Occult Cirrhosis Diagnosed by Transient Elastography
Occult Cirrhosis Diagnosed by Transient Elastography
We found that OC is a frequent clinical entity that is likely under-diagnosed, with significant risk for rapid development of hepatic decompensation. These patients are more likely to be misclassified, and receive the same monitoring schedule as non-cirrhotic patients, potentially delaying essential treatments (i.e. antiviral therapy) or late detection of a treatable complication (i.e. HCC, esophageal varices).
Cirrhosis is a dynamic entity and is better characterized as a disease spectrum rather than a single stage disease. It can remain clinically silent, without any physical, laboratory and imaging findings, until the first clinical decompensation occurs, which impacts on prognosis. The median survival time is >12 years for compensated disease while cumulative death probability per 1 year is 20% or more in decompensated cirrhosis. Earlier diagnosis of compensated cirrhosis remains a major challenge for clinicians when there is absence of any obvious clinical features. We found that the prevalence of OC is considerable, representing 37% of all patients with a liver stiffness ≥13 kPa and 12% of all CLD patients who underwent TE at our center. This proportion suggests up to 1 in 9 individuals with CLD seen at tertiary-referral centers may have their OC missed. Our finding is particularly impressive given that we applied strict criteria for the definition of OC, including absence of any ultrasonographic sign of advanced liver disease, absence of thrombocytopenia and endoscopic signs of portal hypertension.
We employed TE to diagnose OC since it is an accurate and patient friendly non-invasive tool to assess liver cirrhosis. A meta-analysis found that TE has a mean AUC of 0.94 for the diagnosis of cirrhosis using a cut-off value of 13 kPa. The AUC of this optimal diagnostic cut-off value was not significantly influenced by underlying etiologies of CLD. We also applied stringent criteria for the quality of TE to optimize its reliability. In our experience, 25% of all patients were excluded for unreliable readings at first examination. This figure is in line with the unreliable results rates (21%) previously reported with the M probe. Liver biopsy was available only in a minority of patients in our retrospective cohort. In many patients with OC, histology was consistent with advanced liver fibrosis or cirrhosis. Misclassified cases may be due to limitations of liver biopsy, which remains an imperfect standard of reference for staging of hepatic fibrosis. Due to the methodological limitations of liver biopsy and to the limited number of liver biopsies at our disposal, we adopted a prognostic validation of our findings based on longitudinal assessment of outcomes. Because the OC group showed a significantly worse prognosis than patients with non-cirrhotic CLD, we demonstrated that TE could be a useful tool to identify OC. Since liver fibrosis is a dynamic process, validation of surrogate methods by means of longitudinal prognostic studies is more robust than cross-sectional diagnostic studies, which capture a single time point. Moreover, non-invasive fibrosis tools including TE and Fibrotest demonstrated superior prognostic value than liver biopsy. Finally, both APRI and FIB-4 values were significantly higher in the OC group as compared to the non-cirrhotic CLD group. This finding is interesting given that one of the exclusion criteria for definition of OC was the absence of thrombocytopenia, and platelets is used to compute both APRI and FIB-4.
We dissected predictors of OC in order to identify patients who are at higher risk and who should be particularly tested and monitored for OC. Older age was independently associated with OC at baseline. This likely reflects a rising incidence of cirrhosis in elderly due to both a longer duration of liver disease and the fact that cirrhotic patients are living longer. HIV co-infection in patients with chronic viral hepatitis B and C was strongly associated with OC. These patients usually progress more quickly towards cirrhosis than their HCV or HBV mono-infected counterparts due to multiple risk factors for hepatotoxicity, including antiretroviral treatments, oxidative stress, excessive alcohol intake. Higher APRI was also independently associated with OC at baseline, reflecting the correlation between TE and simple biomarkers. It could be argued that simple biomarkers could be a practical solution to screening in settings of limited-resources. However, APRI and FIB-4 indicated cirrhosis or advanced fibrosis only in a minority of patients with OC, suggesting that they have a limited diagnostic accuracy for identification of OC.
During a median of 24 months follow-up, 7% of the OC group, 0% of non-cirrhotic CLD patients and 19% of the clinically evident cirrhosis group developed clinical outcomes. We also investigated the incidence of clinical signs of advanced liver disease in OC (12%) and non-cirrhotic CLD (6%) groups. We were therefore able to appreciate the difference in prognosis between OC and non-cirrhotic CLD, as well as to characterize the faster progression of OC. Because OC and non-cirrhotic CLD groups had similar follow-up, including number of clinical visits and ultrasounds, the difference in clinical outcomes among groups reflect a more rapid disease progression in patients with OC. In the OC group, the cumulative incidences of HCC, ascites, and variceal bleeding were 1%, 5%, and 1% respectively. Of note, in the HCC and variceal bleeding cases, clinical signs of advanced liver disease including thrombocytopenia or ultrasonographic signs, did not precede the clinical outcomes. This was likely due to suboptimal monitoring of patients in the OC group as it was under-staged. This shows that clinicians heavily rely on physical, ultrasonographic and endoscopic findings in order to ensure a diagnosis of cirrhosis and to commence appropriate surveillance program. In our experience, the suboptimal monitoring in OC translated into higher rates of late diagnosis (60%) as compared to the clinically evident cirrhosis group (15%). The patient with HCC in the OC group was diagnosed at 14-month follow-up with intermediate stage by BCLC while the case of bleeding from esophageal varices was diagnosed at 18-month follow-up. With appropriate surveillance, these cases could have been captured earlier. On multivariate analysis, OC was associated with progression to clinical outcomes when compared to non-cirrhotic CLD.
In our study population, the probability of clinical outcomes was increased according to TE measurements. The overall prognostic performance of TE was excellent, with an AUC of 0.90. Our findings confirm recent studies conducted in chronic hepatitis C and B and underline the importance of TE for risk stratification and early determination of prognosis of patients with CLD.
We acknowledge several limitations of our study. First, this study was conducted at a tertiary-care referral center, and although our prevalence of cirrhosis (19%), is comparable to other university centers, this is higher than that reported in primary/secondary care settings and in the general population. Therefore, a parallel over-estimation of the prevalence of OC is possible. Likewise, our study population may have an over-representation of HIV co-infected patients due to possible referral bias. However, our findings reflect the current situation of a tertiary care liver center, with increasing referral of HIV co-infected patients with progressive liver disease. These centers are also most likely to have a TE machine. Second, this was a retrospective study, and as such we were unable to control for potential confounding factors, including the decision to proceed to TE, which could potentially select the sickest patients. Third, liver biopsy was not available for all patients in the OC group. However, advanced fibrosis/cirrhosis was confirmed in the majority of OC patients who had histological evaluation. Moreover, we validated longitudinally our findings by demonstrating that OC group had worse clinical outcomes than the non-cirrhotic CLD group.
In conclusion, this study suggests that OC is frequent in a tertiary-care Liver Unit. Clinicians should be aware that the absence of thrombocytopenia, ultrasonographic signs of advanced liver disease and/or normal gastroscopy cannot rule out the possibility of a compensated cirrhosis in its preclinical stage. Patients with OC can exhibit rapid development of HCC, variceal bleeding and ascites. Those patients are also likely to be under-diagnosed and consequently under-monitored if clinicians base their diagnosis exclusively on physical, ultrasonographic or endoscopic signs. This can in turn translate into delayed diagnosis of complications. The implementation of screening strategies of CLD patients with TE at tertiary care centers may help the early identification of the otherwise clinically occult and unrecognized stage of compensated cirrhosis. It can prompt clinicians to initiate a program of surveillance for HCC and portal hypertension. Patients with OC due to hepatitis C could be given priority for antiviral therapy, in light of concerns about cost-effective use of these potent, new interferon free regimens. These strategies may potentially reduce the burden of liver complications in patients with cirrhosis. Prospective studies are needed to validate our study findings and to confirm the impact of early identification of OC and specific interventions in clinical practice.
Discussion
We found that OC is a frequent clinical entity that is likely under-diagnosed, with significant risk for rapid development of hepatic decompensation. These patients are more likely to be misclassified, and receive the same monitoring schedule as non-cirrhotic patients, potentially delaying essential treatments (i.e. antiviral therapy) or late detection of a treatable complication (i.e. HCC, esophageal varices).
Cirrhosis is a dynamic entity and is better characterized as a disease spectrum rather than a single stage disease. It can remain clinically silent, without any physical, laboratory and imaging findings, until the first clinical decompensation occurs, which impacts on prognosis. The median survival time is >12 years for compensated disease while cumulative death probability per 1 year is 20% or more in decompensated cirrhosis. Earlier diagnosis of compensated cirrhosis remains a major challenge for clinicians when there is absence of any obvious clinical features. We found that the prevalence of OC is considerable, representing 37% of all patients with a liver stiffness ≥13 kPa and 12% of all CLD patients who underwent TE at our center. This proportion suggests up to 1 in 9 individuals with CLD seen at tertiary-referral centers may have their OC missed. Our finding is particularly impressive given that we applied strict criteria for the definition of OC, including absence of any ultrasonographic sign of advanced liver disease, absence of thrombocytopenia and endoscopic signs of portal hypertension.
We employed TE to diagnose OC since it is an accurate and patient friendly non-invasive tool to assess liver cirrhosis. A meta-analysis found that TE has a mean AUC of 0.94 for the diagnosis of cirrhosis using a cut-off value of 13 kPa. The AUC of this optimal diagnostic cut-off value was not significantly influenced by underlying etiologies of CLD. We also applied stringent criteria for the quality of TE to optimize its reliability. In our experience, 25% of all patients were excluded for unreliable readings at first examination. This figure is in line with the unreliable results rates (21%) previously reported with the M probe. Liver biopsy was available only in a minority of patients in our retrospective cohort. In many patients with OC, histology was consistent with advanced liver fibrosis or cirrhosis. Misclassified cases may be due to limitations of liver biopsy, which remains an imperfect standard of reference for staging of hepatic fibrosis. Due to the methodological limitations of liver biopsy and to the limited number of liver biopsies at our disposal, we adopted a prognostic validation of our findings based on longitudinal assessment of outcomes. Because the OC group showed a significantly worse prognosis than patients with non-cirrhotic CLD, we demonstrated that TE could be a useful tool to identify OC. Since liver fibrosis is a dynamic process, validation of surrogate methods by means of longitudinal prognostic studies is more robust than cross-sectional diagnostic studies, which capture a single time point. Moreover, non-invasive fibrosis tools including TE and Fibrotest demonstrated superior prognostic value than liver biopsy. Finally, both APRI and FIB-4 values were significantly higher in the OC group as compared to the non-cirrhotic CLD group. This finding is interesting given that one of the exclusion criteria for definition of OC was the absence of thrombocytopenia, and platelets is used to compute both APRI and FIB-4.
We dissected predictors of OC in order to identify patients who are at higher risk and who should be particularly tested and monitored for OC. Older age was independently associated with OC at baseline. This likely reflects a rising incidence of cirrhosis in elderly due to both a longer duration of liver disease and the fact that cirrhotic patients are living longer. HIV co-infection in patients with chronic viral hepatitis B and C was strongly associated with OC. These patients usually progress more quickly towards cirrhosis than their HCV or HBV mono-infected counterparts due to multiple risk factors for hepatotoxicity, including antiretroviral treatments, oxidative stress, excessive alcohol intake. Higher APRI was also independently associated with OC at baseline, reflecting the correlation between TE and simple biomarkers. It could be argued that simple biomarkers could be a practical solution to screening in settings of limited-resources. However, APRI and FIB-4 indicated cirrhosis or advanced fibrosis only in a minority of patients with OC, suggesting that they have a limited diagnostic accuracy for identification of OC.
During a median of 24 months follow-up, 7% of the OC group, 0% of non-cirrhotic CLD patients and 19% of the clinically evident cirrhosis group developed clinical outcomes. We also investigated the incidence of clinical signs of advanced liver disease in OC (12%) and non-cirrhotic CLD (6%) groups. We were therefore able to appreciate the difference in prognosis between OC and non-cirrhotic CLD, as well as to characterize the faster progression of OC. Because OC and non-cirrhotic CLD groups had similar follow-up, including number of clinical visits and ultrasounds, the difference in clinical outcomes among groups reflect a more rapid disease progression in patients with OC. In the OC group, the cumulative incidences of HCC, ascites, and variceal bleeding were 1%, 5%, and 1% respectively. Of note, in the HCC and variceal bleeding cases, clinical signs of advanced liver disease including thrombocytopenia or ultrasonographic signs, did not precede the clinical outcomes. This was likely due to suboptimal monitoring of patients in the OC group as it was under-staged. This shows that clinicians heavily rely on physical, ultrasonographic and endoscopic findings in order to ensure a diagnosis of cirrhosis and to commence appropriate surveillance program. In our experience, the suboptimal monitoring in OC translated into higher rates of late diagnosis (60%) as compared to the clinically evident cirrhosis group (15%). The patient with HCC in the OC group was diagnosed at 14-month follow-up with intermediate stage by BCLC while the case of bleeding from esophageal varices was diagnosed at 18-month follow-up. With appropriate surveillance, these cases could have been captured earlier. On multivariate analysis, OC was associated with progression to clinical outcomes when compared to non-cirrhotic CLD.
In our study population, the probability of clinical outcomes was increased according to TE measurements. The overall prognostic performance of TE was excellent, with an AUC of 0.90. Our findings confirm recent studies conducted in chronic hepatitis C and B and underline the importance of TE for risk stratification and early determination of prognosis of patients with CLD.
We acknowledge several limitations of our study. First, this study was conducted at a tertiary-care referral center, and although our prevalence of cirrhosis (19%), is comparable to other university centers, this is higher than that reported in primary/secondary care settings and in the general population. Therefore, a parallel over-estimation of the prevalence of OC is possible. Likewise, our study population may have an over-representation of HIV co-infected patients due to possible referral bias. However, our findings reflect the current situation of a tertiary care liver center, with increasing referral of HIV co-infected patients with progressive liver disease. These centers are also most likely to have a TE machine. Second, this was a retrospective study, and as such we were unable to control for potential confounding factors, including the decision to proceed to TE, which could potentially select the sickest patients. Third, liver biopsy was not available for all patients in the OC group. However, advanced fibrosis/cirrhosis was confirmed in the majority of OC patients who had histological evaluation. Moreover, we validated longitudinally our findings by demonstrating that OC group had worse clinical outcomes than the non-cirrhotic CLD group.
In conclusion, this study suggests that OC is frequent in a tertiary-care Liver Unit. Clinicians should be aware that the absence of thrombocytopenia, ultrasonographic signs of advanced liver disease and/or normal gastroscopy cannot rule out the possibility of a compensated cirrhosis in its preclinical stage. Patients with OC can exhibit rapid development of HCC, variceal bleeding and ascites. Those patients are also likely to be under-diagnosed and consequently under-monitored if clinicians base their diagnosis exclusively on physical, ultrasonographic or endoscopic signs. This can in turn translate into delayed diagnosis of complications. The implementation of screening strategies of CLD patients with TE at tertiary care centers may help the early identification of the otherwise clinically occult and unrecognized stage of compensated cirrhosis. It can prompt clinicians to initiate a program of surveillance for HCC and portal hypertension. Patients with OC due to hepatitis C could be given priority for antiviral therapy, in light of concerns about cost-effective use of these potent, new interferon free regimens. These strategies may potentially reduce the burden of liver complications in patients with cirrhosis. Prospective studies are needed to validate our study findings and to confirm the impact of early identification of OC and specific interventions in clinical practice.
Source...