Should Alpha-Fetoprotein (AFP) levels be checked in an adult patient with early cirrhosis, likely due to a history of liver disease such as hepatitis B or C, as indicated by ultrasound?

Should AFP Be Checked in Early Cirrhosis?

Yes, AFP should be checked every 6 months alongside ultrasound in patients with early cirrhosis for hepatocellular carcinoma (HCC) surveillance. 1

Guideline-Based Recommendation

The 2024 British Society of Gastroenterology guidelines explicitly recommend HCC surveillance with 6-monthly ultrasound scan and AFP measurement in all patients with cirrhosis, regardless of stage. 1 This represents the most current, high-quality guideline evidence available and should guide clinical practice.

• All cirrhosis patients warrant surveillance because the annual HCC incidence ranges from 1.3-8% depending on etiology (hepatitis B: 3-8%, hepatitis C: 3-5%, alcohol-related: 1.3-3%, NASH: 1-2%), all exceeding the cost-effectiveness threshold of 0.2-1.5% annual incidence. 1

• The combination of ultrasound plus AFP is superior to either modality alone, despite Western societies previously questioning AFP's value. 1

Why Both Tests Are Necessary

Ultrasound alone misses a significant proportion of HCCs. The complementary nature of these tests is critical:

• Ultrasound has 56-92% sensitivity for HCC detection, meaning it misses 8-44% of cancers when used alone. 2, 3

• AFP adds detection of cases missed by ultrasound, increasing combined sensitivity to 89-99% while maintaining 68-95% specificity. 2, 3

• In one large study of 1,597 cirrhosis patients followed for nearly 5 years, ultrasound detected 92% of HCCs while AFP (at 20 ng/mL cutoff) detected 53%, but the combination detected 99.2% of HCCs. 3

Understanding AFP Limitations and Strengths

AFP has well-documented limitations that explain historical controversy, but these do not negate its surveillance value:

• At 20 ng/mL cutoff: 60% sensitivity, 80-94% specificity—meaning it misses 40% of HCCs but rarely gives false positives. 4

• Up to 20% of HCCs never produce AFP, even when large, and two-thirds of HCCs <4 cm have AFP <200 ng/mL. 1

• However, serial AFP measurements significantly improve performance over single threshold approaches, detecting HCC 1.4-1.9 years earlier. 5

Practical Surveillance Algorithm

For patients with early cirrhosis:

1. Perform ultrasound + AFP every 6 months as baseline surveillance. 1

2. If AFP is elevated but <200 ng/mL with negative ultrasound: Increase ultrasound frequency to every 3 months, as these patients are at particularly high risk. 1

3. Monitor for serial AFP increases: A rise ≥10% over 3-6 months, even if absolute values remain <20 ng/mL, increases HCC risk 12-22 fold and warrants enhanced surveillance. 6

4. If AFP ≥20 ng/mL with ≥10% serial increase: This combination increases HCC risk 41.7-fold within 6 months, mandating immediate advanced imaging (CT/MRI). 6

5. If ultrasound detects a nodule: AFP >200 ng/mL with typical imaging features allows HCC diagnosis without biopsy. 1, 7

Common Pitfalls to Avoid

• Do not rely on ultrasound alone despite older Western guidelines suggesting AFP adds minimal value—the 2024 BSG guidelines explicitly recommend both modalities. 1

• Do not dismiss mildly elevated AFP (7-20 ng/mL) as clinically insignificant—longitudinal trends matter more than single values, and even small increases predict HCC development. 2, 6

• Do not withhold surveillance in "early" cirrhosis—HCC risk begins immediately upon cirrhosis development, and the distinction between "early" and "advanced" cirrhosis is irrelevant for surveillance decisions. 1

• Do not stop surveillance if AFP remains normal—up to 46% of HCC patients have completely normal AFP levels (<20 ng/mL) even at diagnosis. 7

Special Considerations by Etiology

The surveillance recommendation applies regardless of cirrhosis cause, though HCC risk varies:

• Hepatitis B cirrhosis: Highest risk (3-8% annually), and AFP may perform better in this population with demonstrated 37% mortality reduction from AFP-based surveillance. 1, 7

• Hepatitis C cirrhosis: 3-5% annual risk, surveillance remains cost-effective even in patients achieving sustained virological response. 1

• Alcohol-related and NASH cirrhosis: Lower but still significant risk (1.3-3% and 1-2% respectively), warranting identical surveillance. 1

When NOT to Perform Surveillance

Surveillance should be withheld only when patients cannot benefit from treatment:

• Decompensated cirrhosis (Child-Pugh B8 or worse) who are not liver transplant candidates. 1

• Severe comorbidities or poor performance status precluding any cancer-directed therapy. 1

The absolute HCC risk and potential surveillance harms should be discussed individually before enrollment, but for patients with early cirrhosis who are treatment candidates, the evidence strongly supports combined ultrasound and AFP surveillance every 6 months. 1