The diagnostic gap in hepatitis C confirmation and the predictive value of the anti-HCV signal-to-cutoff ratio: a real-world study.

PURPOSE: Guidance requires HCV-RNA confirmation for every anti-HCV-reactive individual, yet many are never tested. We quantified this diagnostic gap and examined whether the anti-HCV signal-to-cutoff (S/CO) ratio can help prioritise which untested patients to recall first.

METHODS

In this retrospective study (STARD 2015) at a tertiary hospital, anti-HCV-positive adults were identified from the laboratory information system. Four immunoassay platforms were used over the record period; because S/CO values are not standardised across assays, the accuracy analysis was restricted to the platform covering most of the cohort, while the diagnostic gap was described for the whole cohort. Performance was assessed by ROC analysis, threshold metrics with 95% confidence intervals (CIs), and logistic regression.

RESULTS

Of 773 anti-HCV-positive patients, HCV-RNA was requested in only 354 (45.8%). Untested patients had lower S/CO ratios (median 2.02 vs 3.50; p<0.001), yet 108 of 419 (25.8%) exceeded 6.5. Ninety-one had died before review; of 268 telephoned, 66 attended and two previously undiagnosed viremic patients began treatment. In the single-platform accuracy set (290 patients, 48 [16.6%] viremic), S/CO ratios were higher in viremic patients (median 13.84 vs 2.46) and predicted viremia with an area under the curve of 0.945 (95% CI 0.919-0.968). The Youden-optimal cut-off was 7.78 (sensitivity 100%, 95% CI 92.6-100; specificity 85.5%). On the preceding platform at the same centre the optimal threshold was 6.50 (area under the curve 0.841).

CONCLUSIONS

More than half of anti-HCV-positive patients never completed confirmatory testing, and many could no longer be reached. The S/CO ratio discriminated viremia well, but the optimal threshold shifted between successive platforms at a single centre, so no locally derived cut-off can safely govern who is tested. These data support universal reflex HCV-RNA testing on every reactive sample, with the S/CO ratio best used to prioritise the recall of patients already missed.