Crohn’s disease is a remitting and chronic disease that often involves the small intestine[18]. Thus, a complication of malnutrition and malabsorption is frequently occurred. However, since a study first reported an association between “an expanded fatty liver” and “ulceration of the colon” in 1873[19], there has been several studies conducted to further study the link between the diseases. The prevalence of NAFLD in CD is reported to vary between 8.2–39.5%[5][20][21], depending on the diagnostic tools and criteria applied. And previous studies have proposed several factors associated with the NAFLD, including ethnic, BMI, disease duration[10], anti-TNF-α therapy[21] and etc. However, these studies were mostly conducted in a western population.
Our study recruited a group of Chinese patients diagnosed with CD, who completed MR during hospitalization, based on which the liver fat was imaging and calculated. Several implications can be highlighted from this research. Firstly, the prevalence of non-alcoholic liver disease is estimated to be 24.8% in CD patients, according to the standard of MR-PDFF > 5.5%, higher than the overall prevalence of 17% in China and Japan area[22]. This is consistent with a previous study reported in Japan, which reported a prevalence of 21.8%[20]. Although the age between groups did not differ significantly, there was a trend that the patients with NAFLD were older than the other group. CD patients with NAFLD presented longer disease duration, which is in accordance with the previous study[6]. This may be explained by the dysbiosis of gut bacteria due to the intestinal inflammation. Besides, the NAFLD group had a higher BMI and more frequent presence of metabolic syndrome.
Another finding is that there is no significant difference in the long-term use of anti-TNF-α between groups. Ambiguous results have been reported in the previous research[23][21]. However, both studies did not evaluate the long-term effect. In this study, we specifically document the times of patients receiving anti-TNF-α therapy, and did not find significantly difference in the occurrence of NAFLD in patients with prolonged anti-TNF-α use.
Liver fat content can be accurately and noninvasively detected through imaging examinations, such as ultrasound, computed tomography or magnetic resonance imaging. Among these tests to diagnose NAFLD, ultrasound and CT share close sensitivity and reliably identify steatosis when the liver comprises ≥20% of fat. Although MRI can identify as little as 5% steatosis, the cost is much more higher[24]. Therefore, a clinical classifier is needed to evaluate the risk for NAFLD in a convenient, cost-effective manner, then to elevate patient compliance.
Applying logistic regression analysis, our model incorporates six parameters: BMI, prealbumin, ALT, AST, GGT and duration. The model has a high AUC value, which is indicative of greater sensitivity and specificity. The Hosmer-Lemeshow test shows that the model’s predicted value does not differ significantly from the actual one. The C-index value is 0.753, indicating good consistency of our model. The model’s calibration slope demonstrates a value of 0.834 after bias correction. The nomogram is a method with superiority widely used by oncologists to predict tumor prognosis. In current research, we have successfully created a visualized nomogram to predict the risk of developing NAFLD.
The major advantages of our study include the diagnosis of NAFLD based on MR-PDFF, use of a large cohort of CD patients, first to evaluate the long-term effect of use of anti-TNF-α on NAFLD. Nevertheless, several limitations need to be addressed. First, due to the retrospective nature, we defined NAFLD through chart review rather than prospectively collected the liver content data. Second, this is a monocentric research, which may limit the application of the diagnostic model. Further studies are warranted to assess if and how the immune dysfunction plays a role in liver steatosis.