AGA Guideline: Transient elastography in liver fibrosis, most used and most accurate

Vibration-controlled transient elastography (VCTE) can accurately diagnose cirrhosis in most patients with chronic liver disease, particularly those with chronic hepatitis B or C, states a new guideline from the AGA Institute, published in the May issue of Gastroenterology (doi: 10.1053/j.gastro.2017.03.017).

However, magnetic resonance elastography (MRE) is somewhat more accurate for detecting cirrhosis in nonalcoholic fatty liver disease, wrote Joseph K. Lim, MD, AGAF, of Yale University in New Haven, Conn., with his associates from the Clinical Guidelines Committee of the AGA. VCTE is convenient but performs unevenly in some liver conditions and is especially unreliable in patients with acute hepatitis, alcohol abuse, food intake within 2-3 hours, congestive heart failure, or extrahepatic cholestasis, the guideline notes. Yet, VCTE remains the most common imaging tool for diagnosing fibrosis in the United States, and the guideline addresses “focused, clinically relevant questions” to guide its use.

When possible, clinicians should use VCTE instead of noninvasive serum tests for cirrhosis in patients with chronic hepatitis C, the guideline asserts. In pooled analyses of 62 studies, VCTE detected about 89% of cirrhosis cases (95% confidence interval, 84%-92%), Fibrosis-4 test (FIB-4) detected 87% (95% CI, 74%-94%), and aspartate aminotransferase to platelet ratio index (APRI) detected 77% (95% CI, 73%-81%). The specificity of VCTE (91%) also equaled or exceeded that of FIB-4 (91%) or APRI (78%), the guideline noted.

For chronic hepatitis C, MRE had “poorer specificity with higher false-positive rates, suggesting poorer diagnostic performance,” compared with VCTE. Lower cost and lower point-of-care availability make VCTE “an attractive solution compared to MRE,” the guideline adds. It conditionally recommends VCTE cutoffs of 12.5 kPa for cirrhosis and 9.5 kPa for advanced (F3-F4) liver fibrosis after patients have a sustained virologic response to therapy. The 9.5-kPa cutoff would misclassify only 1% of low-risk patients and 7% of high-risk patients, but noncirrhotic patients (less than 9.5 kPa) may reasonably choose to continue specialty care if they prioritize avoiding “the small risk” of hepatocellular carcinoma over the “inconvenience and risks of continued laboratory and fibrosis testing.”

For chronic hepatitis B, the guideline conditionally recommends VCTE with an 11.0-kPa cutoff over APRI or FIB-4. In a pooled analysis of 28 studies, VCTE detected cirrhosis with a sensitivity of 86% and a specificity of 85%, compared with 66% and 74%, respectively, for APRI, and 87% and 65%, respectively, for FIB-4. However, the overall diagnostic performance of VCTE resembled that of the serum tests, and clinicians should interpret VCTE in the context of other clinical cirrhosis data, the guideline states.

Among 17 studies of VCTE cutoffs in hepatitis B, an 11.0-kPa threshold diagnosed cirrhosis with a sensitivity of 81% and a specificity of 83%. This cutoff would miss cirrhosis in less than 1% of low-risk patients and about 5% of high-risk patients and would yield false positives in 10%-15% of patients. Thus, its cutoff minimizes false negatives, reflecting “a judgment that the harm of missing cirrhosis is greater than the harm of over diagnosis,” the authors write.

For chronic alcoholic liver disease, the AGA conditionally recommends VCTE with a cirrhosis cutoff of 12.5 kPa. In pooled analyses, this value had a sensitivity of 95% and a specificity of 71%. For suspected compensated cirrhosis, the guideline conditionally suggests a 19.5-kPa cutoff when assessing the need for esophagogastroduodenoscopy (EGD) to identify high-risk esophageal varices. Patients who fall below this cutoff can reasonably pursue screening endoscopy if they are concerned about the small risk of acute variceal hemorrhage, the guideline adds.

The guideline also conditionally recommends a 17-kPa cutoff to detect clinically significant portal hypertension in patients with suspected chronic liver disease who are undergoing elective nonhepatic surgeries. This cutoff will miss about 0.1% of very low-risk patients, 0.8% of low-risk patients, and 7% of high-risk patients. Because the failure to detect portal hypertension contributes to operative morbidity and mortality, higher-risk patients might “reasonably” pursue screening endoscopy even if their kPa is below the cutoff, the guideline states.

The guideline made no recommendation about VCTE versus APRI or FIB-4 in adults with nonalcoholic fatty liver disease (NAFLD), citing “unacceptable bias” in 12 studies that excluded obese patients, used per-protocol rather than intention-to-diagnose analyses, and ignored “unsuccessful or inadequate” liver stiffness measurements, which are relatively common in NAFLD, the guideline notes. It conditionally recommends MRE over VCTE in high-risk adults with NAFLD, including those who are older, diabetic, or obese (especially with central adiposity) or who have alanine levels more than twice the upper limit of normal. However, it cites insufficient evidence to extend this recommendation to low-risk patients who only have imaging evidence of fatty liver.

Overall, the guideline focuses on “routine clinical management issues, and [does] not address comparisons with proprietary serum fibrosis assays, other emerging imaging-based fibrosis assessment techniques, or combinations of more than one noninvasive fibrosis test,” the authors note. They also limited VCTE cutoffs to single thresholds that prioritized sensitivity over specificity. “Additional studies are needed to further define the role of VCTE, MRE, and emerging diagnostic studies in the assessment of liver fibrosis, for which a significant unmet medical need remains, particularly in conditions such as NAFLD/[nonalcoholic steatohepatitis],” they add. “In particular, defining the implications for serial liver stiffness measurements over time on management decisions is of great interest.”

Dr. Muir has served as a consultant for AbbVie, Bristol-Myers Squibb, Gilead, and Merck. Dr. Lim has served as a consultant for Bristol Myers-Squibb, Gilead, Merck, and Boehringer Ingelheim. Dr. Flamm has served as a consultant or received research support from Gilead, Bristol-Myers Squibb, AbbVie, Salix Pharmaceuticals, and Intercept Pharmaceuticals. Dr. Dieterich has presented lectures for Gilead and Merck products. The rest of the authors disclosed no conflicts related to the content of this guideline.

Vibration-controlled transient elastography (VCTE) can accurately diagnose cirrhosis in most patients with chronic liver disease, particularly those with chronic hepatitis B or C, states a new guideline from the AGA Institute, published in the May issue of Gastroenterology (doi: 10.1053/j.gastro.2017.03.017).

However, magnetic resonance elastography (MRE) is somewhat more accurate for detecting cirrhosis in nonalcoholic fatty liver disease, wrote Joseph K. Lim, MD, AGAF, of Yale University in New Haven, Conn., with his associates from the Clinical Guidelines Committee of the AGA. VCTE is convenient but performs unevenly in some liver conditions and is especially unreliable in patients with acute hepatitis, alcohol abuse, food intake within 2-3 hours, congestive heart failure, or extrahepatic cholestasis, the guideline notes. Yet, VCTE remains the most common imaging tool for diagnosing fibrosis in the United States, and the guideline addresses “focused, clinically relevant questions” to guide its use.

When possible, clinicians should use VCTE instead of noninvasive serum tests for cirrhosis in patients with chronic hepatitis C, the guideline asserts. In pooled analyses of 62 studies, VCTE detected about 89% of cirrhosis cases (95% confidence interval, 84%-92%), Fibrosis-4 test (FIB-4) detected 87% (95% CI, 74%-94%), and aspartate aminotransferase to platelet ratio index (APRI) detected 77% (95% CI, 73%-81%). The specificity of VCTE (91%) also equaled or exceeded that of FIB-4 (91%) or APRI (78%), the guideline noted.

For chronic hepatitis C, MRE had “poorer specificity with higher false-positive rates, suggesting poorer diagnostic performance,” compared with VCTE. Lower cost and lower point-of-care availability make VCTE “an attractive solution compared to MRE,” the guideline adds. It conditionally recommends VCTE cutoffs of 12.5 kPa for cirrhosis and 9.5 kPa for advanced (F3-F4) liver fibrosis after patients have a sustained virologic response to therapy. The 9.5-kPa cutoff would misclassify only 1% of low-risk patients and 7% of high-risk patients, but noncirrhotic patients (less than 9.5 kPa) may reasonably choose to continue specialty care if they prioritize avoiding “the small risk” of hepatocellular carcinoma over the “inconvenience and risks of continued laboratory and fibrosis testing.”

For chronic hepatitis B, the guideline conditionally recommends VCTE with an 11.0-kPa cutoff over APRI or FIB-4. In a pooled analysis of 28 studies, VCTE detected cirrhosis with a sensitivity of 86% and a specificity of 85%, compared with 66% and 74%, respectively, for APRI, and 87% and 65%, respectively, for FIB-4. However, the overall diagnostic performance of VCTE resembled that of the serum tests, and clinicians should interpret VCTE in the context of other clinical cirrhosis data, the guideline states.

Among 17 studies of VCTE cutoffs in hepatitis B, an 11.0-kPa threshold diagnosed cirrhosis with a sensitivity of 81% and a specificity of 83%. This cutoff would miss cirrhosis in less than 1% of low-risk patients and about 5% of high-risk patients and would yield false positives in 10%-15% of patients. Thus, its cutoff minimizes false negatives, reflecting “a judgment that the harm of missing cirrhosis is greater than the harm of over diagnosis,” the authors write.

For chronic alcoholic liver disease, the AGA conditionally recommends VCTE with a cirrhosis cutoff of 12.5 kPa. In pooled analyses, this value had a sensitivity of 95% and a specificity of 71%. For suspected compensated cirrhosis, the guideline conditionally suggests a 19.5-kPa cutoff when assessing the need for esophagogastroduodenoscopy (EGD) to identify high-risk esophageal varices. Patients who fall below this cutoff can reasonably pursue screening endoscopy if they are concerned about the small risk of acute variceal hemorrhage, the guideline adds.

The guideline also conditionally recommends a 17-kPa cutoff to detect clinically significant portal hypertension in patients with suspected chronic liver disease who are undergoing elective nonhepatic surgeries. This cutoff will miss about 0.1% of very low-risk patients, 0.8% of low-risk patients, and 7% of high-risk patients. Because the failure to detect portal hypertension contributes to operative morbidity and mortality, higher-risk patients might “reasonably” pursue screening endoscopy even if their kPa is below the cutoff, the guideline states.

The guideline made no recommendation about VCTE versus APRI or FIB-4 in adults with nonalcoholic fatty liver disease (NAFLD), citing “unacceptable bias” in 12 studies that excluded obese patients, used per-protocol rather than intention-to-diagnose analyses, and ignored “unsuccessful or inadequate” liver stiffness measurements, which are relatively common in NAFLD, the guideline notes. It conditionally recommends MRE over VCTE in high-risk adults with NAFLD, including those who are older, diabetic, or obese (especially with central adiposity) or who have alanine levels more than twice the upper limit of normal. However, it cites insufficient evidence to extend this recommendation to low-risk patients who only have imaging evidence of fatty liver.

Overall, the guideline focuses on “routine clinical management issues, and [does] not address comparisons with proprietary serum fibrosis assays, other emerging imaging-based fibrosis assessment techniques, or combinations of more than one noninvasive fibrosis test,” the authors note. They also limited VCTE cutoffs to single thresholds that prioritized sensitivity over specificity. “Additional studies are needed to further define the role of VCTE, MRE, and emerging diagnostic studies in the assessment of liver fibrosis, for which a significant unmet medical need remains, particularly in conditions such as NAFLD/[nonalcoholic steatohepatitis],” they add. “In particular, defining the implications for serial liver stiffness measurements over time on management decisions is of great interest.”

Dr. Muir has served as a consultant for AbbVie, Bristol-Myers Squibb, Gilead, and Merck. Dr. Lim has served as a consultant for Bristol Myers-Squibb, Gilead, Merck, and Boehringer Ingelheim. Dr. Flamm has served as a consultant or received research support from Gilead, Bristol-Myers Squibb, AbbVie, Salix Pharmaceuticals, and Intercept Pharmaceuticals. Dr. Dieterich has presented lectures for Gilead and Merck products. The rest of the authors disclosed no conflicts related to the content of this guideline.

Vibration-controlled transient elastography (VCTE) can accurately diagnose cirrhosis in most patients with chronic liver disease, particularly those with chronic hepatitis B or C, states a new guideline from the AGA Institute, published in the May issue of Gastroenterology (doi: 10.1053/j.gastro.2017.03.017).

However, magnetic resonance elastography (MRE) is somewhat more accurate for detecting cirrhosis in nonalcoholic fatty liver disease, wrote Joseph K. Lim, MD, AGAF, of Yale University in New Haven, Conn., with his associates from the Clinical Guidelines Committee of the AGA. VCTE is convenient but performs unevenly in some liver conditions and is especially unreliable in patients with acute hepatitis, alcohol abuse, food intake within 2-3 hours, congestive heart failure, or extrahepatic cholestasis, the guideline notes. Yet, VCTE remains the most common imaging tool for diagnosing fibrosis in the United States, and the guideline addresses “focused, clinically relevant questions” to guide its use.

When possible, clinicians should use VCTE instead of noninvasive serum tests for cirrhosis in patients with chronic hepatitis C, the guideline asserts. In pooled analyses of 62 studies, VCTE detected about 89% of cirrhosis cases (95% confidence interval, 84%-92%), Fibrosis-4 test (FIB-4) detected 87% (95% CI, 74%-94%), and aspartate aminotransferase to platelet ratio index (APRI) detected 77% (95% CI, 73%-81%). The specificity of VCTE (91%) also equaled or exceeded that of FIB-4 (91%) or APRI (78%), the guideline noted.

For chronic hepatitis C, MRE had “poorer specificity with higher false-positive rates, suggesting poorer diagnostic performance,” compared with VCTE. Lower cost and lower point-of-care availability make VCTE “an attractive solution compared to MRE,” the guideline adds. It conditionally recommends VCTE cutoffs of 12.5 kPa for cirrhosis and 9.5 kPa for advanced (F3-F4) liver fibrosis after patients have a sustained virologic response to therapy. The 9.5-kPa cutoff would misclassify only 1% of low-risk patients and 7% of high-risk patients, but noncirrhotic patients (less than 9.5 kPa) may reasonably choose to continue specialty care if they prioritize avoiding “the small risk” of hepatocellular carcinoma over the “inconvenience and risks of continued laboratory and fibrosis testing.”

For chronic hepatitis B, the guideline conditionally recommends VCTE with an 11.0-kPa cutoff over APRI or FIB-4. In a pooled analysis of 28 studies, VCTE detected cirrhosis with a sensitivity of 86% and a specificity of 85%, compared with 66% and 74%, respectively, for APRI, and 87% and 65%, respectively, for FIB-4. However, the overall diagnostic performance of VCTE resembled that of the serum tests, and clinicians should interpret VCTE in the context of other clinical cirrhosis data, the guideline states.

Among 17 studies of VCTE cutoffs in hepatitis B, an 11.0-kPa threshold diagnosed cirrhosis with a sensitivity of 81% and a specificity of 83%. This cutoff would miss cirrhosis in less than 1% of low-risk patients and about 5% of high-risk patients and would yield false positives in 10%-15% of patients. Thus, its cutoff minimizes false negatives, reflecting “a judgment that the harm of missing cirrhosis is greater than the harm of over diagnosis,” the authors write.

For chronic alcoholic liver disease, the AGA conditionally recommends VCTE with a cirrhosis cutoff of 12.5 kPa. In pooled analyses, this value had a sensitivity of 95% and a specificity of 71%. For suspected compensated cirrhosis, the guideline conditionally suggests a 19.5-kPa cutoff when assessing the need for esophagogastroduodenoscopy (EGD) to identify high-risk esophageal varices. Patients who fall below this cutoff can reasonably pursue screening endoscopy if they are concerned about the small risk of acute variceal hemorrhage, the guideline adds.

The guideline also conditionally recommends a 17-kPa cutoff to detect clinically significant portal hypertension in patients with suspected chronic liver disease who are undergoing elective nonhepatic surgeries. This cutoff will miss about 0.1% of very low-risk patients, 0.8% of low-risk patients, and 7% of high-risk patients. Because the failure to detect portal hypertension contributes to operative morbidity and mortality, higher-risk patients might “reasonably” pursue screening endoscopy even if their kPa is below the cutoff, the guideline states.

The guideline made no recommendation about VCTE versus APRI or FIB-4 in adults with nonalcoholic fatty liver disease (NAFLD), citing “unacceptable bias” in 12 studies that excluded obese patients, used per-protocol rather than intention-to-diagnose analyses, and ignored “unsuccessful or inadequate” liver stiffness measurements, which are relatively common in NAFLD, the guideline notes. It conditionally recommends MRE over VCTE in high-risk adults with NAFLD, including those who are older, diabetic, or obese (especially with central adiposity) or who have alanine levels more than twice the upper limit of normal. However, it cites insufficient evidence to extend this recommendation to low-risk patients who only have imaging evidence of fatty liver.

Overall, the guideline focuses on “routine clinical management issues, and [does] not address comparisons with proprietary serum fibrosis assays, other emerging imaging-based fibrosis assessment techniques, or combinations of more than one noninvasive fibrosis test,” the authors note. They also limited VCTE cutoffs to single thresholds that prioritized sensitivity over specificity. “Additional studies are needed to further define the role of VCTE, MRE, and emerging diagnostic studies in the assessment of liver fibrosis, for which a significant unmet medical need remains, particularly in conditions such as NAFLD/[nonalcoholic steatohepatitis],” they add. “In particular, defining the implications for serial liver stiffness measurements over time on management decisions is of great interest.”

Dr. Muir has served as a consultant for AbbVie, Bristol-Myers Squibb, Gilead, and Merck. Dr. Lim has served as a consultant for Bristol Myers-Squibb, Gilead, Merck, and Boehringer Ingelheim. Dr. Flamm has served as a consultant or received research support from Gilead, Bristol-Myers Squibb, AbbVie, Salix Pharmaceuticals, and Intercept Pharmaceuticals. Dr. Dieterich has presented lectures for Gilead and Merck products. The rest of the authors disclosed no conflicts related to the content of this guideline.