THE OI REPORT: A Critical Review of the Treatment & Prophylaxis of AIDS-Related Opportunistic Infections (OIs)

CYTOMEGALOVIRUS RETINITIS (CMV)
by Michael Marco

INTRODUCTION
Before AIDS, cytomegalovirus (CMV) retinitis was a rare viral disease occurring only among individuals with primary immunodeficiency syndromes or autoimmune disorders, organ transplant patients and immunosuppressed cancer chemotherapy patients (Fiala 1975). In bone marrow transplant patients, there is a greater incidence of CMV pneumonitis than retinitis (Coskuncan 1994).

CMV infection is not the same as CMV disease. CMV infection simply means that CMV organisms live in one's body and that one tests positive for CMV antibodies. Most HIV-positive and HIV-negative people have been exposed to CMV at some time. People may be infected perinatally, in early childhood, or through sexual contact. The initial infection (referred to as primary infection) may present as a mononucleosis-like syndrome which soon resolves; most people then remain latently infected but asymptomatic for life (Jordan 1983). A very high proportion of gay men -- over 90% -- have antibodies to CMV (Drew 1981; Mintz 1983). Primary infection of CMV is thought to be rare in HIV-positive individuals because most have been exposed to CMV prior to their HIV infection (Gallant 1992).

CMV can be transmitted through viral shedding in both oral and genital secretions which allows for primary infection or reinfection (Merigan 1990). Studies conducted at San Francisco's UCSF noted that receptive anal-genital intercourse correlated strongly with acquiring CMV (Mintz 1983) and that condoms -- using an in vitro model -- blocked CMV transmission (Katznelson 1984).

CMV disease occurs when a previously latent CMV infection is activated -- usually when a patient is immunosuppressed -- and causes clinical "end-organ" disease, such as CMV retinitis [affecting the eye(s) and causing a loss of vision], pneumonitis (lungs), encephalitis (brain) or colitis (colon).

From the 1980s to 1996, the incidence of CMV end-organ disease among people with AIDS was estimated to range between 10 and 40% (Peters 1991; Gallant 1992; Hoover 1993; Chan 1995). In the new era of potent protease inhibitor combination therapy, the incidence rate of CMV - for those who have access to protease inhibitors -- is down to under 5% (Baril 1997, Currier 1997). CMV retinitis is by far the most common manifestation, accounting for between 77 and 90% of all CMV disease in people with HIV (Peters 1991; Gallant 1992).

CMV end-organ disease occurs late in the course of AIDS and is associated with extremely low CD4 counts. The average CD4 count in patients with newly diagnosed CMV retinitis is below 30 (SOCA/ACTG 1992; Kuppermann 1993). A 132 patient natural history study noted that of the 26 patients who developed CMV retinitis (20%), none had a CD4 count above 50 (Kuppermann 1983).

The risk of developing CMV retinitis increases markedly as one becomes more immunocompromised and CD4 cells decrease. In a 1002 patient natural history study, Gallant and colleagues found that the probability of CMV disease (at two years) was 21.4% for patients below 100 CD4 cells as compared to 10.3% for patients with CD4 counts above 100 (p<0.001) (Gallant 1992). The incidence of CMV also increased since Pneumocystis carinii pneumonia (PCP) prophylaxis became standard of care. Multicenter AIDS Cohort Study (MACS) data reveals that the cumulative incidence of CMV retinitis in AIDS patients rose from 24.8 to 44.9% after the introduction of primary PCP prophylaxis -- presumably because people who previously died of PCP now lived long enough to develop CMV (Hoover 1993).

The advent of potent protease inhibitor combination therapy (PPICT) has dramatically reduced the incidence of CMV disease, as evidence from both randomized, controlled trials and from epidemiological studies testifies:

[ASDC = Adult Spectrum of Disease Cohort; IDV = indinavir; NRTI = nucleoside reverse transcriptase inhibitor; PI = protease inhibitor; RTV = ritonavir]

While before the protease era virtually all CMV cases presented with CD4 counts below 50, there have been several reports of individuals whose CD4 counts rose after starting PPICT but who nonetheless developed CMV disease - usually in the first weeks or months of PPICT, presumably either before the immune system had a chance to recover, or else reflecting a permanent deletion of CMV-specific T-cell immunity. Jacobson and colleagues reported on five patients with CD4 counts over 195 who developed CMV retinitis just four to seven weeks after initiating PPICT (Jacobson 1997). Each of their CD4 counts had been below 85 five to 24 weeks before initiating PPICT.

To demonstrate how the common CD4 threshold for CMV retinitis has changed has changed after PPICT, Jacobson conducted a retrospective analysis of 76 patients who developed CMV retinitis between July 1995 and August 1996. The data below document the change:

(Jacobson 1997)

Similar results from a prospective cohort study were recently reported by a group from Spain. Mallolas and colleagues documented 21 cases of CMV end-organ disease (nine initial episodes and 12 relapses) in patients on PPICT (Mallolas 1997). Nineteen of 21 (90%) cases occurred within two months of initiating PPICT, and none occurred during the three month median follow-up period (range: 0 - 8 months). The median CD4 count of the 21 cases was 30 (range: 2 to 225), however, 25% of the patients had CD4 counts above 50 at diagnosis.

Thus, CMV retinitis will develop or progress in some patients despite a rise in CD4 counts to levels previously deemed safe. It is possible that these patients had asymptomatic CMV disease or were CMV DNA PCR positive (viremic) and that the CMV virus had already seeded the eye before they started PPICT. Once CMV is in the sanctuary of the eye, PPICT might not effectively halt the clinical progression to retinitis.

PATHOGENESIS
The pathogenesis of CMV retinitis remains incompletely understood. CMV is a herpes virus which infects endothelial cells, monocytes, and granulocytes (Dunn 1996). After 20 years, leading researchers still ask, "What are the key pathogenetic determinants of CMV end-organ disease?" (Jacobson 1995).

CMV DNA PCR
Many clinicians realize that using only a patient's CD4 count to predict the risk of CMV disease is not sensitive or specific enough. Hence, the search for more sensitive markers to predict the development of disease, including polymerase chain reaction (PCR) and antigen testing, is currently in process.

Shinkai and colleagues followed 94 patients in AIDS Clinical Trials Group (ACTG) 181 with quarterly urine and leukocyte (blood) cultures and plasma PCP to determine the risk of developing CMV disease. After a mean follow-up of 12 months, 26 (28%) patients developed CMV end-organ disease. Only two had a negative CMV DNA PCR. Urine cultures, leukocyte cultures and qualitative plasma PCR had a positive predictive value of 31%, 69% and 58%, respectively. However, a CMV DNA PCR copy number of more than 1000/µl (microliter) had a positive predictive value of 100%.

CMV Disease Based on Qualitative & Quantitative CMV DNA PCR
	CMV disease	No CMV Disease	p-value
CMV DNA PCR +	24 patients	17 patients	0.006
>1000 copies/µl	9 (100%)	0 (0%)	0.002 (vs. <100)
100-1000 copies/µl	10 (59%)	7 (41%)	NS
<100 copies/µl	5 (33%)	10 (67%)	0.002 (vs. >1,000)
Median CMV DNA PCR	473 copies/µl	35 copies/µl	0.0007

(Shinkai 1997)

Spector and colleagues conducted a virology sub-study within Syntex/Roche 1654, an oral ganciclovir CMV prophylaxis study. Baseline plasma samples were evaluated in 85% of the 725 patients. About 55% of the patients were CMV DNA PCR positive at entry in both treatment groups (ganciclovir and placebo).

Plasma CMV DNA PCR in Syntex/Roche 1654 Oral GCV Prophylaxis Trial
Baseline Plasma CMV RNA (copies/mL)	Proportion developing CMV end-organ disease
	Ganciclovir	Placebo	N
<50,000	20%	40%
50,000-150,000	75%	75%	16
>150,000	100%	100%	10

(Spector 1996c)

Patients Developing CMV End-Organ Disease
	Ganciclovir	Placebo	p-value
Plasma PCR- at baseline	1%	14%	<0.01
Plasma PCR+ at baseline	26%	43%	0.017

(Spector 1996c)

This viral load sub-study provided useful natural history data about the relationship between CMV viral load and risk of end-organ disease. In the Roche study, about half the patients were CMV DNA positive at baseline. About 15% of the negatives and 45% of the positives developed CMV end-organ disease. Patients with a high CMV viral load were likelier to develop CMV end-organ disease. Only 40% of the CMV PCR DNA positive placebo patients with less than 50,000 copies developed CMV end-organ disease compared to 100% (all ten) of the oral ganciclovir patients with CMV load over 150,000 (Spector 1996c).

Three prospective natural history CMV studies of 90 to 200 patients, all employing different PCR assays, were published in 1997. Approximately one-third of the patients with an average CD4 count of 50 or less were CMV DNA PCR positive and of that one-third, approximately 58% developed CMV end-organ disease (Shinkai 1997, Dodt 1997, Bowen 1997).

Lastly, ACTG 360, "HIV and CMV Viral Burden and the Development of CMV" will help us identify the approximate HIV and CMV viral load threshold in those who develop CMV. Likewise, this study is attempting to validate five different CMV viral load assays. The study has just completed enrollment with 403 patients.

pp65
Several European studies have tested the diagnostic value of CMV pp65 antigenemia in CMV disease (pp65 is a CMV protein) (Torrus 1996a, 1996b; Reynes 1996). In these small (fewer than 75 patient) studies, over 50% of patients positive for the pp65 antigen developed CMV end-organ disease. While pp65 may be more predictive than CMV culture, data are still too preliminary to recommend its wide-scale use, particularly as direct measurement of CMV viral load in plasma and urine with polymerase chain reaction (PCR) technology is becoming more widespread, though neither are commercially available in the United States (Frances Bowen, personal communication). Larger studies should be conducted to standardize these tests and validate their clinical relevance.

CMV PCR as a Prognostic Tool
Bowen and colleagues used PCR (Roche Amplicor™) to detect CMV viral load in the blood and urine of 45 patients on intravenous (IV) ganciclovir therapy. At baseline, 39 (87%) patients with newly diagnosed CMV retinitis were either blood or urine CMV PCR-positive - 46% in blood only, 23% in urine only, and 31% positive for both. The median pre-therapy viral load in the blood was 4.95 log10 copies/µl (range, 3.6 - 7.05) and in urine was 4.90 log10 copies/µl (range, 4.45 - 6.46). After 21 days of IV ganciclovir induction therapy, 33 (85%) of 39 patients became CMV PCR-negative. Six patients were still detectable for CMV DNA: three from blood, two from urine and one from both. After one month of maintenance therapy, five of these six patients became undetectable, and the other one died. The pretreatment viral loads of blood from these six patients who were detectable after their induction therapy were significantly higher than the other 36 patients (median load 6.18 log10 copy/µl versus 4.79 log10 copies/µl; p=0.005). Overall, there were no significant differences in the rate of progression between those patients who were blood or urine PCR-positive or negative at baseline. When restricting the analysis to blood, a loose trend suggested that patients with elevated blood viral loads (more than 4.95 log10 copies/µl) had a shorter time to first progression (p=0.16) and a significantly shorter time to death (p=0.007). The median survival difference between the two groups was 3.9 months (125 days). This relatively small study suggests that CMV PCR measurement may be useful in determining the prognosis of patients (Bowen 1996). Nonetheless, Bowen believes that PCR is not particularly helpful in monitoring PCR negative patients on maintenance therapy, since the majority of them will progress without systemic viremia due to inadequate drug levels in the eye (personal communication). By contrast, apparently disregarding PCR's prognostic utility, Douglas Jabs asks, "Why use a lab test when an eye exam by an ophthalmologist gives answers?" (personal communication).

In a presentation at the 35th IDSA meeting in San Francisco during fall 1997, Finkelstein and colleagues presented an analysis of CMV viremia (blood) and viruria (urine) from 177 patients in ACTG 181. CMV DNA was found in urine samples from 68/177 (58%) patients, and in blood from 22/177 (12%). Just 33/177 patients developed end-organ disease. However, CMV viruria increased the risk of end-organ disease by 3.08-fold, and viremia by 9.2-fold. After shedding CMV, disease occurred at three months in 5% and at six months in 16%. The authors proposed screening individuals with CD4 counts below 50 monthly for CMV viruria. If found positive, they would be screened for viremia. If viremic, they suggested administering pre-emptive CMV treatment, since the risk of end-organ disease in the CMV viremic population with CD4 counts below 50 appears to be in the range of 20% (Finkelstein 1997). Further prospective studies are warranted.

CLINICAL MANIFESTATIONS & DIAGNOSIS
The main symptoms of CMV retinitis include floaters, blurred vision, missing portions of vision and flashing light/sparks. Even subtle changes, such as a minor loss of peripheral vision, can indicate the development of CMV retinitis. There is usually no pain involved. It is important for patients to report any and all visual abnormalities to a primary care physician so that a referral to an ophthalmologist can be made. Some patients can have progressive CMV retinitis and be asymptomatic. Fifteen percent of patients in one CMV treatment study were asymptomatic yet had sight threatening CMV disease (Kupperman 1993).

Because CMV retinitis can be asymptomatic, and because many patients might not understand the sight-threatening implications of their symptoms, routine ophthalmological screening is recommended (Kupperman 1993). Patients with CD4 counts above 100 should see an ophthalmologist yearly; patients with counts between 50 and 100, every six months; and those with counts under 50, every three to four months (Dunn 1996).

Diagnosis must be made by an experienced ophthalmologist. Clinical manifestations of progressive CMV retinitis include a dry appearing, granular border with little vitreous inflammation. The edema and necrosis are also known to cause irregular patches of retinal whitening. A "brushfire" pattern can be visualized once a photograph of the lesion is enlarged.

Ocular zones have been established to define the location of lesions in relation to key parts of the visual apparatus such as the optic nerve and the fovea. Zone 1 disease (defined as 1500 microns from the optic nerve or 3000 microns from the fovea) is considered immediately sight-threatening and requires prompt treatment. Zone 2 disease immediately adjacent to zone 1 also necessitates prompt treatment. Outer Zone 2 and Zone 3 disease are sometimes referred to a "peripheral retinitis" since the lesion is further from the fovea. Untreated peripheral retinitis progresses, usually within a median of two to three weeks, but patients may be able to wait a few days to explore treatment options.

Fundus photographs are often taken to document the extent of a patient's disease and the response to therapy. The ophthalmologist will dilate the patient's eye and photograph the fundus (the interior of the eye). These photographs are then made into slides and kept as a permanent record.

TREATMENTS FOR CMV RETINITIS
There was no FDA-approved drug for CMV retinitis until 1989, when ganciclovir (DHPG, Cytovene™, made by Syntex, now a subsidiary of Hoffmann-La Roche) was licensed. The drug was approved in the absence of controlled data because of strong community pressure and widespread consensus among clinicians, based on five years of experience with open-label ganciclovir distributed under a compassionate use program to thousands of patients, that the drug was effective. Subsequently, foscarnet (Foscavir™, made by Astra) and cidofovir (Vistide™, made by Gilead Sciences) have also been approved by the FDA for this indication. Efficacy for all three drugs was demonstrated by studies comparing immediate versus deferred treatment for non-sight-threatening, peripheral CMV retinitis (Spector 1993; Palestine 1991; Lalezari 1997b). Chiron Vision's Vitrasert™ intraocular ganciclovir implant is now also FDA-approved, although it provides only local treatment and does not protect the other eye or the rest of the body from disseminated CMV disease.

SYSTEMIC THERAPY FOR PRIMARY CMV RETINITIS

Ganciclovir (DHPG, Cytovene™, Syntex/Roche)
In ACTG 071, Spector and colleagues randomized 42 HIV-positive patients with peripheral CMV retinitis to receive either immediate therapy with IV ganciclovir (5 mg/kg twice daily for 14 days followed by 5 mg/kg for 14 weeks) of deferred treatment. Those on deferred treatment were offered ganciclovir when their retinitis progressed. Progression, documented by retinal photography, was defined as any new lesion equal to or greater than 750 microns or an existing lesion expanding by 750 microns or more. Thirteen immediate treatment patients and 22 deferred treatment patients were evaluable.

ACTG 071: Immediate versus Deferred Ganciclovir for Peripheral CMV Retinitis
	Immediate GCV	Deferred GCV	p-value
Progressed	10/13 (76.6%)	20/22 (90.9%)	0.001
Mean time to progression	66.4 days	19.3 days	<0.001
Median time to progression	49.5 days	19.3 days	<0.001
Progressed by 14 days	8.3%	68.2%	<0.0001

(Spector 1993)

About 40% of the ganciclovir patients developed neutropenia and 25% developed thrombocytopenia. There were no differences in the causes of death between the two groups (Spector 1993). The excessively high doses of AZT (1200 mg/day) then recommended may have exacerbated the rate of neutropenia seen in this study, which enrolled between March 1989 and October 1990.

Data from over 300 ganciclovir-treated patients showed that over 90% become urine and blood culture negative within two weeks of starting therapy (Buhles 1988). Maintenance therapy for life is warranted, because relapse is virtually inevitable two to four weeks after discontinuing treatment (Jacobson 1988).

Foscarnet (Foscavir™, Astra Pharmaceuticals)
Foscarnet's efficacy was demonstrated in a similar immediate versus deferred therapy study in which 24 patients with peripheral CMV retinitis were randomized to receive immediate IV foscarnet (60 mg/kg thrice daily for 3 weeks followed by a maintenance regiment of 90 mg/kg once daily) or deferred therapy. The mean time to progression in the 11 evaluable patients on deferred therapy was 3.2 weeks, compared to 13.3 weeks in the 13 foscarnet patients (p<0.001). Over 25% of the foscarnet-treated patients experienced nephrotoxicity (an elevation in serum creatinine) and over 75% developed hypomagnesemia and hypocalcemia (Palestine 1991). In a more recent study, approximately 13% of the foscarnet patients experienced nephrotoxicity (SOCA/ACTG 1995). Foscarnet patients should have their serum creatinine levels monitored closely, particularly when they are also receiving other nephrotoxic drugs such as amphotericin B. In addition, penile and genital ulcers have been reported in uncircumcised males and in patients receiving high-dose foscarnet (Moyle 1993, Katlama 1992). A subsequent study found that a once-daily maintenance dose of 120 mg/kg was more effective at preventing relapse than the dosage of 90 mg/kg (Jacobson 1993).

Ganciclovir versus Foscarnet for CMV Retinitis
From January 1990 through October 1991, the Studies of Ocular Complications of AIDS (SOCA) Research Group, in collaboration with the ACTG, compared both drugs in the Foscarnet-Ganciclovir Cytomegalovirus Retinitis Trial (FGCRT) (SOCA 1992, 1994, 1995, 1996a, 1996b; Holbrook 1996; Jacobson 1996). No significant difference was noted between either drug with regard to visual outcome (SOCA 1994). The fundus photograph reading center documented that the median time to first progression for the foscarnet patients was 53 days as compared with 47 days for the ganciclovir patients (p=0.997). However, patients randomized to receive foscarnet survived approximately

four months longer (12.6 months versus 8.5 months) than those initially assigned to receive ganciclovir (p=0.007) (SOCA 1992), possibly due to foscarnet's anti-HIV activity (Jacobson 1991).

Despite this apparent survival advantage for foscarnet, "Most physicians currently initiate therapy for CMV retinitis with ganciclovir and reserve foscarnet for treatment of retinitis recurrence or ganciclovir intolerance" (Spector 1996). Moreover, this survival advantage occurred between 1990 and 1991 while AZT and ddI were the only available antiviral agents; AZT's bone-marrow toxicities may have increased those of ganciclovir. The availability of new, more potent antiretroviral combinations may render foscarnet's anti-HIV activity, when weighed against its inconvenience, cost, and nephrotoxicity, inadequate for it to replace ganciclovir as first-line CMV therapy.

TREATMENT OF CMV RETINITIS FOLLOWING REACTIVATION
Even with maintenance, relapse and continued progression of retinitis are common when, whether due to inadequate tissue penetration or the evolution of drug resistance, CMV reactivation occurs. To optimize therapy for CMV retreatment, SOCA and the ACTG conducted ACTG 228, the CMV Retinitis Retreatment Trial (SOCA 1996c), a Phase III study which compared ganciclovir alone, foscarnet alone, and ganciclovir and foscarnet together for patients relapsing on either one.

(SOCA 1996c)

While the rate of progression, loss of retinal area and loss of visual field all favored the combination therapy arm, no difference was detected in the rate of loss of visual acuity. Mortality rates were similar among the three treatment groups. For patients originally on monotherapy, switching to the alternate monotherapy was no more effective than staying on the same drug. About 70% of the patients were ganciclovir-experienced and one-third were randomized to receive ganciclovir monotherapy. The initial response and subsequent relapse at a median of 2 months among the ganciclovir monotherapy patients may indicate poor drug penetration to the eye rather than viral resistance.

ACTG 228 was stopped by the SOCA data safety and monitoring board (DSMB) in mid-April 1995, and its results were discussed with the ACTG Executive Committee and all SOCA investigators a few weeks later. These were the only people privy to the treatment results because of SOCA's policy not to publicly present study data until after publication in a peer-reviewed medical journal. Under much pressure from AIDS activists community and some ACTG investigators, SOCA consented to publicly present ACTG 228 in January 1996. In the ten months between study closure and presentation, clinicians were unable to show slides or mention the results of ACTG 228 at lectures or conferences. Many believe that these data were clinically important, and that SOCA's outdated policy hampered the free flow of scientific information. SOCA subsequently changed its ways and has publicly presented data from its most recent studies in a more timely manner.

Cidofovir (HPMPC, Vistide™, Gilead Sciences)
Cidofovir is a nucleoside analogue with broad-spectrum antiviral activity. It has good tissue penetration to the eye and a long half-life, enabling effective anti-CMV activity with once-a-week dosing. However, it is poorly absorbed orally and must be taken intravenously. Moreover, it is highly toxic to the kidneys and requires co-administration of probenecid, which increases the excretion of uric acid and hence improves the kidney's ability to handle cidofovir. However, probenecid increases the elimination rate of other widely-used treatments such as anti-inflammatories and rifamycins, increases concentrations of the cancer chemotherapy drug methotrexate, and has been reported to cause headache, dizziness, liver cell death (hepatic necrosis), vomiting, nausea, anorexia, sore gums, kidney stones, anaphylactic shock, fever, rash, anemia, leukopenia and hair loss. Patients who are sulfa-intolerant should be particularly careful to avoid developing a severe allergic reaction to probenecid (PDR 1996).

Cidofovir was approved by the FDA for the treatment of CMV retinitis in the summer of 1996. Gilead's new drug application (NDA) focused on three studies: 1) a 48 patient Phase II/III immediate versus deferred study of cidofovir treatment for peripheral CMV retinitis; 2) a 100 patient study comparing two maintenance doses (3 mg/kg versus 5 mg/kg) of cidofovir after induction for the treatment of relapsed CMV retinitis (both sponsored by Gilead); 3) a Phase II/III study of immediate versus deferred cidofovir conducted by SOCA and the ACTG. The Gilead 106 study randomized 48 patients with newly diagnosed CMV retinitis to receive either 5 mg/kg of cidofovir immediately or only when progression occurred (Lalezari 1997b).

(Lalezari 1997b)

Some investigators felt that Gilead skewed its analysis by censoring data on patients who died, progressed, or dropped out, thus exaggerating the treatment effect. These investigators were vindicated on March 15, 1996, when the FDA medical officers presented their analysis, reducing the median time to progression in the immediate cidofovir arm from 120 days to 78 days. Adverse events included a 23% rate of proteinuria (protein in the urine, which may indicate kidney damage), neutropenia (15%), and creatinine increases (an increase in serum creatinine of > 0.5 mg/dl from baseline) in 5%.

Gilead's larger 107 study randomized 100 patients to two maintenance regimens of cidofovir (3 versus 5 mg/kg) after induction for the treatment of relapsed CMV retinitis (Lalezari 1996). Fifty-one patients received the 3 mg dose and 49 received the 5 mg dose. These advanced patients had already progressed an average of four times on either ganciclovir or foscarnet.

(Lalezari 1996)

Once again, the FDA's estimate of the drug's efficacy was more conservative than that of the sponsor. Gilead 107 has yet to be published in a peer-reviewed journal, though it has been presented twice at conferences (Lalezari 1995b, 1996c). Each abstract cited the Gilead estimated time to progression rather than the FDA figures. Proteinuria and increased creatinine levels occurred in approximately 16.5% of the patients and 47% experienced mild to moderate reactions (fever, rash, chills) to probenecid, which must be coadministered with cidofovir.

The SOCA cidofovir study had two stages. In stage one, 30 patients were randomized to immediate or deferred induction with 5 mg/kg for two consecutive weekly induction doses, followed by 3 mg/kg every other week for maintenance. In stage two, the remaining patients were randomized to immediate or deferred induction treatment with the same dose, but increased the maintenance dose to 5 mg/kg. Of 64 patients enrolled, 56 were evaluable by the third DSMB interim analysis.

(Meinert 1996)

While full toxicity data are not yet available, four patients had to discontinue cidofovir due to elevated creatinine levels (Curtis L. Meinert, FDA presentation, 15 March 1996).

Clearly, while cidofovir is active against CMV retinitis, and its once-weekly dosing schedule appears convenient, its nephrotoxicity and the requirement to coadminister probenecid to reduce the risk of serious kidney trouble may limit its applicability. In patients who did not receive probenecid during early trials, the dose-limiting cidofovir toxicity was severe kidney damage associated with dose-dependent kidney tubular cell injury resulting in serious blood abnormalities (Lalezari 1995a). The exact mechanisms of cidofovir-induced injury within the tubule cell remain elusive.

In these three Phase II/III studies - and in the Vistide™ treatment IND program - severe nephrotoxicity was minimal due to intensely close monitoring of concomitant probenecid and prehydration of patients. Guidelines state that probenecid dosing must be "clinic witnessed" and "completed" before cidofovir may be given. Even with probenecid, however, approximately 20% of patients still develop proteinuria and approximately 10% have creatinine increases which mandate discontinuation of cidofovir. Unlike with amphotericin B and IV pentamidine, discontinuation of cidofovir does not ensure a rapid reduction in serum creatinine levels, which often continue rising after discontinuation of cidofovir only to plateau weeks later before returning to near-baseline (Mark Jacobson, personal communication).

Before the drug's approval, AIDS treatment activists and many clinicians voiced their concern that the incidence of cidofovir-associated nephrotoxicity would increase when the drug was approved and in the hands of inexperienced clinicians. Gilead, however, spent most of its time emphasizing the "ease," "freedom," and "convenience" of cidofovir as compared to foscarnet and ganciclovir. They would down-play the seriousness of the nephrotoxicity by explaining that probenecid use and careful monitoring would solve everything.

After a few months of approval, Gilead learned the hard way just how nephrotoxic cidofovir was outside of a closely monitored clinical setting. In September 1996, Gilead sent an "Important Drug Warning" letter to health care professionals that read:

Gilead Sciences. Inc., has become aware of several reports of severe renal impairment associated with the use of VISTIDE (cidofovir injection).... The [package] insert has been revised to emphasize further the importance of appropriate patient selection, treatment administration and monitoring when using VISTIDE.

Pronounced probenecid reactions are another concomitant of cidofovir therapy; between 40 to 50% of patients in these studies had allergic reactions to probenecid. Probenecid is a toxic, sulfa-based drug. Patients unable to tolerate probenecid should not be given cidofovir.

Other serious toxicities have recently been identified with intravenous cidofovir use. Davis and colleagues reported the development of 11 cases of iritis (inflamation of the iris) in 46 treated patients, and six cases of hypotony (reduced eyeball tension due to muscle stretching) in four patients (Davis 1997). Five of the patients were noted to have a persistent decrease in visual acuity. These same toxicities have also been reported by Dorothy Friedberg (NYU) and Palau and colleagues from New Orleans (Friedberg 1997; Palau 1997).

Another concern with cidofovir is the threat of extra-ocular CMV disease. No data on cidofovir's preventive effect on extra-ocular CMV disease have been presented. CMV is not solely an ocular disease. Extra-ocular CMV disease - esophagitis, colitis, gastritis, hepatitis, pneumonitis, and encephalitis - accounts for approximately 20-50% of all cases of HIV-related CMV disease (Gallant 1992; Peters 1991). Unlike foscarnet and ganciclovir, which render 90% of patients culture-negative after two weeks, cidofovir renders only approximately 50% (four of eight evaluable patients) culture negative. Does cidofovir cross the blood-brain barrier in order to prevent CMV encephalitis as do ganciclovir and foscarnet, which are 40% permeable?

For those physicians who still wish to use cidofovir in light of its toxicity profile, many still wonder, "What is the correct maintenance dose for primary therapy?" Gilead study 107 for relapsed patients demonstrated that the 5 mg/kg dose was more effective than the 3 mg/kg dose. The SOCA study, however, was not designed to show which dose was more effective, or less toxic, for maintenance therapy after induction for primary CMV disease.

EXPERIMENTAL SYSTEMIC TREatmeNTS FOR PRIMARY CMV RETINITIS

Valganciclovir (RS-790070, Pro-Ganciclovir, Hoffmann-LaRoche, Inc.)
Pro-ganciclovir, a valine ester of ganciclovir, is a pill being studied in for induction therapy of peripheral CMV retinitis. In asymptomatic HIV-positive patients who were CMV seropositive (with no active disease), pro-ganciclovir was compared to IV ganciclovir and oral ganciclovir to test its pharmacokinetics and oral bioavailability. The absolute oral bioavailability of pro-ganciclovir given in a fasted state was 60.9% compared to 5.6% for oral ganciclovir. Plasma concentrations of pro-ganciclovir following the administration of a single oral dose of 360 mg were higher than those obtained with a 1000 mg dose of IV ganciclovir (Jung 1995). A later multiple-dose, dose-ranging study of pro-ganciclovir determined than an approximate dose of 1,800 mg given immediately after food garnered the same systemic exposure as that of a 10 mg/kg daily dose of IV ganciclovir. The 900 mg daily dose of pro-ganciclovir had comparable exposure to that of the standard daily IV maintenance dose of 5mg/kg (Brown 1997).

Hoffmann-La Roche is testing pro-ganciclovir against IV ganciclovir for induction therapy, followed by open-label pro-ganciclovir for maintenance therapy, in patients with peripheral CMV retinitis at trial sites in the US and internationally. This study, WV15376B, will also assess the effects of induction and maintenance level dosing of pro-ganciclovir on CMV viral load, estimated by plasma CMV PCR.

This study, with only 70 patients, is not a true head-to-head comparison. It is only large enough to reliably detect a 30% difference between the two treatment groups. Nonetheless, this is probably one of the most important CMV studies in the past ten years, because it is the first promising an orally-formulated anti-HIV-associated CMV drug to go into a primary induction therapy clinical trial, and it is the first time a drug has been tested against IV ganciclovir for induction since SOCA tested foscarnet vs. ganciclovir (FGCRT) in 1990. If pro-ganciclovir is shown to be safe and effective (comparable to IV ganciclovir) in this study, and if it does well in Hoffmann-La Roche's planned, yet infinitely delayed, immediate versus deferred peripheral CMV retinitis study, it will completely change the standard of care for treating CMV. Likewise, pro-ganciclovir will finally mean the end to Roche's marketing its approved oral ganciclovir for maintenance therapy and prophylaxis. A Roche employee, so excited about the possibility of pro-ganciclovir replacing oral ganciclovir, privately said, "We've all known that it [oral ganciclovir] is shit."

MAINTENANCE THERAPY

Oral Ganciclovir
Oral ganciclovir (GCV) has been studied in many randomized, controlled trials for maintenance therapy of CMV retinitis (Squires 1993; Drew 1995; Oral Ganciclovir European and Australian Cooperative Study Group 1995; Masterson 1996; Lalezari 1996b). The most important and controversial oral GCV maintenance trial was Syntex 1653, which randomized 123 patients with newly-diagnosed and stabilized CMV retinitis to receive either 3 grams daily of oral GCV or 5 mg/kg of IV ganciclovir. CMV progression was monitored by masked fundoscopic photographs and by fundoscopy from a clinical ophthalmologist (Drew 1995).

(Drew 1995)

Survival and changes in visual acuity were similar for both groups. What accounted for the discrepancies between readings by clinical ophthalmologists and those at the fundus photography reading center? In an editorial published alongside the peer-reviewed journal article, Gary Holland wrote, "This discrepancy most likely reflects the fact that clinicians can see the entire retina with indirect ophthalmoscopy, whereas only that portion of the retina posterior to the vortex veins (the "equator" of the eye) can be photographed routinely" (Holland 1995). Oral ganciclovir looked worse according to the median time to progression than according to the mean. Indeed, as Gary Holland noted, "Because this median is only slightly longer than the length of time to progression reported for untreated disease, some investigators have questioned the value of maintenance therapy with oral ganciclovir" (Holland 1995).

In a letter to The New England Journal of Medicine, SOCA statisticians Mark Van Natta and Janet Holbrook took issue with Drew's conclusion that oral and IV ganciclovir demonstrated relative equivalence in Syntex 1654 and that "oral ganciclovir is safe and effective as maintenance therapy for cytomegalovirus retinitis." They based their disagreement on the inadequate sample size and short duration of follow-up, concluding, "Given the above calculations along with the conflicting results of the fundoscopic and photographic evaluation, with respect to the time to progression and the occurrence of new cytomegalovirus retinitis in the previously uninvolved eye, we believe the authors have overstated their conclusion." (Van Natta 1996)

The unconvincing data from this study - as well as data from Syntex study 1774 (Squires 1993) - have made a majority of clinicians decide either not to use oral ganciclovir alone as maintenance therapy or to limit its use to patients without immediately sight-threatening disease (Jabs 1996a; Jacobson 1996a).

Nonetheless, oral ganciclovir at the dose of 3 grams a day is approved and is being used in the community, most often at a patient's request. Syntex/Roche 2226 has now called into question the efficacy of that 3 gram dose (Lalezari 1996b). Syntex 2226 was a randomized, controlled trial comparing three doses of oral ganciclovir (3, 4.5 or 6 grams) with IV ganciclovir (5 mg/kg daily) for maintenance therapy in 281 patients with CMV retinitis. This study, like the earlier studies, enrolled patients whose CMV retinitis was stable after induction therapy, yet, unlike the previous studies, no restriction was placed on the duration of retinitis at entry.

(Lalezari 1996b)

The difference between the 41 days for the 3 gram arm and the 70 days for the IV arm was significant (p=0.0515). Likewise, there was a trend toward better suppression on retinitis with higher oral doses compared to the 3 gram dose (p=0.09). In this study, patients assigned 3 grams of oral ganciclovir fared more poorly than in previous oral ganciclovir studies in which this dose was used. A sub-group analysis which adjusted for duration of retinitis (less than or more than 100 days of retinitis) showed that, when compared to the IV dose, the 25 patients in the 3 gm arm with more than 100 days of retinitis at entry (median 8 months) did far worse than the 48 patients in the 3 gm arm who had less than 100 days at entry (>100 days, p=0.025; <100 days, p=0.499) (Lalezari 1996a).

These disquieting results have led clinicians such as Jacob Lalezari to advocate using oral ganciclovir maintenance only in patients with stable and early peripheral retinitis (Jacob Lalezari, personal communication). It is unknown if the higher oral doses are equivalent to IV. Taking 6 grams of oral ganciclovir would come to 24 pills a day and cost over $50,000 a year.

Is Stopping CMV Maintenance Treatment Sensible in the Protease Era?
Two small retrospective studies in which patients discontinued CMV maintenance treatment were presented at the 37th ICAAC in fall 1997. Torriani and colleagues from UCSD presented data on eight patients who had "voluntarily" discontinued CMV maintenance therapy after long periods of non-progressive disease; five were on IV or oral ganciclovir and three were receiving intraocular cidofovir. At time of discontinuation, the median CD4 count was 172 (range 63-404) and the median HIV RNA was 68,000 copies/µl (range: <200-508,000 copies/µl). None of eight patients progressed after a median follow-up of 146 days (range: 72- 205). Nevertheless, during this follow-up, six of eight patients had detectable HIV RNA (Torriani 1997). This study created some well-deserved controversy. NIAID's Michael Polis was the first up to the microphone asking Torriani how in the world she could discontinue CMV maintenance therapy in a patient with HIV RNA levels of 508,000 copies/µl.

In Spain, Tural and colleagues conducted a similar yet less risky study. Seven patients with newly diagnosed CMV retinitis were given ganciclovir or foscarnet induction and maintenance therapy, and then followed. All were on antiretroviral therapy without protease inhibitors, were CMV DNA PCR positive, and had a median CD4 count of 56. Within 6.3 months, five patients experienced at least one relapse (eight total relapses). When protease inhibitors became available in Spain, all seven patients initiated PPICT. After 3.5 months on PPICT, all patients had CD4 counts of at least 150, and their HIV RNA and CMV DNA levels became undetectable. At this point, patients voluntarily discontinued maintenance therapy and received weekly opthalmologic evaluations for three months, and then every other week. At the time of presentation, no CMV relapses were documented.

To confirm the relevance of their findings, Tural and colleagues planned a prospective study which will use four specific criteria for entry: quiescent CMV, CD4 count over 150, undetectable HIV RNA, and undetectable CMV DNA PCR. Torriani and colleagues from the ACTG and SOCA have also developed a sound, soon to open, 125 patient study, ACTG 379, "The Effects of Stopping CMV Maintenance Therapy."

LOCAL THERAPY

Chiron Vision's Vitrasert™ Intraocular Ganciclovir Implant
Vitrasert is the brand-name of Chiron's intraocular ganciclovir implant. The ganciclovir implant - half the size of a postage stamp, with approximately 7 months' worth of ganciclovir compressed into a pellet - is surgically placed in the eye through an incision at the pars plana. GCVI-606-NEI was the first multi-centered, randomized, controlled study of the ganciclovir implant and used the familiar immediate versus deferred design (Martin 1994). Twenty-six patients (30 eyes) were randomized to receive either immediate treatment with a ganciclovir implant (1 microgram per hour) or deferred treatment with an implant after progression was detected.

(Martin 1994)

Retinal lesion border activity was significantly reduced 4 weeks after surgery in the 76% of the implant group, whereas everyone in the deferred group progressed (p=0.0001). Extraocular CMV (visceral disease) was reported in 8 (31%) of the 26 patients. The estimated median time to visceral disease was 248 days. Contralateral involvement (CMV occurring in the other, non-implanted eye) developed in 14 (67%) of 21 patients who entered the study with unilateral CMV retinitis. About 50% developed contralateral disease by 203 days. Retinal detachment occurred in 7 (17%) eyes with implants. A temporary loss of functional visual acuity (blurriness) occurred in almost all patients following the surgery. It took approximately one month for this to resolve (Martin 1994).

GCVI-601-CMV was a randomized, controlled trial which compared 2 doses of the ganciclovir implant (1 or 2 micrograms per hour) with IV ganciclovir. 188 patients with newly-diagnosed, active CMV retinitis were enrolled.

* Percentages for contralateral disease are taken from Chiron Vison's Intraocular Implant FDA new drug application (NDA) (Chiron Vison 1995), but were not published in the relevant New England Journal of Medicine article (Musch 1997).

There were no statistically significant differences between the implant groups and the IV group in survival (286 implant versus 262 days IV; p=0.80). Complications arising from the implant surgery were pronounced: 21 cases (11.9%) of retinal detachment, 7.8% vitreous hemorrhage and a 1.7% incidence of endophthalmitis (bacterial infection of the eye) in the implant group. Patients with the implant experienced an approximate four-week period of blurry vision (Chiron Vision 1995; Musch 1997).

Hoffmann-La Roche conducted study 2304 in order to determine whether or not adding oral ganciclovir to the ganciclovir intraocular implant would lessen the development of contralateral and extraocular disease. From May 1994 to July of 1996, 377 patients with unilateral CMV retinitis (CMV retinitis in only one eye) were randomized to receive an implant plus oral ganciclovir, implant plus placebo, or IV ganciclovir.

The incidence of biopsy-proven extraocular disease and photographically confirmed contralateral disease at six months was 37.8% for those in the implant plus placebo group, compared to 22.4% for the implant plus oral ganciclovir group (p=0.016) and 17.9% in the IV group (p=0.001) (Martin 1997a). Of interest, there was a protective effect against Kaposi's sarcoma (KS) in those patients receiving oral and IV ganciclovir: eleven percent in the implant plus placebo group developed KS versus 2.7% in the combination group and 1.5% in the IV arm (p=0.008 for both ganciclovir groups). The protective effect seen with ganciclovir may be due to its anti-KSHV/HHV-8 properties.

Martin and colleagues recently reported on the complications arising from multiple exchanges of implants in 22 eyes of 15 patients receiving a second implant and four eyes of four patients who received a third implant. Complications after the second implant procedure included transient vitreous hemorrhage in five eyes, postinterior inflammation in one eye, and retinal detachment in one eye. It took a 42 days for visual acuity to return to 20/20. Three of four eyes had documented dense vitreous hemorrhage after the third procedure. Thus, while multiple implants continue to control long-term CMV, they are also associated with an increased risk of vitreous hemorrhage (Martin 1997). Lastly, it is believed that patients receiving potent protease inhibitor combination therapy (PPICT) will have longer, sustained quiescent disease and thus require few if any implants to be changed (Daniel Martin, personal communication).

While the implants appear effective in protecting the implanted eye, they appear ineffective in protecting the fellow eye and in preventing systemic disseminated CMV disease. Ongoing studies of the implants with oral ganciclovir (or possibly cidofovir) as maintenance might help with these progressions. In addition, ongoing technological developments in implant technology and surgery are warranted in light of the high rate of retinal detachment and the threat of endophthalmitis (inflammation of eye tissue).

Intravitreal Anti-CMV Therapy
Patients failing or intolerant to all IV therapies may be considered for intravitreal injections. Results from single institution, uncontrolled, non-randomized case-series of ganciclovir, foscarnet, and cidofovir intravitreal injections for CMV retinitis have been published (Heinemann 1989; Cochereau-Massin 1991; Diaz-Llopis 1994; Kirsch 1995a, 1995b; Rahhal 1996a, 1996b). These agents are usually injected directly into the vitreous humor using a tuberculin syringe with a 30-gauge needle placed 3 to 4 mm from the corneal limbus. The drug is delivered in a volume of 0.05 to 0.1 milliliter depending on the dose and concentration (Palestine 1996). As with the ganciclovir implant, patients are at increased risk for contralateral and extraocular CMV disease, as well as early retinal detachment, endophthalmitis, vitreous hemorrhaging, and hypotony (reduced eyeball tension due to muscle stretching).

Intravitreal Ganciclovir
Heinemann administered intravitreal ganciclovir injections to 7 patients intolerant of systemic ganciclovir and foscarnet. Five of 7 patients responded to therapy for 14 to 56 weeks (18 to 58 injections). Two patients immediately discontinued therapy, one due to severe thrombocytopenia (platelet reduction), the other due to retinal detachment. One treated patient developed endophthalmitis (Heinemann 1989). Cochereau-Massin and colleagues studied 44 patients (64 eyes) who were intolerant of or refused systemic anti-CMV therapy. Fifty-two of the first 53 injections (98%) led to cicatrization (scarring). The eight-week relapse rate was 53%. Contralateral disease developed in 11% of patients, and 16% developed extraocular disease. Five patients had retinal detachment, and two had intravitreal hemorrhages (Cochereau-Massin 1991).

Intravitreal Foscarnet
Diaz-Llopis and colleagues studied intravitreal foscarnet in 11 patients (15 eyes) who were either refractory or intolerant to systemic anti-CMV therapy. After 3 weeks of induction therapy, complete resolution was noted in 5 of 8 (62.5%) evaluable patients. The 20-week relapse rate was 33%. No local or intraocular toxicities were noted (Diaz-Llopis 1994).

Intravitreal Cidofovir
Four published studies of intravitreal cidofovir treatment have come from William Freeman's group at UCSD (Kirsch 1995a, 1995b; Rahhal 1996a, 1996b). All studies used 20 microgram injections of cidofovir given every 5 to 6 weeks on patients who either refused or were refractory or intolerant to systemic therapy. In one study, 31 cidofovir injections were administered to 24 eyes in 17 patients. The median time to progression was 55 days. Mild to moderate iritis (inflamation of the iris) developed in 4 eyes, and one case of visually significant hypotony (reduced eyeball tension due to muscle stretching) was reported (Kirsch 1995a). Another study (a prospective case series) of 53 eyes in 35 patients documented anti-CMV activity. Of 24 eyes in CMV antiviral-naive patients, none progressed during the study. Four of 29 eyes in the refractory/relapse patients progressed. One patient had retinal detachment and mild inflammation of the iris developed after 14% of the injections. Irreversible visual significant hypotony developed in two eyes (Rahhal 1996b).

In two larger case series looking at the toxicity of intravitreal cidofovir, more cases of hypotony, iritis, and uveitis (an inflammation of the uveal tract of the eye which may lead to blindness) have been documented (Banker 1997; Taskintuna 1997). Taskintuna and colleagues (UCSD) had documented follow-up on 63 patients who had 246 injections in 93 eyes. Three percent of the eyes (1% of injections) experienced permanent visual loss due to chronic hypotony and another 14% with transient hypotony experience mild-to-moderate visual loss. Transient hypotony in the injected eye could predict chronic hypotony in the fellow eye (p=0.02). The incidence of iritis was 32% (Taskintuna 1997).

With such an unsettling and dangerous toxicity profile, many wonder why Freeman's group and others around the country continue to administer intravitreal cidofovir to CMV retinitis patients. Let us hope that the magnitude of possible toxicities is explained to patients before they start therapy.

ISIS 2922
Isis 2922 is an experimental antisense oligonucleotide used only in intravitreal form. Twenty-four patients with refractory CMV retinitis were enrolled in an open-label, dose-ranging study of ISIS 2922. Intravitreal doses of 83, 165, 330, and 496 micrograms were given once weekly for 3 weeks of induction followed by weekly maintenance doses, or after biweekly induction doses of 330 micrograms. Time to progression for patients receiving three doses was about 8 weeks (range 3 to 50 weeks) (Hutcherson 1995). The side effects were pronounced. A majority of patients developed anterior and posterior chamber inflammation. Most responded to topical steroid treatments, but others were left with acute decreases in visual acuity. Some patients treated with doses of 330 micrograms or higher developed extensive destruction of the peripheral retina and retinal pigment epithelium, resulting in severe and irreversible peripheral visual field loss (Palestine 1996).

MSL-109
MSL-109, an anti-CMV monoclonal antibody directed against specific proteins was shown to supplement the antiviral activity of both ganciclovir and foscarnet in a small open-label, Phase I CMV retinitis study. When MSL-109 was added to either ganciclovir or foscarnet, the median time to progression was over 200 days (Tolpin 1993). Attempting to duplicate these results, SOCA and the ACTG conducted ACTG 294, a multicenter, Phase II/III, placebo-controlled study which randomized 209 CMV retinitis patients taking any approved anti-CMV therapy to receive MSL-109 (60 mg IV every two weeks) or placebo. The study was stopped early by the Patient Data and Monitoring Board because about twice as many relapse patients in the MSL-109 group died as compared with those receiving placebo (p=0.016). Moreover, there was no difference documented in the median time to progression in either arm for either newly- diagnosed and relapsed patients - 67 days for the MSL-109 group and 65 days for the placebo recipients (p=0.753) (SOCA 1997). Many believe the mortality difference to be a statistical fluke, citing an exceptionally low mortality rate in the relapse placebo patients which is incongruous with other treatment and natural history studies. Nonetheless, MSL-109 appears to add nothing to an existing anti-CMV regimen and could well be harmful.

Are Immediate versus Deferred Studies Ethical?
As noted, all three drugs approved for the treatment of CMV retinitis were licensed on the basis of studies which randomized patients with peripheral, allegedly "non-sight-threatening" CMV retinitis to receive immediate or deferred therapy. Yet even when this study design was piloted with ganciclovir in ACTG 071, many in the activist community questioned the ethics of the scheme, asking whether there really was such a thing as "non-sight-threatening CMV retinitis" (ACT UP/New York 1989). Indeed, ganciclovir was licensed later in 1989 on the basis of open-label data gathered in a several-thousand patient compassionate use program. Both ganciclovir and foscarnet have significant toxicities which might have strengthened the case for a deferred therapy (or no treatment control) arm, but now that there are three FDA-approved treatments for induction therapy of CMV retinitis such designs appear highly questionable, possibly unethical and definitely anachronistic. While some may claim that there is no ultimate difference in the final visual outcome for patients in deferred therapy, every study demonstrates a dramatic difference in time to progression between immediate and deferred therapy arms:

[Note: the more conservative FDA estimates were used for the Gilead 106 study. Time to progression in the Vitrasert study refers only to the implanted eye.]

In all four studies, the median time to progression in the deferred arm was a mere two to three weeks, demonstrating the rapidity of the progression of untreated CMV retinitis. We know that patients with any type of CMV retinitis will progress. Clear, rational and ethical reasons need to stated for designing immediate versus deferred studies for CMV retinitis. Quicker time to FDA approval is not one of them, and no one has presented a compelling case for continuing to conduct such studies. Alternative trial designs might include randomizing people to receive a new drug plus standard of care versus standard of care alone for induction, maintenance or relapsed CMV retinitis.

CMV PROPHYLAXIS
Considering the debate about the efficacy of oral ganciclovir for CMV maintenance documented above, it is no surprise that the use of oral ganciclovir for CMV prophylaxis is even more controversial, while other potential prophylaxes, such as valacyclovir or high-dose acyclovir, have either demonstrated an unacceptably high trend towards increased mortality (valacyclovir) or a lack of efficacy (acyclovir). Despite FDA approval of oral ganciclovir for this indication, many clinicians are not initiating oral ganciclovir for CMV prophylaxis , choosing instead more frequent ophthalmologic monitoring and CMV PCR testing.

Oral Ganciclovir
Oral ganciclovir is the only drug approved by the FDA for prophylaxis of CMV retinitis. Two randomized, placebo-controlled, Phase III studies have been completed (Spector 1996; Brosgart 1996). Syntex/Roche study 1654 suggested that oral ganciclovir was effective and reduced the risk of CMV end-organ disease by 49%. According to a study carried out by the Community Program for Clinical Research on AIDS , CPCRA 023, however, oral ganciclovir was no better than placebo in preventing CMV. Below is a comparison of the two studies, which both had a 2:1 ganciclovir to placebo design.

* After Syntex 1654 was terminated early, CPCRA 023 patients were given the option of switching to open-label oral ganciclovir. At that time, exposure to oral ganciclovir was 9.3 months for patients randomized to oral ganciclovir and 2.1 months for patients randomized to placebo.

There were significant differences between patient management strategies in the two studies. Syntex 1654 patients were examined by an ophthalmologist at baseline and at every two months, while CPCRA 023 did not require a dilated eye examination by an ophthalmologist at study entry and only required patients to be examined by an ophthalmologist after visual symptoms arose. Thus, Syntex 1654 was likelier to detect CMV retinitis before it became clinically apparent, and CPCRA 023 only afterwards. CPCRA 023 was attempting to study a real-world scenario, equating "usual care" with ophthalmologic exams only after visual symptoms occur. However, this difference in patient monitoring may not have been the only determinant of the discordant results between the two studies. At the IX International Conference on AIDS, the CPCRA's Carol Brosgart presented a post hoc analysis of 023 prepared by the CPCRA Statistical Center, which found an unexpected negative interaction between oral ganciclovir and ddI. Patients who were on placebo and taking ddI had a very low rate of developing CMV, whereas patients in the oral ganciclovir arm on ddI had a high probability of developing CMV disease. The hazard ratio for developing CMV disease with oral ganciclovir versus placebo if on a ddI containing regimen was 7.48 (p=0.02). Of the 63 patients in the placebo arm taking ddI, only 1 patient developed CMV disease. Conversely, when other antivirals were used with oral ganciclovir versus placebo, the hazard ration was 0.62 (p=0.04). This drug interaction between oral ganciclovir and ddI has baffled many. Previously, there have been data indicating an approximate 70% increase in the area under the curve (AUC) of ddI when it is combined with oral ganciclovir. Likewise, the impressive suggested prophylactic results seen here with ddI (in the placebo patients) have not been documented before. In response, Spector reported that the Syntex/Roche 1654 team did a post hoc analysis of all their patients on ddI and did not find the same results seen in CPCRA 023. Brosgart stood by her team's analysis, but cautioned that the results should be interpreted carefully and that the supposed negative ddI and oral ganciclovir interaction should be investigated further (Brosgart 1996).

In a presentation at the 37th ICAAC, Spector presented a meta-analysis of Roche 1654 and CPCRA 023 indicating that oral ganciclovir may improve survival in patients with advanced AIDS (p=0.04). In the Roche study, according to Spector, CMV positivity at baseline increased the risk of CMV disease by 3.4-fold and the risk of mortality by 2.5-fold. 3,000 plasma samples from 553 of 725 patients from baseline and longitudinal visits were evaluated to correlate the impact of treatment on plasma CMV DNA. Spector divided those starting with CMV PCR positivity into responders and non-responders at two months. Responders' CMV PCR became negative, whereas non-responders' CMV PCR stayed positive. Mortality at twelve months was 20% for responders and 48% for non-responders (Spector 1997). A similar analysis should have been done among those in the placebo arm whose CMV PCR became negative.

Does Oral Ganciclovir Prophylaxis Cause Resistance?
Larry Drew presented resistance data on patients from Syntex 1654 who received at least 90 days of oral ganciclovir (Drew 1996b). Of 39 isolates taken from urine culture (mean ganciclovir exposure of 251 days; range 112 to 564 days), two were found to have resistant virus with an IC50 of more than 12 mcM. In this study, the estimated prevalence of resistance to oral ganciclovir after 8.3 months was less than 1%. However, other clinicians are skeptical of this finding. According to Mark Jacobson, 8 months is not long enough to be testing a CMV prophylactic agent now that we have potent antivirals that are allowing patients with CD4 counts below 50 to live longer. He contends, "If there's an exponential increase in resistance that we aren't catching [in eight months], then we're missing something very important" (Jacobson 1996b). Moreover, Drew believes that urine cultures for testing CMV resistance might not correlate with a patient's active retinitis. He acknowledges, "I don't think we know all the significance of PCR positivity and culture negativity. We may have levels of resistance that don't translate into culture positivity yet could be contributing to what's going on in the eye" (Drew 1996a).

Is Oral Ganciclovir Cost Effective?
Many studies have been conducted to determine the cost effectiveness of oral ganciclovir for CMV prophylaxis. In "The Cost-Effectiveness of Preventing AIDS-Related Opportunistic Infections," a study in this month's Journal of the American Medical Association (JAMA), Freedberg and colleagues compared the clinical impact, cost and cost-effectiveness of prophylaxis for PCP, toxoplasmosis, MAC, fungal infections and CMV, and found oral ganciclovir the least cost-effective prophylaxis:

(Freedberg 1998)

It is not surprising that oral ganciclovir is the least cost-effective. In the virology sub-study of Roche's Study 1654, oral ganciclovir was most effective in those patients who were CMV DNA PCR negative (14% developed CMV in the placebo arm vs. 1% in the oral ganciclovir arm). For those who were PCR positive, only those with fewer than 50,000 copies/µl benefitted from oral ganciclovir. There was no statistically significant difference between drug and placebo in those who were PCR positive (Spector 1996c). Using an expensive prophylactic drug that is only effective in patients at least risk will never be cost effective.

Lastly, the U.S. Public Health Service/Infectious Disease Society of America Guidelines Panel for the Prevention of OIs in Persons with HIV recently published its recommendation on the use of oral ganciclovir for CMV prophylaxis. It states:

Prophylaxis with oral ganciclovir may be considered for HIV-infected adults and adolescents who are CMV seropositive and who have a CD4+ T-lymphocyte count of <50 cells/µl Neutropenia, anemia, limited efficacy, lack of improvement in survival, and cost are among the issues that should be considered... (USPHS/IDSA 1997).

EXPERIMENTAL CMV PROPHYLAXIS

Valacyclovir (Valtrex™, Glaxo-Wellcome)
ACTG 204 was a 1227-patient randomized, placebo-controlled, Phase III study of valacyclovir (VACV) versus two doses of acyclovir (ACV) for CMV prophylaxis in patients with advanced HIV disease (Feinberg 1998). VACV recipients had a 33% reduction in CMV endpoints compared to ACV recipients, and the time to confirmed CMV disease was significantly longer (p=0.01) for those in the VACV arm compared to all those pooled in ACV arms. The tables were turned on VACV arm in the final survival analysis where there was a trend (p=0.06) toward earlier mortality for patients in the VACV arm compared to either ACV arm. This trend toward earlier mortality has never been completely understood and is disheartening because valacyclovir may help prevent CMV. Nonetheless, it is only by conducting randomized clinical studies that important - often life threatening - differences such as this one can be determined. With the increased mortality difference noted in ACTG 204, and because of a warning issued against its use in immunocompromised patients due to its association with hemolytic uremic syndrome / thrombotic thrombocytopenic purpura (potentially fatal bleeding in the kidneys) in this population (FDA Data 1996, Bell 1997), valacyclovir has been cast aside by almost everyone as a potential CMV prophylaxis for PWAs.

Valganciclovir (RS-790070, Pro-Ganciclovir, Hoffmann-LaRoche Inc.)
The ACTG is developing a CMV pre-emptive therapy (targeted prophylaxis) study that will use valganciclovir, the oral pro-drug of IV ganciclovir. Approximately 750 patients with high HIV viral loads and low (50 or less) CD4 counts will be enrolled. Patients will be followed with periodic CMV PCR tests (using Roche's assay) and opthalmologic exams. Those who test PCR positive will be randomized to pro-ganciclovir or placebo with the primary endpoint being development of CMV end-organ disease and a secondary endpoint of death. In this era of PPICT, finding 750 patients with high viral load and very low CD4 counts at ACTG trial sites may be a difficult and long process. Nonetheless, it is essential that this study be carried out. Validating the CMV PCR assay, the concept of CMV pre-emptive therapy, and the safety and efficacy of pro-ganciclovir will greatly help all PWAs at risk.

Adefovir dipivoxil (Preveon™, bis(POM) PMEA, Gilead Sciences)
Adefovir dipivoxil (Preveon™) is a nucleotide analog with in vitro activity against HIV, HBV, HSV-1 and -2, EBV, HHV-6 as well as CMV. Adefovir is now in phase III pivotal studies for HIV, as a CMV prophylaxis, and in phase II for HBV. PMEA (9-[2-(phosphonomethoxy)ethyl]adenine) is a nucleotide analog, differing from a nucleoside analog by virtue of a phosphate - the drug is one step closer to activation when it enters a cell, where it is phosphorylated twice, after which it inhibits the CMV polymerase. The fully phosphorylated nucleotide competes with adenosine, one of the building blocks of DNA, for incorporation into DNA. Adefovir or bis(POM) PMEA (bis(pivaloyloxymethyl)-9-[2-(phosphonomethoxy)-ethyl]adenine) is actually the prodrug of PMEA with improved oral bioavailability, tolerability and antiviral activity (by virtue of improved intracellular metabolism). Its side effects are predominantly gastrointestinal, such as nausea and diarrhea. Adefovir depletes levels of L-carnitine, necessitating supplementation with oral L-carnitine 500 mg/day. The drug's long intracellular half-life allows for once-daily dosing. Adefovir has a terminal serum half-life of approximately 5 hours. Its Cmax is dose-proportional at the doses tested. It is renally excreted in unchanged form. There appears to be no drug accumulation over time. It is 30% bioavailable in a fasting state and 40% after eating. The dose under study is 120 mg/day. Two placebo-controlled Gilead studies are assessing the clinical impact of adefovir on HIV progression and the development of CMV disease.

Studies of Adefovir against HIV and CMV
Study	Sites	N	Population	Endpoints
Gilead 407/CPCRA	USA	2,160	CD4<100	AIDS, CMV disease
Gilead 410/ADHOC	Europe/Australia	2,000	CD4<100	Death, CMV disease

(Dietz 1998)

Gilead is moving towards filing for FDA approval of adefovir, at least for the anti-HIV indication, sometime in 1998. In the meantime, the drug is available via expanded access by contacting 1.800.GILEAD.5 (Dietz 1998).

Lobucavir (Cyclobut G, BMS 180194, Bristol-Myers Squibb)
Lobucavir is another nucleoside analog - this one based on guanine - with broad-spectrum in vitro activity against CMV and other herpes viruses, HBV and HIV. Against HIV its activity appears confined to macrophages. Its bioavailability is 30-40%. Doses up to 400 mg orally four times daily seem relatively non-toxic and can render the urine of 60% of CMV viruric subjects CMV negative (Lalezari 1997a, Petty 1995). Despite some encouraging findings, lobucavir does not appear to be moving rapidly through the testing pipeline.

GW1263W94 & BDCRB (Glaxo-Wellcome)
Glaxo-Wellcome is in the early stages of developing GW1263W94 - an oral halogenated benzimidazole derivative with a unique mechanism of action and specific anti-CMV activity - as well as a related anti-CMV compound, BDCRB (Zacny 1997). GW1263W9's clinical development for CMV retinitis is in question because it does not seem to cross the blood-eye barrier well in animals. BDCRB does not appear to have as great a magnitude of anti-CMV activity as GW1263W94, but better crosses the blood-eye barrier in animals (Karen Biron, personal communication).

*

ACT UP/New York (AIDS Coalition to Unleash Power). DHPG Trials: A Monument to Regulatory Hypocrisy. Fact sheet distributed at meeting of the National Committee to Review Current Procedures for Approval of New Drugs for Cancer and AIDS (the Lasagna Committee), Bethesda, MD, 1989.
Banker AS, Arevalo JF, Munguia D, et al.: intraocular pressure and aqueous human dynamics in patients with AIDS treated with intravitreal cidofovir (HPMPC) for cytomegalovirus retinitis. Am J Ophthamol 124:227-33, 1997.
Baril L, Jouan M, Caumes E, et al.: The impact of highly active antiretroviral therapy on the incidence of CMV disease in AIDS patients [abstract: I-31]. 37th Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC), Toronto, 1997.
Bell WR; Chulay JD; Feinberg JE: Manifestations resembling thrombotic microangiopathy in patients with advanced human immunodeficiency virus (HIV) disease in a cytomegalovirus prophylaxis trial (ACTG 204).Medicine 76(5):369-80, 1997.
Bowen EF, Wilson P, Cope A, et al.: Cytomegalovirus retinitis in AIDS patients: influence of cytomegaloviral load on response to ganciclovir, time to recurrence and survival. AIDS 10:1515-20, 1996.
Bowen FE, Sabin CA, Wilson P, et al.:Cytomegalovirus (CMV) viremia detected by polymerase chain reaction identifies a group of HIV-positive patients at high risk of CMV disease. AIDS 11:889-93, 1997.
Brosgart CL, Craig C, Hillman D, et al.: Final results from a randomized, placebo-controlled trial of the safety and efficacy of oral ganciclovir for the prophylaxis of CMV retinal and gastrointestinal mucosal disease [abstract: Th.B. 301]. XI Intl Conf AIDS, Vancouver, 1996.
Brown F, Arum I, Francis G, et al.: Ganciclovir prodrug (RS-79070), multiple dose, dose-ranging study with effect of food [abstract: LB19]. 4th Conference on Retroviruses and Opportunistic Infections, Washington DC, 1997.
Buhles WC, Mastre BJ, Tinker AJ, et al.: Ganciclovir treatment of life-or sight-threatening cytomegalovirus infection: experience in 314 immunocompromised patients. Rev Infect Dis 10(S3):S495-S506, 1988.
Cameron B, Heath-Chiozzi M, Kravcik S, et al.: Prolongation of life and prevention of AIDS in advanced HIV immunodeficiency with ritonavir [abstract LB-6a1]. 3rd Conference on Retroviruses and Opportunistic Infections, Washington, D.C., 1996.
Casado JL, Perez-Elias MJ, Moreno A, et al.: Evolution of cytomegalovirus retinitis in AIDS patients after protease inhibitors introduction [abstract], 37th ICAAC, Toronto, Canada, 1997.
Chan ISF, Neaton JD, Saravolatz LD, et al.: Frequencies of opportunistic diseases prior to death among HIV-infected persons. AIDS 9:1145-1151, 1995.
The Chiron Ganciclovir Implant Study Group: A randomized controlled multicenter clinical trial of a sustained-released intraocular ganciclovir implant in AIDS patients with CMV retinitis [abstract: I215]. 35th ICAAC, San Francisco, 1995.
Cochereau-Massin I, LeHoang P, Lautier-Frau M, et al.: Efficacy and tolerance of intravitreal ganciclovir in cytomegalovirus retinitis in acquired immune deficiency syndrome. Ophthalmology 98:1348-55, 1991.
Coskuncan NM, Jabs DA, Dunn JP, et al.: The eye in bone marrow transplantation. VI. Retinal complications. Acrh Ophthalmology, 112:372-9, 1994.
Currier JS, Williams, and the ACTG 320 Team: Preliminary ACTG 320 OI data analysis [presentation]. 23rd AIDS Clinical Trials Group Meeting, Washington, D.C., July 19-23, 1997.
Davis JL, Taskintuna I, Freeman WR, et al.: Iritis and hypotony after treatment with intravenous cidofovir for cytomegalovirus retinitis. Arch Ophthamol 115:733-37, 1997.
Diaz-Llopis M, Espana E, Munoz G, et al.: High dose intravitreal foscarnet in the treatment of cytomegalovirus retinitis in AIDS. Br J Ophthamol 78:120-124, 1994.
Dieterich DT, Kotler DP, Busch DF, et al.: Ganciclovir treatment of cytomegalovirus colitis in AIDS: a randomized, double-blind, placebo-controlled multicenter study. J Infect Dis 167:278-82, 1993.
Dietz P. Adefovir dipivoxil [chapter]. In: Cox S, Dietz P, Harrington M, Smart T: New Drug Dilemmas: Scientific and Scientific and Regulatory Issues in the Era of Antiretroviral Polytherapy and a Viral Load-Driven Standard of Care, version 2.0, Treatment Action Group (TAG), New York, forthcoming 1998.
Dodt KK, Jocobsen PH, Hoffmann B, et al.: Development of cytomegalovirus (CMV) disease may be predicted in HIV-infected patients by CMV polymerase chain reaction and the antigenemia test. AIDS 11:F21-F28, 1997.
Drew WL, Mintz L, Miner RC, et al.: Prevalence of cytomegalovirus infection in homosexual men. J Infect Dis, 143:188-92, 1981.
Drew WL, et al.: Prevalence of resistance in patients receiving ganciclovir for serious cytomegalovirus infection. J Infect Dis 163:716-9, 1991.
Drew WL, Ives D, Lalezari JP, et al.: Oral ganciclovir as maintenance treatment for cytomegalovirus retinitis in patients with AIDS N Engl J Med 333:615-620, 1995.
Drew WL: The state-of-the-art on CMV retinitis [presentation]. AmFAR CMV Conference, 28 January 1996.
Drew WL, Miner RC, Crager M, et al.: Prevalence of ganciclovir-resistant cytomegalovirus during oral ganciclovir prophylaxis [abstract: LB]. 3rd Conference on Retroviruses and Opportunistic Infections, Washington DC, 1996.
Dunn JP, Jabs DA: Cytomegalovirus retinitis in AIDS: natural history, diagnosis, treatment. In: Volberding PA, Jacobson MA eds. AIDS Clinical Review 1995/1996. New York: Marcel Dekker, Inc., 99-129, 1996.
FDA Data: Bravavir(r) (sorivudine) NDA No. 20-585: Advisory Committee Background Document: p. 55 (Pooled analysis of multinational and US NIH trials), p. 64 (risk benefit analysis), 1996 [unpublished].
Feinberg JE, Hurwitz S, Cooper D, et al.: A randomized, double-blind trial of valacyclovir prophylaxis for cytomegalovirus disease in patients with advanced human immunodeficiency virus infection. AIDS Clinical Trials Group Protocol 204/Glaxo Wellcome 123-014 International CMV Prophylaxis Study Group. J Infect Dis 177:48-56, 1998.
Finkelstein DM, van der Horst C, Bozzette S, et al.: CMV shedding and the risk of end-organ disease [abstract 505]. 35th IDSA, San Francisco, 1997.
Friedberg DN: Hypotony and visual loss with intravenous cidofovir treatment of cytomegalovirus retinitis [letter]. Arch Ophthamol 115:801-02, 1997.
Fiala M, Payne JE, Berne TV, et al.: Epidemiology of cytomegalovirus infection after transplantation and immunosuppression. J Infect Dis 132:421-33, 1975.
Fitzgerald E, Goldman HM, Miller WG, et al.: A penile ulceration in a patient with the acquired immunodeficiency syndrome. Foscarnet-induced penile ulceration. Arch Dermatol 131:1449, 1452, 1995.
Freeman WR: New developments in the treatment of CMV retinitis [Editorial]. Ophthalmology 103(7):999-1000 & 1078-83, 1996.
Gallant JE, Moore RD, Richman DD, et al.: Incidence and natural history of cytomegalovirus disease in patients with advanced human immunodeficiency virus disease treated zidovudine. J Infect Dis, 166:1223-7, 1992.
Griffiths PD, Feinberg J, et al.: Detection of cytomegalovirus by PCR in baseline samples from patients entering a CMV prophylaxis study. Abstract: H131, 35th ICAAC, San Francisco, 1995.
Griffiths PD, Feinberg J, et al.: Detection of cytomegalovirus in samples from patients enrolled in ACTG 204/Glaxo Wellcome 123-014 [abstract]. 3rd Conference on Retroviruses and Opportunistic Infections, Washington DC, 1996.
Heinemann MH: Long-term intravitreal ganciclovir therapy for cytomegalovirus retinotherapy. Arch Ophthamol 107:1767-72, 1989.
Holbrook JT, Davis M, Gilpin AK, et al.: Association of cytomegalovirus (CMV) retinitis characteristics with disease progression and visual loss [abstract: Tu.B. 2190]. XI Intl Conf AIDS, Vancouver, 1996.
Holland GN, Tufail A: New therapies for cytomegalovirus retinitis [Editorial]. N Engl J Med 333:658-9, 1995.
Hoover DR, Saah AJ, Bacellar H, et al: Clinical manifestations in the era of pneumocystis prophylaxis. N Engl J Med, 329:1992-6, 1993.
Hutcherson SL, Palestine AG, Cantrill HL, et al.: Antisense oglionucleotide safety and efficacy for CMV retinitis in AIDS patients [abstract: H143]. 35th ICAAC, San Francisco, 1995.
Jabs DA, Dunn JP, Enger C, et al.: Cytomegalovirus retinitis and viral resistance. Prevalence of resistance at diagnosis, 1994. Cytomegalovirus Retinitis and Viral Resistance Study Group. Arch Ophthamol 114:809-14, 1996.
Jabs DS: Treatment of cytomegalovirus retinitis in patients with AIDS [Editorial]. Ann Intern Med 125:144-5, 1996.
Jacobson MA, O'Donnell JJ, Brodie HR, et al.: Randomized, prospective trial of ganciclovir maintenance therapy for cytomegalovirus retinitis. J Med Virol 25:339-49, 1988.
Jacobson MA, van der Horst C, Causey DM, et al.: In vivo additive antiretroviral effect of combined zidovudine and foscarnet therapy for human immunodeficiency virus infection (ACTG Protocol 053). J Infect Dis 163:1219-22, 1991.
Jacobson MA, Causey D. Polsky B, et al.: A dose-ranging study of daily maintenance intravenous foscarnet therapy for cytomegalovirus retinitis in AIDS. J Infect Dis 168:444-8, 1993.
Jacobson MA: CMV prophylaxis [presentation]. 20th AIDS Clinical Trials Group Meeting, Washington, D.C., 1995.
Jacobson MA: Advances in the prevention and management of CMV disease [abstract: S029]. 36th ICAAC, New Orleans, 1996.
Jacobson MA: The state-of-the-art on CMV retinitis [presentation]. AmFAR CMV Conference, 28 January 1996.
Jacobson MA, Drew WL, Dunn JP, et al.: CMV culture results and clinical outcome in AIDS patients with CMV retinitis treated with either foscarnet or ganciclovir [abstract: Tu.B. 2189]. XI Intl Conf AIDS, Vancouver, 1996.
Jacobson MA, Zegans M, Pavan PR, et al.: Cytomegalovirus retinitis after initiation of highly active antiretroviral therapy. Lancet 349(9036):1443-5, 1997.
Jordan MC: Latent infection and the elusive cytomegalovirus. Ann Intern Med, 103;738-43, 1985.
Jung D, Dorr A: First in humans study: single dose pharmacokinetics and absolute and relative oral bioavailability of RS-79070-194 in HIV and CMV seropositive patients tolerance study of 28 day regimens or oral ganciclovir. Palo Alto (CA): Syntex Study CAN 2661, Final Report, 15 December 1995.
Katlama C, Dohin E, Caumes E, et al.: Foscarnet induction therapy for cytomegalovirus retinitis in AIDS: comparison of twice-daily and three-times-daily regimens. JAIDS 5:S18-S24, 1992.
Katznelson S, Drew WL, Mintz L: Efficacy of the condom as a barrier to the transmission of cytomegalovirus. J Infect Dis, 150:155-7, 1984.
Kirsch LS, Arevalo JF, Chavez de la Paz E, et al.: Intravitreal cidofovir (HPMPC) treatment of cytomegalovirus retinitis in patients with acquired immune deficiency syndrome. Ophthalmology 102:533-43, 1995.
Kirsch LS, Arevalo JF, De Clercq E, et al.: Phase I/II study of intravitreal cidofovir for the treatment of cytomegalovirus retinitis in patients the with acquired immunodeficiency syndrome. Am J Ophthalmology 119:466-76, 1995.
Kuppermann BD, Petty JG, Richman DD, et al.: Correlation between CD4+ counts and prevalence of cytomegalovirus retinitis and human immunodeficiency virus-related noninfectious retinal vasculopathy in patients with acquired immunodeficiency syndrome. Am J Ophthamol 115:575-582, 1993.
Lalezari JP, Drew WL, Glutzer E, et al.: (S)-1-[3 hydroxy-2-(phosphonylmethoxy)propyl]cytosine(cidofovir): results of a phase I/II study of a novel antiviral nucleotide analogue. J Infect Dis 171:788-96, 1995.
Lalezari J, Holland G, Stagg R, et al.: A randomized, controlled study of cidofovir (CDV) for relapsing cytomegalovirus retinitis (CMV-R) in patients with AIDS [abstract: LB-9]. 35th ICAAC, San Francisco, 1995.
Lalezari J: CMV disease - results from studies with oral ganciclovir [abstract]. CMV disease: optimizing prevention and treatment (Hoffmann-La Roche Symposium), Birmingham, England, 1996.
Lalezari J, Friedberg D, Bisset J, et al.: A comparison of the safety and efficacy of 3g, 4.5g, and 6g doses of oral ganciclovir versus IV ganciclovir for the maintenance treatment of CMV retinitis [abstract: TH.B. 305]. XI Intl Conf AIDS, Vancouver, 1996.
Lalezari JP, Kemper C, Stagg R, et al.: A randomized, controlled study of the safety and efficacy of intravenous cidofovir (CDV, HPMPC) for the treatment of relapsing cytomegalovirus retinitis in patients with AIDS [abstract: Th.B.304]. XI Intl Conf AIDS, Vancouver, 1996.
Lalezari JP, Drew WL, Jordan C, et al.: In vivo anti-CMV activity and safety of oral lobucavir in HIV-infected patients [abstract 301]. Fourth Conference on Retroviruses and Opportunistic Infections, Washington, D.C., 1997.
Lalezari JP, Stagg RJ, Kupperman BD, et al.: Intravenous cidofovir for peripheral cytomegalovirus retinitis in patients with AIDS. Ann Intern Med 126:257-63, 1997.
Mallolas J, Arrizabalaga J, Lonca M, et al.: Cytomegalovirus disease in HIV-infected patients treated with protease inhibitor [abstract 295]. 35th IDSA, San Francisco, 1997.
Martin DF, Park DJ, Mellow SD, et al.: Treatment of cytomegalovirus retinitis with an intraocular sustained-release ganciclovir implant. Arch Ophthamol 112:1531-1539, 1994.
Martin DF, Ferris FL, Parks DJ, et al.: Ganciclovir implant exchange. Timing, surgical procedure, and complications. Acrh Ophthamol 115:1389-94, 1997.
Martin D, Kuppermann B, Wolitz R, et al.: Combined Oral Ganciclovir (GCV) and intravitreal GCV implant for treatment of patients with cytomegalovirus (CMV) retinitis: a randomized, controlled study [abstract: LB-9]. 37th ICAAC, Toronto, 1997.
Marx JL, Kapusta MA, Patel SS, et al.: Use of the ganciclovir implant in the treatment of recurrent cytomegalovirus retinitis. Arch Ophthamol 114:815-20, 1996.
Matheron, S, Knospe V, Barton S, et al.: 6g vs. 3g doses of oral ganciclovir in maintenance treatment of CMV retinitis [Abstract: Tu.B.2184]. XI Intl Conf AIDS, Vancouver, 1996.
Merigan T, Resta S: Cytomegalovirus: where have we been and where are we going? Reviews Infect Dis, 12(S7):S693-S700, 1990.
Michaels S, Clark R, Kissenger P: Difference in the incidence rates of opportunistic infections before and after the availability of protease inhibitors [abstract]. 5th Conference on Retroviruses and Opportunistic Infections, Chicago, 1998.
Mintz L, Drew WL, Miner RC, et al.: Cytomegalovirus infection in homosexual men. An epidemiological study. Ann Intern Med, 99:326-9, 1983.
Moyle G, Barton S, Gazzard BG: Penile ulceration with foscarnet therapy. AIDS 7:140-1. 1993.
Mu