NAFLD is regarded as the most common liver disease in the twenty-first century, and it is present if at least 5% of the liver weight is fat without excess alcohol consumption or secondary causes of fat accumulation in the background. Approximately 25% of adults around the world have NAFLD, and the prevalence is still increasing.
The majority of patients in this study had no or mild liver fibrosis [F0: 52 (57.78%), F1: 20 (22.22%)], while 16 patients showed moderate fibrosis [F2: 16 (17.78%)], and only 2 patients showed advanced fibrosis [F4:2 (2.22%)]. These results come against the results of another study done by Fallatah and his colleagues assessing the role of FibroScan compared to other non-invasive assessment scores in 122 Saudi patients with NAFLD. In his study, there was a high percentage of patients showing advanced liver fibrosis by FibroScan [F4: 40 (32.8%)]. These contradicting results can be possibly attributed to demographic differences between patient populations of the two studies, where there is a high prevalence of metabolic syndrome and type 2 diabetes mellitus in the Saudi population, explaining the high prevalence of advanced NAFLD-related liver fibrosis [16].
On the other hand, our results agree with Fallatah et al. study which concluded that there was a significant positive correlation between LSM detected by TE as compared to APRI and FIB-4 results (r = 0.51, r = 0.50, p < 0.001) [16]. This also agrees with Sumida et al. who compared the results of 6 non-invasive markers of liver fibrosis based on data from 576 biopsy-proven NAFLD patients and found the sensitivity and specificity of FIB-4 score for the diagnosis of significant fibrosis was 90% and 64%, respectively, with diagnostic accuracy 87.1% (AUROC 0.871) [22]. Our study also goes with Boursier et al. who found the diagnostic accuracy of FIB-4 score for the diagnosis of significant fibrosis was 70.4% (AUROC 0.704) [23].
Moreover, the current study showed that there was statistically highly significant correlation between NFS score and LSM by TE (r = 0.60, r = 0.53, p < 0.001), which goes with results of another study done by Samy and colleagues who evaluated 60 patients with NAFLD and assessed fibrotest, NFS, FIB-4 score, and LSM by TE in the detection of liver fibrosis depending on liver biopsy and showed there was a statistically significant association between fibrosis and NFS value [24].
We also found that there is statistically highly significant negative correlation between platelet count and LSM by TE (r = − 0.81, r = 0.70, p < 0.001), and this agree with Fallatah et al. who found a strong negative correlation between platelet count and stiffness, as thrombocytopenia in liver disease is associated with advanced fibrosis and even cirrhosis [16].
Moreover, there was a highly significant statistical correlation between ALT, AST, and LSM measured by TE (r = 0.54, r = 0.52, p < 0.001 and r = 0.52, r = 0.59, p < 0.001, respectively), which agrees with Fabrellas and his colleagues who evaluated 215 subjects with metabolic risk factors without known liver disease identified randomly from a primary care center. A control group of 80 subjects matched by age and sex without metabolic risk factors was also studied. CAP and LSM were assessed using TE and found that there was a good statistical correlation between liver transaminases and increased LSM, suggestive of liver fibrosis [25].
As regards GGT, the current study showed a highly significant statistical correlation between GGT and LSM by TE (r = 0.60, r = 0.87, p < 0.001), which goes in accordance with other study done by Mansour et al. who analyzed 108 patients with NAFLD and found a statistically significant correlation between GGT and LSM by TE (r = 0.242, p < 0.05) [26].
As regards liver steatosis grades, the current study showed that most patients had marked hepatic steatosis as demonstrated by CAP [S2: 31 (34.4%), S3: 30 (33.3%)], while the rest showed mild steatosis [S0: 11 (12.2%), S1: 18 (20%)]. This comes against the results of another study by de Lédinghen and his colleagues which concluded that the majority of patients showed no or mild steatosis [S0: 58 (51.8%), S1: 21 (18.8%)], while the rest of the patients showed more advanced steatosis grades [S2: 16 (14.3%), S3: 17 (15.2%)]. The discrepancy in the results between the two studies can be attributed to differences in the study population, where in his study, the mean BMI of patients was 26 kg/m2, while in our current study, the patients had mean BMI of 35.59 ± 5.77. It is clear from this data that the patients in the current study had higher mean body weight and thus are expected to be more liable to hepatic steatosis [17].
We also found that there is a statistically highly significant correlation between FLI and steatosis measured by CAP (r = 0.60, r = 0.53, p < 0.001), and this revealed that FLI has a high discriminatory power in the diagnosis of NAFLD. This result agrees with Motamed and his colleagues who analyzed 5052 subjects and found that there was a significant positive high correlation observed between serum FLI and NAFLD (AUC = 0.8656, 95% CI 0.8548–0.8764) which was also confirmed by binary regression, to the point that a one-unit increase in FLI led to a 5.8% increase in the chance of developing NAFLD and showed good predictive performance in the diagnosis of NAFLD [27]. Additionally, this agrees with Dehnavi et al. who analyzed 212 patients with NAFLD and found that FLI was significantly associated with NAFLD (OR = 1.062, 95%CI 1.042–1.082, p < 0.001), and that mean FLI, BMI, WC, TG, and GGT were all significantly higher in NAFLD patients than in non-NAFLD participants, and that a one unit increase in FLI elevated the chance of developing NAFLD by 6.2% [28].
We also found that there is a statistically significant correlation between GGT and steatosis measured by CAP (r = 0.60, r = 0.53, p < 0.001) which goes with the findings of Dehnavi et al. who concluded that there is a statistically significant correlation between GGT and steatosis (AUC = 0.66, 95%CI = 0.58–0.75, p < 0.001) [28]. This also agrees with Motamed et al. who found that there was a significant positive high correlation was observed between serum GGT and NAFLD (AUC = 0.6927, 95% CI 0.6772–0.7081), p < 0.0001) [27].
We also found that there is a highly significant statistical correlation between TG and serum cholesterol as compared to steatosis measured by CAP (r = 0.56, r = 0.58, p < 0.001, and r = 0.64, r = 0.78, p < 0.001, respectively). This agrees with Kwok et al. who examined 1918 patients with CAP and LSM and found that increased CAP ≥ 222 dB/m was associated with higher body weight, BMI, WC, TG, fasting plasma glucose, and ALT. It was also associated with lower HDL cholesterol [13].
We also found that there is a statistically significant correlation between BMI and WC in comparison to steatosis grades and values obtained by CAP (r = 0.54, r = 0.59, p < 0.028, and r = 0.59, r = 0.62, p < 0.036, respectively), and this agrees with Dehnavi et al. who found that there is a highly significant correlation between BMI and WC and steatosis grades and values (p < 0.001) [28].
The current study also found that there was a statistically significant correlation between DM and steatosis grades and values obtained by CAP (r = 0.46, r = 0.49, p < 0.026), which goes with Kwok et al. who found that there is a significant positive high correlation observed between serum fasting blood glucose and steatosis and that around 32–62% of diabetic patients were found to have NAFLD [13].
We also found that the best cut-off value for fibrosis detection by TE (LSM) vs. NFS is 4.10 KPa, which fulfills the highest sensitivity, specificity, and accuracy (Table 4). A study by Samy and his colleagues found that, depending on liver biopsy, the sensitivity, specificity, and accuracy of NFS to detect liver fibrosis are good, with AUROCs of 0.94. For mild fibrosis, the sensitivity, specificity, and accuracy of NFS was 89.47%, 90.24%, and 94.7%, respectively. On the other hand, the sensitivity, specificity, and accuracy of NFS in cases of severe liver fibrosis were found to be 100%, 89.8%, and 98.1%, respectively [24].
We also found that the best cut-off value for fibrosis detection by TE (LSM) vs. FIB-4 score is 6.95 KPa, which fulfills the highest sensitivity, specificity, and accuracy (Table 4). This also agrees with the results obtained by Samy et al. who concluded that, depending on liver biopsy, the sensitivity, specificity, and accuracy of FIB-4 score to detect liver fibrosis are good, with AUROCs of 0.992 (94.7%, 97.6%, and 99.2%, respectively) [24]. Similarly, another study by Sumida et al. found the sensitivity and specificity of the FIB-4 score for the diagnosis of significant fibrosis was 90% and 64%, respectively, with diagnostic accuracy 87.1% [22].
Our results show that the best cut-off value for fibrosis detection by TE (LSM) vs. APRI score is 4.5 KPa, which fulfills the highest sensitivity, specificity, and accuracy (Table 4). This agrees with Kolhe et al. who analyzed histological and clinical data of 100 consecutive urban slum-dwelling patients with NAFLD and showed that APRI had sensitivity, specificity, accuracy, PPV, NPV, and AUROC of 85.2%, 87.7%, 95%, 58.33%, 96.05%, and 0.95, respectively, with a statistically high significant correlation between APRI and biopsy-proven fibrosis [29].
Our study shows that the best cut-off value for fibrosis detection by TE (LSM) vs. APRI, FIB-4, and NFS scores has an overall average of 5.2 Kpa, which fulfills the highest sensitivity, specificity, and accuracy (85.30%, 47.70%, and 85.48%, respectively, AUC 0.742) (Fig. 1). Similarly, Önnerhag and his colleagues who included 144 patients with biopsy-proven NAFLD showed that FIB-4-index had the highest NPV (91%) and APRI the highest PPV (71%). The AUROC for FIB-4-index, NFS, and APRI acceptably predicted advanced fibrosis with values between 0.81 and 0.86 [30].
Our study results are close to Hashemi et al. who performed a meta-analysis that enrolled the literature published about LSM detected by TE for the diagnosis and staging of NAFLD and found the sensitivity and specificity of FibroScan in the detection of fibrosis to be 87.5% and 78.4%, respectively [31]. This also agrees with Boursier et al. who evaluated the diagnostic accuracy of LSM by TE in a cross-sectional study including 452 NAFLD patients; found that its accuracy was 83.1% [23].
Our study also agrees with Aykut et al. who compared the diagnostic performances of three different non-invasive methods including TE for the detection of liver fibrosis in a total of 88 patients with biopsy-proven NAFLD and found the diagnostic accuracy 90.2% [32].
Our results show that the best cut-off value for steatosis detection by CAP vs FLI score is 220.5 dB/m, which fulfills the highest sensitivity, specificity, and accuracy (Table 5) (Fig. 2). Motamed et al. also showed that FLI showed good performance in the diagnosis of NAFLD with accuracy equal to 86.56% (AUC = 0.8656) and revealed that FLI has a high discriminatory power in the diagnosis of NAFLD [27]. This could be somewhat anticipated due to the fact that FLI is composed of four quantities related to NAFLD, including BMI, WC, GGT, and TG. A high BMI or WC, the main obesity indices, is considered an essential risk factor for NAFLD, and the prevalence of NAFLD substantially increases in obese individuals.
Similarly, Dehnavi et al. investigated the relationship between FLI and NAFLD based on logistic regression and their findings revealed a highly significant positive relationship between FLI and NAFLD, so that even a one unit increase in FLI elevated the chance of developing NAFLD by 6.2% (OR = 1.062, 95%CI 1.042–1.082, p < 0.001). Even after adjusting for confounding factors such as sex, age, diastolic blood pressure (DBP), FBS, ALT, and LDL, the logistic regression analysis showed a significant positive association between FLI and NAFLD (OR = 1.059, 95%CI 1.035–1.083, p < 0.001) [28].
These findings go also with Siddiqui et al. who performed a prospective study of 393 adults with NAFLD who underwent VCTE within 1 year of liver histology analysis and found that the CAP value is positively associated with severity of hepatic steatosis and the cross-validated AUROC is 76% for classifying patients with ≥ 5% steatosis on histology [33]. This also goes with Eddowes et al. who evaluated 450 patients and assessed the diagnostic accuracy of CAP and LSM against liver biopsy and found that CAP by TE is accurate non-invasive methods for assessing liver steatosis in patients with NAFLD with an AUROC of 0.87 (95% CI 0.82–0.92), sensitivity of 0.80, and specificity of 0.83 [34].