RESEARCH ARTICLE
Neonatal Vitamin D Status and Risk of Multiple Sclerosis €rnhielm, MD,3 Peter Ueda, MD, PhD,1 Farshid Rafatnia, MD,2 Maria B€ aa Robin Fr€ obom,3 Greg Korzunowicz,3 Ragnar L€ onnerbro,3 Anna Karin Hedstr€ om, MD,3 Darryl Eyles, PhD,4 Tomas Olsson, MD, PhD,5 and Lars Alfredsson, PhD3 Objective: Low vitamin D status at birth may be associated with risk of adult onset multiple sclerosis, but this link has not been studied directly. We assessed the relation between neonatal vitamin D concentrations, measured in stored blood samples, and risk of multiple sclerosis. Methods: This was a population-based case–control study in Sweden including 459 incident cases of multiple sclerosis and 663 controls, randomly drawn from a national population registry and frequency matched on sex, age, and residential area. Results: There was no association between neonatal 25-hydroxyvitamin D quintile and risk of multiple sclerosis (crude odds ratio 5 1.0, 95% confidence interval 5 0.68–1.44, for the highest quintile compared to the lowest). Adjusting for a number of potential confounding factors in early life (month of birth, latitude of birth, breastfeeding) and in adult life (25-hydroxyvitamin D, sun exposure, vitamin D intake from dairy products, fatty fish consumption, smoking, body mass index at 20 years of age) as well as ancestry, multiple sclerosis heredity, and socioeconomic group did not considerably affect the result. Interpretation: At a broad population level, 25-hydroxyvitamin D at birth was not associated with risk of multiple sclerosis. ANN NEUROL 2014;00:000–000
T
he etiology of multiple sclerosis (MS) is considered an interplay of genetics and environmental factors.1 Epidemiological observations regarding sunlight exposure in childhood,2 dietary vitamin D intake,3 and serum 25hydroxyvitamin D (25[OH]D) concentrations4 have highlighted low vitamin D as a potential risk factor for the disease. A gestational component of MS etiology has been proposed based on a higher concordance rate of MS in dizygotic twins—who are exposed to the same intrauterine environment and timing of birth—as compared with siblings who are not twins.1 Reports from the northern hemisphere indicate that risk of MS is increased among those born during spring, in particular in the months of April and May, compared to the autumn-born.5–8
The pattern is reversed in the southern hemisphere, where individuals born in November and December are at the highest risk.9 It has been suggested that low sunlight exposure and associated low vitamin D status during gestation for individuals born after winter increase susceptibility to MS.1,5–11 The hypothesis has been supported by a study from the United States predicting gestational vitamin D levels in mothers of MS cases using questionnaire data about maternal characteristics and diet during pregnancy, and latitude and season of delivery.12 To date, no study has directly assessed the relation between vitamin D at birth and risk of MS. By measuring vitamin D in stored neonatal blood samples, we had the opportunity to investigate this hypothesis in a large Swedish population-based case–control cohort,
View this article online at wileyonlinelibrary.com. DOI: 10.1002/ana.24210 Received Mar 19, 2014, and in revised form Jun 27, 2014. Accepted for publication Jun 27, 2014. Address correspondence to Dr Ueda, Clinical Epidemiology Unit, T3 Eugeniahemmet, Karolinska University Hospital Solna, 17176 Stockholm, Sweden. E-mail address: [email protected] From the 1Clinical Epidemiology Unit, Department of Medicine, Solna, Karolinska Institutet, Stockholm, Sweden; 2Internal Medicine Department, Karolinska University Hospital, Stockholm, Sweden; 3Institute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden; 4Queensland Centre for Mental Health Research, Queensland Brain Institute, University of Queensland, Brisbane, Queensland, Australia; and 5Neuroimmunology Unit, Department of Clinical Neuroscience and Center for Molecular Medicine, Karolinska Institute, Stockholm, Sweden. Additional supporting information can be found in the online version of this article.
C 2014 American Neurological Association V 1
ANNALS
of Neurology
FIGURE 1: Flow diagram for study population. EIMS 5 Epidemiological Investigation of Multiple Sclerosis.
comprising information about various environmental factors of potential importance for MS risk, including vitamin D status at inclusion in the study. We hypothesized that low neonatal vitamin D status would be associated with increased risk of MS.
Subjects and Methods Study Population Epidemiological Investigation of MS (EIMS) is a population based case–control study using incident cases of MS. The study base comprises the population aged 16 to 70 years in Sweden. Starting from April 2005, cases were recruited from 40 study centers, including all university hospitals in Sweden. Cases were diagnosed by a neurologist according to the McDonald criteria. For each case, 2 controls—frequency matched by age (predetermined 5-year age groups), sex, and residential area—were randomly selected from the National Population Register. A standardized questionnaire was used to obtain data on exposures and other information of relevance. The questionnaires were sent to cases and controls and filled out at home. The Swedish Phenylketonuria (PKU) Register is a near complete biobank containing dried neonatal blood samples from screening of rare metabolic diseases in infants born in Sweden since 1975.
2
Participants of EIMS who were born in Sweden from 1975 and onward, who had been asked to take part in EIMS by October 2011, and who had given consent for the use of blood samples for research purposes were considered for inclusion in the study. Of the subjects born in 1975 and onward, 563 (89%) of the cases and 1,235 (66%) of the controls returned their questionnaire. Of these subjects, 500 cases and 1,070 controls stated that they were born in Sweden. A total of 1,219 participants, 497/500 (99%) cases and 722/1,070 (67%) controls, had given consent for analysis of blood samples. For 16 of these subjects (7 cases and 9 controls), information required for identification of neonatal blood samples was not available, and for 31 cases and 50 controls, neonatal blood samples could not be found in the PKU Register. The final study population thus consisted of 1,122 subjects, of whom 459 were cases and 663 were controls (Fig 1). The study was approved by the regional ethical review board in Stockholm (2012/100-32 and 04-252/1-4). Written consent for participation in EIMS and for analysis of blood samples was obtained from all individuals included in the study.
Measurement of Neonatal Vitamin D Concentrations Two 3.2mm punches were obtained from the participants’ neonatal dried blood spots. Samples were stored at room temperature between 1975 and August 1981, and thereafter in cold
Volume 00, No. 00
Ueda et al: Neonatal Vitamin D and MS
storage (4 C). It is likely that the samples degraded somewhat during the time in storage. 25(OH)D is the metabolite used for assessing vitamin D status13 and can exist in 2 forms: 25(OH)D3 and 25(OH)D2. The latter can only be obtained from certain dietary sources, including supplements. 25(OH)D3 and 25(OH)D2 in dried blood spots were measured with a highly sensitive liquid chromatography tandem mass spectroscopy assay with multiple reactant monitoring.14 The method uses minimal sample cleanup to reduce sample loss during extraction and chemical derivatization to enhance 25(OH)D3 ionization.14 Using samples from Australia and Denmark, it has been shown that the assay can satisfyingly detect seasonal (within year) differences in 25(OH)D3 concentrations and that the measurements are strongly correlated with neonatal cord blood concentrations (r 5 0.86).15 The assay has been used on samples stored up to 35 years16,17 and evaluated with regard to punch position, spot volume, and paper type.18 The punches in this study were consistently sampled from the outer part of the blood spot as outlined by Kvaskoff et al.18 The concentration of 25(OH)D—including 25(OH)D2 and 25(OH)D3—is reported in nanomoles per liter. 25(OH)D2 and 25(OH)D3 are highly protein-bound molecules that are excluded from erythrocytes.18 To make results comparable with existing studies, results are presented as sera concentrations of hematocrit in capillary blood as previously described.15
Other Variables Information about postnatal age of the participant when the blood was sampled and maternal age at birth were recorded from the PKU Register. Age at inclusion in EIMS was categorized into the following 6 groups: 16 to 19, 20 to 24, 25 to 29, 30 to 34, 35 to 39, and 40 to 44 years. Participants were considered as Scandinavian if they had at least 1 parent of Scandinavian (Sweden, Norway, Finland, Denmark, Iceland) origin. Latitude of residential area of birth was categorized into 54 to 58 , 59 , and 60 to 69 north. Body mass index (BMI) at 20 years of age was calculated using self-reported weight and height data from the questionnaire and presented as a continuous variable. Weekly intake of vitamin D through dairy products including milk, yogurt, cheese, and butter was estimated and presented in international units (continuous variable) by multiplying the amount and frequency of the dietary item by its vitamin D content. Consumption of fatty fish (e.g. salmon and herring) 43 or more per month was considered a high fatty fish intake, whereas 33/wk, No. (%)
82 (18)
133 (20)
Sun exposure index, mean [SD]
3.4 [1.7]
3.7 [1.2]
Smoker, No. (%)
213 (47)
258 (39)
65.0 [27.7]
67.8 [26.2]
Manual workers
135 (29)
234(35)
Nonmanual employees
162 (35)
205 (31)
Unknown
162 (35)
223 (34)
No university education
169 (37)
274 (41)
University studies without a degree
148 (32)
188 (28)
University studies with degree
142 (31)
200 (30)
Latitude of birth, No. (%) 54–58 north
59 north
60–69 north
MS heredity, No. (%)
Breastfed, No. (%) 2
a
25(OH)D at inclusion , nmol/l [SD] Socioeconomic group, No. (%)
Education, No. (%)
a
Measurements available for 298 cases and 307 controls 25(OH)D 5 25-hydroxyvitamin D BMI 5 body mass index; MS 5 multiple sclerosis; SD 5 standard deviation.
Comparison with Other Studies The hypothesis that exposure to low vitamin D in utero, in particular during the last trimester, increases risk of developing MS is based on 12 studies5,21 from the northern hemisphere (including a report from Sweden11) and 1 from the southern hemisphere showing an excess of spring births among MS cases.5 However, the observed risk increase of being born in spring has been modest (odds ratio of about 1.1 compared to being born in Month 2014
other seasons), and the season of birth pattern for MS risk has not been consistent over adjacent birth months.11,22,23 Provided the clear seasonality of neonatal vitamin D status as seen in this and other studies,16,17 a more pronounced and consistent month of birth pattern may have been expected should vitamin D at birth be of importance for MS risk. Furthermore, the associations between month of birth and risk of MS have recently been questioned based on a concern of inadequate 5
6
1.0, ref
1.0 (0.68–1.43)
0.9 (0.65–1.37)
1.0 (0.67–1.42)
1.0 (0.68–1.44)
1, lowest
2
3
4
5, highest
0.8 (0.54–1.26)
0.8 (0.55–1.28)
0.8 (0.54–1.25)
0.9 (0.56–1.30)
1.0, ref
Crudeb OR (95% CI)
0.9 (0.54–1.33)
0.9 (0.59–1.42)
0.8 (0.50–1.20)
0.9 (0.56–1.34)
1.0, ref
Adjustedc OR (95% CI)
With Vitamin D–Related Covariates, n 5 885
0.9 (0.51–1.45)
0.9 (0.55–1.57)
1.0 (0.57–1.68)
1.2 (0.71–2.02)
1.0, ref
Cruded OR (95% CI)
0.9 (0.53–1.62)
1.0 (0.59–1.73)
1.0 (0.56–1.72)
1.3 (0.73–2.17)
1.0, ref
Adjustede OR (95% CI)
With Vitamin D–Related Covariates and 25(OH)D at Inclusion, n 5 550
0.9 (0.53–1.39)
1.0 (0.65–1.61)
1.0 (0.63–1.57)
0.8 (0.53–1.33)
1.0, ref
Crudef OR (95% CI)
0.8 (0.52–1.36)
1.0 (0.63–1.56)
1.0 (0.61–1.53)
0.8 (0.52–1.31)
1.0, ref
Adjustedg OR (95% CI)
With Known Socioeconomic Group, n 5 735
0.9 (0.60–1.36)
0.9 (0.62–1.36)
0.9 (0.62–1.38)
1.0 (0.66–1.47)
1.0, ref
Crudeh OR (95% CI)
0.9 (0.58–1.34)
0.9 (0.59–1.32)
0.9 (0.57–1.30)
0.9 (0.62–1.41)
1.0, ref
Adjustedi OR (95% CI)
With Known MS Heredity, n 5 997
b
Adjusted for sex, residential area, and age group. Adjusted for sex, residential area, and age group. Subjects with information regarding month of birth, latitude of birth, Scandinavian parent, breastfeeding, sun exposure, vitamin D intake from dairy products, fatty fish consumption, smoking, and BMI at 20 years of age. c Adjusted for sex, residential area, age group, month of birth, latitude of birth, Scandinavian parent, breastfeeding, sun exposure, vitamin D intake from dairy products, fatty fish consumption, smoking, and BMI at 20 years of age. d Adjusted for sex, residential area, and age group. Subjects with known 25(OH)D at inclusion in the study and information regarding month of birth, latitude of birth, Scandinavian parent, breastfeeding, sun exposure, vitamin D intake from dairy products, fatty fish consumption, smoking, and BMI at 20 years of age. e Adjusted for sex, residential area, age group, month of birth, latitude of birth, Scandinavian parent, breastfeeding, sun exposure, vitamin D intake from dairy products, fatty fish consumption, smoking, BMI at 20 years of age, 25(OH)D at inclusion in the study, and month of 25(OH)D measurement. f Adjusted for sex, residential area, and age group. Subjects with known socioeconomic group. g Adjusted for sex, residential area, age group, and socioeconomic group. h Adjusted for sex, residential area, and age group. Subjects with information about MS heredity. i Adjusted for sex, residential area, age group, and MS heredity. 25(OH)D 5 25-hydroxyvitamin D; BMI 5 body mass index; CI 5 confidence interval; MS 5 multiple sclerosis; OR 5 odds ratio; ref 5 reference.
a
Whole Study Population, n 5 1,121, Crudea OR (95% CI)
Neonatal 25(OH)D Quintile
TABLE 2. Crude and Adjusted ORs and 95% CIs for Risk of MS by Birth Year–Specific Neonatal 25(OH)D Quintile
ANNALS
of Neurology
Volume 00, No. 00
Ueda et al: Neonatal Vitamin D and MS
TABLE 3. ORs with 95% CIs for Risk of Multiple Sclerosis by Birth Year–Specific Neonatal 25(OH)D Quintile and per 10nmol/l Higher Neonatal 25(OH)D
Neonatal 25(OH)D Quintile
Birth Years 1975–1979, n 5 508, ORa (95% CI)
1980–1984, n 5 353, ORa (95% CI)
1985–1989, n 5 214, ORa (95% CI)
1990–1994, n 5 46, ORa (95% CI)
1, lowest
1.0, ref
1.0, ref
1.0, ref
1.0, ref
2
0.9 (0.51–1.55)
0.9 (0.45–1.68)
1.4 (0.57–3.32)
1.4 (0.20–9.84)
3
0.8 (0.47–1.47)
0.7 (0.35–1.36)
1.7 (0.70–4.01)
1.7 (0.19–15.73)
4
1.1 (0.62–1.89)
0.7 (0.34–1.30)
1.4 (0.59–3.43)
0.5 (0.03–9.66)
5, highest
1.2 (0.70–2.15)
0.6 (0.31–1.22)
1.1 (0.47–2.80)
1.4 (0.16–12.90)
Per 10nmol/l higher neonatal 25(OH)D
1.1 (0.95–1.20)
0.9 (0.78–1.01)
1.0 (0.86–1.13)
1.0 (0.67–1.48)
Results are shown by 5-year intervals of birth years. a Adjusted for sex, residential area, and age group. 25(OH)D 5 25-hydroxyvitamin D; CI 5 confidence interval; OR 5 odds ratio; ref 5 reference.
adjustments for potentially confounding factors such as year and place of birth in many but not all9,21 of the studies.24,25 In accordance with our findings, an animal model of MS showed that vitamin D supplementation in the juvenile/adolescence period but not in the gestation/ lactation period attenuated central nervous system inflammation and demyelation.26 In a questionnaire study from the United States, based on the Nurses’ Health Study II, the association between maternal dietary vitamin D and milk intake during pregnancy, and daughters’ risk of developing MS was assessed. In nurses whose mothers reported drinking 2 or 3 glasses of milk per day during pregnancy, a reduced risk of MS was seen compared to nurses whose mothers drank
Neonatal Vitamin D Status and Risk of Multiple Sclerosis €rnhielm, MD,3 Peter Ueda, MD, PhD,1 Farshid Rafatnia, MD,2 Maria B€ aa Robin Fr€ obom,3 Greg Korzunowicz,3 Ragnar L€ onnerbro,3 Anna Karin Hedstr€ om, MD,3 Darryl Eyles, PhD,4 Tomas Olsson, MD, PhD,5 and Lars Alfredsson, PhD3 Objective: Low vitamin D status at birth may be associated with risk of adult onset multiple sclerosis, but this link has not been studied directly. We assessed the relation between neonatal vitamin D concentrations, measured in stored blood samples, and risk of multiple sclerosis. Methods: This was a population-based case–control study in Sweden including 459 incident cases of multiple sclerosis and 663 controls, randomly drawn from a national population registry and frequency matched on sex, age, and residential area. Results: There was no association between neonatal 25-hydroxyvitamin D quintile and risk of multiple sclerosis (crude odds ratio 5 1.0, 95% confidence interval 5 0.68–1.44, for the highest quintile compared to the lowest). Adjusting for a number of potential confounding factors in early life (month of birth, latitude of birth, breastfeeding) and in adult life (25-hydroxyvitamin D, sun exposure, vitamin D intake from dairy products, fatty fish consumption, smoking, body mass index at 20 years of age) as well as ancestry, multiple sclerosis heredity, and socioeconomic group did not considerably affect the result. Interpretation: At a broad population level, 25-hydroxyvitamin D at birth was not associated with risk of multiple sclerosis. ANN NEUROL 2014;00:000–000
T
he etiology of multiple sclerosis (MS) is considered an interplay of genetics and environmental factors.1 Epidemiological observations regarding sunlight exposure in childhood,2 dietary vitamin D intake,3 and serum 25hydroxyvitamin D (25[OH]D) concentrations4 have highlighted low vitamin D as a potential risk factor for the disease. A gestational component of MS etiology has been proposed based on a higher concordance rate of MS in dizygotic twins—who are exposed to the same intrauterine environment and timing of birth—as compared with siblings who are not twins.1 Reports from the northern hemisphere indicate that risk of MS is increased among those born during spring, in particular in the months of April and May, compared to the autumn-born.5–8
The pattern is reversed in the southern hemisphere, where individuals born in November and December are at the highest risk.9 It has been suggested that low sunlight exposure and associated low vitamin D status during gestation for individuals born after winter increase susceptibility to MS.1,5–11 The hypothesis has been supported by a study from the United States predicting gestational vitamin D levels in mothers of MS cases using questionnaire data about maternal characteristics and diet during pregnancy, and latitude and season of delivery.12 To date, no study has directly assessed the relation between vitamin D at birth and risk of MS. By measuring vitamin D in stored neonatal blood samples, we had the opportunity to investigate this hypothesis in a large Swedish population-based case–control cohort,
View this article online at wileyonlinelibrary.com. DOI: 10.1002/ana.24210 Received Mar 19, 2014, and in revised form Jun 27, 2014. Accepted for publication Jun 27, 2014. Address correspondence to Dr Ueda, Clinical Epidemiology Unit, T3 Eugeniahemmet, Karolinska University Hospital Solna, 17176 Stockholm, Sweden. E-mail address: [email protected] From the 1Clinical Epidemiology Unit, Department of Medicine, Solna, Karolinska Institutet, Stockholm, Sweden; 2Internal Medicine Department, Karolinska University Hospital, Stockholm, Sweden; 3Institute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden; 4Queensland Centre for Mental Health Research, Queensland Brain Institute, University of Queensland, Brisbane, Queensland, Australia; and 5Neuroimmunology Unit, Department of Clinical Neuroscience and Center for Molecular Medicine, Karolinska Institute, Stockholm, Sweden. Additional supporting information can be found in the online version of this article.
C 2014 American Neurological Association V 1
ANNALS
of Neurology
FIGURE 1: Flow diagram for study population. EIMS 5 Epidemiological Investigation of Multiple Sclerosis.
comprising information about various environmental factors of potential importance for MS risk, including vitamin D status at inclusion in the study. We hypothesized that low neonatal vitamin D status would be associated with increased risk of MS.
Subjects and Methods Study Population Epidemiological Investigation of MS (EIMS) is a population based case–control study using incident cases of MS. The study base comprises the population aged 16 to 70 years in Sweden. Starting from April 2005, cases were recruited from 40 study centers, including all university hospitals in Sweden. Cases were diagnosed by a neurologist according to the McDonald criteria. For each case, 2 controls—frequency matched by age (predetermined 5-year age groups), sex, and residential area—were randomly selected from the National Population Register. A standardized questionnaire was used to obtain data on exposures and other information of relevance. The questionnaires were sent to cases and controls and filled out at home. The Swedish Phenylketonuria (PKU) Register is a near complete biobank containing dried neonatal blood samples from screening of rare metabolic diseases in infants born in Sweden since 1975.
2
Participants of EIMS who were born in Sweden from 1975 and onward, who had been asked to take part in EIMS by October 2011, and who had given consent for the use of blood samples for research purposes were considered for inclusion in the study. Of the subjects born in 1975 and onward, 563 (89%) of the cases and 1,235 (66%) of the controls returned their questionnaire. Of these subjects, 500 cases and 1,070 controls stated that they were born in Sweden. A total of 1,219 participants, 497/500 (99%) cases and 722/1,070 (67%) controls, had given consent for analysis of blood samples. For 16 of these subjects (7 cases and 9 controls), information required for identification of neonatal blood samples was not available, and for 31 cases and 50 controls, neonatal blood samples could not be found in the PKU Register. The final study population thus consisted of 1,122 subjects, of whom 459 were cases and 663 were controls (Fig 1). The study was approved by the regional ethical review board in Stockholm (2012/100-32 and 04-252/1-4). Written consent for participation in EIMS and for analysis of blood samples was obtained from all individuals included in the study.
Measurement of Neonatal Vitamin D Concentrations Two 3.2mm punches were obtained from the participants’ neonatal dried blood spots. Samples were stored at room temperature between 1975 and August 1981, and thereafter in cold
Volume 00, No. 00
Ueda et al: Neonatal Vitamin D and MS
storage (4 C). It is likely that the samples degraded somewhat during the time in storage. 25(OH)D is the metabolite used for assessing vitamin D status13 and can exist in 2 forms: 25(OH)D3 and 25(OH)D2. The latter can only be obtained from certain dietary sources, including supplements. 25(OH)D3 and 25(OH)D2 in dried blood spots were measured with a highly sensitive liquid chromatography tandem mass spectroscopy assay with multiple reactant monitoring.14 The method uses minimal sample cleanup to reduce sample loss during extraction and chemical derivatization to enhance 25(OH)D3 ionization.14 Using samples from Australia and Denmark, it has been shown that the assay can satisfyingly detect seasonal (within year) differences in 25(OH)D3 concentrations and that the measurements are strongly correlated with neonatal cord blood concentrations (r 5 0.86).15 The assay has been used on samples stored up to 35 years16,17 and evaluated with regard to punch position, spot volume, and paper type.18 The punches in this study were consistently sampled from the outer part of the blood spot as outlined by Kvaskoff et al.18 The concentration of 25(OH)D—including 25(OH)D2 and 25(OH)D3—is reported in nanomoles per liter. 25(OH)D2 and 25(OH)D3 are highly protein-bound molecules that are excluded from erythrocytes.18 To make results comparable with existing studies, results are presented as sera concentrations of hematocrit in capillary blood as previously described.15
Other Variables Information about postnatal age of the participant when the blood was sampled and maternal age at birth were recorded from the PKU Register. Age at inclusion in EIMS was categorized into the following 6 groups: 16 to 19, 20 to 24, 25 to 29, 30 to 34, 35 to 39, and 40 to 44 years. Participants were considered as Scandinavian if they had at least 1 parent of Scandinavian (Sweden, Norway, Finland, Denmark, Iceland) origin. Latitude of residential area of birth was categorized into 54 to 58 , 59 , and 60 to 69 north. Body mass index (BMI) at 20 years of age was calculated using self-reported weight and height data from the questionnaire and presented as a continuous variable. Weekly intake of vitamin D through dairy products including milk, yogurt, cheese, and butter was estimated and presented in international units (continuous variable) by multiplying the amount and frequency of the dietary item by its vitamin D content. Consumption of fatty fish (e.g. salmon and herring) 43 or more per month was considered a high fatty fish intake, whereas 33/wk, No. (%)
82 (18)
133 (20)
Sun exposure index, mean [SD]
3.4 [1.7]
3.7 [1.2]
Smoker, No. (%)
213 (47)
258 (39)
65.0 [27.7]
67.8 [26.2]
Manual workers
135 (29)
234(35)
Nonmanual employees
162 (35)
205 (31)
Unknown
162 (35)
223 (34)
No university education
169 (37)
274 (41)
University studies without a degree
148 (32)
188 (28)
University studies with degree
142 (31)
200 (30)
Latitude of birth, No. (%) 54–58 north
59 north
60–69 north
MS heredity, No. (%)
Breastfed, No. (%) 2
a
25(OH)D at inclusion , nmol/l [SD] Socioeconomic group, No. (%)
Education, No. (%)
a
Measurements available for 298 cases and 307 controls 25(OH)D 5 25-hydroxyvitamin D BMI 5 body mass index; MS 5 multiple sclerosis; SD 5 standard deviation.
Comparison with Other Studies The hypothesis that exposure to low vitamin D in utero, in particular during the last trimester, increases risk of developing MS is based on 12 studies5,21 from the northern hemisphere (including a report from Sweden11) and 1 from the southern hemisphere showing an excess of spring births among MS cases.5 However, the observed risk increase of being born in spring has been modest (odds ratio of about 1.1 compared to being born in Month 2014
other seasons), and the season of birth pattern for MS risk has not been consistent over adjacent birth months.11,22,23 Provided the clear seasonality of neonatal vitamin D status as seen in this and other studies,16,17 a more pronounced and consistent month of birth pattern may have been expected should vitamin D at birth be of importance for MS risk. Furthermore, the associations between month of birth and risk of MS have recently been questioned based on a concern of inadequate 5
6
1.0, ref
1.0 (0.68–1.43)
0.9 (0.65–1.37)
1.0 (0.67–1.42)
1.0 (0.68–1.44)
1, lowest
2
3
4
5, highest
0.8 (0.54–1.26)
0.8 (0.55–1.28)
0.8 (0.54–1.25)
0.9 (0.56–1.30)
1.0, ref
Crudeb OR (95% CI)
0.9 (0.54–1.33)
0.9 (0.59–1.42)
0.8 (0.50–1.20)
0.9 (0.56–1.34)
1.0, ref
Adjustedc OR (95% CI)
With Vitamin D–Related Covariates, n 5 885
0.9 (0.51–1.45)
0.9 (0.55–1.57)
1.0 (0.57–1.68)
1.2 (0.71–2.02)
1.0, ref
Cruded OR (95% CI)
0.9 (0.53–1.62)
1.0 (0.59–1.73)
1.0 (0.56–1.72)
1.3 (0.73–2.17)
1.0, ref
Adjustede OR (95% CI)
With Vitamin D–Related Covariates and 25(OH)D at Inclusion, n 5 550
0.9 (0.53–1.39)
1.0 (0.65–1.61)
1.0 (0.63–1.57)
0.8 (0.53–1.33)
1.0, ref
Crudef OR (95% CI)
0.8 (0.52–1.36)
1.0 (0.63–1.56)
1.0 (0.61–1.53)
0.8 (0.52–1.31)
1.0, ref
Adjustedg OR (95% CI)
With Known Socioeconomic Group, n 5 735
0.9 (0.60–1.36)
0.9 (0.62–1.36)
0.9 (0.62–1.38)
1.0 (0.66–1.47)
1.0, ref
Crudeh OR (95% CI)
0.9 (0.58–1.34)
0.9 (0.59–1.32)
0.9 (0.57–1.30)
0.9 (0.62–1.41)
1.0, ref
Adjustedi OR (95% CI)
With Known MS Heredity, n 5 997
b
Adjusted for sex, residential area, and age group. Adjusted for sex, residential area, and age group. Subjects with information regarding month of birth, latitude of birth, Scandinavian parent, breastfeeding, sun exposure, vitamin D intake from dairy products, fatty fish consumption, smoking, and BMI at 20 years of age. c Adjusted for sex, residential area, age group, month of birth, latitude of birth, Scandinavian parent, breastfeeding, sun exposure, vitamin D intake from dairy products, fatty fish consumption, smoking, and BMI at 20 years of age. d Adjusted for sex, residential area, and age group. Subjects with known 25(OH)D at inclusion in the study and information regarding month of birth, latitude of birth, Scandinavian parent, breastfeeding, sun exposure, vitamin D intake from dairy products, fatty fish consumption, smoking, and BMI at 20 years of age. e Adjusted for sex, residential area, age group, month of birth, latitude of birth, Scandinavian parent, breastfeeding, sun exposure, vitamin D intake from dairy products, fatty fish consumption, smoking, BMI at 20 years of age, 25(OH)D at inclusion in the study, and month of 25(OH)D measurement. f Adjusted for sex, residential area, and age group. Subjects with known socioeconomic group. g Adjusted for sex, residential area, age group, and socioeconomic group. h Adjusted for sex, residential area, and age group. Subjects with information about MS heredity. i Adjusted for sex, residential area, age group, and MS heredity. 25(OH)D 5 25-hydroxyvitamin D; BMI 5 body mass index; CI 5 confidence interval; MS 5 multiple sclerosis; OR 5 odds ratio; ref 5 reference.
a
Whole Study Population, n 5 1,121, Crudea OR (95% CI)
Neonatal 25(OH)D Quintile
TABLE 2. Crude and Adjusted ORs and 95% CIs for Risk of MS by Birth Year–Specific Neonatal 25(OH)D Quintile
ANNALS
of Neurology
Volume 00, No. 00
Ueda et al: Neonatal Vitamin D and MS
TABLE 3. ORs with 95% CIs for Risk of Multiple Sclerosis by Birth Year–Specific Neonatal 25(OH)D Quintile and per 10nmol/l Higher Neonatal 25(OH)D
Neonatal 25(OH)D Quintile
Birth Years 1975–1979, n 5 508, ORa (95% CI)
1980–1984, n 5 353, ORa (95% CI)
1985–1989, n 5 214, ORa (95% CI)
1990–1994, n 5 46, ORa (95% CI)
1, lowest
1.0, ref
1.0, ref
1.0, ref
1.0, ref
2
0.9 (0.51–1.55)
0.9 (0.45–1.68)
1.4 (0.57–3.32)
1.4 (0.20–9.84)
3
0.8 (0.47–1.47)
0.7 (0.35–1.36)
1.7 (0.70–4.01)
1.7 (0.19–15.73)
4
1.1 (0.62–1.89)
0.7 (0.34–1.30)
1.4 (0.59–3.43)
0.5 (0.03–9.66)
5, highest
1.2 (0.70–2.15)
0.6 (0.31–1.22)
1.1 (0.47–2.80)
1.4 (0.16–12.90)
Per 10nmol/l higher neonatal 25(OH)D
1.1 (0.95–1.20)
0.9 (0.78–1.01)
1.0 (0.86–1.13)
1.0 (0.67–1.48)
Results are shown by 5-year intervals of birth years. a Adjusted for sex, residential area, and age group. 25(OH)D 5 25-hydroxyvitamin D; CI 5 confidence interval; OR 5 odds ratio; ref 5 reference.
adjustments for potentially confounding factors such as year and place of birth in many but not all9,21 of the studies.24,25 In accordance with our findings, an animal model of MS showed that vitamin D supplementation in the juvenile/adolescence period but not in the gestation/ lactation period attenuated central nervous system inflammation and demyelation.26 In a questionnaire study from the United States, based on the Nurses’ Health Study II, the association between maternal dietary vitamin D and milk intake during pregnancy, and daughters’ risk of developing MS was assessed. In nurses whose mothers reported drinking 2 or 3 glasses of milk per day during pregnancy, a reduced risk of MS was seen compared to nurses whose mothers drank
