Introduction

Polycystic ovary syndrome (PCOS) affects nearly 3–10% of women in reproductive age [1,2,3]. It is a usual endocrine disorder and is associated with hyperandrogenism, chronic anovulation and the appearance of metabolic disturbances that may have serious implications for long-term health. Thus, there are several clinical definitions for PCOS, but the most broadly accepted is the association of hyperandrogenism with chronic anovulation in women without specific underlying diseases in adrenal or pituitary glands [4, 5].

Nowadays, PCOS is the most frequent cause of anovulation, infertility, endometrial hyperplasia and elevated androgen levels leading to hirsutism, alopecia and acne in women in reproductive age [6,7,8]. Also, PCOS is included within the complex metabolic syndrome (MetS) [9, 10], insulin resistance (IR), obesity and adiposity, obstructive sleep apnea (OSA), hypogonadism, lipodystrophy and microvascular disease [11,12,13,14,15,16,17,18,19,20,21,22,23].

Although the etiology of PCOS is not fully understood, it is considered a multifactorial disorder with genetic, metabolic and endocrine abnormalities [24].
IR with compensatory hyperinsulinism are common PCOS features. This IR seems to be the well-known physiopathological link between PCOS and MetS development, but its underlying mechanisms remain unclear.

The present work is a preliminary study in a small cohort of PCOS patients that precedes an ongoing project focused on the investigation of mechanisms involved in the relationship between PCOS development and MetS establishment. In this fashion, the purpose of the present study was to understand and investigate the impact of current therapies (combined oral contraceptives—COC, metformin and inositols) in PCOS metabolic parameters in a small cohort of patients diagnosed with such disorder.

Materials and methods

Participants and ethical procedures

This protocol was previously approved by the Ethics Committee of the “Fundación de Investigación HM Hospitales de Madrid” (14.11.704-GHM), governed by the basic ethical principles contained in the World Medical Declaration of Helsinki [25].

All diagnosed PCOS patients attending to Gynecology consultations at "Puerta del Sur Hospital (HM)” and “Majadahonda Medical Center” in Madrid (Spain) were offered to participate in the present study. Patients were included after obtaining their written informed consent or their parent/guardian written informed consent (in case of minors). Patients who refused to participate in the study were tracked, unless their refusal meant any change in treatment or patient’s care.

Diagnose criteria and phenotype characterization

PCOS diagnosis and characterization

PCOS was diagnosed according to the Rotterdam criteria [26], including at least two of the following symptoms: 1) oligo or anovulation; 2) clinical or biochemical signs of hyperandrogenism; 3) polycystic ovaries by ultrasound.

Also, four PCOS phenotypes were established according to the Rotterdam criteria [26]: phenotype A with oligoanovulation, clinical or analytical hyperandrogenism and ultrasound signs of PCOS; phenotype B with oligoanovulation and clinical or analytical hyperandrogenism; phenotype C with clinical or analytical hyperandrogenism and ultrasound images; and/or phenotype D with oligoanovulation and PCOS compatible sonographic images. All of them were considered as qualitative nominal variables.

MetS diagnosis

MetS diagnosis was established according to the criteria set forth by the National Cholesterol Education Program Adult Treatment Panel III (NECP-ATP III) [27], meeting at least three of the following: abdominal circumference > 88 cm, triglycerides > 150 mg/dL, HDL cholesterol < 50 mg/dL, blood pressure > 130/85 mmHg or glucose > 110 mg/dL.

Protocol design

An observational, retrospective and transversal study was made in a small cohort of 133 PCOS patients, aged 14–48 years. This cohort was divided into four experimental groups: 1) untreated PCOS patients (n = 51, 38.65%); and 2) PCOS patients treated with (n = 82, 61.45%): a) COC (n = 35, 25.9%); b) metformin (n = 11, 8.25%); and c) inositols (n = 36, 27.3%).

This study also included underaged women since it is in the early years of life where all biochemical and hormonal changes associated with the development of PCOS start, and usually is in the adolescence when PCOS is diagnosed.

After the inclusion of patients in one of the experimental groups, a single blood sample (8-h fasting, obtained in the follicular phase of the menstrual cycle) was acquired to determine hormonal and metabolic parameters.

Treatments and follow-up of patients

Treatments were indicated by the clinician, taking into account contraindications for contraceptive treatment. For inositol treatment, two preparations were used: 1) 2 g myoinositol and 200 μg folic acid (n = 23); 2) 200 mg myoinositol, 400 mg D-chiro-inositol, 10 mg manganese pidolate and 400 μg folic acid (n = 13). Several COC treatments were used, such as sustained release vaginal ring with 11.7 mg etonogestrel and 2.7 mg ethinyl estradiol; 2.5 mg nomegestrol acetate and 1.5 mg estradiol; 3 mg drospirenone and 0.02 mg ethinyl estradiol; 3 mg dienogest and 0.03 mg ethinyl estradiol; 100 mg of levonorgestrel and 0.02 mg ethinyl estradiol; being the preparation of 3 mg drospirenone and 0.03 mg ethinyl estradiol the most widely used in the present study. Ultimately, metformin treatment consisted of a standard dose of 850 mg/day of this drug.

Therapies were considered similar for all patients, with the only exception in PCOS patients treated with inositols, where the treatment was prolonged for either more than three months (n = 17) and less than 3 months (n = 19).

Serum analytical determinations

Follicle-stimulating hormone (FSH), luteinizing hormone (LH), estradiol, thyroid-stimulating hormone (TSH), prolactin, 17OH-progesterone, total testosterone, sex hormone-binding globulin (SHBG), androstenedione and dehydroepiandrosterone sulfate (DHEAs) were assessed by electrochemiluminescence immunoassay (ECLIA) using Elecsys and Cobas autoanalyzers. The rate of free androgens was estimated according to the formula: (total testosterone × 3.47/SHBG) × 100; as well as the ratio LH/FSH.

Total cholesterol, HDL, low-density lipoprotein (LDL), triglycerides, glucose, insulin and glycosylated hemoglobin were also determined. HOMA (homeostasis model assessment), an index of IR, was calculated as insulin (µU/ml) for glucose (mmol/L) / 22.5. IR was considered for values greater than 3. Additionally, sodium, potassium, urea and homocysteine serum levels were also determined. All of them were considered as continuous quantitative variables.

Measurement of additional clinical parameters

Body mass index (BMI, a continuous quantitative variable) was estimated according to the following formula: weight (Kg)/height (m2). BMI was used to classify patients in normal weight (BMI: 5–24.9), overweight (BMI: 25–29.9), obesity grade I (BMI: 30–34.9), obesity grade II (BMI: 35–39.9) and obesity grade III (BMI: > 40) [28].

Blood pressure (a continuous quantitative variable) was determined assessing systolic (SBP) and diastolic blood pressure (DBP) in each patient in sitting position. Mean blood pressure (MBD) was calculated with the formula: 1/3(SBP) + 2/3(DBP).

Statistical analysis

All data are represented as mean ± SMD. Statistical analysis was performed on SPSS 20 (IBM, USA). Qualitative and/or quantitative variables were analyzed with Student T test or \(\chi\)2 test. Correlations were evaluated by Spearman test or "r of Pearson". Significance was estimated by the Kruskal–Wallis ANOVA followed by a post hoc test for distribution-free multiple comparisons (Dunnett’s test) or Mann–Whitney test for unpaired samples. Differences were considered significant at p < 0.05.

Results

Clinical features of the population

As mentioned, the small cohort of 133 women (27.80 ± 6.56 years) was diagnosed with PCOS according to the Rotterdam criteria [26]. Regarding body weight and BMI, a total of 63 patients (47%) were overweight or have obesity (Table 1).

Table 1 General and clinical features of the present population
Full size table

Following Rotterdam criteria, patients were classified into four PCOS phenotypes (A, B, C or D). It was found that 50.8% of patients (67 women) belonged to phenotype A, the most frequent PCOS phenotype in this cohort, followed by phenotype D in 27 patients (20.5%), phenotype C in 25 patients (18.9%) and finally phenotype B in 13 patients (9.8%). Considering only obese patients, the most frequent observed PCOS phenotype was phenotype A (75%), followed by phenotype C (11.5%) and phenotype D (7.7%).

Regarding to clinical features, disturbances in menstrual cycle (81%), hirsutism (75.8%) and ultrasonographic alterations (75%) were the predominant clinical symptoms (Table 1).

Metabolic and endocrinological parameters in PCOS patients

PCOS patients included in this protocol were divided into four groups. Despite the limitations of the present study, treatment with COC in PCOS patients modified the hormonal profile reducing significantly FSH, LH and SHBG serum levels. In addition, COC therapies worsened lipid profile and increased circulating cholesterol and triglyceride levels without modulating glycemic metabolism compared to untreated PCOS patients (Table 2, Fig. 1a, b).

Table 2 Metabolic and endocrinological serum parameters in the four experimental groups of PCOS patients
Full size table
Fig. 1
figure 1

Cholesterol (a) and triglyceride (b) serum levels in untreated PCOS patients, COC-treated PCOS patients and inositol-treated patients. *p < 0.05, ***p < 0.001 COC-treated patients vs untreated patients; &p < 0.05 inositol-treated patients vs COC-treated patients

Full size image

However, as compared to untreated patients, inositol therapies improved glucose metabolism parameters (glycemia and insulinemia), including HOMA, a common index of insulin resistance, and glycosylated hemoglobin (Table 2, Fig. 2a–d). In addition, inositol therapies did not increase cholesterol and triglyceride serum levels, but this therapy reduced significantly both parameters as compared to untreated patients (Fig. 1a, b).

Fig. 2
figure 2

Serum glucose (a), insulin (b), HOMA (c) and glycosylated hemoglobin (d) in untreated PCOS patients, COC-treated PCOS patients and inositol-treated patients. *p < 0.05, **p < 0.01 COC both groups of treated patients vs untreated patients; &p < 0.05 inositol-treated patients vs COC-treated patients

Full size image

Of interest, inositol treatment, administered for more than three months, decreased SBP in PCOS patients (Fig. 3).

Fig. 3
figure 3

Systolic blood pressure (SBP, mmHg) in PCOS patients treated with inositol therapies during < 3 and > 3 months. ***p < 0.001 inositol-treated patients > 3 months vs inositol-treated patients > 3 months

Full size image

Endocrinological and metabolic parameters in obese PCOS patients

47% of patients were overweight or had obesity (Table 1). In the present protocol, obese patients were found in all experimental groups (50% in untreated patients, 26.92% in COC-, 11.53% inositol- and 11.53% metformin-treated patients).

A comparative study of metabolic and endocrinological parameters between obese (BMI > 30) and non-obese (BMI < 25) PCOS patients from the untreated group was performed (Table 3). Significant differences were found in endocrinological profile parameters: obese PCOS patients showed reduced prolactin (p < 0.05) and SHBG levels (p < 0.01) and increased free androgen serum levels (p < 0.05) compared to non-obese PCOS patients. In addition, obese PCOS patients showed reduced HDL (p < 0.01) and increased triglyceride serum levels (p < 0.05), as well as altered glucose metabolism parameters, showing hyperinsulinemia (p < 0.05) and a significant augment in insulin resistance (HOMA, p < 0.05) and glycosylated hemoglobin levels (p < 0.05) compared to non-obese PCOS patients (Table 3).

Table 3 Comparative study between untreated PCOS patients without (BMI < 25) and with obesity (BMI > 30)
Full size table

Additionally, untreated obese PCOS patients exhibit a significant increase in blood pressure (both SBP and DBP) compared to untreated non-obese PCOS patients (p < 0.001) (Table 3).

Finally, these metabolic and endocrinological parameters were compared between non-obese (BMI < 25) and obese (BMI > 30) patients who had received one of the PCOS treatments involved in the present study (COC, metformin and inositols). Most of these parameters were altered in between both groups (Table 4). It is clearly shown that obesity is a sign of poor prognosis that aggravates the progression of PCOS and hinders the effectiveness of any of the studied therapies, particularly in parameters related to dyslipidemia, insulin resistance and blood pressure.

Table 4 Comparative study of endocrine and metabolic parameters between patients without (BMI < 25) and with obesity (BMI > 30) treated for PCOS with combined oral contraceptives, metformin and inositols
Full size table

Correlations

Regarding untreated patients, direct and significant correlations were found between insulin serum levels and BMI, free androgens, triglycerides and MBP (p < 0.001), as well as between circulating glucose levels and glycosylated hemoglobin (r = 0.807, p < 0.001), testosterone (p < 0.005) and MBP (p < 0.03) (Fig. 4a). A direct significant correlation was found between BMI and MBP in obese PCOS patients (Fig. 4b).

Fig. 4
figure 4

a Direct or inverse correlations between circulating levels of glucose or insulin and other studied parameters in untreated PCOS patients. b Direct correlation between Body mass index (BMI) and medium blood pressure (MBP)

Full size image

Discussion

Despite the present limitations, results on this paper showed that COC treatment in PCOS patients had a negative metabolic impact, increasing cholesterol and triglyceride serum levels and insulin resistance, whereas inositol therapies did not affect lipid metabolism but improved insulin resistance, reducing glucose and insulin serum levels, particularly insulin resistance (HOMA) and glycosylated hemoglobin, compared to untreated patients.

Interestingly, this study showed that PCOS management in current clinical practice is not enough directed to solve or prevent the complexity of its endocrine and metabolic alterations. Most of the PCOS treatment guidelines suffer from an integrative view that should address hyperandrogenism and anovulation symptoms, and the vulnerability to metabolic syndrome establishment, overweight/obesity and even type 2 diabetes.

Although < 10% of PCOS patients included in this cohort can be strictly included in MetS, approximately 20% had insulin resistance. Additionally, the comparative study of parameters involved in lipid and glycemic metabolism provide scarce information about the metabolic impact of the different treatments prescribed to PCOS patients, despite the fact that in most patients MetS has not yet been established.

The major finding in this work is that COC treatment in PCOS patients modified the hormonal profile and worsened lipid parameters, and insulin resistance, while inositol therapies improved significantly insulin resistance and glycosylated hemoglobin, reducing cholesterol and triglyceride serum levels as compared to PCOS patients treated with COC. These beneficial effects of inositol therapies are in accordance with those reported by other authors [29,30,31,32,33,34,35,36,37,38,39], some of whom even advise the co-administration of myoinositol and D-chiro-inositol (40:1) to increase the effectiveness in restoring ovary function and metabolic parameters in PCOS [40, 41]. Thus, inositol therapies could be considered an easy, beneficial and integrative treatment for PCOS patients, due to their better tolerability and their diminished risk of adverse effects, compared to metformin treatments for PCOS patients [42]. In this fashion, inositol therapies in the present small cohort induced a significant increase in LH serum levels compared to values found in COC treatments, accordingly to the described effect of inositol in the stimulation of ovulation [43,44,45]. However, more clinical studies are needed in order to confirm this hypothesis.

Results in the present work provide evidences that obesity can be considered a bad prognosis since the presence of obesity (BMI > 30) worsens all parameters of both lipid and glycemic profile as well as blood pressure values (Table 3).

Although it could be considered anecdotal, a COC-treated PCOS patient, included in the present study, only after the loss of 15 kg of body weight became pregnant. These data are in accordance with others previously described where the association of moderate obesity with a poor pregnancy in PCOS was reported, along with the association between weight loss and the correction of gonadotropin and sex steroid alterations in obese anovulatory female [46,47,48,49,50,51]. The impact of polycystic ovaries on the future reproductive function of these women remains unclear, but evidence suggests that an obesity effective treatment is one key to improve fertility in PCOS patients [48,49,50, 52, 53].

When metabolic parameters were analyzed according to the phenotype, it was found that PCOS patients with phenotype A presented greater alterations in lipid metabolism and higher values of glycosylated hemoglobin as well as blood pressure, compared with the other phenotypes.

Interestingly, inositol therapies administered for more than three months decreased systolic blood pressure in PCOS patients. In these women, obesity was associated with greater alterations in lipid and glycemic metabolism and with higher blood pressure levels, as mentioned before.

As aforesaid, the etiology of PCOS is not fully understood. This is a multifactorial disorder associated with genetic, metabolic and endocrine abnormalities [24]. To date, PCOS is currently included in the metabolic syndrome (MetS) alterations [10]. In this way, PCOS patients have higher risks to suffer certain diseases compared to the general population, e.g., type II diabetes, cardiovascular disease, endometrial carcinoma and several gestational complications. These risks expose PCOS patients to high morbidity, being associated with an increase in economic and healthcare impact.

In this context, another potential factor involved in the pathophysiology of PCOS that deserves special mention is insulin-like growth factor 1 (IGF-1) deficiency, which still remains controversial. However, in the last decade several studies have been conducted, revealing the relevant role of IGF-1 deficiency in the development of MetS. Succinctly, an inverse correlation between IGF-1 (IGF-1/IGFBP-3 ratio) circulating levels and several markers for obesity, MetS, type II diabetes, and cardiovascular disease has been found, indicating that low IGF-1 circulating levels can result in MetS, raising the risk for cardiovascular disease and type II diabetes. Nonetheless, more studies are needed to describe the exact mechanism by which IGF-1 deficiency impacts and interacts with other factors and hormones to develop MetS, type II diabetes and its cardiovascular consequences [9].

Accumulated evidence suggests how the GH/IGF-1 axis, together with insulin and IGF-1 binding proteins (IGFBPs), act in a synchronized manner to regulate energy metabolism. Possibly, when this whole system becomes altered by obesity, genetics or environmental factors, several adverse consequences may develop, such as insulin resistance, steatosis, MetS and type II diabetes. A recent review about MetS suggests that IGF-1 acts as the keystone maintaining homeostasis in this system [9].

In summary, PCOS is a gynecological condition in which etiopathogenesis is not fully understood. Until now, therapeutic standard strategies for PCOS have been focused on hirsutism treatment and ovulation restoration. However, it should be taken more into account the prevalence of hyperinsulinemia and insulin resistance, which are often involved in the pathogenesis of this syndrome, in order to establish a better therapeutic strategy for PCOS.