OA has long been regarded a unique consequence of the tearing and abrasion processes that lead to cartilage breakdown. Cartilage loss occurs if there is excessive mechanical stress on the joint, and the creation of osteophytes is considered a reactive process of the bone to protect and stabilize the altered joint [13]. OA is a complex disease that is not purely mechanical or simply caused by inflammation; it not only involves the cartilage but also various other tissues, such as the synovium, subchondral bone, capsule, meniscus, muscle, and tendon [14]. In particular, recent research focused on the activities of hormones and cytokines related to inflammation has revealed that in addition to these mechanisms, metabolic pathways also play a large role in the development of OA [8, 15–17]. Obesity is an important risk factor for OA alongside aging and injury, and the potential involvement of obesity-related mechanisms is revealed by OA affecting not only the weight-bearing joints but also the joints of the hand [15, 18]. OA is a complex disease caused by inflammatory mediators released from cartilages, bones, and synovia [16]. In OA, lipid mediators play a potential role in cartilage degradation, contributing to its pathophysiology [17]. In a prospective population-based study, the incidence of knee OA in participants with metabolic symptoms was attributed to an increased BMI [8]. By analyzing the direct and indirect obesity-related factors associated with the development of OA in mice and other animal models, it has been proposed that a complex interaction between genetic and environmental factors related to obesity contributes to the incidence and severity of OA [19]. Adipose tissue and obesity-related dyslipidemia have been shown to play a central role in obesity-induced OA. Adipose tissue shows signs of inflammation and secretes pro-inflammatory adipokines, in addition to higher levels of cytokines, including TNF-α, IL-6, and vascular endothelial growth factor. This local inflammatory response leads to a low degree of systemic inflammation. In addition, obesity-related dyslipidemia, defined by increased levels of triglycerides, free fatty acids, and oxidative LDLs, can aggravate inflammatory symptoms and increase the production of matrix metalloproteinase in the joint tissue, contributing to the pathogenesis of OA [20].
The basic hypothesis proposed by earlier studies is that obesity is closely related to knee OA. More specifically, in addition to their involvement in energy storage, adipocytes are endocrine cells that secrete hormones and inflammatory cytokines [21–23]. Importantly, the metabolic action of adipose tissue is also further emphasized by its location in the body, as the worse effects are attributed to visceral fat, liver fat, and white fat cells [24–26], which create a framework that is detrimental in chronic diseases [21–23, 27].
In particular, NAFLD adversely affects the prognosis of metabolic diseases [28–30]; thus we designed the study based on this hypothesis. NAFLD is the outcome of obesity, insulin resistance, and metabolic syndrome. However, NAFLD is also responsible for the same conditions [31, 32]. In fact, the pathophysiological mechanisms of NAFLD may lead to the occurrence of extra-hepatic complications, including type 2 diabetes mellitus, cardiovascular disease, and chronic kidney disease [32]. Thus, it can be assumed that NAFLD also affects the occurrence of knee OA and alternatively, knee OA may be associated with the occurrence of NAFLD.
Liver biopsy is the most accurate approach to make a definitive diagnosis of NAFLD [28, 29]. Nonetheless, imaging studies, such as liver ultrasound and computed tomography, are commonly used to diagnose NAFLD in clinical trials [28, 29]. However, liver imaging tests could not be easily performed for patients in the KNHANES, which was a study targeting Korean nationals. Thus, we adopted a simple diagnostic tool, the HSI, which can be easily calculated using routine laboratory tests. To evaluate the effectiveness of the HSI, a cross-sectional study was conducted in 10,704 subjects who underwent health screening [10]. At values of < 30.0 or > 36.0, HSI detected NAFLD with a specificity of 92.4% [10]. Additionally, NAFLD was excluded with a sensitivity of 93.1% [10]. Furthermore, several studies have reported that HSI is a valid tool for diagnosing NAFLD [33–35]. In patients with type 1 diabetes, an HSI > 36 was significantly associated with metabolic syndrome and nephropathy [34]. Several biomarkers, including HSI, can predict liver steatosis and are associated with homeostatic Model Assessment of Insulin Resistance [35].
Previous studies have shown that several metabolic components are closely associated with the development of OA. These include blood pressure, triglycerides, HDL cholesterol, LDL cholesterol, and blood sugar [36, 37]. Hypertension, hypercholesterolemia, and high blood glucose were associated with unilateral and bilateral knee OA in 979 women as noted in our previous study, following an analysis using the Kellgren–Lawrence system [36]. Several factors associated with metabolic syndrome have been inversely related to the minimum joint space width and minimum joint space area [37]. Conversely, another study also showed that OA was not associated with metabolic syndrome; after adjusting for BMI, neither metabolic syndrome or its components were associated with the incidence of OA, rather only with blood pressure [38]. A study investigating the association between OA and metabolic syndrome in Korea also provided evidence supporting the importance of metabolic mechanisms in OA incidence in a large population, using data from the National Health and Nutrition Survey [39].
A key finding of our study is the association between knee OA and NAFLD. Additionally, we noted that the risk of NAFLD differed according to the severity of knee OA—the risk of NAFLD was 6.331 times higher in the moderate OA group than in the normal group.
However, in patients with severe OA, the risk of NAFLD was not higher than that in patients with moderate OA. The reasons for this lack of association could not be determined from this study. This warrants confirmation in a large-scale study in the future.
A limitation of this study is that it had a cross-sectional design; thus, it was not possible to define causal relationships. Further, the most accurate diagnostic method for NAFLD was not used. In the future, further research is needed to supplement these limitations.
In conclusion, the results of this study suggest that metabolic diseases, such as NAFLD, should be considered in the management and therapy of knee OA. Our findings can be used as the rationale for a new multi-faceted approach for treating knee OA.