CON: Noninvasive Modalities Are Preferred to Screen for the Diagnosis of Esophageal and Gastric Varices When the Diagnosis of Cirrhosis Is Made

Watch a video presentation of this article

Watch an interview with the author

Abbreviations

AASLD

American Association for the Study of Liver Diseases

CI

confidence interval

EBL

endoscopic band ligation

HBV

hepatitis B virus

HCV

hepatitis C virus

HRVs

high‐risk varices

LS

liver stiffness

MELD

Model for End‐Stage Liver Disease

MRE

magnetic resonance elastography

NAFLD

nonalcoholic fatty liver disease

NR

not reported

PBC

primary biliary cholangitis

pHTN

portal hypertension

PSC

primary sclerosing cholangitis

pSWE

point shear wave elastography

VCTE

vibration‐controlled transient elastography

Key Points

• Clinicians should integrate noninvasive tests into their practice as the preferred method for variceal screening in select patients.

• Noninvasive tests using elastography and laboratory tests should be used to rule out high‐risk varices (HRVs) in patients with compensated cirrhosis.

• Noninvasive tests have excellent diagnostic performance and represent higher value care.

Bleeding from gastroesophageal varices leads to substantial morbidity and mortality in patients with cirrhosis. Previous American Association for the Study of Liver Diseases (AASLD) guidelines recommended variceal screening using upper endoscopy in all patients newly diagnosed with cirrhosis. ¹ However, endoscopy is invasive and carries procedural and sedation‐related risks. Noninvasive tests using elastography techniques and laboratory markers are safer and more cost‐effective alternatives to predict the presence of varices at high risk for bleeding. Herein we review the rationale for using noninvasive modalities to screen for varices.

Noninvasive Tests Should be Used to Rule Out HRVs in Patients With Compensated Cirrhosis

Clinicians considering variceal screening should understand the role of noninvasive tests and the targeted patient population (Fig. 1). Patients with decompensated cirrhosis, defined by the occurrence of variceal hemorrhage, ascites, or encephalopathy, have severe portal hypertension (pHTN), a high prevalence of varices (85%), and a high mortality rate after variceal hemorrhage. ² , ³ In these patients, endoscopy ought to remain the standard of care for variceal screening. In contrast, patients with compensated cirrhosis, who have mild or clinically significant pHTN, have a lower prevalence rate of varices (30%‐40%) and a low risk for mortality after a variceal bleed. ² , ³ Noninvasive tests are therefore most appropriate for variceal screening in patients with compensated cirrhosis.

Fig 1.

Target population and the role of noninvasive testing in variceal screening.

The goal of variceal screening is to identify patients with HRVs that require treatment with endoscopic band ligation (EBL) or nonselective beta‐blockers for primary prophylaxis. HRVs include medium or large varices, small varices with red wale signs, and any‐size varices in patients with decompensated cirrhosis. ³ All other varices are considered low risk; evidence does not currently support endoscopic or pharmacological prophylaxis for these varices. ³ Noninvasive tests are useful as risk‐stratification tools to rule out patients with HRVs and to safely avoid endoscopy in patients without HRVs. Because pHTN can progress over time, these patients should continue serial noninvasive testing for risk stratification. ⁴ If HRVs cannot be ruled out based on noninvasive testing, endoscopy should be pursued.

Noninvasive Tests Have Excellent Diagnostic Performance

Noninvasive tests generally combine elastography with biochemical parameters and have excellent diagnostic performance in ruling out HRVs (Table 1). The most well‐validated noninvasive modality is the Baveno VI criteria. ⁵ A meta‐analysis of 26 observational studies showed that the Baveno VI criteria had a pooled sensitivity of 97% (95% confidence interval [CI]: 0.95‐0.98) and negative predictive value of 0.09 (95% CI: 0.05‐0.15) for diagnosing HRVs. ⁶ These criteria incorporate liver stiffness (LS) as measured by vibration‐controlled transient elastography (VCTE) and platelet count to identify patients with a very low probability of HRVs who may safely avoid upper endoscopy. ³ , ⁵ The advantages of avoiding endoscopy must be weighed against the risk for missing HRVs. Expert consensus identified 5% as the maximum acceptable false negative (missed HRVs) rate. ⁴ Assuming a 20% prevalence rate of HRVs, the Baveno VI criteria would spare 26% of patients from unnecessary endoscopy and miss HRVs in ≤2%. ⁶ Extensions of the Baveno VI criteria have also been proposed to increase the number of patients spared from endoscopy while maintaining a low rate of missed HRVs. ⁶

Table 1.

Noninvasive Tests for Variceal Screening

Noninvasive Modalities	Method Used	Cutoffs	Diagnostic Performance for Ruling out HRVs	Spared Endoscopies	Missed HRVs
Baveno VI criteria 6	VCTE	LS < 20 kPa and platelet count >150,000/mm3	Pooled sensitivity: 0.97 (95% CI: 0.95‐0.98)	26%	2%
			Pooled negative likelihood ratio: 0.09 (95% CI: 0.05‐0.15)
Expanded Baveno VI criteria 6	VCTE	LS < 25 kPa and platelet count >110,000/mm3	Pooled sensitivity: 0.90 (95% CI: 0.85‐0.93)	43%	5%
			Pooled negative likelihood ratio: 0.20 (95% CI: 0.16‐0.27)
Spleen stiffness 9	pSWE	Spleen stiffness ≤3.30 m/second	Sensitivity: 0.99 (95% CI: 0.94‐1.0)	NR	NR
			Negative predictive value: 0.99 (95% CI: 0.97‐1.0)
			Negative likelihood ratio: 0.018 (95% CI: 0.003‐0.13)
LS by MRE 10	MRE	LS ≤ 5.8 kPa	Sensitivity: 0.96	NR	NR
			Negative predictive value: 0.98
Platelet count and MELD 6 7	Laboratory markers	Platelet count >150,000/mm3 or platelet count <150,000/mm3 but MELD = 6	Sensitivity: 0.94‐1.0	27%‐54%	0%‐1%
			Negative predictive value: 0.97‐1.0
RESIST‐HCV 8	Laboratory markers	Platelet count >120,000/mm3 and albumin >3.6 g/dL	Sensitivity: 0.94	31%	1.6%
			Negative predictive value: 0.98

Cutoff values and diagnostic performance of noninvasive tests from representative studies are shown.

Importantly, noninvasive screening methods based solely on combinations of readily available laboratory parameters (Model for End‐Stage Liver Disease [MELD] score, platelet count, and albumin) have shown to have excellent diagnostic performance, sparing 30% of patients endoscopy while missing HRVs in only 1%. ⁷ , ⁸ Shear wave elastography, magnetic resonance elastography (MRE), and spleen elastography techniques have also shown promise. ⁹ , ¹⁰

Existing studies of noninvasive tests primarily included patients with viral and alcohol‐related liver disease, but evidence is emerging among patients with nonalcoholic fatty liver disease (NAFLD), cholestatic liver diseases, and viral hepatitis after virological suppression ¹¹ , ¹² , ¹³ (Table 2). Further studies are needed, however, before noninvasive modalities can be broadly recommended for all patient subgroups. Patients for whom noninvasive testing is not yet appropriate include those with decompensated cirrhosis, noncirrhotic pHTN, or in whom the chosen noninvasive modality cannot produce a valid measurement (e.g., patients in whom VCTE fails to produce a valid measurement of LS).

Table 2.

Studies of Noninvasive Tests for Variceal Screening in Patients With NAFLD, Cholestatic Liver Disease, and Viral Hepatitis After Viral Suppression

Liver Disease Etiology	Study	No. of Patients	Noninvasive Screening Modality	Diagnostic Performance
NAFLD	Petta et al. (2018) 11	790	Baveno VI	Baveno VI
			Expanded Baveno VI	Sensitivity: 0.89‐0.97
			NAFLD criteria (LS < 30 kPa, platelet count >110,000/mm3)	Negative predictive value: 0.96‐0.99
				Expanded Baveno VI
				Sensitivity: 0.78‐0.82
				Negative predictive value: 0.96
				NAFLD criteria
				Sensitivity: 0.72‐0.78
				Negative predictive value: 0.95‐0.96
Cholestatic liver disease (PBC and PSC)	Moctezuma‐Velazquez et al. (2019) 12	147 (PBC)	Baveno VI	Baveno VI
		80 (PSC)	Expanded Baveno VI*	PBC: 39% spared endoscopy, 0% missed HRVs
				PSC: 30% spared endoscopy, 0% missed HRVs
				Expanded Baveno VI
				PBC: 58% spared endoscopy, 6% missed HRVs
				PSC: 45% spared endoscopy, 3% missed HRVs
Viral hepatitis with viral suppression (mixed HBV and HCV)	Thabut et al. (2019) 13	200 † (HCV: 94, HBV: 98, both: 8)	Baveno VI	40% spared endoscopy
				0% missed HRVs
				228 patients with viral suppression at baseline or during follow‐up had LS and platelet count available at the time of viral suppression. Of these, 64 patients met Baveno VI criteria, of which 0 had progression of pHTN at 1, 3, and 5 years (defined by progression of varices or variceal bleed).

^*Diagnostic performance of several other noninvasive criteria presented in the study is not shown.

^†Refers to patients with viral suppression at the time of inclusion into the study.

Noninvasive Tests Represent Higher Value Care

Noninvasive tests are also preferable due to its lower cost, increased safety, and increased accessibility profile (Fig. 2).

Fig 2.

Noninvasive tests for variceal screening represent higher value care.

AASLD guidelines recommend endoscopy at a minimum of every 2 to 3 years for variceal screening. ³ At our institution, out‐of‐pocket costs for endoscopy with EBL are $3,615 compared with elastography at $417. Assuming a population of 1,000 patients with compensated cirrhosis, noninvasive testing compared with endoscopy would save $3.2 million and $5.98 million in the first and third years of screening, respectively. Although these calculations are not all‐encompassing, such cost savings is substantial.

Noninvasive variceal screening also avoids procedural risks. Endoscopy is invasive and associated with the risk for postbanding ulcer formation at 5% to 15%, aspiration pneumonia and bacterial peritonitis at 1% to 4%, and although rarer, perforation. ¹⁴ Elastography and venipuncture do not have these risks.

Finally, the success of variceal screening programs has been limited by low adherence rates. Among patients with hepatitis C virus (HCV) cirrhosis, only 33.8% of patients received endoscopy within guidelines. ¹⁵ Low rates of adherence are exacerbated by geographical disparities in access to endoscopy because of variable densities of gastroenterology providers in the United States. ¹⁶ In comparison, noninvasive tests are more readily available, often as point‐of‐care tests, and easily accessible at nearby clinics or laboratory sites, making it more likely to promote patient adherence to screening.

In summary, noninvasive modalities are safe, cost‐effective, and feasible alternatives to universal endoscopy for variceal screening. When used in patients with compensated cirrhosis, noninvasive tests can successfully stratify those at risk for development of HRVs, spare unnecessary endoscopies, and minimize undue risks related to procedures and sedation. Furthermore, noninvasive tests are readily available and easily accessible nationwide at a fraction of the cost of endoscopy. In an era when the cost of health care continues to rise and more patients with cirrhosis are diagnosed earlier, noninvasive tests for variceal screening provide clinicians with an opportunity to incorporate higher value care while improving outcomes in patients with cirrhosis.