Up to 45% of patients with AIDS develop CMV disease in the era of PCP prophylaxis, usually when their CD4+ T-lymphocyte count drops below 50 cells/mm 3. CMV retinitis accounts for 80-90% of cases of end-organ CMV disease. There are now two drugs approved by the Food and Drug Administration for the treatment of CMV retinitis--gancilovir (DHPG, Cytovene) and foscarnet (Foscavir). Both drugs are virostatic only, and must therefore be taken indefinitely. Ganciclovir is available in both oral and intravenous forms, while foscarnet must be taken intravenously. In a randomized clinical trial, the two drugs were equivalent in controlling CMV retinitis and preserving vision. The choice of drugs is therefore based on systemic considerations. Foscarnet is associated with significantly increased survival (perhaps due to the antiretroviral properties of the drug), but takes much longer to infuse and is less well-tolerated by patients. Foscarnet is nephrotoxic and often causes nausea, electrolyte abnormalities and anemia. Ganciclovir is marrow toxic, with leukopenia the most common reason for discontinuation of therapy. Anemia and thrombocytopenia may also cause significant problems; however, higher induction (7.5 mg/kg twice daily) and maintenance (10mg/kg/day) doses than those previously used appear to be well-tolerated, and may be more effective in treatment of relapse.
Oral therapy is FDA-approved for maintenance therapy only, following standard induction with intravenous ganciclovir for three weeks. Despite its poor bioavailability (about 6%), oral ganciclovir has been shown to be comparably effective to intravenous ganciclovir. Side effects are similar with the oral and intravenous forms, apart from the absence of catheter-related sepsis with the former. The usual dose of oral ganciclovir is 1 Gm (four capsules) three times daily. Ideal candidates for oral therapy are those patients who have peripheral lesions only, who are active injection drug users, or who are poor candidates psychologically or physically for a permanent in dwelling catheter.
A randomized trial comparing intravenous ganciclovir, foscarnet, and combination ganciclovir and foscarnet for relapsed CMV retinitis found that combination therapy doubles the time to further relapse compared to monotherapy with either drug alone. Combination therapy was not associated with an increased risk of side effects, but was associated with a reduction in the quality of life as a result of the duration of time required for the infusions.
It is not known if relapse more often reflects waning host immunity or true antiviral resistance. Ganciclovir resistance has been associated with mutations in the UL97 gene, which is responsible for the first phosphorylation step of ganciclovir to the active drug ganciclovir triphosphate. We are currently engaged in a prospective epidemiologic trial to determine the frequency and clinical correlations of antiviral resistance in patients with CMV retinitis.
The ganciclovir intraocular device (GIOD, Vitrasert) was approved by an FDA advisory panel in December 1995. The device consists of a 6 mg pellet of ganciclovir encased within a plastic strut which is sutured to the inside of the eye. The device permits a steady-state level of intravitreous ganciclovir to be achieved which is above the ID50 for most CMV isolates. It is designed to last for six months, at which time it can be replaced or a second device can be implanted in the eye. Most patients tolerate the surgery well under local anesthesia. There is a temporary reduction in vision as a result of surgically induced astigmatism, which typically resolves over 4 weeks. Most importantly, the median time to relapse in a national Eye Institute sponsored trial of newly diagnosed CMV retinitis was 209 days in eyes receiving the GIOD, compared to 13 days for deferred-treatment eyes. A study sponsored by the manufacturer showed a time to relapse of 201 days, compared to 72 days in eyes of patients receiving conventional intravenous ganciclovir. Issues that remain undecided at this point include the efficacy of the GIOD in eyes in which the retinitis has already relapsed in which silicone oil has been used as part of retinal detachment repair, the long-term risks of surgical complications such as retinal detachment and infection, and cost. As with all local therapy, the GIOD works only in the eye in which it is used. It is expected that device will be used in combination with oral ganciclovir to reduce the risk of second eye and extraocular CMV disease, which will substantially increase the cost and side effects of therapy. Despite these concerns, however, most experts believe that the combination of the GIOD and oral ganciclovir will become first-line therapy for CMV retinitis when the device is commercially available.
Retinal detachments are a frequent cause of blindness in patients with CMV retinitis. Vitrectomy and silicone oil tamponade is often effective in reattaching the retina. However, visual results are often disappointing, in some cases due to cataract or the farsightedness and other optical effects induced by the silicone oil. Standard phacoemulsification cataract extraction, combined with intraocular lens implantation, can be effective in restoring vision in these cases, although often not to presurgical levels.
Several promising anti-CMV drugs are currently in clinical trials. One is cidofovir (HPMPC, Vistide), a nucleotide analog with a broad spectrum of activity against herpes viruses. Cidofovir does not require phosphorylation to be taken up into cells infected with CMV, so that it may protect uninfected cells. Its long intracellular half-life permits maintenance infusions to be given every two weeks, allowing patients to avoid permanent catheter placement. Cidofovir is nephrotoxic. Probenecid appears to substantially reduce the risk of renal toxicity but is associated with side effects of its own, including fever, nausea, and hypersensitivity reactions. We are currently investigating intravenous cidofovir in the treatment of newly diagnosed peripheral CMV retinitis.
MSL 109 (Protovir) is a monoclonal antibody directed against a CMV antigen. MSL 109 was shown in preliminary studies to double the time to relapse when combined with standard intravenous therapy. Because it is a humanized rather than murine or chimeric antibody, MSL 109 has not induced antibody responses which would block its effect over time, and side effects appear remarkably few. We are investigating MSL 109 in a randomized, placebo-controlled trial in combination with conventional therapy for either newly diagnosed or relapsed CMV retinitis.
Oral ganciclovir is FDA-approved for prophylaxis, based on data from Syntex (now Roche) which has not yet been published except in abstract form. A CPCRA study (also available only in abstract form) failed to show a beneficial effect, but this study has been criticized on methodological grounds. There are concerns about the development of resistance in patients maintained on long-term prophylaxis with oral ganciclovir; the only study available so far on this subject did not indicate resistance to be a problem. Nonetheless, the toxicity of oral ganciclovir and especially the cost of the drug render these issues moot for most patients, unless subsequent studies indicate prophylaxis to be cost-effective in the long term.
While there is no cure for CMV retinitis, therapeutic options have increased enormously since intravenous ganciclovir was first approved in 1987. Most patients are now able to tolerate some form of maintenance therapy indefinitely, and as a result, useful vision is often preserved in at least one eye for the duration of the patient's life.