Hbsag qualitative

hello everyone, could someone tell me how much hbsag is detected to give a positive result in qualitative tests? In other words, how many viruses are necessary to test positive, understanding that a positive is always equal to or greater than 1. Thank you all for your attention.

Hi @Luis,

Good question. The limit of detection for HBsAg in one of the more common qualitative tests (Architect) is ≤ 0.130 IU/mL of HBsAg (https://www.ilexmedical.com/files/PDF/HBsAgQul_ARC.pdf). This is equivalent to ~0.1ng/mL or 0.0000000001 grams per mL of blood. So it’s very sensitive.

It’s hard to say how many viruses because HBsAg is not only there around viruses but can also make up what is called “sub-viral particles”. These are bodies made up of HBsAg, but are not infectious. They actually are the major form of HBsAg in the blood, outweighing viruses by 100000:1. They often contain much more HBsAg too, but if we just take this ratio, then only 0.001pg/mL of HBsAg is associated with virus.

We also know that there’s probably about 360 molecules of HBsAg per virus (Cryo-electron microscopy of hepatitis B virions reveals variability in envelope capsid interactions) and that the molecular weight of the HBsAg is something around 24.0 kDa (I’m using the smallest here to give the highest number). Using these numbers we can say that there’s about 10^-17 g (or 0.00001 pg) of HBsAg per virus. This means that the limit of detection could be around 100 viruses per mL (or ~20 IU per mL).

These are fairly rough, back-of-the-envelope numbers, and I’m happy for my numbers to be checked and confirmed (e.g. @john.tavis). But in all, I think that’s pretty impressive for this test.


Hi all,

Thomas is right, but there’s a much easier way to determine the limit of quantification and limit of detection for HBV viruses in blood than he used (ie, just do quantitative PCR against a standard curve of HBV DNA).

qPCR detection has a limit of detection for the typical clinical assays of around 40 IU virus/mL, with 1 IU of virus = ~2 viruses, so Thomas’ estimate of ~100 is quite good. The most sensitive assays I’ve seen are around 10 IU/mL (about 20 viruses/mL). We can do better than that with research-level assays, but they are only very rarely used in the clinic due to their difficult and expense. The studies I’ve seen indicate that using the more sensitive assays don’t really affect treatment outcome, so the standard ones used in clinical practice are OK.

So in short: Thomas’ estimate is good, the clinical assays are exceptionally sensitive, and the assays give a very good view on what is happening in patients.


Thank you both very much for replying. All the best.

1 Like