While I am not expert in TB treatment, it is fair to point out the among the HIV antivirals, the nucleosides have no known interactions between this class of agents and drugs that induce p450 such as rifampin. So that a combination based solely on nucleoside RT inhibitors should work.
However, recent data on the most popular triple nucleoside combination, called Trizivir (a combination of AZT, 3TC, and abacavir) has been shown to be less effective than other commonly used combinations such as 2 NRTIs plus a nonnucleoside. And for someone with low CD4 counts, this difference would be most important since people with low CD4 counts have often shown lower response rates to most regimens.
What to do? Well, one approach, now undergoing testing, is to add a 4th nucleoside -- or actually a nucleotide RT inhibitor -- the drug tenofovir. Studies are underway adding Trizivir to tenofovir to assess whether this four drug combination is well tolerated and active for those at high viral loads and low CD4 counts. In the absence of trial data, clinicians have reported using this combination and it is reasonable to predict that it should work well, and be particularly worth your consideration in the circumstance you outline.
While you could use another agent from another class, e.g. a nonnucleoside or PI, the drug interactions with rifampin are often significant enough so as to seriously impair the likelihood of getting adequate drug levels of these agents. In fact, most recommendations I've seen suggest using rifabutin instead of rifampin if you need to use a drug from one of these two classes, and then only with reluctance and perhaps even use drug level monitoring if available.
Hope that helps.
