Skip to main content
. 2017 Nov 2;12(11):e0185813. doi: 10.1371/journal.pone.0185813

Table 1. Cross-sectional relationship between serum HMGB1 and various histological features in adults and children with NAFLD.

Histological feature PIVENS (N = 207) P* TONIC (N = 109) P*
No Mean ± SD (ng/mL) No. Mean ± SD (ng/mL)
Fibrosis stage: 0.48 0.31
 None 69 1.2 ± 1.8 38 1.4 ± 2.1
 Mild 65 2.0 ± 2.8 35 2.0 ± 3.6
 Moderate 37 1.5 ± 1.9 17 0.8 ± 1.7
 Bridging 27 1.9 ± 2.5 17 0.9 ± 1.3
 Cirrhosis 9 2.1 ± 2.6 0 --
Steatosis grade: 0.97 0.28
 ≤ 33% 12 1.7 ± 2.4 50 1.1 ± 1.7
 34–66% 52 1.6 ± 2.1 26 1.5 ± 2.3
 > 66% 26 1.4 ± 2.1 31 1.9 ± 3.7
Lobular inflammation: 0.94 0.29
 < 2 foci 16 1.6 ± 2.3 70 1.5 ± 2.9
 ≥ 2 foci 45 1.8 ± 2.4 37 1.3 ± 1.8
Hepatocellular ballooning: 0.94 0.40
 None 93 1.7 ± 2.6 61 1.6 ± 2.9
 Few 47 1.3 ± 2.0 29 1.8 ± 2.4
 Many 65 1.7 ± 2.1 17 0.6 ± 1.0
Portal inflammation: 0.49 0.74
 None 28 1.6 ± 2.3 17 1.0 ± 1.7
 Few 14 1.6 ± 2.3 76 1.5 ± 2.8
 Many 37 1.9 ± 2.3 14 1.4 ± 2.2
Steatohepatitis diagnosis: 0.87 0.87
 None 73 1.7 ± 2.6 46 1.1 ± 1.7
 Borderline suspicious zone 3 45 1.5 ± 2.3 20 1.9 ± 2.4
 Borderline, suspicious zone 1 0 -- 10 3.5 ± 5.9
 Definite 87 1.7 ± 2.1 31 1.0 ± 1.6

* HMGB1 levels assayed from the PIVENS and TONIC participants’ serum collected at 96 weeks.

P values (2-sided) for the association of histological feature and HMGB1 were derived from linear regression of the rank of the HMGB1 value on the histological feature.