Prevalence of Hyperandrogenism in Type 1 Diabetes

Polycystic ovary syndrome (PCOS) is the most common endocrine disorder in women of reproductive age, with an estimated prevalence of 6-15% of the general population worldwide. This heterogeneous syndrome has significant cardio-metabolic, reproductive, and psycho-emotional consequences, and therefore, a prompt recognition and management is of paramount importance for these women. Despite hyperandrogenism is the cornerstone in the pathophysiology of PCOS, this derangement is closely related to insulin resistance, compensatory hyperinsulinemia, and abdominal adiposity. Hyperinsulinemia increases androgen secretion by co-stimulating besides gonadotropins both ovary and adrenal steroidogenesis, which leads to predominant visceral/abdominal fat deposition, and further contributes to insulin resistance and hyperinsulinemia. In addition, PCOS has been classically associated with metabolic alterations such as for overweight/obesity and type 2 diabetes mellitus. However, type 1 diabetes mellitus (T1D) results from autoimmune-mediated destruction of the pancreas, causing a complete insulin lack in most patients. Intensive insulin therapy - a mandatory iatrogenic hyperinsulinism -, while improving chronic glycemic control and prognosis, has led in recent years to the appearance of "new" reproductive consequences in these patients, such as functional hyperandrogenism and menstrual irregularity. This association is expected from the stimulation of ovarian androgen production by exogenous insulin, which reaches the ovary in supraphysiological concentrations. However, these studies present with a high heterogeneity, and prevalence rates significantly vary depending on several variables such as the criteria used for PCOS diagnosis, race/ethnicity, age of the study population, and the prevalence of obesity, among others. In 2016, a systematic review assessing the prevalence of PCOS in T1D was published, including 475 women with T1D from 9 studies. The results showed an overall prevalence of PCOS about 24% in T1D, higher than reported in the general population. Other hyperandrogenic traits such as hirsutism (25%), hyperandrogenaemia (24%), or ovulatory dysfunction (33%) were also common. Although PCOS is one of the most common comorbidities in patients with T1D, there are a limited number of publications in the literature. In summary, PCOS and functional hyperandrogenism remain a condition to be explored thoroughly in these patients.

The investigators hypothesize that the prevalence of functional hyperandrogenism including PCOS in Spanish women with T1D is higher than in women from the general population. Furthermore, signs and symptoms of hyperandrogenism, and hyperandrogenemia may be milder in patients with T1D compared to hyperandrogenic women from the general population. Moreover, the occurrence of PCOS in these women may be influenced by insulin dose, duration of diabetes, and chronic metabolic control.

The main objective of this study is to determine the actual prevalence of PCOS in premenopausal women with T1DM, according to different diagnostic criteria/PCOS phenotypes [classic PCOS (classic NIH criteria), hyperandrogenic PCOS (AES-PCOS criteria), and/or inclusive ESHRE-ASRM/Rotterdam criteria]. As secondary goals, the investigators also aim to describe: i) the hyperandrogenic traits associated with PCOS in women with T1DM; and ii) the metabolic-T1D related parameters in women with or without hyperandrogenism.

Sample size calculation: Sample size analysis used the online sample size and power calculator from the Program of Research in Inflammatory and Cardiovascular Disorders, Institut Municipal d'Investigació Mèdica, Barcelona, Spain (https://www.imim.cat/ofertadeserveis/software-public/granmo/). Considering previous data on prevalence of SOP in adolescents and adult women with T1D according to ESHRE-ASRM/Rotterdam criteria, the investigators concluded that 150 participants would be needed to assume an expected proportion of 40%, with an absolute precision of 5% at both sides of the proportion, and an asymptotic bilateral 95% confidence interval, and with an estimated replacement rate of 10%.

Statistical analysis: Continuous variables will be expressed as mean ± SD with its respective 95% confidence intervals (95%CI). Normality of continuous variables will be checked by the Kolmogorov-Smirnov test, and ensured by applying logarithmic transformations. the investigators will use non-parametric tests to analyse variables that remained skewed even after transformation. The differences in means will be analysed by Student t or Mann-Whitney U tests. Discrete variables will be showed according to their absolute, relative frequency, and 95%CI determined using the Wilson method without continuity correction. The differences between proportions will be estimated using the χ2 or Fisher's exact tests. Correlation analysis will be used to evaluate putative association between continuous variables. Finally, multiple linear an binary logistic regression full and stepwise models (probability for entry ≤0.05, probability for removal ≥0.10) will be performed to ascertain the main determinants of predetermined outcomes. The statistical significance will be set at the P < 0.05 level.

Inclusion criteria

• Age between 18 and 45 years old
• Type 1 diabetes diagnosed at least 1 year before the inclusion in the study. Diagnosis confirmed by positive autoimmunity (GAD-65 or IA2) and insulin deficiency.
• Treatment with subcutaneus insulin therapy (multiple dose or continuous subcutaneous insulin infusion).
• Menarche at least 2 years before the study.

Exclusion criteria

• Honey moon period.
• Altered thyroid hormone or prolactin levels.
• Congenital adrenal hyperplasia.
• Severe chronic disease.
• Oral contraceptive or glucocorticoid therapy in the previous 3 months.