Introduction
PCOS is the most common endocrine disorder in women of reproductive age [ 1 ]. According to the Rotterdam criteria, PCOS is diagnosed by hyperandrogenism (clinical/biochemical), chronic anovulation and/or polycystic ovaries/elevated anti‐müllerian hormone, when other causes have been excluded [ 1 ]. Insulin resistance and low grade inflammation in PCOS are associated with higher risk of type 2 diabetes and infertility, where the metabolic risk is further aggravated by central obesity [ 2 ]. Chronic low grade inflammation and unbalanced estrogen/progesterone secretion in PCOS could result in higher autoimmune activity in PCOS [ 3 , 4 ]. Previous studies reported higher prevalence of thyroid, pancreatic islet cell, anti‐nuclear (ANA), and ovarian autoantibodies in women with PCOS [ 4 ].
In women with type 1 diabetes (T1D), the prevalence of a PCOS diagnosis was 25% compared to an expected prevalence of 10%–15% in women of reproductive age [ 5 , 6 ], which supported high coexistence of T1D and PCOS. HbA1c and/or glucose tolerance test are recommended for metabolic risk screening in women with PCOS, especially due to the increased risk for type 2 diabetes, whereas measurement of pancreatic islet cell antibodies is not routinely performed [ 4 ]. Iatrogenic hyperinsulinemia in women with T1D could lead to stimulation of ovarian hyperandrogenemia, anovulation and development of manifest PCOS [ 5 , 6 ]. The order of PCOS and T1D diagnosis can be used to clarify if the primary mechanism for higher risk of PCOS in T1D is autoimmunity or hyperinsulinemia. If PCOS is diagnosed before T1D, a shared autoimmune pathogenic mechanism is likely, whereas if T1D is diagnosed before PCOS, insulin treatment could be the trigger of PCOS development. We are not aware of previous large scale prospective studies regarding the risk of autoimmunity before and after the diagnosis of PCOS.
Higher risk of thyroid disease in women with PCOS is previously reported [ 7 , 8 , 9 ]. Thyroid dysfunction will to some degree resemble a clinical PCOS phenotype with weight gain in hypothyroidism, and irregular menses in hyper and hypothyroidism and screening for thyroid disorders is part of the recommended work‐up program in women with tentative PCOS suspicion [ 1 ]. Stress and depression are associated with higher risk for autoimmune diseases such as T1D, rheumatoid arthritis, and systemic lupus erythematosus [ 10 , 11 ]. In women with PCOS, factors such as foreign ethnicity, infertility, obesity and comorbidity could be associated with higher risk of autoimmune disease [ 12 ]. Routine screening for other autoimmune diseases than thyroid dysfunction is not widely recommended as part of the evaluation program in PCOS [ 1 ]. A recent publication reported multi comorbidity with higher risk of several autoimmune diseases in 244 women with PCOS at age 50 years compared to women without PCOS from the Northern Finland Birth Cohort [ 13 ]. These findings could support that higher awareness for autoimmune disease is warranted in PCOS.
Our objective was to use national registers to investigate the incidence of autoimmune disease before and after PCOS diagnosis in a Danish study population compared to controls and to investigate possible predictors for autoimmunity in PCOS.
Material and Methods
This study was designed as a register‐based cohort study. In Denmark, citizens receive a social security number at birth or upon immigration, allowing for the linkage of individual‐level data from various national health and social care registers [ 14 ]. The National Patient Register contains records of all somatic hospital contacts since 1977, psychiatric hospital contacts since 1995, and outpatient contacts since 1995 [ 14 ]. The Civil Registration System, established in 1968, provided data on sex, age, date of death, and date of emigration [ 14 ]. The Danish Income Statistics Register includes income data for all economically active individuals in Denmark [ 15 ]. Additionally, the National Prescription Registry holds a complete record of prescriptions filled at Danish pharmacies since 1995, including the Anatomic Therapeutic Chemical (ATC) code, the date the drug was dispensed, and the number of drug packets, regardless of whether the prescriptions are issued by general practitioners or specialists [ 14 ].
Study Population
The PCOS Cohort Included
Women aged ≥ 12 years [ 16 , 17 ] diagnosed with PCOS or hirsutism [ 18 , 19 ] according to the International Classification of Diseases (ICD)−10 (E282 and L680) in the Danish National Patient Register from 1995 to 2018. The index date was defined as the first time of diagnosis E282 or L680.
The Control Population
Included five randomly drawn women from the civil population register for each woman in the PCOS cohort. Controls were matched according to age and were assigned the index date (date of first PCOS diagnosis) of their matched PCOS case and had to be alive and living in Denmark and not diagnosed with PCOS on the index date.
Exclusion Criteria
The PCOS diagnosis requires exclusion of diagnoses that may cause symptoms similar to PCOS [ 20 ]. We excluded women with ICD‐10 diagnoses E220 (acromegaly), E24 (Cushing′s syndrome), E25 (adrenogenital syndrome), and Q96 (Turner syndrome) from the PCOS and control cohorts.
We have previously published data from the study cohort with follow up until 2012 [ 21 , 22 , 23 , 24 ] and until 2018 [ 2 , 25 ].
Definition of Study Parameters
Outcome
The primary study outcome was diagnosis of autoimmune diseases according to ICD10 diagnosis codes [ 26 , 27 ] and/or medicine prescriptions for T1D or thyroid disorder: Diagnosis codes based on ICD‐10, along with the dates of contacts, were extracted from the National Patient Register. Medicine prescriptions including dates of prescription were extracted from National Prescription Registry.
Autoimmune diseases included the following outcomes : Blood: D510 pernicious anemia, D590‐1 autoimmune hemolytic anemia, D690 Henoch Schönlein purpura, D693 idiopathic thrombocytopenic purpura, D86 sarcoidosis. Endocrine: E271 primary adrenal insufficiency. CNS : G35 multiple sclerosis, G61.0 Guillain‐Barre syndrome, G700 myastenia gravis. Eye : H20 iridocyclitis. Circulatory: I00, I01 rheumatic fever, I730 Raynaud′d phenomen. Digestive: K50 Crohn′s disease, K51 ulcerative colitis, K73 autoimmune hepatitis, K743 primary biliary cirrhosis, K900 celiac disease. Skin : L10 pemphigus, L12 pemphigoid, L13.0 dermatitis herpetiformis, L40 psoriasis vulgaris, L52 erythema nodosum, L63 alopecia areata, L80 vilitigo, L93 lupus erythematosus, L940‐3 lichen sclerosus. Musculoskeletal: M023 Reiter′s disease M05‐M06 seropositive rheumatoid arthritis, M08 juvenile arthritis, M300 polyarteritis nodosa, M303 Kawasaki syndrome, M310 Goodpasture′s syndrome, M313 Wegeners granulomatosis, M315‐6, M353 polymyalgia rheumatica, M32 systemic lupus erythematosus, M33 dermatopolymyositis, M34 systemic sclerosis, M32 systemic lupus erytematosis, M350 Sjogrens syndrome, M352 Behcet′s disease, M45 Ankylosing spondylitis, M720 Dupuytren′s disease.
T1D was defined as ICD10 codes E10 type 1 diabetes mellitus and/or prescription of A10A: Insulin.
Thyroid disorder was defined as ICD10 codes E050 Thyrotoxicosis, E063 Autoimmune thyroiditis and/or prescription of H03: Thyroid.
Covariates
Age was calculated at the index date.
Ethnicity was extracted from the Civil Registration System and divided into Danish and Non‐Danish (including immigrants and descendants).
Socioeconomic status was based on family income in the index year, excluding tax and interests from the Income Statistics Register [ 15 ] in the calendar year of the index date (or the first calendar year before in case of missing income information). We assessed family income according to tertiles (high, middle, low) in age groups.
Medical comorbidity : The Charlson comorbidity index is based on 19 conditions [ 28 , 29 ] and presence of comorbidity was defined as Charlson comorbidity index ≥ 1. We applied hospital diagnoses according to the National Patient Register within 5 years before the index date [ 29 , 30 ].
Psychiatric disorder was defined by the ICD‐10 codes F32‐F33 (depression), F40‐F41 (anxiety), X60‐X84 (intentional self‐harm), F10‐F19 (psychoactive substance use), F20‐F29 (psychotic disorders), F30‐F31, and F34‐F39 (bipolar and mood disorders) from hospital contacts and/or u se of psychopharmacological agents defined as more than one prescription and redemption of antidepressants (N06A), anxiolytics (N05B), or antipsychotics (N05A).
Infertility was defined according to ICD10 diagnosis codes: N97, Z35 and fertility treatment: G03GA (gonadotrophin) G03GB (ovulation stimulation).
Number of births were extracted from the Danish Medical Birth Register.
Hormonal anticonception. Oral contraceptives included ATC codes G03AA and G03AB (combined progesterone and estrogen OCP) and G03HB01 (OCP containing cyproterone acetate). Gestagens included oral contraceptives (G03AC) and intrauterine device (G02BA03).
Statistical Analyses
Baseline characteristics of the study cohort were presented as frequencies for categorical variables and medians with interquartile ranges for continuous variables. We used the Wilcoxon Rank Sum test for continuous variables and the chi‐squared test for categorical variables.
Risk time was calculated from the latest of either January 1 1995, date of birth, or date of immigration. It ended at the earliest of either December 31 2018, date of study outcome, emigration, or death.
To estimate incidence rates of autoimmune diseases and compare PCOS cases with controls, we applied generalized estimating equations (GEE). The GEE models used a Poisson distribution with a log link function, an independent within‐group correlation structure, and risk time defined as the amount of exposure.
Cox regression models were used to calculate hazard ratios for autoimmune diseases, incorporating covariates as exposure variables among women with PCOS. Additionally, we conducted separate Cox regressions to examine the impact of PCOS on autoimmune disease while adjusting for each covariate individually. Ethnicity, educational level, and income were treated as fixed covariates, while morbidity, psychiatric conditions, infertility, hormonal contraception use, and births were treated as time‐varying exposures. Risk time was split according to time‐varying exposures, ensuring that individuals contributed person‐time separately as exposed and unexposed. Once an individual became exposed, they remained in the exposed category for the rest of the study period and could not revert to an unexposed state.
We conducted a sensitivity analysis where the study outcome T1D was defined by ICD10 diagnosis code E10 alone as insulin treatment could be used in women with gestational diabetes.
Hirsutism is caused by PCOS in 80%−90% of affected women, but hirsutism may also be idiopathic or a symptom of rare endocrine disorders [ 18 , 19 ]. Therefore, we conducted a sensitivity analysis where women in the PCOS cohort with a diagnosis code of L680 alone were omitted from the analyses.
Data management and data analyses were conducted using Stata MP 18.0.
Ethics
No approval is necessary from the local Ethics committee or Institutional Review Board on register studies according to Danish law [ 14 ]. The study complied with the Helsinki declaration. The study was approved by the Data Protection Agency and by Statistics Denmark, project no 704175.
Results
Figure 1 is a flow chart of included women. The study cohort included 30,340 women with PCOS and 151,520 controls.
Study Characteristics of the Cohort Until End of Follow Up
( Table 1 ). Women with PCOS and controls had a median age of 29 years at study inclusion. At study inclusion, women with PCOS versus controls had higher prevalence of non‐Danish ethnicity (21.2 vs. 14.4%), low family income (35.2% vs. 33.0%), comorbidity (4.7% vs. 3.1%), psychiatric diagnosis/treatment (23.3 vs. 16.4%), infertility (29.8% vs. 12.1%), and lower prevalence of use of hormonal anticonception (65.5% vs. 67.8%), all p < 0.001.
Table 1
Characteristics in women with PCOS and controls until end of follow‐up ( N = 136,657).
Characteristics of the study cohort. Age, ethnicity, family income, comorbidity, psychiatric diagnosis, infertility, and use of hormonal anticonception at baseline. Number of births until end of follow up. ¤First available BMI in women with pregnancy. Comorbidity was defined as a Charlson index ≥ 1. p ‐value Between PCOS and controls. Chi‐square test (for categorical variables) and non‐parametric test on the equality of medians (for continuous variables).
| PCOS | Controls | |||
| ( N = 30,340) | ( N = 151,520) | P ‐value | ||
| Age (y) at baseline, median (Q1–Q3) | 29 (23−36) | 29 (23−36) | 0.92 | |
| Ethnicity | Danish | 23,898 (78.8%) | 129,652 (85.6%) | < 0.001 |
| Non‐Danish | 6442 (21.2%) | 21,868 (14.4%) | ||
| Family income at index year | Low | 10,610 (35.2%) | 49,593 (33.0%) | < 0.001 |
| Middle | 9637 (32.0%) | 50,465 (33.6%) | ||
| High | 9910 (32.9%) | 50,106 (33.4%) | ||
| Comorbidity at index | 1414 (4.7%) | 4637 (3.1%) | < 0.001 | |
| Psychiatric diagnosis and/or treatment | ICD10: F10‐F41, X60‐X84 ATC: N06AA, N06AB, N06AF, N06AG, N06AX | 7072 (23.3%) | 24,850 (16.4%) | < 0.001 |
| Infertility | ICD10: N97, Z35 ATC: G03GA, G03GB | 9041 (29.8%) | 18,314 (12.1%) | < 0.001 |
| Hormonal anticonception | G03AA, G03AB G03HB01, G03AC G02BA03 | 19,869 (65.5%) | 102,756 (67.8%) | < 0.001 |
| Number of births | 0 | 14,532 (47.9%) | 73,182 (48.3%) | < 0.001 |
| 1 | 5275 (17.4%) | 20,758 (13.7%) | ||
| ≥ 2 | 10,533 (34.7%) | 57,580 (38.0%) | ||
| BMI¤ (kg/m 2 )¤ | 25.1 (21.9−30.1) | 22.8 (20.8−25.9) | < 0.001 |
Until end of follow up, women with PCOS had different distribution of number of births. Women with PCOS had comparable presence of nulliparity compared to controls (47.9% vs. 48.3%), but more women with PCOS only had one birth (17.4% vs. 13.7%) and more seldom more than two births (34.7% vs. 38%) compared to controls, p < 0.001. The median BMI was 25.1 versus 22.8 kg/m 2 in women with PCOS compared to controls, p < 0.001.
Autoimmune Disease before PCOS Diagnosis
( Table 2 ): The mean risk time until baseline was 11.3 years in the study cohort. Before PCOS diagnosis, the incidence rate ratio (IRR) of autoimmune disease was 1.52 (1.44; 1.59) in women with PCOS compared to controls. Before baseline, women with PCOS had higher IRR of T1D 2.31 (2.06; 2.59), thyroid disorder 1.72 (1.61; 1.84), skin 1.27 (1.06; 1.52), and musculoskeletal 1.26 (1.08; 1.46) autoimmune diseases.
Table 2
Risk of autoimmunity in women with PCOS and controls before baseline.
Presented as incidence rate (IR) per 1000 PY (person years). Type 1 diabetes: ICD10, E10 Type 1 diabetes and/or prescription of ATC code A10A: Insulin. Thyroid disorder: ICD10: E05.0 Thyrotoxicosis, E06.3 Autoimmune thyroiditis and/or prescription of ATC code H03: Thyroid. Other autoimmune diseases: Diagnosis of any autoimmune disease according to the ICD10 diagnosis codes: Blood: D51.0 pernicious anemia, D59.0‐1 autoimmune hemolytic anemia, D69.0 Henoch Schönlein purpura, D69.3 idiopathic thrombocytopenic purpura, D86 sarcoidosis. Endocrine: E27.1 primary adrenal insufficiency CNS: G35 multiple sclerosis, G61.0 Guillain‐Barre syndrome, G70.0 myastenia gravis Eye: H20 iridycyclitis Circulatory: I00, I01 rheumatic fever, I730 Raynaud′d phenomen. Digestive: K50 Crohn′s disease, K51 ulcerative colitis, K73 autoimmune hepatitis, K74.3 primary biliary cirrhosis, K90.0 celiac disease. Skin: L10 pemphigus, L12 pemphigoid, L13.0 dermatitis herpetiformis, L40 psoriasis vulgaris, L52 erythema nodosum, L63 alopecia areata, L80 vilitigo, L93 lupus erythematosus, L94.0‐3 lichen sclerosus. Musculoskeletal: M02.3 Reiter′s disease M05‐M06 seropositive rheumatoid arthritis, M08 juvenile arthritis, M30.0 polyarteritis nodosa, M30.3 Kawasaki syndrome, M31.0 Goodpasture′s syndrome, M31.3 Wegeners granulomatosis, M31.5‐6, M35.3 polymyalgia rheumatica, M32 systemic lupus erythematosus, M33 dermatopolymyositis, M34 systemic sclerosis, M32 systemic lupus erytematosis, M35.0 Sjogrens syndrome, M35.2 Behcet′s disease, M45 Ankylosing spondylitis, M72.0 Dupuytren′s disease.
| PCOS ( N = 30,340) | Controls ( N = 151,520) | PCOS versus controls | |||||
| Mean risk time (years) | N (%) | Incidence rate per 1,000 PY | N (%) | Incidence rate per 1,000 PY | IRR | P ‐value | |
| Type 1 diabetes | 11.67 | 416 (1.37%) | 1.17 (1.07; 1.29) | 898 (0.59%) | 0.51 (0.48; 0.54) | 2.308 (2.055; 2.592) | < 0.001 |
| Thyroid disorder | 11.59 | 1125 (3.71%) | 3.20 (3.02; 3.39) | 3253 (2.15%) | 1.86 (1.80; 1.92) | 1.721 (1.608; 1.841) | < 0.001 |
| Blood, D | 11.74 | 93 (0.31%) | 0.26 (0.21; 0.32) | 408 (0.27%) | 0.23 (0.21; 0.25) | 1.120 (0.893; 1.404) | 0.326 |
| Endocrine, E | 11.67 | 422 (1.39%) | 1.19 (1.08; 1.31) | 1331 (0.88%) | 0.76 (0.72; 0.80) | 1.576 (1.412; 1.758) | < 0.001 |
| CNS, G | 11.75 | 80 (0.26%) | 0.22 (0.18; 0.28) | 350 (0.23%) | 0.20 (0.18; 0.22) | 1.132 (0.888; 1.444) | 0.316 |
| Eye, H | 11.75 | 45 (0.15%) | 0.13 (0.09; 0.17) | 206 (0.14%) | 0.12 (0.10; 0.13) | 1.082 (0.784; 1.494) | 0.631 |
| Circulatory, I | 11.76 | 24 (0.08%) | 0.07 (0.05; 0.10) | 96 (0.06%) | 0.05 (0.04; 0.07) | 1.239 (0.792; 1.937) | 0.348 |
| Digestive, K | 11.70 | 291 (0.96%) | 0.82 (0.73; 0.92) | 1332 (0.88%) | 0.76 (0.72; 0.80) | 1.084 (0.955; 1.231) | 0.213 |
| Skin, L | 11.73 | 148 (0.49%) | 0.42 (0.35; 0.49) | 578 (0.38%) | 0.33 (0.30; 0.36) | 1.269 (1.059; 1.520) | 0.010 |
| Musculoskeletal, M | 11.71 | 218 (0.72%) | 0.61 (0.54; 0.70) | 861 (0.57%) | 0.49 (0.46; 0.52) | 1.255 (1.082; 1.456) | 0.003 |
| Total | 11.34 | 2281 (7.52%) | 6.63 (6.36; 6.91) | 7527 (4.97%) | 4.38 (4.28; 4.48) | 1.515 (1.445; 1.587) | < 0.001 |
Incidence of Autoimmune Diseases during Follow Up
(Table 3 ): The mean follow up duration was 9.3 years in the study cohort. During follow up, the IRR of autoimmune disease was 1.50 (1.44; 1.57) in women with PCOS compared to controls. Women with PCOS had higher IRR of T1D 3.49 (3.16; 3.87), thyroid disorder 1.48 (1.40; 1.57), endocrine 1.67 (1.51; 1.86), and skin 1.55 (1.33; 1.81) autoimmune diseases during follow up. Regression models adjusting for ethnicity, low family income, morbidity, psychiatric disease, infertility, use of oral contraceptives and number of births, did not significantly change results (appendix, Table 1 ).
Table 3
Risk of autoimmunity in women with PCOS and controls after baseline.
Presented as incidence rate (IR) per 1000 PY (person years). Type 1 diabetes: ICD10, E10 Type 1 diabetes and/or prescription of ATC code A10A: Insulin. Thyroid disorder: ICD10: E05.0 Thyrotoxicosis, E06.3 Autoimmune thyroiditis and/or prescription of ATC code H03: Thyroid. Other autoimmune diseases: Diagnosis of any autoimmune disease according to the ICD10 diagnosis codes: Blood: D51.0 pernicious anemia, D59.0‐1 autoimmune hemolytic anemia, D69.0 Henoch Schönlein purpura, D69.3 idiopathic thrombocytopenic purpura, D86 sarcoidosis. Endocrine: E27.1 primary adrenal insufficiency. CNS: G35 multiple sclerosis, G61.0 Guillain‐Barre syndrome, G70.0 myastenia gravis Eye: H20 iridycyclitis. Circulatory: I00, I01 rheumatic fever, I730 Raynaud′d phenomen. Digestive: K50 Crohn′s disease, K51 ulcerative colitis, K73 autoimmune hepatitis, K74.3 primary biliary cirrhosis, K90.0 celiac disease. Skin: L10 pemphigus, L12 pemphigoid, L13.0 dermatitis herpetiformis, L40 psoriasis vulgaris, L52 erythema nodosum, L63 alopecia areata, L80 vilitigo, L93 lupus erythematosus, L94.0‐3 lichen sclerosus. Musculoskeletal: M02.3 Reiter′s disease M05‐M06 seropositive rheumatoid arthritis, M08 juvenile arthritis, M30.0 polyarteritis nodosa, M30.3 Kawasaki syndrome, M31.0 Goodpasture′s syndrome, M31.3 Wegeners granulomatosis, M31.5‐6, M35.3 polymyalgia rheumatica, M32 systemic lupus erythematosus, M33 dermatopolymyositis, M34 systemic sclerosis, M32 systemic lupus erytematosis, M35.0 Sjogrens syndrome, M35.2 Behcet′s disease, M45 Ankylosing spondylitis, M72.0 Dupuytren′s disease.
| PCOS ( N = 30,340) | Controls ( N = 151,520) | PCOS versus controls | |||||
| Mean risk time (years) | N (%) | incidence rate per 1000 PY | N (%) | incidence rate per 1000 PY | IRR | P ‐value | |
| Type 1 diabetes | 9.81 | 640 (2.11%) | 2.16 (2.00; 2.33) | 910 (0.60%) | 0.62 (0.58; 0.66) | 3.493 (3.156; 3.866) | < 0.001 |
| Thyroid disorder | 9.60 | 1657 (5.46%) | 5.81 (5.54; 6.10) | 5581 (3.68%) | 3.92 (3.82; 4.03) | 1.483 (1.403; 1.567) | < 0.001 |
| Blood, D | 9.95 | 72 (0.24%) | 0.24 (0.19; 0.30) | 313 (0.21%) | 0.21 (0.19; 0.24) | 1.115 (0.862; 1.443) | 0.408 |
| Endocrine, E | 9.84 | 475 (1.57%) | 1.56 (1.43; 1.71) | 1368 (0.90%) | 0.94 (0.89; 0.99) | 1.673 (1.505; 1.859) | < 0.001 |
| CNS, G | 9.94 | 87 (0.29%) | 0.29 (0.23; 0.36) | 413 (0.27%) | 0.28 (0.25; 0.31) | 1.035 (0.821; 1.304) | 0.770 |
| Eye, H | 9.95 | 77 (0.25%) | 0.26 (0.20; 0.32) | 313 (0.21%) | 0.21 (0.19; 0.24) | 1.209 (0.942; 1.552) | 0.135 |
| Circulatory, I | 9.96 | 30 (0.10%) | 0.10 (0.07; 0.14) | 132 (0.09%) | 0.09 (0.08; 0.11) | 1.117 (0.751; 1.660) | 0.585 |
| Digestive, K | 9.90 | 265 (0.87%) | 0.88 (0.78; 0.99) | 1240 (0.82%) | 0.85 (0.80; 0.89) | 1.040 (0.911; 1.189) | 0.560 |
| Skin, L | 9.92 | 213 (0.70%) | 0.70 (0.61; 0.80) | 665 (0.44%) | 0.45 (0.42; 0.49) | 1.548 (1.325; 1.808) | < 0.001 |
| Musculoskeletal, M | 9.90 | 294 (0.97%) | 0.97 (0.87; 1.09) | 1342 (0.89%) | 0.91 (0.87; 0.96) | 1.067 (0.940; 1.211) | 0.317 |
| Total | 9.30 | 2845 (9.38%) | 10.69 (10.30; 11.09) | 9665 (6.38%) | 7.12 (6.98; 7.26) | 1.501 (1.439; 1.565) | < 0.001 |
Predictors of Autoimmune Disease in Women With PCOS
(Table 4 ): In women with PCOS, non‐Danish ethnicity, low family income, morbidity, presence of psychiatric diagnosis/treatment, and infertility increased the hazards of autoimmune disease. Use of hormonal anticonception decreased the hazards of T1D and increased the hazards of thyroid disorder. Parity decreased the hazards of T1D, whereas parity did not affect the hazards of thyroid disorder in PCOS. Higher BMI increased the hazards of total autoimmunity and T1D.
Table 4
Hazard ratios from Cox regression models, predictors of autoimmune disease in women with PCOS.
Cox proportional hazards models in PCOS and controls with autoimmunity (total autoimmunity, T1D and thyroid disorder) entered as dependent variable. Morbidity defined as Charlson comorbidity index ≥ 1 from five years before and until the index date. Psychiatric outcomes defined according to ICD10 F10‐F41, X60‐X84 and ATC N06AA, N06AB, N06AF, N06AG, N06AX with ≥ 2 redemptions of medicine. Infertility defined according to ICD10: N97, Z35, ATC: G03GA, G03GB. Hormonal anticonception defined according to ATC G03AA, G03AB, G03HB01, G03AC, G02BA03.
| Total autoimmunity | T1D | Thyroid disorder | ||
| Ethnicity | Danish | ref. | ref. | ref. |
| Non‐Danish | 1.321 (1.235; 1.413) | 1.307 (1.127; 1.517) | 1.619 (1.484; 1.766) | |
| Family income | Low | ref. | ref. | ref. |
| Middle | 0.889 (0.832; 0.949) | 0.919 (0.795; 1.063) | 0.883 (0.808; 0.965) | |
| High | 0.790 (0.738; 0.845) | 0.845 (0.728; 0.981) | 0.770 (0.702; 0.845) | |
| Morbidity | No | ref. | ref. | ref. |
| Yes | 1.286 (1.196; 1.383) | 1.633 (1.415; 1.885) | 0.801 (0.717; 0.895) | |
| Psychiatric | No | ref. | ref. | ref. |
| Yes | 1.703 (1.598; 1.815) | 1.999 (1.740; 2.296) | 1.559 (1.431; 1.699) | |
| Infertility | No | ref. | ref. | ref. |
| Yes | 1.339 (1.256; 1.429) | 1.654 (1.440; 1.900) | 1.473 (1.353; 1.604) | |
| Hormonal anticonception | No | ref. | ref. | ref. |
| Yes | 1.123 (1.053; 1.197) | 0.713 (0.621; 0.817) | 1.243 (1.137; 1.359) | |
| No of births | 0 | ref. | ref. | ref. |
| ≥ 1 | 0.901 (0.830; 0.977) | 0.511 (0.414; 0.631) | 1.009 (0.909; 1.120) | |
| BMI | Per 1 point increase | 1.025 (1.014; 1.036) | 1.085 (1.066; 1.105) | 1.011 (0.997; 1.025) |
Sensitivity Analyses
When T1D was diagnosed with ICD10 diagnosis code E10 alone, the IRR for T1D before PCOS diagnosis was 1.82 (1.56; 2.14) and the IRR for T1D after the PCOS diagnosis was 2.38 (1.96; 2.90), both p < 0.001.
Analyses were repeated after omitting women with a diagnosis code of hirsutism alone, which resulted in a study cohort of 18,368 women with PCOS. In the sensitivity analysis, women with PCOS had median age of 27.0 (23.0−32.0) years at baseline. At baseline, women with PCOS versus controls had higher prevalence of low family income (29.9% vs. 33.5%), comorbidity (4.3% vs. 2.6%), psychiatric diagnosis/treatment (24.6 vs. 17.4%), and infertility (40.4% vs. 13.1%), all p < 0.001, whereas the prevalence of foreign ethnicity was comparable in women with PCOS and controls (19.9% vs. 16.5%), p = 0.24. BMI in women with PCOS was 25.4 (22.0−30.5) kg/m 2 compared to 22.8 (20.7‐25.9) kg/m 2 in controls. The prevalence of nulliparity in women with PCOS compared to controls was 38.3% versus 41.3% and women with PCOS more seldom had more than two births (38.9% vs. 42.2%) compared to controls, p < 0.001. Results regarding autoimmune disease in the sensitivity analysis were similar to results in the whole study cohort. Before baseline, the IRR of autoimmune disease was 1.56 (1.47; 1.66) in women with PCOS compared to controls, the IRR for T1D was 2.55 (2.19; 2.97) and the IRR for thyroid disorder was 1.94 (1.77; 2.12). During follow up, the IRR of autoimmune disease was 1.69 (1.59; 1.79) in women with PCOS compared to controls, the IRR for T1D was 4.73 (4.07; 5.48), and the IRR for thyroid disorder was 1.67 (1.55; 1.80).
Discussion
In the present national register‐based study including 30,340 women with PCOS and 151,520 controls, we found 1.5 times increased incidence of any autoimmune disease in women after PCOS diagnosis compared to controls, and the incidence for T1D was 3.5 times increased in women with PCOS diagnosis. In women with PCOS, lower socioeconomic status, comorbidity and higher BMI increased the hazards for autoimmune disease. These data outline the importance of autoimmunity in women with PCOS and support that stress and inflammation could be associated with higher autoimmunity in women with PCOS.
Our finding of higher incidence of T1D in PCOS is in accordance with previous studies reporting higher prevalence of PCOS in women with T1D [ 5 , 6 ], however the present study design allowed us to further investigate possible mechanisms for coexistence of PCOS and T1D. We found IRR of 3.45 for T1D in women with PCOS compared to controls after the date for PCOS diagnosis and an IRR of 2.30 before the index date. Our finding that PCOS is often diagnosed before T1D is novel as it contradicts a hypothesis of iatrogenic hyperinsulinemia as a dominant driver for development of PCOS in women with T1D [ 6 ]. It should be considered that women with T1D will attend regular follow up with health care providers and surveillance bias can increase the awareness for diagnosis of PCOS in premenopausal women with T1D. The median age at PCOS diagnosis was 29 years, which supported delay of PCOS diagnosis in our study cohort. However, in the present study, higher incidence for autoimmune disease in PCOS was observed not only for T1D, but also for other autoimmune diseases. Furthermore, PCOS and TID have been reported to share genetic factors [ 31 , 32 ], which further support a common pathogenic mechanism for PCOS and T1D.
The incidence rate for thyroid disease was 1.5 times increased in women with PCOS compared to controls in the present study. In Denmark, treatment of thyroid disorders is often handled by general practitioners, but the study outcome of thyroid disorder included medical treatment for thyroid diseases, which allowed us to identify patients treated for thyroid disorders outside a hospital setting. Our finding of higher incidence for thyroid autoimmunity according to ICD10 and medicine use for thyroid disorders in PCOS is in accordance with previous studies [ 7 , 8 ]. In our previous national study, we reported higher risk for overt hypothyroidism and Graves′ disease in Danish women with PCOS [ 7 ]. A recent meta‐analysis by Bahreiny [ 8 ] included a total of 2,076 women with PCOS and 1,581 controls and found pooled OR of 2.38 (95% CI: 1.63–3.49; p < 0.001) for autoimmune thyroid disease in women with PCOS compared to controls. Autoimmune thyroid disease was defined as elevated TSH, positive Anti‐TPO and/or positive anti‐Tg antibodies, ultrasound findings, and/or clinical symptoms [ 8 ]. Furthermore, higher prevalence of positive TRAb antibodies in PCOS has been described [ 4 ] and support recommendations of routine screening for thyroid disease in PCOS diagnosis [ 1 ]. In adjusted regression analyses, we found that low family income, psychiatric disorder, infertility, and obesity increased the hazards of autoimmune diseases in women with PCOS. These findings are novel and support that several stress factors are associated with higher risk for autoimmunity in women with PCOS. Quality of life is impaired in women with PCOS with comparable mental health scores between women with PCOS and patients with chronic disease e.g. asthma [ 33 , 34 , 35 ], and the risk for depression is increased in PCOS [ 21 ]. Treatment with antidepressants before a PCOS diagnosis could be associated with weight gain and development of a manifest PCOS phenotype [ 36 ], and the link between higher BMI and inflammation could result in higher autoimmune risk. Accordingly, women with PCOS and depression could represent a high risk phenotype for autoimmune disease.
Use of hormonal contraceptives increased the hazards of thyroid disorders and total autoimmunity in the present study. Use of estrogens and progestins affect the immune system and can increase the risk of several autoimmune diseases [ 37 ]. Oral contraceptives can also result in increased inflammation [ 38 ], weight gain [ 39 ], and mood changes [ 40 ], which can be associated with autoimmune disease [ 41 ]. In contrast, use of hormonal contraception and higher number of births were associated with lower incidence for T1D in the present study. Preconception glucose monitoring and treatment of pregnant women with T1D is highly specialized and diabetes dysregulation is associated with infertility and pregnancy complications. These reservations can explain our finding of an inverse association between T1D, use of birth control and birth number.
Strengths and limitations apply to the present study. An important strength was our prospective study design with well validated register data, which allowed us to investigate the order of PCOS and autoimmune diseases. The PCOS diagnosis was obtained by available ICD‐10 codes at hospital contacts. Some women in the control group could have undiagnosed PCOS, which could underestimate the incidence of autoimmune diseases in PCOS. The study population was a relatively lean Nordic study population of women with PCOS (median BMI at first pregnancy 25.1 kg/m 2 in PCOS vs. 22.8 kg/m 2 in controls) and the incidence of autoimmune diseases was associated with ethnicity, comorbidity and BMI. Therefore, the present study results should be validated in study populations of other ethnicities and with higher baseline metabolic risk. PCOS was defined according to ICD‐10 diagnosis code of PCOS and/or hirsutism. Inclusion of women with hirsutism in the data set can be considered a study strength as this allowed the inclusion of different PCOS phenotypes in the data set. However, hirsutism can be idiopathic or hirsutism can be a symptom of rare endocrine diseases [ 18 , 19 ]. Women with PCOS diagnosis code alone were slightly younger, more often of Danish ethnicity and had a higher prevalence of infertility compared to the full study cohort of women with PCOS (including the L680 diagnosis code, hirsutism). When we omitted women with L680 diagnosis alone from the analyses, study results regarding autoimmune disease were unchanged, but the IRR for T1D was higher, which suggested that some PCOS phenotypes could be associated with even higher risk for T1D. This result should be confirmed in future studies.
Conclusion
The incidence of autoimmune disease was significantly higher in women diagnosed with PCOS compared to controls. Our data support continued awareness for autoimmune diseases in women with PCOS not only at diagnosis of PCOS, but also during follow up.
Author Contribution
Dorte Glintborg, Marianne Skovsager Andersen, Jens‐Jakob Kjer Møller, Louise Lehmann Christensen: Idea, design, writing manuscript Jens‐Jakob Kjer Møller: Data analyses, Katrine Hass Rubin: Design, criticism of manuscript.
Disclosure
The authors have nothing to report.
Attestation Statement
Data regarding any of the subjects in the study has not been previously published unless specified.
Capsule
The incidence of autoimmune disease was significantly higher in Danish women diagnosed with PCOS compared to controls.
Supporting information
appendixtable autoimmunity PCOS Rev1.
Data Availability Statement
Danish legislation does not allow sharing of the data set as individual level data are protected according to GDPR (general data protection regulation) rules. Authors can be contacted regarding summarized results, if needed.
References
1 H. J. Teede , C. T. Tay , J. J. E. Laven , et al., “ Recommendations From the 2023 International Evidence‐Based Guideline for the Assessment and Management of Polycystic Ovary Syndrome ,” European Journal of Endocrinology 189 , no. 2 ( 2023 ): G43 – G64 . 37580861 10.1093/ejendo/lvad096
2 D. Glintborg , M. M. Ollila , J. K. Møller , et al., “ Prospective Risk of Type 2 Diabetes in 99892 Nordic Women With Polycystic Ovary Syndrome and 446055 Controls: National Cohort Study From Denmark, Finland, and Sweden ,” Human Reproduction 39 ( 2024 ): 1823 – 1834 . 38859639 10.1093/humrep/deae124
3 J. Wang , T. Yin , and S. Liu , “ Dysregulation of Immune Response in PCOS Organ System ,” Frontiers in immunology 14 ( 2023 ): 1169232 . 37215125 10.3389/fimmu.2023.1169232 PMC10196194
4 H. Mobeen , N. Afzal , and M. Kashif , “ Polycystic Ovary Syndrome May Be An Autoimmune Disorder ,” Scientifica 2016 ( 2016 ): 4071735 . 27274883 10.1155/2016/4071735 PMC4871972
5 A. Bayona , V. Martínez‐Vaello , J. Zamora , L. Nattero‐Chávez , M. Luque‐Ramírez , and H. F. Escobar‐Morreale , “ Prevalence of PCOS and Related Hyperandrogenic Traits in Premenopausal Women With Type 1 Diabetes: A Systematic Review and Meta‐Analysis ,” Human Reproduction Update 28 , no. 4 ( 2022 ): 501 – 517 . 35237802 10.1093/humupd/dmac011
6 M. R. García‐Sáenz , M. Lobaton‐Ginsberg , C. Ramírez‐Rentería , et al., “ Hirsutism and Polycystic Ovarian Morphology Are the Most Frequent Components of Polycystic Ovary Syndrome in Women With Type 1 Diabetes ,” Archives of Medical Research 54 , no. 7 ( 2023 ): 102895 . 37866088 10.1016/j.arcmed.2023.102895
7 D. Glintborg , K. Rubin , M. Nybo , B. Abrahamsen , and M. Andersen , “ Increased Risk of Thyroid Disease in Danish Women With Polycystic Ovary Syndrome ,” Endocrine Connections 8 ( 2019 ): 1405 – 1415 . 31518989 10.1530/EC-19-0377 PMC6826171
8 S. S. Bahreiny , A. Ahangarpour , M. Amraei , et al., “ Autoimmune Thyroid Disorders and Polycystic Ovary Syndrome: Tracing Links Through Systematic Review and Meta‐Analysis ,” Journal of Reproductive Immunology 163 ( 2024 ): 104215 . 38402811 10.1016/j.jri.2024.104215
9 M. Romitti , V. C. Fabris , P. K. Ziegelmann , A. L. Maia , and P. M. Spritzer , “ Association Between PCOS and Autoimmune Thyroid Disease: A Systematic Review and Meta‐Analysis ,” Endocrine Connections 7 , no. 11 ( 2018 ): 1158 – 1167 . 30352422 10.1530/EC-18-0309 PMC6215798
10 G. Frazzei , R. F. van Vollenhoven , B. A. de Jong , S. E. Siegelaar , and D. van Schaardenburg , “ Preclinical Autoimmune Disease: A Comparison of Rheumatoid Arthritis, Systemic Lupus Erythematosus, Multiple Sclerosis and Type 1 Diabetes ,” Frontiers in immunology 13 ( 2022 ): 899372 . 35844538 10.3389/fimmu.2022.899372 PMC9281565
11 D. Glintborg , J. J. K. Møller , K. H. Rubin , et al., “ Mental and Physical Health Among Danish Transgender Persons Compared With Cisgender Persons ,” JAMA Network Open 8 , no. 4 ( 2025 ): e257115 . 40272800 10.1001/jamanetworkopen.2025.7115 PMC12022810
12 D. Glintborg , J. J. K. Møller , K. H. Rubin , L. L. Christensen , and M. S. Andersen , “ Autoimmune Diseases in 3812 Danish Transgender Persons and 38 120 Cisgender Controls before and After Transgender Care: A Register‐Based Cohort Study ,” European Journal of Endocrinology 192 , no. 4 ( 2025 ): 408 – 417 . 40086810 10.1093/ejendo/lvaf051
13 L. Kujanpää , R. K. Arffman , P. Pesonen , et al., “ Women With Polycystic Ovary Syndrome Are Burdened With Multimorbidity and Medication Use Independent of Body Mass Index At Late Fertile Age: A Population‐Based Cohort Study ,” Acta Obstetricia et Gynecologica Scandinavica 101 , no. 7 ( 2022 ): 728 – 736 . 35673942 10.1111/aogs.14382 PMC9564432
14 K. Laugesen , J. F. Ludvigsson , M. Schmidt , et al., “ Nordic Health Registry‐Based Research: A Review of Health Care Systems and key Registries ,” Clinical Epidemiology 13 ( 2021 ): 533 – 554 . 34321928 10.2147/CLEP.S314959 PMC8302231
15 M. Baadsgaard and J. Quitzau , “ Danish Registers on Personal Income and Transfer Payments ,” supplement, Scandinavian Journal of Public Health 39 , no. 7 Suppl ( 2011 ): 103 – 105 . 21775365 10.1177/1403494811405098
16 S. Nader , “ Hyperandrogenism During Puberty in the Development of Polycystic Ovary Syndrome ,” Fertility and Sterility 100 , no. 1 ( 2013 ): 39 – 42 . 23642453 10.1016/j.fertnstert.2013.03.013
17 M. Forman , L. Mangini , R. Thelus‐Jean , and M. D. Hayward , “ Life‐Course Origins of the Ages At Menarche and Menopause ,” Adolescent Health, Medicine and Therapeutics 4 ( 2013 ): 1 – 21 . 24600293 10.2147/AHMT.S15946 PMC3912848
18 D. Glintborg and M. Andersen , “ An Update on the Pathogenesis, Inflammation, and Metabolism in Hirsutism and Polycystic Ovary Syndrome ,” Gynecological Endocrinology 26 , no. 4 ( 2010 ): 281 – 296 . 20141388 10.3109/09513590903247873
19 R. Azziz , M. Amiri , and F. Bril , et al, “ Approach to the Patient: Hirsutism ,” Journal of Clinical Endocrinology and Metabolism , ahead of print, April 9, 2025: dgaf226 , 10.1210/clinem/dgaf226 . 40200552
20 Revised 2003 Consensus on Diagnostic Criteria and Long‐Term Health Risks Related to Polycystic Ovary Syndrome ,” Fertility and Sterility 81 , no. 1 ( 2004 ): 19 – 25 . 10.1016/j.fertnstert.2003.10.004 14711538
21 D. Glintborg , K. Hass Rubin , M. Nybo , B. Abrahamsen , and M. Andersen , “ Morbidity and Medicine Prescriptions in a Nationwide Danish Population of Patients Diagnosed With Polycystic Ovary Syndrome ,” European Journal of Endocrinology 172 , no. 5 ( 2015 ): 627 – 638 . 25656495 10.1530/EJE-14-1108
22 K. H. Rubin , D. Glintborg , M. Nybo , M. Andersen , and B. Abrahamsen , “ Fracture Risk Is Decreased in Women With Polycystic Ovary Syndrome: A Register‐Based and Population‐Based Cohort Study ,” Journal of Bone and Mineral Research 31 , no. 4 ( 2016 ): 709 – 717 . 26542642 10.1002/jbmr.2737
23 K. H. Rubin , D. Glintborg , M. Nybo , B. Abrahamsen , and M. Andersen , “ Development and Risk Factors of Type 2 Diabetes in a Nationwide Population of Women With Polycystic Ovary Syndrome ,” Journal of Clinical Endocrinology & Metabolism 102 , no. 10 ( 2017 ): 3848 – 3857 . 28938447 10.1210/jc.2017-01354
24 D. Glintborg , K. H. Rubin , M. Nybo , B. Abrahamsen , and M. Andersen , “ Cardiovascular Disease in a Nationwide Population of Danish Women With Polycystic Ovary Syndrome ,” Cardiovascular Diabetology 17 , no. 1 ( 2018 ): 37 . 29519249 10.1186/s12933-018-0680-5 PMC5844097
25 D. Glintborg , T. G. Petersen , K. H. Rubin , and M. S. Andersen , “ Diabetes Mellitus Mediates Risk of Depression in Danish Women With Polycystic Ovary Syndrome‐A National Cohort Study ,” Biomedicines 10 , no. 10 ( 2022 ): 2396 . 36289658 10.3390/biomedicines10102396 PMC9598722
26 M. C. Harpsøe , S. Basit , M. Andersson , et al., “ Body Mass Index and Risk of Autoimmune Diseases: A Study Within the Danish National Birth Cohort ,” International Journal of Epidemiology 43 , no. 3 ( 2014 ): 843 – 855 . 24609069 10.1093/ije/dyu045
27 J. Janbek , T. M. Laursen , N. Frimodt‐Møller , et al., “ Hospital‐Diagnosed Infections, Autoimmune Diseases, and Subsequent Dementia Incidence ,” JAMA Network Open 6 , no. 9 ( 2023 ): e2332635 . 37676660 10.1001/jamanetworkopen.2023.32635 PMC10485730
28 M. E. Charlson , P. Pompei , K. L. Ales , and C. R. MacKenzie , “ A New Method of Classifying Prognostic Comorbidity in Longitudinal Studies: Development and Validation ,” Journal of Chronic Diseases 40 , no. 5 ( 1987 ): 373 – 383 . 3558716 10.1016/0021-9681(87)90171-8
29 H. Quan , V. Sundararajan , P. Halfon , et al., “ Coding Algorithms for Defining Comorbidities in ICD‐9‐CM and ICD‐10 Administrative Data ,” Medical Care 43 , no. 11 ( 2005 ): 1130 – 1139 . 16224307 10.1097/01.mlr.0000182534.19832.83
30 H. Quan , B. Li , C. M. Couris , et al., “ Updating and Validating the Charlson Comorbidity Index and Score for Risk Adjustment in Hospital Discharge Abstracts Using Data From 6 Countries ,” American Journal of Epidemiology 173 , no. 6 ( 2011 ): 676 – 682 . 21330339 10.1093/aje/kwq433
31 B. M. Mussa , T. Venkatachalam , A. Srivastava , et al., “ Identification of Novel Differentially Expressed Genes in Type 1 Diabetes Mellitus Complications Using Transcriptomic Profiling of UAE Patients: A Multicenter Study ,” Scientific Reports 12 , no. 1 ( 2022 ): 16316 . 36175575 10.1038/s41598-022-18997-w PMC9523055
32 W. Chen , Q. Yang , L. Hu , et al., “ Shared Diagnostic Genes and Potential Mechanism Between PCOS and Recurrent Implantation Failure Revealed by Integrated Transcriptomic Analysis and Machine Learning ,” Frontiers in immunology 14 ( 2023 ): 1175384 . 37261354 10.3389/fimmu.2023.1175384 PMC10228695
33 C. Castelo‐Branco and I. Naumova , “ Quality of Life and Sexual Function in Women With Polycystic Ovary Syndrome: A Comprehensive Review ,” Gynecological Endocrinology 36 , no. 2 ( 2020 ): 96 – 103 . 31559883 10.1080/09513590.2019.1670788
34 A. Dokras , E. Stener‐Victorin , B. O. Yildiz , et al., “ Androgen Excess‐ Polycystic Ovary Syndrome Society: Position Statement on Depression, Anxiety, Quality of Life, and Eating Disorders in Polycystic Ovary Syndrome ,” Fertility and Sterility 109 , no. 5 ( 2018 ): 888 – 899 . 29778388 10.1016/j.fertnstert.2018.01.038
35 G. L. Jones , J. M. Hall , A. H. Balen , and W. L. Ledger , “ Health‐Related Quality of Life Measurement in Women With Polycystic Ovary Syndrome: A Systematic Review ,” Human Reproduction Update 14 , no. 1 ( 2007 ): 15 – 25 . 17905857 10.1093/humupd/dmm030
36 M. L. Altinok , D. Glintborg , R. dePont Christensen , J. Hallas , and M. Andersen , “ Prescription of Antidepressants Is Increased in Danish Patients With Polycystic Ovary Syndrome and Is Associated With Hyperandrogenism. A Population‐Based Cohort Study ,” Clinical Endocrinology 80 , no. 6 ( 2013 ): 884 – 889 . 24224860 10.1111/cen.12365
37 W. V. Williams , “ Hormonal Contraception and the Development of Autoimmunity: A Review of the Literature ,” Linacre Quarterly 84 , no. 3 ( 2017 ): 275 – 295 . 28912620 10.1080/00243639.2017.1360065 PMC5592309
38 S. F. de Medeiros , M. A. S. Medeiros , N. S. Santos , B. B. Barbosa , and M. M. W. Yamamoto , “ Combined Oral Contraceptive Effects on Low‐Grade Chronic Inflammatory Mediators in Women With Polycystic Ovary Syndrome: A Systematic Review and Meta‐Analysis ,” International Journal of Inflammation 2018 ( 2018 ): 9591509 . 30595838 10.1155/2018/9591509 PMC6286752
39 D. Glintborg , M. L. Altinok , H. Mumm , A. P. Hermann , P. Ravn , and M. Andersen , “ Body Composition Is Improved During 12 Months Treatment With Metformin Alone or Combined With Oral Contraceptives Compared to Treatment With Oral Contraceptives in Polycystic Ovary Syndrome ,” Journal of Clinical Endocrinology and Metabolism 99 ( 2014 ): 2584 – 2591 . 24742124 10.1210/jc.2014-1135
40 C. W. Skovlund , L. S. Mørch , L. V. Kessing , and Ø. Lidegaard , “ Association of Hormonal Contraception With Depression ,” JAMA Psychiatry 73 , no. 11 ( 2016 ): 1154 – 1162 . 27680324 10.1001/jamapsychiatry.2016.2387
41 A. Czarnywojtek , M. Owecki , M. Zgorzalewicz‐Stachowiak , et al., “ The Role of Serum C‐Reactive Protein Measured by High‐Sensitive Method in Thyroid Disease ,” Archivum Immunologiae et Therapiae Experimentalis 62 , no. 6 ( 2014 ): 501 – 509 . 24794233 10.1007/s00005-014-0282-1 PMC4244578