The major finding of this study is the demonstration that lower HMW adiponectin and total adiponectin levels in FF and serum was an important feature in gonadotropin stimulated PCOS patients. Decreased HMW adiponectin levels in the FF was associated with decreased larger follicular diameters in normovulatory and PCOS women under IVF, and this association is independent of overall adiposity. The HMW adiponectin provided a stronger relationship with larger follicular diameters than the total adiponectin among all the women under IVF. Our data suggested that IR, estimated by HOMA-IR, at least partially predicts a reflection of ovarian stimulation in IVF patients.
Findings from previous animal studies show that adiponectin provokes the expression of genes associated with peri-ovulatory remodelling of the ovarian follicle and may be involved in the regulation of follicle maturation and ovulation [12, 21, 22]. Multiple endocrine and intraovarian paracrine interactions may influence the intrafollicular microenvironment for appropriate oocyte development. The concentrations of adipokines in FF correlate with folliculogenesis and oocytes more than with implantation and embryonic development. Recent studies showed that endocrine/paracrine abnormalities and metabolic dysfunction impaired folliculogenesis in PCOS patients and determined the outcome of assisted reproductive technologies (ART) . The findings of our previous study suggested that it is possible that alterations in the adiponectin multimers may contribute to the phenotypic presentation of PCOS via metabolic pathways independent of IR . We suggested that the potential contributions of the alterations in the adiponectin multimers such as HMW adiponectin in FF are implicated in the pathogenesis of ovulatory dysfunction in PCOS women.
Our study directly compared the two compartments (serum and FF) in the same patients at the same time. In this study, we demonstrated for the first time that the levels of HMW adiponectin in FF were significantly lower in the women with PCOS undergoing IVF compared with the BMI-matched normovulatory women. The HMW adiponectin was approximately two times lower in the FF than in the serum (Figure 1). The follicular fluid HMW adiponectin levels, but not the total adiponectin levels, had a significant positive correlation with larger follicular diameters. These findings suggest that the change in an intrafollicular HMW adiponectin environment plays a potential role in folliculogenesis in women undergoing IVF. We did not find that the total adiponectin levels in FF have strong linear relationships with larger follicular diameters. This finding seems to suggest that intrafollicular HMW adiponectin may be more important than total adiponectin in predicting larger follicular diameters in PCOS and non-PCOS women undergoing IVF. Based on our data, causality cannot be established. There should be further investigation of the potential mechanisms of the intrafollicular HMW adiponectin levels that predicted a reflection of ovarian stimulation in IVF patients.
We hypothesize that the detected amount of HMW adiponectin, with its even lower FF-to-serum concentration ratio of 0.5 in the FF, may not be the result of passive diffusion but could be the combination effect of endocrine factors, including those of insulin and the gonadotropins. This finding is further strengthened by the result of this study , which was the presence of a high concentration of insulin in the FF in the non-pregnant cycles of patients with PCOS. It is regretted that we did not detect the concentration of insulin in the FF in the present study. The levels of HOMA-IR demonstrated a strong inverse linear relationship with the HMW adiponectin levels in the FF. The precise mechanisms are unclear, and the lower intrafollicular HMW adiponectin levels should be further investigated.
Previous studies that assessed the relationships of FF and serum total adiponectin levels with the IVF outcomes in women undergoing controlled ovarian hyperstimulation have yielded conflicting results [24–27]. Findings of Bersinger NA  showed higher levels of total adiponectin in serum on the day of oocyte pick-up (OPU) but not in the FF leading to pregnancy, compared with unsuccessful cycles. A recent study  showed that no correlations were observed for total adiponectin or its isoforms in the serum and the FF with estradiol, progesterone, anti-Mullerian hormone, inhibin B, or the total FSH dose administered during the ovarian stimulation phase. They observed a trend towards higher HMW adiponectin serum levels in successful ICSI cycles compared to implantation failure cycles. Takikawa et al.  reported that there was no significant difference in the concentration of adiponectin in the FF between pregnant and non-pregnant cycles; they did find high concentrations of insulin in the FF in non-pregnant cycles of patients with PCOS that suggested the possible involvement of intrafollicular insulin in folliculogenesis. This finding is in contrast to other studies, which found that the leptin to adiponectin ratio (L: A ratio) in the follicular fluid of the preovulatory follicle is related to successful in vitro embryo development and that this action may be independent of FF insulin .
Some findings of the present study contrast with the results of two of these reports [24, 26]. Several explanations for the alleged disparity in these findings can be hypothesised, including differences in the study subject demographics (specifically ethnicity and different PCOS criteria) and different ART and clinical characteristics (i.e., a higher prevalence of obesity in the present study). The indexes for the ART outcome and the methodology used to quantify the adiponectin multimers differed among the studies.
Several limitations of our study should be considered, and the primary weakness was the relatively small sample size. We based the diagnosis of PCOS on the 1990 NIH criteria, and thus the study represents a less heterogeneous group of women. The strengths of our study included its prospective design. All the samples were batched, frozen, and run together in triplicate to minimise error.
The current data demonstrate that low levels of total adiponectin and HMW adiponectin in the serum and FF of PCOS women undergoing IVF and intrafollicular HMW adiponectin had a significant positive correlation with larger follicular diameters. Our results support the possible involvement of intrafollicular HMW adiponectin in folliculogenesis. Further investigation is warranted to determine the mechanisms of the potential contributions of alterations in HMW adiponectin on ovulatory dysfunction in PCOS women.