(1) There is still no cure for HIV infection. The short-term treatment aims are to drive viral load below the current detection limit and to increase the CD4+ T cell count, in order to reduce morbidity and prolong survival. (2) Early initiation of treatment has both advantages and disadvantages. If the patient is symptomatic or if the CD4+ T cell count is below 200 per mm3, antiretroviral treatment should be started immediately. If the patient is asymptomatic, the CD4+ T cell count is above 350 per mm3, and viral load is below 50 000 copies/ml, antiretroviral treatment can often be deferred. Other situations should be considered case by case. (3) The benefits of treating symptomatic primary infection have not been adequately documented, and current recommendations diverge. (4) The advent of new antiretroviral drugs has increased the choice, but cross-resistance often limits treatment options. The new antiretroviral family to have emerged since 1999 is the fusion inhibitor; the only representative of this class, enfuvirtide, is reserved for patients with multiple treatment failure. (5) There is broad agreement on the principles of first-line antiretroviral treatment. It should combine at least two nucleoside (or nucleotide) inhibitors of HIV reverse transcriptase and one non nucleoside inhibitor, or at least one HIV protease inhibitor. Comparative studies have now identified the most effective combinations in terms of virological efficacy and tolerability. The combination should be chosen according to its established efficacy, adverse effects, risks of interactions, and convenience. There is no reference combination suitable for all patients. Among the combinations containing a non nucleoside inhibitor, those based on efavirenz are the most effective after 48 weeks of follow-up. There is less agreement on the optimal treatment of pregnant women. (6) Among the HIV protease inhibitor-based combinations, those containing nelfinavir or the lopinavir + ritonavir combination must be taken during meals. The lopinavir + ritonavir combination showed better virological efficacy than nelfinavir in a comparative trial. Experience and safety evaluation are in favour of nelfinavir, but recent American guidelines issued in November 2003 recommend the lopinavir + ritonavir combination. (7) There is no major difference in virological efficacy between non nucleoside inhibitors and protease inhibitors as first line therapy. (8) In combination with protease inhibitors or non nucleoside inhibitors, the best-assessed nucleoside inhibitors (effective for several years) are lamivudine + zidovudine and lamivudine + stavudine. For first-line treatment, combinations consisting of only three nucleoside (or nucleotide) inhibitors are less effective than combinations containing HIV protease inhibitors or non nucleoside inhibitors. (9) Efficacy is monitored on the basis of changes in viral load and the CD4+ T cell count, one month after the beginning of treatment, then about every three months. Attention must be paid to adverse effects, which may necessitate the replacement of the causative antiviral drug, treatment of the adverse effect, treatment modification or, in extreme cases, treatment withdrawal. (10) Treatment failures must be carefully investigated: the cause(s) may include poor adherence, drug interactions, and inadequate plasma drug concentrations. Ongoing antiretroviral regimens do not always have to be modified if treatment failure occurs. (11) Resistance tests can help to determine the most effective alternative in case of virological failure due to drug resistance. The choice of back-up treatments is complex, and is limited by cross-resistance. Multiple lines of treatment fail in about 3% to 4% of patients.
