To What Degree Does "On Meds" = "Undetectable"?

[viking8x6]

Ed: The paper referred to here has since been refuted - see the below posts. Big thanks to @fskn and @ErosWired for pointing it out so quickly!

Here's an interesting paper (dated 2008!!) that suggests that the chance of transmission at very low ("undetectable") viral loads, while very small, is likely to be significant on a population basis. How this plays out in practice with respect to all the other variables affecting risk, I have no idea, but it seems useful to add to the discussion here (even though it is old), as it includes a quantitative estimate of the relationship between viral load and transmission risk.

[think before following links] [think before following links] https://www.who.int/hiv/events/artprevention/wilson_relation.pdf

Edited March 5, 2022 by viking8x6
Source was refuted

[fskn]

The 2008 paper's HIV transmission estimates were refuted less than a decade later by PARTNER 1 and 2, which, furthermore, were empirical studies (involving real human beings), not a mathematical modeling exercise as in 2008.

WHO should watermark the 2008 paper because, if referenced by itself today, it could support medical misinformation.

PARTNER 1 and 2 refute even the lowest of the lower bounds in Table 2 at the top left corner of the fourth page (labeled p. 317 in the journal volume). The detection threshold at the time was 200 HIV virus copies per millilitre of blood, PARTNER 1 found 0 linked HIV transmission in 58,000 sex acts by almost 900 sero-discordant straight and gay couples who did not use condoms, and PARTNER 2 found 0 linked transmissions in 77,000 sex acts by almost 800 gay couples.

Even though no linked transmission occurred in these large, rigorous studies, upper bounds of risk were calculated just in case some transmission might be expected in even larger or even longer studies. "The new results from PARTNER2 are able to reduce the upper 95% CI [confidence interval] to 0.23/100 CYFU [couple-years of followup] for overall risk in gay couples: equivalent to a worst case when a couple would need to have sex for 400 years".

HIV transmission did occur in PARTNER 1 and 2, but never from HIV-positive, virally-suppressed study participants. Instead, genotyping demonstrated that transmission occurred when HIV-negative participants had sex with people other than their own virally-suppressed partners.

The quote is from i-base, a UK-based clearinghouse that provides HIV treatment information to health professionals and the public. The quote is still more technical than I would have liked, but interested readers will find all sorts of technical and non-technical information about PARTNER 1 and 2, the studies that proved that U=U. [think before following links] https://i-base.info/htb/34604

Edited March 5, 2022 by fskn
Table number

[ErosWired]

2 hours ago, viking8x6 said:

[think before following links] [think before following links] https://www.who.int/hiv/events/artprevention/wilson_relation.pdf

This appears to be a case of statistical analysis attempting to make a point, but the Discussion area of this paper is where you start to find the holes in the study. The phrase “Our key, but ultimately unverifiable, assumption” is kryptonite. Also, the statement “For simplicity we also assumed that heterosexual couples do not engage in penile-to-anal intercourse” indicates extremely sloppy design of research. In more than one place they admit that they’ve based their results on assumptions that are either unverifiable or for which data simply does not yet exist, and that they’re dependent on assumptions. You know the old saying - when you assume you make an ass of u and me. I suspect the lack of rigor in the study may have something to do with why it doesn’t seem to have derailed the U=U guidance in the 14 years since its publication.

It might also be noted that their figures, for male-male transmission, are results for 10,000 couples that each have 1000 sex acts over ten years, or 10,000,000 total sex acts, returning a figure of just over 3,500. That’s an incidence, if their raft of assumptions floats at all, of 0.00035.

You buy your ticket and you take your chances, and you make your decision based on your odds.

Edited to add:

@fskn posted the reply above as I was writing this, and documents just how thoroughly properly conducted science addresses things like this.

Edited March 5, 2022 by ErosWired

[fskn]

I have to thank you both, @viking8x6 and @ErosWired, because we live in a time when misinformation is rampant and can have life-changing consequences.