Antibiotic Exposures and the Likelihood of Developing Pediatric Autoimmune Diseases: a Register-based Matched Case-control Study

Incidences of pediatric autoimmune diseases (ADs) have been increasing without clearly identied environmental risk factors. Using national registers (special reimbursement and drug purchase register), this matched case-control study aims to discover whether antibiotic exposures are related to development of ADs in general. From a cohort of 11,407 children (born 2000-2005, followed-up until December 2018), 242 children with ADs (type 1 diabetes (DM), autoimmune thyroiditis (AIT), juvenile idiopathic arthritis (JIA), and inammatory bowel diseases (IBD)) generated the case group. Each case was matched with one to four child(ren) with similar age, sex, residential area, gestational age, and delivery mode, generating 708 matched controls. Number of antibiotic purchases throughout childhood was unrelated to the development of the studied ADs as one group, nor with DM, AIT, or IBD individually. While antibiotic exposures in infancy did not increase the likelihood of ADs, exposures to macrolides within two years before diagnosis showed minor association (OR 1.24, 95% CI 1.01-1.51). Exposures to cephalosporins, macrolides, and amoxicillin-clavulanic acid throughout childhood may increase the likelihood of JIA (OR 1.25, 95% CI 1.03-1.52; OR 1.21, 95% CI 1.06-1.38; OR 1.19, 95% CI 1.02-1.39, respectively). Regardless of frequent use in childhood (40% of all antibiotics), penicillins were not associated with the onset of any ADs. Conclusion: Occasional use of antibiotics are relatively safe regarding overall development


Introduction
Autoimmune diseases (ADs) are disorders in which the immune system attacks normal tissues. Some ADs such as type 1 diabetes (DM), autoimmune thyroiditis (AIT), juvenile idiopathic arthritis (JIA), and in ammatory bowel diseases (IBD) may have overlapping genetic pathways [1,2] and similarities in their pathogenesis [3][4][5][6][7]. For unknown reasons, the incidences of these four chronic ADs in pediatric population in Finland have been particularly high, presuming a presence of mutual environmental risk factors [8]. Since Finnish children have used more antibiotics in the rst and second year of life compared with Estonian or Russian Karelian children, and since the incidence of ADs in Finland are higher than in eastern Europe, this study focuses on the potential relationship between antibiotics and ADs [9][10][11][12][13].
Previous studies have connected antibiotic exposures (especially in early childhood) with the onset of JIA and Crohn's disease, while ndings regarding DM were controversial [14][15][16][17][18][19]. AIT has been related to tetracycline use in adolescence, but the mechanism remained unspeci ed [20]. Most of previous studies addressed each disease individually in different settings, making it challenging to estimate whether antibiotic exposures could be associated with the development of pediatric autoimmunity in general, yet manifesting as different diseases. The aim of this register-based matched case-control study is to investigate whether ADs (represented by DM, AIT, JIA, and IBD) might be associated with: (1) number of antibiotic exposures during different periods in childhood; (2) exposures to particular types of. antibiotics.

Data sources for the study population
The study population was derived from Finnish Health in Teens (Fin-HIT) cohort -a nationwide prospective school-based cohort to address health behaviors of Finnish children and adolescents, comprising over 11,000 children (born 2000-2005) without speci c exclusion criteria. More details on recruitment process and characteristics of the cohort have been described elsewhere [21]. The cohort represents children from densely populated areas across Finland, with relatively high maternal socioeconomic status [8]. Using a unique personal identity code of every Finnish resident, children in the Fin-HIT cohort were linked to three well established national registers: (1) the Special Reimbursement Register (SRR) -containing records on patients with chronic diseases requiring medication (including entry dates and physician veri ed diagnosis), who are entitled to drug refunds (i.e. special reimbursements) regardless of their socioeconomic status; (2) the Drug Purchase Register (DPR)containing data on all purchased drugs by prescriptions in Finland (including dispensation dates and pharmaceutical information); and (3) the Medical Birth Register (MBR) -containing information on gestational age, delivery modes, and postnatal antibiotic treatment before discharge. These registers are maintained by the Finnish Social Insurance Institution (SRR and DPR) and Finnish Institute for Health and Welfare (MBR). Identifying children eligible for the matched case-control study The outcome of this study was the diagnosis of at least one AD at the end of the follow-up (31 December 2018). DM, JIA, and IBD (including Crohn's disease, ulcerative colitis, and IBD unclassi ed) diagnoses were obtained from the SRR using ICD-10 codes (International Classi cation of Diseases (ICD), 10 th revision): E10 for DM; M08 for JIA; K50 (CD) and K51 (UC/IBDU) for IBD. AIT diagnoses were obtained from the DPR using ATC (Anatomical Therapeutic Chemical) code H03AA01 for thyroxin -the prescription-only drug used for AIT. The excellent coverage of these registers has been described previously [22,23]. DPR was chosen for identifying AIT because thyroxin is inexpensive and not everyone using this medication is registered in the SRR.
Of over 11,000 children in the Fin-HIT cohort, 242 developed a primary AD after rst year of life and generated the case group. Depending on the availability of potential controls, each child in the case group was matched with one to four children from the same cohort with similar age (0-4 days of age difference), sex, and residential area. Since preterm birth has been identi ed as a potential risk factor for ADs, gestational age (preterm/term), and delivery mode (cesarean section/vaginal delivery) were also considered in the matching [8]. Due to the limited number of potential controls, most children born preterm and/or with cesarean section had only one matching control.

Number and types of antibiotic purchases
Data on perinatal antibiotic treatment in birth hospital were collected from the MBR and outpatient antibiotic purchases from the DPR with ATC codes starting with J01. The data was collected from birth until two months prior to index date (date of diagnosis for children with ADs/ compatible date for their matched controls). The two-month period was chosen to reduce the possibility of including antibiotic purchases during exacerbation phase of ADs.
Antibiotic exposures were analyzed in several observation periods based on age distribution of antibiotic purchases (Figure 1 a): (1) throughout childhood -from birth to the date of diagnosis / the index date; (2) infancy -during the rst year of life; (3) toddler phase -from the age 1 up to 3 rd birthday/index date (in those who were diagnosed <3 years old); (4) preschool to adolescence -from the age of 3 years to the index date; and (5) within two years before the index date. Due to observed nonlinear associations between antibiotic exposures and ADs, the numbers of antibiotic purchases in each observation periods were rst categorized into quartiles and then modi ed accordingly: no purchases or 0 to Q1 (when the number of children without antibiotic purchases in that speci c observation period is low), Q2, and combined Q3-Q4. The odds of developing an AD in relation to the number of antibiotic exposures in each observation period (while exposures in earlier periods were disregarded) were analyzed individually. Different types of antibiotics were categorized based on level 4 ATC codes (Supplementary Table 1) into 5 groups: (A) penicillins; (B) macrolides; (C) cephalosporines; (D) amoxicillin-clavulanic acid; and (E) sulphonamides and trimethoprim. Clindamycin, tetracyclines, uoroquinolones and other antibiotics such as nitrofurantoin and metronidazole were considered in the analysis regarding the total number of antibiotic purchases but dismissed from the subgroup analysis due to low frequency of usage.

Statistical methods
The background data of cases and controls are presented as mean and standard deviation (SD) or number/proportion (%). The likelihood of developing an AD was obtained by comparing each case and matched control(s) using conditional logistic regression. Results were presented with Odds Ratio (OR) and 95% con dence interval (CI). The software used was IBM SPSS Statistics 26.0 and a 5% statistical signi cance level was adopted.

Results
The background characteristics of the 242 children who developed ADs (cases) and 708 children who did not (matched controls) are presented in Table 1. Throughout childhood, only 14 (5.8%) children in the case group and 34 (4.8%) in the matched control group had no record of antibiotic purchases (p=0.596).
Age at and the type of the rst antibiotic exposure did not differ between cases and matched controls (Supplementary Table 2). Penicillins were the most common purchased antibiotics (40% of all antibiotics, of which over 80% were amoxicillin), followed by macrolides (20% of all antibiotics, of which over 80% were azithromycin) (Supplementary Table 1). a Cases=children with autoimmune diseases (represented with DM (type 1 diabetes mellitus), AIT (autoimmune thyroiditis), JIA (juvenile idiopathic arthritis), and IBD (in ammatory bowel diseases)). Nine chidren have two diagnoses. b Each child in the case group were matched with one to four children with similar age, sex, residential area, gestational age (preterm/term), delivery mode (cesarean section/vaginal). Due to limited potential controls, most children born preterm/with cesarean section have only one control instead of four. Table 2 Association between numbers of antibiotic purchases in different periods and the development of an autoimmune disease (DM, AIT, JIA, or IBD) a . Total number of antibiotic purchases throughout childhood (from birth to the date of diagnosis/index date) was not associated with the development of the studied ADs as one group, nor with DM, AIT, and IBD individually. Of all purchased antibiotics, 44.3% (n=3,825) were bought before the age of 3 years.
Concerning antibiotic purchases in each age, the only signi cant difference between cases and matched controls was seen during the second year of life: children who later developed ADs purchased more antibiotics than those who did not develop ADs (OR 1.11, 95%CI 1.03-1.20) (Figure 1). As for antibiotic types, penicillins were not associated with the development of any ADs, while purchases of macrolides within two years before the index date were associated with the onset of ADs in general. Moreover, purchases of broader-spectrum antibiotics (macrolides, cephalosporins, or amoxicillin-clavulanic acid) throughout childhood (but particularly during infancy and toddler phase) increased the likelihood of developing JIA.

Number of antibiotic purchases
The relationship between the total number of antibiotic purchases before the index date and the proportion of ADs was non-linear and required categorized analyses ( Figure 2). The highest odds for developing ADs in general (not associated with any speci c diagnosis) were observed in children with frequent antibiotic purchases (≥ 3 courses) when compared with those without antibiotic purchases (OR

Types of antibiotic purchases
When purchases of the ve most common antibiotic types in the study population (penicillins, macrolides, cephalosporins, amoxicillin-clavulanic acid, and sulphonamides and trimethoprim) throughout childhood were analyzed individually, none of them was associated with the onset of the four ADs as one group (Supplementary Table 4

Discussion
Our study is the rst to investigate the association between number and types of antibiotic exposures in different stages of childhood and the onset of four common pediatric ADs (DM, AIT, JIA, or IBD) in a mutual setting. We showed that the relationship between the number of antibiotic exposures and the onset of ADs seems to be non-linear. In addition, the likelihood of developing ADs after antibiotic exposures in infancy remained low, especially regarding DM, AIT, and IBD. Despite of being the most common antibiotic used in childhood, penicillins (predominantly amoxicillin) would be relatively safe to use at any age in relation with the development of ADs, while exposures to macrolides within the two years prior to diagnosis may present minor association. Regarding speci c diagnosis, the number of antibiotic exposures throughout childhood was mainly associated with JIA. Even then, exposures to penicillin group antibiotics did not increase the likelihood of JIA, while exposures to cephalosporin, macrolide, and amoxicillin-clavulanic (especially before the age of three) might moderately do so.
In our study, antibiotic exposures in infancy did not increase the likelihood of developing any ADs. Furthermore, we showed that frequent antibiotic exposures after rst year of life may be related with ADs when compared with occasional exposures, but not when compared with those having no records of antibiotic purchases in each observation period. However, this nding seems to be predominantly related to JIA and DM, but not with AIT and IBD. How do these ndings align with previous studies? A Swedish register-based study presented a connection between prescribed antibiotics during infancy and the onset of DM [18], while studies from which countries have not reported signi cant association between early childhood antibiotic exposures and DM [19,24]. Studies from United Kingdom and Finland have associated both lifetime and early life antibiotic exposures and the onset of JIA [14,15]. As for IBD, early antibiotic exposures have been related with Crohn's disease but not with ulcerative colitis [17,25]. Most of these studies assumed a linear association between antibiotic exposures and the onset of a particular AD, thus assuming that the lowest risk for ADs is retained by those without previous antibiotic exposures.
Our study challenges this presumption. Since infancy is the most susceptible period for common infectious diseases and is the period of most frequent antibiotic use [26], it is reassuring that antibiotics used in early childhood hardly increased the likelihood of developing pediatric ADs. Beyond rst year of life, occasional antibiotic courses (approximately 1-2 courses per year) can generally be seen as a lowrisk approach in relation to the development of ADs.
We showed that exposures to macrolides within the two years before obtaining diagnosis may be related with the overall development of ADs. In addition, exposures to antibiotics with broader spectrum (cephalosporine, macrolide, and amoxicillin-clavulanic acid), particularly before the age of three years, were moderately associated with JIA. On the contrary, penicillins (as the most common antibiotic type used to treat infections throughout childhood) were not related to the development of any ADs at any age. Hence, the relationship between antibiotics and ADs could not solely be explained by number of infections, as often speculated.
Since antibiotics have been shown to have an in uence on gut microbial homeostasis [31], antibiotic exposures could as well be related to ADs through facilitating gut dysbiosis. Variations in magnitude and speci city of gut microbiota modi cation by different antibiotics have been reported, hence requiring variable recovery time after exposures [32][33][34]. For example, macrolides have both a broad spectrum and long-term in uence on gut microbiota that may persist even for several years [35,36]. In our study, azithromycin was the most often used antibiotic among macrolides. Azithromycin has a broad bacteriostatic spectrum, a marked tissue penetration, a high stability and a low clearance rate due to its long half-life, which enable it to reach a higher cellular concentration compared to penicillin [37]. These characteristics may explain the long-term in uences of azithromycin (as a macrolide) on gut microbiota compared with penicillin. Furthermore, previous Fin-HIT study showed that azithromycin presented the strongest inverse association with salivary microbiota diversity [38]. Since dysbiosis of gut and salivary microbiota have also been associated with ADs [39], we suggest that macrolides might catalyze longerterm dysbiosis, explaining their association with the development of ADs, particularly up to two years prior to diagnosis, while penicillin was not associated with any type of ADs at any ages. Further studies to examine the potential link between the use of broad-spectrum antibiotics, the time of their in uence on gut microbiota, gut dysbiosis, and the onset of ADs is warranted.
The strength of our study lies in the comprehensive longitudinal data from national registers. For example, we were able to trace purchased antibiotics as outpatients rather than just prescribed. In addition, we studied several ADs in a mutual setting, using rather homogenous Fin-HIT cohort with small variations in socioeconomic status as the source of study population [8]. The control population was matched with details based on age (with only up to 4 days of age difference), sex, residential area, gestational age, and delivery method to limit the number of potential confounding factors. Therefore, our study setting made it possible to examine the association between childhood antibiotic exposures and onsets of the four pediatric ADs together, and to reliably compare one disease to another in this subject.
As for limitations, we lack the information on the children's genetic susceptibility to infections or to ADs.
We also did not know why the antibiotics were purchased -for treating infections (and if so, for what kind of infection) or for prophylactic purposes -and on whether secondary antibiotic courses for the same infection were needed. In addition, we had no access on the length of purchased antibiotic courses nor on the antibiotics given during inpatient care. Yet, antibiotic treatments during hospitalization are often continued orally after discharge, and our data cover these post-discharge antibiotic purchases.

Conclusion
Use of antibiotics in the rst year of life and occasional use throughout childhood can be considered relatively safe in relation with the development of pediatric autoimmunity. Antibiotics in the penicillin group are unlikely to be associated with the development of any ADs. In contrast, broad-spectrum antibiotics (including macrolides) should be used considerately as they may associate with an increased likelihood of ADs, especially JIA. preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Con ict of interest
The authors have no con icts of interest to declare, nor have any a liations with nancial or nonnancial interest in subjects relevant to the content of this manuscript.

Ethics approval
The Fin-HIT study protocol was approved by the Ethics Committee of the Hospital District of Helsinki and Uusimaa (decision number 169/13/03/00/10).

Consent to participate
The children and a guardian provided a written informed consent for the register linkage.

Consent for publication
All authors approved the nal manuscript as submitted, and agreed to be accountable for all aspects of the work.

Availability of data and material
Data available upon request.
Code availability N/A Contributions LR, EE, HV, KLK conceptualized the study. HV participated in data collection. LR carried out the analyses and interpreted the results with the help of RS and SK, and wrote the initial manuscript, and made the gures and tables. LR and SK searched the literatures. HV and KLK supervised the study. All authors critically reviewed and revised the manuscript for important intellectual content.

Figure 1
Purchased antibiotic courses before index date a in the matched case-control b study. Number of tetracyclin, clindamycin, uoroquinolone, and other antibiotics courses were relatively low.
a Index date= Date of diagnosis for children who developed autoimmune diseases (type 1 diabetes, autoimmune thyroiditis, juvenile idiopathic arthritis, and in ammatory bowel diseases) and compatible date for their matching controls b Cases= 242 children who developed autoimmune diseases by the end of follow-up (31 December 2018).
Each child in the case group was matched with one to four children with similar age, sex, residential area, gestational age (preterm/term), and delivery mode (cesarean section/vaginal), comprising control group of 708 children.

Figure 2
Proportion of autoimmune diseases a in groups of children with different number of antibiotic purchases in different periods of childhood.
a Autoimmune diseases in this study were type 1 diabetes, autoimmune thyroiditis, juvenile idiopathic arthritis, and in ammatory bowel diseases. Index date = Date of diagnosis for children who developed autoimmune diseases and compatible date for their age, sex, residential areas, gestational age (preterm/term), and delivery method (Cesarean section/vaginal delivery) matched controls. Nine children have two diagnoses. OR=Odds ratio, CI=Con dence Interval. Analyses was performed using conditional logistic regression. Index date= age of diagnosis for children who developed autoimmune diseases and compatible date for their matching controls.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. SupplementsAbandautoimmunediseasesRaisanen.docx