Disorders of the Glucose Metabolism Correlate With the Phenotype and the Severity in Women With Polycystic Ovary Syndrome

Josef van Helden; Osman Evliyaoglu; Andreas Küberl; Ralf Weiskirchen


Clin Endocrinol. 2020;93(1):44-51. 

In This Article


In our study with 130 PCOS patients, we found three different states or disorders of the glucose metabolism. A total of 33 patients (IR−) showed no abnormalities, and 22 (IO) showed an insulin overstimulation 60 minutes after oral glucose administration of 75 g in normal basal insulin. The majority of patients (n = 75) had insulin resistance (IR+) according to the criteria described above and thus corresponds to the distribution found in other studies.[19,20] In these three subgroups, the different PCOS phenotypes were distributed differently.[7] The IR+ group showed the highest proportion of patients with chronic anovulation (88.0%). In the IO group, 72.7% were still ovulating and showed normal ovarian cyclic activity. In the group with normal glucose metabolism, the proportion of anovulatory patients was 45.5%. A similar distribution was found with hyperandrogenemia. In the IR+ group, almost every patient was affected by hyperandrogenemia and oligo/amenorrhoea and can therefore be assigned to phenotype A. In the IR− group, it was 66.7% and 45% oligo- or anovulation, so that here, in addition to phenotype A, types B and C also occurred. In the IO group, only 45.5% of the patients showed hyperandrogenemia, but most of the patients had normal cyclic ovarian activity, suggesting that in this subgroup, the phenotype C is prevalent. This finding marks a close correlation between the degree of glucose metabolism impairment and the clinical severity of PCOS. Between these subgroups, there was a significantly different distribution of the examined biomarkers. As expected, the IR+ group showed the highest values for BMI and glucose metabolism parameters such as proinsulin, C-peptide, glucose and insulin before and after exposure. The fact that the highest AMH values were measured in this group also confirms the results of previously published studies that examined a relationship between the level of AMH adjusted for the deviation from the respective year median, PCOS severity and the associated probability of conception.[21–23] Additionally, other studies demonstrated that there are different associations between AMH with clinical or biochemical characteristics between women with and without PCOS[24,25] but also with different phenotypes of PCOS.[26] Taken together, these results provide clear evidence that there is a link between glucose metabolism, PCOS phenotype and clinical severity. Therefore, it is worth looking for biomarkers that can be used to classify and stratify patients.

Adiponectin is decreased in women with PCOS.[14] The lowest adiponectin levels were measured in the subgroup with insulin resistance (IR+). It was suspected that the low adiponectin levels correlate with insulin resistance.[15] We were able to confirm this in our study. Morishita et al[27] have previously shown that successful treatment of insulin resistance in PCOS increases adiponectin levels. This is also an indication that adiponectin is a parameter suitable to measure the severity of PCOS. Numerous studies explain hypoadiponectinemia with an increased incidence of polymorphisms in the adiponectin gene[28] and have shown a significant correlation with the glucose insulin ratio.[29] We can also confirm this in our study. To what extent the adiponectin level also correlates with the pregnancy rate has to be determined in future studies with well-characterized patients with PCOS.

For the first time, however, it could also be shown that the lowest levels of the soluble AMH receptor type 2 were found in the IR+ group. sAMHR2 has not yet been used as a biomarker in patients with PCOS. There are only few studies that could show that this protein in soluble form is measurable in serum.[17,18] Here was shown for the first time that the highest sAMHR2 levels occur in patients with premature ovarian dysfunction or menopausal ovarian failure.[18] In PCOS patients, the average values are lower than in women with a normal cyclic ovarian activity. This difference did not appear to be significant at first. However, our study shows that there is a significant difference in the sAMHR2 level between the IR+ and the IO subgroup. If the quotient of sAMHR2 and AMH is formed, there are even significant differences between all three subgroups. Since the different PCOS phenotypes are represented differently in these subgroups, it could be concluded that the sAMHR2 or sAMHR2/AMH quotient gives an indication of the severity of the glucose metabolism impairment that is associated with the remaining ovarian function. We were able to define optimal cut-offs for these discussed biomarkers or their combinations using receiver-operating characteristics. These included the quotient of age and 60 minutes insulin and, as expected, the HOMA Index for insulin resistance, and the quotient of adiponectin and insulin for insulin overstimulation. Alternatively, the quotient of sAMHR2 and insulin can also be used. We therefore recommend the use of a glucose tolerance test with additional insulin determination and an adiponectin or sAMHR2 determination from the basal sample in patients with PCOS in order to create additional biochemical criteria for classifying the severity of a PCOS in addition to the morphological determination of the phenotype. The ultimate goal is to be able to make a prognosis regarding the possible onset of pregnancy in the patient. However, this requires further prospective studies with these biomarker combinations.