Introduction

Vitamin D status in an individual is dependent on numerous genetic, lifestyle and geographical factors that include age, gender, skin pigmentation, sunlight exposure, latitude, the use of sunscreen, dietary habits and supplement intake [1]. It is best measured by the serum concentration of 25-hydroxyvitamin D levels, also known as 25(OH)D levels [2].

Very low levels of 25(OH)D have been documented in different subgroups of the population worldwide [1, 3–6], which have clinical implications. Vitamin D plays an important role in skeletal growth and development, and in bone health throughout life. It promotes calcium absorption [7] and reduces bone loss through the regulation of parathyroid hormone levels [8]. As a consequence, vitamin D deficiency has been linked to reduced bone mineral density [9, 10] and higher risk of osteoporotic fractures [11]. Although further investigation is necessary, vitamin D supplementation may reduce the risk of other diseases, such as colorectal cancer [12], diabetes [13] and infection [14], and it may help decrease fractures and falls [15, 16]. The loss of muscle mass and strength observed in vitamin D-deficient individuals puts them at higher risk of falls and, therefore, fragility fractures.

The International Osteoporosis Foundation took the initiative to describe the vitamin D status in the general population in different countries based on a systematic review and to present the data on a global map. The aims of the study were to provide a general overview of vitamin D status in countries for which data were available, examine the existing heterogeneities in vitamin D status, and identify research gaps.

Methods

The data used in this project are based on a systematic literature review conducted by the Mannheim Institute of Public Health, Germany. The methods used generally follow the PRISMA Statement (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) [17]. Here, we provide a short summary of the methods used in this review. The methods are described in more detail elsewhere [18].

Eligible criteria

A systematic search was conducted in PubMed/Medline and EMBASE to identify articles on vitamin D status in the global population. Eligible studies included samples of randomly selected persons from the general population in countries throughout the world. The outcome of interest was the mean or median 25(OH)D level measured in serum or plasma. There were no limitations based on the type of assay used. Studies were required to have a cross-sectional design or to include a population-based cohort. Other study types like clinical trials, case–control studies, case reports or case series, reviews or qualitative studies were excluded. Articles had to be written in English and published between 1st January 1990 and 28th February 2011.

Abstract and full publication screening

Two thousand five hundred sixty-six articles were identified from both databases. Two independent researchers screened the articles for excluding studies, with a good agreement (kappa coefficient, 0.719). Disagreements were discussed and resolved. After review, 273 articles were eligible and included in a large database which provided, in part, the following information: the mean or median 25(OH)D levels, population characteristics, study location, assay type, number of participants and age groups.

Data filtering

In a second review process, studies on institutionalized elderly only, those on newborn babies and those having an age range that largely overlapped the two age categories (1–18 and >18 years) were removed. When published repeatedly, the same cohort was not presented more than once. In contrast, studies reporting a sub-analysis of a cohort (number of participants and age range different from the root paper) were retained. Studies originating from England, Northern Ireland and Scotland were grouped as United Kingdom. In the end, our database included 200 studies from 46 countries.

After an examination of the database, two different age categories were selected: children and adolescents (1–18 years) and adults (>18 years). The mean serum or plasma 25(OH)D levels were extracted and reported as gender-specific means weighted by the sample size, where possible. The median levels of 25(OH)D were included when the mean levels were not reported in a study. When data were classified by specific seasons, winter values were chosen. Values in nanograms per millilitre were converted to nanomoles per litre by a multiple of 2.496.

Four colour codes according to the mean (or median) 25(OH)D levels were used:

figure a

For each study, the mean or median vitamin D levels (in nanomoles per litre) were reported from the literature and a study colour code was assigned.

Rationale for the colour coding of countries

For both age categories, the rationale for assigning a colour code to a specific country was based on the following hierarchical selection criteria:

  1. 1.

    Representative of the entire country, population-based and based on a weighted mean of these studies

  2. 2.

    Representative of a region/city of the country, population-based and based on a weighted mean of these studies

  3. 3.

    Based on a weighted mean of multiple studies, non-population-based

  4. 4.

    Based on a single study

Country colour was based on the 25(OH)D level (either the weighted mean or median) of one or more representative studies, if available. If not available, it was based on one or more studies fitting the second criterion cited above, and so forth. A study was considered representative if it represented the entire population for a certain age and sex group in a certain country, region or city. Studies with a selection bias, which excluded individuals for example on the basis of health status, ethnicity, physical abilities, language, smokers and social economic status, were classified as non-representative. However, for some studies, such information was not described in the text.

Design of the figures

The software FlashWorldMap.com was used to produce the maps.

Results

This analysis involves 200 studies from 46 countries. Forty-two of the 200 studies (21 %) were considered representative. Details of these studies for each contributing country are provided in Table 1 (children and adolescents) and Table 2 (adults). The largest number of studies was conducted in Europe (48.0 %), followed by North America (27 %) and the Asia-Pacific region (16.5 %). Of the 46 countries contributing data, 20 (43 %) had at least one study that was classified as representative. Figures 1 and 2 show the vitamin D status in children and adolescents, and adults, respectively, in different countries. The countries are colour-coded according to the serum levels of 25(OH)D reported in Table 1 (children and adolescents) and Table 2 (adults), and the ranges of 25(OH)D represented by each colour are described in the two figures.

Table 1 Country colour codes of vitamin D status in children and adolescents
Full size table
Table 2 Country colour codes of vitamin D status in adults
Full size table
Fig. 1
figure 1

Vitamin D status in children and adolescents (1–18 years) around the world when available; winter values were used to calculate the mean 25(OH)D levels

Full size image
Fig. 2
figure 2

Vitamin D status in adults (>18 years) around the world when available; winter values were used to calculate the mean 25(OH)D levels

Full size image

Discussion

This project provides a ‘snapshot’ or summary of the 25(OH)D levels around the globe, as identified in publications since the year 1990. The maps form a core or platform upon which additional information can and should be added. The number of published papers describing 25(OH)D levels is escalating, and the geographic diversity of incoming data is broadening. As a result, we can anticipate having a more comprehensive picture of global vitamin D status in coming years. Updating of the accompanying tables, in which information from each country is ordered chronologically, will also allow for a qualitative assessment of secular trends in 25(OH)D since 1990, within regions and overall. We can expect to see rises in 25(OH)D levels as awareness and concern about vitamin D deficiency grows and as recommendations for vitamin D supplementation appear in more and more government documents, position statements and clinical practice guidelines for bone health [19–21]. This trend would only be accelerated should vitamin D be proven to modify the risk of non-musculoskeletal diseases, such as diabetes, infection or cancer, as has been suggested by many observational studies [22].

Examination of the current maps enables one to identify regions where information on 25(OH)D levels is lacking. The most striking gap is in children and adolescents. The systematic search did not identify studies in this age range in Central America, the northern and central regions of South America, most of Africa, much of Europe and in Australia. This information gap needs attention in view of the importance of vitamin D in bone and muscle growth and development. In regions where data were available, the predominant colour code for children and adolescents was orange, indicating mean 25(OH)D levels in the 25- to 49-nmol/L range. These values are below those recommended by the Institute of Medicine (50 nmol/L), the International Osteoporosis Foundation and the US Endocrine Society (75 nmol/L) [19–21].

Among adults, most regions offer some data, and their colour codes are approximately evenly split between orange (25–49 nmol/L) and yellow (50–74 nmol/L). Areas where information was not identified include Central America, South America (with the exception of Brazil) and much of Africa. With the known role of vitamin D in preserving bone health, it is important to fill these gaps so that appropriate measures can be implemented to correct inadequate 25(OH)D levels. Information gaps in both age groups would ideally be filled with survey data based on random sampling of a country or region. In the meantime, any and all data from specific regions will make some contribution to defining vitamin D status globally.

Despite using data from a systematic literature review, the maps have limitations. One limitation is that adequate information is not available. An extreme example is that for a few countries, one single small study confined to a limited region of the country and to a narrow age range was used to colour the country (e.g. Argentina). This is of course not a complete picture of the country. Other countries have many studies, representative and non-representative. For example, New Zealand is coloured orange based on the one available representative sample of subjects residing in the city of Auckland. As indicated in the table, other studies from this country involving healthy populations and subjects measured in the summer consistently reported higher 25(OH)D levels in the range of yellow and green. In view of the diversity in the quantity and quality of data used in this study, it is important that the tables be used in conjunction with the maps and that the maps are interpreted with caution.

Several limitations are inherent in the way that the published data were presented. For example, 25(OH)D measurements were sometimes made in specific seasons (i.e. winter or summer) and at other times made without reference to season. When the option was available, we used the winter measurement, representing the worst-case scenario, for the map colouration; however, 25(OH)D levels by season, when available, are provided in the table. There was inconsistency across studies in the age groupings such that we were not always able to break out the levels by our predefined age categories of children and adolescents 1–18 years and adults >18 years. Another limitation is that some of the studies represented small regions within large countries with diverse latitudes; thus, they did not fairly represent the whole nation with respect to the contribution of sun exposure (skin synthesis) to 25(OH)D levels. Additionally, information on body size, clothing habits and skin pigmentation was not consistently available.

An important limitation of this project and of any inter-study comparison of 25(OH)D levels is the well-described variability in 25(OH)D assays. Since the first 25(OH)D assay was developed 30 years ago [23], the analytical options have expanded from the original competitive protein binding assay to include radioimmunoassay, chemiluminescent assay, high-performance liquid chromatography and liquid chromatography–mass spectrometry/mass spectrometry. Unfortunately, the serum 25(OH)D levels vary by up to 20–40 %, depending upon which assay is used [24–27]. Part of the variability can be attributed to the fact that not all of the assays detect 25(OH)D2 as effectively as they detect 25(OH)D3. As a result, in those regions where vitamin D2 is used in most supplements, the total 25(OH)D levels will tend to be underestimated.

To address the assay problem, many laboratories around the world participate in a quarterly quality assurance and surveillance program, the Vitamin D External Quality Assessment Scheme, which we strongly encourage. Standard reference material consisting of known amounts of 25(OH)D2 and 25(OH)D3 in human serum is now available through the U.S. National Institute of Standards and Technology (NIST; SRM972, www.NIST.gov/srm). The use of this material should make inter-laboratory comparisons more readily interpreted and allow for the detection of intra-laboratory changes over time. An important initiative, the vitamin D standardization program (VDSP), is currently underway to make the measurement of 25(OH)D accurate and comparable over time, location and laboratory [28]. The first goal of VDSP is to standardize 25(OH)D values currently being measured in national health surveys to the NIST standards. Australia, Canada, Germany, Ireland, Mexico, South Korea, the UK and the USA are participating in this process. A second goal is to design studies to cross-calibrate data from national surveys in which 25(OH)D measurements have already been completed. The longer range goal is to enable the use of standardized 25(OH)D values in individual research laboratories and in clinical care.

Conclusion

In conclusion, this study provides an overview of 25(OH)D status around the globe. It reveals large gaps in information in children and adolescents and smaller but important gaps in adults. In view of the importance of vitamin D to the overall musculoskeletal health and of its potential importance in other tissues, we strongly encourage new research worldwide to define 25(OH)D status. Deficiency must first be identified before it can be appropriately addressed. Knowledge of specific data gaps may help motivate regional policy makers and granting agencies to define the vitamin D status of their population as they decide how to allocate scarce research resources.