Pakistani Immigrant Children and Adults in Denmark Have Severely Low Vitamin D Status

ORIGINAL ARTICLE Pakistani immigrant children and adults in Denmark have severely low vitamin D status

R Andersen1, C Mølgaard2, LT Skovgaard3, C Brot1, KD Cashman4, J Jakobsen5, C Lamberg-Allardt6 and L Ovesen1

1Department of Nutrition, National Food Institute, Technical University of Denmark, Søborg, Denmark; 2Department of Human Nutrition, Faculty of Life Sciences, University of Copenhagen, Frederiksberg, Denmark; 3Department of Biostatistics, University of Copenhagen, Copenhagen, Denmark; 4Department of Food Science and Technology, University College Cork, Cork, Ireland; 5Department of Chemistry, National Food Institute, Technical University of Denmark, Søborg, Denmark and 6Division of Nutrition, University of Helsinki, Helsinki, Finland

Objective: To determine vitamin D and bone status in adolescent girls, pre-menopausal women and men of Pakistani origin, to single out determinants of vitamin D status and to determine the association between vitamin D status, bone metabolism and bone status. Subjects/Methods: Cross-sectional study, Copenhagen (551N), January-November. Serum 25-hydroxyvitamin D (S-25OHD), serum intact parathyroid hormone (S-iPTH), bone turnover markers and whole body and lumbar spine bone mineral density were measured. Sun, smoking and clothing habits, age, body mass index (BMI), and vitamin D and calcium from food and from supplements were recorded. Thirty-seven girls (median age, range: 12.2 years, 10.1–14.7), 115 women (36.2 years, 18.1–52.7) and 95 men (38.3 years, 17.9–63.5) of Pakistani origin (immigrants or descendants with Pakistani parents) took part in the study. Results: Median concentration of S-25OHD was 10.9, 12.0 and 20.7 nmol/l for girls, women and men, respectively. Forty-seven per cent of the girls, 37% of the women and 24% of the men had elevated S-iPTH, and there was a negative relationship between S-iPTH and S-25OHD. Use of vitamin D-containing supplements had a positive association with S-25OHD for men (P ¼ 0.04) and women (P ¼ 0.0008). Twenty-one per cent of the women and 34% of the men had osteopenia. Neither S-25OHD nor S-iPTH was associated with lumbar spine or whole body bone mineral content. Conclusions: Severely low vitamin D status and elevated S-iPTH is common among Pakistani immigrants in Denmark. The low vitamin D status is not associated with bone markers or bone mass among relatively young Pakistanis. European Journal of Clinical Nutrition (2008) 62, 625–634; doi:10.1038/sj.ejcn.1602753; published online 18 April 2007

Keywords: vitamin D status; 25-hydroxyvitamin D; bone status; bone turnover markers; Pakistani immigrants; Denmark

Introduction facilitate the absorption of calcium and phosphorus in the gut to maintain blood calcium concentrations that are vital Vitamin D is a primary regulator of calcium homoeostasis to proper bone mineralization. Vitamin D deficiency can and bone metabolism. The active metabolites of vitamin D result in rickets in children, and in osteomalacia and/or osteoporosis in adults and elderly (Heaney, 2003; Zitter- Correspondence: R Andersen, Department of Nutrition, National Food Institute, Technical University of Denmark, Mørkhøj Bygade 19, DK-2860 mann, 2003). Since the main source of vitamin D is Søborg, Denmark. ultraviolet light, reduced sun exposure, due to clothing E-mail: [email protected] habits, dark skin pigmentation or living at northern Guarantor: R Andersen. latitudes, can lead to vitamin D deficiency. Diet is a Contributors: RA collected the data, wrote the manuscript and undertook the statistical analyses with advice from LTS, CM and LO. RA, CB, KDC, CL-A, CM secondary source of vitamin D, since only few foods contain and LO designed the study. JJ undertook the measurements of S-25OHD. KDC significant amounts of vitamin D. undertook the measurements of bone turnover markers. CL-A undertook the In the late 1960s and early 1970s Pakistani men moved to measurements of S-iPTH, calcium and phosphate. All contributed to the Denmark as labour immigrants. Many of the immigrants manuscript. Received 27 April 2006; revised 5 March 2007; accepted 8 March 2007; settled down and have now lived in Denmark for about 30 published online 18 April 2007 years with their family and descendants. Presently about Severely low vitamin D status among Pakistanis in Denmark R Andersen et al 626 313 100 people from non-Western countries live in Den- (immigrants or descendants with Pakistani parents) primarily mark, and about 19 250 of these are of Pakistani origin living in the Copenhagen area, Denmark (551N). About 87% (immigrants and descendants) (Statistics Denmark, 2005). of all the Pakistanis in Denmark live in the Copenhagen area People who migrate from sun-rich countries to countries (Statistics Denmark, 2005). that are sun-deprived for half of the year and have low UV- The recruitment began in January 2002. Few subjects were intensity even at summer (e.g., Denmark) may increase their recruited during the first quarter (10%), and the recruitment risk of vitamin D deficiency. In particular, when the habit of strategy was intensified during the spring. Therefore, most avoiding direct sun exposure is maintained and most of the subjects (80%) were recruited during the second and third skin is covered with clothes also during the summer, the quarter of 2002, which is the season in Denmark with deficiency may become severe. vitamin D-producing UV radiation. An extra effort was made In recent studies from Norway (Meyer et al., 2004; Holvik to recruit more men during September–October, and the last et al., 2005) vitamin D deficiency was prevalent among 10% (mainly men) were recruited during these months. The Pakistani adults (mean serum 25-hydroxyvitamin D (S- recruitment ended in November, since fasting blood sam- 25OHD): 21–25 nmol/l), but serious deficiency was not pling is not possible during the Muslim Ramadan (which associated with reduced bone mineral density among the took place during November–December in 2002). Pakistani women (Falch and Steihaug, 2000; Meyer et al., The subjects were recruited through information meetings 2004). Two other studies from Oslo, Norway, found that at schools, mosques, cricket clubs, private organizations, vitamin D deficiency (mean S-25OHD: 15–19 nmol/l) was etc.; adverts in local Pakistani newspapers, local Pakistani common among pregnant Pakistani women (Brunvand and radio and television and posters in relevant places, since the Haug, 1993; Henriksen et al., 1995). Rickets and osteomalacia Danish National Central Offices of Civil Registrations do not have been described among Asian immigrants in the UK contain information about ethnic origin. Consequently, the for several years (Dunnigan et al., 1962; Hunt et al., 1976; sample is not random and the representativeness of the Henderson et al., 1987; Nisbet et al., 1990; Finch et al., 1992; sample could not be assessed due to lack of information Solanki et al., 1995; Serhan and Holland, 2002). In Denmark, about non-acceptors. vitamin D deficiency (mean S-25OHD 7 nmol/l) is found Among a total of 295 subjects signing up for participation among Arab women (Glerup et al., 2000a, b), however people 247 (37 girls, 115 women, 95 men) participated at the of Pakistani origin have not been investigated. Besides, little baseline visit including interview, blood and urine sampling. is known about vitamin D status of immigrant children and The number of subjects corresponds to 4.2, 2.3 and 1.5%, men. respectively, of all the Pakistani girls, women and men living The aim of this study was to determine the vitamin D and in Denmark at corresponding ages (Statistics Denmark, bone status in healthy adolescent girls, pre-menopausal 2005). The setting was the Danish Food and Veterinary women and men of Pakistani origin (immigrants or descen- Administration (now the National Food Institute). dants) living in Denmark, and to determine the association Exclusion criteria were medications and diseases known between the vitamin D status, bone metabolism and bone to affect bone metabolism or S-25OHD concentrations, status. Besides, to identify the most important risk factors for pregnancy and breast-feeding. The women were tested for vitamin D deficiency among the Pakistanis, the determi- pregnancy before dual energy X-ray absorptiometry (DXA) nants of vitamin D status were investigated for the adults. scanning. One pregnant and one post-menopausal woman and one man with lymphoma were excluded. Subjects with incomplete data in one or more of the explanatory variables Subjects and methods were excluded in the multiple regression analyses, and they did not differ significantly from the study population with The study is the baseline part of a 1-year long double-blinded respect to S-25OHD and the available explanatory variables. randomized placebo-controlled intervention study with two doses of vitamin D (10 and 20 mg/day). The local ethic committee approved the study protocol Biochemical sampling and analyses (Reg. No. KA01101g). The study was carried out in accor- Morning blood samples were taken by venipuncture after dance with the Declaration of Helsinki. Written informed an overnight fast. Local anaesthetic patches were offered to consent was obtained from all participants, as well as from the girls to reduce the discomfort of venipuncture. Blood the parents/guardians of the girls. samples were centrifuged (approximately 3000 g for 10 min.) within 2 hours of sampling, and serum was frozen and stored at À801C. Morning second void urine samples were collected Subjects after an overnight fast. Urine samples were frozen and stored The subjects were adolescent girls (median age 12.2 years, at À201C. range 10.1–14.7), pre-menopausal women (median age 36.2 S-25OHD concentrations (vitamin D2 plus D3 are used years, range 18.1–52.7) and men (median age 38.3 years, here, since D2 was only detected in four subjects) were range 17.9–63.5). All subjects were of Pakistani origin analysed by high-performance liquid chromatography

European Journal of Clinical Nutrition Severely low vitamin D status among Pakistanis in Denmark R Andersen et al 627 (HPLC) using a diode array detector for detection and an Waltham, MA, USA). For whole body and lumbar analyses, absorbance detector for quantification. The inter-assay software versions V5.73 and V4.76P, respectively were used. coefficient of variation (CV) is 6.3% and the intra-assay CV Subjects wore shirt and long trousers or were covered by a is 4.3%. The method is described elsewhere (Andersen et al., cotton blanket during the scans. Clothes with metal (mainly 2005). S-25OHD concentrations were analysed at the Danish in buttons or bras) and jewellery were removed. For quality Institute for Food and Veterinary Research, Denmark (now control, a spine phantom was scanned daily. The CVs for the the National Food Institute). Serum intact parathyroid BMC, BA and BMD measurements on the spine phantom hormone (S-iPTH) concentrations were analysed by an over a period of 2 years (n ¼ 429) are 0.35, 0.32 and 0.32%, immuno-radiometric method using a commercial OCTEIA respectively. In adults examined with 8 weeks’ interval assay (IDS, Bolton, UK). The inter-assay CV is 4.0% and the reproducibilities expressed as CVs are 1.6, 2.2 and 0.9% for intra-assay CV is 2.3%. Serum total calcium (S-Ca) and serum whole body BMC, BA and BMD, respectively (Hansen and total phosphate (S-P) concentrations as well as urinary Astrup, 1998, unpublished results). The effective dose for a calcium (U-Ca) and urinary phosphate (U-P) excretion were whole body and a lumbar DXA scan was not more than analysed by an absorptiometry method using a KoneLab 10 mSv in total, equal to about a day’s background radiation spectrophotometer (Thermo Clinical Labsystems Ltd, Espoo, in Denmark. Finland, 2000). The inter- and intra-assay CVs for these parameters are less than 5%, with exception of urinary phosphate with CVs of 10%. S-iPTH, S-Ca, S-P, U-Ca and U-P Dietary intake were analysed at University of Helsinki, Finland. Serum- The subjects answered a food frequency questionnaire (FFQ) 2 þ ionized calcium (S-i-Ca) was analysed by a Ca /pH Analyser covering 1 month that ascertained the food groups con- Ciba Corning 634 (Ciba-Corning Diagnostic Corp., Medfield, tributing to 95% of the vitamin D intake and 75% of the MA, USA). The inter-assay CV is 2.4% on level 1.23 mmol/l calcium intake determined from the Danish national dietary and the intra-assay CV is 0.8% on level 1.16 mmol/l. Extern survey, which, however, does not contain intake data from QC: Labquality/DEKS FHK-LHK. Serum total alkaline phos- Pakistani immigrants (Andersen et al., 1996). The question- phatase was analysed by a multiple-point rate p-nitrophenyl s naire included the following specific food groups: milk and phosphate method using reflection photometry (VITROS milk products (including chai and lassie), cheese, bread 950 Chemistry System, Ortho-Clinical Diagnostics, Rochester, (including chapatti and parathi), fish, meat, egg and dried USA). The inter-assay CV is 2.5% on level 89 U/l and the pulse. The intake calculations were performed using the intra-assay CV is 0.9% on level 123 U/l. Extern QC: General Intake Estimation System described elsewhere Labquality/DEKS FHK-LHK. Ionized calcium and alkaline (Christensen, 2001; Andersen et al., 2005). phosphatase were analysed at the Department of Clinical Biochemistry (Hvidovre Hospital, University of Copenhagen, Denmark). Serum osteocalcin levels were determined by using an enzyme-linked immunosorbent assay (BRI-Diag- Background information nostics, Dublin 9, Ireland). Inter-assay CV is 11% and intra- All the subjects answered a questionnaire that ascertained assay CV is 8%. Urinary pyridinoline and deoxypyridinoline demographic characteristics, chronic diseases, smoking were measured by HPLC with fluorescence detection and habits (adults only), menarche status (girls only), use of quantified by external standardization using a commercially medication, sun habits, clothing habits, holidays and other available pyridinoline/deoxypyridinoline HPLC calibrator lifestyle variables. Intake of vitamin D- and/or calcium- (MetraBiosystems Ltd, Wheatley, Oxon, UK). Inter-assay CV containing supplements was recorded by brand and dosage. is 9 and 11%, respectively. Intra-assay CV is 6 and 7%, Weight and height were recorded without shoes. The body 2 2 respectively. Urinary creatinine was analysed by a colori- mass index (BMI) was calculated (weight (kg)/height (m )). metric procedure using a diagnostic kit (Cat. no. 124, 192 The age of the subjects is the age on the date of blood Boehringer Mannheim GmbH, Germany). The inter-assay sampling and interview. CV is 6.7% and the intra-assay CV is 3.2%. S-osteocalcin, U-pyridinoline and U-deoxypyridinoline were analysed at University College Cork, Ireland. Statistical analysis SAS version 8.02 was used for all statistical analyses. The significance level was chosen as 0.05. All analyses were Bone mineral assessment performed for girls, women and men separately. Analyses Whole body and lumbar (L2–L4) bone mineral content included standard descriptive statistics and Spearman corre- (BMC) measured in grams hydroxy-apatite, bone size lations (do not require normal distribution of and linear expressed as anterior–posterior projected bone area (BA) relationship between x and y). In the multiple regression measured in cm2 and bone mineral density measured in analyses (procedure GLM, general linear model) described grams per cm2 (BMD ¼ BMC/BA) were determined by DXA below S-25OHD, S-iPTH, vitamin D and calcium intake, scan using a Hologic 1000/W scanner (Hologic Inc., bone markers, BMC, BMD, BA, height and weight were

European Journal of Clinical Nutrition Severely low vitamin D status among Pakistanis in Denmark R Andersen et al 628 logarithmically transformed to achieve linearity and nor- come corresponding to one unit increase in the covariate. mally distributed residuals. For the covariates that were themselves logarithmically GLM were performed to describe the relation between transformed (vitamin D and calcium intake from food), this S-25OHD and the possible explanatory variables for the ratio corresponds to the effect of a one unit increase on adults (due to few subjects girls were omitted). The following logarithmic scale, that is, a tenfold increase on the original categorical variables were included: vitamin D-containing scale. Since it is unrealistic to increase the vitamin D and supplements (0 ¼ taking no supplements, 1 ¼ taking supple- calcium intake tenfold, we instead calculated and present ments regularly), calcium-containing supplements (women effects corresponding to a 50% increase for vitamin D and only: 0 ¼ taking no supplements, 1 ¼ taking supplements calcium intake (1.5b). The mathematical calculation of these regularly), sun habits (‘how do you prefer to stay outside interpreted ratio estimates is described elsewhere (Andersen during summer season’: avoid sun, sometimes in sun, or et al., 2005). prefer sun (reference group)), smoking habits (men only: GLM were also performed to describe the relation between smoker, ex-smoker or non-smoker (reference group)), cloth- BMC (lumbar spine and whole body) and vitamin D status. ing habits (women only, ‘how are you normally dressed As recommended by Prentice et al. (1994), it was performed outside during summer’: short sleeves and short trousers/ as size-corrected GLM by including weight, height and BA as skirt, short sleeves and long trousers/skirt or long sleeves and explanatory variables. Moreover, age (years) and S-25OHD long trousers/skirt (reference group)). The following contin- concentrations were included as explanatory variables for uous numerical variables were included: age (years), BMI women and men (for girls only S-25OHD were included, due (kg/m2), vitamin D intake from food (mg/day), calcium intake to the few subjects). The same kind of analysis was from food (mg/day) or total (from supplements þ food) performed to describe the relation between BMC (lumbar calcium intake (men only) (mg/day). spine and whole body) and S-iPTH concentrations. Estimated regression coefficients (b) represent the slope, Age-corrected GLM were performed to investigate the that is the effect on outcome for a one unit increase in the relation between S-25OHD and the bone markers (S- respective covariate. Since outcome (S-25OHD) was logarith- osteocalcin, U-pyridinoline and U-deoxypyridinoline). The mically transformed, we transformed the coefficients to ratio same kind of analysis was performed to describe the relation estimates (10b), indicating a multiplicative increase in out- between the bone markers and S-iPTH concentrations, as

Table 1 Characteristics of the population

Girls Women Men

Number of subjects (n)3711595 Age (years)a 12.2 (10.1, 14.7) 36.2 (18.6, 51.7) 38.3 (18.6, 61.6) Born in Denmark, % (n) 89 (33) 21 (24) 11 (10) Weight (kg)a 48.8 (28.4, 63.8)1 67.9 (43.8, 101.1)2 78.3 (60.7, 99.4)3 Height (cm)a 154.1 (139.0, 166.7)1 157.2 (147.0, 168.3)2 171.2 (156.7, 190.8)3 BMI (kg/m2)a 20.4 (14.6, 26.6)1 26.9 (17.4, 40.0)2 26.8 (19.7, 33.1)3 Body fat (%)a,b 27.5 (15.9, 41.9)4 35.0 (24.5, 47.5)5 23.8 (13.0, 29.3)6 Dietary vitamin D intake (mg/day)a 1.5 (0.4, 8.9) 1.7 (0.5, 6.9) 2.2 (0.8, 11.2) Dietary calcium intake (mg/day)a 548 (140, 2268) 532 (109, 1508) 581 (239, 1801) Taking vitamin D supplements, % (n)c 36 (13)9 22 (25)10 25 (23)11 Taking calcium supplements, % (n)c 8 (3) 12 (11)7 2(2)8 Having started period, % (n) 49 (18) — — Post-menarche time (months) 13 (2, 28) — — Smokers and ex-smokers, % (n) — 3 (4) 45 (43) Package year (smokers) — 7.1 (0.6, 13.5) 5.0 (0.02, 45.1)

Sun habits Prefer to stay in sun % (n)d 11 (4) 43 (50) 45 (43) Prefer sometimes to stay in sun, % (n)d 70 (26) 35 (40) 44 (42) Prefer to avoid sun, % (n)d 16 (6) 22 (25) 11 (10)

Clothing habits Prefer to wear long sleeves/long trousers, % (n)e 8 (3) 43 (50) 14 (13) Prefer to wear short sleeves/long trousers, % (n)e 86 (32) 56 (64) 80 (76) Prefer to wear short sleeves/short trousers, % (n)e 3 (1) 1 (1) 6 (6)

1n ¼ 31, 2n ¼ 108, 3n ¼ 94, 4n ¼ 28, 5n ¼ 90, 6n ¼ 77, 7n ¼ 111, 8n ¼ 93, 9n ¼ 36, 10n ¼ 112, 11n ¼ 92. a 1 1 Median (22,972) percentiles. bMeasured by whole-body DXA scanning. cSubjects taking vitamin D or calcium-containing supplements at baseline. dPreferred habits outside during summer season. Girls: n ¼ 36. ePreferred clothing during summer. Trousers or skirts. Girls: n ¼ 36.

European Journal of Clinical Nutrition Severely low vitamin D status among Pakistanis in Denmark R Andersen et al 629 well as the relation between bone markers and BMC, BMD for girls (P ¼ 0.002) and men (P ¼ 0.01), but not for women and BA. (P ¼ 0.2). The median winter (November–March) values were 4.3–6.2 nmol/l and the median summer (April–October) values were 17.1–21.6 nmol/l for girls and men. Eighty-one Results per cent of the girls and 84% of the women had S-25OHD concentrations below 25 nmol/l (often referred to as defi- Characteristics of the population are shown in Table 1, all ciency), and 46% of the girls and 40% of the women were biochemical parameters are shown in Table 2 and bone below 10 nmol/l (often referred to as severe deficiency). measurements are shown in Table 3. Sixty-five per cent of the men had S-25OHD concentration below 25 nmol/l and 13% were below 10 nmol/l. Almost all persons had S-25OHD concentrations below 50 nmol/l (often Vitamin D status (S-25OHD) referred to as insufficiency) (Table 4). The median concentration of S-25OHD was 10.9, 12.0 and 20.7 nmol/l for girls, women and men, respectively. Eighty per cent of the samples were taken from April to September S-25OHD and S-iPTH (the period in Denmark with vitamin D-producing UV Forty-seven per cent of the girls, 37% of the women and 24% radiation). There was seasonal variation in S-25OHD values of the men had elevated S-iPTH (44.1 pmol/l), and all of

Table 2 Biochemical data

Girls Women Men

Number of subjects (n) 37 115 95 Serum 25OHD (nmol/l)a 10.9 (1.8, 51.2) 12.0 (3.6, 55.9) 20.7 (7.0, 56.0) Serum iPTH (pmol/l)a 3.9 (1.8, 29.7)1 3.7 (2.1, 12.3) 3.3 (1.6, 7.2) Serum ionized calcium (mmol/l)a 1.3 (1.1, 1.4) 1.3 (1.2, 1.4) 1.3 (1.2, 1.4) Serum total calcium (mmol/l)a 2.6 (2.3, 2.8)1 2.5 (2.1, 2.7) 2.5 (2.1, 2.8) Serum phosphate (mmol/l)a 1.5 (0.8, 1.8)1 1.1 (0.9, 1.5) 1.2 (0.8, 1.5) Urine calcium (mmol/l)a 0.8 (0.3, 13.9) 1.4 (0.3, 5.6) 2.0 (0.6, 6.0) Urine phosphate (mmol/l)a 17.4 (2.3, 39.2) 14.0 (2.2, 42.3) 19.1 (2.0, 60.3)4 Serum total alkaline phosphatase (U/l)a 508 (156, 1346) 139 (73, 266) 153 (104, 331) Serum osteocalcin (ng/ml)a 39.7 (13.1, 112)1 7.8 (4.4, 25.4) 10.0 (5.3, 28.2) Urine pyridinoline (nmol/mmol creatinine)a 38.5 (12.4, 100)2 10.1 (5.3, 53.0)3 12.4 (3.8, 74.0)5 Urine deoxypyridinoline (nmol/mmol creatinine)a 25.1 (6.8, 64.0)2 5.5 (2.2, 30.8)3 6.6 (2.0, 42.8)5

Abbreviations: 25OHD, 25-hydroxy vitamin D (D2 and D3); iPTH, intact parathyroid hormone. 1n ¼ 36, 2n ¼ 35, 3n ¼ 106, 4n ¼ 92, 5n ¼ 90. a 1 1 Median (22,972) percentiles.

Table 3 Bone data: DXA scanning

Girls Women Men

Whole body DXA scanning Number of subjects (n)289077 Bone mineral density (g/cm2)a 0.92 (0.74, 1.09) 1.10 (0.92, 1.26) 1.14 (1.03, 1.32) Bone mineral content (g)a 1539 (833, 2244) 2235 (1548, 3317) 2577 (1878, 3421) Bone area (cm2)a 1680 (1132, 2052) 2033 (1641, 2477) 2259 (1825, 2516)

Lumbar spine DXA scanning Number of subjects (n)299986 Bone mineral density (g/cm2)a 0.75 (0.57, 1.21) 1.03 (0.76, 1.24) 1.01 (0.71, 1.35) Bone mineral content (g)a 31 (20, 73) 52 (34, 72) 61 (38, 91) Bone area (cm2)a 41 (33, 60) 51 (43, 61) 61 (52, 74)

Lumbar spine T score, n (%)b pÀ2.5 (osteoporosis) — 3 (3%) 7 (8%) 4À2.5 and pÀ1.0 (osteopenia) — 21 (21%) 29 (34%) 4À1.0 (normal) — 75 (75%) 50 (58%) a 1 1 Median (22,972) percentiles. bPrevalence of subjects by categories of lumbar spine T score.

European Journal of Clinical Nutrition Severely low vitamin D status among Pakistanis in Denmark R Andersen et al 630 Table 4 Prevalence of vitamin D status (%) and median S-iPTH (pmol/l) in categories of S-25OHD

S-25OHD (nmol/l) Girls Women Men

(n) (%) S-iPTHa (n) (%) S-iPTHb (n) (%) S-iPTHc

o10 17 46 5.6 46 40 4.3 12 13 3.7 X10 and o25 13 35 3.1 51 44 3.7 49 52 3.3 X25 and o50 5 14 3.3d 14 12 3.1 29 31 3.2 X50 253.043 2.955 2.4

Abbreviations: 25OHD, 25-hydroxy vitamin D (D2 and D3); iPTH, intact parathyroid hormone. Significant difference between S-25OHD categories, within a column (non-parametric ANOVA) (aP ¼ 0.002, bP ¼ 0.0002, cP ¼ 0.02). dn ¼ 4.

Figure 1 Association between serum 25OHD and iPTH concentrations. The three curves show fitted relations (linear on a double-logarithmic plot) for girls, men and women, respectively.

them had S-25OHD o50 nmol/l. There was a significant The following variables were included for both women and negative correlation (Spearman) between S-25OHD and men: age, BMI, use of vitamin D-containing supplements, S-iPTH for girls (À0.63, Po0.0001) and for women (À0.50, dietary (from food only) vitamin D intake and sun habits. For Po0.0001). For men, the tendency was the same, although women use of calcium-containing supplements and dietary it did not quite reached significance (À0.20, P ¼ 0.053). The calcium intake was also included, but since only two men used association between S-25OHD and S-iPTH is shown in calcium-containing supplements (Table 1), it was not included Figure 1. When comparing different categories of S-25OHD as a separate variable for men. Instead, the dose of calcium (Table 4), there was a significant difference in S-iPTH, for from supplements was added to the dietary calcium intake girls, women and men, and the differences are seen to and the total calcium intake was included for men. Smoking correspond well to the inverse relationship in Figure 1. habits were included for men, but not for women, since only four women were present or former smokers (Table 1). Clothing habits were only included for women, since 80% Determinants of S-25OHD concentrations of the men belonged to the same clothing category (Table 1). To explain the (logarithmically transformed) S-25OHD Use of vitamin D-containing supplements had a positive concentrations multiple regression analysis was performed association with S-25OHD for both men (P ¼ 0.04) and and effects are quantified in Table 5. Owing to the low women (P ¼ 0.0008). The median S-25OHD concentrations number of girls, data are not shown for girls. for supplement users vs non-users were 21.2 vs 10.1 nmol/l

European Journal of Clinical Nutrition Severely low vitamin D status among Pakistanis in Denmark R Andersen et al 631 Table 5 Multiple linear regression with S-25OHDa as dependent variable

Person Explanatory variables Estimate Standard P-value Interpreted ratio Interpreted 95% confidence (b) error estimatec limits

Women Vitamin D-containing supplements 0.32 0.091 0.0008 2.073 ( þ 107%) 1.366, 3.146 (n ¼ 103) Calcium dietary intakea À0.23 0.1139 0.04 0.909 (À9%)d 0.830, 0.996 Clothing habits, short sleeves and trousers 0.50 0.270=> 3.141 ( þ 214%) 0.912, 10.82 Clothing habits, short sleeves and long 0.09 0.057 1.232 ( þ 23%) 0.950, 1.596 trousers ;> 0.08 Clothing habits, long sleeves and trousersb 0.0 —) — — Sun habits, avoid sun À0.10 0.068 0.795 (À21%) 0.582, 1.084 Sun habits, sometimes in sun À0.04 0.061 0.34 0.912 (À9%) 0.690, 1.207 Sun habits, prefer sunb 0.0 — — — BMI 0.004 0.005 0.48 1.009 ( þ 0.9%) 0.984, 1.035 Calcium-containing supplements 0.05 0.117 0.67 1.123 ( þ 12%) 0.659, 1.916 Vitamin D dietary intakea 0.02 0.102 0.81 1.010 ( þ 1%)e 0.931, 1.096 Age 0.0002 0.003 0.94 1.001 ( þ 0.1%) 0.985, 1.016

Men (n ¼ 90) Vitamin D-containing supplements 0.13 0.061 0.04 1.350 ( þ 35%) 1.019, 1.790 Vitamin D dietary intakea 0.12 0.092 0.20 1.050 ( þ 5%)e 0.975, 1.130 Age 0.003 0.002) 0.25 1.007 ( þ 0.7%) 0.995, 1.018 Sun habits, avoid sun 0.08 0.087 1.196 ( þ 20%) 0.804, 1.779 Sun habits, sometimes in sun 0.08 0.056 0.30 1.215 ( þ 21%) 0.938, 1.573 Sun habits, prefer sunb 0.0 — — — BMI 0.002 0.008) 0.85 1.004 ( þ 0.4%) 0.966, 1.042 Smoking habits, non-smoker 0.003 0.057 1.007 ( þ 0.7%) 0.778, 1.305 Smoking habits, ex-smoker 0.05 0.089 0.87 1.111 ( þ 11%) 0.738, 1.673 Smoking habits, smokerb 0.0 — — — Total (diet þ supplement) calcium intakea À0.02 0.126 0.89 0.993 (À0.7%)d 0.897, 1.099

Abbreviation: BMI, body mass index. aLogarithmically transformed. bReference group. cSince outcome (S-25OHD) was logarithmically transformed the estimated regression coefficients (b) was transformed to ratio estimates (10b), indicating a multiplicative increase in outcome corresponding to one unit increase in the covariate. dBy 50% increase of calcium intake. eBy 50% increase of vitamin D intake. and 26.3 vs 19.4 nmol/l for women and men, respectively. per cent of the women and 34% of the men had osteopenia Dietary calcium intake had a negative association (P ¼ 0.04) (À2.5oT score pÀ1.0), and 76% of the women and 58% of with S-25OHD for women. the men had normal bone status (T score 4À1.0). There were no significant correlations (Spearman) between vitamin D status (S-25OHD) and any of the DXA-scanning parameters Interpreted ratio estimates (whole body and lumbar spine: BA, BMD and BMC) for girls, The estimated regression coefficients are not directly com- women and men. In multiple regression analysis S-25OHD parable due to different units in the explanatory variables, concentrations were not significantly related with size- and interpreted ratio estimates are therefore calculated for corrected lumbar spine or with size-corrected whole body women and men (Table 5). The interpreted ratio estimates BMC for girls, women and men. indicate how differently the various explanatory variables In multiple regression analysis S-iPTH concentrations were affect the S-25OHD concentration. One of the covariates of not significantly related with size-corrected lumbar spine or main interest is use of vitamin D-containing supplements. size-corrected whole body BMC for girls, women and men. For this variable, the interpreted ratio estimates are 107% for women and 35% for men. For women it has a massive effect (214%) on S-25OHD concentration to wear short sleeves Bone markers and short trousers or skirt normally during summer, but the In age-corrected multiple regression analysis a significant association is not significant (P ¼ 0.08), since only one negative association between S-osteocalcin and S-25OHD Pakistani woman is dressed this way. was found for women (P ¼ 0.0006, b ¼À0.2), but not for girls and men. Bone status There was a significant positive association between Three per cent of the women and 8% of the men had U-deoxypyridinoline and S-25OHD for girls (P ¼ 0.01, osteoporosis defined by T score pÀ2.5 (Table 3). Twenty-one b ¼ 0.3), but no association for women and men, in

European Journal of Clinical Nutrition Severely low vitamin D status among Pakistanis in Denmark R Andersen et al 632 age-corrected multiple regression analysis. No significant Symposium on Nutritional Aspects of Osteoporosis in 2003 association was found between U-pyridinoline and S-25OHD it was concluded that the optimal S-25OHD concentration for girls, women and men in age-corrected multiple regres- for bone health is between 50 and 80 nmol/l, with five of the sion analysis. six expert opinions clustered between 70 and 80 nmol/l A significant positive association was found between (Dawson-Hughes et al., 2005) (mainly evaluated among S-osteocalcin and S-iPTH for girls (P ¼ 0.02, b ¼ 0.4) and Caucasians; however ethnicity is not always described). A women (Po0.0001, b ¼ 0.4), but not for men. high prevalence of symptomatic osteomalacia could be Neither U-pyridinoline nor U-deoxypyridinoline was sig- expected among these subjects due to the low S-25OHD. nificantly associated with S-iPTH for girls, women and men. However, only one girl, two women and two men had For all groups (girls, women and men) none of the bone elevated serum total alkaline phosphatase, and as shown in markers were associated with either BMC, BMD or BA in Figure 1, PTH is not clearly elevated until S-25OHD levels are age-corrected multiple regression analysis. below 10–12 nmol/l. Besides the biochemical signs there were no other assessment of osteomalacia. Severely impaired muscle function may be present even before biochemical Discussion signs of bone disease develop (Glerup et al., 2000a). It was not within the scope of this project to measure muscle First, very low vitamin D status was found in this study strength. We did ask questions about muscle pain (duration among adolescent girls, pre-menopausal women and men and degree), but the questions were compounded by of Pakistani origin living in Denmark. More than 80% of language difficulties; consequently the questions were not the girls and women had S-25OHD concentration below included in the analysis. 25 nmol/l, and about 40% were below 10 nmol/l. Low Fourth, the severe vitamin D deficiency and elevated PTH vitamin D status was also very common among men, as concentrations found in this group of Pakistanis was not 65% had S-25OHD concentration below 25 nmol/l and about clearly reflected in the bone markers measured. Contrary to 10% below 10 nmol/l. Almost all subjects had S-25OHD an Australian study where a higher urinary excretion of concentrations below 50 nmol/l. Since 80% of the blood deoxypyridinoline was found among vitamin D deficient samples were taken during April to September, the season in Muslim women compared to the non-deficient women Denmark with vitamin D-producing UV radiation, an even (Diamond et al., 2002), we did not find any association for larger proportion of the population may be expected to be women and men. The level of deoxypyridinoline (5–7 nmol/ vitamin D deficient during wintertime. Pakistanis in Norway mmol creatinine) among the Muslim women in Australia had vitamin D status (21–25 nmol/l) similar to our men (Diamond et al., 2002) is comparable to the level found in (median 20.7 nmol/l), but higher than our females (medians our study. The few studies to date investigating the relation- 10.9–12 nmol/l) (Falch and Steihaug, 2000; Meyer et al., ship between S-25OHD and bone markers suggest S-25OHD 2004; Holvik et al., 2005). Pakistani women in UK had higher threshold levels as high as 30 (Diamond et al., 2002 (Muslim S-25OHD concentration (mean 19.8 nmol/l, season not women); Sahota et al., 1999 (UK women)) or 60 nmol/l stated) than the females in our study (Roy et al., 2007). (Jesudason et al., 2002 (white women)) to define vitamin D However, it should be kept in mind that comparing insufficiency based on rise in bone markers. Therefore, our S-25OHD in different studies is problematic due to inter- ability to detect relationship between S-25OHD and bone laboratory variations (Lips et al., 1999; Heaney, 2000; Ovesen markers may be limited by the very low vitamin D levels et al., 2003; Binkley et al., 2004; Carter et al., 2004; Hollis, among the majority of the subjects. 2004). Besides, the non-random recruitment and lack of Fifth, in this study with relatively young healthy Pakista- representativeness is a limitation for this study. nis, bone status (DXA) was neither associated with S-25OHD Second, along with low vitamin D status, 47% of the girls, and PTH nor with bone markers. The lack of association 37% of the women and 24% of the men had elevated S-iPTH between S-25OHD and BMC in our study was also found concentrations. To maintain normal serum calcium concen- in a study from Norway (Falch and Steihaug, 2000). Two trations, the organism will compensate for a reduced vitamin other Norwegian studies found that vitamin D-deficient D status by increasing the secretion of parathyroid hormone Pakistanis living in Oslo had similar BMD (Alver et al., 2004) (PTH), thereby generating secondary hyperparathyroidism and similar level of bone turnover markers (Holvik et al., (Falch and Steihaug, 2000). 2004) as Caucasian Norwegians. In contrast, S-25OHD was Third, there was a clear negative relationship between found in some extent to correlate with BMD among Asian S-iPTH and S-25OHD. However, the S-25OHD level needed Indians in India (Arya et al., 2004), women in Iran (Rassouli for suppression of PTH seems to be much lower (around et al., 2001), and S-25OHD level below 37.5 nmol/l was 10 nmol/l) in this study compared to S-25OHD levels (30– associated with a reduction in bones mass at the hip and 99 nmol/l, with a tendency to cluster in the 75–80 nmol/l wrist among Pakistani women in the UK (Roy et al., 2007). range) estimated to suppress PTH among Caucasians (mainly Whether the very low vitamin D status and the secondary elderly women) (Chapuy et al., 1997; Jesudason et al., 2002; hyperparathyroidism among the Pakistanis in Denmark Dawson-Hughes et al., 2005). At the fifth International would lead to the same skeletal losses seen among

European Journal of Clinical Nutrition Severely low vitamin D status among Pakistanis in Denmark R Andersen et al 633 Caucasians when the Pakistani women and men become Acknowledgements older, we do not know yet. Only about 2% of the Pakistanis actually living in Denmark are above 65 years of age We acknowledge Karin Hess Ygil, Dorte Strange and Nighat (Statistics Denmark, 2005). Racial/ethnic physiological Kwajada for interviewing the participants. Nighat Kwajada is differences regarding bone health (e.g., gut resistance to acknowledged for the blood sampling and the interpretation actions of 1.25(OH)2D and differences in bone turnover) are into Urdu when necessary. Dorte Strange is acknowledged reported for African-Americans compared to whites (Dawson- for the DXA scannings. Birgitte Hermansen is acknowledged Hughes, 2004). Indian/Pakistani women from the USA for assisting the DXA-scannings and assisting designing the had lower BMD values than their Caucasian counterparts, FFQ. Karin Hess Ygil, Tue Christensen and Anders Møller are but it did not appear to be attributable to ethnicity per se acknowledged for the dietary intake calculations. The study (Alekel et al., 1999). Altered vitamin D metabolism was was part of the OPTIFORD-project ‘Towards a strategy for found among Asian Indians in the USA (Awumey et al., optimal vitamin D fortification’, financed by EU, the Fifth 1998). If the vitamin D-deficient Pakistani immigrants are Framework Programme (QLK1-CT-2000-00623). not protected from skeletal loss due to some ethnic/racial differences, large health problems could arise among Pakistani immigrants in the coming decades, after all; 21% References of the women and 34% of the men in this study had osteopenia according to T score. However, it should be Alekel DL, Mortillaro E, Hussain EA, West B, Ahmed N, Peterson CT et al. (1999). Lifestyle and biologic contributors to proximal femur further investigated whether the WHO reference values of T bone mineral density and hip axis length in two distinct ethnic score used to assess osteopenia and osteoporosis are valid for groups of premenopausal women. Osteoporos Int 9, 327–338. all ethnic groups, there might be different reference values Alver K, Meyer HE, Falch JA, Sogaard AJ (2004). Bone mineral density J Bone Miner Res for different ethnic groups. in ethnic Norwegians and Pakistani immigrants. 19, S159 (Abstract). Finally, it was found in this study that the use of vitamin Andersen NL, Fagt S, Groth MV (1996). Danskernes kostvaner 1995 D-containing supplements had a positive association with (Dietary habits of Danes 1995). Publication No. 235, Levnedsmid- S-25OHD for women and men, when different lifestyle delstyrelsen: Søborg, Denmark. Andersen R, Mølgaard C, Skovgaard LT, Brot C, Cashman KD, parameters were included in the statistical analyses Chabros E et al. (2005). Teenage girls and elderly women living in to explain the S-25OHD concentrations. The increase in northern Europe have low winter vitamin D status. Eur J Clin Nutr S-25OHD is about 100% for women taking vitamin 59, 533–541. D-containing supplements regularly compared to women Arya V, Bhambri R, Godbole MM, Mithal A (2004). Vitamin D status and its relationship with bone mineral density in healthy Asian taking no supplements. The dietary vitamin D intake Indians. Osteoporos Int 15, 56–61. was assessed using an FFQ and included in the statistical Awumey EMK, Mitra DA, Hollis BW, Kumar R, Bell NH (1998). analyses. The dietary vitamin D intake (Table 1) was Vitamin D metabolism is altered in Asian Indians in the southern low compared to ethnic Danes (Andersen et al., 1996), and United States: a clinical research center study. J Clin Endocrin Metab the dietary intake was not associated with S-25OHD. 83, 169–173. Binkley N, Krueger D, Cowgill CS, Plum L, Lake E, Hansen KE et al. Unmeasured confounders associated with supplement (2004). Assay variation confounds the diagnosis of hypovitami- intake can be a limitation of this cross-sectional study. nosis D: a call for standardization. J Clin Endocrin Metab 89, Muslim immigrants in Denmark are recommended through 3152–3157. Brunvand L, Haug E (1993). Vitamin D deficiency amongst Pakistani information leaflets to take 10 mg vitamin D supplements women in Oslo. Acta Obestet Gynecol Scand 72, 264–268. per day. Carter GD, Carter R, Jones J, Berry J (2004). How accurate are assays Vitamin D deficiency is suspected to be a contributing for 25-hydroxyvitamin D? Data from the international vitamin D factor to several other diseases, for example, autoimmune external quality assessment scheme. Clin Chem 50, 2195–2197. Chapuy M-C, Preziosi P, Maamer M, Arnaud S, Galan P, Herberg S diseases, some types of cancers and diabetes (mainly among et al. (1997). Prevalence of vitamin D insufficiency in an adult Caucasians) (Holick, 2004), as well as hypovitaminosis D normal population. Osteoporos Int 7, 439–443. myopathy among Arab immigrant women (Glerup et al., Christensen T (2001) Outline of GIES: General Intake Estimation 2000a). System. Fourth International Food Data Conference: New Trends in the Management and Uses of Food Databases. 24–26 August In conclusion, our study shows severely low vitamin D 2001. Bratislava, Slovakia (poster). status and elevated S-iPTH, and a negative relation between Dawson-Hughes B (2004). Racial/ethnic considerations in making S-iPTH and S-25OHD, among Pakistani immigrants and recommendations for vitamin D for adult and elderly men and descendents of both genders living in the Copenhagen area women. Am J Clin Nutr 80S, 1763S–1766S. in Denmark. Use of vitamin D-containing supplements is a Dawson-Hughes B, Heaney RP, Holick MF, Lips P, Meunier PJ, Vieth R (2005). Estimates of optimal vitamin D status. Osteoporos Int 16, determinant for S-25OHD among the adults. Low vitamin D 713–716. status, in this study with relatively young Pakistanis, was not Diamond TH, Levy S, Smith A, Day P (2002). High bone turnover associated with bone markers or bone mass. Randomized in Muslim women with vitamin D deficiency. Med J Aust 177, 139–141. intervention studies are needed to clarify if bone status Dunnigan MG, Paton JPJ, Haase S, McNicol GW, Gardner MD, Smith among immigrants can be improved by increasing vitamin D CM (1962). Late rickets and osteomalacia in the Pakistani status. community in Glasgow. Scot Med J 7, 159–167.

European Journal of Clinical Nutrition Severely low vitamin D status among Pakistanis in Denmark R Andersen et al 634 Falch JA, Steihaug S (2000). Vitamin D deficiency in Pakistani vitamin D and bone resorption markers in vitamin D insuffi- premenopausal women living in Norway is not associated with ciency. Bone 31, 626–630. evidence of reduced skeletal strength. Scand J Clin Lab Invest 60, Lips P, Chapuy M-C, Dawson-Huges B, Pols HAP, Holick MF (1999). 103–110. An international comparison of serum 25-hydroxyvitamin D Finch PJ, Ang L, Eastwood JB, Maxwell JD (1992). Clinical and measurements. Osteoporos Int 9, 394–397. histological spectrum of osteomalacia among Asians in south Meyer HE, Falch JA, Søgaard AJ, Haug E (2004). Vitamin D deficiency London. QJMed83, 439–448. and secondary hyperparathyroidism and the association with Glerup H, Mikkelsen K, Poulsen L, Hass E, Overbeck S, Andersen H bone mineral density in persons with Pakistani and Norwegian et al. (2000a). Hypovitaminosis D myopathy without biochemical background living in Oslo, Norway, The Oslo Health Study. signs of osteomalacic bone involvement. Calcif Tissue Int 66, Bone 35, 412–417. 419–424. Nisbet J, Eastwood JB, Colston KW, Ang L, Flanagan AM, Chambers Glerup H, Mikkelsen K, Poulsen L, Hass E, Overbeck S, Thomsen J TJ et al. (1990). Detection of osteomalacia in British Asians: et al. (2000b). Commonly recommended daily intake of vitamin D a comparison of clinical score with biochemical measurements. is not sufficient if sunlight exposure is limited. J Int Med 247, Clin Sci 78, 383–389. 260–268. Ovesen L, Andersen R, Jakobsen J (2003). Geographical differences in Heaney RP (2000). Vitamin D: how much do we need, and how vitamin D status, with particular reference to European countries. much is too much? Osteoporos Int 11, 553–555. Proc Nutr Soc 62, 813–821. Heaney RP (2003). Long-latency deficiency disease: insights from Prentice A, Parsons TJ, Cole TJ (1994). Uncritical use of bone mineral calcium and vitamin D. Am J Clin Nutr 78, 912–919. density in absorptiometry may lead to size-related artefacts in the Henderson JB, Dunnigan MG, McIntosh WB, Abdul-Motaal AA, identification of bone mineral determinants. Am J Clin Nutr 60, Gettinby G, Glekin BM (1987). The importance of limited 837–843. exposure to ultraviolet radiation and dietary factors in the Rassouli A, Milanian I, Moslemi-Zaden M (2001). Determination of aetiology of Asian rickets: a risk-factor model. QJMed63, 413–425. serum 25-hydroxyvitamin D3 levels in early postmenopausal Henriksen C, Brunvand L, Stoltenberg C, Trygg K, Haug E, Pedersen JI Iranian women: relationship with bone mineral density. Bone 29, (1995). Diet and vitamin D status among pregnant Pakistani 428–430. women in Oslo. Eur J Clin Nutr 49, 211–218. Roy DK, Berry JL, Pye SR, Adams JE, Swarbrich CM, King Y et al. Holick MF (2004). Sunlight and vitamin D for bone health and (2007). Vitamin D status and bone mass in UK South Asian prevention of autoimmune diseases, cancers, and cardiovascular women. Bone 40, 200–204. disease. Am J Clin Nutr 80S, 1678S–1688S. Sahota O, Masud T, San P, Hosking DJ (1999). Vitamin D Hollis BW (2004). The determination of circulating 25-hydroxy- insufficiency increases bone turnover markers and enhances bone vitamin D: no easy task. J Clin Endocrin Metab 89, 3149–3151. loss at the hip in patients with established vertebral osteoporosis. Holvik K, Meyer HE, Falch JA, Søgaard AJ, Haug E (2004). The Oslo Clin Endocrin 51, 217–221. Health Study: biochemical markers of bone turnover and their Serhan E, Holland M (2002). Relationship of hypovitaminosis D and relation to parathyroid hormone levels in persons of Pakistani and secondary hyperparathyroidism with bone mineral density among Norwegian origin living in Oslo. J Bone Min Res 19, S167 (Abstract). UK resident Indo-Asians. Ann Rheum Dis 61, 456–458. Holvik K, Meyer HE, Haug E, Brunvand L (2005). Prevalence and Solanki T, Hyatt RH, Kemm JR, Hughes EA, Cowan RA (1995). Are predictors of vitamin D deficiency in five immigrant groups living elderly Asians in Britain at a high risk of vitamin D deficiency and in Oslo, Norway: the Oslo Immigrant Health Study. Eur J Clin Nutr osteomalacia? Age Ageing 24, 103–107. 59, 57–63. Statistics Denmark, Statbank (2005): http://www.dst.dk BEF3: Hunt SP, O’Riordan JLH, Windo J, Truswell AS (1976). Vitamin D Population by area, age, sex, country of origin and citizenship status in different subgroups of British Asians. BMJ 2, 1351–1354. (1980–2005). Jesudason D, Need AG, Horowitz M, O’Loughlin PD, Morris HA, Zittermann A (2003). Vitamin D in preventive medicine: are we Nordin BEC (2002). Relationship between serum 25-hydroxy- ignoring the evidence? Br J Nutr 89, 552–572.

European Journal of Clinical Nutrition