Skip to main content
. Author manuscript; available in PMC: 2023 Jul 25.
Published in final edited form as: Hepatology. 2021 Aug 25;74(5):2410–2423. doi: 10.1002/hep.31845

TABLE 3.

Risk of Incident Cancer With Increasing NAFLD Histological Severity, Compared With Simple Steatosis*

NAFLD*
Types of Incident Cancer, N. Simple Steatosis
N = 5,939
NASH Without Fibrosis
N = 1,050
Noncirrhotic Fibrosis
N = 1,400
Cirrhosis
N = 503
All incident cancers 1,160 187 233 111
 Incidence RD1 (95% CI) 0 (ref.) 2.0 (−0.3–4.2) 3.4 (1.2–5.6) 12.1 (7.4–16.8)
 20-year absolute risk difference,2 % (95% CI) 0 (ref.) 4.4 (−1.1–10.0) 6.7 (0.8–12.5) 17.3 (5.8–28.9)
 Multivariable aHR3 (95% CI) 1 (ref.) 1.10 (0.94–1.29) 1.08 (0.93–1.26) 1.46 (1.19–1.79)
HCC 72 16 35 30
 Incidence RD1 (95% CI) 0 (ref.) 0.7 (0.0–1.3) 1.5 (0.8–2.3) 5.4 (3.3–7.6)
 20-year absolute risk difference,2 % (95% CI) 0 (ref.) 1.0 (−0.5–2.4) 3.4 (1.5–5.3) 10.2 (5.3–15.2)
 Multivariable aHR3 (95% CI) 1 (ref.) 1.61 (0.91–2.84) 2.45 (1.57–3.84) 4.95 (3.08–7.97)
Non-HCC liver cancers 16 2 5 2
 Incidence RD1 (95% CI) 0 (ref.) 0.0 (−0.2–0.2) 0.2 (−0.1–0.5) 0.2 (−0.3–0.8)
 20-year absolute risk difference,2 % (95% CI) 0 (ref.) 0.0 (−0.6–0.5) 0.1 (−0.5–0.6) 0.7 (−1.0–2.3)
 Multivariable aHR3 (95% CI) 1 (ref.) 0.76 (0.16–3.52) 1.97 (0.66–5.88) 1.56 (0.35–7.08)
EHSO cancers 987 152 175 77
 Incidence RD1 (95% CI) 0 (ref.) 1.0 (−1.0–3.0) 1.2 (−0.7–3.1) 6.3 (2.4–10.1)
 20-year absolute risk difference,2 % (95% CI) 0 (ref.) 3.3 (−1.7–8.4) 2.9 (−2.2–7.9) 9.5 (0.0–19.0)
 Multivariable aHR3 (95% CI) 1 (ref.) 1.04 (0.88–1.25) 0.94 (0.80–1.12) 1.23 (0.97–1.56)
Hematologic cancers 85 17 19 2
 Incidence RD1 (95% CI) 0 (ref.) 0.4 (−0.3–1.0) 0.4 (−0.2–1.0) −0.5 (−1.1–0.1)
 20-year absolute risk difference,2 % (95% CI) 0 (ref.) 0.3 (−0.9–1.5) 1.3 (−0.5–3.0) −0.7 (−2.4–1.0)
 Multivariable aHR3 (95% CI) 1 (ref.) 1.38 (0.80–2.36) 1.39 (0.81–2.37) 0.35 (0.08–1.46)
*

NAFLD was defined by liver histology. For definitions and algorithm, see the Methods section and Supporting Information.

1

Incidence rates were calculated by dividing the number of cases by PYs; CIs for incidence rates and absolute RDs were approximated by the normal distribution (Methods).

2

20-year absolute risks and absolute risk differences (percentage points) were calculated based on Kaplan-Meier estimates (Methods).

3

The multivariable aHR model accounted for the covariates outlined in Table 2.