The last value is at the end of the trial (week 12), not after...

The last value is at the end of the trial (week 12), not after the trial. The error bars overlap so you really can't say much about this value - the error bar on the Atripla control group is absolutely huge.

If BIT225 was uncovering dormant virus in macrophages they should see a spike of viral RNA in the blood then a quick decay in levels as the dormant virus is chased out of the macrophages. BIT see no spike of virus in the blood.

You can come up with hundreds of plausible hypothesis that explain the results. For example, many of the patients who were given BIT225 developed fever and rashes both which can cause the immune system to become activated. It could be that BIT225 causes fever and this fever is what is increasing the level of activated CD4 cells. It could be that BIT225 has an effect on the lining of the gut which results in more bacterial components getting into the blood which are known to activate the immune system. It could just be that a few of the BIT225 treated patients caught a different illness in the last few weeks of the trial. The list of possible hypostheses is almost endless.

There is nothing inherently wrong with BIT hypothesising anything, but they need evidence to rule out the other equally plausible hypothesis. Unfortunately, because the only result that really mattered (plasma viral load) failed to show any difference with the addition of BIT225, all they are left with some immune activation data that doesn't show anything conclusively. For it to be useful they would need data that BIT225 was the cause of the immune activation and that this activation is beneficial to the patient.

In theory, BIT could run another larger and much more complex trial to try to answer what BIT225 is doing to the immune system and if this important for the patient, but I don't think they have either the money or patent life to do this. This is why I think that no pharma company is going to partner with BIT to develop BIT225 - there are just too many unknowns and too little time.

As a BIT shareholder you need to decide what BIT is worth if BIT225 is dead. It is not worth zero, but I don't think it is worth $40 million either.