Polycystic ovary syndrome (PCOS) is a commonly seen endocrine disturbance which is the major cause of anovulatory infertility and affects up to 15% of women of reproduction age. It is characterized by androgenism, ovulatory dysfunction, obesity, menstrual dysfunction, metabolic and psychiatric abnormalities [1, 2]. Hyperandrogenemia is present in 15%–45% of patients with PCOS [3], such as acne, hairiness, and increased level of free testosterone in peripheral blood. Moreover, it is known that women with PCOS are more likely to suffer from psychological problems, especially anxiety and depression [4].

The pathogenesis of PCOS remains unclear. Clinical observations and animal experiment data suggest the assumption that PCOS is inherited and induced by developmental programming of normal genetic mutations. Those genes would be magnified by exposing to in-utero androgen and stimulated by various post-natal life-style and environment factors. Chemistry substances that disrupt endocrine harbor the possibility to affect the developmental programming of PCOS susceptible genes [5].

Depression is a more and more commonly seen disease which constitutes remarkable health-care challenge. In 2008, WHO considered severe depression the 3rd cause of disease burden across the globe and forecasted that the problem will rank first by 2030 [6]. Depression influences females more than males [7] and is the most commonly seen psychological issue for females with PCOS. Females with PCOS present an 8-fold greater incidence of depression in contrast to females with no PCOS [8]. The different pathophysiologic causal links inducing depression are insulin resistance (IR), disturbances in the hypothalamic pituitary adrenal (HPA) axis and hyperandrogenism [2].

Although previous studies provided a suggestive link between PCOS and the risk of depression, residual confounding and reverse causation are difficult to eliminate in observational studies. Recently, there is an alternative method to investigate the potential causal association unbiasedly, two-sample Mendelian randomization, which depends on gene mutations as instrumental variables (IVs) to assess the causal association between an exposure and a result [9]. After random location in the process of meiosis, single-nucleotide polymorphisms (SNPs) used as instrumental variables (IVs) remains stable and non-modified throughout a life-time of environment exposure, which makes them independent of confounding factors or reverse causation [10].

Based on gene mutations, Mendelian randomization (MR)-offered proofs of PCOS-related causality pertaining to several physical disorders have been found [11]. Moreover, previous Mendelian randomized study also explored the causal associations between PCOS and psychiatrical disorder and discovered that PCOS increased the risk of obsessive-compulsive disorder, but was not associated with the other four psychiatric disorders (anxiety disorders, bipolar disorders, severe depression disorders, or schizophrenia) [12]. On the basis of the release of a new database with a larger sample size, we conducted this updated bidirectional Mendelian randomized analyses to assess whether the underlying causality between PCOS and depression could have different outcomes.

Herein, our team completed a two-sample MR and identified a potential causal association of PCOS with depression via the biggest GWAS data set to date. This study suggested that PCOS might increase the risk of depression. In contrast, we did not find evidence that depression may increase the risk of PCOS.

Previously, observational researches have unveiled the underlying correlation between depression and PCOS. Açmaz et al. [23] discovered that PCOS group (n = 86) with infertility had higher depression scores. Altinkaya et al. [24] found that Beck Depression Inventory scoring was remarkably greater in patients with PCOS (n = 83) in contrast to normal controls [24]. An Iranian case control study found that there existed a remarkable diversity between PCOS (n = 742) and controls (n = 798) in depression (18.9% vs. 7.9%; p 0.001) [25]. A cross-sectional study showed that females with PCOS displayed severer anxiety (p = 0.007) and depression (p = 0.048) in contrast to females without PCOS [26]. A study in Southwest China involved 120 out-patients with PCOS and 100 normal controls showed that the prevalence of depression (27.5% vs. 3.0%) was greater in PCOS sufferers in contrast to controls (p 0.05) [27]. Coherent with those aforesaid observation researches, data from this study herein backed the assumption that PCOS might be related to an elevated risk of depression.

Despite the fact that the biofunction of PCOS in the depression progression remains elusive, some researches have offered reasonable elucidation in this regard, one of which is hyperandrogenism [28, 29]. Second, a positive association between insulin resistance and depression was found [30, 31]. The randomized control trial (RCT) of Greenwood et al. [32] (738 PCOS females) potently reveals that insulin resistance is a causation factor for PCOS-related depression. In addition, PCOS is considered a proinflammation status featured by elevated contents of proinflammation biomarkers. Hence, there exists a probability of an inflammation association between depression and PCOS [33]. It is probable that the inflammation biomarkers in PCOS can cross the blood-brain barrier, inducing the progression of depression [33]. While the above theory-wise elucidation is reasonable, more researches are needed to reveal the potential causal link between PCOS and depression.

The major contribution of the present research is the utilization of the MR method, which has been extensively utilized to explore the causality of PCOS with the risks of other diseases. In addition, our team obtained the SNPs of depression and PCOS via the biggest GWAS datasets to date. The SNPs herein were remarkably related to PCOS at genome-wide significance, hence decreasing potential breach of the first hypothesis of MR. In addition, the F-statistics for the IVs all satisfied the liminal value of F-statistics $>$10, revealing that the analyses were not likely to be influenced by weak instrument bias [34]. The MR-Egger regression analyses were completed to study the data and did not observe the existence of directional pleiotropy [22]. Those coherent outcomes revealed the stability of the discoveries in the present paper.

Nonetheless, there are certain limitations of our research. Firstly, as our analyses were limited by European individuals, the results might not be extended to other races. Nevertheless, this remarkably decreased the underlying influences of population stratification bias as well. Secondly, while the MR method may offer a non-biased outcome because of the diminished confounders, the gene–milieu and gene–gene interplay might influence the progression of depression or PCOS inevitably. Thirdly, the MR analyses of PCOS and depression were on the foundation of summary statistics with comparatively smaller sample size, and the underlying side effects of PCOS on the risks of depression ought to be further investigated in bigger sample size. In addition, as the related data of all PCOS phenotypes were unavailable, our team merely investigated the correlation between PCOS and depression, and our team did not stratify the outcomes herein as per the diverse PCOS phenotypes. More researches highlighting the correlation of PCOS with depression are still needed.

To sum up, the present research offered evidence to suggest potential causality between PCOS and an elevated risk of depression amongst European individuals. Nevertheless, the accurate roles and the potential biology processes of PCOS in the progression of depression require deeper explorations.

FQ and YF designed the study. XZ and YT conducted data collection. XZ and YG conducted data analysis. XZ and MD carried out table arrangement and picture drawing. YM conducted the supervision of data analysis. YF conducted supervision. XZ and YT wrote the original draft. YF and FQ reviewed the manuscript. All authors contributed to editorial changes in the manuscript. All authors read and approved the final manuscript.

Not applicable.

Not applicable.

This research was funded by National Natural Science Foundation of China (No. 82074476 and 81874480).

The authors declare no conflict of interest. FQ is serving as one of the Guest editors of this journal. We declare that FQ had no involvement in the peer review of this article and has no access to information regarding its peer review. Full responsibility for the editorial process for this article was delegated to MHD.