Vitamin D metabolic pathways, recommendations for the UK population, intakes and status

The first session set the scene with presentations outlining: (1) vitamin D metabolism and the specific deliberations and recommendations on RNIs for vitamin D from the UK’s SACN (Prof Ann Prentice, SACN Chair, Medical Research Council and Honorary Professor of Global Nutrition and Health, University of Cambridge); (2) vitamin D intakes and vitamin D status in the UK (Gillian Swan, Public Health England) and (3) the UK Government’s response to SACN’s recommendations (Prof Louis Levy, Public Health England). An overview of these three presentations follows.

Vitamin D is a generic name for two different secosteroid compounds, ergocalciferol (vitamin D₂) and cholecalciferol (vitamin D₃). Vitamin D₃ can be obtained from the diet and by endogenous synthesis in the skin via the action of UVB radiation (290–315 nm), which converts 7-dehydrocholesterol to vitamin D₃, whereas vitamin D₂ is obtained solely from the diet. It is naturally present in fungi (e.g. wild mushrooms or UVB-treated cultivated mushrooms, and yeasts). There are relatively few dietary sources of vitamin D₃, the richest being oily fish and egg yolks. Other sources include meat/meat products and fortified foods, such as fat spreads, some breakfast cereals, some dairy products (especially yogurts) and vitamin D-fortified dairy alternatives. In 1940, the fortification of margarine became mandatory in the UK (see DH 1991)⁽⁶⁾ but this requirement ceased in 2013⁽⁸⁾. However, fat spreads (and some other foods) continue to be fortified on a voluntary basis. Infant formula contains vitamin D by law at the level of 1–2·5 µg/100 kcal⁽⁹⁾.

Vitamin D is converted to 25(OH)D in the liver by the 25-hydroxylase enzyme (EC:1.14.14.24) and is the major circulating vitamin D metabolite. In the kidney, 25(OH)D is converted to the active metabolite 1,25-dihydroxyvitamin D, which acts in combination with parathyroid hormone and calcitonin to maintain calcium (Ca) and phosphate homoeostasis (Fig. 1). It should be noted that there is also evidence of extra-renal 1,25(OH)2D synthesis.^{(Reference Bouillon and Bikle11)}

Fig. 1.

Vitamin D metabolic pathway. 1,25(OH)₂D, 1,25-dihydroxyvitamin D; 1αOHase, 1-α-hydroxylase; VDR, vitamin D receptor. Courtesy of Prof Ann Prentice and modified from Prentice (2007)^{(Reference Prentice10)}.

The most commonly used marker of vitamin D status is 25(OH)D because it has a relatively long half-life (several weeks) and its concentration in serum/plasma reflects dietary intakes, UVB exposure and biological reserves of vitamin D. In the UK, a plasma or serum 25(OH)D concentration of 25 nmol/l (10 ng/ml) is considered to be the ‘population protective level’ as a minimum for protecting musculoskeletal health and lowering the risk of osteomalacia (adults) and nutritional rickets (children). Characterised by poor bone mineralisation, pain, deformities and fractures, these conditions are caused by low Ca intakes and/or vitamin D deficiency and cured by nutritional intervention. It should be noted that, elsewhere, different thresholds are used; for example, the US Institute of Medicine set a minimum threshold of 30 nmol/l to define increased risk of deficiency and 50 nmol/l as the minimum threshold for ‘sufficiency’⁽¹²⁾, and these thresholds are used in some of the studies mentioned later.

Synthesis of vitamin D following exposure of the skin to UVB radiation is the predominant source for most people in the UK, especially during spring and summer at UK latitudes. During the ‘vitamin D winter’ period in the UK (ranging from October to February in Southern England and September to April in northern Scotland),^{(Reference MacDonald, Mavroeidi and Fraser13,Reference Darling, Hart and Macdonald14)} UVB availability for dermal synthesis of vitamin D is limited or absent. This places increased emphasis on the dietary supply of the vitamin during the ‘vitamin D winter’. Some ‘at-risk’ groups with little or no skin exposure to sunlight during spring and summer, such as those who do not go outside often or who cover most or all of their skin, are reliant on dietary sources of vitamin D throughout the year. Ethnic minority groups with dark skin may not get enough vitamin D from sunlight exposure during the summer months because melanin present in skin limits the rate of skin synthesis^{(Reference Farrar, Kift and Felton15)}.

Vitamin D recommendations from Scientific Advisory Committee on Nutrition: deliberations and decisions

In 1991, the Committee on Medical Aspects of Food and Nutrition Policy set dietary reference values (DRV) for vitamin D for infants (up to 6 months; 8·5 µg/d), young children (up to 3 years; 7 µg/d), pregnant and lactating women (10 µg/d), older adults (aged 65+ years; 10 µg/d) and those confined indoors or otherwise at-risk of vitamin D deficiency (10 µg/d)⁽⁶⁾. However, no DRV for other age groups was considered necessary at that time⁽⁶⁾. These DRV were based on providing sufficient vitamin D to protect bone health. In 2010, SACN agreed to review the DRV for vitamin D because a substantial amount of published data had accumulated, including evidence of a high prevalence of low status in additional life stage and sex subgroups and new data suggesting a range of other health benefits associated with vitamin D.

SACN’s comprehensive review of the evidence included: consideration of the association between vitamin D status and health outcomes at different stages of life and in different population groups; the relative contributions to vitamin D status made by dietary intakes and cutaneous synthesis in the skin; and the potential adverse effects of high vitamin D intakes. The strongest evidence of health benefits for vitamin D was in the prevention of rickets, osteomalacia and falls, and in benefiting muscle strength and function in adults ≥ 50 years. Insufficient evidence was found for vitamin D in relation to non-musculoskeletal outcomes. As a result, musculoskeletal health outcomes were chosen as the basis for setting DRVs. The RNI established for vitamin D was modelled on the basis of maintaining population protective 25(OH)D concentrations of ≥ 25 nmol/l, at the individual level, throughout the year; below this concentration, the risk of poor musculoskeletal health increases. It was not possible to make any recommendation for the sunlight exposure needed in summertime to ensure winter concentrations remained above 25 nmol/l due to the large and complex variety of factors that influence UVB skin exposure and because of the potential for skin damage/cancer risk.

Using data from intervention studies, an RNI of 10 µg/d (400 IU/d) was established as the amount needed to achieve serum 25(OH)D concentrations ≥ 25 nmol/l during winter in 97·5 % of the population. Although most people would be expected to synthesise vitamin D during summer, between 8 and 53 % of some UK population subgroups have a plasma 25(OH)D concentration below 25 nmol/l during the summer months. Therefore, as a precaution, SACN recommended that the RNI is applicable throughout the year to protect the most vulnerable groups in the population and to ensure coverage of 97·5 % of the population. Since no data were available to set an RNI for young children, a ‘safe intake’ of 8·5–10 μg/d was set for infants from birth until 12 months and 10 μg/d for children aged between 1 and 4 years.

There had been discussion at SACN about possible health concerns for older people with undiagnosed compromised renal function or hyperparathyroidism, who may not be able to excrete or metabolise 25(OH)D sufficiently well to cope with high intakes of supplemental vitamin D, placing them at-risk of hypercalcaemia. Based on the distribution of intakes and 25(OH)D concentrations in the UK, SACN concluded that with an intake at an RNI of 10 µg/d (400 IU/d), the upper tail of the distribution was unlikely to approach 25(OH)D concentrations that might cause concern about toxicity.

In summary, in order to protect those population groups known to be at-risk of a low plasma 25(OH)D concentration due to minimal skin exposure to sunlight, as well as to protect those unidentified individuals at-risk of low status for any other reason, SACN recommended that the RNI and safe intake were applicable throughout the year for the whole population. Thus, pregnant and lactating women, and population groups and individuals at-risk of having a 25(OH)D concentration < 25 nmol/l, are included within the recommendation for the general population (Table 1).

Table 1.

Reference nutrient intake (RNI) and ‘safe intake’ values recommended by the Scientific Advisory Committee on Nutrition

(SACN 2016)

Source: SACN (2016) Vitamin D and health report⁽⁷⁾.

* Recommendations from SACN apply to the whole year and are for total intake (diet plus supplements).

† Includes ‘at-risk’ population subgroups, such as ethnic groups with darker skin and those who do not regularly expose their skin to sunlight during the summer months, who may be at increased risk of having a serum 25(OH)D concentration < 2 5 nmol/l.

Vitamin D intakes and vitamin D status in the UK

Data for children (aged 1·5 years and older) and adults on vitamin D intakes and plasma 25(OH)D status within the UK population are available from the National Diet and Nutrition Survey (NDNS)^{(Reference Roberts, Steer and Maplethorpe16)}. The latest NDNS data on vitamin D intakes and status was published in the Years 9 to 11 report (2016/2017 to 2018/2019)^{(Reference Bates, Collins, Jones, Page, Roberts and Swan17)} and supersedes the Years 7 and 8 report data (2014/2015 to 2015/2016)^{(Reference Roberts, Steer and Maplethorpe16)}, but was not available at the time the Forum was held. Since the Years 7 and 8 data were those presented and discussed at the Forum, we have included data from these years (see Tables 2 and 3).

For children aged 4–18 months, information is provided in the Diet and Nutrition Survey of Infants and Young Children (DNSIYC, 2011)^{(Reference Lennox, Sommerville and Ong18)}. The main dietary contributors for adults and children over 1 year were cereals/cereal products (including fortified breakfast cereals), meat, fish, fortified milk/milk products, eggs and fat spreads, but their relative contributions varied with age (Table 2). Fortified milk/milk products were a major contributor for young children, but this food group made a smaller contribution in older children and adults. Conversely, the contribution from meat and fish was higher in adults. The main dietary contributors for infants under 1 year were infant formula and commercial infant foods, meat, fish, eggs, fat spreads and cereals/cereal products (Table 2).

Table 2.

Main contributors (%) to vitamin D intakes of infants (non-breastfed), children and adults in the UK population*

(Percentages)

N/A, data not available.

* Sources: Diet and Nutrition Survey of Infants and Young Children (indicated by shaded cells; DNSIYC, 2011)^{(Reference Lennox, Sommerville and Ong18)} and the National Diet and Nutrition Survey (years 7 and 8 of the rolling programme; 2014/2015 to 2015/2016)^{(Reference Roberts, Steer and Maplethorpe16)}. The latest NDNS data on vitamin D intakes and status published in the Years 9 to11 report are now available (December 2020), and supersede the Year 7 and 8 report data in this paper but since the Year 7 and 8 data were presented and discussed at the meeting, we have included these here.

At the time the data in each survey were collected, the only UK RNIs in existence were for infants, pregnant/lactating women, older adults, those confined indoors or otherwise at-risk of vitamin D deficiency⁽⁶⁾. With the exception of non-breastfed infants aged 4–12 months in the DNSIYC, mean vitamin D intakes from all sources (food and supplements) were below the RNI subsequently set in 2016 for infants, children and adults of all ages (Table 3). Latest NDNS data (Years 9 to 11; 2016/2017 to 2018/2019)^{(Reference Bates, Collins, Jones, Page, Roberts and Swan17)} indicates that total vitamin D intakes (including dietary supplements) and status have not changed substantially since the recommendations were set in 2016. Total vitamin D intakes remained below the RNI for all age/sex groups, except for women aged 65–74 years (mean intake of 10.1 µg/d). A significant increase in 25(OH)D status was reported for children aged 4–10 years, compared to Years 7 and 8 of the survey, but changes for other age/sex groups were not significant.

Table 3.

Mean vitamin D intakes and vitamin D status of the UK population

(including non-breastfed infants*)

Sources: National Diet and Nutrition Survey (years 7 and 8; 2014/2015–2015/2016) and the Diet and Nutrition Survey of Infants and Young Children (indicated by shaded cells; DNSIYC, 2011)^{(Reference Lennox, Sommerville and Ong18)}.

* Received no breast milk during the 4-d reporting period.

† Includes contribution from dietary supplements.

‡ Sample below n=50. The new NDNS data on vitamin D intakes and status published in the Year 9–11 report are now available (December 2020), and supersede the Year 7 and 8 report data in this paper but since the Year 7 & 8, data were presented and discussed at the meeting, we have included these.

The population mean concentration of 25(OH)D, averaged across the seasons, measured in these surveys was above the ‘population protective level’ of 25 nmol/l in all age groups. The percentage of infants with a plasma 25(OH)D concentration below 25 nmol/l was 6 % (infants aged 5–11 months) and 2 % (12–18 months of age). It was noted that sampling for the DNSIYC was weighted towards the summer months when 25(OH)D concentrations would be expected to be higher^{(Reference Lennox, Sommerville and Ong18)}. In comparison, a greater proportion of children (aged 4 years and over) and adults had plasma 25(OH)D concentrations below 25 nmol/l (Table 3). In particular, 39 % of girls and 15 % of boys aged 11–18 years were below the threshold, while 19 % of men and 16 % of women (aged 19–64 years) and 11 % of men and 15 % of women (aged 65+ years) also had a plasma 25(OH)D below 25 nmol/l (NDNS results are averages of samples collected across all seasons of the year).

In all age groups, circulating 25(OH)D concentrations were lowest during the winter months (January to March) and highest during the summer months (July to September). Although recent time trend analysis revealed some statistically significant changes in vitamin D intakes and status over the past decade (2008–2017), these were small and inconsistent in direction. Vitamin D intake and status increased with income in most age groups^{(Reference Bates, Collins and Cox19)}.

The NDNS shows that a greater proportion of individuals living in Scotland and Northern Ireland had a plasma 25(OH)D concentration below 25 nmol/l than in the UK as a whole, presumably due to the lower availability of UVB for vitamin D synthesis at more northerly latitudes. In addition, although the NDNS is designed to be a representative sample of the UK, owing to under-representation of participants from ethnic minority groups in the sample, it is not possible to provide specific information on the vitamin D intakes and vitamin D status of ethnic minority groups, some of whom may be at greater risk of vitamin D deficiency. However, some recent information, available from other sources, is described later.

In summary, mean intakes during the period 2014–2016, from food and supplements, are below the 2016 RNI in adults and children over 1 year. The main contributors to intake vary a little with age, but in most groups are meat, breakfast cereals, eggs, fat spreads, fortified milk/milk products and oily fish. Infant formula is the highest contributor in young children. While the mean plasma 25(OH)D concentrations measured in 2014–2016 (average of samples collected throughout the year) were above the current RNI in all age groups, a notable proportion had concentrations below 25 nmol/l, for example, 39 % of girls and 15 % of boys aged 11–18 years and 19 % of men and 16 % of women aged 19–64 years. Intakes and plasma concentrations increase with income, and plasma 25(OH)D concentrations are highest in the summer and lowest in the winter months.

Policy and public health implications of Scientific Advisory Committee on Nutrition’s vitamin D recommendations

Public Health England’s advice, based on SACN’s recommendations, can be found in Table 4. Mean intakes of vitamin D from all sources combined (food and supplements) are below the RNI recommended by SACN for most groups of the population. The exception is non-breastfed infants aged 4–9 months consuming infant formula (which contains added vitamin D by law). Data from the Health Survey for England, Scottish Health Survey and cohort studies (as well as data from NDNS, see Table 3) indicate that a notable proportion of some population subgroups in the UK do not achieve a plasma 25(OH)D concentration of 25 nmol/l, even during the summer months. This applies to 17 % of adults living in Scotland; 16 % of adults living in London; 53 % of women of South Asian ethnic origin living in southern England and 29 % of pregnant women in north-west London^{(7,Reference Darling, Hart and Macdonald14)} .

Table 4.

Public Health England advice on vitamin D for different population groups^(1,20)

* PHE re-issued its advice during the COVID-19 pandemic, emphasising the need to consider supplementation, even during the summer months, if time outdoors is limited.

Professor Louis Levy (Public Health England) explained that, based on the findings of the SACN review, government advice for dietary intakes of vitamin D has been provided as a precautionary measure. This decision was made because of the number and complexity of factors influencing endogenous vitamin D synthesis during the summer months (e.g. limited skin exposure, skin pigmentation), the need to avoid excessive sun exposure (because of skin cancer risk) and the lack of sufficient UVB availability for endogenous synthesis during the winter months. Following SACN’s review, several changes to the then existing government advice on vitamin D were announced in July 2016⁽²⁰⁾ (see Table 4). In forming the new public health recommendations, consideration was given to the various factors that may affect vitamin D status, including the seasonality of endogenous vitamin D synthesis at UK latitudes, the frequency of consumption of foods naturally containing, or fortified with, vitamin D, as well as the particular needs of at-risk groups mentioned above. Previous advice⁽⁶⁾ about vitamin D supplementation for infants and children was extended to include the period from birth until 6 months, despite some concerns among policymakers that difficulties in using vitamin D drops for younger infants may adversely affect the uptake of advice on breast-feeding. Professor Levy indicated that the need for somewhat complex and nuanced advice (shown in Table 4) appears to have led to an oversimplified interpretation by health and social care professionals, which may be leading to confusion about the detail of the government’s recommendations. He also noted that current clinical guidelines from the National Institute for Health and Care Excellence do not reflect the recommendation from SACN that pregnant and breast-feeding women should be considered as a part of the general population (i.e. that vitamin D requirements are the same as for non-pregnant women).

In summary, the current RNIs for vitamin D recommended by SACN relate to the maintenance of plasma (or serum) 25(OH)D concentration to ensure musculoskeletal health. Although the majority of the population achieve the plasma 25(OH)D concentration of 25 nmol/l during the summer months, as discussed earlier, plasma concentrations are below this in a sizeable proportion of some population subgroups. It is not possible to predict which individuals within subgroups may be at an increased risk of low vitamin D status owing to the range of factors that determine vitamin status (including dietary intake and skin UVB exposure), and government advice reflects this difficulty.

During the discussion, the pros and cons of vitamin D supplementation versus vitamin D fortification were debated (see Final Discussion). Other points raised were the challenges of delivering advice that is unambiguous and yet appropriately nuanced, an example being how the advice for infants under 1 year to take a supplement ranging from 8·5 to 10 μg/d should be interpreted by parents/carers and supplement manufacturers. Healthy Start vitamin drops provide 10 μg vitamin D₃ per 5 drops (https://www.healthystart.nhs.UK/for-health-professionals/vitamins/).

Vitamin D status in ethnic minority groups

A series of presentations considered vitamin D status in different population subgroups and the prevalence of nutritional rickets. Dr Andrea Darling (University of Surrey) focused on the vitamin D status of ethnic minority groups. The 2006–2007 Diet, Food Intake, Nutrition and Exposure to Sunlight in Southern England (D-FINES) study of South Asian women living in the UK (at a latitude of 52°N) reported that mean 25(OH)D concentrations remained around the 25 nmol/l threshold throughout all seasons of the year, with 53 % of women not achieving a concentration > 25 nmol/l during the summer months^{(Reference Darling, Hart and Macdonald14)}. Data from the 2010 Health Survey for England have also highlighted that Asian participants (over 16 years) had a lower mean 25(OH)D status (20·5 nmol/l) than White participants (45·8 nmol/l)⁽²¹⁾.

Dr Darling presented findings from a secondary analysis of data from the UK Biobank cohort (2006–2010) on the determinants of 25(OH)D status for South Asian adults (Indian, Pakistani and Bangladeshi) living in the UK, including information on serum 25(OH)D concentration as well as vitamin D dietary intakes and supplement use^{(Reference Darling, Blackbourn and Ahmadi22,Reference Darling23)} . The cohort comprises over 500 000 people, aged 40–69 years old, among whom are 8024 participants of South Asian ethnicity. Most of the South Asian participants were recruited from Hounslow, Birmingham, Croydon and Leeds. Serum 25(OH)D concentration was available for 6344 of the South Asian participants, who were: predominantly of Indian ethnicity (75 %); aged 40–60 years (72 %) and living in London (46 %), the north of England (22 %) or the Midlands (20 %). The balance of male (54 %) and female (46 %) participants was approximately even and included individuals with a BMI categorised as under-weight/normal weight (32 %), overweight (45 %) and obese (21 %). Blood sampling for each participant was performed during one season only, although data for all four seasons were available across the sub-cohort. Despite the strengths of this large prospective cohort, care should be taken when interpreting results from the UK Biobank because it contains relatively small numbers of some sub-sections of the UK population and because participants are more health conscious than the general population^{(Reference Fry, Littlejohns and Sudlow24)}.

Regardless of the season when the samples were taken, median serum 25(OH)D concentrations for Indian, Pakistani and Bangladeshi participants were all close to, or below 25 nmol/l, with summer concentrations being only slightly higher than during other seasons^{(Reference Darling23)}. During winter and spring, approximately 60 % of participants had a serum 25(OH)D concentration below 25 nmol/l, whilst over 20 % were below 15 nmol/l. Assessing all seasons combined, 92 % of participants had 25(OH)D < 50 nmol/l; 55 % had 25(OH)D < 25 nmol/l and 20 % had 25(OH)D < 15 nmol/l. Ten percent of participants had missing data as 25(OH)D was under the detection limit (10 nmol/l). Logistic regression analysis revealed several factors that predicted a 25(OH)D concentration <25 nmol/l. These factors included: male sex; Pakistani ethnic origin; higher BMI; being closer to middle age (40–59 years); rarely/never consuming oily fish; vegetarian; little summer sun; not using sun protection (suggesting minimal sun exposure – those who used skin cream presumably spent time in the sun); not using vitamin D-containing supplements and blood sampling during winter/spring. Of note, although 25(OH)D values were slightly lower for those living at more northerly latitudes, these differences were not statistically significant in the model once adjustment had been made for other confounders. Specifically, the logistic regression model suggested that increased summer sunlight exposure (over 3 h/d), having a normal BMI, eating oily fish twice per week (two servings of fish a week, one being an oily fish, is the current recommendation in the UK) and using supplements reduced risk of vitamin D deficiency.

Previously published work on vitamin D intake and supplement use in the UK Biobank cohort revealed low dietary intakes of vitamin D in Indian (mean of 1·0 μg/d) and Pakistani (mean of 1·5 μg/d) participants. Although Bangladeshi participants had slightly higher intakes (3·0 μg/d)^{(Reference Darling, Blackbourn and Ahmadi22)}, these were far below the RNI of 10 μg/d. The majority (>60 %) of south Asian participants of both sexes, and from all three ethnic groups, did not take vitamin D supplements; those who did were, typically, younger, female and living in London^{(Reference Darling, Blackbourn and Ahmadi22)}.

In summary, vitamin D intake is low (1–3 μg/d) and supplement use is also low (only 23 % taking a vitamin D-containing supplement) among UK South Asians. New data suggest that low vitamin D status is almost universal in UK South Asians (e.g. 92 % < 50 nmol/l; 55 % < 25 nmol/l; 20 % < 15 nmol/l)^{(Reference Darling23)}. Year round, median serum 25(OH)D concentrations for Indian, Pakistani and Bangladeshi participants were all close to, or below, 25 nmol/l, with summer concentrations being only slightly higher than for other seasons. About 10 % had blood vitamin D concentrations below the detection limit (10 nmol/l). Since the UK Biobank cohort is likely to be more health conscious than the general population^{(Reference Fry, Littlejohns and Sudlow24)}, vitamin D status in the general South Asian population may be even lower than reported here.

Child and maternal vitamin D requirements

Professor Mairead Kiely reiterated that SACN recommendations are for healthy members of the general population, which includes pregnant women, who are usually under medical supervision. Within the neonatal period (the month following birth), babies are under observation either by their health visitor and/or the primary care team and postnatal staff. After this time, babies and children are considered part of the ‘general population’. As noted above, the current UK government recommendations for vitamin D include a safe intake value of 8·5–10 µg/d throughout the first year of life and 10 µg/d throughout the year for ages 1–4 years (see Table 4). The RNI of 10 µg/d (400 IU/d) for pregnant and lactating women is the same as for non-pregnant adults.

Nutritional rickets in the UK

Dr Suma Uday (University of Birmingham, UK) discussed the burden of rickets in the UK, contrasted UK vitamin D supplementation policy with other European countries and made suggestions for improving adherence to supplementation, particularly in infants. Nutritional rickets and hypocalcaemic complications related to severe vitamin D deficiency are more common among BAME groups residing in the UK than among White children. Dr Uday discussed three clinical cases, from a single UK tertiary centre, of infants (aged 5–6 months) with darker skin who presented in early spring with cardiac arrest, respiratory arrest following seizure or severe respiratory distress^{(Reference Uday, Fratzl-Zelman and Roschger49)}. All three infants were noted to have severe hypocalcaemia (adjusted Ca 1·2–1·96 mmol/l; normal range 2·2–2·7 mmol/l) with elevated alkaline phosphatase (802–1123 μg/l; normal range 105–357 μg/l) and parathyroid hormone levels (219–482 ng/l; normal range 13–29 ng/l) and very low serum 25(OH)D concentrations (<15 nmol/l). Severe underlying rickets on radiographs and dilated cardiomyopathy (enlargement of the heart) were uncovered on further investigation. It was highlighted that none of the infants was on vitamin D supplements and the need for supplements had not been raised during child surveillance visits in any of the cases. One infant sadly died from cardiac arrest. Post-mortem examination revealed florid rickets with: rachitic rosary (‘beading’ of the ribs), abnormally widened rib growth plate on histology (widened, irregular zone of hypertrophic chondrocytes), increased osteoid thickness and volume and extremely low bone mineralisation on histomorphometry^{(Reference Uday, Fratzl-Zelman and Roschger49)}.

The incidence of hospital admissions due to rickets in England increased between 2000 and 2011^{(Reference Goldacre, Hall and Yeates50)}. A survey conducted by the British Paediatric Surveillance Unit between March 2015 and March 2017 suggested an annual overall incidence of 0·48 cases per 100 000 (sixty cases per year) of nutritional rickets, with the greatest number of cases in infants aged 12–23 months, and of BAME origin^{(Reference Julies, Lynn and Pall51)}. The limitations of the survey included a strict case definition leading to exclusion of a significant number of cases, including those of hypocalcaemic seizures^{(Reference Uday and Hogler52)}. The nature of the survey also meant that it did not capture cases seen in primary care or those seen by general practitioners who are not adequately trained in paediatrics. Nonetheless, it is important to note that nearly 80 % of the cases in the survey were not taking the recommended vitamin D supplements. Four cases of dilated cardiomyopathy were identified, of whom two sadly died, highlighting the associated morbidity and mortality. Moreover, nutritional rickets cases presenting to secondary care are only the tip of the iceberg of hidden widespread vitamin D deficiency^{(Reference Goldacre, Hall and Yeates50)}. Depending on the study design and case definition, the incidence of rickets (per 100 000 people) among darker skinned and immigrant populations in the UK (95/100 000), USA (220/100 000), Australia (2300/100 000) and Denmark (60/100 000) have all been reported to exceed the incidence rate for classification as a rare disease (50/100 000)^{(Reference Thatcher, Pludowski and Shaw53)}. In the UK, in addition to the increasing proportion of BAME residents (14 % in the 2011 census), there is a yearly net positive migration of approximately 150–200 000 BAME individuals^{(Reference Uday and Hogler54)}. Hence, the disease burden is likely to rise unless effective strategies for prevention are implemented.

An evidence-based global consensus statement on recommendations for the prevention and management of nutritional rickets was published in 2016^{(Reference Munns, Shaw and Kiely47)}. A group of thirty-three experts, from various fields, developed definitions for sufficiency, insufficiency and deficiency of serum 25(OH)D and daily Ca intakes for the prevention and management of nutritional rickets (Table 5). Universal supplementation of 400 IU/d (10 µg/d) was recommended by the Global Consensus Group for infants (birth to 12 months), a 600 IU/d (15 µg/d) supplement for pregnant women and lifelong supplementation of 600 IU/d (15 µg/d) for those considered to be at-risk, in addition to ensuring adequate Ca intakes. A recent survey (based on questionnaire responses from expert representative members of the Bone and Growth Plate Working group of the European Society of Paediatric Endocrinology) identified that the UK has the lowest adherence (5–20 %) in Europe to infant vitamin D supplementation advice, based on data from twenty-nine countries^{(Reference Uday, Kongjonaj and Aguiar55)}. Factors that were statistically significantly linked to increased adherence in other countries were universal supplementation of breast- and formula-fed infants (P = 0·007), monitoring of adherence at child health surveillance visits (P = 0·001), giving information to parents at discharge from neonatal units (P = 0·02) and providing financial family support (P = 0·005).

Table 5.

Global Consensus Group recommendations for the prevention and management of nutritional rickets in infants, children and adolescents^{(Reference Munns, Shaw and Kiely47)}

* Recommendations for children over 12 months of age. For infants 0–6 and 6–12 months of age, the adequate Ca intake recommended is 200 and 260 mg/d, respectively. For children over 12 months of age, a dietary Ca intake of < 300 mg/d increases the risk of rickets independently of serum 25-hydroxyvitamin D concentration.

Dr Uday argued that the adherence to supplementation advice in the UK could be improved by a combination of some of the above strategies. Universal supplementation of breast- and formula-fed infants would give a simple message to professionals and also the public. There is evidence from the recent British Paediatric Surveillance Unit survey^{(Reference Julies, Lynn and Pall51)} and a previous hypocalcaemic seizure survey^{(Reference Basatemur and Sutcliffe56)} that formula-fed infants can be affected by rickets and that fortified formula alone does not guarantee protection against nutritional rickets^{(Reference Gross, Tenenbein and Sellers57)}. Mandatory monitoring of adherence to supplements at child health surveillance visits will prompt healthcare professionals to check and recommend supplements. Dr Uday suggested that there are simple practical monitoring approaches that could be used.

In summary, the potentially devastating health consequences of vitamin D deficiency in infancy are preventable, yet it is evident that the UK government’s advice on vitamin D supplementation in infancy is not reaching all families. Whilst the government has supplementation policies in place, there is a lack of implementation strategies. Given the significant mortality and morbidity associated with this preventable condition, Dr Uday suggested that the government should consider robust strategies for an infant vitamin D supplementation programme similar to those used in immunisation programmes. In concluding, Dr Uday recommended: (i) universal supplementation and simplified advice; (ii) increased monitoring of adherence to supplementation advice; (iii) financial support or incentives for general practitioners to implement infant vitamin D supplementation and (iv) information at birth and antenatal visits.

Subsequent discussion focused on strategies to improve awareness at ‘grassroots’ level about the clinical significance and long-lasting consequences of inadequate vitamin D intakes. Available evidence suggests that families are unaware of the UK government’s advice, rather than rejecting it, which emphasises the importance of health professionals being aware of the potentially serious consequences of vitamin D deficiency, the population subgroups considered to be at particular risk, and routinely alerting parents and carers to the need for supplementation. Subtle inconsistencies in the communication of the government’s advice, as reflected in the wording of guidelines from several expert bodies, may be leading to confusion among general practitioners and healthcare professionals.

It was suggested that a national publicity campaign might be useful for raising public awareness. However, experience to date suggests that to be effective, this approach needs to be reinforced by other actions. In particular, accurate and consistent messages (in line with government advice) delivered by healthcare professionals in routine contact with parents may be more effective in this case. Reinforcement of government advice through local community groups and initiatives, such as faith and parenting groups, may be particularly helpful in reaching at-risk groups of the population.

Food-based solutions for enhancing vitamin D status

Three potential strategies for addressing poor micronutrient intakes have been identified by the WHO and the FAO of the UN, i.e. increasing the diversity of foods consumed; food fortification and nutrient supplementation⁽⁵⁸⁾; all of which may be relevant to addressing the widespread prevalence of inadequate vitamin D intake.

While often discussed as a possible means of closing the gap between current vitamin D intakes and vitamin D recommendations, increasing supplement usage as an appropriate public health strategy to increase intakes across the population has intrinsic limitations. This is because vitamin D supplements are only effective in those who consume them, and their usage can be low, for example, <20 % use of vitamin D-containing supplements in Ireland and the UK. Novel approaches to supplementation are being developed, such as vitamin D sprays for oral use (discussed further below)^{(Reference Black, Seamans and Cashman59,Reference Kiely and Black60)} .

As a means of addressing low vitamin D intakes, increasing the diversity of foods consumed is particularly challenging because there are very few food sources naturally rich in vitamin D, and these are predominantly of animal origin. This is especially relevant in the context of the recent calls for a radical transformation of the global food system – with specific emphasis on increased consumption of plant-based foods and reductions in animal-derived foods for many, as part of a more sustainable flexitarian-type diet^{(Reference Willett, Rockstrom and Loken61)}. The WHO/FAO have indicated that food fortification has potentially the widest and most sustainable impact and is generally considered more cost-effective than other intervention approaches⁽⁵⁸⁾ (WHO/FAO, 2006). The potential importance of food fortification (and biofortification) as a means of improving vitamin D status at a population level was emphasised during the Forum discussions. A population-level approach was emphasised during this Forum and has been discussed previously,^{(Reference Pilz, Marz and Cashman62)} with Finland leading the way with vitamin D food fortification strategies.

Prof Kevin Cashman from University College Cork (Ireland) explained that a number of RCTs conducted during the winter months, as part of the ODIN project, found that several fortified foods were effective at preventing serum 25(OH)D concentrations below 30 nmol/l in the treatment group (i.e. zero or very low prevalence). These fortified foods included UV-exposed mushrooms (providing around 100 µg D₂/d)^{(Reference Cashman, Kiely and Seamans63)}, eggs biofortified with vitamin D₃ and 25(OH)D₃ (4–5 µg per egg)^{(Reference Hayes, Duffy and O’Grady64)} and fortified low-fat Gouda cheese (about 6 µg D₃/d)^{(Reference Manios, Moschonis and Mavrogianni65)}. The only fortification approach studied that was not found to be effective was the use of UV-irradiated yeast (about 25 µg/d), containing vitamin D₂, used specifically to make bread, which did not raise total 25(OH)D concentrations, presumably due to lower bioavailability^{(Reference Itkonen, Skaffari and Saaristo66)}. However, fortification of bread with vitamin D₃ has been shown to be effective in enhancing vitamin D status^{(Reference Natri, Salo and Vikstedt67)}. Overall, fortification of foods with vitamin D was deemed to be technologically feasible, and sensory data indicated a good level of consumer acceptability^{(Reference Hayes, Duffy and O’Grady64)}.

A dietary modelling analysis has been performed to explore the potential effect of the fortified foods tested within ODIN on raising total vitamin D intakes within several EU countries. For example, data from the National Adult and Nutrition Survey for Ireland were modelled as follows: the contribution from supplements and existing fortified foods was removed to create a ‘baseline’ intake value. When current intakes of individual products were replaced in the modelling by a combination of biofortified foods (beef, pork and eggs), as well as fortified milk and cheese, baseline intakes were estimated to increase from 3·3 µg/d to 8·0 µg/d, thus almost achieving the target intake of 10 µg/d. An integrated predictive model, which accounted for changes in UVB availability during the year, suggested that the prevalence of a 25(OH)D concentration < 30 nmol/l during winter in Irish adults (18·1 %) could be reduced to 6·6 % by the stepwise introduction of an increasing number of fortified foods^{(Reference Cashman, Kazantzidis and Webb68)}. In relation to the efficacy of fortification for at-risk ethnic minority groups, a recent 12-week RCT among women of Pakistani origin, conducted in Denmark, reported that vitamin D-fortified foods, together providing an intake of 20 µg/d, reduced the prevalence of 25(OH)D < 30 nmol/l from 34 % to 3 %^{(Reference Grønborg, Tetens and Christensen69)}.

In the ODIN modelling, consideration was also given to whether consumption of fortified foods could increase the risk of exceeding the upper intake level for vitamin D of 100 µg/d (4000 IU/d) (adults). Consumption of a combination of fortified and biofortified foods did not pose a risk at the 99th percentile of the distribution. Even those taking high dose supplements (50 µg/d) in addition to fortified foods would be unlikely to reach the upper intake level. In summertime, endogenous vitamin D synthesis in individuals with plentiful sun exposure is self-regulating; thus, when combined with fortified foods, this would not be expected to lead to excessive exposure to vitamin D. However, this assumption needs to be tested.

Professor Cashman emphasised that while the RCT evidence suggests a benefit from fortification, one of the greatest challenges is how to take interventions shown to work in a research setting and implement them in a real-world setting; a key issue highlighted by WHO^{(Reference Peters, Tran and Adam70)}. The success of a voluntary, government-driven vitamin D fortification policy in Finland was given as an example. Following introduction of the policy in 2003, the percentage of the population with a 25(OH)D concentration < 30 nmol/l decreased from 12 % in 2000 to <1 % in 2011^{(Reference Jaaskelainen, Itkonen and Lundqvist71)}. Also of note, the introduction of vitamin D food fortification resulting in a much higher increase (about 34 nmol/l) in those with vitamin D deficiency (<30 nmol/l) and the least (about 11 nmol/l) in participants with adequate (>50 nmol/l) serum 25(OH)D concentration at baseline^{(Reference Jaaskelainen, Itkonen and Lundqvist71)}.

In summary, low 25(OH)D concentrations (below 30 nmol/l) remain relatively common in Europe and there is a several-fold higher risk among a number of ethnic minority groups. Animal-feeding trials, food production studies and food-based RCTs, accompanied by dietary modelling experiments, have demonstrated that diverse fortification strategies could increase vitamin D intake across the distribution of population intakes and thus reduce the risk of deficiency. There is an urgent need for follow-up research and implementation per se to scale up interventions into real-world settings, with the ultimate aim of preventing vitamin D deficiency.

Fortification with vitamin D: specific modelling from National Diet and Nutrition Survey

UK government advice to tackle low vitamin D status is described in Table 4 and comprises a blend of advice on UVB exposure, diet and supplement usage (during the winter months for the general population and all year for at-risk groups). As discussed above, fortification of widely consumed foods is a potential means to improve vitamin D status at the population level. Dr Rachel Allen (Rachel Allen Nutrition, London) presented outcomes from a modelling analysis performed to identify the fortification vehicle and concentration of vitamin D most likely to raise intakes safely in at-risk groups of the population (children aged 18–36 months; females aged 15–49 years; adults aged > 65 years), without exceeding conservative estimates for tolerable upper limits for children (25 µg/d) and adults (50 µg/d) as set by EFSA at the time of the study^{(Reference Allen, Dangour and Tedstone75)}. The European Food Safety Authority⁽⁷⁶⁾ and SACN (2016)⁽⁷⁾ subsequently set safe upper limits as follows: adults and children aged 11–17 years (100 µg/d), children aged 1–10 years (50 µg/d) and infants (25 µg/d). Data from the first 2 years of the NDNS Rolling Programme (2008–2010) (the most current data set available at the time) were used in combination with published information on the dose–response relationship between vitamin D intake and status^{(Reference Cashman, Hill and Lucey39,Reference Cashman, Wallace and Horigan40)} . Wheat flour and milk were identified as the primary fortification vehicles of interest, due to their almost universal consumption at that time among the at-risk population subgroups included in the analysis.

The modelling predicted that fortification of wheat flour with 10 µg vitamin D/100 g flour would reduce the proportion of individuals in at-risk groups with intakes below 10 µg/d (which is now the RNI) from 93 % to 50 %, with mean intakes increasing from 3·7 µg/d to 10·8 µg/d, and there was no indication of any individual exceeding the tolerable upper limit. In addition, intakes improved across all socio-economic groups. This scenario was predicted to raise mean wintertime serum 25(OH)D concentration from 39 nmol/l (95 % CI 35, 45 nmol/l) to 51 nmol/l (95 % CI 43, 71 nmol/l). In summary, wheat flour (fortification at 10 µg/100 g flour was the optimal level) appeared to be a more effective fortification vehicle than milk (fortified at concentrations of 0·5 to 7 µg/100 ml of milk), or a combination of wheat flour and milk fortification.

Although analyses such as this^{(Reference Allen, Dangour and Tedstone75)} are hypothetical and have limitations, simulation studies using nationally representative consumption data are important to help inform policymakers about the potential impact of fortification at a population level. Fortification of wheat flour may be an effective method for improving vitamin D status without exceeding the tolerable upper limit set for the UK. Points made during the discussion included: a need to learn from the ongoing consultation in the UK on fortification of wheat flour with folic acid (https://www.gov.UK/government/consultations/adding-folic-acid-to-flour); consideration of other types of flour that may be consumed more extensively by some ethnic minority at-risk groups, such as chapatti and gram flours (which were not included in the modelling); and, if wheat flour was to be adopted as a fortification route, the implications of avoidance of wheat flour by those choosing to avoid gluten-containing foods as a lifestyle choice. It was noted that several countries in the Middle East region (e.g. Saudi Arabia and Jordan; http://www.ffinetwork.org/regional_activity/middle_east.php) have adopted mandatory fortification of flour with vitamin D at approximately 1·4 µg/100 g.

Vitamin D₂ or vitamin D₃: which is the best form for fortification?

Dr Louise Durrant (formerly University of Surrey) presented results from the D₂-D₃ study, which investigated differences in the efficacy of vitamin D₂ and vitamin D₃ supplementation on serum concentrations of total 25(OH)D, 25(OH)D₂ and 25(OH)D₃. Prior to this study, a limited but growing body of evidence suggested that vitamin D₃ was more effective in raising total 25(OH)D concentration than vitamin D₂. This included a systematic review and meta-analysis of RCTs that compared directly the effects of vitamin D₂ and vitamin D₃ on serum 25(OH)D concentration in humans, and which concluded that vitamin D₃ is more effective at raising serum 25(OH)D concentration than is vitamin D₂ ^{(Reference Tripkovic, Lambert and Hart77)}. However, this analysis was limited by the small sample size in most of the included studies and by a high degree of heterogeneity. The D₂-D₃ study compared the efficacy of 15 µg/d of vitamin D₂ or vitamin D₃, consumed for a 12-week period within a fluid (juice) or solid (biscuit) food matrix during wintertime (October to March), on raising vitamin D status in South Asian and White European women, relative to placebo^{(Reference Tripkovic, Wilson and Hart78)}.

Participants (n 335) were recruited and randomised to receive vitamin D₂ juice, vitamin D₂ biscuit, vitamin D₃ juice, vitamin D₃ biscuit or placebo (n 65–70; White European n 48–49; South Asian n 17–19 per group). Although both vitamin D₂ and vitamin D₃ raised total 25(OH)D concentrations, the vitamin D₃ biscuit and D₃ juice groups experienced significantly greater post-intervention increases in total 25(OH)D concentration, relative to the vitamin D₂ biscuit and D₂ juice groups, or the placebo group. Results did not differ according to the type of fortification vehicle (juice v. biscuit), and so the groups were combined in the analyses. Mean post-intervention 25(OH)D concentrations for women given vitamin D₂ within either a biscuit (21·4 nmol/l) or juice (17·0 nmol/l) were not only lower than at baseline (33·7 nmol/l for both groups) but also lower than in the placebo group (24·3 nmol/l). This seemingly adverse influence of vitamin D₂ supplementation on total 25(OH)D concentration was suggested to have contributed to the lower overall 25(OH)D status of the D₂ groups.

Women of South Asian ethnicity had mean plasma total 25(OH)D concentrations below 25 nmol/l at baseline and, in the placebo group, plasma 25(OH)D fell to below 20 nmol/l during the study. While both forms of vitamin D raised mean total 25(OH)D above 25 nmol/l, mean concentrations were above 50 nmol/l only in the combined vitamin D₃ group after the 12-week intervention. A mean total 25(OH)D concentration above 50 nmol/l was not achieved in the combined vitamin D₂ group. The substantial fall in plasma 25(OH) D concentrations in South Asian women in the placebo group prompts ethical questions about the future use of RCTs with a placebo arm for vitamin D intervention studies involving South Asian women.

To our knowledge, the D₂-D₃ study is the largest RCT to date that compared specifically the effects of relatively low doses of vitamin D₂ and vitamin D₃ on serum total 25(OH)D in both White Europeans and South Asians. Several other trials published since 2012 have also concluded that vitamin D₃ is more effective than vitamin D₂ in raising 25(OH)D concentrations, highlighting the need to update the meta-analysis conducted by Tripkovic et al. (2012)^{(Reference Tripkovic, Lambert and Hart77)}. Further research is also required to establish whether the influence of vitamin D₂ supplementation on plasma 25(OH)D₃ concentrations is indeed a reciprocal phenomenon, as emerging evidence suggests may be the case^{(Reference Hammami, Abuhdeeb and Hammami79)}.

In summary, although both vitamin D₂ and vitamin D₃ raise serum 25(OH)D concentrations, the overall body of evidence from intervention trials suggests that the two forms are not equally effective. Considerations about which form of the vitamin to use should be taken into account not only when setting public health recommendations and implementing the most appropriate treatment for deficiency but also when considering the form and dose for any fortification strategies. However, there is a need to understand more about the potentially different pathways through which the two forms are metabolised, and how the body responds to the two forms. Do the two forms have different impacts on health beyond effects on total 25(OH)D concentration?

Points made in the discussion were the inherent biochemical differences in the two forms of vitamin D (e.g. molecular weight) despite the fact that they are still compared on a microgram basis, the need to explore in more depth the relative health benefits of vitamin D₂ and D₃, and the need for further modelling work to establish the amount of vitamin D₂ that would be required as a fortificant to provide equal efficacy with D₃ in achieving a target 25(OH)D status, for example to inform fortification strategies that seek to use a plant-derived form of the vitamin.

The large D₂-D₃ trial described by Tripkovic et al. (2017)^{(Reference Tripkovic, Wilson and Hart78)} also provided a unique opportunity to evaluate whether the two forms of vitamin D influenced global gene expression within whole blood across the 12-week intervention period; blood samples were collected throughout the study for later transcriptome analysis. Professor Colin Smith (formerly of University of Surrey and now at University of Brighton) presented novel results from the transcriptomic analysis of ninety-eight participants (representing equal numbers of D₃-supplemented, D₂-supplemented and placebo-treated individuals). While many changes in gene expression were found between samples taken at baseline and after 12 weeks, they differed between the treatment groups. Although there was some overlap in changes in gene expression between the D₂- and D₃-treated groups, most changes were either specific to the D₃-treated or the D₂-treated subjects. These surprising observations raise the possibility that the physiological effects of vitamins D₃ and D₂ may not be identical in humans^{(Reference Durrant, Bucca and Heskith80)}.

Alternatives for improving vitamin D status: oral spray

Dr Pamela Magee (Ulster University at Coleraine) provided an overview of a commercially available oral spray containing vitamin D and the findings from two human studies investigating its safety and efficacy. The oral spray is designed to be administered by the user under the tongue or on the inside of the cheek to allow rapid absorption via the buccal mucosa, sublingual mucosa and palatal membranes of the oral cavity. It is considered to be advantageous for individuals with intestinal malabsorption (e.g. patients with Crohn’s disease) or those who may have difficulty swallowing (e.g. babies or some elderly patients in care). In the first study, healthy participants (n 22; aged 18+ years) were recruited to receive the vitamin D₃ oral spray and vitamin D₃ capsules at a dose of 75 µg/d (3000 IU), each for a period of 4 weeks, in a randomised crossover design, with a 10-week washout period^{(Reference Todd, McSorley and Pourshahidi81)}. The intention of this trial was to compare the efficacy of the spray v. capsules in increasing vitamin D status. At baseline, approximately 18 % of the study cohort had a 25(OH)D concentration < 30 nmol/l (9 % < 25 nmol/l). Mean 25(OH)D concentration increased by 51 % following supplementation with vitamin D₃ capsules (60 nmol/l to 90 nmol/l) and by 44 % after using the vitamin D₃ oral spray (60 nmol/l to 86 nmol/l). No evidence of hypercalcaemia was reported, indicating that the dose and duration for both methods of supplementation were considered safe. No statistically significant difference between the two supplementation groups was found, in contrast with another study conducted in India, which compared a lower dose (1000 IU/d) vitamin D spray with capsules, and found that the spray raised 25(OH)D concentration significantly more than the capsules^{(Reference Satia, Mukim and Tibrewala82)}. These contrasting findings might be due to differences in study design (sample size, dose and washout period), lower intestinal absorption and membrane permeability in Asian compared with White European participants^{(Reference Menzies, Zuckerman and Nukajam83)}, or genetic variation in vitamin D receptor polymorphisms known to influence vitamin D metabolism^{(Reference Uitterlinden, Fang and Van Meurs84)}. In the study of Satia et al. ^{(Reference Satia, Mukim and Tibrewala82)}, 7 % of the cohort of healthy individuals had plasma vitamin D concentrations < 25 nmol/l at baseline, which was similar (9 %) in the study of Todd et al. (2016)^{(Reference Todd, McSorley and Pourshahidi81)}.

A second study involving Gaelic footballers (mean age 20 (sd 2) years) compared the same vitamin D₃ oral spray used by Todd et al. (2016) in a 12-week randomised double-blinded placebo-controlled trial conducted during wintertime^{(Reference Todd, McSorley and Pourshahidi85)}. At baseline, 22 % of participants had a 25(OH)D status < 30 nmol/l. The mean 25(OH)D concentrations in the intervention (47·4 (sd 13·3) nmol/l; n 22) and placebo groups (43·1 (sd 22·0) nmol/l; n 20) were below 50 nmol/l. Post-intervention 25(OH)D concentrations in the vitamin D₃ oral spray group (83·7 (sd 33·0) nmol/l) were significantly greater than in the placebo group (49·2 (sd 25·4) nmol/l), and there was a significantly greater increase in 25(OH)D concentrations (change from baseline: 36·31 (sd 32·34) v. 6·11 (sd 23·93) nmol/l, respectively). Compliance and acceptance of the mode of application appeared good.

The risk of excessive use (especially with high-dose sprays intended for a single spray dosing regimen) was discussed. Avoiding excessive intake by over-use of the spray would be a particularly important challenge; clear guidance and education on the frequency of use would be critical for all age groups but especially in younger and older populations. Certainly, the above referenced studies showed the spray to be safe and effective. Of specific relevance, the spray may be of key benefit for those with gastro-intestinal malabsorption or swallowing difficulties. The balance between potential difficulties in using a spray for older people with compromised hand movement (grip strength, dexterity) versus the advantage of avoiding adding supplements to the existing medication requirements of care home residents was also discussed.

In summary, oral spray vitamin D₃ appears to be as effective as capsule supplementation at increasing total 25(OH)D concentrations in healthy White UK adults. Further research is needed to confirm the efficacy of vitamin D oral spray in comparison with capsules in other ethnic groups.

Considerations for the implementation of fortification strategies by industry

If the current UK government vitamin D policy, which includes the use of vitamin D supplements during the winter months (and all year for at-risk groups, including those unable to spend time outdoors), fails, other approaches to increase vitamin D intake will be needed urgently. The final session focused on ‘real-world’ issues associated with implementing food fortification strategies. As with all food fortification programmes, there is the risk of excess nutrient intake, but the success of Finland and other countries in their national vitamin D food fortification strategies is a good demonstration that it can be done safely and effectively and make a real difference to population intake^{(Reference Jaaskelainen, Itkonen and Lundqvist71)}.

Gael Delamare (Campden BRI) provided an overview of the technical and consumer considerations associated with potential fortification vehicles for vitamin D, noting that choice of vehicle (and form of the vitamin) may allow targeting to particular consumer groups, according to age, dietary preferences, region or other characteristics. Factors to be considered include: consumer acceptability and preference; shelf-life of the fortified products and stability of the fortificant over time; and the relative healthiness of the food vehicle (to avoid a fortificant being added to a food not considered healthy). To ensure consumer acceptability, changes to the sensory characteristics (e.g. taste, texture and smell) need to be minimised. Processing using heat may affect stability, and different ways of cooking foods (e.g. boiling, frying or baking) may also impact on the final vitamin D content and bioavailability in the prepared food. The ability of a particular approach to raise vitamin D status will be dependent on the amount of the food consumed and on any food matrix components (e.g. fibre or lipids) that affect absorption. The use of UV light to irradiate bread for 1–5 s prior to packaging was given as an example of a means to use processing to enhance vitamin D₂ content; this technique has been approved for use by EFSA⁽⁸⁶⁾.

Another consideration is the legal requirements for foods carrying a nutrition claim. Regulation specifies the amount required to be present in order to permit a claim (15 % of the reference intake, RI (5 µg/d) per 100 g, per 100 ml or per portion of food, or 7·5 % of the RI per 100 ml for beverages). Challenges associated with this are discussed below.

Measuring the vitamin D content of foods presents a challenge in itself as vitamin D is present only in small (µg) quantities. The process includes extraction of vitamin D from the food matrix, followed by analysis using the preferred LC MS/MS method, which is time (about 60 min per sample) and labour intensive.

In summary, consumer attitude is of importance in the choice of a fortificant and food vehicle. Currently, there are no guidelines on the most suitable processes for each food vehicle, e.g. acceptable levels for fortification. Processing can help increase vitamin D concentrations in foods (e.g. UV irradiation). However, bioavailability of vitamin D from different matrices needs further investigation to identify the most effective vehicles. Quantifying the concentration of vitamin D in the food measures the amount present, not the amount available for absorption. Labelling requirements present a challenge.

General discussion

Several themes emerged from the discussion, which are summarised below.

If over time, it transpires that there has been insufficient improvement in the nation’s vitamin D status, what might plan B look like?

Current UK government policy refers to the role of sunlight exposure during the summer months and foods as a source of vitamin D but places emphasis on vitamin D supplementation especially during the winter months and all year round for groups considered to be at risk of low status (see Table 4). Given the limited impact of advice on use of folic acid supplements in the UK and concerns about excessive sun exposure in terms of skin cancer risk, improved understanding of dietary sources of vitamin D and also, potentially, a fortification strategy are the obvious options for improving vitamin D status.

There was discussion about the types of food that might be considered appropriate for fortification. There is some important learning from the RCTs discussed above. The importance of real-world considerations was emphasised, in particular attention given to foods consumed regularly by groups who may be at-risk of low vitamin D intakes, and whether it is inappropriate to add vitamin D to foods that do not contain it naturally, which would severely restrict the options available. Previous modelling studies have focused on bread flour, given the widespread consumption of bread and current UK policy for mandatory addition of several nutrients to bread flour. Discussion focused on whether fortification of wheat flour alone would reach the most ‘at-risk’ groups (who may consume other flours) and the practical and processing challenges already evident from discussions around proposals to add folic acid to bread flour.

Examples were given of vitamin D fortification approaches in Finland, where fortification was ‘promoted’ by the government, and in the USA, where fortification was agreed as a result of coordination between food industry stakeholders. If fortification was to be agreed by industry stakeholders as a desirable approach, then appropriate messaging (e.g. via social media) might then be determined. Another example concerned mandatory fortification of wheat flour in Mongolia (with multiple micronutrients) triggered by the incidence of rickets. An initial public survey prior to fortification indicated that 55 % of urban and rural Mongolians favoured fortification, but after learning about the importance of vitamin D for the prevention of rickets, the percentage favouring fortification increased to 75 %^{(Reference Bromage, Gonchigsumlaa and Traeger87)}.

The discrepancy between the EU labelling requirements for vitamin D (RI of 5 µg/d) and the UK RNI of 10 µg/d (400 IU/d) was discussed as a potential barrier to voluntary fortification by companies. Industry nutritionists considered that whilst it is not a barrier per se, it compromises clear communication. Nevertheless, it has been estimated that if everyone consumed 5 µg/d, then it would reduce the population prevalence of a 25(OH)D concentration < 25 nmol/l to <10 %^{(Reference Cashman, Ritz and Kiely88)}. So, to aim for consolidating current intakes (shown in Table 3) to this level could be a starting point. An Industry representative suggested that establishing a requirement for mandatory fortification would create a more level playing field for all consumers, as well as for industry.

Need for real-world information

To move things forward, the need for ‘real-world’ information on the economic implications of vitamin D fortification for industry was stressed. This is key in order to explore fortification as a viable solution to increasing vitamin D intakes, along with a more collaborative working relationship between the research community, policymakers and industry. For food businesses, the cost of fortification is influenced by: (i) the type of fortificant; (ii) the point at which it is added to the product; (iii) the need for quality control; (iv) the volumes being purchased; (v) the product category and (vi) the profit margin between and within different food categories. For the breakfast cereal industry, vitamin D is one of the most expensive fortificants currently added because of the cost of the technologies required to add the vitamin, rather than the cost of the vitamin per se. However, the cost will be dependent on the nature of the technology needed and whether it is D₂ or D₃ that is added. For example, the cost of adding vitamin D is less for dairy-based foods and so is not perceived to be a major economic hurdle.

For those researchers engaged in modelling, estimations of costs rely on published sources of information where they exist. Examples used were 25(OH)D biofortification of eggs, in which the cost per dozen eggs was deemed minimal (approximately 0·1 p per dozen eggs) and vitamin D fortification of flour, estimated as approximately 1 p per tonne.

Experience from ODIN has demonstrated that modelling the impact of flour fortification is difficult, due to imports/exports and the variety of products in which flour is used. In modelling the effects of fortification, the distribution of intakes was more important than the lower and upper extremes of intake, and the potential use of supplements has to be taken into account.

A need was identified for fortification models that could be responsive to changing patterns of consumption, to ensure that they reflect changes in the level of intake and sources of vitamin D.

How can consumer understanding about vitamin D be improved?

This topic attracted considerable debate and discussion that focused on articulating vitamin D content of foods in a simple, unambiguous way that is also compliant with regulations. It was agreed that communication to consumers remains difficult for the reasons outlined above. Linked with this was the need to understand better the public’s knowledge about the relevance of adequate vitamin D status for health, how this might be achieved, and the role of vitamin D supplementation in achieving this goal. There is a need for such insight for the health and social care community as well as the general public. The Forum took place before the COVID-19 pandemic but the importance of improving the public’s understanding about vitamin D has subsequently been highlighted because opportunities for synthesis of vitamin D following sunlight exposure will have been reduced for those people spending more time indoors during the lockdown.

Greater understanding about where members of the public obtain scientific information might help inform which communication channels may be best utilised. In principle, multiple media avenues should be considered, including for example having a well-known person with a track record of science advocacy to champion and communicate the message. Consideration was also given to what might ‘capture the imagination’ of the public, and whether this might require using a high profile case of a vitamin D deficiency-related death to help people recognise the importance of vitamin D to their own health.

It was agreed that a multi-pronged approach may be needed, involving the food and health sectors, government and other stakeholders, and that there is not necessarily a universal solution. In particular, the benefits of some form of code of practice developed by food manufacturers, ingredients suppliers and retailers in association with other stakeholders were discussed. The code of practice might provide guidance on the most appropriate foods, vitamin D sources and fortification levels and suggest strategies for specified food categories.

The role of the supplements industry was also discussed. Currently, many popular supplements contain more than the recommended 10 µg/d. High intakes of vitamin D through consumption of fortified foods alone are highly unlikely to create a problem in relation to tolerable upper intake levels, and the modelling conducted in the ODIN project indicates that even those people taking high dose supplements (50 µg/d), in addition to fortified foods, would be unlikely to reach the tolerable upper level of intake.