Overview of metabolic risk profile and complications of NAFLD
Excessive hepatic fat accumulation defines the primary event in the continuum of nonalcoholic fatty liver disease (NAFLD) that has become the commonest liver disease in Western populations, affecting 17%–46% of adults . A strong linear correlation is confirmed between the presence of metabolic syndrome (MetS) or its components and the general prevalence of NAFLD . The latter exponentially raises to 50%–70% in individuals with type 2 diabetes mellitus (T2DM) and obesity, to reach a prevalence of 90%–95% in morbidly obese, of whom most will progress to nonalcoholic steatohepatitis (NASH). In the obese population with bariatric surgery, NASH prevalence ranges from 34% to 94% . On the other hand, only 7% of lean subjects, generally young women with presumably ectopic fat deposition develop NAFLD .
Recently, the concept of a new terminology of NAFLD was launched such as MetS- or Metabolic dysfunction-associated fatty liver disease (MAFLD) or even Dys-metabolism-associated fatty liver dysfunction (DALFD), aiming to reflect more accurately the crucial role of associated metabolic dysfunctions in the pathophysiology and as potential targets for future treatments.
Nonetheless, NAFLD is a diagnosis of exclusion of both secondary etiologies of liver disease and a daily alcohol consumption of ≥ 20 g in women ( Table 1 ).
| Disease/group of disorders |
|---|
| Alcoholic fatty liver disease (AFLD) |
| Hepatitis C (genotype 3) |
| Drug-induced fatty liver disease (glucocorticoids, methotrexate, tamoxifen, amiodarone, ibuprofen, aspirin, mipomersen, lomitapide, valproate, carbamazepine, antiretroviral medication, 5-FU, irinotecan) |
| Endocrine disorders |
| Coeliac disease |
| Autoimmune hepatitis |
| Wilson’s disease |
| Hemochromatosis |
| A/hypo-beta-lipoproteinemia |
| Lipodystrophy |
| Inborn errors of metabolism (LCAT-deficiency, cholesterol ester storage disease, lysosomal acid lipase deficiency) |
| Parenteral nutrition |
| Starvation |
| Acute fatty liver of pregnancy |
| HELLP syndrome |
Using a > 5.56% cutoff at proton magnetic resonance spectroscopy ([ 1 H]-MRS) or MR imaging for the hepatic fat content , which corresponds to > 5% hepatocellular triglycerides (TGs) accumulation, steatosis, or fatty liver is defined ( Table 2 ) and regarded as harmless as it is asymptomatic and there is no impact on mortality. The cornerstone in risk stratification and disease prognosis is the onset of NASH defined by the histopathological appearance of steatosis, hepatocyte ballooning, and lobular inflammation with or without fibrosis. Advanced (bridging) fibrosis precedes cirrhosis ( Table 2 ).
| NAFL a (nonalcoholic fatty liver) | ≥ 5% hepatic steatosis | |
| NASH (nonalcoholic steatohepatitis) | Advanced steatosis, hepatocellular ballooning, lobular inflammation; no fibrosis | |
| Mild fibrosis | Perivenular or perisinusoidal or periportal fibrosis | |
| Moderate fibrosis | Perisinusoidal/perivenular and periportal fibrosis | |
| Severe fibrosis | Bridging fibrosis with nodularity | |
| NASH cirrhosis | Cirrhosis | Cirrhosis co-exists with steatosis and inflammation |
There is solid evidence that NASH develops at a substantial rate, that is, 10%–40% of NAFLD cases, of which 15%–25% cases progress to cirrhosis and/or hepatocellular carcinoma, overall increasing the need for liver transplantation. Although NASH correlates with the development of fibrosis, there is no clear prediction of fibrosis progression with NASH and occasionally fibrosis may progress in absence of NASH. Once fibrosis develops, rates of progression from one stage to the next one is doubled by coexistence of inflammation, hypertension, and diabetes, all of these being the major traits of polycystic ovary syndrome (PCOS). In turn, patients with NAFLD are at the risk for T2DM, chronic kidney disease, and extra-hepatic malignancies (i.e., colon, stomach, breast, uterus, and possibly prostate) and are at the forefront of developing increased cardiovascular morbidity and mortality.
Epidemiological aspects of NAFLD in PCOS
Evidence from studies published over the past years demonstrates the increased prevalence of NAFLD in women with PCOS, with an estimated range between 25.4% and 68.8% ; however, numbers may vary considerably as a function of diagnosis criteria, race and ethnicity, body mass and phenotype, and analytical methods. Occasionally, a lower prevalence of 15.8% has been reported in nonobese patients with PCOS .
Conversely, PCOS among premenopausal women with NAFLD seems to be more frequent. In 110 overweight and obese premenopausal women, 64.5% of participants exhibited NAFLD, and 43.7% of these fulfilled 2003 Rotterdam PCOS diagnosis criteria vs only 23.1% of women without NAFLD . A smaller sample of 14 patients with NAFLD found even higher percentages, with 71% of women fulfilling Rotterdam PCOS diagnosis criteria .
A double prevalence of NAFLD in women with PCOS vs non-PCOS was remarked even in age-matched and weight-matched groups , with the most striking differences in the young and obese patients. Large-scale studies and metaanalyses confirmed NAFLD prevalence is almost fourfold in PCOS compared with controls , also depending on the phenotypic aspect and disease severity, in addition to ethnicity and age. Another metaanalysis including 2700 patients and 17 studies stated a 2.54-fold risk of NAFLD in women with PCOS . When stratified by the body mass index (BMI), the NAFLD risk was significantly increased in both obese (odds ratio (OR) 3.01; 95% CI, 1.88–4.82) and nonobese (OR 2.07; 95% CI, 1.12–3.85) individuals . Despite being retrospective, the UK primary care database including ≈ 63.000 patients with PCOS and ≈ 121.000 age-matched, BMI-matched, and location-matched control women provided evidence that PCOS patients had 2.0- to 2.4-fold rate of NAFLD, independently of obesity and dysglycemia . Likewise, in ≈ 50 million women from the United States National Inpatient Sample database 2002–14, multivariate logistic regression revealed that patients with PCOS (≈ 77.000) were at significantly higher risk of developing NAFLD (OR 4.30; 95% CI, 4.11–4.50) . Furthermore, childbearing women with PCOS are at higher risk of developing fatty liver disease during pregnancy.
Taking a closer look, the prevalence of NAFLD in women with PCOS mirrors to a great extent the frequency of MetS in women with PCOS ( Fig. 1 ). More than 50% of patients with PCOS are obese, and even in European countries with a low incidence of obesity (< 9% of the population) and MetS such as Italy, the prevalence of MetS in PCOS is 3.4–6.6 times higher when compared to matched controls, thus, indicating that PCOS status results in a higher risk of MetS . A metaanalysis similarly revealed 2.0–4.0 times OR for impaired glucose tolerance, MetS, and T2DM in premenopausal women with PCOS, independently of the BMI, although with some additive effects of obesity . Therefore, PCOS aggregates a constellation of metabolic dysfunctions driven by insulin resistance (IR), which in turn increases susceptibility for NAFLD ( Fig. 1 ).
We learned from the Study of Women’s Health Across the Nation (SWAN) that as women with PCOS age and transition through menopause, a natural decline in androgen levels occurs and clinical hyperandrogenism in addition to reproductive dysfunction alleviate , whereas the unfavorable metabolic risk profile persists beyond menopause. Postmenopausal PCOS exhibits a worsened lipid profile in terms of high-density cholesterol (HDL-C) and TG, higher concentration of high-sensitivity-C reactive protein (Hs-CRP), and a fivefold risk of developing T2DM compared to age-matched women without PCOS . Unfortunately, we lack consistent data focusing on the prevalence of NAFLD in postmenopausal women with PCOS; however, the current body of evidence supports the conclusion of increased risk of NAFLD in women with PCOS across the lifespan.
Histological proof revealing patients with PCOS tend to develop more rapidly NASH was obtained in liver biopsy studies . In the study of Setji et al., NASH was detected in liver biopsies performed in 29 of 200 (15%) patients of age 29 years, and varying degrees of fibrosis were confirmed in all patients with NASH to certify that PCOS predicts a more severe liver disease.
Putative pathophysiological pathways linking PCOS to NAFLD
Androgen excess
Despite the conflicting results delivered by studies, most opinions favor a key role of hyperandrogenism in promoting fatty liver. Emerging data from experimental models brought convincing evidence to the indisputable pathophysiological role of androgens in women with PCOS and NAFLD ( Fig. 2 ).
Metabolic effects of androgen hormones in PCOS-like animal models
Exposure of female rats to 90-days continuous release pellets of 5α-dihydrotestosterone (DHT) (83 μg/day) or placebo was accompanied by visceral fat deposition and a hypertrophic morphology of visceral adipocytes. Apart from these changes, a typical dyslipidemic lipid profile emerged with increased TG and elevated plasma nonesterified fatty acids (NEFA), paralleled by enhanced gene expression of adipose tissue proteins involved in lipid metabolism and lipogenesis such as steroid regulatory-element binding protein 1 (SREBP-1), peroxisome proliferator-receptor activated (PPAR)-γ, or phosphoenolpyruvate carboxykinase (PEPCK) . Indeed, an intra-adipose mechanism of androgen activation that contributes to adipose cells remodeling and increased de novo lipid synthesis in adipocytes has been identified . When the same hyperandrogenic rat model received liraglutide (0.2 mg/kg), favorable effects on abdominal adiposity, systemic glucose homeostasis, total cholesterol, and blood pressure were obtained . Treating HepG2 cell cultures with various concentrations of DHT resulted in lipid accumulation without affecting cell viability .
Experimental exposure of animals to androgens led to histologically proven fat accumulation in the liver , independently of total body fat . Testosterone (T) effects on the fatty liver were examined using aromatase inhibitor, letrozole, which increases endogenous T levels in the rat. After 90-days of letrozole, T concentration and plasma TG levels increased but aspartate aminotransferase (AST) and alanine aminotransferase (ALT) increased as well to confirm hepatocellular attack evidenced by histopathological examination . Moreover, overexpression of genes related to fatty acid synthesis and—oxidation was observed, including uncoupling protein 3 (UCP3) and carnityl palmitoyltransferase 1α (Cpt1α) genes, in addition to inhibition of AMP-activated protein kinase-α pathway, an aspect reported in DHT-treated HepG2 cells as well . Intriguingly, in a similar PCOS-like rat model, androgen receptor blocker spironolactone was able to prevent hepatic TG accumulation .
Pathophysiological role of hyperandrogenemia in the development of NAFLD in PCOS
A higher liver fat content was detected by [ 1 H]-MRS in hyperandrogenic PCOS vs normo-androgenic counterparts , and some found that serum androgens were higher in patients with NAFLD compared to women with PCOS but without NAFLD . Investigation of the molecular signature in PCOS revealed that DL-receptor expression in the liver, thereby prolonging the half-life of very low-density lipoprotein (VLDL) and LDL and contributing to dyslipidemia and fatty liver, in addition to promoting hepatocellular apoptosis . This is in agreement with the previously reported association between androgen hormone levels and high ALT in PCOS, independently of obesity and IR .
The deleterious effect of androgen excess in women with PCOS appears to fit the recent theory of a sexually dimorphic role of androgens in human metabolic disease, which suggests similar adverse metabolic outcomes in hyperandrogenic women and in men in whom inadequately low androgen levels are present. A systemic lipotoxic metabolome was revealed by nontargeted serum metabolomics performed in women, in agreement with experimental studies .
On top of that, hyperandrogenemia augments IR and the well-described reciprocal interaction between these dominant PCOS traits markedly contributes to the pathogenesis of NAFLD.
Lipotoxicity, glucotoxicity, postprandial dysmetabolism, and insulin resistance
Insulin resistance is a prominent feature of PCOS, affecting 60%–80% of women harboring the syndrome. Obese women with PCOS will progress to IGT and T2DM with a rate of progression from normal glucose tolerance to T2DM as high as 5%–15% ( Fig. 1 ). Conversely, a 45%–50% reduction in glucose disposal, which is a measure of whole-body insulin sensitivity has been documented in nondiabetic patients with NAFLD and interpreted as a hallmark of the “metabolic” NAFLD phenotype to differentiate from the “genetic” NAFLD phenotype that seems not to exhibit IR . Overfeeding studies in men (+ 40-50% diet) have concluded that altered insulin sensitivity may develop within days of exposure to hypercalorie diet, in addition to visceral fat accumulation and that IR constantly precedes NAFLD, thus, underpinning the concept of NAFLD as the hepatic manifestation of IR.
At the time MetS or components of MetS are diagnosed in the patient with PCOS, steatosis develops being attributed to increased NEFA delivery from various sources, a step known as the “first hit” of the pathophysiological chain, resulting in hepatic resistance to insulin and an augmented liver production of TG. A key source of NEFA in PCOS is the skeletal muscle. Postbinding defects in insulin receptor and impaired insulin signaling at the levels of glucose transport in the skeletal muscle are consistent with the molecular signature of IR in PCOS , to explain why muscle glycogen synthesis is reduced, whereas de novo lipogenesis is increased even in lean but insulin-resistant PCOS. Subsequent accumulation of intra-myocellular lipid metabolites such as the long-chain fatty acids acyl-CoA, ceramides, and particularly lipotoxic di-acylglycerols (DAG) leads to IR in the muscle by reducing nonoxidative glucose disposal via activation of protein kinase (PC)-θ . This series of events reflects a nonspecific mechanism as it develops for instance with aging . Redistribution of the unstored carbohydrate load towards the liver results in de novo hepatic lipogenesis, in addition to dyslipidemia as suggested by elevated TG and low Apo-I protein and HDL-C. An acute rise in NEFA load within a lipid infusion lowered the rate of insulin-stimulated glucose disposal in the skeletal muscle by 67% in PCOS vs 49% in controls ( P < .05) thus suggesting that women with PCOS are more susceptible to lipid-induced IR . Gluco- and lipotoxicity further promote IR in both liver and muscle, to initiate a vicious circle. In addition to that, hyperinsulinism on its own augments collagen production by hepatic stellate cells, a pathophysiological process contributing to liver fibrosis .
Postprandial dysmetabolism, a term defining the sum of postprandial hyperglycemia and hyperlipidemia is a putative mechanism governing the pathogenesis of both PCOS and NAFLD and contributes to the development of endothelial dysfunction, oxidative stress, chronic inflammation, and increased cardiovascular morbidity and mortality . Patients diagnosed with PCOS or NAFLD or the combination of the two share postchallenge hyperinsulinemia and various grades of dysglycemia to a greater extent vs controls. Moreover, increased postprandial or postchallenge very low-density lipoprotein (VLDL)-1 accumulation in addition to the increased postchallenge area under the curve (AUC) TG , AUC VLDL , and AUC cholesterol but lower AUC HDL-C was demonstrated in women with PCOS . The high liver fat content predicts metabolic dysfunction in women with PCOS, including impaired lipid and glucose metabolism and IR.
One cannot ignore the fact that IR per se is amplifying hyperandrogenism by various mechanisms, to further promote the development of NAFLD in an indirect manner. Via ovarian insulin receptors, insulin acts as a co-gonadotropin to augment gonadal androgen production in theca cells; in addition to decreasing hepatic sex-hormone binding globulin (SHBG) synthesis and therefore increasing bioavailable T. Hyperinsulinism induces phosphorylation and activation of P 450 -17α-hydroxylase to increase 17α-hydroxy-progesterone and DHEAS synthesis in the adrenal gland through the Δ 4-steroidogenic pathway and stimulates the Δ 5-steroidogenic pathway in the ovaries to increase the production of T, androstendione, and 17α-hydroxy-progesterone.
Metabolic crosstalk between the liver and the pancreas: Role of fetuins
Once constituted, steatosis promotes a fatty degeneration of the pancreas and possibly the kidney via the release of metabolites such as palmitate and the hepatokine fetuin-A by the fatty liver thus triggering a local immune cell infiltration and inflammation and accelerating the development and progression of diabetes. To support the metabolic crosstalk between the liver and the pancreas, glycoprotein fetuin-A was shown to inhibit glucose transporter (GLUT)-4 translocation and interferes with downstream phosphorylation in insulin receptor signaling. The hepatokine level, which in NAFLD increases, is strongly correlated to parameters of IR. Glucolipotoxicity stimulates fetuin-A production as well. Despite conflicting results in smaller studies, high fetuin-A was observed in PCOS in a large cross-sectional study, in addition to a positive correlation between fetuin and BMI, waist-to-hip ratio (WHR), HOMA-IR, androgens, and lipids . Recently, another hepatokine, fetuin-B was increased in a cross-sectional evaluation of 257 women with PCOS and 141 controls and correlated to IR parameters. Six-month glucagon-like peptide (GLP)-1 analogs treatment was able to significantly decrease fetuin-B concentration .
Chronic inflammation and oxidative stress
In response to chronic hepatic lipid accumulation, enhanced mitochondrial β-oxidation ensues, and eventually, the exhaustion of antioxidant mechanisms results in increased oxidative stress. Several inflammatory pathways are activated via insulin signaling including the nuclear factor (NF)-κB, tumor necrosis factor (TNF)-α, and interleukin (IL)-6 pathways and this “second hit” consisting of inflammation, mitochondrial dysfunction, oxidative stress, and lipid peroxidation promotes progression to NASH and liver cell necrosis. Advanced Glycation Endproducts (AGE), which are abundantly found build up in PCOS, have pleiotropic pejorative effects through activation of the receptor for AGE (RAGE), which in the liver is expressed mainly on hepatic stellate and Kupffer cells, thus activating the NF-κB pathway .
It is well described that women with PCOS associate a constellation of elevated inflammatory markers including plasminogen activator inhibitor (PAI)-1, Il-1, IL-6, CRP, and TNF-α, etc., thus being created a microenvironment that may promote steatohepatitis. Hs-CRP and osteopontin were reported to predict NAFLD better than WHR in both women with PCOS and control women . In small-sized non-PCOS studies, IL-8 was linked to the development of NASH, and TNF-α was an independent predictor of fibrosis in NASH . Comparison of inflammatory factors of endothelial dysfunction Intercellular adhesion molecule (ICAM)-1, vascular cell adhesion protein (VCAM)-1, E-selectin, and P-selectin in women with PCOS with and without NAFLD, reported no differences between groups, which questions a more adverse cardiovascular risk profile in the presence of NAFLD at young age . However, the small sample size, possibly not providing adequate power to this study is a limitation.
Obesity, adipose tissue dysfunction, and adipokines
Overweight and obese patients with PCOS have a fourfold prevalence of liver disease compared to lean women with PCOS in several ethnic groups . Excepting two small-sized studies , even lean women with PCOS present higher prevalence of NAFLD compared to lean women without PCOS, to suggest the strong intervention of other pathophysiological determinants besides fat mass.
Once the finite capacity of subcutaneous adipose tissue to expand is exceeded, ectopic fat deposition occurs. This aberrantly functioning fat is resistant to the antilipolytic action of insulin and releases a plethora of pro-inflammatory adipocytokines. Low adiponectin and high visfatin are the key intermediates of deleterious consequences of adipose cell hypertrophy and dysregulation on the liver and vascular system . In particular, high-molecular weight adiponectin is strongly associated with IR. Adiponectin is a stimulus for PPAR-α gene expression, which correlates negatively to the presence and severity of NASH and is independently associated with NASH . However, it seems that when liver disease is limited to steatosis, adiponectin is not different in NAFL of women with or without PCOS . Serum resistin was found to be higher in patients with NASH, but not in patients with NAFL when compared to control women. Knowing that resistin is a pro-inflammatory adipokine, results are possibly in conjunction with inflammatory changes in both liver and adipose tissue. Likewise, in patients with PCOS with biopsy-proven steatosis resistin levels were unchanged in comparison to women of non-PCOS with steatosis .
The concentration of leptin in NAFLD is associated with IR and reflects whole-body fat in both patients with PCOS and patients without PCOS. Changes in leptin predict independently the development of NAFLD in patients with weight gain. Recently, leptin-mediated disease progression through immune cells was clarified in a NAFLD mouse model, demonstrating that leptin exerts pro-inflammatory and pro-apoptotic effects by caspase activation in macrophages and hepatocytes, mediated by CD8 + T lymphocytes, in addition to a pro-fibrogenic effect through stimulation of hepatic stellate cells . Visfatin, a pro-inflammatory adipokine was found to be increased in NAFLD; however, an 11-years population follow-up on 403 patients failed to reveal an association between visfatin and presence or absence of NAFLD . High circulating visfatin was reported in obese women with PCOS; on the contrary, serum visfatin and adipose tissue visfatin mRNA expression was stated as normal in lean women with PCOS . Nonetheless, visfatin levels in women with PCOS are subjected to high variability.
Ghrelin
PCOS associates low levels of ghrelin, a hormone that regulates food intake, the energy balance, and glucose homeostasis, and is negatively related to IR and fat mass. Ghrelin actions on the liver have been studied extensively, to show its ability to regulate autophagy and inflammation and display antifibrotic properties. To date, there is no proof of ghrelin connecting PCOS and NAFLD physiopathologically; however, ghrelin might emerge as a promising therapeutic target in the future.
Genetic and epigenetic risk factors
PCOS and NAFLD are polygenic diseases and inherited genetic variants may have an impact on disease occurrence and progression in an individual. Genome-wide association studies (GWAS) have identified candidate genes closely linked to metabolic traits shared by both PCOS and NAFLD ( Fig. 3 ). Polymorphisms in candidate gene cannabinoid receptor 1, a gene involved in food intake and energy homeostasis are associated with the development of NAFLD in women with PCOS . In view of that, endocannabinoid receptor blockade by rimonabant reduced ALT levels in PCOS, independent of weight loss or inflammation . In the experimental mouse model, the glucokinase regulatory protein (GCKR) gene is involved in glucokinase activity control, and glucose homeostasis. In PCOS, copy-number variations (CNV) of SNP rs126032 (GCKR) are linked to HbA1c, diabetes, and NAFLD . Variation in GCKR locus has been associated with NAFLD in other studies and is explained by uncontrolled glucose uptake by hepatocytes with glycolysis, hepatic fat accumulation, and IR . Polymorphisms in the fat mass and obesity-associated gene (FTO) are closely related to obesity, IR, T2DM, and PCOS, and gain-of-function variants may result in altered NEFA oxidation and hepatic de novo lipogenesis with steatosis . The recent epigenetics input, particularly of dysregulated mi-RNAs , will contribute in the future to a better understanding of the pathogenesis, and possibly, detection of women with PCOS at risk for NAFLD.
Chemicals and endocrine disruptors
Given that the liver has a crucial role in metabolizing many endocrine-disrupting chemicals (EDC) including heavy metals , it was hypothesized that some of these compounds might play a role in the pathogenesis of NAFLD. High serum bisphenol A (BPA) was identified in a subgroup of patients with PCOS with severe IR and androgen excess who also exhibited a higher prevalence of NASH and increased spleen volume, a marker of chronic inflammation . BPA is involved in the multifactorial etiology of PCOS, by promoting both androgen excess and metabolic dysfunction . In the adult mouse liver, BPA has the capacity to induce expression of genes related to lipid synthesis and promote fat accumulation. It inhibits adiponectin production in fat and NEFA oxidation in the skeletal muscle, overall promoting hepatic IR. Exposure to BPA is linked to GM dysbiosis and high endotoxin, subsequently leading to liver inflammation via the toll-like receptor (TLR)-4)/NF-kB pathway . A large U.S. study of 7605 participants stated that high urinary BPA levels were associated with NAFLD . To date, we lack consistent studies to evaluate the implications of BPA in the development or progression of PCOS-related NAFLD.
Gut microbiota
The relationship between gut microbiota (GM) and metabolic diseases is a rapidly expanding field. In the letrozole-induced mouse with PCOS and patients with PCOS, a lower alpha diversity (i.e., species richness and phylogenetic diversity) of the stool microbiome was observed , sharing similarities to changes encountered in human obesity. The findings were positively associated with hyperandrogenism . Furthermore, an altered stool microbiome composition pattern (i.e., dysbiosis) was revealed, with an increase in bacteria belonging to Bacteroides (e.g., Bacteroidaceae and S24-7 families), with a decrease in Firmicutes species (e.g., Lactobacillacae and Ruminococcaceae ). Overall consequences are changes in the production of short-chain fatty acids (SCFA), choline and bile acids, and alterations of the intestinal barrier integrity and immunity . Opportunistic Escherichia/Shigella and Streptococcus increase in obese women with PCOS . Gram-negative bacteria represent a source of bacterial lipopolysaccharides (LPSs) and a cause of metabolic endotoxiemia that may induce chronic inflammation, obesity, and IR, all of these linked to MetS and the development of fatty liver disease. Patients with NASH have an increase in the Bacteroides spp. too, whereas liver fibrosis appears to be linked to an increase in Ruminococcus , although this is controversial. The GM might influence disease severity through its impact on lipid and bile acid metabolism. A gut microbiome-derived signature based on 27 bacterial features has been developed to detect advanced fibrosis and cirrhosis ; however, needs to be validated in further studies. To sum up, alterations in GM observed in PCOS and NAFLD mirror the metabolic dysfunctional background of both diseases. To confirm a causal relationship, further research is warranted.
Independent predictors of NAFL and NASH in PCOS
Age
In women with PCOS, the disorder develops at younger age and results in more severe forms compared to women without PCOS. Prevalence of NAFLD in premenopausal women with PCOS appears to be similar to that reported in women without PCOS but being postmenopausal and higher than in premenopausal women without PCOS, possibly indicating a protective effect of estrogens on the liver.
Free testosterone
A higher free androgen index (FAI) is linked to the presence of NAFLD (mean difference 4.46, 95% CI 3.53–5.39 vs women without NAFLD) and is a crucial contributor to hepatocyte apoptosis and necro-inflammation. Overweight/obese women with PCOS exhibit higher levels of liver enzymes compared to overweight/obese control women , which might be explained by contribution of the hyperandrogenemia. In nonobese women with PCOS, hyperandrogenemia emerges as an independent risk factor for NAFLD . Nonetheless, in a large cohort of 600 patients with PCOS, Macut et al. failed to report differences in the prevalence of NAFLD between hyperandrogenic and nonhyperandrogenic women with PCOS phenotypes .
Sex-hormone binding globulin
A low SHBG is found in patients with PCOS and NAFLD , independently of obesity, and may further promote hyperandrogenism. An impressive hazard ratio of 4.75 in the rate of NAFLD was found in patients with PCOS with SHBG < 30 nmol/L from the UK primary care database . Impaired hepatic production of SHBG due to steatosis was claimed in non-PCOS patients with NAFLD .
Metabolic syndrome
PCOS, MetS, and NAFLD share similar traits and NAFLD even may precede MetS. In PCOS, liver disease is clearly related to the severity of IR, further aggravated by NAFLD. Lean patients with altered body composition or dysfunctional adipose tissue or metabolically healthy obese women with PCOS are also at risk for NAFLD, and in nondiabetic patients, the HOMA-IR might provide an estimate of the risk of steatohepatitis or progression of fibrosis. In 188 patients with PCOS, BMI, and WHR independently predicted NAFLD, in addition to HOMA-IR and TG . In 202 young Italian women with PCOS, IR, expressed by the ISI Matsuda surrogate parameter, emerged as an independent risk factor for steatosis in both obese and nonobese patients . More than 95% of women with PCOS and NAFLD fulfilled IDF criteria for MetS in another study . The presence and severity of MetS are of utmost importance to predict the potential development of hepatitis and fibrosis .
PCOS phenotype A
Patients with phenotype A, hyperandrogenism-anovulation-polycystic ovarian morphology, are at the highest risk of developing NAFLD .
Obstructive sleep apnea
Obstructive sleep apnea (OSA) has been shown in a small cohort to strongly predispose for NAFLD in patients with PCOS .
Diagnosis criteria and risk stratification of NAFLD in PCOS
The general goal in NAFLD is to define biomarkers for diagnosis, risk stratification, and follow-up. The ideal biomarker is highly accurate, simple, noninvasive, readily available, and as priceless as possible. In spite of no biomarkers fulfilling the above-mentioned criteria, a central pillar in the evaluation strategy is represented by the prediction of hepatocellular inflammation and fibrosis as both significantly increase the risk of progression to cirrhosis.
Noninvasive serum laboratory parameters of NAFL and NASH in PCOS
Studies of noninvasive isolated markers or composite algorithms of NAFLD in patients with PCOS are small-sized and include heterogeneous disease phenotypes, many of these trials lacking an adequate control group. Furthermore, the reproducibility of these markers is questioned. However, some have been assessed ( Table 3 ) and deserve attention as predictors of steatosis such as the NAFLD-Liver Fat Score (NAFLD-LFS), an algorithm that incorporates simple variables and was validated against [ 1 H]-MRS with sensitivity and specificity of 95%. Lipid Accumulation Product (LAP) correlates well with NAFLD-LFS in patients with PCOS and has been shown to predict all-cause mortality in patients with high cardiovascular risk . In premenopausal women with PCOS, of whom 70% were diagnosed with hepatic steatosis using FibroScan, LAP was independently associated with NAFL . In obese women with PCOS, the Fatty Liver Index (FLI) predicts NAFLD and MetS and vice versa having PCOS results in 2.5-fold higher FLI levels in age-adjusted analyses .
| Noninvasive index [Reference] | Index equation | Outcomes | Comments |
|---|---|---|---|
| NAFLD liver fat score (NAFLD-LFS) | [22.89 + 1.18 × MetS a (yes = 1/no = 0) + 0.45 × T2DM b (yes = 2/no = 0) + 0.15 × fasting serum INS (mU/L) + 0.04 × AST (U/L) − 0.94 × AST/ALT] | Predictor of hepatic steatosis and MetS in women with PCOS, at NAFLD-LFS cutoff >− 0.64 No information on steatohepatitis or fibrosis | Validated against [ 1 H]-MRS |
| Lipid accumulation product (LAP) | [(WC − 58) × TG] | Independently linked to fatty liver and a good predictor of MetS in PCOS No information on steatohepatitis or fibrosis | Integrates cardio-metabolic risk |
| Fatty liver index (FLI) | e 0.953 × loge(TG) + 0.139 × BMI + 0.718 × loge(GGT) + 0.053 × WC − 15.745 /(1 + e 0.953 × loge(TG) + 0.139 × BMI + 0.718 × loge(GGT) + 0.053 × WC − 15.745 ) × 100 | Independent predictor of hepatic steatosis in obese and nonobese women with PCOS | Validated against FibroScan |
| Hepatic steatosis index (HSI) | 8 × (ALT/AST) + BMI + 2 × (female) + 2 × (T2DM) | Predictor of MetS In PCOS | |
| PCOS-hepatic steatosis (PCOS-HS) index | 1/1 + (exp (−(26.01 + (− 0.3761 × BMI percentile + 0.05781 × WC (cm) + 0.0448 × HOMA-IR + 0.00095519 × HDL (mmol/L) + 0.00005892 × TG (mmol/L) + 0.0964 × ALT (IU/L) + 0.001548 × free testosterone (nmol/L) − 0.06806 × SHBG (nmol/L))) | Predictor of NAFLD with AUROC = 0.81, sensitivity 82% and specificity 90% Positive correlation to the hepatic fat fraction | Validated against MRI Developed in obese, adolescent (12–19 years) PCOS |
| Osteopontin | – | Positive correlation to liver fat content in nonobese PCOS Positive correlation to FAI in nonobese PCOS | Validated against liver ultrasonography |
| Tissue inhibitor of metalloproteinase-1 (TIMP-1) | – | Unchanged after 3-months vitamin D repletion in overweight and obese PCOS | |
| Alanine aminotransferase (ALT) | – | Poor predictor of steatohepatitis with AUROC = 0.61 and low sensitivity and specificity No information on fibrosis | Normal values do not exclude NASH and/or fibrosis |
| Caspase-cleaved cytokeratin-18 fragments (M30) | – | Marker of inflammation, presenting elevated concentration in NASH compared to simple steatosis | High variability of cutoff, therefore optimal cutoff needs validation in the population |
| Fibroblastic growth factor (FGF)-21 | – | Elevated levels in patients with PCOS | Circadian variability Not available for clinical use |
| NAFLD fibrosis score (NFS) | − 1.675 + 0.037 × age (years) + 0.094 × BMI (kg/m 2 ) + 1.13 × IFG/diabetes (yes = 1, no = 0) + 0.99 × AST/ALT ratio − 0.013 × platelet (× 10 9 /L) − 0.66 × albumin (g/dL) | Marker of fibrosis Useful in surveillance of the patient as improvement suggests regression of fibrosis | High specificity to discriminate advanced fibrosis Not specifically calibrated for young population (< 35 years) |
| Fibrosis-4 index (FIB-4) | Age (years) × AST (U/L)/platelets (× 10 9 /L) × √ ALT (U/L) | Marker of fibrosis Unaffected in patients with PCOS | Limited value in incipient fibrosis Related to fibrosis improvement in clinical trials |
| AST-to-platelet ratio index (APRI) | AST/upper limit of normal/platelets (10 9 /L) × 100 | Marker of fibrosis Unaffected in patients with PCOS | |
| BMI-Age-ALT-TG (BAAT) | Age + BMI + TG + ALT | Elevated in PCOS and predictor of MetS Marker of fibrosis | The study lacked multivariate analysis to adjust for differences in BMI ± age |
| BMI-Age-ALT/AST-T2DM (BARD) | BMI + T2DM + 2 × (ALT/AST) | Marker of fibrosis Unaffected in patients with PCOS | |
| Procollagen type III amino-terminal propeptide (PIIINP) | – | Elevated in obese patients with PCOS Predictor of progression of fatty liver to NASH Sensitive marker of fibrosis | |
| Enhanced liver fibrosis score (ELF) | 2.278 + 0.851 ln(HA) + 0.751 ln(PIIINP) + 0.394 ln(TIMP-1) | Decreased after 3-months vitamin D repletion in overweight/obese PCOS | Liver imaging was not available |
| Hyaluronic acid (HA) | Marker of fibrosis Unaffected in patients with PCOS |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
