Hindawi Journal of Diabetes Research Volume 2022, Article ID 6598600, 8 pages https://doi.org/10.1155/2022/6598600 Research Article Prenatal Betamethasone Exposure and its Impact on Pediatric Type 1 Diabetes Mellitus: A Preliminary Study in a Spanish Cohort David Perna-Barrull ,1 Marta Murillo ,2 Nati Real,2 Laia Gomez-Muñoz ,1 Silvia Rodriguez-Fernandez ,1 Joan Bel ,2 Manel Puig-Domingo ,3 and Marta Vives-Pi 1 1 Immunology Service Germans Trias i Pujol Research Institute and University Hospital, Autonomous University of Barcelona, 08916 Badalona, Spain 2 Pediatrics Service Germans Trias i Pujol Research Institute and University Hospital, Autonomous University of Barcelona, 08916 Badalona, Spain 3 Endocrinology Service Germans Trias i Pujol Research Institute and University Hospital, Autonomous University of Barcelona, 08916 Badalona, Spain Correspondence should be addressed to Marta Vives-Pi; mvives@igtp.cat Received 28 October 2021; Revised 24 January 2022; Accepted 18 February 2022; Published 10 March 2022 Academic Editor: Marco Infante Copyright © 2022 David Perna-Barrull et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Background. Betamethasone, a glucocorticoid used to induce lung maturation when there is a risk of preterm delivery, can affect the immune system maturation and type 1 diabetes (T1D) incidence in the progeny. It has been described that prenatal betamethasone protects offspring from experimental T1D development. The main aim of this study was to evaluate the possible association between betamethasone prenatal exposure and T1D in humans. Research Design and Methods. A retrospective case-control study with a total of 945 children, including 471 patients with T1D and 474 healthy siblings, was performed. Participants were volunteers from the Germans Trias i Pujol Hospital and DiabetesCero Foundation. Parents of children enrolled in the study completed a questionnaire that included questions about weeks of gestation, preterm delivery risk, weight at birth, and prenatal betamethasone exposure of their children. Multiple logistic regression was used to detect the association between betamethasone exposure and T1D. Results. We compared T1D prevalence between subjects prenatally exposed or unexposed to betamethasone. The percent of children with T1D in the exposed group was 37.5% (21 of 56), and in the unexposed group was 49.52% (410 of 828) (p = 0:139). The percentage of betamethasone-treated subjects with T1D in the preterm group (18.05%, 13 of 72) was significantly higher than that found in the control group (12.5%, 9 of 72) (p = 0:003). The odds ratio for T1D associated with betamethasone in the univariate logistic regression was 0.59 (95% confidence interval, 0.33; 1.03 [p = 0:062]) and in the multivariate logistic regression was 0.83 (95% confidence interval, 0.45; 1.52 [p = 0:389]). Conclusions. The results demonstrate that the prenatal exposure to betamethasone does not increase T1D susceptibility, and may even be associated with a trend towards decreased risk of developing the disease. These preliminary findings require further prospective studies with clinical data to confirm betamethasone exposure effect on T1D risk. 2 Journal of Diabetes Research 1. Introduction 2. Materials and Methods Type 1 diabetes (T1D) is amongst the most common endo- 2.1. Design. This is a retrospective case-control study con- crine disorder in children and adolescents. Over the last ducted between 2018 and 2021 in Spain. A questionnaire decades, the incidence of T1D during childhood is increas- was carried out to evaluate the possible association between ing with an average of 3–4% per year [1]. T1D is caused by prenatal betamethasone exposure and the T1D develop- the autoimmune destruction of pancreatic β-cells, and ment. A total of 945 children participated in the study. The despite genetic risk factors have been identified, the influ- primary population explored was a group of 471 children ence of environmental factors is of foremost importance being diagnosed with T1D, which was compared with a con- and has been exhaustively studied in large patient cohorts trol group constituted by their sisters and brothers (n = 474). such as TEDDY study (The Environmental Determinants Within these two groups, betamethasone treatment (two of Diabetes in the Young) [2], or DIPP (Type 1 Diabetes Pre- doses of 12 mg, 24 h apart) [13] at the time of late gestation diction and Prevention) study [3]. Among them, was assessed in order to elucidate whether betamethasone environment-related prenatal changes might have an influ- could have a protective effect against T1D development in ence on the development of T1D. In fact, the in utero envi- those children prenatally exposed to the drug. Parents of ronment is critical for disease development, as suggested by children enrolled in the study were requested to complete a the fact that dizygotic twins display an increased concor- questionnaire that included questions about gender, diagno- dance of T1D when compared to nontwin siblings [4]. Envi- sis of T1D, age at diagnosis, weeks of gestation, weight at ronmental factors in utero such as iron overload or low zinc birth, and prenatal betamethasone exposure of their children concentration in drinking water [5, 6] correlate with an (Table 1). The questionnaire was conducted in two settings: increased risk to develop autoimmune diabetes. Moreover, in the Pediatrics Service of the Germans Trias i Pujol Hospi- the gestational age has impact on the T1D risk. In general tal (Badalona, Spain) and in DiabetesCero Foundation terms, preterm newborns showed an increased risk to (Spain), a non-profit organization made up of parents of develop the disease [7, 8] but some controversy exists children with T1D. Questionnaire forms were distributed regarding very preterm birth (birth before 32 weeks of gesta- by email or in person. An electronic database was set-up to tion) that seems to protect against T1D [9]. record entries in submitted questionnaire forms. The ques- Preterm birth, defined as birth occurring at less than 37 tionnaire was approved by the Ethics Committee for weeks of gestation, is a serious and common pregnancy Research at the Germans Trias i Pujol Hospital. All partici- complication reaching 5–11% of live births in developed pants were informed of the nature and purpose of the ques- countries [10, 11]. Administration of antenatal corticoste- tionnaire at the outset, and informed consent was obtained roids is recommended as the standard care for the manage- from legal representatives of all participants. ment of women at risk of imminent preterm delivery [12–14]. However, the benefits and harms of such a strategy 2.2. Sample Size and Statistical Methods. The sample size for are still under discussion, especially when administered after the study was calculated using the Cochran’s formula. Since week 34 of gestation or before cesarean delivery [15, 16]. there are no previous data on similar studies, we calculated Specifically, synthetic glucocorticoids, most often beta- the number of subjects per group with a statistical value (P methasone, are routinely administered to women at risk -value) of 0.05 and a potency value of 0.8. To that end, the of preterm delivery between 24 and 34 weeks of gestation estimated proportions for the control group were obtained to accelerate lung maturation and reduce the severity of from both bibliographic review of Spanish preterm birth rate respiratory distress syndrome, therefore improving the sur- studies (that account for 8.7% of total births) [23] and from vival rates of premature infants. Recent studies in experi- the percentage of antenatal corticosteroids administered in mental models of T1D show that prenatal betamethasone those cases (62% of women at risk of preterm delivery are affects the two main players in this disease, the immune treated with betamethasone) [24]. This resulted in an esti- system and the pancreatic β-cells, and that corticosteroid- mated proportion for the control group of 5.22% subjects related changes may have long-term consequences in the exposed to betamethasone. The estimated proportions for offspring [17–19]. Since both the immune system and the the T1D group (P1 ) were calculated using the formula P1 pancreatic islets are still developing until birth, glucocorti- = OR ∗ P2 /(1-P2) + (OR∗P2 ). In order to obtain the odds coid exposure [20] may affect islet function [21] and T1D ratio (OR), previous results on the offspring of nonobese susceptibility. Data from both experimental and clinical diabetic (NOD) mice after betamethasone exposure were studies suggests a link between prenatal exposure to gluco- used (OR = 0:287) [17]. The P2 was obtained from the corticoids and alterations in immune traits in the offspring already calculated proportion of betamethasone-exposed [22]. In summary, antenatal glucocorticoids exert an anti- individuals in the control group (P2 = 0:0522). The formula inflammatory effect, contributing to a switch towards a offered an estimated proportion for the T1D group of Th2 response. 1.56% subjects exposed to betamethasone. Cochran’s for- The main aim of the current study was to evaluate the mula comprising the calculated proportions (P1 and P2 ) possible association between prenatal betamethasone expo- resulted in a sample size of 378 subjects per group. sure and T1D development in humans, as a previous step Statistical analysis was performed using Prism 9.0 soft- for future studies on human cohorts that will help to eluci- ware (GraphPad Software INC., San Diego, CA, USA) and date the role of this drug in the susceptibility to T1D. statistical software R (Vienna, Austria). Descriptive statistics Journal of Diabetes Research 3 Table 1: Clinical characteristics of the subjects in the study and control groups. Characteristics T1D group Control group p value Number, n (%) 471 (49.8) 474 (50.2) Females, n (%) 192 (47.8)† 193 (47.7)† 0.977‡ Males, n (%) 210 (52.2)† 212 (52.3)† 0.977‡ Age at diagnosis (years, mean ± SD) 7:42 ± 4:94 NA Gestation weeks (mean ± SD) 39:02 ± 1:92 38:56 ± 2:22 0.0011§ Birth weight (kg, mean ± SD) 3:24 ± 0:52 3:21 ± 0:61 0.389§ Betamethasone, n (%) 21 (4.73) 35 (7.77) 0.0166‡ Preterm birth (≤36 weeks of gestation), n (%) 41 (8.70) 60 (13.07) 0.0145‡ † Missing gender accounts for 14.65% (69 patients) in subjects with T1D and 14.56% (69 subjects/participants) in control subjects. Subjects with missing gender have not been considered for gender ratio determination. ‡Chi-squared test; §T-test. NA: not applicable; T1D: type 1 diabetes; SD: standard deviation. were applied to the collected data. Continuous variables with T1D of the exposed group was 37.5% (21 of 56), and (weeks of gestation and birth weight) were analyzed using the percentage of subjects with T1D of the unexposed group a T-test or an analysis of variance (ANOVA) to compare was 49.52% (410 of 828). Despite nonsignificant (p = 0:139), between variables. Categorical variables (T1D, exposure to this result seems to suggest a biological trend. betamethasone and gender) were analyzed using a chi- The percentage of subjects exposed to betamethasone squared test to compare between variables. For correlations, with or without T1D depending on gestational age is shown a nonparametric Spearman’s test was performed. To com- in Figure 1(b). Regarding prematurity (less than 37 weeks of pare T1D prevalence between betamethasone exposed and gestational age), we observed a significantly higher percent- unexposed subjects a Logrank Mantel-Cox test was age of preterm babies in the control group (13.07%, 60 of performed. 474) in comparison to the T1D group (8.70%, 41 of 471) Multiple logistic regression was used to detect the associ- (p = 0:0145). The percentage of subjects with T1D and con- ation between betamethasone exposure and T1D after trols exposed to betamethasone was represented in each sub- adjusting for gender, weeks of gestation, and birth weight. group of gestation duration. As expected, in the control Bioinformatic analysis was carried out by the Statistics and group the highest percentage of betamethasone exposure Bioinformatics Unit (UEB) from Vall d’Hebron Research was found in the very preterm condition. Surprisingly, in Institute (VHIR, Barcelona, Spain). Subjects with missing the T1D group, we observed a higher percentage of beta- values for one or more variables were not added in the statis- methasone exposure in the preterm group (18.05%, 13 of tical analysis. Values with a p < 0:05 were considered statis- 72) when compared to the very preterm group (5%, 1 of tically significant. 20) whereas in the control group this value decreased with the duration of gestation. Moreover, this high percentage 3. Results of betamethasone-treated T1D subjects (18.05%, 13 of 72) in the preterm group is significantly higher than that found Data from 945 children were collected (Table 1) and classi- in the control group (12.5%, 9 of 72) (p = 0:003). The per- fied into two groups: patients with T1D diagnosed during centage of betamethasone exposure in deliveries from 37 childhood or puberty (T1D group) and their siblings (con- weeks of gestation showed a tendency to decrease in the trol group). Males were slightly predominant over females T1D group when compared to controls: early term (1.21 in both groups (52.3% and 47.7% vs. 52.2 and 47.8%, respec- vs. 4.44%, 3 and 11 of 248, respectively) and term (0.85 vs. tively). The gender was missing in 14.6% of total individuals. 1.28%, 4 and 6 of 470, respectively) (Figure 1(b)). As The age at T1D diagnosis was 7:42 ± 4:94 years (mean ± SD expected, a high percentage of preterm newborns received ). Children with T1D were born after a significantly and prenatal betamethasone (Figure 1(c)). A direct correlation slightly longer gestation period than control group was found between weeks of gestation and T1D in the off- (39:02 ± 1:93 weeks vs. 38:56 ± 2:22 weeks; p = 0:0011), but spring (Figure 1(d)). no differences were observed in the birth weight between As detailed in Supplementary Figure 1A, the groups. Only 5.93% (56 of 945) of the subjects in the study predominant age at clinical onset was between 1 and 12 were prenatally exposed to betamethasone, and a signifi- years. With regard to children in the T1D group that were cantly lower number of children who received prenatal beta- exposed to betamethasone in utero, the majority of them methasone were in the T1D group (4.73%, 21 of 471), when were diagnosed at an early age (from 0 to 7 years) (5.88%, compared to the control group (7.77%, 35 of 474) 15 of 255), whereas a minority was diagnosed with the (p = 0:0166). Similarly, a signifanctly lower number of pre- disease at a later age (from 8 to 17 years) (2.97%, 6 of 202) term babies was found in the T1D group when compared (Supplementary Figure 1B). to controls (8.70% vs. 13.07%; p = 0:0145). Then, we analyzed perinatal data depending on the First, we compared the prevalence of T1D between both exposure to betamethasone. As expected, betamethasone- groups. Figure 1(a) shows that the percentage of subjects exposed children displayed a lower birth weight and fewer 4 Journal of Diabetes Research 100 gestation duratin distribution) % of subjects (normalized by 100 80 % subjects with T1D 80 60 60 p = 0.139 40 40 20 20 0 0 Very preterm Preterm Early term Term Post-term 0 2 4 6 8 10 12 14 16 18 Age at T1D diagnosis (years) Prenatal betamethasone Exposed Weeks of gestation Unexposed Bmet-exposed subjects with T1D Bmet-exposed control subjects Bmet-unexposed subjects with T1D Bmet-unexposed control subjects (a) (b) 60 60 % betamethasone-exposed % subjects with T1D 40 40 subjects p = 0.01087 20 20 r = 0.718 R2 = 0.5040 0 0 30 35 40 45 30 35 40 45 Weeks of gestation Weeks of gestation (c) (d) Figure 1: Prenatal betamethasone exposure tends to reduce the percentage of subjects with type 1 diabetes. (a) Percentage of subjects with type 1 diabetes (T1D) in children prenatally exposed to betamethasone (continuous line, n = 56) and in unexposed children (dotted line, n = 828). Logrank Mantel-Cox test was used for statistical analysis. (b) Percentage of children in each gestational period: very preterm (28–32 week gestational age (GA), n = 20); preterm (33–36 week GA, n = 72); early term (37–38 week GA, n = 248); term (week 39–41 GA, n = 470); postterm (≥42 week GA, n = 50). White bars represent the percentage of the betamethasone-unexposed (Bmet-unexposed) control subjects; light grey bars represent the percentage of betamethasone-exposed (Bmet-exposed) control subjects; dark grey bars represent the percentage of Bmet-unexposed subjects with T1D; and black bars represent the percentage of Bmet-exposed subjects with T1D. All percentages were calculated with respect to the total number of subjects in each gestational period. (c) Percentages of children prenatally exposed to betamethasone in relation to weeks of gestation. (d) Correlation between the percentages of children with T1D and weeks of gestation (Spearman’s correlation analysis). weeks of gestation than unexposed children (p = 0:00001) (p = 0:062, Figure 2(a)). Then, we adjusted the betametha- (Table 2). Most children (85.25%, 763 of 895) had appropri- sone effects in a multivariate analysis. The variables weeks ate weight for gestational age, according to the Spanish of gestation and birth weight were considered as confound- birthweight charts [25]. Interestingly, we found that prenatal ing factors. The effect of betamethasone on T1D displayed exposure to betamethasone was more frequent in males than the same trend, although less evidently (OR 0.83 [95% CI, in females (p = 0:0257) (Table 2). 0.45; 1.52] [p = 0:389]). By contrast, the effect of weeks of Finally, we analyzed the data using a logistic regression gestation on T1D showed a statistically significant OR of model. First, we used a univariate approach between beta- 1.11 (95% CI, 1.03; 1.19 [p = 0:0052]) (Figure 2(b)). methasone and T1D development. The odds ratio (OR) for T1D associated with betamethasone was 0.59 (95% confi- 4. Discussion dence interval [95% CI], 0.33; 1.03 [p = 0:062]). Although nonsignificant, a trend towards a protective effect of beta- In this study, we address for the first time the relationship methasone against T1D development was observed between prenatal administration of betamethasone and Journal of Diabetes Research 5 Table 2: Variable distribution depending on betamethasone treatment. Betamethasone Variable Total p-value No Yes Missing Birth at weight‡ n = 822 n = 56 n = 47 n = 925 (kg, mean ± SD) 3:26 ± 0:53 2:63 ± 0:80 3:28 ± 0:53 3:23 ± 0:57 0.00001¶ (P25; P75) (2.94; 3.60) (2.02; 3.19) (2.85; 3.66) (2.90; 3.60) Duration of gestation§ n = 805 n = 56 n = 45 n = 906 (weeks, mean ± SD) 38:97 ± 1:93 35:7 ± 2:87 38:64 ± 1:97 38:79 ± 2:09 0.00001¶ (P25; P75) (38.00;40.00) (34.00;38.00) (37.00;40.00) (38.00;40.00) Female, n (%) 340 (40.4%) 19 (33.9%) 26 (54.2%) 385 (40.8%) Sex Male, n (%) 374 (44.5%) 26 (46.4%) 22 (45.8%) 422 (45.3%) 0.0257|| Missing, n (%) 127 (15.1%) 11 (19.6%) 0 (0%) 138 (13.9%) ‡ Birth weight was missing in 20 subjects (2.1%). § Duration of gestation was missing in 39 subjects (4.1%) ¶: ANOVA (analysis of variance). ||: chi-squared test. Betamethasone OR (95 %CI) No Yes 0.59 (0.33; 1.03) Missing 1.27 (0.71; 2.31) 0.0 0.5 1.0 1.5 2.0 2.5 Odds ratio (a) OR (95 %CI) No Betamethasone Yes 0.83 (0.45;1.52) exposure Missing 1.31 (0.71; 2.43) Female Gender Male 0.99 (0.75;1.31) Missing 0.99 (0.64;1.52) Weeks of gestation 1.11 (1.03;1.19)⁎ 0 1 2 3 Odds ratio (b) Figure 2: Prenatal betamethasone administration and type 1 diabetes risk. (a) Forest plot of the univariate logistic regression odds ratios (OR) and 95% confidence interval (95% CI) for the risk of type 1 diabetes (T1D) according to prenatal betamethasone exposure. (b) Forest plot of the multivariate logistic regression OR and 95% CI for the risk of T1D according to prenatal betamethasone exposure, gender, or week of gestation. T1D OR was adjusted for birth weight, gender, and weeks of gestation. ∗p < 0:01, multiple logistic regression. T1D development in the offspring. Previous results of our seems to be higher in children diagnosed with T1D under 8 group demonstrated a protective role of prenatal betametha- years of age. Because very young children usually have an sone exposure against T1D in NOD mice [17]. The chosen aggressive form of T1D [26], it is reasonable to assume that strategy to evaluate the effect of this drug in humans was a the severity of autoimmunity would be much more difficult questionnaire for parents of children with T1D, whereas to halt or that the genetic component is stronger in these the control group consisted of healthy siblings or twins of children [27]. We have to take into account that the control a subject with the disease. Betamethasone exposure did not group was composed of siblings of patients with T1D, in increase the susceptibility to T1D and was even associated which the risk to develop T1D is more than 10 times higher with a nonsignificant decreased risk of T1D. This association than in the general population [28]. 6 Journal of Diabetes Research The effects of glucocorticoids, administered during preg- We are well aware of the limitations of our study, includ- nancy, on childhood diabetes were previously reported in a ing the use of a questionnaire instead of clinical data registry. Danish cohort study [29]. This study showed that prenatal However, due to technical and ethical issues, it was impossi- exposure to corticosteroids tends to increase T1D incidence ble to link both the medical registries of the mothers and in the offspring. Nevertheless, there are substantial differ- their children in our study population. Another limitation ences between this study and the present work. Whereas was the recall bias in the questionnaire, particularly for the we started with a cohort of T1D children and siblings betamethasone exposure, which is our main focus. We have (n = 945), Greene et al. started with approximately 500,000 determined that 5% of mothers do not remember whether children born in Denmark in a specific period (between 1 they received betamethasone. To avoid it, we have defined January 1997 and 31 December 2004). So, our cohort is a specific target population, and the questionnaire has been smaller than the Danish one but includes 50% of T1D designed as short, precise, and accessible as possible. None- patients, in comparison to less than 1% of Danish children theless, this “missing” group gives similar perinatal data than with T1D. Importantly, in the Danish cohort the authors the unexposed betamethasone group, so it is unlikely that took into account not only prenatal betamethasone adminis- group participants received betamethasone at late preg- tration but also glucocorticoid exposure (topical, inhaled, nancy. In addition, participants were recruited from two and systemic corticosteroids) throughout the pregnancy. sources, but the incidence of T1D in the two geographical Only 4.5% of the prenatally glucocorticoid-exposed children areas where the questionnaire was performed is similar received one or more systemic glucocorticoids (betametha- (questionnaires obtained from Germans Trias i Pujol Hospi- sone and other synthetic glucocorticoids). In this sense, we tal belong to Catalonia region and from DiabetesCero foun- believe that our study was focused on the effect of a single dation belong to all Spanish regions) [37, 38]. The fact that course of prenatal systemic betamethasone exposure in the the control group is composed of siblings of patients with third trimester. The duration of the exposure before giving T1D is a particular feature of the study, aiming at minimiz- birth could prove crucial because betamethasone and its ing the variations due to environmental factors. derivatives have a half-life of approximately 3 days in the However, additional results from the present study fit human body [30]. We hypothesize that repeated doses of well with previous reports, thus validating the study group. this drug in a short span of time would increase exponen- First, a significantly higher number of preterm newborn tially the glucocorticoid concentration, thus inducing side males was observed as compared to their female counter- effects such as hyperglycemia and insulin resistance [31], parts. This finding has already been reported in previous among others [22, 32]. On the other hand, perinatal beta- studies [39, 40]. Second, as expected, betamethasone- methasone effects related to T1D could differ depending exposed newborns had lower birth weight and fewer gesta- on the gestational week of administration. For these reasons, tional weeks than unexposed newborns. we consider that our results point to a different direction. In fact, antenatal glucocorticoids have an important impact on 5. Conclusion immune system ontogeny [22] and modify the T cell recep- tor repertoire, as we demonstrated in NOD mice [17]. Our Our study suggest that the prenatal exposure to betametha- recent results also showed betamethasone effects on β-cell sone does not increase the susceptibility to T1D, and may growth, metabolism, and immunogenicity [19]. Additional even be associated with a trend towards decreased risk of effects of glucocorticoids are epigenetic modifications [33] developing this disease. However, these preliminary findings which could affect both the developing immune system require further prospective studies with clinical data regis- and the β-cells. These effects have positive consequences tries involving a larger sample size to draw definitive conclu- on self-tolerance in NOD mice, in this regard, betametha- sions regarding the effects of prenatal betamethasone sone might even be a factor that could help to tip the balance exposure on the risk of developing T1D in later life. towards β-cell tolerance, thus contributing to prevent auto- immunity in subjects at risk of developing T1D. Data Availability Interestingly, and independently of the effect of beta- methasone, the T1D group of this study shows a longer The data used to support the findings of this study are avail- pregnancy duration than the control group. Other studies able from the corresponding author upon request. do not report any effect [34] or point to the opposite: pre- term and early term delivery correlate to an increased risk Disclosure of T1D, whereas post-term delivery is associated with a reduction in T1D risk [7, 35, 36]. These controversial dif- MVP holds a patent that relate to immunotherapy for T1D ferences can be due to geographical and demographic fac- and is cofounder of Ahead Therapeutics S.L. which aims at tors, but also to the analysis performed in our study. Here, the clinical translation of immunotherapies for the treat- weeks of gestation of the control group and children with ment of autoimmune diseases. SRF is part-time employed T1D were compared, whereas other reported studies ana- at Ahead Therapeutics S.L. lyzed the incidence of T1D in the 4–5 different gestational periods. The reported difference is only of 3 days, and Conflicts of Interest despite it being statistically significant, it could be a coin- cidental finding. The other authors report no conflicts of interest in this work. Journal of Diabetes Research 7 Authors’ Contributions [5] K. Størdal, H. J. McArdle, H. Hayes et al., “Prenatal iron expo- sure and childhood type 1 diabetes,” Scientific Reports, vol. 8, DPB, MPD, and MV designed the questionnaire. MM, NR, no. 1, pp. 1–11, 2018. and JB distributed the questionnaire and collected data. [6] U. Samuelsson, S. Oikarinen, H. Hyöty, and J. Ludvigsson, DPB and LGM analyzed the data. DPB, SRF, and MVP “Low zinc in drinking water is associated with the risk of type wrote the manuscript. All the authors contributed to the dis- 1 diabetes in children,” Pediatric Diabetes, vol. 12, no. 3part1, cussion, revised the manuscript, and gave final approval of pp. 156–164, 2011. this version. [7] K. Huang, S. Si, R. Chen et al., “Preterm birth and birth weight and the risk of type 1 diabetes in Chinese children,” Frontiers in Endocrinology, vol. 12, 2021. Acknowledgments [8] R. R. Goldacre, “Associations between birthweight, gestational age at birth and subsequent type 1 diabetes in children under We are grateful to DiabetesCero Foundation for enabling the 12: a retrospective cohort study in England, 1998–2012,” Dia- distribution of the questionnaire among its members and to betologia, vol. 61, no. 3, pp. 616–625, 2018. the parents who took part in our online poll. We really [9] I. Waernbaum, G. Dahlquist, and T. Lind, “Perinatal risk appreciate Dr. Ricardo Gonzalo and Dr. Santiago Pérez- factors for type 1 diabetes revisited: a population-based reg- Hoyos, from Vall d’Hebron Research Institute, Barcelona, ister study,” Diabetologia, vol. 62, no. 7, pp. 1173–1184, for their help in statistical analysis. Special thanks to Prof. 2019. Eva Tolosa, from University Medical Center Hamburg- [10] J. Zeitlin, K. Szamotulska, N. Drewniak et al., “Preterm birth Eppendorf, Germany, for fruitful discussions. This work time trends in Europe: a study of 19 countries,” British Journal of Obstetrics and Gynaecology, vol. 120, no. 11, pp. 1356–1365, has been funded by DiabetesCero Foundation. 2013. [11] H. Blencowe, S. Cousens, M. Z. Oestergaard et al., “National, regional, and worldwide estimates of preterm birth rates in Supplementary Materials the year 2010 with time trends since 1990 for selected coun- tries: a systematic analysis and implications,” Lancet, Supplementary Figure 1: Prenatal administration of beta- vol. 379, no. 9832, pp. 2162–2172, 2012. methasone and type 1 diabetes according to age. (A) Histo- [12] G. C. Liggins and R. N. Howie, “A controlled trial of antepar- gram of the number of subjects diagnosed with type 1 tum glucoccorticoid treatment for prevention of the respira- diabetes (T1D) in different age groups. White color corre- tory distress syndrome in premature infants,” Pediatrics, sponds to subjects unexposed to prenatal betamethasone, vol. 50, no. 4, pp. 515–525, 1972. black color corresponds to subjects exposed to prenatal beta- [13] D. G. Sweet, V. Carnielli, G. Greisen et al., “European consen- methasone, whereas grey color represents subjects whose sus guidelines on the management of respiratory distress syn- status is missing (regarding the prenatal exposure to beta- drome -2019 update,” Neonatology, vol. 115, no. 4, pp. 432– methasone). (B) Histogram of the percentage of 450, 2019. betamethasone-exposed subjects in the T1D group depend- [14] X. Miracle, G. C. Di Renzo, A. Stark, A. Fanaroff, X. Carbonell- ing on the age at disease diagnosis. Black bar corresponds Estrany, and E. Saling, “Guideline for the use of antenatal cor- to early onset of T1D (from 0 to 7 years, n = 255), while ticosteroids for fetal maturation,” Journal of Perinatal Medi- white bar corresponds to late onset of T1D (from 8 to 17 cine, vol. 36, no. 3, pp. 191–196, 2008. years, n = 202). Chi-squared test was used for statistical anal- [15] K. M. Groom, “Antenatal corticosteroids after 34 weeks' gesta- ysis. (Supplementary Materials) tion: do we have the evidence?,” Seminars in Fetal and Neona- tal Medicine, vol. 24, no. 3, pp. 189–196, 2019. References [16] A. L. Shanks, J. L. Grasch, S. K. Quinney, and D. M. Haas, “Controversies in antenatal corticosteroids,” Seminars in Fetal [1] C. C. Patterson, V. Harjutsalo, J. Rosenbauer et al., “Trends & Neonatal Medicine, vol. 24, no. 3, pp. 182–188, 2019. and cyclical variation in the incidence of childhood type 1 dia- [17] A. Gieras, C. Gehbauer, D. Perna-Barrull et al., “Prenatal betes in 26 European centres in the 25 year period 1989–2013: administration of betamethasone causes changes in the T cell a multicentre prospective registration study,” Diabetologia, receptor repertoire influencing development of autoimmu- vol. 62, no. 3, pp. 408–417, 2019. nity,” Frontiers in Immunology, vol. 8, 2017. [2] M. Rewers, H. Hyöty, Å. Lernmark et al., “The Environmental [18] D. Perna-Barrull, S. Rodriguez-Fernandez, I. Pujol-Autonell Determinants of Diabetes in the Young (TEDDY) Study: 2018 et al., “Prenatal betamethasone interferes with immune system update,” Current Diabetes Reports, vol. 18, no. 12, p. 136, 2018. development and alters target cells in autoimmune diabetes,” [3] M. Mattila, S. Niinistö, H. M. Takkinen et al., “Maternal nitrate Scientific Reports, vol. 9, no. 1, 2019. and nitrite intakes during pregnancy and risk of islet autoim- [19] D. Perna-Barrull, A. Gieras, S. Rodriguez-Fernandez, munity and Type 1 diabetes: the DIPP cohort study,” The Jour- E. Tolosa, and M. Vives-Pi, “Immune system remodelling by nal of Nutrition, vol. 150, no. 11, pp. 2969–2976, 2020. prenatal betamethasone: effects on β-cells and type 1 diabetes,” [4] D. Kumar, N. S. Gemayel, D. Deapen et al., “North-American Frontiers in Endocrinology, vol. 11, p. 540, 2020. twins with IDDM: genetic, etiological, and clinical significance [20] A. E. Coutinho and K. E. Chapman, “The anti-inflammatory of disease concordance according to age, zygosity, and the and immunosuppressive effects of glucocorticoids, recent interval after diagnosis in first twin,” Diabetes, vol. 42, developments and mechanistic insights,” Molecular and Cellu- pp. 1351–1363, 1993. lar Endocrinology, vol. 335, no. 1, pp. 2–13, 2011. 8 Journal of Diabetes Research [21] A. Rafacho, H. Ortsäter, A. Nadal, and I. Quesada, “Glucocor- [36] C. Crump, J. Sundquist, and K. Sundquist, “Preterm birth and ticoid treatment and endocrine pancreas function: Implica- risk of type 1 and type 2 diabetes: a national cohort study,” tions for glucose homeostasis, insulin resistance and Diabetologia, vol. 63, no. 3, pp. 508–518, 2020. diabetes,” Journal of Endocrinology, vol. 223, p. R49, 2014. [37] C. C. Patterson, E. Gyürüs, J. Rosenbauer et al., “Trends in [22] M. E. Solano, M. C. Holmes, P. R. Mittelstadt, K. E. Chapman, childhood type 1 diabetes incidence in Europe during 1989– and E. Tolosa, “Antenatal endogenous and exogenous gluco- 2008: evidence of non-uniformity over time in rates of corticoids and their impact on immune ontogeny and long- increase,” Diabetologia, vol. 55, no. 8, pp. 2142–2147, 2012. term immunity,” Seminars in Immunopathology, vol. 38, [38] S. Conde Barreiro, M. Rodríguez Rigual, G. Bueno Lozano no. 6, pp. 739–763, 2016. et al., “Epidemiology of type 1 diabetes mellitus in children [23] M. García-Reymundo, X. Demestre, M. J. Calvo, G. Ginovart, in Spain,” Anales de Pediatría, vol. 81, no. 3, pp. 189.e1– A. Jiménez, and J. A. Hurtado, “Prematuro tardio en España: 189.e12, 2014. experiencia del Grupo SEN34-36,” Anales de Pediatría [39] J. A. McGregor, M. Leff, M. Orleans, and A. Baron, “Fetal gen- (English Edition), vol. 88, no. 5, pp. 246–252, 2018. der differences in preterm Birth: findings in a north American [24] C. Ramos-Navarro, M. Sánchez-Luna, S. Zeballos-Sarrato, and cohort,” American Journal of Perinatology, vol. 9, no. 1, pp. 43– I. Pescador-Chamorro, “Antenatal corticosteroids and the 48, 1992. influence of sex on morbidity and mortality of preterm [40] I. Ingemarsson, “Gender aspects of preterm birth,” BJOG : An infants,” The Journal of Maternal-Fetal & Neonatal Medicine, International Journal of Obstetrics and Gynaecology, vol. 110, pp. 1–8, 2020. pp. 34–38, 2003. [25] J. M. Terán, C. Varea, C. Bernis, B. Bogin, and A. González- González, “Nuevas curvas de peso al nacer por paridad y tipo de parto para la poblacion española,” Gaceta Sanitaria, vol. 31, no. 2, pp. 116–122, 2017. [26] P. Leete, A. Willcox, L. Krogvold et al., “Differential insulitic profiles determine the extent of β-cell destruction and the age at onset of type 1 diabetes,” Diabetes, vol. 65, no. 5, pp. 1362–1369, 2016. [27] J. R. J. Inshaw, A. J. Cutler, D. J. M. Crouch, L. S. Wicker, and J. A. Todd, “Genetic variants predisposing most strongly to type 1 diabetes diagnosed under age 7 years lie near candidate genes that function in the immune system and in pancreatic β- cells,” Diabetes Care, vol. 43, no. 1, pp. 169–177, 2020. [28] M. A. Atkinson, G. S. Eisenbarth, and A. W. Michels, “Type 1 diabetes,” Lancet, vol. 383, no. 9911, pp. 69–82, 2014. [29] N. H. Greene, L. H. Pedersen, S. Liu, and J. Olsen, “prenatal prescription corticosteroids and offspring diabetes: a national cohort study,” International Journal of Epidemiology, vol. 42, no. 1, pp. 186–193, 2013. [30] A. H. Jobe, M. A. Milad, T. Peppard, and W. J. Jusko, “Pharma- cokinetics and pharmacodynamics of intramuscular and oral betamethasone and dexamethasone in reproductive age women in India,” Clinical and Translational Science, vol. 13, no. 2, pp. 391–399, 2020. [31] J. M. Sheen, C. S. Hsieh, Y. L. Tain et al., “Programming effects of prenatal glucocorticoid exposure with a postnatal high-fat diet in diabetes mellitus,” International Journal of Molecular Sciences, vol. 17, no. 4, p. 533, 2016. [32] E. Tomaszewska and P. Dobrowolski, “Prenatal glucocorti- coids can programme postnatal development,” Journal of Pre-Clinical and Clinical Research, vol. 5, no. 1, p. 21, 2011. [33] A. S. Zannas and G. P. Chrousos, “Epigenetic programming by stress and glucocorticoids along the human lifespan,” Molecu- lar Psychiatry, vol. 22, no. 5, pp. 640–646, 2017. [34] A. Adar, S. Shalitin, O. Eyal et al., “Prevalence of early and late prematurity is similar among pediatric type 1 diabetes patients and the general population,” Diabetes/Metabolism Research and Reviews, vol. 34, no. 5, article e2996, 2018. [35] S. Li, M. Zhang, H. Tian, Z. Liu, X. Yin, and B. Xi, “Preterm birth and risk of type 1 and type 2 diabetes: systematic review and meta-analysis,” Obesity Reviews, vol. 15, no. 10, pp. 804– 811, 2014.