Pretreatment hepatitis C (HCV) RNA levels proved consistently higher with the TaqMan assay than the Abbott assay for patients with high initial HCV loads, according to a 740-sample analysis involving people with HCV genotype 1 infection. Intrapatient variation was substantially greater with TaqMan than with the Abbott test. An international team believes their findings suggest a single HCV RNA measurement may not be a reliable way to pick patients for an eight-week course of direct-acting antiviral (DAAs).
Data from phase 3 DAA trials suggest that some treatment courses may be shortened to eight weeks in DAA-naive noncirrhotic patients with a pretreatment HCV load below 6 million IU/mL or in naive and experienced patients with a pretreatment load below 4 million IU/mL. But the time-consuming assay usually used in DAA approval trials (TaqMan with manual extraction) sees little use in clinical practice. Researchers conducted this study to estimate the reliability of two clinically popular assays in determining the suggested HCV RNA cutoffs of 4 and 6 million IU/mL.
The assays tested are Cobas AmpliPrep/Cobas TaqMan (TaqMan) and Abbott RealTime HCV (Abbott). The investigators conducted two experiments -- one to determine HCV RNA variability between the two assays (intra-assay variability) and one to determine variability in individual patients (intrapatient variability). The intra-assay analyses used random leftover serum samples from HCV genotype 1 patients at two centers in Germany and one in Canada. All samples were tested three times with each assay. The intrapatient analyses involved consecutive noncirrhotic genotype 1 patients from the same centers. All were eligible to receive eight weeks of ledipasvir/sofosbuvir (Harvoni). The investigators collected one or two samples within six months from each patient before anti-HCV therapy began.
The intra-assay analysis involved 501 baseline and on-treatment HCV genotype 1 samples. The difference between the two assays averaged +0.11 log10 IU/mL (P < .001). Viral loads proved consistently higher with TaqMan than Abbott in samples with HCV RNA levels above 1 million IU/mL but significantly lower in samples with HCV RNA levels below 1 million IU/mL (P < .001). For the 4-million IU/mL cutoff, 11% of TaqMan and Abbott samples were discordant; for the 6-million cutoff, 9% of samples were discordant. Whereas 42% of patients had an HCV load below 4 million IU/mL with Abbott, 85% of patients had levels above 4 million with TaqMan. While 73% of patients had an HCV load below 6 million IU/mL with Abbott, 69% had an HCV load above 6 million IU/mL with TaqMan.
The intrapatient analysis involved 239 consecutive previously untreated noncirrhotic patients with HCV genotype 1. Baseline viral load averaged 6.3 log10 IU/mL with TaqMan and 6.0 log10 IU/mL with Abbott. Significantly fewer baseline viral loads were below 4 and 6 million IU/mL with TaqMan (64% and 78%) than with Abbott (87% and 95%) (P < .001 for both comparisons). With the 4-million and 6-million cutoffs, 23% and 18% of samples were discordant between the two assays. The analysis included 86 patients with samples available at different times within six months. TaqMan yielded discordant intrapatient results for 13% of patients at the 4-million cutoff and for 17% at the 6-million cutoff. Respective discordant proportions with Abbott were 2% and 1%.
The researchers conclude that "eligibility for the shortened [eight-week] treatment duration based on [viral load] threshold varies significantly depending on the HCV RNA assay." That discordance suggests "it would be more accurate to establish assay-specific thresholds" rather than using a 4-million or 6-million threshold for all patients. They caution that "current guideline recommendations that support the 8-week [ledipasvir/sofosbuvir] regimen as a first-line treatment option in genotype 1 infection may put patients at risk of misclassification with regard to the optimal treatment duration." The authors believe their intrapatient variability results "indicate that selecting patients for 8 weeks of therapy on the basis of a single HCV RNA level measurement may not be useful in clinical practice." Finally, the researchers suggest that determining treatment duration "may be based on low baseline HCV RNA levels with a safety margin around the proposed clinical cutoff[.]"