It was a frenzied set of hearings at the Food and Drug Administration February 28 to March 1. When the dust had cleared, the agency's Antiviral Drug Advisory Committee had moved toward dethroning both AZT as the centerpiece of first line anti-HIV therapy and CD4 cell counts as the main lab standard for judging drug effectiveness. The advisory committee's concrete moves included recommending the approval of two older nucleoside analogs (ddI and ddC -- the latter only in combination with AZT) and two new protease inhibitors (Merck's and Abbott's) for any stage of HIV infection. The panel, composed of outside experts, largely based its vote on symptom relief and viral load reductions during clinical trials.
The advisory committee worked on the assumption that there is a correlation between these two, that HIV level and physical health are closely linked. It ran into difficulties when that logical assumption was contradicted by the facts. Similarly, the committee assumed that combination antiviral therapy was better than single agent therapy and reacted with considerable consternation when it found that this is not always so.
Meanwhile, any assumption that all barriers to rapid drug approval were down in the agency was countered on March 1, when the metabolic and endocrinology committee voted to reject the Serono Laboratories application for recombinant human growth hormone as a therapy for AIDS-related wasting (see article Growth Hormone Cut Down at the Pass; Mar. 1996). FDA reviewers' disparaging dissection of the company's trials derailed the approval process and underscored that the FDA considers anti-HIV drugs a special, urgent case.
The day after the Abbott protease inhibitor ritonavir (brand name: Norvir) received its approval recommendation, FDA officials announced that they had officially approved the drug. Two business days later, Abbott shipped its first lots, with a one year supply for an individual costing wholesale distributors $6,500 (leading to a retail price of about $8,000). Official approval for Merck's indinavir (brand name: Crixivan) came two weeks later, as that company prepared its distribution system (see article Getting Your Space on the Magic Crixi-van; Mar. 1996). Merck is "only" charging $4,380 wholesale for a year's supply.
Viral load testing, central to optimizing and occasionally adjusting treatment regimens in each person, now seems to be on the FDA fast track, too. The FDA's Blood Products Advisory Committee took up the Roche Amplicor PCR test for viral load on March 21. Roche was originally asking for only a diagnostic approval for this assay (to detect simply whether someone is HIV-positive or -negative). With the FDA's encouragement, the company expanded its request to include approval for monitoring plasma HIV levels for the purposes of predicting disease progression and measuring treatment benefit. Chiron filed for approval of its bDNA test this month, but at press time the advisory committee has yet to schedule a hearing on the application.
For a late report on the unexpectedly muddled Roche PCR hearing, which was held after Treatment Issues went to press, see article FDA Panel Takes Up Viral Load; Mar. 1996.
As this month's Treatment Issues was preparing to go to press, the FDA announced that it had broadly approved Merck's protease inhibitor indinavir (brand name: Crixivan) "for the treatment of HIV infection in adults when antiretroviral therapy is warranted." Anticipating its product launch on March 25, Merck had established a number of procedures to assure a smooth transition when indinavir hits the market. The company's number one problem is that it planned to receive approval next fall, when its production facilities will be completely on line. Because the FDA encouraged Merck to hurry up, the supply of the drug will be somewhat limited until late this year. Just how limited, Merck will not say, terming this figure "proprietary information." (The unofficial word is that there is enough indinavir for 40,000 patients.)
To monitor the sales of indinavir and ensure that an individual will continue to get the drug once he or she starts taking it, the entire retail supply of indinavir will be placed in the hands of Stadtlander's Pharmacy, a mail-order concern. Doctors will have to send or fax a special prescription form to Stadtlander's to obtain indinavir for their patients. Stadtlander's will ship the indinavir after the patient's reimbursement coverage is verified, which should take no longer than 48 hours. (Stadtlander's will ship indinavir COD, too.)
In states where Stadtlander's does not operate or public reimbursement programs have no payment agreement with Stadtlander's, the firm will have to ship the indinavir to an intermediary drug supply company. Further, the Veterans Affairs medical system and perhaps large local hospitals will have their own direct connection to Merck.
To order Crixivan (doctors only) or for further information on the prescription procedure, call the Merck/Stadtlander Crixivan hotline at 800/927-8888. Some activists already are complaining that Merck's program will prove awkward and unworkable, especially for those under the care of public health clinics. Merck has promised to try to adjust the system if there are problems. (Treatment Issues would like to hear about people's experiences obtaining indinavir. Write to us at 129 W. 20th St. New York 10011, fax us at 212/337-3656 or e-mail DAVE_G@gmhc.org).
Before indinavir was approved, Merck would not comment on the price it planned to charge, again citing "proprietary information." There were considerable fears about how much Merck might ask, but the company has now announced that the factory wholesale price for indinavir will be equivalent to $4,380 for a year's supply, making indinavir the cheapest protease inhibitor available.
Merck is offering a financial assistance program for those who are having trouble raising that kind of money. Those eligible for the program must be uninsured and unable to qualify for any publicly financed program. Applicants will also have to pass a financial screening test given over the phone. Again, Merck will not reveal the formula it is using to determine financial need because this is "proprietary information." Alternative ways of qualifying for the program include reaching the annual drug expense cap in your insurance coverage or exhausting your available funds. Financial counselors who will help patients or their representatives obtain third-party payment or Merck's financial assistance also can be reached through the Crixivan hotline (800/927-8888).
The financial assistance program will also provide temporary aid to people whose reimburser is still considering whether or not to include indinavir in its formulary (list of covered drugs). If the third-party payer eventually decides to exclude indinavir, those of its members already benefiting from Merck's assistance will continue to receive free indinavir. But note: Patients whose reimbursers have formally refused to cover indinavir are ineligible for financial assistance (Merck fears if it offered such aid, then every program in the country would reject indinavir, passing the bill back to the company). This quirk is another incentive to start on indinavir as quickly as possible.
Indinavir has an attractive combination of relative potency and safety, and its price is relatively low. Demand is sure to be great. Treatment Issues advises its readers not to tarry if they want the drug any time soon.
It was clear during the advisory committee's second day that ritonavir had a demonstrated ability to extend survival and delay new opportunistic infections in people with advanced disease, at least for six months. Lack of long-term safety or efficacy data, surprisingly meager drug interaction test data (ritonavir's package insert contains a long list of untested "potential" interactions) and the sparseness of the data in early disease strongly influenced the committee's recommendation. While accepting that ritonavir was eligible for full approval for people with advanced HIV infection, the committee recommended that the FDA grant only accelerated approval for use of ritonavir in patients with early HIV infection. It sought further studies to evaluate this patient population.
The case for ritonavir in early disease was not aided by the new and perplexing surrogate marker data in less advanced patients, which indicated that although ritonavir was better than AZT, the combination of ritonavir and AZT was no better than ritonavir monotherapy (see table: New Data from Protease Inhibitor Trials; Mar. 1996). Worse, ritonavir alone suppressed viral load more than the combination arm, and this difference was statistically significant.
Why might ritonavir monotherapy outperform combination therapy? Abbott believes that this was due to lack of compliance, (the study used the old nauseating liquid formulation of ritonavir, now succeeded by capsules) and that adding AZT -- which frequently also causes nausea -- made the drug even harder to tolerate. Plasma levels of ritonavir were 50 to 70 percent lower in the combination arm than in the monotherapy arm, presumably because the people already nauseated by AZT recoiled more from taking the foul elixir. Treatment discontinuation due to adverse events was extremely high in the combination arm, 38.8 percent.
The underlying problem with ritonavir is that its toxicities limit the allowable dose to 600 mg twice a day. This is probably not the optimum dose as far as anti-HIV effect is concerned -- 600 mg suppresses HIV more than 400 mg, and 800 mg could well be more active still, if it could be tolerated. Beside the mostly gastrointestinal side effects listed above, ritonavir is a very potent inhibitor of the liver enzymes that break down many other medications. Ritonavir's possible interactions with the other drugs worried the committee: the magnitude of many of the predicted drug interactions has not been measured in people, and their potential danger should not be underestimated. The committee also expressed concern about increases in liver enzymes and lipids (triglycerides) in the blood since such increases are often predictive of long-term damage to the liver and pancreas.
There are almost no data available on the safety of ritonavir past six months. Nor are there much data about durability of effect beyond this time period. The consensus of the committee was that the goal of therapy now is to safely suppress HIV, by as much and as long as possible, with either single or multiple agents. Higher and possibly more potent doses of ritonavir were too toxic to use in people. If the "tolerable" ritonavir dose turns out to be too low to support lasting antiviral activity, the emergence of ritonavir-resistant HIV could become a pitfall for the drug, particularly in those who begin treatment early.
Representatives from the FDA, clearly poised to approve ritonavir immediately, worked late into the night with Abbott officials to hash out an accord. The next day, when announcing the formal FDA approval of ritonavir (which conformed with the advisory committee's mixed accelerated/full approval vote), David Feigal, head of the FDA's antiviral division, read a letter from Abbott vice-president Andre Pernet making the following commitments:
Actually, the company had already mentioned plans for the second and third points during its presentation to the advisory board the previous day. During that session, the Pediatric AIDS Foundation bitterly denounced Abbott for its neglect in investigating ritonavir for children (with the one small children's trial having enrolled no one under two). As for ritonavir/saquinavir, an initial two-week study in HIV-negative volunteers already is over. The study found that taking ritonavir along with saquinavir increased the body's exposure to saquinavir twenty-fold (in terms of "area under the curve" (AUC), a measure of how well drug levels persist over time). Dr. Feigal of the FDA also noted that the company had expressed a "strong interest" in a ritonavir/ddI/d4T study the day before, but this was not included in the letter.
The advisory committee had less trouble with indinavir on March 1, and smoothly voted to recommend granting it broad accelerated approval for people with HIV infection. Although the application included no results on delay of actual disease progression, Merck provided impressive safety and surrogate marker data for indinavir alone or in combination with nucleoside analogs in both early and late stage disease. Merck's trials included both AZT-experienced and -naive participants, and the studies have had a longer observation period than ritonavir's. Over 2,000 people have taken indinavir, 600 have been on indinavir for more than six months and 250 for more than one year, with some on treatment for as long as two years.
Most of the indinavir studies have been reported previously in Treatment Issues (see February 1996, pages 2, 4-5, and October 1995, pages 5-6). New preliminary surrogate marker data were presented from two studies testing indinavir versus AZT versus the combination -- one a clinical endpoint study in Brazil in AZT-naive people with CD4 cell counts between 50-250, and the other a large surrogate marker study in AZT-naive people with 50 to 500 CD4 cells in the U.S. (see table: New Data from Protease Inihbitor Trials; Mar. 1996).
Neither study can yet confirm that indinavir plus AZT is superior to indinavir monotherapy. Participants in these studies started out at low HIV levels, and thus, the viral load assay's limit of detectability (500 viral particles per milliliter of plasma) served as a barrier to comparing the profound reductions in viral load experienced on single agent and combination therapy. By week 24 in both trials combined, viral load had gone down to below detectable levels in 39 percent of the participants on indinavir alone and in 42 percent of the people on indinavir/AZT.
Resistance to monotherapy also may take longer to occur in people with lower viral loads. Other studies have shown that once resistance to indinavir does occur, viral load rebounds. In further follow-up of the indinavir/3TC/AZT study (see Feb. 1996's Treatment Issues), the median viral load in 20 people on indinavir monotherapy for 24 weeks was returning to baseline (the peak viral load decrease of 1.6 log below baseline, attained at week twelve, slipped to 0.7 log below baseline at week 24). Meanwhile, the nineteen patients on combination therapy to reach week 24 had sustained viral load reductions of more than 2 log (99 percent) below baseline.
Ironically, even though Merck has said more about HIV protease cross-resistance to various inhibitors than any other company, neither the FDA nor Merck brought up the issue during the indinavir hearing. Previously, in hearings on ritonavir and Roche's saquinavir, cross-resistance was considered a serious subject for discussion.
Compared to the Antiviral Drugs Advisory Committee's deliberations on the two protease inhibitors, its discussion February 28, the first day of hearings on new approvals for ddI and ddC might seem anachronistic. The advent of the protease inhibitors taken up in the following two days makes it less likely that ddI or ddC will continue to be used alone or in combination with just AZT. Consideration of ddI and ddC also took place with no account taken of the relative merits of 3TC or the AZT/3TC combination, nor of d4T/ddI or other newer and possibly more potent nucleoside analog regimens.
Yet, the committee deliberations established a pattern of using viral load reductions when evaluating therapeutic regimens. And the committee's dismay at not being able to simply follow the logic of viral load (the deeper the drop, the better) are symptomatic of the problems yet to be resolved.
The day started with a review of various recently completed large trials comparing the effect of different nucleoside regimens on progression of disease symptoms and death (clinical endpoints) as well as viral load and CD4 cell count. All the adult studies, ACTG 175, Delta and CPCRA 007 (NuCombo), have been recapitulated before in these pages (see Treatment Issues, October 1995, pages 2-4 and February, 1996, page 6). The newly analyzed results from ACTG 152, a pediatric trial, are described in this issue (see article Pediaric Treatment Update; Mar. 1996). These trials mostly compared AZT to the AZT/ddI and AZT/ddC combinations. ACTG 175 included a ddI monotherapy arm, as did ACTG 152, which did not test AZT/ddC. Except for ACTG 152 and ACTG 175, trial participants were people with relatively advanced disease.
FDA statistician Paul Flyer, Ph.D., summarized the four trials' results in a statistical analysis of the pooled clinical endpoint data. In the four trials as a whole, AZT/ddC was better than AZT alone only in patients without previous therapy. ddI performed better than AZT while AZT/ddI once again did no better than monotherapy, in this case with ddI rather than protease inhibitor. How can two drugs be no better than one, especially when the effect of the two together on viral load is considerably stronger?
The committee could only conclude this time that ddI was potent enough to cover up the modest antiviral contribution of AZT in the AZT/ddI combination. So, if you are taking ddI, why take AZT too? On the other hand, if you cannot tolerate ddI, you can enhance AZT's effect by including ddC. AZT/ddC presumably does not distinguish itself in AZT-experienced people because their virus has developed resistance to AZT and it is up to the relatively weak ddC to accomplish most of the HIV suppression on its own.
After much agonizing, the committee unanimously recommended approving ddI monotherapy as an initial anti-HIV treatment (it now is approved only to succeed AZT) but could not explicitly endorse AZT/ddI. The vote on AZT/ddI was accompanied by an informal sense of the committee that the preponderance of evidence supports approval of this combination since it seems more effective than AZT or AZT/ddC, a sentiment reflecting the committee's inability to find consistency in the trial data it was reviewing.
The advisory panel voted in favor of full approval of AZT/ddC but only in people with no prior therapy. This recommendation was especially confusing because the combination now has accelerated, or conditional, approval for patients with CD4 cell counts from 150 to 300 whether or not they have had prior AZT. (Below 150, approval is limited to people without past AZT exposure.)
Treatment with ddI will now become more common if the FDA follows its advisors' thinking. Another factor that will increase ddI's popularity is the new, smaller, orange-flavored chewable tablet approved by the FDA on February 27. This tablet is easier to ingest than the older "hockey puck," allowing patients to follow ddI's dosing schedule more rigorously.
The FDA imperative around HIV apparently stems from a growing conviction that a four-drug combination will be highly effective in suppressing the virus. Researchers, like David Ho, are arguing that three drugs will provide long-term suppression of HIV infection but a fourth drug is needed for extra insurance against breakthrough infections. This additional agent, could be a new nonnucleoside reverse transcriptase inhibitor like delavirdine or nevirapine (both soon to be considered by the FDA) or another protease inhibitor. Besides the medical considerations, the FDA is also under considerable political pressure due to the FDA reform movement in Congress. To stave off dilution of its authority, the agency needs to demonstrate that it can act quickly when the situation calls for it.
The more potent combinations now becoming available will no doubt stand out from the single and double nucleoside analog combinations currently available. Precisely how much better the new therapies perform in complex real-life conditions remains to be seen. A treatment's success in individual patients depends on their ability to comply with treatment, the side effects they experience, how their bodies metabolize the compounds, and the susceptibilities of their personal HIV strain to the drugs.
By urging the release of the HIV treatments under consideration with the broadest possible indications, the committee has allowed doctors and individual patients to work out the best therapies on a case by case basis. Viral load monitoring may help make these decisions. One can only hope that using viral load assays will prove more straightforward with the new combinations than it did on February 28 with the traditional antiviral regimens.
The FDA was in classic form on March 1, when Serono Laboratories' version of human growth hormone (HGH) came up for review as a treatment for AIDS-related wasting (see Treatment Issues, May, 1995). The hearing took place at a joint meeting of the endocrinologic and the antiviral drugs advisory committees. By the time the session was over, Serono's case had been demolished point by point by agency reviewers.
Growth hormone is one of several candidate therapies for AIDS wasting syndrome, defined as an involuntary weight loss of at least ten percent of normal body weight. At $1,000 per week, though, it is definitely the most expensive. During the hearing, Serono's case turned out to be built around two twelve-week trials (referred to as trial 5341 and trial 7033) with about 180 participants each.
By the end of twelve weeks, total body weight in both trials' growth hormone arms was up about 1.6 kg, but weight in the placebo arms was up, too -- 0.1 kg in the 5341 study and 0.4 kg in the 7033 study. The differences in weight gain in this second study were not statistically significant, although in the first study they were. Serono was able to supplement the weight gain data in the first study with evidence from that study showing significant increases in the proportion of lean body mass and losses in body fat among those on growth hormone.
Unfortunately, no body composition analyses were done in trial 7033, so Serono could not counter the impression that growth hormone made only a marginal contribution to weight gain -- a gain that could be nearly equaled through the nutritional assessment and counseling the placebo group received along with the HGH group. The benefits of HGH seemed to lie mainly in the first week or two of therapy, according to the FDA reviewer, Saul Malozowksi, M.D., Ph.D. Moreover, his reanalysis (using an "intent-to-treat" approach that included dropouts' experience) showed only a one kilogram weight gain for HGH.
The advisory committee, seriously put off by the lack of confirmation of HGH's benefits by study 7033, was nonetheless impressed by the testimony from patients and doctors of dramatic improvements in weight, strength and energy while on growth hormone. Much time was spent trying to define a subset of people with AIDS who would be the candidate population for "best responders" to HGH.
This salvage effort, too, was nixed by Dr. Malozowksi and the FDA, who produced a slide indicating that weight changes in both the HGH and placebo arms of the trials basically represented a conventional bell curve, with the HGH curve shifted only slightly in a more positive direction than the placebo one. Tellingly, both the greatest weight gainer (+14 kg over the twelve weeks) and the greatest loser (-12 kg over the same period) were receiving growth hormone.
In the end, the advisory committee voted against Serono's application seven to eight. Most of the eight opposing approval supported a suggestion from the nonvoting community representative on the panel, Bill Thorne of ACT UP/Golden Gate in San Francisco: that some kind of accelerated or conditional approval be granted while Serono conducts a third, confirmatory trial of growth hormone.
After the hearing, Mr. Thorne expressed his impatience with the company: "Serono seems completely inept. The data on responders were obscured by Serono's poor trial design and poor analytical technique. It would be dangerous to grant full approval as it is."
In contrast, Donald Kotler, M.D., of Columbia University, who is a world authority on AIDS wasting and who appeared before the advisory committee as a Serono consultant, blamed the FDA: "I was real disappointed. The approval process was based on inaccuracies. Weight gain as the endpoint is wrong because the result is the product of two contradictory factors [loss of fat and gain in lean tissue]. The FDA said it had to be weight, but that's wrong How can they have trashed HGH, when it is obviously an anabolic agent?"
But it is Serono's own fault that it failed do body composition analysis in its second study. "Until 5341 was unblinded, which was after 7033 started, there wasn't as much appreciation of lean body mass," said Robert Kauffman, M.D., who is medical director of metabolic and immune therapy at Serono. "Lean body mass is the best way to look at HGH," he now concludes, with the benefit of hindsight.
Lean body mass results were not the only missing piece of information. Questions were raised during the hearing as to whether the proper dosage and length of administration had been adequately determined. Negotiations are now proceeding between Serono and the FDA over the conditions under which growth hormone might yet be approved. Certainly, some further data will be required to resolve all the uncertainties swirling around the therapy.
ACTG 152: AZT vs. ddI and AZT/ddI
ACTG 152 is a randomized, double blind trial comparing AZT monotherapy and ddI monotherapy with AZT plus ddI in HIV-infected children ages three months up to eighteen years. After an interim analysis in February, 1995, the Data Safety and Monitoring Board (DSMB) recommended closing the AZT monotherapy arm of the study. This recommendation was based on data showing that patients on the AZT monotherapy arm had a significantly increased risk of reaching a primary study endpoint compared with the ddI monotherapy or combination therapy arms. Primary endpoints for the study include survival and HIV disease progression based on growth failure, deterioration in neuropsychological status or the development of more than two new or recurrent opportunistic infections. The interim analysis also revealed an increase in toxicity (low red and white blood cell counts and abnormal blood chemistry) in patients in the AZT
Physicians caring for patients enrolled in the AZT monotherapy arm were advised to offer these patients what seemed the "best available" therapy. Children in the other two arms of the trial were followed through the completion of the study in August 1995.
A full analysis of the completed trial with an unblinding of the other two arms was released at the February 28 FDA advisory committee hearing that reviewed use of AZT, ddI and ddC alone and in combination. The new analysis, completed just days before the presentation, revealed that treatments with either ddI alone or a combination of ddI plus AZT were equally efficacious.
ACTG 152 opened in August 1991 and over a period of two years enrolled 839 children with a mean age of three years nine months at entry. The trial was meant to assess the comparative efficacy of the three antiretroviral regimens as first line therapy for HIV, so children who had been on antiretroviral therapy for more than six weeks were excluded. Just eight percent of patients had received any antiretroviral therapy prior to enrollment.
The trial was otherwise unusually broad in its inclusion criteria. Only asymptomatic children with relatively normal immunological laboratory test values were excluded. The median CD4 counts at study entry for trial participants younger than 30 months and older than 30 months were 1,191 cells/mm3 and 569 cells/mm3 respectively. (Note that children's CD4 counts normally start out higher than adults. The ACTG 152 medians are in the low-to-moderate immune suppression range.)
Interim efficacy analysis of the time to first opportunistic infection or death revealed that the AZT monotherapy arm performed significantly worse than the "best" treatment arm at the time (not further identified to keep the trial blinded). But when survival was evaluated separately, no statistically significant difference was noted.
With the AZT monotherapy arm eliminated, this February's final analysis of the study assessed whether there was a difference between the ddI monotherapy arm and the ddI plus AZT combination therapy arm. This analysis revealed that there was no distinction between the two arms with regard to disease progression (including neuropsychological deterioration) or survival.
Primary disease progression endpoints occurred in 24 percent of children receiving ddI alone and 25 percent of children receiving combination therapy. Of the primary endpoints, 52 percent were growth failure, nineteen percent were neuropsychological deterioration and twelve percent were deaths. (AZT has the reputation of penetrating the central nervous system better than ddI, but even though the incidence of neuropsychological dysfunction was lower in the children receiving ddI/AZT than those on ddI monotherapy, this gap did not reach statistical significance.) The primary endpoint figures are dominated by the children under 30 months old. Far fewer instances of disease progression and death occurred in the older children, and the trial could not distinguish between the performance of the two treatments in the older group of children.
Children receiving ddI monotherapy had significantly fewer hematologic toxicities compared to the other two arms. The increased toxicity with AZT occurred in both those receiving the lower dose (120 mg/m2 of body surface area/dose) on the combination arm and those on the higher dose monotherapy arm (l80 mg/m2). Results of ACTG 128, which directly compared 180 mg/m2 and 90 mg/m2 of AZT revealed that higher and lower AZT doses had similar toxicity. From a clinical standpoint, the two dosages were also equivalent in ACTG 128, with the exception that those children with neurological disease fared better on the higher dose.
After the interim analysis one year ago, many pediatric AIDS specialists continued immunologically and clinically stable children on AZT monotherapy. The recommendations from the National Institute of Allergy and Infectious Diseases (sponsor of ACTG 152) were extremely vague and gave the options of switching or continuing AZT monotherapy. Over time, and with the release of supporting data from ACTG 175 and the European Delta Study (which indicated an advantage to combination therapy and/or ddI monotherapy over AZT monotherapy in adults), physicians started leaning toward using combination antiretrovirals as first line therapy for HIV-infected children.
The final results of ACTG 152 clearly support the results of ACTG 175 (see Treatment Issues, October 1995, pages 2-4) and indicate unequivocally that AZT monotherapy is not the most effective first line therapy. The finding that ddI monotherapy is superior to AZT monotherapy has surprised most clinicians caring for HIV-infected children.
Pneumocystis carinii pneumonia (PCP) is the most common serious opportunistic infection occurring among children with HIV. Most cases occur early in the first year of life with a peak incidence at two to six months of age. Guidelines on PCP prophylaxis for infants and children with HIV were published in the CDC's Morbidity and Mortality Weekly Report (MMWR) in 1991 and followed recommendations developed by an expert panel convened by the National Pediatric and Family HIV Resource Center. Those guidelines were based on the premise that HIV-infected infants and children are at risk for PCP, and therefore should receive prophylaxis, when their CD4 cell count falls below certain age-related thresholds: below 1,500 cells/mm3 for children one to eleven months of age, below 750 cells/mm3 for children twelve to 24 months, less than 500 cells/mm3 for children two to five years, and less than 200 cells/mm3 for children over six years old.
During the past three to four years, it has become evident that in the first six to twelve months of life, CD4 cell count is not as accurate a measure of the risk for PCP as it is in other age groups. Although an infant with a CD4 cell count below 1,500 cells/mm3 in the first year of life is clearly at risk for PCP, there is also a risk for HIV-infected infants with higher CD4 counts. A recent CDC survey of PCP in HIV-infected infants revealed that 22 of 86 HIV-infected infants less than six months of age with PCP had had CD4 counts greater that 1,500 cells/mm3 within two months of the diagnosis.
Subsequent to the publication of the 1991 guidelines, on-going surveillance of AIDS cases by the CDC demonstrated no significant decline in PCP among HIV-infected children. In addition, a CDC survey found that 59 percent of children with PCP who had not received prophylaxis had not been identified as being HIV-exposed soon enough to begin such prophylaxis.
In March, 1994, the National Pediatric and Family HIV Resource Center in collaboration with the CDC convened an expert panel to evaluate the 1991 PCP prophylaxis guidelines in light of this and other information. The consensus of the expert panel was that changes to the 1991 guidelines were appropriate. Revised guidelines were published in the MMWR on April 28, 1995 and are summarized in the table.
The most significant change involves the use of prophylaxis in the first year of life. The new guidelines recommend that all HIV-infected infants start prophylaxis at one month of age regardless of CD4 cell count. Since it is difficult to determine with certainty by one month of age which infants born to HIV-infected women are themselves infected, the new guidelines recommend initiating prophylaxis in all one-month old infants born to HIV-infected mothers and continuing the prophylaxis until HIV infection can be reasonably excluded. Attempts should be made to exclude HIV infection as soon as possible so that uninfected infants can be taken off medication expeditiously to thereby avoid potential side effects.
The guidelines call for the use of HIV culture and PCR in order to diagnose and reasonably exclude HIV infection. It was the consensus of the expert panel that HIV infection can be reasonably excluded with two negative HIV cultures or PCR viral load tests, one obtained at or after one month of age and the other at or after four months. Clinicians in research settings may feel comfortable excluding HIV infection prior to four months of age through serial HIV culture and/or PCR assays.
There are difficulties, though. HIV culture is not universally available outside of major medical centers with virology laboratories, and even then the test is not reimbursable by third party payers including Medicaid. PCR, although available commercially, is not uniformly reimbursable, either, and therefore difficult to use in a practical sense.
Without the use of HIV culture or PCR, it is impossible to rule out HIV infection until at least seven or eight months of age. Although HIV infection can be excluded on the basis of two negative HIV serologies, (the standard HIV antibody blood test) obtained after six months of age, many infants still have significant maternal antibodies present after six months. It is important at this time for third party payers, including the United States government, to reevaluate their reluctance to cover these important diagnostic assays. The New York State Department of Health, in response to the new guidelines, now offers free PCR testing for diagnostic purposes.
Although the substantive changes in the guidelines will be effective, one hopes, in preventing PCP in a larger percentage of infants and children, infants will continue to get PCP if they are not identified as being at risk for HIV infection. It has become increasingly obvious that knowledge of maternal HIV infection has an impact on the newborn infant. There are ample reasons why pregnant women should be counseled with regard to voluntary HIV testing. HIV-infected women need to be aware of the results of ACTG 076 and the possibility of reducing the rate of perinatal transmission with antiretroviral therapy.
HIV-infected women need to know, too, that HIV can be transmitted through breast milk (there is a 14 percent increased risk of transmission) and that breast feeding is contraindicated for HIV-infected women in the United States. Breast feeding is almost universally advocated in this country, and it is difficult to understand how an obstetrician or pediatrician can advise a mother to breast feed her infant without knowing whether or not the woman is infected with HIV.
The rapid changes in the standard of care for HIV infection may have an echo in the expanding options for CMV treatment. As Treatment Issues went to press, a Food and Drug Administration advisory panel recommended approval for Gilead Sciences' intravenous cidofovir (brand name: Vistide) for people with CMV retinitis. In March, also, the FDA granted final marketing approval for the ganciclovir eye implants (sold by Chiron Vision under the name "Vitrasert"). Besides these regulatory developments, recently released clinical data have shed more light on the use of oral ganciclovir prophylaxis and on foscarnet and ganciclovir as treatments, alone and in combination.
The approval for the implants comes roughly three months after the application was reviewed. Chiron Vision and Hoffmann-La Roche are co-marketing and sharing the profits from the ganciclovir implants, which cost $4,000 per device (not including the cost of surgically fastening the device within the eye, which ranges from $3,000 to $6,000). The product's labeling recommends replacing the implants every six months. (The implants' useful lifespan has been a matter of concern due to their variable drug release rate.)
No mention is made in the labeling of the need for backup systemic therapy. Such therapy may be advisable to reduce the risk of developing CMV disease in other parts of the body, including the unaffected eye. Roche is addressing this issue with an ongoing study of oral ganciclovir in combination with the implants.
The advisory panel recommended approving cidofovir as systemic treatment for new or relapsing CMV retinitis in people with AIDS. This application was supported by three studies (see Jan. 1996 Treatment Issues, pages 5-7), two with newly released findings.
One dose-ranging study compared two doses of cidofovir, 3 and 5 mg per kg of body weight once weekly for induction and then every other week for maintenance therapy. All patients had recurrent retinitis (the median number of prior relapses was four). The higher dose worked much better, significantly delaying retinitis progression by 115 days, compared to 49 days on the low dose.
On March 11, recruitment stopped for the latest of two trials comparing immediate cidofovir to deferred therapy in new CMV retinitis. Those in whom therapy was postponed progressed significantly faster than those receiving 3 mg cidofovir/kg of body weight -- a median delay of 69 days for immediate cidofovir arm versus 21 days for deferred therapy. (This trial also had a 5 mg/kg arm, but too little data has been collected to calculate the delay in CMV progression for this dose.)
The FDA's more conservative analysis found shorter times to CMV progression. The agency did not argue with the cidofovir studies' overall conclusions, however.
Until formal FDA approval occurs, patients can still receive cidofovir free through the expanded access programs (more than 250 people now enrolled) or as follow-up to the clinical trials. To be eligible for expanded access, applicants must have experienced a CMV relapse on either ganciclovir or foscarnet. To enroll, call 800/Gilead-5.
Also of interest, Gilead says that it has confirmed that cidofovir has anti-CMV activity when administered by injections into the eye. The company's results do not match the degree of safety or activity reported by William Freeman, M.D., of University of California San Diego. (Dr. Freeman administers intraocular injections only once every six weeks -- see Treatment Issues, July/August, 1995, pages 5-6.) Gilead has placed the study on hold while its Data and Safety Monitoring Board (DSMB) looks at the data. Depending on the DSMB's recommendation, the company may amend its protocol to more closely resemble Dr. Freeman's protocol.
Data from two separate analyses of the Roche oral ganciclovir prophylaxis study presented at the Third Conference on Retroviruses and Opportunistic Infections may ameliorate some of the concerns over the use of this drug. Many fear that the levels of ganciclovir reaching the eye during prophylaxis with oral ganciclovir are suboptimal and that such low-level exposure may induce ganciclovir-resistance in whatever CMV is present. Patients' future response to treatment with intravenous ganciclovir for breakthrough CMV retinitis might then be compromised. Two reports at the Third Conference on Retroviruses argued against this inference.
W. Lawrence Drew, M.D., presented data from 39 CMV isolates cultured from the urine and sometimes the blood or semen of patients on oral ganciclovir. He believes his observations suggest that resistance is a very rare occurrence (abstract LB16). After a mean of 8.3 months on oral ganciclovir in a Roche Bioscience prophylaxis trial, the only resistant isolates were from two patients who subsequently were not responsive to intravenous ganciclovir therapy.
Dr. Drew's study was much criticized, though, because CMV could be cultured from only a very small percentage of the study cohort of 925 people. No one knows how well drug-resistance of CMV in bodily fluids correlates with the state of CMV in the retina, which obviously cannot be sampled for analysis.
Another analysis of the Roche prophylaxis trial (abstract 9) reported on the clinical response to intravenous ganciclovir treatment among trial participants who did develop retinitis. Response to IV therapy was not significantly different between those randomized to prophylaxis and those on placebo. CMV retinitis in the placebo group was more often sight-threatening. Drug resistance, if any, seemed clinically irrelevant here.
Even when there is retinitis progression on maintenance doses of intravenous ganciclovir (which achieves greater blood levels of drug than the poorly absorbed oral ganciclovir), the condition generally responds to the higher doses of ganciclovir used during induction therapy. Data from a study of 279 patients who had failed maintenance therapy (abstract LB15 and also Archives of Ophthalmology, January, 1996, pages 23-33) show that people who had retinitis progression while on ganciclovir maintenance therapy and then were randomized to ganciclovir induction therapy did as well as those that were switched to foscarnet therapy.
Trial participants who were assigned to ganciclovir/foscarnet combination therapy experienced a significantly longer delay of retinitis progression (median time to progression was 60 days for the ganciclovir arm, 39 days for foscarnet and 129 days for the combination). Given the added annoyance, toxicity and cost of combination therapy, it is not clear how many patients will benefit from this finding. (It also is noteworthy that the study found no difference in survival between the study arms, contradicting an earlier study that found a survival benefit for foscarnet.)
A Canadian observational study reported at the Retroviruses Conference found improved results for intravenous ganciclovir or foscarnet by using induction regimens longer than the standard two weeks (abstract 165). One hundred and five patients were treated with open-label foscarnet or ganciclovir until their CMV was deemed stable upon ophthalmologic examination. Of the 79 patients achieving such stability, CMV relapse occurred on maintenance therapy in 51, at an average of 109 days. Comparison of these results with those found in rigorously controlled trials is of course difficult, and again, any added benefit from the extra ganciclovir has to be balanced against the added hazards and expense.
Since most early CMV breakthroughs can be controlled by higher ganciclovir induction doses, progression on maintenance therapy seems to be due to suboptimal dosing of the drug rather than to resistance. This problem could be largely addressed by the localized therapy that ganciclovir implants offer. Resistance more commonly stems from long-term systemic treatment of active CMV disease. It results from repeatedly introducing low-dose ganciclovir maintenance therapy after courses of induction therapy that incompletely suppress the CMV infection. Should widespread ganciclovir resistance occur, new agents such as cidofovir increase the therapeutic alternatives so that it will not entirely derail CMV treatment.
Here is yet another FDA action taken as the March Treatment Issues went to press: the agency's Blood Products Advisory Committee on March 21 reviewed an application for approval of Roche Molecular System's "Amplicor HIV Monitor." This assay measures HIV RNA (viral load) in plasma. The committee appeared dumbfounded as it was given only five hours to digest a large body of data presented by the sponsor, the FDA, outside experts, and representatives from companies with competing products.
Roche is seeking approval of the test "in combination with clinical presentation and other laboratory markers such as CD4 cell counts," said Roche's John Sninsky, Ph.D. Amplicor's several anticipated uses include: "as an indicator of disease prognosis; as an aid in monitoring the effects of antiviral therapy on plasma RNA levels; and as a predictive marker for clinical endpoints in HIV infected individuals." To support the indication, data was presented from several studies of nucleoside analogues and protease inhibitors as well as data from observational cohorts such as the MACS study (see Treatment Issues, November, 1994, pages 4-6, 10-11 and January, 1996, pages 13-14).
The data indicated that:
But the lack of uniformity across the studies (which used different viral load tests and different methods for storing the blood) made it difficult to the determine absolute values or standards needed for using the Amplicor to manage individual patients. The studies presented to the committee were not designed to decide at what viral load treatment should begin in the average patient, nor how great an increase in viral load should trigger a switch in treatment.
Roche has never sponsored such studies and apparently expected that the answers could be inferred from the large drug trials, which do show general trends. As the advisory panel's outside consultant Douglas Mayers, M.D., observed, the problem for the clinician remains, "what do you use for your yardstick?"
The purpose of the advisory meeting, according to Curtis Scribner, M.D., Deputy Director of the FDA's Office of Blood Research and Review, was only to "ask the scientific experts for their opinions on the data supporting the test's possible indications, and what should be the post-marketing requirements for the sponsor."
In the end, the advisory panel didn't take a formal vote. It is unclear what the FDA will make of the deliberations as several members admitted to being "overwhelmed by the data." Charles Flynn, M.D., of the committee acknowledged that "this test measures what it claims to measure," but felt that the sponsor should provide "better data that gives us algorithms for how to use this test." Chairman Scott Swisher, M.D., summed up his group's sense that the "test does have some prognostic value, and does have some value for showing whether a drug is having an antiviral effect," but "on the other hand, I felt completely uninformed about how I would use this test in my clinical practice."