The Predictive Value of Mean Platelet Volume for Liver Fibrosis in Children With Chronic Liver Diseases

Introduction: Almost all causes of chronic liver damage can culminate in liver fibrosis and ultimately cirrhosis. Studies have suggested a relationship between mean platelet volume (MPV) and liver fibrosis; however, this needs confirmation by further studies. We here assessed the predictive value of MPV for liver fibrosis in children with chronic liver diseases. Methods: In this study, children <18 years old with chronic liver diseases referred to the Nemazee Hospital of Shiraz during 2013-2016 were studied. The patients underwent liver biopsy for assessing liver fibrosis. Statistical analyses were conducted in SPSS 23. Results: From 368 studied children, 52.2% were boys. The patients’ mean age was 4.5±3.9 years old. Most patients had grade 6 fibrosis (36.7%). Cryptogenic (42.7%) was the most common cause of chronic liver disease, and jaundice was the most prevalent clinical presentation (53%). There was a significant association between the liver fibrosis and MPV (P = 0.025). Conclusion: MPV was significantly different between patients with different severities of liver fibrosis. However, assigning an appropriate cut off value to distinguish different degrees of fibrosis requires more studies.

presenting clinical symptoms, results of laboratory tests (serum albumin, bilirubin, liver enzymes, sodium, INR creatinine, white blood cells and platelets counts, red cell distribution width and MPV, MELD/PELD (model for end-stage liver disease/pediatric end-stage liver disease) scores, Child-Pugh score, and results of liver biopsy. In addition, non-invasive fibrosis markers (AAR, APRI, and FIB-4) were calculated. These data were recorded using patients' medical archives.
The data were analyzed by SPSS version 23 using descriptive and analytical statistics. Kruskal-Wallis test, Spearman correlation, and ROC curve analysis were used.

Results
In this study, 368 patients with chronic liver diseases who had been undergone liver biopsy were assessed. From these, 192 (52.2%) and 176 (47.8%) were males and females, respectively. The mean age of the patients in this study was 4.5±3.9 years old.
Considering the underlying diseases, 84 (22.8%) of patients had biliary atresia as the second most common reason after unknown causes ( Table 1). The most common presenting clinical symptom was jaundice (n=195; 53%). Table 2 demonstrates clinical symptoms of the patients. A summary on laboratory features of the patients has been demonstrated in Table 3. Most of the patients revealed severe fibrosis with grades 5 (n=33;9%) and 6 (n=135; 36.7%) ( Table 4). There were no association between the severity of fibrosis with underlying causes of liver disease ( Table 5).
Considering the severity of liver fibrosis values based on the underlying disease, MPV showed significant difference only in patients with neonatal hepatitis (P = 0.020) ( Table 7). Statistically significant difference was found between MPV values in children with different Child score (P = 0.002).
There was no significant statistical difference between AAR (AST/ALT ratio) and the grade of fibrosis (P>0.05). However, APRI was significantly correlated with the severity of liver fibrosis (P < 0.001). The FIB-4 was significantly associated with fibrosis grade (P < 0.001) and   Figure 1A). Also, the optima cut off value for diagnosis of moderate and severe fibrosis (Stage ≥3) was obtained MPV ≥3 to 10.9 with AUC=0.548 and sensitivity of 34.8% and specificity of 74.8% ( Figure 1B). The cut off MPV for detecting any fibrosis was obtained >10.8 with AUC=0.544, sensitivity of 35.7% and specificity of 76.1% ( Figure 1C).
The results of ROC curve analysis showed that MPV cut off point of 10.4 rendered AUC=0.582 and sensitivity and specificity of 52.4% and 65% for differentiating severe fibrosis (S≥5) than other stages (S<5). For differentiating    These results show that MPV is not a good marker for predicting liver fibrosis. MPV had no significant association with neither age nor underlying diseases.
In reviewing other markers of liver fibrosis, no significant difference was found in AAR among different severities of liver fibrosis (P>0.05). However, APRI was significantly different among various stages of liver fibrosis (S0=2.09, S1=2.55, S3=2.27, S6=8.36, P < 0.001). Also, APRI was significantly different among patients with different severities of liver fibrosis (mild, moderate and severe) with P < 0.001 that the difference between the mild fibrosis (mild=2.49, moderate=2.34, and severe=7.09, P = 0.001). Therefore, APRI can be used as a marker to predict the severity of liver fibrosis, especially for differentiating mild from severe fibrosis.
The mean Child-Pugh score in our patients was 6.86±1.63 and most of them belonged to class B (53.8%). There was a significant difference in MPV comparing patients with Child class score A and B (P = 0.002). Also, the average MELD/PELD score in patients was 4.68±11.22, and a significant relationship was detected between the MPV and MELD/PELD score (P = 0.022).
The most common causes of chronic liver disease in our study were cryptogenic and biliary atresia. In a study by Dehghani et al, most common causes of liver disease were biliary atresia and Wilson disease. 9,10 In another study in Turkey, cryptogenic was the most common cause of cirrhosis which was in line with our report. 8 The most common clinical symptom in our patients was jaundice which was in accordance to the previous study in Shiraz. 9,10 In this study, a significant difference was observed in MPV values regarding different fibrosis severities which was in parallel to the report of Tahtaci et al on PBC patients, 11 Karagoz et al on patients with chronic hepatitis B, 5 Purnak et al on patients with chronic hepatitis C, 6 and Abdel-Razik et al on AIH patients. 12 Furthermore, in our study only, MPV significantly differed in patients with neonatal hepatitis as well. There was no significant statistical relationship between MPV and age or underlying disease which was in line with the report of Giannini et al in Italy. 7 In this study, there was no significant statistical differences between values and Child-Pugh score which was in oppose to the report of Giannini et al. in Italy, 7 and Erdem et al in Turkey. 8 No significant link was observed between MPV and MELD/ PELD score in our study which was different from the results Giannini et al in Italy, 7 and Hu et al in China. 13 According to the results of ROC curve analysis, an appropriate cut off did not obtained for MPV to differentiate various degrees of fibrosis. On the other hand, Tahtaci et al, 11 Karagoz et al 5 and Purnak et al 6 reported threshold values with better predictability.
In this study, there was no significant statistical difference for AAR in various degrees of fibrosis which was in line with the study of Yang et al in China 14 and Yang et al in South Korea. 15 In this study, there was a statistically significant difference in APRI among different degrees of liver fibrosis which was supported by the reports of Tahtaci et al in Turkey, 11 Kim et al in Japan, 16 Yang et al in South Korea, 15 Yang et al in China, 14 and Shokouhi et al in Iran. 17 FIB-4 index also showed a significant relationship with the degree of fibrosis in our study which was similar to the studies of Purnak et al in Turkey, 6 Yang et al in South Korea, 15 and Yang et al in China. 14

Conclusion
In conclusion, although MPV was significantly associated with the severity of liver fibrosis, an optimal cut off point to distinguish different degrees fibrosis was not obtained. Therefore, MPV seems to not be a good marker for predicting fibrosis stage in children with cirrhosis.

Ethical Approval
This study was approved by the Ethics Committee in Research of Shiraz University of Medical Sciences (IR. SUMS.MED.REC.1399.122).

Conflict of Interest Disclosure
None to declare.