Predictive markers of endometriosis: a future perspective

Abstract

Endometriosis is a chronic gynecological disorder affecting up to 10% of women of reproductive age and is characterized by the presence of endometrial-like tissue outside the uterus. The condition is associated with chronic pelvic pain, dysmenorrhea, dyspareunia, and infertility, and remains difficult to diagnose at early stages. Recent evidence suggests that endometriosis shares genetic and molecular pathways with other chronic pain and inflammatory disorders. Identifying reliable biomarkers is therefore essential to improve risk stratification and support earlier evaluation. Multiple domains have emerged as candidate markers for risk stratification, including genetic and epigenetic alterations, dysmenorrhea, migraine, autoimmune and endocrine disorders, and stress and early-life adversity. These factors have been associated with disease initiation, lesion development, and symptom severity. Understanding their interactions could support multimodal risk stratification approaches and guide preventive strategies. This narrative review highlights the multifactorial nature of endometriosis and synthesizes current evidence on emerging predictive markers. Although progress has been made, large prospective studies are still needed to validate these markers and facilitate the development of targeted interventions.

Introduction

Endometriosis is a chronic gynecological disorder affecting up to 10% of women of reproductive age [1] and is associated with pelvic pain, dysmenorrhea, dyspareunia, and infertility [2]. The burden on quality of life and healthcare systems is substantial, with annual costs comparable to other chronic diseases [3]. Despite its frequency, endometriosis remains underdiagnosed, partly because current diagnostic tools detect disease only after symptoms are established, often through invasive procedures [4–6]. An improved understanding of early biological and clinical markers is therefore essential. Recent evidence highlights shared genetic, inflammatory, hormonal, and neurobiological pathways with other chronic pain and immune-mediated disorders [7]. This supports the hypothesis that certain features, ranging from genetic or epigenetic alterations to migraine, dysmenorrhea, autoimmune conditions, or stress exposure, may function as markers associated with increased risk, rather than consequences of established disease.

This perspective aligns with the broader question of prevention. While endometriosis is not yet preventable, identifying subgroups at increased risk could enable earlier intervention, symptom monitoring, or lifestyle modification. Emerging tools such as polygenic risk scores and machine-learning-based screening models have demonstrated promise in risk stratification and screening-oriented assessment, underscoring the relevance of predictive approaches [8–10].

However, at the current stage, these markers should be viewed primarily as contributors to risk stratification rather than as validated tools for precise individual-level prediction. Their potential clinical value lies in identifying subgroups who may warrant closer monitoring or earlier evaluation, particularly when several clinical and biological risk domains converge.

The purpose of this narrative review is to synthesize current evidence on predictive markers of endometriosis, spanning genetic and epigenetic signatures, clinical symptoms, comorbidities, environmental exposures, and endocrine factors (Fig. 1). By integrating these domains, the review aims to clarify how risk stratification could contribute to future prevention strategies and ultimately improve patient management.

Fig. 1.

Landscape of predictive markers for endometriosis. This schematic illustrates the different domains of markers that may contribute to risk prediction in endometriosis. Each sector represents a family of markers, genetic, epigenetic, clinical symptoms, autoimmune comorbidities, endocrine disruptors, and stress or early-life adversity, with indications of level of evidence, predictive potential, and ease of measurement. No single marker is sufficient on its own, but the combination of multiple domains may support early identification and risk stratification. This figure is conceptual and intended to summarize current evidence and a potential multimodal framework; it is not a validated clinical prediction tool and is not intended for direct clinical decision-making

Genetic markers

Genetic studies have consistently shown that endometriosis is a heritable disorder, with heritability estimates of approximately 50–60% [11, 12]. Genome-wide association studies (GWAS) have identified dozens of risk loci, many of which involve genes implicated in hormone metabolism, inflammation, and neuronal pain pathways. These findings provide a biological rationale for prediction: genetic variants each confer a small increase in risk individually, but their combined effect in polygenic risk scores (PRS) can stratify susceptibility at the population level. Several genes have repeatedly emerged across studies. Variants in ESR1 (estrogen receptor alpha) and FSHB (follicle-stimulating hormone beta) highlight the crucial role of estrogen signaling in disease initiation and progression. Genes involved in cell adhesion and extracellular matrix remodeling, such as WNT4 and VEZT, support mechanisms related to implantation and invasion of ectopic tissue. Immune-related genes, including PTPN22 and selected HLA alleles, suggest that dysregulated immune surveillance may facilitate lesion survival and growth. Importantly, recent large-scale meta-analyses have identified novel risk loci for endometriosis, some shared with chronic inflammatory and pain conditions [11, 12].

The predictive value of genetic markers is strengthened by data on comorbidities sharing biological pathways. Epidemiological studies show an increased prevalence of endometriosis among women with asthma [13, 14] and other immune-mediated conditions [15]. Shared risk variants have been reported between endometriosis and rheumatoid arthritis, although findings with osteoarthritis are less consistent and require further investigation. The overlap suggests that common genetic mechanisms related to hormonal regulation, inflammation, and neuronal signaling may contribute to multiple phenotypes [16, 17].

From a prediction standpoint, these findings are particularly relevant. Polygenic risk models may help stratify genetic susceptibility at the population level, especially when combined with clinical features such as severe dysmenorrhea or migraine. Machine-learning approaches integrating genetic, clinical, and imaging data have already demonstrated high screening accuracy in early studies. As these models mature, they may support risk stratification, earlier diagnosis, and targeted prevention strategies. Genetic markers do not act in isolation, but they provide a foundational layer for risk prediction. Their integration into polygenic scores and multimodal prediction models represents a promising path toward identifying higher risk subgroups who may warrant closer clinical attention.

Epigenetic markers

Epigenetic mechanisms regulate gene expression without altering the underlying DNA sequence. They include DNA methylation, histone modifications, and non-coding RNAs, and they provide a dynamic interface between genetic predisposition and environmental exposures [18]. Epigenetic changes are highly relevant to endometriosis because they influence cell proliferation, inflammation, hormone metabolism, and pain sensitivity [19]. A proposed theory suggests that a combination of genetic and epigenetic factors inherited at birth may explain the hereditary predisposition to endometriosis [20].

Several specific epigenetic alterations have been described in endometriotic tissue. Aberrant DNA methylation of genes involved in steroidogenesis (e.g., ESR2, CYP19A1) and inflammatory pathways (e.g., HOXA10) has been reported, leading to dysregulated estrogen signaling and impaired endometrial receptivity [21]. Increased activity of DNA methyltransferases (DNMTs) is frequently observed in ectopic lesions, supporting a role for transcriptional silencing in disease persistence [22]. In parallel, changes in histone acetylation and methylation affect chromatin structure and gene accessibility. For example, altered expression of histone deacetylases (HDACs) contributes to progesterone resistance and increased cell proliferation [23].

Non-coding RNAs also play an important role. Multiple microRNAs (miRNAs) are differentially expressed in endometriosis, including the miR-200 family, miR-135a/b, and miR-451, which regulate angiogenesis, immune response, and cell migration [24]. Several circulating miRNAs have been proposed as non-invasive biomarkers, detectable in blood or peritoneal fluid, although validation in large cohorts is still needed [25].

These observations have several implications for prediction. Epigenetic alterations occur early in the disease process and are influenced by environmental factors such as inflammation, endocrine disruptors, or early-life stress [24]. Epigenetic signatures could therefore contribute to risk stratification before lesions become clinically detectable. Recent work has shown that combining methylation profiles with clinical features improves diagnostic accuracy and may help predict the severity of symptoms or response to therapy [23]. In addition, the reversibility of epigenetic mechanisms makes them attractive targets for preventive strategies and therapeutic interventions. HDAC inhibitors and other epigenetic drugs have shown promising effects in preclinical models by reducing lesion growth and inflammation [23]. Epigenetic markers bridge genetic susceptibility with environmental influences. Specific alterations in DNA methylation, histone modifications, and miRNA expression may serve as candidate early markers associated with disease risk and contribute to future predictive models aimed at earlier detection and better management of endometriosis.

Dysmenorrhea

Dysmenorrhea is highly prevalent among women of reproductive age and is one of the most common early symptoms [26] that may precede an eventual diagnosis of endometriosis [27]. While primary dysmenorrhea arises without identifiable pelvic pathology, secondary dysmenorrhea is frequently associated with underlying gynecological conditions, among which endometriosis is a leading cause [27]. Women with endometriosis often experience more severe menstrual pain and a higher symptom burden than those without the disease [28].

From a predictive perspective, severe dysmenorrhea in adolescence and young adulthood may represent an early red flag. Multiple longitudinal studies suggest that intense menstrual pain beginning soon after menarche is associated with an increased risk of later endometriosis diagnosis [29]. Because many adolescents and young women consider menstrual pain “normal” and do not seek medical care, opportunities for early identification are often missed [30].

Inflammation plays a key role in the pain associated with dysmenorrhea and endometriosis. Elevated prostaglandin production during menstruation contributes to uterine hypercontractility and ischemia [31], and similar inflammatory mediators are found in peritoneal fluid and ectopic lesions in endometriosis [32, 33]. This shared inflammatory environment may explain why dysmenorrhea and chronic pelvic pain frequently coexist, and why dysmenorrhea can precede other endometriosis symptoms.

Central sensitization provides an additional explanatory link. Persistent nociceptive input can induce structural and functional changes in the central nervous system, lowering pain thresholds and amplifying pain perception [32, 34]. Altered brain metabolism and hypothalamic–pituitary–adrenal axis activity have been documented both in women with primary dysmenorrhea and in those with endometriosis [35]. These changes may make some individuals particularly vulnerable to chronic pelvic pain conditions.

Given these observations, monitoring the severity and persistence of dysmenorrhea, especially during adolescence, could improve risk stratification. Recurrent, intense menstrual pain that interferes with daily activities warrants evaluation for potential secondary causes. Integrating symptom severity with genetic, hormonal, or inflammatory markers may support future prediction models and facilitate earlier referral or intervention (Fig. 2). Dysmenorrhea represents both a common symptom and a potential predictive marker of endometriosis risk. Recognizing severe or early-onset dysmenorrhea as a clinical warning sign may help reduce diagnostic delay and enable more timely management.

Fig. 2.

Clinical trajectory before endometriosis diagnosis. This infographic illustrates the typical clinical trajectory before endometriosis is diagnosed. Early symptoms such as severe dysmenorrhea, migraine and pelvic pain often emerge soon after menarche but are frequently normalized or under-recognized. Repeated consultations in primary care, symptomatic treatment, and normal imaging results contribute to a diagnostic delay of 6–10 years. Recognizing early symptom clusters may provide a window for prediction and earlier evaluation, potentially reducing diagnostic delay. This figure is conceptual and intended to summarize current evidence and a potential multimodal framework; it is not a validated clinical prediction tool and is not intended for direct clinical decision-making

Migraine and endometriosis

Migraine is more prevalent in women with endometriosis than in the general female population, and this association has been observed across age groups, including adolescents [36, 37]. Although the causal direction remains uncertain, several shared biological mechanisms have been proposed. Hormonal fluctuations represent a key link. Both migraine and endometriosis are estrogen-responsive conditions. Estrogen variation can trigger migraine attacks [36, 37] and is involved in lesion growth and inflammation in endometriosis [38]. Excessive estrogen biosynthesis and estrogen-dependent inflammation may therefore contribute to the comorbidity between the two disorders. Inflammation and neuronal sensitization provide a second pathway. Endometriotic lesions induce a local inflammatory response and release pro-inflammatory mediators, which may sensitize peripheral and central neural circuits [38]. Experimental and clinical studies suggest that sensory fibers near ectopic implants can generate hyperexcitability and promote headache or migraine episodes [39]. A shared inflammatory genetic background has also been reported, with overlapping risk loci between migraine and endometriosis [12, 40].

The autonomic nervous system may further contribute. Migraine frequently co-occurs with other pain and hypersensitivity conditions, such as interstitial cystitis, fibromyalgia, and irritable bowel syndrome [41]. These comorbidities are also often present in women with endometriosis, suggesting a common neuroimmune phenotype. From a clinical perspective, migraine alone is not a diagnostic marker of endometriosis, but it may have predictive value when combined with other features, particularly severe dysmenorrhea or chronic pelvic pain [38]. Women presenting with both migraine and intense menstrual pain could represent a subgroup with increased risk and may benefit from earlier evaluation. Migraine and endometriosis share hormonal, inflammatory, and neurobiological pathways. While migraine is not specific enough to serve as a standalone biomarker, its presence, especially in the context of dysmenorrhea, may contribute to risk stratification and earlier suspicion of endometriosis [42].

Stress in endometriosis

Stress is frequently reported by women with endometriosis, but the nature of this relationship remains complex. Ethical limitations prevent controlled studies on untreated symptomatic patients, so it is challenging to determine whether stress contributes to the development of endometriosis or whether it primarily reflects the burden of chronic pain and disease [43]. Current evidence should therefore be interpreted with caution and does not support a direct causal role.

Several findings nonetheless suggest that stressful exposures may influence disease expression. Perinatal and early-life stressors, such as maternal smoking or birth complications, have been associated with later diagnosis of endometriosis [44, 45]. These exposures may affect endocrine or immune function and could modulate susceptibility, although the underlying mechanisms remain unclear. Similar associations with early-life events are discussed in the next section and should not be interpreted as proof of causation.

Stress is also relevant to symptom severity and disease experience. Persistent pain and inflammation may activate the hypothalamic–pituitary–adrenal (HPA) axis, potentially leading to altered cortisol responses and increased pain perception [46]. Women with endometriosis frequently report higher levels of perceived stress, and these levels often correlate with pain intensity and impairment in daily activities [43]. Immune-mediated comorbidities, including inflammatory bowel disease, allergic conditions and autoimmune disorders [13], are more common in endometriosis [47]. These shared associations suggest that neuroendocrine–immune interactions may contribute to symptom chronicity.

Animal studies provide additional insight into biological pathways. In preclinical models, exposure to stressors such as immobilization or predator scent aggravated lesion development and inflammatory parameters [48]. Conversely, stress reduction and β-adrenergic blockade improved outcomes in rodents [49]. These models indicate that stress may modulate disease progression, rather than initiate it. Clinically, stress should be viewed as a modifier of symptoms rather than a primary cause. Stress-reducing strategies, including behavioral therapy, physical activity, or medical treatment for anxiety, may help improve well-being. Whether stress levels could serve as a predictive marker for endometriosis is unclear, but monitoring stress in combination with pain severity may assist in risk stratification and symptom management.

Women with endometriosis often experience high stress levels, but current evidence does not prove that stress causes the disease. Instead, stress appears to interact with pain, inflammation, and neuroendocrine pathways, potentially influencing severity, quality of life, and clinical course.

Early-life abuse and risk of endometriosis

The relationship between early-life abuse and endometriosis has been explored in several studies, but results are mixed and should be interpreted cautiously. Some clinical and epidemiological studies report that adverse childhood experiences, including physical, emotional, or sexual abuse, are associated with an increased risk of chronic pelvic pain conditions, including endometriosis [50]. Case reports also suggest possible links between traumatic experiences and symptom severity [51]. However, other work has not confirmed these associations, particularly when examining surgical cohorts, where no significant relationship was found between childhood physical abuse and laparoscopically confirmed endometriosis [52].

Part of this discrepancy may relate to differences in symptom presentation. Endometriosis can be diagnosed through two predominant routes:

• pain-driven evaluation, and

• infertility-driven evaluation.

Studies show that among women who undergo laparoscopy for pain, those reporting childhood or adolescent abuse are more likely to receive a diagnosis of endometriosis [53]. By contrast, among women investigated primarily for infertility, associations with sexual abuse are not observed, and only small increases in risk are reported with physical or cumulative abuse [54]. These subgroup patterns suggest that early-life adversity may influence pain sensitivity and symptom reporting, rather than lesion formation per se.

Mechanistically, childhood stress and trauma can produce long-lasting alterations in neuroendocrine regulation. Severe abuse or neglect has been linked to hyperactivation of the hypothalamic–pituitary–adrenal (HPA) axis, altered cortisol secretion, and persistent changes in brain regions involved in pain and emotional processing [55]. These neurobiological consequences may lower pain thresholds, amplify nociception, and contribute to the chronic pain phenotype observed in many women with endometriosis. Importantly, this mechanism is consistent with observations in other chronic pain disorders and does not imply that abuse causes endometriosis lesions.

Taken together, current evidence suggests that early-life abuse may be related to symptom severity and the likelihood of pain-driven diagnosis, rather than to the biological development of endometriotic lesions. Abuse could increase vulnerability to pelvic pain through neuroendocrine and central sensitization pathways, thereby influencing when and how women present for medical care. Although early-life abuse is not a predictive marker for endometriosis on its own, a history of trauma may help identify patients at higher risk of chronic pain and may guide supportive management strategies.

Adverse childhood experiences may modulate the clinical expression of endometriosis through neuroendocrine and pain-processing pathways, but evidence remains inconsistent regarding their contribution to disease onset. Further studies are needed to clarify these relationships using standardized definitions of trauma, well-characterized outcomes, and longitudinal designs.

Autoimmune diseases and endometriosis

Immune dysregulation has been consistently observed in women with endometriosis and is considered one of the important biological features of the disease [56]. These alterations may reflect both a response to the presence of ectopic endometrial tissues and a pre-existing predisposition that facilitates their survival and growth. Rather than being a simple consequence of retrograde menstruation, immune abnormalities appear to influence adherence, implantation, and maintenance of ectopic lesions [13].

Epidemiological studies have reported increased rates of autoimmune and immune-mediated disorders in women with endometriosis. Conditions most frequently associated include rheumatoid arthritis, systemic lupus erythematosus, autoimmune thyroid disease, inflammatory bowel disease, and allergic disorders [13, 14, 17, 47]. In addition, asthma is more common in women with endometriosis, and recent genetic studies have identified shared targets between these two conditions [15]. These associations support the idea of overlapping immunological pathways and possibly shared genetic susceptibility.

Several immune abnormalities have been described at the cellular level. Women with endometriosis exhibit altered function of macrophages, natural killer cells, and T lymphocytes, as well as increased levels of cytokines and autoantibodies in peritoneal fluid and blood [57]. These factors contribute to inflammation, angiogenesis, and the invasive properties of endometriotic cells. Although many immune biomarkers have been proposed, their accuracy remains insufficient to replace surgical diagnosis, and none are currently validated for clinical use [25].

Genetic studies provide additional insight. Candidate genes involved in immunity and autoimmunity, such as PTPN22 and specific HLA alleles, have been associated with endometriosis in some cohorts [58]. However, replications in larger genome-wide studies have been inconsistent, and findings do not support the existence of a single shared autoimmune gene signature [59]. This suggests that immune dysfunction in endometriosis is likely multifactorial, involving both genetic predisposition and hormonal and environmental modulators.

From a predictive standpoint, autoimmune comorbidities are not specific enough to serve as standalone markers, but they may contribute to risk stratification. Women who present with chronic pelvic pain and concurrent immune-mediated conditions may have a higher likelihood of endometriosis and could benefit from earlier diagnostic evaluation. The presence of immune disorders may also influence future disease severity or treatment response.

Immune dysregulation is a characteristic feature of endometriosis and is reflected in increased comorbidity with several autoimmune diseases. Although the mechanisms are not fully understood, interactions between hormonal, genetic, and immune factors appear central to lesion development and persistence. Future research may clarify how immune profiles can be integrated into prediction models and personalized treatment strategies.

Environmental endocrine disruptors and dietary factors

Environmental endocrine disruptors

Environmental exposures that interfere with hormone biosynthesis, metabolism, or signaling have been investigated as potential contributors to endometriosis risk [60, 61]. Endocrine disruptors may influence estrogen receptor activity, inflammatory pathways, or immune function, thereby promoting a biological environment that supports the implantation and survival of ectopic endometrial tissues.

Several compounds have been implicated. Dioxins and polychlorinated biphenyls (PCBs) are persistent organic pollutants that accumulate in fatty tissues and may enhance inflammatory and estrogenic responses. Bisphenol A (BPA), commonly found in plastics, has estrogen-mimicking properties and has been associated with altered reproductive function. Phthalates, used in cosmetics, foods, and packaging, have been detected at higher concentrations in women with endometriosis compared to controls [60, 61]. Although evidence is not consistent across all studies, these chemicals may contribute to a susceptible hormonal milieu.

From a predictive perspective, endocrine disruptors are unlikely to serve as individual diagnostic biomarkers. However, they may be relevant as exposure-based risk factors, particularly in combination with genetic or epigenetic susceptibility. High levels of phthalates or other endocrine disruptors may help identify subgroups of women at increased risk who might benefit from preventive strategies or early monitoring.

Dietary and nutritional influences

Dietary factors may also influence endometriosis risk and symptom burden. Polyphenols, bioactive compounds found in plants, have been widely studied because they possess anti-inflammatory and antioxidant properties with relatively low toxicity [61]. Examples include curcumin, epigallocatechin gallate (EGCG), quercetin, and resveratrol, which can modulate estrogen receptors, reduce oxidative stress, and affect angiogenesis and apoptosis [62].

Clinical and preclinical studies have reported improvements in pain, inflammatory markers, and implant size after administration of polyphenols or supplements containing combinations of these substances [63, 64]. Vitamins C, E, and D have also been linked to antioxidant effects and possible symptom relief, although results remain inconsistent [65]. Micronutrients such as calcium [66] and zinc [67] may influence hormonal pathways or immune function, with low zinc levels associated with an increased risk of ovarian endometriotic cysts. However, evidence is stronger for symptom modulation than for primary disease prevention. Nutritional compounds cannot currently be recommended as predictive biomarkers. Their role lies primarily in adjunctive management, with potential benefits for inflammation and pain.

Linking back to prediction and risk

Environmental and nutritional factors are best viewed as modifiers of risk, rather than direct causes or treatments. Endocrine disruptors may create a permissive hormonal and inflammatory environment that increases susceptibility, especially when combined with genetic or epigenetic predisposition. Similarly, diet and micronutrients may influence pain and inflammation and could be incorporated into risk models alongside clinical symptoms, lifestyle, and family history.

Future research should prioritize

• standardized exposure measurements,

• longitudinal designs,

• and integration with genetic, epigenetic, and clinical markers.

This approach is more likely to clarify how exposures and diet contribute to prediction and risk stratification, rather than focusing only on treatment outcomes.

Conclusion

Several categories of markers show promise for risk stratification and for identifying women with a higher likelihood of clinically significant endometriosis or greater symptom burden. Genetic and epigenetic alterations, severe dysmenorrhea, migraine, immune-mediated comorbidities, stress-related neuroendocrine changes, and environmental endocrine disruptors have all been associated with increased risk or symptom burden. Although none of these markers is sufficiently specific to serve as a standalone diagnostic tool, their combination offers a potential pathway toward earlier identification.

Research in this field is still at an early stage. Most evidence comes from cross-sectional or retrospective designs, and large prospective studies are needed to determine the predictive value of these markers and their ability to stratify risk in asymptomatic or minimally symptomatic women. Emerging approaches, including polygenic risk scores and machine-learning models, illustrate how multimodal data could support individualized prediction and guide future clinical decisions (Fig. 3).

Fig. 3.

Conceptual model, polygenic and clinical integration for endometriosis prediction. This conceptual framework illustrates how multiple domains of information can be combined to support risk stratification of endometriosis. Polygenic risk scores, clinical symptoms (including severe dysmenorrhea, migraine, and pelvic pain), comorbidities, environmental endocrine exposures, and epigenetic signatures converge into a multimodal prediction model. This conceptual model illustrates how multimodal data may support risk stratification and earlier evaluation, enabling earlier evaluation of individuals at higher risk of endometriosis. This figure is conceptual and intended to summarize current evidence and a potential multimodal framework; it is not a validated clinical prediction tool and is not intended for direct clinical decision-making

The concept of prevention remains speculative but increasingly relevant. Identifying women at higher risk through genetic susceptibility, early symptoms such as severe adolescent dysmenorrhea, or environmental exposures may allow earlier monitoring, non-invasive investigation, or lifestyle interventions. While endometriosis is not currently preventable, improved prediction could reduce diagnostic delay, lessen disease burden, and support more timely treatment.

In summary, the integration of biological, clinical, and environmental markers represents a promising framework for risk stratification in endometriosis rather than precise individual-level prediction. Although these approaches are not yet ready for direct clinical implementation, they may help identify women who warrant earlier evaluation, particularly those presenting with severe adolescent dysmenorrhea and converging multimodal risk factors. By supporting earlier recognition and more targeted investigation, future predictive models could help reduce the prolonged diagnostic delay that still characterizes endometriosis and ultimately lessen disease burden.

Author contributions

Conceptualization: AV; writing—original draft preparation: AV; writing—reviewing: AV, JMA. The authors have read and agreed to the published version of the manuscript.

Funding

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Data availability

No datasets were generated or analyzed during the current study.

Declarations

Conflict of interest

The authors declare no competing interests.

Ethical approval

Not applicable.

Consent to publish

Not applicable.