Transient Elastography in the Evaluation of Cystic Fibrosis–Associated Liver Disease: Systematic Review and Meta-analysis

Abstract

Background: Cystic fibrosis (CF) is a common autosomal recessive disease. Although pulmonary disease is the main cause of morbidity and mortality, complications of cystic fibrosis associated liver disease (CFLD) are increasingly recognized as a leading non-pulmonary cause of death. Transient elastography (TE), a quick, non-invasive technique has recently been investigated to detect the presence of CFLD in both pediatric and adult CF populations. Objective: To perform a meta-analysis for the use of TE in the detection of CFLD. Diagnostic properties of TE and factors associated with CFLD are summarized. An optimal cut-off for liver stiffness measurement (LSM) as measured by TE will be determined. Methods: Studies evaluating TE to detect CFLD were identified via electronic databases. Two reviewers independently evaluated all relevant studies. The quality of the diagnostic studies was assessed using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2). Published data were summarized in a hierarchical summary receiver operator characteristic (HSROC). Authors of eligible studies submitted data which was normalized to the current accepted diagnostic criteria for CFLD; optimal cut-off, pooled diagnostic properties and factors associated with CFLD were determined. A subgroup analysis for pediatrics was also performed. Results: A total of 69 records were screened after search duplicates were removed. Six studies were used for the meta-analysis. Most studies had low to unclear risk of bias and low concern regarding applicability. Due to heterogeneity, pooling diagnostic properties was not possible. Descriptively, the range of sensitivities and specificities for using TE to detect CFLD were 46 - 100% and 76 - 92% respectively. A sample size weighted HSROC was calculated. Using data submitted by authors, 609 patients were analyzed. The optimal LSM cut-off was >=5.95 kPa yielding a sensitivity, specificity and AUROC of 55% [95% CI 47 - 62%], 87% [95% CI 83 - 89%], 0.76 [95% CI 71 - 80%] respectively. For pediatrics (n = 167), optimal cutoff was >=5.60 kPa with sensitivity, specificity and AUROC of 67% [95% CI 52 - 81%], 86% [95% CI 78 - 92%], and 0.81 [95% CI 74 - 89%] respectively. Other factors associated with CFLD included higher median LSM (4.35 kPa [IQR 3.60 - 5.20 kPa] vs 6.30 kPa [IQR 4.4 - 10.6 kPa]) (p <0.005) and a lower platelet count (302 x 109/L [IQR 251 - 361 x 109/ L] vs 267 x 109/L [IQR 198 - 347 x 109/L]) (p <0.005). Pancreatic insufficiency and sex were not statistically different among those with and without CFLD. Conclusion: To our knowledge, this is the largest cohort of CF patients used to calculate the diagnostic properties and optimal cut-offs for TE in detecting CFLD. This method has low sensitivity but good specificity. We propose that patients above the described LSM cut-off be investigated thoroughly for the presence of CFLD. .