Is Hypovitaminosis D one of the Environmental Risk Factors for Multiple Sclerosis?

Charles Pierrot-Deseilligny; Jean-Claude Souberbielle


Brain. 2010;133(7):1869-1888. 

In This Article

Physiology of Vitamin D

Metabolism and General Effects of Vitamin D

Great advances have recently been made in our knowledge of the physiology of vitamin D (Borradale and Kimlin, 2009; Adams and Hewison, 2010). There are two forms of vitamin D: vitamin D3 (cholecalciferol) i.e. the animal or human vitamin D, and vitamin D2 (ergocalciferol), which is of plant or mushroom origin. Vitamins D2 and D3 are both available in dietary form but only vitamin D3 is synthesized in the skin by ultraviolet B (UVB) radiation from sunlight. Vitamin D and its metabolites are transported in the plasma, bound to the vitamin D binding protein. Vitamin D is transformed in the liver into 25-hydroxyvitamin D [25(OH)D], which is regulated by the supply of synthesized and ingested vitamin D. Under stimulation by parathyroid hormone, this metabolite is transformed in the renal proximal tubule to form 1,25-dihydroxyvitamin D [1,25(OH)2D], which is the active metabolite. 1,25(OH)2D is released into the bloodstream with a half life of several hours, binds to vitamin D receptors in its target tissues and is considered a 'hormone' (Adams and Hewison, 2010). Vitamin D receptors are present not only in the intestine, bone and kidney i.e. the classical target tissues of vitamin D, but also in gonads, breast, pancreas, cardiovascular system, brain (microglia) and circulating immunity cells i.e. macrophages, monocytes and activated lymphocyte T and B cells (Bhalla et al., 1983; Vedman et al., 2000; Mathieu and Adorini, 2002; Holick, 2004; Lips, 2006; Chen et al., 2007; Holick, 2008a, b). All these 'non-classical' target tissues are able to transform 25(OH)D into 1,25(OH)2D, which exerts autocrine/paracine effects within these cells and, possibly, neighbouring cells. The physical link between vitamin D and the basic cells of immunity is of particular interest given the potential immunological role of this vitamin in autoimmunity in general and in multiple sclerosis in particular. Furthermore, single nucleotide polymorphisms of the CYP27B1 and the vitamin D receptor genes influence the metabolism and effects of vitamin D (Uitterlinden et al., 2004) and the risk of multiple sclerosis (Niino et al., 2000; Tajouri et al., 2005; Orton et al., 2008; Smolders et al., 2008a; Torkildsen et al., 2008; Dickinson et al., 2009; Smolders et al., 2009a). Vitamin D binding protein is also genetically influenced, which affects 25(OH)D concentrations (Bouillon et al., 1981; Sinotte et al., 2009; Ahn et al., 2010) and may potentially affect the risk of multiple sclerosis. Finally, a considerable body of literature published during the last 10 years, comprising multiple intervention studies on the effects of vitamin D in bone pathology and numerous association studies on non-classical effects of this vitamin in other organs and pathologies, has revolutionized our knowledge of vitamin D (Holick, 2004, 2007; Vieth, 2007; Cannell et al., 2008; Kimlin, 2008; Borradale and Kimlin, 2009; Bischoff-Ferrari, 2010). The main non-classical effects of vitamin D [via vitamin D receptors and 1,25(OH)2D] appear to be anti-inflammatory, anti-infectious, immunomodulatory, antiproliferative and as a neurotransmitter involving not only many autoimmune diseases—including, among others, multiple sclerosis (see below), type 1 diabetes (Mathieu et al., 2005; Forouhi et al., 2008; Zipitis and Akobeng, 2008; Danescu et al., 2009), rheumatoid arthritis (Merlino et al., 2004; Patel et al., 2007) and systemic lupus erythematosus (Amital et al., 2010)—but also some cancers, in particular colon and breast cancer (Lappe et al., 2007; Abbas et al., 2008; Chen et al., 2009; Yin et al., 2009; Jenab et al., 2010; Kawase et al., 2010), diseases of the cardiovascular system (Dobnig et al., 2008; Forman et al., 2008; Wang et al., 2008), infection (Nnoaham and Clarke, 2008; Ginde et al., 2009b; Urashima et al., 2010; Youssef et al., 2010) and other general symptoms such as muscle weakness and falls (Bischoff-Ferrari et al., 2004a, 2009a; Zhu et al., 2006; Broe et al., 2007; Pfeifer et al., 2009).

Optimal Serum Levels of Vitamin D

25(OH)D, with a half life of several weeks, is representative of the overall vitamin D store in the body (D2 + D3) and is, therefore, the serum component that must be measured to evaluate vitamin D status (Heaney, 2000; Zitterman, 2003; Souberbielle et al., 2008; Zerwekh, 2008). There is not yet a standardized 25(OH)D assay, but according to the UK-based Danish External Quality Assessment Scheme (DEQAS), the main methods give roughly similar mean results, differing by not much more than 7% (Carter et al., 2010). Assays measuring both 25(OH)D2 and D3 (Cavalier et al., 2008) are recommended. According to review and position papers published by many experts during the last decade, the minimum 25(OH)D serum level required to achieve an optimal vitamin D status would be somewhere between 50 and 100 nmol/l (i.e. 20 and 40 ng/mg), though the minimum level most frequently recommended is ~75–80 nmol/l (Lips, 2001; Zitterman, 2003; Holick, 2004; Dawson-Hughes et al., 2005; Hollis, 2005; Vieth, 2005; Biscoff-Ferrari et al., 2006, 2009b; Roux et al., 2008; Souberbielle et al., 2008; Adams and Hewison, 2010; Dawson-Hughes et al., 2010). This limit is not 'population-based', since this has no real sense in countries with limited sunshine, but has been determined using 'health-based reference values' i.e. from both metabolic and pathological bases regarding various outcomes that can grossly be separated into 'bone/calcium-related' and 'not bone/calcium-related' outcomes. When considering 'bone/calcium-related' outcomes, the threshold of 75 nmol/l corresponds to the serum level below which (i) parathyroid hormone secretion is generally stimulated by the lack of vitamin D (Chapuy et al., 1996; Holick, 2007; Durazo-Arvizu et al., 2010); (ii) initial signs of mineralization defect are observed (Premiel et al., 2010); and (iii) calcium absorption by the gut is not yet optimal (Heaney et al., 2003b). Other recent original findings suggest that peak bone density in young adults becomes optimal when 25(OH)D is above the level of 90 nmol/l (Bischoff-Ferrari et al., 2004b) and a recent meta-analysis of 12 placebo-controlled randomized controlled trials concluded that non-vertebral fracture prevention in patients aged 65 and older was optimal in trials with mean 25(OH)D serum levels of 75–110 nmol/l (Bischoff-Ferrari, 2009c). Furthermore, due to a progressive decrease in sensitivity to 1,25(OH)2D and also in the capacity of the kidney to hydroxylate 25(OH)D into 1,25(OH)2D, the minimum optimal 25(OH)D level for bone health probably varies with age and should be higher in the elderly than in the young, namely at least above 75 nmol/l in the former (Baraké et al., 2010; Dawson-Hughes et al., 2010; Whiting and Calvo, 2010). For the 'non-calcium/bone' endpoints, the minimum 25(OH)D target levels are not yet well determined since large randomized controlled trials are still lacking. However, a multitude of epidemiological (association) studies, for example in the cancer and the cardiovascular fields that cannot be reviewed in detail here (see above), suggest a protective effect of vitamin D in people with relatively high 25(OH)D serum levels (usually above 75 or 100 nmol/l) compared to people with low serum levels (usually between 20 and 40 nmol/l) (see Bischoff-Ferrari et al., 2009b). Accordingly, an absolute consensus does not yet exist on the recommended minimum level of 25(OH)D, since some authors recommend a minimum level of 50 nmol/l (Lips et al. , 2009), whereas others argue in favour of at least 80 or 100 nmol/l (Zitterman, 2003; Holick, 2004; Hollis, 2005; Bischoff-Ferrari et al., 2006; Vieth et al., 2007; Niino et al., 2008; Bischoff-Ferrari et al., 2009b, c). However, the question of what 25(OH)D serum level should be defined as the minimum needed to achieve an optimal vitamin D status i.e. between 50 and 100 nmol/l depending on the authors, does not radically change the general problem of vitamin D insufficiency, since currently between a third and a half of the 'normal' population in temperate countries do not even reach the threshold of 50 nmol/l (Mithal et al., 2009; Adams and Hewison, 2010) (see below, 'Vitamin D status in the general population'). Concerning the upper limit for the reference values of 25(OH)D serum level, it must be mentioned that physiologically, in outdoor workers, the serum level is generally between 75 and 175 nmol/l (rarely exceeding 200 nmol/l) (Haddad and Chyu, 1971; Haddock et al., 1982; Barger-Lux and Heaney, 2002), and there is no true risk of vitamin D intoxication up to 375 nmol/l (Hathcock et al., 2007; Burton et al., 2010); this represents a considerable safety margin in cases of simple vitamin D supplementation, assuming a target serum 25(OH)D level between 75 and 125 nmol/l i.e. around 100 nmol/l on average (Bischoff-Ferrari et al., 2009b).


On the basis of these recent metabolic and pathological findings, the daily requirement of vitamin D has been reassessed and is now thought to be far higher than the 300–400 IU/day that, until a few years ago, was estimated to be sufficient. The daily requirement does of course depend on what the optimal target 25(OH)D serum level is considered to be: for a 25(OH)D serum level of 50 nmol/l, 800 IU/day of vitamin D appears sufficient, but to bring most people above the 75 nmol/l level, a dosage of between 1000 and 4000 IU/day (depending upon the individual, but on average 2000 IU/day) is required (Heaney et al., 2003a; Grant and Holick, 2005; Hollis, 2005; Bischoff-Ferrari et al., 2006, 2009b, c; Vieth, 2006; Heaney et al., 2009; Hall et al., 2010; Schwalfenberg et al., 2010; Whiting and Calvo, 2010). Vitamin D intake via (unfortified) food is very marginal in normal Western diets, even in those considered well balanced, and generally provides < 100 IU/day. Even diets that include oily fish, as in traditional Scandinavian food (Mark and Carson, 2006; Kampman and Brustad, 2008), or fortified food (Calvo et al., 2004; Moore et al., 2005; Välimäki et al., 2007; O'Donnell et al., 2008; Vatanparast et al., 2010), rarely exceed a few hundred IU/day and this usually remains markedly below the daily requirement. Sunshine therefore remains the principal natural source of vitamin D, providing 80–90% of the requirement in the absence of fortified food. Although exposing a part of the body (for example the face, trunk and arms) to the sun in summer can provide 10 000 IU of vitamin D in less than half an hour, this supply disappears within a few weeks and cannot readily be replenished throughout the year except in tropical countries (Vieth, 1999; Hollis, 2005; Vieth, 2005; Diffey, 2010). Moreover, elderly and dark-skinned subjects are less able to synthesize vitamin D than young, light-skinned subjects who, if they protect themselves too much from the sun (by clothing or sun-block), may also rapidly find themselves in a state of vitamin D insufficiency (Vieth, 1999; Armas et al., 2007; Binkley et al., 2007).

Geography and Sunshine

Accordingly, a major problem of vitamin D supply exists for many populations, namely those who live beyond the 40th parallels North or South (Holick, 2004, 2008a, b; van Amerongen et al., 2004) (Fig. 2). These geographical parallels mark the line at which the sun at its zenith becomes seasonally so low that for ~4 months of the year UVB levels are insufficient to synthesize vitamin D (Webb et al., 1988). At even higher latitudes, periods without a solar source of vitamin D may reach 6–8 months per year. By contrast, at low latitudes, in particular between the tropics, there is no problem with sunshine. However, it should also be taken into account that UVB is only available for a few hours a day, either side of mid-day i.e. the period during which we are currently advised by dermatologists to limit exposure to the sun, of course for excellent dermatological reasons (Diffey, 2010). Be that as it may, relatively limited amounts of sunshine mainly concern Canada, the Northern half of the USA, almost all of Europe (the 40th parallel passing through the middle of Spain), Russia and a few areas in the Southern hemisphere, such as New Zealand, Tasmania and Patagonia i.e. involving only ~15% of the world's population, the remaining 85% live in regions well endowed with sunshine (Fig. 2). It is well known that, except for Patagonia (Melcon et al., 2008), which is sparsely populated, the regions with limited amounts of sunshine are also those with the highest prevalence of multiple sclerosis (Goodin, 2009) (Fig. 2), even if other environmental risk factors may also be involved in these countries.

Figure 2.

Geographical and historical considerations on vitamin D and multiple sclerosis. MS = multiple sclerosis.


A brief look at the history of humanity suggests that two main events may have been important for vitamin D, the first extremely old and the second quite recent. The first event happened ~1 million years ago when Homo erectus began to migrate from their birthplace in East Africa to Northern regions of the globe, with a much less sunny climate (Fig. 2). In 10 000 centuries, the Homo erectus family and their descendants have had sufficient time to evolve and adapt, in a Darwinian sense, to limited sunshine (Jablonski and Chaplin, 2000, 2010). Evolving into Homo sapiens, humanity has undergone many changes, but one of the most visible alterations in Northern people has been the lightening of their skin (Diamond, 2005; Vieth, 2006; Yuen and Jablonski, 2010). Light skins are remarkably effective at synthesizing vitamin D with only small amounts of sunshine, being about five times more efficient in this respect than dark skins. Even so, light skins still have to be exposed to sunshine in order to synthesize vitamin D. This consideration leads to the second historical event of importance relating to vitamin D, i.e. the so-called 'industrial revolution', which happened only a few generations ago. During the second half of the 19th century, many people in what are now developed countries left an essentially rural way of life—in which they were almost constantly exposed to nature, the climate and sunshine—to colonize towns and live and work indoors. The result has probably been a drastic fall in vitamin D levels, without the possibility of any physiological adaptation in such a short-time scale (Vieth, 2006). Devastating epidemics of rickets were observed in the main Northern industrial cities (e.g. London, Paris, New York), in which it is estimated that ~80% of children were to some extent affected at the end of the 19th century (Hess and Hunger, 1921; Holick, 2007). It was not understood until the beginning of the 20th century that rickets was caused by a lack of sunshine and vitamin D itself was not formally identified until the early 1930s.

Nowadays, in countries with limited sunshine, paediatricians usually prescribe a vitamin D supplement for infants to prevent rickets and geriatricians prescribe it for the elderly to reduce the risk of falls, fractures and osteomalacia. However, nothing is usually done for people between these two extremities of life, although such age groups are just as lacking in vitamin D as infants and the elderly, as shown by recent epidemiological studies (see below). Although there are no apparent bone stigmata suggesting a lack of vitamin D in all these intermediate age groups, a chronic vitamin D insufficiency could have pernicious delayed effects on the development of osteoporosis and a wide range of serious diseases. Therefore, during the past few years, a growing part of the medical community has advocated a systematic supplementation, at least during winter, for the general population living in temperate or Nordic countries (Holick, 2004; Hollis, 2005; Vieth, 2006; Binkley, 2009; Cavalier et al., 2009; Edlich et al., 2009; Grant et al., 2009; Stechschulte et al., 2009; Gillie, 2010; Zittermann et al., 2010). To sum up these historical aspects, it took almost a century to understand that rickets observed in infants in Northern industrial countries was due to vitamin D deficiency, and it has now taken almost another century to realize that all age groups in these countries suffer from a lack of vitamin D.

Vitamin D Status in the General Population

As a probable result of these diverse physiological, geographical and historical considerations, recent epidemiological studies in temperate countries (mainly beyond the 40th parallels) on the adult population (>15–18 years, involving both genders and mostly Caucasian people) have shown that serum levels of 25(OH)D are low, whatever the assays used. For example, in the USA, the mean serum level of 25(OH)D was 74 nmol/l in a large cohort of 15 000 adults (over 18 years) distributed throughout the country and studied between 1988 and 1994, with samples collected all year round (Zadshir et al., 2005). However, in a more recent analogous American cohort of 20 000 adults studied between 2000 and 2004, the mean serum level was 60 nmol/l, which suggests, after accounting for assay differences, a global decrease of ~10 nmol/l in 10 years (Looker et al., 2008; Ginde et al., 2009a). This marked and rapid decrease has mainly been attributed to an increase in the degree of urbanization and in body fat. In the UK, the mean serum level of vitamin D was 51 nmol/l (with 41–60 nmol/l from winter to summer) in a cohort of 7437 British adults, who were 45 years old in 2003, with a North–South gradient existing within the results (Hyppönen and Power, 2007). The authors concluded that there was an urgent need for preventive action. Similarly, low mean serum levels of 25(OH)D were recently reported in normal adults in Australia (51–75 nmol/l depending upon the region, skin colour and lifestyle; van der Mei et al., 2007b), Canada (67 nmol/l; Langlois et al., 2010), New Zealand (mean = 50 nmol/l, with 32 nmol/l in winter and 74 nmol/l in summer; Rockell et al., 2006) and Germany (42 nmol/l in winter and 67 nmol/l in summer; Scharla et al., 1996), with, therefore, serum levels usually 20–40 nmol/l lower in winter than in summer in these countries. In France, a study involving 1579 adults in nine different regions during the winter of 1994–95, found a mean serum level of 61 nmol/l and a North–South gradient (Chapuy et al., 1996); with serum levels of 40–50 nmol/l in the North and 80–90 nmol/l in the South (Fig. 4A). Significant correlations existed in this study between the regional serum levels of vitamin D and both latitude (r = –0.79; P < 0.01) and the annual local amount of sunshine (r = 0.72; P = 0.003) (Fig. 3, links B–C and A–C).

Figure 3.

Environmental climatic risk factors for multiple sclerosis and links between them. The r- and P-values illustrate the example of France and correspond to the Pearson correlation tests reviewed in this article or performed by the authors, based on data for French regions concerning (A) mean latitude, (B) mean global annual sunshine (Suri et al., 2007), (C) mean serum level of vitamin D in normal adults (Chapuy et al., 1996) and (D) multiple sclerosis prevalence in French farmers (Vukusic et al., 2007); r and P in black = data from 22 regions; r and P in red = data from nine regions. Modified from Pierrot-Deseilligny (2009).

Figure 4.

Epidemiological studies on multiple sclerosis prevalence, exposure to sun and serum levels of vitamin D in normal adults in the administrative regions of France. (A) Map of France showing the 22 administrative regions, figures for regional multiple sclerosis prevalence (per 100 000 inhabitants) in the farmer population (Vukusic et al., 2007), the average annual amount of global solar irradiation (yellow spots) per region determined from European environmental data (Suri et al., 2007) and the average serum vitamin D levels in normal adults (red bars) per region from Chapuy et al. (1996). Vukusic et al. (2007) divided France into three main zones of regions (various shades of blue) and showed that a significant gradient existed between the North–East, intermediate and South–West zones in terms of regional multiple sclerosis prevalence. (B) Correlation performed by the authors of the present article between the regional multiple sclerosis prevalence in French farmers (Vukusic et al., 2007) and the average global annual (between 1981 and 1990) solar irradiation in the French regions (Suri et al., 2007), expressed in KWh/m2, using the Pearson test. This correlation is highly significant; note also that the three main zones of regions identified by Vukusic et al. (2007) are still relatively distinct in this comparison (ellipses). (C) Correlation performed by the authors of the present article between the regional multiple sclerosis prevalence in French farmers (Vukusic et al., 2007) and the average serum levels of vitamin D in normal adults living in nine roughly analogous French regions (Chapuy et al., 1996), using the Pearson test (modified from Pierrot-Deseilligny, 2009); this correlation was also significant and the three main zones of regions (as in A and B) were still relatively distinct in this comparison (ellipses). MS = multiple sclerosis.

On a world-wide scale, in a meta-analysis based on 394 studies, a significant correlation existed between 25(OH)D serum levels and latitude in Caucasian subjects (Hagenau et al., 2009). In another meta-analysis, involving Europe and Asia, the factors affecting the 25(OH)D serum levels in adults were (i) age, the synthesis of vitamin D being less efficient in older people; (ii) gender, women generally having lower levels than men; (iii) skin colour, dark skins synthesizing vitamin D less efficiently than light skins; (iv) type of clothing and the extent to which it covers the body; (v) food, whether or not supplemented with vitamin D; and, most importantly, (vi) the degree of urbanization, with nowadays less and less time spent outdoors with exposure to sun (Lips, 2007). In Nordic countries, the serum levels of vitamin D are often lower than those of temperate countries (Välimäki et al., 2004; Andersen et al., 2005), whereas in tropical regions the serum levels are generally higher (Linhares et al., 1984; Chailurkit et al., 1996; Ho-Pham et al., 2010). However, frequent exceptions may be observed to these main trends due to differences in lifestyle or diet with, for example, the possibility of low serum levels of vitamin D in people of sunny countries, if they avoid the sun or, conversely, relatively high serum levels in people of Northern regions, who may take more advantage of the sun in summer and partly compensate the lack of sunshine by a diet rich in vitamin D in winter (van der Wielen et al., 1995; Lips et al., 2001; Lips, 2010). Accordingly, in temperate and Nordic countries, 50–90% of the general population (depending on the cut-off <50 or 75 nmol/l) are more or less permanently in a state of vitamin D insufficiency, a situation that cannot be ignored, whatever the cut-off considered.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: