skin results in synthesis of vitamin D. The first step in this synthesis is the conversion of 7-dehydrocholesterol to pre-vitamin D3, which is then converted to vitamin D3 by a temperature-dependent isomerisation reaction.
Naturally occurring dietary sources of vitamin D include both vitamin D2, found in plants and fungi, and vitamin D3, found in meat, fish and eggs. However, few naturally occurring food sources of vitamin D are considered a rich source of vitamin D. The few foods considered a good source of vitamin D are mostly of animal origin and therefore contain vitamin D3 such as oily fish, egg yolks and meat, liver and kidney.
Other sources include supplements or fortified foods, which will be discussed later in this review. Both vitamins D2 and D3 can be commercially synthesised for use as supplements and to fortify foods, by UVB irradiation of ergosterol from plants and fungi and 7-dehydrocholesterol from sheep's wool, respectively.
In most countries, exposure of skin to UVB radiation or supplement use are the main source of vitamin D and in the UK skin synthesis is the primary source. The latest UK National Diet and Nutrition Survey data collected in 2012/13 and 2013/14 showed mean dietary intakes of 3·1 and 2·5 µg/d in adult men and women, respectively( Reference Bates, Cox and Nicholson 5 ), although lower dietary intakes of 1·25–1·6 µg/d have been reported for South Asian women in the UK( Reference Macdonald, Mavroeidi and Fraser 6 ).
Global vitamin D status and recommendations
Rates of vitamin D deficiency
Data from across the globe have shown that vitamin D deficiency is a worldwide issue. Although different cut-off points have been used to define deficiencies, these data are based on 25(OH)D concentrations below 25 nm/l. Studies have reported the prevalence of deficiency to range from 2 to 30 % across Europe( Reference Spiro and Buttriss 7 ), 38 to 80 % in women of child-bearing age in the Middle East( Reference Bassir, Laborie and Lapillonne 8 , Reference Molla, Al Badawi and Hammoud 9 ), and 3·5 % in the USA, although when ethnicity was considered the prevalence of deficiency were highest among non-Hispanic black individuals at 15 %( Reference Jain 10 ).
In the UK, the latest National Diet and Nutrition Survey( Reference Bates, Cox and Nicholson 5 ) reported that 22 % of men and 15 % of women aged 19–64 years had total 25(OH)D concentrations below 25 nm/l all year-round. Furthermore, seasonal variation in 25(OH)D concentrations exists( Reference Hyppönen and Power 11 ); previous National Diet and Nutrition Survey data reported rates of deficiency increasing to 39 % in men and women between January and March( Reference Bates, Lennox and Prentice 12 ). Sub-groups of the population, particularly those considered at-risk of vitamin D deficiency, have shown even higher rates of deficiency; a longitudinal study (D-FINES) looking at dietary and sunlight contribution to seasonal vitamin D status in South Asian and Caucasian women of child-bearing age reported that 51·4 % of the South Asian women had 25(OH)D concentrations <25 nm/l in the summer months, and this rose to 64·5 % in the winter( Reference Darling, Hart and Macdonald 13 ).
Vitamin D dietary recommendations
In July 2016, the Scientific Advisory Committee on Nutrition published the new UK recommendations for vitamin D( 14 ). Based on the health outcome of musculoskeletal health, the Scientific Advisory Committee on Nutrition concluded that the estimated average nutrient intake required to maintain 25(OH)D concentrations at or above 25 nm/l in winter for 97·5 % of the population aged between 11 years and older is 10 µg/d, and a reference nutrient intake (RNI) of 10 µg/d was set for everyone aged 1 year and above. Prior to this update, there was no RNI for vitamin D set for those aged 4–64 years( 15 ), unless the individual was considered at-risk of deficiency. This was based on the assumption that sufficient vitamin D was made, through skin synthesis, and stored during the summer months to sustain 25(OH)D concentrations during the winter months, which is now known not to be the case( Reference Hyppönen and Power 11 , Reference Darling, Hart and Macdonald 13 ).
This new RNI brings the UK in-line with other European and international recommendations being previously the only country in Europe without a dietary recommendation for all aged 4–6 years( Reference Spiro and Buttriss 7 ).
Vitamins D2 and D3
The chemical structures of vitamins D2 and D3 are similar but not identical; vitamin D3 has an additional double bond and methyl group, and it is postulated that this different structure results in vitamin D3 being the preferred substrate at several stages of the pathway of metabolism of vitamin D.
Both vitamins D2 and D3, irrespective of source, undergo the same metabolism process from the venous circulatory system, a two-step hydroxylation process. Firstly, vitamins D2 and D3 are transported by the vitamin D-binding protein to the liver where they are converted to 25(OH)D2 and 25(OH)D3, respectively, by the action of 25-hydroxylase. These are then transported, again by the vitamin D-binding protein, to the kidneys where they are converted to the biologically active form of vitamin D (1,25(OH)2D), by the action of 1-α-hydroxlyases.
There is data to suggest that the differences in the side chains of the two forms of vitamin D directly affect the rate of vitamin D hydroxylation at the liver, with vitamin D3 thought to be the preferred substrate for hepatic 25-hydroxylase( Reference Holmberg, Berlin and Ewerth 16 , Reference Cheng, Levine and Bell 17 ). Vitamin D3 and its metabolites also have a higher binding affinity to the vitamin D-binding protein than vitamin D2 ( Reference Houghton and Vieth 18 ). In addition to these metabolic differences between the two forms of vitamin D, the degradation of vitamin D3 requires an addition step/process to that of vitamin D2, which suggest that the degradation rate of vitamin D2 may be higher than that of vitamin D3 ( Reference Horst, Reinhardt and Ramberg 19 ).
Therefore, there are several biologically plausible mechanisms by which vitamin D3 may have a greater capacity than vitamin D2 to raise and maintain 25(OH)D concentrations, as reviewed by Houghton and Vieth( Reference Houghton and Vieth 18 ), although randomised-controlled trial (RCT) data have not conclusively supported this theory, as discussed below.
Comparative efficacy at raising 25-hydroxyvitamin D concentrations
Historically, vitamins D2 and D3 were considered as equally effective at raising 25(OH)D concentrations( Reference Park 20 ). However, since the 1980s there have been a number of intervention trials, including RCT, published that have investigated specifically this, as shown in Table 1, and these have shown conflicting results. Although the majority of intervention trials comparing the two forms of vitamin D have provided data suggesting that vitamin D3 is superior to vitamin D2 in raising 25(OH)D concentrations( Reference Armas, Hollis and Heaney 21 – Reference Trang, Cole and Rubin 36 ), there have also been four trials that have produced data supportive of vitamins D2 and D3 being equally effective( Reference Biancuzzo, Young and Bibuld 37 – Reference Nimitphong, Saetung and Chanprasertyotin 40 ). There are no studies which have shown that vitamin D2 is more effective than vitamin D3.
Key: vit, vitamin; M, males; F, females; d, days; wk, weeks; mo, months; OJ, orange juice; 25(OH)D, 25-hydroxyvitamin D; IM, intramuscular.
In 2012, Tripkovic et al. conducted a systematic review and meta-analysis of RCT data comparing the efficacy of vitamins D2 and D3 in raising 25(OH)D concentrations( Reference Tripkovic, Lambert and Hart 41 ). Although ten studies were identified within the systematic review, only seven studies had sufficient and available data to be included within the meta-analysis. The primary analysis, of all seven studies regardless of dosing frequency, showed that vitamin D3 led to a greater absolute change in 25(OH)D concentrations than vitamin D2, with a weighted mean difference of 15·23 (95 % CI 6·12, 24·34; Z = 3·28; I 2 = 81 %; P = 0·001). To determine any confounding effect of dosing frequency, separate analyses were performed on: (a) studies giving bolus doses and (b) studies with daily supplementation. For bolus studies alone vitamin D3 remained significantly more effective than vitamin D2, with a weighted mean difference of 34·10 (95 % CI 16·39, 51·83; Z = 3·77; I 2 = 77 %; P = 0·0002). However, when the analysis was completed on the daily dosing RCT data alone, the differentiation between the two forms of vitamin D was moderated with a non-significant weighted mean difference of 4·83 (95 % CI −0·98, 10·64; Z = 1·63; I 2 = 41 %; P = 0·10).
Key limitations of the meta-analysis were identified and discussed by the authors. Firstly, there were few studies for inclusion within the analysis, and of the studies that were available these were small and unpowered with respect to study population size (n 19–89). There was also substantial between-study heterogeneity, with diverse intervention strategies, including varied doses of vitamin D, frequencies of supplementation and methods of administration, and all the studies used supplementation doses in excess of current recommendations( 14 ). Taking these factors into consideration, the authors concluded that far larger, more robust trials are required to not only measure 25(OH)D concentrations in response to vitamins D2 and D3, but also to explore potential mechanisms behind any differences seen.
Since the 2012 meta-analysis, there have been at least another ten further intervention trials comparing the efficacy of vitamins D2 and D3 in raising 25(OH)D concentrations( Reference Cipriani, Romagnoli and Pepe 23 , Reference Itkonen, Skaffari and Saaristo 26 , Reference Lehmann, Hirche and Stangl 27 , Reference Logan, Gray and Peddie 29 – Reference Oliveri, Mastaglia and Brito 31 , Reference Shieh, Chun and Ma 33 , Reference Stepien, O'Mahony and O'Sullivan 34 , Reference Fisk, Theobald and Sanders 38 , Reference Nimitphong, Saetung and Chanprasertyotin 40 ). However, findings remain equivocal with eight of the studies showing vitamin D3 to be more effective at raising or maintaining 25(OH)D concentrations compared to vitamin D2 ( Reference Cipriani, Romagnoli and Pepe 23 , Reference Itkonen, Skaffari and Saaristo 26 , Reference Lehmann, Hirche and Stangl 27 , Reference Logan, Gray and Peddie 29 – Reference Oliveri, Mastaglia and Brito 31 , Reference Shieh, Chun and Ma 33 , Reference Stepien, O'Mahony and O'Sullivan 34 ), and two of these trials reporting no significant difference between the two forms( Reference Fisk, Theobald and Sanders 38 , Reference Nimitphong, Saetung and Chanprasertyotin 40 ). However, a direct comparison between the total change in 25(OH)D concentrations in response to vitamins D2 and D3 was not reported in the analysis by Fisk et al. which returned neutral findings( Reference Fisk, Theobald and Sanders 38 ).
These more recent studies have addressed some, but not all, of the limitations identified from the 2012 systematic review and meta-analysis. Specifically, over half of the studies since 2012 examined supplementation doses closer to the range of current global recommendations. However, underpowered sample sizes remain an issue; the largest published RCT comparing the efficacy of vitamins D2 and D3 on raising 25(OH)D concentrations had a total of 107 subjects across three intervention groups( Reference Lehmann, Hirche and Stangl 27 ) as shown in Table 1. There remains a need for a large, robust RCT to provide more conclusive evidence in which confidence in results can be sought and the publication of the BBSRC Diet and Health Research Industry Club (DRINC) funded D2–D3 Study (BBSRC DRINC: BB/I006192/1, ISRCTN23421591); an RCT in n 335 healthy white Caucasian and South Asian women, should provide just that( Reference Tripkovic, Wilson and Hart 42 ).
Potential mechanisms from randomised-controlled trial data
Within some of these intervention trials the mechanisms by which vitamins D2 and D3 might lead to different effects on total 25(OH)D concentrations have been explored. Where methods such as LC-MS/MS for measuring 25(OH)D concentration are implemented, 25(OH)D2 and 25(OH)D3 concentrations are measured, which are added together to determine total 25(OH)D concentrations. In studies where these two metabolites have been measured, the vitamin D2 interventions have led to an increase in 25(OH)D2 concentrations and the vitamin D3 interventions have led to an increase in 25(OH)D3 concentrations( Reference Armas, Hollis and Heaney 21 , Reference Glendenning, Chew and Seymour 24 , Reference Lehmann, Hirche and Stangl 27 , Reference Logan, Gray and Peddie 29 , Reference Biancuzzo, Young and Bibuld 37 , Reference Fisk, Theobald and Sanders 38 ), as would be expected. However, a decreasing effect of vitamin D2 interventions on 25(OH)D3 concentrations has been noted in several intervention trials( Reference Armas, Hollis and Heaney 21 , Reference Binkley, Gemar and Engelke 22 , Reference Lehmann, Hirche and Stangl 27 , Reference Tjellesen, Hummer and Christiansen 35 ), although not all( Reference Glendenning, Chew and Seymour 24 , Reference Biancuzzo, Young and Bibuld 37 , Reference Fisk, Theobald and Sanders 38 ). The same has not been shown for vitamin D3; vitamin D3 interventions have not shown a decreasing effect on 25(OH)D2 concentrations; however, baseline 25(OH)D2 concentrations tend to be far lower at baseline (typically <5 nm/l) and so the opportunity for a decreasing effect is not available. Although even where mean baseline 25(OH)D2 concentrations were slightly higher (13·3 nm/l) no change was shown in the vitamin D3 intervention group( Reference Glendenning, Chew and Seymour 24 ).
This decline in 25(OH)D3 concentrations reported in those taking vitamin D2 could explain why vitamin D3 is more effective at raising total 25(OH)D concentrations, and although the exact mechanisms are unknown this could reflect either competitive binding for 25-hydroxylase or the vitamin D-binding protein, or changes in degradation rate as discussed previously. Further research is needed to elucidate the exact mechanisms and to understand the impact of these changes on overall health, and not just total 25(OH)D concentrations.
Across Europe, legislative and voluntary fortification policies and practices vary from country to country. In 1940, vitamin D fortification of margarine and fat spreads became mandatory in the UK and Ireland, although only to bring the vitamin D content up to the level naturally found in butter and not with the aim of improving population intakes. Currently most margarines and fat spreads are still fortified voluntarily despite the mandatory requirement being revoked in 2014( 43 ); however, it is important to note that although these would be considered fortified food, the amount of vitamin D added is 7·5–10 µg per 100 g, and thus the contribution to population intakes of vitamin D is minimal.
Vitamin D, in the form of either vitamin D2 or D3, is legally permitted to be added to foods on a voluntary basis (Annex 1 of Regulation (EC) No 1925/2006, amended by the Commission Regulation (EC) No 1170/2009), and foods which are most commonly fortified include breakfast cereals and more recently, non-dairy milks. However, as the amount of vitamin D added in to fortified products is low, fortified foods still contribute very little (0·8 µg/d) to the dietary intake of the UK adult population( Reference Bates, Lennox and Prentice 12 ).
It has been suggested that additional strategic approaches to fortification, including bio-fortification, of a wider range of foods, have the potential to increase vitamin D intakes in the population( Reference Cashman and Kiely 44 ). Bio-fortification is the process by which nutritional quality is enhanced through agronomic or modern biotechnology techniques, as opposed to being added manually at a later stage of product processing. A thorough review of food-based solutions for vitamin D deficiency has recently been published by Hayes and Cashman earlier this year( Reference Hayes and Cashman 45 ) and includes a review of the need for traditional fortification but also provides an overview of recent advances in the field of bio-fortification, which may have greater consumer appeal. To date, there are several methods of bio-fortification that have begun to be explored, including vitamin D3 enhancement in eggs and meat through addition of vitamin D to animal feeds( Reference Hayes and Cashman 45 ), and the use of UV radiation to enhance the vitamin D2 content of foods such as mushrooms, which has recently been examined in a systematic review and meta-analysis( Reference Cashman, Kiely and Seamans 46 ). Although the majority of these developments have proved successful at improving vitamin D status in RCT, one recent RCT, feeding bread baked with UV-treated yeast, resulted in no significant change in 25(OH)D concentrations, despite the vitamin D2 content of the baked bread being confirmed by HPLC( Reference Itkonen, Skaffari and Saaristo 26 , Reference Lipkie, Ferruzzi and Weaver 47 ). This raises concerns about the bio-accessibility of these bio-fortified foods and highlights the need for human RCT data to provide proof of efficacy and safety prior to products reaching the market.
There are two key projects that have, and continue to, significantly contribute to the evidence base for the potential role of vitamin D fortified foods: (1) the Optimal Fortification with vitamin D (OPTIFORD; www.optiford.org) European project investigated the feasibility of fortification as a strategy for improving vitamin D status; among their findings they concluded that bread was a safe and feasible vehicle for fortification( Reference Natri, Salo and Vikstedt 48 ); (2) the food-based solutions for optimal vitamin D nutrition and health throughout the life cycle project (ODIN; www.odin-vitd.eu), an EU funded project consisting of a multi-disciplinary team across 18 countries, aiming to develop food-based strategies with agri-food producers and the food industry that provide proof of efficacy and safety. To date the research team have published several key studies including a RCT with vitamin D-enhanced eggs( Reference Hayes, Duffy and O'Grady 49 ) and a meta-analysis of studies examining the effects of UV-exposed mushrooms( Reference Cashman, Kiely and Seamans 46 ) on vitamin D status. The final report from the ODIN project is due in 2017.
Supplementation is another potential strategy to support the UK population in achieving the dietary recommendation and subsequently improving vitamin D status. Although there are no data showing the number of people in the UK currently taking vitamin D supplements specifically, National Diet and Nutrition Survey data have shown that 23 % of adults aged 19–64 years and 39 % of adults over 65 years take at least one dietary supplement( Reference Bates, Lennox and Prentice 12 ). Currently the UK advise daily supplements for those considered at-risk of vitamin D deficiency such as young children, pregnant and breastfeeding women, housebound elderly and those with darker skin tones. However, the awareness and adherence to this current guidance is poor( Reference Buttriss 50 ) and so the impact of this advice, and any potential future advice, will only be realised if individuals are willing to take supplements and remember to do so. Universal provision of supplements has been identified as a successful strategy in some European countries where they are recommended for infants (Norway, Germany, Austria and Switzerland) and children up to the age of 5 years (Sweden). The potential universal use of vitamin D supplementation in the UK is therefore worthy of further research, as discussed in greater detail in two recent reviews( Reference Spiro and Buttriss 7 , Reference Buttriss 50 ).
The introduction of 10 µg/d as the new RNI for the UK population aged 1 year and above introduces a new era for vitamin D recommendations( Reference Lanham-New and Wilson 51 ). The Scientific Advisory Committee on Nutrition Vitamin D Working Group recognised that achieving the new RNI of 10 µg/d from natural dietary sources alone would be a challenge and so they have recommended that the UK government, namely Public Health England and the Department of Health, give consideration to strategies to support the UK population with achieving this RNI( 14 ). One of these potential strategies is fortification of foods, via mandatory or voluntary means. There has been significant progress in the research to support the use of vitamin D in food fortification, including the growth of potential bio-fortified foods, but there are outstanding questions and gaps in the research that need to be addressed to ensure the most efficacious and safe fortification practices are put in place. In addition, this review also highlights the need for further clarity as to the relative efficacy of vitamins D2 and D3 in raising total 25(OH)D concentrations, which the publication of the D2–D3 Study results will support. This further research will be key, alongside considerations, including cost, availability and ethics (vitamin D3 is not suitable for vegans), in allowing government, decision-makers, industry and consumers to make informed choices about potential future vitamin D policy and practice.
L. W. would like to acknowledge and thank the D2–D3 Study Team: Professor Susan Lanham-New (Principal Investigator), Dr Kathryn Hart, Dr Laura Tripkovic and Dr Ruan Elliott (University of Surrey), Professor Colin Smith and Dr Giselda Bucca (University of Brighton), Dr Simon Penson and Dr Gemma Chope (Campden BRI), Dr Jacqueline Berry (University of Manchester) and Professor Elina Hyppönen (University of South Australia). L. W. also acknowledges the support of the National Institute of Health Research Clinical Research Network (NIHR CRN) and would like to thank the following parties for their great help and kind assistance in the identification, recruitment and retention of participants onto the D2-D3 Study: Mrs Shahnaz Bano (Surrey County Council, UK), Mrs Fatima Bukhari and Rukhsana Hanjra (Islamic Resource Centre, Kingston, UK). The D2–D3 Study team are extremely grateful to Professor Peter Schroder and Mr James Phillips (BBSRC DRINC Programme), and Professors John Mathers (University of Newcastle) and Professor Hilary Powers (University of Sheffield) for their critical comments in the design of the D2–D3 Study and its implementation.
L. W. would like to acknowledge and thank the BBSRC Diet and Health Research Industry Club (DRINC) for funding the D2–D3 Study (BB/I006192/1) and her studentship. The BBSRC DRINC had no role in the writing of this article.
Conflicts of Interest
L. W. wrote the manuscript. L. T., K. H. and S. L. N. assisted with manuscript editing.