4.1. Main Findings
Ultrasonography had fair-to-good sensitivity and good specificity for the diagnosis of NAFLD in 23-year-old males and females across all ranges of nutritional status when compared with MRS. An area under curve of 86% indicates that this method has good diagnostic performance in the prediction of liver fat infiltration >5% in this age group. Analytical validation also showed that ultrasound was well suited to discriminate individuals with a cardiometabolic profile of high risk. Several authors highlight the potential of ultrasound in clinical contexts, in spite of reckoning its limitations [9
]. Combined with a scoring system, the ultrasound method might serve as an initial screen for fatty liver in those at higher risk of having the condition, for instance people with: overweight or obesity, insulin resistance, prediabetes or type-2 diabetes, and high levels of serum triglycerides [9
]. Second, it may serve to track the progression of the disease as well as grade its severity [10
]. Last, ultrasound may help target and monitor interventions aiming at reducing the degree of liver steatosis and/or the cardiometabolic complications related to NAFLD [9
We used MRS as the gold standard to determine liver fat since this is an accurate method, and liver biopsies could not be performed for ethical reasons in this group of participants without alterations in liver function. A recent study conducted among adolescents in whom a liver biopsy was performed, showed that MRS has a sensitivity and specificity of 92% and 95%, respectively, for the diagnosis of NAFLD [29
]. A meta-analysis by Bohte et al. published in 2010 showed similar results [14
]. Likewise, after revision of 49 studies including n
= 4720 participants, Hernaez et al. found that ultrasound was reliable and accurate in detection of moderate and severe NAFLD, compared to biopsy, with an area under the summary receiving operating characteristics curve of 0.93 (0.91–0.95) [9
Bohte et al. also found that ultrasound has better diagnostic accuracy than computed tomography (CT) scans. Other studies evaluating the accuracy of CT yielded similar results. A meta-analysis comparing CT with liver biopsies in liver donors showed that CT had a sensitivity and specificity of 0.81 and 0.94, respectively, and that the accuracy of CT increased when fat infiltration was higher [30
]. Similarly, CT had a lower concordance with liver biopsies than ultrasound [31
]. A semi-quantitative CT method, comparing liver and spleen radiological attenuation, could increase the precision of the technique [32
]. This method was used in severely obese patients (BMI > 40 kg/m2
). CT liver-spleen index was strongly correlated (−0.8; p
< 0.001) to liver triglycerides, confirming the utility of using CT scanning to non-invasively evaluate the extent of liver fat infiltration in very high BMI patients. Therefore, the evidence shows that CT is not superior to ultrasound for the diagnosis of NAFLD. Considering that the later technique is far less expensive and avoids exposure to radiation, it should be preferred.
Ultrasound has the disadvantage of being operator dependent, since its results may vary according to the amount of gel used, the pressure exerted or the position of the probe [33
]. The Hamaguchi score was designed to provide a numerical value for subjective parameters such as liver brightness or diaphragmatic attenuation. The two observers who determined the score in the present study had a moderate degree of concordance, providing evidence that this ultrasound method to determine liver fat overcomes interobserver variability. However, the concordance with MRS increased, notably when an agreement was reached between the raters. Therefore, having more than one observer may improve the accuracy of ultrasound. A recent report about liver fat assessment by ultrasound also showed a substantial level of concordance between observers when a similar semi-quantitative scoring method was used [34
]. However, since the concordance between observers is not perfect, a good practice, especially for research purposes, is to have more than one assessment of images to reach an agreement when the scores are discordant.
When comparing participants with and without NAFLD, the former had higher BMI, serum lipids and insulin. The differences were more marked among men than women. This is in line with evidence describing a sexual dimorphism in the cardiometabolic profile, with males, generally showing less beneficial profiles [35
4.2. Implications for Practice
Overweight and obesity are associated with a substantial risk of NAFLD in both older and younger age populations. The greater the severity of excess weight, the higher the risk of liver dysfunction, particularly among males. Clinicians should carefully examine patients who are overweight or obese to identify NAFLD early. Although elevated alanine transaminase has been proposed as a surrogate of NAFLD [36
], in youths histologic studies and imaging procedures show that fatty liver is present in individuals with obesity whether the liver enzymes are elevated or not, suggesting that alanine transaminase elevation occurs at more advanced stages of NAFLD or in patients with extreme obesity [38
]. Thus, abdominal ultrasound might be especially useful for screening of NAFLD in mild stages of the disease or in patients with obesity but still normal alanine transaminase levels. In our sample, alanine transaminase was slightly above normal (58.3 IU/L) only in males with NAFLD.
Likewise, significant risk factors for NAFLD are obesity, insulin resistance and the cardiometabolic risk factors that define the Metabolic Syndrome. The prevalence of these conditions has risen dramatically over the past years in both industrialized and non-industrialized countries, even among young adults. In Chile, 40% of 20–29 years-old have obesity, 14% have Metabolic Syndrome, one in three has low high-density lipoprotein cholesterol, and one in five has hypertriglyceridemia [4
]. The number of young adults exposed to NAFLD is growing and, thus, more affordable and accessible methods for NAFLD screening in large groups are needed. This semi-quantitative ultrasound method to determine liver fat infiltration could serve that purpose which might help to discriminate individuals with mild NAFLD, for whom lifestyle interventions may have the potential to improve liver function. Moreover, screening of NAFLD using ultrasound could be done for preventive purposes in subjects with overweight.
4.3. Study Limitations
A major limitation of this study was the use of a relatively small sample (n
= 60) to conduct the clinical validity assessment. Yet the number of participants was sufficient to address our main aim. Similar studies used similar or smaller sample sizes [15
], and the analytical assessment was conducted to tackle this weakness, though the sample had less obesity, insulin resistance and Metabolic Syndrome. A larger sample might further clarify the predictive role of alanine aminotransferase in NAFLD among males. A further limitation was the use of a single ultrasound instrument, which does not allow assessing the variability that would be introduced by using a different ultrasound machine. The restricted age group might also limit the validity of our findings. However, the prevalence of NAFLD has substantially risen in young adults, a group where the disease often goes unrecognized and, if untreated, can progress eventually to steatohepatitis or cirrhosis before the age of 40 [6
]. Lastly, due to the cross-sectional nature of our study, we were not able to assess the progression of NAFLD and confirm its relationship with biochemical and anthropometric markers over time.