If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
To establish a simple economical diagnostic tool for prediction of hepatic steatosis in patients with hepatitis B virus (HBV) infection.
Methods
From January 2006 to January 2015,a total of 1325 consecutive subjects who underwent liver biopsy were enrolled. According to the results of multivariate logistic regression analysis, a new nomogram was conducted. Then discrimination and calibration were conducted to assess the clinical diagnostic value of nomogram.
Results
The nomogram consisted of age, triglyceride (TG), low-density lipoprotein (LDL), uric acid (UA), haemoglobin (HGB). For prediction of hepatic steatosis, the AUROC of nomogram was 0.792 (95%CI: 0.758–0.826). With cut off value of 0.11, 699 (52.8%) of 1325 patients could be free from liver biopsy with a correct rate of 95.3% for diagnosis of hepatic steatosis.
Conclusion
The nomogram for hepatic steatosis has a better clinical diagnostic value for prediction of hepatic steatosis in patients with HBV infection. From the perspective of cost-effectiveness and clinical practice, it is worth considering the use of the nomogram as a mass screening tool before further liver biopsy or imaging examinations.
]. Then the coexistence of HBV infection and NAFLD is becoming more frequent. It was reported that NAFLD patients had an increased overall mortality for cardiovascular disease, extra-hepatic malignancies and liver disease [
]. As an inexpensive noninvasive method, ultrasonography (US) is recommended in detection of hepatic steatosis. However, US is limited by low sensitivity for mild steatosis and inability to differentiate mild fibrosis from steatosis in clinical practice [
]. Magnetic resonance imaging (MRI), computed tomography (CT), and transient elastography (TE) have a better diagnostic value in detecting of hepatic steatosis. However, these methods are limited by several shortages in clinical practice. First, these methods are too expensive for routine health examinations. Second, these methods are not readily available in most hospitals of developing countries. Third, considering cost-effectiveness, MRI, CT, and TE are not suitable for massive outpatient screening in predicting hepatic steatosis. There is an intense need to find an simple, economical, easier practical, and readily available tool for massive health screening as an alternative to imaging examinations or liver biopsy.
To conduct a simple economical method for massive screening for hepatic steatosis as an alternative to imaging examinations or liver biopsy, we performed this retrospective study in patients with HBV infection.
2. Materials and methods
2.1 Patients
Subjects of this study included 1580 consecutive patients who had been diagnosed with HBV infection and had undergone liver biopsy in department of infectious diseases of Shunde First People's Hospital between January 2006 and January 2015. The Patients were enrolled based on the following criteria: chronic hepatitis B defined as hepatitis B surface antigen (HBsAg) positivity for more than 6 months; detectable HBV-DNA with a level >103 copies/ml. The exclusion criteria were as follows: liver cancer or co-infection with hepatitis C virus, hepatitis D virus or human immunodeficiency virus; autoimmune liver diseases such as autoimmune hepatitis, primary biliary cirrhosis, and primary sclerosing cholangitis; alcohol ingestion in excess of 20 g/day. The patients with missing data in terms of age, triglyceride (TG), low-density lipoprotein (LDL), uric acid (UA), and haemoglobin (HGB) were ruled out.
Therefore, there were 255 (16.1% of the total subjects) patients excluded from the study according to above criteria. There were no significant differences in terms of demographic and clinical parameters between patients included and excluded (data not shown). Finally, a total of 1325 patients (944 males and 381 females) were recruited into the study. All patients were informed and written consents were obtained before inclusion. The study protocol was approved by the Ethics Committee of the Shunde First People's Hospital.
2.2 Liver biopsy
Liver biopsies were performed by two experienced physicians using a 16-gauge needle (16G biopsy Menghini's needle, ShangHai). A minimum of 1.5 cm of liver tissue with at least 7 portal tracts was required for diagnosis. The specimens were fixed, paraffin-embedded and stained with haematoxylin and eosin (HE). Histological grading of necro-inflammation (G0–G4) and staging of the liver fibrosis (S0–S4) were carried out according to Scheuer's classification [
] by one experienced pathologist blinded to the clinical data. Hepatic steatosis was graded according to the percent of hepatocytes affected: none (<5%), mild steatosis (5–32%), moderate steatosis (33–65%), and severe steatosis (≥66%) [
]. Steatosis group was defined as steatosis involving more than or equal 5% of hepatocytes and non-steatosis group was defined as steatosis involving less than 5% of hepatocytes (Fig. 1).
Fig. 1Pathological characteristics of hepatic steatosis in patients with hepatitis B virus infection (HE staining). (a) None steatosis (200×). (b) Mild steatosis (200×). (c) Moderate steatosis (200×). (d) Severe steatosis (200×).
All patients systematically underwent complete biochemical workups, US and liver biopsy within 2 days. Blood samples of the subjects were obtained before liver biopsy. Biochemical tests were performed by commercial assays in our hospital laboratory. The serum HBV-DNA level was detected with a Real-Time polymerase chain reaction (PCR) System (ABI7700; Applied Shenzhen city Daeran Biological Engineering Co., Ltd., Shenzhen, Guangdong, CHN).
2.4 Statistical analysis
Continuous data were expressed as mean ± SD or median (quartile range) depending on the normality of the data. Categorical variables were expressed as proportions. Continuous variables were compared with Student's t-test, one-way ANOVA analysis of variance or Kruskal–Wallis H test, depending on the normality of the data; Nominal or ordinal variables were compared with Chi-square test or Kruskal–Wallis H test.
All variables that significantly associated with hepatic steatosis in univariate logistic regression analyses were included in forward stepwise multivariate logistic regression analysis for establishment of nomogram model. According to the results of multivariate logistic regression analysis, a new nomogram was conducted.
The discriminatory ability of the nomogram was measured by means of the area under the receiver operating characteristic curve (AUROC). The optimal cut off value for clinical utility was determined according to positive likelihood ratio (PLR) ≈ 10.0 for confirming diagnosis of steatosis and negative likelihood ratio (NLR) ≈ 0.1 for excluding diagnosis of steatosis [
Users’ guides to the medical literature. III. How to use an article about a diagnostic test. A. Are the results of the study valid? Evidence -Based Medicine Working Group.
]. To furtherly evaluate the clinical utility of new model, the sensitivity (Se), specificity (Sp), positive likelihood ratio (PLR), negative likelihood ratio (NLR), positive predictive value (PPV), and negative predictive value (NPV) were calculated using the ROC curve. Calibration was assessed by plotting the model predicted probability against the observed proportion of hepatic steatosis. Bootstrapping was used to correct for over fitting bias for calibration.
Statistical analyses were performed using SPSS 13.0 (SPSS Inc., Chicago, IL) and software R (http://www.R-project.org) [
], with ‘Hmisc’ and ‘rms’ packages added. P < 0.05 was considered statistically significant.
3. Results
3.1 Baseline characteristics of subjects without and with hepatic steatosis
A total of 1325 patients were included in the study, 197 (14.9%) of whom were diagnosed as hepatic steatosis by mean of liver biopsy. Of the 1325 subjects, 944 (71.2%) were male and 381 (28.8%) were female. All variables were compared according to presence of hepatic steatosis. The baseline characteristics of patients without and with hepatic steatosis were summarized in Table 1.
Table 1Characteristics of subjects with and without hepatic steatosis.
3.2 Performance of US in patients with steatosis diagnosed by biopsy
In all 197 patients with steatosis diagnosed by liver biopsy, fatty liver could be detected by US in 81 (50.9%) patients out of 159 patents with mild steatosis, 18 (72.0%) patients out of 25 patents with moderate steatosis, and 12 (92.3%) patients out of 13 patents with severe steatosis respectively (P < 0.05).
3.3 Multivariate logistic regression analysis and nomogram for hepatic steatosis
Multivariate logistic regression analysis (Table 2) showed that age (OR = 1.043), HGB (OR = 1.040), LDL (OR = 1.866), TG (OR = 2.131) and UA (OR = 1.002) were significantly associated with hepatic steatosis in patients with HBV infection. According to the results of multivariate logistic regression analysis (Table 2), we constructed a new diagnostic model (Fig. 2) named nomogram for hepatic steatosis. The model consisted of age, HGB, LDL, TG and UA, which were significantly correlated with hepatic steatosis.
Table 2Multivariate logistic regression analysis for prediction of hepatic steatosis.
The nomogram could be used by determining predictive values indicated by the variable value scale corresponding to different variables. The values for each predictive variables were determined and summed to calculate the total value. Then the total value was projected to the probability of hepatic steatosis scale locating in the bottom axis to determine the probability of having hepatic steatosis.
4. Diagnostic accuracy of nomogram for prediction of hepatic steatosis
The AUROC of this nomogram was calculated to assess the diagnostic accuracy for prediction of hepatic steatosis (Fig. 3). The AUROC of the nomogram in predicting of hepatic steatosis was 0.792 (95%CI: 0.758–0.826, P < 0.001), which indicated that the nomogram had a fair discrimination for hepatic steatosis in patients with HBV infection.
Fig. 3Receiver operating characteristic (ROC) curve of nomogram in predicting the presence of hepatic steatosis. AUROC, area under the receiver operating characteristics curve.
The further calibration curve was showed in Fig. 4. A calibration plot compares the model's predicted probabilities and observed proportions. The diagonal line reflects the ideal situation (predicted probability = observed proportion). The calibration curve (Fig. 4) showed that the nomogram model appeared to be well-calibrated and there was a good agreement between the observed and predicted probabilities of hepatic steatosis.
Fig. 4Internal calibration of nomogram for hepatic steatosis. Calibration plot showing good agreement between the predicted and observed probabilities for hepatic steatosis in patients with HBV infection.
4.1 Clinical utility and significance of nomogram for prediction of hepatic steatosis
To estimate the clinical utility of the nomogram for prediction of hepatic steatosis in patients with HBV infection, the optimal cut off values were determined according to positive likelihood ratio (PLR) ≈ 10.0 for confirming diagnosis of hepatic steatosis and negative likelihood ratio (NLR) ≈ 0.1 for excluding diagnosis of hepatic steatosis [
Users’ guides to the medical literature. III. How to use an article about a diagnostic test. A. Are the results of the study valid? Evidence -Based Medicine Working Group.
Among the 1325 patients included in the study, 699 (52.8%) patients’ nomogram values were lower than 0.11 and 51 (3.8%) patients’ nomogram values were higher than 0.50. To exclude hepatic steatosis, the cut off value of 0.11 showed a LR− of 0.28, a NPV of 95.3%, and a sensitivity of 83.3%. These patients with a nomogram value lower than 0.11 could be excluded from hepatic steatosis with 4.7% (33/699) missing diagnosis. In the 33 patients with a normogram value <0.11 and diagnosed with steatosis by liver biopsy, 27 (81.8%) patients was diagnosed with mild steatosis, 4 (12.1%) patients with moderate steatosis, and 2 (6.1%) patients with severe steatosis respectively.
To diagnose hepatic steatosis, the cut off value of 0.50 showed a LR+ of 11.1, a PPV of 66.0%, and a specificity of 98.5%. These patients with a nomogram value higher than 0.50 could be considered hepatic steatosis with 33.3% (17/51) misdiagnosed as hepatic steatosis.
With cut off value of 0.11, 699 (52.8%) of 1325 patients could be free from liver biopsy with a correct rate of 95.3% for diagnosis of hepatic steatosis (Fig. 5).
Fig. 5Flow diagram of nomogram for hepatic steatosis before further liver biopsy or imaging examinations in original group.
4.2 Subgroup analysis in patients with different specimen length
The mean length of liver specimen was 2.29 cm with a SD value of 0.50. In all 1325 patients, 1063 (80.2%) patients had a specimen length not shorter than 2.0 cm and 651 (49.1%) patients had a specimen length not shorter than 2.5 cm.
To explore the performance of nomogram in patients with different specimen length, all patients were separated into three groups [group 1 (1.5–1.99 cm), group 2 (2.0–2.49 cm), and group 3 (≥2.5 cm)] according to the length of liver specimen. The reduction in length of liver specimen led to an underestimation of inflammation and fibrosis. Severe inflammation (defined as inflammation grade ≥3) increased from 36.7% in group1 to 43.2% in group 2, and 56.5% in group 3 (P < 0.001). Like grading of inflammation, Severe fibrosis (defined as fibrosis stage ≥3) increased from 40.4% in group1 to 43.2% in group 2, and 59.3% in group 3 (P < 0.001).
There was no significant difference in term of steatosis between three groups (using Kruskal–Wallis H test, P = 0.128). The further analyses between every two groups demonstrated that differences were not statistically significant while group 1 vs group 2 (Hc value = 0.54, P = 0.592) and group 1 vs group 3 (Hc value = 1.344, P = 0.179), while there was significant difference between group 2 and group 3 (Hc value = 2.22, P = 0.026). There was significant difference between patients with liver specimen shorter than 2.5 cm and not shorter than 2.5 cm (Hc value = 1.98, P = 0.048), while there was no significant difference between patients with liver specimen shorter than 2.0 cm and not shorter than 2.0 cm (Hc value = 0.56, P = 0.572) (Table 3).
Table 3Subgroup analysis in patients with different specimen length.
The AUROCs of nomogram for hepatic steatosis were 0.751 (95%CI: 0.652–0.851), 0.779 (95%CI: 0.708–0.850), and 0.820 (95%CI: 0.774–0.866) in group 1, group 2, and group 3 respectively.
5. Discussion
A nomogram consisted of age, HGB, LDL, TG and SUA was established for prediction of hepatic steatosis in patients with HBV infection, with an AUROC of 0.792. With cut off value of 0.11, 699 (52.8%) of 1325 patients could be free from liver biopsy with a correct rate of 95.3% for diagnosis of hepatic steatosis. The subgroup analysis demonstrated that AUROCs of nomogram for hepatic steatosis were 0.751, 0.779, and 0.820 in group 1 (prevalence of steatosis = 14.9%), group 2 (prevalence of steatosis = 13.3%), and group 3 (prevalence of steatosis = 18.7%) respectively.
Katharine Eng et al. developed a non-invasive predictive model with an AUROC of 0.737 for non-alcoholic steatohepatitis in children with NAFLD. By utilizing this nomogram with lower cut off value of 40% and higher cut off value of 75%, 163 (54.0%) subjects out of 302 subjects would have avoided liver biopsy [
]. To our knowledge, our study was the first one that suggested a nomogram for prediction of hepatic steatosis in patients with HBV infection.
This nomogram consisted of age, HGB, LDL, TG and SUA. The influences of these variables for fatty liver had been reported in previous studies. TG had been reported to be an independent factor for hepatic steatosis in general population [
]. Ryan CK et al. reported that US could only detect 55% and 72% of patients with hepatic steatosis 10–19% and 20–29%, respectively; especially, US could not detect hepatic steatosis when steatosis presented in less than 10% of hepatocytes [
]. Our analysis showed that US could only detect 50.9%, 72.0% and 92.3% of patients with mild steatosis, moderate steatosis, and severe steatosis. The conclusion in our study confirmed that US had only a low sensitivity for mild steatosis and Clinicians should be careful in interpreting the accuracy and clinical significance of the results of US.
It was reported that liver specimen size strongly influenced the grading and staging in patients with chronic viral hepatitis [
]. Our analysis showed also that the number of cases with severe inflammation and severe fibrosis significantly decreased in patients with shorter specimens. Therefore, it must be taken into account that the impact of specimen length on inflammation and fibrosis while we interpreted the results of the study.
It had been reported that the length of biopsy specimen was correlated with percentage of patients with definite non-alcoholic steatohepatitis (56% and 65% in biopsies measuring 1.5–2.4 cm and ≥2.5 cm, respectively; P < 0.001) [
]. Our study showed that there was significant difference between patients with liver specimen shorter than 2.5 cm and not shorter than 2.5 cm,whereas there was no significant difference between patients with liver specimen shorter than 2.0 cm and not shorter than 2.0 cm. The conclusion in our study demonstrated that specimen length not shorter than 2.5 cm was more worthy of recommendation to conduct a more convincing conclusion of diagnostic accuracy of clinical predictive model.
In our study, there were 200 (19.8%) patients with a length of liver specimen shorter than 2.0 cm. Taking into account the impact of specimen length on the accuracy of nomogram for hepatic steatosis, we performed a further subgroup analysis according to the length of liver specimen. The subgroup analysis demonstrated that AUROCs of nomogram for steatosis were 0.751, 0.779, and 0.820 in these three groups respectively. The patients with specimen length shorter than 2.0 cm would reduce the diagnostic accuracy of nomogram for hepatic steatosis. The differences of nomogram for steatosis in three groups may be associated with prevalence of steatosis and the sampling error of the liver biopsy. Even if in the patients with specimen length shorter than 2.0 cm, nomogram could gain a AUROC of 0.751, which indicating a moderate diagnostic accuracy for hepatic steatosis. Supposing that all patients had a specimen length not shorter than 2.5 cm, nomogram would gain a AUROC of 0.820, indicating a fair diagnostic accuracy for hepatic steatosis.
We considered that this nomogram was a good choice for massive screening in detecting hepatic steatosis as an alternative to liver biopsy or examinations for the following reasons. Firstly,the nomogram provided a individualized risk prediction,which was easy to obtain from simple clinical parameters for both physicians and patients. Secondly,all relevant parameters of the nomogram were readily available in routine health examination with no additional cost. Thirdly,the nomogram was easily applicable for clinical practice because the nomogram did not need additional equipments such as TE, CT, or MRI, which was of importance for most hospitals in developing countries. Lastly, there were 699 (52.8%) of 1325 patients could be free from liver biopsy with cut off value of 0.11. Patients with nomogram value more than 0.11 would be left undiagnosed and needed further imaging examinations or liver biopsy. It was worth considering utilizing this nomogram as a massive screening tool in selecting patients for further imaging examinations or liver biopsy. In summary, from the perspective of cost-effectiveness and clinical practice, the nomogram was suitable to be a massive screening tool for hepatic steatosis as a simple, economical, easier practical and readily available method, which could reduce the need of liver biopsy and cost of health routine examination.
The present study had two unique feathers: firstly, hepatic steatosis was all diagnosed by liver biopsy in our study, which provided more precise diagnostic results of hepatic steatosis than imagine examinations. Secondly, all subjects of the present study were patients with HBV infection so that the nomogram was more suitable for patients with HBV infection in detecting of hepatic steatosis.
There were three limitations in the present study. The first limitation was that there were 255 (16.1%) patients excluded from this retrospective study for exclusion criteria, which might affect the reliability of the conclusions. The second limitation of our study was that some important variables, for example body mass index (BMI) and homeostasis model assessment-IR (HOMA-IR), did not include in the analysis because that the study was a retrospective study. BMI had been found to be an independent factor for hepatic steatosis in previous studies [
]. Future studies in the field of hepatic steatosis should include these important variables which may strongly affect the diagnostic accuracy of clinical predictive model. The third limitation was that the prevalence of hepatic steatosis in our study was 14.9%, which was lower than that of some previous studies [
]. It must be taken into account that the impact of prevalence on the diagnostic accuracy of clinical predictive model.
In summary, this nomogram for hepatic steatosis has a better clinical diagnostic value for prediction of hepatic steatosis in patients with HBV infection. From the perspective of cost-effectiveness and clinical practice, it is worth considering that nomogram utilizes as a massive screening tool before further liver biopsy or imaging examinations.
Conflict of interest
None declared.
References
Petta S.
Muratore C.
Craxì A.
Non-alcoholic fatty liver disease pathogenesis: the present and the future.
Users’ guides to the medical literature. III. How to use an article about a diagnostic test. A. Are the results of the study valid? Evidence -Based Medicine Working Group.
30270-X&id=gr1.jpg){kind=link}
30270-X&id=gr2.jpg){kind=link}
30270-X&id=gr3.jpg){kind=link}
30270-X&id=gr4.jpg){kind=link}
30270-X&id=gr5.jpg){kind=link}