Author information

Affiliations collapse¹Department of Internal Medicine I, University Medical Centre of the Johannes Gutenberg-University, Mainz, Germany; Cirrhosis Centre Mainz (CCM), University Medical Centre of the Johannes Gutenberg-University, Mainz, Germany. Electronic address: leonard.kaps@unimedizin-mainz.de.

• ²Department of Internal Medicine I, University Medical Centre of the Johannes Gutenberg-University, Mainz, Germany; Cirrhosis Centre Mainz (CCM), University Medical Centre of the Johannes Gutenberg-University, Mainz, Germany.
• ³Department of Internal Medicine I, University Medical Centre of the Johannes Gutenberg-University, Mainz, Germany.
• ⁴Department of Internal Medicine I, University Medical Centre of the Johannes Gutenberg-University, Mainz, Germany; Cirrhosis Centre Mainz (CCM), University Medical Centre of the Johannes Gutenberg-University, Mainz, Germany; Department of Gastroenterology, Hematology, Oncology and Endocrinology, Klinikum Dortmund, Germany.
• ⁵Department of Internal Medicine I, University Medical Centre of the Johannes Gutenberg-University, Mainz, Germany; Cirrhosis Centre Mainz (CCM), University Medical Centre of the Johannes Gutenberg-University, Mainz, Germany. Electronic address: christian.labenz@unimedizin-mainz.de.

Abstract

Background: In contrast to overt hepatic encephalopathy (OHE), the diagnosis of minimal HE (MHE) is challenging in patients with cirrhosis requiring elaborate, specialized testing. The EncephalApp_Stroop is a smartphone-based application established as screening tool for the diagnosis of MHE but has not yet been validated in a German cohort and country specific cut-offs are currently missing.

Methods: 93 patients with cirrhosis were enrolled into this study. Psychometric hepatic encephalopathy score (PHES) was used to detect MHE, and a subset of the patients was tested with critical flicker frequency (CFF). All patients underwent testing with EncephalApp_Stroop. Cut-off thresholds for EncephalApp_Stroop were calculated according to Youden's Index and a separate cut-off was determined with focus on sensitivity.

Results: 24 (26%) patients had MHE according to PHES. EncephalApp_Stroop had a strong correlation with PHES (r=-0.76, p<0.001), while there was only a modest correlation with CFF (r=-0.51, <0.001). On time as well as on+off time discriminated best between patients with and without MHE with AUROCS of 0.87 for both measures. According to Youden's index, a cut-off of >224.7 seconds (sec) (on+off time) discriminated best between patients with and without MHE with a sensitivity of 71% and a specificity of 88%. The adjusted cut-off value for on+off time with focus on sensitivity (sensitivity:specificity weighed 2:1) was 185.1 sec, yielding an optimized sensitivity of 92% and a negative predictive value of 96%. By using this cut-off as a pre-screening test in a stepwise diagnosis algorithm, elaborate testing with PHES could have been avoided in 49% of all patients.

Conclusion: EncephalApp_Stroop may be useful in a stepwise diagnosis algorithm or even as a stand-alone screening tool to detect MHE in German patients with cirrhosis.