Article Text
Abstract
Background Diagnosis of polycystic ovary syndrome (PCOS) depends on 2003 Rotterdam Criteria. According to these criteria there are four possible combinations resulting in various phenotypes. We aimed (i)to confirm that the levels of body mass index (BMI), anti-müllerian hormone (AMH) levels and insulin resistance (IR) are higher in PCOS patients and higher in phenotype-A among PCOS patients, and (ii)to determine cut-off values for the diagnosis of PCOS and phenotype-A.
Materials and methods This study was conducted in an IVF Center, between November 2019 and January 2021. Data of infertile women participating in the study was evaluated retrospectively. Parameters such as menstruation pattern, clinical hyperandrogenism, age, BMI, follicle stimulating hormone (FSH), luteinizing hormone (LH), estradiol, thyroid stimulating hormone (TSH), prolactin, AMH, dehydroepiandrosterone sulphate (DHEA-S), fasting blood glucose, fasting insulin levels, antral follicle counts (AFC) and ovarian volumes were recorded for each patient. Women were grouped as PCOS and non-PCOS, and PCOS group was further divided into 4 sub-groups according to their phenotypes. Data of infertile patients with PCOS patients were compared with infertile non-PCOS patients and PCOS phenotypes were compared among each other.
Results Data of 244 infertile patients was included in the study. BMI, AMH, AFC, and HOMA-IR were statistically higher in PCOS patients, compared to non-PCOS patients. We found the AMH level of >3.105 ng/ml to be having 90.8% sensitivity and 90% specificity to diagnose a patient as PCOS. Among different phenotypes, also BMI, AMH, and insulin resistance index (HOMA-IR) levels were significantly higher in infertile PCOS phenotype-A when compared to other three phenotypes (p:0.003, p:0.000, and p:0.000, respectively). The AMH cut-off value to estimate phenotype-A was found as 6.095 ng/ml with 69.2% sensitivity and 86.7% specificity. We did not found threshold levels of BMI and HOMA-IR with high sensitivity to identify phenotype-A.
Conclusion Properly diagnosing PCOS and determining the phenotype are crucial due to the long-term health conditions. Therefore, we suggest that serum AMH level could be included in PCOS diagnosis criteria, and the value of 3.105 ng/ml would have a 90.8% sensitivity and 90% specificity. Also, to identify phenotype-A, AMH level could be used. Therefore, we speculate that AMH may serve to identify PCOS and PCOS phenotype-A in places where ultrasound imaging is not straightforward to perform or not easily accessible.
Highlights
Based on the standard diagnostic indexes of polycystic ovary syndrome (PCOS), our study investigated BMI, AMH and IR of four clinical phenotypes in infertile patients with PCOS.
The results showed that PCOS Phenotype-A was associated with obesity and insulin resistance (IR). Therefore, A-phenotype was considered to be the most severe type of metabolic abnormality.
AMH was the highest in the Phenotype-A.
BMI, AMH and IR of PCOS A-phenotype were significantly increased, and lifestyle changes such as healthy diet and exercise were recommended for these patients to improve the health problems caused by metabolic abnormalities.
We conclude that BMI, AMH and IR seem to be important parameters in diagnosing PCOS women and diagnostic criteria should be widened with these parameters.
- Polycystic ovarian syndrome
- Body mass index
- Anti-müllerian hormone
- Insulin resistance
- PCOS Phenotypes
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Statistics from Altmetric.com
1 Introduction
Polycystic ovary syndrome (PCOS) is a common cause of anovulatory infertility, and has a wide spectrum of clinical findings.1,2 The prevalence is around 6%–10% among women of reproductive age.3 As PCOS is associated with many comorbidities (e.g., metabolic syndrome, depression, obstructive sleep apnea, endometrial cancer), diagnosing it accurately is crucial for informing the patients about the potential health risks associated with it, for guiding the patients to make appropriate lifestyle changes, and for tailoring the therapy to the patients’ therapeutic goals.4–7
Today, most commonly used diagnostic criteria for PCOS is the Rotterdam criteria; oligo-amenorrhea (OM), hyperandrogenism (HA) and polycystic ovarian morphology (PCOM).8 After exclusion of other etiologies for hyperandrogenism or oligo-amenorrhea, at least 2 out of 3 criteria are required for diagnosis. Based on the Rotterdam criteria, 4 different PCOS phenotypes were identified. Phenotype-A is the most common type and has all three criteria: OM + HA + PCOM. Phenotype-B is characterized with presence of OM + HA, phenotype-C with HA + PCOM and phenotype-D with OM + PCOM.9
Anti-Müllerian hormone (AMH) levels are elevated in women with PCOS and are usually evaluated in PCOS patients with fertility problems as it is regarded as an indirect marker of ovarian reserve.10 Also, PCOS is frequently associated with metabolic abnormalities such as obesity and insulin resistance both of which have become health issues worldwide.11 Even though the current PCOS diagnostic criteria do not include Body Mass Index (BMI), AMH or Insulin Resistance (IR); they are closely interrelated with PCOS.12 Due to the correlations with antral follicle count, ovulatory dysfunction, hyperandrogenism and AMH levels, there is an ongoing debate about the use of AMH as a diagnostic tool for PCOS.13–16 In addition, among PCOS patients, phenotype-A is associated with higher BMI, higher AMH levels and higher IR.17–19
In our study, our aim was (i)to confirm that the levels of BMI, AMH and HOMA-IR are higher in PCOS patients and higher in phenotype-A among PCOS patients, and (ii)to determine cut-off values for the diagnosis of PCOS and phenotype-A. Hence, such cut-off values can be incorporated into PCOS diagnosis criteria. This would be helpful since it might be challenging to perform ultrasound in certain conditions and regions in order to diagnose PCOS.
2 Material and methods
This study was conducted in an IVF Center in Northern Cyprus between November 2019 and January 2021. Data of infertile women was evaluated retrospectively. All patients were Turkish. All procedures were routine standardized procedures performed in the center.
Local Institutional Review Board approval was obtained from Yakin Dogu University (YDU) Ethical Committee with (No:YDU/2021/97–1433).
On the 2nd, 3rd, or 4th day of the cycle, routinely, menstruation pattern, clinical hyperandrogenism, age (years), BMI (kg/m2), follicle stimulating hormone (FSH), luteinizing hormone (LH), estradiol (E2), thyroid stimulating hormone (TSH), prolactin (PRL), AMH, dehydroepiandrosterone sulphate (DHEA-S), fasting blood glucose, fasting insulin levels were recorded. Also via transvaginal ultrasound (DC 60 Exp 2018, Mindray, China), antral follicle counts (AFC) and ovarian volumes on cycle day 2, 3, or 4 were recorded for each patient.20 LH/FSH ratio and HOMA-IR values were calculated for this study from the blood tests that were already taken.21
Rotterdam criteria was used for PCOS diagnosis. Menstruation pattern were classified as having oligo-amenorrhea or not. Oligomenorrhea was defined as the frequency of menstruation to be less than once every 35 days or 8 cycles a year and amenorrhea was defined as no menstruation for over 3 months in women who had previous cycles.8 Clinical hyperandrogenism was diagnosed if a woman had alopecia, acne or hirsutism according to modified Ferriman-Gallwey (mF-G) scores of over 8.22 PCOM is defined as either having ≥20 antral follicles in at least one ovary and/or ovarian volume ≥10 ml at least in one ovary with transvaginal ultrasound.23 Women with PCOS diagnosis was further divided into 4 phenotypes.9
Any infertile women with hyperprolactinemia, thyroid dysfunction, adrenal dysfunction, history of any ovarian surgery, any cystic or solid mass in ovaries on 2nd - 4th days of menstruation detected with USG were excluded from the study. Patients with any missing data were excluded from the study. Flowchart indicating included and excluded patients is shown in Fig. 1.
In total, 244 infertile patients were included in the analysis. There were 174 infertile PCOS patients and 70 infertile non-PCOS patients. Among the PCOS group, 78 were phenotype-A, 42 were phenotype-B, 23 were phenotype-C, and 31 were phenotype-D.
In our study, we aimed to compare BMI, AMH and HOMA-IR values in infertile women with and without PCOS. Infertile women with PCOS were sub-grouped into 4 phenotypes. We also compared BMI, AMH and HOMA-IR values among these four PCOS phenotypes and with non-PCOS patients.
ROC analysis for AMH was done to reveal the best cut-off value for diagnosing PCOS among all 244 patients. ROC analyses for AMH, BMI and HOMA-IR were done to reveal the best cut-off values for diagnosing phenotype-A among all 174 PCOS patients.
3 Statistical analysis
The statistical analysis was done using the SPSS 20 (Statistical Program for Social Sciences, IBM, Chicago, IL). Demographic continuous data were characterized by means and standard deviations and nominal variables were expressed as percentages. Kolmogorov-Smirnov test was used for assessing normality. Parametric tests were used when the variables were normally distributed. Differences in mean values were analyzed with independent t-tests. One-way analysis of variance (ANOVA) used for comparing continuous normally distributed numeric variables between groups. To compare parameters of 4 phenotypes; Kruskal Wallis test was used. Mann-Whitney U test was used to compare more than two phenotypes. Receiver operating characteristic (ROC) curve analysis was performed for determining best cut-off points for detected significant variables to distinguish PCOS among all women for AMH levels. ROC analyses performed for determining best cut-off values of AMH, BMI and HOMA-IR for diagnosing phenotype-A among all 174 PCOS patients. Area under curve (AUC) was calculated for estimation. A p value < 0,05 was considered as statistically significant. For determining best cut-off value for a parameter to distinguish PCOS from all 244 patients or to distinguish Phenotype-A form all PCOS patients; sensitivity and 1-specificity values of the parameter were copied to Excel table from SPSS. On the Excel page, Youden indexes (YI) were calculated using the formula “(sensitivity + specificity) – 1 = YI”. Highest YI value was found, and the corresponding value was considered as the cut-off value; which had highest sensitivity and specificity.
4 Results
Among 244 infertile patients, 174 were PCOS and 70 were non-PCOS. 78 (44.8%) infertile PCOS patients were classified as phenotype-A, 42 (24.1%) were phenotype-B, 23 (13.2%) were phenotype-C, and 31 (17.8%) were phenotype-D (Table 1).
The data of PCOS and non-PCOS infertile patients were compared. The mean age was 27.66 ± 4.68 and 27.62 ± 2.99 in infertile PCOS and non-PCOS patients respectively (p = nonsignificant (NS)). E2, TSH, PRL, fasting glucose and fasting insulin levels were similar between two groups (p = NS). BMI was significantly higher in PCOS patients than non-PCOS patients (p = 0.000). AFC and AMH were significantly higher in PCOS patients than non-PCOS patients (p = 0.000). Although FSH and LH levels were not statistically different between PCOS and non-PCOS patients, LH/FSH ratio was significantly higher in PCOS patients (p = 0.001). DHEA-S was significantly higher in PCOS patients than non-PCOS patients (p = 0.041). HOMA-IR levels were statistically significantly higher in PCOS patients when compared to non-PCOS patients (p = 0.015) (Table 2).
ROC analysis for AMH was done to reveal the best cut-off value for diagnosing PCOS among all patients. An AMH level ≥3.105 ng/ml was found as a cut-off value to diagnose a patient as PCOS; with 90.8% sensitivity and 90% specificity (AUC: 0.947 (p < 0.005, 95% CI: 0.919–0.974)) (Figure-2).
Infertile PCOS patients were classified into 4 phenotypes. Age, E2, FSH, PRL, TSH, fasting glucose levels were similar between 4 phenotypes (p = NS). Among 4 phenotypes; BMI, AMH, LH, LH/FSH, DHEA-S, AFC, fasting insulin and HOMA-IR levels were significantly higher in phenotype- A patients. (p = 0.003, p = 0.000, p = 0.006, p = 0.000, p = 0.000, p = 0.000, p = 0.000 and p = 0.000, respectively) (Table-3).
Comparison of BMI levels among 4 PCOS Phenotypes is given in Figure-3, where phenotype-A had highest values (Figure-3). To estimate best BMI cutoff value among PCOS phenotypes to distinguish phenotype-A, ROC analysis done and best cutoff value of BMI to estimate phenotype-A was found as 26.5 kg/m2, with 41% sensitivity and 92.2% specificity. (AUC: 0.686, 95% CI 0.611–0.765, p = 0.000) (Figure-4).
Comparison of AMH levels among 4 phenotypes is given in Fig. 5, levels were higher in phenotype-A compared to other phenotypes (Figure-5). Best cutoff value of AMH to estimate phenotype-A was found as 6.095 ng/ml with 69.2% sensitivity and 86.7% specificity (AUC: 0.857; 95% CI 0.809–0.905, p = 0.000) (Figure-6).
Comparison of HOMA-IR levels revealed that phenotype-A had significantly higher values among all PCOS (Figure-7). To estimate best HOMA-IR cutoff value among PCOS phenotypes to distinguish phenotype-A, ROC analysis was done and levels above 2.218 with 46.2% sensitivity and 84.3% specificity was considered as best cutoff (AUC: 0.683; 95% CI 0.558–0.717, p = 0.001) (Figure-8).
5 Discussion
In our study, PCOS patients and phenotype-A was associated with significantly higher BMI, AMH levels, and HOMA-IR compared to non-PCOS and other PCOS phenotypes, respectively. The most common infertile PCOS phenotype was found to be Phenotype-A as in other studies in the literature so far.12,18,24
Our study revealed that AMH levels above 3.105 ng/ml was significantly related with PCOS with high sensitivity and specificity (90.8% and 90%, respectively). We also analyzed 4 PCOS phenotypes to find the best cut-off values for AMH, BMI and HOMA-IR to estimate Phenotype-A. We found that; AMH levels of >6.095 ng/ml is having 69.2% sensitivity and 86.7% specificity for identifying phenotype-A. The cut-off values of BMI and HOMA-IR for PCOS phenotype-A were found to be having low sensitivity.
In 2009, Piouka et al. found that, AMH levels reflected the severity of PCOS.25 Sahmay et al. analyzed results of 251 women with PCOS, and revealed that AMH levels were different among phenotypes and significantly higher in phenotype-A.26 Sova et al. analyzed AMH levels among 319 PCOS patients and 109 controls, measured higher AMH levels in PCOS women than controls, and AMH levels were highest in phenotype-A when compared to other phenotypes.18 Jamil et al. revealed that AMH levels were higher in PCOS phenotype-A than controls.19 Ozay et al. studied AMH among PCOS phenotypes and found significantly higher AMH levels in patients with phenotype-A than other phenotypes.27 Similarly, we found out that phenotype-A subgroup had the highest AMH levels.
Bozdağ et al. screened 392 women for determining diagnostic levels of AMH and revealed that phenotype-A PCOS women had significantly high AMH levels and optimal AMH threshold was found to be 4.86 ng/ml for the diagnosis of PCOS.28 Dewailly et al. proposed a simplified PCOS diagnosis based on AMH threshold of 5 ng/ml.29 In our study, we determined the value of 3.105 ng/ml of AMH for PCOS diagnosis had 90.8% sensitivity and 90% specificity.
In the literature, we did not find any study to evaluate the value of AMH to determine phenotype-A. In our study, we determined the cut-off value of 6.095 ng/ml of AMH for phenotype-A with 69.2% sensitivity and 86.7% specificity. This cut-off value can be used in identifying PCOS phenotype-A, since it might be not straightforward to perform ultrasound and evaluate ovaries in virgin, obese patients or in areas where ultrasound is not easily accessible. And since women with PCOS phenotype-A should be counselled for lifelong metabolic effects and lifestyle modifications with healthy dietary habits and regular exercise should be strongly recommended, performing such a quantitative test might be more reliable and accessible compared to performing transvaginal ultrasound.
Obesity and IR are interrelated parameters in the pathophysiology of PCOS.12,17 PCOS phenotype-A is mostly related with obesity and IR, and it is considered as the most severe PCOS type.30 Phenotype-B has similar but lighter metabolic outcomes. Phenotype-C is milder since the predominant problem occurs as subfertility. Phenotype-D is the mildest form where obesity, IR or metabolic adverse effects seem less often.31 PCOS phenotypes A and B have more tendency for obesity and insulin resistance.17 We found out that phenotype-A had higher BMI and HOMA-IR values compared to other phenotypes, in our study. However, we couldn't determine cut-off values to identify different phenotypes with high sensitivity.
A limitation of our study was its retrospective design and also the study group was limited to infertile patients. Prospective studies comparing prognostic factors between all 4 phenotypes in women with and without fertility problems could be performed in the future. Also, this study population was carried out in Turkish patients. Therefore, a large study including different ethnicities are needed in order to generalize these results.
In our study, most common PCOS phenotype was phenotype-A. BMI, AMH and IR parameters were found to be significantly higher in phenotype-A compared to other infertile PCOS phenotypes and also when compared to infertile non-PCOS patients. Women with PCOS phenotype-A should be counselled for lifelong metabolic effects and lifestyle modifications with healthy dietary habits and regular exercise should be strongly recommended. We conclude that BMI, AMH and IR seem to be important in distinguishing women with PCOS, and diagnostic criteria should be reviewed and widened with such parameters accordingly since obesity is a global health issue. AMH levels can be helpful in diagnosing PCOS and identifying PCOS phenotypes, since it might be not straightforward to perform ultrasound and evaluate ovaries in virgin, obese patients or in areas where ultrasound is not easily accessible. In addition, even though BMI and HOMA-IR are interrelated with PCOS phenotype-A, we did not find threshold levels of them with high sensitivity.
6 Conclusion
Properly diagnosing PCOS and determining the phenotype are crucial due to the long-term health conditions. Therefore, we suggest that serum AMH level could be included in PCOS diagnosis criteria, and the value of 3.105 ng/ml would have a 90.8% sensitivity and 90% specificity. Also, to identify phenotype-A, AMH level could be used. Therefore, we speculate that AMH may serve to identify PCOS and PCOS phenotype-A in places where ultrasound imaging is not straightforward to perform or not easily accessible.
Ethics approval
Local Institutional Review Board approval was obtained from Yakin Dogu University (YDU) Ethical Committee with YDU/2021/97–1433 registration number. This study was performed in accordance with ethical standards as settled down in the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards.
Funding
No funding was received for this article.
Consent to participate (from patients)
The patients gave informed consent and the patient anonymity preserved.
Consent for publication (all authors)
All authors of the original manuscript have read and approved the manuscript. We confirm that the inclusion of each author in the authorship list is based only 1) On substantial contributions to a) concept and design or analysis, and interpretation of data and b) drafting the manuscript or revising it critically for important intellectual content and 2) Final approval by each author of the version of the manuscript.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.