Good question. It's not clear. But a lot of research is being
conducted to understand how some people appear to effectively control
HIV infection for years.

Similar things happen with many diseases--herpes viruses, TB, etc.

        George M. Carter

Several things - it depends on the group:

One bunch of people infected in Australia by the same single infected
donor were long-term non-progressors.  The virus the donor was infected
with contained a deletion in the "nef" gene.

Sadly this cohort has gone on to progress, but it highlighted the fact
that the virus could be attenuated by gene deletions.  They should
probably be more properly characterized as slow progressors.

Others have resistance mutations in human genes that the virus uses to
replicate - most famously CCR5.  CCR5 is a protein used in immune
signalling that the virus uses to enter the cells (the same mechanism
targetted by the new drug Fuseon).  Some people have a deletion in that
gene that makes it non-functional and the virus has a much harder time
getting into the cells.  This mutation is more common in Caucasians but
other racial groups have their own resistance mutations.  All genes
come in two copies (one from mom, one from dad) and people with two
defective copies seem to be resistant to infection.  People with a
single defective copy may progress more slowly than usual - but it's
not a reliable effect.

Most commonly it's simply that the LTNP's have far better immune
responses against the virus.  The antibodies don't seem to work very
well (not surprisingly, since antibodies are better against bacteria
than viruses) but the cytotoxic CD8 T cells are used against viral
infections.  People with stronger CD8 T cell responses have lower
amounts of virus in the system, lower amounts of viral replication, and
progress more slowly.  Some may not progress at all...these are the few
true non-progressors.

Whether you make a good or a bad initial response to HIV infection may
be the key, and that is partly defined by your genes (people with
certain combinations of immune-system genes do better than others) and
partly by luck (some people with the "inefficient" gene mixes still get
protection, and some people with the "efficient" gene mixes do still
get infected).  The amount of virus in the innoculum may play a role -
a small amount may not be able to get as much of a replicative foothold
before the CD8 T cells get their act together.  That can take several
weeks, to mount an effective CD8 response.

Contrary to what you may read elsewhere, these LTNPs are not being
ignored or swept under the carpet, but as George says, being actively
researched to find out what makes them so special.  Some of the most
important findings have come from this field - by comparing LTNPs with
normal progressors you can start to work out how HIV causes AIDS and
how the body tries to stop it.

What's striking about HIV is that only 5-10% of people fall into this
category, if you look at long-term untreated cohorts.  Compare that to
Ebola or Marburg, which may have a 50%-75% long-term survival!  Even
the most recent Marburg outbreak in Angola, the worst ever recorded,
has just under 90% mortality - and that's probably due to the fact that
it started in children rather than adults.  HIV may take its time, but
it's extraordiarily lethal.  The drugs do help, but with the specters
of resistance and side effects, they're not a permanent solution for
many.  The best approach is still prevention.

Cheers

Bennett