Steatotic liver disease: use SAFE or FIB-4 to decide who needs cirrhosis follow-up

Design

National retrospective cohort, decision-curve analysis of nine risk scores

Population

853,131 US veterans with imaging-confirmed steatotic liver disease, no cirrhosis

Primary outcome

10-year cirrhosis and hepatocellular carcinoma

Effect

SAFE 29.5 = 2.5% 10-year cirrhosis risk, net benefit 0.019; HCC net benefit 0.0016

A cohort study in JAMA Internal Medicine (21 September) calculated nine risk scores at the first imaging that showed hepatic steatosis in 853,131 US veterans without viral hepatitis or other primary liver disease, then followed them for cirrhosis and hepatocellular carcinoma (HCC). Median age was 61; 32.3% had diabetes; 92.7% were men. Over 10 years, 3.96% developed cirrhosis and 0.35% HCC.

FIB-4, APRI and the SAFE (steatosis-associated fibrosis estimator) score discriminated cirrhosis best, and SAFE gave the greatest net benefit on decision-curve analysis. A SAFE score of 29.5 corresponded to a 10-year cirrhosis risk of 2.5%; using it as the cut-off would identify 1.9 additional future cirrhosis cases per 100 patients classed as at risk, compared with treating no one.

For HCC the answer was the opposite. No score was clinically useful for deciding surveillance in noncirrhotic patients — at a 0.25% 10-year risk threshold, SAFE added 1.6 true positives per 1,000. That supports current practice: HCC surveillance belongs to patients with cirrhosis, and the job in a general clinic is finding the cirrhosis.

• Calculate FIB-4 (or SAFE, where available) whenever imaging shows a fatty liver
• Use the score to decide who needs elastography or hepatology review, and who can simply be rechecked
• Do not start HCC ultrasound surveillance on a risk score alone in a patient without cirrhosis
• Remember the population was mostly older men; performance in women and younger patients is less certain
• In Indian practice FIB-4 needs only routine bloods, so it is the practical first step

Why it matters

Most fatty livers never progress, and this tells a generalist which simple score best separates the few who will.

Don't overread it

The cohort was 93% male veterans, and cirrhosis was identified from diagnosis codes rather than biopsy or elastography.

The statistics, in plain English

Net benefit weighs true positives against false positives at a chosen risk threshold. A net benefit of 0.019 means the score finds about two extra people who will develop cirrhosis per 100 flagged, without the cost of flagging everyone. For HCC the gain is under two per 1,000 — too small to justify surveillance decisions.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.