Validation of the TREAT-B score for hepatitis B treatment eligibility in a large Asian cohort: TREAT-B improves with age

Abstract

We read with interest the articles by Shimakawa et al. outlining a new score to determine treatment eligibility in Africans living with chronic hepatitis B (CHB) 1,2 and by Johannessen et al. 3 validating this score in a hospital-based Ethiopian cohort. The TREAT-B algorithm is a simple score using widely-available alanine aminotransferase (ALT) and HBeAg tests to determine CHB treatment eligibility and to overcome barriers to treatment assessment arising from limited access to HBV DNA nucleic acid testing and transient elastography. 4 Johannessen and colleagues found the TREAT-B algorithm did not perform as well as described previously. 2,5 Both groups found WHO guidelines performed poorly for treatment eligibility. Given the high CHB prevalence in the Asia-Pacific (>115 million people affected) 6 and urgent need for simplified treatment algorithms in lowresource countries within this region, we validated the TREAT-B algorithm in a large Asian-Pacific cohort of patients with CHB. Complete clinical data for 1,358 treatment-naïve patients with CHB from 2 prospective clinical studies conducted in Melbourne, Australia were analysed; 976 (72%) were recruited from a single large metropolitan hospital and 382 (28%) were recruited from primary practice sites. Patients with HIV, HCV or HDV coinfection, hepatocellular carcinoma, decompensated cirrhosis or pregnancy were excluded. We validated the TREAT-B algorithm against the gold standard EASL 2017 guidelines, 7 which replaced the previous EASL 2012 guidelines against which TREAT-B was originally validated. 1 EASL 2017 guidelines recommend treatment in patients with a) cirrhosis and a detectable viral load; b) F2 or greater fibrosis stage and viral load >2,000 IU/L; c) ALT greater than 80 IU/L and viral load >20,000 IU/L; d) Metavir score (liver biopsy) > − A2 and viral load >2,000 IU/ml; e) HBeAg-positive, age >30 years-old and viral load >2,000 IU/L; and f) family history of hepatocellular carcinoma. 1 To define severe fibrosis/ cirrhosis for the purposes of treatment eligibility, we used a transient elastography (TE) cutoff of >12 kPa (Fibroscan®, Echosens, France) for those with elevated ALT and >9 kPa if normal ALT, in accordance with EASL guidelines (2017) 7,8 or cirrhosis confirmed on liver histology. The TREAT-B score was derived by addition of HBeAg status (positive 1 point, negative 0 points) to ALT level (<20 IU/L, 0 points; 20-39 IU/L, 1 point; 40-79 IU/L, 2 points; > − 80 IU/L, 3 points) 1 ; an overall score > − 2 was deemed eligible for treatment. Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of TREAT-B was compared to EASL 2017 guidelines. 7

Assessment and validation of the TREAT-B score to assess the treatment eligibility of patients with chronic hepatitis B virus infection

“…The sensitivity and specificity of TREAT-B were reported as 53% and 83.4% in Ethiopia ( 16 ) and 69.8% and 70.4% in Burkina Faso ( 17 ), respectively. A multicenter study in Australia revealed that TREAT-B had greater sensitivity (91%) but lesser specificity (63%) in the hospital than in the community cohort (sensitivity, 70%; specificity, 88%) ( 18 ). TREAT-B of ≥2 demonstrated a sensitivity of 94.6% and specificity of 46.7% in a cohort of Vietnamese patients with chronic HBV ( 14 ), a finding consistent with this current study showing that using a cut-off of TREAT-B of ≥2 had excellent sensitivity (96.3%) but low specificity (41.4%) and fair accuracy (AUROC, 0.69; 95% CI, 0.58–0.66).…”

“…A previous study from an Australian cohort found that 42% of patients with advanced fibrosis or cirrhosis and 20% of patients with cirrhosis with treatment eligibility were missed using the TREAT-B algorithm ( 18 ). In contrast to our study, 21.7% of patients with ≥F3 fibrosis and 17.5% of patients with cirrhosis were missed in the TREAT-B of ≥2.…”

“…TREAT-B score of ≥2 had an accuracy with an area under the receiver operating characteristic curve (AUROC), sensitivity, and specificity of 0.85, 85%, and 77%, respectively, for selecting treatment-eligible patients with chronic HBV in African cohorts (13). The performance of TREAT-B for identification of HBV treatment eligibility has been evaluated in several subsequent studies in East and West Africa, Australia, Europe, and Vietnam (14)(15)(16)(17)(18). Several factors might affect the performance of TREAT-B, such as age (18), type of HBeAg assay (15), and the presence of HBeAg (19).…”

“…The performance of TREAT-B for identification of HBV treatment eligibility has been evaluated in several subsequent studies in East and West Africa, Australia, Europe, and Vietnam (14)(15)(16)(17)(18). Several factors might affect the performance of TREAT-B, such as age (18), type of HBeAg assay (15), and the presence of HBeAg (19). However, the results were conflicting among the studies.…”

A Novel and Simplified Score for Determining Treatment Eligibility for Patients with Chronic Hepatitis B

“…The WHO’s simplified criteria and the TREAT-B scoring system were validated in African, European, Australian, and Vietnamese populations. They showed variations in performance across geographic areas [ 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 ]. The variations were explained by differences in viral genotype, modes of HBV transmission, and rates of spontaneous loss of HBeAg [ 18 , 19 ].…”