Are antiviral cocktails increasing the virulence of HIV?

I came across an article published by ScienceDaily about a new, highly virulent recombinant HIV virus that seems to be able to go from initial infection to AIDS in a matter of only three years, as opposed to the average of ten (assuming no treatment). Although this isn't the first time an extra virulent HIV strain has been found (this strain leading to the development of AIDS in only five years) such a small window of time seems to produce serious issues with detection in infected individuals.
        A study published by Ndawinz et al. in 2001 seems to suggest an estimated mean infection to diagnosis time of 53 months for heterosexual males and 37 months for homosexual males in a sample of men in France. Although it is unlikely that such a study is entirely generalization to the populations of other developed countries, there's no reason to believe a significant variance in times from infection to diagnosis in these areas. As such, this strain seems to be able to cause significant increases in viral load before any sort of detection can occur, allowing uncontrolled spread before the infection can be treated.
The reasons for such a strain evolving is unclear, and it may simply be a random mutation that will be out of the headlines permanently in the next few weeks (as was the fate of the 'five year strain' above). It's hard to be sure, though it has been suggested that HIV may be getting more virulent throughout since its initial outbreak 30 years ago. The selective pressure that pushes for the increased virulence of HIV is a result of a plethora of factors, though the driving force is likely the widespread use of drug cocktails for treatment of infections in developed countries. Increases in viral load (concentration of HIV virus in blood) lead to increased probability of infection from a host to a sexual partner in any one sexual encounter. After treatment with a antiviral cocktail, viral load in infected individuals drops significantly, and as a result, so does infectivity, an incredibly unfavorable outcome for the virus.
Larger spikes in viral load over a shorter amount of time seem to counteract this effect if such a spike occurs before detection and administration of antiviral drugs. Such a significant spike in virulence naturally leads to more infections in less time, and as a side effect of the pronounced increase in virulence, more deaths.
The above speculation should not be taken as fact, as a study by Oxford University as well as one released by Nature seem to suggest there is a lot more to the story, with reduced virulence seen over time by the Oxford study as a result of antiviral drug cocktails.
The Nature study, on the other hand, seems to suggest a 'bottleneck' after each moment of person to person spread, where the virus must evolve to be better suited for the new hosts immune system leading to reduced overall fitness for the beginning stages of infection. It is also suggested by the study that less virulent infections may very well result in a higher spread of HIV virus, as a longer living host may have more opportunities to spread the virus. The relevant quote is found below:

" Increased transmission efficiency is associated with higher viral loads, which are highest during acute infection. However, the short time interval of acute infection provides limited opportunity for transmission. As the disease progresses and viral load increases, the efficiency of transmission will increase even though opportunities for transmission can remain constant. Consequently, more transmission events might result from prolonged, compared with rapid, disease progression ... even when accounting for the high transmission efficiencies during acute infection^149, 150. In this model, owing to the fitness bottlenecks at transmission and the subsequent loss in fitness caused by CTL escape^{111, 112, 113}, the circulating HIV-1 isolate will start to lose virulence. This loss in virulence would decrease pathogenesis, prolong the time to AIDS or even result in a non-symptomatic infection. The complete clearance of HIV-1 is unlikely because of the stable integration of the provirus into memory T cells with long half lives and possible re-activation of virus replication in these cells"

With widespread use of antiviral cocktails in more developed nations, it seems the benefit of the weaker, longer infection is strongly countered. The virus is forced to either develop resistance to the drug cocktail, or infect as many potential hosts as possible in the period pre-diagnosis. The suggestion that a longer term infection can lead to more opportunities for infection seems to be a difficult one to accept, especially when strains are beginning to emerge that can more strongly counter the early infection bottleneck.

So is HIV getting more virulent over time? It's a very difficult question to form a definitive opinion on, and lots of contradictory research exists on the topic. There seems to be pressures for increased virulence in areas with greater access to antiviral drugs, while countries with poor health infrastructure seem to favor a more mild virus with a longer progression. The virulence pattern of HIV will prove to be an important consideration in the future, and as the sixth largest killer of humans overall, it is a truly important consideration in the fight against human disease globally.


Above image from the following article: http://www.dddmag.com/news/2014/08/protein-family-can-inhibit-hiv-release