The incidence of cytomegalovirus (CMV) retinitis in AIDS has declined significantly due to the use of highly active antiretroviral therapy (HAART). However, patients with HIV, especially those failing HAART, may still suffer with CMV retinitis, which can lead to significant loss of vision and blindness. Ganciclovir has traditionally been considered the recommended treatment for CMV retinitis; however, due to side effects and the possibility of developing viral resistance, other agents may be preferred in certain situations. Foscarnet, which has similar efficacy to ganciclovir but a different side effect profile, is more difficult to administer and is less well-tolerated. Intravenous cidofovir, which may be more effective than either iv. ganciclovir or foscarnet, can also be used as a first line agent; however, it is associated with toxicity (renal and ocular) and thus needs careful use. Local therapy for CMV retinitis has been a significant advance. The intraocular ganciclovir implant has the highest efficacy of the approved agents and is well-tolerated. Fomivirsen, an oligonucleotide injected intravitreally, is a newly approved agent which offers alternative treatment. Intravitreal ganciclovir or foscarnet, although not approved, have been used successfully in some patients especially those with recurrent or refractory disease. The development of new anti-CMV agents has been stalled by the decreased incidence of the disease. Valganciclovir, a prodrug of ganciclovir, offers excellent oral bioavailability and is the closest to approval of all the new anti-CMV drugs. High ganciclovir blood levels are achieved without the complications associated with the requirement for long-term iv. access. The monoclonal antibody (mAb) MSL-109, did not offer a significant advantage when added to traditional anti-CMV therapy. Development plans of other agents such as cyclic HPMPC and lobucavir have been put on hold by their respective manufacturers. Adefovir is a nucleotide analogue that possesses anti-CMV activity, but is currently only being pursued for the treatment of hepatitis B virus. Other compounds possessing significant anti-CMV activity, including BAY 38-4766 and GW1263W94 are still in the early stages of development.