Studies analyzing the effect of gestational exposure to vitamin D fortified margarine on the risk of gestational diabetes [
17], pre-eclampsia [
18], childhood asthma [
20], adult celiac disease [
21], and inflammatory bowel disease (IBD) [
22] later in life were subsequently published. The termination of the Danish fortification program in June 1985 constituted the basis of the study designs in these studies. The risk of the selected outcomes in birth cohorts from two years before (i.e., exposed) and after (i.e., unexposed) the fortification termination, separated by 15 months of washout period, was compared. Exposed and unexposed cohorts were followed for equal amounts of time in the Danish National Patient Register [
30], to detect specified outcomes (see
Figure 1).
For gestational diabetes and pre-eclampsia outcomes, the focus was on the first diagnosis of the relevant outcomes in women from the selected cohorts, who gave birth for the first time at age 14–27 years for pre-eclampsia [
18] or at age 20–27 years for gestational diabetes [
17]. These women were identified using information from the Danish Medical Birth Register [
31], where information on covariates (e.g., mother’s smoking during pregnancy and pre-pregnancy BMI, offspring’s gestational age at birth and sex) was also obtained. For asthma outcomes, the focus was on the time of the first inpatient asthma diagnosis admission at age 0–9 years [
20]. For celiac disease and IBD, the focus was on the first admission with a diagnosis at age 0–30 years [21,22. Due to the design, where individuals from entire adjacent birth cohorts were examined, unmeasured covariates were considered to be equally distributed between adjacent exposed and unexposed birth cohorts. Depending on the type of the outcome (i.e., time to the hospital-based first diagnosis, or the presence of the hospital-based diagnosis), Cox or logistic regression models were run, consequently running stratified analyses considered relevant (e.g., stratified by smoking status during pregnancy for pre-eclampsia [
18] or by sex for asthma and IBD [
20,
22]). The effect of the season of birth was investigated by adjusting the analyses by month of birth (for asthma [
20]), including exposure and season (or month) of birth interaction into regression modeling (for asthma [
20], celiac disease [
20], and IBD [
22]), or running analyses stratified by season of birth (for gestational diabetes [
17]). Prenatally exposed women born in spring, when pregnant later in life, had a lower risk of developing gestational diabetes compared to unexposed women born in the same season (OR 0.68, 95% CI 0.50, 0.94) [
17]. Prenatally exposed women, if smoking when pregnant later in life, had a lower risk of developing pre-eclampsia compared to unexposed and smoking women (OR 0.49, 94% CI 0.34, 0.72) [
18]. A decreased risk for asthma was shown in boys under 3 years of age exposed to the fortification, compared to boys of the same age not exposed to fortification (HR 0.78, 95% CI 0.67, 0.92) with no modification effect by month of birth [
20]. ORs for developing celiac disease were similar in exposed and unexposed cohorts, with no significant modification effect by season of birth [
21]. Lower odds of IBD were found for individuals under the age of 30 exposed to the fortification prenatally, compared to individual not exposed (OR 0.87, 95% CI 0.79, 0.95), and no significant effect of season of birth was detected [
22]. The specific feature of these studies was the use of hospital-based diagnoses, thus leaving out mild and moderate asthma cases diagnosed in primary care [
20] and raising some questions regarding the validity of the gestational diabetes diagnosis [
17]. Another specific feature of all these studies (except the one on IBD, where the secular trend in disease incidence could partly explain the results [
22]) was that no secular trends in the incidence of the investigated outcomes were identified in Denmark based on the obtained data from 1983 to 1988. It was not known, however, whether there were any secular changes over the longer period. Moreover, secular changes could occur in alternative exposure sources (i.e., vitamin D intake via food), and potential confounders (e.g., mother’s socioeconomic status, mother’s exposure to infections and physical activity), and the latter was explored if the data was available from the Danish Medical Birth Register [
31]. The strengths of these studies were their novelty, long follow-ups, and large sample sizes with up to 200,000 individuals.